Integrated Science Assessment
for Particulate Matter
Second External Review Draft
ISA: EPA/600/R-08/139B
ANNEXES: EPA/600/R-08/139BA
National Center for Environmental Assessment-RTP Division
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC

-------
Disclaimer
This document is the second external review draft for review purposes only and does not
constitute U.S. Environmental Protection Agency policy. Mention of trade names or
commercial products doesnot constitute endorsement or recommendation for use.

-------
Table of Contents
List of Tables	xiii
List of Figures	xxi
PM ISA Project Team	xu
Authors, Contributors, Reviewers	xliv
Clean Air Scientific Advisory Committee for Particulate Matter IMAAQS	l
Acronyms and Abbreviations	i
Chapter 1.Introduction	1-1
1.1. Legislative Requirements	1-4
1.2. History of Reviews of the NAAQS for PM	1 -5
1.3. ISA Development	1-12
1.4. Document Organization	1-16
1.5. EPA Framework for Causal Determination	1-18
1.5.1.	Scientific Evidence Used in Establishing Causality	1-19
1.5.2.	Association and Causation 	1-19
1.5.3.	Evaluating Evidence for Inferring Causation	1-20
1.5.4.	Application of Framework for Causal Determination 	1-25
1.5.5.	First Step-Determination of Causality	1-27
1.5.6.	Second Step-Evaluation of Response	1-30
1.5.6.1.	Effects on Fluman Populations	1-30
1.5.6.2.	Effects on Public Welfare	1-31
1.5.7.	Concepts in Evaluating Adversity of Health Effects	1 -32
1.6. Summary	1-33
Chapter 2. Integrative Health and Welfare Effects Overview	2-1
2.1. Concentrations and Sources of Atmospheric PM	2-2
2.1.1.	Ambient PM Variability and Correlations	2-2
2.1.1.1.	Spatial Variability across the U.S. 	2-3
2.1.1.2.	Spatial Variability on the Urban and Neighborhood Scales	2-4
2.1.2.	Trends and Temporal Variability	2-4
2.1.3.	Correlations between Copollutants	2-5
2.1.4.	Measurement Techniques	2-5
2.1.5.	PM Formation in the Atmosphere and Removal	2-6
2.1.6.	Source Contributions to PM	2-7
2.1.7.	Policy-Relevant Background 	2-8
2.2. Human Exposure	2-8
2.2.1.	Spatial Scales of PM Exposure Assessment	2-9
2.2.2.	Exposure to PM Components and Copollutants	2-10
2.2.3.	Implications for Epidemiologic Studies	2-11
2.3. Health Effects	2-12
July 2009	i	DRAFT-DO NOT CITE OR QUOTE

-------
2.3.1.	Exposure to PM2.5	2-13
2.3.1.1.	Effects of Short-Term Exposure to PM2.5	2-13
2.3.1.2.	Effects of Long-Term Exposure to PM2.5	2-16
2.3.2.	Integration of PM2.5 Health Effects	2-19
2.3.3.	Exposure to PM10 2.5	2-25
2.3.3.1. Effects of Short-Term Exposure to PM10 2.5 	2-25
2.3.4.	Integration of PM102.5 Effects	2-27
2.3.5.	Exposure to Ultrafine PM	2-30
2.3.5.1. Effects of Short-Term Exposure to UFPs	2-30
2.3.6.	Integration of UFP Effects	2-31
2.4. Policy Relevant Considerations	2-32
2.4.1.	Potentially Susceptible Subpopulations	2-32
2.4.2.	Lag Structure of PM—Morbidity and PM-Mortality Associations	2-34
2.4.2.1.	PM-Cardiovascular Morbidity Associations	2-34
2.4.2.2.	PM—Respiratory Morbidity Associations	2-35
2.4.2.3.	PM-Mortality Associations	2-35
2.4.3.	PM Concentration-Response Relationship 	2-36
2.5. Ecological and Welfare Effects	2-37
2.5.1.	Summary of Effects on Visibility 	2-37
2.5.2.	Summary of Effects on Climate	2-39
2.5.3.	Summary of Ecological Effects of PM	2-41
2.5.4.	Summary of Effects on Materials	2-43
Chapter 3. Source to Human Exposure	3-1
3.1. Introduction	3-1
3.2. Overview of Basic Aerosol Properties	3-2
3.3. Sources, Emissions and Deposition of Primary and Secondary PM	3-6
3.3.1.	Emissions of Primary PM and Precursors to Secondary PM	3-9
3.3.2.	Formation of Secondary PM 	3-12
3.3.2.1.	Formation of Nitrate and Sulfate	3-12
3.3.2.2.	Formation of Secondary Organic Aerosol	3-12
3.3.2.3.	Formation of new particles	3-15
3.3.3.	Mobile Source Emissions	3-16
3.3.3.1.	Emissions from Gasoline Fueled Engines	3-16
3.3.3.2.	Emissions from Diesel Fueled Engines	3-17
3.3.4.	Deposition of PM	3-19
3.3.4.1.	Deposition Forms	3-21
3.3.4.2.	Methods for Estimating Dry Deposition	3-23
3.3.4.3.	Factors Affecting Dry Deposition Rates and Totals	3-24
3.4. Monitoring of PM	3-27
3.4.1.	Ambient Measurement Techniques	3-27
3.4.1.1.	PM Mass	3-27
3.4.1.2.	PM Speciation 	3-31
3.4.1.3.	Multiple-Component Measurements on Individual Particles	3-39
3.4.1.4.	Ultrafine PM: Mass, Surface Area, and Number	3-39
3.4.1.5.	PM Size Distribution 	3-40
3.4.1.6.	Satellite Measurement	3-41
3.4.2.	Ambient Network Design 	3-42
3.4.2.1.	Monitor Siting Requirements	3-42
3.4.2.2.	Spatial and Temporal Coverage	3-44
3.4.2.3.	Network Application for Exposure Assessment with Respect to Susceptible
Sub-populations 	3-48
3.5. Ambient PM Concentrations	3-52
3.5.1. Spatial Distribution	3-54
July 2009
DRAFT-DO NOT CITE OR QUOTE

-------
3.5.1.1.	Variability across the U.S.	3-54
3.5.1.2.	Urban-Scale Variability	3-74
3.5.1.3.	Neighborhood-Scale Variability	3-98
3.5.2.	Temporal Variability	3-106
3.5.2.1.	Regional Trends 	3-106
3.5.2.2.	Seasonal Variations	3-111
3.5.2.3.	Hourly Variability	3-113
3.5.3.	Statistical Associations with Copollutants	3-117
3.5.4.	Summary	3-120
3.6. Mathematical Modeling of PM	3-122
3.6.1.	Estimating Source Contributions to PM Using Receptor Models	3-122
3.6.1.1. Receptor Models	3-123
3.6.2.	Chemistry Transport Models	3-130
3.6.2.1.	Global Scale	3-132
3.6.2.2.	Regional Scale 	3-132
3.6.2.3.	Local or Neighborhood Scale	3-135
3.6.3.	Air Quality Model Evaluation for Air Concentrations 	3-136
3.6.3.1.	Ground-based Comparisons of Photochemical Dynamics	3-142
3.6.3.2.	Predicted Chemistry for Nitrates and Related Compounds	3-142
3.6.4.	Evaluating Concentrations and Deposition of PM Components with CTMs	3-149
3.6.4.1.	Global CTM Performance	3-149
3.6.4.2.	Regional CTM Performance	3-151
3.7. Background PM	3-162
3.7.1. Contributors to PRB concentrations of PM	3-163
3.7.1.1.	Estimates of PRB Concentrations in Previous Assessments	3-164
3.7.1.2.	Chemistry Transport Models (CTMs) for Predicting PRB Concentrations	3-167
3.8. Issues in Exposure Assessment for PM and its Components	3-178
3.8.1.	General Exposure Concepts	3-179
3.8.2.	Exposure Modeling	3-182
3.8.2.1.	Time-Weighted Microenvironmental Models	3-182
3.8.2.2.	Stochastic Population Exposure Models 	3-184
3.8.2.3.	Dispersion Models	3-186
3.8.2.4.	Land Use Regression (LUR) and GIS-Based Models	3-187
3.8.3.	Personal and Microenvironmental Exposure Monitoring 	3-190
3.8.3.1.	New Developments in Personal Exposure Monitoring Techniques	3-190
3.8.3.2.	New Developments in Microenvironmental Exposure Monitoring Techniques	3-191
3.8.4.	Exposure Assessment Studies at Different Spatial Scales	3-192
3.8.4.1.	Urban Scale Ambient PM Exposure	3-193
3.8.4.2.	Micro-to-Neighborhood Scale Ambient PM Exposure	3-197
3.8.4.3.	Indoor Exposure to Ambient: Infiltration and Differential Infiltration	3-201
3.8.5.	Multicomponent and Multipollutant PM Exposures 	3-203
3.8.5.1.	Exposure Issues Related to PM Composition 	3-203
3.8.5.2.	Exposure to PM and Copollutants	3-209
3.8.6.	Implications of Exposure Assessment Issues for Interpretation of Epidemiologic Studies	3-211
3.8.6.1.	Measurement Error	3-211
3.8.6.2.	Model-Related Errors	3-212
3.8.6.3.	Spatial Variability	3-215
3.8.6.4.	Temporal Variability	3-219
3.8.6.5.	Use of Surrogates for PM Exposure	3-220
3.8.6.6.	Compositional Differences 	3-222
3.8.6.7.	Conclusions	3-223
3.9. Summary and Conclusions	3-224
3.9.1. Concentrations and Sources of Atmospheric PM	3-224
3.9.1.1.	Ambient PM Variability and Correlations	3-224
3.9.1.2.	Temporal Variability	3-227
3.9.1.3.	Correlations between Copollutants	3-228
July 2009
DRAFT-DO NOT CITE OR QUOTE

-------
3.9.1.4.	Measurement Techniques	3-228
3.9.1.5.	PM Source Characteristics	3-229
3.9.1.6.	Source Contributions to PM 	3-229
3.9.1.7.	Policy-Relevant Background 	3-230
3.9.2. Human Exposure	3-231
3.9.2.1.	Characterizing Human Exposure	3-231
3.9.2.2.	Spatial Scales of PM Exposure Assessment	3-232
3.9.2.3.	Multicomponent and Multipollutant PM Exposures	3-233
3.9.2.4.	Implications for Epidemiologic Studies	3-234
Chapter 4. Dosimetry	4-1
4.1. Introduction	4-1
4.1.1.	Size Characterization of Inhaled Particles 	4-2
4.1.2.	Structure of the Respiratory Tract	4-3
4.2. Particle Deposition	4-6
4.2.1.	Mechanisms of Deposition	4-7
4.2.2.	Deposition Patterns	4-9
4.2.2.1.	Total Respiratory Tract Deposition	4-11
4.2.2.2.	Extrathoracic Region	4-11
4.2.2.3.	Tracheobronchial and Alveolar Region	4-12
4.2.2.4.	Localized Deposition Sites	4-13
4.2.3.	Interspecies Patterns of Deposition 	4-14
4.2.4.	Biological Factors Modulating Deposition	4-15
4.2.4.1.	Physical Activity	4-15
4.2.4.2.	Age	4-17
4.2.4.3.	Gender	4-19
4.2.4.4.	Anatomical Variability	4-19
4.2.4.5.	Respiratory Tract Disease	4-20
4.2.4.6.	Hygroscopicity of Aerosols	4-22
4.2.5.	Summary	4-22
4.3. Clearance of Poorly Soluble Particles 	4-23
4.3.1.	Clearance Mechanisms and Kinetics	4-24
4.3.1.1.	Extrathoracic Region	4-24
4.3.1.2.	Tracheobronchial Region	4-24
4.3.1.3.	Alveolar Region	4-26
4.3.2.	Interspecies Patterns of Clearance and Retention	4-27
4.3.3.	Particle Translocation	4-29
4.3.3.1.	Alveolar Region	4-29
4.3.3.2.	Olfactory Region	4-31
4.3.4.	Factors Modulating Clearance	4-33
4.3.4.1.	Age	4-33
4.3.4.2.	Gender	4-34
4.3.4.3.	Respiratory Tract Disease	4-34
4.3.4.4.	Particle Overload	4-36
4.3.5.	Summary	4-36
4.4. Clearance of Soluble Materials	4-37
4.4.1.	Clearance Mechanisms and Kinetics	4-38
4.4.2.	Factors Modulating Clearance	4-40
4.4.2.1.	Age	4-40
4.4.2.2.	Physical Activity	4-40
4.4.2.3.	Disease	4-41
4.4.2.4.	Concurrent Exposures 	4-42
4.4.3.	Summary	4-43
July 2009
iv
DRAFT-DO NOT CITE OR QUOTE

-------
Chapter 5. Possible Pathways/ Modes of Action	5-1
5.1. Pulmonary Effects	5-2
5.1.1.	Reactive Oxygen Species	5-2
5.1.2.	Activation of Cell Signaling Pathways	5-4
5.1.3.	Pulmonary Inflammation	5-6
5.1.4.	Respiratory Tract Barrier Function	5-7
5.1.5.	Antioxidant Defenses and Adaptive Responses	5-8
5.1.6.	Pulmonary Function	5-10
5.1.7.	Allergic Disorders 	5-10
5.1.8.	Impaired Lung Defense Mechanisms	5-11
5.1.9.	Resolution of Inflammation/Progression or Exacerbation of Disease	5-11
5.1.9.1.	Factors Affecting the Retention of PM	5-11
5.1.9.2.	Factors Affecting the Balance of Pro/Anti-Inflammatory Mediators,
Qxidants/Anti-Qxidants and Proteases/Anti-Proteases	5-12
5.1.9.3.	Pre-Existing Disease	5-13
5.1.10.	Pulmonary DNA Damage	5-14
5.1.11.	Epigenetic Changes	5-14
5.1.12.	Lung Development	5-16
5.2. Systemic Inflammation	5-16
5.2.1.	Endothelial Dysfunction and Altered Vasoreactivity	5-18
5.2.2.	Activation of Coagulation and Acute Phase Response	5-19
5.2.3.	Atherosclerosis	5-20
5.2.4.	Activation of the Autonomic Nervous System by Pulmonary Reflexes	5-22
5.3. Translocation of Ultrafine PM or Soluble PM Components 	5-24
5.4. Disease of the Cardiovascular and Other Organ Systems	5-26
5.5. Acute and Chronic Responses	5-27
5.6. Results of New Inhalation Studies which Contribute to Modes of Action	5-28
Chapter 6. Integrated Health Effects of Short-Term PM Exposure	6-1
6.1. Introduction	6-1
6.2. Cardiovascular and Systemic Effects 	6-2
6.2.1.	Heart Rate and Heart Rate Variability	6-2
6.2.1.1.	Epidemiologic Studies	6-3
6.2.1.2.	Controlled Human Exposure Studies	6-12
6.2.1.3.	Toxicological Studies	6-14
6.2.2.	Arrhythmia	6-18
6.2.2.1.	Epidemiologic Studies	6-19
6.2.2.2.	Toxicological Studies	6-26
6.2.3.	Ischemia 	6-29
6.2.3.1.	Epidemiologic Studies	6-29
6.2.3.2.	Controlled Human Exposure Studies	6-32
6.2.3.3.	Toxicological Studies	6-32
6.2.4.	Vasomotor Function	6-34
6.2.4.1.	Epidemiologic Studies	6-35
6.2.4.2.	Controlled Human Exposure Studies	6-37
6.2.4.3.	Toxicological Studies	6-41
6.2.5.	Blood Pressure	6-47
6.2.5.1.	Epidemiologic Studies	6-48
6.2.5.2.	Controlled Human Exposure Studies	6-51
6.2.5.3.	Toxicological Studies	6-52
6.2.6.	Cardiac Contractility	6-53
6.2.6.1. Toxicological Studies	6-54
6.2.7.	Systemic Inflammation	6-55
July 2009
v
DRAFT-DO NOT CITE OR QUOTE

-------
6.2.7.1.	Epidemiologic Studies	6-55
6.2.7.2.	Controlled Human Exposure Studies	6-59
6.2.7.3.	Toxicological Studies	6-61
6.2.8.	Hemostasis, Thrombosis and Coagulation Factors	6-63
6.2.8.1.	Epidemiologic Studies	6-63
6.2.8.2.	Controlled Human Exposure Studies	6-65
6.2.8.3.	Toxicological Studies	6-67
6.2.9.	Systemic and Cardiovascular Oxidative Stress	6-70
6.2.9.1.	Epidemiologic Studies	6-70
6.2.9.2.	Controlled Human Exposure Studies	6-72
6.2.9.3.	Toxicological Studies	6-73
6.2.10.	Hospital Admissions and ED Visits	6-75
6.2.10.1.	All Cardiovascular Disease	6-81
6.2.10.2.	Cardiac Diseases	6-87
6.2.10.3.	Ischemic Heart Disease 	6-88
6.2.10.4.	Acute Ml	6-90
6.2.10.5.	Congestive Heart Failure	6-92
6.2.10.6.	Cardiac Arrhythmias	6-94
6.2.10.7.	Cerebrovascular Disease	6-96
6.2.10.8.	Peripheral Vascular Disease	6-98
6.2.10.9.	Copollutant Models	6-100
6.2.10.10.Concentration	Response 	6-101
6.2.10.11.Out of Hospital Cardiac Arrest 	6-103
6.2.11.	Short-term Exposure to PM and Cardiovascular Mortality	6-105
6.2.12.	Summary and Causal Determinations	6-106
6.2.12.1.	PMzb	6-106
6.2.12.2.	PMio-2.5	6-111
6.2.12.3.	UltrafinePM	6-113
6.3. Respiratory Effects	6-115
6.3.1.	Respiratory Symptoms and Medication Use	6-115
6.3.1.1.	Epidemiologic Studies	6-116
6.3.1.2.	Controlled Human Exposure Studies	6-127
6.3.2.	Pulmonary Function	6-128
6.3.2.1.	Epidemiologic Studies	6-129
6.3.2.2.	Controlled Human Exposure Studies	6-134
6.3.2.3.	Toxicological Studies	6-136
6.3.3.	Pulmonary Inflammation	6-138
6.3.3.1.	Epidemiologic Studies	6-138
6.3.3.2.	Controlled Human Exposure Studies	6-143
6.3.3.3.	Toxicological Studies	6-146
6.3.4.	Oxidative Responses 	6-153
6.3.4.1.	Controlled Human Exposure Studies	6-153
6.3.4.2.	Toxicological Studies	6-154
6.3.5.	Pulmonary Injury	6-158
6.3.5.1.	Epidemiologic Studies	6-158
6.3.5.2.	Controlled Human Exposure Studies	6-158
6.3.5.3.	Toxicological Studies	6-159
6.3.6.	Allergic Responses	6-168
6.3.6.1.	Epidemiologic Studies	6-168
6.3.6.2.	Controlled Human Exposure Studies	6-168
6.3.6.3.	Toxicological Studies	6-169
6.3.7.	Host Defense	6-178
6.3.7.1.	Epidemiologic Studies	6-178
6.3.7.2.	Toxicological Studies	6-179
6.3.8.	Respiratory ED Visits, Hospital Admissions and Physician Visits 	6-183
6.3.8.1.	All Respiratory Diseases	6-184
6.3.8.2.	Asthma	6-190
6.3.8.3.	COPD	6-197
July 2009
vi
DRAFT-DO NOT CITE OR QUOTE

-------
6.3.8.4.	Pneumonia and Respiratory Infections	6-198
6.3.8.5.	Copollutant Models	6-204
6.3.9.	Short-term Exposure to PMand Respiratory Mortality	6-205
6.3.10.	Summary and Causal Determinations	6-206
6.3.10.1.	PM2.5	6-206
6.3.10.2.	PM10-2.5	6-210
6.3.10.3.	UltrafinePM	6-212
6.4. Central Nervous System Effects	6-214
6.4.1.	Epidemiologic Studies	6-214
6.4.2.	Controlled Human Exposure Studies	6-215
6.4.3.	Toxicological Studies	6-215
6.4.3.1.	Urban Air	6-215
6.4.3.2.	CAPS 	6-216
6.4.3.3.	Diesel Exhaust 	6-217
6.4.4.	Summary and Causal Determination	6-219
6.5. Mortality Associated with Short-Term Exposure	6-219
6.5.1.	Summary of Findings from 2004 PM AQCD	6-220
6.5.2.	Associations of Mortality and Short-Term Exposure to PM	6-221
6.5.2.1.	PM10	6-223
6.5.2.2.	PMzb	6-241
6.5.2.3.	Thoracic Coarse Particles (PM10 2.5)	6-252
6.5.2.4.	Ultrafine PM	6-259
6.5.2.5.	Chemical Components of PM	6-261
6.5.2.6.	Source-Apportioned PM Analyses	6-267
6.5.2.7.	Investigation of Concentration-Response Relationship	6-268
6.5.3.	Summary and Causal Determinations	6-271
6.5.3.1.	PMzb	 6-271
6.5.3.2.	PM10-2.5	6-272
6.5.3.3.	Ultrafine PM	6-273
6.6. Attribution of Ambient PM Health Effects to Specific Constituents or Sources	6-274
6.6.1.	Evaluation Approach 	6-275
6.6.2.	Findings	6-276
6.6.2.1.	Epidemiologic Studies	6-276
6.6.2.2.	Controlled Human Exposure Studies	6-280
6.6.2.3.	Toxicological Studies	6-281
6.6.3.	Summary by Health Effects	6-285
Chapter 7. Integrated Health Effects of Long-Term PM Exposure	7-1
7.1. Introduction	7-1
7.2. Cardiovascular and Systemic Effects 	7-1
7.2.1.	Atherosclerosis	7-2
7.2.1.1.	Epidemiologic Studies	7-2
7.2.1.2.	Toxicological Studies	7-6
7.2.2.	Venous Thromboembolism	7-9
7.2.2.1. Epidemiologic Studies	7-9
7.2.3.	Diabetes 	7-10
7.2.3.1. Toxicological Studies	7-10
7.2.4.	Systemic Inflammation, Immune Function, and Blood Coagulation	7-10
7.2.4.1.	Epidemiologic Studies	7-10
7.2.4.2.	Toxicological Studies	7-12
7.2.5.	Renal and Vascular Function	7-13
7.2.5.1.	Epidemiologic Studies	7-14
7.2.5.2.	Toxicological Studies	7-15
7.2.6.	Autonomic Function	7-16
7.2.6.1. Toxicological Studies	7-16
July 2009
DRAFT-DO NOT CITE OR QUOTE

-------
7.2.7.	Cardiac changes	7-17
7.2.7.1. Toxicological studies	7-17
7.2.8.	Left Ventricular Mass and Function	7-18
7.2.9.	Clinical Outcomes in Epidemiologic Studies 	7-18
7.2.10.	Cardiovascular Mortality	7-23
7.2.11.	Summary and Causal Determinations	7-25
7.2.11.1.	PM2.5	7-25
7.2.11.2.	PMio"2.5	7-27
7.2.11.3.	UltrafinePM	7-27
7.3. Respiratory Effects	7-28
7.3.1.	Respiratory Symptoms and Disease Incidence 	7-29
7.3.1.1. Epidemiologic Studies	7-29
7.3.2.	Pulmonary Function	7-37
7.3.2.1.	Epidemiologic Studies	7-37
7.3.2.2.	Toxicological Studies	7-43
7.3.3.	Pulmonary Inflammation	7-45
7.3.3.1.	Epidemiologic Studies	7-45
7.3.3.2.	Toxicological Studies	7-45
7.3.4.	Pulmonary Oxidative Response	7-49
7.3.4.1. Toxicological Studies	7-49
7.3.5.	Pulmonary Injury	7-49
7.3.5.1. Toxicological Studies	7-49
7.3.6.	Allergic Responses	7-55
7.3.6.1.	Epidemiologic Studies	7-55
7.3.6.2.	Toxicological Studies	7-56
7.3.7.	Host Defense	7-57
7.3.7.1.	Epidemiologic Studies	7-57
7.3.7.2.	Toxicological Studies	7-58
7.3.8.	Respiratory Mortality	7-59
7.3.9.	Summary and Causal Determinations	7-59
7.3.9.1.	PMzb	7-59
7.3.9.2.	PM10-2.5	7-62
7.3.9.3.	Ultrafine PM	7-63
7.4. Reproductive, Developmental, Prenatal and Neonatal Outcomes 	7-63
7.4.1.	Epidemiologic Studies	7-63
7.4.2.	Toxicological Studies	7-84
7.4.2.1.	Female Reproductive Effects	7-85
7.4.2.2.	Male Reproductive Effects	7-86
7.4.2.3.	Multiple Generation Effects	7-90
7.4.2.4.	Receptor Mediated Effects	7-91
7.4.2.5.	Developmental Effects	7-92
7.4.3.	Summary and Causal Determinations	7-97
7.4.3.1.	PMzb	7-97
7.4.3.2.	PMiO"2.5	7-98
7.5. Cancer, Mutagenicity, and Genotoxicity	7-99
7.5.1.	Epidemiologic Studies	7-101
7.5.1.1.	Lung Cancer Mortality and Incidence 	7-101
7.5.1.2.	Other Cancers	7-105
7.5.1.3.	Markers of Exposure or Susceptibility	7-105
7.5.2.	Toxicological Studies	7-108
7.5.2.1.	Mutagenesis and Genotoxicity	7-109
7.5.2.2.	Carcinogenesis	7-114
7.5.3.	Epigenetic Studies and Other Heritable DNA mutations	7-115
7.5.4.	Summary and Causal Determinations	7-117
7.6. Mortality Associated with Long-term Exposure	7-118
7.6.1. Recent Studies of Long-Term Exposure to PM and Mortality	7-120
July 2009
DRAFT-DO NOT CITE OR QUOTE

-------
7.6.2.	Composition and Source-Oriented Analyses of PM	7-129
7.6.3.	Within-City Effects of PM Exposure	7-130
7.6.4.	Effects of Different Long-term Exposure Windows 	7-132
7.6.5.	Summary and Causal Determinations	7-135
7.6.5.1.	PMzb	7-135
7.6.5.2.	PMio-2.5	7-138
Chapter 8. Susceptible Subpopulations	8-1
8.1. Potentially Susceptible Subpopulations	8-1
8.1.1.	Age	8-3
8.1.1.1.	Older Adults	8-3
8.1.1.2.	Children	8-6
8.1.2.	Pregnancy and Developmental Effects	8-7
8.1.3.	Gender	8-7
8.1.4.	Race/Ethnicity	8-8
8.1.5.	Gene-Environment Interaction	8-9
8.1.6.	Pre-Existing Disease	8-12
8.1.6.1.	Cardiovascular Diseases	8-12
8.1.6.2.	Respiratory Illnesses	8-16
8.1.6.3.	Respiratory Contributions to Cardiovascular Effects	8-18
8.1.6.4.	Diabetes and Obesity	8-19
8.1.7.	Socioeconomic Status	8-21
8.1.8.	Summary	8-22
Chapter 9. Welfare Effects	9-1
9.1. Introduction	9-1
9.2. Effects on Visibility	9-1
9.2.1.	Introduction	9-1
9.2.2.	Background 	9-3
9.2.2.1.	Non-PM Visibility Effects	9-6
9.2.2.2.	PM Visibility Effects 	9-7
9.2.2.3.	Direct Optical Measurements	9-11
9.2.2.4.	Value of Good Visual Air Quality	9-13
9.2.3.	Monitoring and Assessment	9-14
9.2.3.1.	Aerosol Properties	9-15
9.2.3.2.	Spatial Patterns	9-22
9.2.3.3.	Urban and Regional Patterns	9-29
9.2.3.4.	Temporal Trends	9-37
9.2.3.5.	Causes of Haze	9-44
9.2.4.	Urban Visibility Valuation and Preference	9-74
9.2.4.1.	Urban Visibility Preference Studies	9-75
9.2.4.2.	Denver, Colorado Urban Visibility Preference Study	9-77
9.2.4.3.	Phoenix, Arizona Urban Visibility Preference Study	9-78
9.2.4.4.	British Columbia, Canada Urban Visibility Preference Study	9-79
9.2.4.5.	Washington, DC Urban Visibility Pilot Preference Study	9-79
9.2.4.6.	Urban Visibility Valuation Studies	9-81
9.2.5.	Summary of Effects on Visibility 	9-83
9.3. Effects on Climate	9-86
9.3.1.	The Climate Effects of Aerosols	9-87
9.3.2.	Overview of Aerosol Measurement Capabilities	9-94
9.3.2.1.	Satellite Remote Sensing	9-94
9.3.2.2.	Focused Field Campaigns	9-100
9.3.2.3.	Ground-Based In Situ Measurement Networks	9-101
9.3.2.4.	In Situ Aerosol Profiling Programs	9-103
9.3.2.5.	Ground-Based Remote Sensing Measurement Networks	9-107
9.3.2.6.	Synergy of Measurements and Model Simulations	9-108
July 2009
ix
DRAFT-DO NOT CITE OR QUOTE

-------
9.3.3.	Assessments of Aerosol Characterization and Climate Forcing	9-111
9.3.3.1.	The Use of Measured Aerosol Properties to Improve Models	9-111
9.3.3.2.	Intercomparisons of Satellite Measurements and Model Simulation of Aerosol
Optical Depth	9-114
9.3.3.3.	Satellite-Based Estimates of Aerosol Direct Radiative Forcing	9-116
9.3.3.4.	Satellite-Based Estimates of Anthropogenic Component of Aerosol Direct
Radiative Forcing	9-123
9.3.3.5.	Aerosol-Cloud Interactions and Indirect Forcing 	9-124
9.3.3.6.	Remote Sensing of Aerosol-Cloud Interactions and Indirect Forcing	9-125
9.3.3.7.	In Situ Studies of Aerosol-Cloud Interactions	9-128
9.3.4.	Outstanding Issues 	9-129
9.3.5.	Concluding Remarks	9-132
9.3.6.	Modeling the Effect of Aerosols on Climate	9-134
9.3.6.1.	Introduction	9-134
9.3.6.2.	Modeling of Atmospheric Aerosols	9-136
9.3.6.3.	Calculating Aerosol Direct Radiative Forcing 	9-141
9.3.6.4.	Calculating Aerosol Indirect Forcing	9-150
9.3.6.5.	Aerosol in the Climate Models	9-157
9.3.6.6.	Impacts of Aerosols on Climate Model Simulations	9-165
9.3.6.7.	Outstanding Issues 	9-169
9.3.6.8.	Conclusions	9-170
9.3.7.	Fire as a Special Source of PM Welfare Effects 	9-171
9.3.8.	Radiative Effects of Volcanic Aerosols 	9-173
9.3.8.1. Explosive Volcanic Activity	9-173
9.3.9.	Other Special Sources and Effects	9-178
9.3.9.1.	Glaciers and Snowpack	9-181
9.3.9.2.	Radiative Forcing by Anthropogenic Surface Albedo Change: BC in Snow and
Ice	9-184
9.3.9.3.	Effects on Local and Regional Climate	9-185
9.3.10.	Summary of Effects on Climate	9-187
9.4. Ecological Effects of PM	9-188
9.4.1.	Introduction	9-188
9.4.1.1.	Ecosystem Scale, Function, and Structure	9-191
9.4.1.2.	Ecosystem Services	9-192
9.4.2.	Deposition of PM	9-192
9.4.2.1.	Forms of Deposition	9-192
9.4.2.2.	Components of PM Deposition	9-194
9.4.2.3.	Magnitude of Dry Deposition	9-198
9.4.3.	Direct Effects of PM on Vegetation 	9-203
9.4.3.1. Effects of Coarse-mode Particles	9-204
9.4.4.	PM and Diffuse Light Effects	9-205
9.4.5.	Effects of Trace Metals on Ecosystems	9-206
9.4.5.1.	Direct Effects of Metals	9-207
9.4.5.2.	Effects on Soil Chemistry	9-209
9.4.5.3.	Effects on Soil Microbes and Plant Uptake via Soil	9-210
9.4.5.4.	Plant Response to Metals	9-215
9.4.5.5.	Effects on Aquatic Ecosystems	9-218
9.4.5.6.	Effects on Animals 	9-219
9.4.5.7.	Biomagnification across trophic levels	9-221
9.4.5.8.	Effects near Smelters and Roadsides	9-223
9.4.6.	Organic Compounds	9-225
9.4.7.	Summary of Ecological Effects of PM	9-230
9.5. Effects on Materials	9-232
9.5.1.	Effects on Paint 	9-234
9.5.2.	Effects on Metal Surfaces	9-234
9.5.3.	Effects on Stone	9-235
9.5.4.	Summary of Effects on Materials	9-236
July 2009
x
DRAFT-DO NOT CITE OR QUOTE

-------
Annex A. Atmospheric Science	A-1
A.1. Ambient Air Particle Monitoring	A-1
A.1.1. Measurements and Analytical Specifications	A-1
A.1.2. Networks	A-49
A.1.3. Monitor Distribution with Respect to Population Density	A-54
A.2. Ambient PM Concentration	A-82
A.2.1. Speciation Trends Network Site Data	A-82
A.2.2. Intraurban Variability	A-88
A.2.3. Speciation	A-170
A.2.4. Diel Trends	A-192
A.2.5. Copollutant Measurements	A-201
A.3. Source Apportionment	A-215
A.3.1. Type of Receptor Models	A-215
A.3.2. Source Profiles	A-228
A.3.3. Receptor Model Results	A-231
A.4.	Exposure Assessment	A-232
A.4.1. Exposure Assessment Study Findings	A-232
Annex B. Dosimetry	B-1
B.1.	Ultrafine Disposition	B-1
B.2. Olfactory Translocation 	B-3
B.3. Clearance and Age	B-4
Annex C. Controlled Human Exposure Studies	C-1
Annex D.Toxicological Studies	D-1
D.1.	Carcinogenisis, Mutagenesis, Genotoxicity	D-172
Annex E. Epidemiologic Studies	E-1
E.1.	Short-Term Exposure and Cardiovascular Outcomes	E-1
E.1.1. Cardiovascular Morbidity Studies	E-1
E.1.2. Cardiovascular Emergency Department Visits and Hospital Admissions	E-84
E.2. Short-Term Exposure and Respiratory Outcomes	E-146
E.2.1. Respiratory Morbidity Studies	E-146
E.2.2. Respiratory Emergency Department Visits and Hospital Admissions	E-276
E.3. Short-Term Exposure and Mortality	E-351
E.4. Long-Term Exposure and Cardiovascular Outcomes	E-468
E.5. Long-Term Exposure and Respiratory Outcomes	E-490
E.6. Long-Term Exposure and Cancer	E-555
E.7. Long-Term Exposure and Reproductive Effects	E-563
E.8. Long-Term Exposure and Mortality	E-630
E.9. Long-Term Exposure and Mortality	E-639
Annex F. Source Apportionment Studies	F-1
July 2009
xi
DRAFT-DO NOT CITE OR QUOTE

-------
List of Tables
Table 1-1.
Summary of NAAQS promulgated for PM, 1971-2006.
1-6
Table 1-2
Aspects to aid in judging causality-
1-26
Table 1-3.
Weight of evidence for causal determination.
1-29
Table 2-1
Summary of causal determinations for short-term exposure to PM2.5.
2-13
Table 2-2.
Summary of causal determinations for long-term exposure to PM2.5.
2-16
Table 2-3.
Summary of causal determinations for short-term exposure to PM10-2.5.
2-25
Table 2-4.
Summary of causal determinations for short-term exposure to UFPs.
2-30
Table 2-5.
Summary of causality determination for welfare effects.
2-37
Table 3-1.
Characteristics of ambient fine (ultrafine plus accumulation-mode) and coarse
particles.
3-4
Table 3-2.
Constituents of atmospheric particles and their major sources.
3-7
Table 3-3.
Proximity to PM10 and PM2.5 monitors for total population3 by city.
3-47
Table 3-4.
Proximity to PM10 and PM2.5 monitors for children 0-4 y, children 5-17 y, and adults 65
y and older.aThe figures presented here are cumulative for the 15 CSAs/CBSAs
examined in Chapter 3.
3-49
Table 3-5.
Proximity to PM10 and PM2.5 monitors for adults aged 65 and older3 by city.
3-50
Table 3-6.
Proximity to PM10 and PM2.5 monitors based on the population identified as white,
black, Hispanic, or non-Hispanic
3-51
Table 3-7.
Proximity to PM10 and PM2.5 monitors based on the population below or above the
poverty line
3-52
Table 3-8.
PM2.5 distributions derived from AQS data (concentration in /yg/m3).
3-56
Table 3-9.
PM10 2.5 distributions derived from AQS data (concentration in /yg/m3).
3-59
Table 3-10.
PM10 distributions derived from AQS data (concentration in /yg/m3).
3-60
Table 3-11.
Inter-sampler comparison statistics for each pair of 24-h PM2.5 monitors reporting to
AQS for Boston, MA.
3-78
Table 3-12.
Inter-sampler comparison statistics for each pair of 24-h PM2.5 monitors reporting to
AQS for Pittsburgh, PA.
3-81
Table 3-13.
Inter-sampler comparison statistics for each pair of 24-h PM2.5 monitors reporting to
AQS for Los Angeles, CA.
3-83
Table 3-14.
Inter-sampler comparison statistics for each pair of 24-h PM10 monitors reporting to
AQS for Boston, MA.
3-88
Table 3-15.
Inter-sampler comparison statistics for each pair of 24-h PM10 monitors reporting to
AQS for Pittsburgh, PA.
3-90
Table 3-16.
Inter-sampler comparison statistics for each pair of 24-h PM10 monitors reporting to
AQS for Los Angeles, CA.
3-93
Table 3-17.
Example of emissions factors (ng/kg) for trace elements under variable speed and
steady speed driving conditions for PM emitted by diesel and gasoline engines.
3-125
July 2009
xii DRAFT-DO NOT CITE OR QUOTE

-------
Table 3-18.
Table 3-19.
Table 3-20.
Table 3-21.
Table 3-22.
Table 3-23.
Table 3-24.
Table 4-1.
Table 6-1.
Table 6-2.
Table 6-3.
Table 6-4.
Table 6-5.
Table 6-6.
Table 6-7.
Table 6-8.
Table 6-9.
Table 6-10.
Table 6-11.
Table 6-12.
Table 6-13.
Table 6-14.
Table 6-15.
Table 6-16.
Table 6-17.
July 2009
Estimates of annual average natural background concentrations of PM2.5 and PM10	3-165
Annual and quarterly mean PM2.5 concentrations (pg/rn3) measured at IMPROVE sites
in 2004.	3-166
Annual and quarterly mean PM2.5 concentrations (pg/rn3) for the CMAQ "base case" at
IMPROVE sites in 2004.	3-176
Annual and quarterly mean PM2.5 concentrations (pg/rn3) for the CMAQ PRB
simulations at IMPROVE sites in 2004.	3-177
Annual and quarterly mean of the CMAQ-predicted base case PM2.5 concentrations
(Mg/m3) in the U.S. EPA CONUS regions in 2004.	3-177
Annual and quarterly mean of the CMAQ-predicted PRB PM2.5 concentrations (|jg/m3)
in the U.S. EPA CONUS regions in 2004.	3-177
Examples of studies comparing near-road personal exposures with fixed site ambient
concentrations.	3-199
Breathing patterns with activity level in adult human male.	4-16
Characteristics of epidemiologic/panel studies investigating associations between PM
and changes in HRV.	6-9
Studies of ventricular arrhythmia and ambient PM concentration, in patients with
implantable cardioverter defibrillators.	6-22
Median particle concentrations.	6-48
Ambient concentrations in six European cities.	6-56
Description of ICD-9 and ICD-10 codes for diseases of the circulatory system.	6-77
Characterization of ambient PM concentrations in studies of hospital admission and ED
visits for cardiovascular diseases.	6-86
Characterization of ambient PM concentrations from studies of respiratory morbidity
and short-term exposures in asthmatic children and adults. All concentrations are for
the 24-h avg unless otherwise noted.	6-119
PAMCHAR PM10 2.5 inflammation results with ambient PM.	6-164
Other ambient PM - in vivo PM10 2.5 studies - BALF results, 18-24 h post-IT exposure	6-164
Description of ICD-9 and ICD-10 codes for diseases of the respiratory system.	6-183
PM concentrations in studies of respiratory diseases published since 2002.	6-204
Overview of U.S. and Canadian multicity PM studies of mortality analyzed in the 2004
PM AQCD and the PM ISAb	6-222
NMMAPS national and regional percentage increase in all-cause, cardio-respiratory,
and other-cause mortality associated with a 10//g/m3 increase in PM10 at lag 1 day
for the periods 1987-1994,1995-2000, and 1987-2000.	6-226
Key to Figure 6-24 	6-246
Key for Figure 6-29	6-256
Effect modification of composition on the estimated percent increase in mortality with
a 10 |jg/m3 increase in PM2.5.	6-264
Study-specific PM2.5 factor/source categories associated with health effects.	6-282
DRAFT-DO NOT CITE OR QUOTE

-------
Table 7-1.	Characterization of ambient PM concentrations from studies of subclinical measures of
cardiovascular diseases.	7-15
Table 7-2.	Characterization of ambient PM concentrations from studies of clinical cardiovascular
diseases.	7-18
Table 7-3.	Characterization of ambient PM concentrations from studies of respiratory
symptoms/disease and long-term exposures.	7-31
Table 7-4.	Characterization of ambient PM concentrations from studies of FEVi and long-term
exposures.	7-37
Table 7-5.	Characterization of ambient PM concentrations from studies of reproductive,
developmental, prenatal and neonatal outcomes and long-term exposure.	7-65
Table 7-6.	Characterization of ambient PM concentrations from recent studies of cancer and
long-term exposures to PM. 	7-102
Table 7-7.	Associations* between ambient PM concentrations from select studies of lung cancer
mortality and incidence.	7-103
Table 7-8.	Characterization of ambient PM concentrations from studies of mortality and long-
term exposures to PM.	7-119
Table 7-9:	Comparison of results from ACS intra-urban analysis of Los Angeles and new york city
using kriging or land use regression to estimate exposure	7-131
Table 7-10. Distribution of the effect of a hypothetical reduction of 10//g/m3 PM10 in 2000 on all-
cause mortality 2000-2009 in Switzerland.	7-134
Table 8-1.	Definitions of susceptible and vulnerable in the PM literature. 	8-1
Table 8-2.	Susceptibility Factors.	8-3
Table 8-3.	Percent of the U.S. population with respiratory diseases, cardiovascular diseases, and
diabetes.	8-16
Table 9-1.	Regional Planning Organization websites with visibility characterization and source
attribution assessment information.	9-23
Table 9-2.	Summary of urban visibility preference studies.	9-76
Table 9-3.	Top-of-atmosphere, cloud-free, instantaneous direct aerosol radiative forcing
dependence on aerosol and surface properties.	9-93
Table 9-4.	Summary of major satellite measurements currently available for the tropospheric
aerosol characterization and radiative forcing research.	9-95
Table 9-5.	List of major intensive field experiments that are relevant to aerosol research in a
variety of aerosol regimes around the globe conducted in the past two decades.	9-105
Table 9-6.	Summary of major U.S. surface in situ and remote sensing networks for the
tropospheric aerosol characterization and radiative forcing research.	9-106
Table 9-7. Summary of approaches to estimating the aerosol direct radiative forcing in three
categories: (A) satellite retrievals; (B) satellite-model integrations; and (C) model
simulations.	9-117
Table 9-8.	Summary of seasonal and annual average clear-sky DRF (W/m2) at the TOA and the
surface (SFC) over global OCEAN derived with different methods and data.	9-121
Table 9-9.	Summary of seasonal and annual average clear-sky DRF (W/m2) at the TOA and the
surface (SFC) over global LAND derived with different methods and data. 	9-122
July 2009
xiv
DRAFT-DO NOT CITE OR QUOTE

-------
Table 9-10.
Estimates of anthropogenic components of aerosol optical depth (Tam) and clear-sky
DRF at the TOA from model simulations
9-124
Table 9-11.
Anthropogenic emissions of aerosols and precursors for 2000 and 1750.
9-138
Table 9-12.
Summary of statistics of AeroCom Experiment A results from 16 global models.
9-139
Table 9-13.
SO42" mass loading, MEE and AOD at 550 nm, shortwave radiative forcing at the top
of the atmosphere, and normalized forcing with respect to AOD and mass.
9-143
Table 9-14.
Particulate organic matter (POM) and BC mass loading, AOD at 550 nm, shortwave
radiative forcing at the top of the atmosphere, and normalized forcing with respect to
AOD and mass.
9-144
Table 9-15.
Differences in present day and pre-industrial outgoing solar radiation (W/m2) in the
different experiments.
9-153
Table 9-16.
Forcings used in IPCC AR4 simulations of 20th century climate change.
9-157
Table 9-17.
Olimate forcings (1880-2003) used to drive GISS climate simulations, along with the
surface air temperature changes obtained for several periods.
9-166
Table 9-18.
Overview of the different aerosol indirect effects and their sign of the net radiative
flux change at the top of the atmosphere (TOA).
9-180
Table 9-19.
Overview of the different aerosol indirect effects and their implications for the global
mean net shortwave radiation of the surface Fsfc (columns 2-4) and for precipitation
(columns 5-7).
9-180
Table 9-20.
Recent studies highlighting POP occurrence and fate in the major arctic
compartments.
9-182
Table 9-21.
Factors potentially important in estimating mercury exposure.
9-195
Annex Tables
Table A-1.
Summary of integrated and continuous samplers included in the field comparison.
A-1
Table A-2.
Summary of PM2.5 and PM10 FRM and FEM samplers.
A-4
Table A-3.
Measurement and analytical specifications for filter analysis of mass, elements, ions,
and carbon.
A-6
Table A-4.
Measurement and analytical specifications for filter analysis of organic species.
A-8
Table A-5.
Measurement and analytical specifications for continuous mass and mass surrogate
instruments.
A-10
Table A-6.
Measurement and analytical specifications for continuous elements.
A-12
Table A-7.
Measurement and analytical specifications for continuous NO3.
A-12
Table A-8.
Measurement and analytical specifications for continuous SO42.
A-14
Table A-9.
Measurement and analytical specifications for ions other than NO3 and SO42".
A-16
Table A-10.
Measurement and analytical specifications for continuous carbon.
A-17
Table A-11.
Summary of mass measurement comparisons.
A-19
Table A-12.
Summary of element and liquid water content measurement comparisons.
A-27
Table A-13.
Summary of PM2.5 NO3 measurement comparisons.
A-28
Table A-14.
Summary of PM2.5 SO42 measurement comparisons
A-32
July 2009
XV DRAFT -
DO NOT CITE OR QUOTE

-------
Table A-15.
Summary of PM2.5 carbon measurement comparisons.
A-35
Table A-16.
Summary of particle mass spectrometer measurement comparisons.
A-40
Table A-17.
Summary of particle mass spectrometer measurement comparisons.
A-44
Table A-18.
Summary of key parameters for TD-GC/MS and pyrolysis-GC/MS.
A-47
Table A-19.
Relevant Spatial Scales for PM10, PM2.5, and PM10-2.5 Measurement
A-49
Table A-20.
Major routine operating air monitoring networks'1
A-51
Table A-21.
Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Atlanta, GA.
A-90
Table A-22.
Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Birmingham, AL.
A-94
Table A-23.
Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Chicago, IL.
A-97
Table A-24.
Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Denver, CO.
A-101
Table A-25.
Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Detroit, Ml.
A-105
Table A-26.
Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Houston, TX.
A-109
Table A-27.
Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
New York City, NY.
A-113
Table A-28.
Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Philadelphia, PA.
A-118
Table A-29.
Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Phoenix, AZ.
A-121
Table A-30.
Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Riverside, CA.
A-125
Table A-31.
Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Seattle, WA.
A-127
Table A-32.
Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
St. Louis, MO.
A-131
Table A-33.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Atlanta, GA.
A-133
Table A-34.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Birmingham, AL.
A-136
Table A-35.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Chicago, IL.
A-140
Table A-36.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Denver, CO.
A-143
Table A-37.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Detroit, Ml.
A-146
Table A-38.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Houston, TX.
A-149
July 2009	xvi	DRAFT - DO NOT CITE OR QUOTE

-------
Table A-39. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
New York City, NY.	A-150
Table A-40. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Philadelphia, PA.	A-153
Table A-41. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Phoenix, AZ.	A-157
Table A-42. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Riverside, CA.	A-162
Table A-43. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Seattle, WA. 	A-165
Table A-44. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
St. Louis, MO. 	A-168
Table A-45. Correlation coefficients of hourly and daily average particle number, surface and
volume concentrations in selected particle size ranges.	A-169
Table A-46. Different receptor models used in the Supersite source apportionment studies:
chemical mass balance.	A-215
Table A-47. Different receptor models used in the Supersites source apportionment studies: factor
analysis.	A-216
Table A-48. Different receptor models used in the Supersites source apportionment studies: tracer-
based methods. 	A-220
Table A-49. Different receptor models used in the Supersites source apportionment studies:
meteorology-based methods.	A-222
Table A-50.	Source Profiles	A-228
Table A-51.	PM10 receptor model results	A-231
Table A-52.	PM2.5 receptor model results	A-232
Table A-53.	Exposure Assessment Study Summaries	A-232
Table A-53. Examples of studies showing developments with UFP sampling methods since the
2004 PM AQCD.	A-279
Table A-54.	Summary of in-vehicle studies of exposure assessment.	A-280
Table A-55.	Summary of personal PM exposure studies with no indoor source during 2002-2008. 	A-283
Table A-56.	Summary of PM species exposure studies.	A-287
Table A-57.	Summary of personal PM exposure source apportionment studies. 	A-299
Table A-54.	Summary of PM infiltration studies.	A-301
Table A-55.	Summary of PM - copollutant exposure studies. 	A-312
Table A-56.	Summary of studies relating PM, SES, and mortality and/or morbidity.	A-316
Table B-1.	Ultrafine disposition in humans.	B-1
Table B- 2.	Ultrafine disposition in animals.	B-2
Table B- 3.	In vitro studies of ultrafine disposition.	B-2
Table B-4.	Olfactory particle translocation. 	B-3
Table B-5.	Studies of respiratory tract mucosal and macrophage clearance as a function.	B-4
July 2009
xvii
DRAFT-DO NOT CITE OR QUOTE

-------
Table C-1.	Cardiovascular Effects	C-1
Table C-2.	Respiratory effects	C-8
Table C- 3.	Central Nervous System Effects	C-13
Table D-1.	Cardiovascular effects.	D-1
Table D-2.	Respiratory effects: in vitro studies.	D-29
Table D-3.	Respiratory effects: in vivo studies.	D-78
Table D-4.	Effects related to immunity and allergy.	D-123
Table D-5.	Effects of the central nervous system.	D-165
Table D-6.	Reproductive and developmental effects.	D-167
Table D-7.	Mutagenic/genotoxic effects in bacterial cultures.	D-172
Table D-8.	Mutagenicity and genotoxicity data summary: in vitro studies.	D-175
Table D-9.	Mutagenicity and genotoxicity data summary: in vivo studies.	D-181
Table E-1	Short-term exposure - cardiovascular morbidity outcomes - PM10 	E-1
Table E-2.	Short-term exposure - cardiovascular morbidity studies: PM10 2.5.	E-18
Table E-3.	Short-term exposure - cardiovascular morbidity studies: PM2.5 (including PM
components/sources). 	E-20
Table E-4.	Short-term exposure - cardiovascular morbidity studies - other size fractions.	E-77
Table E-5.	Short-term exposure-cardiovascular-ED/HA PM10	E-84
Table E-6.	Short-term exposure-cardiovascular-ED/HA - PM10 2.5.	E-109
Table E-7.	Short-term exposure - cardiovascular-ED/HA - PM2.5 (including PM
components/sources)	E-112
Table E-8.	Short-term exposure-cardiovascular-ED/HA-other size fractions.	E-140
Table E-9.	Short-term exposure-respiratory morbidity outcomes -PM10	E-146
Table E-10.	Short-term exposure - respiratory morbidity outcomes - PM10 2.5.	E-194
Table E-11.	Short-term exposure - respiratory morbidity outcomes - PM2.5 (including
components/sources).	E-198
Table E-12.	Short-term exposure—respiratory—ED/HA-PMio	E-276
Table E-13.	Short-term exposure—respiratory—ED/HA-PMio 2.5	E-308
Table E-14.	Short-term exposure—respiratory—ED/HA-PM2.5 (including PM components/sources).	E-316
Table E-15.	Short-term exposure—respiratory—ED/HA-Other Size Fractions	E-340
Table E-16.	Short-term exposure - mortality - PM10.	E-351
Table E-17.	Short-term exposure - mortality - PM10 2.5. 	E-433
Table E-18.	Short-term exposure - mortality - PM2.5 (including PM components/sources). 	E-439
Table E-19.	Short-term exposure - mortality - other PM size fractions.	E-464
Table E-20.	Long-term exposure - cardiovascular morbidity outcomes - PM10	E-468
Table E-21.	Long-term effects-cardiovascular- PM2.5 (including PM components/sources)	E-476
Table E-22.	Long-term exposure - respiratory morbidity outcomes - PM10.	E-490
July 2009
xviii
DRAFT -
DO NOT CITE OR QUOTE

-------
Table E-23.	Long-term exposure ¦ respiratory morbidity outcomes ¦ PM10 2.5.	E-521
Table E-24.	Long-term exposure ¦ respiratory morbidity outcomes ¦ PM2.5 (including PM
components/sources).	E-528
Table E-25.	Long-term exposure ¦ respiratory morbidity outcomes ¦ other PM size fractions. 	 E-553
Table E-26.	Long-term exposure - cancer outcomes - PM10.	E-555
Table E-27.	Long-term exposure - cancer outcomes - PM2.5 (including PM components/sources).	 E-559
Table E-28.	Long-term exposure - cancer outcomes - other PM size fractions.	E-562
Table E-29.	Long-term exposure - reproductive outcomes - PM10.	E-563
Table E-30.	Long-term exposure - mortality - PM10.	E-630
Table E-31.	Long-term exposure - mortality - PM10.	E-639
Table E-32.	Long-term exposure - mortality - PM10 2.5.	E-640
Table E-33.	Long-term exposure - mortality - PM2.5 (including PM components/sources).	E-641
Table E-34.	Long-term exposure - central nervous system outcomes - PM. 	E-672
Table F-1.	Epidemiologic studies of ambient PM sources, factors, or constituents.	F-1
Table F-2.	Human clinical studies of ambient PM sources, factors, or constituents. 	F-4
Table F-3.	Toxicological studies of ambient PM sources, factors, or constituents	F-5
July 2009
xix
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 1-1
Figure 2-1.
Figure 2-2.
Figure 2-3.
Figure 3-1.
Figure 3-2.
Figure 3-3.
Figure 3-4.
Figure 3-5.
Figure 3-6.
Figure 3-7.
Figure 3-8.
Figure 3-9.
Figure 3-10.
Figure 3-11.
Figure 3-12.
Figure 3-13.
Figure 3-14.
Figure 3-15.
Figure 3-16.
Figure 3-17.
July 2009
List of Figures
Identification of studies for inclusion in the ISA.	1-13
Excess risk estimates from epidemiologic studies of PM2.5 ordered by mean 24-h avg
concentration as reported by the investigator.	2-20
Summary of U.S. studies examining the association between long-term exposure to
PM2.5 and CVD morbidity/mortality, respiratory morbidity/mortality, and all-cause
mortality conducted in locations where the mean annual PM2.5 concentration ranged
from 10.7-29 //g/m3. All effect estimates have been standardized to reflect a 10
//g/m3 increase in mean annual PM2.5 concentration.	2-21
Effect estimates from epidemiologic studies of PM10-2.5 ordered by mean 24-h avg
concentration as reported by the investigator.	2-28
Particle size distributions by number and volume.	3-3
X-ray spectra and scanning electron microscopy images of individual particles 	3-5
Detailed source categorization of anthropogenic emissions of primary PM2.5, PM10 and
gaseous precursor species SO2, NOx, NFI3 and VOCs for 2002 in units of million metric
tons (MMT). (EGUs = Electricity Generating Units)	3-11
Primary emissions and formation of SOA through gas, cloud and condensed phase
reactions. 	3-15
Schematic of the resistance-in-series analogy for atmospheric deposition.	3-20
The relationship between particle diameter and Vd for particles.	3-25
PM10 monitor distribution in comparison with population density, Boston CSA.	3-45
PM2.5 monitor distribution in comparison with population density, Boston CSA.	3-46
Three-yr avg 24-h PM2.5 concentration by county derived from FRM or FRM-like data,
2005-2007.	3-55
Three-yr avg 24-h PM10-2.5 concentration by county derived from co-located low volume
FRM PM10 and PM2.5 monitors, 2005-2007. 	3-58
Three-yr avg 24-h PM10 concentration by county derived from FRM or FEM monitors,
2005-2007.	3-60
Three-yr avg 24-h PM2.5 OC concentrations measured at CSN sites across the U.S.,
2005-2007.	3-65
Three-yr avg 24-h PM2.5 EC concentrations measured at CSN sites across the U.S.,
2005-2007.	3-66
Three-yr avg 24-h PM2.5 SO42"concentrations measured at CSN sites across the U.S.,
2005-2007.	3-67
Three-yr avg 24-h PM2.5 NO3" concentrations measured at CSN sites across the U.S.,
2005-2007.	3-68
Three-yr avg 24-h PM2.5 NFU+ concentrations measured at CSN sites across the U.S.,
2005-2007.	3-69
Three-yr avg PM2.5 speciation estimates for 2005-2007 derived using the SANDWICH
method	3-70
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 3-18.
Figure 3-19.
Figure 3-20.
Figure 3-21.
Figure 3-22.
Figure 3-23.
Figure 3-24.
Figure 3-25.
Figure 3-26.
Figure 3-27.
Figure 3-28.
Figure 3-29.
Figure 3-30.
Figure 3-31.
Figure 3-32.
Figure 3-33.
Figure 3-34.
Figure 3-35.
Figure 3-36.
Figure 3-37.
Figure 3-38.
Figure 3-39.
Figure 3-40.
Figure 3-41.
Figure 3-42.
July 2009
Seasonally-stratified three-yr avg PM2.5 speciation estimates for 2005-2007 derived
using the SANDWICH method	3-71
Locations of PM2.5 monitors and major highways, Boston, MA.	3-76
Seasonal distribution of 24-h avg PM2.5 concentrations by site for Boston, MA, 2005-
2007. Box plots show the median and interquartile range with whiskers extending to
the 5th and 95th percentiles at each site during (1) winter (December-February), (2)
spring (March-May), (3) summer (June-August) and (4) fall (September-November). 	3-77
Locations of PM2.5 monitors and major highways, Pittsburgh, PA.	3-79
Seasonal distribution of 24-h avg PM2.5 concentrations by site for Pittsburgh, PA,
2005-2007.	3-80
Locations of PM2.5 monitors and major highways, Los Angeles, CA.	3-82
Seasonal distribution of 24-h avg PM2.5 concentrations by site for Los Angeles, CA,
2005-2007.	3-83
Inter-sampler correlations for 24-h PM2.5 as a function of distance between monitors in
Boston, MA.	3-84
Inter-sampler correlations for 24-h PM2.5 as a function of distance between monitors in
Pittsburgh, PA.	3-84
Inter-sampler correlations for 24-h PM2.5 as a function of distance between monitors in
Los Angeles, CA.	3-85
Seasonal distribution of 24-h avg PM10-2.5 concentrations by site 	3-86
Locations of PM10 monitors and major highways, Boston, MA.	3-87
Seasonal distribution of 24-h avg PM10 concentrations by site for Boston, MA,
2005-2007.	3-88
Locations of PM10 monitors and major highways, Pittsburgh, PA.	3-89
Seasonal distribution of 24-h avg PM10 concentrations by site for Pittsburgh, PA,
2005-2007.	3-90
Locations of PM10 monitors and major highways, Los Angeles, CA.	3-92
Seasonal distribution of 24-h avg PM10 concentrations by site for Los Angeles, CA,
2005-2007.	3-93
Inter-sampler correlations for 24-h PM10 as a function of distance between monitors in
Boston, MA.	3-95
Inter-sampler correlations for 24-h PM10 as a function of distance between monitors in
Pittsburgh, PA.	3-95
Inter-sampler correlations for 24-h PM10 as a function of distance between monitors in
Los Angeles, CA.	3-96
Bin-wise Spearman correlation coefficients in aerosol particle number concentrations	3-97
Dimensionless concentration as a function of height at windward and leeward
locations and street canyon aspect ratios	3-99
Inter-sampler correlations for 24-h PM2.5 and PM10 as a function of distance between
monitors for samplers located within 4 km (neighborhood scale). 	3-100
Particle size distributions measured at various distances from the 710 freeway	3-102
Mass distributions for BaP	3-105
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 3-43.	Mass distributions for 16 PAHs at a high traffic city center in Seville, Spain.	3-105
Figure 3-44.	Ambient 24-h PM2.5 concentrations in the U.S., 1999-2007, 	3-107
Figure 3-45.	Ambient annual PM2.5 concentrations in the U.S., 1999-2007,	3-108
Figure 3-46.	Ambient 24-h PM10 concentrations in the U.S., 1988-2007,	3-109
Figure 3-47. Regional and seasonal trends in annual PM2.5 compostion from 2002 to 2007 derived
using the SANDWICH method.	3-110
Figure 3-48. Ultrafine particle size distribution	3-112
Figure 3-49. Diel plot generated from hourly FRM-like PM2.5 data	3-114
Figure 3-50. Diel plots generated from hourly FEM PM10 data 	3-115
Figure 3-51. Average diel variation 	3-116
Figure 3-52. Distribution of correlations between 24-h avg PM2.5 and co-located 24-h avg PM10,
PM10-2.5, SO2, N02andC0	3-117
Figure 3-53. Distribution of correlations between 24-h avg PM10 and co-located 24-h avg PM2.5,
PM102.5, SO2, NO2 and CO	3-118
Figure 3-54. Schematic of organic composition of particulate emissions from gasoline-fueled
vehicles.	3-124
Figure 3-55. Source category contributions to PM2.5 at a number of sites in the East derived using
PMF.	3-128
Figure 3-56. Pearson correlation coefficients for source category contributions to PM2.5 between
the ten Regional Air Pollution Study/Regional Air Monitoring System (RAPS/RAMS)
monitoring sites in St. Louis.	3-129
Figure 3-57. Pearson correlation coefficients for source contributions to PM10 2.5 between the ten
Regional Air Pollution Study/Regional Air Monitoring System (RAPS/RAMS) monitoring
sites in St. Louis. 	3-129
Figure 3-58. Eight km southeast U.S. CMAQ-UCD domain zoomed over Tampa Bay, FL.	3-137
Figure 3-59. Two km southeast U.S. CMAQ-UCD domain zoomed over Tampa Bay, FL.	3-137
Figure 3-60. Hourly average CMAQ-UCD predictions and measured observations of NO (top), NO2
(middle), and total NOx (bottom) concentrations for May 1-31, 2002. 	3-138
Figure 3-61. CMAQ-UCD predictions and measured observations of ethene concentrations at
Sydney, FL for May 1-31, 2002.	3-139
Figure 3-62. CMAQ-UCD predictions and measured observations of isoprene concentrations at
Sydney, FL for May 1-31, 2002.	3-140
Figure 3-63. CMAQ-UCD predictions and measured observations of PM2.5 concentrations at St.
Petersburg, FL for May 1-31, 2002.	3-141
Figure 3-64. CMAQ-UCD predictions of HNO3" concentrations and corresponding measured
observations at Sydney, FL	3-144
Figure 3-65. CMAQ-UCD predictions of NH3 concentrations and corresponding measured
observations at Sydney, FL, for May 1-31, 2002.	3-145
Figure 3-66. CMAQ-UCD predictions of pN03~ concentrations and corresponding measured
observations at Sydney, FL, for 1-31 May, 2002.	3-146
Figure 3-67. CMAQ-UCD predictions of the ratio of HNO3 to total NO3 and corresponding measured
observations at Sydney, FL, for May 1-31, 2002.	3-147
July 2009
xxii
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 3-68. CMAQ-UCD predicted size and chemical-form fractions of total NO3" for days in May
2002 with measured observations. Measured concentrations (top panel); 8 km solution
(middle panel); 2 km solution (bottom panel).	3-148
Figure 3-69. Scatter plot of total nitrate (FINO3 plus pl\l03~) wet deposition (mg N/m2/yr) of the
model mean versus measurements	3-150
Figure 3-70. Scatter plot of total SO42" wet deposition (mg S/m2/yr) of the model mean versus
measurements	3-150
Figure 3-71.	CMAQ modeling domains for the OAQPS risk and exposure assessments	3-151
Figure 3-72.	12-km EUS Summer sulfate PM.	3-152
Figure 3-73.	12-km EUS Winter nitrate PM. 	3-153
Figure 3-74.	12-km EUS Winter total nitrate (HNQ3 + total pl\l03~).	3-154
Figure 3-75.	12-km EUS annual sulfate wet deposition. 	3-155
Figure 3-76.	12-km EUS annual nitrate wet deposition.	3-156
Figure 3-77. CMAQ vs. measured air concentrations from east-coast sites in the IMPROVE, CSN
(labeled STN), and CASTNet sites in the summer of 2002 for sulfate (left) and
ammonium (right). Solid lines indicate a factor of 2 around the 1:1 line shown between
them.	3-157
Figure 3-78. Comparison of CMAQ-predicted and NADP-measured NFU+ wet deposition	3-157
Figure 3-79. CMAQ-predicted (red symbols and lines) and 12-h measured (blue symbols and lines)
NFI3 and SO42" surface concentrations at high and low concentration grid cells	3-158
Figure 3-80. Surface grid cell (layer 1) analysis of the sensitivity of l\IHx deposition and transport to
the change in NFI3 Vd in CMAQ.	3-159
Figure 3-81. Total column analysis for NFI3 (left) and l\IHx (right) showing modeled NFI3 emissions,
transformation, and transport throughout the mixed layer and up to the free
troposphere.	3-160
Figure 3-82. Range of influence (where 50% of emitted NFI3 deposits) from the high concentration
Sampson County grid cell in the June 2002 CMAQ simulation of Vd sensitivities.	3-161
Figure 3-83. Areal extent of the change in l\IHx range of influence as predicted by CMAQ	3-162
Figure 3-84. IMPROVE monitoring site locations.	3-165
Figure 3-85. 12-km EUS Summer SO42" PM.	3-169
Figure 3-86 12-km EUS Winter NOs PM.	3-170
Figure 3-87. 12-km EUS Winter total nitrate (HNQ3 + total particulate NO3").	3-170
Figure 3-88. Monthly average of PM2.5 concentrations measured at IMPROVE sites in the East and
Midwest for 2004.	3-172
Figure 3-89. Monthly average of PM2.5 concentrations measured at IMPROVE sites in the West for
2004.	3-173
Figure 3-90. Distribution of PM2.5 concentrations measured at IMPROVE sites in the East and
Midwest for 2004.	3-174
Figure 3-91. Distribution of PM2.5 concentrations measured at IMPROVE sites in the West for 2004.	3-175
Figure 3-92. Model of total personal exposure to PM as a function of ambient and nonambient
sources.	3-180
July 2009
xxiii
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 3-93. Distribution of time sample population spends in various environments, from the
National Human Activity Pattern Survey.	3-186
Figure 3-94. Total exposure to SCU2" as a function of measured ambient SCU2" concentration	3-194
Figure 3-95. Estimated ambient exposure to PM2.5 as a function of measured ambient PM2.5
concentration	3-195
Figure 3-96.	Total exposure to PM2.5 as a function of measured ambient PM2.5 concentration.	3-195
Figure 3-97.	Fm as a function of particle size.	3-203
Figure 3-98.	Apportionment of aliphatic carbon, carbonyl and SCU2" components	3-206
Figure 3-99.	Apportionment of infiltrated mechanically-generated	3-207
Figure 3-100.	Results of the positive matrix factorization model	3-208
Figure 3-101. Grid resolution of the CMAQ model in Philadelphia compared with distribution of
census tracts in which exposure assessment is performed. 	3-213
Figure 4-1. Diagrammatic representation of respiratory tract regions in humans.	4-4
Figure 4-2. Structure of lower airways with progression from the large airways to the alveolus.	4-5
Figure 4-3. Comparison of total and regional deposition results from the ICRP and MPPD models
for a resting breathing pattern (Vt = 625 mL, f = 12 min1) and corrected for particle
inhalability.	4-10
Figure 4-4. Comparison of total and regional deposition results from the ICRP and MPPD models
for a light exercise breathing pattern (Vt = 1250 mL, / = 20 min1) and corrected for
particle inhalability.	4-10
Figure 4-5. Total lung deposition measured in healthy adults (ultrafine, 11 M, 11 F, 31 ±4 years;
fine and coarse, 11 M, 11 F, 25 ± 4 years) during controlled breathing on a
mouthpiece.	4-11
Figure 4-6. Total deposition of hygroscopic sodium chloride and hydrophobic aluminosilicate
aerosols during oral breathing (Vt = 1.0 L; f = 15 min1).	4-22
Figure 4-7. Retention of poorly soluble particles (0.5-5 //m) in the alveolar region of the lung over
various mammalian species.	4-28
Figure 5-1.	PM oxidative potential. 	5-2
Figure 5-2.	PM stimulates pulmonary cells to produce ROS/RNS.	5-4
Figure 5-3.	PM activates cell signaling pathways leading to pulmonary inflammation. 	5-5
Figure 5-4.	Potential pathways for the effects of PM on the respiratory system.	5-9
Figure 5-5.	Potential pathways for the effects of PM on the cardiovascular system.	5-17
Figure 6-1. Excess risk estimates per 10//g/m3 increase in PM10, PM2.5 and PM10 2.5 for studies of
CVD ED visits* and hospitalizations. Studies represented in the figure include all
multicity studies. Single-city studies conducted in the U.S. or Canada are also included.	6-85
Figure 6-2. Excess risk estimates per 10//g/m3 increase in PM10, PM2.5, PM10 2.5 for studies of
EDvisits * and hospitalizations for IHD and Ml. 	6-89
Figure 6-3.	Excess risk estimates per 10//g/m3 increase in PM10, PM10 2.5 and PM10 2.5 for studies
of CHF ED visits* and hospitalizations.	6-94
Studies represented in the figure include all multicity studies. Single-city studies conducted in the U.S. and
Canada are also included.	6-94
July 2009
xxiv
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 6-4. Excess risk estimates per 10 |jg/m3 increase in PM10 and PM2.5 for studies of ED
visits* and hospitalizations for CBVDs.	6-99
Figure 6-5. Excess risk estimates per 10 |jg/m3 increase in PM2.5, and PM10 2.5 for studies of ED
visits* and hospitalizations for CVDs.	6-101
Figure 6-6. Combined random-effect estimate of the dose-response relationship between Ml
emergency hospital admissions and PM10, computed by fitting a piecewise linear
spline, with slope changes at 20 |jg/m3 and 50 (jg/m3.	6-103
Figure 6-7. Respiratory symptoms and/or medication use among asthmatic children following
acute exposure to PM2.5. 	6-117
Figure 6-8. Respiratory symptoms and/or medication use among asthmatic adults following acute
exposure to particles. 	6-125
Figure 6-9. Respiratory symptoms following acute exposure to particles and additional criteria
pollutants.	6-127
Figure 6-10. Excess risk estimates per 10 |jg/m3 24-h avg PM concentration for studies of ED visits
and hospitalizations for respiratory diseases in children.	6-189
Figure 6-11. Excess risks estimates per 10 |jg/m3 increase in 24-h avg PM for studies of ED visits
and hospitalizations for respiratory diseases among adults.	6-192
Figure 6-12. Excess risk estimates per 10 |jg/m3 increase in 24-h avg PM10, PM2.5 and PM10 2.5 for
studies of asthma ED visits* and hospitalizations.	6-195
Figure 6-13. Excess risks estimates per 10 |jg/m3 increase in 24-h avg PM10, PM2.5 and PM10 2.5 for
studies of COPD ED visits* and hospitalizations among older adults (65+ yr,
unlessother age group is noted). 	6-199
Figure 6-14. Excess risks estimates per 10 |jg/m3 increase in 24-h avg PM10, PM2.5 and PM10 2.5
forstudies of respiratory infection ED visits* and hospitalizations. Studies represented
in the figure include all multicity studies. Single-city studies conducted in the U.S. are
also included.	6-202
Figure 6-15. Excess risk estimates per 10//g/m3 increase in PM2.5, and PM10 2.5 for studies of ED
visits* and HAs for respiratory diseases. 	6-203
Figure 6-16. National and regional estimates of smooth seasonal effects for PM10 at a 1 -day lag
and their sensitivity to the degrees of freedom assigned to the smooth function of time
in the updated NMMAPS data 1987-2000.	6-224
Figure 6-17. Percent increase in the daily number of deaths, for all ages, associated with a
10-jtyg/m3 increase in PM10	6-231
Figure 6-18.	Effect modification by city characteristics in 20 U.S. cities.	6-234
Figure 6-19.	PM10 risk estimates (per 10 //g/m3) by individual-level characteristics.	6-236
Figure 6-20.	PM10 risk estimates (per 10 //g/m3) by location of death and by season.	6-237
Figure 6-21.	PM10 risk estimates (per 10 //g/m3) by contributing causes of deaths.	6-238
Figure 6-22. Summary of PM10 risk estimates (per 10 //g/m3) for all-cause mortality from recent
multicity studies. 	6-240
Figure 6-23. Percent increase in mortality for 10 |jg/m3 increase in the average of 0- and 1-day
lagged PM2.5, combined by climatic regions.	6-244
Figure 6-24. Empirical Bayes-adjusted city-specific effect estimates for total, cardiovascular, and
respiratory mortality for 10 |jg/m3 increase in the average of 0- and 1-day lagged
PM2.5 by decreasing mean 24-h avg PM2.5 concentrations. 	6-245
July 2009
XXV
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 6-25. Summary of all-cause mortality PM2.5 risk estimates per 10 //g/m3 by various effect
modifiers. 	6-248
Figure 6-26. Summary of PM2.5 risk estimates per 10//g/m3 for major underlying causes of death. 	6-250
Figure 6-27. Summary of PM2.5 risk estimates (per 10//g/m3) for cause-specific mortality for all
U.S.- and Canadian-based studies.	6-251
Figure 6-28. Percent increase in mortality for 10 |jg/m3 increase in the average of 0- and 1-day
lagged PM10 2.5, combined by climatic regions.	6-254
Figure 6-29. Empirical Bayes-adjusted city-specific effect estimates for total, cardiovascular, and
respiratory mortality for 10 |jg/m3 increase in the average of 0- and 1-day lagged
PMio 2.5 by decreasing 98th percentile of mean 24-h avg PM2.5 concentrations. 	6-255
Figure 6-30. Summary of PM10 2.5 risk estimates (per 10//g/m3) for cause-specific mortality for all
U.S.-, Canadian-, and international-based studies.	6-258
Figure 6-31. Percent increase in PM10 risk estimates (point estimates and 95% confidence intervals)
associated with a 5th-to-95th percentile: increase in PM2.5 and PM2.5 chemical
components. 	6-262
Figure 6-32. Sensitivity of the percent increase in PM10 risk estimates (point estimates and 95%
confidence intervals) associated with an interquartile increase in Ni.	6-262
Figure 6-33. Excess risk (CI) of total mortality per IQR of concentrations.	6-266
Figure 6-34. Relative risk and CI of cardiovascular mortality associated with estimated PM2.5
source contributions.	6-268
Figure 6-35. Concentration-response curves (spline model) for all-cause, cardiovascularrespiratory
and other cause mortality from the 20 NMMAPS cities. 	6-269
Figure 6-36. Percent increase in the risk death on days with PM10 concentrations in the ranges of
15-24, 25-34, 35-44, and 45 //g/m3 and greater, compared to a reference of days
when concentrations were below 15 //g/m3.	6-270
Figure 6-37. Combined concentration-response curves (spline model) for all-cause, cardiovascular,
and respiratory mortality from the 22 APHEA cities.	6-271
Figure 7-1. Risk estimates for the associations of clinical outcomes with long-term exposure to
ambient PM2.5 and PM10. 	7-21
Figure 7-2. Adjusted ORs and 95% CIs of symptoms and respiratory diseases associated with a
decline of 10 //g/m3 PM10 levels in Swiss Surveillance Program of Childhood Allergy
and Respiratory Symptoms	7-32
Figure 7-3. Effect of PM2.5 on the association of lung function with asthma.	7-34
Figure 7-4. Proportion of 18-year olds with a FEV1 below 80% of the predicted value plotted
against the average levels of pollutants 	7-40
Figure 7-5. Percent increase in postneonatal mortality per 10 //g/m3 in PM10, comparing risk for
total and respiratory mortality.	7-83
Figure 7-6. Mortality risk estimates associated with long-term exposure to PM2.5 from the Harvard
Six Cities Study (SCS) and the American Cancer Society Study (ACS).	7-123
Figure 7-7. Mortality risk estimates associated with long-term exposure to PM2.5 in recent cohort
stud	7-124
Figure 7-8. Plots of the relative risk of death from cardiovascular disease from the Women's
Health Initiative study	7-132
July 2009
xxvi
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 7-9.
The model-averaged estimated effect of a 10-//g/m3 increase in PM2.5 on all-cause
mortality
7-133
Figure 7-10.
Time course of relative risk of death after a sudden decrease in air pollution exposure
during the year 2000, assuming a steady state model (solid line) and a dynamic model
(bold dashed line). The thin dashed line refers to the reference scenario.
7-134
Figure 7-11.
Experts' mean effect estimates and uncertainty distributions for the PM2.5 mortality
7-137
Figure 9-1.
Important factors involved in seeing a scenic vista are outlined. Image-forming
information from an object is reduced (scattered and absorbed) as it passes through
the atmosphere to the human observer.
9-4
Figure 9-2.
Schematic of remote-area (top) and urban (bottom) nighttime sky visibility showing the
effects of PM and light pollution.
9-5
Figure 9-3.
Effect of relative humidity on light scattering by mixtures of ammonium nitrate and
ammonium sulfate.
9-8
Figure 9-4.
Estimated fractions of total particulate nitrate during each field campaign comprised
of ammonium nitrate, reacted sea salt nitrate (shown as NaNOs), and reacted soil dust
nitrate (shown as CafNOah).
9-17
Figure 9-5.
A scatter plot of the original IMPROVE algorithm estimated particle light scattering
versus measured particle light scattering.
9-21
Figure 9-6.
Scatter plot of the revised algorithm estimates of light scattering versus measured
light scattering.
9-21
Figure 9-7.
IMPROVE network PM species estimated light extinction for 2000 (left) and for 2004
(right).
9-24
Figure 9-8.
Mean estimated light extinction from PM speciation measurements for the first (top
left), second (top right), third (bottom left), and fourth (bottom right) calendar quarters
of 2004.
9-24
Figure 9-9.
Percent contributions of ammonium nitrate (left column) and ammonium sulfate (right
column) to particulate light extinction for each calendar quarter of 2004 (first through
fourth quarter arranqed from top to bottom).
9-26
Figure 9-10.
Percent contributions of organic mass (left column) and EC (right column) to
particulate light extinction for each calendar quarter of 2004 (first through fourth
quarter arranqed from top to bottom).
9-27
Figure 9-11.
Percent contributions of coarse mass (left column) and fine soil (right column) to
particulate light extinction for each calendar quarter of 2004 (first through fourth
quarter arranqed from top to bottom).
9-28
Figure 9-12.
IMPROVE Mean PM2.5 mass concentration determined by summing the major
components for the 2000-2004.
9-30
Figure 9-13.
IMPROVE and CSN (STN) mean PM2.5 mass concentration determined by summing the
major components for 2000-2004
9-30
Figure 9-14.
IMPROVE mean ammonium nitrate concentrations for 2000-2004.
9-31
Figure 9-15.
IMPROVE and CSN (STN) mean ammonium nitrate concentrations for 2000-2004.
9-31
Figure 9-16.
IMPROVE mean ammonium sulfate concentrations for 2000-2004.
9-32
Figure 9-17.
IMPROVE and CSN (STN) mean ammonium sulfate concentrations for 2000-2004.
9-32
Figure 9-18.
IMPROVE monitored mean organic mass concentrations for 2000-2004.
9-34
July 2009	xxvii	DRAFT - DO NOT CITE OR QUOTE

-------
Figure 9-19.	IMPROVE and CSN (STN) mean organic mass concentrations for 2000-2004.	9-35
Figure 9-20.	IMPROVE mean EC concentrations for 2000-2004. 	9-35
Figure 9-21.	IMPROVE and CSN (STN) mean EC concentrations for 2000-2004.	9-36
Figure 9-22.	IMPROVE mean fine soil concentrations for 2000-2004.	9-36
Figure 9-23.	IMPROVE and CSN (STN) fine soil concentrations, 2000-2004.	9-37
Figure 9-24. Regional and local contributions to annual average PM2.5 by particulate SO42", nitrate
and total carbon (i.e., organic plus EC) for select urban areas based on paired
IMPROVE and CSN monitoring sites. 	9-38
Figure 9-25. IMPROVE mean coarse mass concentrations for 2000-2004.	9-39
Figure 9-26. Ten-yr (1995-2004) haze trends for the mean of the 20% best annual haze conditions.	9-40
Figure 9-27. Ten-yr (1995-2004) haze trends for the mean of the 20% worst annual haze
conditions.	9-40
Figure 9-28. Ten-yr trends in the 80th percentile particulate SO42" concentration based on
IMPROVE and CASTNet monitoring and net SO2 emissions from the National Emissions
Trends (NET) data base by region of the U.S.	9-42
Figure 9-29. Map of 10-yr trends (1994-2003) in haze by particulate nitrate contribution to haze
for the worst 20% annual haze periods.	9-43
Figure 9-30. Contributions of the Pacific Coast area to the ammonium sulfate (|jg/m3) at 84
remote-area monitoring sites in western U.S. based on trajectory regression for all
sample periods from 2000-2002 (dots denote locations of the IMPROVE aerosol
monitoring sites). 	9-46
Figure 9-31. Shows the IMPROVE monitoring sites in the WRAP region with at least three yr of
valid data and identifies the six sites selected to demonstrate the apportionment tools.	9-48
Figure 9-32. Particulate SO42" (a) and nitrate (b) source attribution by region using CAMx modeling
for six western remote area monitoring sites	9-50
Figure 9-33. Monthly averaged model predicted organic mass concentration apportioned into
primary and anthropogenic and biogenic secondary PM categories for the Olympic NP
(top) and San Gorgonio Wilderness (bottom) monitoring sites.	9-51
Figure 9-34. Monthly averaged model predicted organic mass concentration apportioned into
primary and anthropogenic and biogenic secondary PM categories for the Yellowstone
NP (top) and Grand Canyon (Hopi Point) (bottom) monitoring sites.	9-52
Figure 9-35. Monthly averaged model predicted organic mass concentration apportioned into
primary and anthropogenic and biogenic secondary PM categories for the Badland NP
(top) and Salt Creek Wilderness (bottom) monitoring sites.	9-53
Figure 9-36. Comparison of carbon concentrations between Seattle (Puget Sound site) and Mt.
Rainer (left) and between Phoenix and Tonto (right) showing the background site
concentration (gray) and the urban excess concentration (black) for total, fossil and
contemporary carbon during the summer and winter studies.	9-54
Figure 9-37. Average contemporary fraction of PM2.5 carbon for the summer (top) and winter
(bottom), estimated from IMPROVE monitoring data (6/04 to 2/06) based on EC/TC
ratios.	9-55
Figure 9-38. Results of the weighted emissions potential tool applied to primary OC emissions (top)
and EC emissions (bottom) for the baseline and projected 2018 emissions inventories
for Olympic NP	9-57
July 2009
xxviii
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 9-39. Results of the weighted emissions potential tool applied to primary OC emissions (top)
and EC emissions (bottom) for the baseline and projected 2018 emissions inventories
for San Gorgonio W.	9-58
Figure 9-40. Results of the weighted emissions potential tool applied to primary OC emissions (top)
and EC emissions (bottom) for the baseline and projected 2018 emissions inventories
for Yellowstone NP. 	9-60
Figure 9-41. Results of the weighted emissions potential tool applied to primary OC emissions (top)
and EC emissions (bottom) for the baseline and projected 2018 emissions inventories
for Grand Canyon NP.	9-61
Figure 9-42. Results of the weighted emissions potential tool applied to primary OC emissions (top)
and EC emissions (bottom) for the baseline and projected 2018 emissions inventories
for Badlands NP.	9-62
Figure 9-43. Results of the weighted emissions potential tool applied to primary OC emissions (top)
and EC emissions (bottom) for the baseline and projected 2018 emissions inventories
for Salt Creek W.	9-63
Figure 9-44. BRAVO study haze contributions for Big Bend NP, TX during a 4-mo period in 1999.	9-65
Figure 9-45. Maps of spatial patterns for average annual particulate nitrate measurements (top),
and for ammonia emissions for April 2002 from the WRAP emissions inventory
(bottom).	9-66
Figure 9-46. Maps of spatial patterns of annual NO (left) and NO2 (right) emissions for 2002 from
the WRAP emissions inventory.	9-67
Figure 9-47. Midwest ammonia monitoring network. 	9-69
Figure 9-48. Upwind transport probability fields associated with high particulate
nitrateconcentrations 	9-70
Figure 9-49. Trajectory probability fields for periods with high particulate SO42" measured at
Underhill, VT and Brigantine, NJ (shown as white stars) associated with oil-burning
trace components (left) and with coal-burning trace components (right).	9-71
Figure 9-50. Scatter plots of particulate SO42" (left) and particulate nitrate and organic mass (right)
versus nephelometer measured particle light scattering for Acadia NP, ME.	9-72
Figure 9-51. CMAQ air quality modeling projections of visibility responses on the 20% worst haze
days at Great Smoky Mountains NP, NC (top) and Swanquarter Wilderness, NC
(bottom) to 30% reductions from a projected 2009 emission inventory of
visibility-reducing pollutants by source category and geographic areas.	9-73
Figure 9-52. Aerosol radiative forcing.	9-88
Figure 9-53. Global average radiative forcing (RF) estimates and uncertainty ranges in 2005,
relative to the pre-industrial climate. 	9-90
Figure 9-54. Probability distribution functions (PDFs) for anthropogenic aerosol and GHG RFs.	9-91
Figure 9-55. The clear-sky forcing efficiency Et	9-92
Figure 9-56. A composite of MODIS/Terra observed aerosol optical depth (at 550 nm, green light
near the peak of human vision) and fine-mode fraction that shows spatial and seasonal
variations of aerosol types.	9-99
Figure 9-57. Oregon fire on September 4, 2003 as observed by MISR: 	9-100
Figure 9-58. Global maps at 18 km resolution showing monthly average	9-101
July 2009
xxix
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 9-59. A dust event that originated in the Sahara desert on 17 August 2007 and was
transported to the Gulf of Mexico.	9-103
Figure 9-60. A constellation of five spacecraft that overfly the Equator	9-104
Figure 9-61. Geographical coverage of active AERONET sites in 2006.	9-107
Figure 9-62. Comparison of the mean concentration (pg/rn3) and standard deviation of the modeled
(STEM) aerosol chemical components	9-110
Figure 9-63. Location of aerosol chemical composition measurements with aerosol mass
spectrometers.	9-112
Figure 9-64. Scatterplots of the submicrometer POM measured during NEAQS versus A) acetylene
and B) iso-propyl nitrate. 	9-113
Figure 9-65. Comparison of annual mean aerosol optical depth (AOD)	9-116
Figure 9-66. Percentage contributions of individual aerosol components 	9-116
Figure 9-67. Geographical patterns of seasonally (MAM) averaged aerosol optical depth 	9-119
Figure 9-68. Summary of observation- and model-based (denoted as OBS and MOD, respectively)
estimates of clear-sky, annual average DRF at the TOA and at the surface.	9-120
Figure 9-69. Scatter plots showing mean cloud drop effective radius (re) vs. aerosol extinction
coefficient	9-127
Figure 9-70. Sampling the Arcic Haze. Pollution and smoke aerosols can travel long distances, from
mid-latitudes to the Arctic, causing "Arctic Flaze."	9-135
Figure 9-71. Global annual averaged AOD (upper panel) and aerosol mass loading (lower panel)	9-142
Figure 9-72. Aerosol direct radiative forcing in various climate and aerosol models. Observed values
are shown in the top section.	9-146
Figure 9-73. Aerosol optical thickness and anthropogenic shortwave all-sky radiative forcing from
the AeroCom study.	9-148
Figure 9-74. Radiative forcing from the cloud albedo effect (1st aerosol indirect effect) in the global
climate models used from	9-149
Figure 9-75. Anthropogenic impact on cloud cover, planetary albedo, radiative flux at the surface
(while holding sea surface temperatures and sea ice fixed) and surface air temperature
change from the direct aerosol forcing (top row), the first indirect effect (second row)
and the second indirect effect (third row).	9-151
Figure 9-76. Global average present-day short wave cloud forcing at TOA (top) and change in whole
sky net outgoing shortwave radiation (bottom) between the present-day and pre-
industrial simulations for each model in each experiment.	9-152
Figure 9-77. Direct radiative forcing by anthropogenic aerosols in the GISS model (including
sulfates, BC, OC and nitrates).	9-161
Figure 9-78. Percentage of aerosol optical depth in the GISS, left, based on Liu et al. (2006,
190422), provided by A. Lacis, GISS, and GFDL, right, from Ginoux et al. (2006,
190582). Models associated with the different components 	9-163
Figure 9-79. Most probable aerosol altitude (in pressure, hPa) from the GISS model in January (top)
and July (bottom).	9-165
Figure 9-80. Time dependence of aerosol optical thickness (left) and climate forcing (right).	9-167
Figure 9-81. Change in global mean ocean temperature (left axis) and ocean heat content (right
axis) for the top 3000 m due to different forcings in the GFDL model.	9-168
July 2009
XXX
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 9-82. Visible (wavelength 0.55 |jm) optical depth estimates of stratospheric SCk2 aerosols
formed in the aftermath of explosive volcanic eruptions that occurred between 1860
and 2000.	9-176
Figure 9-83. The transfer of POPs between the major abiotic compartments of the Arctic. Shaded
arrows represent inputs/outputs of POPs to the Arctic.	9-184
Figure 9-84. Relationship of plant nutrients and trace metals with vegetation.	9-207
Annex Figures
Figure A-1.	PM2.5 monitor distribution in comparison with population density, Atlanta, GA.	A-54
Figure A-2.	PM10 monitor distribution in comparison with population density, Atlanta, GA.	A-55
Figure A-3.	PM2.5 monitor distribution in comparison with population density, Birmingham, AL.	A-56
Figure A-4.	PM10 monitor distribution in comparison with population density, Birmingham, AL.	A-57
Figure A-5.	PM2.5 monitor distribution in comparison with population density, Chicago, IL.	A-58
Figure A-6.	PM10 monitor distribution in comparison with population density, Chicago, IL.	A-59
Figure A-7.	PM2.5 monitor distribution in comparison with population density, Denver, CO. 	A-60
Figure A-8.	PM10 monitor distribution in comparison with population density, Denver, CO.	A-61
Figure A-9.	PM2.5 monitor distribution in comparison with population density, Detroit, Ml.	A-62
Figure A-10.	PM10 monitor distribution in comparison with population density, Detroit, Ml.	A-63
Figure A-11.	PM10 monitor distribution in comparison with population density, Detroit, Ml.	A-64
Figure A-12.	PM10 monitor distribution in comparison with population density, Houston, TX.	A-65
Figure A-13.	PM2.5 monitor distribution in comparison with population density, Los Angeles, CA.	A-66
Figure A-14.	PM10 monitor distribution in comparison with population density, Los Angeles, CA.	A-67
Figure A-15.	PM2.5 monitor distribution in comparison with population density. New York City, NY.	A-68
Figure A-16.	PM10 monitor distribution in comparison with population density. New York City, NY. 	A-69
Figure A-17.	PM2.5 monitor distribution in comparison with population density, Philadelphia, PA.	A-70
Figure A-18.	PM10 monitor distribution in comparison with population density, Philadelphia, PA. 	A-71
Figure A-19.	PM2.5 monitor distribution in comparison with population density. Phoenix, AZ.	A-72
Figure A-20.	PM10 monitor distribution in comparison with population density. Phoenix, AZ.	A-73
Figure A-21.	PM2.5 monitor distribution in comparison with population density, Pittsburgh, PA.	A-74
Figure A-22.	PM10 monitor distribution in comparison with population density, Pittsburgh, PA. 	A-75
Figure A-23.	PM2.5 monitor distribution in comparison with population density. Riverside, CA.	A-76
Figure A-24.	PM10 monitor distribution in comparison with population density. Riverside, CA.	A-77
Figure A-25.	PM2.5 monitor distribution in comparison with population density, Seattle, WA.	A-78
Figure A-26.	PM2.5 monitor distribution in comparison with population density, Seattle, WA.	A-79
Figure A-27.	PM2.5 monitor distribution in comparison with population density, St. Louis, MO.	A-80
Figure A-28.	PM10 monitor distribution in comparison with population density, St. Louis, MO.	A-81
Figure A-29.	Three-yr avg of 24-h PM2.5 Cu concentrations measured at CSN sites across the U.S.,
2005-2007.	A-82
July 2009
xxxi
DRAFT-DO NOT CITE OR QUOTE

-------
Figure A-30.
Three-yr avg of 24-h PM2.5 iron concentrations measured at CSN sites across the U.S.,
2005-2007
A-83
Figure A-31.
Three-yr avg of 24-h PM2.5 nickel concentrations measured at CSN sites across the
U.S., 2005-2007
A-84
Figure A-32.
Three-yr avg of 24-h PM2.5 lead concentrations measured at CSN sites across the U.S.,
2005-2007
A-85
Figure A-33.
Three-yr avg of 24-h PM2.5 selenium concentrations measured at CSN sites across the
U.S., 2005-2007
A-86
Figure A-34.
Three-yr avg of 24-h PM2.5 vanadium concentrations measured at CSN sites across the
U.S., 2005-2007
A-87
Figure A-35.
PM2.5 monitor distribution and major highways, Atlanta, GA.
A-89
Figure A-36.
Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Atlanta, GA.
A-90
Figure A-37.
PM2.5 inter-sampler correlations as a function of distance between monitors for
Atlanta, GA.
A-91
Figure A-38.
PM2.5 monitor distribution and major highways, Birmingham, AL.
A-92
Figure A-39.
Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Birmingham, AL.
A-93
Figure A-40.
PM2.5 inter-sampler correlations as a function of distance between monitors for
Birmingham, AL.
A-94
Figure A-41.
PM2.5 monitor distribution and major highways, Chicago, IL.
A-95
Figure A-42.
Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Chicago, IL.
A-97
Figure A-43.
PM2.5 inter-sampler correlations as a function of distance between monitors for
Chicago, IL.
A-99
Figure A-44.
PM2.5 monitor distribution and major highways, Denver, CO.
A-100
Figure A-45.
Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Denver, CO.
A-101
Figure A-46.
PM2.5 inter-sampler correlations as a function of distance between monitors for
Denver, CO.
A-102
Figure A-47.
PM2.5 monitor distribution and major highways, Detroit, Ml.
A-103
Figure A-48.
Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Detroit, Ml.
A-104
Figure A-49.
PM2.5 inter-sampler correlations as a function of distance between monitors for
Detroit, Ml.
A-106
Figure A-50.
PM2.5 monitor distribution and major highways, Houston, TX.
A-107
Figure A-51.
Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Houston, TX.
A-108
Figure A-52.
PM2.5 inter-sampler correlations as a function of distance between monitors for
Houston, TX.
A-109
Figure A-53.
PM2.5 monitor distribution and major highways. New York City, NY.
A-110
July 2009
xxxii DRAFT-DO NOT CITE OR QUOTE

-------
Figure A-54. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
New York City, NY.	A-112
Figure A-55 PM2.5 inter-sampler correlations as a function of distance between monitors for New
York City, NY. 	A-115
Figure A-56. PM2.5 monitor distribution and major highways, Philadelphia, PA.	A-116
Figure A-57. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Philadelphia, PA.	A-117
Figure A-58. PM2.5 inter-sampler correlations as a function of distance between monitors for
Philadelphia, PA.	A-119
Figure A-59. PM2.5 monitor distribution and major highways. Phoenix, AZ.	A-120
Figure A-60. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Phoenix, AZ.	A-121
Figure A-61. PM2.5 inter-sampler correlations as a function of distance between monitors for
Phoenix, AZ.	A-122
Figure A-62. PM2.5 monitor distribution and major highways. Riverside, CA.	A-123
Figure A-63. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Riverside, CA.	A-124
Figure A-64. PM2.5 inter-sampler correlations as a function of distance between monitors for
Riverside CA.	A-125
Figure A-65. PM2.5 monitor figudistribution and major highways, Seattle, WA. 	A-126
Figure A-66. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Seattle, WA. 	A-127
Figure A-67. PM2.5 inter-sampler correlations as a function of distance between monitors for
Seattle, WA. 	A-128
Figure A-68. PM2.5 monitor distribution and major highways, St. Louis, MO.	A-129
Figure A-69. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for St.
Louis, MO.	A-130
Figure A-70 PM2.5 inter-sampler correlations as a function of distance between monitors for St.
Louis, MO.	A-131
Figure A-71. PM10 monitor distribution and major highways, Atlanta, GA.	A-132
Figure A-72. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Atlanta, GA.	A-133
Figure A-73. PM10 inter-sampler correlations as a function of distance between monitors for
Atlanta, GA.	A-134
Figure A-74. PM10 monitor distribution and major highways, Birmingham, AL.	A-135
Figure A-75. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Birmingham, AL.	A-136
Figure A-76 PM10 inter-sampler correlations as a function of distance between monitors for
Birmingham, AL.	A-137
Figure A-77. PM10 monitor distribution and major highways, Chicago, IL.	A-138
Figure A-78. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Chicago, IL.	A-139
July 2009
xxxiii
DRAFT-DO NOT CITE OR QUOTE

-------
Figure A-79. PM10 inter-sampler correlations as a function of distance between monitors for
Chicago, IL.	A-141
Figure A-80. PM10 monitor distribution and major highways, Denver, CO.	A-142
Figure A-81. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Denver, CO.	A-143
Figure A-82. PM10 inter-sampler correlations as a function of distance between monitors for Denver,
CO.	A-144
Figure A-83. PM10 monitor distribution and major highways, Detroit, Ml.	A-145
Figure A-84. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Detroit, Ml.	A-146
Figure A-85. PM10 inter-sampler correlations as a function of distance between monitors for Detroit,
A-147
Figure A-86. PM10 monitor distribution and major highways, Houston, TX.	A-148
Figure A-87. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Houston, TX.	A-149
Figure A-88. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
New York City, NY.	A-150
Figure A-89. PM10 inter-sampler correlations as a function of distance between monitors for New
York City, NY. 	A-151
Figure A-90. PM10 monitor distribution and major highways, Philadelphia, PA.	A-152
Figure A-91. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Philadelphia, PA.	A-153
Figure A-92. PM10 inter-sampler correlations as a function of distance between monitors for
Philadelphia, PA.	A-154
Figure A-93. PM10 monitor distribution and major highways. Phoenix, AZ. 	A-155
Figure A-94. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Phoenix, AZ.	A-157
Figure A-95. PM10 inter-sampler correlations as a function of distance between monitors for
Phoenix, AZ.	A-159
Figure A-96. PM10 monitor distribution and major highways. Riverside, CA. 	A-160
Figure A-97. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Riverside, CA.	A-161
Figure A-98. PM10 inter-sampler correlations as a function of distance between monitors for
Riverside, CA.	A-163
Figure A-99. PM10 monitor distribution and major highways, Seattle, WA.	A-164
Figure A-100. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Seattle, WA. 	A-165
Figure A-101. PM10 inter-sampler correlations as a function of distance between monitors for
Seattle, WA. 	A-166
Figure A-102. PM10 monitor distribution and major highways, St. Louis, MO.	A-167
Figure A-103. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for St.
Louis, MO.	A-168
July 2009
xxxiv
DRAFT-DO NOT CITE OR QUOTE

-------
Figure A-104. PM10 inter-sampler correlations as a function of distance between monitors for St.
Louis, MO.	A-169
Figure A-105. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter,
c) spring, d) summer and e) fall derived using the SANDWICH method in Atlanta, GA.	A-170
Figure A-106. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Birmingham, AL.	A-171
Figure A-107. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Boston, MA.	A-172
Figure A-108. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Chicago, IL.	A-173
Figure A-109. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Denver, CO.	A-174
Figure A-110. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Detroit, Ml.	A-175
Figure A-111. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Houston, TX.	A-176
Figure A-112. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Los Angeles, CA.	A-177
Figure A-113. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in New York City,
NY.	A-178
Figure A-114. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Philadelphia.	A-179
Figure A-115. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Phoenix, AZ.	A-180
Figure A-116. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Pittsburgh, PA.	A-181
Figure A-117. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Riverside, CA.	A-182
Figure A-118. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Seattle, WA.	A-183
Figure A-119. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in St. Louis, MO.	A-184
Figure A-120. Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for Atlanta, GA, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.	A-185
Figure A-121. Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for Birmingham, AL, 2005-2007. The gray line represents the difference in
OCM calculated using material balance and blank corrected OC x 1.4.	A-185
Figure A-122. Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for Boston, MA, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.	A-186
July 2009
XXXV
DRAFT-DO NOT CITE OR QUOTE

-------
Figure A-123.
Figure A-124.
Figure A-125.
Figure A-126.
Figure A-127.
Figure A-128.
Figure A-129.
Figure A-130.
Figure A-131.
Figure A-132.
Figure A-133.
Figure A-134.
Figure A-135.
Figure A-136.
Figure A-137.
July 2009
Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for Chicago, IL, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.	A-186
Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for Denver, CO, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.	A-187
Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for Detroit, Ml, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.	A-187
Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for Houston, TX, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.	A-188
Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for Los Angeles, CA, 2005-2007. The gray line represents the difference in
OCM calculated using material balance and blank corrected OC x 1.4.	A-188
Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for New York City, NY, 2005-2007. The gray line represents the difference in
OCM calculated using material balance and blank corrected OC x 1.4.	A-189
Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for Philadelphia, PA, 2005-2007. The gray line represents the difference in
OCM calculated using material balance and blank corrected OC x 1.4.	A-189
Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for Phoenix, AZ, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.	A-190
Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for Pittsburgh, PA, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.	A-190
Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for Riverside, CA, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.	A-191
Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for Seattle, WA, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.	A-191
Seasonal patterns in PM2.5 chemical composition from city-wide monthly average
values for St. Louis, MO, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.	A-192
Diel plots generated from all available hourly FRM-like PM2.5 data, stratified by
weekday (left) and weekend (right), in Atlanta, GA. Included are the number of monitor
days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.	A-192
Diel plots generated from all available hourly FRM-like PM2.5 data, stratified by
weekday (left) and weekend (right), in Chicago, IL. Included are the number of monitor
days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.	A-193
Diel plots generated from all available hourly FRM-like PM2.5 data, stratified by
weekday (left) and weekend (right), in Houston, TX. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each
hour.	A-193
DRAFT-DO NOT CITE OR QUOTE

-------
Figure A-138
Figure A-139
Figure A-140
Figure A-141
Figure A-142.
Figure A-143.
Figure A-144.
Figure A-145.
Figure A-146.
Figure A-147.
Figure A-148
Figure A-149
Figure A-150.
July 2009
Diel plots generated from all available hourly FRM-like PM2.5 data, stratified by
weekday (left) and weekend (right), in New York City, NY. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each
hour.	A-194
Diel plots generated from all available hourly FRM-like PM2.5 data, stratified by
weekday (left) and weekend (right), in Pittsburgh, PA. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each
hour.	A-194
Diel plots generated from all available hourly FRM-like PM2.5 data, stratified by
weekday (left) and weekend (right), in Seattle, WA. Included are the number of monitor
days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.	A-195
Diel plots generated from all available hourly FRM-like PM2.5 data, stratified by
weekday (left) and weekend (right), in St. Louis, MO. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each
hour.	A-195
Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by
weekday (left) and weekend (right), in Atlanta, GA. Included are the number of monitor
days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.	A-196
Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by
weekday (left) and weekend (right), in Chicago, IL. Included are the number of monitor
days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.	A-196
Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by
weekday (left) and weekend (right), in Denver, CO. Included are the number of monitor
days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.	A-197
Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by
weekday (left) and weekend (right), in Detroit, Ml. Included are the number of monitor
days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.	A-197
Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by
weekday (left) and weekend (right), in Los Angeles, CA. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each
hour.	A-198
Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by
weekday (left) and weekend (right), in Philadelphia, PA. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each
hour.	A-198
Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by
weekday (left) and weekend (right), in Phoenix, AZ. Included are the number of monitor
days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.	A-199
Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by
weekday (left) and weekend (right), in Pittsburgh, PA. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each
hour.	A-199
Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by
weekday (left) and weekend (right), in Riverside, CA. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each
hour.	A-200
DRAFT-DO NOT CITE OR QUOTE

-------
151.
152.
153.
154.
155.
156.
157.
158.
159.
160.
161.
162.
163.
164.
165.
Diel plot generated from all available hourly FRM/FEM PMio data, stratified by
weekday (left) and weekend (right), in Seattle, WA. Included are the number of monitor
days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.	A-200
Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by
weekday (left) and weekend (right), in St. Louis, MO. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each
hour.	A-201
Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10 2.5, SO2, NO2
and CO and daily maximum 8-h avg O3 for Atlanta, GA, stratified by season (2005-
2007). One point is included for each available monitor pair. 	A-201
Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10 2.5, SO2, NO2
and CO and daily maximum 8-h avg O3 for Birmingham, AL, stratified by season (2005-
2007). One point is included for each available monitor pair. 	A-202
Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10 2.5, SO2, NO2 and
CO and daily maximum 8-h avg O3 for Boston, MA, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-202
Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10 2.5, SO2, NO2 and
CO and daily maximum 8-h avg O3 for Chicago, IL, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-203
Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10 2.5, SO2, NO2 and
CO and daily maximum 8-h avg O3 for Denver, CO, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-203
Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10 2.5, SO2, NO2 and
CO and daily maximum 8-h avg O3 for Houston, TX, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-204
Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10 2.5, SO2, NO2 and
CO and daily maximum 8-h avg O3 for Los Angeles, CA, stratified by season (2005-
2007). One point is included for each available monitor pair. 	A-204
Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10 2.5, SO2, NO2 and
CO and daily maximum 8-h avg O3 for Philadelphia, PA, stratified by season (2005-
2007). One point is included for each available monitor pair. 	A-205
Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10 2.5, SO2, NO2 and
CO and daily maximum 8-h avg O3 for Phoenix, AZ, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-205
Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10 2.5, SO2, NO2 and
CO and daily maximum 8-h avg O3 for Pittsburgh, PA, stratified by season (2005-
2007). One point is included for each available monitor pair. 	A-206
Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10 2.5, SO2, NO2 and
CO and daily maximum 8-h avg O3 for Riverside, CA, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-206
Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10 2.5, SO2, NO2
and CO and daily maximum 8-h avg O3 for St. Louis, MO, stratified by season (2005-
2007). One point is included for each available monitor pair. 	A-207
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10 2.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for Atlanta, GA, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-207
xxxviii
DRAFT-DO NOT CITE OR QUOTE

-------
166.
167.
167.
168.
169.
170.
171.
172.
173.
174.
175.
176.
177.
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10 2.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for Birmingham, AL, stratified by season (2005-
2007). One point is included for each available monitor pair. 	A-208
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM102.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for Boston, MA, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-208
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10 2.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for Chicago, IL, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-209
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10 2.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for Denver, CO, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-209
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10 2.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for Detroit, Ml, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-210
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10 2.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for Houston, TX, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-210
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10 2.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for Los Angeles, CA, stratified by season (2005-
2007). One point is included for each available monitor pair. 	A-211
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10 2.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for New York City, NY, stratified by season (2005-
2007). One point is included for each available monitor pair. 	A-211
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10 2.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for Philadelphia, PA, stratified by season (2005-
2007). One point is included for each available monitor pair. 	A-212
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10 2.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for Phoenix, AZ, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-212
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10 2.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for Pittsburgh, PA, stratified by season (2005-
2007). One point is included for each available monitor pair. 	A-213
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10 2.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for Riverside, CA, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-213
Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10 2.5 SO2, NO2 and
CO and daily maximum 8-h avg O3 for St. Louis, MO, stratified by season (2005-2007).
One point is included for each available monitor pair.	A-214
xxxix
DRAFT-DO NOT CITE OR QUOTE

-------
PM ISA Project Team
Executive Direction
Dr. John Vandenberg (Director)—National Center for Environmental Assessment-RTP
Division, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Ila Cote (Acting Director)—National Center for Environmental Assessment-RTP
Division, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Debra Walsh (Deputy Director)—National Center for Environmental Assessment-RTP
Division, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Mary Ross (Branch Chief)—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Scientific Staff
Dr. Lindsay Wichers Stanek (PM Team Leader)—National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeffrey Arnold—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Christal Bowman—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. James S. Brown—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Allen Davis—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jean-Jacques Dubois— Oak Ridge Institute for Science and Education, Postdoctoral
Research Fellow to National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Steven J. Dutton—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Erin Hines—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Douglas Johns—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Ellen Kirrane—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Dennis Kotchmar—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Thomas Long—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
July 2009
xli
DRAFT-DO NOT CITE OR QUOTE

-------
Dr. Thomas Luben—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Qingyu Meng—Oak Ridge Institute for Science and Education, Postdoctoral Research
Fellow to National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Kristopher Novak—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Joseph Pinto—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jennifer Richmond-Bryant—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Mary Ross—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. David Svendsgaard—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Lisa Vinikoor—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. William Wilson—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Lori White—National Institute for Environmental Health Sciences, Research Triangle
Park, NC
Technical Support Staff
Mattie Arnold—Senior Environmental Employee Program, National Center for
Environmental Assessment, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Laeda Baston—Senior Environmental Employee Program, National Center for
Environmental Assessment, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Kimberly Branch—Student Services Contractor, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ken Breito—Senior Environmental Employee Program, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Eleanor Jamison—Senior Environmental Employee Program, National Center for
Environmental Assessment, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Ryan Jones—Oak Ridge Institute for Science and Education, at National Center for
Environmental Assessment, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Erica Lee—Oak Ridge Institute for Science and Education, at National Center for
Environmental Assessment, U.S. Environmental Protection Agency, Research Triangle
Park, NC
July 2009
xlii
DRAFT-DO NOT CITE OR QUOTE

-------
Barbara Liljequist—Senior Environmental Employee Program, National Center for
Environmental Assessment, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Ellen Lorang—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Kelsey Matson—Student Services Contractor, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Sandy Pham—Student Services Contractor, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Olivia Phillpott—Senior Environmental Employee Program, National Center for
Environmental Assessment, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Deborah Wales—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Erica Wilson—Oak Ridge Institute for Science and Education, at National Center for
Environmental Assessment, U.S. Environmental Protection Agency, Research Triangle
Park, NC
Richard Wilson—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Barbara Wright—Senior Environmental Employee Program, National Center for
Environmental Assessment, U.S. Environmental Protection Agency, Research Triangle
Park, NC
July 2009
xliii
DRAFT-DO NOT CITE OR QUOTE

-------
Authors, Contributors, Reviewers
AUTHORS
Dr. Lindsay Wichers Stanek (PM Team Leader)—National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeffrey Arnold—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Christal Bowman—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. James S. Brown—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Allen Davis—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jean-Jacques Dubois—Oak Ridge Institute for Science and Education, Postdoctoral
Research Fellow to National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Steven J. Dutton—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Tara Greaver—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Erin Hines—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Douglas Johns—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Ellen Kirrane—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Dennis Kotchmar—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Thomas Long—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Thomas Luben—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Qingyu Meng—Oak Ridge Institute for Science and Education, Postdoctoral Research
Fellow to National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Kristopher Novak—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Joseph Pinto—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
July 2009
xliv
DRAFT-DO NOT CITE OR QUOTE

-------
Dr. Jennifer Richmond-Bryant—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Mary Ross—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Timothy J. Sullivan—E&S Environmental Chemistry, Inc., Corvallis, OR
Dr. David Svendsgaard—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Lisa Vinikoor—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. William Wilson—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Lori White—National Institute for Environmental Health Sciences, Research Triangle
Park, NC
Dr. Christy Avery—University of North Carolina, Chapel Hill, NC
Dr. Kathleen Belanger —Center for Perinatal, Pediatric and Environmental Epidemiology,
Yale University, New Haven, CT
Dr. Michelle Bell—School of Forestry & Environmental Studies, Yale University, New
Haven, CT
Dr. William D. Bennett—Center for Environmental Medicine, Asthma and Lung Biology,
University of North Carolina, Chapel Hill, NC
Dr. Matthew J. Campen—Lovelace Respiratory Research Institute, Albuquerque, NM
Dr. Leland B. Deck— Stratus Consulting, Inc., Washington, DC
Dr. Janneane F. Gent—Center for Perinatal, Pediatric and Environmental Epidemiology,
Yale University, New Haven, CT
Dr. Yuh-Chin Tony Huang—Department of Medicine, Division of Pulmonary Medicine,
Duke University Medical Center, Durham, NC
Dr. Kazuhiko Ito—Nelson Institute of Environmental Medicine, NYU School of Medicine,
Tuxedo, NY
Mr. Marc Jackson—Integrated Laboratory Systems, Inc., Research Triangle Park, NC
Dr. Michael Kleinman—Department of Community and Environmental Medicine,
University of California, Irvine
Dr. Sergey Napelenok—National Exposure Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Marc Pitchford—National Oceanic and Atmospheric Administration, Las Vegas, NV
Dr. Les Recio—Genetic Toxicology Division, Integrated Laboratory Systems, Inc., Research
Triangle Park, NC
Dr. David Quincy Rich—Department of Epidemiology, University of Medicine and Dentistry
of New Jersey, Piscataway, NJ
Dr. Timothy Sullivan— E&S Environmental Chemistry, Inc., Corvallis, OR
July 2009
xlv
DRAFT-DO NOT CITE OR QUOTE

-------
Dr. George Thurston—Department of Environmental Medicine, NYU, Tuxedo, NY
Dr. Gregory Wellenius—Cardiovascular Epidemiology Research Unit, Beth Israel
Deaconess Medical Center, Boston, MA
Dr. Eric Whitsel—Departments of Epidemiology and Medicine, University of North
Carolina, Chapel Hill, NC
CONTRIBUTORS
Dr. Philip Bromberg—Department of Medicine, University of North Carolina, Chapel Hill,
NC
Mr. Michael Burr—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Turhan Carroll—Student Services Contractor, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Rosana Datti—Student Services Contractor, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Neil Frank—Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Jonathan Krug—Student Services Contractor, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Katie Lane—Student Services Contractor, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Phil Lorang—Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Christina Miller—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Irina Mordukhovich—Oak Ridge Institute for Science and Education, at National
Center for Environmental Assessment, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Elizabeth Oesterling Owens—Oak Ridge Institute for Science and Education,
Postdoctoral Research Fellow to National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Adam Reff—Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Vicki Sandiford— Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Mark Schmidt—Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Angelina Schultz—Student Services Contractor, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Kirsten Simmons—Student Services Contractor, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Genee Smith—Oak Ridge Institute for Science and Education, at National Center for
Environmental Assessment, U.S. Environmental Protection Agency, Research Triangle
Park, NC
July 2009
xlvi
DRAFT-DO NOT CITE OR QUOTE

-------
Mr. Kurt Susdorf—Student Services Contractor, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Rebecca Yang—Student Services Contractor, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
PEER REVIEWERS
Dr. Sara Dubowsky Adar, Department of Epidemiology, University of Washington, Seattle,
WA
Mr. Chad Bailey, Office of Transportation and Air Quality, Ann Arbor, MI
Mr. Richard Baldauf, Office of Transportation and Air Quality, Ann Arbor, MI
Dr. Prakash Bhave, National Exposure Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. George Bowker, Office of Atmospheric Programs, U.S. Environmental Protection
Agency, Washington, D.C.
Dr. Judith Chow, Division of Atmospheric Sciences, Desert Research Institute, Reno, NV
Dr. Dan Costa, Office of Research and Development, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Ila Cote, National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Robert Devlin, National Health and Environmental Effects Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. David DeMarini, National Health and Environmental Effects Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Neil Donahue, Department of Chemical Engineering, Carnegie Mellon University,
Pittsburgh, PA
Dr. Aimen Farraj, National Health and Environmental Effects Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Mark Frampton, Department of Environmental Medicine, University of Rochester
Medical Center, Rochester, NY
Mr. Neil Frank, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Tyler Fox, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jim Gauderman, Department of Environmental Medicine, Department of Preventive
Medicine, University of Southern California, Los Angeles, CA
Dr. Barbara Glenn, National Center for Environmental Research, U.S. Environmental
Protection Agency, Washington, D.C.
Dr. Terry Gordon, School of Medicine, New York University, Tuxedo, NY
Mr. Tim Hanley, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jack Harkema, Department of Pathobiology and Diagnostic Investigation, Michigan
State University, East Lansing, MI
July 2009
xlvii
DRAFT-DO NOT CITE OR QUOTE

-------
Ms. Beth Hassett-Sipple, Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Amy Herring, Department of Biostatistics, University of North Carolina, Chapel Hill,
NC
Dr. Israel Jirak, Department of Meteorology, Embry-Riddle Aeronautical University,
Pre scott, AZ
Dr. Mike Kleeman, Department of Civil and Environmental Engineering, University of
California, Davis, CA
Dr. Petros Koutrakis, Exposure, Epidemiology and Risk Program, Harvard School of Public
Health, Boston, MA
Dr. Sagar Krupa, Department of Plant Pathology, University of Minnesota, St. Paul, MN
Mr. John Langstaff, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Meredith Lassiter, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Phil Lorang, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Karen Martin, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Connie Meacham—National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Mr. Tom Pace, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jennifer Peel, Department of Environmental and Radiological Health Sciences, College
of Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins,
CO
Dr. Zackary Pekar, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Rob Pinder, National Exposure Research Laboratory, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Norm Possiel, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Sanjay Rajagopalan, Division of Cardiovascular Medicine, Ohio State University,
Columbus, OH
Dr. Pradeep Rajan, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Venkatesh Rao, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Joann Rice, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Mr. Harvey Richmond, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Victoria Sandiford, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
July 2009
xlviii
DRAFT-DO NOT CITE OR QUOTE

-------
Dr. Stefanie Sarnat, Department of Environmental and Occupational Health, Emory
University, Atlanta, GA
Dr. Frances Silverman, Gage Occupational and Environmental Health, University of
Toronto, Toronto, ON
Mr. Steven Silverman, Office of General Council, U.S. Environmental Protection Agency,
Washington, D.C.
Dr. Barbara Turpin, Department of Environmental Sciences, Rutgers University, New
Brunswick, NJ
Dr. Robert Vanderpool, National Exposure Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Alan Vette, National Exposure Research Laboratory, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Tim Watkins, National Exposure Research Laboratory, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Christopher Weaver, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Mr. Lewis Weinstock, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Ms. Karen Wesson, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Jason West, Department of Environmental Sciences and Engineering, University of
North Carolina, Chapel Hill, NC
Mr. Ronald Williams, National Exposure Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. George Woodall, National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Antonella Zanobetti, Department of Environmental Health, Harvard School of Public
Health, Boston, MA
July 2009
xlix
DRAFT-DO NOT CITE OR QUOTE

-------
Clean Air Scientific Advisory Committee for
Particulate Matter NAAQS
CHAIRPERSON
Dr. Jonathan Samet, Department of Preventive Medicine, Keck School of Medicine,
University of Southern California, Los Angeles, CA
MEMBERS
Dr. Lowell Ashbaugh, Crocker Nuclear Lab, University of California, Davis, CA
Dr. Ed Avol, Department of Preventive Medicine, Keck School of Medicine, University of
Southern California, Los Angeles, CA
Dr. Joseph Brain*, Department of Environmental Health, Harvard School of Public Health,
Harvard University, Boston, MA
Dr. Wayne Cascio, Brody School of Medicine, East Carolina University, Greenville, NC
Dr. Ellis B. Cowling*, Colleges of Natural Resources and Agriculture and Life Sciences,
North Carolina State University, Raleigh, NC
Dr. James Crapo*, Department of Medicine, National Jewish Medical and Research Center,
Denver, CO
Dr. Douglas Crawford-Brown, Department of Environmental Sciences and Engineering,
University of North Carolina at Chapel Hill, Chapel Hill, NC
Dr. H. Christopher Frey*, Department of Civil, Construction and Environmental
Engineering, College of Engineering, North Carolina State University, Raleigh, NC
Dr. David Grantz, Botany and Plant Sciences and Air Pollution Research Center, Riverside
Campus and Kearney Agricultural Center, University of California, Parlier, CA
Dr. Joseph Helble, Thayer School of Engineering, Dartmouth College, Hanover, NH
Dr. Rogene Henderson**, Lovelace Respiratory Research Institute, Albuquerque, NM
Dr. Philip Hopke, Department of Chemical Engineering, Clarkson University, Potsdam, NY
Dr. Donna Kenski*, Lake Michigan Air Directors Consortium, Rosemont, IL
Dr. Morton Lippmann, Nelson Institute of Environmental Medicine, New York University
School of Medicine, Tuxedo, NY
Dr. Helen Suh Macintosh, Environmental Health, School of Public Health, Harvard
University, Boston, MA
Dr. William Malm, National Park Service Air Resources Division, Cooperative Institute for
Research in the Atmosphere, Colorado State University, Fort Collins, CO
Mr. Charles Thomas (Tom) Moore, Jr., Western Regional Air Partnership, Western
Governors'Association, Fort Collins, CO
July 2009
DRAFT-DO NOT CITE OR QUOTE

-------
Dr. Robert F. Phalen, Center for Occupation & Environment Health, College of Medicine,
Department of Community and Environmental Medicine, Air Pollution Health Effects
Laboratory, University of California Irvine, Irvine, CA
Dr. Kent Pinkerton, Center for Health and the Environment, University of California,
Davis, CA
Mr. Richard L. Poirot, Air Pollution Control Division, Department of Environmental
Conservation, Vermont Agency of Natural Resources, Waterbury, VT
Dr. Armistead (Ted) Russell*, Department of Civil and Environmental Engineering, Georgia
Institute of Technology, Atlanta, GA
Dr. Frank Speizer, Channing Laboratory, Harvard Medical School, Boston, MA
Dr. Sverre Vedal, Department of Environmental and Occupational Health Sciences, School
of Public Health and Community Medicine, University of Washington, Seattle, WA
*Members of the statutory Clean Air Scientific Advisory Committee (CASAC) appointed by
the EPA Administrator.
**As immediate past CASAC Chair, Dr. Henderson is invited to participate in CASAC
advisory activities for FY 2009.
SCIENCE ADVISORY BOARD STAFF
Dr. Holly Stallworth, Economist and Designated Federal Officer, Clean Air Scientific
Advisory Committee, Environmental Economics Advisory Committee, Washington, D.C.
July 2009
li
DRAFT-DO NOT CITE OR QUOTE

-------
Acronyms and Abbreviations
a	alpha, ambient exposure factor (varies between 0 and l)
a-HCH	alpha-hexachlorocyclohexane
A	Angstrom exponent! determined by the contrast between the AOD at
two or more different wavelengths and is related to aerosol particle size
A	surface albedo (varies between 0 and l), the fraction of the total light
striking a surface that gets reflected back off that surface. An object
that has a high albedo (near l) is very bright! an object that has a low
albedo (near 0) is dark. The Earth's albedo is about 0.37. The Moon's is
about 0.12.
A549	human epithelial cell line
AAC	abdominal aortic calcium
AAS	atomic absorption spectrophotometry
AB	Alcian Blue stain
ABC	Asian Brown Cloud- (ABC- Climate and Other Environmental Impacts!
UNEP Report)
ABI	ankle-arm or resting blood pressure index
AC	air conditioning
Ace	acenaphthene
ACE-1	angiotensin converting enzyme
ACEAsia	(Asian Pacific Regional) Aerosol Characterization Experiment (study of
radiative forcing due to anthropogenic aerosols over the Asian Pacific
region)
ACGIH	American Conference of Governmental Industrial Hygienists
ACh	acetylcholine
Acl	acenaphthylene
ACP	accumulation mode particle
ACS	American Cancer Society
actinomycetes	bacteria, gram-positive, anaerobic, filamentous/branching growth
pattern
adrenal-4-binding protein (also known as steroidogenic factor-1 (SF-l),
Ad4BP/SF-l, or NR5A1)
Advanced Earth Observing Satellite-1
automobile diesel exhaust particles
angular distribution model(s), angular dependence model
asymmetric dimethylarginine
Asian Dust Network
AERONET
Ad4BP
ADEOS-1
A-DEP
ADM
ADMA
AD-Net
Ae
July 2009
i
DRAFT-DO NOT CITE OR QUOTE

-------
AeroCom	Aerosol Comparisons between Observations and Models
AERONET	NASA AERosol RObotic NETwork, aerosol observation system
Aerosols 99	1999 NOAA aerosol research cruise across the Atlantic
aethalometer	instrument that measures mass concentration of soot (BC or EC)
particles in aerosols or in an air stream
AF	atrial fibrillation
AGA	appropriate for gestational age
AGE	advanced glycation end product
AHR	airway hyperresponsiveness, airway hyperreactivity
AhR	arylhydrocarbon receptor
AHSMOG	California Seventh Day Adventist study
AI	aerosol index
AIC	Akaike's information criterion
AIM	ambient ion monitor
AIOP	2003 Aerosol Intensive Operating Period, a DOE program
air light	light scattered or diffused in the air by dust, haze, etc., especially as it
limits the visibility of distant, dark objects by causing them to blend
with the background sky
AIRS	Aerometric Information Retrieval System
Al	aluminum
albedo	a measure of light reflectance, See "A" abbreviation
ALT	air liquid interface
AM	alveolar macrophage(s)
AM, AMF	arbuscular mycorrhizal (fungi)
AMAP	Arctic Monitoring and Assessment Programme
AMDP	annual maximum of daily precipitation
AMI	acute myocardial infarction
AMS	aerosol mass spectrometry
Ang II	angiotensin II
ANOVA	analysis of variance
ANP	atrial natriuretic peptide
ANS	autonomic nervous system
Ant	anthracene
AOD	aerosol optical depth
AP-1	activator protein 1
APC	antigen presenting cell(s)
APCS	Absolute Principal Components Scores
July 2009
DRAFT-DO NOT CITE OR QUOTE

-------
APEX	Air Pollutants Exposure Model
APHEA	Air pollution and Health- a European Approach (study)
APHEA2	extended analysis of APHEA
APO	apocynin
ApoE	apolipoprotein E
ApoE-/'	mouse strain devoid of ApoE protein
APS	aerodynamic particle sizer, aerosol polarimetry sensor
aPTT	activated partial thromboplastin time
AQCD	Air Quality Criteria Document
AQI	Air Quality Index
AQM	air quality model
AQS	U.S. EPA Air Quality System database
Aqua	NASA satellite to study Earth's water cycle, radiative energy fluxes,
aerosols, vegetation cover on the land, phytoplankton and dissolved
organic matter in the oceans, and temperatures
AR4	Fourth Assessment Report (AR4) from the IPCC
ARCTAS	Arctic Research of the Composition of the Troposphere from Aircraft
and Satellites
ARD	Air Resources Division (U.S. Dept. of Interior, National Park Service)
ARDS	adult respiratory distress syndrome
ARI	acute respiratory infection
ARIC	Atherosclerosis Risk in Communities study
ARIES	Aerosol Research and Inhalation Epidemiology Study
ARM	DOE's Atmospheric Radiation Measurement program
ARQM	Air Quality Research Branch (Meteorological Service of Canada
Toronto)
ARS	Air Resource Specialists
As	arsenic
ASDNN5	mean of the standard deviation in all 5-min segments of a EKG 24 h
recording
ASOS	Automated Surface Observing System
ATOFMS	aerosol time-of-flight mass spectrometry
ATP	adenosine triphosphate
A-Train	a group of 5 afternoon overpass satellites (Aura, PARASOL, CALIPSO,
CloudSat, Aqua) that overfly the Equator, carrying sensors to study
aerosols
ATS	American Thoracic Society
AURA	(Latin for breeze) NASA satellite to study the Earth's ozone, air quality
and climate
July 2009
DRAFT-DO NOT CITE OR QUOTE

-------
avg	average
AVHRR	Advanced Very High Resolution Radiometer
J3	beta, beta coefficient, slope
6-HCH	beta-hexachlorocyclohexane(s)
36HSD	36-hydroxysteroid dehydrogenase
6TGF	6 transforming growth factor
bag	absorption by gases coefficient
BAP	absorption by particles coefficient
BEXT	light extinction coefficient, in units of Mm-1
BSG	scattering by gases coefficient
BSP	sum of light scattering by (aerosol) particles coefficient, aerosol light
scattering
Ba	barium
BaA	benz[a]anthracene
BAD	bronchial artery diameter
BAL	bronchoalveolar lavage
BALB/c	albino inbred mouse strain
BALF	bronchoalveolar lavage fluid
BALT	bronchus-associated lymphoid tissues
BAM	beta attenuation monitor
BaP	benzo[a]pyrene
BASIC	Brain Attack Surveillance in Corpus Christi (project)
BASE-A	(Biomass) Burning Airborne and Spaceborne Experiment - Amazon and
Brazil
BbF	benzo[b]fluoranthene
BC	black carbon
BCC-CMI	Beijing Climate Center — Carbon Mitigation Initiative
BCCR	Bjerknes Centre for Climate Research, Univ. of Bergen, Norway
BCM2.0	BCCR's Bergen Climate Model, Version 2
BEAS-2B	human bronchial epithelial cell line
BeP	benz[e]pyrene
BghiP, BpPe	benzo[g,h,i]perylene
BGT	beta-gauge technique
BH4	tetrahydrobiopterin
bhp	brake horsepower
BkF	benzo[k]fluoranthene
July 2009
iv
DRAFT-DO NOT CITE OR QUOTE

-------
BMI	body mass index
BMP	bone morphogenetic protein (e.g., BMP-6, BMP-15)
BN/BR	Brown Norway rat strain
BNP	brain natriuretic peptide, B-type natriuretic peptide
BOSS	BYU Organic Sampling System! multichannel diffusion denuder
sampling system
BP	blood pressure
BPM	blowing PM2.5
BpPe, BghiP	benzo[ghi]perylene
BPQ	benz(ayrene (BaP)-quinone
Br	bromine
BRAVO	Big Bend Regional Aerosol and Visibility Observational (Study)
BrdU	bromodeoxyuridine
BS	black smoke
BUC	bucillamine (N-[2-mercapto-2-methylpropionyl]-L-cysteine)
BVAIT	B-Vitamin Atherosclerosis Intervention Trial
BW, bw	body weight
BYU	Brigham Young University
C	carbon
C4	Center of Clouds, Chemistry and Climate
12C	carbon-12
13C	carbon-13
14C	carbon-14
Ceo(OH)24	water-soluble fullerene
CeoCs Ceo	fullerenes
Ca	calcium
CAA	Clean Air Act
CAAA	1977 Clean Air Act Amendments
CAAM	continuous ambient mass monitor
CAC	coronary artery calcification
CaC03	calcium carbonate
CAD	coronary artery disease
CALINE	California Line Source Dispersion Model
CALIOP	Cloud and Aerosol Lidar with Orthogonal Polarization
CALIPSO	Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations
July 2009
v
DRAFT-DO NOT CITE OR QUOTE

-------
CAM	Community Atmosphere Model (replaced NCAR CCM3 atmospheric
model)
CAMM	continuous ambient mass monitor
CAMP	Childhood Asthma Management Program
CAMx	comprehensive air quality model with extensions (modeling system)
Ca(N03)2	calcium nitrate
CAP	concentrated ambient particle
CAPMoN	Canadian Air and Precipitation Monitoring Network
CASAC	Clean Air Scientific Advisory Committee
CaS04	calcium sulfate
CASTNet	Clean Air Status and Trends Network
CATS	cumulative air toxics surface
CB	carbon black, chronic bronchitis
CB-Fe	carbon black particles artificially coated with Fe(II) salt.
CB(P)	carbon black (particles)
CBSA	Core-Based Statistical Area based on the 2000 U.S. Census
CB-V	carbon black particles artificially coated with a targeted concentration
of Vanadium (IV) salt
CBVD	cerebrovascular disease
CC16	Clara cell protein, Clara cell 16 protein
CCCma	Canadian Centre for Climate Modeling and Analysis
CCM3	NCAR Community Climate Model
CCM-MATCH	general circulation model (NCAR CCM), in tandem with a related
chemical transport model (MATCH), and observations
CCN	cloud condensation nuclei! cloud seed, small particles about which cloud
droplets coalesce (A typical raindrop is about 2 mm in diameter, a
typical cloud droplet is on the order of 0.02 mm, and a typical cloud
condensation nucleus (aerosol) is on the order of 0.0001 mm or 0.1
micrometer or greater in diameter.)
CCPM	continuous coarse particle monitor
CCSM3	NCAR community climate system model, version 3
CCSP	U.S. Climate Change Science Program
Cd	cadmium
CD1	albino outbred mouse strain
CDC	Centers for Disease Control and Prevention
CDE	conjugated diene
CDNC	cloud droplet number concentration
CDPHE	Colorado Department of Public Health and Environment
Ce	cerium
July 2009
vi
DRAFT-DO NOT CITE OR QUOTE

-------
CEN	European Committee for Standardization
CenRAP	Central Regional Air Planning Association
CERES	Clouds and the Earth's Radiant Energy System
CERFACS	European Centre for Research and Advanced Training in Scientific
Computation
CF	coronary flow, cystic fibrosis
CFA	coal fly ash
CFD	cystic fibrosis disease
CFR	Code of Federal Regulations
CGCM3.1	Canada's CCCma third generation coupled global climate model, runs
at 2 resolutions (T47 and T63)
cGMP	cyclic guanosine monophosphate
CH2CI2	methylene chloride
CH2O	formaldehyde
CH4	methane
CHAD	Consolidated Human Activity Database
CHD	chronic heart disease
CHF	congestive heart failure
CHL	crown heel length
CHO	Chinese hamster ovary cells
Chr	chrysene
CHS	Children's Health Study
CI	confidence interval
CIF	carbon-impregnated charcoal filter
CUT	The Chemical Industry Institute of Toxicology
CIMT	carotid intimal-medial thickness
CI	chlorine
CL	chemiluminescence
CLAMS	ARM's Chesapeake Lighthouse and Aircraft Measurements for
Satellites
CloudSat	satellite to provide observations necessary to advance understanding of
cloud abundance, distribution, structure, and radiative properties
CM	conditioned medium, cell culture medium
CMAQ	Community Multi-scale Air Quality modeling system
CMAR	CSIRO Marine and Atmospheric Research! houses Australia's leading
regional climate change modeling research
CMB	chemical mass balance
CMD	count median diameter
July 2009
DRAFT-DO NOT CITE OR QUOTE

-------
CNES	Centre National dMEtudes Spatiales, or National Space Study Center
Toulouse, France
CNP	carbon nano particle
CNRM	Centre National de Recherches Meteorologiques, Meteo France, France
CNRM-CM3	Center National Weather Research ¦ global coupled system, third
version (of the ocean-atmosphere model initially developed at
CERFACS (Toulouse, France)
CNS	central nervous system
Co	cobalt
CO	carbon monoxide
CO2	carbon dioxide
COD	coefficient of divergence
COH, CoH	coefficient of haze (a measurement of visibility interference in the
atmosphere, as the quantity of dust and smoke in a theoretical 1,000
linear feet of air).
Cong	U.S. Congress
CONUS	continental United States
COO'	carboxyl group
COPD	chronic obstructive pulmonary disease
CoPP	cobalt protoporphyrin a potent inhibitor of HO-1 (heme oxygenase)
COX-2	cyclooxygenase 2 enzyme
CPC	condensation particle counter
CPZ	capsazepine
Cr	chromium
C-R	concentration-response (relationship)
CRP	C-reactive protein
cryosphere	land or sea covered by snow or ice
Cs	cesium
137Cs	cesium-137
CS	cigarette smoke
CSA	Combined Statistical Area based on the 2000 U.S. Census
CSC	cigarette smoke condensates
CSE	cigarette smoke extract
cSHMT	cytosolic serine hydroxymethyltransferase gene
CSIRO	Commonwealth Scientific and Industrial Research Organization
(Australia National Science Agency)
CSIRO-Mk3.x	CSIRO Mark 3.x coupled climate system model
CSN	Chemical Speciation Network
July 2009
DRAFT-DO NOT CITE OR QUOTE

-------
CTM	chemistry-transport model, chemical transport model
Cu	copper
C11SO4	copper sulfate
Cu/Zn SOD	Cu/Zn superoxide dismutase
CUP	Current Use Pesticide (excluding DDT)
CV	cardiovascular, coefficient of variation
CVD	cardiovascular disease(s)
CVM	contingent valuation method — used in urban visibility valuation
studies
CYP	cytochrome P450
CYP 1A1	cytochrome P450 1A1
A	delta, change, difference
AFEVi	change in forced expiratory volume in one second
dso	50 percent cut point or 50 percent diameter
dae	aerodynamic diameter of a particle
D	diameter
Da	Dalton (measure of molecular weight)
DAAC	Distributed Active Archive Center
DAASS	Dry Ambient Aerosol Size Spectrometer
DABEX	Dust and Biomass-burning Experiment (in West Africa)
DAR	denuded aortic ring
DAX-1	x-chromosome gene-1
DBA	dibenzo(a,hnthracene
DBP	diastolic blood pressure
DC	dendritic cell
DC	diesel exhaust particles + cigarette smoke condensates
D.C.	District of Columbia
DC8	Douglas aircraft, originally designed as an 80 passenger airliner
DCF	direct climate forcing, 2',7'-dichlorofluorescin
DDT	dichlorodiphenyltrichloroethane, an insecticide
DE	diesel exhaust
July 2009
ix
DRAFT-DO NOT CITE OR QUOTE

-------
deciview haze index See dv, log transformation of light extinction, similar in many ways to
the decibel index for acoustic measurements. The name deciview is
used because of the similarity of the decibel scale in acoustics. Both use
10 times the logarithm of a ratio of a measured physical quantity to a
reference value to create scales that are approximately linear with
respect to changes as perceived by human senses. Because the index
increases from zero as haze increases, it is characterized as a haziness
index. Expressed in terms of extinction coefficient (bext) and visual
range (vr)- haziness (dv) = 10 In (bext/0.01 km'1) = 10 In (391 km/vr)
DEE	diesel exhaust extract
DEP	diesel exhaust particle
DEPAL	diesel exhaust particles methylene chloride extracts aliphatic (hexane)
DEPAR	diesel exhaust particles methylene chloride extracts aromatic
(hexane/methlene chloride)
DEPE	5 grams diesel exhaust particles in 5 mL PBS containing 0.05% Tween
80
DEPM	diesel exhaust particles methanol extract
DEPME	diesel exhaust particles methylene chloride extracts
DEPPO	diesel exhaust particles methylene chloride extracts polar (methylene
chloride/methanol)
Dex	dexamethasone
dFld	change fold, unit change in property
DFO	desferrioxamine (Desferral) an iron chelator
DFX	deferasirox (Exjade) an oral iron chelator
DHR	dihydrorhodamine 123
diel	a 24-hour period, usually involving a day and the adjoining night
DLCO	carbon monoxide diffusing capacity
DMEM	Dulbecco's modified Eagle's medium (culture medium)
DMSO	dimethyl sulfoxide
DMT1	divalent metal transporter-1 protein (transport and detoxification of
metals)
DMTU	dimethylthiourea
DNA	deoxyribonucleic acid
DOE	U.S. Department of Energy
downwelling	when ocean winds cause the surface water to move toward a coastline,
the nutritionally-depleted warmer surface water will move deeper
downward, thereby creating a downwelling current
dpc	days post conception
DPC	dodecylphosphocholine
DPCC	l,2-dipalmitoyl-SN-glycero-3-phosphocholine
DPI	diphenyleneiodonium
DPM	diesel particulate matter
July 2009
x
DRAFT-DO NOT CITE OR QUOTE

-------
DPPC
DRE
DRF
DRUM
DS
DSP
DTMA
DTPA
DU
dv
DYT
EAD
EANET
EARLINET
EarthCARE
EAST-AIRE
EBCT
EC
ECE-1
ECG, EKG
ECHAM5
ECHAMS5 -HAM
ECHO-G
ECRHS
EC/TC
ED
EDGAR
dipalmitoylphosphatidylcholine, the major phospholipid constituent of
pulmonary surfactant
direct radiative effects
direct radiative forcing
Davis Rotating Uniform size-cut Monitor — UC Davis aerosol sampling
technique
diffusion screens
daily sperm production
Dynamic mechanical thermal analysis
diethylene triamine pentaacetic acid
dust
deciview(s) unit, convenient, numerical method for presentation of
visibility values. The log scale of this visibility index, expressed in
deciview(s) (dv), is linear with respect to humanly-perceived changes in
visual air quality over its entire range, analogous to the decibel scale
for sound. The dv scale is near zero for a pristine environment and
increases as visibility degrades. One deciview is about a 10% change in
light extinction. See deciview haze index.
deep vein thrombosis
electrical aerosol detector
Acid Deposition Monitoring Network in East Asia
European Aerosol Research Lidar Network
Earth Clouds, Aerosols and Radiation Explorer (European Space
Agency satellite)
East Asian Study for Tropospheric Aerosols- An International Regional
Experiment
electron beam computed tomography
elemental carbon
endothelin converting enzyme
electrocardiogram, electrical activity of the heart over time, measured
by an electrocardiograph
European Centre Hamburg with Hamburg Aerosol Module, 5th
generation model of the ECHAM general atmosphere circulation model,
studies the climate of the troposphere
European Centre Hamburg, with Hamburg Aerosol Module;
(ECHAM4 + HOPE-G): Global climate model used at MPI, and
Meteorological Institute of the University of Bonn (Germany) and
Institute of KMA (Korea)
European Community Respiratory Health Survey
ratio of elemental carbon to total carbon
emergency room, emergency department
Emissions Database for Global Atmospheric Research, version 2
July 2009
xi
DRAFT-DO NOT CITE OR QUOTE

-------
EDTA	ethylenediaminetetraacetic acid
ED-XRF	energy dispersive X-ray fluorescence
EGM	electrogram
EGU	electricity-generating unit
EHC-93	Ottawa dust; urban air particulate matter PMio, collected in 1993 in
Ottawa, Canada
EKG, ECG	electrocardiogram
ELISA	enzyme-linked immunosorbent assay
EMECAS	Spanish Multi-centric Study on the Relation between Air Pollution and
Health
EMEP	European Monitoring and Evaluation Programme
eNO	exhaled nitric oxide
eNOS	endothelial nitric oxide synthase
EOS	Earth Observing System
EPA	U.S. Environmental Protection Agency
ER	estrogen receptor
ERBS	Earth Radiation Budget Satellite
ERK1/2	ERK-1 (MAPKp42) and ERK-2 (MAPKp44) (extracellular signal-
regulated kinases [in cell signaling pathway])
ESRL	NOAA Earth System Research Laboratory
ESTR	expanded simple tandem repeat
Et	forcing efficiency
ET	extrathoracic region of respiratory tract
ET	endothelin
ET-1	endothelin-1
ET-2	endothelin-2
ET-3	endothelin-3
ETa	endothelin A receptor subtype
ETb	endothelin B receptor subtype
ETS	environmental tobacco smoke
EU	endotoxin units
EXPOLIS	EXPOLIS (exposure + polis [Greek for city]); six-city European air
pollution study
F	breathing frequency
f	the ratio of ambient aerosol mass (wet) to dry aerosol mass M.
FO	function of variable inside parentheses
fosp(RH)	total light scattering coefficient at given relative humidity(RH) values
July 2009
xii
DRAFT-DO NOT CITE OR QUOTE

-------
faf	anthropogenic fraction of fine-mode fraction
ff	fine mode fraction
f(RH)	the unitless water growth term that depends on relative humidity
F	fine particles, aerosol direct Radiative Forcing (RF) at the Top of the
Atmosphere (TOA)
F344	Fisher 344 strain of rats
Fa.	adjusted forcings
FA	filtered air
FAC	ferric ammonium citrate
FBI	Federal Bureau of Investigation
FBS	fetal bovine serum
FCS	fetal calf serum
FDMS	Filter Dynamics Measurement System, a self referencing airborne
particulate monitor that provides a measurement of the mass of
nonvolatile and semivolatile airborne particulate matter
FDMS-TEOM	Filter Dynamics Measurement System ¦ Tapered Element Oscillating
Microbalance
Fe	effective (F e) forcings
Fe	iron
Fe2(S04)3	ferric sulfate
FeCl3	ferric chloride
FEF	forced expiratory flow
FEF25 75	mean forced expiratory flow over the middle half of the forced vital
capacity
FEF5o%	mid-expiratory flow
FEM	Federal Equivalent Method
FeNO	fractional exhaled nitric oxide
FERA	Fire and Environmental Research Applications Team (Pacific
Northwest Research Station, U.S. Forest Service)
FEVi	forced expiratory volume in one second
FGA	one fibrinogen alpha chain
FGB	one fibrinogen beta chain
FGOALS-gl.O	Flexible Global Ocean-Atmosphere-Land System Model, a global
climate system model developed at the Laboratory of Numerical
Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics
(LASG), Institute of Atmospheric Physics, Chinese Academy of
Sciences, Beijing, China
Fi	instantaneous forcing, simplest measure of radiative climate forcing
FID	flame ionization detection
FIMS	fast integrated mobility scanners
July 2009
xiii
DRAFT-DO NOT CITE OR QUOTE

-------
FINF	infiltration factors
FKHR	Proapoptotic Factor F0X01
Fie	fluorine
Flu	fluoranthene
FMD	flow-mediated dilation
forcing	changes in composition of the Earth's atmosphere, leading to changes in
the global energy balance! "forcing" the climate to change
FPG	formamidopyrimidine-DNA glycosylase
f-PM, FPM	fine particulate matter
FR	Federal Register
FRM	Federal Reference Method
FROSTFIRE	The landscape-scale prescribed research burn in the boreal forest of
interior Alaska, July 1999! conducted by FERA.
Fs	SST forcing(s), forcing driven by sea surface temperature (SST)
Fsfc	mean net solar flux at the (Earth) surface
FT	free troposphere
FTIR	Fourier transform infrared spectrometry
F/UFP	mix of fine and ultrafine particles, all < 2.5 jim
FVC	forced vital capacity
yGCS	gamma glutamylcysteine sythetase
Ga	gallium
GAM	generalized additive model
GATOR	Gas, Aerosol, Transport, and Radiation model
GATORG	Gas, Aerosol, Transport, Radiation, and General circulation model
GAW	Global Atmospheric Watch network
GBS	group B streptococcus
GC	gas chromatography
GCM(s)	general circulation model(s), global climate model
GCMOM	General Circulation, Mesoscale and Ocean Model
GC/MS	gas chromatography/mass spectrometry
GCS	gamma glutamylcysteine sythetase
GD	gestational day
GDF	growth differentiation factor (e.g., GDF-9)
GEE	generalized estimating equations, gasoline engine exhaust
GEIA	Global Emissions Inventory Activity
GEM	gaseous elemental mercury
July 2009
xiv
DRAFT-DO NOT CITE OR QUOTE

-------
GEOS-Chem	NASA Goddard Earth Observing System-CHEMistry (global 3-D
chemical transport model (CTM) for atmospheric composition)
GFAAS	graphite furnace atomic absorption spectrometry
GFAP	glial fibrillatory acidic protein
GFDL	NOAA's Geophysical Fluid Dynamics Laboratory
GFDL-CM2.X	GFDL Climate Models
GFED	Global Fire Emission Database (U.S. Oak Ridge National Laboratory)
GGT	gamma-glutamyltranspeptidasel a marker of epithelial injury
GHG	greenhouse gas
GIS	Geographic Information System
GISS	NASA Goddard Institute for Space Studies
GISS-AOM	GISS Atmosphere-Ocean Model climate prediction model
GISS-EH	GISS AOM for sea ice model
GISS-ER	GISS AOM for liquid sea model
GLAS	Geoscience Laser Altimeter System
GLM	generalized linear models
GM	geometric mean
GM-CSF	granulocyte macrophage colony-stimulating factor
GMD	NOAA Global Monitoring Division of the Earth System Research
Laboratory
GMS	Greater Mekong Subregion (Cambodia, the People's Republic of China,
Lao People's Democratic Republic, Myanmar, Thailand, and Viet Nam)
Core Environment Program
GOCART	NASA Goddard Chemistry Aerosol Radiation and Transport; a model
simulation of major tropospheric aerosol components
GOES	Geostationary Operational Environmental Satellite
GoMACCS	Texas Air Quality Study (TexAQS) - Gulf of Mexico Atmospheric
Composition and Climate Study
GPS	Global Positioning System
GSD	geometric standard deviation
GSFC	NASA Goddard Space Flight Center
GSH	glutathione
GSH-GSSG	ratio of reduced glutathione to glutathione disulfide (oxidized
glutathione)
GSO, GSNO	S-Nitrosoglutathione
GSSG	glutathione disulfide! oxidized glutathione
GST	glutathione-S-transferase
GSTM1	glutathione S-transferase polymorphism Ml
GSTP1	glutathione-S-transferase polymorphism PI
July 2009
xv
DRAFT-DO NOT CITE OR QUOTE

-------
GSTT1
GWP
h
H
H+
HR
H2
H2CO
H2O
H2O2
h2s
H2S04
H9c2
HA
HAEC
HAPC
haze index
HBE, HBEC
HC
HCB
HCH
HDL
HEAPSS
HEI
HEPA
HERO
HF
HFCD
HFE
Hg
Hg(0)
Hg(II)
HH
glutathione-S-transferase polymorphism T1
global warming potential
hour
atomic hydrogen, hydrogen radical, height, heart rate, high dose, high
exposure
hydrogen ion
heart rate
molecular hydrogen
formaldehyde
water
hydrogen peroxide
hydrogen sulfide
sulfuric acid
rat embryonic cardiomyocytes cell line
hospital admission
Human Aortic Endothelial Cell
Harvard ambient particle concentrator
expressed in deciview (dv) units ¦ log transformation of light extinction,
similar in many ways to the decibel index for acoustic measurements
(see dv)
Human Bronchial Epithelial cells
hydrocarbon(s); head circumference
hexachlorobenzene
hexachlorocyclohexane(s) (e.g. a-HCH, 6-HCH)
high density lipoprotein
Health Effects of Air Pollution among Susceptible Subpopulations
study
Health Effects Institute
high efficiency particle air (filter)
Health and Environmental Research Online, NCEA Database System
heart failure, high frequency (HRV parameter), high (dose/exposure)
filtered
High-Fat Chow Diet
HFE gene, HFE protein
mercury
gaseous elemental mercury
gaseous divalent (oxidized) mercury
hereditary hemochromatosis
July 2009
xvi
DRAFT-DO NOT CITE OR QUOTE

-------
HNRS	Hans Nixdorf Recall Study
HOI	heme oxygenase-1
hOGGl	8-hydroxyguanine DNA-glycosylase
HOPE-G	Hamburg Atmosphere-Ocean Coupled Circulation Model
hPA	hectopascal (unit of pressure); 1 hPA = 1 millibar, = 100 Pascals
hPAEC	human pulmonary artery endothelial cells
hPBMC	human peripheral blood mononuclear cells
HPLC	high pressure liquid chromatography
HPMF	high particulate matter filtered
hPMVEC	human pulmonary microvascular endothelial cells
HR	heart rate, hazard ratio, high level DE, Hunter College, NY
HRV	heart rate variability
HSD	176-hydroxysteroid dehydrogenase
HSP-70	heat shock protein
HSPH	Harvard School of Public Health
HSRL	NASA's High Spectral Resolution Lidar
HUVEC	human umbilical vein endothelial cells
hv	photon
H/W	height to width ratio
HWS	hardwood smoke
Hz	hertz
IC	ion chromatography
ICAM-1	intercellular adhesion molecule-1
ICARTT	International Consortium for Atmospheric Research on Transport and
Transformation
I CAS	Inner-City Asthma Study
ICD	implantable/implanted cardioverter defibrillator
ICD-9	International Classification of Disease 9th revision
ICD-10	International Classification of Disease 10th revision
ICESat	NASA Ice, Cloud and land Elevation Satellite
ICP-AES	inductively coupled plasma-atomic emission spectroscopy
ICP-MS	inductively-coupled plasma-mass spectrometry
ICR	imprinting control region, mouse strain
ICRP	International Commission on Radiological Protection
ID	identification number
IDP	indeno[l,2,3-c,d]pyrene
July 2009
xvii
DRAFT-DO NOT CITE OR QUOTE

-------
IFN-y
IFS
Ig
IGS
IHD
IIASA
IL
IM
iMDDC
IMPACT
IMPROVE
IN
INAA
INCA
index of refraction
INDOEX
INGV-SXG
INM-CM3.0
iNOS
INTEX-A
INTEX-B
INTEX-NA
I/O
IOM
i.p.
IP
IPCC
IPSL-CM4
interferon ¦ gamma
Integrated Forest Study
immunoglobulin (e.g., IgE)
International Genetic Standard
ischemic heart disease
International Institute for Applied Systems Analysis, Luxemburg
Austria
interleukin (e.g., IL-4, IL-5, IL-6, IL-8)
IMPACT (Michigan, USA)
immature monocyte-derived dendritic cells
NASA Langley Research Center's Interactive Modeling Project for
Atmospheric Chemistry and Transport (model)
Interagency Monitoring of Protected Visual Environment, operated by
the National Park Service Air Resources Division
ice nuclei
instrumental neutron activation analysis
Interactions between Chemistry and Aerosol, a LMDz model
refractive index ¦ a measure of how much the speed of light is seemingly
reduced by a medium (in a vacuum = 1.0J through air at STP =
1.000029, through ice = 1.31, etc.)
NOAA 1999 Indian Ocean Experiment
Istituto Nazionale di Geofisica e Vulcanologia, Italy! SINTEX-G model
(A Coupled Atmosphere Ocean Sea-Ice General Circulation Climate
Model evolving from the the SINTEX and SINTEX-F models)
Institute of Numerical Mathematics climate model, Russian Academy
of Science, Russia
inducible nitric oxide synthase
INTEX-Phase A (2004)
INTEX-Phase B (2006)
NASA Intercontinental Chemical Transport Experiment — North
America
indoor-outdoor ratio
Institute of Medicine
intraperitoneal (injection)
inhalable particle
Intergovernmental Panel on Climate Change
climate model developed at the Institut Pierre Simon Laplace (IPSL), a
federation of five research laboratories in the Paris area studying
terrestrial and planetary environments (Des Sciences de
L'Environnement)
IQR
interquartile range
July 2009
xviii
DRAFT-DO NOT CITE OR QUOTE

-------
Ir
IR
IRE
IRMS
ISA
ISO
ISO
IT
IUGG
IUGR
i.v.
JNK
kB
K
KC
kHz
kJ
KLH
km
km1
Kow
L, dL, mL, pL
L
La
LAC
LACE98
lag
lagO
lag 0-3
LBA-SMOCC
LBW
LC
iridium
incidence rate, infrared radiation
iron responsive element
isotope ratio mass spectrometer
Integrated Science Assessment
International Standards Organization
isoprene, 2-methyl analog of 1,3-butadiene
intratracheal, intratracheally (region or installation)
International Union of Geodesy and Geophysics
intrauterine growth restriction, intrauterine growth retardation
intravenous
c-jun N-terminal kinase
kappa B transcription factor
potassium
local neutrophil chemoattractant protein, the murine analog of
interleukin-8
kilohertz
kilojoules
keyhole limpet hemocyanin
kilometer
inverse kilometer
octanol-water partition coefficient
Liter, deciLiter, milliLiter, microLiter
low (dose / exposure)
lanthanum
light-absorbing carbon
Lindenberg Aerosol Characterization Experiment 1998 (The
Lindenberg Meteorological Observatory, Berlin)
time between one event and another
same day as the death, test, hospital, ED, clinic, physician visit! that
occurs on the same day as the exposure to the pollution
all the deaths, tests, hospital, ED, clinic, physician visits! that occured
on the same day as the exposure to the pollution and the two days
following the day of exposure
Brazil's Large-Scale Atmosphere-Biosphere Experiment in Amazon —
European Commission's Smoke Aerosols, Clouds, Rainfall and Climate-
Aerosols from Biomass Burning Perturb Global and Regional Climate
low birth weight
lethal concentration
July 2009
xix
DRAFT-DO NOT CITE OR QUOTE

-------
LC50
LDH
LDL
LDLR
LDYP
LES
LF
LF/HF
LIBS
lidar
LIF
light extinction
LITE
litterfall
LMD
LMDz
LMDZ-INCA
LMDZ-LOA
In
L-NAME
L-NMMA
LnRMSSD
InSDNN
LOA
LOESS
log
longwave emissivity
LOSU
Lpm
median lethal concentration
lactate dehydrogenase
low-density lipoprotein
low-density lipoprotein receptor
left developing ventricular pressure
large eddy simulations model
low frequency an HRV parameter
ratio of LF to HF an HRV parameter
laser induced breakdown spectroscopy
(light detection and ranging) A lidar instrument uses short pulses of
laser light to detect particles or gases in the atmosphere, like a radar
bounces radio waves off rain drops in clouds. The resulting reflected
laser radiation is measured and used to determine the location,
distribution and nature of the atmospheric particles.
leukemia inhibitory factor
fractional loss of intensity in a light beam per unit distance due to
scattering and absorption by the gases and particles in the air
NASA Lidar In-space Technology Experiment (1994)
leaves, branches, and other organic fuel deposition that fall to the
ground and decompose allow nutrients and organic matter to transfer
back to the soil
Laboratoire de Meteorologie Dynamique
Laboratoire de Meteorologie Dynamique with Zoom
IPSL Laboratoire de Meteorologie Dynamique's INteractive Chemistry
and Aerosols model
IPSL-LMDZ model with model from Laboratoire d'Optique
Atmospherique — Universite des Sciences et Technologies de Lille,
France
natural logarithm
arginine analog! N(G)-nitro-L- arginine methyl ester
N (G) -mono -methyl ¦ L ¦ ar ginine
natural log of RMSSD; measure of HRV
natural log of the standard deviation of NN intervals in an EKG
Laboratoire d'Optique Atmospherique, Universite des Sciences et
Technologies de Lille, France
locally weighted scatterplot smoothing
logarithm of a number raised to a given base (e.g., logio)
a material's ability to emit or release the longwave (thermal) energy
which it has absorbed
level of scientific understanding
liters per minute (L/min)
July 2009
xx
DRAFT-DO NOT CITE OR QUOTE

-------
LPMF	low particulate matter filtered
LPO	plasma lipid peroxides
LPS	lipopolysaccharide
LRAT	long range atmospheric transport
LROT	long range oceanic transport
LSCE	Laboratoire des Sciences du Climat et de l'Environnement, Gif sur
Yvette, France
LSDF	low-sulfur diesel fuel
LTB4	leukotriene B4
LTE4	leukotriene E4
LUA NRW	The North Rhine-Westphalia State Environment Agency
LUDEP	LUng Dose Evaluation Program
LUR	land use regression (model)
LV	left ventricle
LVEDP	left-ventricular end-diastolic pressure,
LVSP	left-ventricular systolic pressure, left ventricular developed pressure
LAV	ratio of lumen to wall
LWC	liquid water content
LWDE	Low Whole Diesel Exhaust
LWP	liquid water path
jig	microgram
jig/m3	micrograms per cubic meter (unit of chemical concentration in air)
jim	micrometer, micron
yM	microMolar (10 6 Molar)
m, cm, jim, nm	meter(s), centimeter(s), micrometer(s) = micron(s), nanometer(s)
M, mM, pM, nM, pM Molar, milliMolar (103 Molar), microMolar (10 6 Molar), nanoMolar (10 9
Molar), picoMolar (10 12 Molar)
M	dry aerosol mass, medium dose/exposure
ma	moving average
MAM	March-April-May
MAN	Maritime Aerosol Network
MANE-VU	Mid-Atlantic/Northeast Visibility Union
MAP	mitogen-activated protein, mean arterial pressure
MAPK	mitogen-activated protein kinase(s), MAP kinase
MARAMA	Mid Atlantic Regional Air Management Association
MATCH	NCAR Model of Atmospheric Transport and Chemistry
July 2009
xxi
DRAFT-DO NOT CITE OR QUOTE

-------
maximum
MBP	major basic protein
MCAPS	Medicare Air Pollution Study
Mch	methacholine
MCN	mixed carbon nanoparticle
MCP-1	monocyte chemoattractant protein 1
MCV	mean corpuscular volume
MD	mineral dust
MDA	malondialdehyde
MDCT	multidetector computed tomography
ME	Multilinear Engine
MEE	mass extinction efficiency, (in units of m2/g), a parameter linking the
particle mass concentration to light scattering
MEF	maximal expiratory flow
MEF50	maximum expiratory flow rate at 50% of vital capacity
MeHg	methyl mercury
MENTOR	Modeling Environment for Total Risk Studies
MEP	motorcycle exhaust particulate(s)
MEPE	motorcycle exhaust particulate extract (particle-free)
MESA	Multi-Ethnic Study of Atherosclerosis
MFFSR	multifilter rotating shadowband radiometer
mg/m3	milligrams per cubic meter
Mg	magnesium
MI	myocardial infarction
MIROC3.X	Model for Interdisciplinary Research on Climate, Center for Climate
System Research, University of Tokyo
Mie solution	solves Maxwell's equations for light scattering; valid for all possible
ratios of diameter to wavelength, used by optical scattering
measurements to accurately calculate light scattering and absorption of
particles
MILAGRO	Megacity Initiative- Local and Global Research Observations, study of
air pollution in Mexico City
min	minute (s), minimum
MINOS	MPI Mediterranean INtensive Oxidant Study
MIP-2	macrophage inflammatory protein-2
MIRAGE	Megacities Impact on Regional and Global Environment program
MIS	mullerian inhibiting substance
MISR	Multi-angle Imaging SpectroRadiometer
July 2009
xxii
DRAFT-DO NOT CITE OR QUOTE

-------
Mm	megameter [l million meters]
Mm'1	inverse megameter
Mm'Vfug/m3)	units used in light scattered per unit of mass concentration (mass
scattering efficiency), reduces to m2/g
MM	monocyte-derived macrophages
MM5	PSU/NCAR mesoscale model used to predict mesoscale atmospheric
circulation. (Pennsylvania State University / National Center for
Atmospheric Research, Mesoscale and Microscale Meteorology Division)
MMAD	mass median aerodynamic diameter
MMD	mass median diameter
MMEF	maximal mid-expiratory flow! synonymous with FEF2575
mmHg	millimeters of mercury
MMP	mitochondria membrane potential
MMP(2,9)	matrix metalloproteinase (2, or 9)
MMT	million metric tons
Mn	manganese
MN	micronuclei
MnS04	manganese sulfate
MnSOD	manganese superoxide dismutase
MnTBAP	manganese tetrakis (4-benzoic acid) porphyrin (membrane-permeable
SOD mimetic)
mo	month
MO_GO	satellite-model integration, MODIS over ocean and GOCART simulated
AOD
MO_MI_GO	satellite-model integration, MODIS over ocean and MISR over land and
GOCART simulated AOD
MOA	mode(s) of action
MODIS	MODerate resolution Imaging Spectroradiometer, an aerosol sensor
MOUDI	Micro-Orifice Uniform Deposit Impactor (a University of Minnesota
sampling technique)
MOZART	MOdel for Ozone and Related chemical Tracers! a chemical transport
model
MP	mid polar, myelopeptide
MPC	mean platelet component
MPF	median power frequency
MPG	synthetic aminothiol, N-(2-mercaptopropionyl) glycine
MPI	Max Planck Institute for Meteorology in Hamburg, Germany
MPLNET	NASA Micro-Pulse Lidar Network
MPO	myeloperoxidase
July 2009
xxiii
DRAFT-DO NOT CITE OR QUOTE

-------
MPPD	Multiple-Path Particle Dosimetry model
MPV	mean platelet volume
MRI	Meteorological Research Institute, Japan
MRI-CGCM	MRI coupled general circulation model
mRNA	messenger RNA
MRPO	Midwest Regional Planning Organization
ms	millisecond
MSA	metropolitan statistical area based on the 1990 U.S. Census
MSH	melanocyte stimulating hormone
MSHA	Mount St. Helen ash
MSU	monosodium urate crystals
MT	metric ton (1000 kg)
MTHFR	methylenetetrahydrofolate reductase
MTT	methyl thiazol tetrazolium
MV	motor vehicle
MWNT	multiplewall nanotube
M/Z	mass-to-charge ratio
N	nitrogen
N, n	number of observations
N2O	nitrous oxide
Na	sodium
Na2S04	sodium sulfate
NAAQS	National Ambient Air Quality Standards
NAC	N-acetylcysteine, a thiol antioxidant
NaCl	sodium chloride
NADPH	reduced form of nicotinamide adenine dinucleotide phosphate
NAG	N-acetyl-6-D-glucosaminidase
Na,K-ATPase	sodium-potassium adenosine triphosphatase
NAMS	National Ambient Monitoring Stations
NaN3	sodium azide
NaN03	sodium nitrate
nano-BAM	low pressure-drop ultrafine particle impactor coupled with a Beta
Attenuation Monitor
NAPAP	National Acid Precipitation Assessment Program
July 2009
xxiv
DRAFT-DO NOT CITE OR QUOTE

-------
NAPCA	National Air Pollution Control Administration, formerly in the
Environmental Health Service (Environmental Control Administration
and National Air Pollution Control Administration), Public Health
Service (PHS), U.S. Department of Health, Education, and Welfare
(HEW)
NAS	National Academy of Sciences
NASA	U.S. National Aeronautics and Space Administration
NASDA	National Space Development Agency, Japan
NATA	U.S. EPA's National Air Toxics Assessment
2-NB	2-nitrobenzanthrone
NC	total (particle) number concentration
NCo.oi o.i	10—100 nm number concentrations
NCAR	National Center for Atmospheric Research (Pennsylvania State
University)
NCC-MPSP	negatively charged carboxylate-modified polystyrene particle(s)
NCD	Normal Chow Diet
NCEA	National Center for Environmental Assessment
NCHS	National Center for Health Statistics
NCICAS	National Cooperative Inner-City Asthma Study
NCore	National Core (a multi pollutant network of measurement systems for
ambient particles, pollutant gases and meteorology)
Nd	drop number concentration (cloud droplets/cm3)
Nd-YAG	neodymium-doped yttrium aluminum garnet laser
NDDN	National Dry Deposition Network
NEAQS	NOAA New England Air Quality Study
nephelometer	an instrument for measuring the concentration of a suspension by its
scattering of a beam of light
NEI	National Emissions Inventory
NESCAUM	Northeast States for Coordinated Air Use Management
NET	National Emissions Trends database
net irradiance	the difference between incoming and outgoing radiation energy in a
climate system, measured in watts per square meter (W/m3)
NFkB	nuclear factor kappa-BJ transcription factor, light-chain enhancer of B
cells
NG	neutrophil granulocytes
NH	northern hemisphere
NH3	ammonia
NH4+	ammonium ion
NH4NO3	ammonium nitrate
July 2009
XXV
DRAFT-DO NOT CITE OR QUOTE

-------
(NH4)2S04	ammonium sulfate
NHANES	National Health and Nutrition Examination Survey
NHBE(C)	normal human bronchial epithelial cells
NHPAE	normal human pulmonary artery endothelial cells
NHS	Nurses' Health Study
Ni	nickel
NIOSH	National Institute for Occupational Safety and Health
NIST	National Institute of Standards and Technology
nm	nanometers
NMHC	non-methane volatile hydrocarbon
NMMAPS	U.S. National Morbidity, Mortality, and Air Pollution Study
NN intervals	normal-to-normal (NN or RR, sinus) time interval between each QRS
complex in the EKG
NO	nitric oxide
N02) NO2	nitrogen dioxide, nitrogen dioxide radical
NO3-	nitrate
NOAA	National Oceanic and Atmospheric Administration
NOAEL	no observed adverse effect level
NOS	nitric oxide synthase
NOS3	nitric oxide synthase 3
NOx	nitrogen oxides, oxides of nitrogen (NO + NO2)
NP	National Park
NPM	non-blowing PM2.5
NPOESS	National Polar-orbiting Operational Environment Satellite System
NPS	National Park Service, U.S. Department of the Interior
NR	not reported
NR5A1	nuclear receptor subfamily 5, group A, member 1 (previously known as
AD4BP/SF-1 or SF-l)
NRC	National Research Council
NRPB	National Radiological Protection Board
NSA	North Slope Alaska
NT	neurotrophin, nitrotyrosine
NWS	National Weather Service
NYHA	New York Heart Association (NYHA) Classification Scale
NYHA I	Class I- No symptoms at any level of exertion
NYHA II	Class II- Mild symptoms and slight limitation during regular activity
July 2009
xxvi
DRAFT-DO NOT CITE OR QUOTE

-------
NYHA III	Class III- Noticeable limitation due to symptoms, even during minimal
activity
NYHA IV	Class IV- Severe limitations. Experience symptoms even while at rest
O	oxygen
02	molecular oxygen
03	ozone
OAQPS	Office of Air Quality Planning and Standards
OC	organic carbon
OCM	organic carbon mass
OE	organic extracts
OGG1	8 oxo-guanine repair enzyme
OH, OH •	hydroxyl group, hydroxyl radical
8-OHdG	8-hydroxydeoxyguanosine
OM	organic matter
OMI	Ozone Monitoring Instrument
OMM	organic molecular marker
OR	odds ratio(s)
Orographic fog	formed as the air rises up a slope and will often envelope the summit.
When the humid rising air expands and cools, the cloud cannot hold
moisture as well as a warm cloud, and some of the moisture will fall as
rain on the windward slope and on the summit.
oncostatin M, a cytokine
Operational Street Pollution Model
ovalbumin
oxidation of LDL, marker of oxidative stress
8 -oxo ¦ 7 -hydrodeoxyguanosine
oxidized l-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphorylcholine (an
atherogenic oxidized phospholipid)
cytochrome P450
cytochrome P450 17-crhydroxylase
cytochrome P450 cholesterol side chain cleavage enzyme
90th percentile of the absolute difference in concentrations
Printex 90, industrial (Degussa) carbon black ultrafine particles
probability value
phosphorus
photoacoustic analyzer, physical activity, plasminogen activator,
pulmonary arterial, alveolar pressure
platelet-activating factor
OSM
OSPM
OVA
oxLDL
8-oxodG
ox-PAPC
P450
P450cl7
P450scc
P90
P90
P
P
PA
PAF
July 2009
xxvii
DRAFT-DO NOT CITE OR QUOTE

-------
PAH
PAI
PALMS
PAMCHAR
PAMS
PAR
PAR(s)
PARASOL
PARP
PAS
Pb
2°7pb
PBDE
PBL
PBMC
PBMM
PBP
PBS
PC
PCA
PCA-MPSP
PCB
PCDD
PCIS
PCM
PCPSP
PCR
PDF
pDR
PE
PEACE
PEC
PECAM-1
pedogenesis
polycyclic aromatic hydrocarbon(s)
plasminogen activator inhibitor, (e.g. PAI-l)
NOAA Particle Analysis by Laser Mass Spectrometry instrument
Chemical and Biological Characterisation of Ambient Air Coarse, Fine,
and Ultrafine Particles for Human Health Risk Assessment in Europe
Photochemical Assessment Monitoring Stations network
photosynthetically active radiation
Pulmonary Artery Rings
Polarization and Directionality of the Earth's Reflectances, coupled
with observations from a Lidar, a CNES satellite
poly(ADP-ribose) polymerase
Periodic Acid Schiff stain
lead
lead-207
polybrominated diphenyl ether
planetary boundary layer
peripheral blood mononuclear cell
peripheral blood monocyte-derived macrophages
primary biological particle(s)
phosphate buffered saline
synthetic carboxylate-modified particles
principal component analysis
positively-charged amine modified polystyrene particle
polychlorinated biphenyl(s)
polychlorinated dibenzo-p-dioxin
Personal Cascade Impactor Sampler
NCAR Parallel Climate Model
positively charged polystyrene particle
polymerase chain reaction
probability distribution functions
personal DataRam
post exposure, post exercise, phenylephrine
Pollution Effects on Asthmatic Children in Europe study
particulate elemental carbon
platelet endothelial cell adhesion molecule 1
soil evolution or soil formation ¦ the process in which soil is created
July 2009
xxviii
DRAFT-DO NOT CITE OR QUOTE

-------
PEF	peak expiratory flow (L/min)
PEFR	peak expiratory flow rate
PEFT	time to peak flow
PEM	personal exposure monitor
PEM-West	NASA Pacific Exploratory Missions in the western Pacific
Penh	enhanced pause (altered ventilatory timing)
Per	perylene
PESA	particle elastic scattering analysis
PFDE	particle free diesel exhaust
PGE2	prostaglandin E2
PGI2	prostacyclin
pH	scale of acidity (log of hydrogen ion concentration); decreased breath pH
is a biomarker for airway inflammation
Phe	phenanthrene
photoacoustic	based on the photoacoustic effect (thin discs emit sound when exposed
to laser beams, with the sound proportional to the light intensity)
photoacoustic spectroscopy is a sensitive technique to study
concentrations of gases down to the ppb or ppt levels
photopic	photopic vision, the vision of the eye under well-lit conditions
Phytochelatins	oligomers of glutathione, produced by the enzyme phytochelatin
synthase. They are found in plants, fungi, nematodes, and all groups of
algae, including all groups of algae, including cyanobacteria.
Phytochelatins act as chelators and are important for heavy metal
detoxification.
PI	post instillation, posterior interval, pulmonary inflammation
PICT	pollution-induced community tolerance
PILS	Particle Into Liquid Sampler
PILS-IC	Particle Into Liquid Sampler-Ion Chromatography
PIXE	Particle Induced X-ray Emission
PKA	protein kinase A/cAMP-dependent protein kinase
planet brightening an increase in the amount of sunlight reaching the planet's land surface
planet dimming	a decrease in the amount of sunlight reaching the
planet's land surface
PLS	partial least squares, projection to latent structures
PM	particulate matter
PMx	particulate matter of a specific size range. X refers to the diameter at
which the sampler collects 50% of the particles and rejects 50% of the
particles. Collection efficiency increases for particles with smaller
diameters and decreases for particles with larger diameters. The
variation of collection efficiency with size is given by a collection
efficiency curve. The definition of PMx is frequently abbreviated as
"particles with a nominal mean aerodynamic diameter less than or
equal to x jim.
July 2009
xxix
DRAFT-DO NOT CITE OR QUOTE

-------
PMx-y	particulate matter with a nominal mean diameter greater than x jj,m
and less than y (im where x and y are the numeric mean aerodynamic
or mobility diameters (jo.m).
PMo.i	particulate matter with a nominal mean mobility diameter less than or
equal to 0.1 (im (referred to as ultrafine PM)
PM2.5	particulate matter with a nominal mean aerodynamic diameter less
than or equal to 2.5 pm (referred to as fine PM)
PM10	particulate matter with a nominal mean aerodynamic diameter less
than or equal to 10 pm
PMio-2.5	particulate matter with a nominal mean aerodynamic diameter greater
than 2.5 pm and less than or equal to 10 pm (referred to as thoracic
coarse particulate matter or the course fraction of PM10) Concentration
may be measured with a dichotomous sampler or calculated as the
difference between measured PM10 and measured PM2.5 concentrations.
PMA	phorbol 12-myristate 13-acetate
PMF	particulate matter filtrate, positive matrix factorization
PM-HD	particulate matter at high concentration
PM-LD	particulate matter at low concentration
PMN	polymorphonuclear leukocytes
PN	particle number
PNC	particle number concentration, particle number count
PND, pnd	post-natal day
PNMD	particle number median diameter
PNN	proportion of interval differences of successive normal-beat intervals in
EKG
pNN50	proportion of interval differences of successive normal-beat intervals
greater than 50 ms in an EKG
PNNL	DOE Pacific Northwest National Laboratory
PNO3	particulate nitrate
POA	primary organic aerosol
POC	particulate organic carbon
polarimeter	a laboratory instrument used to determine the angle of optical rotation
of plane-polarized light passing through a sample of material
POLDER	CNES satellite - POLarization and Directionality of the Earth's
Reflectance
polymorphism	an inherited genetic variation occurring in a population
POM	particulate organic matter
POP	persistent organic pollutant
Pp	particle density
PP	pulse pressure
ppb	parts per billion
July 2009
XXX
DRAFT-DO NOT CITE OR QUOTE

-------
PPFL
ppm
ppt
PRB
PRE
PRELC
PRIDE
PS
PSAS
PSO
pS04
PSS
PSU
PT
PTT
PTV
PVD
Pyr
Q
Q
QAI
QBQ
QEEG
Qext
r
R2
radiative forcing
RAIN
RAMS
RANTES
percent predicted lung function
parts per million
parts per trillion
policy-relevant background
AeroCom Experiment
Primary Rat Epithelial Lung Cells
NASA Puerto Rico Dust Experiment
public school
The French National Program on Air Pollution Health Effects
Public Service Company of Oklahoma (subsidiary of American Electric
Power)]
particulate sulfate
physiologic saline solution
Pennsylvania State University
prothrombin time
partial thomboplastin time
programmable temperature vaporization
peripheral vascular disease
pyrene
cardiac output
coronary flow of the heart
QA interval (simple systolic time interval)
backup quartz-fiber filter behind a quartz-fiber filter
quantitative electroencephalography
the extinction coefficient (a function of particle size distribution and
refractive index)
correlation coefficient
coefficient of determination
a way to compare different causes of perturbations in a climate system,
radiative forcing measures change in net irradiance (W/m2) at the
tropopauseJ atmospheric aerosols and particles can scatter and bounce
incoming solar radiation back into space, resulting in negative radiative
forcing (or absorb solar and infrared radiation, resulting in positive
radiative forcing) on the climate system.
Regional Aerosol Intensive Network, established by MANE-VU
real-time total ambient mass sampler
regulated upon activation, normal T cell expressed and secreted (a
chemotactic cytokine)
RAPS/RAMS
Regional Air Pollution Study/ Regional Air Monitoring Study
July 2009
xxxi
DRAFT-DO NOT CITE OR QUOTE

-------
RAK
RASMC
RAW 264.7
Rayleigh scattering
RBC
RD
re
REALM
RF
restinga
RFL
RH
RHMYE
rho(O)
RHR
RLF
RMC
RMEr
RMSSD
RMY
RNA
RNS
RO
ROCK
ROFA
ROFA-L
ROI
ROS
rapidly activating receptor(s)
rat aortic smooth muscle cells
mouse macrophage cell line
describes the elastic scattering of light by particles much smaller than
the wavelength of the light. Rayleigh scattering of sunlight in a clear
atmosphere is the main reason why the sky appears to be blue.
red blood cell
respiratory disease
cloud drop effective radius re
Regional East Atmospheric Lidar Mesonet (REALM) in North America
radiative forcing(s) measured in Watts per square meter (W/m2) ¦ the
change in net irradiance at the tropopause. A positive forcing (more
incoming energy) tends to warm the system, while a negative forcing
(more outgoing energy) tends to cool it. Affected by variations in the
amount of radiatively active gases and aerosols present.
particle effective radius
a distinct type of ecoregion ¦ coastal tropical and subtropical moist
broadleaf forest found in Brazil. Restingas form on sandy, acidic, and
nutrient-poor soils, and are characterized by medium sized trees and
shrubs adapted to the drier and nutrient-poor conditions.
Fetal Lung Fibroblasts
relative humidity
rat heart micro-vessel endothelial cell
rho(O) cells (cell lacking mitochondrial DNA)
Regional Haze Rule
rat lung fibroblasts
rat cardiomyocyte(s)
rapeseed oil methyl ester
root mean squared differences of successive normal-beat to normal-beat
(NN or RR) time intervals between each QRS complex in the EKG, also
referred to as^ r-MSSD and rMSSD
respiratory minute volume
ribonucleic acid
reactive nitrogen species
residual oil
rho associated kinase
residual oil fly ash (particles)
residual oil fly ash leachate
reactive oxygen intermediates
reactive oxygen species
July 2009
xxxii
DRAFT-DO NOT CITE OR QUOTE

-------
RPO
RR
RS
RSY
RTI
RTM
RTP
RY
RYCFB
RVCM
o
lo
S
SAB
SAFARI
SAGE
SALIA
SAM
SAMUM
SAP2.3
Sb
SB
SBL
SBP
Sc
sc
SCAB
SCAR-A
SCAR-B
SCARPOL
Regional Planning Organizations
risk ratio, relative risk, normal-to-normal (NN or RR) time interval
between each QRS complex in the EKG, using the R-wave peak as the
reference point.
resuspended soil
respiratory syncytial virus
respiratory tract infection
Radiative Transfer Model
Research Triangle Park, North Carolina
right ventricular
right ventricular cardio fibroblasts
right ventricular cardiomyopathy, rat ventricular cardiomyocytes,
reduced volume culture medium
sigma, standard deviation
one sigma! one standard deviation
sigma-g! geometric standard deviation
second
sulfur
(EPA) Science Advisory Board
South African Fire-Atmosphere Research Initiative
Stratospheric Aerosol and Gas Experiment
German study on the Influence of Air Pollution on Lung Function,
Inflammation, and Aging
Stratospheric Aerosol Measurement
NASA Saharan Mineral Dust Experiment
(CCSP Final Report) Synthesis and Assessment Product 2.3,
Atmospheric Aerosol Properties and Climate Impacts (CCSP Final
Report)
antimony
strand breaks
stable boundary layer
systolic blood pressure
scandium
summer curbside particles
California South Coast Air Basin
NASA Smoke/Sulfates, Clouds and Radiation - America experiment
NASA Sulfates, Clouds and Radiation - Brazil
Swiss Study on Childhood Allergy and Respiratory Symptoms with
Respect to Air Pollution
July 2009
xxxiii
DRAFT-DO NOT CITE OR QUOTE

-------
sCD40L	soluble CD40 ligand
SCE	sister chromatid exchange
SCS	Harvard Six Cities Study
Sd	standard deviation
SD	Sprague-Dawley rat
SDANN5	standard deviation of the average of normal to normal (N-N) intervals
in all 5-min intervals in a 24-h period
SDNN	standard deviation normal-to-normal (NN or RR) time interval between
each QRS complex in the EKG
SDNN24HR	standard deviation of the average of all normal to normal intervals in a
24-h period
Se	selenium
se	standard error
SEARCH	Southeastern Aerosol Research and Characterization
sem	standard error of mean
SEM	scanning electron microscopy
SES	socioeconomic status, sample equilibration system
Sess	Session of U.S. Congress
SF-1	steroidogenic factor-1
SF-UFID	suspension, particle free ultrafine industrial exhaust
SGA	small for gestational age
sGC	soluble guanylate cyclase
SGP	Southern Great Plains
-SH	sulfhydryl group
SH	Mount Saint Helen's ash
SH, SHR	spontaneously hypertensive disease model rat
SHADE	Saharan Dust Experiment
SHEDS	Stochastic Human Exposure and Dose Simulation model
shortwave albedo	the fraction of shortwave (solar) radiation reflected from the earth back
into space
shortwave emissivity	a material's ability to emit or release the shortwave (solar) energy
which it has absorbed
Si	silicon
sICAM-1	soluble intercellular adhesion molecule
SIDS	sudden infant death syndrome
Si02	silicone dioxide
SIPS	State Implementation Plan
SJV	San Joaquin Valley
July 2009
xxxiv
DRAFT-DO NOT CITE OR QUOTE

-------
Sky glow
SLAMS
SME
SMOCC
SMOKE
SMPS
SMPS-APS
SMRA
SNP
SNS
S02
so3
so42"
SOA
SOC
SOD
SOPHIA
SOx
SP
SPA
SPD
SPEW
SPG
SPM
SPRINT ARS
SRM-154b
SRM1648
SRM-1649
SRM-1650
SRM-1879
A kind of light pollution, human-made electrical lighting contributes to
sky glow, visible by the "glowing" effect seen in the skies over many
cities and towns as a dome of light. Light can be emitted directly
upward or reflected from the ground and is scattered by dust and gas
molecules in the atmosphere, producing a luminous background. It has
the effect of reducing one's ability to view the stars.
State and Local Air Monitoring Stations
soybean oil methyl ester
European Commission's Smoke Aerosols, Clouds, Rainfall and Climate
(Aerosols from Biomass Burning Perturb Global and Regional Climate)
Spare-Matrix Operator Kernel Emissions system
scanning mobility particle sizer
scanning mobility particle sizer— aerodynamic particle sizer (in tandem)
small mesenteric rat arteries
single-nucleotide polymorphism, sodium nitroprusside
sympathetic nervous system
sulfur dioxide
sulfur trioxide
sulfate
secondary organic aerosol
semi-volatile organic compound
superoxide dismutase
Study of Particulates and Health in Atlanta
sulfur oxides, oxides of sulfur
surfactant protein (e.g., SPA, SPD)
surfactant protein A (present in lung surfactant)
surfactant protein D (present in lung surfactant)
Speciated Pollutant Emission Wizard
Southern Great Plains site established by DOE's ARM Program,
suspended particulate matter
Spectral Radiation-Transport Model for Aerosol Species
NIST standard reference material 154b; (Ti02 Titanium dioxide)
NIST standard reference material 1648! (urban particulate matter)
NIST standard reference material 1649 (Washington, D. C. urban air
particulate matter, urban dust)
NIST standard reference material 1650 (diesel exhaust particulate
matter)
NIST standard reference material 1859! (silicone dioxide, respirable
cristobalite [respirable crystalline silica])
July 2009
XXXV
DRAFT-DO NOT CITE OR QUOTE

-------
SRM-2975	NIST standard reference material 2975 (diesel exhaust particulate
matter)
s-ROFA	soluble portion of residual oil fly ash
SS	secondary sulfate, sea salt
SSA	single-scattering albedo (the ratio of scattering efficiency to total light
extinction, or a sum of scattering and absorption)
SSR	standardized sex ratio
SST	sea surface temperature
STEM	Sulfur / Sulfate Transport Eulerian Model
stemflow	during precipitation (rain, snow, etc.) when the plant, forest system,
ecosystem can hold no more, the water will drip from the plant
(throughfall) or run down the stem (stemflow) before reaching the
ground
STN	EPA Speciation Trend Network
Stokes velocity	the terminal velocity at which a sphere of a given density will sink (or
rise) in a medium of a given density.
STP	standard temperature and pressure
STZ	Streptozotocin
SUB	summer urban background particles
Sunglint	an area of a satellite image viewing the ocean surface, which appears
smooth and silver, due to the sun reflecting off the surface of the ocean
at the same angle as the satellite
SURFRAD	NOAA GMD Surface Radiation network
SVA	supraventricular arrhythmia
sVCAM-1	soluble vascular adhesion molecule 1
SVEB	supraventricular ectopic beats
SWNT	singlewailed nanotube
SXRF	Synchroton X-ray fluorescence
SZA	solar zenith angle
x	photochemical lifetime
T	body temperature
TAR	IPCC 3rd Assessment Report
TARC	thymus and activation-regulated chemokine
TARFOX	Tropospheric Aerosol Radiative Forcing Observational Experiment
TAT	thrombin-anti-thrombin complexes
TB	tracheobronchial region of the respiratory tract
TBA	thiobarbituric acid
TBAP	tetrakis(4-benzoic acid) porphyrin
TBARS	thiobarbituric acid reactive substances
July 2009
xxxvi	DRAFT - DO NOT CITE OR QUOTE

-------
TBQ	backup quartz-fiber filter behind a Teflon-membrane filter
99mTc	Technetium-99m
99mTc-DMTA	99MTc Dynamic mechanical thermal analysis
99MTc-DTPA	99MTc-diethylenetriaminepentaacetic acid
Tco	core temperature
TD	thermal desorption, tire debris extracted in methanol
TD-GC/MS	thermal desorption-gas chromatography/mass spectrometry
TEAC	Trolox Equivalent Antioxidant Capacity assay
TEOM	Tapered Element Oscillating Microbalance
Terra	multi-national, multi-disciplinary satellite mission involving
partnerships between NASA and the aerospace agencies of Canada and
Japan
TexAQS	Texas Air Quality Field Study
TF	tissue factor
TFPI	tissue factor pathway inhibitor
Tg	teragram (one trillion grams 1 x 1012 grams! one billion kilograms 1 x
109 kg! one million metric tons)
TG	terminal ganglion (neurons)
TGF	transforming growth factor
TGF 6	6 transforming growth factor
Th	thorium
Thl	T helper cell type 1
Th2	T helper cell type 2
tHcy	total homocysteine
throughfall	during precipitation (rain, snow, etc.) when the plant, forest system,
ecosystem can hold no more, the water will drip from the plant
(throughfall) or run down the stem (stemflow) before reaching the
ground
titanium
transient ischemic attack
iron-loaded fine titanium oxide
tissue inhibitor of MMP
titanium dioxide
thymidine kinase
transition metals
Thematic Mapper, a sensor on Landsat5 satellite, for mapping tempo-
spatial dynamics
tetramethylthiourea
tumor necrosis factor alpha
Ti
TIA
TiFe
TIMP-2
Ti02
TK
TM
TM5
TMTU
TNF-a
July 2009
xxxvii
DRAFT-DO NOT CITE OR QUOTE

-------
TOA	top of the atmosphere
TOF-SIMS	time-of-flight ¦ secondary ion mass spectrometry
TOMS	Total Ozone Mapping Spectrometer
TOT/GC	thermal optical transmission analyzer coupled with gas
chromatography
TOYS	TIROS-N Operational Vertical Sounder
toxaphene	an insecticide used before 1982, not used in U.S. after 1990
tPA, t-PA	tissue plasminogen activator
TRACE	Transition Region and Coronal Explorer (satellite)
TRACE-A	Transport and Chemical Evolution Over the Atlantic
TRACE-P	Transport and Chemical Evolution Over the Pacific mission model
Transmissometer an instrument for approximating the visual range, by measuring the
extinction coefficient of the atmosphere at the middle of the visible
waveband (550 nm). Also called hazemeter, or transmittance meter.
TRP	transient receptor potential family of ion channels
TRPV1	transient receptor potential vanilloid-1 receptor
TR-XRF	total reflection X-ray fluorescence
TSA	trichostatin A
TSP	total suspended particulate
TSP-10	total suspended particulates up to 10 jj,m
TSS	WRAP Technical Support System website
TYOC	total YOC
Twomey effect	first "indirect' climate effect! an increase in cloud brightness (describes
how aerosols and cloud condensation nuclei from anthropogenic
pollution may increase the amount of solar radiation reflected by
clouds, changing cloud brightness)
TWP	Tropical West Pacific island
TXB2	thromboxane B-2
U	uranium
UACR	urinary albumin / creatinine ratio
UAE2	United Arab Emirates Unified Aerosol Experiment
UAP	urban ambient particle
UF	ultrafine, uncertainty factor
UFAA	ultrafine ambient air
UFC	ultrafine carbon
UfCB	ultrafine carbon black
UFDG	ultrafine diesel engine exhaust
UFID	ultrafine industrial exhaust
July 2009
xxxviii
DRAFT-DO NOT CITE OR QUOTE

-------
UFP
UFPM
UFTi02
UIO
U.K.
UKMO
ULAQ
ULTRA
UMI
UNEP
UP
UPM
UPSP
upwelling
URI
URS
U.S.
USC, U.S.C.
UY
Y
V, mV, iiV
VAQ
YCAM-1
Vd
YEAPS
VEGF
VIEWS
VISTAS
VOC
VOSO4
VPB
VR
ultrafine particle
ultrafine particulate matter
ultrafine titanium dioxide
University of Oslo
United Kingdom
United Kingdom Meteorological Office
University of lL'Aquila.
Exposure and Risk Assessment for Fine and Ultrafine Particles in
Ambient Air (study)
University of Michigan
United Nations Environmental Programme
urban particle
ultrafine particulate matter
unmodified polystyrene particle(s)
when ocean winds cause surface water to move away from a coastline
the deeper, colder water will move upward to the surface, creating a
upwelling current; which brings replenishing nutritional components to
the surface
upper respiratory infection
upper respiratory symptoms
United States of America
U.S. Code
ultraviolet radiation
vanadium
volt, millivolt, microvolt
visual air quality, used here to refer to the visibility effects caused
solely by air quality conditions, so for example it excludes the reduced
visibility caused by fog
vascular adhesion molecule 1
deposition velocity
Vitamin E Atherosclerosis Progression Study
vascular endothelial growth factor
Visibility Information Exchange Web Site, ambient monitoring data
system,
Visibility Improvement State and Tribal Association of the Southeast
volatile organic compound
vanadyl sulfate
ventricular premature beat
visual range
July 2009
xxxix	DRAFT - DO NOT CITE OR QUOTE

-------
VK1
YSCC
YSMC
Yt
vWF
W
WACAP
WBC
WC
WHI
WHI OS
WinHaze
wk
WKY
W/m2, W m"2
WMO
Wnt
WRAP
WRF
WS
WSOC
WUB
XAD
XPS
Y
yr
Z
Zn
ZnO
ZnS
ZnS04
Zr
vanilloid receptor 1
very sharp cut cyclone
Vascular Smooth Muscle Cells
tidal volume
von Willebrand factor
Wilderness
Western Airborne Contaminates Assessment Project
white blood cell(s)
winter curbside particles
Women's Health Initiative
Women's Health Initiative Observational Study
imaging software from ARS, simulates visual air quality differences of
various scenes
week(s)
Wistar-Kyoto rat strain
watts per square meter
World Meteorological Organization
wingless gene family, encoding oncogenesis signaling pathways (e.g.,
Wnt-4, Wnt-7a [MMTV integration site family, member 4 or 7a])
Western Regional Air Partnership
Weather Research and Forecasting model
wood smoke
water soluble organic carbon
winter urban background particles
polystyrene-divinyl benzene
X-ray photoelectron spectroscopy
yttrium
year
radar reflectivity (measured in dBZ [decibels of Z, where Z represents
the energy reflected back to the radar.])
zinc
zinc oxide
zinc sulfide
zinc sulfate
zirconium
July 2009
xl
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Chapter 1. Introduction
The second external review draft Integrated Science Assessment (ISA) is a review,
synthesis, and evaluation of the most policy-relevant evidence, and communicates critical
science judgments relevant to the National Ambient Air Quality Standards (NAAQS)
review. As such, the ISA forms the scientific foundation for the review of the primary
(health-based) and secondary (welfare-based) NAAQS for particulate matter (PM). The ISA
accurately reflects "the latest scientific knowledge useful in indicating the kind and extent
of identifiable effects on public health which may be expected from the presence of [a]
pollutant in ambient air" (42 U.S.C. 7408). Key information and judgments formerly
contained in an Air Quality Criteria Document (AQCD) for PM are incorporated in this
assessment. Additional details of the pertinent literature published since the last review, as
well as selected older studies of particular interest, are included in a series of annexes. This
ISA thus serves to update and revise the evaluation of the scientific evidence available at
the time of the previous review of the NAAQS for PM that was concluded in 2006.
The Integrated Review Plan for the National Ambient Air Quality Standards for
Particulate Matter identifies a series of policy-relevant questions that provide a framework
for this assessment of the scientific evidence (U.S. EPA, 2008, 157072). These questions
frame the entire review of the NAAQS for PM, and thus are informed by both science and
policy considerations. The ISA organizes and presents the scientific evidence such that,
when considered along with findings from risk analyses and policy considerations, will help
the EPA address these questions during the NAAQS review for PM. In evaluating the
health evidence, the focus of this assessment will be on scientific evidence that is most
relevant to the following questions that have been taken directly from the Integrated
Review Plan:
¦	Has new information altered the body of scientific support for the occurrence of
health effects following short- and/or long-term exposure to levels of fine and
thoracic coarse particles found in the ambient air?
¦	Has new information altered conclusions from previous reviews regarding the
plausibility of adverse health effects associated with exposures to PM2.5, PM10, PM10-
2.5, or alternative PM indicators that might be considered?
¦	What evidence is available from recent studies focused on specific size fractions,
chemical components, sources, or environments (e.g., urban and non-urban areas) of
Note- Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http V/ep a. gov/her o. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009
1-1
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
¦	To what extent is key scientific evidence becoming available to improve our
understanding of the health effects associated with various time periods of PM
exposures, including not only short-term (daily or multi-day) and chronic (months to
years) exposures, but also peak PM exposures (less than 24 hours)? To what extent
is critical research becoming available that could improve our understanding of the
relationship between various health endpoints and different lag periods (e.g., less
than one day, single day, multi-day distributed lags)?
¦	What data are available to improve our understanding of spatial and/or temporal
heterogeneity of PM exposures considering different size fractions and/or
components?
¦	At what levels of PM exposure do health effects of concern occur? Is there evidence
for the occurrence of adverse health effects at levels of PM lower than those observed
previously? If so, at what levels and what are the important uncertainties associated
with that evidence? What is the nature of the dose-response relationships of PM for
the various health effects evaluated?
¦	What evidence is available linking particle number concentration with adverse
health effects of ultrafine particles?
¦	Do risk/exposure estimates suggest that exposures of concern for PM-induced health
effects will occur with current ambient levels of PM or with levels that just meet the
current standards? If so, are these risks/exposures of sufficient magnitude such that
the health effects might reasonably be judged to be important from a public health
perspective? What are the important uncertainties associated with these
risk/exposure estimates?
¦	To what extent is key evidence becoming available that could inform our
understanding of subpopulations that are particularly sensitive or vulnerable to PM
exposures? In the last review, sensitive or vulnerable subpopulations that appeared
to be at greater risk for PM-related effects included individuals with pre-existing
heart and lung diseases, older adults, and children. Has new evidence become
available to suggest additional sensitive subpopulations should be given increased
focus in this review (e.g., fetuses, neonates, genetically susceptible subpopulations)?
¦	To what extent is key evidence becoming available to inform our understanding of
populations that are particularly vulnerable to PM exposures? Specifically, is there
July 2009
1-2
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
¦	To what extent have important uncertainties identified in the last review been
reduced and/or have new uncertainties emerged?
¦	To what extent is new information available to inform our understanding of non-PM-
exposure factors that might influence the associations between PM levels and health
effects being considered (e.g., weather-related factors! behavioral factors such as
heating/air conditioning use! driving patterns! and time-activity patterns)?
In evaluating evidence on welfare effects of PM, the focus will be on evidence that can
help inform these questions from the Integrated Review Plan:
¦	What new evidence is available on the relationship between PM mass/size fraction
and/or specific PM components and visibility impairment and climate-related and
other welfare effects?
¦	To what extent has key scientific evidence now become available to improve our
understanding of the nature and magnitude of visibility, climate, and ecosystem
responses to PM and the variability associated with those responses (including
ecosystem type, climatic conditions, environmental effects and interactions with
other environmental factors and pollutants)?
¦	Do the evidence, the air quality assessment, and the risk/exposure assessment
provide support for considering alternative averaging times?
¦	At what levels of ambient PM do visibility impairment and/or environmental effects
of concern occur? Is there evidence for the occurrence of adverse visibility and other
welfare-related effects at levels of PM lower than those observed previously? If so, at
what levels and what are the important uncertainties associated with the evidence?
¦	Do the analyses suggest that PM-induced visibility impairment and/or other
welfare-effects will occur with current ambient levels of PM or with levels that just
meet the current standards? If so, are these effects of sufficient magnitude and/or
frequency such that these effects might reasonably be judged to be important from a
public welfare perspective? What are the uncertainties associated with these
estimates?
¦	What new evidence and/or techniques are available to quantify the benefits of
improved visibility and/or other welfare-related effects?
July 2009
1-3
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
¦ To what extent have important uncertainties identified in the last review been
reduced and/or have new uncertainties emerged?
1.1. Legislative Requirements
Two sections of the Clean Air Act (CAA, the Act) govern the establishment and
revision of the NAAQS. Section 108 of the Act (42 U.S.C. 7408) directs the Administrator to
identify and list "air pollutants" that "in his judgment, may reasonably be anticipated to
endanger public health and welfare" and whose "presence... in the ambient air results from
numerous or diverse mobile or stationary sources" and to issue air quality criteria for those
that are listed (42 U.S.C. 7408). Air quality criteria are intended to "accurately reflect the
latest scientific knowledge useful in indicating the kind and extent of identifiable effects on
public health or welfare which may be expected from the presence of [a] pollutant in
ambient air..." 42 U.S.C. 7408(b).
Section 109 of the Act (42 U.S.C. 7409) directs the Administrator to propose and
promulgate "primary" and "secondary" NAAQS for pollutants listed under Section 108. 42
U.S.C. 7409(a). Section 109(b)(1) defines a primary standard as one "the attainment and
maintenance of which in the judgment of the Administrator, based on such criteria and
allowing an adequate margin of safety, are requisite to protect the public health."1 42 U.S.C.
7409(b)(1). A secondary standard, as defined in Section 109(b)(2), must "specify a level of air
quality the attainment and maintenance of which, in the judgment of the Administrator,
based on such criteria, is required to protect the public welfare from any known or
anticipated adverse effects associated with the presence of [the] pollutant in the ambient
air."2 42 U.S.C. 7409(b)(2).
The requirement that primary standards include an adequate margin of safety was
intended to address uncertainties associated with inconclusive scientific and technical
information available at the time of standard setting. It was also intended to provide a
reasonable degree of protection against hazards that research has not yet identified. See
Lead Industries Association v. EPA, 647 F.2d 1130, 1154 (D.C. Cir 1980), cert, denied, 449
U.S. 1042 (1980); American Petroleum Institute v. Costle, 665 F.2d 1176, 1186 (D.C. Cir.
1981), cert, denied, 455 U.S. 1034 (1982); American Farm Bureau Federation v. EPA, 559 F.
3d 512, 533 (D.C. Cir. 2009). Both kinds of uncertainties are components of the risk
associated with pollution at levels below those at which human health effects can be said to
occur with reasonable scientific certainty. Thus, in selecting primary standards that include
1	The legislative history of section 109 indicates that a primary standard is to be set at "the maximum permissible ambient air
level... which will protect the health of any [sensitive] group of the population," and that for this purpose "reference should be
made to a representative sample of persons comprising the sensitive group rather than to a single person in such a group"
[S. Rep. No. 91-1196, 91st Cong., 2d Sess. 10 (1970)].
2	Welfare effects as defined in Section 302(h) [42 U.S.C. 7602(h)] include, but are not limited to, "effects on soils, water, crops,
vegetation, man-made materials, animals, wildlife, weather, visibility and climate, damage to and deterioration of property,
and hazards to transportation, as well as effects on economic values and on personal comfort and well-being."
July 2009
1-4
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
an adequate margin of safety, the Administrator is seeking not only to prevent pollution
levels that have been demonstrated to be harmful but also to prevent lower pollutant levels
that may pose an unacceptable risk of harm, even if the risk is not precisely identified as to
nature or degree.
In selecting a margin of safety, the EPA considers such factors as the nature and
severity of the health effects involved, the size of the sensitive population(s) at risk, and the
kind and degree of the uncertainties that must be addressed. The selection of any particular
approach to providing an adequate margin of safety is a policy choice left specifically to the
Administrator's judgment. See Lead Industries Association v. EPA, supra, 647 F.2d 1161-62.
In setting standards that are "requisite" to protect public health and welfare, as
provided in Section 109(b), the Administrator's task is to establish standards that are
neither more nor less stringent than necessary. In so doing, EPA may not consider the costs
of implementing the standards. See generally Whitman v. American Trucking Associations,
531 U.S. 457, 465-472, 475-76 (2001).
Section 109(d)(1) requires that "not later than December 31, 1980, and at 5-yr
intervals thereafter, the Administrator shall complete a thorough review of the criteria
published under Section 108 and the national ambient air quality standards... and shall
make such revisions in such criteria and standards and promulgate such new standards as
maybe appropriate..." 42 U.S.C. 7409(d)(1). Section 109(d)(2) requires that an independent
scientific review... committee "shall complete a review of the criteria and the national
primary and secondary ambient air quality standards... and shall recommend to the
Administrator any new standards and revisions of existing criteria and standards as may
be appropriate..." 42 U.S.C. 7409(d)(2). Since the early 1980s, this independent review
function has been performed by the Clean Air Scientific Advisory Committee (CASAC).
1.2. History of Reviews of the NAAQS for PM
PM is the generic term for a broad class of chemically and physically diverse
substances that exist as discrete particles (liquid droplets or solids) over a wide range of
sizes. Particles originate from a variety of anthropogenic stationary and mobile sources as
well as from natural sources. Particles may be emitted directly or formed in the atmosphere
by transformations of gaseous emissions such as sulfur oxides (SOx), nitrogen oxides (NOx),
and volatile organic compounds (VOC). The chemical and physical properties of PM vary
greatly with time, region, meteorology, and source category, thus complicating the
assessment of health and welfare effects. Table 1-1 summarizes the NAAQS that have been
promulgated for PM to date. These reviews are briefly described below, and further details
are provided in the Integrated Review Plan (U.S. EPA, 2008, 157072).
July 2009
1-5
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Table 1-1. Summary of NAAQS promulgated for PM, 1971-2006.
Final Rule
Indicator
Avg Time
Level
Form
1971 (36 FR 8186)
TSP (Total
Suspended
Particulates)
24-h
260 /yglm3 (primary)
150 /yglm3 (secondary)
Not to be exceeded more than once per yr


Annual
75/yg/m3 (primary)
Annual geometric mean
1987 (52 FR 24634)
PMio
24-h
150 /yglm3
Not to be exceeded more than once per yr on average over a 3-yr period


Annual
50/yg/m3
Annual arithmetic mean, averaged over 3 yrs
1997 (62 FR 38652)
PM2.6
24-h
65/yg/m3
98th percentile, averaged over 3 yrs


Annual
15/yg/m3
Annual arithmetic mean, averaged over 3 yrs1

PMio
24-h
150/yg/m3
Initially promulgated 99th percentile, averaged over 3 yrs; when 1997 standards were
vacated in 1999, the form of 1987 standards remained in place (not to be exceeded more
than once per yr on average over a 3-yr period)


Annual
50/yg/m3
Annual arithmetic mean, averaged over 3 yrs
2006(71 FR 61144)
PM2.6
24-h
35/yg/m3
98th percentile, averaged over 3 yrs


Annual
15/yg/m3
Annual arithmetic mean, averaged over 3 yrs2

PMio
24-h
150 /yglm3
Not to be exceeded more than once per yr on average over a 3-yr period
Note: When not specified, primary and secondary standards are identical.
EPAfirst established NAAQS for PM in 1971 (36 FR 8186, April 30, 1971), based on
the original criteria document (NAPCA, 1969, 014684). The reference method specified for
determining attainment of the original standards was the high-volume sampler, which
collects PM up to a nominal size of 25 to 45 micrometers (lim) (referred to as total
suspended particulates or TSP). The primary standards (measured by the indicator TSP)
were 260 |ug/m3 (ug/m3), 24-h avg, not to be exceeded more than once per year, and 75 ug/m3,
annual geometric mean. The secondary standard was 150 |ug/m3, 24-h avg, not to be
exceeded more than once per year. In October 1979 (44 FR 56730, October 2, 1979), EPA
announced the first periodic review of the air quality criteria and NAAQS for PM, and
significant revisions to the original standards were promulgated in 1987 (52 FR 24634, July
1, 1987). In that decision, EPA changed the indicator for particles from TSP to PMio, the
latter including particles with a mean aerodynamic diameter3 less than or equal to 10 um,
which delineated that subset of inhalable particles small enough to penetrate to the
thoracic region (including the tracheobronchial and alveolar regions) of the respiratory tract
1 The level of the 1997 annual PM2.5 standard was to be compared to measurements made at the community-oriented
monitoring site recording the highest level, or, if specific constraints were met, measurements from multiple
community-oriented monitoring sites could be averaged ("spatial averaging"). This approach was judged to be consistent
with the short-term epidemiologic studies on which the annual PM2.5 standard was primarily based, in which air quality data
were generally averaged across multiple monitors in an area or were taken from a single monitor that was selected to
represent community-wide exposures, not localized "hot spots" (62 FR 38672). These criteria and constraints were intended
to ensure that spatial averaging would not result in inequities in the level of protection afforded by the PM2.5 standards.
Community-oriented monitoring sites were specified to be consistent with the intent that a spatially averaged annual
standard provide protection for persons living in smaller communities, as well as those in larger population centers.
2In the revisions to the PM NAAQS finalized in 2006, EPA tightened the constraints on the spatial averaging criteria by
further limiting the conditions under which some areas may average measurements from multiple community-oriented
monitors to determine compliance (see 71 FR 61165-61167, October 17, 2006).
3 The more precise term is 50% cut point or 50% diameter (dso). This is the aerodynamic particle diameter for which the
efficiency of particle collection is 50%. Larger particles are not excluded altogether, but are collected with substantially
decreasing efficiency and smaller particles are collected with increasing (up to 100%) efficiency.
July 2009
1-6
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
(referred to as thoracic particles). EPA also revised the level and form of the primary
standards by (l) replacing the 24-h TSP standard with a 24-h PMio standard of 150 |ug/m3
with no more than one expected exceedence per year! and (2) replacing the annual TSP
standard with a PMio standard of 50 ug/m3, annual arithmetic mean, averaged over three
years.
The secondary standard was revised by replacing it with 24-h and annual standards
identical in all respects to the primary standards. The revisions also included a new
reference method for the measurement of PMio in the ambient air and rules for determining
attainment of the new standards. On judicial review, the revised standards were upheld in
all respects. Natural Resources Defense Council v. Administrator, 902 F. 2d 962 (D.C. Cir.
1990), cert, denied, 498 U.S. 1082 (1991).
In April 1994, EPA announced its plans for the second periodic review of the air
quality criteria and NAAQS for PM, and promulgated significant revisions to the NAAQS in
1997 (62 FR 38652, July 18, 1997). In that decision, EPA revised the PM NAAQS in several
respects. Most significantly, EPA determined that the fine and coarse1 fractions of PMio
should be considered separately. The Administrator's decision to modify the standards was
based on evidence that serious health effects were associated with short- and long-term
exposure to fine particles in areas that met the existing PMio standards. EPA accordingly
added new standards, using PM2.5 as the indicator for fine particles (with PM2.5 referring to
particles with a nominal mean aerodynamic diameter less than or equal to 2.5 urn), and
PMio as the indicator for thoracic coarse particles or coarse-fraction particles (generally
including particles with a nominal mean aerodynamic diameter greater than 2.5 urn and
less than or equal to 10 |um, or PM10-2.5). The EPA established two new PM2.5 standards: an
annual standard of 15 |ug/m3, based on the 3-yr avg of annual arithmetic mean PM2.5
concentrations from single or multiple community-oriented monitors! and a 24-h standard
of 65 |ug/m3, based on the 3-yr avg of the 98th percentile of 24-h PM2.5 concentrations at
each population-oriented monitor within an area. Also, EPA established a new reference
method for measuring PM2.5 in the ambient air and adopted protocols for determining
attainment of the new standards. To continue to address thoracic coarse particles, EPA
retained the annual PMio standard, while revising the form, but not the level, of the 24-h
PMio standard to be based on the 99th percentile of 24-h PMio concentrations at each
monitor in an area. The EPA revised the secondary standards by making them identical in
all respects to the primary standards.
Following promulgation of the 1997 PM NAAQS, petitions for review were filed by a
large number of parties, addressing a broad range of issues. In May 1999, a three-judge
panel of the U.S. Court of Appeals for the District of Columbia Circuit issued an initial
decision that upheld EPA's decision to establish fine particle standards, holding that "the
growing empirical evidence demonstrating a relationship between fine particle pollution
1 See definitions of "fine" and "coarse" particles in Section 3.2.
July 2009
1-7
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
and adverse health effects amply justifies establishment of new fine particle standards."
American Trucking Associations v. EPA (175 F. 3d 1027, 1055-56 (D.C. Cir. 1999); rehearing
granted in part and denied in part, 195 F. 3d 4 (D.C. Cir. 1999), affirmed in part and
reversed in part, Whitman v. American Trucking Associations 5,°) ] U.S. 457 (2001). The
panel also found "ample support" for EPA's decision to regulate coarse particle pollution, but
vacated the 1997 PMio standards, concluding that EPA had not provided a reasonable
explanation justifying use of PMio as an indicator for coarse particles (175 F. 3d at 1054-55).
Pursuant to the court's decision, EPA removed the vacated 1997 PMio standards from the
Code of Federal Regulations. The pre-existing 1987 PMio standards remained in place (65
FR 80776, December 22, 2000). The Court also upheld EPA's determination not to establish
more stringent secondary standards for fine particles to address effects on visibility (175 F.
3d at 1027).
More generally, the panel held (over one judge's dissent) that EPA's approach to
establishing the level of the standards in 1997, both for the PM and ozone (O3) NAAQS
promulgated on the same day, effected "an unconstitutional delegation of legislative
authority" (id. at 1034-40). Although the panel stated that "the factors EPA uses in
determining the degree of public health concern associated with different levels of ozone
and PM are reasonable," it remanded the rule to EPA, stating that when EPA considers
these factors for potential non-threshold pollutants "what EPA lacks is any determinate
criterion for drawing lines" to determine where the standards should be set. Consistent
with EPA's long-standing interpretation and D.C. Circuit precedent, the panel also
reaffirmed its prior holdings that in setting NAAQS EPA is "not permitted to consider the
cost of implementing those standards" (id. at 1040-41).
On EPA's petition for rehearing, the panel adhered to its position on these points.
American Trucking Associations v. EPA, 195 F. 3d 4 (D.C. Cir. 1999). The full Court of
Appeals denied EPA's suggestion for rehearing en banc, with five judges dissenting (id. at
13).
Both sides filed cross appeals on these issues to the U.S. Supreme Court, and the
Court granted certiorari. In February 2001, the Supreme Court issued a unanimous
decision upholding EPA's position on both the constitutional and cost issues. Whitman v.
American Trucking Associations, 531 U.S. 457, 464, 475-76. On the constitutional issue, the
Court held that the statutory requirement that NAAQS be "requisite" to protect public
health with an adequate margin of safety sufficiently guided EPA's discretion, affirming
EPA's approach of setting standards that are neither more nor less stringent than
necessary. The Supreme Court remanded the case to the Court of Appeals for resolution of
any remaining issues that had not been addressed in that court's earlier rulings. Id. at
475-76. In March 2002, the Court of Appeals rejected all remaining challenges to the
standards, holding under the traditional standard of judicial review that PM2.5 standards
July 2009
1-8
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
were reasonably supported by the administrative record and were not "arbitrary and
capricious" American Trucking Associations v. EPA, 283 F. 3d 355, 369-72 (D.C. Cir. 2002).
In October 1997, EPA published its plans for the third periodic review of the air
quality criteria and NAAQS for PM (62 FR 55201). After CASAC and public review, EPA
finalized the 2004 PM AQCD (U.S. EPA, 2004, 056905) and 2005 Staff Paper (U.S. EPA,
2005, 090209). For the primary fine particle standards, most CASAC PM Panel members
favored the option of revising the level of the 24-h PM2.5 standard in the range of 35 to 30
ug/ma with a 98th percentile form, in concert with revising the level of the annual PM2.5
standard in the range of 14 to 13 jug/m3 (Henderson, 2005, 188316). Most of the members of
the CASAC PM Panel also strongly supported establishing a new, secondary PM2.5 standard
to protect urban visibility and recommended establishing a sub-daily (4- to 8-h averaging
time) PM2.5 standard within the range of 20 to 30 pg/m3 with a form within the range of the
92nd to 98th percentile (Henderson, 2005, 188316). For thoracic coarse particles, there was
general concurrence among CASAC PM Panel members to revise the PM10 standards by
establishing a primary standard specifically targeted to address particles in the size range
of 2.5 to 10 pm. The CASAC PM Panel was also in general agreement "that coarse particles
in urban or industrial areas are likely to be enriched by anthropogenic pollutants that tend
to be inherently more toxic than the windblown crustal material which typically dominates
coarse particle mass in arid rural areas." Based on its review of the Staff Paper, there was
general agreement among the CASAC PM Panel members that a 24-h PM10-2.5 standard
with a level in the range of 50 to 70 |ug/m3, with a 98th percentile form, was reasonably
justified and that a PM10-2.5 standard with an annual averaging time was not warranted
(Henderson, 2005, 156537). On January 17, 2006, EPA proposed to revise the NAAQS for
PM (71 FR 2620). For fine particles, EPA proposed to retain PM2.5 as the indicator, to retain
standards for 24-h and annual exposures, and to revise the form of the annual standard to
tighten conditions for demonstrating compliance using spatially averaged monitoring. EPA
also proposed to revise the level of the 24-h PM2.5 standard to 35 jug/m3to provide increased
protection against health effects associated with short-term PM2.5 exposures, including
premature mortality and increased hospital admission and emergency room visits, but
proposed to retain the level of the annual PM2.5 standard at 15 ug/m3, continuing protection
against health effects associated with long-term exposure including premature mortality
and development of chronic respiratory disease. With regard to the primary standards for
thoracic coarse particles, EPA proposed to revise the 24-h PM10 standard in part by
establishing a new indicator for thoracic coarse particles (particles generally between 2.5
and 10 um in PM10-2.5 diameter), qualified so as to include any ambient mix of PM10-2.5 that
was dominated by resuspended dust from high density traffic on paved roads and PM
generated by industrial sources and construction sources, and proposed to exclude any
ambient mix of PM10-2.5 that was dominated by rural windblown dust and soils and PM10-2.5
generated by agricultural and mining sources. EPA also proposed a detailed monitoring
July 2009
1-9
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
regime in conjunction with this proposed indicator. 71 FR 2710, 2731-42. The EPA proposed
to set a 24-h standard (using the proposed indicator) at a level of 70 ug/m3 to continue to
provide a level of protection against health effects associated with short-term exposure
(including hospital admissions for cardiopulmonary diseases, increased respiratory
symptoms and possibly premature mortality) in those areas where the proposed indicator
was found, generally equivalent to the level of protection provided by the existing 24-h PMio
standard. Also, EPA proposed to revoke, upon finalization of a primary 24-h standard for
thoracic coarse particles, the 24-h PMio standard as well as the annual PMio standard.
EPA proposed to revise the secondary standards by making them identical to the suite
of proposed primary standards for fine and coarse particles, providing protection against
PM-related public welfare effects including visibility impairment, effects on vegetation and
ecosystems, and materials damage and soiling. EPA also solicited comment on adding a new
sub-daily PM2.5 secondary standard to address visibility impairment in urban areas.
CASAC provided additional advice to EPA in a letter to the Administrator requesting
reconsideration of CASAC's recommendations for both the primary and secondary PM2.5
standards as well as standards for thoracic coarse particles (Henderson, 2006, 156538).
On September 21, 2006, EPA announced its final decisions to revise the primary and
secondary NAAQS for PM to provide increased protection of public health and welfare,
respectively (71 FR 61144). With regard to the primary and secondary standards for fine
particles, EPA revised the level of the 24-h PM2.5 standard to 35 ug/m3, retained the level of
the annual PM2.5 standard at 15 ug/m3, and revised the form of the annual PM2.5 standard
by narrowing the constraints on the optional use of spatial averaging. EPA established the
secondary standard for fine particles identical to the primary standards. With regard to the
primary and secondary standards for thoracic coarse particles, EPA retained PMio as the
indicator for coarse particles, retained the level and form of the 24-h PMio standard (so the
standard remains 150 ug/m3 with a one expected exceedence form) and revoked the annual
standard because available evidence generally did not support a link between long-term
exposure to current ambient levels of coarse particles and health or welfare effects.
Following promulgation of the revised PM NAAQS in 2006, several parties filed
petitions for review with respect to: (l) selecting the level of the annual primary PM2.5
standard! (2) setting the secondary PM2.5 standards identical to the primary standards! (3)
retaining PMio as the indicator for coarse particles and retaining the level and form of the
PMio 24-h standard, and (4) revoking the PMio annual standard. On judicial review, the
D.C. Circuit remanded the annual standard for fine particles to EPA because EPA failed to
adequately explain why the annual PM2.5 standard provided the requisite protection from
both short- and long-term exposures to fine particles including protection for vulnerable
subpopulations. With respect to protection from short-term exposures, in 1997 EPA
determined that the annual standard was the generally controlling standard for lowering
both short- and long-term PM2.5 concentrations and the 24-h standard was set to "provide
July 2009
MO
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
an adequate margin of safety against infrequent or isolated peak concentrations that could
occur in areas that attain the annual standard" (62 FR 38676-77, July 18, 1997). In the
2006 decision, the Administrator considered it appropriate to use a somewhat different
evidence-based approach from that used in 1997 to set the level of the 24-h and annual
PM2.5 standards. In that decision, the Administrator relied upon evidence from the short-
term exposure PM2.5 studies as the principal basis for selecting the proposed level of the 24-
h standard and relied upon evidence from the long-term exposure PM2.5 studies as the
principal basis for selecting the level of the annual standard. The court found EPA failed to
adequately explain this change in approach in light of CASAC and staff's recommendations
to do otherwise. The court also found that EPA had failed to adequately explain why a
short-term 24-h standard by itself would provide the protection needed from short-term
exposures. American Farm Bureau Federation v. EPA, 559 F. 3d 512, 520-24 (D.C. Cir.
2009). With respect to protection from long-term exposure, the court found that EPA failed
to adequately explain how the current standard provided "an adequate margin of safety for
vulnerable subpopulations, such as children, the elderly, or those with conditions that
exposure them to greater risk from from fine particles". Specifically, EPA did not provide a
reasonable explanation of why certain studies, including a study of children in Southern
California showing lung damage from long-term exposure, did not call for a more stringent
annual standard. Id. at 522-23.
The court also remanded the secondary standard for fine particles, based on EPA's
failure to adequately explain why setting the secondary NAAQS equivalent to the primary
standards provided the required protection for public welfare including protection from
visibility impairment. The court found that EPA failed to identify a target level of visibility
impairment that would be requisite to protect public welfare. This was contrary to the
statute and resulted in a lack of a reasoned basis for the final decision. In addition, EPA's
near exclusive reliance on a comparison of numbers of counties that would be in
nonattainment under various types of standards was an inadequate basis for making a
decision. It did not take into account the relative visibility protection of different standards,
as well as the failure of a 24-hstandard to address regional differences in humidity and its
effect on visibility. Id. at 528-31.
The court upheld EPA's decision to retain the 24-hPMio standard to provide protection
for coarse particle exposures and to revoke the annual PM10 standard. The court found that
EPA reasonably included all coarse PM within the standard, both urban and non-urban, to
provide nationwide protection for exposure to coarse PM. It rejected arguments that the
evidence showed there are no risks from exposure to non-urban coarse PM. Id. at 531-33.
The court further found that EPA had a reasonable basis to not set separate standards for
urban and non-urban coarse PM, namely the inability to reasonably define what ambient
mixes would be included under either "urban" or "non-urban." In addition, the court found
that record evidence supported EPA's cautious decision to provide "some protection from
July 2009
1-11
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
exposure to thoracic coarse particles... in all areas." The court also upheld EPA's decision to
use PMio as the indicator for coarse particles, and to retain the level of the standard at 150
jug/m3. EPA's final rule acknowledged that evidence of harm from urban-type coarse PM is
stronger than for other types, and targeted protection at areas where urban-type coarse PM
is most likely present. The targeting is done by using the indicator PMio for coarse
particles. PMio includes both coarse PM and fine PM. Urban and industrial areas tend to
have higher levels of fine PM than rural areas, so that in those areas less coarse PM is
allowed - the desired targeting. Conversely, fine PM levels tend to be lower in rural areas,
so more coarse particles are allowed in those areas - again the desired targeting. Likewise,
the court concluded that the EPA's choice of the level for the PMio standard was reasonable,
for many of the same reasons. Id. at 533-36. The court also upheld EPA's decision to revoke
the annual PMio standard. Id. at 537-38.
1.3. ISA Development
EPA initiated the current formal review of the NAAQS for PM on June 28, 2007 with
a call for information from the public (72 FR 35462). In addition to the call for information,
publications were identified through an ongoing literature search process that includes
extensive computer database mining on specific topics. Literature searches were conducted
routinely to identify studies published since the last review, focusing on publications from
2002 to May 2009. Search strategies were iteratively modified in an effort to optimize the
identification of pertinent publications. Additional papers were identified for inclusion in
several ways: review of pre-publication tables of contents for journals in which relevant
papers may be published; independent identification of relevant literature by expert
authors! and identification by the public and CASAC during the external review process.
Generally, only information that had undergone scientific peer review and had been
published or accepted for publication was considered. All relevant epidemiologic, controlled
human exposure, and animal toxicological studies, including those related to
exposure-response relationships, mode(s) of action (MOA), or susceptible or vulnerable
subpopulations, and welfare effects studies published since the last review were considered.
July 2009
1-12
DRAFT-DO NOT CITE OR QUOTE

-------
KEY DEFINITIONS
INFORMATIVE studies are well-designed,
properly implemented, thoroughly described.
HIGHLY INFORMATIVE studies reduce
uncertainty on critical issues, may include
analyses of confounding or effect modification
by copollutants or other variables, analyses of
concentration-response or dose-response
relationships, analyses related to time
between exposure and response, and offer
innovation in method or design.
POLICY-RELEVANT studies may include
those conducted at or near ambient concen-
trations and studies conducted in U.S. and
Canadian airsheds.
/ Studies are
evaluated for inclusion
in the ISA and/ >
\ or Annexes. /
Informative
studies
are identified.
Yes
Continuous,
comprehensive
literature review
of peer-reviewed
journal articles
Studies added
to the docket
during public
comment period.
Studies identified
during EPA
sponsored kickoff
meeting (including
studies in
preparation).
Studies that do
not address
exposure and/or
effects of air
pollutant(s) under
review are
excluded.
Selection of
studies
discussed and
additional studies
identified during
CASAC peer
review of draft
document.
All newly identified informative studies are included in the Annexes. Older, key
studies included in previous assessments may be included as well.
ANNEXES
Policy relevant and highly informative studies discussed in the ISA text include
those that provide a basis for or describe the association between the criteria
pollutant and effects. Studies summarized in tables and figures are included
because they are sufficiently comparable to be displayed together. A study
highlighted in the ISA text does not necessarily appear in a summary table or
figure.
ISA
Figure 1-1 Identification of studies for inclusion in the ISA.
July 2009
1-13
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
In general, in assessing the scientific quality and relevance of health and
environmental effects studies, the following considerations have been taken into account
when selecting studies for inclusion in the ISA or its annexes. The selection process for
studies included in this ISA is shown in Figure 1-1.
¦	Are the study populations, subjects, or animal models adequately selected and are
they sufficiently well defined to allow for meaningful comparisons between study or
exposure groups?
¦	Are the statistical analyses appropriate, properly performed, and properly
interpreted? Are likely covariates adequately controlled or taken into account in the
study design and statistical analysis?
¦	Are the PM aerometric data, exposure, or dose metrics of adequate quality and
sufficiently representative of information regarding ambient PM?
¦	Are the health or welfare effect measurements meaningful and reliable?
In selecting epidemiologic studies, EPA considered whether a given study contained
information on associations with short- or long-term PM exposures at or near ambient
levels of PM; evaluated health effects of PM size fractions, components or source-related
indicators; considered approaches to evaluate issues related to potential confounding and
modification of effects by other pollutants! and evaluated important methodologic issues
(e.g., lag or time period between exposure and effects, model specifications, thresholds,
mortality displacement) related to interpretation of the health evidence. Among the
epidemiologic studies selected, particular emphasis was placed on those studies most
relevant to the review of the NAAQS. Specifically, studies conducted in the United States
(U.S.) or Canada were discussed in more detail than those from other geographical regions.
Particular emphasis was placed on: (l) recent multicity studies that employ standardized
analysis methods for evaluating effects of PM and that provide overall estimates for effects
based on combined analyses of information pooled across multiple cities, (2) studies that
help understand quantitative relationships between exposure concentrations and effects,
(3) new studies that provide evidence on effects in susceptible or vulnerable populations,
and (4) studies that consider and report PM as a component of a complex mixture of air
pollutants.
Criteria for the selection of research evaluating controlled human exposure or animal
toxicological studies included a focus on studies conducted using relevant pollutant
exposures. For both types of studies, relevant pollutant exposures are considered to be
those generally within one or two orders of magnitude of ambient PM concentrations.
Studies in which higher doses were used may also be considered if they provide information
relevant to understanding MO As or mechanisms, as noted below.
July 2009
1-14
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Evaluation of controlled human exposure studies focused on those that approximated
expected human exposure conditions in terms of concentration and duration. In the
selection of controlled human exposure studies, emphasis is placed on studies that (l)
investigate potentially susceptible populations such as people with cardiovascular diseases
or asthmatics, particularly studies that compare responses in susceptible individuals with
those in age-matched healthy controls! (2) address issues such as concentration-response or
time-course of responses! (3) investigate exposure to PM separately and in combination
with other pollutants such as ().;! (4) include control exposures to filtered air! and (5) have
sufficient statistical power to assess findings.
For selecting toxicological studies for highlighting in the text, emphasis is placed on
inhalation studies conducted at concentrations <2 mg/m3 and those studies that
approximate expected human dose conditions in terms of concentration, size distributions,
and duration, which will depend on the toxicokinetics and biological sensitivity of the
particular laboratory animals examined. Studies that elucidated mechanisms of action
and/or susceptibility, particularly if the studies were conducted under atmospherically
relevant conditions, were emphasized whenever possible. A limited number of toxicological
studies were included that employed intratracheal instillation (IT) techniques, mainly for
PMio-2.5 studies in rodents, that explored new emerging areas of investigation (e.g.,
vasomotor function), or that evaluated specific potential MOAor mechanisms of response.
The sources, transport, and fate of fibers and unique nano-materials (viz., dots, hollow
spheres, rods, fibers, tubes) are not reviewed herein because the in vivo disposition of these
unique nanomaterials is not necessarily relevant to the behavior of ultrafine aerosols in the
urban environment that are created by combustion sources and photochemical formation of
secondary organic aerosols. In considering the potential effects of different components of
PM, EPA has focused on studies that have assessed effects for a range of PM sources or
components, including those using source apportionment methods or comparing effects for
numerous PM components, and not on studies of individual constituents or species.
These criteria provide benchmarks for evaluating various studies and for focusing on
the policy relevant studies in assessing the body of health and welfare effects evidence.
Detailed critical analysis of all PM health and welfare effects studies, especially in relation
to the above considerations, is beyond the scope of this document. Of most relevance for
evaluation of studies is whether they provide useful qualitative or quantitative information
on exposure-effect or exposure-response relationships for effects associated with current
ambient air concentrations of PM that can inform decisions on whether to retain or revise
the standards.
In developing the PM ISA, EPA began by reviewing and summarizing the evidence on
(l) atmospheric sciences and exposure! (2) the welfare effects of PM, including visibility,
climate, and ecological effects! and (3) the health effects evidence from in vivo and in vitro
animal toxicology, controlled human exposure, and epidemiology studies. In June 2008, EPA
July 2009
1-15
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
held a workshop to obtain review of the scientific content of initial draft materials or
sections for the draft ISA and its annexes, that primarily contain summary information.
The purpose of the initial peer review workshop was to ensure that the ISA is up-to-date
and focused on the most policy-relevant findings, and to assist EPA with integration of
evidence within and across disciplines. Following the peer review workshop, EPA addressed
comments from the peer review workshop and completed the initial integration and
synthesis of the evidence.
The integration of evidence on health or welfare effects involves collaboration between
scientists from various disciplines. As described in the section below, the ISA organization is
based on health or welfare effect categories. As an example, an evaluation of health effects
evidence would include summaries of findings from epidemiologic, controlled human
exposure, and toxicological studies, and integration of the results to draw conclusions based
on the causal framework described below. Using the causal framework described in
Section 1.5, EPA scientists consider aspects such as strength, consistency, coherence and
biological plausibility of the evidence, and develop draft judgments on the whether the
relationships are causal. The draft integrative synthesis sections and conclusions are
reviewed by EPA internal experts and, as appropriate, by outside expert authors. In
practice, causality determinations often entail an iterative process of review and evaluation
of the evidence. The draft ISA is released for review by the CASAC and the public.
Comments on the characterization of the science as well as the implementation of the
causal framework are carefully considered in revising and completing the ISA.
1.4. Document Organization
This ISA is composed of nine chapters. This introductory chapter presents background
information, and provides an overview of EPA's framework for making causal judgments.
Key findings and conclusions for consideration in the review of the NAAQS for PM from the
atmospheric sciences, ambient air data analyses, exposure assessment, dosimetry, health
and welfare effects, including judgments on causality for the health and welfare effects of
PM exposure, are presented in Chapter 2. More detailed summaries, evaluations and
integration of the evidence are included in Chapters 3 through 9.
Chapter 3 highlights key concepts or issues relevant to understanding the
atmospheric chemistry, sources, and exposure of and to PM following a "source-to-exposure"
paradigm. Chapter 4 summarizes key concepts and recent findings on the dosimetry of PM
and Chapter 5 discusses possible pathways and MOAfor the effects of PM. Chapters 6 and
7 evaluate and integrate epidemiologic, controlled human exposure, and animal
toxicological information relevant to the review of the primary NAAQS for PM. Health
effects related to short-term exposures (hours to days) to PM are the focus of Chapter 6.
Chapter 7 evaluates health evidence related to long-term exposures (months to years) to
July 2009
1-16
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
PM. Chapters 6 and 7 are organized by health outcome categories, such as cardiovascular
or respiratory effects, and each section includes effects of the various types of PM studied.
For each health outcome category, summary sections then integrate the findings to draw
conclusions on the evidence for the main size classes of PM (i.e., PM2.5, PM10-2.5, and
ultrafine particles). PM10 studies are included in this assessment where they provide
insights into relationships between PM and effects, but the ISA draws no conclusions
regarding causality between exposure to PM10 and effects! to the extent possible, the
findings of these studies are considered insofar as they provide information relevant to the
review of the standards for fine and thoracic coarse particles. Chapter 6 also includes a
summary and synthesis of the recent health evidence that uses apportionment methods to
assess health effects of sources and constituents of ambient PM; most such studies have
evaluated effects of short-term exposure. Chapter 8 evaluates evidence related to
populations potentially susceptible to PM-related effects.
Chapter 9 evaluates welfare effects evidence that is relevant to the review of the
secondary NAAQS for PM. This chapter includes consideration of effects of PM on visibility
impairment, materials damage, effects of PM on climate, and ecological effects of PM that
were not addressed in the Integrated Science Assessment for Oxides of Nitrogen and
Sulfur—Ecological Critera (NOxSOx ISA) (U.S. EPA, 2008, 157074). The chapter also
presents key conclusions and scientific judgments regarding causality for welfare effects of
PM. In 2008, EPA completed the NOxSOx ISA (U.S. EPA, 2008, 157074). that focused on
ecological effects related to the deposition of nitrogen (N)- and sulfur (S)-containing
compounds. The 2008 NOxSOx ISA included ecological effects from particle-phase
compounds (e.g., nitrates and sulfates), primarily effects from acidification and N-nutrient
enrichment and eutrophication. In the PM ISA, EPA focuses on recent data for direct
welfare effects of particle-phase NOx and SOx in the ambient air — primarily visibility
impairment, damage to materials, and positive and negative climate interactions — not the
welfare effects related to deposition of particle-phase NOx and SOx.
A series of annexes supplement this ISA. The annexes provide additional details of
the pertinent literature published since the last review, as well as selected older studies of
particular interest. These annexes contain information on:
¦	atmospheric chemistry of PM, sampling and analytic methods for measurement of
PM, concentrations, emissions, sources and human exposure to PM (Annex A)
¦	studies on the dosimetry of PM (Annex B)
¦	controlled human exposure studies of health effects related to exposure to PM
(Annex C)
¦	toxicological studies of health effects in laboratory animals (Annex D); and
July 2009
1-17
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
¦	epidemiologic studies of health effects from short- and long-term exposure to PM
(Annex E);
¦	studies that evaluate PM-induced health effects attributable to specific constituents
or sources.
Within the annexes, detailed information about methods and results of health studies
is summarized in tabular format, and generally includes information about: concentrations
of PM and averaging times! study methods employed! results! and quantitative results for
relationships between effects and exposure to PM. As noted in the section above, the most
pertinent results of this body of studies are brought into the ISA.
1.5. EPA Framework for Causal Determination
The EPA has developed a consistent and transparent basis to evaluate the causal
nature of air pollution-induced health or environmental effects. The framework described
below establishes uniform language concerning causality and brings more specificity to the
findings. It drew standardized language from across the federal government and wider
scientific community, especially from the recent National Academy of Sciences (NAS)
Institute of Medicine (IOM) document, Improving the Presumptive Disability
Decision-Making Process for Veterans (IOM, 2008, 156586). the most recent comprehensive
work on evaluating causality.
This introductory section focuses on the evaluation of health effects evidence! while
focusing on human health outcomes, the concepts are also generally relevant to causality
determination for welfare effects. This section:
¦	describes the kinds of scientific evidence used in establishing a general causal
relationship between exposure and health effects!
¦	defines cause, in contrast to statistical association!
¦	discusses the sources of evidence necessary to reach a conclusion about the existence
of a causal relationship!
¦	highlights the issue of multifactorial causation!
¦	identifies issues and approaches related to uncertainty! and
¦	provides a framework for classifying and characterizing the weight of evidence in
support of a general causal relationship.
Approaches to assessing the separate and combined lines of evidence (e.g.,
epidemiologic, controlled human exposure, and animal toxicological studies) have been
July 2009
1-18
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
formulated by a number of regulatory and science agencies, including the IOM of the NAS
(IOM, 2008, 156586), International Agency for Research on Cancer (IARC, 2006, 093206),
EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005, 086237), Centers for
Disease Control and Prevention (CDC, 2004, 056384), and National Acid Precipitation
Assessment Program (NAPAP) (NAPAP, 1991, 095894). These formalized approaches offer
guidance for assessing causality. The frameworks are similar in nature, although adapted
to different purposes, and have proven effective in providing a uniform structure and
language for causal determinations. Moreover, these frameworks have supported
decision-making under conditions of uncertainty.
1.5.1.	Scientific Evidence Used in Establishing Causality
Causality determinations are based on the evaluation and synthesis of evidence from
across scientific disciplines! the type of evidence that is most important for such
determinations will vary by assessment. The most direct evidence of a causal relationship
between pollutant exposures and human health effects comes from controlled human
exposure studies. This type of study experimentally evaluates the health effects of
administered exposures in human volunteers under highly-controlled laboratory conditions.
In most epidemiologic or observational studies of humans, the investigator does not
control exposures or intervene with the study population. Broadly, observational studies
can describe associations between exposures and effects. These studies fall into several
categories: cross-sectional, prospective cohort, and time-series studies. "Natural
experiments" offer the opportunity to investigate changes in health with a change in
exposure! these include comparisons of health effects before and after a change in
population exposures, such as the closure of a pollution source.
Experimental animal data complement the controlled human exposure and
observational data! these studies can help characterize effects of concern,
exposure-response relationships, susceptible subpopulations, MOAs and enhance
understanding of biological plausibility of observed effects. In the absence of clinical or
epidemiologic data, animal data alone may be sufficient to support a likely causal
determination, assuming that similar responses are expected in humans.
1.5.2.	Association and Causation
"Cause" is a significant, effectual relationship between an agent and an effect on
health or public welfare. "Association" is the statistical dependence among events,
characteristics, or other variables. An association is prima facie evidence for causation!
alone, however, it is insufficient proof of a causal relationship between exposure and disease
or health effect. Determining whether an observed association is causal rather than
spurious involves consideration of a number of factors, as described below. Much of the
July 2009
1-19
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
newly available health information evaluated in this ISA comes from epidemiologic studies
that report a statistical association between ambient exposure and health outcomes.
Many of the health and environmental outcomes reported in these studies have
complex etiologies. Diseases such as asthma, coronary artery disease or cancer are typically
initiated by a web of multiple agents. Outcomes depend on a variety of factors, such as age,
genetic susceptibility, nutritional status, immune competence, and social factors (Gee and
Payne-Sturges, 2004, 093070; IOM, 2008, 156586). Effects on ecosystems are also
multifactorial with a complex web of causation. Further, exposure to a combination of
agents could cause synergistic or antagonistic effects. Thus, the observed risk represents
the net effect of many actions and counteractions.
1.5.3. Evaluating Evidence for Inferring Causation
Moving from association to causation involves elimination of alternative explanations
for the association. In estimating the causal influence of an exposure on health or
environmental effects, it is recognized that scientific findings include uncertainty.
Uncertainty can be defined as a state of having limited knowledge where it is impossible to
exactly describe an existing state or future outcome! the lack of knowledge about the correct
value for a specific measure or estimate. Uncertainty characterization and uncertainty
assessment are two activities that lead to different degrees of sophistication in describing
uncertainty. Uncertainty characterization generally involves a qualitative discussion of the
thought processes that lead to the selection and rejection of specific data, estimates,
scenarios, etc. The uncertainty assessment is more quantitative. The process begins with
simpler measures (e.g., ranges) and simpler analytical techniques and progresses, to the
extent needed to support the decision for which the assessment is conducted, to more
complex measures and techniques. Data will not be available for all aspects of an
assessment and those data that are available may be of questionable or unknown quality.
In these situations, evaluation of uncertainty can include professional judgment or
inferences based on analogy with similar situations. The net result is that the assessments
will be based on a number of assumptions with varying degrees of uncertainty.
Uncertainties commonly encountered in evaluating health evidence for the criteria air
pollutants are outlined below for epidemiologic and experimental studies. Various
approaches to characterizing uncertainty include classical statistical methods, sensitivity
analysis, or probabilistic uncertainty analysis, in order of increasing complexity and data
requirements. The ISA generally evaluates uncertainties qualitatively in assessing the
evidence from across studies! in some situations quantitative analysis approaches, such as
meta-regression may be used.
It is important to note here that, although the following discussion refers primarily to
health effect studies, many parallels exist with welfare effects studies. Controlled exposure
July 2009
1-20
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
studies have been conducted in which plant species have been directly exposed to air
pollutants, and the strengths and limitations of that body of studies mirror those of the
controlled human exposure studies discussed below. Ecological field or natural gradient
studies are similar to epidemiologic studies, for example, in the study of free-living
populations and in the challenges faced in distinguishing effects of pollutants within a
mixture.
Controlled human exposure studies evaluate the effects of exposures to a variety of
pollutants in a highly-controlled laboratory setting. Also referred to as human clinical
studies, these experiments allow investigators to expose subjects to fixed concentrations of
air pollutants under carefully regulated environmental conditions and activity levels.
Controlled human exposures to PM typically involve exposing subjects either at rest or
while engaged in intermittent exercise in a whole-body exposure chamber, although
mouthpiece and facemask systems can also be used. A variety of different types of particles
are used in these studies including ambient outdoor particles, concentrated ambient
particles (CAPs), diesel exhaust (DE) from a diesel engine, wood smoke (WS) generated in a
wood stove, laboratory generated model particles (e.g., elemental carbon [EC] or zinc oxide
[ZnO]), or particles collected on a filter, resuspended in saline, and administered either
through instillation or inhalation (aerosolized and delivered using a nebulizer). The
recovery of particles on filters is variable and some components, such as organics, may be
too volatile to be collected. Exposures to artificially generated particles may provide
important information on the health effects of PM, but are not truly representative of
ambient air pollution particles. The direct exposure of humans to ambient air pollution
particles may be complicated by factors that cannot be controlled such as coexposures to
other air pollutants (e.g., O3, SO2, and NO2). In concentrating ambient particles, gaseous
copollutants are not proportionately concentrated and interactions between PM and the
copollutants cannot be investigated unless the latter are re-introduced. These limitations as
well as daily variability in concentration and composition can make it difficult to compare
the results from controlled human exposure studies employing particles from different
sources.
In some instances, controlled human exposure studies can also be used to characterize
concentration-response relationships at pollutant concentrations relevant to ambient
conditions. Controlled human exposures are typically conducted using a randomized
crossover design with subjects exposed both to PM and a clean air control. In this way,
subjects serve as their own controls, effectively controlling for many potential confounders.
However, controlled human exposure studies are limited by a number of factors including a
small sample size and short exposure time. These laboratory studies are often conducted at
PM concentrations much higher than those typically observed under ambient conditions,
which may result in an overestimate of the acute response to exposure in the general
population. Although the repetitive nature of ambient PM exposures may lead to
July 2009
1-21
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
cumulative health effects, this type of exposure is not practical to replicate in a laboratory
setting. In addition, while subjects do serve as their own controls, personal exposure to
pollutants in the hours and days preceding the controlled exposures may vary significantly
between and within individuals. Finally, controlled human exposure studies require
investigators to adhere to stringent health criteria for a subject to be included in the study,
and therefore the results cannot necessarily be generalized to an entire population.
Although some controlled human exposure studies have included health comprised
individuals such as asthmatics or individuals with chronic obstructive pulmonary disease
(COPD) or coronary artery disease, these individuals must also be relatively healthy and do
not represent the most sensitive individuals in the population. Thus, a lack of observation
of effects from controlled human exposure studies does not necessarily mean that a causal
relationship does not exist. While controlled human exposure studies provide important
information on the biological plausibility of associations observed between air pollutant
exposure and health outcomes in epidemiologic studies, observed effects in these studies
may underestimate the response in certain subpopulations.
Epidemiologic studies provide important information on the associations between
health effects and exposure of human populations to ambient air pollution. In the
evaluation of epidemiologic evidence, one important consideration is potential confounding.
Confounding is "... a confusion of effects. Specifically, the apparent effect of the exposure
of interest is distorted because the effect of an extraneous factor is mistaken for or mixed
with the actual exposure effect (which may be null)" (Rothman and Greenland S, 1998,
086599). One approach to remove spurious associations from possible confounders is to
control for characteristics that may differ between exposed and unexposed persons! this is
frequently termed "adjustment." Scientific judgment is needed regarding likely sources and
magnitude of confounding, together with consideration of how well the existing
constellation of study designs, results, and analyses address this potential threat to
inferential validity. One key consideration in this review is evaluation of the potential
contribution of PM to health effects when it is a component of a complex air pollutant
mixture. Reported PM effect estimates in epidemiologic studies may reflect independent
PM effects on respiratory and cardiovascular health. Ambient PM may also be serving as an
indicator of complex ambient air pollution mixtures that share the same source as PM
(i.e., combustion of S-containing fuels or motor vehicle emissions). Alternatively,
copollutants may mediate the effects of PM or PM may influence the toxicity of
copollutants.
Multivariable regression models constitute one tool for estimating the association
between exposure and outcome after adjusting for characteristics of participants that might
confound the results. The use of multipollutant regression models has been the prevailing
approach for controlling potential confounding by copollutants in air pollution health effects
studies. Finding the likely causal pollutant from multipollutant regression models is made
July 2009
1-22
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
difficult by the possibility that one or more air pollutants may be acting as a surrogate for
an unmeasured or poorly measured pollutant or for a particular mixture of pollutants. In
addition, more than one pollutant may exert similar health effects, resulting in
independently observed associations for multiple pollutants. Further, the correlation
between the air pollutant of interest and various copollutants may make it difficult to
discern associations between different pollutant exposures and health effects. Thus, results
of models that attempt to distinguish gaseous and particle effects must be interpreted with
caution. The number and degree of diversity of covariates, as well as their relevance to the
potential confounders, remain matters of scientific judgment. Despite these limitations, the
use of multipollutant models is still the prevailing approach employed in most air pollution
epidemiologic studies, and provides some insight into the potential for confounding or
interaction among pollutants.
Another way to adjust for potential confounding is through stratified analysis,
i.e., examining the association within homogeneous groups with respect to the confounding
variable. Stratified analysis can also be used to examine potential effect modification. The
use of stratified analyses has an additional benefit: it allows examination of effect
modification through comparison of the effect estimates across different groups. If
investigators successfully measured characteristics that distort the results, adjustment of
these factors help separate a spurious from a true causal association. Appropriate
statistical adjustment for confounders requires identifying and measuring all reasonably
expected confounders. Deciding which variables to control for in a statistical analysis of the
association between exposure and disease or health outcome depends on knowledge about
possible mechanisms and the distributions of these factors in the population under study.
Identifying these mechanisms makes it possible to control for potential sources that may
result in a spurious association.
Adjustment for potential confounders can be influenced by differential exposure
measurement error. There are several components that contribute to exposure
measurement error in epidemiologic studies, including the difference between true and
measured ambient concentrations, the difference between average personal exposure to
ambient pollutants and ambient concentrations at central monitoring sites, and the use of
average population exposure rather than individual exposure estimates. Previous AQCDs
have examined the role of measurement error in time-series epidemiologic studies using
simulated data and mathematical analyses and suggested that "transfer of effects" would
only occur under unusual circumstances (i.e., "true" predictors having high positive or
negative correlation! substantial measurement error! or extremely negatively correlated
measurement errors) (U.S. EPA, 2004, 056905).
Confidence that unmeasured confounders are not producing the findings is increased
when multiple studies are conducted in various settings using different subjects or
exposures! each of which might eliminate another source of confounding from consideration.
July 2009
1-23
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Thus, multicity studies which use a consistent method to analyze data from across locations
with different levels of covariates can provide insight on potential confounding in
associations. Intervention studies, because of their quasi-experimental nature, can be
particularly useful in characterizing causation.
In addition to clinical and epidemiologic studies, the tools of experimental biology
have been valuable for developing insights into human physiology and pathology. Animal
toxicological studies explore the effects of pollutants on human health, especially through
the study of model systems in other species. These studies evaluate the effects of exposures
to a variety of pollutants in a highly-controlled laboratory setting, and allow exploration of
MO As or mechanisms by which a pollutant may cause effects. Background knowledge of the
biological mechanisms by which an exposure might or might not cause disease can prove
crucial in establishing, or negating, a causal claim. There are, however, uncertainties
associated with quantitative extrapolations between laboratory animals and humans on the
pathophysiological effects of any pollutant. Animal species can differ from each other in
fundamental aspects of physiology and anatomy (e.g., metabolism, airway branching,
hormonal regulation) that may limit extrapolation. The differences between humans and
rodents with regard to pollutant absorption and distribution profiles based on breathing
pattern, exposure dose, and differences in lung structure and anatomy all have to be taken
into consideration.
A relatively new tool available for experimental studies of PM exposure is the particle
concentrator. Particle concentrators enable human subjects, animals, or cell culture
systems to be exposed to atmospheric PM at concentrations much greater than that
observed under ambient conditions. As ambient PM is just one component of a complex
mixture that interacts with gases and other aerosols, CAPs systems provide a method of
exposing subjects to the particle phase. Gases (such as O3 and SO2) are not concentrated
nor are thoracic coarse particles (except for the coarse particle concentrator) and only
certain systems are capable of concentrating ultrafine PM. In ultrafine CAPs systems,
increased number fraction of organic carbon and PAHs, along with decreased relative
percentage of EC particles have been reported in concentrated PM compared to ambient PM
(Su et al., 2006, 157021). These data suggest that for ultrafine concentrators, the CAPs do
not accurately reflect atmospheric ultrafine composition. There are several instrument
systems used to concentrate ambient PM in controlled human or animal exposure studies
(Gordon et al., 1999, 001176; Maciejczyk and Chen, 2005, 087456; Sioutas et al., 1995,
001629; Sioutas et al., 1999, 001633).
The ability to extrapolate between species has not generally changed since the 2004
PM AQCD but some considerations related to coarse particles merit attention. The
inhalability of particles greater than 2.5 pm in diameter is considerably lower in rats than
in humans! however, once inhaled, deposition in the extrathoracic region is near 100%
percent for particles greater than 5 pm for most laboratory animal species (rat, mouse,
July 2009
1-24
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
hamster, guinea pig, and dogs). By contrast, penetration of thoracic coarse particles into the
lower respiratory tract is greater in humans than rodents due to the moderately less
efficient nasal deposition of humans and oronasal breathing (especially during exercise).
The extent to which coarse particle deposition in the lower respiratory tract differs between
the species is highly dependent on the activity level of the human exposure scenario in
contrast with the resting exposure conditions common to rodent exposures. Endotracheal
exposures of rodents may be needed to achieve coarse particle tissue doses in the lower
respiratory tract of rodents similar to those experienced by humans. For particles less than
1 pm, including ultrafine particles, deposition is expected to be relatively similar between
the species.
There are also differences between species in both the rates of particle clearance from
and retention in the lung. The clearance rate of particles from the ciliated airways of rats is
considerably greater than humans. There is also evidence of prolonged particle retention in
the smaller bronchioles of humans that does not appear to exist or has not been observed in
rats. Under most circumstances, clearance from the alveolar region of rats is also more
rapid than observed in humans. Thus, these combined effects contribute to a greater
particle burden in the lower respiratory tract of humans relative to rats. An important
consideration in studies where are rats chronically exposed to high concentrations of
insoluble particles, is the potential for "overload conditions." Rats, unlike other laboratory
animals or humans, may experience a reduction in their alveolar clearance rates and an
accumulation of interstitial particle burden and under these conditions, the relevance of
tissue burdens and responses to humans is questionable. Considering interspecies
differences in both deposition and clearance, greater exposure concentrations are required
to achieve coarse particle tissue doses in the lower respiratory tract of rodents similar to
those experienced by humans.
1.5.4. Application of Framework for Causal Determination
EPA uses a two-step approach to evaluate the scientific evidence on health or
environmental effects of criteria pollutants. The first step determines the weight of
evidence in support of causation and characterizes the strength of any resulting causal
classification. The second step includes further evaluation of the quantitative evidence
regarding the concentration-response relationships and the loads or levels, duration and
pattern of exposures at which effects are observed.
July 2009
1-25
DRAFT-DO NOT CITE OR QUOTE

-------
Table 1-2 Aspects to aid in judging causality.
	Aspect	Description	
Consistency of the	An inference of causality is strengthened when a pattern of elevated risks is observed across several
observed association	independent studies, conducted in multiple locations by multiple investigators. The reproducibility of findings
constitutes one of the strongest arguments for causality. If there are discordant results among
investigations, possible reasons such as differences in exposure, confounding factors, and the power of the
study are considered.
An inference of causality from epidemiologic associations may be strengthened by other lines of evidence
(e.g., controlled human exposure and animal toxicological studies) that support a cause-and-effect
interpretation of the association. Causality is also supported when epidemiologic associations are reported
across study designs and across related health outcomes. Evidence on ecological or welfare effects may be
drawn from a variety of experimental approaches (e.g., greenhouse, laboratory, and field) and subdisciplines
of ecology (e.g., community ecology, biogeochemistry and paleological/ historical reconstructions). The
coherence of evidence from various fields greatly adds to the strength of an inference of causality. The
absence of other lines of evidence, however, is not a reason to reject causality
An inference of causality tends to be strengthened by consistency with data from experimental studies or
other sources demonstrating plausible biological mechanisms. A proposed mechanistic linking between an
effect, and exposure to the agent, is an important source of support for causality, especially when data
establishing the existence and functioning of those mechanistic links are available. A lack of biological
understanding, however, is not a reason to reject causality.
A well characterized exposure-response relationship (e.g., increasing effects associated with greater
exposure) strongly suggests cause and effect, especially when such relationships are also observed for
duration of exposure (e.g., increasing effects observed following longer exposure times). There are, however,
many possible reasons that a study may fail to detect an exposure-response relationship. Thus, although the
presence of a biological gradient may support causality, the absence of an exposure-response relationship
does not exclude a causal relationship.
The finding of large, precise risks increases confidence that the association is not likely due to chance, bias,
or other factors. However, given a truly causal agent, a small magnitude in the effect could follow from a
lower level of exposure, a lower potency, or the prevalence of other agents causing similar effects. While
large effects support causality, modest effects therefore do not preclude it.
Experimental evidence The strongest evidence for causality can be provided when a change in exposure brings about a change in
occurrence or frequency of health or welfare effects.
Temporal relationship of Evidence of a temporal sequence between the introduction of an agent and appearance of the effect
the observed association constitutes another argument in favor of causality.
Specificity of the	As originally intended, this refers to increased inference of causality if one cause is associated with a single
observed association	effect or disease (Hill, 1965, 071664). Based on the current understanding this is now considered one of the
weaker guidelines for causality; for example, many agents cause respiratory disease and respiratory disease
has multiple causes. At the scale of ecosystems, as in epidemiology, complexity is such that single agents
causing single effects, and single effects following single causes, are extremely unlikely. The ability to
demonstrate specificity under certain conditions remains, however, a powerful attribute of experimental
studies. Thus, although the presence of specificity may support causality, its absence does not exclude it.
Structure activity relationships and information on the agent's structural analogs can provide insight into
whether an association is causal. Similarly, information on mode of action for a chemical, as one of many
structural analogs, can inform decisions regarding likely causality.
Coherence
Biological plausibility
Biological gradient
(exposure-response
relationship)
Strength of the
observed association
Analogy
July 2009
1-26
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
To aid judgment, various "aspects"1 of causality have been discussed by many
philosophers and scientists. The most widely cited aspects of causality in epidemiology, and
public health, in general, were articulated by Sir Austin Bradford Hill in 1965 and have
been widely used (CDC, 2004, 056384; Hill, 1965, 071664; I ARC, 2006, 093206; IOM, 2008,
156586; NRC, 2004, 156814; U.S. EPA, 2005, 086237). Several adaptations of the Hill
aspects have been used in aiding causality judgments in the ecological sciences (Adams,
2003, 156192; Collier, 2003, 155736; Fox, 1991, 156444; Gerritsen et al„ 1998, 156465).
These aspects (Hill, 1965, 071664) have been modified (Table 1-2) for use in causal
determinations specific to health and welfare effects or pollutant exposures.2 Some aspects
are more likely than others to be relevant for evaluating evidence on the health or
environmental effects of criteria air pollutants. For example, the analogy aspect does not
always apply and specificity would not be expected for multi-etiologic health outcomes such
as asthma or cardiovascular disease, or ecological effects related to acidification. Aspects
that usually play a larger role in determination of causality are consistency of results across
studies, coherence of effects observed in different study types or disciplines, biological
plausibility, exposure-response relationship, and evidence from "natural" experiments.
Although these aspects provide a framework for assessing the evidence, they do not
lend themselves to being considered in terms of simple formulas or fixed rules of evidence
leading to conclusions about causality (Hill, 1965, 071664). For example, one cannot simply
count the number of studies reporting statistically significant results or statistically
nonsignificant results and reach credible conclusions about the relative weight of the
evidence and the likelihood of causality. In addition, it is important to note that the aspects
in Table 1-2 cannot be used as a strict checklist, but rather to determine the weight of the
evidence for inferring causality. While these aspects are particularly salient in this
assessment, it is also important to recognize that no one aspect is either necessary or
sufficient for drawing inferences of causality.
1.5.5. First Step-Determination of Causality
In the ISA, EPA assesses the results of recent relevant publications, building upon
evidence available during the previous NAAQS review, to draw conclusions on the causal
relationships between relevant pollutant exposures and health or environmental effects.
This ISA uses a five-level hierarchy that classifies the weight of evidence for causation, not
1	The "aspects" described by Hill (1965, 071664) have become, in the subsequent literature, more commonly described as
"criteria." The original term "aspects" is used here to avoid confusion with criteria' as it is used, with different meaning, in
the Clean Air Act.
2	The Hill aspects were developed for interpretation of epidemiologic results. They have been modified here for use with a
broader array of data, i.e., epidemiologic, controlled human exposure, and animal toxicological studies, as well as in vitro
data, and to be more consistent with EPA's Guidelines for Carcinogen Risk Assessment.
July 2009
1-27
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
just association1. In developing this hierarchy, EPA has drawn on the work of previous
evaluations, most prominently the IOM's Improving the Presumptive Disability
Decision-Making Process for Veterans (IOM, 2008, 156586). EPA's Guidelines for
Carcinogen Risk Assessment (U.S. EPA, 2005, 086237) and the U.S. Surgeon General's
smoking reports (CDC, 2004, 056384). This weight of evidence evaluation is based on
various lines of evidence from across the health and environmental effects disciplines.
These separate judgments are integrated into a qualitative statement about the overall
weight of the evidence and causality. The five descriptors for causal determination are
described in Table 1-3.
For PM, this determination of causality step involved a rather complex evaluation of
evidence for different PM indices, different types of health or environmental effects, and for
short- and long-term exposure periods. Determination of causality was made for both the
PM measure (pm2.5, PM10-2.5, and ultrafine particles, to the extent evidence was available for
each measure) and for the overall effect category. As noted above, to the extent possible,
results of PM10 studies are considered in causality determinations for PM2.5 and PM10-2.5. In
the evaluation of health effects findings in Chapter 6 (for short-term exposure) and Chapter
7 (for long-term exposure), evidence was evaluated for health outcome categories, such as
cardiovascular effects, and then conclusions were drawn based upon the integration of
evidence from across disciplines (e.g., epidemiology, controlled human exposure, and
toxicology) and also across the suite of related individual health outcomes. These chapters
initially summarize and evaluate findings for individual health outcomes, then integrate
the results in summary sections to draw conclusions on causality for each PM indicator. The
causality narratives present the weight of evidence that highlights the quality and breadth
of the data, including any limitations or uncertainties. In the integrative synthesis and
conclusions in Chapter 2, the ISA presents causality determinations and a summary of the
underlying basis for those determinations for the PM indicator (e.g., PM2.5), for the
exposure time period (e.g., short- and long-term exposure) and for the major health effect
categories.
1 It should be noted that the CDC and IOM frameworks use a four-category hierarchy for the strength of the evidence. A
five-level hierarchy is used here to be consistent with the EPA Guidelines for Carcinogen Risk Assessment and to provide a
more nuanced set of categories.
July 2009
1-28
DRAFT-DO NOT CITE OR QUOTE

-------
Table 1-3. Weight of evidence for causal determination.
Determination
Health Effects
Ecological and Welfare Effects
Causal	Evidence is sufficient to conclude that there is a causal
relationship relationship with relevant pollutant exposures. That is, the
pollutant has been shown to result in health effects in studies
in which chance, bias, and confounding could be ruled out
with reasonable confidence. For example: a) controlled human
exposure studies that demonstrate consistent effects; or b)
observational studies that cannot be explained by plausible
alternatives or are supported by other lines of evidence (e.g.,
animal studies or mode of action information). Evidence
includes replicated and consistent high-quality studies by
multiple investigators.
Evidence is sufficient to conclude that there is a causal
relationship with relevant pollutant exposures. That is, the
pollutant has been shown to result in effects in studies in
which chance, bias, and confounding could be ruled out with
reasonable confidence. Controlled exposure studies
(laboratory or small- to medium-scale field studies) provide
the strongest evidence for causality, but the scope of
inference may be limited. Generally, determination is based
on multiple studies conducted by multiple research groups,
and evidence that is considered sufficient to infer a causal
relationship is usually obtained from the joint consideration
of many lines of evidence that reinforce each other.
Likely to be a Evidence is sufficient to conclude that a causal relationship is
causal	likely to exist with relevant pollutant exposures, but important
relationship uncertainties remain. That is, the pollutant has been shown to
result in health effects in studies in which chance and bias
can be ruled out with reasonable confidence but potential
issues remain. For example: a) observational studies show an
association, but copollutant exposures are difficult to address
and/or other lines of evidence (controlled human exposure,
animal, or mode of action information) are limited or
inconsistent; or b) animal toxicological evidence from multiple
studies from different laboratories that demonstrate effects,
but limited or no human data are available. Evidence generally
includes replicated and high-quality studies by multiple
investigators.
Evidence is sufficient to conclude that there is a likely causal
association with relevant pollutant exposures. That is, an
association has been observed between the pollutant and the
outcome in studies in which chance, bias and confounding
are minimized, but uncertainties remain. For example, field
studies show a relationship, but suspected interacting
factors cannot be controlled, and other lines of evidence are
limited or inconsistent. Generally, determination is based on
multiple studies in multiple research groups.
Suggestive of
a causal
relationship
Evidence is suggestive of a causal relationship with relevant
pollutant exposures, but is limited because chance, bias and
confounding cannot be ruled out. For example, at least one
high-quality epidemiologic study shows an association with a
given health outcome but the results of other studies are
inconsistent.
Evidence is suggestive of a causal relationship with relevant
pollutant exposures, but chance, bias and confounding
cannot be ruled out. For example, at least one high-quality
study shows an effect, but the results of other studies are
inconsistent.
Inadequate to
infer a causal
relationship
Evidence is inadequate to determine that a causal relationship
exists with relevant pollutant exposures. The available studies
are of insufficient quantity, quality, consistency or statistical
power to permit a conclusion regarding the presence or
absence of an effect.
The available studies are of insufficient quality, consistency
or statistical power to permit a conclusion regarding the
presence or absence of an effect.
Not likely to Evidence is suggestive of no causal relationship with relevant
he a causal pollutant exposures. Several adequate studies, covering the
relationship 'u" ran9e °' 'eve's °' exposure that human beings are known
to encounter and considering susceptible or vulnerable
subpopulations, are mutually consistent in not showing an
effect at any level of exposure.
Several adequate studies, examining relationships with
relevant exposures, are consistent in failing to show an
effect at any level of exposure.
July 2009
1-29
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
1.5.6. Second Step-Evaluation of Response
Beyond judgments regarding causality are questions relevant to quantifying health or
environmental risks based on our understanding of the quantitative relationships between
pollutant exposures and health or welfare effects.
1.5.6.1. Effects on Human Populations
Once a determination is made regarding the causal relationship between the
pollutant and outcome category, important questions regarding quantitative relationships
include^
¦	What is the concentration-response or dose-response relationship in the human
population?
¦	What exposure conditions (dose or exposure, duration and pattern) are important?
¦	What subpopulations appear to be differentially affected i.e., more
susceptible/vulnerable to effects?
To address these questions, in the second step of the EPA framework, the entirety of
quantitative evidence is evaluated to best quantify those concentration-response
relationships that exist. This requires evaluation of levels of pollutant and exposure
durations at which effects were observed for exposed populations including potentially
susceptible subpopulations. This integration of evidence results in identification of a study
or set of studies that best approximates the concentration-response relationships between
health outcomes and PM indicators for the U.S. population or subpopulations, given the
current state of knowledge and the uncertainties that surrounded these estimates.
To accomplish this, evidence from multiple and diverse types of studies is considered.
To the extent available, the ISA evaluates results from across epidemiologic studies that
use various methods to evaluate the form of relationships between PM and health
outcomes, and draws conclusions on the most well-supported shape of these relationships.
Animal data may also inform evaluation of concentration-response relationships,
particularly relative to MOAs, and characteristics of susceptible subpopulations. For some
health outcomes, the probability and severity of health effects and associated uncertainties
can be characterized. Chapter 2 presents the integrated findings informative for evaluation
of population risks.
An important consideration in characterizing the public health impacts associated
with exposure to a pollutant is whether the concentration-response relationship is linear
across the full concentration range encountered, or if nonlinear relationships exist along
any part of this range. Of particular interest is the shape of the concentration-response
curve at and below the level of the current standards. The shape of the
July 2009
1-30
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
concentration-response curve varies, depending on the type of health outcome, underlying
biological mechanisms and dose. At the human population level, however, various sources of
variability and uncertainty tend to smooth and "linearize" the concentration-response
function (such as the low data density in the lower concentration range, possible influence
of measurement error, and individual differences in susceptibility to air pollution health
effects). In addition, many chemicals and agents may act by perturbing naturally occurring
background processes that lead to disease, which also linearizes population
concentration-response relationships (Clewell and Crump, 2005, 156359; Crump et al.,
1976, 003192; Hoel, 1980, 156555). These attributes of population dose-response may
explain why the available human data at ambient concentrations for some environmental
pollutants (e.g., PM, O3, lead [Pb], environmental tobacco smoke [ETS], radiation) do not
exhibit evident thresholds for health effects, even though likely mechanisms include
nonlinear processes for some key events. These attributes of human population
dose-response relationships have been extensively discussed in the broader epidemiologic
literature (Rothman and Greenland S, 1998, 086599).
Publication bias is a source of uncertainty regarding the magnitude of health risk
estimates. It is well understood that studies reporting non-null findings are more likely to
be published than reports of null findings, and publication bias can also result in
overestimation of effect estimate sizes (Ioannidis, 2008, 188317). For example, effect
estimates from single-city epidemiologic studies have been found to be generally larger than
those from multicity studies (Anderson et al., 2005, 087916).
Finally, identification of the susceptible or vulnerable population groups contributes
to an understanding of the public health impact of pollutant exposures. Epidemiologic
studies can help identify susceptible or vulnerable subpopulations by evaluating health
responses in the study population. Examples include stratified analyses for subsets of the
population under study, or testing for interactions or effect modification by factors such as
gender, age group, or health status. Experimental studies using animal models of
susceptibility or disease can also inform the extent to which health risks are likely greater
in specific population subgroups. Further discussion of these groups is in Chapter 8.
1.5.6.2. Effects on Public Welfare
Key questions for understanding the quantitative relationships between exposure (or
concentration or deposition) to a pollutant and risk to the public welfare (e.g., ecosystems,
visibility, materials, climate):
¦	What elements of the ecosystem (e.g., types, regions, taxonomic groups, populations,
functions, etc.) appear to be affected, or are more sensitive to effects?
¦	Under what exposure conditions (amount deposited or concentration, duration and
pattern) are effects seen?
July 2009
1-31
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
¦ What is the shape of the concentration-response or exposure-response relationship?
Evaluations of causality typically consider the probability of welfare effects changing
in response to exposure. A challenge to the quantification of exposure-resonse relationships
for ecological effects is the variability across ecosystems. Ecological responses are evaluated
within the range of observations, so a quantitative relationshi may be determined for a
given ecological system and scale. However, there is great regional and local variability in
ecosystems. Thus, exposure-response relationships are often available site by site, rather
than at the national or even regional scale. Quantitative relationships therefore are
available site by site. For example, an ecological response to deposition ofa given pollutant
can differ greatly between ecosystems. Where results from greenhouseor nimal ecotox-
icological studies are available, they may be used to aid in character-zing exposure-response
relations, particularly relative to mechanisms of action, and characteristics of sensitive
biota.
1.5.7. Concepts in Evaluating Adversity of Health Effects
In evaluating the health evidence, a number of factors can be consiered in deter-
mining the extent to which health effects are "adverse" for health outcomes such as changes
in lung function. What constitutes an adverse health effect may vary between populations.
Some changes in healthy individuals may not be considered adverse while those of a similar
type and magnitude are potentially adverse in more susceptible individuals.
The American Thoracic Society (ATS) published an official statement titled What
Constitutes an Adverse Health Effect of Air Pollution? (ATS, 2000, 011738). This statement
updated the guidance for defining adverse respiratory health effects that had been
published 15 years earlier (ATS, 1985, 006522), taking into account new investigative
approaches used to identify the effects of air pollution and reflecting concern for impacts of
air pollution on specific susceptible groups. In the 2000 update, there was an increased
focus on quality of life measures as indicators of adversity and a more specific consideration
of population risk. Exposure to air pollution that increases the risk of an adverse effect to
the entire population is viewed as adverse, even though it may not increase the risk of any
identifiable individual to an unacceptable level; estimated mean population effects do not
reflect more severe effects in individuals. For example, a population of asthmatics could
have a distribution of lung function such that no identifiable individual has a level
associated with significant impairment. Exposure to air pollution could shift the
distribution such that no identifiable individual experiences clinically-relevant effects! this
shift toward decreased lung function, however, would be considered adverse because
individuals within the population would have diminished reserve function and, therefore,
would be at increased risk to further environmental insult.
July 2009
1-32
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
1.6. Summary
This second external review draft ISA is a review, synthesis, and evaluation of the
most policy-relevant science, and communicates critical science judgments relevant to the
NAAQS review. It reviews the most policy-relevant evidence frenvironmend environmenttal
effects studies and includes information on atmospheric chemistry, PM sources and
emissions, exposure, and dosimetry. This draft ISA incorporates clarification and revisions
based on advice and comments provided by EPA's CASAC (Samet, 2009, 190992). Annexes
to the ISA provide additional details of the literature published since the last review. A
framework for making critical judgments concerning causality appears in this chapter. It
relies on a widely accepted set of principles and standardized language to xpress evaluation
of the evidence. This approach can bring rigor and clarity to the current and future
assessments. This ISA should assist EPA and others, now and in the future, to accurately
represent what is presently known—and what remains unknown—concerning the effects of
PM on human health and public welfare.
July 2009
1-33
DRAFT-DO NOT CITE OR QUOTE

-------
Chapter 1 References
Adams SM. (2003). Establishing causality between environmental stressors and effects on aquatic ecosystems.
Human and Ecological Risk Assessment. , 9- 17-35. 156192
Anderson HR; Atkinson RW! Peacock JL; Sweeting MJ; Marston L. (2005). Ambient particulate matter and
health effects' Publication bias in studies of short-term associations. Epidemiology, 16- 155-163. 087916
ATS. (1985). Guidelines as to what constitutes an adverse respiratory health effect, with special reference to
epidemiologic studies of air pollution. American Thoracic Society . Am Rev Respir Dis, 131- 666-668.
006522
ATS. (2000). What constitutes an adverse health effect of air pollution? Official statement of the American
Thoracic Society. Am J Respir Crit Care Med, 161- 665-673. 011738
CDC. (2004). The health consequences of smoking- A report of the Surgeon General. Centers for Disease Control
and Prevention, U.S. Department of Health and Human Services. Washington, DC. 056384
Clewell HJ; Crump KS. (2005). Quantitative estimates of risk for noncancer endpoints. Risk Anal, 25- 285-289.
156359
Collier TK. (2003). Forensic ecotoxicology- Establishing causality between contaminants and biological effects in
field studies. Human and Ecological Risk Assessment. , 9- 259 - 266. 155736
Crump KS; Hoel DG; Langley CH; Peto R. (1976). Fundamental carcinogenic processes and their implications
for low dose risk assessment. Cancer Res, 36- 2973-2979. 003192
Fox GA. (1991). Practical causal inference for ecoepidemiologists. J Toxicol Environ Health, 33- 359-373. 156444
Gee GC; Payne-Sturges DC. (2004). Environmental health disparities' A framework integrating psychosocial
and environmental concepts. Environ Health Perspect, 112' 1645-1653. 093070
Gerritsen J; Carlson RE; Dycus DL; Faulkner C; Gibson GR. (1998). Lake and reservoir bioassessment and
biocriteria' Technical guidance document. US Environmental Protection Agency, Office of Water.
Washington, DC. EPA 841-B-98-007. httpV/www.epa.gov/owow/monitoring/tech/lakes.html . 156465
Gordon T; Gerber H; Fang CP; Chen LC. (1999). A centrifugal particle concentrator for use in inhalation
toxicology. Inhal Toxicol, 11' 71-87. 001176
Henderson R. (2005). Clean Air Scientific Advisory Committee (CASAC) review of the EPA staff
recommendations concerning a Potential Thoracic Coarse PM standard in the Review of the National
Ambient Air Quality Standards for Particulate Matter' Policy Assessment of Scientific and Technical
Information. U.S. Environmental Protection Agency. RTP, NC. 156537
Henderson R. (2005). EPA's Review of the National Ambient Air Quality Standards for Particulate Matter
(Second Draft PM Staff Paper, January 2005)' A review by the Particulate Matter Review Panel of the
EPA Clean Air Scientific Advisory Committee. U.S. Environmental Protection Agency. Washington D.C..
188316
Henderson R. (2006). Clean Air Scientific Advisory Committee recommendations concerning the proposed
National Ambient Air Quality Standards for particulate matter. U.S. Environmental Protection Agency.
Washington, DC. 156538
Hill AB. (1965). The environment and disease' Association or causation. J R Soc Med, 58' 295-300. 071664
Hoel DG. (1980). Incorporation of background in dose-response models. FedProc, 39' 73"75. 156555
IARC. (2006). IARC monographs on the evaluation of carcinogenic risks to humans' Preamble. International
Agency for Research on Cancer. Lyon,France .
http'//monographs.iarc.fr/ENG/Preamble/CurrentPreamble.pdf . 093206
Ioannidis JPA. (2008). Why most discovered true associations are inflated. Epidemiology, 19' 640-648. 188317
IOM. (2008). Improving the Presumptive Disability Decision-Making Process for Veterans; Committee on
Evaluation of the Presumptive Disability Decision-Making Process for Veterans, Board on Military and
Veterans Health. Washington, DC' Institue of Medicine of the National Academies, National Academies
Press. 156586
Maciejczyk P; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in
mice' VIII source-related daily variations in in vitro responses to CAPs. Inhal Toxicol, 17' 243-253.
087456
NAPAP. (1991). The experience and legacy of NAPAP report of the Oversight Review Board. National Acid
Precipitation Assessment Program. Washington, DC. 095894
NAPCA. (1969). Air quality criteria for particulate matter. National Air Pollution Control Administration.
Washington, DC. 014684
NRC. (2004). Research priorities for airborne particulate matter' IV Continuing research progress. National
Academies Press . Washington, DC. 156814
July 2009
1-34
DRAFT-DO NOT CITE OR QUOTE

-------
Rothman KJ; Greenland S eds. (1998). Modern epidemiology. Philadelphia, PA- Lippincott-Raven Pubhshers.
086599
Samet JM. (2009). Review of EPA's Integrated Science Assessment for Particulate Matter (First External
Review Draft, December 2008). Clean Air Scientific Advisory Committee (CASAC) and members of the
CASAC Particulate Matter (PM) Review Panel, Science Advisory Board, U.S. Environmental Protection
Agency. Washington, D.C.. EPA-CASAC-09-008. 190992
Sioutas C> Kim Si Chang M. (1999). Development and evaluation of a prototype ultrafine particle concentrator.
J Aerosol Sci, 30: 1001-1017. 001633
Sioutas Di Koutrakis Pi Ferguson STi Burton RM. (1995). Development and evaluation of a prototype ambient
particle concentrator for inhalation exposure studies. Inhal Toxicol, T- 633-644. 001629
Su Y> Sipin MF; Spencer MT; Qin X; Moffet RC; Shields LG; Prather KA; Venkatachari P; Jeong C-H> Kim E>
Hopke PE Gelein RM; Utell MJ; Oberdorster G> Berntsen J; Devlin RB; Chen LC. (2006). Real-time
characterization of the composition of individual particles emitted from ultrafine particle concentrators.
Aerosol Sci Technol, 40: 437 - 455. 157021
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA/600/P"99/002aF"bF. 056905
U.S. EPA. (2005). Guidelines for carcinogen risk assessment, Risk Assessment Forum Report. U.S.
Environmental Protection Agency. Washington, DC. EPA/630/P-03/001F.
http://cfpub.epa.gov/ncea/index.cfm . 086237
U.S. EPA. (2005). Review of the national ambient air quality standards for particulate matter: Policy
assessment of scientific and technical information OAQPS staff paper. Office of Air Quality Planning and
Standards, U.S. EPA. Research Triangle Park, North Carolina. EPA/452/R"05"005a.
http://www.epa.gov/ttn/naaqs/standards/pm/data/pmstaffpaper_20051221.pdf . 090209
U.S. EPA. (2008). Integrated review plan for the national ambient air quality standards for particulate matter.
U.S. Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment. Research Triangle Park, NC. 157072
U.S. EPA. (2008). Integrated science assessment for oxides of nitrogen and sulfur: Ecological criteria. EPA.
Research Triangle Park, NC. EPA/600/R-08/082F. 157074
July 2009
1-35
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Chapter 2. Integrative Health and
Welfare Effects Overview
The subsequent chapters of this ISA will present the most policy-relevant information
related to this review of the NAAQS for PM. This chapter integrates the key findings from
the disciplines evaluated in this current assessment of the PM scientific literature, which
includes the atmospheric sciences, ambient air data analyses, exposure assessment,
dosimetry, health studies (e.g., toxicological, controlled human exposure, and
epidemiologic), and ecological and welfare effects. The EPA framework for causal
determinations described in Chapter 1 has been applied to the body of scientific evidence in
order to collectively examine the health effects, or ecological and welfare effects attributed
to PM exposure in a two-step process.
As described in Chapter 1, EPA assesses the results of recent relevant publications,
building upon evidence available during the previous NAAQS review, to draw conclusions
on the causal relationships between relevant pollutant exposures and health or
environmental effects. This ISA uses a five-level hierarchy that classifies the weight of
evidence for causation:
¦	Causal relationship
¦	Likely to be a causal relationship
¦	Suggestive of a causal relationship
¦	Inadequate to infer a causal relationship
¦	Not likely to be a causal relationship
Beyond judgments regarding causality are questions relevant to quantifying health or
environmental risks based on our understanding of the quantitative relationships between
pollutant exposures and health or welfare effects. Once a determination is made regarding
the causal relationship between the pollutant and outcome category, important questions
regarding quantitative relationships include^
¦	What is the concentration-response or dose-response relationship?
¦	Under what exposure conditions (amount deposited, dose or concentration, duration
and pattern) are effects seen?
Note- Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health
and Environmental Research Online) at http • I lev a. go v/her o. HERO is a database of scientific literature used by U.S. EPA in
the process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009
2-2
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
¦ What subpopulations appear to be differentially affected i.e., more susceptible to
effects? What elements of the ecosystem (e.g., types, regions, taxonomic groups,
populations, functions, etc.) appear to be affected, or are more sensitive to effects?
To address these questions, in the second step of the EPA framework, the entirety of
quantitative evidence is evaluated to identify and characterize potential
concentration-response relationships. This requires evaluation of levels of pollutant and
exposure durations at which effects were observed for exposed populations including
potentially susceptible subpopulations.
This chapter summarizes and integrates the newly available scientific evidence that
best informs consideration of the policy relevant questions that frame this assessment,
presented in Chapter 1. Section 2.1 discusses the trends in ambient concentrations and
sources of PM and provides a brief summary of ambient air quality for short- and long-term
exposure durations. Section 2.2 presents the evidence regarding personal exposure to
ambient PM in outdoor and indoor microenvironments, and it discusses the relationship
between ambient PM concentrations and exposure to PM from ambient sources. Section 2.3
integrates the evidence for studies that examine the health effects associated with short-
and long-term exposure to PM and discusses important uncertainties identified in the
interpretation of the scientific evidence. In addition, this section presents the evidence for
potentially susceptible subpopulations to PM exposure. Section 2.4 provides discussion of
policy-relevant considerations, such as potentially susceptible subpopulations, lag
structure, and the PM concentration-response relationship. Finally, Section 2.5 summarizes
the evidence for ecological and welfare effects related to PM exposure.
2.1. Concentrations and Sources of Atmospheric PM
2.1.1. Ambient PM Variability and Correlations
Recently, advances in understanding the spatiotemporal distribution of PM mass and
its constituents have been made, particularly with regard to PM2.5 and its components as
well as ultrafine particles (UFP). Emphasis in this ISA is placed on the period from 2005-
2007 incorporating the most recent validated EPAAir Quality System (AQS) data. The AQS
is EPA's repository for ambient monitoring data reported by the national, and state and
local air monitoring networks. Measurements of PM2.5 and PM10 are reported into AQS
while PM10-2.5 concentrations are obtained as the difference between PM10 and PM2.5 (after
converting PM10 concentrations from STP to local conditions! Section 3.5). Note, however,
that a majority of U.S. counties were not represented in AQS because their population fell
below the regulatory monitoring threshold. Moreover, monitors reporting to AQS were not
uniformly distributed across the U.S. or within counties, and conclusions drawn from AQS
July 2009
2-3
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
data may not apply equally to all parts of a geographic region. Furthermore, biases can
exist for some PM constituents (and hence total mass) owing to volatilization losses of
nitrates and other semi-volatile compounds, and, conversely, to retention of particle-bound
water by hygroscopic species. The degree of spatial variability in PM was likely to be
region-specific and strongly influenced by local sources and meteorological and topographic
conditions.
2.1.1.1. Spatial Variability across the U.S.
AQS data for concentrations of PM2.5 for 2005-2007 showed considerable variability
across the U.S. (Section 3.5.1.1). Counties with the highest average concentrations of PM2.5
(>18 |ug/m3) were reported for several counties in the San Joaquin Valley and inland
southern California as well as Jefferson County, AL (containing Birmingham) and
Allegheny County, PA (containing Pittsburgh). Relatively few regulatory monitoring sites
have the appropriate co-located monitors for computing PM10-2.5, resulting in poor
geographic coverage on a national scale (Figure 3-10). Although the general understanding
of PM differential settling leads to an expectation of greater spatial heterogeneity in the
PM10-2.5 fraction, deposition of particles as a function of size depends strongly on local
meteorological conditions. Better geographic coverage is available for PM10, where the
highest reported annual average concentrations (>50 ug/m3) occurred in southern
California, southern Arizona and central New Mexico. The size distribution of PM varied
substantially by location, with a generally larger fraction of PM10 mass in the PM10-2.5 size
range in western cities (e.g., Phoenix and Denver) and a larger fraction of PM10 in the PM2.5
size range in eastern U.S. cities (e.g., Pittsburgh and Philadelphia). UFPs are not measured
as part of AQS or any other routine regulatory network in the U.S. Therefore, limited
information is available regarding regional variability in the spatiotemporal distribution of
UFPs.
Spatial variability in PM2.5 components obtained from the Chemical Speciation
Network (CSN) varied considerably by species from 2005-2007 (Figures 3-12 through 3-18).
The highest annual average organic carbon (OC) concentrations were observed in the
western and southeastern U.S. OC concentrations in the western U.S. peaked in the fall
and winter, while OC concentrations in the Southeast peaked anytime between spring and
fall. Elemental carbon (EC) exhibited less seasonality than OC and showed lowest seasonal
variability in the eastern half of the U.S. The highest annual average EC concentrations
were present in Los Angeles, Pittsburgh, New York, and El Paso. Concentrations of sulfate
(SO42-1 were higher in the eastern U.S. as a result of higher sulfur dioxide (SO2) emissions
in the East compared with the West. There is also considerable seasonal variability with
higher S( ) r concentrations in the summer months when the oxidation of SO2 proceeds at a
faster rate than during the winter. Nitrate (NO3-1 concentrations were highest in California
and during the winter in the Upper Midwest. In general, NOa was higher in the winter
July 2009
2-4
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
across the country, in part as a result of temperature "driven partitioning and volatilization.
Exceptions existed in Los Angeles and Riverside, CA, where high N03-concentrations
appeared year-round. There is variation in both PM2.5 mass and composition among cities,
some of which might be due to regional differences in meteorology, sources, and topography.
2.1.1.2. Spatial Variability on the Urban and Neighborhood Scales
In general, PM2.5 has a longer atmospheric lifetime than PM10-2.5. As a result, PM2.5 is
more homogeneously distributed than PM10-2.5, whose concentrations more closely reflect
proximity to local sources (Section 3.5.1.2). Since PM10 incorporates PM10-2.5 in addition to
PM2.5, it also exhibits more spatial heterogeneity than PM2.5. Urban- and neighborhood-
scale variability in PM mass and composition was examined by focusing on fifteen
metropolitan areas, which were chosen based on their geographic distribution and coverage
in recent health effects studies. The urban areas selected were Atlanta, Birmingham,
Boston, Chicago, Denver, Detroit, Houston, Los Angeles, New York, Philadelphia, Phoenix,
Pittsburgh, Riverside, Seattle and St. Louis. Inter-monitor correlation remained higher
over long distances for PM2.5 as compared with PM10 in these 15 urban areas. To a large
extent, greater variation in PM2.5 and PM10 concentrations within cities was observed in
areas with lower ratios of PM2.5 to PM10. When the data was limited to only sampler pairs
with less than 4 km separation (i.e., on a neighborhood scale), inter-sampler correlations
remained higher for PM2.5 than for PM10. Average correlation was maintained at 0.93 for
PM2.5, while it dropped to 0.70 for PM10 (Section 3.5.1.3). Insufficient data were available in
the 15 metropolitan areas to perform similar analyses for PM10-2.5 using co-located, low
volume FRM monitors.
As previously mentioned, UFPs are not measured as part of AQS or any other routine
regulatory network in the U.S. Therefore, information about the spatial variability of UFPs
is sparse! however, their number concentrations are expected to be highly spatially and
temporally variable. This has been shown on the urban scale in studies in which UFP
number concentrations drop off quickly with distance from roads compared to accumulation
mode particle numbers.
2.1.2. Trends and Temporal Variability
Overall, PM2.5 concentrations decreased from 1999 (the beginning of nationwide
monitoring for PM2.5) to 2007 in all 10 EPA Regions, with the 3-yr avg of the 98th percentile
of 24-h PM2.5 concentrations dropping 10% over this time period. However from 2002 to
2007, concentrations of PM2.5 were nearly constant with decreases observed in only some
EPA Regions (Section 3.5.2.1). Concentrations of PM2.5 components were only available for
2002-2007 using CSN data and showed little decline over this time period. This trend in
PM2.5 components is consistent with trends in PM2.5 mass concentration observed after 2002
July 2009
2-5
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
(shown in Figures 3-44 through 3-47). Concentrations of PMio also declined from 1988 to
2007 in all 10 EPA Regions.
Using hourly PM observations in the 15 metropolitan areas, diel variation showed
average hourly peaks that differ by size fraction and region (Section 3.5.2.3). For both PM2.5
and PM10, a morning peak was typically observed starting at approximately 6^00 a.m.,
corresponding with the start of morning rush hour. There was also an evening
concentration peak that was broader than the morning peak and extended into the
overnight period, reflecting the concentration increase caused by the usual collapse of the
mixing layer after sundown. The magnitude and duration of these peaks varied
considerably by metropolitan area investigated.
UFPs were found to exhibit similar two-peaked diel patterns in Los Angeles and the
San Joaquin Valley of CA, Rochester, NY as well as in Kawasaki City, Japan and
Copenhagen, Denmark. The morning peak in UFPs likely represents primary source
emissions, such as rush-hour traffic, while the afternoon peak likely represents the
combination of primary source emissions and nucleation of new particles.
2.1.3.	Correlations between Copollutants
Correlations between PM and gaseous copollutants including SO2, nitrogen dioxide
(NO2), carbon monoxide (CO) and ozone (O3) varied both seasonally and spatially between
and within metropolitan areas (Section 3.5.3). On average, PM2.5 and PM10 were correlated
with each other better than with the gaseous copollutants. There was relatively little
seasonal variability in the mean correlation between PM in both size fractions and SO2 and
NO2. CO, however, showed higher correlations with PM2.5 and PM10 on average in the
winter compared with the other seasons. This seasonality results in part because a larger
fraction of PM is primary in origin during the winter. To the extent that this primary
component of PM is associated with common combustion sources of NO2 and CO, then
higher correlations with these gaseous copollutants are to be expected. Increased
atmospheric stability in colder months also results in higher correlations between primary
pollutants (Section 3.5).
The correlation between daily maximum 8"h avg O3 and PM2.5 showed the highest
degree of seasonal variability with positive correlations on average in the spring, summer
and fall, and negative correlations on average in the winter. This situation arises as the
result of seasonal differences in sources and photochemical production of secondary PM2.5
and O3. However, this relationship was not present in Birmingham, Boston and St. Louis.
2.1.4.	Measurement Techniques
The federal reference methods (FRMs) for PM2.5 and PM10 are based on criteria
outlined in the Code of Federal Regulations. They are, however, subject to several
July 2009
2-6
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
limitations that should be kept in mind when using compliance monitoring data for health
studies. For example, FRM techniques are subject to the loss of semi-volatile species such
as organic compounds and ammonium nitrate (especially in the West). Since FRMs based
on gravimetry use 24-h integrated filter samples to collect PM mass, no information is
available for variations over shorter averaging times from these instruments. However,
reliable methods have been developed to measure real-time PM mass concentrations. Real-
time (or continuous and semi-continuous) measurement techniques are also available for
PM species, such as particle into liquid sampler (PILS) for multiple ions analysis and
aerosol mass spectrometer (AMS) for multiple components analysis (Section 3.4.1).
Advances have also been achieved in PM organic speciation. New 24-h FRMs and Federal
Equivalent Methods (FEMs) based on gravimetry and continuous FEMs for PM10-2.5 are
available. FRMs for PM10-2.5 rely on calculating the difference between co-located PM10 and
PM2.5 measurements and a dichotomous sampler is designated as an FEM.
2.1.5. PM Formation in the Atmosphere and Removal
PM in the atmosphere contains both primary (i.e., emitted directly by sources) and
secondary components, which can be anthropogenic or natural in origin. Secondary PM
components can be produced by the oxidation of precursor gases such as SO2 and NOx to
acids followed by neutralization with ammonia (NH3) and the partial oxidation of organic
compounds. In addition to being emitted as primary particles, UFPs are produced by the
nucleation of H2SO4 vapor and perhaps NH3 and certain organic compounds. Over most of
the earth's surface, nucleation is probably the major mechanism forming new UFPs. New
UFP formation has been observed in environments ranging from relatively unpolluted
marine and continental environments to polluted urban areas as an ongoing background
process and during nucleation events. However, as noted above a large percentage of UFPs
come from combustion-related sources such as mobile sources.
Developments in the chemistry of formation of secondary organic aerosol (SOA)
indicate that oligomers are likely a major component of OC in aerosol samples. Recent
observations also suggest that small but significant quantities of SOA are formed from
isoprene oxidation of terpenes and organic hydrocarbons with six or more carbon atoms.
Gasoline engines have been found to emit a mix of nucleation-mode heavy and large
polycyclic aromatic hydrocarbons on which unspent fuel and trace metals can condense,
while diesel particles are composed of a soot nucleus on which sulfates and hydrocarbons
can condense. To the extent that the primary component of organic aerosol is overestimated
in emissions from combustion sources, the semi-volatile components are underestimated.
This situation results from the lack of capture of evaporated semi-volatile components upon
dilution in common emissions tests. As a result, near-traffic sources of precursors to
secondary organic aerosol would be underestimated. The oxidation of these precursors
July 2009
2-7
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
results in more oxidized forms of SOAthan previously considered, both in near source
urban environments and further downwind. Organic peroxides constitute a significant
fraction of SOA and represent an important class of reactive oxygen species (ROS) that
have high oxidizing potential. More information on sources, emissions and deposition of PM
are included in Section 3.3.
Wet and dry deposition are important processes for removing PM and other pollutants
from the atmosphere on urban, regional, and global scales. Wet deposition includes
incorporation of particles into cloud droplets that fall as rain (rainout) and collisions with
falling rain (washout). Other hydrometeors (snow, ice) can also serve the same purpose. Dry
deposition involves transfer of particles through gravitational settling and/or by impaction
on surfaces by turbulent motions. The effects of deposition of PM on ecosystems and
materials are discussed in Section 2.5 and in Chapter 9.
2.1.6. Source Contributions to PM
Results of receptor modeling calculations indicate that PM2.5 is produced mainly by
combustion of fossil fuel, either by stationary sources or by transportation. A relatively
small number of broadly defined source categories, compared to the total number of
chemical species that typically are measured in ambient monitoring source receptor
studies, account for the majority of the observed PM mass. Some ambiguity is inherent in
identifying source categories. For example, quite different mobile sources such as trucks
and locomotives rely on diesel engines and ancillary data is often required to resolve these
sources. A compilation of study results shows that secondary S( ) r (mainly from Electric
Generating Units [EGUs]), NOa (from the oxidation of NOx emitted mainly from
transportation sources and EGUs), and primary mobile source categories constitute most of
PM2.5 (and PM10) in the East. PM10-2.5 is mainly a primary pollutant, having been emitted
from pollution sources as fully formed particles derived from abrasion and crushing
processes, soil disturbances, desiccation of marine aerosol emitted from bursting bubbles,
hygroscopic fine PM expanding with humidity to coarse mode, and/or gas condensation
directly onto preexisting coarse particles. Suspended primary coarse PM may contain iron,
silica, aluminum, and base cations from soil, plant and insect fragments, pollen, fungal
spores, bacteria, and viruses, as well as fly ash, brake lining particles, debris, and
automobile tire fragments. Quoted uncertainties in the source apportionment of
constituents in ambient aerosol samples typically range from 10 to 50%. An
intercomparison of source apportionment techniques indicated that the same major source
categories of PM2.5 were consistently identified by several independent groups working with
the same data sets. Soil-, sulfate-, residual oil-, and salt-associated mass were most clearly
identified by the groups. Other sources with more ambiguous signatures, such as vegetative
burning and traffic-related emissions were less consistently identified.
July 2009
2-8
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Spatial variability in source contributions across urban areas is an important
consideration in assessing the likelihood of exposure error in epidemiologic studies relating
health outcomes to sources. Concepts similar to those for using ambient concentrations as
surrogates for personal exposures apply here. Some source attribution studies for PM2.5
indicate that intra-urban variability increases in the following order: regional sources (e.g.,
secondary SO r originating from EGUs) < area sources (e.g., on-road mobile sources)
< point sources (e.g., stacks). Although limited information was available for PM10-2.5, it does
indicate a similar ordering, but without a regional component (resulting from the short
lifetime of PM10-2.5 compared to transport times on the regional scale). More discussion on
source contributions to PM is available in Section 3.6.
2.1.7. Policy-Relevant Background
The background concentrations of PM that are useful for risk and policy assessments,
which inform decisions about the NAAQS are referred to as policy-relevant background
(PRB) concentrations. PRB concentrations have historically been defined by EPA as those
concentrations that would occur in the U.S. in the absence of anthropogenic emissions in
continental North America defined here as the U.S., Canada, and Mexico. For this
document, PRB concentrations include contributions from natural sources everywhere in
the world and from anthropogenic sources outside continental North America. Background
concentrations so defined facilitated separation of pollution that can be controlled by U.S.
regulations or through international agreements with neighboring countries from those
that were judged to be generally uncontrollable by the U.S. Over time, consideration of
potential broader ranging international agreements may lead to alternative determinations
of which PM source contributions should be considered by EPA as part of PRB.
Contributions to PRB concentrations of PM include both primary and secondary natural
and anthropogenic components. For this document, PRB concentrations of PM2.5 for the
continental U.S. were estimated using EPA's Community Multi-scale Air Quality (CMAQ)
modeling system, a deterministic, chemical-transport model (CTM), using output from
GEOS-Chem a global-scale model for CMAQ boundary conditions. PRB concentrations of
PM2.5 were estimated to be less than 1 ug/m3 on an annual basis, with maximum daily
average values in a range from 3.1 to 20 |u,g/m3 and having a peak of 63 ug/m3 at the nine
national park sites across the U.S. used to evaluate model performance for this analysis. A
description of the models and evaluation of their performance is given in Section 3.6 and
further details about the calculations of PRB concentrations is given in Section 3.7.
2.2. Human Exposure
This section summarizes the findings from the recent exposure assessment literature,
which includes the assessment of exposure to ambient PM, infiltration of ambient PM to
July 2009
2-9
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
indoor environments, and source apportionment of exposure. This summary is intended to
support the interpretation of the findings from epidemiologic studies (Section 3.8).
2.2.1. Spatial Scales of PM Exposure Assessment
Assessing population-level exposure at the urban scale is particularly relevant for
time-series epidemiologic studies, which provide information on the relationship between
health effects and community-average exposure, rather than an individual's exposure. PM
concentrations measured at a central-site ambient monitor are used as surrogates for
personal PM exposure. However, the correlation between the PM concentration measured
at central-site ambient monitor(s) and the unknown true community average concentration
depends on the spatial distribution of PM, the location of the monitoring site(s) chosen to
represent the community average, and division of the community by terrain features or
local sources into several sub-communities that differ in the temporal pattern of pollution.
Concentrations of SO r and some components of SOA measured at central-site monitors are
expected to be uniform in urban areas because of the regional nature of their sources.
However, this is not true for primary components like EC whose sources are strongly
spatially variable in urban areas.
At micro-to-neighborhood scales, heterogeneity of sources and topography are sources
of variability in exposure. This is particularly true for PM10-2.5 and for UFPs, which have
spatially variable urban sources and loss processes (mainly gravitational settling for PM10-
2.5 and coagulation for UFPs) that also limit their transport from sources more readily than
for PM2.5. Personal activity patterns also vary across urban areas and across regions. Some
studies, conducted mainly in Europe, have found personal PM2.5 and PM10 exposures for
pedestrians in street canyons to be higher than ambient concentrations measured by urban
central site ambient monitors. Likewise, microenvironmental UFP concentrations were
observed to be substantially higher in near-road environments, street canyons, and tunnels
when compared with background concentrations. In-vehicle UFP and PM2.5 exposures can
also be important. As a result, ambient monitors located at background, central urban, road
side, or near-residential sites might not reflect the contributions in UFP or PM2.5 exposures
to individuals who commute.
There is significant variability within and across regions of the country with respect
to indoor exposures to ambient PM. Infiltrated ambient PM concentrations depend in part
on the ventilation properties of the building or vehicle in which the person is exposed. PM
infiltration factors depend on particle size, chemical composition, season, and region of the
country. Infiltration can best be modeled dynamically rather than being represented by a
single value. Season is important to PM infiltration because it affects the ventilation
practices (e.g., open windows) used. In addition, ambient temperature and humidity
conditions affect the transport, dispersion, and size distribution of PM. Residential air
July 2009
2-10
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
exchange rates have been observed to be higher in the summer for regions with low air
conditioning usage. Regional differences in air exchange rates (Southwest < Southeast
< Northeast < Northwest) also reflect ventilation practices. Differential infiltration occurs
as a function of PM size and composition (the latter of which is described below). PM
infiltration is larger for accumulation mode particles than for UFPs and PM10-2.5.
Differential infiltration by size fraction can affect exposure estimates if not accurately
characterized.
2.2.2. Exposure to PM Components and Copollutants
Emission inventories and source apportionment studies suggest that sources of PM
exposure vary by region. Comparison of studies performed in the eastern U.S. with studies
performed in the western U.S. suggest that the contribution of SC>42~ to exposure is higher
for the East (16-46%) compared with the West (-4%) and that motor vehicle emissions and
secondary NO; are larger sources of exposure for the West (-9%) as compared with the
East (-4%). Results of source apportionment studies of exposure to SO r indicate that SO r
exposures are mainly attributable to ambient sources. Source apportionment for OC and
EC is difficult because they originate from both indoor and outdoor sources. Exposure to OC
of indoor and outdoor origin can be distinguished by the presence of aliphatic (Ml groups
generated indoors, since outdoor concentrations of aliphatic C-H are low. Studies of
personal exposure to ambient trace metal have shown significant variation among cities
and over seasons. This is in response to geographic and seasonal variability in sources
including incinerator operation, fossil fuel combustion, biomass combustion (wildfires), and
the resuspension of crustal materials in the built environment. Differential infiltration is
also affected by variations in particle composition and volatility. For example, EC infiltrates
more readily than OC. This can lead to outdoor-indoor differentials in PM composition.
Some studies have explored the relationship between PM and copollutant gases and
suggested that certain gases can serve as surrogates for describing exposure to other air
pollutants. The findings indicate that ambient concentrations of gaseous copollutants can
act as surrogates for personal exposure to ambient PM. Several studies have concluded that
ambient concentrations of O3, NO2, and SO2 are associated with the ambient component of
personal exposure to total PM2.5 as opposed to the ambient component of personal
exposures to the gases. However, in some of these studies this result may have arisen in
part because personal exposure to the gases was often beneath the detection limits of the
personal monitoring devices. If associations between ambient gases and personal exposure
to PM2.5 of ambient origin exist, such associations are complex and vary by season and
location.
July 2009
2-11
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
2.2.3. Implications for Epidemiologic Studies
In epidemiologic studies, exposure may be estimated using various approaches most
of which rely on measurements obtained using central site monitors. The magnitude and
direction of the biases introduced through error in exposure measurement depend on the
extent to which the error is associated with the measured PM concentration. In general,
when exposure error is not strongly correlated with the measured PM concentration, bias is
toward the null and effect estimates are underestimated. Moreover, lack of information
regarding exposure measurement error can also add uncertainty to the health effects
estimate.
One important factor to be considered is spatial variation in PM concentrations. The
degree of urban-scale spatial variability in PM concentrations varies across the country and
with size fraction. PM2.5 concentrations are relatively well-correlated across monitors in the
urban areas examined for this assessment. The limited available evidence indicates that
there is greater spatial variability in PM10-2.5 concentrations than PM2.5 concentrations,
resulting in increased exposure error for the larger size fraction. Likewise, studies have
shown UFP to be more spatially variable across urban areas compared to PM2.5. Even if
PM2.5, PM10-2.5, or UFP concentrations measured at sites within an urban area are generally
highly correlated, significant spatial variation in their concentrations can occur on any
given day. In addition, there can be differential exposure errors for PM components (e.g.,
sulfates, OC, EC). Current information suggests that UFP, PM10-2.5, and some PM
components are more spatially variable than PM2.5. For these PM indicators their spatial
variability adds uncertainty to exposure estimates.
Overall, recent studies generally confirm and build upon the key conclusions of the
2004 PM AQC I K separation of total PM exposures into ambient and nonambient
components reduces potential uncertainties in the analysis and interpretation of PM health
effects data! and ambient PM concentration can be used as a surrogate for ambient PM
exposure in community time-series epidemiologic studies because the change in ambient
PM concentration should be reflected in the change in the health risk coefficient. The use of
the community average ambient PM2.5 concentration as a surrogate for the community
average personal exposure to ambient PM2.5 is not expected to change the principal
conclusions from time-series and most panel epidemiologic studies that use community
average health and pollution data. Several recent studies support this by showing how the
ambient component of personal exposure to PM2.5 could be estimated using various tracer
and source apportionment techniques and by showing that the ambient component is highly
correlated with ambient concentrations of PM2.5. These studies show that the non-ambient
component of personal exposure to PM2.5 is largely uncorrelated with ambient PM2.5
concentrations. A few panel epidemiologic studies have included personal as well as
ambient monitoring data, and generally reported associations with all types of PM
July 2009
2-12
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
measurements. Epidemiologic studies of long-term exposure typically exploit the differences
in PM concentration across space as well as time to estimate the effect of PM on the health
outcome of interest. Long-term exposure estimates are most accurate for pollutants that do
not vary substantially within the geographical area studied.
2.3. Health Effects
This section evaluates the evidence from toxicological, controlled human exposure,
and epidemiologic studies that examined the health effects associated with short- and
long-term exposure to PM (i.e., PM2.5, PM10-2.5 and UFPs). The results from the health
studies evaluated in combination with the evidence from atmospheric chemistry and
exposure assessment studies contribute to the causal determinations made for the health
outcomes discussed in this assessment (Section 1.5.4). In the following sections a discussion
of the causal determinations will be presented by PM size fraction and exposure duration
(i.e., short- or long-term exposure) for the health effects for which sufficient evidence was
available to conclude a causal, likely to be causal, or suggestive relationship. To the extent
possible, results of PM10 studies were considered in causality determinations for PM2.5 and
PM10-2.5. Although an extensive amount of research has been conducted to examine
PM-related health effects, a limited body of evidence is currently available to examine the
presence or absence of associations between some health outcomes and PM size fractions.
The evaluation of the aforementioned factors together has resulted in evidence that is
inadequate to infer whether a causal relationship exists for the following exposure
durations, PM size fractions, and health categories^
¦	Short-term exposure to UFPs and mortality
¦	Short-term exposure to all PM size fractions and central nervous system (CNS)
effects
¦	Long-term exposure to PM10-2.5 and all health effects and mortality
¦	Long-term exposure to UFPs and all health effects and mortality
Although not presented in depth in this chapter, a detailed discussion of the
underlying evidence used to formulate each causal determination can be found in Chapters
6 and 7.
July 2009
2-13
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
2.3.1. Exposure to PM2.5
2.3.1.1. Effects of Short-Term Exposure to PM2.5
Table 2-1 Summary of causal determinations for short-term exposure to PM2.5.
Size Fraction
Outcome
Causality Determination
PM2.5
Cardiovascular Effects
Causal
Respiratory Effects
Likely to be causal
Mortality
Likely to be causal
Cardiovascular Effects
Epidemiologic studies that examined the effect of PM2.5011 cardiovascular emergency
department (ED) visits and hospital admissions (HA) reported consistent positive
associations (predominantly for ischemic heart disease [IHD] and congestive heart failure
[CHF]), with the majority reporting increases ranging from 0.5 to 3.4% per 10 pg/m3
increase in PM2.5. These effects were observed in study locations with mean1 24-h avg PM2.5
concentrations ranging from 7" 18 pg/m3 (Section 6.2.10), with effects becoming more precise
and consistently positive in locations with mean PM2.5 concentrations of 13 pg/m3 and above
(Figure 2-1). Toxicological studies have provided biologically plausible mechanisms (e.g.,
increased right ventricular pressure and diminished cardiac contractility) for the
associations observed between pm25 and CHF in epidemiologic studies. The largest U.S.-
based multicity study evaluated, Medicare Air Pollution Study (MCAPS), provided evidence
of regional heterogeneity (e.g., the largest excess risks occurred in the Northeast [1.08%])
and seasonal variation (e.g., the largest excess risks occurred during the winter season
[1.49%]) in PM2.5 CVD risk estimates, which is consistent with the null findings of several
single-city studies conducted in the Western U.S. These associations are supported by
multicity epidemiologic studies that observed consistent positive associations between
short-term exposure to PM2.5 and cardiovascular mortality as well as regional and seasonal
variability in risk estimates. The multicity studies evaluated reported consistent, precise
increases in cardiovascular mortality ranging from 0.47 to 0.85% in study locations with
mean 24-h avg PM2.5 concentrations above 13 pg/m3 (Table 6-12).
1 In this context mean represents the arithmetic mean of 24-h average PM concentrations.
July 2009	2-14	DRAFT - DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Controlled human exposure studies have demonstrated PIVh.r,-induced changes in
various measures of cardiovascular function among healthy and health-compromised
adults. The most consistent evidence is for altered vasomotor function following exposure to
diesel exhaust (DE) or concentrated ambient particles (CAPs) with ozone (O3) (Section
6.2.4.2). Although these findings provide biological plausibility for the observations from
epidemiologic studies, the fresh DE used in the controlled human exposure studies
evaluated contains gaseous components (e.g., CO, NO*), and therefore, the possibility that
some of the changes in vasomotor function might be due to gaseous components cannot be
ruled out. Furthermore, the prevalence of UF particles in fresh DE limits the ability to
conclusively attribute the observed effects to either the UF fraction or PM2.5 as a whole. An
evaluation of toxicological studies found evidence for altered vessel tone and microvascular
reactivity, which provide coherence and biological plausibility for the vasomotor effects that
have been observed in both the controlled human exposure and epidemiologic studies
(Section 6.2.4.3). However, most of these toxicological studies exposed animals via
intratracheal instillation (IT) or using relatively high inhalation concentrations.
In addition to the effects observed on vasomotor function, myocardial ischemia has
been observed across disciplines through PM2.5 effects on ST-segment depression, with
toxicological studies providing biological plausibility by demonstrating reduced blood flow
during ischemia (Section 6.2.3). There is also a growing body of evidence from controlled
human exposure and toxicological studies demonstrating PIVkr,-induced changes on
markers of systemic oxidative stress and heart rate variability (HRV) (Section 6.2.1 and
Section 6.2.9). Additional, but inconsistent effects of PM2.5 on BP, blood coagulation
markers, and markers of systemic inflammation have also been reported across disciplines.
Together, the collective evidence from epidemiologic, controlled human exposure, and
toxicological studies is sufficient to conclude that a causal relationship exists between short-
term exposures to PM2.sand cardiovascular effects.
Respiratory Effects
The recent epidemiologic studies evaluated report consistent positive associations
between short-term exposure to PM2.5 and respiratory ED visits and HAs for chronic
obstructive pulmonary disease (COPD) and respiratory infections (Section 6.3). Positive
associations were also observed for asthma ED visits and HAs for adults and children
combined, but effect estimates are imprecise and not consistently positive for children
alone. Most effects were in the range of ~1% to 4% and were observed in study locations
with mean 24-h avg PM2.5 concentrations ranging from 6.1 - 19.2 ug/ma, with effects
becoming more precise and consistently positive in locations with mean PM2.5
concentrations of 13 pg/m3 and above (Figure 2-1). Additionally, multicity epidemiologic
studies observed consistent positive associations between short-term exposure to PM2.5 and
respiratory mortality as well as regional and seasonal variability in risk estimates. The
July 2009
2-15
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
multicity studies evaluated reported consistent, precise increases in respiratory mortality
ranging from 1.67 to 2.20% in study locations with mean 24-h avg PM2.5 concentrations
above 13 |ug/m3 (Table 6-12). Evidence for PM^r,-related respiratory effects was also
observed in panel studies, which indicate associations with respiratory symptoms,
pulmonary function, and pulmonary inflammation among asthmatic children. Although not
consistently observed, some controlled human exposure studies have reported small
decrements in various measures of pulmonary function following controlled exposures to
PM2.5 (Section 6.3.2.2).
Controlled human exposure studies using adult volunteers have demonstrated
increased markers of pulmonary inflammation following exposure to a variety of different
particle types! oxidative responses to DE and WS; and exacerbations of allergic responses
and allergic sensitization following exposure to DE particles (Section 6.3). Toxicological
studies have provided additional support for PIVb r,-related respiratory effects through
inhalation exposures of animals to CAPs, DE, other traffic-related PM and WS. These
studies reported an array of respiratory effects including altered pulmonary function, mild
pulmonary inflammation and injury, oxidative responses, airway hyperresponsiveness
(AHR) in allergic and non-allergic animals, exacerbations of allergic responses, and
increased susceptibility to infections (Section 6.3). Overall, there is limited coherence across
disciplines for the PM2.5"induced respiratory outcomes observed. Epidemiologic studies have
reported variable results among specific respiratory outcomes, specifically in asthmatics
(e.g., increased respiratory symptoms in asthmatic children, but not increased asthma HAs
and ED visits) (Section 6.3.8). Additionally, respiratory effects have not been consistently
demonstrated following controlled exposures to PM2.5 among asthmatics or individuals with
COPD. Collectively, the studies evaluated demonstrate a wide range of respiratory
responses, and although results are not fully consistent and coherent across studies the
evidence is sufficient to conclude that a causal relationship is likely to exist between short-
term exposures to PM2.5 and respiratory effects.
Mortality
An evaluation of the epidemiologic literature indicates consistent positive associations
between short-term exposure to PM2.5 and all-cause, cardiovascular-, and respiratory-
related mortality (Section 6.5.2.2.). The evaluation of multicity studies found that risk
estimates for all-cause (non-accidental) mortality ranged from 0.29% to 1.21% per 10 ug/m3
increase in PM2.5 at lags of 1 and 0-1 days. These consistent, precise effects were observed
in study locations with mean 24-h avg PM2.5 concentrations of 13 ug/m3 and above (Table 6-
12). Cardiovascular-related mortality risk estimates were found to be similar to those for
all-cause mortality whereas, the risk estimates for respiratory-related mortality were
consistently larger: 1.01-2.2% using the same lag periods and averaging indices. Regional
and seasonal patterns in PM2.5 risk estimates were observed with the greatest effect
July 2009
2-16
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
estimates occurring in the eastern U.S. and during the spring. Of the studies evaluated, no
U.S.-based multicity studies conducted a detailed analysis of the potential confounding of
PM2.5 risk estimates by gaseous pollutants but Burnett et al. (2004, 086247) found mixed
results, with possible confounding by NO2 when analyzing gaseous pollutants in a multicity
Canadian-based study (Section 6.5.2.1). However, it should be noted that U.S.-based
multicity studies that focused on the association between PM10 and mortality found that
gaseous pollutants are not likely to confound the PM-mortality relationship. An
examination of effect modifiers (e.g., demographic and socioeconomic factors), specifically
air conditioning use as an indicator for decreased pollutant penetration indoors, has
suggested that PM2.5 risk estimates increase as the percent of the population with access to
air conditioning decreases. Collectively, the epidemiologic literature provides evidence that
a causal relationship is likely to exist between short-term exposures to PM2.5 and mortality.
2.3.1.2. Effects of Long-Term Exposure to PM2.5
Table 2-2. Summary of causal determinations for long-term exposure to PM2.5.
Size Fraction	Outcome	Causality Determination
PM2.5	Cardiovascular Effects	Causal
Respiratory Effects
Likely to be causal
Mortality
Likely to be causal
Reproductive and Developmental
Suggestive
Cancer, Mutagenicity, and Genotoxicity
Suggestive
Cardiovascular Effects
The strongest evidence for CVD health effects related to long-term exposure to PM2.5
comes from large, multicity U.S.-based studies, which provide consistent evidence of an
association between long-term exposure to PM2.5 and cardiovascular mortality (Section
7.2.10). These associations are supported by a large U.S.-based epidemiologic study (i.e.,
Women's Health Initiative [WHI] study) that reports associations between PM2.5 and CVDs
among post-menopausal women using a 1-yr avg PM2.5 concentration (mean = 13.5 |ug/m3)
(Section 7.2). Epidemiologic studies that examined subclinical markers of CVD report
inconsistent findings. In addition, epidemiologic studies have provided some evidence for
potential modification of the PM2.5-CVD association when examining individual-level data,
specifically smoking status and the use of anti-hyperlipidemics. Although epidemiologic
July 2009
2-17
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
studies have not consistently detected effects on markers of atherosclerosis due to long-
term exposure to PM2.5, toxicological studies have provided strong evidence for accelerated
development of atherosclerosis in ApoE '" mice exposed to CAPs and have shown effects on
coagulation, experimentally-induced hypertension, and vascular reactivity (Section 7.2.1.2).
Evidence from toxicological studies provides biological plausibility and coherence with
studies of short-term exposure and CVD morbidity and mortality, as well as with studies
that examined long-term exposure to PM2.5 and CVD mortality. Taken together, the
evidence from epidemiologic and toxicological studies is sufficient to conclude that a causal
relationship exists between long-term exposures to PM2.5 and cardiovascular effects.
Respiratory Effects
Recent epidemiologic studies conducted in the U.S. and abroad provide evidence of
associations between long-term exposure to PM2.5 and decrements in lung function growth,
increased respiratory symptoms, and asthma development in study locations with mean
PM2.5 concentrations ranging from 13.8 to 30 ug/m3 during the study periods (Section 7.3.1.1
and 7.3.2.1), with effects becoming more precise and consistently positive in locations with
mean PM2.5 concentrations of 14 pg/m3 and above (Figure 2-2). These results are supported
by studies that observed associations between long-term exposure to PM10 and an increase
in respiratory symptoms and reductions in lung function growth in areas where PM10 is
dominated by PM2.5. However, the evidence to support an association with long-term
exposure to PM2.5 and respiratory mortality is limited (Figure 7-8). Subchronic and chronic
toxicological studies of CAPs, DE, roadway air and woodsmoke provide coherence and
biological plausibility for the effects observed in the epidemiologic studies. These
toxicological studies have presented some evidence for altered pulmonary function, mild
inflammation, oxidative responses, immune suppression, and histopathological changes
including mucus cell hyperplasia (Section 7.3). Exacerbated allergic responses have been
demonstrated in animals exposed to DE and WS. In addition, pre- and postnatal exposure
to ambient levels of urban particles was found to affect lung development in an animal
model. This finding is important because impaired lung development is one mechanism by
which PM exposure may decrease lung function growth in children. Collectively, the
evidence from epidemiologic and toxicological studies is sufficient to conclude that a causal
relationship is likely to exist between long-term exposures to PM2.5 and respiratory effects.
Mortality
The recent epidemiologic literature reports associations between long-term PM2.5
exposure and increased risk of mortality in areas with mean PM2.5 concentrations during
the study period ranging from 13.2 to 29 jug/m3 (Section 7.6), with effects becoming more
precise and consistently positive in locations with mean PM2.5 concentrations of 13.5 pg/m3
and above (Figure 2-2). When evaluating cause-specific mortality, the strongest evidence
July 2009
2-18
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
can be found when examining associations between PM2.5 and cardiovascular mortality, and
positive associations were also reported between PM2.5 and lung cancer mortality
(Figure 7"8). The cardiovascular mortality association has been confirmed further by the
extended Harvard Six Cities and ACS studies, which both report strong associations
between long-term exposure to PM2.5 and cardiopulmonary and IHD mortality (Figure 7" 7).
The most recent evidence for the association between long-term exposure to PM2.5 and
CVD-mortality is particularly strong for women. Fewer studies evaluate the respiratory
component of cardiopulmonary mortality, and the evidence to support an association with
long-term exposure to PM2.5 and respiratory mortality is limited (Figure 7-8). The evidence
for cardiovascular and respiratory morbidity due to short- and long-term exposure to PM2.5
discussed above provides biological plausibility for cardiovascular- and respiratory-related
mortality. Collectively, the evidence is sufficient to conclude that a causal relationship is
likely to exist between long-term exposures to PM2.5 and mortality.
Reproductive and Developmental
Evidence is accumulating for PM2.5 effects on low birth weight and infant mortality,
especially due to respiratory causes during the post-neonatal period. The mean PM2.5
concentrations during the study periods ranged from 5.3-27.4 ug/ma (Section 7.4), with
effects becoming more precise and consistently positive in locations with mean PM2.5
concentrations of 15 pg/m3 and above (Section 7.4). Exposure to PM2.5 was usually
associated with greater reductions in birth weight than exposure to PM10. The evidence
from a few U.S. studies that investigated PM10 effects on fetal growth, which reported
similar decrements in birth weight, provide consistency for the PM2.5 associations observed
and strengthen the interpretation that particle exposure may be causally related to
reductions in birth weight. The epidemiologic literature does not consistently report
associations between long-term exposure to PM and preterm birth, growth restriction, birth
defects or decreased sperm quality. Toxicological evidence supports an association between
PM2.5 and PM10 exposure and adverse reproductive and developmental outcomes, but
provided little mechanistic information or biological plausibility for an association between
long-term PM exposure and adverse birth outcomes (e.g., low birth weight or infant
mortality). New evidence from animal toxicological studies on heritable mutations is of
great interest, and warrants further investigation. Overall, the epidemiologic and
toxicological evidence is suggestive of a causal relationship between long-term exposures to
PM2.5 and reproductive and developmental outcomes.
Cancer, Mutagenicity, and Genotoxicity
Multiple epidemiologic studies have shown a consistent positive association between
PM2.5 and lung cancer mortality, but studies have generally not reported associations
between PM2.5 and lung cancer incidence (Section 7.5). Animal toxicological studies have
July 2009
2-19
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
examined the potential relationship between PM and cancer, but have not focused on
specific size fractions of PM. Instead they have examined ambient PM, WS, and DEP. A
number of recent studies indicate that ambient urban PM, emissions from wood/biomass
burning, emissions from coal combustion, and gasoline and DE are mutagenic, and that
PAHs are genotoxic. These findings are consistent with earlier studies that concluded that
ambient PM and PM from specific combustion sources are mutagenic and genotoxic and
provide biological plausibility for the results observed in the epidemiologic studies. A
limited number of epidemiologic and toxicological studies examined epigenetic effects, and
demonstrate that PM induces some changes in methylation. However, it has yet to be
determined how these alterations in the genome could influence the initiation and
promotion of cancer. Collectively, the evidence from epidemiologic studies, primarily those
of lung cancer mortality, along with the toxicological studies that show some evidence of the
mutagenic and genotoxic effects of PM is suggestive of a causal relationship between long-
term exposures to PM2.5 and cancer.
2.3.2. Integration of PM2.5 Health Effects
In epidemiologic studies, short-term exposure to PM2.5 is associated with a broad
range of respiratory and cardiovascular effects, as well as mortality. For effects on the
cardiovascular system, the evidence supports the existence of a causal relationship with
short-term PM2.5 exposure! the evidence indicates that a causal relationship is likely to exist
between short-term PM2.5 exposure and effects on the respiratory system. The effect
estimates from U.S. and Canadian-based epidemiologic studies (Figure 2.1) have found
consistent positive associations in study areas with mean PM2.5 concentrations ranging
from 6.1 to 22 jug/m3.
Long-term exposure to PM2.5 has been associated with health outcomes similar to
those found in the short-term exposure studies, specifically for respiratory and
cardiovascular morbidity and mortality. As found for short-term PM2.5 exposure, the
evidence indicates that a causal relationship exists between long-term PM2.5 exposure and
cardiovascular effects and that a causal relationship is likely to exist between long-term
PM2.5 exposure and effects on the respiratory system. Figure 2.2 highlights these findings,
which show a range of health effects and outcomes occurring in studies with long-term
mean PM2.5 concentrations ranging from 10.7 to 29 ug/m3 during the study periods.
Additional studies not included in this figure that focus on subclinical outcomes, such as
changes in lung function or atherosclerotic markers also report effects in areas with similar
concentrations (Sections 7.2 and 7.3). The long-term exposure studies provide additional
evidence for reproductive and developmental effects (i.e., LBW) and cancer (i.e., lung cancer
mortality) in response to PM2.5.
July 2009
2-20
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Outcome
Mean; 98,n
l/uglm3)
Effect Estimates
Chimonas & Gessne (2007. 093261)
Asthma
61
NR

A. 1


Lower Respiratory Infection
61
NR

A 1

Lisabeth et al. (2008,155939)
Ischemic Stroke/TIA
7*-
NR



Slauqhter et al. (2005, 073854)
Asthma Exacerbation
73
NR




Chen et al. (2004, 089899)
COPD
7 7
NR




Chen et al. (2005, 087942)
Respiratory Disease
77
NR




Funq et al. (2006, 089789)
Respiratory Disease
7.7
NR



Rich et al. (2004,055631)
ICD Shock
8 7
NR

1

Villeneuve et al. (2006, 090191)
Hemmhoraqic Stroke: Cool Seas.
8.5
NR

T



Hemmhoraqic Stroke: Warm Seas.
8.5
NR

i

	>

Ischemic Stroke: Cool Seas.
8.5
NR

	m	I	



Ischemic Stroke: Warm Seas.
8.5
NR




TIA: Cool Seas.
8.5
NR
<	



TIA: Warm Seas.
8.5
NR



Lin et al. (2005, 087828)
Respiratory Tract Infection
96
NR




Mar et al. (2004, 057309)
Asthma Symptoms
8.1-11.0+; NR
/ m



Dockerv et al. (2005, 078995)
Ventricular Arrhythmia
10.3*; NR




Mar et al. (2004, 057309)
Symptoms (Any)
10.3; NR


A

Pope et al. (2006, 091246)
IHD
10.7: NR




Rabinovitch et al. .(2006, 088031)
Asthma Medication Use
10
8; 23.4



Rabinovitch et al. (2004, 096753)
Asthma Exacerbation
10.
8; 29.3
<	
—A	1	

Pope et al. (2008,191969)
CHF
114
NR


-%	
Zanobetti & Schwartz (2009,188462)
Ml
13.3; 34.3




Pneumonia
13.3; 34.3




Ml
12.1:28.2

I ¦

Slaughter et al. (2005, 073854)
CVD
12.2: NR

i


Asthma
12.2: NR




COPD
12.2: NR

•i


Respiratory Disease
12.2: NR



Sullivan et al. (2005, 050854)
Ml
12
S; NR



Sullivan et al. (2003, 043156)
Cardiac Arrest
12
3; NR



Zanobetti & Schwartz (2009,188462)
CV Mortality
13.2; 34.33

i#


Respiratory Mortality
13.2: 34.33



Dominici et al. (2006, 088398)
CVD
13.4: NR




PVD
13.4: NR

c


IHD
13.4: NR

b


Dysrhythmia
13.4: NR

b-


CHF
13.4: NR

i m-


COPD
13.4; NR




Respiratory Tract Infection
13.4; NR



Bell et al. (2008,156266)
Respiratory Disease
13.4:34.16

p


CVD
13.4:34.16

.i

Zhang et al. (2009,191970)
ST Seqment Depression
13.9; NR



O'Connor et al. (2008,156818)
WheezelCouqh
14; NR

4 1 *

Itoetal. (2007,091262)
Asthma
15; NR

1 u

Delfino et al. (1997,082687)
Respiratory Disease
15.4: NR

i _

Symons et al. (2006, 091258)
CHF
16-
NR



Rich et al. (2006,089814)
Ventricular Arrhythmia
16.2*; NR
4*


Metzqeret al. (2007, 092856)
Ventricular Arrhythmia
16.2*; NR



Sheppard (2003, 042826)
Asthma
16.7:46.6



Burnett et al. (1997, 082676)
Respiratory Disease
16
!; NR



Gent et al. (2009,180399)
Wheeze
17;
NR

	JL	


Persistent Couqh
1/;
NR

	I	


Shortness of Breath
1/;
NR

	1	A	


Medication Use
1/;
NR

	A	

Tolbert et al. (2007,090316)
Respiratory Disease
1/.
I; 38.7




CVD
1/
I; 38.7



Metzqeret al. (2004, 044222)
CVD
17.
!*; 39.8

i —•—


CHF
M.
!*; 39.8

l	~	


IHD
17.
B*; 39.8

I •	


PVD CVD
17.
B*; 39.8

:—•—

Lin et al. (2002, 026067)
Asthma: Boys
17.99; NR

	A	


Asthma: Girls
17.99: NR

	1—~	

Burnett et al. (1999, 017269)
IHD
18;
55.2

1 —•—

Ito (2003, 042856)
CHF
18;
55.2



Lippmann (2000, 024579)
Dysrhythmia
18;
55.2




IHD
18;
55.2

—!-¦—


Stroke
18;
55.2

	P	

Ito (2003, 042856)
COPD
18-
NR

—i—'

Lippmann (2000, 024579)
Pneumonia
18;
NR

T—¦—


Thurston et al. (1994, 043921)
Respiratory Disease
18.6; NR

1 •


Ostro et al. (2009,191971)
Respiratory Disease
19;
61.3

1 A
+ Range of Means

Peel et al. (2005, 056305)
Asthma
19.2:39.8

—m—
5 Estimated


COPD
19.2:39.8

	f	
~ Children


Pneumonia
19.2:39.8

—r#	
• All Ages


URI
19.2:39.8

-f-9—
¦ 65 +

Moolqavkar et al. (2003, 042864)
COPD
??*
; NR

¦

Ostro et al. (2006, 087991)
CV Mortality
19.
88; 68.2

!•


Respiratory Mortality
19.
88; 68.2

. . . i . . . 	

All effect estimates have been standardized to reflect a TO/yg/rrFincrease in mean PM2.5	' ' 9^ ' Jo ' 1 1)4 ' 1 ^8 ' 1 2 ' ' 1 *18 ' ' 1 ^4 ' "p28
concentration.	"	.....	.	.	.
Figure 2-1. Excess risk estimates from epidemiologic studies of PM2.5 ordered by mean 24-h avg
concentration as reported by the investigator.
July 2009
2-21
DRAFT-DO NOT CITE OR QUOTE

-------


Mean

Study
Endpoint
Concentration
Effect Estimate


l/jglm'l

Zegeret al. (2008,191951)
All Cause Mortality
10.7
¦ Central U.S.
i
Kim et al. (2004, 087383)
Bronchitis (Children)
12
i
i ,

i
Zegeret al. (2008,191951)
All Cause Mortality
13.1
J Western U.S.
-r
Miller et al. (2007, 090130)
CVD Morbidity or Mortality
13.5
i

1
Eftim et al. (2008,099104)
All Cause Mortality
13.6
i ACS Sites
i
Goss et al. (2004, 055624)
All CauseMortality
13.7
i

i
McConnell et al. (2003, 049490)
Bronchitis (Children)
13.8
i

i •
Zegeret al. (2008,191951)
All Cause Mortality
14.0
i Eastern U.S.
i
Krewski et al. (2009,191193)
All Cause Mortality
14.0
¦
¦
Krewski et al. (2009,191193)
IHD Mortality
14.0
i


Krewski et al. (2009,191193)
Lung Cancer Mortality
14.0
i

i "
Eftim et al. (2008,099104)
All Cause Mortality
14.1
i Six Cities Study Sites

i
Lipfert et al. (2006, 088756)
All Cause Mortality
14.3
i

i
Dockerv et al. (1996, 077269)
Bronchitis (Children)
14.5
i m

i
Woodruff et al. (2008, 098386)
Infant Mortality (Respiratory)
14.8
i
i •
Laden et al. (2006, 087605)
All Cause Mortality
16
¦
i 	*
i
Laden et al. (2006, 087605)
CVD Mortality
16
i
i
Laden et al. (2006, 087605)
Lung Cancer Mortality
16
i
i
Woodruff et al. (2008, 098386)
Infant Mortality (Respiratory)
19.2
i
1 •
i
Enstrom (2005, 087356)
All Cause Mortality
23.4
i
i
Chen et al. (2005, 087942)
CHD Mortality
29.0
Females
i
Chen et al. (2005, 087942)
CHD Mortality
29.0
i Males
1
2
3
4
5
l	1——l	1	1	1	1	1
0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1
Figure 2-2. Summary of U.S. studies examining the association between long-term exposure to
PM2.5 and CVD morbidity/mortality, respiratory morbidity/mortality, and all-cause
mortality conducted in locations where the mean annual PM2.5 concentration ranged
from 10.7-29 //g/m3. All effect estimates have been standardized to reflect a 10yL/g/m3
increase in mean annual PM2.5 concentration.
The observations from both the short- and long-term exposure studies are supported
by experimental findings of PM2.5"induced subclinical and clinical cardiovascular effects.
Epidemiologic studies have shown an increase in ED visits and HAs for IHD upon exposure
to PM2.5. These effects are coherent with the changes in vasomotor function and STsegment
depression observed in both toxicological and controlled human exposure studies. It has
July 2009	2-22	DRAFT - DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
been postulated that exposure to PM2.5 can lead to myocardial ischemia through an effect on
the autonomic nervous system or by altering vasomotor function. PM-induced systemic
inflammation, oxidative stress and/or endothelial dysfunction may contribute to altered
vasomotor function. These effects have been demonstrated in recent animal toxicological
studies, along with altered microvascular reactivity, altered vessel tone, and reduced blood
flow during ischemia. Toxicological studies demonstrating increased right ventricular
pressure and diminished cardiac contractility also provide biological plausibility for the
associations observed between PM2.5 and CHF in epidemiologic studies.
Thus, the overall evidence from the short-term epidemiologic, controlled human
exposure, and toxicological studies evaluated provide coherence and biological plausibility
for cardiovascular effects related to myocardial ischemia and congestive heart failure.
Coherence in the cardiovascular effects observed can be found in long-term exposure
studies, especially for CVDs among post-menopausal women. Additional studies provide
limited evidence for subclinical measures of atherosclerosis in epidemiologic studies with
stronger evidence from toxicological studies that have demonstrated accelerated
development of atherosclerosis in ApoE"'" mice exposed to PM2.5 CAPs along with effects on
coagulation, experimentally-induced hypertension, and vascular reactivity. Repeated acute
responses to PM may lead to cumulative effects that manifest as chronic disease, such as
atherosclerosis. Contributing factors to atherosclerosis development include systemic
inflammation, endothelial dysfunction, and oxidative stress all of which are associated with
PM2.5 exposure. However, it has not yet been determined whether PM initiates or promotes
atherosclerosis. Collectively the evidence from both short- and long-term exposure studies
on cardiovascular morbidity is consistent with the cardiovascular mortality effects observed
when examining both exposure durations. In addition, CVD HA and mortality studies that
examined the PM10 concentration-response relationship found evidence of a log-linear no-
threshold relationship between PM exposure and CVD-related morbidity (Section 6.2) and
mortality (Section 6.5).
Epidemiologic studies have also reported respiratory effects related to short-term
exposure to PM2.5, which include increased ED visits and HAs, as well as alterations in lung
function and respiratory symptoms in asthmatic children. These respiratory effects were
found to be generally robust to the inclusion of gaseous pollutants in copollutant models
with the strongest evidence from the higher powered studies (Figures 6-9 and 6-15). The
respiratory effects observed in epidemiologic studies are consistent with those from studies
that examined the association between short-term exposure to PM2.5 and respiratory
mortality. However, the effect of gaseous pollutants on PM2.5 respiratory mortality risk
estimates has not been extensively examined in PM2.5 mortality studies. Important new
findings, which support the PM2.5"induced respiratory effects mentioned above, include
associations with post-neonatal (between 1 month and 1 year of age) mortality. Controlled
human exposure studies provide some support for the respiratory findings from
July 2009
2-23
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
epidemiologic studies, with demonstrated increases in pulmonary inflammation following
short-term exposure. However, there is limited and inconsistent evidence of effects in
response to controlled exposures to PM2.5 on respiratory symptoms or pulmonary function
among healthy adults or adults with respiratory disease. Long-term exposure epidemiologic
studies provide additional evidence for PM2.5"induced respiratory morbidity, but little
evidence for an association with respiratory mortality. These epidemiologic morbidity
studies have found decrements in lung function growth, as well as increased respiratory
symptoms, and asthma. Toxicological studies provide coherence and biological plausibility
for the respiratory effects observed in response to short and long-term exposures to PM by
demonstrating a wide array of biological responses including: altered pulmonary function,
mild pulmonary inflammation and injury, oxidative responses, and histopathological
changes in animals exposed by inhalation to PM2.5 derived from a wide variety of sources.
In some cases, prolonged exposures led to adaptive responses. Important evidence was also
found in an animal model for altered lung development following perinatal exposure to
urban air, which may provide a mechanism to explain the reduction in lung function growth
observed in children in response to long-term exposure to PM.
Additional respiratory-related effects have been tied to allergic responses.
Epidemiologic studies have provided evidence for increased HAs for allergic symptoms (e.g.,
allergic rhinitis) in response to short- and long-term exposure to PM2.5. Panel studies also
positively associate long-term exposure to PM2.5 and PM10 with indicators of allergic
sensitization. Controlled human exposure and toxicological studies provide coherence for
the exacerbation of allergic symptoms, by showing that PM2.5 can promote allergic
responses and intensify existing allergies. Allergic responses require repeated exposures to
antigen over time and co-exposure to an adjuvant (possibly DEP or ultrafine CAPs) can
enhance this response. Allergic sensitization often underlies allergic asthma, characterized
by inflammation and AHR. In this way, repeated or chronic exposures involving
multifactorial responses (immune system activation, oxidative stress, inflammation) can
lead to irreversible outcomes. Epidemiologic studies have also reported evidence for
increased HAs for respiratory infections in response to both short- and long-term exposures
to PM2.5. Toxicological studies suggest that PM impairs innate immunity, which is the first
line of defense against infection, providing coherence for the respiratory infection effects
observed in epidemiologic studies.
The difference in effects observed across studies and between cities may be attributed,
at least in part, to the differences in PM composition across the U.S.; however, the available
evidence and the limited amount of city-specific speciated PM2.5 data does not allow
conclusions to be drawn that specifically differentiate effects of PM in different locations.
Regional differences in PM2.5 composition are outlined briefly in Section 2.1 above and in
Section 3.5 in more detail. Although PM2.5 is produced mainly by the combustion of fossil
fuels, either by stationary sources or by transportation, there is a large degree of regional
July 2009
2-24
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
variability in PM mass concentration, composition, and sources. It remains a challenge to
determine relationships between specific constituents, combination of constituents, or
sources of PM2.5 and the various health effects observed. Source apportionment studies of
PM2.5 have attempted to decipher some of these relationships and in the process have
identified associations between multiple sources and various respiratory and cardiovascular
health effects, as well as mortality. Although different source apportionment methods have
been used across these studies, the methods used have been validated and found to
generally identify the same sources and associations between sources and health effects
(Section 6.6). While uncertainty remains, it has been recognized that many components of
PM2.5may contribute to health effects. Overall, the results displayed in Table 6-17 indicate
that many constituents of PM2.5 can be linked with multiple health effects and the evidence
is not yet sufficient to allow differentiation of those constituents or sources that are more
closely related to specific health outcomes.
Variability in the associations observed, specifically across epidemiologic studies may
be due in part to exposure error related to the use of air quality data at the county level.
Because western U.S. counties tend to be much larger than eastern U.S. counties and more
diverse in topography and population characteristics, the day-to-day variations in
concentration at one site, or even for the average of several sites, may not correlate well
with the day-to-day variations in all parts of the county. For example, site-to-site
correlations as a function of distance between sites (Section 3.5.1.2) fall off rapidly with
distance in Los Angeles, but high correlations extend to larger distances in cities such as
Boston and Pittsburgh. These differences cannot be attributed solely to topographic
differences between East and West because some eastern cities (e.g., Pittsburgh) are located
in complex topography. Regional differences in climate which can lead to more time
outdoors or indoors along with more or less air conditioning, and housing stock (e.g., new
homes tend to be tighter with lower infiltration ratios than older homes) may also cause
regional differences in effect estimates. Overall, the various differences between eastern
and western U.S. counties can result in exposure misclassification and an underestimation
of effects in western counties.
The new evidence reviewed in this ISA greatly expands upon the evidence available in
the 2004 PM AQCD particularly in providing greater understanding of the underlying
mechanisms for PM2.5 induced cardiovascular and respiratory effects for both short- and
long-term exposures. Recent studies have provided new evidence linking long-term
exposure to PM2.5 with cardiovascular outcomes that has expanded upon the continuum of
effects ranging from the more subtle subclinical measures to cardiopulmonary mortality.
July 2009
2-25
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
2.3.3. Exposure to PM10 2.5
2.3.3.1. Effects of Short-Term Exposure to PM102.5
Table 2-3. Summary of causal determinations for short-term exposure to PM10 2.5.
Size Fraction
Outcome
Causality Determination
PM10-2.5
Cardiovascular Effects
Suggestive
Respiratory Effects
Suggestive
Mortality
Suggestive
Cardiovascular Effects
Of the epidemiologic studies evaluated, generally positive associations were reported
between short-term exposure to PM10-2.5 and HAs or ED visits for CVDs. These results are
supported by a large U.S. multicity study of older adults that reported PM10-2.5 associations
with CVD HAs, and only a slight reduction in the PM10-2.5 risk estimate when included in a
copollutant model with PM2.5 (Section 6.2.10). The PM10-2.5 associations with cardiovascular
HAs and ED visits were observed in study locations with mean 24-h avg PM10-2.5
concentrations ranging from 7.4 to 13 |u,g/m3, with effects becoming more precise and
consistently positive in locations with mean PM10-2.5 concentrations of 10 pg/m3 and above
(Figure 2-3). These results are supported by the associations observed between PM10-2.5 and
cardiovascular mortality in areas with similar 24-h avg PM10-2.5 concentrations ranging
from 6.1-16.4 jug/m3 (Section 6.2.11), with effects becoming more precise and consistently
positive in locations with mean PM10-2.5 concentrations of 12 pg/m3 and above (Figure 2-3).
The results of the epidemiologic studies were further confirmed by studies that examined
dust storm events, which contain high concentrations of crustal material, and found an
increase in CVD ED visits and HAs. Other epidemiologic studies have reported PM10-2.5
associations with other cardiovascular health effects including supraventricular ectopy and
changes in HRV (6.2.1.1). Although limited in number, studies of controlled human
exposures provide some evidence to support the alterations in HRV observed in the
epidemiologic studies (Section 6.2.1.2). The few toxicological studies that examined the
effect of PM10-2.5 on cardiovascular health effects used IT instillation due to the technical
challenges in exposing rodents via inhalation to PM10-2.5, and, as a result, provide only
limited evidence on the biological plausibility of PM10-2.5 induced cardiovascular effects. The
potential for coarse particles to elicit an effect is supported by dosimetric studies, which
July 2009	2-26	DRAFT - DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
show that a large proportion of inhaled coarse particles in the 3"6 micron (dae) range can
reach and deposit in the lower respiratory tract, particularly the tracheobronchial (TB)
airways (Figures 4-3 and 4-4). Collectively, the evidence from epidemiologic studies, along
with the more limited evidence from controlled human exposure and toxicological studies is
suggestive of a causal relationship between short-term exposures to PM10 2.5 and cardiovascular
effects.
Respiratory Effects
A number of recent epidemiologic studies conducted in Canada and France found
consistent, positive associations between respiratory ED visits and hospital admissions and
short-term exposure to PMio-2.5in studies with mean 24-h avg concentrations ranging from
5.6-13.5 pg/m3 (Section 6.3.8), with effects becoming more precise and consistently positive
in locations with mean PM10-2.5 concentrations of 10 pg/m3 and above (Figure 2-3). In these
studies, the strongest relationships were observed among children, with less consistent
evidence for adults and older adults (i. e., > 65). In a large multicity study of older adults,
PM10-2.5 was positively associated with respiratory HAs in both single and copollutant
models with PM2.5. In addition, a U.S.-based multicity study found evidence for an increase
in respiratory mortality upon short-term exposure to PM10-2.5, but these associations have
not been consistently observed in single-city studies (Section 6.3.9). A limited number of
epidemiologic studies have focused on specific respiratory morbidity outcomes, and found no
evidence of an association with lower respiratory symptoms, wheeze, and medication use
(Section 6.3.1.1). While controlled human exposure studies have not observed an effect on
lung function or respiratory symptoms in healthy or asthmatic adults in response to
exposure to PM10-2.5, healthy volunteers have exhibited an increase in markers of
pulmonary inflammation. Toxicological studies using inhalation exposures are still lacking,
but pulmonary injury has been observed in animals after IT exposure (Section 6.3.5.3). In
some cases, PM10-2.5 was found to be more potent than PM2.5 and effects were not
attributable to endotoxin. Both rural and urban PM10-2.5 have induced inflammation and
injury responses in rats or mice exposed via IT instillation, making it difficult to distinguish
effects of PM10-2.5 from different environments. Overall, epidemiologic studies, along with
the limited number of controlled human exposure and toxicological studies that examined
PM10-2.5 respiratory effects provide evidence that is suggestive of a causal relationship
between short-term exposures to PM10 2.5 and respiratory effects.
Mortality
The majority of studies evaluated in this review provide some evidence for mortality
associations with PM10-2.5. Anew U.S.-based multicity study, which estimated PM10-2.5 by
calculating the difference between the county-average PM10 and PM2.5, found associations
July 2009
2-27
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
between PM10-2.5 and mortality across the U.S., including evidence for regional variability in
PM10-2.5 risk estimates (Section 6.5.2.3). Additionally the U.S.-based multicity study
provides preliminary evidence for greater effects during the warmer months (i.e., spring
and summer). A multicity Canadian study provides additional evidence for an association
between short-term exposure to PM10-2.5 and mortality (Section 6.5.2.3). Uncertainty
surrounds the PM10-2.5 associations reported in the studies evaluated due to the limited
number of PM10-2.5 studies that have investigated confounding by gaseous copollutants or
the influence of model specification on PM10-2.5 risk estimates. Overall, the consistent
positive association between short-term exposure to PM10-2.5 and mortality observed in the
U.S. and Canadian-based multicity studies, along with the positive associations from
single-city studies conducted in these locations, provides evidence that is suggestive of a
causal relationship between short-term exposures to PM10 2.5 and mortality.
2.3.4. Integration of PM10 2.5 Effects
Epidemiologic, controlled human exposure, and toxicological studies have provided
evidence that is suggestive for relationships between short-term exposure to PM10-2.5 and
cardiovascular effects, respiratory effects, and mortality. Conclusions regarding causation
for the various health effects and outcomes were made for PM10-2.5 as a whole regardless of
origin, since PMio-2 5-related effects have been demonstrated for a number of different
environments. These effects have been observed in locations with mean PM10-2.5
concentrations ranging from 5.6 to 13 ug/m3 (Figure 2-3). To date, a sufficient amount of
evidence does not exist in order to draw conclusions regarding the health effects and
outcomes associated with long-term exposure to PM10-2.5.
In epidemiologic studies, associations between short-term exposure to PM10-2.5 and
cardiovascular outcomes (i.e., IHD HAs, supraventricular ectopy, and changes in HRV) have
been found that are similar in magnitude to those observed in PM2.5 studies. Controlled
human exposure studies have also observed alterations in HRV, providing consistency and
coherence for the effects observed in the epidemiologic studies. To date, only a limited
number of toxicological studies have been conducted to examine the effects of PM10-2.5 on
cardiovascular outcomes. All of these studies involved IT instillation due to the technical
challenges of using PM10-2.5 for rodent inhalation studies. As a result, the toxicological
studies evaluated provide limited biological plausibility for the PM10-2.5 effects observed in
the epidemiologic and controlled human exposure studies.
July 2009
2-28
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Outcome
Mean; Max*
Effect Estimates
Chen et al. (2004,147143)
Fung et al. (2006, 089789)
Chen et al. (2005, 087942)
Yang et al. (2004, 087488)
Peters et al. (2001, 016546)
Metzgeret al. (2007, 092856)
Tolbert et al. (2007,090316)
Metzger et al. (2004, 044222)
Peel et al. (2005, 056305)
Maret al. (2004, 057309)
Lin et al. (2005, 087828)
Zanobetti & Schwartz (2009,188462)
Lin et al. (2002, 026067)
Burnett et al. (1999, 017269)
Ito (2003, 042856);
Lippmann (2000, 024579)
Sheppard (2003, 042826)
COPD
COPD
RD
RD
RD
Ml
Ventricular Arrhythmia
CVD
RD
CHF
IHD
Asthma
COPD
RD
Pneumonia
URI
5.6; 24.6
5.6; 24.6
5.6; 27.07
5.6; 24.6
5.6; 24.6
7.4; NR
9.6; 50.3
9; 50.3
9; 50.3
9.1**; NR
9.1**; NR
9.7
9.7
9.7
9.7
9.7
NR
NR
NR
NR
NR
Peng et al. (2008,156850)
RD
9.8*
*; NR
i
#¦

CVD
9.8*
*; NR
i
¦
Symptoms (Any)
Asthma Symptoms
10.8; NR
10.8; NR «-
Slaughter et al. (2005, 073854)
Asthma
NR; NR


COPD
NR; NR
	i.	

RD
NR; NR
1
RTI
Respiratory Mortality
CVD Mortality
Asthma- Boys
Asthma- Girls
CHF
IHD
CHF
IHD
COPD
Pneumonia
Asthma
10.86; 45
11.79; 88.32
11.79; 88.32
12.17:68.00
12.17:68.00
12.2
12.2
13.3
13.3
13.3
13.3
16.2
—A—j.
*//g(m3
** Median
~ - Children
• - All Ages
¦ - Older Adults
I	1	1	1	1	1	1	1	1	1	1	1
-12 -6 -4 0 4 8 t2 16 20 24 28 32
Figure 2-3.
Effect estimates from epidemiologic studies of PM10 2.5 ordered by mean 24-h avg
concentration as reported by the investigator.
1	Limited evidence is available from epidemiologic studies for respiratory health effects
2	and outcomes in response to short-term exposure to PM10-2.5. An increase in respiratory HA
3	and ED visits has been observed, but primarily in studies conducted in Canada and Europe.
4	In addition, associations are not reported for lower respiratory symptoms, wheeze, or
5	medication use. Controlled human exposure studies have not observed an effect on lung
6	function or respiratory symptoms in healthy or asthmatic adults, but healthy volunteers
7	have exhibited pulmonary inflammation. The toxicological studies (all IT) provide evidence
July 2009
2-29
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
of pulmonary injury and inflammation. In some cases, PM10-2.5 was found to be more potent
than PM2.5 and effects were not solely attributable to endotoxin.
Currently a national network is not in place to monitor PM10-2.5 concentrations. As a
result, uncertainties surround the concentration at which the observed associations occur.
Ambient concentrations of PM10-2.5 are generally determined by the subtraction of PM10 and
PM2.5 measurements, using various methods. For example, some epidemiologic studies
estimate PM10-2.5 by taking the difference between collocated PM10 and PM2.5 monitors while
other studies have taken the difference between county average PM10 and PM2.5
concentrations. Therefore, there is greater error in ambient exposure to PM10-2.5 compared to
PM2.5. This would tend to increase uncertainty and make it more difficult to detect effects of
PM10-2.5 in epidemiologic studies. In addition, the various differences between eastern and
western U.S. counties can lead to exposure misclassification, and the potential
underestimation of effects in western counties (as discussed for PM2.5 in Section 2.3.2).
It is also important to note that the chemical composition of PM10-2.5 can vary
considerably by location, but city-specific speciated PM10-2.5 data are limited. However, PM10-
2.5 may contain iron, silica, aluminum, and base cations from soil, plant and insect
fragments, pollen, fungal spores, bacteria, and viruses, as well as fly ash, brake lining
particles, debris, and automobile tire fragments.
The 2004 PM AQCD presented the limited amount of evidence available that
examined the potential association between exposure to PM10-2.5 and health effects and
outcomes. The current evidence, primarily from epidemiologic studies, builds upon the
results from the 2004 PM AQCD and indicates that short-term exposure to PM10-2.5 is
associated with effects on both the cardiovascular and respiratory systems. However,
variability in the chemical and biological composition of PM10-2.5, limited evidence regarding
effects of the various components of PM10-2.5, and lack of clearly defined biological
mechanisms for PMio-2 5-related effects are important sources of uncertainty.
July 2009
2-30
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
2.3.5. Exposure to Ultrafine PM
2.3.5.1. Effects of Short-Term Exposure to UFPs
Table 2-4. Summary of causal determinations for short-term exposure to UFPs.
Size Fraction
Outcome
Causality Determination
Ultrafine Particles
Cardiovascular Effects
Suggestive
Respiratory Effects
Suggestive
Cardiovascular Effects
Controlled human exposure studies provide the majority of the evidence for
cardiovascular health effects in response to short-term exposure to UFP. While there are a
limited number of studies that have examined the association between UFP and
cardiovasc.ular morbidity, there is a larger body of evidence from studies that exposed
subjects to fresh DE, which is typically dominated by UFPs. These studies have
consistently demonstrated effects on vasomotor function (Section 6.2.4.2). Markers of
systemic oxidative stress have also been observed to increase after exposure to various
particle types that are predominantly in the UFP size range. In addition controlled human
exposure studies have observed alterations in HRV parameters in response to exposure to
ultrafine CAPs, with inconsistent evidence for changes in markers of blood coagulation
following exposure to ultrafine CAPs and DE (Sections 6.2.1.2 and 6.2.8.2). A few
toxicological studies have also observed consistent changes in vasomotor function, which
provides coherence with the effects demonstrated in the controlled human exposure studies
(Section 6.2.4.3). Additional UFP-induced effects observed in toxicological studies include
alterations in HRY with less consistent effects observed for systemic inflammation and
blood coagulation. Only a few epidemiologic studies have examined the effect of UFP on
cardiovascular morbidity and collectively they found inconsistent evidence for an
association between UFP and CVD hospital admissions, but some positive associations for
subclinical measures of CVD (i.e., arrhythmias and supraventricular beats) (Section
6.2.2.1). These studies were conducted in the U.S. and Europe at mean particle number
concentration ranges of-8,500-36,000 particles/cm3. However, UFP number concentrations
are highly dependent on monitor location (i.e., concentrations drop off quickly from the road
compared to accumulation mode particles), and therefore, more subject to exposure error
than accumulation mode particles. In conclusion, the evidence from the studies evaluated is
July 2009	2-31	DRAFT - DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
suggestive of a causal relationship between short-term exposures to UFP and cardiovascular
effects.
Respiratory Effects
A limited number of epidemiologic studies have examined the potential association
between short-term exposure to UFP and respiratory morbidity. Of the studies evaluated,
there is limited, and predominately inconsistent evidence for an association between short-
term exposure to UFPs and respiratory symptoms, as well as asthma hospital admissions
at a median particle number concentration of-6,200 to a mean of 38,000 particles/cm3
(Section 6.3.8). The spatial and temporal variability of UFPs also affects these associations.
Although controlled human exposure studies have not extensively examined the effect of
UFP on respiratory outcomes, a few studies have observed small UFP-induced decreases in
pulmonary function. However, these studies have not reported an increase in respiratory
symptoms and the observed effects on pulmonary inflammation are not consistent.
Toxicological studies have also reported mixed results when examining the effect of UFP on
respiratory effects, but several studies demonstrate oxidative, inflammatory, and allergic
responses (Section 6.3). Some effects, such as inflammation or pulmonary histopathology,
are only observed when using particular animal models (e.g., immature or compromised).
Additionally, although a number of controlled human exposure and toxicological studies
that used controlled exposures to fresh DE report respiratory effects, the relative
contributions of gaseous copollutants to the health effects observed remain unresolved.
Thus, the current collective evidence is suggestive of a causal relationship between short-term
exposures to UFP and respiratory effects.
2.3.6. Integration of UFP Effects
The controlled human exposure studies evaluated have consistently demonstrated
effects on vasomotor function and systemic oxidative stress with additional evidence for
alterations in HRV parameters in response to exposure to ultrafine CAPs. The toxicological
studies provide coherence for the changes in vasomotor function observed in the controlled
human exposure studies. Epidemiologic studies are limited because a national network is
not in place to measure UFP in the U.S. UFP concentrations are spatially variable, which
would increase uncertainty and make it difficult to detect associations between health
effects and UFPs in epidemiologic studies. In addition, data on the composition of UFPs and
potential effects of UFP constituents are sparse.
More limited evidence is available regarding the effect of UFP on respiratory effects.
Controlled human exposure studies have not extensively examined the effect of UFPs on
respiratory measurements, but a few studies have observed small decrements in pulmonary
function. Additional effects including oxidative, inflammatory, and pro-allergic outcomes
July 2009
2-32
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
have been demonstrated in toxicological studies, but the lack of coherence with the
controlled human exposure studies limits the interpretation of these findings.
Overall, a limited number of studies have examined the association between exposure
to UFP and morbidity and mortality. Of the studies evaluated, controlled human exposure
studies provide the most evidence for UFP-induced cardiovascular and respiratory effects!
however, these studies focus on exposure to DE. As a result, it is unclear if the effects
observed are due to UFP, larger particles (i.e., PM2.5), or the gaseous components of DE.
Additionally, ultrafine CAPs systems are limited as the atmospheric ultrafine PM
composition is modified when concentrated, which adds uncertainty to the health effects
observed in controlled human exposure studies (Chapter l).
2.4. Policy Relevant Considerations
2.4.1. Potentially Susceptible Subpopulations
Upon evaluating the association between short- and long-term exposure to PM and
various health outcomes, studies also attempted to identify subpopulations that are more
susceptible to PM (i.e., populations that have a greater likelihood of experiencing health
effects related to PM exposure). These studies did so by: conducting stratified analyses!
examining individuals with an underlying health condition! or developing animal models
that mimic the physiological conditions associated with an adverse health effect. These
studies identified a multitude of factors that could potentially contribute to whether an
individual is susceptible to PM (Table 8-2). Although studies have primarily used exposures
to PMa.sor PM10, the available evidence suggests that the identified factors may also
enhance susceptibility to PM10-2.5. The examination of susceptible subpopulations to PM
exposure allows for the NAAQS to provide an adequate margin of safety for both the
general population and for sensitive subpopulations.
During specific periods of life (i.e., childhood and advanced age), individuals maybe
more susceptible to environmental exposures, which in turn can render them more
susceptible to PM-related health effects. An evaluation of age-related health effects
suggests that older adults have heightened responses for cardiovascular morbidity with PM
exposure. In addition, epidemiologic and toxicological studies provide evidence that
indicates children are at an increased risk of PM-related respiratory effects. It should be
noted that the health effects observed in children could be initiated by exposures to PM that
occurred during key windows of development, such as in utero. Epidemiologic studies that
focus on exposures during development have reported inconsistent findings (Section 7.4),
but a recent toxicological study suggests that inflammatory responses during pregnancy
due to exposure to PM could result in health effects in the developing fetus.
July 2009
2-33
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Epidemiologic studies have also examined whether additional factors, such as gender,
race, or ethnicity modify the association between PM and morbidity and mortality
outcomes. Although gender and race do not seem to modify PM risk estimates, limited
evidence from two studies conducted in California suggest that Hispanic ethnicity may
modify the association between PM and mortality.
Recent epidemiologic and toxicological studies provided evidence that individuals with
null alleles or polymorphisms in genes that mediate the antioxidant response to oxidative
stress (i.e., GSTMl), regulate enzyme activity (i.e., MTHFR and cSHMT), or regulate
physiological levels of inflammatory markers (i.e., fibrinogen) are more susceptible to PM
exposure. However, some studies have shown that polymorphisms in genes (e.g., HFE) can
have a protective effect against effects of PM exposure. Additionally, preliminary evidence
suggests that PM exposure can impart epigenetic effects (i.e., DNAmethylation); however,
this requires further investigation.
Collectively, the evidence from epidemiologic and toxicological, and to a lesser extent,
controlled human exposure studies indicate increased susceptibility of individuals with
underlying CVDs and respiratory illnesses, particularly asthma, to PM exposure.
Controlled human exposure and toxicological studies provide additional evidence for
increased PM-related cardiovascular effects in individuals with underlying respiratory
health conditions.
Recently studies have begun to examine the influence of preexisting chronic
inflammatory conditions, such as diabetes and obesity, on PM-related health effects. These
studies have found some evidence for increased associations for cardiovascular outcomes
along with physiological alterations in markers of inflammation, oxidative stress, and acute
phase response. However, more research is needed to thoroughly examine the affect of PM
exposure on obese individuals and to identify the biological pathway(s) that could increase
the susceptibility of diabetic and obese individuals to PM.
There is also evidence that socioeconomic status (SES), measured using surrogates
such as educational attainment or residential location, modifies the association between
PM and morbidity and mortality outcomes. In addition, nutritional status, another
surrogate measure of SES, has been shown to have protective effects against PM exposure
in individuals that have a higher intake of some vitamins and nutrients.
Overall, the epidemiologic, controlled human exposure, and toxicological studies
evaluated in this review provide evidence for increased susceptibility for various
subpopulations, including children and older adults, people with pre-existing
cardiopulmonary diseases, and people with lower SES.
July 2009
2-34
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
2.4.2. Lag Structure of PM-Morbidity and PM-Mortality
Associations
Epidemiologic studies have attempted to identify the time-frame in which exposure to
PM can impart a health effect. Although PM exposure-response relationships have
traditionally been examined using air quality data for a defined lag period (e.g, 1 day or
average of 0-1 days), the relationship can potentially be influenced by a multitude of
factors, such as the underlying susceptibility of an individual (e.g., age, pre-existing
diseases), which could increase or decrease the lag times observed.
An attempt has been made to identify whether certain lag periods are more strongly
associated with specific health outcomes. The epidemiologic evidence evaluated in the 2004
PM AQCD supported the use of lags of 0-1 days for cardiovascular effects and longer
moving averages or distributed lags for respiratory diseases (U.S. EPA, 2004, 056905).
However, currently, little consensus exists as to the most appropriate a priori lag times to
use when examining morbidity and mortality outcomes. As a result, many investigators
have chosen to examine the lag structure of associations between PM concentration and
health outcome instead of focusing on a priori lag times. This approach is informative
because if effects are cumulative, higher overall risks may exist than would be observed for
any given single day lag. An examination of lag times used in the epidemiologic studies
evaluated in this assessment can provide further insight on the relationship between PM
exposure and morbidity and mortality outcomes.
2.4.2.1. PM-Cardiovascular Morbidity Associations
Most of the studies evaluated that examined the association between cardiovascular
HAs and ED visits report strong associations with short-term PM exposure at lags 0- to 2-
days, with more limited evidence for shorter durations (i.e., hours) between exposure and
response for some health effects (e.g., onset of MI) (Section 6.2.10). However, these studies
have rarely examined alternative lag structures. Controlled human exposure and
toxicological studies provide biological plausibility for the health effects observed in the
epidemiologic studies at immediate or concurrent day lags. Although the majority of the
evidence supports shorter lag times for cardiovascular health effects, a recent study has
provided preliminary evidence suggesting that longer lag times (i.e., 14-day distributed lag
model) maybe plausible for non-ischemic cardiovascular conditions (Section 6.2.10). Panel
studies of short-term exposure to PM and cardiovascular endpoints have also examined lag
times using a wide range of lag times. Studies of ECG changes, indicating ischemia, show
effects at lags from several hours to 2 days, while lag times ranging from hours to several
week moving averages have been observed in studies of arrhythmias, ECG measures of
arrhythmia, vasomotor function and blood markers of inflammation, coagulation and
oxidative stress (Section 6.2). The longer lags observed in these panel studies may be
July 2009
2-35
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
explained if the effects of PM are cumulative. Although few studies of cumulative effects
have been conducted, toxicological studies have demonstrated PM-dependent progression of
atherosclerosis. It should be noted that PM exposure could also lead to an acute event (e.g.,
infarction or stroke) in individuals with atherosclerosis that may have progressed in
response to cumulative PM exposure. Therefore, effects have been observed at a range of
lag periods from a few hours to several days with no clear evidence for any lag period
having stronger associations then another.
2.4.2.2.	PM-Respiratory Morbidity Associations
Generally, recent studies of respiratory HAs that evaluate multiple lags, have found
effect sizes to be larger when using longer moving averages or distributed lag models. For
example, when examining HAs for all respiratory diseases among older adults, the
strongest associations were observed when using PM concentrations 2 days prior to the HA
(Section 6.3.8). Longer lag periods were also found to be most strongly associated with
asthma HAs and ED visits children (3-5 days) with some evidence for more immediate
effects in older adults (lags of 0 and 1 day), but these observations were not consistent
across studies (Section 6.3.8). These variable results could be due to the biological
complexity of asthma, which inhibits the identification of a specific lag period. The longer
lag times identified in the epidemiologic studies evaluated are biologically plausible
considering that PM effects on allergic sensitization and lung immune defenses have been
observed in controlled human exposure and toxicological studies. These effects could lead to
respiratory illnesses over a longer time course (e.g., within several days respiratory
infection may become evident, resulting in respiratory symptoms or hospital visits).
However, inflammatory responses, which contribute to some forms of asthma, may results
in symptoms requiring medical care within a shorter time frame (e.g., 0" 1 days).
2.4.2.3.	PM-Mortality Associations
Epidemiologic studies that focused on the association between short-term PM
exposure and mortality (i.e., all-cause, cardiovascular, and respiratory) mostly examined a
priori lag structures of either 1 or 0*1 days. Although mortality studies do not often
examine alternative lag structures, the selection of the aforementioned a priori lag days has
been confirmed in additional studies, with the strongest PM-mortality associations
consistently being observed at lag 1 and 0- 1-days (Section 6.5). However, of note is recent
evidence for larger effect estimates when using a distributed lag model.
Epidemiologic studies that examined the association between long-term exposure to
PM and mortality have also attempted to identify the latency period from PM exposure to
death (Section 7.6.4). Results of the lag comparisons from several cohort studies indicate
that the effects of changes in exposure on mortality are seen within five years, with the
July 2009
2-36
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
strongest evidence for effects observed within the first two years. Additionally, there is
evidence, albeit from one study, that the mortality effect had larger cumulative effects
spread over the followup year and 3 preceding years.
2.4.3. PM Concentration-Response Relationship
An important consideration in characterizing the PM-morbidity and mortality
association is whether the OR relationship is linear across the full concentration range
that is encountered or if there are concentration ranges where there are departures from
linearity (i.e., nonlinearity). In this ISA studies have been identified that attempt to
characterize the shape of the PM OR curve along with possible PM "thresholds" (i.e., levels
which PM concentrations must exceed in order to elicit a health response). The
epidemiologic studies evaluated that examined the shape of the OR curve and the potential
presence of a threshold have focused on cardiovascular HAs and ED visits and mortality
associated with short-term exposure to PMio and mortality associated with long-term
exposure to PM2.5.
A limited number of studies have been identified that examined the shape of the PM-
cardiovascular HA and ED visits OR relationship. Of these studies, some conducted an
exploratory analysis during model selection to determine if a linear curve most adequately
represented the OR relationship! whereas, only one study conducted an extensive analysis
to examine the shape of the OR curve at different concentrations (Section 6.2.10.10).
Overall, the limited evidence from the studies evaluated supports the use of a no-threshold,
log-linear model, which is consistent with the observations made in studies that examined
the PM-mortality relationship.
Although multiple studies have previously examined the PM-mortality OR
relationship and whether a threshold exists, more complex statistical analyses continue to
be developed to analyze this association. Using a variety of methods and models, most of
the studies evaluated support the use of a no-threshold, log-linear model! however, one
study did observe heterogeneity in the shape of the OR curve across cities (Section 6.5).
Overall, the studies evaluated further support the use of a no-threshold log-linear model,
but additional issues such as the influence of heterogeneity in estimates between cities, and
the effect of seasonal and regional differences in PM on the OR relationship still require
further investigation.
In addition to examining the OR relationship between short-term exposure to PM
and mortality, Schwartz et al. (2008, 156963) conducted an analysis of the shape of the OR
relationship associated with long-term exposure to PM. Using a variety of statistical
methods, the OR curve was found to be indistinguishable from linear, and, therefore, little
evidence was observed to suggest that a threshold exists in the association between
long-term exposure to PM2.5 and the risk of death (Section 7.6).
July 2009
2-37
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
2.5. Ecological and Welfare Effects
This section presents key conclusions and scientific judgments regarding causality for
welfare and ecological effects of PM as discussed in Chapter 9. The effects of particulate
NOx and SOx have recently been evaluated in the ISA for Oxides of Nitrogen and Sulfur -
Ecological Criteria (U.S. EPA, 2008, 157074). That ISA focused on the effects from
deposition of gas- and particle-phase pollutants related to ambient NOx and SOx
concentrations that can lead to acidification and nutrient enrichment. Thus, emphasis in
Chapter 9 is placed on the effects of airborne PM on visibility and climate, and on the
effects of deposition of PM constituents other than NOx and SOx, primarily metals and
carbonaceous compounds. EPA's framework for causality, described in Chapter 1, was
applied and the causal determinations are highlighted.
Table 2-5. Summary of causality determination for welfare effects.
Welfare Effects
Causality Determination
Effects on Visibility
Causal
Effects on Climate
Causal
Ecological Effects
Likely to be causal
Effects on Materials
Causal
2.5.1. Summary of Effects on Visibility
Visibility impairment is caused by light scattering and absorption by suspended
particles and gases. There is strong and consistent evidence that PM is the overwhelming
source of visibility impairment in both urban and remote areas. EC and some crustal
minerals are the only commonly occurring airborne particle components that absorb light.
All particles scatter light, and generally light scattering by particles is the largest of the
four light extinction components (i.e., absorption and scattering by gases and particles).
Although a larger particle scatters more light than a similarly shaped smaller particle of
the same composition, the light scattered per unit of mass is greatest for particles with
diameters from -0.3-1.0 pm.
For studies where detailed data on particle composition by size are available, accurate
calculations of light extinction can be made. However, routinely available PM speciation
data can be used to make reasonable estimates of light extinction using relatively simple
algorithms that multiply the concentrations of each of the major PM species by its dry
July 2009
2-38
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
extinction efficiency and by a water growth term that accounts for particle size change as a
function of relative humidity for hygroscopic species (e.g., sulfate, nitrate, and sea salt).
This permits the visibility impairment associated with each of the major PM components to
be separately approximated from PM speciation monitoring data.
Direct optical measurement of light extinction measured by transmissometer, or by
combining the PM light scattering measured by integrating nephelometers with the PM
light absorption measured by an aethalometer, offer a number of advantages compared to
algorithm estimates of light extinction based on PM composition and relative humidity
data. The direct measurements are not subject to the uncertainties associated with
assumed scattering and absorption efficiencies used in the PM algorithm approach. The
direct measurements have higher time resolution (i.e., minutes to hours), which is more
commensurate with visibility effects compared with calculated light extinction using
routinely available PM speciation data (i.e., 24-h duration).
Particulate sulfate and nitrate have comparable light extinction efficiencies (haze
impacts per unit mass concentration) at any relative humidity value. Their light scattering
per unit mass concentration increases with increasing relative humidity, and at sufficiently
high humidity values (RH>85%) they are the most efficient particulate species contributing
to haze. Particulate sulfate is the dominant source of regional haze in the eastern U.S.
(>50% of the particulate light extinction) and an important contributor to haze elsewhere in
the country (>20% of particulate light extinction). Particulate nitrate is a minor component
of remote-area regional haze in the non-California western and eastern U.S., but an
important contributor in much of California and in the upper Midwestern U.S., especially
during winter when it is the dominant contributor to particulate light extinction.
EC and OC have the highest dry extinction efficiencies of the major PM species and
are responsible for a large fraction of the haze, especially in the northwestern U.S., though
absolute concentrations are as high in the eastern U.S. Smoke plume impacts from large
wildfires dominate many of the worst haze periods in the western U.S. Carbonaceous PM is
generally the largest component of urban excess PM2.5 (i.e., the difference between urban
and regional background concentration). Western urban areas have more than twice the
average concentrations of carbonaceous PM than remote areas sites in the same region. In
eastern urban areas PM2.5 is dominated by about equal concentrations of carbonaceous and
sulfate components, though the usually high relative humidity in the East causes the
hydrated sulfate particles to be responsible for about twice as much of the urban haze as
that caused by the carbonaceous PM.
PM2.5 crustal material (referred to as fine soil) and PM10-2.5 are significant contributors
to haze for remote areas sites in the arid southwestern U.S. where they contribute a
quarter to a third of the haze, with PM10-2.5 usually contributing twice that of fine soil.
Coarse mass concentrations are as high in the Central Great Plains as in the deserts
though there are no corresponding high concentrations of fine soil as in the Southwest. Also
July 2009
2-39
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
the relative contribution to haze by the high coarse mass in the Great Plains is much
smaller because of the generally higher haze values caused by the high concentrations of
sulfate and nitrate PM in that region.
Visibility has direct significance to people's enjoyment of daily activities and their
overall sense of wellbeing. A number of social science disciplines have attempted to link
perceived urban visibility to an array of effects reflecting the overall desire for good visual
air quality (VAQ), and the benefits of improving currently degraded VAQ. For example,
psychological research has demonstrated that people are emotionally affected by poor VAQ
such that their overall sense of wellbeing is diminished. Urban visibility has been examined
in two types of studies directly relevant to the NAAQS review process^ urban visibility
preference studies and urban visibility valuation studies. Both types of studies are designed
to evaluate individuals' desire for good VAQ where they live, using different metrics. Urban
visibility preference studies examine individuals' preferences by investigating the amount
of visibility degradation considered unacceptable, while economic studies examine the value
an individual places on improving VAQ by eliciting how much the individual would be
willing to pay for different amounts of VAQ improvement.
There are three urban visibility preference studies and one additional pilot study
(designed as a survey instrument development project) that have been conducted to date
that provide useful information on individuals' preferences for good VAQ in the urban
setting. The completed studies were conducted in Denver, Colorado, two cities in British
Columbia, Canada, and Phoenix, Arizona. The pilot study was conducted in Washington,
DC. One notable finding of the three visibility preference studies and the one pilot study is
the general degree of consistency in the median preferences for an acceptable amount of
visibility degradation. The range of median acceptable visibility preference values from the
four studies is 19-25 deciviews (dv). Measured in terms of visual range (VR), these median
acceptable values are between 59 km and 32 km.
The economic importance of urban visibility has been examined by a number of
studies designed to quantify the benefits (or willingness to pay) associated with potential
improvements in urban visibility. Urban visibility valuation research was described in the
2004 PM AQCD and the 2005 OAQPS PM NAAQS Staff Paper. Since the mid-1990s, little
new information has become available regarding urban visibility valuation (Section 9.2.4).
Collectively, the evidence is sufficient to conclude that a causal relationship exists
between PM and visibility impairment.
2.5.2. Summary of Effects on Climate
Aerosols affect climate through direct and indirect effects. The direct effect is
primarily realized as planet brightening when seen from space because most aerosols
scatter most of the visible spectrum light that reaches them. The IPCC AR4 reported that
July 2009
2-40
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
the radiative forcing from this direct effect was -0.5 (±0.4) W/m2 and identified the level of
scientific understanding of this effect as "medium-low." The global mean direct radiative
forcing from individual aerosol components varies from strongly negative for sulfate to
positive for black carbon with weaker positive or negative effects for other components, all
of which can vary strongly over space and time and with aerosol size. The indirect effects
are primarily realized as an increase in cloud brightness (termed the "first indirect" or
"Twomey" effect), changes in precipitation, and possible changes in cloud lifetime. The IPCC
AR4 reported that the radiative forcing from the Twomey effect was -0.7 (range: -1.1 to +4)
and identified the level of scientific understanding of this effect as "low." The other indirect
effects from aerosols were not considered to be radiative-forcing.
Taken together, direct and indirect effects from aerosols increase Earth's shortwave
albedo or reflectance, thereby reducing the radiative flux reaching Earth's surface from the
Sun. This produces net climate cooling from aerosols. The current scientific consensus
reported by IPCC AIM is that the direct and indirect radiative forcing from anthropogenic
aerosols computed at the top of the atmosphere, on a global average, is about -1.3 (range: -
2.2 to -0.5) W/m2. Although the magnitude of this negative radiative forcing appears large
in comparison to the analogous IPCC AIM estimate of positive radiative forcing from
anthropogenic GHG of about 2.9 (±0.3) W/m2, the spatial and temporal distributions of
these two very different radiative forcing agents are dissimilar! therefore, they do not
simply cancel and regional differences can be large. These differences result from the much
shorter atmospheric lifetime of aerosols than for the radiatively important trace gases,
implying that the radiative effects of aerosols respond much more quickly to changes in
emissions than do the effects from the gas-phase forcing agents Moreover, the effect of
present-day aerosols is to cool Earth's surface but, on average, to heat the atmosphere
itself within the atmospheric column, the radiative forcing effect from aerosols is estimated
to range from +0.8 to +2 W/m2.
Considerable progress has been made with in situ and remotely-sensed aerosols
concentrations including the MODIS, MISR, POLDER, and OMI satellite instruments
(Section 3.4.1.6). The accuracy for aerosol optical depth (AOD) measured with these global-
coverage remote sensing instruments is on the order of 0.05 or 20% of the AOD, but is still
much lower than for the limited-area surface-based sun photometers which have accuracy
in the range of 0.01-0.02. The differences remaining between surface and remotely sensed
AOD and between estimates computed from measurements and from numerical model
predictions are important because AOD is a significant element in determining radiative
forcing. Hence, uncertainty and error in AOD measurements and modeling propagate into
the range of estimates for total radiative forcing reported here.
Numerical modeling of aerosol effects on climate has also sustained remarkable
progress since the 2004 PM AQCD, though model solutions still display large heterogeneity
in their estimation of the direct radiative forcing effect from anthropogenic aerosols.
July 2009
2-41
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Differences among models are due in large measure to differences in: emissions of PM and
precursors to secondary PM formation! the representation of aerosol microphysical and
optical processes! and regional- and global-scale transport and transformation! as well as in
the effects of aerosol radiative forcing. The clear-sky direct radiative forcing over ocean due
to anthropogenic aerosols is estimated from satellite instruments to be on the order of -1.1
(±0.37) W/m2 while model estimates are -0.6 W/m2. The models' low bias over ocean is
carried through for the global average: global average direct radiative forcing from
anthropogenic aerosols is estimated from measurements to range from -0.9 to -1.9 W/m2,
larger than the estimate of -0.8 W/m2 from the models. Spatial heterogeneity in radiative
forcing is expected to exert significant effects on regional climate, but because the effects of
climate warming and cooling are not strictly co-located spatially or temporally with
radiative forcing or with emissions in particular for precursors for secondary PM formation,
assessments of effects for sub-global domains are even more uncertain than the global
averages reported here.
Overall, the evidence is sufficient to conclude that a causal relationship exists between
PM and effects on climate, including both direct effects on radiative forcing and indirect effects
that involve cloud feedbacks that influence precipitation formation and cloud lifetimes.
2.5.3. Summary of Ecological Effects of PM
Ecological effects of PM include direct effects to metabolic processes of plant foliage!
contribution to total metal loading resulting in alteration of soil biogeochemistry, plant
growth and animal growth and reproduction! and contribution to total organics loading
resulting in bioaccumulation and biomagnification across trophic levels. These effects were
well-characterized in the 2004 PM AQCD. Thus, the summary below builds upon the
conclusions provided in that review.
PM deposition comprises a heterogeneous mixture of particles differing in origin, size,
and chemical composition. Exposure to a given concentration of PM may, depending on the
mix of deposited particles, lead to a variety of phytotoxic responses and ecosystem effects.
Moreover, many of the ecological effects of PM are due to the chemical constituents (e.g.,
metals and organics) and their contribution to total loading within an ecosystem.
Investigations of the direct effects of PM deposition on foliage have suggested little or
no effects on foliar processes, unless deposition levels were higher than is typically found in
the ambient environment. However, consistent and coherent evidence of direct effects of PM
has been found in heavily polluted areas adjacent to industrial point sources such as
limestone quarries, cement kilns, and metal smelters (Sections 9.4.3. and 9.4.5.8.). Where
toxic responses have been documented, they generally have been associated with the
acidity, trace metal content, surfactant properties, or salinity of the deposited materials.
July 2009
2-42
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
An important characteristic of fine particles is their ability to affect the flux of solar
radiation passing through the atmosphere directly, by scattering and absorbing solar
radiation, and, indirectly, by acting as cloud condensation nuclei (CCN) that, in turn,
influence the optical properties of clouds. Regional haze has been estimated to diminish
surface solar visible radiation. Crop yields can be sensitive to the amount of sunlight
received, and crop losses have been attributed to increased airborne particle concentrations
in some areas of the world. PM has been observed to cause a decrease in photosynthetically
active radiation (PAR) via thick haze occurring in China that decreases plant growth in the
diffuse light portion of PAR. However, a global model showed that PM can increase the
diffuse light fraction of PAR. On a global scale, the diffuse light fraction of PAR has been
shown to increase growth. Consequently, it was shown that when PM is decreased, plant
growth and C storage are also decreased. Further research is needed to determine net
effects of PM alteration of light conditions on the growth of vegetation in the U.S.
The deposition of PM onto vegetation and soil, depending on its chemical composition,
can produce responses within an ecosystem. The ecosystem response to pollutant deposition
is a direct function of the level of sensitivity of the ecosystem and its ability to ameliorate
resulting change. Many of the most important ecosystem effects of PM deposition occur in
the soil. Upon entering the soil environment, PM pollutants can alter ecological processes of
energy flow and nutrient cycling, inhibit nutrient uptake, change ecosystem structure, and
affect ecosystem biodiversity. The soil environment is one of the most dynamic sites of
biological interaction in nature. It is inhabited by microbial communities of bacteria, fungi,
and actinomycetes, in addition to plant roots and soil macro-fauna. These organisms are
essential participants in the nutrient cycles that make elements available for plant uptake.
Changes in the soil environment can be important in determining plant and ultimately
ecosystem response to PM inputs.
There is strong and consistent evidence from field and laboratory experiments that
metal components of PM alter numerous aspects of ecosystem structure and function.
Changes in the soil chemistry, microbial communities and nutrient cycling, can result from
the deposition of trace metals. Exposures to trace metals are highly variable, depending on
whether deposition is by wet or dry processes. Although metals can cause phytotoxicity at
high concentrations, few heavy metals (e.g., Cu, Ni, Zn) have been documented to cause
direct phytotoxicity under field conditions. Exposure to coarse particles and elements such
as Fe and Mg are more likely to occur via dry deposition, while fine particles are more likely
to contain elements such as Ca, Cr, Pb, Ni, and V. Ecosystems immediately downwind of
major emissions sources can receive locally heavy deposition inputs. Phytochelatins
produced by plants as a response to sublethal concentrations of heavy metals are indicators
of metal stress to plants. Increased concentrations of phytochelatins across regions and at
greater elevation have been associated with increased amounts of forest injury in the
northeastern U.S.
July 2009
2-43
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Overall, the ecological evidence is sufficient to conclude that a causal relationship is
likely to exist between deposition of PM and a variety of effects on individual organisms and
ecosystems, based on information from the previous review and limited new findings in this
review. However, in many cases, it is difficult to characterize the nature and magnitude of
effects and to quantify relationships between ambient concentrations of PM and ecosystem
response due to significant data gaps and uncertainties as well as considerable variability
that exists in the components of PM and their various ecological effects.
2.5.4. Summary of Effects on Materials
Building materials (metals, stones, cements, and paints) undergo natural weathering
processes from exposure to environmental elements (wind, moisture, temperature
fluctuations, sunlight, etc.). Metals form a protective film of oxidized metal (e.g., rust) that
slows environmentally induced corrosion. However, the natural process of metal corrosion is
enhanced by exposure to anthropogenic pollutants. For example, formation of hygroscopic
salts increases the duration of surface wetness and enhances corrosion.
A significant detrimental effect of particle pollution is the soiling of painted surfaces
and other building materials. Soiling changes the reflectance of opaque materials and
reduces the transmission of light through transparent materials. Soiling is a degradation
process that requires remediation by cleaning or washing, and, depending on the soiled
surface, repainting. Particulate deposition can result in increased cleaning frequency of the
exposed surface and may reduce the usefulness of the soiled material.
Attempts have been made to quantify the pollutant exposure levels at which
materials damage and soiling have been perceived. However, to date, insufficient data are
available to advance the knowledge regarding perception thresholds with respect to
pollutant concentration, particle size, and chemical composition. Nevertheless, the evidence
is sufficient to conclude that a causal relationship exists between PM and effects on materials.
July 2009
2-44
DRAFT-DO NOT CITE OR QUOTE

-------
Chapter 2 References
Bell ML; Ebisu E Peng RD; Walker J; Samet JM; Zeger SL; Dominic F. (2008). Seasonal and regional short-
term effects of fine particles on hospital admissions in 202 U.S. counties, 1999-2005. Am J Epidemiol,
168: 1301-1310. 156266
Burnett RT; Brook JR; Yung WT; Dales RE; Krewski D. (1997). Association between ozone and hospitalization
for respiratory diseases in 16 Canadian cities. Environ Res, 72: 24-31. 082676
Burnett RT; Smith-Doiron M; Stieb D; Cakmak S; Brook JR. (1999). Effects of particulate and gaseous air
pollution on cardiorespiratory hospitalizations. Arch Environ Occup Health, 54: 130-139. 017269
Burnett RT; Stieb D; Brook JR; Cakmak S; Dales R; Raizenne M; Vincent R; Dann T. (2004). Associations
between short-term changes in nitrogen dioxide and mortality in Canadian cities. Arch Environ Occup
Health, 59: 228-236. 086247
Chen B; Hong C; Kan H. (2004). Exposures and health outcomes from outdoor air pollutants in China.
Toxicology, 198: 291-300. 089899
Chen LH; Knutsen SF; Shavlik D; Beeson WL; Petersen F; Ghamsary M; Abbey D. (2005). The association
between fatal coronary heart disease and ambient particulate air pollution: Are females at greater risk?.
Environ Health Perspect, 113: 1723-1729. 087942
Chen SJ; Hsieh LT; Kao MJ; Lin WY; Huang KL; Lin CC. (2004). Characteristics of particles sampled in
southern Taiwan during the Asian dust storm periods in 2000 and 2001. Atmos Environ, 38: 5925-5934.
147143
Chimonas MA; Gessner BD. (2007). Airborne particulate matter from primarily geologic, non-industrial sources
at levels below National Ambient Air Quality Standards is associated with outpatient visits for asthma
and quick-relief medication prescriptions among children less than 20 years old enrolled in Medicaid in
Anchorage, Alaska. Environ Res, 70: 2021-2026. 093261
Delfino RJ; Murphy-Moulton AM; Burnett RT; Brook JR; Becklake MR. (1997). Effects of air pollution on
emergency room visits for respiratory illnesses in Montreal, Quebec. Am J Respir Crit Care Med, 155:
568-576. 082687
Delfino RJ; Zeiger RS; Seltzer JM; Street DH. (1998). Symptoms in pediatric asthmatics and air pollution:
differences in effects by symptom severity, anti-inflammatory medication use and particulate averaging
time. Environ Health Perspect, 106: 751-761. 051406
Dockery DW! Cunningham J; Damokosh AI; Neas LM; Spengler JD; Koutrakis P; Ware JH; Raizenne M;
Speizer FE. (1996). Health effects of acid aerosols on North American children: respiratory symptoms.
Environ Health Perspect, 104: 500-505. 077269
Dockery DW; Luttmann-Gibson H; Rich DQ; Link MS; Mittleman MA; Gold DR; Koutrakis P; Schwartz JD;
Verrier RL. (2005). Association of air pollution with increased incidence of ventricular tachyarrhythmias
recorded by implanted cardioverter defibrillators. Environ Health Perspect, 113: 670-674. 078995
Dominici F; Peng RD; Bell ML; Pham L; McDermott A; Zeger SL; Samet JL. (2006). Fine particulate air
pollution and hospital admission for cardiovascular and respiratory diseases. JAMA, 295: 1127-1134.
088398
Eftim SE; Samet JM; Janes H; McDermott A; Dominici F. (2008). Fine particulate matter and mortality: a
comparison of the six cities and American Cancer Society cohorts with a medicare cohort. Epidemiology,
19: 209-216. 099104
Enstrom JE. (2005). Fine particulate air pollution and total mortality among elderly Californians, 1973-2002.
Inhal Toxicol, 17: 803-816. 087356
Fung KY; Khan S; Krewski D; Chen Y. (2006). Association between air pollution and multiple respiratory
hospitalizations among the elderly in Vancouver, Canada. Inhal Toxicol, 18: 1005-1011. 089789
Gent JF; Koutrakis P; Belanger K; Triche E; HolfordTR; Bracken MB; Leaderer BP. (2009). Symptoms and
medication use in children with asthma and traffic-related sources of fine particle pollution. Environ
Health Perspect, In Press: 1-41. 180399
Goss CH; Newsom SA; Schildcrout JS; Sheppard L; Kaufman JD. (2004). Effect of ambient air pollution on
pulmonary exacerbations and lung function in cystic fibrosis. Am J Respir Crit Care Med, 169: 816-821.
055624
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa. go v/her o. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009	2-45	DRAFT - DO NOT CITE OR QUOTE

-------
Ito K. (2003). Associations of particulate matter components with daily mortality and morbidity in Detroit,
Michigan. 042856
Ito E Thurston GD; Silverman RA;. (2007). Association between coarse particles and asthma emergency
department (ED) visits in New York City. 091262
Kim JJ; Smorodinsky Si Lipsett Mi Singer BCi Hodgson ATi Ostro B. (2004). Traffic-related air pollution near
busy roads' the East Bay children's Respiratory Health Study. Am J Respir Crit Care Med, 170- 520-526.
087383
Krewski Di Jerrett Mi Burnett RTi Ma Ri Hughes Ei Shi Yi Turner MCi Pope AC Illi Thurston Gi Calle EEi
Thun MJ. (2009). Extended follow-up and spatial analysis of the American Cancer Society study linking
particulate air pollution and mortality. Health Effects Institute. Cambridge, MA. 140. 191193
Laden Fi Schwartz Ji Speizer FEi Dockery DW. (2006). Reduction in fine particulate air pollution and mortality-
extended follow-up of the Harvard Six Cities study. Am J Respir Crit Care Med, 173: 667-672. 087605
Lin Mi Chen Yi Burnett RTi Villeneuve PJi Krewski D. (2002). The influence of ambient coarse particulate
matter on asthma hospitalization in children: case-crossover and time-series analyses. Environ Health
Perspect, 110: 575-581. 026067
Lin Mi Stieb DMi Chen Y. (2005). Coarse particulate matter and hospitalization for respiratory infections in
children younger than 15 years in Toronto: a case-crossover analysis. , 116: 235-240. 087828
Lipfert FWi Baty JDi Miller JPi Wyzga RE. (2006). PM2.5 constituents and related air quality variables as
predictors of survival in a cohort of U.S. military veterans. Inhal Toxicol, 18: 645-657. 088756
Lippmann M. (2000). Environmental toxicants: human exposures and their health effects. 024579
Lisabeth LDi Escobar JDi Dvonch JTi Sanchez BNi Majersik JJi Brown DLi Smith MAi Morgenstern LB. (2008).
Ambient air pollution and risk for ischemic stroke and transient ischemic attack. , 64: 53"59. 155939
Mar TFi Larson TV; Stier RA; Claiborn C; Koenig JQ. (2004). An analysis of the association between respiratory
symptoms in subjects with asthma and daily air pollution in Spokane, Washington. Inhal Toxicol, 16:
809-815. 057309
McConnell R; Berhane K; Gilliland F; Molitor J; Thomas D; Lurmann F; Avol E; Gauderman WJ; Peters JM.
(2003).	Prospective study of air pollution and bronchitic symptoms in children with asthma. Am J Respir
Crit Care Med, 168: 790-797. 049490
Metzger KB; Klein M; Flanders WD; Peel JL; Mulholland JA; Langberg JJ; Tolbert PE. (2007). Ambient air
pollution and cardiac arrhythmias in patients with implantable defibrillators. Epidemiology, 18: 585-592.
092856
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd KH; Mulholland JA; Ryan PB; Frumkin H.
(2004).	Ambient air pollution and cardiovascular emergency department visits. , 15: 46"56. 044222
Miller KA; Siscovick DS; Sheppard L; Shepherd K; Sullivan JH; Anderson GL; Kaufman JD. (2007). Long-term
exposure to air pollution and incidence of cardiovascular events in women. , 356: 447-458. 090130
Moolgavkar SH. (2003). Air pollution and daily deaths and hospital admissions in Los Angeles and Cook
counties. 042864
O'Connor GT; Neas L; Vaughn B; Kattan M; Mitchell H; Crain EF; Evans R, 3rd; Gruchalla R; Morgan W; Stout
J; Adams GK; Lippmann M. (2008). Acute respiratory health effects of air pollution on children with
asthma in US inner cities. J Allergy Clin Immunol, 121: 1133-1139 e!131. 156818
Ostro B; Broadwin R; Green S; Feng W"Y; Lipsett M. (2006). Fine particulate air pollution and mortality in nine
California counties: results from CALFINE. Environ Health Perspect, 114: 29"33. 087991
Ostro B; Roth L; Malig B; Marty M. (2009). The effects of fine particle components on respiratory hospital
admissions in children. Environ Health Perspect, 117: 475. 191971
Peel JL; Tolbert PE; Klein M; Metzger KB; Flanders WD; Knox T; Mulholland JA; Ryan PB; Frumkin H. (2005).
Ambient air pollution and respiratory emergency department visits. , 16: 164-174. 056305
Peng RD; Chang HH; Bell ML; McDermott A; Zeger SL; Samet JM; Dominici F. (2008). Coarse particulate
matter air pollution and hospital admissions for cardiovascular and respiratory diseases among Medicare
patients. JAMA, 299: 2172-2179. 156850
Peters A; Dockery DW; Muller JE; Mittleman MA. (2001). Increased particulate air pollution and the triggering
of myocardial infarction. , 103: 2810-2815. 016546
Pope C; Renlund D; Kfoury A; May H; Home B. (2008). Relation of heart failure hospitalization to exposure to
fine particulate air pollution. , 102: 1230-1234. 191969
Pope CA III; Muhlestein JB; May HT; Renlund DG; Anderson JL; Home BD. (2006). Ischemic heart disease
events triggered by short-term exposure to fine particulate air pollution. Circulation, 114: 2443-2448.
091246
Rabinovitch N; Strand M; Gelfand EW. (2006). Particulate levels are associated with early asthma worsening in
children with persistent disease. Am J Respir Crit Care Med, 173: 1098-1105. 088031
July 2009
2-46
DRAFT-DO NOT CITE OR QUOTE

-------
Rabinovitch N; Zhang LN; Murphy JR; Vedal S; Dutton SJ; Gelfand EW. (2004). Effects of wintertime ambient
air pollutants on asthma exacerbations in urban minority children with moderate to severe disease. J
Allergy Clin Immunol, 114- 1131-1137. 096753
Rich DQ; Kim MH; Turner JR; Mittleman MA; Schwartz J; Catalano PJ; Dockery DW. (2006). Association of
ventricular arrhythmias detected by implantable cardioverter defibrillator and ambient air pollutants in
the St Louis, Missouri metropolitan area. Occup Environ Med, 63- 591-596. 089814
Rich KE; Petkau J; Vedal S; Brauer M. (2004). A case-crossover analysis of particulate air pollution and cardiac
arrhythmia in patients with implantable cardioverter defibrillators. Inhal Toxicol, 16- 363-372. 055631
Schwartz J; Coull B; Laden F; Ryan L. (2008). The effect of dose and timing of dose on the association between
airborne particles and survival. Environ Health Perspect, 116- 64"69. 156963
Sheppard L. (2003). Ambient air pollution and nonelderly asthma hospital admissions in Seattle, Washington,
1987-1994. 042826
Slaughter JC; Kim E; Sheppard L; Sullivan JH; Larson TV; Claiborn C. (2005). Association between particulate
matter and emergency room visits, hospital admissions and mortality in Spokane, Washington. J Expo
Sci Environ Epidemiol, 15: 153-159. 073854
Sullivan J; Ishikawa N; Sheppard L; Siscovick D; Checkoway H; Kaufman J. (2003). Exposure to ambient fine
particulate matter and primary cardiac arrest among persons with and without clinically recognized
heart disease. Am J Epidemiol, 157: 501-509. 043156
Sullivan J; Sheppard L; Schreuder A; Ishikawa N; Siscovick D; Kaufman J. (2005). Relation between short-term
fine-particulate matter exposure and onset of myocardial infarction. , 16: 41-48. 050854
Symons JM; Wang L; Guallar E; Howell E; Dominici F; Schwab M; Ange BA; Samet J; Ondov J; Harrison D;
Geyh A. (2006). A case-crossover study of fine particulate matter air pollution and onset of congestive
heart failure symptom exacerbation leading to hospitalization. Am J Epidemiol, 164: 421-33. 091258
Thurston GD; Ito K; Hayes CG; Bates DV; Lippmann M. (1994). Respiratory hospital admissions and
summertime haze air pollution in Toronto, Ontario: consideration of the role of acid aerosols. Environ
Res, 65: 271-290. 043921
Tolbert PE; Kleina M; Peelb JL; Sarnata SE; Sarnata JA. (2007). Multipollutant modeling issues in a study of
ambient air quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 24: 938"
945. 090316
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA/600/P-99/002aF"bF. 056905
U.S. EPA. (2008). Integrated science assessment for oxides of nitrogen and sulfur: Ecological criteria. EPA.
EPA. EPA/600/R-08/082F. 157074
Villeneuve PJ; Chen L; Stieb D; Rowe BH. (2006). Associations between outdoor air pollution and emergency
department visits for stroke in Edmonton, Canada. Eur J Epidemiol, 21: 689-700. 090191
Woodruff TJ; Darrow LA; Parker JD. (2008). Air pollution and postneonatal infant mortality in the United
States, 1999-2002. Environ Health Perspect, 116: 110-5. 098386
Yang Q; Chen Y; Krewski D; Shi Y; Burnett RT; McGrail KM. (2004). Association between particulate air
pollution and first hospital admission for childhood respiratory illness in Vancouver, Canada. Arch
Environ Occup Health, 59: 14-21. 087488
Zanobetti A; Schwartz J. (2009). The effect of fine and coarse particulate air pollution on mortality: A national
analysis. Environ Health Perspect, 117: 1-40. 188462
Zeger S; Dominici F; McDermott A; Samet J. (2008). Mortality in the Medicare population and chronic exposure
to fine particulate air pollution in urban centers (2000-2005). Environ Health Perspect, 116: 1614.
191951
Zhang Z; Whitsel E; Quibrera P; Smith R; Liao D; Anderson G; Prineas R. (2009). Ambient Fine Particulate
Matter Exposure and Myocardial Ischemia in the Environmental Epidemiology of Arrhythmogenesis in
the Women?s Health Initiative (EEAWHI) Study. Environ Health Perspect, 117: 751. 191970
July 2009
2-47
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Chapter 3. Source to Human Exposure
3.1. Introduction
This chapter describes basic concepts and new and established findings in
atmospheric sciences and human exposure assessment relevant to PM to establish a
foundation for the health and ecological effects discussed in subsequent chapters.
Information in this chapter builds on previous AQCDs for PM using new data and re-
interpretations of extant studies as well. This includes new knowledge of PM chemistry, the
latest developments in monitoring methodologies, recent national and local measurements
and trends in PM concentrations as a function of size range and composition, advances in
receptor and chemistry-transport modeling, revised estimates of policy-relevant background
PM, and recent work on exposure assessment.
The chapter and its associated annex material are organized as follows: Section 3.2
presents an overview of basic information related to the size distribution and composition of
airborne particles. Section 3.3 provides a brief description of the sources, emissions, and
deposition of PM, including discussions of possible mechanisms of secondary PM formation
from gaseous precursors and of the atmospheric processes that deposit PM to the earth's
surface. Issues related to the measurement of PM and its chemical components and to
monitors and networks in the U.S. are covered in Section 3.4; supplementary material on
these topics is contained in Annex A, Section A. 1. Analyses of data for ambient
concentrations of PM and its components are characterized in Section 3.5, and
supplementary information can be found in Annex A, Section A.2. Section 3.6 describes
methods for determining source contributions to ambient samples by receptor models and
presents results from recent receptor modeling studies. In addition, the construction of
chemistry-transport models (CTMs) to determine pollutant concentrations is described in
Section 3.6. Supplementary information about receptor model methods and results is given
in Annex A, Section A.3. Policy relevant background concentrations of PM, i.e., those
concentrations defined to result from natural sources everywhere in the world together
Note- Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health
and Environmental Research Online) at http • I lev a. go v/her o. HERO is a database of scientific literature used by U.S. EPA in
the process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009
3-1
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
with anthropogenic sources outside of Canada, the United States, and Mexico, are
presented in Section 3.7. Issues related to human exposure assessment including sources of
exposure and implications for epidemiologic studies are discussed in Section 3.8.
Supplementary information on exposure studies is included in Annex A, Section A.4.
Finally, the summary and conclusions from Chapter 3 are presented in Section 3.9.
3.2. Overview of Basic Aerosol Properties
Unlike gas-phase pollutants such as SO2, CO, H2CO and O3, which are well-defined
chemical entities, atmospheric PM varies in size, shape, and chemical composition.
Atmospheric chemical and microphysical processing of direct emissions of PM and its
precursors together with mechanical generation of particles tend to produce distinct
lognormal modes (Whitby, 1978, 071181) as shown in Figure 3-1. To the extent that
information is available, discussions in this and subsequent chapters will focus on particles
in specific size ranges (i.e., PM2.5, PM10-2.5 and PM10). The subscripts after PM refer to the
aerodynamic diameter1 (dae) in micrometers (um) of 50% cut points of sampling devices. For
example, EPA defines PM10 as particles collected by a sampler with an upper 50% cut point
of 10 um dae and a specific, fairly sharp, penetration curve, as defined in the Code of Federal
Regulations (40 CFR Part 58). PM2.5 is defined in an analogous way. Ultrafine particles,
defined here as particles with a diameter <0.1 um (typically based on physical size, thermal
diffusivity, or electrical mobility), will also be discussed.
The terms "fine particles" and "coarse particles" have lost the precise meaning as
defined in Whitby (1978, 071181). where "fine particles" refers to all particles in the
nucleation, Aitken, and accumulation modes! and "coarse particles" characterizes all
particles larger than these. Ultrafine particles correspond loosely to the nucleation plus
Aitken modes (in earlier literature, these modes were not separated and the combination,
unresolved by older instruments, was called the Aitken mode). Now, the term "fine
particles" is most often associated with the PM2.5 fraction, which includes the nucleation,
1 Aerodynamic diameter is the diameter of a unit density (l g/cm3) sphere that has the same gravitational settling velocity as
the particle of interest and is a useful metric for characterizing particles >~ 1 jj,m. For sub-micron particles, forces other than
gravity increase in importance in determining a particle's motion and air can no longer be considered a continuum.
Regardless, aerodynamic diameter is frequently reported down to ~ 0.1 (im where the assumptions used in its derivation no
longer hold. A useful metric for characterizing particles <~ 0.5 (im is the mobility diameter defined as the diameter of a
particle having the same diffusivity or electrical mobility in air as the particle of interest. In the region between ~ 0.5—1.0
jim, aerodynamic and mobility diameters are not necessarily the same. The question of how best to merge these diameters is
unresolved and depends on the particle properties of interest.
July 2009
3-2
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
Aitken and accumulation modes and some particles from the lower-size tail of the coarse
particle mode between about 1 and 2.5 um aerodynamic diameter. "Thoracic coarse" is
frequently used in reference to PM10-2.5, which does not include the low-end tail of the coarse
particle mode. With high relative humidity larger particles in the accumulation mode could
also extend into the 1 to 3 um size range. These relationships can be seen in Figure 3-1,
which shows the number distribution for ultrafine particles and the volume distribution (or
mass distribution if particle density is constant across the size range) for fine and (thoracic)
coarse particles. The figure is arranged this way because particle number is most highly
concentrated in the ultrafine size range but volume (or mass) is most concentrated in the
larger size ranges.
	Fine Particles	 Coarse particles
Ultrafine Particles
40
Nucleation
Mode
30 —
Aitken
Mode
40—
Droplet
Submode
Accumulation
Mode _
Coarse
Mode
Condensation
Submode
rrffrrj-
0.01	0.1	1	10
Diameter (micrometers)
Source: Pandis (2004,156838).
Figure 3-1. Particle size distributions by number and volume. Dashed lines refer to values in
individual modes and solid lines to their sum. Note that ultrafine particles are a subset
of fine particles.
Characterizing particle size is important because different size particles penetrate to
different regions of the human respiratory tract. Thoracic particles refer to particles that
July 2009
3-3
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
travel past the larynx to reach the lung airways and the gas-exchange region of the lung,
and respirable particles are those that reach the gas-exchange region. The selection of PMio
as an indicator of thoracic particles was based in large part on dosimetry (U.S. EPA, 1996,
079380). However, the selection of PM2.5 to characterize respirable particles was driven
mainly by considerations related to measurement techniques available at the time rather
than dosimetry. Currently, cut points other than 2.5 um are attainable and frequently put
into use. For example, the American Conference of Governmental Industrial Hygienists
(ACGIH, 2005, 156188), the International Standards Organization, and the European
Standardization Committee have adopted a 50% cut point of 4 um as an indicator of
respirable particles. Most commonly, however, PM2.5 is used as an indicator of respirable
particles, PM10-2.5 is used as an indicator of the thoracic component of coarse particles that
is sometimes referred to as thoracic coarse (noting that it excludes some coarse particles
below 2.5 um and above 10 um), and PM10 is used as an indicator of thoracic particles.
As can be seen from Table 3-1, particles in individual size modes are characterized by
rather distinct sources, composition, chemical properties, lifetimes in the atmosphere (x)
and distances over which they can travel. Whereas particles in the smaller size modes are
formed mainly by combustion processes and by nucleation and condensation of gases,
coarse particles are generated mainly by mechanical activity, such as by the action of wind
on either the ground or the sea surface or by construction or by resuspension by traffic.
Particles in the ultrafine size range are either emitted directly to the atmosphere or are
formed by nucleation of gaseous constituents in the atmosphere as shown in Table 3-1. The
properties of fibers and engineered nano-objects and nanometer scale products (e.g., dots,
hollow spheres, rods, fibers, and tubes) are not reviewed in this chapter because these
classes of objects are found mainly in certain occupational settings and not in ambient air.
Table 3-1.
Characteristics of ambient fine (ultrafine plus accumulation-mode) and coarse particles.

Fine


Ultrafine Accumulation

Formation
Processes
Combustion, high-temperature processes, and atmospheric reactions
Break up of large solids/droplets
Formed by
Nucleation of atmospheric gases including Condensation of gases
H2SO4, NHs and some organic compounds Coagu|ation of smaNer partic|es
Condensation of gases Reactions of gases in or on particles
Evaporation of fog and cloud droplets in which ga;
dissolved and reacted
Mechanical disruption (crushing, grinding,
abrasion of surfaces)
Evaporation of sprays
>es have Suspension of dusts
Reactions of gases in or on particles
July 2009
3-4
DRAFT-DO NOT CITE OR QUOTE

-------
Fine
Composed of Sulfate
Sulfate, nitrate, ammonium, and hydrogen ions
Nitratesfchloridesfsulfates from HNO3IHCIISO2
EC
EC
reactions with coarse particles
Metal compounds
Large variety of organic compounds
Oxides of crustal elements (Si, Al, Ti, Fe)
Organic compounds with very low
Metals: compounds of Pb, Cd, V, Ni, Cu, Zn, Mn, Fe, etc.
CaC03, CaSOi, NaCI, sea salt
saturation vapor pressure at ambient
Particle-bound water
Bacteria, pollen, mold, fungal spores, plant and
temperature
animal debris
Bacteria, viruses

Solubility
Not well characterized
Largely soluble, hygroscopic, and deliquescent
Largely insoluble and nonhygroscopic
Sources	High temperature combustion
Atmospheric reactions of primary, gaseous
compounds.
Combustion of fossil and biomass fuels, and high
temperature industrial processes, smelters, refineries, steel
mills etc.
Atmospheric oxidation of NO2, SO2, and organic
compounds, including biogenic organic species
(e.g., terpenes)
Resuspension of particles deposited onto roads
Tire, brake pad, and road wear debris
Suspension from disturbed soil (e.g., farming,
mining, unpaved roads)
Construction and demolition
Fly ash from uncontrolled combustion of coal,
oil, and wood
Ocean spray
Atmospheric
half-life
Minutes to hours
Days to weeks
Minutes to hours
Removal
Processes
Grows into accumulation mode
Diffuses to raindrops
Forms cloud droplets and rains out
Dry deposition
Dry deposition by fallout
Scavenging by falling rain drops
Travel distance
< 1 to 10s of km
100s to 1000s of km
< 1 to 10s of km (100s to 1,000s of km in
dust storms for the small size tail)
Source: Wilson and Suh (1997, 077408) (adapted).
July 2009
3-5
DRAFT-DO NOT CITE OR QUOTE

-------

i m
i iL 1 i
O All JjTJj
i'AlLls ,t

Cl
Na l
Li ...1
B 1
1L 3
? S K
Source: National Exposure Research Laboratory.
Figure 3-2. X-ray spectra and scanning electron microscopy images of individual particles
including:(1) an aluminum-silicate fly ash sphere emitted from a coal-fired power plant,
(2) an iron oxide sphere emitted from a steel manufacturing facility, (3) An aluminum-
silicate particle, probably of crustal origin, (4) a carbon soot aggregate from a diesel
engine consisting of many sub-micron size carbon particles, (5) a sodium chloride
crystal, potentially of marine origin, and (6) a partially collapsed pollen particle. The
polycarbonate filter substrate used to collect the particles contributes to the carbon
peak in each spectrum.
1	Particles appear in a wide variety of shapes such as spheres, ellipsoids, cubes, and
2	irregular or fractal geometries. This is one reason why a standard metric such as
3	aerodynamic diameter is useful for describing the mechanical properties of the particles.
4	The shape of particles is important for determining the optical properties of the particles.
5	The directionality of sunlight scattered by certain shapes of particles, such as plates, also
July 2009
3-6
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
depends strongly on their physical orientation while suspended. The shape of particles also
affects the surface area of the particles in contact with the surface it is deposited on,
including cell membranes.
Images of six types of individual particles obtained using scanning electron
microscope (SEM) and their corresponding x-ray spectra showing their major elemental
composition are shown in Figure 3-2. The images show particles sitting on a thin
polycarbonate film with pores and a 1 um scale bar for size reference. Air is pulled through
the filter pores with a vacuum pump and the particles are left behind on the surface. The
metal dominated spherical particles (image 2) were formed at high temperatures and were
quickly cooled. Particles which are liquid such as sulfate are also spherical. Sodium chloride
(NaCl) crystals are cubic (image 5); this particular particle could be marine sea salt due to
the proximity to the ocean where the sample was collected. Other particles, such as the
carbon chain agglomerates from diesel engines have much more irregular and complex
shapes (image 4). Note that these particles were placed under vacuum resulting in
volatilization of water and other volatile components and partial collapse of the pollen grain
(image 6). Changes in composition and possibly morphology could occur during the
sampling, collection and analysis of aerosol samples. For example, particles maybe coated
with semi-volatile material that evolves off the particles under vacuum and under the
electron beam. Similarly, particles in ambient air are generally mixtures or agglomerates of
particles coming from multiple sources and have a diverse chemical make-up.
3.3. Sources, Emissions and Deposition of Primary and
Secondary PM
PM is composed of both primary (derived directly from emissions) or secondary
(derived from atmospheric reactions involving gaseous precursors) components. Table 3-2
summarizes anthropogenic and natural sources for the major primary and secondary
aerosol constituents of fine and coarse particles. Anthropogenic sources can be further
divided into stationary and mobile sources. Stationary sources include fuel combustion for
electrical utilities, residential space heating and cooking! industrial processes! construction
and demolition; metal, mineral, and petrochemical processing! wood products processing!
mills and elevators used in agriculture! erosion from tilled lands! waste disposal and
recycling! and biomass combustion. Biomass combustion encompasses many emission
July 2009
3-7
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
activities including burning of wood for fuel, burning of vegetation to clear land for
agriculture and construction, to dispose of agricultural and domestic waste, to control the
growth of animal or plant pests, and to manage forest resources (prescribed burning).
Wildlands also burn due to lightning strikes and arson. Mobile or transportation-related
sources include direct emissions of primary PM and secondary PM precursors from highway
vehicles and non-road sources as well as fugitive dust from paved and unpaved roads. Also
shown in Table 3-2 are sources for several precursor gases, the oxidation of which can form
secondary PM. An overview of estimates of emissions of primary PM and precursors to
secondary PM from major sources is given in Section 3.3.1. The transformations from
gaseous precursors shown in Table 3-2 to secondary PM are described in Section 3.3.2.
In general, the sources of fine PM are very different from those of coarse PM. Some of
the mass in the fine size fraction forms during combustion from material that has
volatilized in combustion chambers and then recondensed before emission to the
atmosphere. Some ambient PM2.5 forms in the atmosphere from photochemical reactions
involving precursor gases. Included in this category is the formation of new ultrafine
particles by homogeneous nucleation of precursor gases in addition to the condensation of
gases on pre-existing particles. PM formed by the first mechanism is referred to as primary,
and PM formed by the second mechanism is referred to as secondary. Biological material
also exists in the fine fraction including many types of microorganisms, especially viruses
and bacteria and fragments of pollens and fungal spores. PM10-2.5 is mainly primary in
origin, as it is produced by surface abrasion or by suspension of biological material and
fragments of living things (e.g., plant and insect debris). In addition, atmospheric reaction
products condense on coarse particles. Because precursor gases undergo mixing during
transport from their sources and reactions in the atmosphere can produce the same
products, it is difficult to identify individual sources of secondary PM. Transport and
transformation of precursors can occur over distances of hundreds of kilometers. PM10-2.5
has a shorter lifetime in the atmosphere, so its effects tend to be more localized. However,
intercontinental transport of dust from African and Asian deserts occurs and some of this
material is in the PM10-2.5 size range. Major events are highly episodic but much smaller
contributions can be made at other times (see Section 3.7).
July 2009
3-8
DRAFT-DO NOT CITE OR QUOTE

-------
Table 3-2. Constituents of atmospheric particles and their major sources.
Primary (PM <2.5//m)
Primary (PM >2.5//m)
Secondary PM Precursors
(PM < 2.5 //m)
Aerosol species
Natural
Anthropogenic
Natural
Anthropogenic
Natural
Anthropogenic
Sulfate (SO42)
Sea spray
Fossil fuel combustion
Sea spray
_
Oxidation of reduced
Oxidation of SO2



sulfur gases emitted
emitted from fossil fuel





by the oceans and
combustion





wetlands and SO2 and






H2S emitted by






volcanism and forest






fires

Nitrate (NO3 )

Mobile source
_
_
Oxidation of NOx
Oxidation of NOx


exhaust


produced by soils,
emitted from fossil fuel





forest fires, and
combustion and in motor





lighting
vehicle exhaust
Minerals
¦ ¦ and re-
Fugitive dust from
Erosion and re-
Fugitive dust, paved
_
_

entrainment
paved and unpaved
entrainment
and unpaved road




roads, agriculture,

dust, agriculture,




forestry,

forestry,




construction, and

construction, and




demolition

demolition


Ammonium (NH4+)
_
Mobile source
_

Emissions of NH3
Emissions of NH3 from


exhaust


from wild animals,
motor vehicles, animal





and undisturbed soil
husbandry, sewage, and






fertilized land
Organic carbon (0C)
Wildfires
Prescribed burning,
Soil humic matter
¦¦ and asphalt
Oxidation of
Oxidation of

wood burning, motor

wear, paved and
hydrocarbons emitted
hydrocarbons emitted


vehicle exhaust,

unpaved road dust
by vegetation
by motor vehicles,


cooking, tire wear and


(terpenes, waxes) and
prescribed burning,


industrial processes


wild fires
wood burning, solvent






use and industrial






processes
EC
Wildfires
Mobile source

Tire and asphalt

_


exhaust (mainly

wear, paved and




diesel), wood biomass

unpaved road dust




burning, and cooking




Metals
Volcanic activity Fossil fuel	Erosion, re-
combustion, smelting entrainment, and
and other	organic debris
metallurgical
processes, and brake
wear
Bioaerosols
Viruses and
bacteria
Plant and insect
fragments, pollen,
fungal spores, and
bacterial
Dash (—) indicates either very minor source or no known source of component.
Source: U.S. EPA (2004, 056905).
Only major sources for each constituent within each broad category shown at the top
of Table 3-2 are listed. Not all sources are equal in magnitude. Chemical characterizations
of primary particulate emissions for a wide variety of natural and anthropogenic sources (as
shown in Table 3-2) were given in Chapter 5 of the 1996 PM AQCD (U.S. EPA, 1996,
079380). Summary tables of the composition of source emissions presented in the 1996 PM
AQCD (U.S. EPA, 1996, 079380) and updates to that information are provided in Appendix
3D to the 2004 PM AQCD (U.S. EPA, 2004, 056905). Source composition profiles are
July 2009
3-9
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
archived by the EPA at http://www.epa.gov/ttn/chief/software/speciate/. The profiles of
source composition were based in large measure on the results of studies that collected
source signatures for use in source apportionment studies.
Natural sources of primary PM include windblown dust from undisturbed land, sea
spray, and biological material. The oxidation of a fraction of terpenes emitted by vegetation
and reduced sulfur species from anaerobic environments leads to secondary PM formation.
Ammonium (NH4"1") ions, which play a major role in regulating the pH of particles, are
derived from emissions of NH3 gas. Source categories for NH3 have been divided into
emissions from undisturbed soils (natural) and emissions that are related to human
activities (e.g., fertilized lands, domestic and farm animal waste). There is ongoing debate
about characterizing emissions from wildfires (i.e., unwanted fire) as either natural or
anthropogenic. Wildfires have been listed in Table 3-2 as natural in origin, but land
management practices and other human actions affect the occurrence and scope of
wildfires. For example, fire suppression practices allow the buildup of combustible fuels and
increase the susceptibility of forests to more severe and infrequent fires from whatever
cause, including lightning strikes. Similarly, prescribed burning is listed as anthropogenic,
but can be viewed as a substitute for wildfires that would otherwise occur eventually on the
same land.
3.3.1. Emissions of Primary PM and Precursors to
Secondary PM
U.S. national average emissions of primary PM2.5, PM10 and gaseous precursor species
(SO2, NOx, NH3 and VOCs) from different source categories are shown in Figure 3-3. Note
that the entries refer mainly to anthropogenic sources, with little information about natural
sources. However, for categories such as VOCs, the contribution from biogenic emissions of
isoprene and terpenes can be quite large. The entries are continually undergoing revision
and are subject to varying degrees of uncertainty. For example, almost all of the sulfur in
fuel is released as volatile components (SO2 or SO3) during combustion. Hence, sulfur
emissions can be calculated on the basis of sulfur content in fuels to a greater accuracy than
can be done for other pollutants like nitrogen oxides or primary PM. There have been
notable downward revisions to the inventories since 2002 in the emissions of dust from
roads. These have resulted in large measure from incorporation of emissions test data that
July 2009
3-10
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
relies on updated measurement methods. Also, the spatial and temporal characterization of
wildfire emissions has improved since 2002 by integrating satellite-derived fire detection
and state-of-the-art fuels characterization and consumption models (Pouliot et al., 2008,
156883). Emission measurements from high-temperature combustion sources are sensitive
to the dilution, temperature, and pre-treatment of dilution air (England et al., 2007,
156420; England et al., 2007, 156421; Sheya et al., 2008, 156977).
To a large extent, especially with regard to the contribution of road dust to PM,
refinements in emission estimates have been guided by the use of receptor modeling. See
Section 3.6.1 in this chapter for a description of receptor modeling techniques and the last
PM AQCD (U.S. EPA, 2004, 056905) for the role of receptor models in refining emissions
estimates. Note that since the estimates given in Figure 3-3 are U.S. national averages,
they may not accurately reflect the contribution of specific local sources determining a
person's exposures to PM at any given time and location.
As can be seen from a comparison of the total U.S. emissions (in million metric tons)
shown in Figure 3"3, estimates of total emissions of potential precursors to secondary PM
formation including SO2, NOx, NH3 and VOCs are considerably larger than those for
primary PM sources. The emissions of SO2, NOx, and NH3 that are converted to PM should
be multiplied by factors of 1.5, 1.35, and 1.07, respectively, to account for their chemical
form in the aerosol phase. Estimating a factor for VOCs is somewhat less straightforward.
In addition, OC mass, whether primary or secondary, must be adjusted to account for aging
in the atmosphere. Turpin and Lim (2001, 017093) recommend factors ranging from 1.4 to
2.0 to account for the conversion of OC to oxygen- and nitrogen-containing compounds in
the aerosol phase. In general, however, the emissions of precursors cannot be translated
directly into rates of PM formation. Dry deposition and precipitation scavenging of some of
these gaseous precursors and their intermediate oxidation products occur before they are
converted to PM in the atmosphere, and most of the VOCs are oxidized to carbon dioxide
(CO2) rather than PM. In addition, some fraction of these gases is transported outside the
domain of the continental United States before being oxidized. Likewise, emissions of these
gases from areas outside the United States can result in the transport of their oxidation
products into the United States.
July 2009
3-11
DRAFT-DO NOT CITE OR QUOTE

-------
Fertilizer &
..v-.lt-
[>sposa
So vent
Livestock
Road Dust
Residential
V.C:X!
Incust
ConrrVRes
Industrial
On-Road
2%
Non-Road
5%
Msc.
16%
PM2 5 (5-4 MMT)
On-Road
Non-Road 2%
3%
Msc.
0% "
Residential
Road uist
Sorvent|
'.Vaot*
I/scosa
Industrial
hdust
Ccrrv;-
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
3.3.2. Formation of Secondary PM
Precursors to secondary PM have natural and anthropogenic sources, just as primary
PM has natural and anthropogenic sources. A substantial fraction of the fine particle mass,
especially during the warmer months of the year, is secondary in nature, formed as the
result of atmospheric reactions involving both inorganic and organic gaseous precursors.
The major atmospheric chemical transformations leading to the formation of particulate
nitrate (pNOa) and sulfate (pSC>4) are relatively well understood; whereas those involving
the formation of secondary organic aerosol (SOA) are less well understood and are subject
to much current investigation. A large number of organic precursors are involved and many
of the kinetic details still need to be determined. Also, many of the products of the oxidation
of hydrocarbons have yet to be identified. However, there has been substantial progress
made in understanding the chemistry of SOA formation in the past few years.
3.3.2.1.	Formation of Nitrate and Sulfate
The basic mechanism of the gas and aqueous phase oxidation of NO2 and SO2 has long
been studied and can be found in numerous texts on atmospheric chemistry, e.g., Seinfeld
and Pandis (1998, 018352), Finlayson-Pitts and Pitts (2000, 055565), Jacob (1999, 091122),
and Jacobson (2002, 090667). The reader is referred to the 2004 PM AQCD (U.S. EPA, 2004,
056905) where these processes are described in great detail, as well as the 2008 NOx ISA
(U.S. EPA, 2008, 157073) and the 2008 SOx ISA (U.S. EPA, 2008, 157071).
3.3.2.2.	Formation of Secondary Organic Aerosol
Some key new findings have altered perceptions of SOA formation since the 2004 PM
AQCD (see especially the reviews by Kroll and Seinfeld (2008, 155910) and Rudich et al.
(2007, 156059). New measurement techniques for estimating the speciation and water
solubility of organic aerosols have noted the dominant contribution of oxygenated species in
atmospheric particles. Recent measurements show that the abundance of oxidized SOA
exceeds that of more reduced hydrocarbon like organic aerosol in Pittsburgh (Zhang et al.,
2005, 157185) and in about 30 other cities across the Northern Hemisphere (Zhang et al.,
2007, 101119). Based on aircraft and ship-based sampling of organic aerosols in coastal
July 2009
3-13
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
waters downwind of northeastern U.S. cities, de Gouw et al. (2008, 191757) reported that
40-70% of measured organic mass was water soluble and estimated that approximately 37%
of SOAis attributable to aromatic precursors using PM mass yields estimated for NOx-
limited conditions, while 63% of mass was unexplained, possibly due to oxidation of
semivolatile precursors not measured by standard gas chromatography. Aerosol yields from
the oxidation of aromatic compounds have been reported to be higher when reactions with
NOx are not dominant, suggesting that transport of less reactive compounds (e.g., benzene)
out of source regions with high NOx levels could result in greater overall SOA yields than
previously estimated (Ng et al., 2007, 090983). Furthermore, Zhang et al. (2007, 157186)
noted that the most common mass spectrum of oxygenated OA measured in ambient air
resembles mass spectra measured in irradiated diesel exhaust reported by Robinson et al.
(2007, 156053) and Sage et al. (2008, 191758).
Typical dilution sampling of combustion sources employ dilution rate, temperature,
pressure, and background aerosol concentrations that can differ substantially from ambient
conditions. Lipsky and Robinson (2006, 189891) and Robinson et al. (2007, 156053) showed
that under higher dilution conditions, the fraction of diesel engine organic emissions that
volatilizes is higher than that measured using common test methods.
Murphy and Pandis (2009, 190095) pointed out the importance of characterizing the
volatility distribution of emissions of organic species from combustion sources for more
accurately predicting the abundance and oxidation state of SOA in both urban and
surrounding regional background environments. They note that in urban areas, volatile
emissions can be photochemically oxidized to more non-volatile compounds which then
condense, forming oxidized SOA. Braun (2009, 189997) suggested that the weathering of
diesel exhaust particles involves the desorption of semi-volatile organic compounds followed
by the decomposition and reaction of the amorphous non-volatile carbon. These reactions
would result in the formation of a number of functional groups on the surface of the carbon
core including quinones, carbohydroxide and carboxyl groups as well as sulfate. In general,
all the above studies underscore the importance of accurately describing the phase
distribution of semivolatile organic compounds emitted by combustion sources under
atmospheric conditions, and of atmospheric photochemical reactions in modifying the
composition of emissions.
Until a few years ago, the oxidation of terpenes and aromatic compounds were
considered as sources of SOA, and the oxidation of isoprene was not considered a source of
July 2009
3-14
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
SOA. However, observations of 2-methyl tetrols in ambient samples from a number of
different environments suggest that small but not insignificant quantities of SOA are
formed (Claeys et al., 2004, 058608). Laboratory studies also indicate the formation of
2-methyl tetrols from isoprene oxidation (Edney et al., 2005, 155760; Kleindienst et al.,
2006, 156650). Xia and Hopke (2006, 179947) observed the seasonal variations for the two
major diastereoisomers produced during the oxidation of isoprene with highest
concentrations occurring during summer and lowest concentrations occurring during
winter. During summer, their maximum contribution to OC was 2.8%.
Kroll and Seinfeld (2008, 155910) and Rudich et al. (2007, 156059) noted that the
composition of SOA evolves from repeated cycles of volatilization and condensation of
chemical reaction products in both the particle and gas phases. Rudich et al. (2007, 156059)
focused on the oxidation of particle phase species by reaction with gas phase oxidants. Kroll
and Seinfeld (2008, 155910) identified three factors that determine the SOA forming
potential of organic compounds in the atmosphere:
1.	Oxidation reactions of gas-phase organic species. These species include alkanes, alkenes,
aromatics, cyclic olefins, isoprene and terpenes. Note that oxidation reactions can either lower
volatility by addition of functional groups or increase volatility by cleavage of carbon-carbon
bonds;
2.	Reactions in the particle, or condensed, phase that can change volatility either by oxidation or
formation of high-molecular-weight species. These reactions can lead to the formation of
oligomers, thereby decreasing volatility or to the formation of more volatile products; and
3.	Ongoing reactions that result from the varied volatility of oxidation products.
Other detailed work has focused on the formation of higher molecular weight
particle-phase oligomers (Gao et al., 2004, 156460; Kalberer et al., 2004, 156619; Tolocka et
al., 2004, 087578) the importance of cloud processing in the evolution of SOA (Blando and
Turpin, 2000, 155692; Gelencser and Varga, 2005, 156463) and the role of acid seeds in
oligomer formation (Tolocka et al., 2004, 087578). These results imply that ambient samples
could contain mixtures of SOA from different sources at different stages of processing, some
with common reaction products making source identification of SOA problematic.
Figure 3-4 shows a schematic of processes involved in the formation of SOA.
July 2009
3-15
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
PRIMARY	SECONDARY
products
products
products
Droplet
evaporation
Organic
Compounds
Aerosol
phase
reactions
Gas/particle partitioning
Nucleation, sorption, condensation/evaporation
Gas phase
reactions
of alkanes,
aromatics, alkenes,
olefins, etc
and OH, 03, N02
Aqueous
phase
reactions
of dicarbonyls,
organic acids
and OH, N02, 03
Figure 3-4. Primary emissions and formation of SOA through gas, cloud and condensed phase
reactions.
It should be noted that many of the products of terpene oxidation are oxidative in
nature, and are not merely nonreactive oxidation products. Organic peroxides represent an
important class of reactive oxygen species (ROS) that have high oxidizing potential and
could cause oxidative stress in cells on which these species deposit. For example, Docherty
(2005, 087613) found that 47 to 85% of SOA is composed of peroxides. Biogenic terpenoids
can be oxidized outdoors by ozone to form SOA. Various terpenoid compounds are used in a
number of household products and can be oxidized by ozone that has infiltrated from
outdoors. The oxidation of terpenoid compounds indoors produces ultrafine particles as
described in the 2006 Ozone AQCD (U.S. EPA, 2006, 088089).
3.3.2.3. Formation of new particles
In addition to emission by high temperature combustion sources, atmospheric gases
nucleate to form new particles. New particle formation has been observed in environments
ranging from relatively unpolluted marine and continental environments to polluted urban
areas (Kulmala et al., 2004, 089159). These new, nucleation mode particles are formed from
molecular clusters. Competition between condensation of gases and clusters onto existing
July 2009
3-16
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
particles and coagulation of clusters determines which process will dominate (McMurry et
al., 2005, 191759). Because of this competition, it is expected that particle number
concentrations are dominated by primary anthropogenic emissions in highly polluted
settings and by nucleation in remote continental sites. However, nucleation still occurs in
urban environments and can still be the major source under certain conditions. The
composition of ultrafine particles will differ depending on the nature of their sources.
Nucleation is observed in the morning and extending into the afternoon, and occurs at
higher rates during summer than during winter, consistent with a photochemical process.
New particle formation events have been observed to occur over distances of several
hundred kilometers in what have been called regional nucleation events (Shi et al., 2007,
191760).
The major gas phase nucleating species involved are sulfuric acid vapor and water
vapor. Kuang et al. (2008, 191196) suggested that the rate of nucleation is 2nd order with
respect to H2SO4 vapor depending on mechanism. However, other studies (e.g., Kulmala et
al., 2007, 191761) have suggested that the nucleation rate is first order with respect to
H2SO4 vapor. These differences imply that a number of mechanistic details still remain to
be determined, including the interactions with other species. However, this disparity is
small compared to classical thermodynamic binary nucleation theory involving H2SO4 and
water vapor, in which the nucleation rate is given by H2SO4 vapor to at least the 10th power
(Kulmala et al., 1998, 129111). H2SO4 vapor is produced by the gas phase oxidation of SO2
by OH radicals (U.S. EPA, 2008, 157075). Ammonia (Gaydos et al., 2005, 191762) and
organic acids and bases (amines) are also involved to some extent (Smith et al., 2008,
190037). The formation of ultrafine particles indoors from the oxidation of terpenoids by O3
as mentioned above also indicates that nucleation occurs indoors as well (U.S. EPA, 2006,
088089).
3.3.3. Mobile Source Emissions
3.3.3.1. Emissions from Gasoline Fueled Engines
PM emitted from gasoline fueled engines is a mix of OC, EC and small quantities of
trace metals and sulfates, with OC constituting anywhere from 26-88% of PM (Cadle et al.,
1999, 007636; (idler et al., 2006, 139644; Schauer et al., 2002, 035332). Most of the
compounds in OC have yet to be characterized, though constituents of unburned or
July 2009
3-17
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
partially oxidized fuel and lube oil are the dominant components of OC emitted by diesel
engines. High molecular weight and large PAHs have been identified in gasoline fueled
vehicle emissions (Phuleria et al., 2006, 156867; Riddle et al., 2007, 115272). EPA exhaust
emission standards do not control PM from gasoline vehicles as stringently as diesel
vehicles. PM emissions from gasoline fueled vehicles decreased greatly as other exhaust
emissions (primarily hydrocarbons and carbon monoxide) were controlled by improvements
in the catalytic converter and better control of air-to-fuel mixture ratios for the engine
intake. When leaded gasoline was used in pre-1975 model year vehicles, gasoline engine
PM emissions were relatively large (about 300 mg/mile) and consisted largely of lead salts
from combustion of the lead additive. A current gasoline fueled vehicle emits far lower PM,
about 1-10 mg/mile. Emissions of gasoline PM increase at colder ambient temperatures and
recent rulemaking for air toxics will result in significant reduction of gasoline PM at colder
temperatures. Further details about the composition of motor vehicle emissions in the
context of source apportionment modeling can be found in Section 3.6.1.
3.3.3.2. Emissions from Diesel Fueled Engines
Matti Maricq (2007, 155973) presents a conceptual model of diesel PM as a mix of
nucleation-mode SO r and hydrocarbons from unspent fuel and soot embedded with trace
metals on which SC>42~ and hydrocarbons condense. PM emissions from pre 2007 diesels
consist largely of EC (about 70% by mass) and OC (high molecular weight compounds
derived from both diesel fuel and lubricating oil) which is responsible for about 25% of the
PM (Matti Maricq, 2007, 155973). The elemental carbon is non-volatile, while the organic
material present exhibits temperature-dependent evaporation in similar fashion to a
mixture of C24-C32 alkanes (Sakurai et al., 2003, 113924). Sulfates constitute about 5% of
the PM. A small fraction of the diesel fuel sulfur (typically about 1-2%) is oxidized to
sulfate. Trace elements (such as zinc and halogens, mainly from lubricating oil! and others)
are also present. Mass spectra of organic diesel particles from pre-2007 engines appear to
be largely similar to engine lube oil with minor contributions from unburned diesel fuel.
Effective with the 2007 model year for on-road diesel heavy-duty highway truck
engines, the new EPA PM standard (0.01 g per brake horsepower-hour, g/bhp-hr) reduced
PM emission limits by 90% from the prior standard (0.10 g/bhp-hr). By comparison,
uncontrolled heavy-duty diesels (pre-1988 model years) emitted about 1-2 g/bhp-hour of
July 2009
3-18
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
PM. The 2007 standard resulted in the introduction of new emission control technology,
mainly the diesel particulate filter (DPF). Other elements of the new control technology also
include water-cooled exhaust gas recirculation (mostly for control of NOx), a diesel oxidation
catalyst (DOC) used in some vehicles and improved fuel injection systems.
Besides the large reduction in diesel PM on a mass basis, the composition of diesel
PM changed greatly EPA regulations required that diesel fuel for on-road vehicles contain
no more than 15 ppm sulfur as of January 2007 (Lim et al., 2007, 155931). Prior to that, the
limit on diesel fuel sulfur established in 1995 for on-road vehicles was 500 ppm. The HEI-
ACES study characterizes emissions from four engines and shows that PM emissions are
about 0.001 g/BHP-hr or 90% below the level of the current (2007) emission standard
(Shimpi et al., 2009, 189888). This study also characterizes the composition of the much
lower mass of PM emitted with this new technology. PM samples collected over a composite
type test consisted of 53% sulfate, 30% OC, 13% EC and 4% other components, including
metals. A substantial fraction of sulfur is converted to sulfate over the diesel particulate
filter resulting in the higher fractional content of sulfate emissions. However, due to the
much lower mass of PM being emitted (over a 90% reduction compared to earlier diesels) as
well as the low sulfur content of the fuel, the total mass of sulfate emitted is somewhat less
than that from earlier diesels. This work also shows that ultrafine particle number
emissions are lower (about 90% lower) and that a number of other emissions are also
controlled, including PAHs, nitro-PAHs, carbonyls (such as aldehdyes), and metals and that
their emissions are lower than those from earlier diesels.
Pre-2007 engines can be retrofitted with exhaust after treatment devices, including
DPFs, DOCs, and selective catalytic reduction (SCR) systems to reduce emissions
(U.S. EPA, 2009, 189885). Hu et al. (2009, 189886) examined emissions of various metals
(vanadium, platinum) from various diesel retrofit systems including those using vanadium -
SCR and a zeolite-based SCR with a DPF. Pakbin et al. (2009, 189893) shows significant
reductions in emissions of various PAHs for diesels with SCR retrofit systems for NOx
control. Biswas et al. (2008, 189969) examined PM size distribution and composition
(including semi-volatiles and non-volatiles) from several advanced technology diesels
including those with SCR. They showed major reductions in PM number in most driving
conditions but did not show a reduction in PM number under cruise conditions. Biswas et
al. (2009, 189880) examined four heavy-duty diesel vehicles with various retrofits showing,
July 2009
3-19
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
in general, large reductions in PM but, in some cases, somewhat higher emissions (or
smaller decreases than expected) for EC and OC.
In general, under light load conditions such as idle, diesel PM has a higher percentage
of OC emissions than at high load conditions. Under lighter loads, organic compounds are
not oxidized as effectively as under high loads. Under higher loads, PM contains more EC
than under light loads. Also, newer model year diesels through the 2006 model year tend to
have a higher fraction of PM that is EC than older models.
Emissions have been measured with the new technology engines under a number of
driving cycles besides the Federal Test Procedure. The HEI ACES study examined
emissions during a range of test procedures. In general, the lowload test cycle resulted in
lower exhaust temperatures and higher emission than did the high-load cycles.
Regeneration events also produced short-term increases in particle emissions. However,
particle emission measurements on the ACES engines were consistently lower than those
on a typical 2004 engine (Shimpi et al., 2009, 189888).
EPA standards will result in non-road diesels also having technology like catalyzed
diesel particulate filters starting in 2012. Similar standards have also been promulgated for
locomotives powered by diesel engines. Some work has been done with prototype SCR
systems for diesel NOx control such as would be used for the 2010 diesel NOx standard.
This standard will also result in reductions for NOx similar to those seen for PM in the
2007 standard.
There is no information on emissions from diesel engines with this new technology, at
temperatures of < 10° C. In general, the ratio of emissions under cold start conditions at low
temperatures to emissions at 24° C is significantly higher for diesel engines with new
technology compared to emissions from non-catalyst systems. Note that the engines with
the newer technology require time to allow the catalyst to reach normal operating
temperatures for full emission reductions. During the period of catalyst heating, particles
will be trapped in the DPF, but volatile components can pass through. As a result of this
particle-trapping, post-2007 model year engines still emit less than the older engines, even
under cold start conditions.
July 2009
3-20
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
3.3.4. Deposition of PM
Wet and dry deposition are important processes for removing PM and other pollutants
from the atmosphere on urban, regional, and global scales. The conceptual model for dry
deposition is to view the flux deposited on a surface as the product of a concentration (mass
or moles of a pollutant/m3) times a deposition velocity (Vd) (m/s). Therefore, deposition has
the units of mass per unit time per unit area, or flux. The general approach used to
estimate Vd is the resistance-in-series method represented by Equation 3.1:
Vd=l/(Ra+Rb+Rc)
Equation 3-1
where Ra, Rb, and H, represent the resistance due to atmospheric turbulence, transport in
the fluid sublayer very near the elements of surface such as leaves or soil, and the
resistance to uptake of the surface itself, respectively. These processes are shown
schematically in Figure 3"5. This approach works for a range of substances, although it is
inappropriate for species with substantial re-emissions from the surface or for species
where deposition to the surface depends on concentrations at the surface itself. The
approach is also modified somewhat for aerosols in that the terms Rb and Rc are replaced
with a surface Vd to account for gravitational settling.
July 2009
3-21
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
aerodynamic
Atmospheric
Resistances
'laminar' sub-layer
cuticular
Canopy ^
Resistances
soil
c3
Resistance analogy for the deposition of atmospheric pollutants
Source: Courtesy of T. Pierce, USEPAI ORDI NERL / Atmospheric Modeling Division.
Figure 3-5.	Schematic of the resistancein-series analogy for atmospheric deposition.
Function of wind speed, solar radiation, plant characteristics, precipitation/moisture,
and soil/air temperature.
Wesley and Hicks (2000, 025018) listed several shortcomings of the then-current
knowledge of dry deposition. Among those shortcomings were difficulties in representing
dry deposition over varying terrain where horizontal advection plays a significant role in
determining the magnitude of Ra and difficulties in adequately determining Vd for
extremely stable conditions such as those occurring at night! see the discussion by Mahrt
(1998, 048210). Under optimal conditions, when a model is exercised over a relatively small
area where dry deposition measurements have been made, models still generally showed
uncertainties on the order of ± 30% (see, e.g.,Brook et al., 1996, 024023; Massman et al.,
1994, 043681; Padro, 1996, 052446; Wesely and Hicks, 2000, 025018).
Wesley and Hicks (2000, 025018) concluded that an important result of those
comparisons was that the level of sophistication of most dry deposition models was
July 2009
3-22
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
relatively low, and that deposition estimates, therefore, must rely heavily on empirical data.
Still larger uncertainties exist when the surface features in the built environment are not
well known or when the surface comprises a patchwork of different surface types, as is
common in the eastern U.S.
3.3.4.1. Deposition Forms
Wet Deposition
Wet deposition results from the incorporation of atmospheric particles and gases into
cloud droplets and their subsequent precipitation as rain or snow, or from the scavenging of
particles and gases by raindrops or snowflakes as they fall (Lovett, 1994, 024049). Wet
deposition depends on precipitation amount and ambient pollutant concentrations.
Receptor (i.e., vegetation) surface properties have little effect on wet deposition, although
leaves can retain liquid and solubilized PM. In terrain containing extensive vegetative
canopies, any material deposited via precipitation to the upper stratum of foliage is likely to
be intercepted by several foliar surfaces before reaching the soil. This allows such processes
as foliar uptake, chemical transformation, and re-suspension into the atmosphere to occur.
Landscape characteristics can affect wet deposition via orographic effects and by the
closer aerodynamic coupling to the atmosphere of tall forest canopies as compared to the
shorter shrub and herbaceous canopies. Following wet deposition, humidity and
temperature conditions further affect the extent of drying versus concentrating of solutions
on foliar surfaces, which influence the rate of metabolic uptake of surface solutes (Swietlik
and Faust, 1984, 046678). The net consequence of these factors on direct physical effects of
wet deposited PM on leaves is not known (U.S. EPA, 2004, 056905).
Rainfall introduces new wet deposition and also redistributes throughout the canopy
previously dry-deposited particles (Peters and Eiden, 1992, 045277). The concentrations of
suspended and dissolved materials are typically highest at the onset of precipitation and
decline with duration of individual precipitation events (Hansen et al., 1994, 046634).
Sustained rainfall removes much of the accumulation of dry-deposited particles from foliar
surfaces, reducing direct foliar effects and combining the associated chemical burden with
the wet-deposited material (Lovett, 1994, 021019) for transfer to the soil. Intense rainfall
may contribute substantial total particulate inputs to the soil, but it also removes
bioavailable or injurious pollutants from foliar surfaces. This washing effect, combined with
July 2009
3-23
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
differential foliar uptake and foliar leaching of different chemical constituents from
particles, alters the composition of the rainwater that reaches the soil and the pollutant
burden that is taken-up by plants. Once in the soil, these particle constituents may affect
biogeochemical cycles of major, minor, and trace elements. Low intensity precipitation
events, in contrast, may deposit significantly more particulate pollutants to foliar-surfaces
than high intensity precipitation events. Additionally, lowintensity events may enhance
foliar uptake through the hydrating of some previously dry-deposited particles (U.S. EPA,
2004, 056905)
Dry Deposition
Dry particulate deposition, especially of heavy metals, base cations, and organic
contaminants, is a complex and poorly characterized process. It appears to be controlled
primarily by such variables as atmospheric stability, macro- and micro-surface roughness,
particle diameter, and surface characteristics (Hosker RP and Lindberg, 1982, 019118). The
range of particle sizes, the diversity of canopy surfaces, and the variety of chemical
constituents in airborne particles have made it difficult to predict and to estimate dry
particulate deposition (U.S. EPA, 2004, 056905).
Dry deposition of atmospheric particles to plant and soil surfaces affects all exposed
surfaces. Larger particles >5 pm diameter are dry-deposited mainly by gravitational
sedimentation and inertial impaction. Smaller particles, especially those with diameters
between 0.2 and 2.0 pm, are not readily dry-deposited and may travel long distances in the
atmosphere until their eventual deposition, most often via precipitation. Plant parts of all
types, along with exposed soil and water surfaces, receive steady deposits of dry dusts, EC,
and heterogeneous secondary particles formed from gaseous precursors (U.S. EPA, 1982,
017610).
Estimates of regional particulate dry deposition infer fluxes from the product of
variable and uncertain measured or modeled particulate concentrations in the atmosphere
and even more variable and uncertain estimates of Vd parameterized for a variety of
specific surfaces (e.g.,Brook et al., 1996, 024023). Even for specific sites and well-defined
particles, uncertainties are large. Modeling the dry deposition of particles to vegetation is at
a relatively early stage of development, and it is not currently possible to identify a best or
most generally applicable modeling approach (U.S. EPA, 2004, 056905).
July 2009
3-24
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Deposition from Clouds and Fog
The occurrence of cloud and fog deposition tends to be geographically restricted to
coastal and high mountain areas. Several factors make it particularly effective for the
delivery of dissolved and suspended particles to vegetation. Concentrations of particulate-
derived materials are often many-fold higher in cloud or fog water than in precipitation or
ambient air due to orographic effects and gas-liquid partitioning. In addition, fog and cloud
water deliver particulate chemical species in a bioavailable-hydrated form to foliar surfaces.
This enhances deposition by sedimentation and impaction of submicron aerosol particles
that exhibit low Vd before fog droplet formation (Fowler et al., 1989, 002515). Deposition to
vegetation in fog droplets is proportional to wind speed, droplet size, concentration, and fog
density. In some areas, typically along foggy coastlines or at high elevations, this deposition
represents a substantial fraction of total deposition to foliar surfaces (Fowler et al., 1991,
046630).
3.3.4.2. Methods for Estimating Dry Deposition
Methods for estimating dry deposition of particles are more restricted than for
gaseous species and fall into two major categories: surface analysis methods, which include
all types of estimates of contaminant accumulation on surfaces of interest! and atmospheric
deposition rate methods, which use measurements of contaminant concentrations in the
atmosphere and descriptions of surrounding surface elements to estimate deposition rates
(Davidson and Wu, 1990, 036799). Surface extraction or washing methods characterize the
accumulation of particles on natural receptor surfaces of interest or on experimental
surrogate surfaces. These techniques rely on methods designed specifically to remove only
surface-deposited material. Total surface rinsate may be equated to accumulated deposition
or to the difference in concentrations in rinsate between exposed and control (sheltered)
surfaces and may be used to refine estimates of deposition. Foliar extraction techniques
may underestimate deposition to leaves because of uptake and translocation processes that
remove pollutants from the leaf surface (Garten CT and Hanson, 1990, 036803; Taylor GE
et al., 1988, 019289). Foliar extraction methods also cannot distinguish gas from particle-
phase sources (Bytnerowicz et al., 1987, 036493; Dasch, 1987, 036496; Kelly, 1988, 037379;
Lindberg and Lovett, 1985, 036530; Van Aalst, 1982, 036481).
July 2009
3-25
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
The National Dry Deposition Network was established in 1986 to document the
magnitude, spatial variability, and trends in dry deposition across the United States.
Currently, the network operates as a component of the CASTNet (Clarke et al., 1997,
025022). A significant limitation on current capacity to estimate regional effects of NOx and
SOx deposition is inadequate knowledge of the mechanisms and factors governing particle
dry deposition to diverse surfaces (U.S. EPA, 2004, 056905).
Dry deposition cannot be directly measured. Deposition rates and totals are often
calculated as the product of measured ambient concentration and a modeled Vd. This
method is widely used because atmospheric concentrations are easier to measure than are
dry deposition rates, and models have been developed to estimate Vd. Ambient pollutant
concentrations and meteorological conditions required for application of inferential models
are routinely collected at CASTNet dry deposition sites. Monitored chemical species are
limited to O3, SO42 , NCV, NH4"1", SO2, and HNO3. The temporal resolution for the ambient
concentration measurements and dry deposition flux calculations is hourly for O3 and
weekly for all the other species (Clarke et al., 1997, 025022).
Collection and analysis of stem flow and throughfall can also provide useful estimates
of particulate deposition when compared to directly sampled precipitation. The method is
most precise for particle deposition when gaseous deposition is a small component of the
total dry deposition and when leaching or uptake of compounds of interest out of or into the
foliage is not a significant fraction of the deposition because these lead to positive and
negative artifacts in the calculated totals.
Foliar washing, whether using precipitation or experimental lavage, is one of the best
available methods to determine dry deposition to vegetated ecosystems. Major limitations
include the site specificity of the measurements and the restriction to elements that are
largely conserved within the vegetative system. Surrogate surfaces have not been found
that can adequately replicate essential features of natural surfaces! and therefore do not
produce reliable estimates of particle deposition to the landscape.
Micrometeorological methods employ eddy covariance, eddy accumulation, or flux
gradient protocols for quantifying dry deposition. These techniques require measurements
of particulate concentrations and of atmospheric transport processes. They are currently
well developed for ideal conditions of flat, homogeneous, and extensive landscapes and for
chemical species for which accurate and rapid sensors are available. Additional studies are
needed to extend these techniques to more complex terrain and more chemical species.
July 2009
3-26
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
The eddy covariance technique measures vertical fluxes of gases and fine particles
from calculations of the mean covariance between the vertical component of wind velocity
and pollutant concentration (Wesely et al., 1982, 036564) using sensors acquiring
concentration data at 5 to 20 Hz. For the flux gradient or profile techniques, vertical fluxes
are calculated from a concentration difference and an eddy exchange coefficient determined
at discrete heights (Erisman et al., 1988, 036510; Huebert et al., 1988, 036569). Most
measurements of eddy transport of PM have used chemical sensors (rather than mass or
particle counting) to focus on specific PM components. These techniques have not been well
developed for generalized particles and may be less suitable for coarse particles that are
transported efficiently in high frequency eddies (Gallagher et al., 1988, 046631).
3.3.4.3. Factors Affecting Dry Deposition Rates and Totals
In the size range of-0.1-1.0 pm where Vd is relatively independent of particle
diameter as shown in Figure 3"6, particulate deposition is controlled by roughness of the
surface and by the stability and turbulence of the atmospheric surface layer. Impaction and
interception dominate over diffusion as dry deposition processes, and the Vd is considerably
lower than for particles that are either smaller than ~0.1 um or larger than ~1.0 um (Shinn,
1978, 071426).
Deposition of particles between 1 and 10 pm diameter is strongly dependent on
particle size (Shinn, 1978, 071426). Larger particles within this size range are collected
more efficiently at typical wind speeds than are smaller particles (Clough, 1975, 070850).
suggesting the importance of impaction. Impaction is related to wind speed, the square of
particle diameter, and the inverse of receptor diameter as a depositing particle fails to
follow the streamlines of the air in which it is suspended around the receptor. When
particle trajectory favors a collision, increasing either wind speed or the ratio of particle
size to receptor cross-section increases the probability of collision.
July 2009
3-27
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
1,000
Stokes Law
Brownian Diffusion
Peters and Enden (1992)
Lritle and Wffen (1977)
100 —
in
£
u
10 —
o
o
2
e
o
1
0.1 —
0.01 —
w
o
CL
O)
o
0.001 —
0.000
0.001
0.01
0.1
1
10
100
Particle Diameter (Mm)
Source: U.S. EPA (2004,056905).
Figure 3-6. The relationship between particle diameter and Vd for particles. Values measured in
wind tunnels by Little and Wiffen (1977, 070869) over short grass with wind speed of
2.5 mi/s closely approximate the theoretical distribution determined by Peters and Eiden
(1992, 045277) for a tall spruce forest. These distributions reflect the interaction of
Brownian diffusivity (descending dashed line), which decreases with particle size and
sedimentation velocity (ascending dotted line derived from Stokes Law), which
increases with particle size. Intermediate-sized particles (0.1-1.0 Dm) are influenced
strongly by both particle size and sedimentation velocity, and deposition is relatively
independent of size.
Empirical estimates of Vd for fine particles under wind tunnel and field conditions are
often several-fold greater than predicted by theory (Unsworth and Wilshaw, 1989, 046682).
A large number of transport phenomena, including streamlining of foliar obstacles,
turbulence structure near surfaces, and various phoretic transport mechanisms are not well
characterized (U.S. EPA, 2004, 056905). The discrepancy between estimated and predicted
values of Vd may reflect model limitations or experimental limitations in the specification of
the effective size and number of receptor obstacles. Previous reviews (e.g., U.S. EPA, 1996,
079380; U.S. EPA, 2004, 056905) suggest the following generalizations: (l) particles >10 pm
exhibit variable Vd between 0.5 and 1.1 cm/s depending on friction velocities, whereas a
minimum particle Vd of 0.03 cm/s exists for particles in the size range 0.1 to 1.0 pm, (2) the
Vd of particles is approximately a linear function of friction velocity, and (3) deposition of
particles from the atmosphere to a forest canopy is from 2 to 16 times greater than
deposition in adjacent open terrain like grasslands or other low vegetation.
July 2009
3-28
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Leaf Surface Effects on Vd
The chemical composition of a particle is not usually considered to be a primary
determinant of its Vd. Rather, the plant leaf surface has an important influence on the Vd of
particles, and therefore on the flux of dry deposition to the terrestrial environment.
Relevant leaf surface properties include stickiness, microscale roughness, and cross-
sectional area. These properties affect the probability of impaction and particle bounce. The
efficiency of deposition to vegetation also varies with leaf shape. Particles impact more
frequently on the adaxial (upper) surface than on the abaxial (lower) surface. Most particles
accumulate in the midvein, central portion of leaves. The greatest particle loading on
dicotyledonous leaves is frequently on the adaxial surface at the base of the blade, just
above the petiole junction. Precipitation washing probably plays an important role in this
distribution pattern (U.S. EPA, 2004, 056905).
Lead particles have been shown to accumulate to a greater extent on older as
compared with younger needles and twigs of white pine, suggesting that wind and rain may
be insufficient to fully wash the foliage. Fungal mycelia (derived from windborne spores)
were frequently observed in intimate contact with other particles on leaves, which may
reflect minimal re-en train men I of the spore due to shelter by the particles, mycelia
development near sources of soluble nutrients provided by the particles, or simply co-
deposition (Smith and Staskawicz, 1977, 046675).
Leaves with complex shapes tend to collect more particles than do those with shapes
that are more regular. For example, conifer needles are more efficient than broad leaves in
collecting particles by impaction as a result of the small cross-section of the needles relative
to the larger leaf laminae of broadleaves allowing for greater penetration of wind into
conifer canopies than broadleaf ones (U.S. EPA, 2004, 056905).
Canopy Surface Effects on Vd
Surface roughness increases particulate deposition, and Vd is usually greater for a
forest than for a nonforested area and greater for a field than for a water surface. Different
size particles have different transport properties and Vd. The upwind leading edges of
forests, hedgerows, and individual plants are primary sites of coarse particle deposition.
Impaction at high wind speed and the sedimentation that follows the reduction in wind
speed and carrying capacity of the air in these areas lead to preferential deposition of larger
particles (U.S. EPA, 2004, 056905).
July 2009
3-29
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Air movement is slowed in proximity to vegetated surfaces. Canopies of uneven age or
with a diversity of species are typically aerodynamically rougher and receive larger inputs
of dry deposited pollutants than do smooth, low, or monoculture vegetation (Garner et al.,
1989, 042085; U.S. EPA, 2004, 056905). Canopies on slopes facing the prevailing winds
receive larger inputs of pollutants than more sheltered, interior canopy regions.
All foliar surfaces within a forest canopy are not equally exposed to particle
deposition. Upper canopy foliage tends to receive maximum exposure to coarse and fine
particles, but foliage within the canopy tends to receive primarily fine aerosol exposures.
The dry deposition of fine-mode particles and unreactive gases tends to be more evenly
distributed throughout the canopy.
Both uptake and release of PM constituents can occur within the canopy. The leaf
surface is a region of leaching and uptake. Exchange also occurs with epiphytic organisms
and bark and through solubilization of previously dry-deposited PM. Vegetation emits a
variety of particles and particulate precursor materials.
3.4. Monitoring of PM
3.4.1. Ambient Measurement Techniques
3.4.1.1. PM Mass
Federal Reference Methods (FRMs) and Federal Equivalent Methods (FEMs) for PM
were discussed in detail in the 2004 PM AQCD (U.S. EPA, 2004, 056905). Issues discussed
there include the definition or description of FRMs and FEMs for PMio, PM2.5, and
measurement methods for PM10-2.5 with detailed descriptions of the WINS impactor, virtual
and cascade multistage impactors for PM10-2.5 measurement, high-volume and low-volume
PM10 samplers, and real-time or continuous methods for PM10 and PM2.5 including:
¦	Tapered Element Oscillating Microbalance (TEOM) operated at various
temperatures;
¦	Sample Equilibration System (SES)-TEOM;
¦	Differential TEOM;
¦	B-Gauge Techniques (BGT);
July 2009
3-30
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
¦	Piezoelectric Microbalance!
¦	Real-Time Total Ambient Mass Sampler (RAMS);
¦	Continuous Ambient Mass Monitor (CAMM);
¦	Continuous Coarse Particle Monitor (CCPM);
¦	Micro-orifice Uniform Deposit Impactor (MOUDI);
¦	Multichannel diffusion denuder sampling system (BOSS); and
¦	Light scattering photometric instruments.
In this section, FRMs and FEMs for PMio, PM2.5, and PM10-2.5 will be revisited and
evaluated based on the cumulative understanding of these methods with a focus on
evaluations performed following the 2004 PM AQCD, followed by the discussion of new
techniques under development or evaluation.
Federal Reference Method and Federal Equivalent Method
The FRM and FEM are designed to measure the mass concentrations of ambient
particles. The FRMs for measuring PM10, PM2.5, and PM10-2.5 are specified in CFR 40 Part
50, Appendices J, L, and O, respectively. The PM10 FRM is a performance-based method in
which particles are inertially separated with a penetration efficiency of 50% at 10 ± 0.5 pm
aerodynamic diameter. The collection efficiency specified in the CFR approaches 100% as
particle size decreases and approaches 0% as particle size increases. Particles are collected
on filters for which mass concentrations are determined gravimetrically. In contrast, the
FRM for PM2.5 is a design-based method that specifies certain details of the sampler design,
as well as of sample handling and analysis, whereas other aspects (e.g., flow control) have
performance specifications (U.S. EPA, 2004, 056905) PM10-2.5 concentration is computed as
the difference between concurrent and co-located PM10 and PM2.5 concentrations obtained
from co-located FRM samplers of the same make and model. It should be noted that the
FRM for PM10 operates and reports data corrected to standard temperature (298K) and
pressure (101.3 kPa) (STP), while the FRM for PM2.5 and PM10-2.5 operates and reports data
under local conditions.
Avery sharp cut cyclone (VSCC) was approved in 2004 as a Class II PM2.5 FEM
method (Kenny et al., 2004, 155895). The VSCC provides superior performance over long
sampling periods under heavy loading and was also incorporated as an optional second-
July 2009
3-31
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
stage separator for the PM2.5 FRM (71 FR 61214, October 17, 2006). In 2006, EPAfinalized
new performance criteria (40 CFR Part 53) for the approval of FEMs as Class II equivalent
methods when based on integrated filter sampling and as Class III equivalent methods
when based on continuous technologies that can provide at least hourly data reporting. The
performance criteria include evaluating additive bias (intercept) and multiplicative bias
(slope) as well as correlation from co-located candidate and FRM methods at field studies
covering multiple seasons and locations. As a result of these new performance criteria, EPA
has recently approved two PM2.5 and two PM10-2.5 Class II FEMs based on the virtual
impactor techniques (dichotomous sampler), and five PM2.5 and one PM10-2.5 Class III FEMs
based on BGT or TEOM techniques. The most recent list of FRMs and FEMs can be found
in Annex A, Table A-2 and on the following EPA web site:
http://www.epa.gov/ttn/amtic/files/ambient/criteria/reference-eauivalent-methods-list.pdf.
Evaluations of FRMs and FEMs were conducted both in super site studies and in other
research studies (Ayers, 2004, 097440; Brown et al., 2006, 097665; Butler et al., 2003,
156313; Cabada et al., 2004, 148859; Chang and Tsai, 2003, 155718; Charron et al., 2004,
053849; Cowburn et al., 2008, 191142; Grover et al., 2005, 156500; Hains et al., 2007,
091039; Hering et al., 2004, 155837; Jaques et al., 2004, 155878; Krieger et al., 2007,
129657; Lee et al., 2005, 156679; Lee et al., 2005, 155924; Price et al., 2003, 098082; Rees et
al., 2004, 097164; Russell et al., 2004, 156061; Salminen and Karlsson, 2003, 156070;
Schwab et al., 2004, 098450; Schwab et al., 2006, 098449; Solomon et al., 2003, 156994; Tsai
et al., 2006, 098312; Vega et al., 2003, 105974; Wilson et al., 2006, 091142; Yi et al., 2004,
156169; Zhu et al., 2007, 098367) (see Annex A, Tables A-3, A-5 and A-11). In general, the
co-located FRMs showed very good precision with coefficient of variation (CV) <5%. For
different co-located FRMs, the regression slope of one sampler on another is commonly close
to unity and with R2 >0.95. The PM2.5 and PM10 concentrations measured by dichotomous
samplers were within 10% of the FRM methods, and the differences can be attributed to the
sampling artifacts of semi-volatile components! see Section 3.4.1.2 for details. The precision
of various TEOMs ranges from 10-30%. The concentration measured by the TEOM operated
at 50°C was consistently lower than those measured by the TEOM operated at 30°C. The
differences between these monitors were also found to be a function of season and location.
BGTs were highly correlated with FRMs but BGT mass could be higher than the FRM mass
(30% higher at the Fresno supersite) (Chow et al., 2008, 156355). CAMMs did not show a
consistent pattern when compared with FRMs. These differences could be largely attributed
July 2009
3-32
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
to the sampler operating principles and design, ambient conditions (copollutants and
meteorological parameters), and the built-in default calibration factors for non-FRM/FEM
instruments. Additionally, a number of techniques have been developed to reduce positive
and negative sampling artifacts. These are described in the ISA for NOx and SOx -
Environmental Criteria (U.S. EPA, 2008, 157074).
For PM2.5 and PM10, it has long been known that FRMs are subject to sampling
artifacts, i.e., the loss of semi-volatile components of PM (e.g., NH4NO3, and some organics).
Therefore, the FRMs cannot provide the "true" PM mass concentrations. The accuracy of
the method cannot be directly evaluated due to the lack of standard reference materials.
However, in comparison with other sampling techniques that can measure both semi-
volatile and nonvolatile PM, FRMs reported PM2.5 or PM10 mass concentrations biased low
by 10-30%. The bias of the FRMs depends on the components of ambient PM and the
sampling conditions (e.g., ambient temperature and relative humidity), which vary from
day to day and from season to season. Another limitation of the current FRM sampling
frequency is that filter samples are often collected every three or six days (approximately
150 of the 900 plus PM2.5 FRMs operating in 2007 were scheduled to sample each day).
Under this operating condition, the concentration-response relationship in air pollution
health studies (especially in time-series studies) cannot be fully evaluated in terms of lag
structures and distributed lags between ambient concentration and health outcome
(Lippmann, 2009, 190083; Solomon and Hopke, 2008, 156997). Despite these issues, the
precision of the FRMs are usually quite high (CV <5%), thereby providing consistent and
reliable PM measurements.
Development and Evaluation of New Techniques
Several new innovations have recently emerged to measure both fine and coarse PM
fractions in the ambient air. These techniques include the Filter Dynamics Measurement
System-TEOM (FDMS-TEOM) (Grover et al., 2006, 138080) for real-time measurement of
PM2.5 or PM10 and several new methods for measurement of PM10-2.5. In addition, several
new techniques exist for measuring ultrafine particles (discussed later in Section 3.4.1.4)
and for estimating PM mass concentration indirectly using particle size (discussed later in
Section 3.4.1.5).
July 2009
3-33
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Real-time Measurement of PM2.5 or PM10 using the FDMS-TEOM
The FDMS-TEOM incorporates self-referencing capability to the traditional TEOM by
alternating measurement of ambient air and chilled clean air (particles and semivolatile
gases are removed by filtration at 4°C after which the clean air is reheated to 30°C) in
6-min intervals. As clean air flows over the sample filter, the semivolatile PM on the sample
filter is evaporated. Thus, the instrument provides direct measurements of the monvolatile
particle mass and incorporates an adjustment for the semivolatile NH4NO3 and organic
material. In a comparison between the TEOM, FDMS-TEOM, and FRM mass, the PM2.5
concentration measured by the TEOM operated at 50°C was consistently lower than those
measured by the TEOM operated at 30°C. The TEOM operated at 30°C provided
concentrations 50% lower than the FDMS-TEOM, and the FDMS-TEOM provided
concentrations 10-30% higher than the FRM mass (Chow et al., 2008, 156355; Schwab et
al„ 2006, 098449).
Techniques for Measurement of PM10 2.5
Methods developed to measure PM10-2.5 are based on three measurement techniques^
l) virtual impactors using lowvolume, high-volume, and real-time techniques! 2) cascade
impactors! and 3) passive samplers.
A lowvolume dichotomous sampler (operated at 16.7 L/min), based on virtual
impaction, was described in the 2004 AQCD (U.S. EPA, 2004, 056905). Since then, a high-
volume dichotomous sampler was developed to operate at 1,000 L/min (Sardar et al., 2006,
156071). The sampler was evaluated in the field by comparison with a MOUDI sampler,
and the measured PM10-2.5 mass concentrations were within 10%. The high-volume
dichotomous sampler provides sufficient mass collection for comprehensive standard
chemical analyses over short sampling intervals. Using the virtual impactor technique in
conjunction with a TEOM or BGT as the detector, continuous PM10-2.5 measurement
techniques were developed (Miller GT, 1975, 018988; Misra et al., 2003, 180217; Solomon
and Sioutas, 2008, 190139). The TEOM method was highly correlated with the PM10-2.5
FRM, but mass concentrations measured by the TEOM were 20-30% lower (Solomon and
Sioutas, 2008, 190139). The BGT method also agreed well with the PM10-2.5 FRM (slopes =
0.88-1.17, and R2 >0.95) (Solomon and Sioutas, 2008, 190139).
Case et al. (2008, 155149) evaluated a cascade PM sampler designed to collect PM10-2.5
on a foam impactor. Particle bouncing on impactors has long been a concern for PM
July 2009
3-34
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
collection. Porous foam was used to serve as the impactor substrate to reduce particle
bounce and to collect relatively large amounts of particles (Demokritou et al., 2004, 186901;
Demokritou et al., 2004, 190115; Huang et al., 2005, 186991; Kuo et al., 2005, 186997). The
sampler was operated at 5 L/min, and it agreed with a lowvolume dichotomous sampler
within ±20%. The precision of the sampler was 20% as determined by the CV.
An inexpensive passive sampler for PM10-2.5 was also developed (Leith et al., 2007,
098241; Ott et al., 2008, 191765; Wagner and Leith, 2001, 190153; Wagner and Leith, 2001,
190154). The passive sampler collects particles by gravity, diffusion, and convective
diffusion onto a glass coverslip, and then an image analysis is conducted on the collected
particles to estimate mass flux as a function of aerodynamic diameter. Leith et al. (2007,
098241) conducted a field evaluation of the passive sampler, and the difference between an
FRM and the co-located passive sampler was within lo of concentrations measured with
PM10-2.5 FRM samplers. Ott et al. (2008, 191765) reported the precision of the sampler was
11.6% (CV), and the detection limit was 2.3 |a,g/m3 for a 5-day sample.
3.4.1.2. PM Speciation
The following sections describe recent developments regarding measurement
techniques to ascertain quantities of particle-bound water, cations and anions, elemental
composition, carbon, and organic species.
Particle-Bound Water
Particle-bound water is an important component of ambient PM (U.S. EPA, 2006,
157071). Recently, a differential method was developed to measure particle-bound water
(Santarpia et al., 2004, 156944; Stanier et al., 2004, 095955). The dry ambient aerosol size
spectrometer (DAASS) can measure particle-bound water in the particle size range from 3
nm-10 |um (Stanier et al., 2004, 095955), by alternatively measuring ambient PM size
distribution at low relative humidity (RH) and ambient RH. A comparison of the two size
distributions provides information on the water absorption and change in particle size due
to RH. Khlystov et al. (2005, 156635) reported that the particle-bound water, measured by
DAASS, was underestimated for particles <200 nm and overestimated for particles >200
nm compared with thermodynamic models. The loss of semi-volatile components during
measurement may bias the particle-bound water measurement results. Methods and
analytical specifications for particle-bound water are listed in Annex A, Table A-12.
July 2009
3-35
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Cations and Anions
The measurement of cations and anions including SO42 , NO3", NH4"1", CI", Na+, and K+
still relies primarily on filter-based collection, water based extraction and ion
chromatography (IC) based chemical speciation and quantification. In addition, denuders
are frequently used in the sampling system to adjust for sampling artifacts. These methods
have been reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905). Filter-denuder based
integrated sampling methods for SO42 , NOa , and NH4"1" have been detailed in the 2008 SOx
ISA and 2008 NOxSOx ISA (U.S. EPA, 2008, 157075; U.S. EPA, 2008, 157074).
Recent developments in multiple ion measurements have focused on the coupling of
IC and a sample dissolution system, represented by the Particle into Liquid Sampler-Ion
Chromatography (PILS-IC) and the Ambient Ion Monitor (AIM) (Dasgupta et al., 2007,
156383; Orsini et al., 2003, 156008; Weber et al., 2001, 024640). When ambient PM passes
through the PILS-IC system, water droplets are generated by mixing ambient PM with
saturated water vapor and collected by impaction. The resulting liquid stream is then
introduced into the IC system for ion speciation and quantification. Hourly concentrations
of multiple ions can be obtained with the system, with a CV of 10%. For the AIM system, a
parallel plate denuder is used to remove the interfering gases, and then particles enter a
super-saturation chamber to form droplets. The collected droplets are then introduced into
the IC for analysis. The AIM system can provide hourly concentrations for multiple ions.
The particle mass spectrometer is another advance in multiple PM component
measurements, but most of these types of measurements are semi-quantitative and will be
detailed later in Section 3.4.1.3. Note that measurement and analytical specifications for
ions other than SOr and NOa are listed in Annex A, Table A-9.
Sulfate
Methods used for continuous (sampling interval of minutes) measurements of SO r
include Aerosol Mass Spectrometry (AMS) (Drewnick et al., 2003, 099160; Hogrefe et al.,
2004, 156560), PILS-IC (Weber et al., 2001, 024640), flash volatilization techniques (Bae,
2007, 155669; Stolzenburg and Hering, 2000, 013289) and the Harvard School of Public
Health (HSPH) tube furnace to convert SO r to SO2 for detection by a SO2 analyzer (Allen
et al., 2001, 156205). These methods are described in detail by Drewnick et al. (2003,
099160), along with an inter-sampler comparison that found overall agreement within 2.9%
for all continuous instruments with R2 of 0.87 or better. When compared with filter
July 2009
3-36
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
samples, Drewnick et al. (2003, 099160) showed differences were less than 25% for the
AMS, PILS, flash vaporization, and HSPH continuous SC>42~ monitors. The Thermo 5020
particulate sulfate analyzer (based on the HSPH technique) compared within 80% of 24-h
filter-based measurement at a rural site in New York (Schwab et al., 2006, 098449). Annex
A, Tables A-8 and A-14, list detailed methods and analytical specifications for sampling
S042-.
N it rate
In addition to the nylon filter-based method and the new developments mentioned for
SO42-, methods based on flash volatilization-chemiluminescence analysis and catalytic
conversion-chemiluminescence analysis have also been developed for continuous NO3
measurement (averaging time 30 seconds to 10 minutes). For the flash volatilization system
(Fine et al., 2003, 155775; Stolzenburg and Hering, 2000, 013289; Stolzenburg et al., 2003,
156102), particles are collected by a humidified impaction process and analyzed in place by
flash vaporization and chemiluminescent detection of the evolved NOx. For the catalytic
conversion-chemiluminescence analysis system (Weber et al., 2003, 157129), NC>3~ was
measured by conversion of particle N( )a into NO, and then detected with the
chemiluminescence method. Field and lab comparisons were conducted to compare the
different instruments mentioned above. Although the R&P 8400N ambient particulate NO:;
monitor, which is based on the Stolzenburg flash vaporization technique, could provide
10-min resolution data and showed excellent precision (with a CV <10%) (Harrison et al.,
2004, 136787; Hogrefe et al., 2004, 156560; Long and McClenny, 2006, 098214; Rattigan et
al., 2006, 115897), it consistently reported N03~ concentrations -30% lower than the
denuder-filter systems in both the Baltimore supersite and the multiyear field study in New
York (Harrison et al., 2004, 136787; Hogrefe et al., 2004, 156560; Rattigan et al., 2006,
115897). In the New York measurement campaign, an AMS was also co-located with other
instruments to obtain the real-time NOa information. AMS did not always agree well with
the denuder-filter system for reasons not entirely apparent. However, Bae et al. (2007,
156244) reported that some organic compounds can also produce signals at mass-to-charge
ratio m/z- 30, which is one of the characteristic m/z for NOa . Therefore, the disagreement
between the AMS and the filter-based method could be a result of the interference of
organic compounds using the AMS. Annex A, Tables A-7 and A-13, list methods and
analytical specifications for sampling NO3-.
July 2009
3-37
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Ammonium
Several continuous and semi-continuous instruments can be used to monitor ambient
ammonium concentrations (Al-Horr et al., 2003, 153951; Bae, 2007, 155669) including
many listed above for SO42- and NO3-. Bae et al. (2007, 155669) conducted an
inter-comparison of three semi-continuous instruments during the New York multiyear air
sampling campaign: a PILS-IC, an AMS, and a wet scrubbing-long path absorption
photometer. Bae et al. (2007, 155669) reported the inter-sampler coefficients of
determination (R2) between these instruments were above 0.75, and the slopes (with zero
intercept) were between 0.71 and 1.04. Annex A, Table A-9 describes measurement of ions
other than N( )a and S() r , including NH4"1".
Elemental Composition
Techniques for measuring the elemental composition of PM samples were reviewed in
the 2004 PM AQCD (U.S. EPA, 2004, 056905). These methods include:
¦	Energy dispersive X-ray fluorescence (ED-XRF);
¦	Synchrotron X-ray fluorescence (SXRF);
¦	Particle-induced X-ray emission (PIXE);
¦	Particle elastic scattering analysis (PESA);
¦	Total reflection X-ray fluorescence (TR-XRF);
¦	Instrumental neutron activation analysis (INAA);
¦	Atomic absorption spectrophotometry (AAS);
¦	Inductively-coupled plasma-atomic emission spectroscopy (ICP-AES);
¦	Inductively-coupled plasma-mass spectrometry (ICP-MS); and
¦	Scanning electron microscopy (SEM).
Recent development in this area focused on the semi-continuous measurement
methods, in which elements were analyzed in the lab using the methods mentioned above
on time-resolved and/or size resolved samples (Kidwell and Ondov, 2004, 155898). The
concentrated slurry/graphite furnace atomic absorption spectrometry (GFAAS) method
collects ambient PM as a slurry using impactors, and then the collected PM is analyzed by
AAS in the lab. Laser induced breakdown spectroscopy (LIBS) was used to measure seven
July 2009
3-38
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
metals at the Pittsburgh supersite. LIBS concentrates ambient PM using a virtual imp actor
into a sample cell, and then a Nd^YAG laser-spectrometer is used to identify and quantify
different elements. A full listing of measurement techniques and analytical specifications
for trace elements is provided in Annex A, Table A-6.
Elemental and Organic Carbon
The large variety of aspects of carbon analyses were reviewed in the 2004 PM AQCD
(U.S. EPA, 2004, 056905). Measurement and analytical specifications for carbon
measurements are listed in Annex A, Tables A-10 and A-15. Aspects of the measurements
include sampling artifacts associated with the integrated filter-based OC and EC sampling
methods, the IMPROVE vs. CSN thermal optical protocols (i.e., different thermal optical
methods) and optical techniques to measure light-absorption or BC. One significant change
taking place in the CSN is that the method for carbon measurements is being changed from
the CSN method to a method designed to be consistent with the IMPROVE carbon analysis
protocol. This is a phased process that began in May of 2007 with the conversion of 56
stations. Phase 2 of the carbon sampler conversion occurred in April of 2009 with another
62 stations. The balance of the CSN is scheduled to be converted to IMPROVE-like
sampling and IMPROVE analysis in late 2009 (Henderson, 2005, 156537). The CSN
network was implemented to support the PM2.5 NAAQS and provides data for PM2.5 mass,
SO42-, NOa , NH4"1", Na, K, EC, OC, and select trace elements (A1 through Pb) at many sites
across the U.S. This conversion will increase consistency between these two networks. Also,
since the release of the 2004 PM AQCD, more studies have been conducted to extend the
understanding of sampling artifact issues (Chow et al., 2008, 156355; Watson et al., 2005,
157125), evaluate different thermal and optical procedures (Chen et al., 2004, 147143;
Chow et al., 2004, 156347; Chow et al., 2005, 155728; Chow et al., 2007, 156354; Conny et
al., 2003, 145948; Han et al., 2007, 155823; Subramanian et al., 2004, 081203; Watson et
al., 2005, 157125), develop reference materials (Klouda et al., 2005, 130382; Lee, 2007,
155926), create water soluble organic carbon (WSOC) measurement techniques (Andracchio
et al., 2002, 155657; Yang et al., 2003, 156167), develop
semi-continuous/continuous/real-time carbon measurement techniques (Chow et al., 2008,
156355; Watson et al., 2005, 157125), and introduce isotope identification into the OC/EC
measurement (Huang et al., 2006, 097654).
July 2009
3-39
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
OC sampling artifact issues were further addressed in various studies (Arhami et al.,
2006,	156224; Bae et al., 2004, 156243; Chow et al., 2005, 155728; Fan et al., 2003, 058628;
Fan et al., 2004, 155770; Grover et al., 2008, 156502; Lim et al., 2003, 156697; Mader et al.,
2003, 155955; Matsumoto et al., 2003, 124293; Muller et al., 2004, 097109; Offenberg et al.,
2007,	156822; Olson and Norris, 2005, 156005; Park et al., 2006, 098104; Rice, 2004,
156049; Subramanian et al., 2004, 081203; ten Brink et al., 2004, 097110; ten Brink et al.,
2005, 156115; Viana et al., 2006, 096384), and were well summarized by Watson et al.
(2005, 157125) and Chow et al. (2008, 156355). There are two commonly used methods to
correct OC sampling artifacts: the filter with backup filter system (TBQ: placing a backup
quartz-fiber filter behind the front Teflon-membrane filter! QBQ: placing a backup
quartz-fiber filter behind the front quartz-fiber filter); and the denuder-filter*adsorbent
system. Subramanian et al. (2004, 081203) and Chow et al. (2006, 099031) reported that
during the Pittsburgh and Fresno supersite studies the positive artifact (organic gases
condensed on filters) from TBQ (24-34%, up to 4 jug/m3 OC) was nearly twice that from QBQ
(13-17%). With the denuder-filter-adsorbent system, the negative artifact (OC evaporating
from the filter) was 5-10%. Watson and Chow (2002, 037873) reported that the XAD-coated
denuder could function as efficiently as a parallel plate denuder using carbon-impregnated
charcoal filters (CIF) with frequent denuder changes. Huebert and Charlson (2000, 156577)
reported that using tandem filter packs may hinder a quantitative analysis of the artifacts.
Different temperature protocols and optical correction methods in thermal-optical
analyses were further evaluated by Watson et al. (2005, 157125), Chow et al. (2004, 156347;
2005, 155728; 2007, 156354), Subramanian et al. (2006, 156107), Conny et al. (2005,
155728; 2003, 145948), Han et al. (2007, 155823), Chen et al. (2004, 147143), and (Conny et
al., 2009, 191999). Solomon et al. (2003, 156994) reported a 20-50% difference for OC and a
20-200% difference for EC using 11 filter samples and 4 different analytical protocols. In an
assessment of the different thermal-optical analysis protocols used around the world,
Watson et al. (2005, 157125) reported that differences of a factor of 2 to 7 in EC between
different methods could be observed, and a factor of 2 was common, while the relative
differences in OC between different methods were relatively small. As Watson et al. (2007,
157127) stated, there are 12 major differences among the thermal methods: (l) analysis
atmosphere; (2) temperature ramping rates! (3) temperature plateaus! (4) residence time at
each plateau! (5) optical pyrolysis monitoring configuration and wavelength; (6)
standardization; (7) oxidation and reduction catalysts! (8) sample aliquot and size! (9)
July 2009
3-40
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
evolved carbon detection method; (10) carrier gas flow through or across the sample! (ll)
location of the temperature monitor relative to the sample! and (12) oven flushing
conditions. Chow et al. (2004, 156347) and Chen et al. (2004, 147143) addressed the
difference between optical transmission and optical reflectance methods for charring
correction, and they reported that the charring OC on the surface of or inside a filter
dominated the differences between these two correction methods. The differences between
different sampling and measurement methods are also applied to the
in-situ/semi-continuous methods, since most of these methods are also based on
thermal-optical analysis of collected filters. Most of these methods agree with integrated
filter methods within 30%.
The differences observed between methods for OC and EC come largely from how OC
and EC are defined. They are defined on an operational basis, as there are no standard
reference materials. Initial efforts have been made to produce OC/EC reference materials at
the National Institute of Science and Technology (NIST) (Klouda et al., 2005, 130382; Lee,
2007, 155926). Klouda et al. (2005, 130382) described the development of Reference
Material 8785: Air Particulate Matter on Filter Media. Each reference filter is uniquely
identified by its air PM number and its gravimetrically determined mass of fine Standard
Reference Material (SRM) 1649a, and each filter has values assigned for total carbon, EC,
and organic carbon mass fractions measured according to both IMPROVE and NIOSH
protocols. Lee et al. (2007, 155926) reported a method to create a reference filter with a
known amount of OC (as potassium hydrogen phthalate), and EC (as carbon black
hydrosol).
Measurement methods for WSOC have been developed recently (Greenwald et al.,
2007, 155809; Miyazaki et al., 2006, 156767; Sullivan and Weber, 2006, 157031; Sullivan et
al., 2004, 157029; Sullivan et al., 2006, 157030; Sullivan et al., 2007, 100083; Yu et al.,
2004, 156172). WSOC can be measured on integrated filter samples, or in-situ
measurement can be conducted by coupling with the PILS-IC (Sullivan et al., 2004,
157029). For integrated filter samples, filters are extracted with deionized water and
followed by oxidation of total WSOC to CO2. CO2 can then be detected by either infrared
spectroscopy (IR) (Decesari et al., 2000, 155748; Kiss et al., 2002, 156646; Yang et al., 2003,
156167), FID (Yang et al., 2003, 156167), or pyrolysis gas chromatography/mass
spectrometry (GC/MS) (Gelencser et al., 2000, 155785). A correlation coefficient of 0.84 was
July 2009
3-41
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
reported by Sullivan et al. (Sullivan et al., 2004, 157029) between in-situ and filter based
measurement of WSOC.
Further development and evaluation has been conducted on the measurement of BC
with light absorption instruments (Andreae and Gelencser, 2006, 156215; Arnott et al.,
2003, 037711; Bae et al., 2004, 156243; Borak et al., 2003, 156284; Cyrys et al., 2003,
049634; Kurniawan and Schmidt-()((, 2006, 098823; Park et al., 2006, 098104; Saathoff et
al., 2003, 156066; Sadezky et al., 2005, 097499; Slowik et al., 2007, 096177; Taha et al.,
2007, 096277; Virkkula et al., 2007, 157098; Wallace, 2000, 000803; Weingartner et al.,
2003, 156149; Williams et al., 2006, 157148; Wu et al., 2005, 157155). These instruments
include the aethalometer, particle absorption photometer, and photoacoustic analyzer.
However, these instruments are subject to interferences by particle scattering, interactions
with the filter substrate, particle loading on filters, and other pollutants (e.g., NO2).
Uncertainties of up to 50% were observed in the studies mentioned above by comparing
these methods with integrated filter methods and thermal analysis methods.
Huang et al. (2006, 097654) reported the measurement of a stable isotope, 13C, in OC
and EC with a thermal optical transmission analyzer coupled with gas
chromatography-isotope ratio mass spectrometer (TOT-GOIRMS). The ratio of 13C/12C in
OC and EC can provide useful information on OC/EC source categories and origin. The
method was applied to Pacific2001 aerosol samples from the Greater Vancouver area in
Canada and produced a precision of-0.03%. Gustafsson et al. (2009, 192000) applied the
radiocarbon measurement technique and quantified the source contributions of
carbonaceous aerosols to the Indian Ocean "brown cloud," with particular relevance for
understanding and mitigating the climate effects of EC/BC.
Organic Speciation
Organic matter makes up a substantial fraction of PM in all regions of the U.S.
(U.S. EPA, 2004, 056905), and 10 to 40% of the total organic matter is currently
quantifiable at the individual compound level (Poschl, 2005, 156882). Recent advacements
in traditional solvent extraction gas chromatography/mass spectrometry (GC/MS) and high
pressure liquid chromatography (HPLC) as well as application of thermal desorption (TD)
techniques are helping to expand the understanding of the composition of organic matter as
well as improving detection limits for quantification of organic molecular marker (OMM)
compounds (Robinson et al., 2006, 156918; Schnelle-Kreis et al., 2005, 112944; Sheesley et
July 2009
3-42
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
al., 2007, 112017; Shrivastava et al., 2007, 111594). In addition, information about organic
functional groups can be obtained with Fourier transform infrared spectrometry (FTIR)
(Tsai and Kuo, 2006, 156127).
Recent advancements in GC/MS technology including inert electron ionization sources
and improved instrument sensitivity and scan rates for better OMM quantification, have
increased its application in organic aerosol characterization studies (Cass, 1998, 155716;
Fraser et al., 2003, 042231; Graham et al., 2003, 156489; Hays et al., 2002, 026104;
Robinson et al., 2006, 156918; Schauer et al., 1996, 051162; Sheesley et al., 2007, 112017;
Subramanian et al., 2006, 156107; Watson et al., 1998, 012257; Zheng et al., 2002, 026100;
Zheng et al., 2006, 157189). Incorporation of high volume injection using programmable
temperature vaporization (PTV) (Engewald et al., 1999, 155765) has further lowered
detection limits for trace level OMM compounds. High volume injection has the added
benefit of preventing the loss of semivolatile compounds (Swartz et al., 2003, 157035), and
has been applied for analysis of PAHs using low volume samplers (down to 5 Lpm), allowing
for smaller required mass loadings (Bruno et al., 2007, 155706; Crimmins and Baker, 2006,
097008). Since last review, HPLC analysis with fluorescence detection has also been used
frequently for quantification of semivolatile organic compounds in both the particle and gas
phase (Albinet et al., 2007, 154426; Barreto et al., 2007, 155676; Chow, 2007, 157209;
Eiguren-Fernandez et al., 2003, 142609; Goriaux et al., 2006, 156484; Murahashi, 2003,
096539; Ryno et al., 2006, 156065; Stracquadanio et al., 2005, 156104; Temime-Roussel et
al., 2004, 098530; Temime-Roussel et al., 2004, 098521). Lengthy extraction and analysis
times remain a limiting factor for these methods.
TD techniques bypass one of the time consuming steps in traditional solvent
extraction analysis for nonpolar organic compounds (n-alkanes, branched alkanes,
cyclohexanes, hopanes, steranes, alkenes, phthalates and PAHs). This is achieved by
vaporizing and analyzing organic constituents directly from the collection substrate,
thereby bypassing the extraction step (Chow et al., 2007, 156354). Methods exist for both
off-line TD analysis of previously collected filter samples and semi-continuous TD analysis.
Annex A, Table A-17 is adapted from Chow et al. (2007, 157209) and summarizes recent
TD-GC/MS studies. The most common off-line method is thermal desorption-gas
chromatography/mass spectrometry (TD-GC/MS) (Hays and Lavrich, 2007, 155831).
Continuous or semi-continuous methods have been developed for direct analysis of
individual organic constituents by coupling TD with various forms of mass spectrometry
July 2009
3-43
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
(Smith et al., 2004, 156090; Tobias and Ziemann, 1999, 157053; Tobias et al., 2000, 156121;
Voisin et al., 2003, 156141; Williams et al., 2006,156157). A comparison of measurement
and analytical specifications for filter analysis using solvent extraction and TD methods for
organic speciation are summarized in Annex A, Table A-17.
3.4.1.3. Multiple-Component Measurements on Individual Particles
The 2004 PM AQCD discussed the aerosol time-of-flight mass spectrometry
(ATOFMS). Recently, the ATOFMS and several other aerosol mass spectrometry methods
have been further developed. Both lab and field comparisons have been conducted to
evaluate the reliability of these types of instruments.
There are four types of commonly used aerosol mass spectrometry: (l) particle
analysis by laser MS (PALMS; National Oceanic and Atmospheric Administration [NOAA]);
(2) rapid single particle mass spectrometer (RSMS; University of Delaware); (3) aerosol
time-of-flight MS (ATOFMS; TSI, Inc.); and (4) AMS (Aerodyne) (Chow et al., 2008, 156355;
Nash et al., 2006, 156795). The differences between these instruments primarily come from
the particle sizing methods of mass spectrometers, as shown in Annex A, Table A-16.
Although the technique varies, the underlying principle is to fragment each particle into
ions, using either a high-power laser or a heated surface, and then a mass spectrometer to
measure the mass to charge ratio of each ion fragment in a vacuum.
These instruments were evaluated at the Atlanta, Houston, Fresno, Pittsburgh, New
York, and Baltimore supersites (Bein et al., 2005, 156265; Drewnick et al., 2004, 155755;
Drewnick et al., 2004, 155754; Hogrefe et al., 2004, 156560; Jimenez et al., 2003, 156611;
Lake et al., 2003, 156669; Lake et al., 2004, 088411; Middlebrook et al., 2003, 042932;
Phares et al., 2003, 156866; Qin and Prather, 2006, 156895; Wenzel et al., 2003, 157139;
Zhang et al., 2008, 155144). Measurements of the gross composition and abundance of
particles by these instruments were generally semi-quantitative, with the exception of
AMS. Particles of similar composition (e.g., OC/SO42-, Na/K/SO r , soot/hydrocarbon, and
mineral particle types) were characterized by these instruments during the studies
mentioned above. NOa and SO r concentrations measured with AMS were comparable
with other continuous and filter-based methods, as mentioned in Section 3.4.1.2. In
addition, concentrations of different particle types can be obtained by the co-location of
July 2009
3-44
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
these aerosol mass spectrometers and other particle sizing instruments, such as particle
counters or the MOUDI.
3.4.1.4. Ultrafine PM: Mass, Surface Area, and Number
Instruments for measuring ultrafine PM developed during the past decade permit
measurement of size distributions of particles down to 3 nm in diameter with mobility
particle sizers. Concentrations down to this size range can be obtained by a Micro-Orifice
Uniform Deposit Impactor (MOUDI). The recently developed low pressure-drop ultrafine
particle impactor coupled with a 6 Attenuation Monitor (nano-BAM) can also provide
ultrafine PM (<150 nm) mass concentrations (Chakrabarti et al., 2004, 147867). A high
correlation coefficient was observed between MOUDIs and nano-BAMs, with a correlation
of 0.96. A 50% cutpoint (d 50) of 13*200 nm can be achieved by a high-volume slot-type
ultrafine PM virtual impactor (Middha and Wexler, 2006, 155982).
Methods are also being developed to measure the surface area of ultrafine particles.
Particle surface area is usually measured by attaching labeled (radioactive or electrical
labeling) molecules to particles and detect the radioactive or electrical properties of the
attached molecules. Wilson et al. (2007, 157149) suggested that the electrical aerosol
detector (EAD, based on diffusion charging) measurement might be a useful indicator of the
particle surface area deposited in the lung. This method can be potentially useful for
examining the association between health effects and particle surface areas.
Developments involving the condensation particle counter include use of de-ionized
water as a condensation media in lieu of butanol or n-propanol in condensation particle
counters (Hering et al., 2005, 155838; Hermann et al., 2007, 155840; Petaja, 2006, 156021).
This development makes the condensation particle counter (CPC) easier to use in field
studies because water does not have some of the same chemical properties (with respect to
hazard and odor) as butanol or n-propanol. The performance of this CPC was reported to be
similar to the conventional butanol based CPC (Hering et al., 2005, 155838). Use of a
battery of water and butanol-based CPCs was demonstrated to detect a range of solubilities
in nucleation-mode particles (Kulmala et al., 2007, 155911). Additionally, CPCs have been
used to measure particles in the smaller end of the ultrafine scale through adjustment of
CPC cut-off diameters through tuning the temperature difference between the CPC
saturator and condenser (Kulmala et al., 2007, 155911) and improved charge reduction
July 2009
3-45
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
techniques (Winkler et al., 2008, 156160). The latter method was effective in reducing the
size of particles detected by a CPC to less than 2 nm. These studies include assessment of
errors related to these developments with the CPC and generally show that counting
efficiencies with these devices is upwards of 95% (Hermann et al., 2007, 155840).
Additionally, recent advancements have been made in development of fast scanning
methods for ultrafine particle size distributions, including diffusion screens (DS)
(Feldpausch et al., 2006, 155773) and fast integrated mobility scanners (FIMS) (Olfert et
al., 2008, 156004).
3.4.1.5.	PM Size Distribution
Along with particle density and shape (U.S. EPA, 2004, 056905), the particle size
distribution can be used to estimate PM mass concentrations. For particles >0.1 urn, several
instruments, including DRUM, MOUDIs, and aerodynamic particle sizer (APS), are
available to measure mass-based or count-based particle size distribution. An APS
incorporating very sharp cut points between 0.1 and 10 pm is now available (Peters, 2006,
156860; Zeng, 2006, 098375). For particles in this range, inertial forces are used to separate
particles based on impaction. For particles <0.1 um, particles can be separated by their
electrical mobility, and as a result, electrical mobility diameter is often used to describe
ultrafine PM size distribution in lieu of aerodynamic diameter. It has been necessary to
develop techniques to convert mobility diameters, measured by the scanning mobility
particle sizer (SMPS) or the Engine Exhaust Particle Sizer (EEPS), to aerodynamic
diameters, measured by the APS, or vice versa, in order to merge the distributions
spanning the ultrafine, accumulation, and coarse modes. A variety of techniques for
combining SMPS and APS diameters have been reported in the literature (Hand and
Kreidenweis, 2002, 155824; Khlystov et al., 2004, 155897; Morawska et al., 1999, 007609;
Morawska et al., 2007, 155990; Shen et al., 2002, 156086; TSI, 2005, 157196). However,
each of these techniques incurs some uncertainty of which the user must be aware.
3.4.1.6.	Satellite Measurement
Instruments sensing back scattered solar radiation on satellites have made it possible
to derive information about tropospheric aerosol properties on the global scale. The satellite
borne instruments vary in their complexity and in the aerosol properties they can measure.
July 2009
3-46
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Satellite instruments measure radiance (or brightness temperature) that can then be used
to provide information on the aerosol column amount, or the aerosol optical depth (AOD).
Depending on the wavelengths sampled and the spectral resolution of the instruments,
information about the composition of particles of diameter <2 um and particles of diameter
>2jum can be obtained. Data from two main instruments, the moderate resolution imaging
spectroradiometer (MODIS) and the multiangle imaging spectroradiometer (MISR) have
been used to estimate surface PM in the U.S. MODIS measures the intensity of back
scattered sunlight at seven wavelengths through the visible to the near infrared at one
viewing direction! and MISR measures the intensity at four wavelengths (from the visible
to the near IR) and the same ground pixel at nine viewing angles. The spatial resolution of
reported AOD is 17.6 x 17.6 km for MISR and either 10 x 10 or 1 x 1 km for MODIS,
depending on retrieval algorithm. Since both instruments are located on the same satellite,
their times of overpass are the same, about 1330 local time. Due to precession of the
satellite's orbit, the satellite does not pass over the same path every day, and instruments
cannot sense aerosol properties beneath cloud tops.
The problem of using satellite data to retrieve properties of the atmospheric aerosol is
complex because the surface contribution to satellite measured reflectance must be
separated from the aerosol signal. Difficulties can arise when attempting to derive aerosol
information over land surfaces because of uncertainties in surface reflectivity, similarities
between aerosol and surface composition, and high signal-to-noise ratio when viewing AOD
over reflective surfaces such as desert and snow. To overcome this difficulty, data from
MODIS have been applied over dark land surfaces and ongoing improvements in retrieval
algorithms are being developed. Instruments such as the multiangle imaging
spectroradiometer (MISR) that sense at multiple viewing angles can better cope with the
problems over land surfaces because they can use the information on the angular
dependence of reflection from the surface and the atmosphere to distinguish between their
signals. Not only can total AODs be derived, but fractional AODs that reflect external
mixtures characterized by particle shape, effective radius, and single scattering albedo can
also be derived. These properties can then be used to infer particle composition. Retrievals
over the oceans have had less difficulty because the optical properties of sunlight reflected
from the sea surface are much better known, and reflectivities are low over most zenith
angles at less than gazing incidence.
July 2009
3-47
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Kokhanovsky et al. (2007, 190009) examined the errors associated with MODIS,
MISR, and a number of other satellite instruments with respect to associated retrieval
algorithms for retrievals over Central Europe. They found a correlation coefficient between
MODIS and MISR AOD of 0.62. Both MODIS and MISR AOD tended to underestimate
ground-based AOD measurements from AERONET (NASA's AErosol RObotic NETwork)
slightly with MODIS generally retrieving higher AODs than MISR. Chu et al. (2003,
190049) found correlations between MODIS and AERONET AODs ranging from 0.82-0.91;
and Kahn et al. (2005, 189961) found correlations of 0.7-0.9 between MISR and AERONET
AODs.
Further complexity is added when attempting to relate surface PM2.5 to aerosol optical
depths. The detailed comparisons of surface measurement and satellite measurements are
given in Chapter 9.
3.4.2. Ambient Network Design
3.4.2.1. Monitor Siting Requirements
The AQS contains measurements of air pollutant concentrations in the 50 states, plus
the District of Columbia, Puerto Rico, and the Virgin Islands, for the 6 criteria air
pollutants as well as a more limited dataset of hazardous air pollutants. In 2007, there
were 4,693 PM10 monitors and 2,194 PM2.5 monitors reporting values to the EPA Air Quality
System database (AQS). Where SLAMS PM10 and PM2.5 monitoring is required, at least one
of the sites must be a maximum concentration site for that specific area. The appropriate
spatial scales for PM10, PM2.5, and PM10-2.5 monitoring differ given the contrasting spatial
gradients of coarse PM relative to fine. The relevant scales for each size classification are
provided in Annex A, Table A-18.
Criteria for siting ambient monitors for PM at national monitoring networks are
summarized below by PM size, and details are given in the CFR 40 Part 58 Appendix D,
and SLAMS/NAMS/PAMS Network Review Guidance (U.S. EPA, 1998, 083151). Table A-19
in Annex A provides a summary of the number of sites and operating specifications of these
networks. Probing and monitoring path siting criteria for any specific monitoring site are
given in CFR 40 Part 58 Appendix E, including horizontal and vertical placement, spacing
from minor source, spacing from obstructions, spacing from trees, and spacing from
roadways.
July 2009
3-48
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
PMio
Metropolitan Statistical Areas (MSAs) with populations in excess of one million are
required to have between 2 and 10 monitors (depending on concentration), while MSAs with
populations less than one million are required to have between 0 and 8 monitors (40 CFA
Part 58, Appendix D). Except from some circumstances where microscale (<100 m, for
maximum PMio exposure) monitoring may be appropriate, the most important scales to
characterize the emissions of PMio effectively from both mobile and stationary sources are
the middle (for short-term public exposure) and neighborhood scales (for trends and
compliance with standards). PMio measurements are obtained at standard temperature
and pressure across the NAMS/SLAMS networks (40 CFR Part 58).
PM2.5
Monitor requirements for PM2.5 based on MSA population are similar to those for
PMio above, but also include pollutant concentration as a factor in determining the number
of required stations. Continuous PM2.5 monitors must be operated in no fewer than one-half
of the minimum required sites in each area. Most PM2.5 monitoring in urban areas should
be representative of a neighborhood scale (for trends and compliance with standards).
Urban or regional scale sites are located to characterize regional transport of PM2.5. In
certain instances where population-oriented micro- or middle-scale PM2.5 monitoring are
determined by the Regional Administrator to represent many such locations throughout a
metropolitan area, these smaller scales can be considered to represent community-wide air
quality. PM2.5 measurements are obtained at local temperature and pressure across the
NAMS/SLAMS networks (40 CFR Part 58).
PM species are monitored at both mostly urban (CSN) and mostly rural (IMPROVE)
locations. PM2.5 chemical speciation monitoring is currently conducted at 197 CSN sites
(http://www.epa.gov/ttn/amtic/specgen.html). Within the CSN network, 53 locations are
recognized as the Speciation Trends Network (STN) operating on a sample schedule of one
in every three days, while the rest of the CSN typically operates every sixth day.
PM102.5
PM10-2.5 has not been required to be monitored at SLAMS sites, but will be required at
NCore1 Stations (which is a sub-set of the SLAMS) by January 1, 2011. Middle and
neighborhood scale measurements are the most important station classifications for PM10-2.5
1 For more information on NCore, see the NCore web site at- http7/www.epa.gov/ttn/amtic/ncore/index.html
July 2009
3-49
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
to assess the variation in coarse particle concentrations that would be expected across
populated areas that are in proximity to large emissions sources. PM10-2.5 chemical
speciation monitoring and analyses is required at NCore sites, also by January 1, 2011. EPA
has already approved FRMs and FEMs for PM10-2.5 mass! however, methods for PM10-2.5
speciation are still being developed (Henderson, 2009, 192001). PM10-2.5 measurements are
obtained at local temperature and pressure by recalculating the co-located PM10 for local
conditions.
3.4.2.2. Spatial and Temporal Coverage
Locations of PM2.5 and PM10 Monitors in Selected Metropolitan Areas in the U.S.
Fifteen metropolitan regions were chosen for closer investigation of monitor siting
based on their distribution across the nation and relevance to health studies analyzed in
subsequent chapters of this ISA. These regions were: Atlanta, Birmingham, Boston,
Chicago, Denver, Detroit, Houston, Los Angeles, New York City, Philadelphia, Phoenix,
Pittsburgh, Riverside, Seattle, and St. Louis. Core-Based Statistical Areas (CBSAs) and
Combined Statistical Areas (CSAs), as defined by the U.S. Census Bureau
(http7/www.census/gov/). were used to determine which counties, and hence which
monitors, to include for each metropolitan region.1 Figures 3-7 and 3"8 display PM10 and
PM2.5 monitor density with respect to population density for Boston. Annex A includes
similar information for the remaining fourteen metropolitan regions (Figures A-l-A-28).
1 A CBSA represents a county-based region surrounding an urban center of at least 10,000 people determined using 2000
census data and replaces the older Metropolitan Statistical Area (MSA) definition from 1990. The CSA represents an
aggregate of adjacent CBSAs tied by specific commuting behaviors. The broader CSA definition was used when selecting
monitors for the cities listed above with the exception of Los Angeles, Riverside and Phoenix. Los Angeles and Riverside are
contained within the same CSA, so the smaller CBSA definition was used to delineate these two cities. Phoenix is not
contained within a CSA, so the smaller CBSA definition was used for this city as well.
July 2009
3-50
DRAFT-DO NOT CITE OR QUOTE

-------
~ Kilometers
0 5 10 20 30 40 50
A —1
0/
r
2005 Population Density
Boston PM10 Monitors (15 km buffer)
Population per Sq Km
¦¦ 0 - 251
| 252-502
503 - 2508
2509-5016
5017- 12539
¦¦ 12540-50155
] Kilometers
0 15 30 60 90 120 150
Figure 3-7. PMio monitor distribution in comparison with population density, Boston CSA.
July 2009
3-51
DRAFT-DO NOT CITE OR QUOTE

-------
~ Kilometers
0 5 10 20 30 40 50
r
r>
2005 Population Density
Boston PM2.5 Monitors (15km buffer)
Population per Sq Km
0-251
~ 252 - 502
503 - 2508
2509-5016
5017-12539
I 12540-50155
] Kilometers
0 15 30 60 90 120 150
Figure 3-8. PM2.5 monitor distribution in comparison with population density, Boston CSA.
July 2009
3-52
DRAFT-DO NOT CITE OR QUOTE

-------
Table 3-3. Proximity to PM10 and PM2.5 monitors for total population3 by city.
Proximity to PM Monitors'1
Region
Total CSA/CBSA

< 1 km

< 5 km

<10 km
<15 km

N

N %

N %

N "/.
> N
%
PROXIMITY TO PMn MONITORS
Atlanta
5316742
30973
0.58
416440
7.83
1090497
20.51
1837983
34.57
Birmingham
1166100
23943
2.05
251310
21.55
473054
40.57
638472
54.75
Boston
7502707
63614
0.85
1090172
14.53
2087770
27.83
2939870
39.18
Chicago
9754262
55642
0.57
844714
8.66
2374972
24.35
3844297
39.41
Denver
2952039
38449
1.30
521201
17.66
1146286
38.83
1799187
60.95
Detroit
5553465
14050
0.25
309623
5.58
748971
13.49
1300995
23.43
Houston
5503320
36795
0.67
832767
15.13
2227314
40.47
3141150
57.08
Los Angeles
13061361
52052
0.40
1404389
10.75
4899254
37.51
9075863
69.49
New York
22050940
19842
0.09
292105
1.32
592631
2.69
773962
3.51
Philadelphia
6388913
23988
0.38
376966
5.90
1091532
17.08
2238309
35.03
Phoenix
3818147
99520
2.61
1255430
32.88
2615738
68.51
3416682
89.49
Pittsburgh
2515383
65906
2.62
706413
28.08
1291700
51.35
1705451
67.80
Riverside
3781063
61356
1.62
895615
23.69
2360272
62.42
2922799
77.30
Seattle
3962434
4851
0.12
220539
5.57
709887
17.92
1211430
30.57
St. Louis
2869955
27872
0.97
380411
13.25
891695
31.07
1212543
42.25
PROXIMITY TO PM2.5 MONITORS
Atlanta
5316742
23461
0.44
581461
10.94
1990477
37.44
3179844
59.81
Birmingham
1166100
12925
1.11
240383
20.61
666926
57.19
848447
72.76
Boston
7502707
185457
2.47
1877180
25.02
3356019
44.73
4641175
61.86
Chicago
9754262
177076
1.82
3091573
31.69
6473463
66.37
8185010
83.91
Denver
2952039
40601
1.38
649953
22.02
1548976
52.47
2252657
76.31
Detroit
5553465
54997
0.99
1174733
21.15
2791555
50.27
3845190
69.24
Houston
5503320
11586
0.21
213708
3.88
905007
16.44
1599079
29.06
Los Angeles
13061361
115477
0.88
2579809
19.75
7544466
57.76
10792727
82.63
New York
22050940
717094
3.25
8107764
36.77
13493867
61.19
16571764
75.15
Philadelphia
6388913
117389
1.84
1878373
29.40
3517321
55.05
4393136
68.76
Phoenix
3818147
37133
0.97
490072
12.84
1099069
28.79
1739542
45.56
Pittsburgh
2515383
40574
1.61
587148
23.34
1331230
52.92
1883301
74.87
Riverside
3781063
43739
1.16
723829
19.14
1855296
49.07
2344394
62.00
Seattle
3962434
13723
0.35
287373
7.25
931630
23.51
1561792
39.41
St. Louis
2869955
37329
1.30
563176
19.62
1338349
46.63
1760985
61.36
8Based on 2005 population totals.
bPercentages are given with respect to the total population per city provided.
Table 3"3 shows the population density around PM2.5 and PM10 monitors for the total
population for each CSA/CBSA individually. Population totals within various distances of
July 2009
3-53
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
PM monitors were calculated assuming equal internal population distribution for
individual census blocks. Between-city disparities in population density were large and
were dependent primarily on the location and number of PM monitoring sites per
CSA/CBSA. For PM2.5, Los Angeles (83%) and Denver (76%) had the largest proportion of
the total population within 15 km of a monitor. Houston (29%) had the least population
coverage with their PM2.5 monitors. For PM10, Phoenix (89%) had the largest proportion of
the total population within 15 km of a monitor. Detroit (23%), Boston (39%), Seattle (31%),
and Philadelphia (35%) had the smallest proportions of the population within 15 km of a
PM10 monitor. Proximity to monitoring stations is considered further in Sections 3.5 and 3.7
regarding spatial variability within cities. Figure 3"8 shows that the PM2.5 network more
closely samples near population centers in the Boston CSA compared with the PM10
network shown in Figure 3" 7, although both PM10 and PM2.5 networks place at least one
monitor in the city center.
3.4.2.3. Network Application for Exposure Assessment with Respect to Susceptible
Sub-populations
Subject Age
Table 3-4 breaks down the population density around PM2.5 and PM10 monitors for
sub-populations of children aged 0-4, children aged 5-17, and elderly adults aged 65 and
over cumulatively for the fifteen CSAs/CBSAs examined, and Table 3-5 shows the
distribution for adults aged 65 and over for each CSA/CBSA individually. This detail of
information is not provided for the 0-4 y and 5-17 y age groups because variation in
percentage within a certain radius of the monitor was generally fairly low for each city
across the child age groups when compared to total population. In the cases of Denver,
Detroit, Phoenix, Riverside, and St. Louis for PM2.5 and Birmingham, Denver, Riverside,
and St. Louis for PM10, the elderly population's distribution around the samplers varied
more from the total population compared to other age groups. When all CSAs/CBSAs were
considered cumulatively, the percentage of the population within 15 km of a monitor was
similar for all age groups for both PM2.5 and PM10. Between-city disparities in elderly
population density within a sampler radius were larger. For PM2.5, Chicago (86%) and
Denver (84%) had the largest proportion of the total population within 15 km of a monitor.
Houston (31%) had the least population coverage with their PM2.5 monitors. For PM10,
Phoenix (90%) had the largest proportion of the total population within 15 km of a monitor.
July 2009
3-54
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
New York (4%), Detroit (27%), Seattle (32%), and Philadelphia (38%) had the smallest
proportions of the population within 15 km of a PMio monitor. These differences may reflect
overall density of the samplers within a given city, with PM2.5 monitors more numerous
than PM10 monitors in most of the CS As/CBS As, and retirement and settlement trends
among elderly adults.
Table 3-4. Proximity to PM10 and PM2.5 monitors for children 0-4 y, children 5-17 y, and adults 65 y
and older.aThe figures presented here are cumulative for the 15 CSAs/CBSAs examined in
Chapter 3.
Proximity to PM Monitors'1
Afle Total CSA/CBSA
< 1 km
< 5 km

< 10 km
< 15 km
N
N
%
N
%
N
%
N
%
PROXIMITY TO PMio MONITORS
0-4 6400785
44384
0.69
695120
10.86
1725419
26.96
2636782
41.19
5-17 17212825
110882
0.64
1756246
10.20
4441239
25.80
6942001
40.33
>64 10391023
68367
0.66
1056375
10.17
2631243
25.32
4041802
38.90
PROXIMITY EAR GLASS IN WOONPDSAL TO PMzs MONITORS
0-4 6400785
109466
1.71
1603000
25.04
3361922
52.52
4462403
69.72
5-17 17212825
275427
1.60
4164132
24.19
8814179
51.21
11813997
68.63
>64 10391023
175113
1.69
2570909
24.74
5483776
52.77
7288049
70.14
8Based on 2000 population totals.
bPercentages are given with respect to the total population per city provided.
Race and Hispanic Origin
Table 3-6 shows the percent of the population self-identified as white or black and
having a Hispanic or non-Hispanic origin within 1, 5, 10, or 15 kilometer distances of PM2.5
and PM10 monitors cumulatively across the fifteen CSAs/CBSAs. For PM10, Hispanics (54%)
represented the subpopulation with the largest percentage of total population within 15 km
of a monitor across the fifteen CSAs/CBSAs studied. The percentage of blacks within that
same distance was marginally lower (48%), whereas the percentage of whites and non-
Hispanics within 15 km of a monitor was approximately two-thirds that of Hispanics (35%
and 37%, respectively). For PM2.5, blacks and Hispanics had similar percentages of the
population within 15 km of a monitor (86% and 82%, respectively), while a smaller
proportion of whites and non-Hispanics were within that same distance (63% and 67%,
respectively), across the fifteen CSAs/CBSAs studied. Higher percentages of individual
ethnic subpopulations within 15 km of a PM2.5 monitor most likely represents the fact that
July 2009
3-55
DRAFT-DO NOT CITE OR QUOTE

-------
more PM2.5 monitors are currently deployed compared with PM10 monitors. While no ethnic
subpopulation appears to be well represented at the neighborhood scale, greater
percentages of the black and Hispanic populations (1% each) are within 1 km of a PM10
monitor than the corresponding white and non-Hispanic populations (0.5% each). Likewise,
2.5% of the black population and 2.8% of the Hispanic population reside within 1 km of a
PM2.5 monitor compared to 1.4% of the white population and 1.5% of the non-Hispanic
population. Furthermore, it is notable that at any scale shown in Table 3"6, for both PM10
and PM2.5 monitors, those self-identified as black or Hispanic actually have greater
representation by the monitors than those identified as white or non-Hispanic.
Table 3-5.
Proximity to PM10 and PM2.5 monitors for adults aged 65 and older3 by city.






Proximity to PM Monitors'1




Region
Total CSA/CBSA

< 1 km

< 5 km

<10 km

<15 km

N

N "/

N %

N %
N
%
PROXIMITY TO PMio MONITORS
Atlanta
362201
2115
0.58
35448
9.79
93903
25.93
139240
38.44
Birmingham
145905
3663
2.51
35628
24.42
66839
45.81
86299
59.15
Boston
945790
6852
0.72
124911
13.21
262854
27.79
385046
40.71
Chicago
1018983
7619
0.75
107540
10.55
291705
28.63
441771
43.35
Denver
232974
3675
1.58
43658
18.74
107548
46.16
168447
72.30
Detroit
626216
1555
0.25
41833
6.68
99680
15.92
167760
26.79
Houston
377586
2085
0.55
57413
15.21
166715
44.15
219615
58.16
Los Angeles
1207436
4693
0.39
126696
10.49
422725
35.01
810078
67.09
New York
2710675
2463
0.09
37580
1.39
80222
2.96
104951
3.87
Philadelphia
834110
2740
0.33
49413
5.92
154535
18.53
322700
38.69
Phoenix
388150
8605
2.22
119306
30.74
267456
68.91
348464
89.78
Pittsburgh
449544
13302
2.96
133285
29.65
243723
54.22
314941
70.06
Riverside
342334
4181
1.22
65499
19.13
182615
53.34
236900
69.20
Seattle
390372
503
0.13
22333
5.72
72979
18.69
123054
31.52
St. Louis
358747
4316
1.20
55833
15.56
117743
32.82
172535
48.09
PROXIMITY TO PM2.5 MONITORS
Atlanta
362201
1757
0.49
36772
10.15
136179
37.60
207122
57.18
Birmingham
145905
1619
1.11
29952
20.53
84223
57.72
106488
72.98
Boston
945790
18821
1.99
224628
23.75
438920
46.41
606231
64.10
Chicago
1018983
18539
1.82
348656
34.22
713194
69.99
883112
86.67
Denver
232974
3891
1.67
59625
25.59
140523
60.32
196361
84.28
Detroit
626216
5765
0.92
138672
22.14
345808
55.22
469462
74.97
Houston
377586
1010
0.27
14911
3.95
66741
17.68
117661
31.16
Los Angeles
1207436
9653
0.80
229893
19.04
688844
57.05
984889
81.57
July 2009
3-56
DRAFT-DO NOT CITE OR QUOTE

-------
Proximity to PM Monitors'1
New York
2710675
78918
2.91
921599
34.00
1619177
59.73
2048842
75.58
Philadelphia
834110
13323
1.60
251459
30.15
487003
58.39
605663
72.61
Phoenix
388150
2738
0.71
39833
10.26
90304
23.27
142084
36.61
Pittsburgh
449544
8933
1.99
111050
24.70
249269
55.45
347711
77.35
Riverside
342334
3024
0.88
50901
14.87
129836
37.93
170933
49.93
Seattle
390372
1721
0.44
29429
7.54
101223
25.93
156562
40.11
St. Louis
358747
5401
1.51
83528
23.28
192532
53.67
244929
68.27
+Based on 2000 population totals.
bPercentages are given with respect to the total population per city provided.
Table 3-6. Proximity to PM10 and PM2.5 monitors based on the population identified as white, black,
Hispanic, or non-Hispanica. The figures presented here are cumulative for the 15
CSAs/CBSAs examined in Chapter 3.
Proximity to PM Monitorsb
Race or
Total CSA/CBSA

< 1 km
< 5 km
< 10 km
< 15 km
Hispanic
Origin
N
N
"/
i N
%
N
%
N
%
PROXIMITY TO PMn MONITORS
White
61936855
325771
0.53
5554906
8.97
14041215
22.67
21913907
35.38
Black
12668004
134174
1.06
1611263
12.72
3867436
30.53
6020348
47.52
Hispanic
15916208
169305
1.06
2496959
15.69
5905322
37.10
8589819
53.97
Non-Hispanic
74611962
421917
0.57
6767187
9.07
17261734
23.14
27254421
36.53
PROXIMITY TO PM2.5 MONITORS
White
61936855
863823
1.39
12257978
19.79
27553900
44.49
39030037
63.02
Black
12668004
320447
2.53
4780620
37.74
9241172
72.95
10906346
86.09
Hispanic
15916208
445126
2.80
5782482
36.33
10661947
66.99
13094618
82.27
Non-Hispanic
74611962
1135999
I 1.52
16553574
22.19
36318474
48.68
49629054
66.52
8Based on 2000 population totals
Percentages are given with respect to the total population per city provided.
Socioeconomic Status
Table 3-7 shows the percent of the population below and above the poverty level and
the percent of the population over age 25 with less than high school, high school or more,
and college graduate education levels that reside within 1 km, 5 km, 10 km, and 15 km of a
PM10 and PM2.5 monitor cumulatively across the fifteen CSAs/CBSAs for which PM
concentration data were examined. For PM10, 47% of the population below the poverty level
and 45% of the population with less than a high school education are within 15 km of a
monitor, whereas the percentage of those above the poverty line (39%), those with a high
July 2009
3-57
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
school or more education (38%), and those with a college education (35%) were slightly less.
For PM2.5, 80% of the population below poverty level and 77% of the population with less
than high school education are within 15 km of a monitor for the fifteen CSAs/CBSAs
studied. Again, populations of those above the poverty line, those with a high school or more
education, and those with a college education within 15 km of a monitor were less than the
low SES groups (67% for each). Higher percentages of individual SES subpopulations
within PM2.5 monitors likely reflect the fact that more PM2.5 monitors are currently
deployed within 15 CSAs/CBSAs studied compared with PM10 monitors. Lower SES groups
are not shown to be well-represented at the neighborhood scale, with 1.2% of the population
below the poverty level and 1.0% of the population with less than a high school education
residing within 1 km of a PM10 monitor. Likewise, 3.1% of the population below the poverty
level and 2.4% of the population with less than high school education reside within 1 km of
a PM2.5 monitor. However, the populations of low SES groups are more represented at the
neighborhood scale than those for higher SES groups. For example, only about 1.5-1.7% of
those above the poverty line, those with a high school or more education, or those with a
college education are within 1 km of a PM2.5 monitor. Moreover, it is notable that at any
scale shown in Table 3-7 and for both PM10 and PM2.5, those living under the poverty line
and those with less than high school education have greater representation by the monitors
than those above the poverty line or those age 25 and older with high school or college
education.
Table 3-7. Proximity to PM10 and PM2.5 monitors based on the population below or above the poverty
line, population over age 25 with less than high school education, population over 25 with
high school education, and population over 25 with college education or morea. The figures
presented here are cumulative for the 15 CSAs/CBSAs examined in Chapter 3.
Proximity to PM Monitors'1
SES
Total
< 1 km
< 5 km
< 10 km
< 15 km

CSA/CBSA








N
N
%
N
%
N
%
N %
PROXIMITY TO PMio MONITORS
Below poverty line
10645411
132979
1.25
1626694
15.28
3504957
32.92
5024714 47.20
Above poverty line
85551420
485874
0.57
8171398
9.55
21096615
24.66
33034279 38.61
Less than HS
education
11606042
112901
0.97
1544594
13.31
3537414
30.48
5186441 44.69
July 2009
3-58
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Proximity to PM Monitors'1
HS education
30583598
185439
0.61
2975200
9.73
7580000
24.78
11747653
38.41
College education
16433811
59892
0.36
1229885
7.48
3447148
20.98
5722347
34.82
PROXIMITY TO PM2.5
MONITORS








Below poverty line
10645411
330970
3.11
3951549
37.12
7107192
66.76
8528731
80.12
Above poverty line
85551420
1297591
1.52
19094985
22.32
41736460
48.79
57070312
66.71
Less than HS
education
11606042
276942
2.39
3806208
32.80
7225291
62.25
8930174
76.94
HS education
30583598
444262
1.45
6940261
22.69
15152047
49.54
20489904
67.00
College education
16433811
280810
1.71
3451717
21.00
7776218
47.32
11000917
66.94
8Based on 2000 population totals
bPercentages are given with respect to the total population per city provided.
3.5. Ambient PM Concentrations
This section describes measurements of ambient PM mass concentration and
composition made since the 2004 PM AQCD (U.S. EPA, 2004, 056905) including analyses
using EPAAir Quality System (AQS) data as well as published findings. Emphasis is placed
on the period from 2005-2007 which incorporates the most recent validated AQS data
available at the time this document was prepared.
When the 2004 PM AQCD was written, the full nationwide PM2.5 compliance
monitoring network had only recently been deployed, providing three years (1999 to 2001)
of completed measurements. Based on observations from these first three years, the 2004
PM AQCD found that PM2.5 in eastern cities was generally more highly correlated across
monitoring sites than in western cities. The higher spatial correlations in the eastern cities
resulted from the more regionally dispersed sources of PM2.5 in the East. Although PM2.5
concentrations at sites within an urban area can be highly correlated, significant
differences in concentrations can occur on any given day. The ratio of PM2.5 to PM10 was
found to be higher in the East than in the West in general, and values for this ratio are
consistent with those found in numerous earlier studies presented in the 1996 PM AQCD
(U.S. EPA, 1996, 079380). Differences in the composition of PM2.5 between eastern and
western cities were also found to be consistent with differences found in the 1996 PM
AQCD. Much more limited data were available for describing the spatial variability of
coarse particulate mass measured as PM10-2.5, ultrafine PM, and PM composition. The 2004
AQCD noted that components produced by area (e.g., traffic) and point sources are more
July 2009
3-59
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
spatially variable than regionally dispersed components (e.g., secondary SO r ). Spatial
variability will affect estimates of community scale human exposure and caution should be
exercised in extrapolating conclusions from one area to another, particularly on a regional
scale.
For this PM ISA, the PM2.5 monitoring network has been active for eight or nine years
depending on location. Observations and analyses based on PM2.5 measurements reported
to AQS are included in this chapter. Furthermore, by selecting locations where PM10 and
PM2.5 measurements are co-located, information about the spatiotemporal distribution of the
PMio-2.5 size fraction is investigated. Given the form of the current standard and the relative
abundance of PM10 monitors in the AQS network, PM10 mass concentrations are also
included in this section with the understanding that PM10 includes mass contributions from
the smaller size fractions and therefore overlaps with PM2.5, PM10-2.5 and ultrafine particle
mass concentrations. Although compliance monitoring does not apply for ultrafine particles
because there is no ambient standard for them, new observational information is available
from detailed studies in several cities. Similarly, advancements have been made in
understanding PM composition from the CSN and IMPROVE networks. Descriptions of
ultrafine particles and speciated PM are covered where information is available.
Unless otherwise specified, the PM2.5, PM10-2.5 and PM10 data utilized in this
section comes from the AQS. Based on the population and exposure requirements for
monitor siting of 40 CFR Part 58 described in Section 3.4.2, monitors reporting to the AQS
are not uniformly distributed across the U.S. Monitors are far more abundant in urban
areas than rural ones, so actual rural spatiotemporal distributions may differ considerably
from those reported here. Furthermore, biases exist for some PM constituents (and hence,
total mass) owing to volatilization losses of NO3" and other semi-volatile compounds and,
conversely, retention of particle-bound water with hygroscopic species. The magnitude of
these effects is likely to be region-specific.
Spatial distributions of PM across a range of geographic scales are covered in
Section 3.5.1. Temporal behavior including trends, seasonality and hourly variability are
covered in Section 3.5.2. The section ends with statistical associations between different
size fractions of PM and copollutants including CO, NO2, O3 and SO2 in Section 3.5.3.
July 2009
3-60
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
3.5.1. Spatial Distribution
Spatial scales of interest for PM range from global and continental scales (>1000 km)
down to micro scale (~5-100m). Variation in PM concentration depends on the spatial scale
and magnitude of PM sources, formation and removal mechanisms, and transport and
dispersion of PM. These different sources and processes can cause substantial variation in
particle size distribution and chemistry. This section addresses the spatial variability of PM
by focusing primarily on AQS data across three different scales: variability across the U.S.,
urban-scale variability and neighborhood-scale variability. These sections are further
subdivided to the extent possible into PM size fractions and composition.
3.5.1.1. Variability across the U.S.
PM2.5
Figure 3"9 shows the three-yr mean of the 24-h PM2.5 concentrations by county across
the U.S. for 2005-2007. The data used in generating this map are from the AQS database
after applying a completeness criterion of 75% per quarter (or 11 out of 15 quarterly
measurements for a one-in-six-day sampling schedule). Counties shown in white did not
contain sufficient PM2.5 data meeting the completeness criterion for inclusion. Of the 3,225
U.S. counties, 540 (17%) had PM2.5 data meeting the completeness criterion in all three
years (2005-2007) and have been included in Figure 3-9. These 540 counties represent
roughly 63% of the U.S. population. The fraction of the population residing within each
county-average concentration range is shown on the left-hand margin of the figure. Given
the number of counties with no data, the varying size of counties, and the non-uniform
spacing of the monitors and population within each reporting county, this should only be
taken as a rough estimate of the relationship between population and average ambient
concentrations. Kern County, CA reported the highest 3-yr avg 24-h PM2.5 concentration in
excess of 20 pg/m3. Average concentrations between 18 and 20 |ug/m3 were reported for
several counties in the San Joaquin Valley and inland southern California as well as
Jefferson County, AL containing Birmingham and Allegheny County, PA containing
Pittsburgh.
July 2009
3-61
DRAFT-DO NOT CITE OR QUOTE

-------
Concentration Range
a > 20.1 pg/m3 [1 county]
18.1-20.0 ^g/m3 [7 counties]
15.1 - 18.1 ^g/m3 [53 counties]
12.1 - 15.0 pg/m3 [242 counties]
m £ 12.0 ^ig/m3 [237 counties]
~ No data
Population
(mil ions)
Figure 3-9. Three-yr avg 24-h PM2.5 concentration by county derived from FRM or FRM-like data,
2005-2007. The population bar shows the number of people residing within counties
that reported county-wide average concentrations within the specified ranges.
1	Table 3-8 contains summary statistics for PM2.5 reported to AQS for the period 2005-
2	2007. All 24-h FRM and 1-h FRM-like1 data reported to AQS and meeting the completeness
3	criterion outlined above are included in the table. The table provides a distributional
4	comparison between annual, 24-h and 1-h averaging times, calendar years (2005, 2006 and
1 FRM-like refers to PM2.5 concentration data associated with the parameter code "88502 ¦ Acceptable PM2.5 AQI and
Speciation Mass" in the EPA Air Quality System. These data were collected by continuous instruments which are not
approved as FRM or FEM, and consequently EPA does not use these data for regulatory purposes. These data are denoted as
"FRM-like" because state and local monitoring agencies have individually decided that the continuous instruments reporting
these data have a degree of agreement with FRM/FEM methods that is sufficient in their opinion for the data to be used in
public advisories regarding current air quality. In some cases, these data include statistical adjustments by the state/local
monitoring agency based on one-time or ongoing correlation analysis with co-located FRM/FEM monitors, intended to
improve the "FRM-likeness" of the continuous concentration data (see, for example, Bortnick et al. (2002, 156285) State/local
monitoring agency decisions about whether to adjust continuous PM2.5 data and whether their raw or adjusted continuous
PM2.5 data should be associated with parameter code 88502 were informed by non-binding EPA guidance issued in 2006
(Technical Note on Reporting PM2.5 Continuous Monitoring and Speciation Data to AQS
http V/www. epa. gov/ttn/n rn ti n/fi 1 flg/flmhi em t/pm 2R/H atamany/contrept .pdf).
July 2009
3-62
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
2007) and seasons: winter (December-February), spring (March-May), summer
(June-August), and fall (September-November). In addition, fifteen CSAs/CBSAs were
chosen for their importance in recent PM health studies, as described in Section 3.4, and
have been included individually in the table.
The distribution of PM2.5 annual averages—calculated without seasonal weighting
and presented in Table 3-8—was generated from 2382 individual annual means reported by
794 24-h FRM monitors reporting to AQS between 2005 and 2007. The mean of the annual
averages was 12 |ug/m3, equivalent to the mean of the individual 24-h avg The maximum
annual average PM2.5 concentration calculated from 24-h FRM data over these three years
was 23 |ug/m3 in Bakersfield, CA(AQS monitor ID: 060290010) during 2007. This site is
located in the heavily populated portion of the San Joaquin Valley where air pollution
frequently becomes trapped at ground level due to local topography. The distribution of the
24-h and 1-h avg, both generated from the same 1-h FRM-like data, are comparable up to
the 90th percentile. The 1-h avg is 3 |ug/m3 higher than the 24-h avg at the 95th percentile
and 7 ug/m3 higher at the 99th percentile. This deviation between 1-h and 24-h averaging
times is a result of short duration spikes in PM2.5 mass lasting long enough to influence the
upper percentiles of the 1-h distribution but not the 24-h avg distribution. Exceptional
events were not removed from this data set and are responsible for at least some of the
higher concentrations observed. For example, the maximum 1-h reading of 828 ug/m3 was
reported by a monitor in Boise, ID (AQS monitor ID: 160010011) on July 4, 2007. Nine of
the top 12 1-h PM2.5 concentrations reported across the country also occurred on July 4th,
implicating fireworks as the common source for these high values.
The distribution of the 24-h FRM PM2.5 data was similar across the three years (2005-
2007) investigated. Summer (June-August) had the highest mean and median relative to
other seasons, but only by a small margin. At the 99th percentile, winter (December-
February) was slightly higher than the other seasons. This is consistent with wintertime
stagnation events resulting in short-term elevated PM2.5 concentrations. Of the 15
CSAs/CBSAs investigated, the highest mean of 24-h PM2.5 concentrations was reported for
Riverside (17 |u.g/m3), Birmingham (16 ug/m3) and Pittsburgh (16 ug/m3); the lowest was
reported for Denver (9 |ug/m3) and Seattle (9 ug/m3).
July 2009
3-63
DRAFT-DO NOT CITE OR QUOTE

-------
Table 3-8. PM2.5 distributions derived from AQS data (concentration in //g/m3).


Mean




Percentiles





n
1
5
10
25
50
75
90
95
99
Max
20052007PM2.5 FOR DIFFERENT A VERAGING PERIODS
Annual avg" (24-h FRM)
2382
12
5
7
8
10
12
14
16
17
19
23
24-h avg (24-h FRM)
349,028
12
2
4
4
7
10
16
23
28
39
193
24-h avg (1-h FRM-like)
183,057
10
1
2
3
5
8
13
19
24
35
126
1-havg (1-h FRM-like)
4,403,817
10
0
1
2
4
8
13
21
27
42
828
PM2.5 ANNUAL AND SEASONAL STRA TIFICA HON USING 24 H AVG FRM DA TA
2005
114,346
13
2
4
5
7
11
17
24
30
42
133
2006
113,197
12
2
4
4
7
10
15
21
26
36
193
2007
121,485
12
2
4
4
7
10
16
22
27
40
145
Winter (December-February)
86,286
12
2
4
5
7
10
15
22
27
44
193
Spring (March-May)
88,489
11
2
3
4
6
9
14
20
24
33
145
Summer (June-August)
86,830
14
2
4
5
8
12
19
26
31
40
133
Fall (September-November)
87,423
12
2
3
4
6
10
15
22
26
39
126
20052007PM2.5 IN INDIVIDUAL CSAS/CBSAS USING 24 H AVG FRM DATA
Atlanta
4,939
15
4
6
7
10
14
19
25
29
37
145
Birmingham
4,869
16
4
6
7
10
15
21
29
34
47
64
Boston
8,464
10
2
3
4
5
9
13
20
24
32
50
Chicago
10,308
14
3
4
6
8
13
18
25
31
42
65
Denver
4,192
9
2
3
4
6
8
10
14
18
31
61
Detr oit
5,223
14
2
3
5
7
12
19
26
31
45
82
Houston
1,342
15
4
6
8
10
14
18
23
26
34
44
Los Angeles
6,600
15
3
5
6
9
13
18
25
32
50
133
New York
15,826
13
2
4
4
6
10
17
24
29
39
58
Philadelphia
7,541
14
3
4
5
8
12
18
25
30
38
63
Phoenix
1,634
10
2
3
4
6
9
12
17
21
32
77
Pittsburgh
5,783
16
3
5
6
9
13
20
29
36
52
101
Riverside
2,751
17
3
5
6
10
14
21
31
40
58
106
Seattle
1,297
9
2
3
3
4
7
10
20
29
43
68
St. Louis
6,887
14
3
5
6
9
13
18
24
29
40
50
All 15 CSAs/CBSAs
87,656
14
2
4
5
7
12
17
25
30
42
145
Not in the 15 CSAs/CBSAs
261,372
12
2
3
4
6
10
15
22
27
38
193
Straight annual average without quarterly weighting.
PM102.5
Since PM10-2.5 is not routinely measured and reported to AQS, co-located PM10 and
PM2.5 measurements from the AQS network were used to investigate the spatial
distribution in PM10-2.5. Only low-volume FRM or FRM-like samplers were considered in
July 2009
3-64
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
calculating PM10-2.5 to avoid complications with vastly different sampling protocols (e.g.,
flow rates) between the independent PM10 and PM2.5 measurements. The same 11+ days per
quarter completeness criterion discussed above was applied to the PM10 and PM2.5
measurements. The PM2.5 concentrations are reported to AQS at local conditions whereas
the PM10 concentrations are reported at standard conditions. Therefore, prior to calculating
PM10-2.5 by subtraction, the PM10 AQS data were adjusted to local conditions on a daily basis
using temperature and pressure measurements from the nearest National Weather Service
station. Figure 3" 10 shows the three-yr mean of the 24-h pmio-2.5 concentration by county
across the U.S. for 2005-2007. There is considerably less coverage than for PM2.5 or PM10
alone since only a small subset of PM monitors are co-located and low-volume. The 40
counties included in Figure 3-10 incorporate less than 5% of the U.S. population. Of the
3,225 U.S. counties, only 40 (1%) met the completeness and co-location criteria in all three
years (2005-2007), and therefore the available measurements do not provide sufficient
information to adequately characterize regional-scale coarse PM spatial concentration
distributions.
July 2009
3-65
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
Concentration Range
" > 26 pg/m3 [1 county]
21-25 iig/m3 [3 counties]
16-20 Lig/m3 [3 counties]
11-15 ^ig/m3 [17 counties]
B < 10 pg/m3 [16 counties]
~ No data
Population
(millions)
Figure 3-10. Three-yr avg 24-h PM102.5 concentration by county derived from co-located low volume
FRM PM10 and PM2.5 monitors, 2005-2007. The population bar shows the number of
people residing within counties that reported county-wide average concentrations
within the specified ranges.
Table 3"9 contains summary statistics for PM10-2.5 for the period 2005-2007 similar to
those reported in Table 3-8 for PM2.5. Only six of the 15 CSAs/CBSAs had sufficient data for
inclusion in Table 3-9. Although fewer monitoring sites within these CSAs/CBSAs were
used for PM10-2.5 than for PM2.5, Table 3-8 and 3-9 provide a rough comparison of the PM
present in the fine and thoracic coarse modes for these six cities. The eastern cities
including Atlanta, Boston, Chicago and New York all had a higher fraction in the fine mode
with the greatest ratio of fine to thoracic coarse in Chicago (14 jug/m3 PM2.5, 5 |ug/m3 PM10-2.5,
ratio = 2.8). In contrast, Denver (9 |ug/m3 PM2.5, 20 |ug/m3 PM10-2.5, ratio = 0.45) and Phoenix
(10 |ug/m3 PM2.5, 22 jug/m3 PM10-2.5, ratio = 0.45) had a higher fraction in the thoracic coarse
mode. Given the limited information available from AQS for PM10-2.5 and the current
National Ambient Air Quality Standard for PM10, the next section characterizes the more
July 2009
3-66
DRAFT-DO NOT CITE OR QUOTE

-------
1	prevalent PMio data, acknowledging that PMio incorporates both thoracic coarse and fine
2	particles.
Table 3-9. PM10 2.5
distributions derived from AQS data (concentration in //g/m3).











Percentiles





n
Mean
1
5
10
25
50
75
90
95
99
Max
20052007PM,0-2.5 FOR DIFFERENT A VERA GING PERIODS
Annual avga(low volume FRM)
130
12
3
5
6
9
11
14
19
23
39
43
24-h avg (low volume FRM)
12,027
13
¦3
1
2
6
10
17
26
33
54
246
PM,0-2.5 ANNUAL AND SEASONAL STRA TIFICA TION USING 24 H A VG LOW VOLUME FRM DA TA
2005
3,990
12
¦5
0
2
5
10
16
26
33
52
246
2006
4,037
13
¦2
1
2
6
10
17
27
34
56
182
2007
4,000
13
¦2
1
3
6
11
18
26
33
56
148
Winter (December-February)
2,942
11
¦5
¦1
1
4
8
15
27
34
56
246
Spring (March-May)
3,088
13
¦2
1
2
5
10
17
26
33
62
151
Summer (June-August)
2,968
14
¦2
3
5
8
12
18
25
31
44
93
Fall (September-November)
3,029
14
¦2
1
3
6
11
18
28
34
60
148
20052007 PM,0-2.5 IN INDIVIDUAL CSAS/CBSAS USING 24 H AVG LOW VOLUME FRM DATA1
Atlanta
167
10
¦4
1
2
5
9
13
18
21
30
46
Boston
340
7
¦2
1
2
4
6
9
12
16
25
27
Chicago
161
5
¦8
¦4
¦3
1
4
8
14
19
37
37
Denver
353
20
0
4
6
11
19
28
36
42
59
78
New York
338
9
¦16
¦2
1
5
8
12
17
23
34
56
Phoenix
163
22
¦3
8
11
16
20
29
35
46
67
70
All 6 CSAs/CBSAs
1,522
12
¦6
0
2
5
10
17
27
34
51
78
Not in the 6 CSAs/CBSAs
10,505
13
¦2
1
2
6
10
17
26
33
56
246
Straight annual average without quarterly weighting.
bNo co-located FRM PMio and FRM-like PM2.5 monitors present in Birmingham, Detroit, Houston, Los Angeles, Philadelphia, Pittsburgh, Riverside, Seattle or St. Louis.
PM10
3	Figure 3" 11 shows the 3-yr mean of the 24-h PMio concentrations by county across the
4	U.S. for 2005-2007. Both FRM and FEM PMio data reported to AQS were included and the
5	same 11+ days per quarter completeness criterion described above for PM2.5 was applied.
6	The highest 3-yr avg for PMio (>50 ug/m3) occurred in inland southern California and the
7	populous counties of southern Arizona and central New Mexico. Of the 3,225 U.S. counties,
8	676 (12%) contained PMio data meeting the completeness criterion in all three years! these
9	676 counties incorporate approximately 43% of the U.S. population.
July 2009
3-67
DRAFT-DO NOT CITE OR QUOTE

-------
Concentration Range
m > 51 Lig/m3 [7 counties]
41 - 50 ^ig/m3 [15 counties]
31-40 ^ig/rn3 [378 counties]
21-30 iig/m3 [162 counties]
m < 20 ^ig/m3 [114 counties]
~ No data
Population
(m llions)
Figure 3-11. Three-yr avg 24-h PM10 concentration by county derived from FRM or FEM monitors,
2005-2007. The population bar shows the number of people residing within counties
that reported county-wide average concentrations within the specified ranges.
1	Table 3" 10 contains summary statistics for PMio reported to AQS for the period 2005-
2	2007. Both 24-h FRM and 1-h FEM data are included in the table. To facilitate a
3	distributional comparison between averaging times, annual, 24-h and l"h averaging times
4	using the FRM and FEM data have been included separately in Table 3-10. As in the earlier
5	tables, the data is also stratified by year and season and includes the 15 CSAs/CBSAs
6	individually.
July 2009
3-68
DRAFT-DO NOT CITE OR QUOTE

-------
Table 3-10. PM10 distributions derived from AQS data (concentration in//g/m3).

n
Mean 1




Percentiles





1
5
10
25
50
75
90
95
99

20052007PMio FOR DIFFERENT A VERA GING PERIODS
Annual avg® (24-h FRM and 1-h FEM)
2022
25
10
14
16
19
23
28
35
44
60
85
24-h avg (24-h FRM and 1-h FEM)
326,675
26
3
6
9
14
21
32
46
59
97
8299
24-h avg (24-h FRM)
167,310
25
2
6
9
14
21
31
45
57
91
8299
24-h avg (1-h FEM)
156,931
26
4
7
9
14
21
32
48
62
105
979
1-h avg (1-h FEM)
3,767,533
27
1
4
6
11
19
32
51
69
145
8540
PMio ANNUAL AND SEASONAL STRA TIFICA HON USING 24 H A VG FRM AND FEM DA TA
2005
107,524
25
2
6
9
13
21
31
46
58
93
1441
2006
109,505
26
3
6
9
13
21
32
46
59
101
8299
2007
109,646
26
4
7
9
14
21
32
47
60
99
2253
Winter (December-February)
80,959
23
2
5
7
11
17
27
42
57
99
8299
Spring (March-May)
82,772
25
2
6
8
13
20
31
45
58
96
2253
Summer (June-August)
81,351
29
6
10
12
18
25
35
49
60
92
1839
Fall (September-November)
81,593
26
3
7
9
14
21
32
48
62
102
1212
20052007PMio IN INDIVIDUAL CSAS/CBS AS USING 24 H AVG FRM AND FEM DA TA
Atlanta
1,868
24
6
9
11
16
23
31
39
44
57
108
Birmingham
5,478
34
6
9
12
19
28
43
64
82
120
241
Boston
1,412
17
2
5
7
10
15
22
30
36
50
58
Chicago
6,165
26
6
9
11
16
23
32
45
55
78
214
Denver
4,706
28
5
10
12
18
25
35
47
54
75
118
Detroit
1,407
30
7
10
12
18
26
38
53
64
81
182
Houston
1,397
31
7
10
12
17
23
34
56
80
137
248
Los Angeles
2,020
27
4
8
11
18
25
33
42
51
74
489
New York
514
19
2
6
7
11
17
25
35
40
51
83
Philadelphia
4,207
19
4
7
9
12
17
24
34
40
52
84
Phoenix
12,005
52
7
14
19
29
44
65
91
112
166
2253
Pittsburgh
12,677
24
4
7
9
13
19
31
45
57
83
157
Riverside
4,327
35
4
8
11
19
30
45
64
75
111
1212
Seattle
2,136
19
5
7
9
12
17
23
31
37
52
79
St. Louis
2,464
33
6
10
12
18
28
42
59
74
114
315
All 15 CSAs/CBSAs
62,783
32
5
8
10
16
25
39
60
77
120
2253
Not in the 15 CSAs/CBSAs
263,892
24
2
6
8
13
20
30
43
54
88
8299
Straight annual average without quarterly weighting
The maximum annual average PM10 concentration calculated from 24-h FRM data
over these three years was 85 jug/m3 in Stanfield, AZ (AQS monitor ID: 040213008) during
2007. Stanfield is a small agricultural town (2007 population = 1074) approximately 64 km
south of Phoenix and is in a region heavily influenced by windblown dust. Many of the
July 2009
3-69
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
maximum 24-h and 1-h avg PMio concentrations in Table 3-10 exceed 1,000 |ug/m3, but
these represent rare events given the much lower 99th percentiles. Exceptional events were
not removed from this data set and are responsible for at least some of the higher
concentrations observed.
The distribution of the 24-h FRM and FEM PMio data was similar across the three
years (2005-2007) investigated. Summer (June-August) had the highest mean and median
relative to other seasons, consistent with PM2.5 observations. Of the 15 CSAs/CBSAs
investigated, the highest mean of 24-h PMio concentrations was reported for Phoenix (52
ug/ma), considerably higher than the means for the other CSAs/CBSAs investigated. The
lowest was reported for Boston (17 |ug/m3) with New York, Philadelphia and Seattle only
slightly higher (19 ug/ma).
On average using the 2005-2007 data for PM2.5 in Table 3-8 and PMio in Table 3-10,
the distribution between fine and coarse PM varies substantially by location. A larger
fraction of PM mass is present in the thoracic coarse mode in Phoenix and Denver (3-yr
mean PM2.5/PM10 ratios of 0.19 and 0.32, respectively). In contrast, a larger fraction is
present in the fine mode in Philadelphia (0.74), New York (0.68) and Pittsburgh (0.67).
Comparisons of PM2.5 to PMio as reported to AQS should be used with caution, however,
since PM2.5 concentrations are reported for local conditions while PMio concentrations are
converted to STP before reporting. Nevertheless, these findings are consistent with those in
Table 3-9 for PM10-2.5 in the subset of 6 cities with available co-located low-volume PM data
that have been properly adjusted for temperature and pressure. These findings are also
consistent with those reported in the previous PM AQCD (U.S. EPA, 2004, 056905) where
ratios of PM2.5 to PMio were observed to be highest in the northeast (0.70), southeast (0.70),
and industrial Midwest (0.70) and lower in the upper Midwest (0.53), northwest (0.50),
southern California (0.47) and southwest (0.38).
Ultrafine Particles
Little is known about the spatiotemporal distribution or composition of ultrafine
particles on a regional scale. New particle formation has been observed in environments
ranging from relatively unpolluted marine and continental environments to polluted urban
areas as an ongoing background process and during nucleation events (Kulmala et al.,
2004, 089159). During nucleation events, which may last for several hours, ultrafine
particle number concentrations can exceed 104cm3 over distances of several hundred
July 2009
3-70
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
kilometers (Kulmala et al., 2004, 089159; Qian et al., 2007, 116435). These events occur
throughout the year on 5-40% of days, depending on location (Qian et al., 2007, 116435).
Cloud condensation nuclei, with diameters between 10 and -100 nm have been monitored
for several years at a number of nonurban sites in the U.S.
(http://cmdl.noaa.gov/aero/data/). Average particle number counts at these sites in the U.S.
range from several hundred to several thousand per cm3. The particles are formed by
nucleation of atmospheric gases with a secondary contribution from primary emissions in
these environments (Pierce and Adams, 2009, 191189).
In an urban setting, a large percentage of ultrafine particles come from combustion-
related emissions from mobile sources (Sioutas et al., 2005, 088428). Ultrafine particle
number concentrations drop off quickly with distance from the roadway (Reponen et al.,
2003, 088425; Zhu et al., 2005, 157191; Levy et al., 2003, 052661), and therefore
concentrations can be highly heterogeneous in the near-road environment depending on
traffic, meteorological and topographic conditions (Baldauf et al., 2008, 190239). Studies
characterizing spatial variability in ultrafine particles are currently limited to a handful of
close proximity locations and therefore are discussed in Sections 3.5.1.2 and 3.5.1.3 in the
context of urban- and neighborhood-scale variability. Further elaboration on the
composition of ultrafine particles is included below.
PM Constituents
Only PM2.5 is collected routinely at CSN network sites so the majority of this
section on PM constituents is devoted to PM2.5. PM10-2.5 and ultrafine PM composition is
discussed to the extent possible below. Figures 3-12 through 3-16 contain U.S.
concentration maps for OC, EC, SO42-, NO3", and NH4"1" mass from PM2.5 measurements
taken as part of the CSN network for the period 2005-2007. Data used in these figures are
as reported to AQS: no correction was applied to OC for non-carbon mass and NO3"
represents total particulate NO3-. Figure 3-12 shows regions of high PM2.5 OC mass
concentration with annual average concentrations greater than 5 ug/m3 in the western and
the southeastern U.S. Concentrations at the western monitors peak in the fall and winter
while those in the Southeast peak anywhere from spring through fall. The central and
northeastern portions of the U.S. generally contain lower measured OC. Bell et al. (2007,
155683) present a similar map for estimated organic carbon mass (OCM) from 2000-2005
calculated by multiplying the blank corrected OC measurement by 1.4 to account for
July 2009
3-71
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
non-carbon mass. This differs from the raw OC values shown in Figure 3" 12. There are a
range of estimates in the literature for suggested scaling factors (Turpin and Lim, 2001,
017093), depending predominantly on how highly oxygenated the aerosol is (Pang et al.,
2006, 156012). Turpin and Lim (2001, 017093) recommended ratios of 1.6 ± 0.2 for urban
and 2.1 ± 0.2 for non-ruban aerosols. Fresh PM, more common in urban regions, has
undergone limited chemical transformation. As the aerosol is transported to rural regions,
it becomes more oxygenated. As a result, the necessary correction factor to account for non-
carbon mass (e.g., oxygen) is higher in rural locations compared to urban locations, with an
average of 1.9 and estimates ranging from 1.6 to 2.6 for rural IMPROVE monitors (El-
Zanan et al., 2005, 155764). Therefore, applying one correction factor of 1.4 across the
entire U.S. will lead to an underestimate of the OCM in rural regions. Therefore, the data
presented in Figure 3" 12 represent OC as measured and with a national blank correction,
but no adjustment to OCM.
Figure 3" 13 contains a similar map for PM2.5 EC mass concentration that exhibits
smaller seasonal variability than OC, particularly in the eastern half of the U.S. There are
isolated monitors spread throughout the country that measure high annual average EC
concentrations. These EC 'hot spots' are primarily associated with larger metropolitan
areas such as Los Angeles, Pittsburgh, and New York, but El Paso, TX, also reported high
annual average EC concentrations (driven by a wintertime average concentration greater
than 2 jug/m3). In a similar analysis for EC by Bell et al. (2007, 155683) for 2000-2005 data,
there were also high wintertime EC concentrations in eastern Kentucky and western
Montana. These particular locations do not stand out in the 2005-2007 data in Figure 3-13.
Figure 3-14 contains a map for PM2.5 SO 1- mass concentration which shows that
SO42- is more prevalent in the eastern U.S. owing to the strong west-to-east gradient in SO2
emissions. This gradient is magnified in the summer months when more sunlight is
available for photochemical formation of SO42In contrast, PM2.5 NOa mass concentration
in Figure 3-15 is highest in the west, particularly in California. There are also elevated
concentrations of N03~ in the upper midwest. The seasonal plots show generally higher
NOs- in the wintertime as a result of temperature driven partitioning. Exceptions exist in
Los Angeles and Riverside where high NOa readings appear year-round. The PM2.5 NH4"1"
mass concentration concentration maps in Figure 3-16 shows spatial patterns related to
both SO42- and N03~ resulting from its presence in both (NH4)2S04 and NH4NO3. Figures A-
29 through A"34 in Annex A show similar U.S. concentration maps for PM2.5 Cu, Fe, Ni, Pb,
July 2009
3-72
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Se and V mass concentrations as measured by XRF. There is considerably less seasonal
variation in the concentration profile for these metals than OC or the ions.
For the fifteen metropolitan areas identified earlier, the contribution of the major
component classes to total PM2.5 mass was derived using the measured sulfate, adjusted
NO3", derived water, inferred carbonaceous mass approach (SANDWICH) (Frank, 2006,
098909). This approach uses the measured FRM PM2.5 mass and co-located CSN chemical
constituents to perform a mass balance-based estimation of the PM2.5 mass fraction
attributed to SO42 , NO3 , EC, OCM, and crustal material. SO r and NOa include
associated NH4"1" mass and estimated particle-bound water. Furthermore, NOa is assumed
to be fully neutralized as NH4NO3 and has been adjusted to represent the amount retained
by the FRM monitor. EC is taken as measured, and the crustal component is derived from
common oxides contained in the Earth's crust (Pettijohn, 1957, 156862). but can also
include significant anthropogenic contributions, such as coal fly ash that are unrelated to
soil resuspension. Finally, OCM is estimated using mass balance by subtracting the sum of
all other constituents from the FRM PM2.5 mass. The SANDWICH method takes into
account passive collection of semi-volatile or handling-related mass on the FRM filters in
the mass balance calculation. The magnitude of this artifact is assigned a nominal value of
0.5 |ug/m3, which is derived from limited analysis of FRM field blanks. Other constituents
such as salt and other metallic oxides, however, are not included in these calculations and
therefore the OCM fraction estimated by mass balance represents an upper bound on the
FRM retained OCM. The calculations and assumptions that go into the SANDWICH
method are discussed in detail in Frank (2006, 098909)with further information available
on EPA's AirExplorer web site
(http://www.epa.gov/cgi-bin/htmSQL/mxplorer/querv spe.hsal).
July 2009
3-73
DRAFT-DO NOT CITE OR QUOTE

-------
oc
Annual
^ ••
o° • *
r
0	° * A o ^
^ ©V-. JOF
O
o o
1.
• ©
® °	® 6°^?®° "e
9
°° 'S
V
• . . o L+w*^.
®	o	O	®	•
Concentration (ng/m3): <1.25 >1.25-2.50 >2.50-3.75 >3.75- 5.30 >5.00
Spring	Summer
. °. . ' «° v«
©
o<^
o
o O
#¦
° *r«r °
%	° 0 f « w5S/	% ¦ '. ; - • 0 (8°	° ^ o^o®1
Roo o
Fall	Winter
¦ i %• - 		v	r\ ® ••
e • * • • •	• " At •* © * * ' •
• "	• *	„o' •„$_	• °
ta. H r °	*.	'** —*
\V/ "» ° v	/ " k
I t - V «f oW	1 •
% • "	° O • °°£ rf>.	\ ©
ig^l* "O
•	Tfc © ° ° « 0 °°80v
0T% °	^ o	a A °
• 9 n«jrT	•~• * r>. •*
bo • — *s\ .=. . .?• .V- • o. rw*9
-
Figure 3-12. Three-yr avg 24h PM2.5 0C concentrations measured at CSN sites across the U.S., 2005-
2007.
July 2009
3-74
DRAFT-DO NOT CITE OR QUOTE

-------
EC
Annual
§
o
0
o
oo
O Q ©
o
O O
8 o ©
o*
„ °
6
<3 O * oi°
If
° - >
* 8 <£00°*
O o 
 • ° •
fl
o -k#
o
?• 0
Concentration (ng/'m3): £-0.5
Spring
>0.5-10
>1.0-1.5. >1.5-2.:
>2.0
Summer

o o
1>0 J
e .
I *
o
• »
® ~4f
O	o
O	O
O	©
S>
o°
t f"
J—®
T
e 0
r*AS
O oO
V5* * °d#
0
* *8 * %v
' Vv 0
O CP cfi D
0 •
«¦
> 8
0 0
%
0 0
o°
0
&
9 0 0
0
°o 0 0 ojr 0
0 '"S.V
« 0. ° .
o0fi.°. 0
+\J/IZ A \ • '—¦** *•"
0
9 o ®
0
0

\ •

Fall

O
©
® ¦
n .	.
9>
o
& o ©
O ,e
* 3^*V
' o CP/
a	^ O
o a (P rtj i
)	^9 O o °{F
o a
0 °°om
® *o
o a
Oo O
* * Vi
. . ri	<,
0 
• o
Winter
o
o °
« o
« 1
o°
$ 9-L.l
%
O
8
«o «
*g
£° 0
O
«0 n o
" 'iV
'<> V'

Figure 3-13. Three-yr avg 24-h PM2.5 EC concentrations measured at CSN sites across the U.S., ZOOS-
ZOO?.
July 2009
3-75
DRAFT-DO NOT CITE OR QUOTE

-------
so42
Annual
o
o o
o \m>j
<*>o
cP
j \Jjte
tr* io
! o	'«*
19 » dioo°° •
°$ of-®0-
: °0 ¦£*•>
'a°
3	O O	0
o o * Cy*	\ \
o
o
Concentration (ng/m3): <2	>2-4
Spring
o
>4-6
>6-8
>8
Summer
o
OO
%
o
° o
OjjC^9

o	<*« °
° o *6
I > <8°°!? / o
°<« O?qg°o
• ° °°8D0^
® 0°o 0 °
o oOQ o
°°e
Q	• 1
%
V- ° Bo M
W
°° 0 • «v
3 o *iW
°e
Fall
Winter
o
00
o
%
°n «fc'«
cO
; Airir
o *P
ff
°s °
0 n^o °
B° O
o
OO
°o
o
*•
• 4
0o «o
¦TT*
S °0
» °»s °»°
Figure 3-14. Three-yr avg 24-h PM2.5 SO*2 concentrations measured at CSN sites across the U.S.
2005-2007.
July 2009
3-76
DRAFT-DO NOT CITE OR QUOTE

-------
N03-
Annual

e
9* -V .
o
o cpg>u " O
* « $©-00
o _0	C cpn
°e o o °J< o
o -*TT$ •
S	0 9	o
_ ° o „
.«<• o
Concentration (ng/m3): <1	>1-2
Spring
o	•
>2-3	>3-4
>4
Summer

V
°ov
® 0 jgo© 5
~ 0
fe
0 c
ft® >
• c o y
W
. . •. *>°* *vl
. -*• r,v
41 1	•
10	**>0
L	r W -f—2«Sr~» »
*	«0 0 B Of 0
o 0-0 ° „o
©	0 O	®
• • •
Fall
Winter
o o
°«fc
\
> • " *i
<%G Oe
0 f
a o-ff	„ "
* « <¦•;* 0
-VP—
0 • • °°s„ v
.1
o
\
Figure 3-15. Three-yr avg 24-h PM2.5 N(h concentrations measured at CSN sites across the U.S.
2005-2007.
July 2009
3-77
DRAFT-DO NOT CITE OR QUOTE

-------
nh4+
Annual
8
b
o
¦bo
B
% c£ ,Q o ° a •»
O °o	^%a, °«s^
<9
o •
e 300-
,0 „¦ - 85 op
°e o 0 off
° °og °o°
jgS^-T"
Concentration (jig/m3): <1
Spring
>1- 2	>2-3	>3-4	>4
Summer

"X5
a .
» « "
» Hv°s
o  "•« °0°
* %
0
©o
•k
a a
<*» !
6°
o " oJ
• 9 <800°° °
«„ " °c6> o®qS rj°
„ o °o8 °T°
Fall
Winter
o
9 ,
ty#"
. T. 8 kr& °
-v£—SS2*s
O ° °°S °C°
* r* JC )
o ."O •
o
a*" o
£f"C« ^0^:.
0	0 . O
Figure 3-16. Three-yr avg 24-h PM2.5 rJ!ii concentrations measured at CSN sites across the U.S.
2005-2007.
July 2009
3-78
DRAFT-DO NOT CITE OR QUOTE

-------
1	Figure 3-17 shows the PM2.5 compositional breakdown for the fifteen CSAs/CBSAs. All
2	available monitoring sites with co-located FRM PM2.5 and CSN speciation data reporting in
3	all four seasons for at least one calendar year from 2005-2007 were included. Furthermore,
4	each season was required to contain five reported values for mass and the major PM2.5
5	constituents. This resulted in a varying number of sites (ranging from one to seven, as
6	indicated in the caption to Figure 3-17) used to create the averages shown in the figure.
7	Variability in PM2.5 composition within each CSA/CBSA where multiple monitors were
8	available and trends in composition over time are discussed in subsequent sections.
July 2009
3-79
DRAFT-DO NOT CITE OR QUOTE

-------
Legend
] Sulfate
I Nitrate
I EC
] OCM
I Crustal
Figure 3-17. Three-yr avg PM2.5 speciation estimates for 2005-2007 derived using the SANDWICH
method for the following fifteen CSAs/CBSAs (with the number of sites per CSA/CBSA
listed in parenthesis): Atlanta, GA (1); Birmingham, AL (3); Boston, MA (4); Chicago, IL (7);
Denver, CO (2); Detroit, Ml (4); Houston, TX (1); Los Angeles, CA (1); New York City, NY
(7); Philadelphia, PA (6); Phoenix, AZ (2); Pittsburgh, PA (4); Riverside, CA (1); Seattle,
WA (4); and St. Louis, MO (3). SO42" and NO3 estimates include NH4+ and particle bound
water and the circles are scaled in proportion to FRM PM2.5 mass as indicated in the
legend.
1	On an annual average basis, SO42- is a dominant PM component in the eastern U.S.
2	cities. For the presented cities, this includes everything east of Houston where the S() r
3	fraction of PM2.5 ranges from 42% in Chicago to 56% in Pittsburgh on an annual average
4	basis. OCM is the next largest component in the east ranging from 27% in Pittsburgh to
5	42% in Birmingham. In the west, OCM is the largest constituent on an annual basis,
6	ranging from 34% in Los Angeles to 58% in Seattle. SO42 , NOs~ and crustal material are
7	also important in many of the included western cities. In the west, fractional S() r ranges
8	from 18% in Denver to 32% in Los Angeles while fractional NOs~ is relatively large in
9	Riverside (22%), Los Angeles (19%) and Denver (15%) and less important on an annual
July 2009
3-80
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
basis in Phoenix (1%) and Seattle (2%). Crustal material is particularly prevalent in
Phoenix (28%). EC makes up a smaller fraction of the PMa.s (4 to 11%), but it is consistently
present in all included cities regardless of region.
I I Sulfate . 20
~^¦Nitrate \^J
[==10CM	s 10
CU Crustal
Spring
Summer
Legend
Winter
Figure 3-18. Seasonally-stratified three-yr avg PM2.5 speciation estimates for 2005-2007 derived
using the SANDWICH method for the following fifteen CSAs/CBSAs: Atlanta, GA;
Birmingham, AL; Boston, MA; Chicago, IL; Denver, CO; Detroit, Ml; Houston, TX; Los
Angeles, CA; New York City, NY; Philadelphia, PA; Phoenix, AZ; Pittsburgh, PA;
Riverside, CA; Seattle, WA; and St. Louis, M0. SO42" and NO3 estimates include NH4+ and
particle bound water and the circles are scaled in proportion to FRM PM2.5 mass as
indicated in the legend.
4	The seasonal variation in PM2.5 composition across the fifteen CSAs/CBSAs is shown
5	in Figure 3" 18 where the seasons are defined as before. S()¦-¦ dominates in most metropoli-
6	tan areas in the summertime, while NO3- becomes important in the colder wintertime
7	months. Notable summertime exceptions include Denver, Phoenix, and Seattle, where SO12
8	makes up a smaller fraction of the PM2.5 mass. Likewise, NO:; is less pronounced in the
9	wintertime in Atlanta, Birmingham, Houston, Phoenix, and Seattle. Los Angeles and
10	Riverside exhibit elevated NO.t from fall through spring. Crustal material is a substantial
July 2009
3-81
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
summertime component in Houston (26%), and is generally low elsewhere in the East in all
seasons. In the West, crustal material represents a substantial component year-round in
Phoenix and Denver.
Only speciated PM2.5 is collected routinely at CSN network sites, resulting in far less
information on speciated PM10-2.5. Edgerton et al. (2005, 088686; 2009, 180385) published
speciated measurements for PM2.5 and PM10-2.5 obtained using dichotomous samplers from
four locations included in the Southeastern Aerosol Research and Characterization
(SEARCH) study: Yorkville, GA, Centreville, AL, Birmingham, AL and Atlanta, GA.
Samples were collected between 1999 and 2003 on a l-in-3 day or l-in-6 day schedule,
depending on site. Speciated measurements for both PM2.5 and PM10-2.5 included SO42 , NO3"
, NH4"1", and major metal oxides (MMO). In addition, OC and either black carbon (BC) or EC
were reported for PIVk.sover the entire study period and for PM10-2.5 for a subset of samples
extending from April 2003 to April 2004.
For the Atlanta and Birmingham SEARCH sites, the annual average NO3" mass
fraction was approximately equal for PM2.5 (5.6% and 5.0%, respectively, for Atlanta and
Birmingham) and PM10-2.5 (4.9% and 3.3%). Likewise, the OC mass fraction was
approximately equal for PM2.5 (26% and 26%) and PM10-2.5 (24% and 27%). MMO
contributed an order of magnitude smaller mass fraction to PM2.5 (2.6% and 4.7%) than
PM10-2.5 (38% and 35%). In contrast, S042~ contributed an order of magnitude greater mass
fraction to PM2.5 (25.1% and 24.1%) than PM10-2.5 (2.8 and 2.1%). BC also represented a
slightly larger mass fraction of PM2.5 (8.6% and 10.5%) than EC did for PM10-2.5 (2.9% &
2.4%). Based on these findings, MMO are present primarily in the thoracic coarse mode
while SO42- and EC/BC are present primarily in the fine mode. NO3" and OC are present in
both modes in approximately equal mass fractions. These results are specific to Atlanta and
Birmingham and may not represent other geographic regions. However, they are consistent
with the current understanding of sources and formation of these constituents and
therefore likely resemble the general compositional split between fine and thoracic coarse
mode particles.
Information about the composition of ambient ultrafine particles directly emitted by
sources is still sparse compared to that for the larger size modes. However, their
composition is expected to reflect that of their sources. As noted in Section 3.3 (and
references therein), particle number emissions from motor vehicles are dominantly in the
ultrafine size range. The composition of gasoline vehicle emissions consists mainly of a mix
July 2009
3-82
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
of OC, EC and small quantities of trace metals and sulfates, with OC constituting
anywhere from 26-88% of PM. Diesel PM is generally comprised of an EC and trace metal
ash core onto which organic material condenses, and nucleation-mode SOr . With the
introduction of new diesel emissions standards in 2007, total emissions have decreased
dramatically, particularly for carbon. In areas where nucleation is the dominant source of
ultrafine particles, sulfate along with ammonium, and secondary organic compounds are
the likely major components of ultrafine particles.
In a study conducted at several urban sites in Southern California, Cass et al. (2000,
020680) found that the composition of ultrafine particles ranged from 32-67% organic
compounds, 3.5-17.5% elemental carbon, 1-18% sulfate, 0-19% NO3", 0-9% ammonium,
1-26% metal oxides, 0-2% sodium, and 0-2% chloride. Thus carbon, in various forms, was
found to be the major contributor to the mass of ultrafine particles. However, ammonium
was found to contribute 33% of the mass of ultrafine particles at one site in Riverside. Iron
was the most abundant metal found in the ultrafine particles. Chung et al. (2001, 017105)
found that carbon was the major component of the mass of ultrafine particles in a study
conducted during January of 1999 in Bakersfield, CA. However, in the study of Chung et
al., the contribution of carbonaceous species (OC and EC; typically 20-30%) was much lower
than that found in the cities in Southern California. They found that calcium was the
dominant cation, accounting for about 20% of the mass of ultrafine particles in their
samples. Sizable contributions from silicon (0*4%) and aluminum (6-14%) were also found.
MOUDIs are used to collect size-segregated filter samples in the ultrafine particle
compositional analyses described above. Coarse particle bounce is a concern when using
MOUDIs and further studies, including scanning electron microscopy, may be needed to
quantify the effect of this sampling artifact on ultrafine particle compositional analyses.
Herner et al. (2005, 135983) reported a gradual increase in OC mass fraction as
particle size decreases from 1 um (20% OC) to 100 nm (80% OC) in the San Joaquin Valley
of California. Sardar et al. (2005, 180086) found OC to be the major component of ultrafine
particles at four locations in California, with higher OC mass fraction in the wintertime
relative to summertime. EC and S() r were also present in the ultrafine samples but at
much smaller mass fractions! EC was present year-round whereas S042~ had a summertime
preference. More detailed chemical characterization of the OC fraction of ambient ultrafine
particles is extremely limited, but recent studies have identified specific organic molecular
markers affiliated with motor vehicle emissions including hopanes and polycyclic aromatic
July 2009
3-83
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
hydrocarbons (Fine et al., 2004, 141283; Ning et al., 2007, 156809; Phuleria et al., 2007,
117816).
As noted in the 2004 PM AQCD (U.S. EPA, 2004, 056905), primary biological aerosol
particles (PBAP), which include microorganisms, fragments of living things, and organic
compounds of miscellaneous origin in surface deposits on filters, are not distinguishable in
analyses of total OC. A clear distinction should be made between PBAP and primary OC
produced by organisms (e.g., waxes coating the surfaces of organisms) and precursors to
secondary OC such as isoprene and terpenes. Indeed, the fields of view of many
photomicrographs of PM samples obtained by scanning electron microscopy are often
dominated by large numbers of pollen spores, plant and insect fragments, and
microorganisms. Bioaerosols such as pollen, fungal spores, and most bacteria are expected
to be found mainly in the coarse size fraction (see Figure 3-2 for an illustrative example of a
pollen particle). However, allergens from pollens can also be found in respirable particles
(Edgerton et al., 2009, 180385; Taylor, 2002, 025693). Matthias-Maser et al. (2000, 155972)
summarized information about the size distribution of PBAP in and around Mainz,
Germany in what is perhaps the most complete study of this sort. Matthias-Maser found
that PBAP constituted up to 30% of total particle number and volume in the approximate
size range from 0.35-50 urn on an annual basis. Additionally, whereas the contribution of
PBAP to the total aerosol volume did not change appreciably with season, the contribution
of PBAP to total particle number ranged from about 10% in December and March to about
25% in June and October. Bauer et al. (2008, 189986) measured contributions of fungal
spores to OC at an urban and a suburban site in Vienna, Austria in spring and summer.
Fungal spores at the suburban site contributed on average 10% to OC in PMio and 5% at
the urban site. At the suburban site, in summer, fungal spores accounted on average for
60% of the OC (0.56 |ug/m3) in PM10-2.1 (2.6 ug/m"). The contribution to PM2.1 was estimated
to be about 10% that in PM10-2.1. Womiloju et al. (2003, 179954) estimated that fungal
spores contribute 14-22% of OC in PM2.5 in and around Toronto.
Edgerton et al. (2009, 180385) found that PBAP contributed 60-70% of OC (average ~
1.7 |ug/m3) in PM10-2.5 at an urban and a rural site in Alabama in fall of 2000 and spring of
2001. The percentage contributions were similar at both sites and higher concentrations
were found in spring than in fall. Although results for the U.S. are more limited, they are
broadly consistent with the results of the other studies in illustrating the importance of
PBAP, at least for fungal spores in OC.
July 2009
3-84
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
3.5.1.2. Urban-Scale Variability
PM2.5
Data from the fifteen CSAs/CBSAs were used to investigate urban-scale variability in
PM reported to AQS. PM2.5 has a longer residence time in the atmosphere compared to
PMio'2.5 resulting from a slower Vd. As a result, PM2.5 exhibits increased spatial
homogeneity with less localized influence from point sources. Maps of PM2.5 monitor
locations and box plots of seasonal PM2.5 mass concentration data are provided for Boston
(Figure 3-19 and 3-23), Pittsburgh (Figure 3-21 and 3-25), and Los Angeles (Figure 3-23 and
3"27). Figures A-35 through A-70 in Annex A contain similar information for the remainder
of the 15 CSAs/CBSAs under investigation. With very few exceptions, the PM2.5
concentration is quite uniformly distributed across the monitors. Los Angeles has one
monitor (Site I) that reported noticeably less PM2.5 in all four seasons than the rest of the
monitors in the region. This monitor is located at Lancaster CA, separated from the rest of
the Los Angeles region by the San Gabriel Mountains. In general, however, PM2.5 varies
approximately the same magnitude between monitors as it does between seasons for the 15
selected cities.
Tables 3-11 through 3-13 contain pair-wise statistics for PM2.5 in Boston, Pittsburgh,
and Los Angeles, respectively. Tables A-20 through A-31 in Annex A contain pair-wise
statistics for PM2.5 measured within the remaining CSAs/CBSAs. Comparison statistics
shown include the Pearson correlation coefficient (R), the 90th percentile of the absolute
difference in concentrations (P90), the coefficient of divergence (COD) and the number of
paired observations (N). The COD provides an indication of the variability across the
monitoring sites in each CSA/CBSA and is defined as follows^
COD]k =.

p i=l x
Equation 3-2
where Xv- and X,/< represent observed concentrations averaged over some measurement
averaging period i (hourly, daily, etc.) at sites j and k, and p is the number of paired
observations. A COD of 0 indicates there are no differences between concentrations at
paired sites (spatial homogeneity), while a COD approaching 1 indicates extreme spatial
heterogeneity.
July 2009
3-85
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Temporal correlations between 24-h PM2.5 concentrations in Boston range from 0.61 to
0.97 in Table 3-11. The lowest correlation in this CSA was between Site A located in Fall
River, MA 1 km from the Narragansett Bay and Site L located in Nashua, NH on the bank
of the Merrimack River, 120 km north. The highest correlation was between Sites P and R,
located less than a kilometer apart in Providence, RI.
In Pittsburgh, 24-h PM2.5 correlations range from 0.65 to 0.97. The lowest correlation
in this CSA was between Sites B and D, located diametrically opposite downtown
Pittsburgh and 33 km apart. The highest correlation was for Sites K and D, located 21 km
apart and both west of downtown. The prevailing wind in Pittsburgh is from the west,
which explains the higher correlation between the two upwind sites.
In Los Angeles, 24-h PM2.5 correlations range from 0.21 to 0.96. The lowest correlation
was between Sites I and J located 123 km apart and separated by the San Gabriel
Mountains as discussed earlier. The highest correlation was for sites G and H, located 3.7
km apart and both in Long Beach, CA. Therefore, while distance between monitors plays an
important role in how well any two monitors correlate, other factors such as meteorology
and topography can be important.
July 2009
3-86
DRAFT-DO NOT CITE OR QUOTE

-------
A
Figure 3-19.
PM2.5 Monitors
Iriterstates
Major Highways
100
	1 Kilometers
Locations of PM2.5 monitors and major highways, Boston, MA.
• Boston
Boston
Boston
Boston
July 2009
3-87
DRAFT-DO NOT CITE OR QUOTE

-------
Site A
Site B
Site C
Site D
Site E
Site F
SiteG
Site H
Site I
Site J

A
B
C
D
E
F
G
H
I
J
Mean
9.1
9.1
8.9
9.4
9.6
11.7
11.6
10.5
121
10.7
Obs
341
350
342
355
357
349
398
349
1015
335
SD
6.0
6.5
6.3
6.6
6.6
7.0
6.8
6.9
6.9
7.2
AQS Site ID
25-005-1004
25-009-2006
25-009-5005
25-009-6001
25-023-0004
25-025-0002
25-025-0027
25-025-0042
25-025-0043
25-027-0016
1= winter
2=spring
3=summer
4=fall
S ite K
Site L
SiteM
Site N
Site 0
Site P
Site Q
Site R
SiteS
AQS Site ID
33-001-2004
33-011-1015
33-013-1006
33-015-0014
44-007-0022
44-007-0026
44-007-0028
44-007-1010
1 234 1 234 1 234 1 234 1 234 1 234 1 234 1 234 1 234 1 234

K
L
M
N
0
P
Q
R
S
Mean
11.4
7.2
10.0
9.7
8.9
10.1
11.9
10.5
9.7
Obs
346
183
361
362
362
1027
321
313
998
SD
7.4
5.3
6.7
6.0
5.8
6.6
6.9
6.5
6.5
1 =winter
2=spring
3=summer
4=fall
1 2 3 4 1 2 3'
1234 1234 1234 1234 1234 1234 1234
Figure 3-20. Seasonal distribution of 24-h avg PM2.5 concentrations by site for Boston, MA, ZOOS-
ZOO?. Box plots show the median and interquartile range with whiskers extending to
the 5th and 95th percentiles at each site during (1) winter (December-February), (2)
spring (March-May), (3) summer (June-August) and (4) fall (September-November).
July 2009
3-88
DRAFT-DO NOT CITE OR QUOTE

-------
Table 3-11. Inter-sampler comparison statistics for each pair of 24-h PM2.5 monitors reporting to AQS
for Boston, MA.
Site
A
B
C
D
E
F
G
H
I
J
A
1.00
0.80
0.77
0.71
0.84
0.79
0.78
0.79
0.79
0.77

(0.0,0.00)
(6.6,0.21)
(6.2,0.22)
(6.9,0.23)
(4.8,0.19)
(8.1,0.23)
(7.7,0.24)
(6.8,0.22)
(7.9,0.25)
(7.5,0.24)

341
326
318
323
329
318
319
325
338
310
B

1.00
0.92
0.87
0.87
0.90
0.90
0.90
0.90
0.85


(0.0,0.00)
(4.1,0.17)
(4.1,0.18)
(4.7,0.19)
(6.3,0.21)
(6.2,0.23)
(4.9,0.19)
(7.1,0.26)
(5.5,0.21)


350
328
331
339
326
323
333
343
317
C


1.00
0.90
0.85
0.90
0.89
0.90
0.88
0.86



(0.0,0.00)
(3.5,0.17)
(5.3,0.21)
(6.3,0.23)
(6.3,0.24)
(5.0,0.20)
(6.8,0.26)
(6.2,0.21)



342
321
331
316
318
326
336
311
D



1.00
0.80
0.88
0.88
0.86
0.86
0.87

KEY


(0.0,0.00)
(5.6,0.20)
(5.8,0.21)
(5.8,0.22)
(4.6,0.19)
(7.0,0.26)
(5.8,0.19)

Pearson R


355
336
324
329
332
345
313
E
(P90, COD)



1.00
0.90
0.90
0.89
0.87
0.87

n



(0.0,0.00)
(5.9,0.19)
(5.8,0.21)
(5.0,0.19)
(6.9,0.24)
(5.4,0.20)





357
330
333
340
350
322
F





1.00
0.94
0.94
0.92
0.92






(0.0,0.00)
(3.8,0.14)
(3.5,0.15)
(4.5,0.17)
(5.4,0.18)






349
324
324
339
310
G






1.00
0.94
0.94
0.89







(0.0,0.00)
(4.0,0.16)
(4.3,0.15)
(5.7,0.20)







398
325
338
308
H







1.00
0.93
0.89








(0.0,0.00)
(4.7,0.19)
(5.0,0.17)








349
342
318
I 1.00 0.86









(0.0,0.00)
(6.9,0.23)









1015
330
J









1.00
(0.0,0.00)
335

Site
K
L
M
N
0

P
Q
R
S
A
0.77
0.61
0.71
0.68
0.73

0.87
0.81
0.85
0.86

(8.1,0.23)
(8.3,0.29)
(8.0,0.23)
(7.9,0.23)
(7.0,0.22)

(5.3,0.18)
(7.2,0.23)
(5.6,0.20)
(5.2,0.18)

320
173
324
334
331

326
292
285
306
B
0.86
0.80
0.87
0.83
0.88

0.86
0.80
0.85
0.85

(6.6,0.21)
(6.2,0.23)
(5.3,0.19)
(6.0,0.21)
(4.7,0.18)

(5.6,0.19)
(7.9,0.26)
(5.7,0.21)
(6.0,0.19)

329
175
331
341
336

335
300
288
314
C
0.86
0.89
0.93
0.90
0.93

0.83
0.79
0.81
0.82

(6.9,0.21)
(4.8,0.23)
(4.4,0.17)
(4.6,0.19)
(3.8,0.18)

(5.9,0.21)
(7.8,0.26)
(6.2,0.23)
(6.0,0.21)

321
173
323
335
328

329
290
281
309
D
0.88
0.79
0.91
0.85
0.86

0.80
0.75
0.79
0.80

(6.4,0.19)
(5.7,0.25)
(3.5,0.16)
(4.7,0.19)
(4.2,0.18)

(6.2,0.20)
(7.8,0.25)
(6.2,0.21)
(5.8,0.20)

325
174
329
339
334

342
300
287
321
E
0.87
0.72
0.83
0.79
0.84

0.91
0.86
0.88
0.91

(6.3,0.20)
(8.3,0.27)
(5.8,0.17)
(6.3,0.20)
(4.8,0.18)

(4.5,0.17)
(6.3,0.22)
(4.9,0.18)
(3.9,0.17)

333
179
338
347
343

343
306
295
324
F
0.91
0.78
0.90
0.85
0.85

0.89
0.86
0.88
0.89

(4.7,0.17)
(9.6,0.33)
(5.3,0.18)
(6.4,0.20)
(7.5,0.22)

(5.2,0.16)
(6.0,0.16)
(4.9,0.16)
(5.5,0.17)

323
168
323
334
330

336
295
281
316
G
0.90
0.77
0.90
0.85
0.87

0.88
0.86
0.87
0.88

(5.0,0.19)
(9.0,0.33)
(5.3,0.19)
(6.3,0.20)
(7.0,0.22)

(5.5,0.17)
(5.3,0.17)
(5.2,0.17)
(5.7,0.19)

320
172
326
335
329

383
296
282
356
H
0.90
0.75
0.88
0.83
0.84

0.89
0.86
0.87
0.88

(4.4,0.17)
(9.4,0.30)
(4.9,0.18)
(5.6,0.21)
(6.8,0.21)

(4.5,0.16)
(6.0,0.19)
(4.5,0.16)
(5.1,0.17)

327
175
332
341
336

335
299
289
314
I
0.87
0.75
0.86
0.82
0.83

0.88
0.84
0.85
0.87

(6.1,0.20)
(10.0,0.36)
(6.7,0.22)
(7.2,0.23)
(8.2,0.25)

(6.1,0.20)
(6.0,0.16)
(6.0,0.18)
(6.3,0.21)

341
181
352
356
357

957
314
306
936
J
0.95
0.73
0.87
0.84
0.80

0.90
0.86
0.87
0.88

(3.0,0.14)
(9.2,0.28)
(5.2,0.18)
(5.9,0.20)
(7.5,0.22)

(5.0,0.17)
(5.9,0.20)
(5.3,0.17)
(5.2,0.18)

316
167
314
326
323

321
283
272
302
K
1.00
0.71
0.88
0.85
0.81

0.89
0.86
0.87
0.88

(0.0,0.00)
(10.3,0.31)
(6.0,0.16)
(6.5,0.19)
(8.2,0.22)

(5.2,0.16)
(5.8,0.18)
(5.5,0.16)
(5.5,0.18)

346
170
326
337
332

331
296
286
313
L

1.00
0.89
0.91
0.90

0.68
0.63
0.72
0.69


(0.0,0.00)
(6.7,0.24)
(5.9,0.23)
(4.8,0.21)

(10.0,0.29)
(12.1,0.35)
(9.1,0.30)
(9.8,0.29)


183
176
181
177

181
153
149
164
M


1.00
0.94
0.90

0.83
0.81
0.82
0.84



(0.0,0.00)
(3.8,0.13)
(4.6,0.16)

(5.5,0.16)
(7.4,0.20)
(5.8,0.17)
(5.1,0.16)



361
341
336

345
300
288
326
N



1.00
0.90

0.77
0.75
0.78
0.78




(0.0,0.00)
(4.4,0.17)

(6.7,0.19)
(8.1,0.22)
(6.4,0.20)
(6.2,0.19)




362
346

347
309
297
327
0




1.00

0.80
0.75
0.79
0.80





(0.0,0.00)

(5.8,0.19)
(8.8,0.25)
(6.8,0.21)
(6.0,0.19)





362

348
304
292
330
P






1.00
0.95
0.97
0.97







(0.0,0.00)
(3.6,0.14)
(2.0,0.09)
(2.1,0.08)







1027
307
299
943
July 2009
3-89
DRAFT-DO NOT CITE OR QUOTE

-------
Site
K I
M
N
0
P
Q
R
S
Q





1.00
0.92
0.94






(0.0,0.00)
(3.1,0.13)
(4.0,0.16)






321
268
290
R






1.00
0.94







(0.0,0.00)
(2.7,0.12)







313
280
S







1.00
(0.0,0.00)
908
Pittsburgh PM2.5 Monitors
	 Pittsburgh Interstates
/•		 Pittsburgh Major Highways
Pittsburgh
0 10 20	40	60	80	100
i Kilometers
Figure 3-21. Locations of PM2.5 monitors and major highways, Pittsburgh, PA.
July 2009
3-90
DRAFT-DO NOT CITE OR QUOTE

-------
A	B	C	D	E	F
Mean 15.1
19.8
13.2
13.6
15.1
16.4
Obs
1063
1066
306
165
332
337
SD
8.9
14.7
8.0
8.5
9.0
9.5
Site A
Site B
SiteC
SiteD
SiteE
Site F
SiteG
SiteH
Site I
Site J
Site K
Site L
AQS Site ID
42-003-0008
42-003-0064
42-003-0067
42-003-0095
42-003-1008
42-003-1301
1=winter
2=spring
3=summer
4=fall
AQS Site ID
42-003-3007
42-007-0014
42-125-0005
42-125-0200
42-125-5001
42-129-0008
1 = winter
2=spring
3-summer
4=fall


III
1234 1234 1234 1234 1234 1234

G
H
I
J
K
L
Mean
15.3
16.4
15.5
14.8
13.4
15.4
Obs
171
328
354
345
966
350
SD
8.3
9.3
8.6
8.0
8.6
8.7
1234 1234 1234 1234 1234 1234
Figure 3-22. Seasonal distribution of 24-h avg PM2.5 concentrations by site for Pittsburgh, PA, 2005-
2007. Box plots show the median and interquartile range with whiskers extending to
the 5th and 95th percentiles at each site during (1) winter (December-February), (2)
spring (March-May), (3) summer (June-August) and (4) fall (September-November).
July 2009
3-91
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Table 3-12. Inter-sampler comparison statistics for each pair of 24-h PM2.5 monitors reporting to AQS
for Pittsburgh, PA.

A
B
C







K
L
A
1.00
0.79
0.95
0.92
0.93
0.95
0.95
0.85
0.90
0.93
0.91
0.88

(0.0,0.00)
(15.9,0.19)
(5.6,0.13)
(4.7,0.11)
(4.7,0.11)
(4.9,0.10)
(3.8,0.10)
(6.4,0.13)
(6.4,0.13)
(5.0,0.12)
(6.0,0.13)
(5.6,0.12)

1063
1035
298
164
323
329
170
319
344
337
934
340
B

1.00
0.71
0.65
0.80
0.85
0.76
0.69
0.71
0.68
0.68
0.67


(0.0,0.00)
(16.9,0.24)
(17.4,0.25)
(14.4,0.19)
(12.5,0.14)
(15.7,0.20)
(17.0,0.19)
(15.7,0.21)
(17.8,0.23)
(19.3,0.25)
(15.9,0.21)


1066
303
165
329
335
171
324
350
341
938
346
C


1.00
0.93
0.90
0.91
0.94
0.80
0.93
0.96
0.95
0.91



(0.0,0.00)
(2.8,0.09)
(6.6,0.16)
(8.7,0.17)
(6.0,0.14)
(9.4,0.19)
(6.7,0.15)
(4.6,0.12)
(4.5,0.10)
(6.5,0.15)



306
144
282
282
148
268
290
286
270
286
D



1.00
0.84
0.87
0.91
0.79
0.89
0.91
0.97
0.85




(0.0,0.00)
(6.4,0.15)
(8.5,0.16)
(5.8,0.13)
(9.2,0.17)
(5.9,0.13)
(4.6,0.11)
(3.1,0.08)
(6.5,0.15)




165
153
161
158
156
158
155
146
157
E




1.00
0.90
0.90
0.84
0.85
0.86
0.88
0.83





(0.0,0.00)
(6.4,0.13)
(6.5,0.13)
(6.8,0.14)
(8.3,0.16)
(7.7,0.16)
(7.6,0.15)
(7.3,0.15)



KEY

332
313
157
295
320
315
290
318
F

Pe
arson R


1.00
0.91
0.82
0.88
0.88
0.89
0.86


(P90, COD)


(0.0,0.00)
(6.7,0.13)
(7.4,0.14)
(7.1,0.15)
(7.9,0.15)
(8.8,0.17)
(7.0,0.14)



n


337
167
302
327
319
296
322
G






1.00
0.78
0.94
0.93
0.90
0.91







(0.0,0.00)
(7.3,0.16)
(4.0,0.10)
(5.0,0.11)
(6.6,0.15)
(5.0,0.13)







171
159
163
159
149
161
H







1.00
0.80
0.78
0.82
0.70








(0.0,0.00)
(8.4,0.15)
(8.2,0.17)
(9.0,0.18)
(9.2,0.18)








328
317
309
288
314
1








1.00
0.93
0.89
0.88









(0.0,0.00)
(5.0,0.11)
(7.2,0.16)
(6.0,0.13)









354
334
310
339
J









1.00
0.93
0.88










(0.0,0.00)
(5.5,0.12)
(5.9,0.13)










345
302
331
K










1.00
0.86
(0.0,0.00) (6.9,0.15)
	966	306
_L	1.00
	(0.0,0.00)
	350
To further investigate the relationship between correlation and distance, Figures 3-28
through 3-30 plot inter-sampler correlation as a function of distance between monitors for
PM2.5 in Boston, Pittsburgh, and Los Angeles. These three cities were selected to illustrate
how this relationship varies across urban areas with different topography and climatology
as well as different PM2.5 sources, compositions and monitor densities. Plots are provided in
Annex A for the remainder of the 15 CSAs/CBSAs under investigation beginning with
Figure A-l. The Boston data exhibit the strongest relationship between inter-sampler
correlation and distance, with average inter-sampler correlation remaining higher than
80% when samplers are 95 km apart (R2 = 0.55). This small amount of variability is
expected given the consistency between distributions shown in the corresponding box plots
(Figure 3"20). The Pittsburgh data show some reductions in inter-sampler correlations at
short distances, with the samplers at Sites B and G having only 76% correlation with a
distance of less than 4 km. Site B is located in Liberty, PA, a mountainous suburb of
Pittsburgh where emissions from steel manufacturing and frequent stable conditions in the
planetary boundary layer cause localized events of elevated concentration. In contrast, Site
July 2009
3-92
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
G is in the neighboring town of Clairton, PA, located at a lower elevation on the bank of the
opposite side of the Monongahela River from Liberty. On average, inter-sampler correlation
remained higher than 80% when samplers were separated by 61 km, but in this case with
much greater scatter (R2 = 0.22) than observed in the Boston data. This scatter is driven by
the measurements at Site B; Figure 3-22 shows an elevated mean and variability for this
site compared with other monitors situated around the Pittsburgh CSA. When data from
Site B are removed, the inter-sampler correlation vs. distance plot for Pittsburgh PM2.5
resembles the one from Boston (with R2 increasing to 0.68). The Los Angeles data exhibit a
much steeper slope, with average inter-sampler correlation remaining higher than 80%
when samplers are 29 km apart (R2 = 0.74). This suggests that other factors, such as
mountainous topography separating monitors, the distribution of traffic and suspension of
crustal components, and occurrence of stable boundary layers, may cause more spatial
variation in the PM2.5 concentration profile within the Los Angeles region when compared
with other parts of the country. The Site I monitor, separated from the rest of the Los
Angeles region by the San Gabriel Mountains as mentioned above, provides the low
correlations grouped in the lower right portion of Figure 3-27. It should also be noted in
examining Figures 3-19 through 3-23 that some monitors are often located close to major
interstate highways while others in the same urban area are not. These differences in
proximity of monitors to nearby major roads will also result in lower inter-monitor
correlations.
July 2009
3-93
DRAFT-DO NOT CITE OR QUOTE

-------
o)
Los Angelas PM2.5 Monitors
- Los Angeles Interstates
Los Angetes Major Highways
Los Angetes
80
100
—i Kilometers
Figure 3-23. Locations of PM2.5 monitors and major highways, Los Angeles, CA.
July 2009
3-94
DRAFT-DO NOT CITE OR QUOTE

-------
Site A
SiteB
SrteC
Site D
Sitt?E
SrteF
Srte G
SiteH
Site I
Site J
SitpK
Mcdn
Obs
SD
60
AQS Site ID
06-037-0002
06037-1002
06-037 1103
06-037-1201
06037-1301
06-037-2005
06-037-4002
06-037-4004
06037 9033
06059-0007
06059-2022
1 =winter
2=spring
3-summer
4=fdll
161
862
10.8
B
17.0
308
10.2
C
16.7
1004
9.8
O
13.3
291
7.5
16.7
327
9.3
14.3
334
8.9
G
14.7
946
8.4
14.2
990
7.7
221
3.8
J
14.4
999
8.5
10.9
318
64
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
Figure 3-24. Seasonal distribution of 24-h avg PM2.5 concentrations by site for Los Angeles, CA,
2005-2007. Box plots show the median and interquartile range with whiskers extending
to the 5th and 95th percentiles at each site during (1) winter (December-February), (2)
spring (March-May), (3) summer (June-August) and (4) fall (September-November).
Table 3-13. Inter-sampler comparison statistics for each pair of 24-h PM2.5 monitors reporting to AQS
for Los Angeles, CA.

A
B
C
D
E
F
G
H
I
J
K
A
1.00
0.86
0.87
0.81
0.80
0.88
0.68
0.64
0.30
0.70
0.82

(0.0,0.00)
(9.0,0.18)
(7.7,0.16)
(9.0,0.19)
(9.7,0.21)
(5.8,0.14)
(11.5,0.22)
(12.4,0.23)
(18.0,0.36)
(10.5,0.21)
(11.4,0.23)

862
252
803
238
262
269
761
793
179
804
259
B

1.00
0.92
0.87
0.83
0.88
0.77
0.73
0.31
0.74
0.71


(0.0,0.00)
(5.5,0.11)
(9.1,0.19)
(9.0,0.15)
(7.6,0.15)
(9.8,0.17)
(11.6,0.18)
(24.1,0.38)
(11.9,0.19)
(15.0,0.27)


308
293
250
278
279
268
282
177
292
277
C


1.00
0.80
0.89
0.92
0.84
0.79
0.29
0.82
0.78



(0.0,0.00)
(9.6,0.20)
(5.8,0.11)
(6.4,0.13)
(9.0,0.15)
(10.0,0.17)
(18.6,0.38)
(9.4,0.16)
(13.2,0.25)



1004
274
315
319
880
913
213
920
305
D



1.00
0.69
0.77
0.63
0.60
0.41
0.64
0.60




(0.0,0.00)
(10.9,0.23)
(7.4,0.18)
(11.3,0.22)
(11.1,0.22)
(14.8,0.31)
(9.6,0.21)
(11.6,0.23)




291
263
263
256
268
164
274
261
E




1.00
0.79
0.95
0.92
0.34
0.88
0.76





(0.0,0.00)
(9.1,0.19)
(5.9,0.11)
(7.6,0.13)
(19.7,0.39)
(8.2,0.15)
(13.7,0.27)





327
301
289
301
192
307
291
F





1.00
0.70
0.70
0.33
0.69
0.72


KEY



(0.0,0.00)
(10.5,0.18)
(9.2,0.19)
(14.8,0.34)
(9.8,0.19)
(9.9,0.21)


Pearson R



334
290
302
184
311
293
G

(P90, COD)




1.00
0.96
0.23
0.92
0.78


n




(0.0,0.00)
(4.0,0.09)
(17.0,0.35)
(5.4,0.12)
(11.0,0.21)







946
859
194
882
277
H







1.00
0.26
0.91
0.78








(0.0,0.00)
(15.3,0.34)
(5.9,0.12)
(9.5,0.21)








990
208
914
294
I








1.00
0.21
0.31









(0.0,0.00)
(18.3,0.35)
(9.7,0.28)









221
205
180
J









1.00
0.84










(0.0,0.00)
(9.8,0.19)
ggg 298
K	1.00
[0.0,0.00!
July 2009
3-95
DRAFT-DO NOT CITE OR QUOTE

-------
~ ~
+*r :
* * * * >
« 4» ~« * » ~ ~	* ^	4» fa.t - .~	«	. .
< ••. V . * •
~	~	~	~	4 ~
	i	A . a .	aJ	~ ~
40	50	60	70	80	90	100
Distance Between Samplers (km)
Figure 3-25. Inter-sampler correlations for 24-h PM2.5 as a function of distance between monitors in
Boston, MA.
0.6
0.4
0.2
~ f A	~~ ~	~
~ * ~ *
~	~ * C # ~ ~
f ~	. 4	. 4	4	<
~	~
10	20	30	40	50	60
Distance Between Samplers (km)
70	80	90	100
Figure 3-26.
Inter-sampler correlations for 24-h PM2.5 as a function of distance between monitors in
Pittsburgh, PA.
July 2009
3-96
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
0.8
0.6
O
o
0.4
0.2
~
~	4*
~
~ ~
~
~ %
~
~
10	20	30	40	50	60
Distance Between Samplers (km)
70
80
90
100
Figure 3-27. Inter-sampler correlations for 24-h PM2.5 as a function of distance between monitors in
Los Angeles, CA.
PM102.5
Given the limited number of co-located low volume FRM PM10 and FRM PM2.5
monitors, only a very limited investigation into the intra "urban spatial variability of PM10-
2.5 was possible using AQS data. Of the 15 cities under investigation, only six (Atlanta,
Boston, Chicago, Denver, New York and Phoenix) contained data sufficient for calculating
PM10-2.5 according to the data completeness and monitor specification requirements
discussed earlier. Figure 3-28 contains box plots of PM10-2.5 for one or two available sites per
CSA/CBSA providing adequate PM10-2.5 concentration data. For Boston, the correlation
between the two sites for PM10-2.5 was 0.45 compared with 0.73 for PM2.5 alone and 0.84 for
PM10 alone (using the same two monitoring sites). For New York, the correlation was
slightly higher for the two sites: 0.74 for PM10-2.5 compared with 0.93 for PM2.5 alone and
0.82 for PM10 alone. The COD for PM10-2.5 also increases in both cities compared with PM2.5
and PM10 alone, suggesting less spatial homogeneity for thoracic coarse particles compared
with fine particles. Wilson and Suh (1997, 077408) reported PM10-2.5 correlations between
eight sites in Philadelphia ranging from 0.14 to 0.63 with an average of 0.38. This was
considerably less than the corresponding average correlation for PM2.5 (r = 0.90) and PM10
(r = 0.87) from the same study. Thornburg et al. (2009, 190999) also reported a high degree
July 2009
3-97
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
of spatial variability in PM10-2.5 in Detroit with between-monitor correlations ranging from
0.03 to 0.76. These results suggest that local sources can have a substantial impact on
PM10-2.5 concentrations, resulting in a higher degree of spatial variability in PM10-2.5 relative
to PM2.5 or PM10.
Atlanta
Boston
Chicago
Denver
New York
Phoenix
AOS SilsiQ
anB Siso A	13-121-0032
Sinn Site A	25-025-ffiW
SfeB	33-01S-0014
icago Site A	18-127-3Q24
nyer SifeA	08-5301-0008
w Yotli Site A	09-001-0010
Site B	09-C09-Z133
earns Sits A	04-013-399?
12 3 4
1 2 3 4 1 2 3 4
12 8 4
1 2 3 4 1 2 3
12 3 4
Figure 3-28.
Seasonal distribution of 24-h avg PM 10-2.5 concentrations by site for Atlanta, GA; Boston,
MA; Chicago, IL; Denver, CO; New York City, NY; and Phoenix, AZ; 2005-2007. Box plots
show the median and interquartile range with whiskers extending to the 5th and 95th
percentiles at each site during (1) winter (December-February), (2) spring (March-May),
(3) summer (June-August) and (4) fall (September-November). Note the different
concentration scales on the y-axes.
PM10
PM10 mass concentration has been shown to vary as much as a factor of five over
urban-scale distances of 100 km or less, and by a factor of 2 or more on scales as small as 30
km in an analysis of California air quality (Alexis et al., 2001, 079886). This can be
attributed to the rapid Vd and resulting short atmospheric lifetime of the coarse-mode
particles making up much of PM10 mass. As a result, local emission sources often dominate
PM10 annual average mass at certain monitors. Data from the fifteen CSAs/CBSAs were
used to investigate urban variability in PM10 reported to the AQS database.
Maps of PM10 monitor locations and box plots of seasonal PM10 mass concentration
data are provided for Boston (Figures 3-32 and 3-33), Pittsburgh (Figures 3-34 and 3-35),
and Los Angeles (Figures 3-36 and 3-37) similar to the PM2.5 maps and box plots shown
earlier in Figures 3"38 through 3*40 Annex A, Figures A-71 through A-106 incorporate
July 2009
3-98
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
similar information for the remainder of the 15 CSAs/CBSAs. Tables 3-14 through 3-16
contain pair wise, within-city comparison statistics (R, P90, COD and N, as defined above)
for PMio measured at the available monitors in Boston (Table 3-14), Pittsburgh (Table 3-15)
and Los Angeles (Table 3-16); the remainder of the 15 CSAs/CBSAs are included in Annex
A, Tables A-32 through A-43.
Boston is an example of a city with a wide range in concentrations measured at
different sites. Inter-monitor variation in PMio is frequently larger than the seasonal
variation measured at any given site. Pairwise correlations between monitors in Boston
range from 0.45 to 0.95 in Table 3-14. Pittsburgh is an example of a city with a large
number of PMio monitors providing consistent values with a select few reporting higher
concentrations (sites D, H, I and K in Figure 3-32). This illustrates the potential influence
of localized point or area sources or topography. Correlations between monitors in
Pittsburgh range from 0.47 to 0.97 in Table 3-15. Los Angeles shows a high degree of
between-season and within-season variability, which is on the order of the between-monitor
variation. Correlations between monitors in Los Angeles range from 0.29 to 0.93 in
Table 3-16. Once again, the lowest correlations are with the monitor separated from the
other monitors by the San Gabriel Mountains (Site E in Figure 3"33).
July 2009
3-99
DRAFT-DO NOT CITE OR QUOTE

-------
60 80 100
i Kilometers
• Boston PM10 Monitors
	Boston Interstate^
Boston Major Highways
Boston
Figure 3-29. Locations of PM10 monitors and major highways, Boston, MA.
July 2009
3-100
DRAFT-DO NOT CITE OR QUOTE

-------
Site A
SiteB
SlteC
SiteD
SiteE
SiteF
SiteG
Site H
AQS Site ID
25-025-0042
25-027-0023
33-011-0020
33-015-0014
44-003-0002
44-007-0022
44-007-0026
44-007-0027
Mean
16.4
21.6
16.5
14.8
10.7
17.0
21.3
18.8
Obs
191
174
182
175
171
182
169
168
SD
7.7
11.9
9.2
8.2
7.0
8.7
10.8
9.4
1=winter
2=spring
3=summer
4=fall
60 ¦
E
U)
c
o
c
o
20 •
1234 1234 1234 1234 1234 1234 1234 1234
Figure 3-30. Seasonal distribution of 24-h avg PM10 concentrations by site for Boston, MA,
2005-2007. Box plots show the median and interquartile range with whiskers extending
to the 5th and 95th percentiles at each site during (1) winter (December-February), (2)
spring (March-May), (3) summer (June-August) and (4) fall (September-November).
Table 3-14. Inter-sampler comparison statistics for each pair of 24-h PM10 monitors reporting to AQS
for Boston, MA.
Site A
B
C
D
E
F
G
H
A 1.00
0.69
0.69
0.73
0.71
0.84
0.70
0.79
(0.0,0.00)
(15.0,0.22)
(12.0,0.20)
(10.0,0.22)
(13.0,0.30)
(8.0,0.14)
(15.0,0.20)
(10.0,0.17)
191
169
179
173
171
182
169
167
B
1.00
0.66
0.56
0.45
0.69
0.77
0.65

(0.0,0.00)
(17.0,0.24)
(19.0,0.28)
(24.0,0.39)
(15.0,0.21)
(12.0,0.17)
(16.0,0.20)

174
167
161
158
169
156
154
C

1.00
0.72
0.47
0.62
0.64
0.59


(0.0,0.00)
(10.0,0.22)
(17.0,0.33)
(12.0,0.21)
(16.0,0.26)
(16.0,0.24)


182
170
168
179
166
164
D


1.00
0.63
0.68
0.59
0.69



(0.0,0.00)
(11.0,0.29)
(10.0,0.23)
(19.0,0.30)
(13.0,0.26)



175
163
173
161
158
E
KEY


1.00
0.84
0.58
0.80

Pearson R


(0.0,0.00)
(13.0,0.29)
(22.0,0.38)
(15.0,0.33)

(P90, COD)


171
171
161
157
F
n



1.00
0.81
0.95





(0.0,0.00)
(11.0,0.16)
(5.0,0.11)





182
169
167
G





1.00
0.79






(0.0,0.00)
(10.0,0.13)






169
154
H






1.00
(0.0,0.00)
168
July 2009
3-101
DRAFT-DO NOT CITE OR QUOTE

-------
60	80	100
i Kilometei
• Pittsburgh PM10 Monitors
	Pittsburgh Interstates
Pittsburgh Major Highways
Pittsburgh
Figure 3-31. Locations of PMio monitors and major highways, Pittsburgh, PA.
July 2009
3-102
DRAFT-DO NOT CITE OR QUOTE

-------

A
B
C
D
E
F
G
H
1
Mean
21.2
18.2
21.8
27.7
23.4
19.6
184
29.7
25.2
Obs
1077
1019
1083
1087
176
179
978
182
1022
SD
12.9
11.4
123
20.3
11.1
9.9
11.7
16.8
19.3
AQS Site ID
Site A	42-003-0002
Site B	42-003-0021
SiteC	42-003-0031
Site D	42-003-0064
Site E	42-003-0092
Site F	42-003-0095
SiteG	42-003-0116
SiteH	42-003-1301
E 60 ¦
Ol
50'
c
o
% 40'
c
v an-
1=winter
2=spring
3=summer
4=fall

2 3 4
J
1 2 3 4 1
K
2 3 4 1
L
2 3 4 12 3 4
M
12 3 4
N
12 3 4
0
12 3 4
P
12 3 4
Q
Mean
20.1
36.5
26.3
26.4
21.7
19.7
27.3
21.1
Obs
177
1061
1051
1069
1092
167
178
1079
SD
10.3
26.7
16.3
15.0
12.4
11.0
12.0
11.4
Site I
S ite J
SiteK
Site L
Site M
Site N
SiteO
Site P
SiteQ
AQS Site ID
42-003-3006
42-003-3007
42-003-7004
42-007-0014
42-073-0015
42-125-0005
42-125-5001
42-129-0007
42-129-0008
120
110
100
£T- 90
£
™ 80 '
Oi
3.
C 70
o
ro 60'
30 -
20 -
1=winter
2=sprir»g 1 0 -
3=summer
4=fall	0 -
1234 1234 1234 1234 1234 1234 1234 1234
Figure 3-32. Seasonal distribution of 24-h avg PM10 concentrations by site for Pittsburgh, PA, 2005-
2007. Box plots show the median and interquartile range with whiskers extending to
the 5th and 95th percentiles at each site during (1) winter (December-February), (2)
spring (March-May), (3) summer (June-August) and (4) fall (September-November).
Table 3-15. Inter-sampler comparison statistics for each pair of 24-h PM10 monitors reporting to AQS
for Pittsburgh, PA.
Site	A	§	C	D	I	F	G	H	i
July 2009
3-103
DRAFT-DO NOT CITE OR QUOTE

-------
Site
A
B
C
D
E
F
G
H
I
A
1.00
0.93
0.93
0.80
0.92
0.89
0.93
0.79
0.86

(0.0,0.00)
(9.0,0.15)
(8.0,0.14)
(23.0,0.21)
(8.0,0.12)
(14.0,0.18)
(8.0,0.14)
(16.0,0.17)
(18.0,0.18)
T
1077
1002
1065
1070
175
178
960
181
1005
B

1.00
0.96
0.80
0.91
0.92
0.97
0.81
0.89


(0.0,0.00)
(8.0,0.15)
(29.0,0.24)
(11.0,0.20)
(6.0,0.16)
(5.0,0.10)
(25.0,0.29)
(22.0,0.20)


1019
1007
1012
163
166
911
169
954
C


1.00
0.81
0.94
0.93
0.94
0.77
0.87



(0.0,0.00)
(23.0,0.20)
(6.0,0.11)
(7.0,0.12)
(8.0,0.13)
(21.0,0.22)
(19.0,0.17)



1083
1075
173
176
966
179
1010
D



1.00
0.72
0.66
0.76
0.83
0.88




(0.0,0.00)
(21.0,0.20)
(26.0,0.24)
(27.0,0.24)
(14.0,0.18)
(16.0,0.14)




1087
176
179
970
182
1014
E




1.00
0.90
0.90
0.78
0.77





(0.0,0.00)
(10.0,0.14)
(10.0,0.17)
(20.0,0.20)
(20.0,0.19)

KEY



176
173
154
175
166
F
Pearson R




1.00
0.94
0.70
0.74

(P90, COD)




(0.0,0.00)
(7.0,0.12)
(25.0,0.27)
(25.0,0.22)

n




179
157
178
168
G






1.00
0.70
0.87







(0.0,0.00)
(22.0,0.28)
(20.0,0.19)







978
160
910
H







1.00
0.76








(0.0,0.00)
(17.0,0.20)








182
171
I 1.00
(0.0,0.00)
1022


J
K

L
M
N
0
P
Q
A
0.84
0.76

0.88
0.85
0.86
0.77
0.78
0.86

(14.0,0.20)
(40.0,0.30)

(15.0,0.18)
(16.0,0.19)
(11.0,0.16)
(16.0,0.22)
(15.0,0.19)
(11.0,0.15)

176
1044

1033
1052
1074
166
177
1061
B
0.93
0.76

0.88
0.81
0.91
0.76
0.83
0.88

(7.0,0.16)
(43.0,0.36)

(19.0,0.23)
(20.0,0.26)
(10.0,0.16)
(12.0,0.19)
(18.0,0.28)
(10.0,0.18)

164
986

982
994
1016
157
165
1003
C
0.90
0.75

0.88
0.83
0.89
0.78
0.88
0.90

(8.0,0.13)
(39.0,0.30)

(14.0,0.17)
(15.0,0.19)
(9.0,0.12)
(12.0,0.18)
(13.0,0.19)
(9.0,0.12)

174
1049

1039
1057
1080
164
175
1067
D
0.73
0.84

0.80
0.78
0.76
0.57
0.64
0.74

(24.0,0.22)
(24.0,0.22)

(20.0,0.18)
(20.0,0.20)
(25.0,0.20)
(28.0,0.26)
(20.0,0.25)
(26.0,0.21)

177
1055

1043
1061
1084
167
178
1071
E
0.86
0.65

0.83
0.80
0.84
0.77
0.84
0.85

(10.0,0.16)
(36.0,0.29)

(16.0,0.16)
(14.0,0.17)
(12.0,0.14)
(14.0,0.19)
(13.0,0.16)
(11.0,0.15)

171
169

169
172
176
161
172
174
F
0.90
0.57

0.82
0.75
0.86
0.83
0.84
0.86

(7.0,0.12)
(41.0,0.34)

(20.0,0.20)
(19.0,0.22)
(11.0,0.14)
(9.0,0.15)
(16.0,0.22)
(9.0,0.14)

174
172

172
175
179
164
175
177
G
0.92
0.73

0.87
0.78
0.89
0.81
0.84
0.86

(7.0,0.13)
(45.0,0.35)

(18.0,0.21)
(19.0,0.24)
(9.0,0.15)
(11.0,0.17)
(17.0,0.26)
(10.0,0.16)

156
955

938
952
975
146
157
967
H
0.74
0.68

0.77
0.78
0.74
0.60
0.65
0.76

(23.0,0.26)
(26.0,0.22)

(15.0,0.18)
(17.0,0.18)
(21.0,0.22)
(27.0,0.29)
(19.0,0.22)
(21.5,0.24)

176
175

175
178
182
167
177
180
I
0.79
0.83

0.82
0.78
0.81
0.66
0.69
0.78

(22.0,0.20)
(30.0,0.25)

(16.0,0.17)
(18.0,0.20)
(20.0,0.17)
(26.0,0.24)
(21.0,0.25)
(22.0,0.19)

166
992

978
998
1019
158
167
1009
J
1.00
0.66

0.79
0.72
0.88
0.78
0.86
0.86

(0.0,0.00)
(44.5,0.33)

(18.0,0.20)
(18.0,0.22)
(8.0,0.13)
(11.0,0.17)
(16.0,0.21)
(8.0,0.15)

177
170

170
173
177
163
173
175
K

1.00

0.74
0.75
0.70
0.47
0.58
0.68


(0.0,0.00)

(31.0,0.26)
(33.0,0.24)
(40.0,0.30)
(44.0,0.36)
(34.0,0.30)
(43.0,0.30)


1061

1017
1035
1058
160
171
1048
L



1.00
0.87
0.85
0.70
0.74
0.80




(0.0,0.00)
(13.0,0.16)
(16.0,0.17)
(22.0,0.24)
(17.0,0.21)
(18.0,0.19)




1051
1025
1048
160
171
1035
M




1.00
0.74
0.64
0.67
0.77





(0.0,0.00)
(18.0,0.21)
(19.0,0.26)
(17.0,0.22)
(18.0,0.19)





1069
1067
163
174
1053
N





1.00
0.72
0.86
0.86






(0.0,0.00)
(13.0,0.18)
(14.0,0.20)
(10.0,0.14)






1092
167
178
1076
0






1.00
0.75
0.69







(0.0,0.00)
(18.0,0.25)
(14.0,0.19)







167
163
165
P







1.00
0.84








(0.0,0.00)
(15.0,0.21)








178
176
Q








1.00
(0.0,0.00)
1079
July 2009
3-104
DRAFT-DO NOT CITE OR QUOTE

-------
• Los Angeles
	 Los Angeles
60	80	100
i Kilometers
PM10 Monitors
Interstates
Los Angeles Major Highways
Los Angeles
Figure 3-33. Locations of PMio monitors and major highways, Los Angeles, CA.
July 2009
3-105
DRAFT-DO NOT CITE OR QUOTE

-------
AOS Site ID
Site A	0&037-0002
Stei	06-037-1103
SifeC	06-037-4002
SrteD	06-037-6012
SteE	06-037-9033
Site F	06-059-0007
Site 6	06-059-2022
Mean
Ote
SO
90-
80 "
70 ¦
E
60 ¦
O 50'
C 40-
w
w
c
o
u
30 ¦
1=winter
2 k spring
3=summer
4=fall
20 '
10'
A
35.3
169
19,8
B
31.1
175
13,3
C
31. S
178
19.6
D
27.3
176
18,1
E
23.7
985
12.1
F
33.5
175
37.6
G
21.6
162
9,4
1234 1234 1234 1234 1234 1234 1234
Figure 3-34. Seasonal distribution of 24-h avg PM10 concentrations by site for Los Angeles, CA, ZOOS-
ZOO?. Box plots show the median and interquartile range with whiskers extending to
the 5th and 95th percentiles at each site during (1) winter (December-February), (Z)
spring (March-May), (3) summer (June-August) and (4) fall (September-November).
Table 3-16. Inter-sampler comparison statistics for each pair of Z4-h PM10 monitors reporting to AQS
for Los Angeles, CA.
Site
A
B
C
D
E
F
G
A
1.00
0.73
0.44
0.73
0.47
0.41
0.65

(0.0,0.00)
(17.0,0.17)
(27.0,0.24)
(24.0,0.22)
(28.0,0.26)
(29.0,0.24)
(30.0,0.28)

169
153
154
157
169
155
143
B

1.00
0.61
0.57
0.52
0.42
0.73


(0.0,0.00)
(14.0,0.14)
(21.0,0.24)
(23.0,0.23)
(15.0,0.16)
(20.0,0.23)


175
159
159
173
162
149
C


1.00
0.65
0.43
0.93
0.73

KEY

(0.0,0.00)
(27.0,0.28)
(22.0,0.24)
(11.0,0.11)
(21.0,0.22)

Pearson R

178
158
176
159
148
D
(P90,tC0D)


1.00
0.70
0.65
0.57

n


(0.0,0.00)
(16.0,0.20)
(26.0,0.28)
(19.5,0.24)




176
175
161
150
E




1.00
0.29
0.38





(0.0,0.00)
(26.0,0.25)
(20.0,0.24)





985
173
159
F





1.00
0.65






(0.0,0.00)
(21.5,0.22)






175
150
G






1.00
(0.0,0.00)
162
July 2009
3-106
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Figures 3-38 through 3-42 illustrate the relationship between inter-sampler
correlation and distance between sites for PMio measurements obtained in Boston,
Pittsburgh and Los Angeles. Annex A contain similar plots for the remainder of the fifteen
CSAs/CBSAs under investigation beginning with Figure A-2. In each plot, substantially
more scatter is observed when compared to those for PM2.5 (Figure 3-25 through 3-30). This
is consistent with the variability observed in the seasonal box plots of concentration shown
in Figures 3-33, 3-35, and 3-37. The Boston data exhibit the strongest relationship between
inter-sampler correlation and distance, with average inter-sampler correlation remaining
higher than 80% when samplers are 44 km apart (R2 = 0.61). The lowest correlations on
this plot originate from comparisons between Site B (rural Worcester, MA) and samplers
located at Sites E (West Greenwich, RI) and G (Providence, RI). Boston is subject to long
range transport of SO42", which is a regional pollutant and is a major component of PM2.5
and PM10 in the eastern U.S. The Pittsburgh data shows some lower inter-sampler
correlations, with one sampler pair having only 66% correlation within a distance of 2 km.
On average, inter-sampler correlation remained higher than 80% when samplers were also
separated by 44 km, but in this case with much greater scatter (R2 = 0.28) than observed in
the Boston data. As seen for the Pittsburgh PM10 box plots in Figure 3-32, sites D, H, I, and
K have elevated means and high variability that is driving the observed scatter. These four
sites are all located in mountainous suburbs of Pittsburgh (North Braddock, PA, Liberty,
PA, Lincoln Boro, PA, and Beaver Falls, PA, respectively), where emissions from steel
manufacturing and frequent stable conditions in the planetary boundary layer cause
localized events of elevated concentration. When those four sites are removed, scatter
decreases greatly (R2 = 0.56). The Los Angeles data exhibit a much steeper slope, with
average inter-sampler correlation remaining higher than 80% when samplers are only 30
km apart (R2 = 0.56). The lower inter-sampler correlations in part reflect the fact that some
of these monitoring sites are separated from each other by hills or, in the case of one sited
at Lancaster, CA (Site I), by the San Gabriel Mountains. The Los Angeles data exhibit
greater scatter than the Pittsburgh data. However, the smallest inter-sampler separation
distance is 23 km, and there are relatively fewer PM10 samplers. It is not possible to judge
how data would correlate on smaller spatial scales.
July 2009
3-107
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 3-35.
40	50	60
Distance Between Samplers (km)
100
Inter-sampler correlations for 24-h PM10 as a function of distance between monitors in
Boston, MA.
~ •» ~ ~ ~ «,
~	44 ~	~	V	+	~
~	#	4 a O 4 / ~	~
% ~ ~ *,~~~~~*
	.	~	A	~ ~	4		
0.2
10
20
30
40
50
60
70
80
90
100
Distance Between Samplers (km)
Figure 3-36. Inter-sampler correlations for 24-h PM10 as a function of distance between monitors in
Pittsburgh, PA.
July 2009
3-108
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
1
0.8
0.6
C
o
JS
£
o
o
0.4
0.2
0
0	10	20	30	40	50	60	70	80	90	100
Distance Between Samplers (km)
Figure 3-37. Inter-sampler correlations for 24-h PM10 as a function of distance between monitors in
Los Angeles, CA.
Ultrafine Particles
Relatively few studies compare ultrafine measurements at multiple locations within
an urban center. An early study by Buzorius et al. (1999, 081205) suggested spatial
homogeneity in total particle number concentrations between multiple locations in
Helsinki, Finland. They found correlations in 10-min avg at three sites within the city as
high as 0.84. The sites, however, were relatively close together (2 km) and all near the same
roadway. There was a high degree of correlation between traffic intensity and total aerosol
number concentrations, suggesting that traffic was the primary source of the measured
particles and the driving force behind the high correlations. Weekend correlations
(0.28-0.47) and correlations with a fourth monitor located 22 km outside the city (0.05-0.64)
were much lower.
Tuch et al. (2006, 157060) found more spatial heterogeneity in ultrafine particle
concentrations measured for an entire year at two locations 1.5 km apart in Leipzig,
Germany. Figure 3"38 shows the correlation as a function of particle size (mobility
diameter) dropping off as the particle size decreases from 0.5 at 100 nm down to 0.2 at 3
nm. Table A-44 in Annex A contains correlation coefficients of hourly and daily average
July 2009
3-109
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
particle number, surface area and volume concentrations as a function of particle diameter
adapted from the Tuch et al. (2006, 157060) study. For all days (N = 5481 hourly
observations), the correlation between ultrafine particles (10-100 nm) measured at the two
sites was 0.31.
1 o
a 08
£
a>
Q
o
c o 6
0,4
Q
02
W
0.0

10
100
diameter [nm]
Source: Tuch et al. (2006,157060)
Figure 3-38. Bin-wise Spearman correlation coefficients in aerosol particle number concentrations
between the Ift (urban background) and the Eisenbahn-strasse (city/urban center) sites in
Leipzig, Germany.
The two sites represented in Figure 3-38 and Table A-44 were relatively close to each
other, but one was located in a mixed semi-industrial region while the other was in a street
canyon in a residential neighborhood near busy roadways. This suggests a high degree of
spatial heterogeneity in ultrafine particles driven primarily by differences in nearby source
characteristics. Sioutas et al. (2005, 088428) reviewed studies of the distribution of
ultrafine particles and came to the similar conclusion that mobile sources make a large
contribution to ultrafine particles and therefore ultrafine concentrations can exhibit
substantial variability in space and time. This is to be expected since ultrafine particle
concentrations drop off much quicker with distance from roadways than larger particle
sizes (Levy et al., 2003, 156688; Levy et al., 2003, 052661; Reponen et al., 2003, 088425;
Zhu et al., 2005, 157191). Hagler et al. (2009, 191185) showed similar exponential decreases
in ultrafine particle number concentrations with distance from the road for multiple
locations in the U.S. Neighborhood-scale variability and near-roadway concentration
gradients for ultrafine particles are discussed further in Section 3.1.1.3.
July 2009
3-110
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
PM Constituents
The pie charts showing PM2.5 composition that were generated using the SANDWICH
method for the fifteen CSAs/CBSAs presented earlier in Figures 3-17 and 3-18 represent
the average of all available monitors within each region. Individual pie charts for each
monitor are included in Figures A-107 through A-121 in Annex A and provide an indication
of the urban-scale spatial variability in PM2.5 composition. In the instances where multiple
monitors were available, there was a fair degree of spatial homogeneity in PM2.5 bulk
chemistry within each metropolitan area. Some notable exceptions exist, however.
Birmingham and Detroit show variation in the amount of crustal material, both spatially
and seasonally. Denver exhibits some spatial variation in NO:; during the winter, the
season with the highest measured PM2.5 mass. Several sites in New York and one in
Pittsburgh have elevated fractions of EC relative to the other sites within the respective
cities. In Phoenix, high winter PM2.5 mass is site specific and appears to be associated with
high organic carbon! the crustal component also varies and is inversely proportional to total
measured mass.
3.5.1.3. Neighborhood-Scale Variability
Neighborhood scale spatial variability in the particle concentration profile is affected
by land and building topography, meteorology, particle size distribution, particle
composition, and particle volatility. Population density at the neighborhood scale is also an
important determinant of the spatial distribution of PM concentration because density
impacts source prevalence, source magnitude, topographical-driven ventilation, and heat
island effects (Crist et al., 2008, 156372; Makar et al., 2006, 155959; Mfula et al., 2005,
123359; Rigby and Toumi, 2008, 156050). These considerations are crucial to understanding
variability in community monitoring data and for monitor deployment planning.
A number of computational and wind tunnel modeling street canyon studies have
demonstrated the potential variability in pollutant concentrations within a street canyon
(Borrego et al., 2006, 155697; Chang and Meroney, 2003, 090298; Chang and Tsai, 2003,
155718; Kastner-Klein and Plate, 1999, 001961; So et al., 2005, 110746; Xiaomin et al.,
2006, 156165). Influential parameters include street canyon height to width ratio (H/W),
source positioning, wind speed and direction, building shape and upstream configuration of
buildings. Figure 3"39 shows pollutant concentrations obtained from wind tunnel and
July 2009
3-111
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
computational fluid dynamics simulations of transport and dispersion in an infinitely long
street canyon with a line source centered at the bottom of the canyon (Xiaomin et al., 2006,
156165). When the canyon height was equal to the street width (typical of moderate density
suburban or urban fringe residential neighborhoods) and lower background wind speed
existed, concentrations on the leeward canyon wall were four times those of the windward
wall near ground level. When the canyon height was twice the street width (typical of
higher-density urban planning) and background winds were somewhat higher, near ground
level concentrations on the windward canyon wall were roughly three times higher than
those measured at the leeward wall. Baldauf et al. (2008, 191017; 2009, 191766) noted that
the presence of noise barriers, vegetation, or changes in topography adjacent to the road
can also alter particle dispersion characteristics. Specifically, depressed road sections,
where the road bed is below the surrounding terrain, leads to increased air turbulence and
pollutant dispersion. These results suggest that micro- and neighborhood-scale variation
related to urban topography may have a significant impact on pollutant concentrations at
this scale.
July 2009
3-112
DRAFT-DO NOT CITE OR QUOTE

-------
^ windward,
measured
	 windward,
simulated
leeward,
measured
leeward,
simulated
Dimensionless concentration
Dimensionless concentration
Source: Xiaomin et al. (2006,156165).
Figure 3-39. Dimensionless concentration as a function of height at windward and leeward locations
and street canyon aspect ratios (H/W). (a) Dimensionless concentration on the windward
and leeward sides of the canyon when H/W = 1 and wind speed = 3 m/s. (hi
Dimensionless concentration on the windward and leeward sides of the canyon when
H/W = 2 and wind speed = 5 m/s. Computational fluid dynamics modeling was
performed, and measurements were obtained in wind tunnel simulations.
PM2.5 and PMto
1	Knowledge of neighborhood-scale variability is important for interpreting data from
2	PM2.5 and PM10 community monitors. Figure 3-40 shows data derived from the fifteen
3	CSAs/CBSAs for PM2.5 and PMiq discussed in Section 3.5.1.2. This figure is limited to the
4	inter-sampler correlations obtained for sampler pairs located within a distance of 4 km (i.e.,
5	neighborhood scale). PM2.5 data exhibit a flatter slope, with average correlation maintained
6	at 93% within 4 km (R2 = 0.22). There is more scatter and variability among the PM10 data,
7	with an average correlation of 70% within 4 km (R2 = 0.03). The degree of variability in
8	PM10 compared with PM2.5 relates to transport and dispersion of the PM10-2.S component of
July 2009
3-113
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
PMio compared with PM2.5. However, differences in composition, source location,
topography, and monitor height—all of which could affect concentrations—could drive the
relatively high degree of scatter for both size classes, considering the low computed R2
values for each of these curves.
1.00
0.90
0.80
0.70
0.60
C
O
| 0.50
o
o
0.40
0.30
0.20
0.10
0.00
0	0.5	1	1.5	2	2.5	3	3.5	4
Distance Between Samplers (km)
Figure 3-40. Inter-sampler correlations for 24-h PM2.5 and PM10 as a function of distance between
monitors for samplers located within 4 km (neighborhood scale).
Isakov et al. (2007, 156588) compared PM2.5 concentrations from a central monitoring
site in Wilmington, DE with PM0.3 measurements taken on a mobile platform driven
through mostly quite residential streets within a 4 km x 4 km grid containing the central
monitor. Correlations were generally high (average r = 0.87) over all time periods and
locations monitored, consistent with the range of correlations for PM2.5 shown in
Figure 3-40.
PM102.5
Neighborhood-scale variability in PM10-2.5 was investigated by Chen et al. (2007,
147318) in the Raleigh/Durham area of NC. The average correlation between 26 residential
pm2 5 —0.0163D + 1
o R2 = 0.2178
OO
io = -0.0748D + 1
R2 = 0.0346
• PM2.5
O PM10
^—Linear (PM2.5)
- - Linear (PM10)
July 2009
3-114
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
monitors located throughout the region and a centrally located monitor representing a
maximum inter-sampler range of 60 km was found to be 0.75 for PM10-2.5 compared with
0.92 and 0.94 for PM2.5 and PM10, respectively. Based on this study, neighborhood-scale
variability is greater for PM10-2.5 than for PM2.5 or PM10, matching the conclusion drawn
above on the broader urban-scale.
Ultrafine Particles
Moore et al. (2009, 191002) monitored ultrafine PM concentrations throughout the
Ports of Los Angeles and Long Beach, through which an interstate highway runs and found
that concentrations varied by a factor of 5-7 across sites with substantial differences in the
daily concentration time series at each site. Such variability reflects diversity of the sources
(some near the Interstate, some near the Port), and the influence of changing meteorology
over an urban area. In a mobile platform sampling study, Westerdahl et al. (2005, 086502)
and Fruin et al. (2008, 097183) also reported substantial peaks in ultrafine PM
concentration when sampling at highways in comparison with a background site (the
University of Southern California) using the same data set.
Near roadway environments can exhibit high concentration gradients, particularly for
ultrafine particles. Ntziachristos et al. (2007, 089164) observed that the near-road particle
size distribution was substantially higher in the ultrafine mobility diameter range and that
these results were very sensitive to meteorology (rain) and time of day. Baldauf et al. (2008,
190239) reported elevated ultrafine particle number concentrations downwind of a highway
in Raleigh, North Carolina when compared to measurements approximately 100 m upwind
of the road. Hagler et al. (2009, 191185) noted a 5-12% decrease in number concentrations
per 10 m distance from the road for a number of studies in the U.S. with unobstructed air
flow.
After initial emission from a motor vehicle, the evolution of the PM distribution
within the plume is a function of (l) the turbulence that dilutes the plume and (2)
evaporation or condensation of the volatile portion of the aerosol that results from rapid
cooling of the exhaust. Figure 3-41 shows the size distribution measured by Zhu and Hinds
(2002, 011552) at distances of 17-300 m away from the roadway (in this case, highway 710
in Los Angeles) and at an upwind site. It can be seen that a mode originally measured
around 9 nm increases in diameter and decreases in magnitude as distance from the
highway increases. Smaller secondary modes appear around 30 m from the roadway with
July 2009
3-115
DRAFT-DO NOT CITE OR QUOTE

-------
1	multiple modes at some particle sizes. By 150 m away from the highway, the size
2	distribution flattens with a small mode around 50 nm. It is clear from the bottom
3	figure that the larger particles are better represented by one sampler than ultrafine
4	particles because the 100-220 nm number concentration is fairly constant with distance up
5	to at least 300 m whereas the <100 nm particle sizes measured show more variation with
6	distance, in both an absolute and relative sense.
17m
3.0e+5 -
E
S 2.0e+5
Cu
a
Q0
o
§
Z
T3
1.0e+5 -
20 m
150 m
Up Wind^^/ ^
300 m
10	100
Particle Diameter, Dp (nm)
1000
a
0
1
£3
£ 0.6
e
0.4
25-50 nm
T3
n 0.2
1
o
2
0.0 -
6-25 nm
50-100 nm
100-220 nm
100	200	300
Distance down wind from the 710 freeway (m)
400
Source: Zhu and Hinds (2002,041552).
Figure 3-41. Particle size distributions measured at various distances from the 710 freeway in Los
Angeles, CA (top), and particle number concentration as a function of distance from the
710 freeway (bottom).
July 2009
3-116
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Zhou and Levy (2007, 098633) performed a meta-analysis of traffic-related air
pollution literature and found that background pollution and meteorology can have
important impacts on the size of the elevated concentration region around the highway. Zhu
and Hinds (2002, 041552) and Zhang et al. (2005, 157185) noted in field measurements of
ultrafine PM that small particles can be lost due to evaporation or to coagulation during
Brownian diffusion to form bigger particles, resulting in an upward shift in mode diameter
with distance from the roadway Studies of particle sizes on roads (Kittelson et al., 2006,
156649; Kittelson et al., 2006, 156648), in tunnels (Venkataraman et al., 1994, 002475), and
upwind and downwind of roads (Wilson and Suh, 1997, 077408; Zhu and Hinds, 2002,
041552) suggest that for well-maintained spark-ignition vehicles, a large fraction of the
mass of particles emitted from the vehicles are in the nuclei mode (i.e., smaller than
accumulation mode). High-speed highway driving may be associated with a larger fraction
of particle mass being emitted in the ultrafine size range, while lower speed operation
results in a higher mass fraction in the accumulation mode (Cadle et al., 2001, 017192).
Traffic may also generate some coarse mode particles (Wilson and Suh, 1997, 077408)
(Wilson and Suh, 1997, 077408) from material resuspended from the road and brake and
tire wear. In situations in which the dilution rates are lower than in a short tunnel or
downwind of a road way, condensation of vapors can give rise to particles in the
accumulation mode (Kittelson, 1998, 051098; Wilson and Suh, 1997, 077408). Diesel
engines, in particular, emit elemental (black) carbon in the lower end of the accumulation
mode, with number emissions dominated by semi-volatile material in the nuclei mode
(Kittelson, 1998, 051098; Kittelson et al., 2006, 156649; Kittelson et al., 2006, 156648).
Sharp gradients in black carbon mass have been observed along roadways with high diesel
traffic (Zhu and Hinds, 2002, 011552). As the traffic pollution moves downwind, the
ultrafine particles may grow into the accumulation mode by coagulation or condensation. In
addition to Gaussian dispersion and wind eddies caused by the presence of natural and
anthropogenic barriers, Sahlodin et al. (2007, 114058) demonstrated that turbulence
produced by vehicles can result in modification of the plume emanating from the highway.
Hence, on-road turbulence could potentially alter the aerosol size distribution. This added
turbulence could cause some evaporation of tiny nucleation particles that have not absorbed
or adsorbed onto soot nuclei, which may affect the rate of coagulation (Jacobson et al., 2005,
191187). The roadway configuration may also affect particle transport and dispersion.
Depressed road sections, where the road bed is below the surrounding terrain, leads to
July 2009
3-117
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
increased air turbulence and mixing as air flows up and out of the road depression. This
configuration can result in lower particulate concentrations and flatter concentration decay
curves away from the road. On the other hand, configurations with the road bed at-grade
with surrounding terrain, or elevated above the surrounding terrain with solid fill material
resulted in the highest pollutant concentrations and sharpest concentration gradients
downwind from the road.
PM Constituents
The composition of PM will also vary on the neighborhood-scale in response to local
sources and differential dispersion, resulting in variable spatial distribution of individual
components. Krudysz et al. (2008, 190064) investigated spatial variation in size-
fractionated (<0.25 |um, 0.25-2.5 |u,m, >2.5 |um) PM composition data at four sites located
within 3-6 km of each other in the Long Beach, CA area. Inter-site R2 values in the 0.25-2.5
jum size range were higher for mass (ranging from 0.56-0.91) than for EC (0.02-0.71) for pair
wise site comparisons. Spatial heterogeneity in all size ranges investigated was also found
for several elements associated with motor vehicle emissions and resuspended road dust
including Cu, Mg, Ba, Ca and Al. Viana et al. (2008, 156135) observed higher
concentrations of crustal elements in PM2.5 and PM10 samples in rural neighborhoods and
higher concentrations of combustion-derived PM2.5 and PM10, such as EC and NO:; , in
higher density urban areas. Gutierrez-Daban et al. (2005, 155818) examined the mass
distribution of various PAHs under different traffic and urban density conditions.
Figure 3-42 displays the distributions for benz[a]pyrene (BaP) at high and low traffic sites
at the urban center, periphery, and industrial areas found in Gutierrez-Daban et al. (2005,
155818). It can be seen that concentrations were nearly an order of magnitude lower for the
low traffic urban periphery location when compared with the high traffic or industrial
locations. Particles smaller than ~ 600 nm had roughly an order of magnitude higher
concentration than those at larger sizes and tended to have a larger spread in
concentrations among sampling sites. Figure 3-43 shows the distributions for sixteen PAHs
at a high traffic location at the city center from Gutierrez-Daban et al. (2005, 155818). PAH
species varied in concentration by up to two orders of magnitude for each particle size bin,
and the highest concentrations of individual PAHs were generally found for particles
smaller than approximately 600 nm. Olson and McDow (2009, 191188) reported decreases
by a factor of 1.04 to 2.37 in select PAH and organic source marker concentrations when
July 2009
3-118
DRAFT-DO NOT CITE OR QUOTE

-------
1	comparing measurements 10 m and 275 m from a highway in Raleigh, North Carolina.
2	Phuleria et al. (2006, 156867) sampled ultrafine PM and PM2.5 concentrations and PAH
3	species at the mouth of the Caldecott Tunnel in Orinda, CA and found that the two size
4	classes were highly correlated (R2 = 0.97). Given the size differentials of each size bin
5	presented in the Gutierrez-Daban et al. (2005, 155818) study, it is possible that the PM2.5
6	sampled at the tunnel mouth represented secondary PM2.5 that grew from ultrafine PM
7	emissions trapped within the tunnel.

X
~
¦
~
X
~
X
¦
A
X
~
X
¦
~
X
~
X
¦
~
X ~
¦
X I




* 1

~ HTC

¦ HTP

A LTC

X LTP

X LTIP
<0.6	1.3-0.6	2.7-1.3	4.9-2.7
Size bin (^m)
Source: Gutierrez-Daban et al. (2005,155818).
Figure 3-42. Mass distributions for BaP at a high traffic urban center (HTC), high traffic urban
periphery (HTP), low traffic urban center (LTC), low traffic urban periphery (LTP), and
low traffic industrial urban periphery (LTIP) in Seville, Spain.
July 2009
3-119
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
~	Naph
¦ Ace
A Acey
X Flu
X Phen
•	Ant
+ Flua
-Pyr
-BaA
~	Chry
~ BbF
ABkF
x BaP
XlnP
ODbA
+ Bper
1.3-0.6	2.7-1.3	4.9-2.7
Size bin (^.m)
Source: Gutierrez-Daban et al. (2005,155818).
Figure 3-43. Mass distributions for 16 PAHs at a high traffic city center in Seville, Spain.
3.5.2. Temporal Variability
Temporal variability is another important factor in characterizing PM. This
section addresses trends as well as seasonal and hourly variability. Trends in PMio and
PM2.5 are addressed in Section 3.5.2.1 based on AQS data. Seasonality is coupled with
spatial variability and has been discussed in the regional context above. Section 3.5.2.2
below briefly investigates the seasonality on a finer time scale, thereby addressing issues
relating to the seasonal definitions used earlier. Section 3.5.2.3 addresses hourly patterns,
an issue particularly important to understanding the behavior of PM concentrations in
reference to sources, human activity patterns and exposure. Hourly patterns are
investigated using AQS data on a national basis for PM2.5 and PM10. Data for ultrafine
particles and PM constituents are presented where available.
3.5.2.1. Regional Trends
This section summarizes available information on trends in PM mass and
composition. Mass concentration trends are based on AQS data and incorporate nine years
(1999-2007) of PM2.5 data and 20 years (1988-2007) of PM10 data. Composition trends are
July 2009
3-120
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
based on six years of available CSN data (2002-2007). Several monitoring sites were
excluded from the following trend analyses to provide a consistent basis for comparison over
the desired years of monitoring. This included exclusion of sites when there was no
corresponding site in later or earlier years. Region-average trends were calculated to
facilitate presentation and extrapolation of the results. These region-averages, however,
may not necessarily represent the trends that are being observed at any individual monitor
or geographical location within the specified region.
PM2.5
Figure 3-44 shows trends in U.S. ambient 24-h PM2.5 concentrations from 1999 to
2007. In the period 2005-2007, the three-yr avg of the 98th percentile of 24-h PM2.5
concentrations fell 10% relative to the 1999-2001 period (see Figure 3-44a). The number of
sites reporting values greater than the 24-h NAAQS was shown to decline 40% in
Figure 3"44b. Figure 3"44c illustrates the downward trend in the 98th percentile of 24-h
PM2.5 concentrations for three consecutive calendar years in all U.S. EPA regions. This
trend is most pronounced in Region 9 incorporating Arizona, California and Nevada where
this value dropped 25% from the 1999-2001 period to the 2005-2007 period.
July 2009
3-121
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
60
j= 45
A) Ambient Concentrations
i_ 03 on
'¦go™
15
0
90% of sites have
NAAQS = 35 pg/m3
/
concentrations below this line
/ Median
Y

A

Average
10% of sites have

concentrations below this line
1	1	I	
¦ 1
:sr-60
45
gl 30
C) Trends by EPA Region3
15
"99-01 "00-'02
'01-'03 02- 04 '03-'05
Averaging period
'04-'O6 05-07
NAAQS = 35 pg/m3

—R1

— R2

— R3

-R4

—R5

R6

R7

R8

— R9

— R10

—Natl
'99-01 '00-'02 01 '03 '02-'04 '03-'05 '04-'06 '05-'07
Averaging period
400
B) Number of Trend Sites Above NAAQS
Co r;
° £300
f o
•3
; 200
5 ® 'S
1 o s 100
'Coverage: 697 monitoring sites
in the EPA Regions (out of a total
of 831 sites measuring PM2.5 in
2007) that have sufficient data to
assess PM25 trends since 1999.
EPA Regions
©
©
JH
G
00

'99-01 '00-'02 '01-'03 '02-04 '03-05 '04-'06 '05-'07
Averaging period
Source: U.S. EPA (2008, 157076)
Figure 3-44. Ambient 24-h PM2.5 concentrations in the U.S., 1999-2007, showing A) ambient
concentrations, B) number of trends sites above the 24-h NAAQS and C) trends by U.S.
EPA Region.
Figure 3"45 contains similar trend information for the annual PM2.5 NAAQS. The
seasonally weighted 3-yr avg PM2.5 concentrations for the years 2005 to 2007 were at the
lowest since national monitoring began in 1999 (see Figure 3"45a). The seasonally weighted
3-yr avg fell 10% between the 1999-2001 averaging period and the 2005-2007 averaging
period. The number of sites reporting concentrations above the annual average PM2.5
NAAQS fell 56% over these same periods in Figure 3"45b. Figure 3-45c illustrates the
annual trends in PM2.5 by U.S. EPA region. Declines were the greatest in Region 9 again
where PM2.5 concentrations fell 20% from the 1999-2001 averaging period to the 2005-2007
averaging period.
July 2009
3-122
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
A) Ambient Concentrations
90% of sites have concentrations below this line
NAAQS = 15 p
Median
Y
verage
K. 10
10% of sites have concentrations below this line
20
C) Trends by EPA Region3
 
CO
CO E
^ O O)
15
^ r s 10
?
ID r ffl
5 o "o
CO	r
sis5
o
'99-'01 '00-'02 '01-'03 '02-'Q4 '03-05
Averaging period
'04-'06 '05-'07
NAAQS = 15 MO/rn3
0
'99-'01 '00-02 '01-'03 '02-'04 03-05 *04-'06 '05-'07
Averaging period

— R1

— R2

— R3

— R4

— R5

-R6

R7

R8

— R9

—R10

—Nat'l
¦ 300
CO W
CO
C "O
c cz i_
CO CI) c
¦o CO V
CD o o)
gj.i 5 250
as 00
75 g § 200
isi
& S>
S a s 150
¦§ & £
s = S
g .2 ® 100
+=: ro 
-------
1
2
3
4
5
6
7
8
9
0
1
A) Ambient Concentrations
NAAQS = 150 pg/m3
o of sites have concentrations below this line
veiage
*
Median
" o 50
C) Trends by EPA Region
10% of sites have
concentrations below this line
"I	1	1	1	1	1	1—
'90 '92 '94 '96 '98 '00 '02 '04 '06
Year
B) Number of Trend Sites Above NAAQS
NAAQS = 150 pg/m
S " 40
—	R1
R2
—R3
-M
—	R5
—	R6
R7
R8
—	R9
—	R10
—Nat'l
'90 '92 '94 '96 "98 '00 '02 '04 '06
Year
'Coverage: 2/4 monitoring sites
in the EPA Regions (out of a total
of 879 sites measuring PMm in
2007) that have sufficient data to
assess PM10 trends since 1988.
EPA Regions
0
fell
©0
0

0
'90 '92 '94 '96 '98 '00 '02 '04 '06
Year
Figure 346.
Source: U.S. EPA (2008,157076).
Ambient 24-h PM10 concentrations in the U.S., 1988-2007, showing A) ambient
concentrations, B) number of trends sites above the 24-h NAAQS and C) trends by U.S
EPA Region.
PM Constituents
The SANDWICH method discussed in Section 3.5.1.1 for estimating PM2.6 composition
from FRM mass measurements and CSN bulk composition measurements was used to
evaluate trends in PM2.5 constituents. Figure 3-47 includes stacked bar charts of PM2.5
composition from 2002 to 2007 stratified by region and season. The regions used in
Figure 3-47 were selected based on common aerosol characteristics including trends,
seasonality, size distributions and/or composition as described in chapter 6 of the 1996
AQCD and differ from the EPA regions used in the preceding figures. Figure 3"4 7 is based
on 42 monitoring locations reporting complete CSN data with 2002 being the first year with
sufficient speciation data. The Southwest region incorporating Arizona, New Mexico and
parts of Texas and Oklahoma did not contain any complete data and therefore is not
represented in this analysis. Two seasons representing different temperature ranges—cool
July 2009
3-124
DRAFT-DO NOT CITE OR QUOTE

-------
(October-April) and warm (May-September)—were considered in the figure since many
PM2.5 components exhibit strong temperature dependence.
Cool
Warm
Cool
Warm
Northwest
20
£16
I12
 8
03
2 4
0





20-
¦
—




	
pzi
=

—
16-
12-

m
—


m
¦
-
.
u
¦
8-
4 -
—
—

—

¦
U"TJ



*-r
0-
1_rl

	
	
1-r'
Lrl
02 03 04 05 06 07
02 03 04 05 06 07
02 03 04 05 06 07
02 03 04 05 06 07
~ Sulfate ~ Nitrate
I I Organic Carbon
Elemental Carbon
¦ Crustal
Northwest
North Central
Midwest
Southern •
California Southwest
* • Northeast
Southeast
Source: U.S. EPA (2008,191190).
Figure 3-47. Regional and seasonal trends in annual pm2.. compostion from 2002 to 2007 derived using
the SANDWICH method. Data from the 42 monitoring locations shown on the map were
stratified by region and season including cool months (October-April) and warm months
(May-September). SO*2" and N(h estimates include NH4* and particle bound water.
July 2009
3-125
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Most of the components showed little discernable trend over the 6-yr period. SO r
showed a peak during the warm months of 2005 in the Southeast, Northeast and Midwest,
partly due to atypical weather conditions (U.S. EPA, 2008, 191190). However, no trend over
the 6-yr time period is present for SC>42~ in any of the regions or seasons. The same is true
for EC and crustal material. A slight decline in OC was observed for the Northeast during
warm months and in Southern California year-round. The largest decline was for NO3" in
Southern California during both cool and warm months. A smaller decline in NO3" is also
observed in the other regions with the exception of the Northwest where no discernible
trend is present. This analysis is limited in time and space by the availability of CSN data
so a high degree of uncertainty remains regarding PM2.5 compositional trends. However,
with the exception of NO3" concentrations in Southern California, no major changes in PM2.5
composition are evident based on available CSN data from 2002-2007. This is consistent
with Figures 3-44 through 3-47 where the downward trend in PM2.5 mass begins to level off
after 2002.
3.5.2.2. Seasonal Variations
Many of the figures and tables presented in the preceding sections have included a
seasonal break-down based on the following climatological seasons: winter
(December-February), spring (March-May), summer (June-August) and fall
(September-November). Figures A-122 through A-136 in Annex A show bar charts of PM2.5
composition by individual month, illustrating intra-annual variability on a finer time scale.
The same 15 CSAs/CBSAs are investigated and included in these plots! they are generated
from the same data used in the seasonal and annual pie charts based on the SANDWICH
method discussed in Section 3.5.1.1 and illustrated in Figures 3-17 and 3-18.
Monthly plots for most of the areas reveal heterogeneity in PM composition within the
3-month long seasonal bins defined earlier. This is especially true in the spring and fall
when daily average weather conditions (e.g., temperature) are changing most rapidly,
driving fluctuation in PM2.5 composition on relatively short timescales in many cities. For
example, the NCV mass in Los Angeles (Figure A-129) and Riverside (Figure A-134) can
vary from a small fraction to the most prevalent fraction of PM2.5 mass in a month's time
based on the 3-yr aggregate data. Therefore, selecting a different delineation point for the
July 2009
3-126
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
seasons can have an influence on the seasonal composition analysis, specifically for
constituents that fluctuate rapidly (e.g., NO3").
Relatively little is known about the seasonal variability in ultrafine particles. Kuhn et
al. (2005, 129448) and Zhu et al. (2004, 156184) found that the concentrations in the
ultrafine mode in Los Angeles, CAcan be much higher during winter, particularly during
evenings, because atmospheric dilution is reduced in response to lower mixing heights. This
can be seen in Figure 3-48. Jeong et al. (2004, 180350) made similar observations in
Rochester, NY, suggesting an inverse relationship between temperature and ultrafine
particle formation in the 11-470 nm size range. Singh et al. (2006, 190136) reported higher
particle number concentrations during winter months, relative to summer and spring, at
urban sites in Southern California, and that afternoon particle number concentrations in
warm months either occurred during a peak in ozone concentrations or followed shortly
thereafter, suggesting a role for photochemistry in addition to meteorological changes in the
formation of aerosols. The study also reported increased concentrations of 60-200 nm
particles during a labor strike at the Port of Long Beach, suggesting contributions from
idling ships. Moore et al. (2009, 191001) also report higher particle number concentrations
during cooler months at 14 sites in Long Beach, San Pedro, and Wilmington, CA, a location
with diverse industrial and transportation sources. However, they noted substantial
heterogeneity in seasonal trends between sites with seasonal numerical size distributions
not generalizable across the study area with a maximum monitor separation of under 10
km.
July 2009
3-127
DRAFT-DO NOT CITE OR QUOTE

-------
I
13
»
81
200000
180000
'60000
'40300
'20000
OOOO'O
bOOOO
60DC0
-iOOCO
20000
0
25000
20000
15000
10000
5000
0
Z te A 'A si ter Eve i.
Site - n*er Dc /
See - Curr

10
20 30 50
100
Site E Airier Ever
"ite S Wrier Day
5 re B SuTTier
\
\
y
10	20 30 50
Particle diameter dp [nm]
100
Source: Kuhr et al. (2005, 129448).
Figure 3-48. Ultrafine particle size distribution at highway (site A) and background (site B) sites in
Los Angeles, CA, during summer and winter seasons, with winter broken into day and
evening distributions.
1	Studies reporting higher cold-season particle number concentrations are consistent
2	with vehicle emission studies that found particle emission rates elevated during lower
3	ambient temperatures (Baldauf et al., 2005, 191184; Mathis et al., 2005; 155970; U.S. EPA,
4	2008, 191767). Mathis et al. (2005, 155970) found that cold-start conditions produce roughly
5	an order of magnitude greater PM number emissions in gasoline engines and more than
6	two orders of magnitude higher PM number emissions in diesel engines when compared
7	with warm start conditions.
3.5.2.3. Hourly Variability
Hourly PMio and PM2.5 measurements are conducted at many sites using beta gauge
or TEOM monitors. Many of the hourly measurements for PM10 have FRM or FEM status.
July 2009
3-128
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
All available hourly data from FRM, FEM and FRM-like monitors in the fifteen
CSAs/CBSAs discussed earlier were used to investigate diel variation in PM. Of the fifteen
CSAs/CBSAs, Atlanta, Chicago, Pittsburgh, Seattle and St. Louis had qualifying hourly
PM2.5 and PM10 data available. Houston and New York had only qualifying PM2.5 data.
Denver, Detroit, Los Angeles, Philadelphia, Phoenix, and Riverside had only qualifying
hourly PM10 data. Birmingham and Boston had no qualifying hourly PM2.5 or PM10 data.
Diel plots for PM2.5 stratified by weekdays and weekends for seven of the fifteen
CSAs/CBSAs with available data between 2005 and 2007 are included in Annex A, Figures
A-137 through A-143. In most cities investigated, a morning PM2.5 peak is present starting
at approximately 6^00 a.m., corresponding with the start of the morning rush hour just
before the break-up of overnight stagnation. In Pittsburgh, dispersion behavior during the
night results in elevated PM2.5 concentrations throughout the night that blend in with any
morning peak. With the exception of Pittsburgh, all seven metropolitan areas show two
distinct daily peaks on both the weekdays and weekends. The evening PM2.5 concentration
peak is broader than the morning peak and extends to overnight hours, reflecting the
concentration increase caused by a drop in boundary layer height at night. Figure 3-49
compares the two-peak diel distribution in PM2.5 for Seattle with the one-peak distribution
in PM2.5 for Pittsburgh Since these figures represent the distribution of hourly observations
over a 3-yr period, any fluctuations or changes in the timing of the daily peaks would result
in a broadening of the curves shown in the diel plot.
July 2009
3-129
DRAFT-DO NOT CITE OR QUOTE

-------
Pittsburgh
77
« 58
CO
=L
38
19
Weekday (N = 981)
12 18 24
77
58
38
19
Weekend (N = 407)
12 18 24
¦ Median
	Mean
	90th & 10th
	95th & 5th
Seattle
CO
3.
77
58
38
19
Weekday (N = 5775)
12
18
24
77
58
38
19
Weekend (N = 2332)
12
18
24
¦	Median
¦	Mean
	90th & 10th
	95th & 5th
Figure 3-49. Diel plot generated from hourly FRM-like PM2.5 data (^ig/m3) stratified by weekday (left)
and weekend (right) for Pittsburgh, PA, and Seattle, WA, from 2005 to 2007. Included
are the number of monitor days (N) and the median, mean, 5th, 10th, 90th and 95th
percentiles for each hour of the day shown on the horizontal axis.
1	Annex A, Figures A-144 through A-154 show diel patterns for PM10 stratified by
2	weekdays and weekends for eleven of the fifteen CSAs/CBSAs with available data between
3	2005 and 2007. All cities show a gradual morning increase in mean PM10 starting at
4	approximately 6^00 a.m. on weekdays, corresponding with the start of the morning rush
5	hour before the break-up of overnight stagnation. The magnitude and duration of this peak,
6	however, varies considerably by area. Phoenix shows the most pronounced morning PM10
July 2009
3-130
DRAFT-DO NOT CITE OR QUOTE

-------
1	peak concentration, which drops off during the day and reappears in the evening. In
2	contrast, Chicago shows a less pronounced peak with the PMio concentration remaining
3	elevated throughout the day. Figure .'>-50 shows the diel plots of PMio for Chicago and
4	Phoenix. In both instances, the weekend diel pattern is similar in shape to the weekday
5	pattern with less pronounced peaks. Once again, any fluctuations in the timing of the daily
6	peaks could result in a broadening of the peaks in the 3-yr composite diel figures.
Chicago
291 i

OS
145
Si 73 i
Weekday (N = 1971)
12
18
291
218
145
73 -
24
Weekend (N = 793)
12
18
¦	Median
¦	Mean
	90th & 10th
	95th & 5th
24
Phoenix
291 i
« 218 -|
E
O)
=*• 145
Weekday (N = 1532)
73 -
12 18 24
Weekend (N = 618)
291
218 -
145
73 -
¦ Median
	Mean
¦ - - 90th & 10th
	95th & 5th
12 18 24
July 2009
3-131
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Figure 3-50. Diel plots generated from hourly FEM PM10 data (|ig/m3) stratified by weekday (left) and
weekend (right) for Chicago, IL, and Phoenix, AZ, from 2005 to 2007. Included are the
number of monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles
for each hour of the day shown on the horizontal axis.
Ultrafine particles in urban environments have been shown to exhibit a similar
two-peaked diel pattern in Los Angeles (Moore et al., 2007, 122445; Sardar et al., 2005,
180086) and the San Joaquin Valley (Herner et al., 2005, 135983) in CA, Rochester, NY
(Jeong et al., 2004, 180350), Raleigh, NC (Baldauf et al., 2008, 190239) as well as in
Kawasaki City, Japan (Hasegawa et al., 2005, 157355) and Copenhagen, Denmark (Ketzel
et al., 2003, 131251). Figure 3-51 from the Denmark study shows a large peak in total
particle number (dominated by ultrafine particles) corresponding with the morning rush
hour. The morning peak is absent on Sundays, however. Many studies also show a broad
afternoon ultrafine concentration peak, which likely originates from a combination of
evening rush-hour traffic, decreased atmospheric dilution and formation of ultrafine
particles through nucleation involving products of active photochemistry. Nucleation likely
plays an important role since the afternoon peak is present on weekends whereas the
morning traffic related peak is absent. This is consistent with observations of particle
counts in Atlanta peaking during the mid-afternoon for particles less than 10 nm (Woo et
al., 2001, 011702) resulting from nucleation.
July 2009
3-132
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
Weekdays
Sundays
x
40000
35000
3DDOO
;50oo
20DEM
5000
. 0000
sow
I
A Jagtv
J. /. V
—#—RC0
1
J
J j
1 i 1 m
s




1
18
24
3DOOO
5000
20000

>
Q
Jaastv
HC0

-HC0
Source: Ketzel et al. (2003,131251)
Figure 3-51. Average diel variation in total particle number (ToN) and total particle volume (ToV) on
weekdays (left column) and Sundays (right column) from two sites in Denmark: one in a
busy street canyon (Jagtv) and one measuring urban background (HC0).
Hourly variability in particle-phase OC and EC were investigated by Bae et al. (2004,
156243) in the urban St. Louis atmosphere. OC diel patterns were similar during weekdays
and weekends with a broad morning and evening concentration peak most likely reflecting
daily fluctuations in atmospheric mixing height. Weekend EC diel patterns were similar to
those for OC, but the weekday patterns showed more abrupt EC concentration peaks in the
morning and afternoon, coinciding with rush-hour traffic. The divergent weekday patterns
between OC and EC suggests motor vehicles or other EC sources with temporal profiles
tracking traffic patterns are primarily responsible for the daily fluctuations in EC
concentrations in St. Louis.
3.5.3. Statistical Associations with Copollutants
Associations between PM and other copollutants including SO2, NO2, CO and O3 are
investigated in this section. AQS data were obtained from all available co-located monitors
across the U.S. after application of an 11 or more observations per quarter completeness
July 2009
3-133
DRAFT-DO NOT CITE OR QUOTE

-------
criterion. Pearson correlation coefficients (R) were calculated using 2005-2007 data. The
results are displayed graphically in Figure 3"52 for correlations with PM2.5 and Figure 3"53
for correlations with PM10.
Winter
Spring
PM10 (daily avg)
PM10-2.5 (daily avg) -
S02 (daily avg)
N02 (daily avg) -
CO (daily avg)
03 (daily max 8-h) -
J_i	i	i_l	i_l	1	L-
t—«—1—«—1—«—1—>—r~
	
¦•4"
•f-
ll| |—^

»|
-1—"—1—"—1—"—1—"—r
	
1—«—1—«—1—«—1—«—r~

-Qlhf
~i—1—1—1—1—1—1—¦—r
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 -1.0 -0.8-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Summer
Fall
PM10 (daily avg) -
pmio-2.5 (dai|y av9)
S02 (daily avg) -
N02 (daily avg)
CO (daily avg) -
03 (daily max 8-h)
	
t—«—i—«—i—«—i—'—I-
	
•|f
'•••"I	l*l hi—
-i—'—i—'—i—'—i—'—r
	
-h
T	'	1—'	1—'	1	'—I-

»l |~4»
-TW
*
n—'—i—'—i—'—i—'—r
-1.0 -0.8 -0.6 -0.4-0.2 0.0 0.2 0.4 0.6 0.8 1.0 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
correlation (r) with PM25 (daily avg)	correlation (r) with PM25 (daily avg)
Figure 3-52. Distribution of correlations between 24-h avg PM2.5 and co-located 24-h avg PMio,
PM102.5, SO2, NO2 and CO and daily max 8-h avg O3 for the U.S. stratified by season
(2005-2007). Statistics shown include the mean (green star), median (red line), inner
quartile range (box), 5th/95th percentiles (whiskers) and outliers (black circles).
July 2009
3-134
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Winter
Spring
PM25 (daily avg)
PMio.2.5 (daily avg) -
S02 (daily avg)
N02 (daily avg) -
CO (daily avg) -
03 (daily max 8-h) -
J_i	i	i_l	i_l	I	L
~i—«—i—«—i—«—i—>—r~
	
1
»•••«•!	1 #| |—4*
-
• .-J—
. ^—r*n-4
-
•
1 4 \ + • •
• «-|	1 -*¦ | |—^
-
wm



•»
-
H—1 1*
	
	
•I—rr~H»
• 'H~n
T	"	1	"	1	"	1	'	1-
-1.0 -0.8 -0.6 -0.4-0.2 0.0 0.2 0.4 0.6 0.8 1.0 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
Summer
Fall
PM25 (daily avg) -
PM10-2.5 (dai|y aV9)
S02 (daily avg) -
N02 (daily avg)
CO (daily avg) -
03 (daily max 8-h)
j	I	I	I	I	I	I	I	L
t—«—i—«—i—«—i—'—r
	

• "H	L*L_Ff

. . 1—1 *1 hf
• +-
H > 1—•



1 * 1 t*
-1-
-1 [jf 1	^ ••
	
—1 *1 1—
~i—1—i—1—i—1—i—1—r
	
t—«—i—«—i—«—i—«—r
	

M* «m4	1 #1

•• |	l - l K
-4
—I 4 |	
• «
1 »i 1 f"
1—1 4 \-b'
••—1	
\ 4 1—
t—'—i—'—i—'—i—'—r
-1.0-0.8 -0.6 -0.4-0.2 0.0 0.2 0.4 0.6 0.8 1.0 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
correlation (r) with PM10 (daily avg)
correlation (r) with PM10 (daily avg)
Figure 3-53. Distribution of correlations between 24-h avg PM10 and co-located 24-h avg PM2.5,
PM102.5, SO2, NO2 and CO and daily max 8-h avg O3 for the U.S. stratified by season
(2005-2007). Statistics shown include the mean (green star), median (red line), inner
quartile range (box), 5th/95th percentiles (whiskers) and outliers (black circles).
For both PM2.5 and PM10 national composite copollutant correlations, there is
considerable spread in the observed correlations in all four seasons. On average, PM2.5 and
PM10 correlate with each other more than with the gaseous copollutants. The correlations
between PM2.5 and PM10 are all positive but span the range from just above zero to near
one. This illustrates the wide variability in correlation between these two PM metrics.
Fewer points are available for correlation with PM10-2.5 because only data from lowvolume
FRM/FRM-like samplers were used to calculate PM10-2.5. The available data suggest a
stronger correlation between PM10 and PM10-2.5 than between PM2.5 and PM10-2.5 on a
national basis.
The correlation between PM and the gaseous pollutants included in Figures 3.5-47
and 3.5-48 also have a large range in values based on the national composite data. There is
little seasonal variability in the mean correlation between PM and SO2. NO2 and CO,
however, show higher correlations with PM on average in the wintertime than in the other
seasons. This is possibly driven by meteorology with increased frequency of stagnation
July 2009
3-135
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
events in colder months as well as potential concurrent increases in emissions of these
compounds from motor vehicles with colder temperatures. The correlation between daily
max 8"h avg O3 and PM shows high seasonal variability with positive correlations on
average in the spring, summer and fall and negative correlations on average in the winter.
The highest positive correlations are in the summer, likely driven by favorable
photochemical formation conditions for both O3 and secondary aerosols (Joseph, 2008,
155219; Meng et al., 1997, 083324). The mean correlation drops below zero (-0.3 for PM2.5
and -0.2 for PM10) in the wintertime. As discussed in Chapter 3 of the last AQCD for Ozone
and other Photochemical Oxidants (U.S. EPA, 2006, 088089), this situation arises because
photochemical production of ozone in the planetary boundary layer is much smaller during
the winter than summer, while primary PM concentrations are elevated in many areas as a
product of heating emissions and lower mixing heights. Ozone in the boundary layer is
mainly associated with the subsidence from above the boundary layer following the passage
of cold fronts. This subsiding air has much lower PM concentrations than were present in
the boundary layer. Therefore, a negative association between O3 and PM2.5 is frequently
observed in the wintertime. Bell et al. (2007, 155683) also observed a wintertime minima in
same-day correlations between 24-h avg PM10 and O3 using data from 98 U.S. urban
communities over a 14-yr period (1987-2000). The average correlations were not negative in
wintertime, however, as seen in Figure 3"53. Furthermore, the highest national average
correlations were in spring and fall in the Bell et al. (2007, 155683) analysis rather than
summer as observed in Figure 3-53. This discrepancy could be a result of the different
averaging times used for O3 or the selection of different monitoring networks and/or time
periods.
Correlations among copollutants for individual CSAs/CBSAs are included in Annex A,
Figures A-155 through A-166 for PM2.5 and Figures A-167 through A-180 for PM10. The 15
CSAs/ CBSAs were chosen for further investigation, but several had an insufficient amount
of co-located data to be included. As can be seen from the individual CSAs/CBSAs in these
figures with multiple pairs of co-located monitors per pollutant, there can be considerable
variation in the correlations even within an individual urban area. Birmingham, Boston,
and St. Louis all show positive wintertime correlations between PM10 and daily maximum
8-h avg O3; Denver, Detroit, Houston, Los Angeles and Phoenix show negative wintertime
correlations. The remaining seven CSAs/CBSAs have insufficient data. For PM2.5, all
selected cities with sufficient data show negative correlations in the wintertime with daily
July 2009
3-136
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
max 8"h avg O3 (including Birmingham, Boston, Chicago, Denver, Houston, Los Angeles,
Philadelphia, Phoenix, Pittsburgh, Riverside and St. Louis). The remaining four
CSAs/CBSAs have insufficient data. In Baltimore (not one of the fifteen CSAs/CBSAs
included in this investigation), Sarnat et al. (2001, 019401) found a significant (at the
p <0.05 level) positive (0.67) and negative (-0.72) correlation between daily PM2.5 and O3 in
the summer (June 19" August 23, 1998) and winter (February 2-March 13, 1999),
respectively. These copollutant correlations illustrate the importance of considering
seasonality when assessing temporal relationships between air pollutants, particularly PM
and O3.
3.5.4. Summary
Many analyses in this section were based on 2005-2007 AQS data, which has varying
degrees of availability depending on the PM size fraction or component of interest. Overall,
PM2.5 mass has the broadest geographic coverage with PM10 having slightly less coverage.
PM10-2.5 mass is not routinely measured and reported to AQS and therefore was calculated
by difference using data from co-located PM10 and PM2.5 monitors. After applying data
completeness criterion and limiting the calculation to lowvolume FRM and FRM-like
monitors for measurement consistency, only 1% of U.S. counties, incorporating less than 5%
of the U.S. population, had qualifying PM10-2.5 data. Therefore, insufficient information was
available using AQS data to adequately characterize regional-scale PM10-2.5 concentrations
for this review. In all cases, monitors reporting to AQS were not uniformly distributed
across the U.S. or within counties, and conclusions drawn from AQS data may not apply
equally to all parts of a geographic region.
In general, for the eastern metropolitan areas investigated including Atlanta, Boston,
Chicago and New York, most of the mass of PM10 was in the PM2.5 size fraction, with the
highest ratio of PM2.5 to PMio-2.5in Chicago. In contrast, Denver and Phoenix contained
most of PM10 mass as PM10-2.5. This is consistent with the conclusion drawn in the 1996 PM
AQCD (U.S. EPA, 1996, 079380) where ratios of PM2.5 to PM10 were generally found to be
higher in the east than the west. However, expanded monitoring specifically designed for
PM10-2.5 is needed before a more conclusive regional characterization of PM10-2.5 can be
made.
July 2009
3-137
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Chemical Speciation Network (CSN) data was used to characterize PM2.5 composition
by region and CSA/CBSA. The highest annual average OC concentrations (>5 ug/m3) were
observed in the western and southeastern U.S. EC exhibited less seasonal variability than
OC with annual average EC concentrations greater than 1.5 ug/m3 in Los Angeles,
Pittsburgh, New York and El Paso. Concentrations of SC>42~ were higher in the eastern U.S.
resulting from higher SO2 emissions in the East, compared with the West. There was also
considerable seasonal variability with higher S() r concentrations in the summer months
when the oxidation of SO2 proceeds at a faster rate than during the winter. NO3
concentrations were highest in California, with annual averages >4 ug/m3 at many
monitoring locations. There were also elevated concentrations of NCV in the Midwest
(>2 ug/m3), with wintertime concentrations exceeding 4 ug/m3. In general, NOa was higher
in the winter across the country, resulting from a number of factors including (l) lower
temperatures favoring partitioning into particles, (2) higher relative humidity, particularly
in arid and semi-arid regions, (3) lower sulfate concentrations allowing higher uptake of
NOa- and (4) increased residential wood burning in specific areas of the U.S., especially in
the Northwest. Exceptions existed in Los Angeles and Riverside, where high NO3 readings
appeared year-round. Crustal material constituted a substantial fraction of PM2.5 year-
round in Phoenix (28%) and Denver (16%), and during the summer in Houston (26%).
In general, PM2.5 has a longer atmospheric lifetime than PM10-2.5 because larger
particles have a higher gravitational settling velocity. For PM2.5, most metropolitan areas
investigated exhibited high inter-monitor correlations (generally >0.75) out to a distance of
100 km. PM10 data from the same metropolitan areas, however, showed sharper declines in
inter-monitor correlations as a function of distance. Insufficient data were available in the
15 metropolitan areas to perform similar analyses for PM10-2.5 using co-located, low volume
FRM or FRM-like monitors. Although the general understanding of PM differential settling
leads to an expectation of greater spatial heterogeneity in the PM10-2.5 fraction relative to
the PM2.5 fraction in urban areas, deposition of particles as a function of size depends
strongly on local meteorological conditions, in particular on the degree of turbulence in the
mixing layer. Therefore, the findings from the 15 CSAs/CBSAs investigated may not apply
to all locations or at all times. Population density and associated building density are also
important determinants of the spatial distribution of PM concentrations.
Few studies performed direct comparisons of ultrafine particle measurements at
multiple locations within an urban area. A decrease in the number of ultrafine particles was
July 2009
3-138
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
demonstrated with shifts from a dominant mode at around 10 nm within 20 m of a freeway
to a flattened dominant mode at around 50 nm fat a distance of roughly 100-150 m. At the
same time, accumulation mode particle number concentration remained relatively constant
to within -300 m from the freeway. These findings suggest a high degree of spatial
heterogeneity in ultrafine particles compared with accumulation mode particles on the
urban scale.
Using hourly PM2.5 and PM10 observations in the 15 metropolitan areas, diel variation
showed a morning peak starting at approximately G:00 a.m., corresponding with the start of
morning rush hour. There was also an evening concentration peak that was broader than
the morning peak and extended into the overnight period, likely reflecting a combination of
evening rush hour and the concentration increase caused by the usual collapse of the mixed
layer after sundown. The magnitude and duration of these peaks varied considerably by
size fraction and metropolitan area. Studies indicate that ultrafine particles in urban
environments exhibit similar two-peaked diel patterns. Comparison between weekdays and
weekends as well as between urban street canyon and urban background sites suggest
traffic is a major source of ultrafine particles during the morning rush hour. The afternoon
peak in ultrafine particles likely represents the combination of primary source emissions
such as evening rush-hour traffic and photochemical formation of secondary organic and
sulfate aerosols.
Correlations between PM and gaseous copollutants varied both seasonally and
spatially between and within metropolitan areas. On average, PM10 and PM2.5 were
correlated with each other better than with the gaseous copollutants. There was relatively
little seasonal variability in the mean correlation between PM in both size fractions and
SO2 and NO2. CO, however, showed higher correlations with PM10 and PM2.5 on average in
the winter compared with the other seasons. This seasonality results in part because a
larger fraction of PM is primary in origin during the winter. To the extent that this primary
component of PM is associated with common sources of NO2 and CO, then higher
correlations with these gaseous copollutants are to be expected. Increased atmospheric
stability in colder months would also reinforce these associations. The correlation between
daily maximum 8"h avg O3 and PM2.5 showed the highest degree of seasonal variability with
positive correlations on average in the spring, summer and fall, and negative correlations
on average in the winter. This situation arises as the result of seasonal differences in PM
primary emissions and photochemical production of secondary PM2.5 and O3.
July 2009
3-139
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
3.6. Mathematical Modeling of PM
There are two main classes of models used to study atmospheric PM, receptor models
and chemistry-transport models (CTMs). Receptor models are statistical models whereas
CTMs are numerical models, i.e., they approximate derivatives by finite difference
approximations. Finite element models are also numerical models but have not been used
as extensively for applications described here, and so are not discussed further. Receptor
models are diagnostic in their approach, in that they try to derive source contributions at
monitoring locations using either ambient data alone or in combination with data for the
chemical composition of sources or in combination with meteorological data. Three-
dimensional chemistry and transport models are formulated in a prognostic, or predictive
manner, that is, they attempt to predict species concentrations by solving a set of coupled,
non-linear partial differential equations (continuity equations) for chemical species that
include terms based on emissions inventories, atmospheric transport, chemical
transformations, and deposition. Monitoring data is used to evaluate the performance of
CTMs. Each of these approaches has its own advantages and disadvantages.
3.6.1. Estimating Source Contributions to PM Using Receptor
Models
Methods for analyzing the composition of ambient PM samples in terms of
contributions from different sources are reviewed in this section. Associations between
exposures to ambient PM, as represented by ambient monitors, and health outcomes have
been extensively studied. Some health studies, described in Section 6.6, have used source
apportionment modeling to evaluate relationships between health outcomes and PM
(mainly PM2.5) from different sources. This section is intended to provide background
concerning the uses of source apportionment techniques in such studies. Understanding the
contribution of different emissions sources to ambient PM has also been used extensively in
evaluating air quality data for use in developing control strategies.
3.6.1.1. Receptor Models
Receptor models have been used mainly as part of the development of air quality
management plans. However, there have been several publications relating apportioned
July 2009
3-140
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
source types based on receptor models to human health effects. Discussions in this
section will focus mainly on those methods that have been used to relate health outcomes to
sources. More complete descriptions of a large number of types of receptor models currently
in use are given in Watson et al. (2008, 157128), who summarize the properties of these
methods, including the strengths and weaknesses. This compilation of receptor models,
broken down into different approaches (i.e., chemical mass balance, factor analysis,
tracer-based, meteorology based) is included in Tables A-45 through A-48 in Annex A.
Receptor models such as the chemical mass balance (CMB) model (Watson et al.,
1990, 004848) relate source category contributions to ambient PM concentrations based on
analyses of the compositional profiles of ambient and source emissions samples. It uses as
its basis a mass balance equation that represents all chemical species in an aerosol sample
as linear combinations of contributions from a fixed number of independent sources plus an
error term representing the portion of the measurement that cannot be fit by the model.
The compositional profiles used in receptor models can be extensive (see for example
the SPECIATE database, (http://www.eDa.gov/ttnchiel/software/sDeciate/index.html) for a
comprehensive collection of results from a large number of studies. As an example, several
studies have identified EC and over 100 organic carbon compounds in gasoline PM
emissions, including alkanes, PAHs, oxyPAHs, steranes, hopanes, and organic acids (Matti
Maricq, 2007, 155973; Schauer et al., 1999, 010582; Schauer et al., 2002, 035332). This
breakdown in identifiable groups of organic compounds is illustrated in Figure 3"54 and
Table 3-17 shows emissions factors for trace elements. Data for the compositional profiles
for several other important sources of PM that could be used for CMB modeling are shown
in Table A-49 in Annex A.
Source categories are amenable to refinement and to analysis as information on
tracers becomes available. For example, primary biological aerosol particles (PBAP) have
long been known to be significant constituents of the atmospheric aerosol, but not many
studies have evaluated their contributions, largely because of the lack of suitable tracers
and the additional equipment needs for sampling and analysis of bioaerosols. Bauer et al.
(2008, 189986) reviewed studies estimating the contribution of fungal spores to PM2.5 and
PM10-2.5 as fungal spores were expected to be major contributors to PBAP. They proposed the
use of arabitol and mannitol as unique tracers with an estimated accuracy of ± 50% to
apportion the contribution of fungal spores to OC in both PM2.5 and PM10-2.5. They estimated
July 2009
3-141
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
24-h avg contributions of - 40% to OC in PM10-2.5 during spring and summer in Vienna, with
a smaller contribution to PM2.5.
Fine Organic Extractabie and	Resolved	identified
Compounds	Elutable	Organics	Organics
S mg/l
18 mg/I
142 mg/l
118 mg/l
Not Extractabie
or Elutable
Extractabie
and Elutable
Organic Acids
Alkanes
PAH
Unresolved
Resolved
Unidentified
Identified
Source: Fraseret al. (1999. 010819)
Figure 3-54. Schematic of organic composition of particulate emissions from gasoline-fueled
vehicles.
One recently-identified concern in the application of CMB-based receptor models with
detailed organic marker compounds is the photochemical stability of those species.
Robinson et al. (2006, 156918) reported evidence of significant summertime photooxidation
of hopanes and long-chain alkenoic acids, low-volatility compounds often used as mobile
source and cooking emissions, respectively. Seasonal differences in hopanes/EC ratios
differed in a manner consistent with oxidation. Photochemical loss of particle-phase marker
species mass complicates the interpretation of model results, as long-range transport and
photochemistry may result in the loss of markers for distant sources. Furthermore,
photochemical breakdown of organic marker species may cause losses in CMB model
performance criteria and possible bias in source contribution estimates. The photooxidation
of condensed-phase organic compounds also may affect the polarity and volatility of these
compounds
In other methods, various forms of factor analysis are used that rely on the varying
mix of species present in ambient observations of compositional data to derive the source
contributions. Standard factor analytic approaches such as Principal Component Analysis
(PCA) have been used but PCA alone can apportion only the variance, not the mass, in an
aerosol composition data set. Additional steps such as those applied in Absolute Principal
Components Scores (APCS) are required to apportion mass from PCA (Miller et al., 2002,
July 2009
3-142
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
030661; Thurston and Spengler, 1985, 056074). In Positive Matrix Factorization (PMF)
(Paatero and Tapper, 1994, 086998), the ambient compositional data matrix is decomposed
into the product of a matrix representing the source contributions and one representing the
source profiles. Solutions are obtained by minimizing an object function with respect to
these two matrices, and solutions are subject to non-negativity constraints. PMF also allows
for the treatment of missing data and data near or below detection limits by weighting
elements inversely according to their uncertainties. The PMF approach requires a large
number of samples (n typically >50) and are most often applied to time series data, whereas
CMB can be applied to a single sample. Both the CMB and the PMF approaches find
solutions based on least squares fitting and minimization of an object function. Both
methods provide error estimates for the solutions based on estimates of the errors in the
input parameters. It should be noted, though, that the error estimates for both methods
often contain subjective judgments about the magnitude of the analytical and monitoring
errors.
Table 3-17. Example of emissions factors (ng/kg) for trace elements under variable speed and steady
speed driving conditions for PM emitted by diesel and gasoline engines. Note that
emissions are highly variable.
Element

Diesel

Gasoline
Transient
Steady State
Transient
Steady State
Al
9108 (5224)
2706
2273 (545)
252
Ca
69,443 (23,640)
16,128
18,247 (3044)
2324
Fe
22,910(21,448)
2036
10,266 (9928)
138
K
4672 (752)
1191
1935(558)
117
Mg
3087 (461
997
5183 (1706)
183
Na
7736 (1751)
1945
2237 (1125)
321
Ba
583 (349)
73
331 (55)
4.8
Be
26(12)
23
6.7 (1.1)
1.5
Cr
634 (354)
93
138 (6.7)
8.6
Cu
1944(679)
627
1745 (1803)
16
Li
13(0.2)
7.9
3.0(1.4)
0.9
Mn
368 (183)
76
152 (85)
3.4
Ni
2310(656)
644
107 (0.7)
21
Pb
793 (593)
79
237 (2.3)
11
S
23,750 (5295)
6713
8705 (3375)
349
Ti
2036 (320)
345
118(9.3)
24
July 2009
3-143
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29


Diesel

Gasoline
V
28 (9.4)
11
15(11)
1.8
Zn
21,118(4422)
5620
4650 (1225)
198
Source: Geller and Solomon (2006,139645).
The nature of the solutions in terms of source categories is different in the CMB and
PMF approaches. In the CMB approach, the composition of the source emissions is assumed
to be known based on measurements. These assumptions may or may not reflect the
composition of emissions affecting a particular site at any given time or place. However,
there may be variations in the composition of individual source categories (e.g., soils, motor
vehicle emissions) across a given airshed and even in the composition of the same source
with time. Source profiles can also be altered between emission and receptor locations
resulting from atmospheric reactions, depending on the source type and species under
analysis. The CMB technique was developed for apportioning source categories of primary
PM and was not formulated to include sources of secondary PM. CMB mighty not explain
all the mass or produce a valid result unless there is information for the composition of all
major sources affecting a given site, and there is confidence that the existing source profiles
are specific to those sources. For example, Volckens et al. (2008, 105465) describe PAH
emission profiles from hand-held gasoline lawn and garden equipment as found in some
CMB source profiles for motor vehicles.
In PMF, the source solutions are more general in that they contain information about
the entrainment of emissions from additional sources during transport, the time
dependence of the composition of emissions from particular sources, the formation of
secondary species and local differences in source compositions. PMF differs from CMB
because it derives the mix of factors from measured data. However, the procedure used to
find a solution results in some rotational ambiguity (Paatero and Tapper, 1994, 086998).
The assignment of sources to PMF factors depends largely on past experience and
judgments. Judgments are based to large extent on comparison with data for source profiles
and also on the factors that could modify the assignments. These issues are alleviated to
some degree by incorporating information about local wind fields and other physical
parameters.
The UNMIX model takes a geometric approach that exploits the covariance of the
ambient data to determine the number of independent sources, the composition and
contributions of the sources, and corresponding uncertainties (Henry, 1997, 020941).
July 2009
3-144
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
UNMIX uses PCAto find edges in m-dimensional space, where m is the number of ambient
species. Success of the UNMIX model hinges on the ability to find these "edges" in the
ambient data from which the number of source types and the source compositions are
extracted. In simplest terms, the approach can be seen to be similar to that for deriving
ternary mixing diagrams, except there is extension to higher dimensionality. Measurement
errors in the ambient data "fuzz" the edges, making them difficult to find. UNMIX employs
an "edge-finding" algorithm to find the best edges in the presence of error. UNMIX does not
make explicit use of errors or uncertainties in the ambient concentrations, unlike the
methods outlined above. Rather they are implicitly incorporated into the analyses. PMF
and UNMIX have also used data for particle size distributions to obtain further information
about sources.
Partial least squares (PLS) is another mathematical model related to PCA which has
been used in a limited number of PM toxicology studies to establish a relationship between
PM constituents and health outcomes (McDonald et al., 2004, 087158: Seagrave et al., 2006,
091291; Veranth et al., 2006, 087479). Unlike PCA and other receptor models discussed in
this section, PLS incorporates both predictor variables (e.g., PM component concentrations)
and outcome variables (e.g., toxicological responses) into one coupled regression model. Like
PCA, PLS groups the observable variables into a reduced number of latent variables,
thereby reducing the dimensionality of the model. Typically, PM toxicology studies have
been limited to two-component models (two latent variables on the predictor side compared
with two on the outcome side), thereby producing a 2x2 loading plot revealing relationships
between predictors and outcomes. PLS is particularly useful when there are more predictor
variables than observations, which is a situation that other multivariate factor analysis
approaches do not handle well. However, since PLS is a variance based approach, it shares
the same shortcomings discussed earlier for PCA. PLS has also traditionally been limited to
two-component applications even though this is not a strict mathematical limitation.
Results from Receptor Models
Results from receptor modeling calculations indicate that PM2.5 is most often
produced mainly by fossil fuel combustion. Fugitive dust, found mainly in the PM10-2.5 size
range, represents the largest source of measured ambient PM10 in many locations in the
western U.S. Quoted uncertainties in the source apportionment of constituents in ambient
aerosol samples typically range from 10 to 50%. It is apparent that a relatively small
July 2009
3-145
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
number of broadly defined source categories, compared to the total number of chemical
species that typically are measured in ambient monitoring-source receptor model studies,
are needed to account for the majority of the observed mass of PM in these studies. Trying
to be more specific about contributions from source categories could result in ambiguity. For
example, some stationary sources (e.g., agriculture use engines) and quite different mobile
sources (e.g., trucks and locomotives) rely on diesel power and ancillary data is required to
resolve contributions from these sources. Compilations of source attribution studies using
CMB for PMio have appeared in the PM AQCD (U.S. EPA, 2004, 056905) and using PMF
for PM2.5 in Engel'Cox and Weber (2007, 156419). Results of the compilation by Engel-Cox
and Weber (2007, 156419) for the eastern U.S. are shown in Figure 3"53. There are only
three main source categories in the figure constituting most of the PM2.5mass. Two of these
are predominantly secondary and not identified by sources of precursors. Tables A-50 and
A"51 in Annex A list results of other receptor modeling studies for PM10 and PM2.5, many of
which are in the western U.S.
23 "
TOTAL Sulfate	Nitrate	Mobile Bwiuag Industrial Oiii.«ts»lnlt Other
Source
Source: Engel-Cox and Weber (2007,156419).
Figure 3-55. Source category contributions to PM2.5 at a number of sites in the East derived using
PMF.
July 2009
3-146
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
Spatial Variability in Source Contributions to PM Based on Receptor Models
Spatial variability in source contributions across urban areas is an important
consideration in assessing the likelihood of exposure measurement error in epidemiologic
studies relating health endpoints to sources. Arguments similar to those for using ambient
concentrations as surrogates for personal exposures apply here. Studies for PM2.5 (Kim et
al., 2005, 083181; Wongphatarakul et al., 1998, 049281) indicate that intra-urban
variability increases in the following order: regional sources (e.g., secondary S() r
originating from EGUs) < area sources (e.g., on-road mobile sources) < point sources (e.g.,
stacks). This point is illustrated in Figure 3"56. The only study available for PM10-2.5
(Hwang et al., 2008, 134420) indicates a similar ordering, but without a regional component
(resulting from the short lifetime of coarse PM compared to transport times on the regional
scale) as shown in Figure 3"57.
1.0
0.8
c
Q>
O 0.6
it
©
O
o
TO
© 0.4
O
o
0.0
-0.2
Source: Kim et al. (2005, 083181)
Figure 3-56. Pearson correlation coefficients for source category contributions to PM2.5 between the
ten Regional Air Pollution Study/Regional Air Monitoring System (RAPS/RAMS)
monitoring sites in St. Louis.
July 2009
3-147
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
0.8
£
O
O
0)
o
o
c
o
_ro
v
—
k.
o
O
0.6 -
0.4 -
0.2 -
0.0 -
-0.2
T
T
T
T
T
T


&

6°x

#



Source: Hwange et al. (2008,134420)
Figure 3-57. Pearson correlation coefficients for source contributions to PM102.5 between the ten
Regional Air Pollution Study/Regional Air Monitoring System (RAPS/RAMS) monitoring
sites in St. Louis.
3.6.2. Chemistry Transport Models
Chemistry Transport Models (CTMs) are the prime tools used to compute the
interactions among atmospheric pollutants and their transformation products, the
production of secondary aerosols, the evolution of particle size distribution, and transport
and deposition of pollutants. CTMs are driven by emissions inventories for primary species
such as NOx, SOx, NHs, VOCs, and primary PM, and by meteorological fields produced by
other numerical prediction models. Values for meteorological state variables such as winds
and temperatures are taken from operational analyses, reanalyses, or weather circulation
models. In most cases, these are off-line meteorological analyses, meaning that they are not
modified by radiatively active species generated by the air quality model (AQM). Work to
integrate meteorology and chemistry was done in the mid 1990s by Lu et al. (1997, 048202
and references therein! 1997, 191768) although limits to computing power prevented their
wide-spread application. More recently, new, integrated models of meteorology and
chemistry are now available as well! see, for example, (Binkowski et al., 2007, 090563) and
July 2009
3-148
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
the Weather Research and Forecast model with chemistry (WRF-Chem)
(http: IIr uc. fsl. no a a. go v/wr f/W Gil).
CTMs have been developed for application over a wide range of spatial scales ranging
up from neighborhood to global. CTMs are used to: (l) obtain better understanding of the
processes controlling the formation, transport, and destruction of gas- and particle-phase
criteria and hazardous air pollutants! (2) understand the relations between concentrations
of secondary pollutant products and concentrations of their precursors! (3) understand
relations among the concentration patterns of various pollutants that may exert adverse
effects! and (4) evaluate how changes in emissions propagate through the atmospheric
system to secondary products and deposition.
Emissions of precursor compounds can be divided into anthropogenic and natural
source categories. Natural sources can be further divided into biogenic from vegetation,
microbes, and animals, and abiotic from biomass burning, lightning, and geogenic sources.
However, the distinction between natural sources and anthropogenic sources is often
difficult to make in practice, as human activities affect directly or indirectly emissions from
what would have been considered natural sources during the preindustrial era. Thus,
emissions from plants and animals used in agriculture have been referred to as
anthropogenic or biogenic in different applications. Wildfire emissions may be considered
natural, except that forest management practices can lead to buildup of fuels on the forest
floor, thereby altering the frequency and severity of forest fires.
The initial conditions, or starting concentration fields of all species computed by a
model, and the boundary conditions, or concentrations of species along the horizontal and
upper boundaries of the model domain throughout the simulation, must be specified at the
beginning of the simulation. Both initial and boundary conditions can be estimated from
models or data or, more generally, model + data hybrids. Because data for vertical profiles
of most species of interest are very sparse, results of model simulations over larger, usually
global, domains are often used. As might be expected, the influence of boundary conditions
depends on the lifetime of the species under consideration and the time scales for transport
from the boundaries to the interior of the model.
Each of the model components described above has associated uncertainties and the
relative importance of these uncertainties varies with the modeling application. The largest
errors in photochemical modeling are still thought to arise from the meteorological and
emissions inputs to the model (Russell and Dennis, 2000, 035563). While the effects of
July 2009
3-149
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
poorly specified boundary conditions propagate through the model's domain, the effects of
these errors remain undetermined. Because many meteorological processes occur on spatial
scales smaller than the model's vertical or horizontal grid spacing and thus are not
calculated explicitly, parameterizations of these processes must be used. These
parameterizations introduce additional uncertainty. Because the chemical production and
loss terms in the continuity equations for individual species are numerically coupled, the
chemical calculations must be performed iteratively until calculated concentrations
converge to within some preset criterion. The number of iterations and the convergence
criteria chosen also can introduce error.
CTMs in current use mostly have one of two forms. The first, grid-based or Eulerian
air quality models subdivide the region to be modeled, the modeling domain, into a three-
dimensional array of grid cells. Spatial derivatives in the species continuity equations are
cast in finite-difference form over this grid and a system of equations for the concentrations
of all the chemical species in the model are solved numerically at each grid point. Finite
element Eulerian models also exist and have been exercised, but less frequently. Time-
dependent continuity or mass conservation equations are solved for each species including
terms for transport, chemical production and destruction, and emissions and deposition (if
relevant), in each grid cell. Chemical processes are simulated with ordinary differential
equations, and transport processes are simulated with partial differential equations.
Because of a number of factors such as the different time scales inherent in different
processes, the coupled, nonlinear nature of the chemical process terms, and computer
storage limitations, not all of the terms in the equations are solved simultaneously in three
dimensions. Instead, operator splitting, in which terms in the continuity equation involving
individual processes are solved sequentially, is used.
In the second common CTM formulation, trajectory or Lagrangian models, a number
of hypothetical air parcels are specified as though following wind trajectories. In these
models, the original system of partial differential equations is transformed into a system of
ordinary differential equations.
A less common approach is to use a hybrid Lagrangian-Eulerian model, in which
certain aspects of atmospheric chemistry and transport are treated with a Lagrangian
approach and others are treaded in an Eulerian manner (see, e.g., Stein et al., 2000,
048341).
July 2009
3-150
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Each approach has advantages and disadvantages. The Eulerian approach is more
general in that it includes processes that mix air parcels and allows integrations to be
carried out for long periods during which individual air parcels lose their identity. There
are, however, techniques for including the effects of mixing in Lagrangian models such as
FLEXPART (Zanis et al., 2003, 053423), ATTILA (Reithmeier and Sausen, 2002, 053447),
and CLaMS (McKenna et al., 2002, 053445). Because both the accuracy and the
computational intensity of Eulerian models depend strongly on the size of the horizontal
and vertical grid spacing, speed and fidelity to actual atmospheric conditions much
sometimes be traded-off! that is to say, while finer grid spacing will often capture effects
missed at larger grid intervals, models set up in this way require longer to solve. In a
similar manner, the accuracy of Lagrangian models depends on the number of air parcels
deployed! thus they, too, become computationally intensive when higher-order accuracy is
desired. More detailed discussion of CTM applications appears in the 2008 ISA for NOx and
SOx ¦ Ecological Criteria (U.S. EPA, 2008, 157074).
3.6.2.1. Global Scale
Global-scale CTMs are used to address issues associated with climate change and
stratospheric O3 depletion to characterize long-range air pollution transport, and to provide
boundary conditions for the regional-scale models. The CTMs include parameterizations of
atmospheric transport! the transfer of solar radiation through the atmosphere! chemical
reactions! and removal to the surface by turbulent motions and precipitation for emitted
pollutants. The upper boundaries of the CTMs extend anywhere from the tropopause (~8
km at the poles to ~16 km in the tropics) to the mesopause at ~80 km in order to obtain
more realistic boundary conditions for problems involving stratospheric dynamics and
chemistry.
Global simulations are typically conducted with a horizontal grid spacing of 200 km or
more, although some models such as GEOS-Chem have been run at grid spacings of about
100 km (see e.g.,Wu et al., 2008, 190039) and efforts are being made to go to even higher
spatial resolution. Simulations of the effects of long-range transport at particular locations
link multiple horizontal resolutions from the global to the local scale. Finer resolution can
only improve scientific understanding to the extent that the governing processes are more
July 2009
3-151
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
accurately described at that scale. Consequently, there is a crucial need for observations at
the appropriate scales to evaluate the scientific understanding represented by the models.
3.6.2.2. Regional Scale
Most major regional-scale air-related modeling efforts at EPA use the Community
Multi-scale Air Quality modeling system (CMAQ) (Byun and Ching, 1999, 156314; Byun
and Schere, 2006, 090560). A number of other modeling platforms using Lagrangian and
Eulerian frameworks were reviewed in the 2006 AQCD for O3 (U.S. EPA, 2006, 088089) and
in Russell and Dennis (Russell and Dennis, 2000, 035563). The capabilities of a number of
CTMs designed to study local- and regional-scale air pollution problems were summarized
by Russell and Dennis (2000, 035563). Evaluations of the performance of CMAQ are given
in Arnold et al. (2003, 087579) Eder and Yu (2005, 089229) Appel et al. (2005, 089227), and
Fuentes and Raftery (2005, 087580). CMAQ's horizontal domain can extend from a few
hundred kilometers on a side to the entire hemisphere. CMAQ is most often driven by the
MM5 mesoscale meteorological model (Seaman, 2000, 035562), though it may be driven by
other meteorological models including WRF and the Regional Atmospheric Modeling
System (RAMS); see http://atmet.com. Simulations of pollution episodes over regional
domains have been performed with a horizontal resolution as low as 1 km! see the
application and general survey results reported in Ching et al. (2006, 090300). However,
simulations at such high resolutions require better parameterizations of meteorological
processes such as boundary layer fluxes, deep convection and clouds (Seaman, 2000,
035562). Finer spatial resolution is necessary to resolve features such as urban heat island
circulation; sea, bay, and land breezes! mountain and valley breezes! and the nocturnal low-
level jet, all of which can affect pollutant concentrations.
The most common approach to setting up the horizontal domain is to nest a finer grid
within a larger domain of coarser resolution. However, there are other strategies such as
the stretched grid (see, e.g., Fox-Rabinovitz et al., 2002, 047806) and the adaptive grid (see,
e.g., Hansen et al., 1994, 046634). In a stretched grid, the grid's resolution continuously
varies throughout the domain, thereby eliminating any potential problems with the sudden
change from one resolution to another at the boundary. Caution should be exercised in
using such a formulation because certain parameterizations like those for convection might
be valid on a relatively coarse grid scale but may not be valid on finer scales. Adaptive grids
July 2009
3-152
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
are not fixed at the start of the simulation, but instead adapt to the needs of the simulation
as it evolves. They have the advantage that they can resolve processes at relevant spatial
scales. However, they can be very slow if the situation to be modeled is complex.
Additionally, if adaptive grids are used for separate meteorological, emissions, and
photochemical models, there is no reason a priori why the resolution of each grid should
match, and the gains realized from increased resolution in one model will be wasted in the
transition to another model. The use of finer horizontal resolution in CTMs will necessitate
finer-scale inventories of land use and better knowledge of the exact paths of roads,
locations of factories, and, in general, better methods for locating sources and estimating
their emissions.
The vertical resolution of these CTMs is variable and usually configured to have more
layers in the PBL and fewer higher up. Because the height of the boundary layer is of
critical importance in simulations of air quality, improved resolution of the boundary layer
height would likely improve air quality simulations. Additionally, current CTMs do not
adequately resolve fine-scale features such as the nocturnal low-level jet in part because
little is known about the nighttime boundary layer.
CTMs require time-dependent, three-dimensional wind fields for the period of
simulation. The winds may be generated either by a model using initial fields alone or with
four-dimensional data assimilation to improve the model's performance! i.e., model
equations can be updated periodically to bring results into agreement with observations.
Modeling series durations can range from simulations of several days duration, the typical
time scale for individual O3 episodes, to several months or multiple seasons of the year. The
current trend in modeling applications is towards annual simulations. This trend is driven
in part by the need to improve understanding of observations of periods of high wintertime
PM (Blanchard et al., 2002, 047598) and the need to simulate O3 episodes occurring in
spring, fall, and winter.
Chemical kinetics mechanisms representing the important reactions occurring in the
atmosphere are used in CTMs to estimate the rates of chemical formation and destruction
of each pollutant simulated as a function of time. Mechanisms that treat the reactions of all
individual reactive species explicitly are computationally too demanding to be incorporated
into CTMs for regulatory use. Similarly, very extensive "master mechanisms" (Derwent et
al., 2001, 047912) that include approximately 10,500 reactions involving 3603 chemical
species (Derwent et al., 2001, 047912) can be combined into mechanisms that group
July 2009
3-153
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
together compounds with similar chemistry. Because of different approaches to the lumping
of organic compounds into surrogate groups for computational efficiency, chemical
mechanisms can produce different results under similar conditions. The Carbon Bond
chemical mechanisms starting with CB-IV (Gery et al., 1989, 043039), the RADM II
mechanism (Stockwell et al., 1990, 043095), the SAPRC (see, e.g., Carter, 1990, 042893;
Wang et al., 2000, 048357; Wang et al., 2000, 048365), and the RACM mechanisms can be
used in CMAQ. Jimenez et al. (Jimenez et al., 2003, 156611) provided brief descriptions of
the features of the main mechanisms in use and compared concentrations of several key
species predicted by seven chemical mechanisms in a box-model simulation over 24 h.
CMAQ and other state-of-the-science CTMs incorporate processes and interactions of
aerosol-phase chemistry (Binkowski and Roselle, 2003, 191769; Gaydos et al., 2007, 139738)
Zhang and Wexler (2008, 191770). There have also been several attempts to study the
feedbacks of chemistry on atmospheric dynamics using meteorological models like MM5
and WRF (Grell et al., 2000, 048047; Liu et al., 2001, 048201; Lu et al., 1997, 048202; Park
et al., 2001, 156841). This coupling is necessary to accurately simulate feedbacks which
may be caused by the heavy aerosol loading found in forest fire plumes (Lu et al., 1997,
048202; Park et al., 2001, 156841) or in heavily polluted areas. Photolysis rates in CMAQ
can now be calculated interactively with model produced O3, NO2, and aerosol fields
(Goldstein et al., 1973, 015674).
Spatial and temporal characterizations of anthropogenic and biogenic precursor
emissions must be specified as inputs to a CTM. Emissions inventories have been compiled
on grids of varying resolution for many HCs, aldehydes, ketones, CO, NH3, and NOx.
Emissions inventories for many species require the application of algorithms for calculating
the dependence of emissions on physical variables such as temperature and to convert the
inventories into formatted emission files which can be used by a CTM. For example,
preprocessing of emissions data for CMAQ often is done by the Spare-Matrix Operator
Kernel Emissions (SMOKE) system (httt>://smoke-model.org). For many species,
information concerning the temporal variability of emissions is lacking, so long-term
annual averages are used in short-term, episodic simulations. Annual emissions estimates
are often modified by the emissions model to produce emissions more characteristic of the
time of day and season. Significant errors in emissions can occur if inappropriate time
dependence is used. Additional complexity arises in model calculations because different
July 2009
3-154
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
chemical mechanisms can include different species, and inventories constructed for use
with one mechanism must be adjusted to reflect these differences in another.
3.6.2.3. Local or Neighborhood Scale
The grid spacing in regional CTMs, usually between 1 and 12 km2, is usually too
coarse to resolve spatial variations on the neighborhood scale. The interface between
regional scale models and models of smaller exposure scales is provided by smaller scale
dispersion models. Several models could be used to simulate concentration fields near
roads, each with its own set of strengths and weaknesses. The California Department of
Transportation's most recent line dispersion model is CALINE4; see
http://www.dot.ca.gov/hq/env/air/pages/calinesw.htm. The CALINE family of models is not
supported by the California Department of Transportation for modeling of highway source
PM, however, but only for roadway CO, although PM work with CALINE has performed for
more than ten years! see Wu et al. (2009, 191773) and references therein.
In addition, AERMOD (http://www.epa.gov/scrarnOOl/dispersion prefrec.htm) is a
steady-state plume model formulated as a replacement to the ISC3 dispersion model. In the
stable boundary layer (SBL), it assumes the concentration distribution to be Gaussian in
both the vertical and horizontal dimensions. In the convective boundary layer, the
horizontal distribution is also assumed to be Gaussian, but the vertical distribution is
described with a bi-Gaussian probability density function (pdf). AERMOD has provisions to
be applied to flat and complex terrain and multiple source types (including, point, area and
volume sources) in both urban and rural areas. It incorporates air dispersion based on PBL
turbulence structure and scaling concepts and is meant to treat both surface and elevated
sources and simple and complex terrain in rural and urban areas. The dispersion of
emissions from line sources like highways in AEROMOD is handled as a source with
dimensions set using an area or volume source algorithm in the model! however, actual
emissions are usually not in steady state and there are different functional relationships
between buoyant plume rise in point and line sources. Moreover, most simple dispersion
models including AERMOD are designed without chemical mechanisms and so cannot
produce secondary pollutants from their primary emissions .
There are also non-steady state models that incorporate plume rise explicitly from
different types of sources. For example, CALPUFF
July 2009
3-155
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
(httpV/www.src.com/c alpuff/calnul'l'l.htm), which is EPA's recommended dispersion model
for transport in ranges >50 km, is a non-steady-state puff dispersion model that simulates
the effects of time- and space-varying meteorological conditions on pollution transport,
transformation, and removal and has provisions for calculating dispersion from surface
sources. However, CALPUFF was not designed to treat the dispersion of emissions from
roads, and like AERMOD does not include production of secondary pollutants. The
distinction between a steady-state and time varying model could be unimportant for long
time scales! however, at short time scales, the temporal variability in traffic emissions could
result in underestimation of peak concentration and exposures.
3.6.3. Air Quality Model Evaluation for Air Concentrations
Urban and regional air quality is determined by a complex system of coupled chemical
and physical processes including emissions of pollutants and pollutant precursors, complex
chemical reactions, physical transport and diffusion, and wet and dry deposition. NOx in
these systems has long been known to act nonlinearly in the production of O3 and other
secondary pollutants (Dodge, 1977, 038646) to extend over multiple spatial and temporal
scales! and to involve complicated cross-media environmental issues such as acidic or
nutrient deposition to sensitive biota and degradation of visibility.
NOy species emitted and transformed from emissions control the production and fate
of both O3 and aerosols by sustaining or suppressing OH cycling. Correctly characterizing
the interrelated NOy and OH dynamics for O3 formation and fate in the polluted
troposphere depends on new techniques using combinations of several NOy species for
diagnostically probing the complex atmospheric dynamics in typical urban and regional
airsheds.
Evaluation results from a recent EPA exercise of CMAQ in the Tampa Bay, FL,
airshed are presented here as an example of the present level of skill of state-of-the-science
AQMs for predicting atmospheric concentrations of some of the relevant species for this PM
NAAQS assessment. This modeling series exercised CMAQ version 4.4 and with the
University of California at Davis (UCD) sectional aerosol module in place of the standard
CMAQ modal aerosol module and was driven by meteorology from MM5 v3.6 and with NEI
emissions as augmented by continuous emissions monitoring data where available. The
UCD size-segregated module was preferred for this application because of the importance of
July 2009
3-156
DRAFT-DO NOT CITE OR QUOTE

-------
1	sea salt particles in the bay airshed. Testing of this new engineering extension to CMAQ
2	(termed CMAQ-UCD below) revealed that its performance was very similar to that of
3	CMAQ's standard modal module! hence, model behavior and performance reported here can
4	stand as a general indication of CMAQ's skill.
5	The CTM was run with 21 vertical layers for the month of May 2002. For this
6	evaluation, CMAQ-UCD was run in a one-way nested series of three domains with 32 km, 8
7	km, and 2 km horizontal grid spacings from the CONUS (32 km) to central Florida and the
8	eastern Gulf of Mexico (2 km). Depictions of the 8 km and 2 km domains used here zoomed
9	over the central Tampa area are shown in Figure 3-58 and Figure 3-59.
8KM RESOLUTION; ZOOMED2
• Sydney
Gulf of Mexico
Figure 3-58. Eight km southeast U.S. CMAQ-UCD domain zoomed over Tampa Bay, FL.
RESOLUTION; ZOOMED2
2KM
• Sydney
Gulf of Mexico
Figure 3-59. Two km southeast U.S. CMAQ-UCD domain zoomed over Tampa Bay, FL.
July 2009	3-157	DRAFT - DO NOT CITE OR QUOTE

-------
5 T
4 - -
Z2--
Hours (EST)
Hours (EST)
Hours (EST)
Figure 3-60. Hourly average CMAQ-UCD predictions and measured observations of NO (top), NO2
(middle), and total NOx (bottom) concentrations for May 1-31, 2002. Green squares = 8
km solution, red diamonds = 2 km solution, blue circles = observations.
July 2009
3-158
DRAFT-DO NOT CITE OR QUOTE

-------
4 T
3 - -
¦Q
Q.
a.

LU
1 -May 2-May 3-May 4-May 5-May 6-May 7-May 8-May 9-May 10-
May
Hours (date mark = 0000 EST)
4 r
3 - -
2 - -
11- 12- 13- 14- 15- 16- 17- 18- 19- 20-
May May May May May May May May May May
Hours (date mark = 0000 EST)
4 T
3 - -
v 2 --
21- 22- 23- 24- 25- 26- 27- 28- 29- 30- 31-
May May May May May May May May May May May
Hours (date mark = 0000 EST)
Figure 3-61. CMAQ-UCD predictions and measured observations of ethene concentrations at Sydney,
FL for May 1-31, 2002. Green squares = 8 km solution, blue circles = observations .
July 2009
3-159
DRAFT-DO NOT CITE OR QUOTE

-------
8 T
6 - -
¦Q
Q.
a.

o
c
1
0
21- 22- 23- 24- 25- 26- 27- 28- 29- 30- 31-
May May May May May May May May May May May
Hours (date mark = 0000 EST)
Figure 3-62. CMAQ-UCD predictions and measured observations of isoprene concentrations at
Sydney, FL for May 1-31, 2002. Green squares = 8 km solution, blue circles =
observations.
July 2009
3-160
DRAFT-DO NOT CITE OR QUOTE

-------
60 x
50 --
^ V #

1 -May 2-May 3-May 4-May 5-May 6-May 7-May 8-May 9-May 10-
May
Hours (date mark = 0000 EST)
11 -May 12-May 13-May 14-May 15-May 16-May 17-May 18-May 19-May 20-May
Hours (date mark = 0000 EST)
I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I
21- 22- 23- 24- 25- 26- 27- 28- 29- 30- 31-
May May May May May May May May May May May
Hours (date mark = 0000 EST)
Figure 3-63. CMAQ-UCD predictions and measured observations of PM2.5 concentrations at St.
Petersburg, FL for May 1-31, 2002. Green squares = 8 km solution, blue circles =
observations.
July 2009
3-161
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
3.6.3.1.	Ground-based Comparisons of Photochemical Dynamics
Errors in the NOx concentrations in the model, most likely from on-road emissions,
affected NOx predictions shown in Figure .'WiO, but CMAQ-UCD's general responses were
reasonable. The model also replicated anthropogenic and biogenic VOC emissions well; see
Figure 3"61 and Figure .'Hi2, respectively. After initial errors leading to underprediction in
the first 21 days, CMAQ-UCD's predictions of hourly PM2.5 concentrations and trends over
the whole month also replicated the observed concentrations well; see Figure 3-63.
3.6.3.2.	Predicted Chemistry for Nitrates and Related Compounds
Particulate NOa (pNOa ) plays a crucial and complex role in the health of aquatic and
estuarine ecosystems and human drinking water systems. Gas-phase NC>3~ replacement of
CI on sea salt particles is often favored thermodynamically and the Vd of the coarse pNOa
formed through this replacement is more than an order of magnitude greater than for fine
pN03_. Over open bodies of salt water such as the Gulf of Mexico and Tampa Bay, FL,
pNOa from this reaction dominates dry deposition and is estimated to be of the same order
as pN03~wet deposition.
However, total NO3 concentrations are driven, buffered, and altered by a wide range
of photochemical gas-phase reactions, heterogeneous reactions, and aerosol dynamics, mak-
ing them especially difficult to model. Because pNOs is derived mostly from gas-phase
HNO3 and will interact with Na+, NH4"1", CI , and SO r , all these species and the physical
parameters governing their creation, transport, transformation, and fate must be accu-
rately replicated to predict pNOa with high fidelity. This has historically been a difficult
problem for numerical process models, owing not least to the pervasive dearth of reliable
ambient measurements of N03~in its various forms. Normalized mean error (NME) for the
large-scale Eulerian CTM-predicted pN03~ has typically been on the order of a factor of 3
greater than the NME for particulate SO42- (pS042~) (Odman et al., 2002, 092474; Pun et al.,
2003, 047775).
SO42", NH4"1", Na+, and CI were all predicted to within a factor of 2 and with no signifi-
cant bias during the photochemical day in the 8 km CMAQ-UCD solution, although a
significant bias in Na+ and CI was evident in the 2 km solution for two near-water sites.
This grid-size dependent bias is still being explored. Size segregation maxima were correct
to within two size bins every day for which there were observations for both S() r and NH4"1"
July 2009
3-162
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
(0.2 to 1.0 pm), and Na+ and CI (2.0 to 10.0 pm). CI-concentrations were greatly overpre-
dicted during dark hours, but were nearer to observed values during the photochemical day.
CMAQ-UCD performance for HNO3 and NH3 are shown in Figure 3"64 and Figure 3"65, re-
spectively.
Figure 3"66 shows that CMAQ-UCD systematically underp re dieted the hourly time
series of measured pNOa concentrations at the Sydney supersite, the only location with
discrete pNCV data. These time series data establish that CMAQ-UCD's largest errors
were on four days in the first two weeks of the month, but that the total peak pNO.3 concen-
trations were nearly all underp re dieted.
Since pNCV is derived in large part from gas-phase HNO3, its underprediction may be
due to an underprediction of HNO3 concentrations or an underrepresentation of the gas- to
aerosol-phase change. At Sydney, FL, in fact, both these conditions held. Figure 3-64 depicts
the model's bias for HNO3 underprediction in both the 8 km and 2 km solutions, excepting
four days of very large peak overp re dictions. This trend was especially true overnight! on
eight other days the model overpredicted the one hour peak concentration as well, though
not so substantially, but the chief effect was still one of an artificial and inappropriate N
limitation in the model.
A time series molar equivalent ratio of HNO3 to total NO3 depicts which phase stores
the NCV and how that storage ratio changes over time. Figure 3-67 shows that at Sydney,
FL, CMAQ-UCD stored too much NO3- in the gas phase as HNO3 (and recall that the day-
time HNO3 concentrations were sometimes overpredicted by the model) and too little in the
gas phase overnight, when the model was regularly low against the measurements! com-
pare Figures 3-64 and 3-65. Note again here the self-similarity of the 8 km and 2 km solu-
tions in this comparison.
Interestingly, the 23"h integrated data did not reveal this important difference in NO3"
form between the model and measurements as Figure 3-68 shows in the stacked bar
percentage plots of fine and coarse pNCV together with gas-phase HNO3. Both the 8 km
(Figure 3-68, middle panel) and the 2 km (Figure 3-68, bottom panel) solutions predicted
distributions between the two general ranges of aerosol size, and between gas and aerosol
phases, with good fidelity to the daily observations (Figure 3-68, top panel) at Sydney, FL.
This result illustrates that while discrete time series data are crucial for diagnosing model
behavior, on the integrated total daily and longer basis used for computing total annual N
July 2009
3-163
DRAFT-DO NOT CITE OR QUOTE

-------
1	loads, CMAQ-UCD predicted approximately the correct distributions for pNOa , even
2	though the total NOa concentration prediction was biased low.
o
1-May 2-May 3-May 4-May 5-May 6-May 7-May 8-May 9-May 10-
May
Hours (Date Mark = 0000 EST)
11-May 12-May 13-May 14-May 15-May 16-May 17-May 18-May 19-May 20-May
Hours (Date Mark = 0000 EST)
0
21-May22-May23-May24-May25-May26-May27-May28-May29-May30-May31-May
Hours (Date Mark = 0000 EST)
Figure 3-64. CMAQ-UCD predictions of HhKh' concentrations and corresponding measured
observations at Sydney, FL, for May 1-31, 2002. Green x = 8 km solution, red diamonds
= 2 km solution, blue circles = observations.
July 2009
3-164
DRAFT-DO NOT CITE OR QUOTE

-------
-May 2-May 3-May 4-May 5-May 6-May 7-May 8-May 9-May 10-
May
Hours (date mark = 0000 EST)
11- 12- 13- 14- 15- 16- 17- 18- 19- 20-
May May May May May May May May May May
Hours (date mark = 0000 EST)
-M ay22-May23-M ay24-M ay25-May26-M ay27-May28-May29-May30-May31-May
Hours (date mark = 0000 EST)
Figure 3-65. CMAQ-UCD predictions of IUH3 concentrations and corresponding measured observations
at Sydney, FL, for May 1-31, 2002. Green squares = 8 km solution, red
July 2009
3-165
DRAFT-DO NOT CITE OR QUOTE

-------
"re 2
Q-
i2
21-May22-May23-May24-May25-May26-May27-May28-May29-May30-May31-IVlay
Hours (date mark = 0000 EST)
10-
May
11-May 12-May 13-May 14-May 15-May 16-May 17-May 18-May 19-May 20-May
Hours (date mark = 0000 EST)
Hours (date mark = 0000 EST)
3-May 4-May 5-May 6-May 7-May 8-May 9-May
E 5 - -
ra 3 - -
0
1-May 2-May
Figure 3-66. CMAQ-UCD predictions of phl03~ concentrations and corresponding measured
observations at Sydney, FL, for 1-31 May, 2002. Green x = 8 km solution, red diamonds
= 2 km solution, blue circles = observations.
July 2009
3-166
DRAFT-DO NOT CITE OR QUOTE

-------
o
Eo.6
0.8
00.4
0.0 I T I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I T I ¦ I ¦ I T I ¦ I ¦ I T I T I ¦ I ¦ I T I ¦ I
11-May 12-May 13-May 14-May 15-May 16-May 17-May 18-May 19-May 20-May
Hours (date mark = 0000 EST)
0.8 --
0.2 --
0.0 ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I ¦ I
21-May22-May23-May24-May25-May26-May27-May28-May29-May30-May31-lVIay
Hours (date mark = 0000 EST)
1-May 2-May 3-May 4-May 5-May 6-May 7-May 8-May 9-May 10-May
Hours (date mark = 0000 EST)
Figure 3-67. CMAQ-UCD predictions of the ratio of HhKh to total IUO3 and corresponding measured
observations at Sydney, FL, for May 1-31, 2002. Green x = 8 km solution, red diamonds
= 2 km solution, blue circles = observations.
July 2009
3-167
DRAFT-DO NOT CITE OR QUOTE

-------
MOUDI aerosol N03 and ARA HN03 (23 h dailymeai)
J
5/1 5/3 5/5 5/7 5/9 5/11 5/13 5/15 5/17 5/19 5/21 5/23 5/25 5/27 5/29 5/31
¦ < 25 irn 125 -10 un ~ HN03
CMAQ-UCD 8 km solution (23 h daily meai)
1
5/1 5/3 5/5 5/7 5/9 5/11 5/13 5/15 5/17 5/19 5/21 5/23 5/25 5/27 5/29 5/31
I < 25 un 125- 10um CIHN03
CMAQ-UCD 2 km solution (23hotivteans)
1
5/1 S3 5/5 5/7 5/9 5/11 5/13 5/15 5/17 5/19 5/21 5/23 5/25 5/27 5/29 5/31
Date. 2002
¦ <25 um 125-10 urn ~ HN03
Figure 3-68. CMAQ-UCD predicted size and chemical-form fractions of total N03~ for days in May
2002 with measured observations. Measured concentrations (top panel); 8 km
solution (middle panel); 2 km solution (bottom panel). Red bars = phl03~ < 2.5 Dm; blue
bars = pl\l03~ 2.5-10 Dm; green bars = HhKh.
July 2009
3-168
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
While inorganic aerosol anion totals were dominated by NO:; in the coarse fraction
and by S( )r in the fine fraction, there was sufficient NHx (NHx = NH3 + NH4"1") at Sydney,
FL, to form fine aerosol NH4NO3 in some circumstances. Figure 3"65 depicts the hourly
mass concentration of NH3 at Sydney, FL, showing again the strong self-similarity of the 8
km and 2 km solutions. Each solution, however, underpredieted the measured NH3
concentrations consistently, and especially for the nine very large excursions of 10 to 20
pg/m3 during the month.
Overall, CMAQ-UCD was found to be operationally sound in this evaluation of its 8
km and 2 km solutions for the Tampa Bay airshed using the ground-based and aloft data
(not shown here) from the May 2002 field intensive. Moreover, results from diagnostic tests
of the model's photochemical dynamics were generally in excellent agreement with results
from the ambient atmosphere. However, CMAQ-UCD was biased low in this application for
total NO3 and for NO3 present as gas-phase HNO3. In addition, the model was biased low
for the HOx radical reservoir species CH2O and H2O2 (not shown here), though this bias
appeared to have been limited to these species. Performance of the new UCD aerosol
module was judged to be entirely adequate, allocating aerosols by chemical makeup to the
appropriate size fractions. Model performance for fine-mode aerosols was also judged to be
fully adequate.
3.6.4. Evaluating Concentrations and Deposition of PM
Components with CTMs
3.6.4.1. Global CTM Performance
The wet and dry deposition processes described in Section 3.3 are of necessity highly
parameterized in all CTMs. While all current models implement resistance schemes for dry
deposition, the Vd generated from different models can vary highly across terrain types
(Stevenson et al., 2006, 089222) .The accuracy of wet deposition in global CTMs is tied to
spatial and temporal distribution of model precipitation and the treatment of chemical
scavenging. Dentener et al. (2006, 088434) compared wet deposition across 23 models with
available measurements around the globe. Figures 3-69 and 3-70 extract results of a
comparison of the 23-model mean versus observations over the eastern U.S. for pNC>3~ and
pSC>42- deposition, respectively. The mean model results were strongly correlated with the
July 2009
3-169
DRAFT-DO NOT CITE OR QUOTE

-------
1	observations (r >0.8), and usually captured the magnitude of wet deposition to within a
2	factor of two over the eastern U.S. Dentener et al. (2006, 088434) concluded that 60 to 70%
3	of the participating models captured the measurements to within 50% in regions with
4	quality controlled observations.
600
i "I.
400
03
~o
o
2
* ¦
200
Percentage within ± 50%: 74.i
o
200
400
600
Measurement
Source: Dentener et al. (2006, 088434)). Reprinted with permission.
Figure 3-69. Scatter plot of total nitrate (HNO3 plus phl03~) wet deposition (mg hl/m2/yr) of the model
mean versus measurements for the North American Deposition Program (NADP)
network. Dashed lines indicate a factor of two. The gray line is a linear regression
through zero.
July 2009
3-170
DRAFT-DO NOT CITE OR QUOTE

-------
1000
800 -
600 -
o
T3
O
S
¦ s #
/ ¦ i"i Ay i" ¦
/	¦ pf ¦	Bi
¦/ . H- * .
I /, ' ¦ it" '
' V'ii ¦ " 1
/¦ *¦ ¦# ¦ ¦ ''
, l 1i r" i -'
; ii . ¦
Lit*
* l'! ¦ ^	
Ave. model: 383 Ave. Meas: 322 r: 0.87 n = 226
2 param fit: y = 114.0 + 0.77x
1 param fit: y = 1.00x
percentage within 150%: 66.0	
_i_
-L-
400 600
Measurement
800 1000
Source: Dentener et al. (2006, 088434). Reprinted with permission.
Figure 3-70. Scatter plot of total S042~ wet deposition (mg S/m2/yr) of the model mean versus
measurements for the National Atmospheric Deposition Program (IUADP) network.
Dashed lines indicate a factor of two. The gray line is a linear regression through zero.
3.6.4.2. Regional CTM Performance
1	Regional CTM performance for concentration and deposition of some of the most
2	relevant PM species is illustrated here with examples from CMAQ version 4.6.1 as
3	configured and run for exposure and risk assessments reported in the Draft Risk and
4	Exposure Assessment for the Review of the Secondary National Ambient Air Quality
5	Standards for Oxides of Nitrogen and Oxides of Sulfur (U.S. EPA, 2009, 191774); additional
6	details on the model configuration and application are found there. A map of the 36 km
7	parent domain and two 12 km (east and west) progeny domains appears in Figure 3-71.
July 2009
3-171
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Figure 3-71. CMAQ modeling domains for the OAQPS risk and exposure assessments: 36 km outer
parent domain in black; 12 km western U.S. (WUS) domain in red; 12 km eastern U.S.
(EUS) domain in blue.
Comparisons from the 2002 annual run of CMAQ for the exposure assessment are
shown here against measured concentrations and deposition totals from nodes in three
networks: IMPROVE, CSN (labeled STN in the plots) and CASTNet. Comparisons were
made as model-observation pairs at all sites having sufficient data for the seasonal or the
2002 annual time period in the two 12 km east and west domains and were evaluated with
the following descriptive statistics: correlation, root mean square error, normalized mean
bias, and normalized mean error.
Summertime pS() r concentrations are well predicted by CMAQ, to within a factor of
2 at nearly every point, and with R2 >0.8 across all three networks! see Figure 3"72. This
result tracks the generally well-predicted S() r concentrations found in earlier CMAQ
evaluations: see Eder and Yu (2005, 089229; Mebust et al., 2003, 156749), Mebust et al.
(2003, 156749) and Tesche et al. (2006, 157050). Since pSC>42~ concentrations are strongly a
function of precipitation, care must be taken to ensure that the meteorological solution
driving individual CMAQ chemical applications produces precipitation fields with low bias
as discussed by Appel et al. (2008, 155660).
July 2009
3-172
DRAFT-DO NOT CITE OR QUOTE

-------
2002ac_met2v33_12kmE S04 for June to August 2002

2Q02ac
met2v33 12kmE

CORR
RMSE NMB NME
IMPROVE
0.96
0.84 -11.7 17.2
STN
0.83
1.57 0.4 20.1
CASTNet
0.97
0.83 -8.6 12.5
Q IMPROVE (2002ac met2v33 12kmE) a
A STN (2002ac_met2v33_12kmE)
CASTNet (2002ac_met2v33_12km E)
Monthly Average
S04 (ug/m3)
Observation
Figure 3-72. 12-km EUS Summer sulfate PM. Each data point represents a paired monthly averaged
(June/July/August) observation and CMAQ prediction at a particular IMPROVE, STN, and
CASTNet site. Solid lines indicate a factor of two around the 1:1 line shown between
them.
1	Wintertime pNCV (Figure 3-73) and total NOs 0 INOi;+pNO:;) (Figure 3-74)
2	concentrations are predicted less well by CMAQ; but NOs is a pervasively difficult species
3	to measure and model. Still, at the CASTNet nodes where the total NOs concentrations are
4	higher than they are at all but a few of the remote IMPROVE sites, CMAQ predicts
5	concentrations for nearly every node to within a factor of 2 and with an R2 >0.8. These
6	CMAQ-predicted concentrations, coupled with modeled cloud and precipitation fields
7	produce wet deposition fields for SO42- and NOs- in the east domain as shown in Figures
8	3-75 and 3-76, respectively.
July 2009
3-173
DRAFT-DO NOT CITE OR QUOTE

-------
2002ac_met2v33_12kmE N03 for December to February 2002
~ IMPROVE (2002ac_met2v33_12kmE)
A STN (2002ac_met2v33_12kmE)
oo
CO
AA
O
<
2
O
AA,
Monthly Average
a N03 ( ug/m3 )
A A
,/ CP
*2VP ~
C\J
2002ac met2v33 12kmE
CORR RMSE NMB NME
0.5 47
-6.9 35.7
IMPROVED 0.73 0.69
STN 0.67 1.32
O
0
2
6
8
4
Observation
Figure 3-73. 12 km EUS Winter nitrate PM. Each data point represents a paired monthly averaged
(December/January/February) observation and CMAQ prediction at a particular IMPROVE
and STN site. Solid lines indicate a factor of two around the 1:1 line shown between
theem.
July 2009
3-174
DRAFT-DO NOT CITE OR QUOTE

-------
2002ac_met2v33_12kmE TN03 for December to February 2002
~ CASTNet (2002ac_met2v33_12kmE)
CO
in
]~ SI
qnn
m
~
~~
~ ~
CO
Monthly Average
TN03 ( ug/m3)
~~
OJ
~ ~
2002ac met2v33 12kmE
CORR RMSE NMB NME
CASTNet 0.85 0.88 -1.6 21.4
O
0
1
2
3
4
5
6
7
Observation
Figure 3-74. 12-km EUS Winter total nitrate (HhKh + total phl03~). Each data point represents a paired
monthly averaged (December/January/February) observation and CMAQ prediction at a
particular CASTNet site. Solid lines indicate a factor of two around the 1:1 line shown
between them.
July 2009
3-175
DRAFT-DO NOT CITE OR QUOTE

-------
2002ac met2v33 12km E S04 for 20020101 to 20021231
~ NADPdep (2002ac_met2v33_12kmE)
NMB
NME
NMdnB
NMdnE
FB
FE
RMSE
RMSEs
RMSEu
MB
ME
MdnB
MdnE
Period Accumulated
S04 (kg/ha)
Observation
Figure 3-75. 12-km EUS annual sulfate wet deposition.Each data point represents an annual average
paired observation and CMAQ prediction at a particular NADP site. Solid lines indicate
the factor of 2 around the 1:1 line shown between them.
July 2009
3-176
DRAFT-DO NOT CITE OR QUOTE

-------
2002ac_met2v33_12kmE N03 for 20020101 to 20021231
~ NADP dep (2002ac met2v33 12kmE}
(kg/ha)
NMB
NME
NMdnB
NMdnE
FB
FE
RMSE
RMSEs
RMSEu
MB
ME
MdnB
MdnE
Period Accumulated
N03 (kg/ha)
Observation
Figure 3-76. 12-km EUS annual nitrate wet deposition. Each data point represents an annual average
paired observation and CMAQ prediction at a particular NADP site. Solid lines indicate a
factor of two around the 1:1 line shown between them.
1	Importantly, CMAQ captured the chief spatial patterns and magnitudes of air
2	concentrations and wet deposition relevant to computing concentration and deposition
3	budgets, as shown in Figures 3-77 for concentrations and 3-78 for deposition. More
4	specifically, CMAQ s predictions of NHs and S(for both high and low concentration sites
5	are well within the range of measurements there; see Figure 3-79.
July 2009
3-177
DRAFT-DO NOT CITE OR QUOTE

-------
J4c_emteul7 S04 lor June to August 2002	J4c«nlsv17 NH4 for June to August 2002
CJ
o
o
2
CO
monthly average
NH4 (ug/m3)
o
0
2
4
10
0
1
2
3
4
5
Observation	Observation
Figure 3-77. CMAQ vs. measured air concentrations from east-coast sites in the IMPROVE, CSN
(labeled STN), and CASTNet sites in the summer of 2002 for sulfate (left) and ammonium
(right). Solid lines indicate a factor of 2 around the 1:1 line shown between them.
CMAQ Ammonium Ion Wet
WET MM UzHJHIlUN jK3>*9
OMQtMt
VR N«P Ca»T—2KB M3MGB3
LMTEDTO 3STB& N THE EMCIBH US.
JWNUN.
NADP Ammonium Ion Wet Deposition
200 2003 200
Ammonium *»
NH* (kg/ha)
Figure 3-78. Comparison of CMAQ-predicted and NADP-measured l\IH4+ wet deposition: (top left)
CMAQ prediction; (bottom left) NADP-measurements; (right) regression and smoothed
median line through CMAQ predictions and NADP measurements with sites in the
Chesapeake Bay watershed highlighted.
July 2009
3-178
DRAFT-DO NOT CITE OR QUOTE

-------
Kenansville Ammonia July 2004
12-hour Averages: 6am-6pm
190 192 194 196 198 200 202 204 206 208 210 212 214
Julian Day: TickMark at Midnight (July 2004)
I-~-Kenansville NH3 -B-CMAQ-J4c NH3 I
Millbrook (Raleigh) Sulfate July 2004
12-hour Averages: 6am-6pm
3 10
190 192 194 196 198 200 202 204 206 208 210 212 214
Julian Day: TickMark at Midnight (July 2004)
!-~-Millbrook S04 CMAQ-J4c AS04 I
Kenansville Sulfate July 2004
12-hour Averages: 6am-6pm











If


t . tK t I


~


\
I





T
"11 w
1

190 192 194 196 198 200 202 204 206 208 210 212 214
Julian Day: TickMark at Midnight (July 2004)
I Kenansville S04 CMAQ-J4c AS04 I
Millbrook (Raleigh) Ammonia July 2004
12-hour Averages: 6am-6pm
190 192 194 196 198 200 202 204 206 208 210 212 214
Julian Day: TickMark at Midnight (July 2004)
I Millbrook NH3 CMAQ-J4c NH3 I
Figure 3-79. CMAQ-predicted (red symbols and lines) and 12-h measured (blue symbols and lines) IUH3
and SO42- surface concentrations at high and low concentration grid cells in North
Carolina in July 2004. (top left) High concentration IUH3 in Kenansville; (top right) high
concentration S042~ in Kenansville; (bottom left) low concentration IUH3 in Raleigh;
(bottom right) low concentration S042~ in Raleigh.
1	Deposition velocities are difficult to estimate for reasons described in Section 3.3.3.
2	Recent work in EPA's Atmospheric Modeling and Analysis Division with CMAQ showed
3	that the original Vd for NHs was very likely too high and should be nearer to the values for
4	SO2 deposition, or even lower over some land use surface types. A sensitivity study with the
5	model was performed to test the effects of changing Vd for NHs on the fraction of NHs
6	available for transport away from grid cells with high emissions concentrations.
7	Comparisons were made for the surface grid cells and total column NHs concentrations.
8	In the highest emissions grid cells during June 2002, the surface NHx budget was
9	dominated by turbulent transport or vertical mixing moving a majority of the surface NHs
10 emissions up and away from the surface into the mixed layer. Figure 3"80 depicts the NHx
July 2009
3-179
DRAFT-DO NOT CITE OR QUOTE

-------
1	budget under the base case (Base Vd) and the sensitivity case (SO2 Vd) for which the NHs Vd
2	was set equal to the SO2 Vd. Lower NHs Vd decreased NHx deposition to the surface from 15
3	to 8 %, leaving more NHx for transport horizontally, 22% up from 20% in the base case, and
4	vertically, 69% up from 64% in the base case. Typically, -67% of surface emissions were
5	moved aloft where most was advected away from the high emissions grid cell, with a small
6	fraction converted to pNH4+ and an even smaller fraction wet-deposited to the surface. The
7	total column analyses for NHs and NHx are shown in Figure 3"81.
lop
TTyf^IoileT- TTHim
Free Troposphere
: 69-64%:
jVertical
¦Diffusiort
Mixed Layer (-2 km)
0.04%
38m Qas
Surface partiriP
TT
c
22-20%
Horizontal
*\dvection
NHj
Emissions
8-15%
Dry Deposition
Right Number: BaseVd
Left Number: SC^Vd
NH.< Layer 1 Analysis
Figure 3-80. Surface grid cell (layer 1) analysis of the sensitivity of NHx deposition and transport to
the change in IUH3 Vd in CMAQ.
July 2009
3-180
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
2.1-2.2%
Gas to Particle
Conversion
(unusually
low)
'Tojwf fyfocfeT- HHun
Free Troposphere
89-82%
Horizontal
Advection Mixed Layer(-2 km)
f 0.55-0.54%
Wet Deposition
13 Dry Deposition
Emissions	8-15%
NH
91-84%
Horizontal
Advection
Xffj Column Analysis
Right Number: BaseVd
Left Number: S02Vd
0.58-0.57%
Wet Deposition
13 Dry Deposition
Emissions	8-15%
NH
NHx Column Analysis
Figure 3-81. Total column analysis for IUH3 (left) and NHx (right) showing modeled IUH3 emissions,
transformation, and transport throughout the mixed layer and up to the free
troposphere.
Local total deposition (wet + dry) is a significant but not dominant loss pathway for
surface NHs emissions. In these simulations, CMAQ deposited -25% of the NHs emissions
from the single high concentration grid cell in Sampson County, NC, back into that grid cell.
By far, the largest contribution to the local deposition total was dry deposition! dry-to-wet
deposition ratios for the Sampson County high emissions grid cell and surrounding surface
grid cells ranged from 2 to 10.
Deposition to grid cells farther away from the high concentration, immediately
surrounding grid cells was significantly affected by the change in NHs Vd tested in this
case. Figure 3-82 depicts the range of influence of the high concentration grid cell, where
that range is defined to be the distance by which 50% of the emissions attributable to that
grid cell have deposited. The range of influence of the high concentration Sampson County
grid cell was extended in the Vd sensitivity tested here from -180 km in the base case to
-400 km in the case using the lower, more realistic Vd for NHs. The areal extent of this
difference in range of influence is mapped in Figure 3"83.
July 2009
3-181
DRAFT-DO NOT CITE OR QUOTE

-------
Distance from Center (km)
June 2002 NHx Range of Influence: BaseVd vs. S02Vd
Sampson County (single cell)
BaseVd
S02Vd
-180km
-400km
- -Wet+Dry Dep BaseVd
—o—Wet+Dry Dep S02Vd
Advection BaseVd
Advection S02Vd
Figure 3-82. Range of influence (where 50% of emitted IUH3 deposits) from the high concentration
Sampson County grid cell in the June 2002 CMAQ simulation of Vd sensitivities. Base
case and sensitivity case total deposition (blue symbols and lines); base case and
sensitivity case advection totals (red symbols and lines).
July 2009
3-182
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 3-83. Areal extent of the change in NHx range of influence as predicted by CMAQ for the
Sampson County high concentration grid cell (center of range circles) in June 2002 using
the base case and sensitivity case Vd.
3.7. Background PM
1	The background concentrations of PM that are useful for risk and policy assessments
2	informing decisions about the NAAQS are referred to as policy-relevant background (PRB)
3	concentrations. PRB concentrations have historically been defined by EPA as those
4	concentrations that would occur in the U.S. in the absence of anthropogenic emissions in
5	continental North America defined here as the U.S., Canada, and Mexico. For this
6	document, PRB concentrations include contributions from natural sources everywhere in
7	the world and from anthropogenic sources outside continental North America. Background
8	concentrations so defined facilitated separation of pollution that can be controlled by U.S.
9	regulations or through international agreements with neighboring countries from those
July 2009
3-183
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
that were judged to be generally uncontrollable by the U.S. Over time, consideration of
potential broader ranging international agreements may lead to alternative determinations
of which PM source contributions should be considered by EPA as part of PRB.
3.7.1. Contributors to PRB concentrations of PM
Contributions to PRB concentrations of PM include both primary and secondary
natural and anthropogenic components. Natural sources include wind erosion of natural
surfaces (Gillette and Hanson, 1989, 030212); volcanic production of SO42-, primary
biological aerosol particles (PBAP); wildfires producing EC, OC, and inorganic and organic
PM precursors; and SOA produced by oxidation of biogenic hydrocarbons such as isoprene
and terpenes. However, human intervention can be involved in the formation of SOA, as
production of natural SOA depends to a large extent on the presence of anthropogenic NOx.
As described earlier in Section 3.3, prescribed fires are considered part of PRB. In addition
to emissions from forest fires in the U.S., emissions from forest fires in other countries can
be transported to the U.S. For example, Boreal forest fires in Canada (Mathur, 2008,
156742) and Siberia (Generoso et al., 2007, 155786) and tropical forest fires in the Yucatan
Peninsula and Central America (Wang et al., 2006, 157109) have affected PM
concentrations in the U.S. PRB PM varies across the contiguous Unites States (CONUS) by
region and season as a function of complex mechanism of transport, dispersion, deposition,
and reentrainment.
Dust from the Sahara desert and the Sahel in North Africa (Chiapello et al., 2005,
156339) affects mainly the eastern U.S.; dust from the Gobi and Taklimikan deserts in Asia
(VanCuren and Cahill, 2002, 157087; Yu et al., 2008, 157168) have the largest effects in the
western U.S. but also affect air quality in the eastern U.S. Husar et al. (2001, 024947)
report that the average PM10 concentration at 25 reporting stations throughout the
northwestern U.S. reached 65 pg/m3 during an episode in the last week in April 1998,
compared to an average of 10-25 pg/m3 during the rest of April and May. This was
accompanied by visual reports of milky-white discoloration of the normally blue sky in non-
urban areas along the west coast.
PRB contributions to PM2.5, PM10-2.5, and PM10 can also be viewed as coming from two
conceptually separate components: a reasonably consistent "baseline" component and an
episodic component. The baseline component consists of contributions that are generally
July 2009
3-184
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
well characterized by a reasonably consistent distribution of daily values each year,
although there is variability by region and season. The episodic component consists of
infrequent, sporadic contributions from natural high-concentration events occurring over
shorter periods of time (e.g., hours to several days) both within North America (e.g.,
volcanic eruptions, large forest fires, dust storms) and outside North America (e.g.,
transport related to dust storms from deserts in North Africa and China and storms at sea).
These episodic natural events, as well as events like the uncontrolled biomass burning in
Central America, are essentially uncontrollable and do not necessarily occur in all years.
In-situ measurements provide evidence for the transport of anthropogenic PM from
Asia on Mt. Batchelor, OR by Jaffe et al. (2003, 041957). These data show sporadic but well
correlated increases in CO, O3, total Hg, and aerosol backscatter associated with air coming
from Asia. The ITCT-2K2 campaign also found evidence for the oxidation of SO2 to H2SO4
during trans-Pacific transport of Asian emissions. If particulate S042~ were to be formed in
the polluted boundary layer where it originated, it would likely be deposited prior to
transport across the Pacific Ocean (Brock et al., 2004, 156295). Thus primary species
emitted directly and secondary species formed during transport contribute to PRB
concentrations. Satellite data have provided images to track clouds of dust and pollution
across the oceans and have been used for some quantitative estimation of the flux of
material leaving continents. Yu et al. (2008, 157168) used optical thickness data to estimate
column loadings from the Moderate Resolution Imaging Spectroradiometer (MODIS) along
with satellite assimilated wind fields to estimate the transport of PM from Asia.
Three-dimensional, global-scale chemistry-transport models have also been used to
estimate intercontinental transport of PM pollution (UNCEC, 2007, 157078) and
trans-Pacific transport of mineral dust from Asian deserts (Fairlie et al., 2007, 141923) and
the Sahara Desert (McKendry et al., 2007, 156748).
Estimates for the contribution of PBAP are highly problematic. Heald and Spracklen
(2009, 19001 1) estimated the contribution of fungal spores to PM2.5 based on GEOS-Chem
simulations of mannitol, considered to be a unique tracer for fungal spores (Bauer et al.,
2008, 189986). They estimated an annual mean contribution of fungal spores to OC ranging
from <0.1 ug/ma in the desert Southwest to ~ 0.5 ug/m3 in the more humid Southeast. It
should be noted that these are model derived estimates that still require evaluation against
measurements in the U.S. They do, however, provide the only quantitative estimates of
PBAP concentrations across the continental U.S.
July 2009
3-185
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
3.7.1.1. Estimates of PRB Concentrations in Previous Assessments
Estimates of PRB concentrations reported in the 1996 AQCD for PM (U.S. EPA, 1996,
079380) and earlier AQCDs were based in large measure on estimates by Trijonis et al.
(1990, 157058) for the national Acid precipitation Assessment Program as shown in
Table 3-18. Different approaches for estimating PRB concentrations in the western and
eastern U.S. were taken in the 2004 PM AQCD (U.S. EPA, 2004, 056905). Data obtained at
IMPROVE monitoring sites in the western U.S. shown in Figure 3"84 were chosen as
estimates of the distribution of daily average PRB concentrations in the West because they
were thought to be among the least likely influenced by regional pollution sources
especially at the upper end of the concentration distribution. This conclusion was drawn
from back trajectory analyses and examination of the trace elemental composition at
IMPROVE sites. Because of likely unresolved contamination from pollution sources at other
IMPROVE sites, it was recommended to use averaged data from these sites throughout the
West. Concentrations distributions from 1988 through 2001 can be found in Appendix 3E of
the 2004 PM AQCD. Median concentrations were around 3 ug/m3. Little interannual
variability was observed below the 90th percentile values. However, at the upper end of the
concentration distribution substantial interannual variability was observed due mainly to
forest fires and dust transport from Asian deserts. It was also recognized that this method
would likely overestimate PRB concentrations.
Table 3-18.
Estimates of annual average natural background concentrations of PM2.5 and PM10 (//g/m3)

from Trijonis et al. (1990,157058). Estimates of PM10 2.5 were obtained by subtraction.
PM2.5 PM10 PM102.5
East
2-5 5-11 ~ 1-9
West
1-4 4-8 ~ 1-7
July 2009
3-186
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 3-84. IMPROVE monitoring site locations.
1	Table 3-19 shows annual and quarterly average PM2.5 concentrations measured at the
2	IMPROVE sites shown in Figure 3"82 for 2004. Annual average concentrations tend to be
3	slightly higher in the East, particularly in Brigantine and Dolly Sods. When the data are
4	broken down by season, a more complex picture emerges. Highest values in the East and
5	Midwest are found during the 3rd calendar quarter, whereas in the West highest quarterly
6	averages can occur during other quarters. As can also be seen from a comparison with
7	values shown in Table 3-18, PM2.5 values measured in the East are much higher than the
8	PRB concentration estimates by Trijonis et al. (1990, 157058) for the NAPAP.
Table 3-19.
Annual and quarterly mean PM2.5 concentrations (Dg/m3) measured at IMPROVE sites in
2004.

Mean
January-March
April-June
July-September
October-December
EAST
Acadia
4.5
3.9
4.6
6.0
3.5
Brigantine
9.5
8.1
11.3
11.6
7.3
Dolly Sods
9.5
6.7
9.8
15.5
5.7
MIDWEST
Voyageurs
3.8
4.1
3.1
4.2
3.6
WEST
Bridger
2.1
1.2
3.1
2.8
1.3
July 2009
3-187
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

Mean
January-March
April-June
July-September
October-December
Canyonlands
2.6
2.2
3.2
2.9
2.1
Gila
2.9
2.0
4.0
3.8
1.8
Glacier
4.8
4.6
4.2
5.3
5.0
Redwood
3.5
2.7
3.6
3.7
3.9
Thus, estimating daily average PRB concentrations in the eastern U.S. using
observations is highly problematic because of the widespread mixing of precursors and
anthropogenic PM generated in the East. Therefore, results from receptor modeling studies
using PMF by Song et al. (2001, 036064) were used in the 2004 PM AQCD for the East to
separate contributions from likely regional pollution sources from natural and imported
pollution. The "background" sources contribute about 7% to annual average PM2.5
concentrations at Underhill, VT and about 12% at Brigantine, NJ, i.e., values between 1
and 2 jug/m3. These sites were chosen because they were outside of urban areas making it
easier to separate pollution from background components. However, some contribution of
regional anthropogenic pollution was still present.
The PM Staff Paper (U.S. EPA, 2006, 157071) adopted a different approach for
estimating PRB concentrations. This approach separated out components mainly thought to
be emitted by regional pollution sources such as SO42", which are obtained directly from
observations at many more IMPROVE sites than are shown in Figure 3"84, and to use the
remaining PM components in both the East and the West to estimate PRB. Removing
regional pollution from data obtained at IMPROVE sites is problematic as it involves
assumptions about the relative contributions of regional pollution and background sources.
Although sulfate in the East is mainly produced by regional pollution sources, it is not the
only component with a regional pollution source. By comparison, sulfate is a very minor
component of PM2.5 in the West, leading to substantial overestimates in populated states
like California. Annual mean estimates in the continental U.S. ranged from 2.5 ug/m3 in the
Central West (ID, MT, WY, ND, SD, CO, UT, NZ, AZ) to 5.2 ug/m3 for the Southwest Coast
(most of CA), the latter value likely reflecting contributions from local, non-sulfate
pollution.
In general, the methods outlined in both these documents are of limited utility for two
reasons: (l) they lack detailed spatial coverage across the whole U.S., since both methods
rely on monitoring data that are limited both spatially and temporally! and (2) PM
measurements from even the limited, remote sites used in the previous estimates of PRB
July 2009
3-188
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
can not be completely devoid of contributions from anthropogenic PM. Because of these
limitations, numerical modeling can provide superior PM background estimates, as
described just below.
3.7.1.2. Chemistry Transport Models (CTMs) for Predicting PRB Concentrations
CTMs can be used to estimate the PRB concentrations of atmospheric components
including PM using a "zero-out" approach in which anthropogenic emissions inside
continental North America are set to zero while global biogenic emissions and
anthropogenic emissions outside continental North America remain. Numerical modeling
can provide more precision in the estimate of PRB PM than measurements since even the
most remote measurement sites like some of those in the IMPROVE network (see the
discussion in Section 6.1.1. above) will necessarily be affected by non-local non-biogenic
pollution, thereby confusing the contributions from these sources. Numerical models are
also capable of supplying estimates of PRB concentrations at much higher spatial and
temporal resolution than can be obtained by relying on measurements obtained at even the
most remote monitoring sites. In this approach, the monitoring data are used to evaluate
the CTM's performance.
For this assessment, the global-scale circulation model GEOS-Chem was coupled with
the regional scale air quality model CMAQ (see the discussion of CMAQ in Section 3.6.2) to
simulate 1-yr of air quality data over the CONUS in two series of runs, the first annual
series with all anthropogenic and biogenic emissions included and the second annual series
with the zero-out approach employed.
The global-scale scale circulation model was set up as follows. GEOS-Chem version 7
was used, with modifications to include aromatic and biogenic SOA formation! emissions
were computed from a variety of sources including the Global Emissions Inventory Activity
(GEIA) (Benkovitz et al., 1996, 156267), and Emissions Database for Global Atmospheric
Research, version 2 (EDGAR) (Olivier et al., 1996, 156828; Olivier et al., 1999, 156829).
Particularized emissions in specific areas used the European Monitoring and
Evaluation Program (EMEP) for Europe (Auvray and Bey, 2005, 156237), BRAVO (Kuhns
and Knipping, 2005, 156663) for Mexico, Streets et al. (2006, 157019) for Asia, Martin et al.
(2002, 089380) for additional NOx emissions from biofuels, lightning, and ship traffic, Bond
et al. (2004, 056389) for global primary organic aerosols, and Cooke et al. (1999, 156365)
July 2009
3-189
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
and Park et al. (2003, 156842) for U.S. primary organic aerosols. Biomass burning
emissions are not climatological but were computed with GFEDv2 (Giglio et al., 2006,
156469; van der Werf et al., 2006, 157084) monthly values using active fire observations
from MODIS; global dust fields were computed off-line using GOCART (see emissions from
DEAD: http://dust.ess.uci.edu/dead/) to make annual adjustments to photolysis rates and
heterogeneous-phase chemistry.
The regional CTM was set up as follows. CMAQ version 4.7 (excluding the dynamic
coarse mode updates) was used with the SAPRC_99 chemical mechanism and AER05
aerosol module! emissions were processed through SMOKE (http://smoke-model.org)
version 2.4 based on the 2004 projections from the NEI with specific CEM, biogenics, and
fire updates! MM5 version 3.7.4 was used with the Asymmetric Convective Mixing, version
2.2, PBL scheme! and data nudging was used to analyze fields for winds and temperature.
Model Evaluation
Details from evaluations of the performance of a number of CMAQ applications are
given in Arnold et al. (2003, 087579). Eder and Yu (2006, 1 12721). Appel et al. (2005,
089227), and Fuentes and Raftery (2005, 087580).
In an annual simulation series for 2002 using CMAQ v4.6.1 in two 12 km domains for
the CONUS (see Figure 3-85), predicted concentrations of summertime particulate SO4,
often a major determinant of surface-layer PM concentrations, were well-predicted by
CMAQ at a 12 km grid spacing, to within a factor of 2 at nearly every point of comparison
and with R2 >0.8 across all three national networks (CASTNet, IMPROVE and CSN); a
more detailed description is included in the 2008 NOxSOx ISA (U.S. EPA, 2008, 157071).
This result for CMAQ v4.6.1 for 2002 tracks the generally well-predicted SO r
concentrations found in most earlier CMAQ evaluations: see Mebust et al. (2003, 156749),
Eder and Yu (2006, 112721). and Tesche et al. (2006, 157050). Since particulate S042~
concentrations are strongly a function of precipitation, care must be taken to ensure that
the meteorological solution driving individual CMAQ chemical applications produces
precipitation fields with low bias as discussed by Appel et al. (2008, 155660).
July 2009
3-190
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
2002ae_mel2v33_12kmE S04 for June to August 2002
o
< <°
2
o
<£>
Monthly Average
S04 ( ug>'m3}
CM
O
0
2
6
8
10
12
14
4
Figure 3-85. 12 km EUS Summer SO*2- PM. Each data point represents a paired monthly averaged
{June/July/August) observation and CMAQ prediction at a particular IMPROVE, STN, and
CASTNet site. Solid lines indicate the factor of 2 around the 1:1 line shown between
them.
Wintertime particulate N(V (Figure.3-86) and total NO;, (I [NO.-; + particulate NO;, )
(Figure 3-87) concentrations are predicted as well by CMAQ; but NQ3- is a pervasively
difficult species to measure and model. Still, at the CASTNet nodes where the total NOs~
concentrations are higher than they are at all but a few of the remote IMPROVE sites,
CMAQ predicts concentrations for nearly every node to within a factor of 2 and with an R2
>0.8.
A "base case" in which conditions for 2004 including all the anthropogenic and natural
sources both within and outside of continental North America was run for comparison with
measurements. A PRB si m illation was also run by shutting off the anthropogenic sources of
primary PM and precursors to secondary PM inside continental North America.
July 2009
3-191
DRAFT-DO NOT CITE OR QUOTE

-------
20DZic_m«l?i31_1 SmE NCU for (lucarnb*i la Ftbnury 2D07
IMPROVE rSOUSK ma12v33 1»mE|
& STN-;SDaZac_mrt?v33 I2km£;
Monfly An/wage
aW09 j u^'ffiS )
'* iV'" *
* %¦¦ ^
N
0
2
6
a
4
QfeSfflVattan
Figure 3-86 12 km EUS Winter N(h PM. Each data point represents a paired monthly averaged
(December/January/February) observation and CMAQ prediction at a particular IMPROVE
and STN site. Solid lines indicate the factor of 2 around the 1:1 line shown between
them.
2lMZ*c_mi?t2v33_12fcmlL TNOcj tor December to February 2D0Z
MnrtMy Avenge
tkOS r ua	-\« 21.4
QASTNhI |20O?h::j ihI?v33 12hrrE".
Ob.scrva.1ian
Figure 3-87. 12-km EUS Winter total nitrate (HIUO3 + total particulate NO3-). Each data point
represents a paired monthly averaged (December/January/February) observation and
CMAQ prediction at a particular CASTNet site. Solid lines indicate the factor of 2 around
the 1:1 line shown between them.
1	Figures 3-88 through 3-90 show monthly average concentrations, and Figures 3-91
2	through 3.7-10 show 24-h avg concentration distributions for 2004 predicted by CMAQ for
July 2009
3-192
DRAFT - DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
the base case and for PRB. Measurements are also included for the five Western and four
Eastern/Midwestern IMPROVE sites shown in Figure 3-84. Wildfires could have affected
the grid cell containing the Midwestern Voyageurs site resulting in the high PRB values
found for the July average compared to the PRB values for the rest of the year. The "base
case" simulations tend to underestimate concentrations throughout most western sites as
shown in Figures 3-89 and 3-90. These underestimates are still within the range of a
few |ug/m3. However, the base case simulation also greatly over-predicts PM2.5
concentrations at the upper end of the distribution at the Redwoods site (Figure 3-90). This
over-prediction results from emissions from wildfires in northern California that are
included in the grid cell containing the Redwoods site, but may not have affected the site.
However, wildfires indicated by MODIS would have affected other areas either close to
these sites or could have affected other locations in between the IMPROVE sites. The
simulated monthly average PRB concentrations in the East/Midwest range from a
minimum of 0.6 |ug/m3 at Acadia NP in July to 3.7 ug/m3 at Voyageurs NP in July. However,
most values are <1 ug/m3. The monthly average PRB concentrations calculated for the West
tend to be lower than for the East and range from 0.2 jug/m3 at Bridger and Glacier NPs in
January and February, respectively, to 8.7 |ug/m3 at Redwoods NP in November. Excluding
values at Redwoods NP which greatly exceed measurements, the highest monthly average
concentration was 3.7 |ug/m3 at Voyageurs NP in the East/Midwest and 2.4 ug/m3 at Gila NP
in the West.
July 2009
3-193
DRAFT-DO NOT CITE OR QUOTE

-------
Acidia
Brigantine
20
16
"I 12
3 8
s
4
—i—i—i i i—i—i—i	1—i—i
1 2 3 4 6 6 7 0 9 10 11 12
Month
I
20
16
12
8
4
0
—i—i i i—i i i i—i—i—i
2 3 4 6 6 7 0 9 10 11 12
Month
Dolly Soda
Voyagoura
J
j
20
16
12
8
4
0
1 2 3 4 5 6 7 0 9 10 11 12
Month
-0IM
2 3 4 6 6 7 0 9 10 11 12
Month
PRB
Figure 3-88. Monthly average of PM2.5 concentrations measured at IMPROVE sites in the East and
Midwest for 2004. Also shown are distributions of PM2.5 concentrations calculated by
CMAQ for the base case and for PRB.
July 2009
3-194
DRAFT-DO NOT CITE OR QUOTE

-------
Brldgar
Canyonlanda
Olla
Qlaolar
1 2 3 4 6 6 7 8 9 10 11 12
Month
Monitored —m— Bats
PRB
Figure 3-89. Monthly average of PM2.5 concentrations measured at IMPROVE sites in the West for
2004. Also shown are distributions of PM2.5 concentrations calculated by CMAQ for the
base case and for PRB.
July 2009
3-195
DRAFT-DO NOT CITE OR QUOTE

-------
Acaaia
Bngantine
110 20 30 40 60 60 70 60 80 96 99m«
mln 10 20 30 40 60 60 70 60 90 96 99mu
PareantJIa
Dolly Soda
110 20 30 40 60 60 70 60 90 96 99mnc
Pwoantto
Voyagaun
¦ 6ai
10 20 30 40 60 00 70 60 60 96 99mu
PareantJIa
PRB
Figure 3-90. Distribution of PM2.5 concentrations measured at IMPROVE sites in the East and
Midwest for 2004. Also shown are distributions of PM2.5 concentrations calculated by
CMAQ for the base case and for PRB.
July 2009
3-196
DRAFT-DO NOT CITE OR QUOTE

-------
Brldgar
Canyonlanda
Radwooda
00]
min 10 20 30 40 60 60 70 SO 90 96 99 max
ParoanUla
ParcanUla
Qlaelar
ParcanUla
mln 10 20 30 40 60 60 70 60 80 96 99 max
ParoanUla
mln 10 20 30 40 60 60 70 80 90 96 99 max
mln 10
20 30 40 60 60 70 80 90 96 99 max
mln 10 20 30 40 60 80 70 80 90 96 99 max
Paroantila
—*¦ Monitored —Baaa -*-PRB
Figure 3-91. Distribution of PM2.5 concentrations measured at IMPROVE sites in the West for 2004.
Also shown are distributions of PM2.5 concentrations calculated by CMAQ for the base
case and for PRB. Note the scale change on the y-axis for Redwoods RIP.
Table 3-20 gives the annual and quarterly average CMAQ predictions at IMPROVE
sites for the "base case" and the ratio of CMAQ predictions to the measured concentrations
at those sites in 2004. CMAQ performance for the annual average concentrations and most
of the seasonal averages is very good in the East and Midwest, generally falling within 35%
July 2009
3-197
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
In the West, CMAQ's prediction of PM2.5 mass averages at these remote sites is generally
too low in all seasons, often by 50%. Air quality model predictions in the mountainous West
are often not as good as those over the flatter terrain in the East and Midwest because the
model's grid spacing (36 km in this case) smoothes over significant variation at the surface
which results in differences at such remote sites. However, the model's trend relative to the
geospatial difference is correct: the predicted PM2.5 concentrations are lower at the western
sites than they are at the eastern sites, just as the measurements are. Table 3-21 shows
corresponding annual and quarterly mean PRB PM2.5 concentrations at IMPROVE sites.
Table 3-20.
Annual and quarterly mean PM2.5 concentrations (Dg/m3) for the CMAQ "
IMPROVE sites in 2004.
base case" at

Annual; mod/obs
Jan-March; mod/obs
Apr-Jun; mod/obs
Jul-Sep; mod/obs
Oct-Dec; mod/obs
EAST
Acadia
4.7; 1.04
5.6; 1.44
4.0; 0.87
4.6; 0.77
4.6; 1.31
Brigantine
10.2; 1.07
10.9;1.35
10.3; 0.91
10.2; 0.88
9.4; 1.29
Dolly Sods
9.8; 1.03
8.3; 1.24
8.6; 0.88
14.0; 0.90
8.0; 1.40
MIDWEST
Voyageurs
4.0; 1.05
4.9; 1.19
2.2; 0.71
3.9; 0.93
4.9; 1.36
WEST
Bridger
1.6; 0.76
1.3; 1.08
1.6; 0.52
1.8; 0.64
1.7; 1.30
Canyonlands
1.6; 0.62
1.9; 0.86
1.4; 0.44
1.5; 0.52
1.6; .76
Gila
1.6; 0.55
1.4; 0.70
2.2; 0.55
1.7; 0.45
1.1; 0.61
Glacier
2.2; 0.45
1.8; 0.39
2.1; 0.50
2.1; 0.40
2.8; 0.56
Redwood
4.6; 1.31
4.0; 1.48
3.0; 0.83
2.9; 0.78
8.4; 2.15
Table 3-22 illustrates CMAQ predictions of seasonal variation in the base case PM2.5
concentrations across regions of the CONUS, while Table 3-23 shows CMAQ predictions of
the seasonal variation in regional PRB PM2.5 concentrations. Highest base case PM2.5
concentrations were observed for the Northeast, Southeast, and Industrial Midwest with
highest concentrations during the fall and winter (and comparably high concentrations in
the summer for the Industrial Midwest). PRB PM2.5 concentrations were highest on an
annual basis in the Southeast, and peaking during the winter. In the summer, PRB PM2.5 is
roughly comparable for the Northwest, and Southern California and elevated but slightly
lower for the Southwest. These results also likely indicate the influence of sources that are
more strongly related to hot and dry conditions such as wildfires and dust suspension.
July 2009
3-198
DRAFT-DO NOT CITE OR QUOTE

-------
Table 3-21. Annual and quarterly mean PM2.5 concentrations (Dg/m3) for the CMAQ PRB simulations at
IMPROVE sites in 2004.

Annual

January-March
April-June

July-September
October-December
EAST
Acadia
0.70
0.76

0.76

0.65

0.65
Brigantine
0.77
0.86

0.91

0.70

0.63
Dolly Sods
0.79
0.88

0.83

0.75

0.66
MIDWEST
Voyageurs
1.2
0.83

0.91

2.0

0.93
WEST
Bridger
0.57
0.33

0.57

0.76

0.61
Canyonlands
0.49
0.38

0.54

0.68

0.35
Gila
0.74
0.42

1.4

0.80

0.32
Glacier
0.91
0.36

0.87

1.1

1.3
Redwood
2.8
2.4

1.5

1.1

6.1

Table 3-22.
Annual and quarterly mean of the CMAQ-predicted base case PM2.5
in the U.S. EPA CORIUS regions in 2004.
concentrations (Dg/m3)

Annual

January-March

April-June

July-September
October-December
Northeast
9.76

10.74

8.38

9.55
10.38
Southeast
10.05

12.28

7.72

9.78
10.42
Industrial Midwest
11.38

12.22

9.37

11.89
12.00
Upper Midwest
6.70

8.83

4.95

5.34
7.67
Southwest
3.30

4.08

2.77

3.31
3.03
Northwest
2.72

2.49

2.21

2.71
3.44
Southern California
4.43

4.64

3.93

4.34
4.82
Table 3-23. Annual and quarterly mean of the CMAQ-predicted PRB PM2.5 concentrations (Dg/m3) in the
U.S. EPA CORIUS regions in 2004.

Annual
January-March
April-June
July-September
October-December
Northeast
0.74
0.85
0.78
0.67
0.68
Southeast
1.72
2.43
1.41
1.41
1.64
Industrial Midwest
0.86
0.89
0.89
0.94
0.73
Upper Midwest
0.84
0.79
0.93
0.99
0.66
July 2009
3-199
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

Annual
January-March
April-June
July-September
October-December
Southwest
0.62
0.61
0.76
0.70
0.40
Northwest
1.01
0.48
0.81
1.42
1.32
Southern California
0.84
0.54
0.92
1.21
0.67
3.8. Issues in Exposure Assessment for PM and its
Components
The purpose of this section is to present the latest exposure assessment studies to
characterize the exposure of individuals and populations to PM of ambient origin. Such
information will aid the interpretation of epidemiologic studies described in subsequent
chapters of this Integrated Science Assessment. This section includes descriptions of
modeling and monitoring techniques used to capture personal PM exposure, observations
reported in the literature at various relevant spatial scales, observations related to PM
composition and PM in a mix of copollutants, and the effect of exposure estimates on
epidemiologic results. Attention is given to use of community-based monitors at urban
spatial scales and use of personal and microenvironmental exposure data to present how
each metric can be used in exposure assessment and what errors and uncertainties exist for
each approach. Understanding of exposure errors is important because exposure error can
potentially bias an estimate of a health effect endpoint, or increase the size of confidence
intervals around a health effect estimate. Typically, exposure error biases analyses of
health effects towards the null (i.e., no relationship between exposure and health effect).
The information presented in this section builds upon the key findings of the 2004 PM
AQCD (U.S. EPA, 2004, 056905). One key finding was that separation of total PM
exposures into ambient and nonambient components reduces potential uncertainties in the
analysis and interpretation of PM health effects data. At the time of the 2004 PM AQCD
(U.S. EPA, 2004, 056905), one study had reported that individual daily values of the total
and nonambient personal PM exposure were both poorly correlated with the daily ambient
PM concentrations while individual daily values of ambient PM exposure and daily ambient
PM concentrations were highly correlated. In pooled studies (different subjects measured on
different days), individual, daily values of the total PM exposure were generally shown to
be poorly correlated with the daily ambient PM concentrations. In longitudinal studies
July 2009
3-200
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
(each subject measured for multiple days), individual, daily values of the total PM personal
exposure and the daily ambient PM concentrations were found to be highly correlated for
some, but not all, subjects. Using the PTEAM study data, the 2004 PM AQCD (U.S. EPA,
2004, 056905) also analyzed exposure measurement errors in the context of time-series
epidemiology to show that the error introduced by using ambient PM concentrations as a
surrogate for ambient PM exposure negatively biases the estimation of health risk
coefficients by the ratio of ambient PM exposure to ambient PM concentration (called a, the
ambient exposure factor). However, it was concluded that the health risk coefficient
determined using ambient PM concentrations provides the correct information on the
change in health risks that would be produced by a change in ambient concentrations.
The material in this section is designed to help interpret findings from epidemiologic
studies presented in Chapters 6 and 7 of this ISA. The chapter is structured as follows. A
conceptual model of PM exposure is presented in Section 3.8.1, followed by exposure
modeling techniques in Section 3.8.2. Next, new developments in techniques for measuring
personal and indoor PM are presented. In Section 3.8.4, exposure assessment field studies
in the literature are presented. This section is divided into urban and neighborhood-to-
micro spatial scales of ambient exposure with attention to near-road, in-vehicle, and indoor
environments. Section 3.8.5 presents issues related to PM composition and PM in
multipollutant mixtures. The section culminates with implications of exposure assessment
issues for epidemiologic studies in Section 3.8.6.
3.8.1. General Exposure Concepts
A theoretical model of personal exposure is presented to highlight what is measurable
and what uncertainties exist in this framework. An individual's time-integrated total
exposure to airborne PM can be described based on a compartmentalization of the person's
activities throughout a given time period:
Iir = | Cjdt
Equation 3-3
where Et= total exposure over a time-period of interest, Q = airborne PM concentration at
microenvironment j, and (H, = portion of the time-period spent in microenvironment /.
Equation 3"3 can be decomposed into a microenvironmental model that accounts for
exposure to PM of ambient (Ea) and nonambient (Ena) origin of the form:
July 2009
3-201
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
Et = Ea+ Ena
1	a	na
Equation 3-4
Figure 3"92 illustrates Equation 3-4. Examples of ambient PM sources include
industrial and mobile source emissions, resuspended dust, biomass combustion, and
secondary formation. Examples of nonambient sources include smoking, cooking, home
heating, cleaning, and indoor air chemistry. PM concentrations generated by ambient and
nonambient sources are subject to spatial and temporal variability that can affect estimates
of exposure and resulting health effects. Exposure factors affecting interpretation of
epidemiology are discussed in detail in Section 3.8.6.
Ambient
Exposure
Ambient
PM, C white
Outdoors
Ambient
PM that
Infiltrates
Indoors
Total Personal
Exposure to FM
4^
REM
.•in1* 2L
Nonambient
Exposure
Indoor Sources
(cooking,
cleaning)
Personal Sources
(smoking,
hobby)
Source: Wilson and Brauer (2008).
Figure 3-92. Model of total personal exposure to PM as a function of ambient and nonambient
sources.
This assessment focuses on the ambient component of exposure because this is more
relevant to the NAAQS review. Ea can be expressed in terms of the fraction of time spent in
various outdoor and indoor microenvironments (Wallace et al., 2006, 089190; Wilson et al.,
2000, 010288):
July 2009
3-202
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Equation 3-5
where /= fraction of the relevant time period (equivalent to dt in Equation 3-2),
subscript o = index of outdoor microenvironments, subscript i- index of indoor
microenvironments, subscript o,i— index of outdoor microenvironments adjacent to a given
indoor microenvironment i, and Fmf,i = infiltration factor for indoor microenvironment /.
Equation 3-5 is subject to the constraint ~Lfo + 21£ = 1, and each term on the right hand side
of the equation has a summation because it reflects various microenvironmental exposures.
Here, "indoors" refers to being inside any aspect of the built environment, e.g., home, office
buildings, enclosed vehicles (automobiles, trains, buses), and/or recreational facilities
(movies, restaurants, bars). "Outdoor" exposure can occur in parks or yards, on sidewalks,
and on bicycles or motorcycles.
represents the equilibrium fraction of the PM concentration outside the
microenvironment that penetrates inside the microenvironment and remains suspended. It
is a function of the microenvironmental air exchange characteristics and the particle
properties. Assuming steady state ventilation conditions, the infiltration factor is a function
of the penetration, PCPis the ratio of indoor to outdoor PM), of PM, the air exchange rate,
a, of the indoor microenvironment, and the rate of PM loss, k, within the indoor
microenvironment: Fmf = I'nl(;&!<). Determination of Ea can be complicated by PM loss
through chemical and physical processes in microenvironments and the composition of PM
can be modified during infiltration of outdoor air into microenvironments (Sarnat et al.,
2006, 089166; Meng et al., 2007, 091197)
In the context of interpreting epidemiologic studies of the effects of ambient
pollutants on human health, the association between Ea and concentrations from a central
site monitor, Ca, is more relevant than the association between Et (which includes both the
ambient and nonambient component of PM exposure) and Ca because nonambient PM is
uncorrelated with Ca, as shown in Section 3.8.4. In ecologic studies of large panels or
cohorts, Ca is often used in lieu of outdoor microenvironmental data to represent these
exposures based on the availability of data. Thus it is often assumed that Co = and that
the fraction of time spent outdoors can be expressed cumulatively as fm the indoor terms
still retain a summation because infiltration differs among different microenvironments.
Under these assumptions, an individual's exposure to ambient PM, first given in Equation
July 2009
3-203
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
3-5, can be re-expressed as a function of ( The following approximation has been
employed in the literature to describe ambient exposure based on these assumptions
(Wallace et al., 2006, 089190; Wilson and Brauer, 2006, 088933; Wilson et al., 2000,
010288):
Equation 3-6
Particle size, particle composition, meteorology, urban and natural topography, and
other factors determine whether or not Equation 3"6 is a reasonable approximation for
Equation 3"5. Errors and uncertainties inherent in use of Equation 3-6 in lieu of Equation
3-5 are described in Section 3.8.6 with respect to implications for epidemiology. If
concentration measured at a central site monitor is used to represent ambient
concentration, then a, the ratio between personal exposure to ambient PM and the ambient
concentration of PM, can be defined as:
f
a = -2-
Equation 3-7
If the assumptions forming the basis for Equation 3"6 are valid, then a is the
proportionality factor in Equation 3"6:
« = /0+Z/^nf,!
Equation 3-8
a varies between 0 and 1. If a person's exposure occurs in a single microenvironment, the
ambient component of a microenvironmental PM concentration can be represented as the
product of the ambient concentration and Fmf. Wallace et al. (2006, 089190) note that time-
activity data and corresponding estimates of Fmt for each microenvironmental exposure are
needed to compute an individual's a with accuracy. If local sources and sinks exist and are
significant but not captured by central site monitors, then the ambient component of
outdoor air must be estimated using dispersion models, land use regression models,
receptor models, fine scale chemistry-transport models or some combination of these
techniques. These techniques are described in Section 3.8.2.
July 2009
3-204
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
3.8.2. Exposure Modeling
This section describes a variety of techniques used to model PM exposure. Many of
these methods are used in combination to link ambient PM levels in the atmosphere or
source characteristics to human exposure among individuals or sample populations. Recent
developments in exposure modeling are described in this subsection, and errors and
uncertainties of these approaches are described in the Section 3.8.6.2.
3.8.2.1. Time-Weighted Microenvironmental Models
An individual's exposure is dictated by his or her activity patterns, as modeled by £>
and fi in Equation 3"5. A number of panel studies have tracked subject exposures using
questionnaires, time-activity diaries, or global positioning systems (e.g.Cohen et al., 2009,
190639; Elgethun et al., 2003, 190640; Johnson et al., 2000, 001660; Olson and Burke,
2006, 189951; Wallace et al., 2006, 089190). In many cases, the time-activity tracking is
performed in conjunction with personal exposure and/or indoor and outdoor PM
concentration monitoring to estimate overall PM exposure. Wu (2005, 086397) described a
microenvironmental model of total personal exposure:
ET = f*Co + focP-a + ff,
Equation 3-9
where foh = fraction of time spent outdoors at home, = fraction of time spent
outdoors away from home, Co = PM concentration outside the home, and G = indoor PM
concentration. In Equation 3-9, AVcan be calculated based on time-activity diary data and
time-resolved PM concentration measurements, i^Vcan be expressed as a time-resolved
value or cumulatively over a time period of interest using this formulation. In this model,
ambient and nonambient exposure cannot be separated because it incorporates indoor
concentrations that are a function of both ambient and nonambient sources. Additionally,
Equation 3"9 distinguishes ambient concentration measured at a monitor from that
measured immediately outside the home. Liu et al. (2003, 073841) found that this model
predicted elderly exposures adequately but was a poor predictor of PM2.5 exposure for
asthmatic children. Wu et al. (2005, 086397) point out that this may be due to lack of
availability of the children's time-activity data in school where children spend a substantial
portion of their day. In a study of school children's exposure patterns, DeCastro et al. (2007,
July 2009
3-205
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
190996) computed odds ratios of a panel subject's location within a given microenvironment
using multivariate logistic models of the indoor school, indoor home, and outdoor
microenvironments and found that city of residence was a significant predictor of being
indoors at school, having an afterschool job was a significant predictor of being indoors at
home, and age and having an afterschool job were significant predictors of being outdoors.
The results of the DeCastro et al. (2007, 190996) study were designed to predict £ and £> in
exposure modeling.
A second approach proposed by Wu et al. (2005, 086397) is similar in formulation to
Equation 3"5 because it computes ambient PM exposure by considering the amount of
outdoor PM infiltrated indoors. This version also incorporates Co and Ca-
Ea = foh^o + focPa + infQ
Equation 3-10
Equation 3" 10 differs from 3-5 because it accounts for concentrations immediately
outside the building rather than considering all outdoor exposures to be a function of that
measured at a community monitor. Factors influencing the contribution of Ea to Et may
include sample population characteristics, location of a site for microenvironmental
monitoring, seasonal trends in PM concentration, and regional differences affecting
ambient concentration and infiltration.
Regression based approaches can also be incorporated into time-weighted
microenvironmental modeling. Chang et al. (2003, 053789) used data from the Scripted
Activity Study and the Older Adults Study, both conducted in Baltimore in 1998 and 1999,
to compute total personal exposure based on time-weighting microenvironmental exposures
for each panel subject:
Er =	+MEJ^ftCt
k	k
Equation 3-11
where ME= microenvironment (indoor or outdoor), 6= regression coefficient
reflecting the accuracy of the exposure estimate for a given microenvironment, fk = fraction
of time performing an activity k, and Ck = personal exposure while performing activity k. In
this work, Chang et al. (2003, 053789) tested the models with hourly personal exposure
data, hourly ambient concentration data, and daily ambient concentration data. The study
found that time-activity data improved estimates of 24-h PM2.5 exposure in comparison with
July 2009
3-206
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
using 24-h ambient PM2.5 data but use of hourly ambient data was comparable to personal
microenvironmental data in estimating exposure. When using ambient concentration data,
reflected infiltration for the indoor microenvironmental estimates for a sample
population. Wallace et al. (2006, 089190) also used a multivariate regression to assess the
impact of various factors on total and ambient PM2.5 personal exposure and found several
building- and activity-related variables were significant predictors of total and ambient
PM2.5 exposure. Using a similar activity-based exposure modeling approach for a panel
study in Seattle from 1999-2002, Allen et al. (2004, 190089) found that subjects' PM
exposures were best represented by modeling ambient and nonambient exposures
separately because ambient PM exposure was well correlated with PM concentration at
central site monitors while nonambient PM exposure was not.
3.8.2.2. Stochastic Population Exposure Models
Population-based methods, such as the Air Pollution Exposure (APEX), Stochastic
Human Exposure and Dose Simulation (SHEDS), and EXPOLIS (exposure in polis, or
cities) models, involve stochastic treatment of the model input factors
(http://www.epa.gov/ttn/fer a/human anex.html (Burke et al., 2001, 014050; Kruize et al.,
2003, 156661). These are described in detail in Annex 3.7 of the 2008 NOx ISA (U.S. EPA,
2008, 157073). Stochastic models utilize distributions of pollutant-related and
individual-level variables, such as ambient and local PM concentration source contributions
and breathing rate respectively, to compute the distribution of individual exposures across
the modeled population. Using distributions of input parameters in the model framework
rather than point estimates allows the models to explicitly incorporate uncertainty and
variability into exposure estimates (Zidek et al., 2007, 190076). These models estimate
time-weighted exposure for modeled individuals by summing exposure in each
microenvironment visited during the exposure period. The models also have the capability
to estimate received dose through a dosimetry model. The initial set of input data for
population exposure models is ambient air quality data, which may come from a monitoring
network or model estimates. Estimates of concentrations in a set of microenvironments are
generated either by mass balance methods or microenvironmental factors.
Microenvironments modeled include residential indoor microenvironments! other indoor
microenvironments, such as schools, offices, and public buildings! vehicles! and outdoor
July 2009
3-207
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
microenvironments. The sequence of microenvironments and exertion levels during the
exposure period is determined from characteristics of each modeled individual. The APEX
model does this by generating a profile for each simulated individual by sampling from
distributions of demographic variables such as age, gender, and employment; physiological
variables such as height and weight; and situational variables such as living in a house
with a gas stove or air conditioning. Activity patterns from a database such as CHAD are
assigned to the simulated individual using age, gender, and biometric characteristics.
Breathing rates are calculated for each activity based on exertion level, and the
corresponding received dose is then computed. For APEX, the PM dosimetry algorithm is
based on the International Commission on Radiological Protection's Human Respiratory
Tract Model for Radiological Protection (ICRP, 1994, 006988), and calculates the rate of
mass deposition of PM in the respiratory system (U.S. EPA, 2008, 191775). Summaries of
individual- and population-level metrics are produced, such as maximum exposure or dose,
number of individuals exceeding a specified exposure/dose threshold, and number of person-
days at or above certain exposure levels. The models also consider the non-ambient
contribution to total exposure. Nonambient source terms are added to the infiltration of
ambient pollutants to calculate the total concentration in the microenvironment. Output
from model runs with and without nonambient sources can be compared to estimate the
ambient contribution to total exposure and dose.
Recent larger-scale human activity databases, such as those developed for the
Consolidated Human Activity Database (CHAD) or the National Human Activity Pattern
Survey (NHAPS), have been designed to characterize exposure patterns among much larger
population subsets than can be examined during individual panel studies (Klepeis et al.,
2001, 002437; McCurdy et al., 2000, 000782). CHAD consists of a consolidation of human
activity data obtained during several panel studies in which diary or retrospective activity
data were obtained, while NHAPS acquired sample population time activity data through
surveys about human activity (Klepeis et al., 2001, 002437). The complex human activity
patterns across the population (all ages) are illustrated in Figure 3-93 (Klepeis et al., 2001,
002137). This figure is presented to illustrate the diversity of daily activities among the
entire population as well as the proportion of time spent in each microenvironment.
Different patterns would be anticipated when breaking down activity patterns for
subgroups such as children or the elderly. With data for average PM concentrations in each
July 2009
3-208
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
microenvironment, population exposures can be estimated from this break-down of time-
activity data.
100
Other Outdoor
Residence-Indoors
Residence-Outdoors
©
r-j
Near Vehicle
(Outdoors)
Bar/Restaurant
' ">
Other
Indoor
Mall/Store
School/
Public Bldg.
Q
In
a>
Oh
T Mill l~m^rnTTTTT-TJTTT

p
P3
C G K
ooooooooooooooooooooooooo
OOOOOOOOOOOOOOOOOOOOOOOOO
Time of Day
Source: Klepeis et al. (2001,002437).
Figure 3-93. Distribution of time sample population spends in various environments, from the
National Human Activity Pattern Survey.
Stochastic and deterministic methods are often combined, as described below.
Recently SHEDS has been linked with the Modeling Environment for Total Risk Studies
(MENTOR) model to expand population exposure assessment to individual risk assessment
(Georgopoulos et al., 2005, 080269). In this formulation, CMAQ was used to predict initial
concentrations at a coarse scale, and then a spatiotemporal random field method (Vyas and
Christakos, 1997, 156142) or Bayesian max entropy method (Serre and Christakos, 1999,
156968) was applied to interpolate the concentration to census tract scale in which
exposure estimates are made. CHAD can also be incorporated into MENTOR so that
estimates of exposure are related to dose and metabolic distributions to estimate risk of
specific health impacts.
July 2009
3-209
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
3.8.2.3. Dispersion Models
Dispersion models have been used both for direct estimation of exposure and as
inputs for stochastic modeling systems, as described above. Location-based exposures have
been predicted using a model such as California Line Source Dispersion Model (CALINE),
the American Meteorological Society/Environmental Protection Agency Regulatory Model
(AERMOD), CALPUFF (long-range plume transport model created by the California Air
Resources Board), or the Operational Street Pollution Model (OSPM) for estimation of
street-level PM pollution coupled with infiltration models to represent indoor exposure to
ambient levels (Gilliam et al., 2005, 056749; Hering, 2007, 155839; Mensink et al., 2008,
155980; Wilson and Zawar-Reza, 2006, 088292). For instance, CALPUFF was used to model
transport and dispersion in lower Manhattan following the September 11, 2001 World
Trade Center collapse (Gilliam et al., 2005, 056750) to determine average location-based
exposures. Wilson and Zawar-Reza (2006, 088292) used The Air Pollution Model (TAPM),
which integrated an emissions model with a mesoscale meteorological driver, to assess PMio
dispersion and potential for exposure in Christchurch, New Zealand. Gulliver and Briggs
(2005, 191079) used the Atmospheric Dispersion Modeling System (ADMS) to model
dispersion of "line-source" traffic emissions in an urban environment. In a method similar
to that employed by Georgopoulos et al. with SHEDS, Wu et al. (2005, 157155) used
CALINE to predict street-level concentrations of pollutants and input the results of that
dispersion model into an individual exposure model that accounts for infiltration of specific
building characteristics. Wu et al. (2005, 086397) employed CHAD to estimate the
time-basis of exposures from the CALINE predictions. With an individualized exposure
approach, the model is deterministic. However, population exposures can be estimated by
performing repeated simulations using various housing characteristics and then computing
a posterior probability distribution function for exposure. Isakov et al. (2009, 191192)
developed a methodology to link a chemical transport model, used to compute regional scale
spatiotemporally-varying concentration in an urban area, with stochastic population
exposure models to predict annual and seasonal variation in urban population exposure
within urban microenvironments.
July 2009
3-210
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
3.8.2.4. Land Use Regression (LUR) and GIS-Based Models
Land use regression (LUR) models have also been developed to describe pollution
levels as a function of source characteristics (Briggs et al., 1997, 025950; Gilliland et al.,
2005, 098820; Ryan and LeMasters, 2007, 156063). LUR develops a regression from
monitored concentration values as a function of data from a combination of factors such as
land use designation, traffic counts, home heating usage, point source strength, and
population density and then computes the regression at multiple locations based on the
independent variables at those particular locations without monitors. At the census tract
level, a LUR is a multivariate description of pollution as a function of traffic, land use, and
topographic variables (Briggs et al., 1997, 025950). Originally, LUR was used for NO2
dispersion, but it was adapted for PM2.5 exposure estimation by Brauer et al. (2003, 155702)
for Stockholm, Sweden, Munich, Germany, and throughout The Netherlands. This study
found a measure of traffic density to be the most significant variable predicting PM2.5
exposure. Ryan et al. (2008, 156064) reported on a LUR model for childhood exposure to
traffic-derived EC for the Cincinnati Allergy and Air Pollution Study and also found traffic
to be the most important determinant of diesel exhaust particle exposure. In this case, wind
direction was also factored into the model as a determinant of EC mixing. Like
deterministic dispersion models, LUR can be performed over wide areas to develop a
posterior probability distribution function of exposure at the urban scale. However, Hoek et
al. (2008, 156554) warn of several limitations of LUR, including distinguishing real
associations between pollutants and covariates from those of correlated copollutants,
limitations in spatial resolution from monitor data, applicability of the LUR model under
changing temporal conditions, and introduction of confounding factors when LUR is used in
epidemiologic studies.
A GIS platform is typically used to organize the independent variable data and map
the results. The GIS software creates numerous lattice points for the regression of
concentration as a function of the covariates. For instance, Krewski et al. (2009, 191193)
computed PM2.5 concentrations for the New York City and Los Angeles metropolitan areas.
For the Los Angeles analysis, the LUR was applied at 18,000 points in the simulation
domain, and an inverse distance weighting kriging method was applied to interpolate the
predicted concentration. In New York City, the LUR was applied at 49 monitors for a 3-yr
model and 36 monitors for a model of winter 2000; kriging was employed only for the
July 2009
3-211
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
purpose of visualizing the concentration between monitors. The models explained 69% and
66% of the variation in PM2.5 in Los Angeles and New York City, respectively.
GIS-based spatial smoothing models can be used to estimate PM concentration levels
where monitors are not located. Yanosky et al. (2008, 099467) described an approach to
estimate concentrations using a combination of reported AQS data and GIS-based and
meteorological covariates. Temporally stationary covariates included distance to nearest
road for different PM size fractions, urban land use, population density, point source
emissions within 1 and 10 km buffers, and elevation above sea level. Time-varying
covariates included area source emissions, precipitation, and wind speed. In this analysis,
the GIS-based covariates were temporally stationary, while the meteorological and PM
monitored concentration inputs were time-varying. This approach was applied to estimate
PM10, PM10-2.5, and PM2.5 exposures for the Nurse's Health Study and provided estimates of
concentration at approximately 70,000 nodes with PM10 and/or PM2.5 data input from more
than 900 AQS sites with good validation of the PM10 and PM2.5 models (Paciorek et al.,
2008, 190090; Yanosky et al., 2008, 099467; Yanosky et al., 2009, 190111).
GIS-based methods can also be applied to integrate exposures over different
microenvironments. For example, Gulliver and Briggs (2005, 191079) described
development of the Space-Time Exposure Modeling System (STEMS) to model PM10
concentration. STEMS is a multipronged model that links traffic emissions estimates,
dispersion, background PM estimates, and time-activity data within a GIS framework to
create exposure estimates. Traffic emissions estimates and meteorological parameters are
input into the ADMS dispersion model, which along with background PM measurements
are used to create hourly point estimates of PM concentration. Based on the time-activity
data, the concentration estimates were then used to calculate exposures to traffic-related
PM10 along a commuting path while an individual is in transit. PM10 was used by Gulliver
and Briggs (2005, 191079) because ADMS had not yet been validated for smaller size
fractions. In an analysis of the sensitivity of included variables, Gulliver and Briggs (2005,
191079) showed the model to be most sensitive to fluctuations in local meteorology followed
by sudden speed reductions of 10 km/h. The STEMS model was primarily designed to model
exposures during transit, but the authors state that this technique can be applied to
modeling other microenvironmental exposures.
Source proximity is sometimes used as a covariate in GIS-based regression models.
For instance, Baxter et al. (2007, 092725) predicted indoor exposure to PM2.5, EC, and NOx
July 2009
3-212
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
based on distance to traffic sources, indoor source characteristics, window opening, and
ambient concentrations in the Boston metropolitan area. In this effort, Baxter et al.
examined a variety of factors estimated using GIS including roadway density, roadway
length, average daily traffic, and population density to determine which variables were
significant predictors. They found that point estimates of PM2.5 were largely influenced by
regional ambient PM2.5 while EC estimates were more influenced by local traffic sources.
However, Baxter et al. (2008, 1!) 11 91) found no association between distance to a bridge toll
booth station and indoor EC concentration in Detroit homes when studying the impact of
diesel emissions from traffic on the Ambassador Bridge as part of the Detroit Exposure and
Aerosol Research Study (DEARS). Being located downwind of the booth, however, was a
significant predictor of indoor EC concentration. Corburn (2007, 155738) tested two distinct
modeling approaches, the cumulative air toxics surface (CATS) and the U.S. EPA's National
Air Toxics Assessment (NATA) to determine how these approaches can yield estimates of
human exposure to diesel exhaust and 33 air toxics for environmental impact assessment.
The CATS approach included an exposure term incorporating source density and distance
to source, and the sources could include traffic as well as bus depots and transfer stations,
airports, and industrial point sources. Corburn's paper demonstrated that robust land use
data can provide an approximation for urban exposures, although he cautioned that such
estimates should not supersede environmental monitoring. In using these approaches,
Huang and Batterman (2000, 156572) warn that geographic divisions must be sufficiently
small to avoid inter-zone variability in source and exposure characteristics. Moreover, the
HEI Report on Traffic Related Health Effects (2009, 1!) 100!)) discourages use of source
proximity as a surrogate for traffic exposure in epidemiologic studies because it is not
specific to particular pollutants and can be subject to confounding factors such as
socioeconomic status.
3.8.3. Personal and Microenvironmental Exposure Monitoring
The purpose of this section is to present new discoveries related to measuring aerosol
concentration. A review of over 200 personal and microenvironmental PM exposure studies
published since 2002 (see Table A-52 of Annex A) reveals that the majority of the
monitoring techniques in use were previously reviewed in the 2004 PM AQCD (U.S. EPA,
2004, 056905) for PM. Detailed descriptions of these methodologies are provided in the 2004
July 2009
3-213
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
PM AQCD and therefore will not be repeated in this document. The following sections will
include only findings from 2002 or later regarding monitoring and modeling methodologies
in common use and significant advancements in understanding the capabilities and
limitations of these methods for assessment of PM exposure.
3.8.3.1. New Developments in Personal Exposure Monitoring Techniques
Current personal exposure sampling methodology consists largely of integrated filter
sampling using a cyclone or Personal Exposure Monitor (PEM) to achieve a cut-point at a
desired particle size (Hopke et al., 2003, 095544; Larson et al., 2004, 098145). This method
of sampling facilitates speciation work because the filters can be archived for chemical and
gravimetric analysis. Additionally, light scattering aerosol detection instruments, such as
the personal DataRam (pDR) and SidePak personal aerosol monitor have seen some use in
personal PMio, PM2.5, and PMi monitoring (e.g.Lewne et al., 2006, 090556; Wallace et al.,
2006, 088211). Several researchers have noted that relative humidity causes overestimation
of the estimated particle mass in light-scattering personal exposure monitors
(e.g.Lowenthal et al., 1995, 045134; Ramachandran et al., 2003, 191197); a correction factor
has been applied to concentrations to address this issue. In the DEARS, Williams et al.
(2008, 191198) attempted to reduce humidity of the sample stream by placing a drying
column upstream of the pDR's detector. Although this addressed the humidity issue,
Williams et al. (2008, 191198) also found that the drying column occasionally released
particles and therefore caused artificial concentration peaks. For this reason, Williams et
al. (2008, 191198) determined that the humidity correction approach is preferable.
One area of further development is in personal sampling of the thoracic and
respirable particle size distribution. Variations of the cascade impactor have been developed
for personal sampling and tested for use in field studies (Case et al., 2008, 155149; Lee et
al., 2006, 098249; Singh et al., 2003, 156088). The model developed and tested by Lee et al.
(2006, 098249) operates with a 1 pm cut point and therefore can characterize respirable
particles well. Case et al. (2008, 155149) two-stage cascade impactor separated PM10-2.5 from
PM2.5 for personal monitoring and sampled within ± 20% of a reference method. Hsiao et al.
(2009, 191001) developed a mini-cyclone with a 1 pm or 0.3 pm cutpoint for sampling
accumulation mode and ultrafine PM. The Personal Cascade Impactor Sampler (PCIS) has
the capability to sample down to a cutpoint of 250 nm (Singh et al., 2003, 156088). For
July 2009
3-214
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
PM2.5, the difference between the PCIS and MOUDI cascade impactor was 11%, while
difference between the PCIS and SMPS-APS was only 2%. Difference in PM2.5 species
compared with the MOUDI was generally higher: 11% for SO4222% for NO3", 19% for EC,
and 94% for OC. Mass was overestimated by 3%, 16%, and 31% for PM1-0.5, PM0.5-0.25, and
PMo .25, respectively, when compared with the SIVtPS'APS. Similarly, Case et al. (2008,
155149) found a difference ranging from -11 to +10% for PM10-2.5 with the Personal
Respirable Particulate Sampler (PRPS), and Lee et al. (2006, 098249) found a difference of
¦6 to 0% for PM2.5 and "6 to -1% for PM10 when comparing results from this device with
those from the PEM. Leith et al. (2007, 098241) redesigned the Wagner-Leith passive
sampler for measuring PM10-2.5. In this work,
Leith et al. (2007, 098241) tested the passive PM10-2.5 sampler and found the
difference between a PM10-2.5 FRM and the co-located passive sampler was within 1
standard deviation of concentrations measured by the FRM samplers.
A number of personal PM monitors are under development as part of the National
Institutes of Health Genes, Environment, and Health Initiative
(http://www.gei.nih.gov/exposurebiology/ program/sensor.asp). Funded projects for
miniature personal monitors include a platform that records real-time BC and PM
concentrations and archives PM for further analysis, a badge containing a sensor array that
detects several compounds found in diesel PM, a micro-nephelometer recording PM and
endotoxin exposure, a complimentary metal-oxide-semiconductor (CMOS) fitting in the nose
to measure allergen PM, and a micro-thermofluidic nanoparticle sensor. The mini-cyclone
cited above was designed to operate upwind of the micro-thermofluidic sensor (Hsiao et al.,
2009, 191001). LeVine et al. (1999, 156687) and Schwartz et al. (2008, 156963) described
use of the CMOS technology for real-time DNA detection. Mulchandani et al. (2001, 191003)
reviewed amperometric biosensors used for organophosphate pesticide detection that are
the basis of the diesel detection badge! several articles have been published by that group
before and after Mulchandani et al's (2001, 191003) review to describe applications of
amperometric sensors.
3.8.3.2. New Developments in Microenvironmental Exposure Monitoring Techniques
The majority of developments since the 2004 PM AQCD (U.S. EPA, 2004, 056905)
regarding microenvironmental PM characterization have involved real-time
July 2009
3-215
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
instrumentation in the ultrafine PM size range. Because these methods are also used for
ambient sampling, they are described in Section 3.4.
New developments in microenvironmental sampling for exposure assessment have
also included construction, testing, and implementation of mobile environmental sampling
laboratories for PM mass, particle count, and composition, as well as other criteria
pollutants (CO, SO2, NO2, O3). These mobile laboratories typically contain a suite of
real-time equipment with short sampling intervals (e.g., 1-10 minutes), such as an SMPS
with CPC, APS, laser photometers, and aethalometers for aerosols! monitors for the gaseous
criteria pollutants! weather station for meteorological variables, and a Global Positioning
System (GPS) for position. Videotape or journal observations are sometimes logged
simultaneously to track local on-road sources of pollution in this way. One key application
of mobile laboratory technology is assessment of the outdoor microscale environments and
in-vehicle microenvironments on roadways for determining exposure during on-road
transportation (Pirjola et al., 2004, 117564! Weijers et al., 2004, 104186! Westerdahl et al.,
2005, 086502! Sabin et al., 2005, 087728). For instance, Sabin et al. (2005, 087728) used
videotape records to determine whether BC detected on a school bus was the result of local
outdoor sources from other vehicles or "self-pollution" from the school bus's own engine
exhaust. Westerdahl et al. (2005, 086502) used the GPS time series to determine that
minima in the ultrafine PM time series corresponded to passage through residential areas
of Long Beach and Pasadena, in contrast to the pollution spikes observed along highways.
Studies have also shown that detection of PM from vehicle exhaust could be improved
through use of combined measurement results to improve statistical analysis (Ntziachristos
and Samaras, 2006, 116722).
3.8.4. Exposure Assessment Studies at Different Spatial
Scales
A number of exposure studies have been published since 2002. Table A-55 in Annex A
lists those studies performed in the U.S. by region of the country with personal,
microenvironmental, and ambient mass concentrations presented (note that chemical
speciation data where available are presented and discussed below). The majority of urban-
scale studies focus on PM2.5 because PM2.5 concentrations are more homogeneously
distributed. Studies of microscale to neighborhood scale dispersion more commonly include
July 2009
3-216
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
data on ultrafine and thoracic coarse PM in addition to PM2.5 because they travel over
shorter distances from the site of generation, as described in Section 3.5. Some of these
studies present the outdoor concentration measured outside the test building, while others
use ambient concentration obtained from a community site monitor. As would be expected,
there is considerable variability within and across regions of the country with respect to
indoor exposures and ambient concentrations. Furthermore, some regions are represented
by only one or two studies, while other regions have many studies! in most regions, studies
have been conducted in only one or two metropolitan areas. Thus, the results presented
may not be broadly representative.
Results of these studies highlight the uncertainties surrounding various estimates of
the ambient contribution to personal exposure. This variation can be attributed to a
number of factors, including PM size distribution, scope and magnitude of
microenvironmental sources, proximity to microenvironmental sources, ambient
concentrations of PM, percentages of time spent in various microenvironments, the age and
condition of indoor microenvironments, natural and urban topography, and outdoor
meteorology. Errors in exposure estimation are linked to the spatial scale of concentration
measurements because pollutant transport and dispersion varies over different spatial
scales as a function of the many factors mentioned in the previous sentence. Findings
related to identifying the ambient components of personal exposure and modes of PM
infiltration indoors are discussed in the subsequent subsections with respect to urban,
neighborhood, and micro spatial scales.
3.8.4.1. Urban Scale Ambient PM Exposure
The following paragraphs describe assessment of personal exposure to ambient PM at
the urban scale. As shown in Figure 3"93, the large majority of an individual's time is spent
indoors. Although calculation of infiltration and indoor personal exposure is an important
part of this assessment, such exposures are described with respect to central site monitors.
Assessing population-level exposure at the urban scale is particularly relevant for
epidemiologic studies, which typically provide information on the relationship between
health effects and community average exposure, rather than individual exposure.
In the context of determining the effects of ambient pollutants on human health, the
association between Ea and Ca is more relevant than the association between Et (which
July 2009
3-217
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
includes both the ambient and nonambient component of PM exposure) and Ca. If there are
no indoor or other nonambient sources of a pollutant, the total personal exposure is equal to
the ambient personal exposure. However, indoor or other nonambient sources could
significantly affect a person's total exposure to many pollutants. Wilson and Brauer (2006,
088933) noted that exposure to nonambient PM will not affect the relationship between Ca
and Ea, but "the difference between ambient concentration and ambient exposure will bias
the relative risk derived from epidemiologic studies."
Some studies have used PM components to estimate the infiltration of ambient PM to
indoor environments. Wilson et al. (2000, 010288) first proposed that SO42- could be used as
a tracer of the ambient PM2.5 infiltration rate. Sarnat et al. (2002, 037056) also noted that it
is reasonable to assume that the size distribution of ambient SO r particles is sufficiently
similar to the size distribution of ambient PM2.5, and therefore that the ambient SO r to
personal S( ) r ratio is an acceptable surrogate for the ratio of the ambient PM2.5 exposure
to the ambient PM2.5 concentration. Sulfate has been used this way in several studies,
including Ebelt et al. (2005, 056907), Wallace and Williams (2005, 057485; 2006, 089190)
and Wilson and Brauer (2006, 088933). For this method to be successful, indoor or other
nonambient sources of the tracer must be small compared to ambient sources over the
period of sampling. Wilson and Brauer (2006, 088933) noted that environmental tobacco
smoke and tap water used in showers or humidifiers are indoor sources of SO42Other
concerns in using SO42- as a tracer for PM2.5 arise because SO42- tends to be concentrated in
smaller particles and thus it might be a better tracer for fine mode particles than for total
PM2.5, which can include larger particles in the tail end of the coarse mode (Wallace and
Williams, 2005, 057485). Strand et al. (2006, 157017) suggested that Fe be used as an
additional tracer to correct for the infiltration of larger PM2.5 particles. Their study took
place in Denver, where indoor sources of Fe were small. However, there could be more
substantial contributions from tracking iron in soil indoors in other locations. The spatial
variability of Fe is also larger than that of PM2.5 across urban areas. Volatilization of nitrate
or organic compounds after infiltration of PM2.5 indoors could lead to bias in exposure
estimates (Lunden et al., 2003, 081201). This could be a large problem in areas in which
PM contains a large semi-volatile component.
Figure 3-94 shows total exposure to SO42- as a function of measured ambient SO42-
concentration. Figure .'>"95 shows estimated ambient exposure to PM2.5 as a function of
measured ambient PM2.5 concentration, where ambient personal exposure is calculated
July 2009
3-218
DRAFT-DO NOT CITE OR QUOTE

-------
from the ambient exposure factor for SO42Close agreement between these figures can be
observed. Figure 3-96 shows total exposure to PM2.5 as a function of measured ambient
PM2.5 concentration. However, the total exposure to PM2.5 shows virtually no association
with ambient PM2.5 because it contains nonambient contributions to PM2.5.
«¦ 0.74 C«>«- 0.01
R2 - 0..75
1	2 3 4 5 6
Ambient Concentration, (pgAm3)
Source: Wilson and Brauer (2006, 088933)
Figure 3-94. Total exposure to S042~ as a function of measured ambient S042~ concentration , from
the Vancouver study (Vancouver British Columbia, April-September 1998, with 16 non-
smoking subjects aged 54-86).
A" =0.76 C5 0 093
•n 20
m - 0.62
5 10 15 20 25 30
Ambient Concentration, Czs (pg/nr*)
Source: Wilson and Brauer (2006, 088933)
Figure 3-95. Estimated ambient exposure to PM2.5 as a function of measured ambient PM2.5
concentration , from the Vancouver study (Vancouver British Columbia, April-September
1998, with 16 non-smoking subjects aged 54-86).
July 2009
3-219
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Bt F"s = G,T7C« + 8,24
R2.5 = 007
«
0 5 10 15 20 26 30
Ambient Concentration, Cf5 (ijg/m3)
Source: Wilson and Brauer (2006, 088933)
Figure 3-96. Total exposure to PM2.5 as a function of measured ambient PM2.5 concentration.
The estimated ambient exposure to PM2.5 is well correlated with measured ambient
PM2.5 concentration with zero intercept implying that nonambient sources were minor. This
technique works well in areas where sulfate is a regional pollutant insuring that its spatial
variations are small (Kim et al., 2005, 156640; U.S. EPA, 2004, 056905). Wilson and Brauer
(2006, 088933) reported that the pooled Pearson correlation coefficient was 0.79 for
personal ambient exposures (estimated by tracer element method) versus ambient
concentrations, and was 0.001 for personal nonambient exposure vs. ambient
concentrations. Strand et al. (2007, 157018) conducted an exposure study in Denver (from
2002 to 2004) for 6-12-year-old school children. Up to 10 personal exposure samples were
collected on each day, and ambient concentrations were measured simultaneously at a fixed
site located at the school. The daily average personal sulfate exposure was strongly
associated with ambient sulfate concentration (r = 0.96, 120 >N >100). Koutrakis et al.
(2005, 095800) reported the median Spearman correlation coefficients between personal
sulfate exposure and ambient sulfate concentration were above 0.60 during both winter and
summer in Boston and Baltimore (15 subjects with 12 consecutive measurements during
each season in both Boston and Baltimore). For another Baltimore cohort (15 senior
subjects with up to 23 consecutive measurements for each person), Hopke et al. (2003,
095544) reported that the median Pearson correlation coefficient between personal
exposure to sulfate factor and ambient sulfate factor was 0.91 (ranging from 0.56 to 0.95 for
different subjects), while the median Pearson correlation coefficients were 0.34 for the
crustal factor (ranging from -0.05 to 0.62), and 0.31 for a factor whose origin was not
July 2009
3-220
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
identified (ranging from -0.01 to 0.88), respectively. The inferences drawn from using the
S042- component of PM2.5 as an indicator for personal exposure to ambient PM2.5 may apply
in areas where SOr is a minor component of PM2.5 or in the absence of significant
nonambient sources of SC>42~ (Sarnat et al., 2001, 019401).
Source apportionment techniques could also be used, in principle, to derive ambient
personal PM2.5 concentrations. They would be especially useful in areas where the
application of a tracer method might be problematic. Strand et al. (2007, 157018) noted that
the four outdoor factors (nitrate-sulfate, sulfate, OC, motor vehicle exhaust) would
constitute an estimate of the personal ambient PM2.5 concentration. However, the data used
in this portion of the analysis were obtained only with fixed monitors and did not include
measurements made by PEMs. They also used the Multilinear Engine (ME) to derive
factors that were required to contribute jointly to central indoor and outdoor, individual
apartment and individual PEM samples of a panel of residents. Hopke et al. (2003, 095544)
used PMF to derive source contributions to community, outdoor, and indoor PM exposures
at a retirement facility in Towson, MD. Hopke et al. (2003, 095544) found three sources^
sulfate, unknown (perhaps combustion related, according to the authors) and soil, jointly
contributing 46%, 13% and 4% of PM2.5 to the PEM samples. Further source resolution was
not possible because of a lack of data for a number of components in the PEM samples. The
largest and most clearly identified contribution to personal exposure was from the sulfate
factor. This study also determined that a few minor indoor and personal activity sources
contributed <10% of the ambient sulfate source to personal exposures.
Wilson and Brauer (2006, 088933) presented an adaptation of the SO42- method for
estimating exposure to PM10-2.5. a is computed based on the S( ) r method as the ratio of
exposure to ambient SO42-, as measured by a personal monitor, to ambient SO r
concentration. Then, knowing an individual subjects' time diary and the penetration and
loss properties of SO42", the air exchange rate for an individual location can be calculated
from Equations 3-5 and 3"7. Finally, the penetration and loss rates of PM10-2.5 from the
PTEAM database (Ozkaynak et al., 1996, 073986) can be input into the individual exposure
model along with the individual activity pattern and residential air exchange rate to
compute the ambient PM10-2.5 exposure factor and ambient PM10-2.5 exposure knowing PM10-
2.5 concentration. Given that PM10-2.5 deposits more readily and therefore disperses over a
shorter distance than PM2.5, it is possible that use of ambient PM10-2.5 concentration may
incur more error than in using this method for PM2.5. Ebelt et al. (2005, 056907) observed in
July 2009
3-221
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
a Vancouver, Canada panel study that ambient PM10-2.5 exposure was less correlated with
ambient PM10 exposure (r = 0.72) than was ambient PM2.5 (r = 0.92) exposure. In this study,
PM10-2.5 mass concentration was calculated from the difference between ambient PM10 and
PM2.5 mass concentration. This is attributed to both a smaller Amr for PM10-2.5 and PM2.5
comprising a greater fraction of the PM10 for the Vancouver study.
Wilson and Brauer (2006, 088933) state that their methodology for computing the
ambient exposure factor based on the PM2.5 sulfate method can be applied to PM in the
0.1-0.5 |um size range. Little SO42- mass is found below 0.1 |um, so the sulfate tracer method
would not be applicable for ultrafine PM. Given the short atmospheric lifetime of ultrafine
PM resulting from particle growth and evaporation processes, primary ultrafine PM is most
prevalent at microscale rather than urban scale (Sioutas et al., 2005, 088428). Moore et al.
(2009, 191004) found substantial spatial, hourly, and daily variability in ultrafine PM
concentration in a saturation study of Los Angeles. Moore et al. (2009, 191001) and
Harrison and Jones (2005, 191005) also found that ultrafine PM and PM2.5 measurements
were poorly correlated at the monitoring sites.
3.8.4.2. Micro-to-Neighborhood Scale Ambient PM Exposure
Near-Road Exposures
Sections 3.3 and 3.5 describe the physical and chemical composition of traffic
emissions as well as evolution of the plume away from roads. Table 3-24 contains data from
recent studies comparing outdoor personal exposure to fixed site monitors. Only studies
where samples were obtained outdoors and compared with a community-based ambient
monitoring site were included because indoor microenvironments have penetration losses
that impact the comparability of the results. Note that some of these studies included
enclosed transportation microenvironments (e.g., cars, buses, subways), but all studies
examined personal exposure in the outdoor microscale environment. Also note that studies
must be reviewed cautiously because most used different instrumentation for personal,
microenvironmental, and ambient measurements so that measurement artifacts related to
each instrument may differ. The Violante et al. (2006, 156140) study showed that outdoor
personal exposure to PM10 was significantly higher than fixed community-based ambient
monitoring site measurements in downtown Bologna, Italy. Likewise, the Kaur et al. (2005,
088175), Kaur et al. (2005, 086504), and Adams et al. (2001, 019350) studies showed PM2.5
July 2009
3-222
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
measurements to be significantly higher than fixed community-based ambient monitoring
site measurements in central London, U.K. However, Kinney et al. (2000, 001774) showed
that on-street PM2.5 concentrations were not significantly different from ambient PM2.5
measurements in a study of PM2.5 exposure in New York City! this is more consistent with
the urban-scale homogeneity in concentration of PM2.5.
Morwaska et al. (2008, 191006) stated that ultrafine PM number concentrations in
the near-road environment were roughly 18 times higher than in a non-urban background
environment, while measured concentrations in street canyons and tunnels were 27 and 64
times higher, respectively than background. This suggests that trapping of sources in a
semi-enclosed environment can lead to higher ultrafine PM exposures. Additionally, fresh
emission of short-lived ultrafine PM would explain substantially higher concentrations near
the site of emission. By sampling ultrafine PM count at multiple sites in Los Angeles,
Moore et al. (2009, 191004) demonstrated five-to-seven-fold difference between
concentrations measured directly next to a freeway and an oceanside site during morning
rush hour with substantial variability among sites throughout the day. When comparing
sampling campaign data on clear weather and rainy days next to the I "710 freeway in Los
Angeles, Ntziachristos et al. (2007, 089164) found that particle number concentration
obtained with an CPC was 2.4 times higher on clear weather than when raining! particle
surface area was 3.7 times higher in clear weather! and, black carbon concentration was 1.7
times higher in clear weather. However, SMPS data reported for rainy day particle number
concentrations were almost 29 times higher in this study. Likewise, Zhou and Levy (2007,
098633) noted in a meta-analysis of near-road studies that the zone of influence of the near-
road environment was generally within 300-400 m for EC and ultrafine PM counts. Kinney
et al. (2000, 001774) showed EC to increase linearly with increasing traffic counts with
large spatial variations where two sites had concentrations significantly higher than
ambient measurements. These observations suggest caution should be taken regarding the
representativeness of community averaged monitoring data for assessing exposures.
Particle chemistry is also an important consideration, because exposure may differ for
PM components. Farmer et al. (2003, 156431) found that exposure to particle-bound PAHs
including benzo[a]pyrene can be 2-3 times higher among those routinely exposed to outdoor
traffic emissions (e.g., police, bus drivers) compared with control subjects. Kinsey et al.
(2007, 190073) estimated from continuous idling and restart school bus operating
conditions (without retrofitting) that over a 10-min period of waiting at a bus stop,
July 2009
3-223
DRAFT-DO NOT CITE OR QUOTE

-------
continuous idling resulted in exposure to 33% more particle "bound PAH than in the case
where the bus was restarted 2 min into the simulation and idled for 8 min. Continuous
idling produced 3380% more particle-bound PAH than the case where the bus was off for 10
min then restarted and left at the end of the simulation.
Table 3-24. Examples of studies comparing near-road personal exposures with fixed site ambient
concentrations.
Reference
Ambient
monitors
Personal monitors
Microenvironment,
other variables
Ambient v. Personal
Association
Primary Findings
Violante et al.
(2006,
156140)
Bologna, Italy
Fixed PMio and
benzene monitoring
station (method not
specified).
Active pump with PMio
PEM, passive sample for
benzene desorbed and
analyzed by GC-MS.
Localized traffic density
(vehiclesfh);
Meteorology (wind speed,
wind direction, visibility,
relative humidity).
Personal: 185.10 ± 38.52/yg/m3
Fixed: 43.56 ± 24.10/yg/m3
(p< 0.0001); small but significant
correlation observed (R2 - 0.19,
p - 0.035) but disappeared after
outlier removal (R2 - 0.09, p - 0.165).
Fixed PMio correlated with
multivariate model of traffic and
meteorology but not personal
PMio; relationship between
benzene and PMio not explored.
Kaur et al.
(2005,
086504)
London, U.K.
Fixed TE0M for
PM2.6 and fixed CO
monitor at ambient
and curbside sites.
High flow personal
samplers for PM2.6, P-Trak
monitors for UFP, Langan
T15 and T15v for CO.
Exposures stratified by mode
of transport (walk, cycle,
bus, car, taxi).
Average PM2.6 at TE0MS was 3 times
lower than average personal PM2.6
sample, and 8 times lower than max
personal PM2.6 sample.
PM2.E exposures during walking
significantly lower than during
car and taxi rides, UFP
exposures during walking
significantly lower than bus and
car rides, cycling exposures to
PM2.6 and UFP not significantly
different from those on bus, car,
or taxi.
Kaur et al.
(2005,
088175)
London, U.K.
Fixed TE0M for
PM2.6 and fixed CO
monitor at ambient
and curbside sites.
High flow personal
samplers for PM2.6
analyzed post-sample for
reflectance for EC, P-Trak
monitors for UFP, Langan
T15 and T15v for CO.
Volunteers walking at set
times and directions along
Marylborne Rd in London.
Fixed vs. personal PM2.6: slope - 0.29,
R - 0.6; personal PM2.6 measurements
were >2 times background levels and
more than 15/yg/m3 greater than
curbside measurements.
Pedestrian exposures were
significantly higher than fixed
site curbside (or ambient
measurements. Results indicate
that exposure decline up to 10%
from curb-side to building edge
within a street canyon.
Adams et al.
(2001,
019350)
London, U.K.
Fixed TE0M for
PM2.6 and fixed CO
monitor at ambient
and curbside sites.
High flow personal
samplers for PM2.6.
Exposures stratified by mode
of transport (cycle, bus, car,
subway).
Median values: (//g/m3i

Summer Winter
Cycle
34.5 23.5

Bus
39.0 38.9

Car
37.7 33.7

Subway
247.2 157.3
Fixed 15
13

Curb
24 37

Mean values: (/yg/m3i


PM2.6
EC
Site 1
45.7 (10.1)
6.2(1.9)
Site 2
47.1 (16.4)
3.7 (0.6)
Site 3
36.6(10.8)
2.3 (0.9)
Ambient
38.7 (10.9)
1.5(0.5)
Exposures were 2.3-16.5 times
higher than ambient and
1.4-10.3 times higher than
curbside during summer. During
winter, only subway exposures
were appreciably higher (4.3
times) than curbside.
Kinney et al.
(2000,
001774)
New York
City, NY
(Harlem)
Ambient site filter	Three high traffic sites
in greased impactor	filter in greased impactor
with pump;	with pump; absorbance
absorbance testing	testing on filter for EC.
on filter for EC.
Traffic counts per h.
PM2.6 at high traffic sites was
not significantly higher than
ambient; EC was significantly
higher than ambient at 2 sites.
EC increased linearly with traffic
counts.
In-Vehicle and In-Transit Exposures
In-vehicle pollution has been identified in various studies as a source of exposure to
PMio, PM2.5, and ultrafine PM (Briggs et al., 2008, 156294; Diapouli et al., 2007, 156397;
Fruin et al., 2008, 097183; Gomez-Perales et al., 2004, 054418; Gomez-Perales et al., 2007,
July 2009
3-224
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
138816; Gulliver and Briggs, 2004, 053238; Gulliver and Briggs, 2007, 155814; Rossner et
al., 2008, 156927; Sabin et al., 2005, 088300). Results from recent studies are provided in
Table A-54 of Annex A. In many of these studies, in-vehicle exposures are shown to be
comparable or less than that of walkers on the same route. Typically, in-vehicle exposures
were still higher than community-based ambient monitor concentrations for TSP and PMio
(Diapouli et al., 2008, 190119). Curbside measurements of ultrafine PM and PM2.5 obtained
at a fixed site in the Kaur et al. (2005, 088175; 2005, 086504) studies were generally lower
than exposures during transit (including during walking and cycling), but in the Adams et
al. (1987, 019356) study PM2.5 exposures were higher than curbside during the summer and
lower than curbside during the winter. As particle size decreased to the fine and ultrafine
range, less difference between in-vehicle and ambient concentrations was observed for PM
mass or count, with the exception of the Diapouli et al. (2008, 190119) study where in-bus
ultrafine PM concentrations were several times higher than indoor or outdoor residential
and school concentrations.
Fruin et al. (2008, 097183) and Westerdahl et al. (2005, 086502) observed that
in-vehicle ultrafine PM concentrations increased for freeways in comparison with arterial
roads. They estimated that 36% of exposure to ultrafine PM occurred during a total daily
commuting time of 1.5 hours (6% of the day); 22% of total exposure occurred during 0.5
hours spent on freeways. Gong et al. (2009, 190121) demonstrated that ultrafine PM
deposition rate increased with decreasing particle size (down to -30 nm) and increasing
surface area inside the vehicle, where deposited PM on the seats and dashboard can be
resuspended and then inhaled or ingested. Ultrafine PM deposition also rose slightly with
increased number of passengers. Zhu et al. (2007, 179919) found that in-vehicle ultrafine
PM counts were 85% lower than outdoors when the fan was operating and recirculation was
on. They estimated that a 1-hcommute (4% of the day) accounts for 10-50% of daily
exposure to ultrafine PM generated by traffic. Based on the American Time Use Survey
estimation of 70.2 minutes spent in vehicles each day (U.S. Bureau of Labor Statistics
http://www.bls.gov/tus/), cumulative in-vehicle exposure can be found that PM2.5 on school
buses was two times higher than on-road levels and four times higher than central site
measurements. Sabin et al. (2005, 087728) demonstrated for school buses that emissions
control technologies had a significant impact on in-bus concentrations of black carbon mass,
and Hammond et al. (2007, 190135) demonstrated significant reductions of particle number
concentration measured for 0.02-1 pm particles when comparing buses using clean diesel or
July 2009
3-225
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
retrofits compared with non-retrofitted buses. Although not tested here for other vehicle
types with respect to PM, these findings suggest that a portion of in-vehicle concentrations
are due to self-pollution, defined by Behrentz et al. (2004, 155682) as the fraction of a
vehicle's own exhaust entering the vehicle microenvironment. Behrentz et al. (2004,
155682) tested self-pollution with school buses using SF6 tracer gas and demonstrated that
0.3% of in-vehicle air comes from self-pollution, and that this number was roughly ten times
greater than in-vehicle concentrations related to self-pollution on newer buses. The
Behrentz et al. (2004, 155682) study also measured EC and particle-bound PAH and found
that 25% of the variability in EC concentration was related to self-pollution. Adar et al.
(2008, 1!) 1200) estimated that 7 pg/m3 of PM2.5 mass on school buses was from self-
pollution. These findings are important for exposure estimation when partitioning local and
ambient sources of pollution during transport in vehicles.
3.8.4.3. Indoor Exposure to Ambient: Infiltration and Differential Infiltration
Fmf varies substantially given a vast array of conditions, and it can best be modeled
dynamically based on a distribution of air exchange and deposition or other ultrafine,
accumulation mode, fine, and coarse PM loss rates rather than a single value (Bennett and
Koutrakis, 2006, 089184; Wallace et al., 2006, 089190), Given that air exchange rates
within a building vary as a function of ambient temperature and pressure, Fmf is subject to
seasonal and regional changes (Meng et al., 2005, 058595; Sarnat et al., 2006, 089166;
Wallace and Williams, 2005, 057485). These factors make Fmf a more accurate descriptor of
infiltration than a simple I/O ratio because the I/O ratio also includes contributions from
indoor sources in addition to PM that infiltrates from outdoors. Wallace et al. (2006,
089190) identified several significant factors affecting Finf, including window opening, age of
an indoor microenvironment, number of occupants, location on a dirt road, dryer usage, and
air conditioning usage). This complex term becomes even more complicated when one
considers transformation of the size distribution and chemical composition of the PM
through chemical reactions on the particle surface, agglomeration, growth, and evaporation
given that Fmf depends on particle size (Keller and Siegmann, 2001, 025881). Fmf for PM is
influenced by physical mechanisms, such as Brownian diffusion, thermophoresis, and
impaction, all of which are functions of particle size (Bennett and Koutrakis, 2006, 089184;
Tung et al., 1999, 049003). These differential effects are summarized below. Recent studies
July 2009
3-226
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
on infiltration are summarized in Table A "58 of Annex A. Fu\ and I/O are listed where
available, although it is recognized that I/O is not as meaningful a descriptor but provides
an approximation of Fmf.
A number of studies have examined the impact of season on PM infiltration. Season is
important because it impacts the ventilation practices used (e.g., open windows, air
conditioning or heating use) and the ambient temperature and humidity conditions affect
the transport, dispersion, and size distribution of the PM. Pandian et al. (1998, 090552)
found that residential air exchange rates vary by season as: summer > spring > winter
> fall with summer air exchange roughly 1.5-2 times greater than average air exchange
rate for the entire year because the rates are driven by home air conditioning and heating
usage. Allen (2003, 053578) gave information on the range and distribution of Fmf at 44
residences in Seattle. The mean Mnf (± SD) measured by light scattering for all sampling
days was 0.65 ± 0.21. Differences in infiltration were observed for the heating season
(0.53 ± 0.16), when windows would be expected to be closed, versus non-heating season
(0.79 ± 0.18). Residences with open windows had a mean Fnf of 0.69 vs. 0.58 for residences
with closed windows. The authors combined the light scattering results with indoor and
outdoor sulfur measurements to estimate that 79 ± 17% of indoor PM2.5 was generated
outdoors. This study provides important data on the distribution of residential Fmf values
and illustrates the magnitude of the effect of season and window position on infiltration
rates. Barn et al. (2008, 156252) and Baxter et al. (2007, 092725) both also noted that
window opening was an important variable. Barn et al. (2008, 156252) found Fnf of
0.61 ± 0.27 for 13 homes during summer and 0.27 ± 0.18 for 19 homes during winter in
Canada. Likewise, location could impact residential ventilation practices and infiltration.
Cohen et al. (2009, 190639) noted differences in median infiltration among eight areas
(among this eight were three comprising the Los Angeles region and two comprising the
New York City region). Indoor-outdoor S( ) r ratio was noted to be highest in New York City
(median: 0.85) and Los Angeles (median: 0.84) and lowest in St. Paul (median: 0.54).
Pandian et al. (1998, 090552) observed that residential air exchange rates vary by region
as: southwest > southeast > northeast > northwest, which reflects regional use of air
conditioning. Sarnat et al. (2006, 089166) noted differences in infiltration between coastal
and inland residences, although variability in these datasets made the differences not
statistically significant.
July 2009
3-227
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Differential infiltration as a function of particle size has been observed to occur.
Infiltration factors for particle diameters ranging from 20 nm to 10 pm were reported in
Boston by Long et al. (2001, 011526) during summer and fall for nighttime periods, when
personal activity patterns would be less likely to generate indoor PM. The maximum
infiltration factor was reported for particles between 80 and 500 nm to range from 0.8-1.0.
Summer values were uniformly higher than fall values, consistent with higher observed air
exchange rates. The infiltration factor decreased with size above 500 nm, reaching 0.1-0.2
for 6-10 pm diameter particles. Particles smaller than 80 nm also were reported to have
lower infiltration factors. This demonstrates the size dependence of PM infiltration, which
has been further studied by recent investigators. Sarnat et al. (2006, 089166) examined
infiltration as a function of particle size and found that Fmf varies by particle diameter, as
measured by a SMPS-APS system to estimate particle volume. Figure 3"97 presents /•'i,r
values for size fractions ranging from 0.02-10 pm. The maximum infiltration factors were
observed around the accumulation mode (0.1-0.5 pm), with Fint = 0.7-0.8. Reduced
infiltration was observed for coarse-mode particles (0.1-0.2 for DP = 5-10 pm) and, to a
lesser extent, ultrafine particles (0.5-0.7 for DP = 0.02-0.1 pm). This is consistent with
increased removal mechanisms for those size fractions: deposition caused by settling for
coarse-mode particles and diffusion for ultrafine particles! for ultrafine PM, diffusion leads
to deposition by agglomeration into larger particles and settling as well as losses to walls.
July 2009
3-228
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
so 0.6
o 0.5

0.01
0.1	1
Particle Size (ym)
10
Source: Sarnat et al. (2006, 089166).
Figure 3-97. Fm as a function of particle size.
3.8.5. Multicomponent and Multipollutant PM Exposures
3.8.5.1. Exposure Issues Related to PM Composition
Annex A presents exposure studies that include chemical speciation data in Table A-
56. Some of these studies focused on SO42 , NO3" or carbonaceous aerosols (EC, OC,
particle-bound PAHs), while others measured concentrations of trace elements from crustal
(Ca, Fe, Mn, K, Al, S, CI in salt), mobile (Al, Ca, Fe, K, Mg, Na, Ba, Cr, Cu, Mn, Ni, Pb, S,
Ti, V, and Zn), or industrial (particle-bound Hg, CI, V, Zn, Ti, Cu, Pb) sources. A number of
source apportionment studies have been performed over the last five years to determine the
contribution of outdoor sources to indoor and personal PM constituents.
Several source apportionment studies by Kim et al. (2005, 156640), Hopke et al.
(2003, 095544) and Zhao et al. (2006, 156181) have shown that secondary SO42- provides
the largest ambient contribution to personal and indoor exposures. These studies took place
on the east coast in Baltimore and Raleigh/Chapel Hill, NC. In Larson et al. (2004, 098145)
source apportionment study in Seattle, vegetative burning was the most significant source
July 2009
3-229
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
of outdoor origin. Zhao et al. (2007, 156182) performed a source apportionment study of
personal exposure to PM2.5 among residents in Denver and also saw lower contributions
from secondary SC>42~ in comparison with motor vehicle emissions and secondary NO3". This
suggests that personal exposure to SC>42~ in parts of the West is lower than in the
Mid-Atlantic. These observations are consistent with the composition distribution shown in
Figures 3-17 and 3-18. Viana et al. (2008, 155269) analyzed PM10 and PM2.5 samples for
several species along urban-to-rural gradients centered in Valencia, Spain and found
gradients for both size fractions in anthropogenically-generated SOr , OC, EC, NO3", Fe,
and NH4"1" but not in mineral species. Combined, these findings suggest urban- and regional-
scale variation in species composition can influence exposure estimates. Personal PM
exposure source apportionment studies are presented in Annex A, Table A-57.
Source apportionment for carbonaceous aerosols is complicated by the fact that they
can be derived from indoor and outdoor combustion sources. Carbonaceous aerosols are
difficult to trace to specific indoor and outdoor sources because combustion is widespread.
Sorensen et al. (2005, 089428), Ho et al. (2004, 056804), Larson et al. (2004, 098145), and
Jansen et al. (2005, 082236) all found that personal and microenvironmental exposure to
total carbon or BC was lower than that measured outdoors, while Sarnat et al. (2006,
089166) showed significant associations between personal and ambient measurements of
EC for measurements taken during the fall (slope = 0.66-0.73) and summer under high
ventilation conditions (slope = 0.41). Wu et al. (2006, 157156), Delfino et al. (2006, 090745),
Olson and Norris et al. (2005, 156005) and Turpin et al. (2007, 157062) all demonstrated
much higher levels of OC compared with EC in personal samples, possibly due to indoor
sources of OC from cooking and home heating. Reff et al. (2007, 156045) and Meng et al.
(2007, 091197) both reported findings from the Relationships between Indoor, Outdoor, and
Personal Air (RIOPA) study in Los Angeles, Houston, and Elizabeth, NJ. Results from Reff
et al. (2007, 156045) are shown in Figure 3-98. These reveal significantly higher detection
of aliphatic C-H functional groups indoors and in personal samples compared with
outdoors. This information may help to distinguish carbonaceous compounds of indoor and
outdoor origin in future source apportionment studies of PM exposure. Little regional
variation in the aliphatic, carbonyl, or SO r groups tested were reported in this study. In
Meng et al. (2007, 091197) indoor exposures were shown to decrease for secondary
formation aerosols including S() r but excluding NOa (not tested) when compared with
outdoor concentrations, and indoor exposures to mechanically generated aerosols increased
July 2009
3-230
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
in comparison with outdoors. Differences in infiltration based on source category are shown
in Figure 3 "99.
Trace metal studies have shown variable results regarding personal exposure to
ambient constituents. For instance, Molnar et al. (2006, 156773) found that personal
exposure was higher than outdoor and ambient concentrations for mostly crustal CI, K, Ca,
Ti, Fe, and Cu. However, Adgate et al. (2007, 156196) found that personal exposures were
higher than ambient for Fe, Mg, K, Zn, Cu, Pb, and Mn but lower than ambient for Al, Na,
and Ti. Larson et al. (2004, 098145) found that personal exposure to Ca and CI were higher
than concentrations measured at ambient (central site) and residential outdoor monitors,
lower for Fe, K, Mn, and As and the same for Al, Br, Cr, Cu. Source apportionment for trace
metals can vary significantly among cities and over seasons. For instance, in a Baltimore
source apportionment study, exposure to Mn could be attributed nearly equally to the
Quebec wildfires, roadway wear, and soil, while Pb exposure was largely found to be due to
a local incinerator (Ogulei et al., 2006, 119973). In this case, the Quebec wildfires were a
transient episodic source while roadway wear and incineration were continuous. However,
in Larson et al. (2004, 098145), Mn and Pb exposures in Seattle were largely attributable to
mobile source and stationary source emissions. For this reason, source composition
behavior cannot be generalized for characterizing exposures and resulting health effects
across multiple locations or times.
July 2009
3-231
DRAFT-DO NOT CITE OR QUOTE

-------
Los Angeles Co., CA
I
J? ^ J"
Elizabeth, NJ
ro 10
d y
_ 25
CO
¦JL 20
CD
3.
c 15
o
c
° y
^o5*
0 \y
* jfjP ^ ^

Source: Reff et a\. (2007,156045)
Figure 3-98. Apportionment of aliphatic carbon, carbonyl and S042" components of outdoor, indoor,
and personal PM2.5 samples, for Los Angeles (top), Elizabeth (center), and Houston
(bottom).
July 2009
3-232
DRAFT-DO NOT CITE OR QUOTE

-------
o
o
-o
e
30
20
Mechanical Finf = 0.04
y = 0.04X +0.51, RJ = 0.51
/
/
/
/
/
/
/
/
/ •	AER: < 0.5 h1
V	AER: 0.5-1.5 h''
/	~	AER: > 1.5 h'*
10 7
&4
30

/
/
/
/
~ v
3D—	,V V,
10 15 20 25 30 35
m 25 ¦
Primary Fln( = 0.51
y = 0.51x + 0.19, R2 = 0.61
Qfi
=L
20
/
/
/
/
/
/
/
/
/
S • AER: < 0.5 h 1
X V AER: 0.5-1.5 h'1
~ AER: > 1.5 h'
50
40 ¦
30
20 -
Secondary Finf = 0.78
y = 0.78x - 0.53, RJ = 0.75
/
/
/
/ V
/
/
/
/ • AER: *0.5 h-'
V AER: 0.5 -1.5 h"1
~ AER: > 1.5 h'1
30
40
50
Outdoor Concentration (jig/m )
Source: Meng et al. (2007, 091197).
Figure 3-99. Apportionment of infiltrated mechanically-generated (top), primary combustion (center),
and secondary combustion (bottom).
A number of chemical factors influence the tendency for differential infiltration in PM.
Lunden et al. (2003, 156718) studied infiltration of BC and OC aerosols and found that JSsf
July 2009
3-233
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
can vary substantially as a function of gas transport properties within differing air
exchange rates. This study and Sarnat et al. (2006, 090489) also showed that BC aerosol
infiltration is considerably higher than infiltration of OC, and that carbonaceous aerosol
infiltration differed substantially from N( )a and S( ) r aerosols under the same building air
exchange conditions. These differences are likely related to differences in the particle size
distribution of PM components. As shown in Figure 3-100, the composition of indoor PM
that has infiltrated from outdoors is different from that of outdoor PM. In this case, the
particles containing photochemical products (primarily accumulation mode) have a higher
infiltration rate than the larger (primarily coarse mode) mechanically generated particles
or the smaller primary combustion particles (probably mostly ultrafine particles in the
nucleation or Aitken nuclei mode).
7
8
5
CO
E
g 4
3
2
1
0
Source: Meng et al. (2007, 091197).
Figure 3-100. Results of the positive matrix factorization model showing differences in the mass of
outdoor PM and PM that has infiltrated indoors based on source category.
PM species enriched in the accumulation mode, such as SO42~, will infiltrate more
efficiently than components with larger size distributions, such as iron (Strand et al., 2007,
Photochemical
Secondary
(~ Accumulation Mode)
Primary
Combustion
Particles
(- Ultrafine)
6,8
6.4
Mechically
Generated
(- Coarse Mode)
5.1
0.13
3.3
Out In
Out In
Out In
July 2009
3-234
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
157018). Lunden et al. (2008, 155949) also compared I/O ratios for PM2.5, total carbon in
PM, OC in PM, and BC in PM in an unoccupied house and found the lowest ratio for PM2.5
(0.41 ± 0.2), the highest for BC (0.61 ± 0.2), and intermediate values for total carbon
(0.50 ± 0.2) and OC (0.47 ± 0.2). The authors attributed the lower PM2.5 I/O ratio to indoor
loss of NH4NO3 aerosol. The authors note that their BC I/O of 0.6 is somewhat lower than
BC ratios measured in occupied spaces (Polidori et al., 2007, 156877). Conversely, indoor
sources in occupied residences contribute to observed OC I/O ratios greater than 1 in other
studies (Polidori et al., 2006, 156876; Sawant et al., 2008, 156949). Analytical results for
PM2.5 components from the Baxter et al. (2007, 092726) study found Fmf of 0.95 ± 0.07 for S
and 0.60 ± 0.04 for V, the two components identified as having no indoor sources and which
had I/O ratios significantly less than 1 (Baxter et al., 2007, 092726). It is possible that
association of V with larger particles of lower penetration efficiency could contribute to a
lower infiltration rate. Meng et al. (2005, 058595) also noted that the lack of indoor sources
of S and V result in much lower variability in penetration and loss rates.
Volatilization of PM during infiltration can cause differences between the composition
of indoor and outdoor PM. NO3 , a prevalent PM component in the western U.S. year-round
and during winter throughout the mid-western and northeastern states, has a decreased
Fmf due to volatilization of' NOa indoors. Sarnat et al. (2006, 089166) calculated Mnf values
for NO3", PM2.5, and BC, and found the values to increase in that order. NOa was low
(median = 0.18, IQR = 0.12-0.33), while BC was high (median = 0.84, IQR = 0.70-0.96); the
intermediate value of PM2.5 (median = 0.48, IQR = 0.39-0.57) reflected its composition as a
mixture of those two components (among others). Indoor volatilization of N03~ enriches
ambient PM in other components, creating differences in toxicity between indoor ambient
and outdoor ambient PM. The high infiltration of non-volatile BC creates additional
sorption sites for organics, including indoor-generated compounds. Meng et al. (2007,
091197) found that secondary formation accounts for 55% of indoor aerosols of outdoor
origin, while primary combustion accounts for 43% and mechanical generation for 2%.
Meng et al. (2007, 091197) noted that secondary formation processes often result in more
accumulation mode particles so that diffusion losses are not as great as for primary
combustion particles, composed primarily of nucleation and condensation modes (see
Figure 3-97). Likewise, Polidori et al. (2007, 156877) suggest that similarities in the EC and
OC size distributions and infiltration factors reflect low vapor pressure secondary organic
aerosols in the OC component. Sioutas et al. (2005, 088428) suggest that volatilization of
July 2009
3-235
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
ultrafine PM while crossing the building envelope may hamper infiltration in this size
range. Variations in the presence of outdoor PM indoors, and resulting changes in removal
behavior once indoors, relate to the species composition and makeup of PM.
3.8.5.2. Exposure to PM and Copollutants
Analysis of personal exposure to multipollutant mixtures is an area of growing
research. Several multipollutant studies involving ultrafine, fine, and coarse PM are
presented in Table A-59. Understanding the health impacts of complex multipollutant
mixtures, including multiple PM species, can have a substantial impact on the
interpretation of health effects data. Challenges are presented in accurately estimating the
components of a mixture, their concentrations, and personal exposure to those species.
One question that has been raised is whether copollutants act as confounders in PM
exposure assessments. Sarnat et al. (2001, 019401) explored the relationship between PM
and copollutant gases and suggested that certain gases can serve as surrogates for
describing exposure to other air pollutants. Sarnat et al. (2001, 01!) 101) found significant
associations between personal exposure to PM2.5 and ambient concentrations of O3, NO2, CO
(significant only for winter), and SO2 in a panel study conducted in Baltimore. Personal
exposures to PM2.5 and personal exposures to the gases were not correlated in this study.
This result may have arisen in part because personal exposures to the gases were often
beneath detection limits of the personal monitoring devices. Schwartz et al. (2007, 090220)
also used data from the Baltimore panel study to simulate distributions of personal
exposures and ambient concentrations of PM2.5, PM10, SO42 , NO2, and O3. They found that
personal exposure to ambient PM2.5 was significantly associated with ambient
concentrations of PM2.5, NO2, and O3 (O3 in an inverse relationship). They also reported that
personal exposure to SO r was significantly positively associated with ambient PM2.5 and
O3 concentrations.
Seasonality of the associations could be a result of seasonal variability in
photochemistry, source generation, and building ventilation. Sarnat et al. (2005, 087531)
observed associations between personal PM2.5 and ambient O3, NO2, and SO2 exposure for a
group of healthy senior citizens and school children in Boston for summer but not for
winter. In this study, significant associations between personal exposure to ambient PM2.5
and personal O3 exposures were observed in summer and between personal PM2.5 and
July 2009
3-236
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
personal NO2 in winter and summer, unlike the Baltimore study in which only summertime
personal PM2.5 and personal NO2 were associated (Sarnat et al., 2001, 019101). In their
study of personal exposure to ambient air pollutants in Steubenville, OH, Sarnat et al.
(2006, 090489) found that, in the summer, low but significant associations existed for
ambient O3 with personal PM2.5, SO r , and EC. In the summer, low but significant
associations between ambient SO2 and personal PM2.5 and between ambient NO2 and
personal EC were also observed. In the fall, ambient O3 had a weak but significant
association with personal EC, and SO2 had a weak but significant association with personal
SO42-. Ambient NO2 was significantly associated with personal PM2.5, SO42-, and EC with
somewhat higher coefficient of determination (R2 = 0.25-0.49). There is evidence that
associations between ambient gases and personal exposure to PM2.5 of ambient origin exist
but are complex and vary by season and region.
3.8.6. Implications of Exposure Assessment Issues for
Interpretation of Epidemiologic Studies
Environmental epidemiologic study designs vary by many factors, including study
sample size, measurement time interval, study duration, monitor type, and spatial
distribution of the study sample. A panel epidemiology study consists of a relatively small
sample (typically tens) of study participants followed over a period of days to months. Time-
activity diary studies are examples of panel studies (e.g.Cohen et al., 2009, 190639;
Elgethun et al., 2003, 190640; Johnson et al., 2000, 001660; Olson and Burke, 2006,
189951), and a microenvironmental model might be applied to represent exposure in this
case. Community time-series studies may involve millions of people whose exposure and
health status is estimated over the course of a few years using a short monitoring interval
(hours to days). Because so many people are involved, community-averaged concentration is
typically used as a surrogate for exposure in community time-series studies. Exposures and
health effects are spatially aggregated over the time intervals of interest because they are
designed to examine health effects and their potential causes at the community level
(e.g.Dominici et al., 2000, 005828; Peng et al., 2005, 087463). A longitudinal cohort
epidemiology study typically involves hundreds or thousands of subjects followed over
several years or decades. Concentrations are generally aggregated over time and by
community to estimate exposures (e.g.Dockery et al., 1993, 044457; Krewski et al., 2000,
July 2009
3-237
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
012281). The importance of exposure misclassification varies with study design based on
the spatial and temporal aspects of the design. Other factors that could influence exposure
estimates in PM epidemiologic studies include source characteristics, particle size
distribution, and particle composition. Potential issues that could influence estimates of PM
exposure include measurement, modeling, spatial variability, temporal variability, use of
surrogates for PM exposure, and compositional differences. These are described in detail in
the following sections.
3.8.6.1. Measurement Error
Measurement Error at Community-Based Ambient Monitors and Exposure Assessment
Community-based ambient monitors are employed for time-series and longitudinal
studies, although they can be used for panel studies as well. Section 3.4 discusses potential
errors in measuring ambient PM in detail. Because there will likely be some random
component to instrumental measurement error, the correlation of the measured PM mass
with the true PM mass will likely be less than 1. Sheppard et al. (2005, 079176) indicate
that instrument error in the hourly or daily average concentrations has "the effect of
attenuating the estimate of a." However, Zeger et al. (2000, 001949) suggest that in order
for this error to cause substantial bias in later estimation of a health outcome, the
measurement error must be strongly correlated with the measured concentrations. Positive
and negative artifacts resulting from sampling volatile PM may therefore lead to lack of
association with health endpoints in time-series and longitudinal studies. In multicity
longitudinal studies, PM composition and associated artifacts may vary across cities.
Measurement Error for Personal Exposure Monitors
Personal exposure monitors (PEM) are primarily used in panel exposure studies to
measure total exposure to PM (e.g.Cohen et al., 2009, 190639; Elgethun et al., 2003,
190640; Johnson et al., 2000, 001660; Olson and Burke, 2006, 189951). PEMs are
specialized monitors that, because people must carry them, have to be small, light, quiet,
and battery operated or passive. As a result, they may have lower face velocities across the
filter and pressure drops than ambient based filter measurements of PM, which typically
sample at much higher flow rates, and consequently at much higher face velocities. Light
scattering measurements are biased when relative humidity is high and are sensitive to
size distribution (Lowenthal et al., 1995, 045134; Sioutas et al., 2000, 025223). Positive
July 2009
3-238
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
artifacts resulting from adsorption of vapor-phase organic compounds and negative
artifacts from evaporation of semi-volatile PM also creates challenges for personal exposure
(e.g.Pang et al., 2002, 030353). Olson and Norris et al. (2005, 156005) attributed more OC
particle mass collection using PEMS than when using FRMs! this was attributed to face
velocity differences. Accuracy of PEMs is described in much greater detail in the 2004 PM
AQCD (U.S. EPA, 2004, 056905). The artifacts listed here could result in either negative or
positive sampling bias.
3.8.6.2. Model-Related Errors
When models are used in lieu of or to supplement measurements of ambient PM
exposure or community-based ambient PM concentration, it is important to identify errors
and uncertainties that could affect estimates of PM-related health effects. Model-related
errors are determined by four factors^ representativeness of the mathematical model,
accuracy of model inputs, scale of model resolution, and model sensitivity. If verification
errors related to these four factors are minimized, then the model can be evaluated against
physical data to determine how well the model truly captures a real situation (Roache,
1998, 156915). Detail of the model design and inputs can have significant impact on
validation, as demonstrated in Meng et al.(2005, 058595) and Hering (2007, 155839). Meng
et al. (2005, 058595) demonstrated how use of an increasingly more detailed mathematical
model decreases the variability of the results with respect to modeled indoor PM2.5
concentration of outdoor origin and infiltration factor. Hering (2007, 155839) compared
infiltration model results for PM-based EC, NO3", and SO42Model inputs were from a
central site monitor only, central site monitor with air exchange data, and detailed inputs
related to initial outdoor (outside test building) and indoor concentrations. Use of more
detailed inputs resulted in significant reductions in error for indoor EC concentration,
smaller improvements for indoor SO42", and negligible improvement in model results for
indoor N03_.This illustrates the impact of differential infiltration discussed in
Sections 3.8.4 and 3.8.5.
July 2009
3-239
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
v 4430
~ 4425
36 km CMAQ grid
12 km CMAQ grid
4 km CMAQ grid
Census tract
centroids
in Philadelphia
county:
Tracts in each
4x4km
grid cells are shown
in different colors
4405
465 470 475 480 485 490 495 500 505 510 515
West-East distance (km)
Source: Isakov et al. (2007,156588).
Figure 3-101. Grid resolution of the CMAQ model in Philadelphia compared with distribution of census
tracts in which exposure assessment is performed.
For any spatial interpolation models, grid resolution is another source of error
addressed. Isakov et al. (2007, 156588) linked CMAQ with the Hazardous Air Pollutant
Exposure Model for exposure assessment in Philadelphia. Their simulation was
implemented on a 4 km nested grid within 12 km and 36 km grids to bring the scale of their
model from national to urban scale. However, the census tracts in which Isakov et al. (2007,
156588) sought to describe exposure were distributed on a much finer scale (see
Figure 3-101). They were required to supplement the CMAQ model with an Industrial
Source Complex Short Term (ISCST) dispersion model to resolve the subgrid scale behavior.
If concentrations were averaged across the cell in lieu of a more detailed subgrid
representation, (Isakov et al., 2007, 156588) found that exposures were overestimated by a
factor of two. Appelet al. (2008, 155660) noted that their 36 km simulations provided a
closer estimate of SO** aerosol concentration than did their 12 km nested simulation,
which overestimated concentrations. Hogrefe et al. (2007, 156561) also noted
overestimation of the CMAQ model at the 12 km scale, where multiple point interpolation
was used to obtain subgrid estimates. Model convergence theory would suggest that the 36
July 2009
3-240
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
km simulation is not actually more accurate but coincidentally closer to the physical values
(Roache, 1998, 156915). It is possible that if secondary pollutants are more regionally
dispersed, lower spatial resolution would be required to attain a converged solution of the
spatial concentration distribution. However, higher spatial resolution in the simulation
should produce very similar results if the solution is convergent.
Use of geospatial statistical methods for grid interpolation, as performed in the
SHEDS/MENTOR simulation by Georgopoulos et al. (2005, 080269), provides another
methodology for grid interpolation. Similar to Isakov et al. (2007, 156588), Georgopoulos et
al. (Georgopoulos et al., 2005, 080269) linked CMAQ with an exposure model for estimation
of neighborhood-scale exposures. The authors found that CMAQ underestimated PM2.5
concentration at many times during the simulation. Kim et al. (2009, 188446) compared
results from an exposure model using six different levels of spatial resolution. Kim et al.
(2009, 188116)'s model predicted PM2.5 at monitor locations as a function of the mean
concentration and spatial and random errors. The six levels of spatial resolution were
simulated through assignment of the spatial and random error terms and by defining the
distance over which spatial errors are correlated. Between monitors, Kim et al. (2009,
188446) compared results from assigning nearest monitor values and from kriging. They
found that model prediction error and bias in health effects estimates decreased with
increased spatial resolution of the error terms as well as with kriging over a nearest
monitor scheme.
For GIS-based models designed to improve spatial resolution of exposure estimates,
Yanosky et al. (2008, 099467) described three sources from which they derived an estimate
of total model uncertainty: transient model components, stationary model components, and
residual spatial and temporal components of variance. When analyzing relative
contributions to uncertainty, they found that unexplained local spatial variability was the
largest contributor. With model inputs from PM10 monitors for this study, Yanosky et al.
(2008, 099467) observed poor model performance and high uncertainty where monitors
were sparsely located. High uncertainties were also calculated in a few select urban areas
(New York City, Detroit, Cleveland, and Pittsburgh) where concentrations tend to be higher,
although the latter may have been related to the Taylor series approximation Yanosky et al.
(2008, 099467) used for the residual uncertainty term. Spatial and temporal uncertainties
were also reduced when temporal resolution was increased in the model implementation.
July 2009
3-241
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
In his review of various exposure assessment modeling techniques, Jerrett et al.
(2005, 092864) reviewed source proximity and LUR for application to exposure assessment.
The literature contains mixed evidence of the association between health effects and source
proximity (Langholz et al., 2002, 191771; Maheswaran and Elliott, 2003, 125271; e.g.,Venn
et al., 2000, 007895; Venn et al., 2001, 023644). Jerrett et al. (2005, 092864) contend that
source proximity modeling is limited because other confounding covariates, such as
socioeconomic status, may be related to source proximity. Additionally, subjects' time-
activity patterns may vary from locations modeled through source proximity. Wind direction
and topography may bring PM plumes away from a site located even in very close proximity
to the source so that high concentrations would be found at distances far downwind of a
source. Jerrett et al. (2005, 092864) state that LUR is an adaptable framework allowing
adaptation to localized conditions, but they caution that LUR is limited to fairly
homogeneous spatial regions. They point to Briggs' (2000, 191772) simulation of Amsterdam
as an example of LUR surfaces produced with little spatial variability.
LUR and kriging were both used in the ACS data reanalysis by Krewski et al. (2009,
191193) to study mortality as a function of spatial variability in PM2.5 in New York and Los
Angeles. The LUR predicted 66% and 69% of the variability in PM2.5 concentration in New
York City and Los Angeles, respectively. The LUR solution produced some observed
overp re dictions near freeways. Kriged results were compared with LUR for both cities. For
New York City, kriging produced slightly attenuated mortality risk estimates, while for Los
Angeles, kriging did not exhibit as much spatial variability as LUR. The latter may be due
to the fact that the monitoring network in Los Angeles was not situated to capture spatial
variability in PM2.5 concentration occurring near areas of high traffic. Despite similarities
in the LUR performance, health effects predictions were quite different for New York City
and Los Angeles, with increased hazard ratios of 1.56 and 1.39, respectively using the same
LUR covariates. Krewski et al. (2009, 191193) noted that in New York City, the healthiest
(and wealthiest) segment of the population lived in the most polluted areas, while in Los
Angeles, there was a strong association between pollution and mortality. This finding
implies that, because geographic regions may differ by multiple factors such as building
and power plant fuel use, roadway design, traffic patterns, building design, traffic patterns,
and building design, significant variables in an LUR analysis may also differ by region.
July 2009
3-242
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
3.8.6.3. Spatial Variability
For PM, spatial and temporal distribution as a function of particle size and
composition also plays a large role in the selection of an exposure model. For instance, use
of Equation 3-6 might be employed for a study of ambient PM2.5 exposure because spatial
variability in PM2.5 concentration can be low over urban to regional scales. PM10-2.5 tends to
deposit and become resuspended over a neighborhood scale as a result of gravitational
forces. Ultrafine PM is generally limited to micro-to-neighborhood scale environments
because it tends to diffuse rapidly and then coagulate with other particles or serve as
condensation media. These phenomena cause rapid growth of ultrafine PM that limits their
lifetime and spatial extent. Spatial issues leading to exposure misclassification are
discussed below.
Spatial variability of ambient PM concentration can occur when there are local
sources, particularly for ultrafine PM and coarse PM with shorter spatial scales. In panel
studies, an exposure error will be introduced if the ambient PM concentration measured at
the central site monitor is used as an exposure surrogate and differs from the actual
ambient PM concentration outside a subject's residence and/or worksite. Filleul et al. (2006,
089862) computed exposure based on varying contributions of community-based ambient
monitors (deemed background) and proximal monitors (to represent a receptor) in Le
Havre, France for black smoke measurements. Filleul et al. (2006, 089862) found that
increasing the weight of proximal monitors resulted in non-significant but increased mean
concentrations. Moore et al.'s (2009, 191002) findings of high variability in ultrafine PM
across Los Angeles also suggest that exposure error would occur from using one or a few
ultrafine PM monitors. In an example using AQS data, PM10 monitors in the Chicago CSA
are south of the most populated areas within Chicago (see Figure A-6 in Annex A), and
intersampler correlations for urban scale PM10 data for several monitor pairs are below 0.4
(see Table A-22 in Annex A). In another example from AQS data, PM2.5 and PM10 monitors
in the Riverside CBSA are shown to correspond more closely with higher population density
areas (see Figures A-23 andA-24 in Annex A), and urban scale intersampler correlations for
both PM2.5 and PM10 are below 0.4 for several monitor pairs there. For most cities,
intersampler correlation is much higher for PM2.5 than for PM10. This is consistent with the
findings of Sarnat et al. (2009, 180084) where, in a time-series study of the effect of spatial
variation in concentration on epidemiologic associations, spatially homogeneous PM2.5 and
July 2009
3-243
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
O3 were found to be associated with emergency room visits while spatially heterogeneous
CO and NO2 were not well associated. Considering results reported in the literature along
with inter-sampler correlations reported in Annex A, Section A.2 for PM10 and PM2.5, and
their corresponding monitor locations shown in Annex A, Section A. 1, the magnitude of
spatial exposure error likely depends on particle size as well as monitor location, source
location, and characteristics such as urban and natural topography and meteorological
trends.
Spatial variability among various studies further suggests that use of a single or
small number of ambient monitors introduces uncertainty in exposure assessment panel
studies. Violante et al. (2006, 156140) studied personal exposures to traffic of parking police
in Bologna, Italy to determine how personal exposure to outdoor PM10 and benzene
compares with that measured at a community-based monitor. This study found that
personal exposures to PM10 were significantly higher than at the community-based monitor,
although the authors were not able to demonstrate significant effects of meteorology or
traffic on those exposures. Spatial heterogeneity of personal exposures to metals in PM10
and PM2.5, with higher levels found near high-traffic and industrial areas, was observed by
Nerriere et al. (2007, 156801). In a Bayesian hierarchical model analysis of personal
exposure and ambient data from the pilot Baltimore Epidemiology-Exposure Panel Study of
16 subjects, McBride et al. (2007, 124058) showed that community monitors overestimated
personal exposures for the panel subjects, and that these results were not sensitive to
model selection.
For community time-series epidemiology, the community averaged concentration, not
the concentration at each fixed monitoring site, is the concentration variable of concern
(Zeger et al., 2000, 001!) 1!)). Because variation in trends is of interest, bias in the central
site monitor data will not affect health effects estimates unless the central site monitor is
not correlated with the community averaged concentration. The latter condition will cause
the health effect estimate to be biased towards the null (Sheppard et al., 2005, 079176). The
correlation between the concentration at a central community ambient monitor and the
community averaged concentration depends on homogeneity of the spatial distribution and
representativeness of the central site monitor location. Regional pollutants such as S( )r
will be more spatially homogeneous than point source pollutants. Traffic emissions might
show spatial heterogeneity near sources but more homogeneous distribution farther
downwind from sources. Kim et al. (2005, 083181) noted that spatial variability among PM
July 2009
3-244
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
species can add uncertainty to exposure estimates in community time-series epidemiology
studies exploring source contributions to health effects. The monitoring site is selected to
represent the community average of the PM characteristic (mass and/or species) of interest.
If the selected site is far from PM sources, then the average measurement may be lower
than actual ambient PM concentrations. Likewise, if the site is selected to measure a "hot
spot" or pollution from a nearby source, exposure estimates across the community could be
skewed upwards.
Intra-urban spatial heterogeneity could affect health effects estimates derived from
community time-series studies if a community is divided by urban or natural topographic
features or by source locations into several sub-communities that differ in the temporal
pattern of pollution. Intra-urban spatial heterogeneity is discussed in detail in Section 3.5.
Community exposure may not be well-represented when monitors cover large areas with
several sub-communities having different sources and topographies. This point is
illustrated for Los Angeles in Figures 3-27 and 3-37. Using zip code classified mortality
data in a study of socioeconomic status and acute cardiovascular mortality in Phoenix, high
risk ratios were computed when a small area near the monitoring site was studied, (Mar et
al., 2003, 156731; Wilson et al., 2007, 157149) while use of larger-area county-wide data
produced non-significant associations (Moolgavkar, 2000, 010305; Smith et al., 2000,
010335). At least part of the heterogeneity found between cities in multicity studies may be
due to the use of a large geographic area that is composed of several sub-communities that
differ in the spatiotemporal distribution of air pollutants. Note that when zip codes cover
large areas (e.g., in western mountain states) or when counties cover small areas (e.g., the
northeast), then assumptions change regarding use of zip code- or county-level data for
epidemiology studies. For all metropolitan areas investigated in this assessment, the PMio
data have significantly more scatter. This suggests that the uncertainty of the community
average concentration would increase in the coarse PM range. Metrics have been developed
and used to compare the spatial variability of air pollutants (Wongphatarakul et al., 1998,
049281). These metrics are useful in assessing the potential for exposure error in the
epidemiologic studies, especially when different monitors are used on different days to
construct city-wide averages.
Epidemiologic studies of long-term exposure rely on differences among communities
in long-term average ambient concentrations. If exposure errors are different in the
different communities, the differences in long-term ambient concentrations among
July 2009
3-245
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
communities may not represent the differences in long-term average exposures (Dockery et
al., 1993, 044457). For example, there maybe community to community differences in
measurement error if exposure to fresh pollutants generated by vehicular traffic or
pollutants from other localized sources differed among the spatial areas. Thus, in a
regression of health effects against average concentration as an indicator for average
exposure, there could be a different magnitude and direction of error in the exposure
indicator for each spatial area. This could bias the slope up or down. The following long-
term mortality studies, described in Section 7.6, are cited here to illustrate the effect of
spatial exposure error on health effects estimates! the reader is referred to Section 7.6 for a
more detailed description of these results. The Harvard Six-City Study dealt with this issue
by design, where the members of the cohort in each city were located in a relatively small
area near the monitor (Dockery et al., 1993, 044457). In the ACS study, the spatial area was
the Metropolitan Statistical Area (MSA) (Krewski et al., 2000, 012281); other studies have
used counties as the spatial area (Enstrom, 2005, 087356; Lipfert et al., 2000, 004087). In a
comparison of several of the larger long-term cohort studies (Enstrom, 2005, 087356), those
using county level spatial areas (Enstrom, 2005, 087356; Lipfert et al., 2000, 004087)
sometimes did not find significant associations whereas those using MSAs (Pope CAet al.,
1995, 045159; Pope CAet al., 2002, 024689) or cities (Dockery et al., 1993, 044457) did find
significant associations. Jerrett et al. (2005, 189405) used smaller zip code areas within Los
Angeles County and found effects that were both significant and largest in magnitude
among those that have been reported for long-term cohort studies. Krewski et al. (2009,
191193) suggested that significant associations between cardiovascular health effects
estimates and PM2.5 observed in Low Angeles but not in New York City were related to
spatial homogeneity of PM2.5 concentration in New York City. The Nurses' Health Study
examined associations of mortality with PM10 found higher and more significant
associations when using estimated concentrations at subjects' individual residences (Puett
et al., 2008, 156891) in lieu of county-level concentrations (Fuentes et al., 2006, 097647).
These considerations suggest that studies that include large U.S. counties as spatial areas
and find no significant associations of health effects with pollution cannot be considered
definitive, because the likelihood of exposure error increases in this situation. Reducing the
exposure error by using concentrations based on residence address or small zip code areas
is associated with larger relative risk than those obtained with county-wide averages of
concentrations.
July 2009
3-246
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
3.8.6.4. Temporal Variability
Temporal Correlation
The time series analyzed together for community time-series epidemiology studies can
include those obtained from several central site monitors, a single central site monitor with
the true community average exposure, and the representative monitor and health effect
endpoint. Within a city, lack of correlation of relevant time series at various sites results in
smoothing the exposure/surrogate concentration function over time and resulting loss of
peak structure from the data series. Burnett and Goldberg (2003, 042798) found that
community time-series epidemiology results reflect actual population dynamics only when
five conditions are met: environmental covariates are fixed spatially but vary temporally!
the probability of the health effect estimate is small at any given time! each member of the
population has the same probability of the health effect estimate at any given time after
adjusting for risk factors! each member of the population is equally affected by
environmental covariates! and, if risk factors are averaged across members of the
population, they will exhibit smooth temporal variation. Note that for this study, Burnett
and Goldberg (2003, 042798) analyzed mortality, but the results are generalized here.
Dominici et al. (2000, 005828) note that ensuring correlation between ambient and
community average exposure time series is made difficult by limitations in availability and
duration of detailed ambient concentration and exposure time series data and so this is
often a source of uncertainty. Sheppard et al. (2005, 079176) also add that the health effect
estimate can be biased by time-dependent error in a in a time-series study, provided the
spatial variation in PM concentration is not significant and thus Equation 3"6 can be
assumed a valid exposure model. The direction of bias is related to seasonal correlation
between a and ambient concentration.
Seasonality
Community time-series studies can be designed to investigate seasonal effects by
incorporating seasonal interaction terms for the exposure surrogate and/or meteorology
(e.g.Dominici et al., 2000, 005828). Studies from Section 6.5 are cited here to illustrate how
seasonal exposures can influence health effects estimates! the reader is referred to
Section 6.5 for a more detailed review of the results. Peng et al. (2005, 087463) and Bell et
al. (2008, 156266) observed higher health effects estimates and stronger seasonal
July 2009
3-247
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
dependence in the northeast than the rest of the country for PMio and PM2.5 respectively.
Peng et al. (2005, 087463) stated that these results generated three hypotheses. First, the
PM composition and resulting toxicity might vary with season. Bell et al. (2008, 156266)
results showed seasonal differences between respiratory and cardiovascular effects
estimates that the authors hypothesize relate to seasonal differences in dominance of a
given PM species. Second, Peng et al. (2005, 087463) suggested that less seasonality in
regions other than the northeast may reflect regional tendencies for spending more or less
time outdoors. Exposure estimates for time spent outdoors may be less subject to exposure
error because uncertainties related to infiltration are not a factor during that time. At the
same time, air conditioning usage, which is more common in the summer and in warm
climates Pandian et al. (1998, 090552), has been associated with decreased association
between PM2.5 and cardiovascular morbidity (Bell et al., 2009, 191007). Third, infectious
diseases are more prevalent during winter and so may influence health outcomes. However,
it would be expected that regions other than the northeast would be affected by influenza in
winter. Uncertainty in sources of seasonal bias may also indicate other unknown factors.
Data Frequency
Most panel studies look at the associations of health outcomes only with exposure (or
exposure surrogates) on the day of exposure (lag 0). Zanobetti et al. (2000, 004133) suggest
that health effects may not occur until subsequent days or be distributed over several days.
When PM measurements are obtained every three or every six days, it is difficult to refine
the study lag structure down to the day-level. In studies of the effects of short-term PM2.5
exposure on cardiovascular and respiratory hospitalization in >200 urban U.S. counties,
Bell et al. (2008, 156266) and Dominici et al. (2006, 088398) worked with a combination of
measurements obtained daily and those obtained every three days, while their
hospitalization data were daily. Time lags of 0, 1, and 2 days were applied so that PM2.5
data obtained only on day 0 would be applied as lag 0 for the corresponding day's
hospitalization record, as lag 1 for the next day's hospitalization, and as lag 2 for the
following day. No lag 0 data would be available for day 1, and no lag 0 or 1 data would be
available for day 2 in this example. This analysis could be performed with sufficient
statistical power despite the reduction in days of PM data because a large number of
counties were analyzed. Single city studies using data obtained every three or six days
could employ the same approach but would lose statistical significance compared with daily
July 2009
3-248
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
data because fewer data points would exist for each lag. Likewise, Dominici et al. (2006,
088398) only applied distributed lag analysis to the daily hospitalization data where daily
PM2.5 concentration data were also available.
3.8.6.5. Use of Surrogates for PM Exposure
Surrogates for Infiltration Tracers
In panel studies, a tracer can be used for PM, such as sulfate as a surrogate for PM2.5,
as described in Section 3.8.4. For this method to be successful, indoor or other nonambient
sources of the tracer must be small compared to ambient sources over the period of
sampling. Wilson and Brauer (2006, 088933) noted that environmental tobacco smoke and
tap water used in showers or humidifiers are indoor sources of SO42Wallace and Williams
(2005, 057485) observed that, because SO42- particles are typically smaller than other PM
contributing to measurable PM2.5 mass, /'mr may be biased by using this term to describe
PM2.5 infiltration. Other concerns in using SO42- as a tracer for PM2.5 arise because SO42-
tends to be concentrated in smaller particles and thus it might be a better tracer for fine
mode particles than for all PM2.5 particles. Stand et al. (2006, 157017) suggested that Fe be
used as an additional tracer to correct for the infiltration of larger PM2.5 particles. In their
study, they noted that indoor sources of Fe were small. However, in other environments
there could be more substantial contributions from tracking iron in soil indoors. The spatial
variability of Fe is also larger than that of PM2.5 across urban areas. Volatilization of nitrate
or organic compounds after infiltration of PM2.5 indoors results in these components being
poor surrogates for ambient PM in exposure estimates Lunden et al. (2003, 081201).
Use of Ambient PM Concentration in Lieu of Ambient PM Exposure
Ambient PM concentration is often used as a surrogate for exposure to ambient PM in
epidemiologic studies. Based on the information presented in Section 3.8.4 related to urban-
scale PM distribution, there is less exposure error for accumulation mode PM because it has
a more homogeneous spatial distribution and higher filtration indoors, compared with for
coarse or ultrafine PM. The ambient concentration may be based on measurements made
just outside the primary microenvironment, at the nearest community monitor, at a single
community monitor, or as the average of several community monitors. If appropriate
measurements are made, it is also possible to estimate the ambient and nonambient
components of total personal exposure and use four exposure surrogates in panel
July 2009
3-249
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
epidemiologic studies: Ca, Et, Ea, and Ena (Ebelt et al., 2005, 056907; Koenig et al., 2005,
087384; Strand et al., 2006, 157017; Wilson and Brauer, 2006, 088933). Results from Wilson
and Brauer (Wilson and Brauer, 2006, 088933) showed for their panel that exposure error is
introduced by l) using Ca instead of Ea and 2) assuming Ea and Ena have the same effects on
health outcomes. There was essentially no association of the effect with Et or Ena. Strand et
al. (2006, 157017) also noted that inclusion of nonambient PM2.5 would not be expected to
change health effects estimates because ambient and nonambient PM2.5 calculations were
not correlated.
Zeger et al. (2000, 001949) pointed out that for community time-series epidemiology,
which analyzes the association between health effects and potential causal factors at the
community level rather than the individual level, it is the correlation of the daily
community-averaged personal exposure to the ambient concentration with daily community
average concentration that is important, not the correlation of each individual's daily
exposure with the daily community average concentration. Thus, the low correlation of
individual daily exposure with the daily community average concentration, as frequently
found in pooled panel exposure studies, is not relevant to error in community time-series
epidemiologic analysis. Sheppard et al. (2005, 079176) also notes that an insufficient
number of total personal exposure samples used in a time-series design would introduce
large classical measurement errors related to high variability in Ena Sheppard et al. (2005,
079176) further maintain that these errors can be minimized by using the average of
concentration measured at community-based ambient monitors. However, overestimation of
the community-averaged exposure by substituting Ca for Ea leads to underestimation of the
effect estimate per unit mass of ambient PM. City-to-city variations in the indoor air
exchange rate, related to differences in climate or housing stock, will cause city-to-city
differences in the health effect endpoint estimate obtained from the study using Ca even if
the endpoint remained the same using the community-averaged exposure.
Relationship between PM and Copollutants
Section 3.8.5 describes studies exploring associations of PM2.5 and SC>42~ with
copollutants O3, NO2, CO, and SO2 observed in Sarnat et al. (2001, 019101) and Schwartz et
al. (2007, 090220). Strong associations between ambient O3, NO2, CO, and SO2 with
personal PM2.5 and S042~ were observed by Sarnat et al. (2001, 019101: 2005, 087531). For
example, O3 may be an indicator of photochemical oxidation products including organic
July 2009
3-250
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
particulate matter. SO2 may be an indicator of Ni emissions from smelters, V from oil fired
power plants, or As, Se or Hg from coal fired power plants. In another example, the HEI
Report on traffic-related health effects (2009, 191009) lists CO, NO2, PM10 and PM2.5 mass,
ultrafine PM count, EC, benzene, and traffic metrics (e.g., count, fuel consumption) all as
potential surrogates for traffic or for the mix of PM and gaseous pollutants in traffic
because all of these pollutants are found in mobile source emissions. Furthermore, in a
multipollutant model, transfer of association can occur by an increase in the slope of a
confounding copollutant and a concurrent decrease in the slope of the truly causal
covariate. This can occur when copollutants are highly correlated with larger error for the
true copollutant, smaller error for the confounder, and correlation between the copollutant
measurement errors (U.S. EPA, 2004, 056905; Zeger et al., 2000, 001949; Zidek et al., 1996,
051879). For these reasons, this is an important area of uncertainty for interpretation of
the multipollutant models discussed in Chapter 6.
3.8.6.6. Compositional Differences
Differences between the composition of ambient PM and the ambient PM that has
infiltrated indoors may affect exposure estimates. Numerous differential infiltration studies
related to indoor-outdoor changes in size distribution and chemical composition are cited in
Sections 3.8.4 and 3.8.5, respectively. Differential infiltration can be caused by impaction of
coarse PM, diffusion of ultrafine PM, and evaporation of semi-volatile components during
infiltration for particles of different sources and sizes. Local meteorology and building air
exchange also influence differential infiltration rates for individual buildings. Differential
infiltration results in differences in PM size distribution and chemical composition between
indoor-ambient PM and outdoor-ambient PM. For instance, Adgate et al. (2007, 156196)
found that outdoor concentrations underestimate personal exposures to trace elements.
Baxter et al. (2007, 092726) showed that V tends to have lower penetration efficiency,
perhaps because metals exist more in the coarse range, while S has penetration efficiency
close to unity. Epidemiologic studies cited in Section 6.6 indicate that significant
associations between health effects estimates and trace metal exposures exist and may be
modified by season. Trace metal penetration efficiency estimates are thus relevant to those
findings. Section 6.6 also discusses significant associations between health effects endpoints
and exposure to EC and OC. After initial emission, traffic-related PM is generally in the
July 2009
3-251
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
accumulation mode with volatile components! accumulation mode PM tends to have the
highest infiltration factors but volatile components may be lost during infiltration (Sarnat
et al., 2006, 089166). If outdoor residential or central-site measurements are used for an
exposure surrogate, differences in indoor and outdoor PM composition related to infiltration
could introduce uncertainty into effects estimates.
Ebelt et al. (2005, 056907) illustrated that exposure error occurs when the PM on one
or more days is not representative of the normal community PM. Section 6.3.2.1 cites Ebelt
et al.'s (2005, 056907) panel study of the association between chronic obstructive pulmonary
disease and PM2.5, PM10-2.5, and PM10 In their analysis, one day of dust from the Gobi
Desert caused an increase in the concentration of fine and coarse PM. When this day was
deleted from the analysis, the relationships were changed and the associations of chronic
obstructive pulmonary disease with PM10 and especially with PM10-2.5, became larger and
more significant. Similar peaks in PM10 concentration have been observed on the Iberian
Peninsula as a result of high transport events carrying dust from the Sahara Desert that
could affect epidemiologic associations on given days (Artinano et al., 2001, 190099).
3.8.6.7. Conclusions
This section presents considerations for exposure assessment and the exposure
misclassification issues that can potentially affect health effects estimates. These issues can
be categorized into six areas: measurement, modeling, spatial variability, temporal
variability, use of surrogates for PM exposure, and compositional differences. Potential
influences of each of these sources on health effect estimates derived from panel, time-
series, and longitudinal epidemiologic studies are described above. Additionally, error
sources often interact with each other and are driven by particle size distribution. For
example, fresh diesel-generated PM is characterized by ultrafine PM that dynamically grow
and change in chemical composition over short time and spatial scales, and lack of spatial
and temporal resolution in measurements or models can result in misclassifying this
exposure (Moore et al., 2009, 191002). For this reason, conclusions regarding ultrafine PM
exposure cannot be drawn from PM2.5 concentration data in part because PM2.5
concentration is more spatially homogeneous across a city. In most circumstances, exposure
error tends to bias a health effect estimate downward (Sheppard et al., 2005, 079176; Zeger
et al., 2000, 001949). Insufficient spatial or temporal resolution to capture true variability
July 2009
3-252
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
and correlation of PM with copollutants are examples of sources of uncertainty that could
widen confidence intervals and so reduce the significance of health effects estimates.
3.9. Summary and Conclusions
3.9.1. Concentrations and Sources of Atmospheric PM
This section summarizes concentrations and sources of atmospheric PM. The
following summaries cover ambient variability and correlations, temporal variability and
copollutant correlations from Section 3.5, measurement techniques from Section 3.4, source
characteristics from Section 3.3, source contributions from Section 3.6 and policy relevant
background concentrations from Section 3.7.
3.9.1.1. Ambient PM Variability and Correlations
Advances in understanding the spatiotemporal distribution of PM mass and
constituents have recently been made, particularly with regard to PM2.5 mass and chemical
composition and ultrafine concentrations. Emphasis in this ISA was on the period from
2005-2007 so that the most recent validated EPAAir Quality System (AQS) data were used.
Note, however, that a majority of U.S. counties were not represented in AQS data, since
their population fell below the regulatory monitoring threshold for PM. Moreover, monitors
reporting to AQS were not uniformly distributed across the U.S. or within counties, and
conclusions drawn from AQS data may not apply equally to all parts of a geographic region.
Furthermore, biases can exist for some PM constituents (and hence total mass) owing to
volatilization losses of nitrates and other semi-volatile compounds, and, conversely, to
retention of particle-bound water by hygroscopic species. The degree of spatial variability in
PM is likely to be region-specific and strongly influenced by region-specific sources and
meteorological and topographic conditions.
Spatial Variability across the U.S.
County-scale, 24-h avg concentration data for PM2.5 between 2005-2007 showed
considerable variability across the U.S. (see Figure 3-9). The highest reported 3-yr avg
concentrations were reported for six counties within the San Joaquin Valley and inland
southern California, as well as Jefferson County, AL (containing Birmingham) and
Allegheny County, PA (containing Pittsburgh). The lowest reported annual average PM2.5
July 2009
3-253
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
concentrations were contained within 237 counties distributed throughout many western
and northern states as well as Florida and the Carolinas. Of the 15 individual CSAs/CBSAs
selected for detailed investigation based on their geographic distribution and importance in
recent health effect studies, the highest mean 24-h PM2.5 concentrations were reported for
Riverside (17 |u.g/m3), Birmingham (16 ug/m3) and Pittsburgh (16 ug/m3); the lowest were
reported for Denver (9 |ug/m3) and Seattle (9 ug/m3).
Since PM10-2.5 is not routinely measured and reported to AQS, co-located lowvolume
PM10 and PM2.5 measurements from the AQS network were used to investigate the spatial
distribution in PM10-2.5. Current data coverage (see Figure 3-10) and measurement errors
limit the ability to draw any meaningful conclusions regarding the large-scale spatial
distribution of PM10-2.5 in urban areas. Only six of the 15 CSAs/CBSAs chosen for closer
investigation had sufficient data for calculating PM10-2.5. In general, in the eastern
metropolitan areas including Atlanta, Boston, Chicago and New York. Most of the mass of
PM10 was in the PM2.5 size fraction, with the highest ratio of PM2.5 to PM10-2.5UI Chicago (14
jug/m3 PM2.5, 5 |ug/m3 PM10-2.5, ratio = 2.8). In contrast, in Denver (9 |ug/m3 PM2.5, 20 |ug/m3
PM102.5, ratio = 0.45) and Phoenix (10 |ug/m3 PM2.5, 22 jug/m3 PM10-2.5, ratio = 0.45) most of
PM10 was in the thoracic coarse mode.
Given the limited information available from AQS for PM10-2.5 and the current
National Ambient Air Quality Standard for PM10, analyses were performed on the more
prevalent PM10 data (see Figure 3-11), acknowledging that PM10 incorporates both thoracic
coarse and fine particles. The highest reported 3-yr avg PM10 concentrations (>51 ug/m3)
occurred in two counties in southern California and five counties in southern Arizona and
central New Mexico. The lowest reported annual average PM10 concentrations (< 20 ug/m3)
were within 114 counties distributed fairly uniformly across the U.S. Of the 15
CSAs/CBSAs investigated, the highest mean 24-h PM10 concentrations was reported for
Phoenix (52 ug/m3), considerably higher than the means for the other CSAs/CBSAs
investigated. The lowest was reported for Boston (17 |ug/m3) with New York, Philadelphia
and Seattle only slightly higher (19 |ug/m3).
Spatial variability in PM2.5 components obtained from the Chemical Speciation
Network (CSN) varied considerably by species. The highest annual average OC
concentrations (>5 ug/m3) were observed in the western and southeastern U.S. (see
Figure 3-12) Concentrations in the West peaked in the fall and winter, while concentrations
in the Southeast peaked anytime between spring and fall. Of the 15 CSAs/CBSAs
July 2009
3-254
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
investigated, OC was the dominant PM2.5 component on an annual basis in the western
cities, ranging from 34% of PM2.5 mass in Los Angeles to 58% in Seattle (see Figure 3-17).
EC exhibited less seasonal variability than OC and was particularly stable in the eastern
half of the U.S. (see Figure 3-13). Annual average EC concentrations greater than 1.5 ug/m3
were present in Los Angeles, Pittsburgh, New York and El Paso. Concentrations of SC>42~
were higher in the eastern U.S. (see Figure 3" 14) resulting from higher SO2 emissions in
the East, compared with the West. There is also considerable seasonal variability with
higher SO r concentrations in the summer months when the oxidation of SO2 proceeds at a
faster rate than during the winter. Of the 15 CSAs/CBSAs selected, sulfate was the
dominant PM2.5 component on an annual basis in the eastern cities, ranging from 42% of
PM2.5 mass in Chicago to 56% in Pittsburgh (see Figure 3" 17). NOa concentrations were
highest in California, with annual averages >4 ug/m3 at many monitoring locations (see
Figure 3-15). There were also elevated concentrations of N03~ in the Midwest (>2 ug/m3),
with wintertime concentrations exceeding 4 ug/m3. In general, NOa was higher in the
winter across the country, resulting from a number of factors including lower temperatures
which favor partitioning into particles! higher relative humidity, mainly in dry areas! lower
sulfate, allowing higher uptake of NO3 ! and residential wood burning in specific areas of
the U.S., especially in the Northwest. Exceptions existed in Los Angeles and Riverside,
where high NOa readings appeared year-round. Crustal material constituted a substantial
fraction of PM2.5 year-round in Phoenix (28%) and Denver (16%), and during the summer in
Houston (26%) (see Figure 3-17).
Clearly there are variations in both PM2.5 mass and composition by city. Such
variability results from numerous controlling variables (e.g., meteorology, the nature of
sources, proximity to sources, topography) that are too poorly characterized on a broad scale
to allow conclusions to be drawn regarding PM2.5 mass and composition across all cities
within a given geographic region.
Spatial Variability on the Urban and Neighborhood Scales
In general, PM2.5 has a longer atmospheric lifetime than PM10-2.5 because larger
particles have a higher gravitational settling velocity. For PM2.5, most metropolitan areas
exhibited high correlations (generally >0.75) out to a distance of 100 km (e.g., Figures 33-25
through 3-27). Notable exceptions were Denver, Los Angeles and Riverside where
correlations dropped below 0.75 somewhere between 20 and 50 km. Insufficient data were
July 2009
3-255
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
available in the 15 metropolitan areas to perform similar analyses for PM10-2.5 using co-
located, low volume FRM monitors. More abundant PM10 data, however, showed larger
declines in inter-monitor correlations as a function of distance (e.g., Figures 3"35 through
3-37) relative to PM2.5. Atlanta, Boston, Denver, Los Angeles, New York City, Philadelphia,
Phoenix, Pittsburgh and Riverside all showed an average correlation of 0.75 at 40 km or
greater monitor separation while Birmingham, Chicago, Detroit, Houston and St. Louis had
correlations that dropped off much more quickly with distance (average correlation of 0.75
at 6 km or less monitor separation). Furthermore, correlations between PM10
concentrations exhibited substantially more scatter relative to PM2.5. Shorter atmospheric
lifetimes for PM10 can result in local emission sources dominating PM10 annual average
mass concentrations at particular monitors. Although the general understanding of PM
differential settling leads to an expectation of greater spatial heterogeneity in the PM10-2.5
fraction relative to the PM2.5 fraction in urban areas, deposition of particles as a function of
size depends strongly on local meteorological conditions, in particular on the degree of
turbulence in the mixing layer. Therefore, the findings from these 15 CSAs/CBSAs may not
apply to all locations or at all times.
Population density and associated building density are also important determinants
of the spatial distribution of PM concentrations. Inter-sampler correlations as a function of
distance between monitors obtained for sampler pairs located less than 4 km apart (i.e., on
a neighborhood scale) showed a shallower slope for PM2.5 than for PM10. The average
correlation was 0.93 for PM2.5, but it dropped to 0.70 for PM10 (see Figure 3-43).
Few studies performed direct comparisons of ultrafine particle measurements at
multiple locations within an urban area. A decrease in the number of ultrafine particles was
demonstrated with shifts from a dominant mode at around 10 nm within 20 m of a freeway
to a flattened dominant mode at around 50 nm at a distance of roughly 100-150 m. At the
same time, accumulation mode particle number concentration remained relatively constant
to within -300 m from the freeway. These findings suggest a high degree of spatial
heterogeneity in ultrafine particles compared with accumulation mode particles on the
urban scale.
July 2009
3-256
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
3.9.1.2.	Temporal Variability
A steady decrease in PM2.5 concentrations from 1999 (the beginning of nationwide
monitoring for PM2.5) to 2007 was observed in all 10 EPA Regions, with the three-yr avg of
the 98th percentile of 24-h PM2.5 concentrations dropping 10% over this time period.
Similar trends in PM10 concentrations show a steady decline from 1988 to 2007 in all 10
EPA Regions.
Using hourly PM observations in the 15 metropolitan areas, diel variation showed
peaks that differ by PM size fraction and region. For PM2.5, a morning peak was observed
starting at approximately f>-()() a.m., corresponding with the start of morning rush hour
(e.g., Figure 3-49b). There was also an evening PM2.5 concentration peak that was broader
than the morning peak and extended into the overnight period, likely reflecting a
combination of evening rush hour and the concentration increase caused by the usual
collapse of the mixed layer after sundown. PM2.5 concentrations in Pittsburgh remained
elevated throughout the night, obscuring the morning peak (see Figure 3"49a).For PM10, all
areas showed a morning and afternoon peak in mean concentrations. The magnitude and
duration of this peak varied considerably by metropolitan area (e.g., Figure 3-50). Since
these figures represent the distribution of hourly observations over a 3-yr period, any
fluctuations or changes in the timing of the daily peaks would result in a broadening of the
curves shown in the diel plot.
Studies indicate that ultrafine particles in urban environments exhibit similar two-
peaked diel patterns in Los Angeles and the San Joaquin Valley as well as in Kawasaki
City, Japan and Copenhagen, Denmark. The afternoon peak in ultrafine particles likely
represents the combination of primary source emissions such as evening rush-hour traffic
and photochemical formation of secondary organic aerosol and sulfate. Several example diel
patterns are shown in Figure 3"51 for the study in Denmark. Comparison between
weekdays and Sundays as well as an urban street canyon site and an urban background
site in this figure suggest traffic is a major source of ultrafine particles within a street
canyon during the morning rush hour.
3.9.1.3.	Correlations between Copollutants
Correlations between PM and gaseous copollutants including SO2, NO2, CO and O3
varied both seasonally and spatially between and within metropolitan areas. On average,
July 2009
3-257
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
PMio and PM2.5 were correlated with each other better than with the gaseous copollutants.
There was relatively little seasonal variability in the mean correlation between PM in both
size fractions and SO2 and NO2. CO, however, showed higher correlations with PM10 and
PM2.5 on average in the winter compared with the other seasons. This seasonality results in
part because a larger fraction of PM is primary in origin during the winter. To the extent
that this primary component of PM is associated with common sources of NO2 and CO, then
higher correlations with these gaseous copollutants are to be expected. Increased
atmospheric stability in colder months would also reinforce these associations.
The correlation between daily maximum 8"h avg O3 and PM showed the highest
degree of seasonal variability with positive correlations on average in the spring, summer
and fall, and negative correlations on average in the winter. This situation arises as the
result of seasonal differences in PM primary emissions and photochemical production of
secondary PM2.5 and O3. However, this relationship is not found in Birmingham, Boston and
St. Louis.
3.9.1.4. Measurement Techniques
The federal reference methods for PM2.5 and PM10 are based on criteria outlined in the
Code of Federal Regulations. They are, however, subject to several limitations that should
be kept in mind when using compliance monitoring data for health outcome studies. FRM
methods are subject to the loss of semi-volatile species such as organic compounds and
ammonium nitrate (especially in the West). Since FRM gravimetric methods involve 24 -h
integrated filter samples, no information is available for variations over shorter averaging
times. However, reliable methods have been developed to measure real-time PM2.5 or PM10
mass concentrations (e.g., FDMS-TEOM). New FRMs and FEMs are available for PM10-2.5
and various methods (dichotomous samplers, cascade impactors, and passive sampling
techniques) are under evaluation to improve PM10-2.5 measurements. Techniques are
available to characterize ultrafine PM mass, surface area, and number concentrations.
Continuous and semi-continuous measurement techniques are also available for PM
species, such as PILS for multiple ion analysis and AMS for multiple component analysis.
Advances have also been achieved in PM organic speciation (e.g., TD-GC-MS). For
additional information see Section 3.4.1.
July 2009
3-258
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
3.9.1.5.	PM Source Characteristics
PM in the atmosphere contains both primary (i.e., emitted directly by sources) and
secondary components, which can be anthropogenic or natural in origin. Secondary
components are produced by the oxidation of precursor gases such as SO2, NOx; and
reactions of acidic products with NH3; and organic compounds.
Developments in the chemistry of formation of SOA indicate that oligomers are likely
a major component of OC in aerosol samples. Recent observations suggest that small, but
still significant quantities of SOA are formed from isoprene oxidation. Gasoline engines
have been found to emit a mix of OC, EC, and nucleation-mode heavy and large polycyclic
aromatic hydrocarbons on which unspent fuel and trace metals condense, while diesel
particles are composed of a soot nucleus on which SO r and hydrocarbons condense. Data
from standard emissions tests in which there is insufficient dilution of fresh exhaust from
combustion sources tend to overestimate the primary component of organic aerosol at the
expense of the semi-volatile components. These semi-volatile components are precursors to
secondary organic aerosol formation and their oxidation results in more oxidized forms of
SOA, both in near source urban environments and further downwind, than previously
considered, both in near source urban environments and further downwind.
3.9.1.6.	Source Contributions to PM
Results of receptor modeling calculations indicate that PM2.5 is produced mainly by
combustion of fossil fuel, either by stationary sources or by transportation. It is apparent
that a relatively small number of source categories, compared to the total number of
chemical species that typically are measured in ambient monitoring source receptor model
studies, are needed to account for the majority of the observed mass of PM in these studies.
Trying to be more specific about contributions from source categories could result in
ambiguity. For example, quite different mobile sources (e.g., trucks and locomotives) rely on
diesel power and ancillary data is required to resolve contributions from these sources. A
compilation of study results shows that secondary sulfate (mainly from EGUs), nitrate
(from the oxidation of NOx emitted mainly from transportation and EGUs), and primary
mobile source categories constitute most of PM2.5 (and PM10) in the East. Fugitive dust,
found mainly in the PM10-2.5 size range, represents the largest source of ambient PM10 in
many locations in the western U.S. Quoted uncertainties in the source apportionment of
July 2009
3-259
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
constituents in ambient aerosol samples typically range from 10 to 50%. A comparison of
source apportionment techniques indicated that the same major source categories of PM2.5
were consistently identified by several independent groups working with the same data
sets. Soil-, sulfate-, residual oil-, and salt-associated mass were most clearly identified by
the groups. Other sources with more ambiguous signatures, such as vegetative burning and
traffic-related emissions were less consistently identified.
Spatial variability in source contributions across urban areas is an important
consideration in assessing the likelihood of exposure error in epidemiologic studies relating
health endpoints to sources. Concepts similar to those for using ambient concentrations as
surrogates for personal exposures apply here. Studies for PM2.5 indicate that intra-urban
variability increases in the following order: regional (e.g., secondary SO r from EGUs)
< area (e.g., on-road mobile sources) < point (e.g., stacks) sources. Only one study was
available for PM10-2.5, indicating a similar ordering, but without a regional component
(resulting from the short lifetime of PM10-2.5 compared to transport times on the regional
scale). Descriptions of receptor models and three dimensional chemistry transport models
are given in Section 3.6.
3.9.1.7. Policy-Relevant Background
The background concentrations of PM that are useful for risk and policy assessments
informing decisions about the NAAQS are referred to as policy-relevant background (PRB)
concentrations. PRB concentrations have historically been defined by EPA as those
concentrations that would occur in the U.S. in the absence of anthropogenic emissions in
continental North America defined here as the U.S., Canada, and Mexico. For this
document, PRB concentrations include contributions from natural sources everywhere in
the world and from anthropogenic sources outside continental North America. Background
concentrations so defined facilitated separation of pollution that can be controlled by U.S.
regulations or through international agreements with neighboring countries from those
that were judged to be generally uncontrollable by the U.S. Over time consideration of
potential broader ranging international agreements may lead to alternative determinations
of which PM source contributions should be considered by EPA as part of PRB.
Contributions to PRB concentrations of PM include both primary and secondary natural
and anthropogenic components. For this document, PRB concentrations for the continental
July 2009
3-260
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
U.S. were estimated using EPA's Community Multi-scale Air Quality (CMAQ) modeling
system, a deterministic, chemical-transport model (CTM), and with GEOS-Chem, a global-
scale model for CMAQ boundary conditions. PRB concentrations of PM2.5 were estimated to
be less than 1 ug/ma on an annual basis, with maximum daily average values in a range
from 3.1 to 20 |ug/m3 and having a peak of 63 ug/ma at the nine national park sites across
the U.S. used to evaluate model performance for this analysis. For further information on
methods used in modeling of PRB concentrations see Section 3.6, and for further
information on the results of calculation of PRB concentrations see Section 3.7.
3.9.2. Human Exposure
This section summarizes the findings from the recent exposure assessment literature,
which include the assessment of exposure to ambient PM, infiltration of ambient PM to
indoor environments, and source apportionment of PM exposure. This summary is intended
to support the interpretation of the findings from epidemiologic studies. For more detailed
explication see Section 3.8.
3.9.2.1. Characterizing Human Exposure
A number of techniques have been applied in the literature to model human exposure
to PM. Several studies have used time-weighted microenvironmental models to define total
or ambient PM exposure. Time-activity diaries or global positioning systems have been
employed to capture the time-basis for those models. Stochastic population exposure
models, such as APEX and SHEDS, are applied for PM exposure risk assessment among
the population. Concentrations from chemistry transport models have also been used to
provide input to the stochastic exposure models at particular locations. LUR models applied
for individual exposures at the intra-urban scale to examine exposure to pollution
surrogates, such as traffic counts, land use, or topographic variables. Source proximity and
kriging have also been applied. GIS-based models have been used to model exposure over
large regions (e.g., for the Nurse's Health Study) using spatial smoothing models of AQS
data and incorporating GIS-based and meteorological covariates. GIS approaches have also
been used for intra-urban scale exposure studies. These methods all have their own uses
and caveats. LUR is an adaptable framework allowing adaptation to localized conditions
but might best be applied in relatively spatially homogeneous areas. In a comparison of
July 2009
3-261
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
LUR with kriging, kriging produced slightly attenuated mortality risk estimates for New
York City, while for Los Angeles, kriging did not exhibit as much spatial variability as LUR.
Source proximity modeling is relatively simple to apply but is limited because other
confounding covariates, such as socioeconomic status, may be related to source proximity
Additionally, subjects' activities may take them away from residences located close to source
New advancements in personal and microenvironmental monitoring techniques have
been reported. Personal monitoring developments include new models of cascade impactors
and cyclones to sample in the ultrafine PM size range and miniature monitors for species
detection. Additionally, new work on microenvironmental modeling using mobile platforms
and GPS technology has been reported. The reader is referred to the 2004 AQCD for
descriptions of most real-time and filter-based personal and microenvironmental PM
monitors currently available (U.S. EPA, 2004, 056905).
3.9.2.2. Spatial Scales of PM Exposure Assessment
Assessing population-level exposure at the urban scale is particularly relevant for
time-series epidemiologic studies, which provide information on the relationship between
health effects and community-average exposure, rather than variations in individual
exposure. The correlation between the PM concentration measured at a central-site
community ambient monitor and the true community average concentration depends on the
spatial distribution of the PM, location of the monitoring site chosen to represent the
community average, and division of the community by terrain features or source locations
into several sub-communities that differ in the temporal pattern of pollution.
Concentrations of S( ) r and some components of SOA measured at central-site monitors are
expected to be uniform in urban areas because of the regional nature of their sources.
However, this is not true for primary components like EC whose sources are strongly
spatially variable in urban areas. Given that roughly 90% of an individual's day is spent
indoors, assessment of exposure to infiltrated ambient SO#-, whose formation and
dispersion also occurs over urban-to-regional scales and whose size distribution is in the
accumulation mode, is commonly used to assess ambient PM2.5 exposure. This technique
has also been applied to assess PM10-2.5 exposure but is likely to have more error than for
PM2.5 because PM10-2.5 is more highly spatially variable than PM2.5. Source apportionment
July 2009
3-262
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
techniques have also been applied to assess urban-scale PM2.5 exposures using community-
based ambient monitoring, outdoor, and indoor samples.
At micro-to-neighborhood scales, heterogeneity of sources and topography may cause
more variability in exposure. This is particularly true for PM10-2.5 and for ultrafine PM, both
of which are more highly spatially variable than PM2.5. Particle chemistry and source
behavior also contribute to spatial heterogeneity of PM concentration. Some studies,
conducted mainly in Europe, have found personal PM2.5 and PM10 exposures for pedestrians
in street canyons to be higher than ambient concentrations measured by urban background
ambient monitors. Likewise, microenvironmental ultrafine PM concentrations were
observed to be substantially higher in near-road environments, street canyons, and tunnels
when compared with other environments in urban areas. In-vehicle ultrafine PM exposures
can also be important. As a result, ambient monitors located at background, central urban,
road side, or near-residential sites might not reflect peak exposures to individuals who
commute.
PM infiltration factors, Fmf, depend on particle size, chemical composition, season, and
region of the country. Infiltration can best be modeled dynamically based on a distribution
of air exchange and deposition PM loss rates rather than being represented by a single
value. There is significant variability within and across regions of the country with respect
to indoor exposures to ambient PM. Infiltrated ambient PM concentrations depend in part
on the ventilation properties of the building or vehicle in which the person is exposed.
Season is important to PM infiltration because it affects the ventilation practices used, and
ambient temperature and humidity conditions affect the transport, dispersion, and size
distribution of PM. Residential air exchange rates have been observed to be higher in
summer for regions with low air conditioning usage, and regional differences in air
exchange rates (Southwest < Southeast < Northeast < Northwest) also reflect ventilation
practices. Differential infiltration occurs as a function of PM size and composition. PM
infiltration is largest for accumulation mode particles, and decreases for ultrafine PM lost
to diffusion and for coarse PM lost through inertial impaction mechanisms. Differential
infiltration by size fraction can affect exposure estimates if not properly characterized.
July 2009
3-263
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
3.9.2.3.	Multicomponent and Multipollutant PM Exposures
Emission inventories and source apportionment studies suggest that sources of PM
exposure vary by region. Comparison of studies performed in the eastern U.S. with studies
performed in the western U.S. suggest that the contribution of SO42- to personal exposure is
higher for the East (16-46%) compared with the West (~4%) and that motor vehicle
emissions and secondary NO3" are larger sources of personal exposure for the West (~9%) as
compared with the East (~4%). Results of source apportionment studies of personal
exposure to SO r indicate that personal S( ) r exposures are mainly attributable to ambient
sources. Source apportionment for OC and EC is difficult because they originate from both
indoor and outdoor sources. Exposure to OC of indoor and outdoor origin can be
distinguished by the presence of aliphatic (Ml groups generated indoors, since outdoor
concentrations of aliphatic CM I are low. Trace metal studies have shown variable results
regarding personal exposure to ambient constituents with significant variation among cities
and over seasons that can be related to incinerator operation, fossil fuel combustion,
biomass combustion (wildfires), and presence of crustal materials in the built environment,
among other sources. Differential infiltration is also affected by variations in particle
composition and volatility. For example EC infiltrates more readily than OC. This can lead
to outdoor-indoor differentials in PM toxicity.
A number of studies have examined whether gaseous copollutants could act as
surrogates for exposure to ambient PM. Several studies have concluded that ambient
concentrations of O3, NO2, and SO2 are associated with the ambient component of personal
exposure to total PM2.5 as opposed to the ambient component of personal exposures to the
gases. However, in some studies this result may have arisen in part because personal
exposure to the gases was often beneath the detection limits of the personal monitoring
devices. Thus, the evidence that ambient gases can be considered surrogates of PM2.5
exposure is mixed. It is likely that associations between ambient gases and personal
exposure to PM2.5 of ambient origin exist, but they are complex and vary by season and
location.
3.9.2.4.	Implications for Epidemiologic Studies
The importance of exposure error varies with study design based on the spatial and
temporal aspects of the design. For PM epidemiology studies, source characteristics,
July 2009
3-264
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
particle size distribution, and particle composition are also important factors in
interpreting exposure error for an epidemiology study. Potential sources of error that could
influence estimates of PM exposure include measurements, surrogacy, modeling, spatial
variability, temporal variability, and compositional differences.
PM exposure estimates are subject to monitoring and modeling errors. Ambient and
personal exposure monitoring errors can bias health effects estimates if the error is
strongly correlated with the measurements of concentration. This can be an issue for
sampling semi-volatile organic compounds in PM, especially where PM exposures in cities
with different PM composition are compared. Ambient monitor height also affects estimates
of exposure because PM concentration varies as a function of height. Within a street
canyon, changes in wind direction and speed cause significant variability over a small
distance, with findings showing up to a two order of magnitude change in benzo[a]pyrene
concentrations within the depth of a street canyon. Wind tunnel studies have shown street
canyon effects exist for suburban and not just for downtown, heavily urbanized settings.
Additionally, model-based exposure estimates are subject to errors related to the spatial
resolution of the modeling technique and the measurement-based inputs used.
Variations in PM and its components could lead to errors in using ambient PM
measures as surrogates for exposures to PM. PM2.5 concentrations are relatively well-
correlated across monitors in the urban areas examined. Correlation coefficients tend to be
lower, and concentration differences tend to be higher between PM10 monitoring sites than
between PM2.5 monitoring sites. Likewise, studies have shown ultrafine PM to be more
spatially variable across urban areas. Even if PM2.5, PM10-2.5, and ultrafine PM
concentrations measured at sites within an urban area are highly correlated, significant
differences in their concentrations can occur on any given day. The degree of urban-scale
spatial variability in PM concentrations varies across the country and with size fraction.
Current information suggests that ultrafine PM, PM10-2.5, and many PM components are
more spatially variable than PM2.5. These factors should be considered in using data
obtained from monitoring networks to estimate community-scale human exposure to
ambient PM, and caution should be exercised in extrapolating conclusions obtained from
one urban area to another.
Community time-series epidemiologic studies use the average community PM
concentration as a surrogate for the average personal exposure to ambient PM. The
resulting health effect risk estimate, based on the average community ambient
July 2009
3-265
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
concentration, differs from the risk that would be estimated if the average community
ambient exposure were used in the epidemiologic study. However, the risk estimate based
on the ambient concentration gives the change in health effects resulting from a change in
ambient PM concentration and is, therefore, an appropriate measure for risk assessment
and risk management. Variations in ambient concentrations across a community, variations
in individual ambient exposures around the community average, and seasonal or daily
variation in the ambient exposure estimate may increase standard errors of PM health
effects estimates, making it more difficult to detect a true underlying association between
the correct exposure metric and the health outcome studied. Likewise, sampling time
interval and lag time selection both determine whether an epidemiologic model captures
the phenomena of interest with sufficient resolution. The use of the community average
ambient PM2.5 concentration as a surrogate for the community average personal exposure
to ambient PM2.5 is not expected to change the principal conclusions from PM2.5
epidemiologic studies that use community average health and pollution data. Several
recent studies support this by showing how the ambient component of personal exposure to
PM2.5 could be estimated using various tracer and source apportionment techniques and
that it is highly correlated with ambient concentrations of PM2.5. These studies also show
that the non-ambient component of personal exposure to PM2.5 is basically uncorrelated
with ambient PM2.5 concentrations. For long-term studies that use differences in long-term
community average ambient PM concentrations as an exposure metric, the effect of possible
community-to-community differences in the average ambient exposure factor or in the
average non-ambient exposure are less understood. For panel epidemiologic studies, the
most appropriate exposure metric may depend on the health outcome measured. However,
sufficient information should be obtained to enable determining the association of the
health outcome with ambient concentration, ambient exposure, non-ambient exposure, and
total personal exposure.
Health effect estimates are often not clearly associated with a given PM constituent
because questions persist whether the measured surrogate or a different component are
actually responsible for the effect. Two studies have suggested that if PM composition data
are accurate, resulting health effects estimates are more significant than when the PM are
poorly characterized. Differences between composition of outdoor and indoor ambient PM
may also cause error in exposure assessment related either to differential losses of ultrafine
or coarse PM from diffusion, evaporation of semi-volatile PM, or impaction. The resulting
July 2009
3-266
DRAFT-DO NOT CITE OR QUOTE

-------
1	differences in PM size distribution and chemical composition between indoor-ambient PM
2	and outdoor-ambient PM are expected to cause differences in toxicity that could affect
3	health effects outcomes. Lack of information regarding these relationships adds uncertainty
4	to the health effects estimate.
July 2009
3-267
DRAFT-DO NOT CITE OR QUOTE

-------
Chapter 3 References
ACGIH. (2005). TLVs and BEIs: Based on the documentation of the threshold limit values for chemical
substances and physical agents and biological exposure indices. 156188
Adams HS; Nieuwenhuijsen MJ; Colvile RN; McMullen MAS; Khandelwal P. (2001). Fine particle (PM25)
personal exposure levels in transport microenvironments, London, UK. Sci Total Environ, 279- 29"44.
019350
Adar SDi Davey Mi Sullivan JRi Compher Mi Szpiro Ai Liu L-J. (2008). Predicting airborne particle levels
aboard Washington State school buses . Atmos Environ, 42- 7590-7599. 191200
Adgate JLi Mongin SJi Pratt GCi Zhang Ji Field MPi Ramachandran Gi Sexton K. (2007). Relationships
between personal, indoor, and outdoor exposures to trace elements in PM(2.5). Sci Total Environ, 386- 21"
32.156196
Al-Horr Ri Samanta Gi Dasgupta PK. (2003). A Continuous Analyzer for Soluble Anionic Constituents and
Ammonium in Atmospheric Particulate Matter. Environ Sci Technol, 37- 5711-5720. 153951
Albinet Ai Leoz-Garziandia Ei Budzinski Hi Villenave E. (2007). Sampling precautions for the measurement of
nitrated polycyclic aromatic hydrocarbons in ambient air. Atmos Environ, 41' 4988-4994. 154426
Alexis Ai Garcia Ai Nystrom Mi Rosenkranz K. (2001). The 2001 California almanac of emissions and air
quality. Retrieved , from . 079886
Allen GAi Harrison Di Koutrakis P. (2001). A New Method for Continuous Measurement of Sulfate in the
Ambient Atmosphere. Presented at American Association for Aerosol Research, 20th Annual Conference.
156205
Allen Ri Larson Ti Sheppard Li Wallace LLiu L-JS. (2003). Use of real-time light scattering data to estimate the
contribution of infiltrated and indoor-generated particles to indoor air. Environ Sci Technol, 37- 3484"
3492. 053578
Allen Ri Wallace Li Larson Ti Sheppard Li Liu LJ. (2004). Estimated hourly personal exposures to ambient and
nonambient particulate matter among sensitive populations in Seattle, Washington. , 54- 1197-1211.
190089
Andracchio A; Cavicchi Ci Tonelli Di Zappoli S. (2002). A new approach for the fractionation of water-soluble
organic carbon in atmospheric aerosols and cloud drops. Atmos Environ, 36- 5097-5107. 155657
Andreae MO; Gelencser A. (2006). Black carbon or brown carbon? The nature of light-absorbing carbonaceous
aerosols. , 6: 3131-3148. 156215
Appel K; Bhave PV; Gilliland AB; Sarwar G; Roselle SJ. (2008). Evaluation of the community multiscale air
quality (CMAQ) model version 4.5- Sensitivities impacting model performance! Part IIa€"particulate
matter. Atmos Environ, 42- 6057-6066. 155660
Appel KWi Gilliland A; Eder B. (2005). An operational evaluation of the 2005 release of models-3 CMAQ version
45. National Oceanic and Atmospheric Administration—Air Resources. Washington DC. 089227
Arhami Mi Kuhn Ti Fine PMi Delfino RJi Sioutas C. (2006). Effects of Sampling Artifacts and Operating
Parameters on the Performance of a Semicontinuous Particulate Elemental Carbon/Organic Carbon
Monitor. Environ Sci Technol, 40- 945-954. 156224
Arnold JRi Dennis RLi Tonnesen GS. (2003). Diagnostic evaluation of numerical air quality models with
specialized ambient observations' testing the Community Multiscale Air Quality modeling system
(CMAQ) at selected SOS 95 ground sites. Atmos Environ, 37- 1185-1198. 087579
Arnott WPi Moosmiiller Hi Sheridan PJi Ogren JAi Raspet Ri Slaton WVi Hand JLi Kreidenweis SMi Collett JL
Jr. (2003). Photoacoustic and filter-based ambient aerosol light absorption measurements' Instrument
comparisons and the role of relative humidity. J Geophys Res, 108' AAC 15-1. 037711
Artinano Bi Querol Xi Salvador Pi Rodriguez Si Alonso DGi Alastuey A. (2001). Assessment of airborne
particulate levels in Spain in relation to the new EU-directive . Atmos Environ, 35' S43-S53. 190099
Auvray Mi Bey I. (2005). Long-range transport to Europe' Seasonal variations and implications for the
European ozone budget. J Geophys Res, 110' D11303.l-D11303.22. 156237
Ayers GP. (2004). Potential for simultaneous measurement of PM10, PM25 and PMI for air quality monitoring
purposes using a single TEOM. Atmos Environ, 38: 3453-3458. 097440
Note^ Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health
and Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in
the process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009	3-268	DRAFT - DO NOT CITE OR QUOTE

-------
Bae MS. (2007). Intercomparison of Real Time Ammonium Measurements at Urban and Rural Locations in New
York. Aerosol Sci Technol, 41' 329-341. 155669
Bae MS; Schauer JJ; Deminter JT; Turner JR. (2004). Hourly and daily patterns of particle-phase organic and
elemental carbon concentrations in the urban atmosphere. J Air Waste Manag Assoc, 54- 823-833.
156243
Bae MS; Schwab JJ; Zhang Q; Hogrefe O; Demerjian KL; Weimer S; Rhoads K; Orsini D; Venkatachari P;
Hopke PK. (2007). Interference of organic signals in highly time resolved nitrate measurements by low
mass resolution aerosol mass spectrometry. J Geophys Res, 112- D22305-D22305 (l"16). 156244
Baldauf R; Thoma E; Hays M; Shores R; Kinsey J; Gullett B; Kimbrough S; Isakov V; Long T; Snow R; Khlystov
A; Weinstein J; Chen FL; Seila R; Olson D; Gilmour I; Cho SH; Watkins N; Rowley P; Bang J. (2008).
Traffic and meteorological impacts on near-road air quality- Summary of methods and trends from the
raleigh near-road study. , 58: 865-878. 190239
Baldauf R; Thoma E; Khlystov A; Isakov V; Bowker G; Long TJ Snow R. (2008). Impacts of noise barriers on
near-road air quality. Atmos Environ, 42: 7502-7507. 191017
Baldauf R; Watkins N; Heist D; Bailey C; Rowley P; Shores R. (2009). Near-road air quahty monitoring: Factors
affecting network design and interpretation of data. , 2: 1"9. 191766
Baldauf RW; Gabele P; Crews W', Snow R; Cook JR. (2005). Criteria and air-toxic emissions from in-use
automobiles in the national low-emission vehicle program. , 55' 1263-1268. 191184
Barn P; Larson T; Noullett M; Kennedy S; Copes R; Brauer M. (2008). Infiltration of forest fire and residential
wood smoke: an evaluation of air cleaner effectiveness. J Expo Sci Environ Epidemiol, 18: 503-511.
156252
Barreto RP; Albuquerque FC; Netto ADP. (2007). Optimization of an improved analytical method for the
determination of 1-nitropyrene in milligram diesel soot samples by high-performance liquid
chromatography—mass spectrometry. J Chromatogr, 1163: 219-227. 155676
Bauer H; Schueller E; Weinke G; Berger A; Hitzenberger R; Marr I; Puxbaum H. (2008). Significant
contributions of fungal spores to the organic carbon and to the aerosol mass balance of the urban
atmospheric aerosol. Atmos Environ, 42: 5542-5549. 189986
Baxter LK; Barzyk TM; Vette VF; Croghan C; Williams RW. (2008). Contributions of diesel truck emissions to
indoor elemental carbon concentrations in homes in proximity to Ambassador Bridge . Atmos Environ,
42: 9080-9086 . 191194
Baxter LK; Clougherty JE; Laden F; Levy JI. (2007). Predictors of concentrations of nitrogen dioxide, fine
particulate matter, and particle constituents inside of lower socioeconomic status urban homes. J Expo
Sci Environ Epidemiol, 17: 433-444. 092726
Baxter LK; Clougherty JE; Paciorek CJ; Wright RJ; Levy JI. (2007). Predicting residential indoor
concentrations of nitrogen dioxide, fine particulate matter, and elemental carbon using questionnaire
and geographic information system based data. Atmos Environ, 41: 6561-6571. 092725
Behrentz E; Fitz DR; Pankratz DV; Sabin LD; Colome SD; Fruin SA; Winer AM. (2004). Measuring self-
pollution in school buses using a tracer gas technique. Atmos Environ, 38: 3735-3746. 155682
Bein KJ; Zhao Y; Wexler AS; Johnston MV. (2005). Speciation of size-resolved individual ultrafine particles in
Pittsburgh, Pennsylvania. J Geophys Res, 110: D07S05. 156265
Bell ML; Dominici F; Ebisu K; Zeger SL; Samet JM. (2007). Spatial and Temporal Variation in PM2. 5 Chemical
Composition in the United States for Health Effects Studies. , 115: 989. 155683
Bell ML; Ebisu K; Peng RD; Dominici F. (2009). Adverse health effects of particulate air pollution: modification
by air conditioning. Epidemiology, in press: in press. 191007
Bell ML; Ebisu K; Peng RD; Walker J; Samet JM; Zeger SL; Dominic F. (2008). Seasonal and regional short-
term effects of fine particles on hospital admissions in 202 U.S. counties, 1999-2005. Am J Epidemiol,
168: 1301-1310. 156266
Benkovitz CM; Scholtz MT; Pacyna J; Tarrason L; Dignon J; Voldner EC; Spiro PA; Logan JA; Graedel TE.
(1996). Global gridded inventories of anthropogenic emissions of sulfur and nitrogen: North Atlantic
Regional Experiment(NARE). J Geophys Res, 101: 29239-29253. 156267
Bennett DH; Koutrakis P. (2006). Determining the infiltration of outdoor particles in the indoor environment
using a dynamic model. J Aerosol Sci, 37: 766-785. 089184
Binkowski F; Roselle S. (2003). Models-3 Community Multiscale Air Quality(CMAQ) model aerosol component
1. Model description. , 108: np. 191769
Binkowski FS; Arunachalam S; Adelman Z; Pinto JP. (2007). Examining photolysis rates with a prototype
online photolysis module in CMAQ. , 46: 1252-1256. 090563
July 2009
3-269
DRAFT-DO NOT CITE OR QUOTE

-------
Biswas S> Hu Si Verma Vi Herner Ji Robertson Wi Ayala Ai Sioutas C. (2008). Physical properties of particulate
matter (PM) from late model heavy-duty diesel vehicles operating with advanced PM and NOx emission
control technologies. Atmos Environ, 42: 5622-5634. 189969
Biswas Si Verma Vi Schauer JJi Sioutas C. (2009). Chemical speciation of PM emissions from heavy-duty diesel
vehicles equipped with diesel particulate filter (DPF) and. Atmos Environ, 43- 1917-1925. 189880
Blanchard Pi Brook JRi Brazil P. (2002). Chemical characterization of the organic fraction of atmospheric
aerosol at two sites in Ontario, Canada. J Geophys Res, 107: ICC10.1-ICC10.8 . 047598
Blando JDi Turpin BJ. (2000). Secondary organic aerosol formation in cloud and fog droplets: a literature
evaluation of plausibility. Atmos Environ, 34: 1623-1632. 155692
Bond TCi Streets DGi Yarber KF; Nelson SMi Woo J-Hi KLimont Z. (2004). A technology-based global inventory
of black and organic carbon emissions from combustion. J Geophys Res, 109: D14203. 056389
Borak Ji Sirianni Gi Cohen HJi Chemerynski Si Wheeler R. (2003). Comparison of NIOSH 5040 Method versus
Aethalometer™ to Monitor Diesel Particulate in School Buses and at Work Sites. AIHA J, 64: 260-268.
156284
Borrego C> Tchepel Oi Costa AM; Martins Hi Ferreira Ji Miranda AI. (2006). Traffic-related particulate air
pollution exposure in urban areas. Atmos Environ, 40: 7205-7214. 155697
Bortnick SMi Coutant BWi Eberly SI. (2002). Using continuous PM2.5 monitoring data to report an air quality
index. J Air Waste Manag Assoc, 52: 104-112. 156285
Brauer Mi Hoek Gi van Vhet Pi Meliefste Ki Fischer Pi Gehring Ui Heinrich Ji Cyrys Ji Bellander Ti Lewne Mi
Brunekreef B. (2003). Estimating long-term average particulate air pollution concentrations: application
of traffic indicators and geographic information systems. , 14: 228-239. 155702
Braun A. (2009). Two-process model for the atmospheric weathering, oxidation and ageing of diesel soot.
Geophys Res Lett, 36: L07810. 189997
Briggs D. (2000). Exposure assessment. In Spatial Epidemiology: Methods and Applications (pp. 335-359).
Oxford: Oxford University Press. 191772
Briggs DJi Collins Si Elliott Pi Fischer Pi Kingham Si Lebret Ei Pryl Ki Van Reeuwijk Hi Smallbone Ki Van Der
Veen A. (1997). Mapping urban air pollution using GIS: a regression-based approach. Int J Geogr Inform
Sci, 11: 699-718. 025950
Briggs DJi de Hoogh Ki Morris C> Gulliver J. (2008). Effects of travel mode on exposures to particulate air
pollution. Environ Int, 34: 12-22. 156294
Brock CAi Hudson PK; Lovejoy ER; Sullivan A; Nowak JB; Huey LG; Cooper OR; Cziczo DJ; de Gouw J;
Fehsenfeld FC. (2004). Particle characteristics following cloud-modified transport from Asia to North
America. J Geophys Res, 109: D23S26. 156295
Brook JR; Sirois A; Clarke JF. (1996). Comparison of dry deposition velocities for S02, HNO3, and S04(2")
estimated with two inferential models. Water Air Soil Pollut, 87: 205-218. 024023
Brown AS; Yardley RE; Quincey PG; Butterfield DM. (2006). Studies of the effect of humidity and other factors
on some different filter materials used for gravimetric measurements of ambient particulate matter.
Atmos Environ, 40: 4670-4678. 097665
Bruno P; Caselli M; de Gennaro G; Tutino M. (2007). Determination of polycyclic aromatic hydrocarbons (PAHs)
in particulate matter collected with low volume samplers. Talanta, 72: 1357-1361. 155706
Burke JM; Zufall MJ; Ozkaynak H. (2001). A population exposure model for particulate matter: case study
results for PM25 in Philadelphia, PA. J Expo Sci Environ Epidemiol, 11: 470-489. 014050
Burnett RT; Goldberg MS. (2003). Size-fractionated particulate mass and daily mortality in eight Canadian
cities. 042798
Butler AJ; Andrew MS; Russell AG. (2003). Daily sampling Of PM2. 5 in Atlanta: Results of the first year of the
Assessment of Spatial Aerosol Composition. J Geophys Res, 108: 8415. 156313
Buzorius G; Hameri K; Pekkanen J; Kulmala M. (1999). Spatial variation of aerosol number concentration in
Helsinki city. Atmos Environ, 33: 553-565. 081205
Bytnerowicz A; Miller PR; Olszyk DM. (1987). Dry deposition of nitrate, ammonium and sulfate to a Ceanothus
crassifolius canopy and surrogate surfaces. Atmos Environ, 21: 1749-1757. 036493
Byun D; Schere KL. (2006). Review of the governing equations, computational algorithms, and other
components of the models-3 community multiscale air quality (CMAQ) modeling system. Appl. Mech.
Rev., 59: 51-77. 090560
Byun DW; Ching JKS. (1999). Science Algorithms of the EPA Models-3 Community Multiscale Air Quality
(CMAQ) Modeling System. 156314
Cabada JC; Rees S; Takahama S; Khlystov A; Pandis SN; Davidson CI; Robinson AL. (2004). Mass size
distributions and size resolved chemical composition of fine particulate matter at the Pittsburgh
supersite. Atmos Environ, 38: 3127-3141. 148859
July 2009
3-270
DRAFT-DO NOT CITE OR QUOTE

-------
Cadle SH; Mulawa P; Groblicki P; Laroo C> Ragazzi RA; Nelson Ki Gallagher G> Zielinska B. (2001). Irruse
light-duty gasohne vehicle particulate matter emissions on the FTP, REP05, and UC cycles [final report].
017192
Cadle SH; Mulawa PA; Hunsanger EC; Nelson K; Ragazzi RA; Barrett R; Gallagher GL; Lawson DR; Knapp
KT; Snow R. (1999). Composition of light-duty motor vehicle exhaust particulate matter in the Denver,
Colorado area. Environ Sci Technol, 33- 2328-2339. 007636
Carter WPL. (1990). A detailed mechanism for the gas-phase atmospheric reactions of organic compounds.
Atmos Environ, 24: 481-518. 042893
Case MW; Williams R; Yeatts K; Chen F"L; Scott J; Svendsen E; Devlin RB. (2008). Evaluation of a direct
personal coarse particulate matter monitor. Atmos Environ, 42(19)' 4446-4452. 155149
Cass GR. (1998). Organic molecular tracers for particulate air pollution sources. Trends Analyt Chem, 17- 356"
366. 155716
Cass GR; Hughes LA; Bhave P; Kleeman MJ; Allen JO; Salmon LG. (2000). The chemical composition of
atmospheric ultrafine particles. Proc Biol Sci, 358: 2581-2592. 020680
Chakrabarti B; Fine PM; Delfino R; Sioutas C. (2004). Performance evaluation of the active-flow personal
DataRAM PM25 mass monitor (Thermo Anderson pDR-1200) designed for continuous personal exposure
measurements. Atmos Environ, 38: 3329-3340. 147867
Chang C-H; Meroney RN. (2003). Concentration and flow distributions in urban street canyons: wind tunnel
and computational data. J Wind Eng IndAerod, 91: 1141-1154. 090298
Chang CT; Tsai CJ. (2003). A model for the relative humidity effect on the readings of the PM10 beta-gauge
monitor. J Aerosol Sci, 34: 1685-1697. 155718
Chang L-T; Koutrakis P; Catalano PJ; Suh HH. (2003). Assessing the importance of different exposure metrics
and time-activity data to predict 24"h personal PM25 exposures. J Toxicol Environ Health A, 66: 1825"
1846. 053789
Charron A; Harrison RM; Moorcroft S; Booker J. (2004). Quantitative interpretation of divergence between
PM10 and PM25 mass measurement by TEOM and gravimetric (Partisol) instruments. Atmos Environ,
38: 415-423. 053849
Chen FL; Williams R; Svendsen E; Yeatts K; Creason J; Scott J; Terrell D; Case M. (2007). Coarse particulate
matter concentrations from residential outdoor sites associated with the North Carolina Asthma and
Children's Environment Studies (NC-ACES). Atmos Environ, 41: 1200-1208. 147318
Chen SJ; Hsieh LT; Kao MJ; Lin WY; Huang KL; Lin CC. (2004). Characteristics of particles sampled in
southern Taiwan during the Asian dust storm periods in 2000 and 2001. Atmos Environ, 38: 5925-5934.
147143
Chiapello I; Moulin C; Prospero JM. (2005). Understanding the long-term variability of African dust transport
across the Atlantic as recorded in both Barbados surface concentrations and large-scale Total Ozone
Mapping Spectrometer (TOMS) optical thickness. J Geophys Res, 110: D18S10. 156339
Ching J; Herwehe J; Swall J. (2006). On joint deterministic grid modeling and sub-grid variability conceptual
framework for model evaluation. Atmos Environ, 40: 4935-4945. 090300
Chow JC. (2007). The application of thermal methods for determining chemical composition of carbonaceous
aerosols: A review. , 42: 1521-1541. 157209
Chow JC; Doraiswamy P; Watson JG; Chen LW; Ho SS; Sodeman DA. (2008). Advances in integrated and
continuous measurements for particle mass and chemical composition. J Air Waste Manag Assoc, 58:
141-163. 156355
Chow JC; Watson JG; Chen LW; Chang MC; Robinson NF; Trimble D; Kohl S. (2007). The IMPROVE_A
temperature protocol for thermal/optical carbon analysis: maintaining consistency with a long-term
database. J Air Waste Manag Assoc, 57: 1014-1023. 156354
Chow JC; Watson JG; Chen LWA; Arnott WP; Moosmuller H; Fung K. (2004). Equivalence of elemental carbon
by thermal/optical reflectance and transmittance with different temperature protocols. Environ Sci
Technol, 38: 4414-4422. 156347
Chow JC; Watson JG; Louie PKK; Chen LWA; Sin D. (2005). Comparison of PM2.5 carbon measurement
methods in Hong Kong, China. , 137: 334-344. 155728
Chow JC; Watson JG; Park K; Lowenthal DH; Robinson NF; Magliano KA. (2006). Comparison of particle light
scattering and fine particulate matter mass in central California. J Air Waste Manag Assoc, 56: 398-410.
099031
Chu D; Kaufman Y; Zibordi G; Chern J; Mao J; Li C; Holben B. (2003). Global monitoring of air pollution over
land from the Earth Observing System-Terra Moderate Resolution Imaging Spectroradiometer (MODIS).
, 108: ACH4.1-ACH4.18. 190049
July 2009
3-271
DRAFT-DO NOT CITE OR QUOTE

-------
Chung A; Herner JD; KLeeman MJ. (2001). Detection of alkaline ultrafine atmospheric particles at Bakersfield,
California. Environ Sci Technol, 35: 2184-2190. 017105
Claeys M> Graham B; Vas G> Wang Wi Vermeylen R; Pashynska V; Cafmeyer J; Guyon P; Andreae MO; Artaxo
P> Maenhaut W. (2004). Formation of secondary organic aerosols through photooxidation of isoprene. ,
303: 1173-1176. 058608
Clarke JF; Edgerton ES; Martin BE. (1997). Dry deposition calculations for the clean air status and trends
network. Atmos Environ, 31: 3667-3678. 025022
Clough WS. (1975). The deposition of particles on moss and grass surfaces. Atmos Environ, 9: 1113-1119.
070850
Cohen M; Adar S; Allen R; Avol E; Curl C; Gould T; Hardie D; Ho A; Kinney P; Larson T. (2009). Approach to
estimating participant pollutant exposures in the multi-ethnic study of atherosclerosis and air pollution
(MESA Air). , x: 709-742. 190639
Conny JM; Khnedinst DB; Wight SA; Paulsen JL. (2003). Optimizing thermal-optical methods for measuring
atmospheric elemental (black) carbon: A response surface study. Aerosol Sci Technol, 37: 703-723.
145948
Conny JM; Norris GA; Gould TR. (2009). Factorial-based response-surface modeling with confidence intervals
for optimizing thermal-optical transmission analysis of atmospheric black carbon. Anal Chim Acta, 635:
144-156. 191999
Cooke WF; Liousse C; Cachier H; Feichter J. (1999). Construction of a 1 0 x 1 0 fossil fuel emission data set for
carbonaceous aerosol and implementation and radiative impact in the ECHAM 4 model. J Geophys Res,
104: 22137-22162. 156365
Corburn J. (2007). Urban land use, air toxics and public health: Assessing hazardous exposures at the
neighborhood scale. Environ Impact Assess Rev, 27: 145-160. 155738
Cowburn AS; Condliffe AM; Farahi N; Summers C; Chilvers ER. (2008). Advances in neutrophil biology: clinical
implications. , 134: 606-12. 191142
Crimmins BS; Baker JE. (2006). Improved GC/MS methods for measuring hourly PAH and nitro-PAH
concentrations in urban particulate matter. Atmos Environ, 40: 6764-6779. 097008
Crist KC; Liu B; Kim M; Deshpande SR; John K. (2008). Characterization of fine particulate matter in Ohio:
Indoor, outdoor, and personal exposures. Environ Res, 106: 62-71. 156372
Cyrys J; Heinrich J; Hoek G; Meliefste K; Lewne M; Gehring U; Bellander T; Fischer P; Van Vliet P; Brauer M;
Wichmann H"E; Brunekreef B. (2003). Comparison between different traffic-related particle indicators:
elemental carbon (EC), PM25 mass, and absorbance. J Expo Sci Environ Epidemiol, 13: 134-143. 049634
Dasch JM. (1987). Measurement of dry deposition to surfaces in deciduous and pine canopies. , 44: 261-277.
036496
Dasgupta PK; Idowu AD; Li J. (2007). Method and Apparatus for Analyzing Arsenic Concentrations Using Gas
Phase Ozone Chemiluminescence. 156383
Davidson CI; Wu Y"L. (1990). Dry deposition of particles and vapors. In Lindberg, S. E.; Page, A. L.; Norton, S.
A. (Ed.),Acidic precipitation: v. 3, sources, deposition, and canopy interactionsNew York, NY: Springer-
Verlag. 036799
de Gouw JA; Brock CA; Atlas EL; Bates TS; Fehsenfeld FC; Goldan PD; Holloway JS; Kuster WC; Lerner BM;
Matthew BM. (2008). Sources of particulate matter in the northeastern United States in summer: 1.
Direct emissions and secondary formation of organic matter in urban plumes. , 113: np. 191757
deCastro BR; Sax AN; Chillrud SN; Kinney PL; Spengler JD. (2007). Modeling time-location patterns of inner-
city high school students in New York and Los Angeles using a longitudinal approach with generalized
estimating equations. J Expo Sci Environ Epidemiol, 17: 233-247. 190996
Decesari S; Facchini MC; Fuzzi S; Tagliavini E. (2000). Characterization of water-soluble organic compounds in
atmospheric aerosol: Anew approach. J Geophys Res, 105: 1481-1489. 155748
Delfino RJ; Staimer N; Gillen D; Tjoa T; Sioutas C; Fung K; George SC; KLeinman MT. (2006). Personal and
ambient air pollution is associated with increased exhaled nitric oxide in children with asthma. Environ
Health Perspect, 114: 1736-1743. 090745
Demokritou P; Lee S; Koutrakis P. (2004). Development and Evaluation of a High Loading PM2.5 Speciation
Sampler. Aerosol Sci Technol, 38: 111-119. 186901
Demokritou P; Lee SJ; Ferguson ST; Koutrakis P. (2004). A compact multistage (cascade) impactor for the
characterization of atmospheric' aerosols. J Aerosol Sci, 35: 281-299 . 190115
July 2009
3-272
DRAFT-DO NOT CITE OR QUOTE

-------
Dentener F; Stevenson D; Ellingsen K; Van Noije T; Schultz M; Amann M; Atherton C; Bell N; Bergmann D;
Bey L Bouwman L; Butler Ti Cofala J; Collins B; Drevet J; Doherty R; Eickhout B; Eskes H; Fiore A;
Gauss Mi Hauglustaine Di Horowitz Li Isaksen ISA; Josse B; Lawrence M; Krol MJ Lamarque JF;
Montanaro V; Muller JF; Peuch VH; Pitari G; Pyle J; Rast S; Rodriguez J; Sanderson M; Savage NH;
Shindell D; Strahan Si et al. (2006). The global atmospheric environment for the next generation.
Environ Sci Technol, 40: 3586-3594. 088434
Derwent RGi Collins WJi Johnson CEi Stevenson DS. (2001). Transient behaviour of tropospheric ozone
precursors in a global 3-D CTM and their indirect greenhouse effects. Clim Change, 49: 463-487. 047912
Diapouli Ei Chaloulakou Ai Spyrellis N. (2007). Levels of ultrafine particles in different microenvironments--
implications to children exposure. Sci Total Environ, 388: 128-136. 156397
Diapouli Ei Grivas Gi Chaloulakou Ai Spyrellis N. (2008). PM10 and ultrafine particles counts in-vehicle and
on-road in the Athens area . , 8- 89"97. 190119
Docherty KSi Wu Wi Lim YBi Ziemann PJ. (2005). Contributions of organic peroxides to secondary aerosol
formed from reactions of monoterpenes with 03. Environ Sci Technol, 39: 4049-4059. 087613
Dockery DWi Pope CA III; Xu X; Spengler JD; Ware JH; Fay ME; Ferris BG Jr; Speizer FE. (1993). An
association between air pollution and mortality in six US cities. , 329: 1753-1759. 044457
Dodge MC. (1977). Combined use of modeling techniques and smog chamber data to derive ozone-precursor
relationships. 038646
Dominici F; Peng RD; Bell ML; Pham L; McDermott A; Zeger SL; Samet JL. (2006). Fine particulate air
pollution and hospital admission for cardiovascular and respiratory diseases. JAMA, 295: 1127-1134.
088398
Dominici F; Zeger SL; Samet JM. (2000). A measurement error model for time-series studies of air pollution and
mortality. , l: 157-175. 005828
Drewnick F; Jayne JT; Canagaratna M; Worsnop DR; Demerjian KL. (2004). Measurement of Ambient Aerosol
Composition During the PMTACS-NY 2001 Using an Aerosol Mass Spectrometer. Part II: Chemically
Speciated Mass Distributions. Aerosol Sci Technol, 38: 104-117. 155755
Drewnick F; Schwab J; Jayne J; Canagaratna M; Worsnop D; Demerjian K. (2004). Measurement of Ambient
Aerosol Composition During the PMTACS-NY 2001 Using an Aerosol Mass Spectrometer. Part I: Mass
Concentrations. Aerosol Sci Technol, 38: 92-103. 155754
Drewnick F; Schwab JJ; Hogrefe O; Peters S; Husain L; Diamond D; Weber R; Demerjian KL. (2003).
Intercomparison and evaluation of four semi-continuous PM2.5 sulfate instruments. Atmos Environ, 37:
3335-3350. 099160
Ebelt ST; Wilson WE; Brauer M. (2005). Exposure to ambient and nonambient components of particulate
matter: a comparison of health effects. Epidemiology, 16: 396-405. 056907
Eder B; Kang D; Mathur R; Yu S; Schere K. (2006). An operational evaluation of the Eta-CMAQ air quality
forecast model. Atmos Environ, 40: 4894-4905. 142721
Eder B; Yu S. (2005). A performance evaluation of the 2004 release of Models-3 CMAQ. Atmos Environ, 40:
4811-4824. 089229
Edgerton ES; Casucclo GS; Saylor RD; Lersch TL; Hartsell BE; Jansen JJ; Hansen DA. (2009). Measurements
of OC and EC in coarse particulate matter in the southeastern United States. J Air Waste Manag Assoc,
59: 78-90. 180385
Edgerton ES; Hartsell BE; Saylor RD; Jansen JJ; Hansen DA; Hidy GM. (2005). The Southeastern Aerosol
Research and Characterization Study: Part II Filter-based measurements of fine and coarse particulate
matter mass and composition. J Air Waste Manag Assoc, 55: 1527-1542. 088686
Edney EO; Lewandowski M; Offenberg JH; Wang W; Claeys M; Kleindienst TE; Jaoui M. (2005). Formation of
2-methyl tetrols and 2-methylglyceric acid in secondary organic aerosol from laboratory irradiated
isoprene/NOX/SO 2/air mixtures and their detection in ambient PM2.5 samples collected in the eastern
United States. Atmos Environ, 39: 5281-5289. 155760
Eiguren-Fernandez A; Miguel AH; Jaques PA; Sioutas C. (2003). Evaluation of a denuder-MOUDI-PUF
sampling system to measure the size distribution of semi-volatile polycyclic aromatic hydrocarbons in the
atmosphere. Aerosol Sci Technol, 37: 201-209. 142609
El-Zanan HS; Lowenthal DH; Zielinska B; Chow JC; Kumar N. (2005). Determination of the organic aerosol
mass to organic carbon ratio in IMPROVE samples. Chemosphere, 60: 485-496. 155764
Elgethun K; Fenske R; Yost M; Palcisko G. (2003). Time-location analysis for exposure assessment studies of
children using a novel global positioning system instrument. Environ Health Perspect, 111: 115-122.
190640
July 2009
3-273
DRAFT-DO NOT CITE OR QUOTE

-------
Engel-Cox JA; Weber SA. (2007). Compilation and assessment of recent positive matrix factorization and
UNMIX receptor model studies on fine particulate matter source apportionment for the eastern United
States. J Air Waste Manag Assoc, 57- 1307-1316. 156419
Engewald W> Teske J; Efer J. (1999). Programmed temperature vaporisers-based large volume injection in
capillary gas chromatography. J Chromatogr, 842- 143-161. 155765
England GC; Watson JG; Chow JC; Zielinska B; Chang M. (2007). Dilution-Based Emissions Sampling from
Stationary Sources' Part 2" Gas-Fired Combustors Compared with Other Fuel-Fired Systems. J Air
Waste Manag Assoc, 57- 79"93. 156421
England GC; Watson JG; Chow JC; Zielinska B; Oliver Chang MC; Loos KR; Hidv GM. (2007). Dilution-based
emissions sampling from stationary sources' Part 1-Compact sampler methodology and performance. J
Air Waste Manag Assoc, 57' 65"78. 156420
Enstrom JE. (2005). Fine particulate air pollution and total mortality among elderly Californians, 1973-2002.
Inhal Toxicol, 17: 803-816. 087356
Erisman J-W; Vermetten AWM; Asman WAH; Waijers-IJpelaan A; Slanina J. (1988). Vertical distribution of
gases and aerosols: the behaviour of ammonia and related components in the lower atmosphere. Atmos
Environ, 22: 1153-1160. 036510
Fairlie TD; Jacob DJ; Park RJ. (2007). The impact of transpacific transport of mineral dust in the United States.
Atmos Environ, 41: 1251-1266. 141923
Fan X; Brook JR; Mabury SA. (2003). Sampling atmospheric carbonaceous aerosols using an integrated organic
gas and particle sampler. Environ Sci Technol, 37: 3145-3151. 058628
Fan X; Lee PKH; Brook JR; Mabury SA. (2004). Improved Measurement of Seasonal and Diurnal Differences in
the Carbonaceous Components of Urban Particulate Matter Using a Denuder-Based Air Sampler.
Aerosol Sci Technol, 38: 63-69. 155770
Farmer PB; Singh R; Kaur B; Sram RJ; Binkova B; Kalina I; Popov TA; Garte S; Taioli E; Gabelova A;
Cebulska-Wasilewska A. (2003). Molecular epidemiology studies of carcinogenic environmental
pollutants. Effects of polycyclic aromatic hydrocarbons (PAHs) in environmental pollution on exogenous
and oxidative DNA damage. , 544: 397-402. 156431
Feldpausch P; Fiebig M; Fritzsche L; Petzold A. (2006). Measurement of ultrafine aerosol size distributions by a
combination of diffusion screen separators and condensation particle counters. J Aerosol Sci, 37: 577-597.
155773
Filleul L; Zeghnoun A; Cassadou S; Declercq C; Eilstein D; Le Tertre A; Medina S; Pascal L; Prouvost H; Saviuc
P; Quenel P. (2006). Influence of set-up conditions of exposure indicators on the estimate of short-term
associations between urban pollution and mortality. Sci Total Environ, 355: 90-97. 089862
Fine PM; Chakrabarti B; Krudysz M; Schauer JJ; Sioutas C. (2004). Diurnal Variations of Individual Organic
Compound Constituents of Ultrafine and Accumulation Mode Particulate Matter in the Los Angeles
Basin. Environ Sci Technol, 38: 1296-1304. 141283
Fine PM; Jaques PA; Hering SV; Sioutas C. (2003). Performance Evaluation and Use of a Continuous Monitor
for Measuring Size-Fractionated PM 2.5 Nitrate. Aerosol Sci Technol, 37: 342 - 354. 155775
Finlayson-Pitts BJ; Pitts JN Jr. (2000). Chemistry of the upper and lower atmosphere: theory, experiments and
applications. 055565
Fowler D; Cape JN; Unsworth MH. (1989). Deposition of atmospheric pollutants on forests. Proc Biol Sci, 324:
247-265. 002515
Fowler D; Duyzer JH; Baldocchi DD. (1991). Inputs of trace gases, particles and cloud droplets to terrestrial
surfaces. Proc Roy Soc Edinb B Biol, 97: 35-59. 046630
Fox-Rabinovitz MS; Takacs LL; Govindaraju RC. (2002). A variable -resolution stretched"grid general circulation
model and data assimilation system with multiple areas of interest: studying the anomalous regional
climate events of 1998. J Geophys Res, 107D24. 047806
Frank NH. (2006). Retained nitrate, hydrated sulfates, and carbonaceous mass in federal reference method fine
particulate matter for six eastern U.S. cities. J Air Waste Manag Assoc, 56: 500-11. 098909
Fraser MP; Cass GR; Simoneit BRT. (1999). Particulate organic compounds emitted from motor vehicle exhaust
and in the urban atmosphere. Atmos Environ, 33: 2715-2724. 010819
Fraser MP; Yue ZW; Buzcu B. (2003). Source apportionment of fine particulate matter in Houston, TX, using
organic molecular markers. Atmos Environ, 37: 2117-2123. 042231
Fruin S; Westerdahl D; Sax T; Sioutas C; Fine PM. (2008). Measurements and predictors of on-road ultrafine
particle concentrations and associated pollutants in Los Angeles. Atmos Environ, 42: 207-219. 097183
Fuentes M; Raftery AE. (2005). Model evaluation and spatial interpolation by Bayesian combination of
observations with outputs from numerical models. Biometrics, 61: 36"45. 087580
July 2009
3-274
DRAFT-DO NOT CITE OR QUOTE

-------
Fuentes M> Song HR; Ghosh SK; Holland DM; Davis JM. (2006). Spatial association between speciated fine
particles and mortality. Biometrics, 62- 855-63. 097647
Gallagher MWi Choularton TW; Morse AP; Fowler D. (1988). Measurements of the size dependence of cloud
droplet deposition at a hill site. , 114- 1291-1303. 046631
Gao S> Ng NL; Keywood Mi Varutbangkul V; Bahreini R; Nenes A; He J; Yoo KY; Beauchamp JL; Hodyss RP;
Flagan RC; Seinfeld JH. (2004). Particle phase acidity and oligomer formation in secondary organic
aerosol. Environ Sci Technol, 38- 6582-6589. 156460
Garner JHB; Pagano Ti Cowling EB. (1989). An evaluation of the role of ozone, acid deposition, and other
airborne pollutants in the forests of eastern North America. 042085
Garten CT Jrl Hanson PJ. (1990). Foliar retention of 15N-nitrate and 15N-ammonium by red maple (Acer
rubrum) and white oak (Quercus alba) leaves from simulated rain. Environ Exp Bot, 30- 333-342. 036803
Gaydos TM; Pinder R; Koo Bi Fahey KM; Yarwood GJ Pandis SN. (2007). Development and application of a
three-dimensional aerosol chemical transport model, PMCAMx. Atmos Environ, 41- 2594-2611. 139738
Gaydos TM; Stanier CO; Pandis SN. (2005). Modeling of in situ ultrafine atmospheric particle formation in the
eastern United States. , 110- D07S12. 191762
Gelencser A; Varga Z. (2005). Evaluation of the atmospheric significance of multiphase reactions in atmospheric
secondary organic aerosol formation. , 5- 2823—2831. 156463
Gelencser A; Meszaros T; Blazso M; Kiss G; Krivacsy Z; Molnar A; Meszaros E. (2000). Structural
Characterisation of Organic Matter in Fine Tropospheric Aerosol by Pyrolysis-Gas Chromatography-
Mass Spectrometry. , 37- 173-183. 155785
Geller M; Solomon P. (2006). Special issue of Aerosol Science and Technology for particulate matter supersites
program and related studies. Aerosol Sci Technol, 40- 735-736. 139645
Geller MD! Ntziachristos L; Mamakos A; Samaras Z; Schmitz DA; Froines JR; Sioutas C. (2006).
Physicochemical and redox characteristics of particulate matter (PM) emitted from gasoline and diesel
passenger cars. Atmos Environ, 40- 6988-7004. 139644
Generoso S; Bey I; Attie J-L; Breon F"M. (2007). A satellite- and model-based assessment of the 2003 Russian
fires' impact on the arctic region. J Geophys Res, 112- 5302. 155786
Georgopoulos PG; Wang S"W; Vyas VM; Sun Q; Burke J; Vedantham R; McCurdy T; Ozkaynak H. (2005). A
source-to-dose assessment of population exposures to fine PM and ozone in Philadelphia, PA, during a
summer 1999 episode. J Expo Sci Environ Epidemiol, 15: 439-457. 080269
Gery MW; Whitten GZ; Killus JP; Dodge MC. (1989). A photochemical kinetics mechanism for urban and
regional scale computer modeling. J Geophys Res, 94: 12,925-12,956. 043039
Giglio L; van der Werf GR; Randerson JT; Collatz GJ; Kasibhatla P. (2006). Global estimation of burned area
using MODIS active fire observations. , 6: 957-974. 156469
Gillette DA; Hanson KJ. (1989). Spatial and temporal variability of dust production caused by wind erosion in
the United States. J Geophys Res, 94: 2197-2206. 030212
Gilliam RC; Childs PP; Huber AH; Raman S. (2005). Metropolitan-scale transport and dispersion from the New
York World Trade Center following September 11, 2001 Part I: an evaluation of the CALMET
meteorological model. , 162: 1981-2003. 056749
Gilliam RC; Huber AH; Raman S. (2005). Metropolitan-scale transport and dispersion from the New York World
Trade Center following September 11, 2001 Part IP an application of the CALPUFF plume model. , 162:
2005-2028. 056750
Gilliland F; Avol E; Kinney P; Jerrett M; Dvonch T; Lurmann F; Buckley T; Breysse P; Keeler G; de Villiers T;
McConnell R. (2005). Air pollution exposure assessment for epidemiologic studies of pregnant women and
children: lessons learned from the Centers for Children's Environmental Health and Disease Prevention
Research. Environ Health Perspect, 113: 1447-54. 098820
Goldstein E; Eagle MC; Hoeprich PD. (1973). Effect of nitrogen dioxide on pulmonary bacterial defense
mechanisms. Arch Environ Occup Health, 26: 202-204. 015674
Gomez-Perales JE; Colvile RN; Fernandez-Bremauntz AA; Gutierrez "Ave doy V; Paramo-Figueroa VH; Bianco-
Jimenez S; Bueno-Lopez E; Bernabe-Cabanillas R; Mandujano F; Hidalgo "Navarro M; Nieuwenhuijsen
MJ. (2007). Bus, minibus, metro inter-comparison of commuters' exposure to air pollution in Mexico City.
Atmos Environ, 41: 890-901. 138816
Gomez-Perales JE; Colvile RN; Nieuwenhuijsen MJ; Fernandez-Bremauntz A; Gutierrez-Avedoy VJ; Paramo-
Figueroa VH; Bianco-Jimenez S; Bueno-Lopez E; Mandujano F; Bernabe-Cabanillas R; Ortiz-Segovia E.
(2004). Commuters' exposure to PM25, CO, and benzene in public transport in the metropolitan area of
Mexico City. Atmos Environ, 38: 1219-1229. 054418
Gong L; Xu B; Zhu Y. (2009). Ultrafine particles deposition inside passenger vehicles . Aerosol Sci Technol, 43:
544-553. 190124
July 2009
3-275
DRAFT-DO NOT CITE OR QUOTE

-------
Goriaux M> Jourdain B; Temime B; Besombes JL; Marchand N> Albinet A; Leoz-Garziandia E> Wortham H.
(2006). Field comparison of particulate PAH measurements using a low-flow denuder device and
conventional sampling systems. Environ Sci Technol, 40- 6398-6404. 156484
Graham B; Guyon P; Taylor PE; Artaxo P; Maenhaut W', Glovsky MM; Flagan RC; Andreae MO. (2003). Organic
compounds present in the natural Amazonian aerosol- Characterization by gas chromatography-mass
spectrometry. J Geophys Res, 108- 4766. 156489
Greenwald R; Bergin MH; Weber R; Sullivan A. (2007). Size -resolved, real-time measurement of water "in soluble
aerosols in metropolitan Atlanta during the summer of 2004. Atmos Environ, 41- 519-531. 155809
Grell GA; Emeis Si Stockwell WR; Schoenemeyer T> Forkel R; Michalakes J; Knoche R; Seidl W. (2000).
Application of a multiscale, coupled MM5/chemistry model to the complex terrain of the VOTALP valley
campaign. Atmos Environ, 34: 1435-1453. 048047
Grover BD; Eatough NL; Eatough DJ; Chow JC; Watson JG; Ambs JL; Meyer MB; Hopke PK; Al-Horr R; Later
DW; Wilson WE. (2006). Measurement of Both Nonvolatile and Semi-Volatile Fractions of Fine
Particulate Matter in Fresno, CA. Aerosol Sci Technol, 40: 811-826. 138080
Grover BD; Hopke PK; Long RW; Wilson WE; Meyer MB; Ambs JL; Kleinman M; Eatough NL; Eatough DJ.
(2005). Measurement of total PM2.5 mass (nonvolatile plus semivolatile) with the Filter Dynamic
Measurement System tapered element oscillating microbalance monitor. J Geophys Res, 110: 1-9. 156500
Grover BD; Kleinman M; Eatough NL; Eatough DJ; Cary RA; Hopke PK; Wilson WE. (2008). Measurement of
fine particulate matter nonvolatile and semi-volatile organic material with the Sunset Laboratory
Carbon Aerosol Monitor. J Air Waste Manag Assoc, 58: 12-11. 156502
Gulliver J; Briggs DJ. (2004). Personal exposure to particulate air pollution in transport microenvironments.
Atmos Environ, 38: 1-8. 053238
Gulliver J; Briggs DJ. (2005). Time—space modeling of journey "time exposure to traffic-related air pollution
using GIS. Environ Res, 97: 10-25. 191079
Gulliver J; Briggs DJ. (2007). Journey-time exposure to particulate air pollution. Atmos Environ, 41: 7195-7207.
155814
Gustafsson O; Krusa M; Zencak Z; Sheesley RJ; Granat L; Engstrom E; Praveen PS; Rao PS; Leek C; Rodhe H.
(2009). Brown clouds over South Asia: biomass or fossil fuel combustion?. , 323: 495-498. 192000
Gutierrez-Daban A; Fernandez-Espinosa AJ; Ternero-Rodriguez M; Fernandez-Alvarez F. (2005). Particle-size
distribution of polycyclic aromatic hydrocarbons in urban air in southern Spain. Anal Bioanal Chem, 381:
721-736. 155818
Hagler GSW; Baldauf RW; Thoma ED; Long TR; Snow RF; Kinsey JS; Oudejans L; Gullett BK. (2009). Ultrafine
particles near a major roadway in Raleigh, North Carolina: Downwind attenuation and correlation with
traffic-related pollutants. Atmos Environ, 43: 1229-1234. 191185
Hains JC; Chen L"WA; Taubman BF; Doddridge BG; Dickerson RR. (2007). A side-by-side comparison of filter-
based PM25 measurements at a suburban site: a closure study. Atmos Environ, 41: 6167-6184. 091039
Hammond DM; Lalor MM; Jones SL. (2007). In-vehicle measurement of particle number concentrations on
school buses equipped with diesel retrofits . , 179: 217-225. 190135
Han Y; Cao J; Chow JC; Watson JG; An Z; Jin Z; Fung K; Liu S. (2007). Evaluation of the thermal/optical
reflectance method for discrimination between char-and soot-EC. Chemosphere, 69: 569-574. 155823
Hand JL; Kreidenweis SM. (2002). A New Method for Retrieving Particle Refractive Index and Effective Density
from Aerosol Size Distribution Data. Aerosol Sci Technol, 36: 1012-1026. 155824
Hansen K; Draaijers GPJ; Ivens WPMF; Gundersen P; van Leeuwen NFM. (1994). Concentration variations in
rain and canopy throughfall collected sequentially during individual rain events. Atmos Environ, 28:
3195-3205. 046634
Harrison D; Shik Park S; Ondov J; Buckley T; Roul Kim S; Jayanty RKM. (2004). Highly time resolved fine
particle nitrate measurements at the Baltimore Supersite. Atmos Environ, 38: 5321-5332. 136787
Harrison RM; Jones AM. (2005). Multisite study of particle number concentrations in urban air. , 39: 6063-6070.
191005
Hasegawa S;Hirabayashi M;Kobayashi S;Moriguchi Y;Kondo Y;Tanabe EWakamatsu. (2005). Size Distribution
and Characterization of Ultrafine Particles in Roadside Atmosphere. , 39: 2671-2690. 157355
Hays MD; Geron CD; Linna KJ; Smith ND; Schauer JJ. (2002). Speciation of gas-phase and fine particle
emissions from burning of foliar fuels. Environ Sci Technol, 36: 2281-2295. 026104
Hays MD; Lavrich RJ. (2007). Developments in direct thermal extraction gas chromatography-mass
spectrometry of fine aerosols. Trends Analyt Chem, 26: 88-102. 155831
Heald C; Spracklen D. (2009). Atmospheric budget of primary biological aerosol particles from fungal sources.
Geophys Res Lett, 36: L09806. 190014
July 2009
3-276
DRAFT-DO NOT CITE OR QUOTE

-------
Health Effects Institute. (2009). Traffic-related air pollution- A critical Review of the literature on emissions,
exposure, and health effects . Health Effects Institute. Boston. 191009
Henderson R. (2005). Clean Air Scientific Advisory Committee (CASAC) Review of the EPA Staff
Recommendations Concerning a Potential Thoracic Coarse PM Standard in the Review of the National
Ambient Air Quality Standards for Particulate Matter- Policy Assessment of Scientific and Technical
Information. U.S. Environmental Protection Agency. RTP, NC. 156537
Henderson R. (2009). Clean Air Scientific Advisory Committee (CASAC) letter the the Administrator-
Consultation on monitoring issues related to the NAAQS for particulate matter. U.S. Environmental
Protection Agency . Washington, DC. EPA-CASAC-09-006. 192001
Henry RC. (1997). History and fundamentals of multivariate air quality receptor models. , 37- 37"42. 020941
Hering S> Fine PM; Sioutas Ci Jaques PA; Ambs JL; Hogrefe O; Demerjian KL. (2004). Field assessment of the
dynamics of particulate nitrate vaporization using differential TEOM® and automated nitrate monitors.
Atmos Environ, 38: 5183-5192. 155837
Hering SV. (2007). Using Regional Data and Building Leakage to Assess Indoor Concentrations of Particles of
Outdoor Origin. Aerosol Sci Technol, 41: 639-654. 155839
Hering SV; Stolzenburg MR; Quant FR; Oberreit DR; Keady PB. (2005). A Laminar-Flow, Water-Based
Condensation Particle Counter (WCPC). Aerosol Sci Technol, 39: 659-672. 155838
Hermann M; Wehner B; Bischof O; Han HS; Krinke T; Liu W', Zerrath A; Wiedensohler A. (2007). Particle
counting efficiencies of new TSI condensation particle counters. J Aerosol Sci, 38: 674-682. 155840
Herner JD; Aw J; Gao O; Chang DP; Kleeman MJ. (2005). Size and composition distribution of airborne
particulate matter in Northern California: I - Particulate mass, carbon, and water-soluble ions. J Air
Waste Manag Assoc, 55: 30-51. 135983
Ho KF; Cao JJ; Harrison RM; Lee SC. (2004). Indoor/outdoor relationships of organic carbon (OC) and elemental
carbon (EC) in PM25 in roadside environment of Hong Kong. Atmos Environ, 38: 6327-6335. 056804
Hoek G; de Hartog J; Meliefste K; ten Brink H; Katsouyanni K; Karakatsani A; Lianou M; Kotronarou A;
Kavouras I; Pekkanen J; Vallius M; Kulmala M; Puustinen A; Thomas S; Meddings C; Ayres J; van
Wijnen J; Hameri K; Kos G; Harrison R. (2008). Indoor-outdoor relationships of particle number and
mass in four European cities. Atmos Environ, 42: 156-169. 156554
Hogrefe C; Hao W; Civerolo K; Ku JY; Sistla G; Gaza RS; Sedefian L; Schere K; Gilliland A; Mathur R. (2007).
Daily Simulation of Ozone and Fine Particulates over New York State: Findings and Challenges. , 46:
961-979. 156561
Hogrefe O; Schwab JJ; Drewnick F; Lala GG; Peters S; Demerjian KL; Rhoads K; Felton HD; Rattigan OV;
Husain L. (2004). Semicontinuous PM 2.5 sulfate and nitrate measurements at an urban and a rural
location in New York: PMTACS-NY Summer 2001 and 2002 Campaigns. J Air Waste Manag Assoc, 54:
1040-1060. 156560
Hopke PK; Ramadan Z; Paatero P; Norris GA; Landis MS; Williams RW; Lewis CW. (2003). Receptor modeling
of ambient and personal exposure samples: 1998 Baltimore Particulate Matter Epidemiology-Exposure
Study. Atmos Environ, 37: 3289-3302. 095544
Hosker RP Jr; Lindberg SE. (1982). Review: atmospheric deposition and plant assimilation of gases and
particles. Atmos Environ, 16: 889-910. 019118
Hsiao T-C; Chen D-R; Son SY. (2009). Development of mini-cyclones as the size-selective inlet of miniature
particle detectors. J Aerosol Sci, 40: 481-491. 191001
Hu S; Herner JD; Shafer M; Robertson W', Schauer JJ; Dwyer H; Collins J; Huai T; Ayala A. (2009). Metals
emitted from heavy-duty diesel vehicles equipped with advanced PM and NO"X emission controls. Atmos
Environ, 43: 2950-2959. 189886
Huang C; Chang C; Chang S; Tsai C; Shih T; Tang D. (2005). Use of porous foam as the substrate of an impactor
for respirable aerosol sampling. J Aerosol Sci, 36: 1373-1386. 186991
Huang L; Brook JR; Zhang W', Li SM; Graham L; Ernst D; Chivulescu A; Lu G. (2006). Stable isotope
measurements of carbon fractions (OC/EC) in airborne particulate: A new dimension for source
characterization and apportionment. Atmos Environ, 40: 2690-2705. 097654
Huang YL; Batterman S. (2000). Selection and evaluation of air pollution exposure indicators based on
geographic areas. Sci Total Environ, 253: 127-144. 156572
Huebert BJ; Charlson RJ. (2000). Uncertainties in data on organic aerosols. Tellus B Chem Phys Meteorol, 52:
1249-1255. 156577
Huebert BJ; Luke WT; Delany AC; Brost RA. (1988). Measurements of concentrations and dry surface fluxes of
atmospheric nitrates in the presence of ammonia. J Geophys Res, 93: 7127-7136. 036569
July 2009
3-277
DRAFT-DO NOT CITE OR QUOTE

-------
Husar RB; Tratt DM; Schichtel BA; Falke SR; Li F; Jaffe D; Gasso S> Gill T> Laulainen NS; Lu F> Reheis MC;
Chun Y; Westphal D; Holben BN; Gueymard C; McKendry I; Kuring N; Feldman GC; McClain C; Frouin
RJ; Merrill J; DuBois D; Vignola F; Murayama T; Nickovic S; Wilson WE; Sassen K; Sugimoto N; Malm
WC. (2001). Asian dust events of April 1998. J Geophys Res, 106: 18,317-18,330. 024947
Hwang K-W; Lee J-H; Jeong D"Y; Lee C-H; Bhatnagar A; Park J-M; Kim S"H. (2008). Observation of difference
in the size distribution of carbon and major inorganic compounds of atmospheric aerosols after the long-
range transport between the selected days of winter and summer. Atmos Environ, 42: 1057-1063. 134420
ICRP. (1994). Human respiratory tract model for radiological protection: a report of a task group of the
International Commission on Radiological Protection. Ann ICRP, 24: 1-482. 006988
Isakov V; Touma JS; Burke J; Lobdell DT; Palma TJ Rosenbaum A; Ozkaynak H. (2009). Combining regional"
and local-scale air quality models with exposure models for use in environmental health studies. , 59:
461-472. 191192
Isakov V; Touma JS; Khlystov A. (2007). A method of assessing air toxics concentrations in urban areas using
mobile platform measurements. J Air Waste Manag Assoc, 57: 1286-1295. 156588
Jacob DJ. (1999). Introduction to atmospheric chemistry. 091122
Jacobson MZ. (2002). Atmospheric pollution: history, science, and regulation. 090667
Jacobson MZ; Kittelson DB; Watts WF. (2005). Enhanced coagulation due to evaporation and its effect on
nanoparticle evolution. , 39: 9486-9492. 191187
Jaffe DH; Singer ME; Rimm AA. (2003). Air pollution and emergency department visits for asthma among Ohio
Medicaid recipients, 1991-1996. Environ Res, 91: 21-28. 041957
Jansen KL; Larson TV; Koenig JQ; Mar TF; Fields C; Stewart J; Lippmann M. (2005). Associations between
health effects and particulate matter and black carbon in subjects with respiratory disease. Environ
Health Perspect, 113: 1741-1746. 082236
Jaques PA; Ambs JL; Grant WL; Sioutas C. (2004). Field evaluation of the differential TEOM monitor for
continuous PM 2.5 mass concentrations. Aerosol Sci Technol, 38: 49"59. 155878
Jeong CH; Hopke PK; Chalupa D; Utell M. (2004). Characteristics of nucleation and growth events of ultrafine
particlesmeasured in Rochester, NY. Environ Sci Technol, 38: 1933-1940. 180350
Jerrett M; Arain A; Kanaroglou P; Beckerman B; Potoglou D; Sahsuvaroglu T; Morrison J; Giovis C. (2005). A
review and evaluation of intraurban air pollution exposure models. J Expo Sci Environ Epidemiol, 15:
185-204. 092864
Jerrett M; Burnett RT; Ma R; Pope 3rd CA; Krewski D; Newbold KB; Thurston G; Shi Y; Finkelstein N; Calle
EE; Thun MJ. (2005). Spatial analysis of air pollution and mortality in Los Angeles. Epidemiology, 16:
727-36. 189405
Jimenez JL; Jayne JT; Shi Q; Kolb CE; Worsnop DR; Yourshaw I; Seinfeld JH; Flagan RC; Zhang X; Smith KA.
(2003). Ambient aerosol sampling using the Aerodyne Aerosol Mass Spectrometer. J Geophys Res, 108:
8425. 156611
Johnson T; Long T; Ollison W. (2000). Prediction of hourly microenvironmental concentrations of fine particles
based on measurements obtained from the Baltimore scripted activity study. J Expo Sci Environ
Epidemiol, 10: 403-411. 001660
Johnston IDA; Bland JM; Anderson HR. (1987). Ethnic variation in respiratory morbidity and lung function in
childhood. Thorax, 42: 542-548. 019356
Joseph PM. (2008). Can fine particulate matter explain the paradoxical ozone associations?. Environ Int, 34(8):
1185-1191. 155219
Kahn R; Gaitley B; Martonchik J; Diner D; Crean K; Holben B. (2005). Multiangle Imaging Spectroradiometer
(MISR) global aerosol optical depth validation based on 2 years of coincident Aerosol Robotic Network
(AERONET) observations. J Geophys Res, 110: D10S04. 189961
Kalberer M; Paulsen D; Sax M; Steinbacher M; Dommen J; Prevot ASH; Fisseha R; Weingartner E; Frankevich
V; Zenobi R. (2004). Identification of Polymers as Major Components of Atmospheric Organic Aerosols. ,
303: 1659-1662. 156619
Kastner-KLein P; Plate EJ. (1999). Wind-tunnel study of concentration fields in street canyons. Atmos Environ,
33: 3973-3979. 001961
Kaur S; Nieuwenhuijsen M; Colvile R. (2005). Personal exposure of street canyon intersection users to PM25,
ultrafine particle counts and carbon monoxide in central London, UK. Atmos Environ, 39: 3629-3641.
086504
Kaur S; Nieuwenhuijsen MJ; Colvile RN. (2005). Pedestrian exposure to air pollution along a major road in
Central London, UK. Atmos Environ, 39: 7307-7320. 088175
Keller A; Siegmann HC. (2001). The role of condensation and coagulation in aerosol monitoring. J Expo Sci
Environ Epidemiol, 11: 441-448. 025881
July 2009
3-278
DRAFT-DO NOT CITE OR QUOTE

-------
Kelly JM. (1988). Annual elemental input/output estimates for two forested watersheds in eastern Tennessee. J
Environ Qual, 17: 463-468. 037379
Kenny LC; Merrifield T> Mark D; Gussman R; Thorpe A. (2004). The Ddevelopment and Designation Testing of
a New USEPA-Approved Fine Particle Inlet: A Study of the USEPA Designation Process. Aerosol Sci
Technol, 38: 15-22. 155895
Ketzel M> Wahlin P; Berkowicz R; Palmgren F. (2003). Particle and trace gas emission factors under urban
driving conditions in Copenhagen based on street and roof-level observations. Atmos Environ, 37: 2735"
2749. 131251
Khlystov A; Stanier C> Pandis SN. (2004). An Algorithm for Combining Electrical Mobility and Aerodynamic
Size Distributions Data when Measuring Ambient Aerosol. Aerosol Sci Technol, 38: 229-238. 155897
Khlystov A; Stanier CO; Takahama Si Pandis SN. (2005). Water content of ambient aerosol during the
Pittsburgh Air Quality Study. J Geophys Res, 110: D07S10. 156635
Kidwell CBi Ondov JM. (2004). Elemental Analysis of Sub-Hourly Ambient Aerosol Collections. Aerosol Sci
Technol, 38: 205-218. 155898
Kim Di Sass-Kortsak A; Purdham JT; Dales RE; Brook JR. (2005). Sources of personal exposure to fine particles
in Toronto, Ontario, Canada. J Air Waste Manag Assoc, 55: 1134-1146. 156640
Kim E; Hopke PK; Pinto JP; Wilson WE. (2005). Spatial variability of fine particle mass, components, and
source contributions during the regional air pollution study in St Louis. Environ Sci Technol, 39: 4172"
4179. 083181
Kim SY; Sheppard L; Kim H. (2009). Health effects of long-term air pollution: influence of exposure prediction
methods. Epidemiology, 20: 442-50. 188446
Kinney PL; Aggarwal M; Northridge ME; Janssen NAH; Shepard P. (2000). Airborne concentrations of PM25
and diesel exhaust particles on Harlem sidewalks: a community-based pilot study. Environ Health
Perspect, 108: 213-218. 001774
Kinsey JS; Williams DC; Dong Y; Logan R. (2007). Characterization of fine particle and gaseous emissions
during school bus idling. , 41: 4972-4979. 190073
Kiss G; Varga B; Galambos I; Ganszky I. (2002). Characterization of water-soluble organic matter isolated from
atmospheric fine aerosol. J Geophys Res, 107: 8339. 156646
Kittelson DB. (1998). Engines and nanoparticles: a review. J Aerosol Sci, 29: 575-588. 051098
Kittelson DB; Schauer JJ; Lawson DR; Watts WF; Johnson JP. (2006). On-road and laboratory evaluation of
combustion aerosols-Part 2'. Summary of spark ignition engine results. J Aerosol Sci, 37: 931-949.
156648
Kittelson DB; Watts WF; Johnson JP. (2006). On-road and laboratory evaluation of combustion aerosols—Part
l: Summary of diesel engine results. J Aerosol Sci, 37: 913-930. 156649
KLeindienst TE; Edney EO; Lewandowski M; Offenberg JH; Jaoui M. (2006). Secondary Organic Carbon and
Aerosol Yields from the Irradiations of Isoprene and-Pinene in the Presence of NO x and SO 2. Environ
Sci Technol, 40: 3807-3812. 156650
KLepeis NE; Nelson WC; OttWR; Robinson JPTsang AM; Switzer P; Behar JV; Hern SC; Engelmann WH.
(2001). The National Human Activity Pattern Survey (NHAPS): a resource for assessing exposure to
environmental pollutants. J Expo Sci Environ Epidemiol, 11: 231-252. 002437
KLouda GA; Filliben JJ; Parish HJ; Chow JC; Watson JG; Cary RA. (2005). Reference material 8785: Air
particulate matter on filter media. Aerosol Sci Technol, 39: 173-183. 130382
Koenig JQ; Mar TF; Allen RW; Jansen K; Lumley T; Sullivan JH; Trenga CA; Larson T; Liu LJ. (2005).
Pulmonary effects of indoor- and outdoor-generated particles in children with asthma. Environ Health
Perspect, 113: 499-503. 087384
Kohsaka A; Watanobe H; Kakizaki Y; Suda T. (1999). A comparative study of the effects of nitric oxide and
carbon monoxide on the in vivo release of gonadotropin-releasing hormone and neuropeptide Y from rat
hypothalamus during the estradiol "induced luteinizing hormone surge: estimation by push-pull
perfusion. Neuroendocrinology, 69: 245-253. 191000
Kokhanovsky A; Breon F; Cacciari A; Carboni E; Diner D; Di Nicolantonio W; Grainger R; Grey W; Holler R; Lee
K. (2007). Aerosol remote sensing over land: a comparison of satellite retrievals using different
algorithms and instruments. , 85: 372-394. 190009
Koutrakis P; Suh HH; Sarnat JA; Brown KW; Coull BA; Schwartz J. (2005). Characterization of particulate and
gas exposures of sensitive subpopulations living in Baltimore and Boston. , 131: 1-65. 095800
Krewski D; Burnett RT; Goldberg MS; Hoover K; Siemiatycki J; Jerrett M; Abrahamowicz M; White WH.
(2000). Reanalysis of the Harvard Six Cities study and the American Cancer Society study of particulate
air pollution and mortality: a special report of the Institute's Particle Epidemiology Reanalysis Project.
Health Effects Institute. Cambridge, MA. 012281
July 2009
3-279
DRAFT-DO NOT CITE OR QUOTE

-------
Krewski D; Jerrett M; Burnett RT; Ma R; Hughes E; Shi Y; Turner MC; Pope AC IIL Thurston G; Calle EE;
Thun MJ. (2009). Extended follow-up and spatial analysis of the American Cancer Society study linking
particulate air pollution and mortality. Health Effects Institute. Cambridge, MA. 140. 191193
Krieger UK; Rupp S; Hausammann E; Peter T. (2007). Simultaneous measurements of PM10 and PM1 using a
single TEOM. Aerosol Sci Technol, 41' 975-980. 129657
Kroll JH; Seinfeld JH. (2008). Chemistry of secondary organic aerosol- Formation and evolution of low-volatility
organics in the atmosphere. Atmos Environ, 42- 3593-3624. 155910
Krudysz M; Froines J; Fine P; Sioutas C. (2008). Intra-community spatial variation of size-fractionated PM
mass, OC, EC, and trace elements in the Long Beach, CA area. Atmos Environ, 42- 5374-5389. 190064
Kruize H; Hanninen O; Breugelmans CK Lebret E; Jantunen M. (2003). Description and demonstration of the
EXPOLIS simulation model- two examples of modeling population exposure to particulate matter. , 13-
87-99. 156661
Kuang C; McMurry PH; McCormick AV; Eisele FL. (2008). Dependence of nucleation rates on sulfuric acid
vapor concentration in diverse atmospheric locations. J Geophys Res, 113- D10209. 191196
Kuhn T; Biswas S; Sioutas C. (2005). Diurnal and seasonal characteristics of particle volatility and chemical
composition in the vicinity of a light-duty vehicle freeway. Atmos Environ, 39- 7154-7166. 129448
Kuhns H; Knipping EM. (2005). Development of a United States-Mexico emissions inventory for the Big Bend
Regional Aerosol and Visibility Observational (BRAVO) Study. J Air Waste Manag Assoc, 55: 677-692.
156663
Kulmala M; Laaksonen A; Pirjola L. (1998). Parameterizations for sulfuric acid/water nucleation rates. J
Geophys Res, 103: 8301-8307. 129411
Kulmala M; Mordas G; Petaja T; Gronholm T; Aalto PP; Vehkamaki H; Hienola AL Herrmann E; Sipila M;
Riipinen I. (2007). The condensation particle counter battery (CPCB): A new tool to investigate the
activation properties of nanoparticles. J Aerosol Sci, 38: 289-304. 155911
Kulmala M; Riipinen I; Sipila M; Manninen HE; Petaja T; Junninen H; Maso MD; Mordas G; Mirme A; Vana M;
Hirsikko A; Laakso L; Harrison RM; Hanson I; Leung C; Lehtinen KE; Kerminen VM. (2007). Toward
direct measurement of atmospheric nucleation. , 318: 89"92. 191761
Kulmala M; Vehkamaki H; Petajda T; Dal Maso M; Lauri A; Kerminen V"M; Birmili W; McMurry PH. (2004).
Formation and growth rates of ultrafine atmospheric particles: a review of observations. J Aerosol Sci,
35: 143-176. 089159
Kuo Y; Huang S; Shih T; Chen C; Weng Y; Lin W. (2005). Development of a size-selective inlet-simulating ICRP
lung deposition fraction. Aerosol Sci Technol, 39: 437-443. 186997
Kurniawan A; Schmidt-Ott A. (2006). Monitoring the soot emissions of passing cars. Environ Sci Technol, 40:
1911-5. 098823
Lake DA; Tolocka MP; Johnston MV Wexler AS. (2003). Mass Spectrometry of Individual Particles between 50
and 750 nm in Diameter at the Baltimore Supersite. Environ Sci Technol, 37: 3268-3274. 156669
Lake DA; Tolocka MP; Johnston MV; Wexler AS. (2004). The character of single particle sulfate in Baltimore.
Atmos Environ, 38: 5311-5320. 088411
Langholz B; Ebi K; Thomas D; Peters J; London S. (2002). Traffic density and the risk of childhood leukemia in
a Los Angeles case-control study. Ann Epidemiol, 12: 482-487. 191771
Larson T; Gould T; Simpson C; Liu LJ; Claiborn C; Lewtas J. (2004). Source apportionment of indoor, outdoor,
and personal PM2.5 in Seattle, Washington, using positive matrix factorization. J Air Waste Manag
Assoc, 54: 1175-87. 098145
Lee HM. (2007). Fabrication of Reference Filter for Measurements of EC (Elemental Carbon) and OC (Organic
Carbon) in Aerosol Particles. Aerosol Sci Technol, 41: 284-294. 155926
Lee HM; Kim CS; Shimada M; Okuyama K. (2005). Effects of Mobility Changes and Distribution of Bipolar Ions
on Aerosol Nanoparticle Diffusion Charging. J Chem Eng Jpn, 38: 486-496. 156679
Lee HM; Soo Kim C; Shimada M; Okuyama K. (2005). Bipolar diffusion charging for aerosol nanoparticle
measurement using a soft X-ray charger. J Aerosol Sci, 36: 813-829. 155924
Lee SJ; Demokritou P; Koutrakis P; Delgado-Saborit JM. (2006). Development and evaluation of personal
respirable particulate sampler (PRPS). Atmos Environ, 40: 212-224. 098249
Leith D; Sommerlatt D; Boundy MG. (2007). Passive sampler for PM10-2.5 aerosol. J Air Waste Manag Assoc, 57:
332-6. 098241
LeVine AM; Gwozdz J; Stark J; Bruno M; Whitsett J; Korfhagen T. (1999). Surfactant protein-A enhances
respiratory syncytial virus clearance in vivo. J Clin Invest, 103: 1015-1021. 156687
Levy JI; Bennett DH; Melly SJ; Spengler JD. (2003). Influence of traffic patterns on particulate matter and
polycyclic aromatic hydrocarbon concentrations in Roxbury, Massachusetts. J Expo Sci Environ
Epidemiol, 13: 364-371. 052661
July 2009
3-280
DRAFT-DO NOT CITE OR QUOTE

-------
Levy JI; Wilson AM; Evans JS; Spengler JD. (2003). Estimation of primary and secondary particulate matter
intake fractions for power plants in Georgia. Environ Sci Technol, 37- 5528-5536. 156688
Lewne Mi Nise Gi Lind MLi Gustavsson P. (2006). Exposure to particles and nitrogen dioxide among taxi, bus
and lorry drivers. Int Arch Occup Environ Health, 79- 220-226. 090556
Lim HJi Allen Gi Maring Hi Solomon Pi Turpin BJi Edgerton E> Hering SV. (2003). Semicontinuous aerosol
carbon measurements' Comparison of Atlanta supersite measurements. J Geophys Res, 108- SOS-SOS.
156697
Lim MCHi Ayoko GA; Morawska L; Ristovski ZD; Jayaratne ER. (2007). The effects of fuel characteristics and
engine operating conditions on the elemental composition of emissions from heavy duty diesel buses. , 86-
1831-1839. 155931
Lindberg SE; Lovett GM. (1985). Field measurements of particle dry deposition rates to foliage and inert
surfaces in a forest canopy. Environ Sci Technol, 19- 238-244. 036530
Lipfert FW; Perry HM Jr; Miller JP; Baty JD; Wyzga RE; Carmody SE. (2000). The Washington University-
EPRI veterans' cohort mortality study- preliminary results. Inhal Toxicol, 4- 41-73. 004087
Lippmann M. (2009). Semi-continuous speciation analyses for ambient air particulate matter- an urgent need
for health effects studies. J Expo Sci Environ Epidemiol, 19: 235-247. 190083
Lipsky E; Robinson A. (2006). Effects of dilution on fine particle mass and partitioning of semivolatile organics
in diesel exhaust and wood smoke. , 40: 155-162. 189891
Little P; Wiffen RD. (1977). Emission and deposition of petrol engine exhaust Pb"I deposition of exhaust Pb to
plant and soil surfaces. Atmos Environ, 11: 437-447. 070869
Liu L-J; Box M; Kalman D; Kaufman J; Koenig J; Larson T; Lumley T; Sheppard L; Wallace L. (2003). Exposure
assessment of particulate matter for susceptible populations in Seattle. Environ Health Perspect, 11:
909-918. 073841
Liu XH; Hegg DA; Stoelinga MT. (2001). Numerical simulation of new particle formation over the northwest
Atlantic using the MM5 mesoscale model coupled with sulfur chemistry. J Geophys Res, 106: 9697-9715.
048201
Long CM; Suh HH; Catalano PJ; Koutrakis P. (2001). Using time- and size-resolved particulate data to quantify
indoor penetration and deposition behavior. Environ Sci Technol, 35: 2089-2099. 011526
Long RW; McClenny WA. (2006). Laboratory and field evaluation of instrumentation for the semicontinuous
determination of particulate nitrate (and other water-soluble particulate components). J Air Waste
Manag Assoc, 56: 294-305. 098214
Lovett GM. (1994). Atmospheric deposition of nutrients and pollutants in North America: an ecological
perspective. Ecol Appl, 4- 629-650. 024049
Lowenthal DH; Rogers CF; Saxena P; Watson JG; Chow JC. (1995). Sensitivity of estimated light extinction
coefficients to model assumptions and measurement errors. Atmos Environ, 29: 751-766. 045134
Lu R; Turco RP; Jacobson MZ. (1997). An integrated air pollution modeling system for urban and regional
scales: 1 Structure and performance. J Geophys Res, 102: 6063-6079. 048202
Lu R; Turco RP; Jacobson MZ. (1997). An integrated air pollution modeling system for urban and regional
scales: 2. Simulations for SCAQS 1987 . J Geophys Res, 102: 6081-6098. 191768
Lunden MM; Kirchstetter TW; Thatcher TL; Hering SV; Brown NJ. (2008). Factors affecting the indoor
concentrations of carbonaceous aerosols of outdoor origin. Atmos Environ, 42: 5660-5671. 155949
Lunden MM; Revzan KL; Fischer ML; Thatcher TL; Littlejohn D; Hering SV; Brown NJ. (2003). The
transformation of outdoor ammonium nitrate aerosols in the indoor environment. Atmos Environ, 37:
5633-5644. 081201
Lunden MM; Thatcher TL; Hering SV; Brown NJ. (2003). The use of time-and chemically resolved particulate
data to characterize the infiltration of outdoor PM2.5 into a residence in the San Joaquin Valley. , 37:
4724-4732. 156718
Mader BT; Schauer JJ; Seinfeld JH; Flagan RC; Yu JZ; Yang H; Lim HJ; Turpin BJ; Deminter JT; Heidemann
G. (2003). Sampling methods used for the collection of particle-phase organic and elemental carbon
during ACE-Asia. Atmos Environ, 37: 1435-1449. 155955
Maheswaran R; Elliott P. (2003). Stroke mortality associated with living near main roads in England and wales:
a geographical study. , 34: 2776-80. 125271
Mahrt L. (1998). Stratified atmospheric boundary layers and breakdown of models. ,11: 263-279. 048210
Makar PA; Gravel S; Chirkov V; Strawbridge KB; Froude F; Arnold J; Brook J. (2006). Heat flux, urban
properties, and regional weather. Atmos Environ, 40: 2750-2766. 155959
Mar TF; Norris GA; Larson TV; Wilson WE; Koenig JQ. (2003). Air pollution and cardiovascular mortality in
Phoenix, 1995-1997. 156731
July 2009
3-281
DRAFT-DO NOT CITE OR QUOTE

-------
Martin RV; Chance K; jacob DJ; Kurosu TP; Spurr RJD; Bucsela E; Gleason JF; Palmer PI; Bey I; Fiore AM; Li
Q; Yantosca RM; Koelemeijer RBA. (2002). An improved retrieval of tropospheric nitrogen dioxide from
GOME. J Geophys Res, 107: 1-21. 089380
Massman WJ; Pederson J; Delany A; Grantz D; Denhartog G; Neumann HH; Oncley SP; Pearson R; Shaw RH.
(1994). An evaluation of the regional acid deposition model surface module for ozone uptake at 3 sites in
the San-Joaquin valley of California. J Geophys Res, 99: 8281-8294. 043681
Mathis U; Mohr M; Forss AM. (2005). Comprehensive particle characterization of modern gasoline and diesel
passenger cars at low ambient temperatures. Atmos Environ, 39: 107-117. 155970
Mathur R. (2008). Estimating the impact of the 2004 Alaskan forest fires on episodic particulate matter
pollution over the eastern United States through assimilation of satellite-derived aerosol optical depths
in a regional air quality model. J Geophys Res, 113: D17302. 156742
Matsumoto K; Hayano T; Uematsu M. (2003). Positive artifact in the measurement of particulate carbonaceous
substances using an ambient carbon particulate monitor. Atmos Environ, 37: 4713-4717. 124293
Matthias-Maser S; Bogs B; Jaenicke R. (2000). The size distribution of primary biological aerosol particles in
cloud water on the mountain Kleiner Feldberg/Taunus (FRG). , 54: 1-13. 155972
Matti Maricq M. (2007). Chemical characterization of particulate emissions from diesel engines: A review. J
Aerosol Sci, 38: 1079-1118. 155973
McBride SJ; Williams RW! Creason J. (2007). Bayesian hierarchical modeling of personal exposure to
particulate matter. Atmos Environ, 41: 6143-6155. 124058
McCurdy T; Glen G; Smith L; Lakkadi Y. (2000). The National Exposure Research Laboratory's Consolidated
Human Activity Database. J Expo Sci Environ Epidemiol, 10: 566-578. 000782
McDonald JD; Eide I; Seagrave J; Zielinska B; Whitney K; Lawson DR; Mauderly JL. (2004). Relationship
between composition and toxicity of motor vehicle emission samples. Environ Health Perspect, 112: 1527"
1538. 087458
McKendry IG; Strawbridge KB; 0?Neill NT; Macdonald AM; Liu PSK; Leaitch WR; Anlauf KG; Jaegle L; Fairlie
TD; Westphal DL. (2007). Trans-Pacific transport of Saharan dust to western North America: A case
study. J Geophys Res, 112: D01103. 156748
McKenna DS; Konopka P; Grooss J-U; Gunther G; Muller R; Spang R; Offermann D; Orsolini Y. (2002). A new
chemical Lagrangian model of the stratosphere (CLaMS) 1 formulation of advection and mixing. J
Geophys Res, 107. 053445
McMurry PH; Fink M; Sakurai H; Stolzenburg MR; Mauldin RL, III; Smith J; Eisele F; Moore K; Sjostedt S;
Tanner D; Huey LG; Nowak JB; Edgerton E; Voisin D. (2005). A criterion for new particle formation in
the sulfur-rich Atlanta atmosphere. J Geophys Res, 110: np. 191759
Mebust MR; Eder BE Binkowski FS; Roselle SJ. (2003). Models-3 community multiscale air quality (CMAQ)
model aerosol component 2. Model evaluation. J Geophys Res, 108: 4184. 156749
Meng QY; Turpin BJ; Korn L; Weisel CP; Morandi M; Colome S; Zhang J; Stock T; Spektor D; Winer A; Zhang
L; Lee JH; Giovanetti R; Cui W', Kwon J; Alimokhtari S; Shendell D; Jones J; Farrar C; Maberti S.
(2005). Influence of ambient (outdoor) sources on residential indoor and personal PM2.5 concentrations:
analyses of RIOPA data. J Expo Sci Environ Epidemiol, 15: 17-28. 058595
Meng QY; Turpin BJ; Lee JH; Polidori A; Weisel CP; Morandi M; Colome S; Zhang JF; Stock T; Winer A. (2007).
How does infiltration behavior modify the composition of ambient PM2.5 in indoor spaces? An analysis of
RIOPA data. Environ Sci Technol, 41: 7315-7321. 091197
Meng YY; Wilhelm M; Rull RP; English P; Ritz B. (2007). Traffic and outdoor air pollution levels near
residences and poorly controlled asthma in adults. Ann Allergy Asthma Immunol, 98: 455-63. 093275
Meng Z; Dabdub D; Seinfeld JH. (1997). Chemical coupling between atmospheric ozone and particulate matter. ,
277: 116-119. 083324
Mensink C; De Ridder K; Deutsch F; Lefebre F; Van de Vel K. (2008). Examples of scale interactions in local,
urban, and regional air quality modelling. , 89: 351-357. 155980
Mfula AM; Kukadia V; Griffiths RF; Hall DJ. (2005). Wind tunnel modelling of urban building exposure to
outdoor pollution. Atmos Environ, 39: 2737-2745. 123359
Middha P; Wexler A. (2006). Design of a Slot Nanoparticle Virtual Impactor. Aerosol Sci Technol, 40: 737-743.
155982
Middlebrook AM; Murphy DM; Lee S"H; Thomson DS; Prather KA; Wenzel RJ; Liu D"Y; Phares DJ; Rhoads KP;
Wexler AS; Johnston MV Jimenez JL; Jayne JT; Worsnop DR; Yourshaw I; Seinfeld JH; Flagan RC.
(2003). A comparison of particle mass spectrometers during the 1999 Atlanta Supersites project. J
Geophys Res, 108D7. 042932
July 2009
3-282
DRAFT-DO NOT CITE OR QUOTE

-------
Miller GT Jr. (1975). Carbon and oxygen cycles showing chemicals cycling and energy flow. In Living in the
Environment Concepts, Problems, and AlternativesBelmont, CA- Wadsworth Publishing Company, Inc.
018988
Miller SL; Anderson MJ; Daly EP; Milford JB. (2002). Source apportionment of exposures to volatile organic
compounds I Evaluation of receptor models using simulated exposure data. Atmos Environ, 36- 3629"
3641. 030661
Misra C> Geller MD; Sioutas C> Solomon PA. (2003). Development and evaluation of a PM10 impactor-inletfor a
continuous coarse particle monitor. Aerosol Sci Technol, 37- 271-281. 180217
Miyazaki Y> Kondo Y> Takegawa N> Komazaki Y> Fukuda Mi Kawamura Ki Mochida Mi Okuzawa Ki Weber RJ.
(2006). Time -resolved measurements of water-soluble organic carbon in Tokyo. J Geophys Res, 111-
D23206.156767
Molnar Pi Johannesson Si Boman Ji Barregard Li Sallsten G. (2006). Personal exposures and indoor, residential
outdoor, and urban background levels of fine particle trace elements in the general population. J Environ
Monit, 8: 543-551. 156773
Moolgavkar SH. (2000). Air pollution and hospital admissions for diseases of the circulatory system in three US
metropolitan areas. J Air Waste Manag Assoc, 50- 1199-1206. 010305
Moore Ki Krudysz Mi Pakbin Pi Hudda Ni Sioutas C. (2009). Intra-community variability in total particle
number concentrations in the San Pedro Harbor Area (Los Angeles, California). Aerosol Sci Technol, 43-
587-603. 191002
Moore Ki Krudysz Mi Pakbin Pi Hudda Ni Sioutas C. (2009). Intra-community variability in total particle
number concentrations in the San Pedro Harbor Area (Los Angeles, California). Aerosol Sci Technol, 43-
587-603. 191004
Moore KFi Ning Zi Ntziachristos Li Schauer JJi Sioutas C. (2007). Daily variation in the properties of urban
ultrafine aerosol-Part I' Physical characterization and volatility. Atmos Environ, 41- 8633-8646. 122445
Morawska Li Keogh DUi Thomas SBi Mengersen K. (2007). Modality in ambient particle size distributions and
its potential as a basis for developing air quality regulation. Atmos Environ. 155990
Morawska Li Ristovski Zi Jayaratne ERi Keogh DUi Ling X. (2008). Ambient nano and ultrafine particles from
motor vehicle emissions' Characteristics, ambient processing and implications on human exposure . , 42-
8113-8138. 191006
Morawska Li Thomas Si Jamriska Mi Johnson G. (1999). The modality of particle size distributions of
environmental aerosols. Atmos Environ, 33- 4401-4411. 007609
Mulchandani Ai Chen Wi Mulchandani Pi Rogers KR. (2001). Biosensors for direct determination of
organophosphate pesticides . , 16: 225-230. 191003
Muller Ki Spindler Gi Maenhaut Wi Hitzenberger Ri Wieprecht Wi Baltensperger Ui ten Brink H. (2004).
INTERCOMP2000, a campaign to assess the comparability of methods in use in Europe for measuring
aerosol composition. Atmos Environ, 38: 6459-6466. 097109
Murahashi T. (2003). Determination of mutagenic 3-nitrobenzanthrone in diesel exhaust particulate matter by
three-dimensional high-performance liquid chromatography. Analyst, 128: 42-5. 096539
Murphy BNi Pandis SN. (2009). Simulating the formation of semivolatile primary and secondary organic aerosol
in a regional chemical transport model. , in press: in press. 190095
Nash DGi Tolocka MPi Baer T. (2006). The uptake of 03 by myristic acid-oleic acid mixed particles: evidence for
solid surface layers. Phys Chem Chem Phys, 8: 4468-4475. 156795
Nerriere E> Guegan H; Bordigoni B; Hautemaniere A; Momas L Laclner J; Target A; Lameloise P; Delmas V;
Personnaz MB; Koutrakis P; Zmirou-Navier D. (2007). Spatial heterogeneity of personal exposure to
airborne metals in French urban areas. Sci Total Environ, 373: 49"56. 156801
Ng LJ; Stuhmiller LM; Stuhmiller JH. (2007). Incorporation of acute dynamic ventilation changes into a
standardized physiologically based pharmacokinetic model. Inhal Toxicol, 19: 247-263. 090983
Ning Z; Geller MD; Moore KF; Sheesley R; Schauer JJ; Sioutas C. (2007). Daily Variation in Chemical
Characteristics of Urban Ultrafine Aerosols and Inference of Their Sources. Environ Sci Technol, 41:
6000-6006. 156809
Ntziachristos L; Ning Z; Geller MD; Sioutas C. (2007). Particle concentration and characteristics near a major
freeway with heavy-duty diesel traffic. Environ Sci Technol, 41: 2223-2230. 089164
Ntziachristos L; Samaras Z. (2006). Combination of aerosol instrument data into reduced variables to study the
consistency of vehicle exhaust particle measurements. Atmos Environ, 40: 6032-6042. 116722
Odman MT; Boylan JW; Wilkinson JG; Russell AG; Mueller SF; Imhoff RE; Doty KG; Norris WB; McNider RT.
(2002). SAMI air quality modeling: final report. 092474
July 2009
3-283
DRAFT-DO NOT CITE OR QUOTE

-------
Offenberg JH; Lewis CW; Lewandowski M; Jaoui M; Kleindienst TE; Edney EO. (2007). Contributions of
toluene and alpha-pinene to SOA formed in an irradiated toluene/alpha-pinene/NO(x)/ air mixture-
comparison of results using 14C content and SOA organic tracer methods. Environ Sci Technol, 41' 3972"
3976. 156822
Ogulei D; Hopke PK; Zhou L; Patrick Pancras J; Nair Ni Ondov JM. (2006). Source apportionment of Baltimore
aerosol from combined size distribution and chemical composition data. Atmos Environ, 40- 396-410.
119973
Olfert JS; Kulkarni P; Wang J. (2008). Measuring aerosol size distributions with the fast integrated mobility
spectrometer. J Aerosol Sci, 39- 940-956. 156004
Olivier JGJ; Bouwman AF; Berdowski JJM; Veldt C; Bloos JPJ; Visschedijk AJH; van der Maas CWM;
Zandveld PYJ. (1999). Sectoral emission inventories of greenhouse gases for 1990 on a per country basis
as well as on l°x 1°. , 2: 241-263. 156829
Olivier JGJ; Bouwman AF; Berdowski JJM; Veldt C; Bloos JPJ; Visschedijk AJH; Zandveld PYJ; Haverlag JL.
(1996). Description of EDGAR Version 2.0- A set of global emission inventories of greenhouse gases and
ozone-depleting substances for all anthropogenic and most natural sources on a per country basis and on
1 degree x 1 degree grid. 156828
Olson DA; Burke JM. (2006). Distributions of PM2.5 source strengths for cooking from the Research Triangle
Park particulate matter panel study. Environ Sci Technol, 40- 163-169. 189951
Olson DA; Mcdow SR. (2009). Near roadway concentrations of organic source markers. Atmos Environ, 43- 2862"
2867. 191188
Olson DA; Norris GA. (2005). Sampling artifacts in measurement of elemental and organic carbon- Low-volume
sampling in indoor and outdoor environments. Atmos Environ, 39- 5437-5445. 156005
Orsini DA; Ma Y; Sullivan A; Sierau B; Baumann K; Weber RJ. (2003). Refinements to the particle-into-liquid
sampler (PILS) for ground and airborne measurements of water soluble aerosol composition. Atmos
Environ, 37: 1243-1259. 156008
Ott DK; Kumar N; Peters TM. (2008). Passive sampling to capture spatial variability in PM10-2.5. Atmos
Environ, 42: 746-756. 191765
Ozkaynak H; Xue J; Zhou H; Raizenne M. (1996). Associations between daily mortality and motor vehicle
pollution in Toronto, Canada. 073986
Paatero P; Tapper U. (1994). Positive matrix factorization: a non-negative factor model with optimal utilization
of error estimates of data values. Environmetrics, 5' 111-126. 086998
Paciorek CJ; Yanosky JD; Suh HH. (2008). Practical Large-Scale Spatio-Temporal Modeling of Particulate
Matter Concentrations. , unknown: unknown. 190090
Padro J. (1996). Summary of ozone dry deposition velocity measurements and model estimates over vineyard,
cotton, grass and deciduous forest in summer. Atmos Environ, 30: 2363-2369. 052446
Pakbin P;Ning ZiSchauer JiSioutas C. (2009). Characterization of Particle Bound Organic Carbon from Diesel
Vehicles Equipped with Advanced Emission Control Technologies. , In Press: 5622-5634. 189893
Pandian MD! Behar JV; Ott WR; Wallace LA; Wilson AL; Colome S D. (1998). Correcting errors in the
nationwide data base of residential air exchange rates. J Expo Sci Environ Epidemiol, 8: 577-586. 090552
Pandis SN. (2004). Atmospheric aerosol processes. In McMurry PH; Shepard, M; Vickery JS (Ed.), Particulate
Matter Science for Policy Makers: A NARSTO AssessmentCambridge, UK: Cambridge University Press.
156838
Pang Y; Eatough NL; Modey WK; Eatough DJ. (2002). Evaluation of the RAMS continuous monitor for
determination of PM2.5 mass including semi-volatile material in Philadelphia, PA. J Air Waste Manag
Assoc, 52: 563-572. 030353
Pang Y; Turpin BJ; Gundel LA. (2006). On the Importance of Organic Oxygen for Understanding Organic
Aerosol Particles. Aerosol Sci Technol, 40: 128-133. 156012
Park J; Mitchell MJ; McHale PJ; Christopher SF; Myers TP. (2003). Interactive effects of changing climate and
atmospheric deposition on N and S biogeochemistry in a forested watershed of the Adirondack
Mountains, New York State. , 9: 1602-1619. 156842
Park K; Chow JC; Watson JG; Trimble DL; Doraiswamy P; Arnott WP; Stroud KR; Bowers K; Bode R; Petzold
A; Hansen AD. (2006). Comparison of continuous and filter-based carbon measurements at the Fresno
supersite. J Air Waste Manag Assoc, 56: 474-91. 098104
Park RJ; Stenchikov GL; Pickering KE; Dickerson RR; Allen DJ; Kondragunta S. (2001). Regional air pollution
and its radiative forcing- Studies with a single-column chemical and radiation transport model. J
Geophys Res, 106: 28. 156841
Peng RD; Dominici F; Pastor-Barriuso R; Zeger SL; Samet JM. (2005). Seasonal analyses of air pollution and
mortality in 100 US cities. Am J Epidemiol, 161: 585-594. 087463
July 2009
3-284
DRAFT-DO NOT CITE OR QUOTE

-------
Peters Ki Eiden R. (1992). Modelling the dry deposition velocity of aerosol particles to a spruce forest. Atmos
Environ, 26: 2555-2564. 045277
Peters TM. (2006). Use of the Aerodynamic Particle Sizer to Measure Ambient PM 10"2. 5' The Coarse Fraction
of PM 10. J Air Waste Manag Assoc, 56: 411-416. 156860
Pettijohn FJ. (1957). Sedimentary rocks. 156862
Petaja T. (2006). Detection Efficiency of a Water-Based TSI Condensation Particle Counter 3785. Aerosol Sci
Technol, 40: 1090-1097. 156021
Phares DJ; Rhoads KP; Johnston MV; Wexler AS. (2003). Size -resolved ultrafine particle composition analysis
2. Houston. J Geophys Res, 108: 8420. 156866
Phuleria HC; Geller MD; Fine PM; Sioutas C. (2006). Size-Resolved emissions of organic tracers from light-and
heavy-duty vehicles measured in a California roadway tunnel. Environ Sci Technol, 40: 4109-4118.
156867
Phuleria HC; Sheesley RJ; Schauer JJ; Fine PM; Sioutas C. (2007). Roadside measurements of size-segregated
particulate organic compounds near gasoline and diesel-dominated freeways in Los Angeles, CA. Atmos
Environ, 41: 4653-4671. 117816
Pierce JR; Adams PJ. (2009). Uncertainty in global CCN concentrations from uncertain aerosol nucleation and
primary emission rates. , 9: 1339-1356. 191189
Pirjola L; Parviainen H; Hussein T; Valli A; Haemeri K; Aaalto P; Virtanen A; Keskinen J; Pakkanen TA;
Maekelae T; Hillamo RE. (2004). 'Sniffer'; a novel tool for chasing vehicles and measuring traffic
pollutants. Atmos Environ, 38: 3625-3635. 117564
Polidori A; Arhami M; Sioutas C; Delfino RJ; Allen R. (2007). Indoor/Outdoor relationships, trends, and
carbonaceous content of fine particulate matter in retirement homes of the Los Angeles Basin. J Air
Waste Manag Assoc, 57: 366-379. 156877
Polidori A; Turpin B; Meng QY; Lee JH; Weisel C; Morandi M; Colome S; Stock T; Winer A; Zhang J; Kwon J;
Alimokhtari S; Shendell D; Jones J; Farrar C; Maberti S. (2006). Fine organic particulate matter
dominates indoor-generated PM2.5 in RIOPAhomes. J Expo Sci Environ Epidemiol, 16: 321-331. 156876
Pope CA III; Burnett RT; Thun MJ; Calle EE; Krewski D; Ito K; Thurston GD. (2002). Lung cancer,
cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA, 287: 1132-
1141. 024689
Pope CA III; Thun MJ; Namboodiri MM; Dockery DW; Evans JS; Speizer FE; Heath CW Jr. (1995). Particulate
air pollution as a predictor of mortality in a prospective study of US adults. Am J Respir Crit Care Med,
151: 669-674. 045159
Pouliot G; Pace TG; Roy B; Pierce T; Mobley D. (2008). Development of a biomass burning emissions inventory
by combining satellite and ground-based information. , 2: 021501-021517. 156883
Price M; Bulpitt S; Meyer MB. (2003). A comparison of PM10 monitors at a Kerbside site in the northeast of
England. Atmos Environ, 37: 4425-4434. 098082
Puett RC; Schwartz J; Hart JE; Yanosky JD; Speizer FE; Suh H; Paciorek CJ; Neas LM; Laden F. (2008).
Chronic particulate exposure, mortality, and coronary heart disease in the nurses' health study. Am J
Epidemiol, 168: 1161-1168. 156891
Pun BE Seigneur C; White W. (2003). Day-of-week behavior of atmospheric ozone in three US cities. J Air
Waste Manag Assoc, 53: 789-801. 047775
Poschl U. (2005). Atmospheric aerosols: composition, transformation, climate and health effects. Angew Chem
Weinheim Bergstr Ger, 44: 7520-7540. 156882
Qian S; Sakurai H; McMurry PH. (2007). Characteristics of regional nucleation events in urban East St Louis.
Atmos Environ, 41: 4119-4127. 116435
Qin X; Prather KA. (2006). Impact of biomass emissions on particle chemistry during the California Regional
Particulate Air Quality Study. Int J Mass Spectrom, 258: 142-150. 156895
Ramachandran G; Adgate JL; Pratt GC; Sexton K. (2003). Characterizing indoor and outdoor 15 minute average
PM 2.5 concentrations in urban neighborhoods . Aerosol Sci Technol, 37: 33"45. 191197
Rattigan OV; Hogrefe O; Felton HD; Schwab JJ; Roychowdhury UK; Husain L; Dutkiewicz VA; Demerjian KL.
(2006). Multi-year urban and rural semi-continuous PM25 sulfate and nitrate measurements in New
York state: Evaluation and comparison with filter based measurements. Atmos Environ, 40: 192-205.
115897
Rees SL; Robinson AL; Khlystov A; Stanier CO; Pandis SNSN. (2004). Mass balance closure and the Federal
Reference Method for PM2.5 in Pittsburgh, Pennsylvania. Atmos Environ, 38: 3305-3318. 097164
Reff A; Weisel CP; Zhang J; Morandi M; Stock T; Colome S; Winer A; Turpin BJ; Offenberg JH. (2007). A
functional group characterization of organic PM2.5 exposure: Results from the RIOPA study. Atmos
Environ, 41: 4585-4598. 156045
July 2009
3-285
DRAFT-DO NOT CITE OR QUOTE

-------
Reithmeier Ci Sausen R. (2002). ATTILA- atmospheric tracer transport in a Lagrangian model. , 54B- 278-299.
053447
Reponen Ti Grinshpun SA; Trakumas S> Martuzevicius D; Wang Z"M; LeMasters G> Lockey JE; Biswas P.
(2003). Concentration gradient patterns of aerosol particles near interstate highways in the Greater
Cincinnati airshed. J Environ Monit, 5- 557-562. 088425
Rice J. (2004). Comparison of Integrated Filter and Automated Carbon Aerosol Measurements at Research
Triangle Park, North Carolina. Aerosol Sci Technol, 38- 23"36. 156049
Riddle SG; Jakober CA; Robert MA; Cahill TM; Charles MJ; KLeeman MJ. (2007). Large PAHs detected in fine
particulate matter emitted from light-duty gasoline vehicles. Atmos Environ, 41- 8658-8668. 115272
Rigby M> Toumi R. (2008). London air pollution climatology- Indirect evidence for urban boundary layer height
and wind speed enhancement. Atmos Environ, 42- 4932-4947. 156050
Roache PJ. (1998). Verification and validation in computational science and engineering. 156915
Robinson AL; Donahue NM; Rogge WF. (2006). Photochemical oxidation and changes in molecular composition
of organic aerosol in the regional context. J Geophys Res, 111- D03302. 156918
Robinson AL; Donahue NM; Shrivastava MK Weitkamp EA; Sage AM; Grieshop AP; Lane TE; Pierce JR;
Pandis SN. (2007). Rethinking organic aerosols' semivolatile emissions and photochemical aging. , 315:
1259-1262. 156053
Rossner P; Svecova V; Milcova A; Lnenickova Z; Solansky N; Sram RJ. (2008). Seasonal variability of oxidative
stress markers in city bus drivers - Part I. Oxidative damage to DNA. , 642- 14-20. 156927
Rudich Y; Donahue NM; Mentel TF. (2007). Aging of Organic Aerosol- Bridging the Gap Between Laboratory
and Field Studies. Annu Rev Phys Chem, 58- 321. 156059
Russell A; Dennis R. (2000). NARSTO critical review of photochemical models and modeling. Atmos Environ, 34-
2283-2324. 035563
Russell M; Allen D; Collins D; Fraser M. (2004). Daily, Seasonal, and Spatial Trends in PM2. 5 Mass and
Composition in Southeast Texas. Aerosol Sci Technol, 38- 14-26. 156061
Ryan PH; LeMasters GK. (2007). A Review of Land-use Regression Models for Characterizing Intraurban Air
Pollution Exposure. Inhal Toxicol, 19- 127. 156063
Ryan PH; LeMasters GK; Levin L; Burkle J; Biswas P; Hu S; Grinshpun S; Reponen T. (2008). Aland-use
regression model for estimating microenvironmental diesel exposure given multiple addresses from birth
through childhood. Sci Total Environ. 156064
Ryno M; Rantanen L; Papaioannou E; Konstandopoulos AG; Koskentalo T; Savela K. (2006). Comparison of
pressurized fluid extraction, Soxhlet extraction and sonication for the determination of polycyclic
aromatic hydrocarbons in urban air and diesel exhaust particulate matter. J Environ Monit, 8- 488-493.
156065
Saathoff H; Naumann KH; Schnaiter M; Schock W', Weingartner E; Baltensperger U; Kramer L; Bozoki Z;
Poschl U; Niessner R. (2003). Carbon mass determinations during the AIDA soot aerosol campaign 1999.
J Aerosol Sci, 34: 1399-1420. 156066
Sabin LD; Kozawa K; Behrentz E; Winer AM; Fitz DR; Pankratz DV; Colome SD; Fruin SA. (2005). Analysis of
real-time variables affecting children's exposure to diesel-related pollutants during school bus commutes
in Los Angeles. Atmos Environ, 39: 5243-5254. 087728
Sabin LD; Lim JH; Stolzenbach KD; Schiff KC. (2005). Contribution of trace metals from atmospheric deposition
to stormwater runoff in a small impervious urban catchment. Water Res, 39: 3929-3937. 088300
Sadezky A; Muckenhuber H; Grothe H; Niessner R; Poschl U. (2005). Raman microspectroscopy of soot and
related carbonaceous materials: Spectral analysis and structural information. Carbon N Y, 43: 1731-
1742. 097499
Sage AM; Weitkamp EA; Robinson AL; Donahue NM. (2008). Evolving mass spectra of the oxidized component
of organic aerosol: results from aerosol mass spectrometer analyses of aged diesel emissions. , 8: 1139-
1152. 191758
Sahlodin AM; Sotudeh-Gharebagh R; Zhu Y. (2007). Modeling of dispersion near roadways based on the vehicle-
induced turbulence concept. Atmos Environ, 41: 92-102. 114058
Sakurai H; Tobias HJ; Park K; Zarling D; Docherty KS; Kittelson DB; McMurry PH; Ziemann PJ. (2003). On-
line measurements of diesel nanoparticle composition and volatility. Atmos Environ, 37: 1199-1210.
113924
Salminen K; Karlsson V. (2003). Comparability of low-volume PM10 sampler with β-attenuation monitor
in background air. Atmos Environ, 37: 3707-3712. 156070
Santarpia JL; Li RJ; Collins DR. (2004). Direct measurement of the hydration state of ambient aerosol
populations. J Geophys Res, 109: D18209. 156944
July 2009
3-286
DRAFT-DO NOT CITE OR QUOTE

-------
Sardar SB; Fine PMJ Mayo PR; Sioutas C. (2005). Size-fractionated measurements of ambient ultrafine particle
chemicalcomposition in Los Angeles using the NanoMOUDI. Environ Sci Technol, 39: 932-944. 180086
Sardar SB; Fine PM; Sioutas C. (2005). Seasonal and spatial variability of the size-resolved chemical
composition of particulate matter (PM10) in the Los Angeles Basin. J Geophys Res, 110D07S08. 089165
Sardar SB; Solomon PA; Geller MD! Sioutas C. (2006). Development and evaluation of a high-volume
dichotomous sampler for chemical speciation of coarse and fine particles. J Aerosol Sci, 37: 1455-1466.
156071
Sarnat JA; Brown KW; Schwartz J; Coull BA; Koutrakis P. (2005). Ambient gas concentrations and personal
particulate matter exposures: implications for studying the health effects of particles. , 16: 385-395.
087531
Sarnat JA; Long CM; Koutrakis P; Coull BA; Schwartz J; Suh HH. (2002). Using sulfur as a tracer of outdoor
fine particulate matter. Environ Sci Technol, 36: 5305-5314. 037056
Sarnat JA; Schwartz J; Catalano PJ; Suh HH. (2001). Gaseous pollutants in particulate matter epidemiology:
confounders or surrogates?. Environ Health Perspect, 109: 1053-1061. 019401
Sarnat SE; Coull BA; Ruiz PA; Koutrakis P; Suh HH. (2006). The influences of ambient particle composition and
size on particle infiltration in Los Angeles, CA residences. J Air Waste Manag Assoc, 56: 186-196. 089166
Sarnat SE; Klein M; Sarnat JA; Flanders WD; Waller LA; Mulholland JA; Russell AG; Tolbert PE. (2009). An
examination of exposure measurement error from air pollutant spatialvariability in time-series studies. J
Expo Sci Environ Epidemiol, In Press: 1-12. 180084
Sarnat SE; Suh HH; Coull BA; Schwartz J; Stone PH; Gold DR. (2006). Ambient particulate air pollution and
cardiac arrhythmia in a panel of older adults in Steubenville, Ohio. Occup Environ Med, 63: 700-706.
090489
Sawant AA; Cocker DR, 3rd; Miller JW; Taliaferro T; Diaz-Sanchez D; Linn WS; Clark KW; Gong H, Jr. (2008).
Generation and characterization of diesel exhaust in a facility for controlled human exposures. J Air
Waste Manag Assoc, 58: 829-837. 156949
Schauer JJ; KLeeman MJ; Cass GR; Simoneit BRT. (1999). Measurement of emissions from air pollution sources
2 CI through C30 organic compounds from medium duty diesel trucks. Environ Sci Technol, 33: 1578"
1587. 010582
Schauer JJ; KLeeman MJ; Cass GR; Simoneit BRT. (2002). Measurement of emissions from air pollution sources
5 CI - C32 organic compounds from gasoline-powered motor vehicles. Environ Sci Technol, 36: 1169-
1180. 035332
Schauer JJ; Rogge WF; Hildemann LM; Mazurek MA; Cass GR. (1996). Source apportionment of airborne
particulate matter using organic compounds as tracers. Atmos Environ, 30: 3837-3855. 051162
Schnelle-Kreis J; Sklorz MJ Peters A; Cyrys J; Zimmermann R. (2005). Analysis of particle-associated semi-
volatile aromatic and aliphatic hydrocarbons in urban particulate matter on a daily basis. Atmos
Environ, 39: 7702-7714. 112944
Schwab JJ; Felton HD; Rattigan OV; Demerjian KL. (2006). New York State urban and rural measurements of
continuous PM2.5 mass by FDMS, TEOM, and BAM: Evaluations and Comparisons with the FRM . J Air
Waste Manag Assoc, 56: 372-83. 098449
Schwab JJ; Hogrefe O; Demerjian KL; Ambs JL. (2004). Laboratory characterization of modified tapered
element oscillating microbalance samplers. J Air Waste Manag Assoc, 54: 1254-63. 098450
Schwartz J; Coull B; Laden F; Ryan L. (2008). The effect of dose and timing of dose on the association between
airborne particles and survival. Environ Health Perspect, 116: 64"69. 156963
Schwartz J; Sarnat JA; Coull BA; Wilson WE. (2007). Effects of exposure measurement error on particle matter
epidemiology: a simulation using data from a panel study in Baltimore, MD. J Expo Sci Environ
Epidemiol, 17: S2-S10. 090220
Seagrave JC; McDonald JD; Bedrick E; Edgerton ES; Gigliotti AP; Jansen JJ; Ke L; Naeher LP; Seilkop SK;
Zheng M; Mauderley JL. (2006). Lung toxicity of ambient particulate matter from southeastern US sites
with different contributing sources: relationships between composition and effects. Environ Health
Perspect, 114: 1387-93. 091291
Seaman NL. (2000). Meteorological modeling for air quality assessments. Atmos Environ, 34: 2231-2259. 035562
Seinfeld JH; Pandis SN. (1998). Atmospheric chemistry and physics: from air pollution to climate change.
018352
Serre ML; Christakos G. (1999). Modern geostatistics: computational BME analysis in the light of uncertain
physical knowledge-the Equus Beds study. Stoch Environ Res Risk Assess, 13: 1-26. 156968
Sheesley RJ; Schauer JJ; Meiritz M; Deminter JT; Bae M"S; Turner JR. (2007). Daily Variation in Particle-
Phase Source Tracers in an Urban Atmosphere. Aerosol Sci Technol, 41: 981-993. 112017
July 2009
3-287
DRAFT-DO NOT CITE OR QUOTE

-------
Shen S; Jaques PA; Zhu Y; Geller MD; Sioutas C. (2002). Evaluation of the SMPS-APS system as a continuous
monitor for measuring PM2. 5, PM10 and coarse (PM2. 5" 10) concentrations. Atmos Environ, 36- 3939"
3950. 156086
Sheppard L; Slaughter JC; Schildcrout J; Liu L-JS; Lumley T. (2005). Exposure and measurement contributions
to estimates of acute air pollution effects. J Expo Sci Environ Epidemiol, 15- 366-376. 079176
Sheya SA; Glowacki C; Chang MC; Chow JC; Watson JG. (2008). Hot filter/impinger and dilution sampling for
fine particulate matter characterization from ferrous metal casting processes. J Air Waste Manag Assoc,
58: 553-561. 156977
Shi Q; Sakurai H; McMurry PH. (2007). Climatology of Regional Nucleation Events in Urban East St. Louis.
Atmos Environ, 41: 4119-4127 . 191760
Shimpi S; Khalek I; Bougher T; Tennant C. (2009). Number count measurements of particulate trap-equipped
diesel truck exhaust from engines that meet 2007 US on-highway emission regulations. Presented at
A&WMA's 102nd Annual Conference and Exhibition, June 16-19, 2009, Pittsburgh, PA. 189888
Shinn JH. (1978). A critical survey of measurements of foliar deposition of airborne sulfates and nitrates.
Presented at Presented at: 71st annual meeting of the Air Pollution Control Association; June! Houston,
TX. Pittsburgh, PA: Air Pollution Control Association; paper no. 78"7.2. 071426
Shrivastava MK Subramanian R; Rogge WF; Robinson AL. (2007). Sources of organic aerosol: Positive matrix
factorization of molecular marker data and comparison of results from different source apportionment
models. Atmos Environ, 41: 9353-9369. 111594
Singh M; Misra C; Sioutas C. (2003). Field evaluation of a personal cascade impactor sampler (PCIS). Atmos
Environ, 37: 4781-4793. 156088
Singh M; Phuleria HC; Bowers K; Sioutas C. (2006). Seasonal and spatial trends in particle number
concentrations and size distributions at the children's health study sites in Southern California. J Expo
Sci Environ Epidemiol, 16: 3-18. 190136
Sioutas C; Delfino RJ; Singh M. (2005). Exposure assessment for atmospheric ultrafine particles (UFPs) and
implications in epidemiologic research. Environ Health Perspect, 113: 947-955. 088428
Sioutas C; Kim S; Chang M; Terrell LL; Gong H Jr. (2000). Field evaluation of a modified DataRAM MIE
scattering monitor for real-time PM25 mass concentration measurements. Atmos Environ, 34: 4829"
4838. 025223
Slowik JG; Cross ES; Han J-H; Davidovits P; Onasch TB; Jayne JT; Williams LR; Canagaratna MR; Worsnop
DR; Chakrabarty RE Moosm; uuml; ller H; Arnott WP; Schwarz JP; Gao R-S; Fahey DW; Kok GL;
Petzold A. (2007). An Inter-Comparison of Instruments Measuring Black Carbon Content of Soot
Particles. Aerosol Sci Technol, 41: 295 - 314. 096177
Smith JN; Moore KF; McMurry PH; Eisele FL. (2004). Atmospheric Measurements of Sub"20 nm Diameter
Particle Chemical Composition by Thermal Desorption Chemical Ionization Mass Spectrometry. Aerosol
Sci Technol, 38: 100-110. 156090
Smith JR; Bailey MR; Etherington G; Shutt AL; Youngman MJ. (2008). Effect of particle size on slow particle
clearance from the bronchial tree. Exp Lung Res, 34: 287 - 312. 190037
Smith RL; Spitzner D; Kim Y; Fuentes M. (2000). Threshold dependence of mortality effects for fine and coarse
particles in Phoenix, Arizona. J Air Waste Manag Assoc, 50: 1367-1379. 010335
Smith WH; Staskawicz BJ. (1977). Removal of atmospheric particles by leaves and twigs of urban trees: some
preliminary observations and assessment of research needs. J Environ Manage, l: 317-330. 046675
So ESP; Chan ATY; Wong AYT. (2005). Large-eddy simulations of wind flow and pollutant dispersion in a street
canyon. Atmos Environ, 39: 3573-3582. 110746
Solomon P; Baumann K; Edgerton E; Tanner R; Eatough D; Modey W', Marin H; Savoie D; Natarajan S; Meyer
MB. (2003). Comparison of integrated samplers for mass and composition during the 1999 Atlanta
supersites project. J Geophys Res, 108: 8423. 156994
Solomon PA; Hopke PK. (2008). A Special I ssue of JA&WMA Supporting Key Scientific and Policy-and Health-
Relevant Findings from EPA's Particulate Matter Supersites Program and Related Studies: An
Integration and Synthesis of Results. J Air Waste Manag Assoc, 58: 137. 156997
Solomon PA; Sioutas C. (2008). Continuous and semicontinuous monitoring techniques for particulate matter
mass and chemical components: a synthesis of findings from EPA's Particulate Matter Supersites
Program and related studies. , 58: 164-195. 190139
Song CH; Carmichael GR. (2001). A three-dimensional modeling investigation of the evolution processes of dust
and sea-salt particles in east Asia. J Geophys Res, 106: 18,131-18,154. 036064
Sorensen M; Loft S; Andersen HV; Raaschou-Nielsen 0; Skovgaard LT; Knudsen LE; Nielsen IV; Hertel O.
(2005). Personal exposure to PM25, black smoke and N02 in Copenhagen: relationship to bedroom and
outdoor concentrations covering seasonal variation. J Expo Sci Environ Epidemiol, 15: 413-422. 089428
July 2009
3-288
DRAFT-DO NOT CITE OR QUOTE

-------
Stanier CO; Khlystov AY; Chan WR; Mandiro MJ Pandis SN. (2004). A Method for the In Situ Measurement of
Fine Aerosol Water Content of Ambient Aerosols' The Dry-Ambient Aerosol Size Spectrometer (DAASS).
Aerosol Sci Technol, 38: 215 - 228. 095955
Stein AF; Lamb D; Draxler RR. (2000). Incorporation of detailed chemistry into a three-dimensional
Lagrangian-Eulerian hybrid model: application to regional tropospheric ozone. Atmos Environ, 34: 4361"
4372. 048341
Stevenson D; Dentener FJ; Schultz MG; Ellingsen K; Van Noije TPC; Wild Oi Zeng G> Amann Mi Atherton CSi
Bell Ni Bergmann DJi Bey Ii Butler Ti Cofala Ji Collins WJi Derwent RGi Doherty RMi Drevet Ji Eskes
HJi Fiore AM; Gauss M; Hauglustaine DA; Horowitz LW; Isaksen ISA; Krol MC; Lamarque J-F;
Lawrence MG; Montanaro V; Muller J-F; Pitari G; Prather MJ; Pyle JA; Rast S; Rodriguez JM;
Sanderson MG. (2006). Multimodel ensemble simulations of present-day and near-future tropospheric
ozone. J Geophys Res, 111. 089222
Stockwell WR; Middleton P; Chang JS; Tang X. (1990). The second generation Regional Acid Deposition Model
chemical mechanism for regional air quality modeling. J Geophys Res, 95: 16,343-16,367. 043095
Stolzenburg MR; Dutcher DD; Kirby BW; Hering SV. (2003). Automated Measurement of the Size and
Concentration of Airborne Particulate Nitrate. Aerosol Sci Technol, 37: 537-546. 156102
Stolzenburg MR; Hering SV. (2000). Method for the automated measurement of fine particle nitrate in the
atmosphere. Environ Sci Technol, 34: 907-914. 013289
Stracquadanio M; Bergamini D; Massaroli E; Trombini C. (2005). Field evaluation of a passive sampler of
polycyclic aromatic hydrocarbons (PAHs) in an urban atmosphere (Bologna, Italy). J Environ Monit, 7'
910-915. 156104
Strand M; Hopke PK; Zhao W; Vedal S; Gelfand E; Rabinovitch N. (2007). A study of health effect estimates
using competing methods to model personal exposures to ambient PM2.5. J Expo Sci Environ Epidemiol,
17: 549-558. 157018
Strand M; Vedal S; Rodes C; Dutton SJ; Gelfand EW; Rabinovitch N. (2006). Estimating effects of ambient
PM2.5 exposure on health using PM2.5 component measurements and regression calibration (vol 16, pg
30, 2006). J Expo Sci Environ Epidemiol, 16: 471-471. 157017
Streets DG; Zhang Q; Wang L; He K; Hao J; Wu Y; Tang Y; Carmichael GR. (2006). Revisiting China?s CO
emissions after the Transport and Chemical Evolution over the Pacific (TRACE-P) mission: Synthesis of
inventories, atmospheric modeling, and observations. J Geophys Res, 111: D14306. 157019
Subramanian R; Khlystov A; Robinson A. (2006). Effect of Peak Inert-Mode Temperature on Elemental Carbon
Measured Using Thermal-Optical Analysis. Aerosol Sci Technol, 40: 763-780. 156107
Subramanian R; Khlystov AY; Cabada JC; Robinson AL. (2004). Positive and negative artifacts in particulate
organic carbon measurements with denuded and undenuded sampler configurations. Aerosol Sci Technol,
l: 27-48. 081203
Sullivan AP; Peltier RE; Brock CA; de Gouw JA; Holloway JS; Warneke C> Wollny AG; Weber RJ. (2006).
Airborne measurements of carbonaceous aerosol soluble in water over northeastern United States:
Method development and an investigation into water-soluble organic carbon sources. J Geophys Res, 111:
1-14. 157030
Sullivan AP; Weber RJ. (2006). Chemical characterization of the ambient organic aerosol soluble in water: 1.
Isolation of hydrophobic and hydrophilic fractions with a XAD-8 resin. J Geophys Res, 111: D05314.
157031
Sullivan AP; Weber RJ; Clements AL; Turner JR; Bae MS; Schauer JJ. (2004). A method for on-line
measurement of water-soluble organic carbon in ambient aerosol particles: Results from an urban site.
Geophys Res Lett, 31: L13105. 157029
Sullivan JH; Hubbard R; Liu SL; Shepherd K; Trenga CA Koenig JQ; Chandler WL; Kaufman JD. (2007). A
community study of the effect of particulate matter on blood measures of inflammation and thrombosis in
an elderly population. Environ Health Perspect, 6: 3. 100083
Swartz E; Stockburger L; Gundel LA. (2003). Recovery of Semivolatile Organic Compounds during Sample
Preparation: Implications for Characterization of Airborne Particulate Matter. Environ Sci Technol, 37:
597-605. 157035
Swietlik D; Faust M. (1984). Foliar nutrition of fruit crops. , 6: 287-355. 046678
Taha G; Box GP; Cohen DD; Stelcer E. (2007). Black Carbon Measurement using Laser Integrating Plate
Method. Aerosol Sci Technol, 41: 266 - 276. 096277
Taylor DA. (2002). Dust in the wind. Environ Health Perspect, 110: A80-A87. 025693
July 2009
3-289
DRAFT-DO NOT CITE OR QUOTE

-------
Taylor GE Jri Hanson PJi Baldocchi DD. (1988). Pollutant deposition to individual leaves and plant canopies-
sites of regulation and relationship to injury. In Heck, W. W.i Taylor, O. C.i Tingey, D. T.
(Ed.),Assessment of crop loss from air pollutants (pp. 227-257). New York, NY- Elsevier Applied Science.
019289
Temime-Roussel Bi Monod A; Massiani Ci Wortham H. (2004). Evaluation of an annular denuder for
atmospheric PAH partitioning studies--2- evaluation of mass and number particle losses. Atmos Environ,
38: 1925-1932. 098530
Temime-Roussel Bi Monod A; Massiani Ci Wortham H. (2004). Evaluation of an annular denuder tubes for
atmospheric PAH partitioning studies--i: evaluation of the trapping efficiency of gaseous PAHS. Atmos
Environ, 38: 1913-1924. 098521
ten Brink Hi Hoek Gi Khlystov A. (2005). An approach to monitor the fraction of elemental carbon in the
ultrafine aerosol. Atmos Environ, 39: 6255-6259. 156115
ten Brink Hi Maenhaut Wi Hitzenberger Ri Gnauk Ti Spindler Gi Even Ai Chi Xi Bauer Hi Puxbaum Hi Putaud
J-Pi Tursic Ji Berner A. (2004). INTERCOMP2000: the comparability of methods in use in Europe for
measuring the carbon content of aerosol. Atmos Environ, 38: 6507-6519. 097110
Tesche TWi Morris Ri Tonnesen Gi McNally Di Boylan Ji Brewer P. (2006). CMAQ/CAMx annual 2002
performance evaluation over the eastern US. Atmos Environ, 40: 4906-4919. 157050
Thornburg Ji Rodesa CEi Lawless PAi Williams R. (2009). Spatial and Temporal Variability of Outdoor Coarse
Particulate Matter Mass Concentrations Measured with a New Coarse Particle Sampler During the
Detroit Exposure and Aerosol Research Study . Atmos Environ, 0: 0. 190999
Thurston GDi Spengler JD. (1985). A quantitative assessment of source contributions to inhalable particulate
matter pollution in metropolitan Boston. Atmos Environ, 19: 9-25. 056074
Tobias HJi Kooiman PMi Docherty KSi Ziemann PJ. (2000). Real-Time Chemical Analysis of Organic Aerosols
Using a Thermal Desorption Particle Beam Mass Spectrometer. Aerosol Sci Technol, 33: 170-190. 156121
Tobias HJi Ziemann PJ. (1999). Compound Identification in Organic Aerosols Using Temperature-Programmed
Thermal Desorption Particle Beam Mass Spectrometry. Anal Chem, 71: 3428-3435. 157053
Tolocka Mi Jang PMi Ginter JMi Cox FJi Kamens RMi Johnston MV. (2004). Formation of oligomers in
secondary organic aerosol. Environ Sci Technol, 38: 1428-1434. 087578
Trijonis JCi Malm WCi Pitchford Mi White WHi Charlson R. (1990). Acidic deposition: State of science and
technology. Report 24. Visibility: Existing and historical conditions-causes and effects. Final report.
157058
Tsai CJi Chang CTi Huang CH. (2006). Direct field observation of the relative humidity effect on the beta-gauge
readings. J Air Waste Manag Assoc, 56: 834-40. 098312
Tsai Yli Kuo SC. (2006). Development of diffuse reflectance infrared Fourier transform spectroscopy for the
rapid characterization of aerosols. Atmos Environ, 40: 1781-1793. 156127
TSI. (2005). Data Merge Software Module for merging and fitting of SMPA and APS Data Files, P/N 1930074,
User's Guide, Model 390069, Revision A. St. Paul, MN. 157196
Tuch TMi Herbarth Oi Franck Ui Peters Ai Wehner Bi Wiedensohler Ai Heintzenberg J. (2006). Weak
correlation of ultrafine aerosol particle concentrations< 800 nm between two sites within one city. J Expo
Sci Environ Epidemiol, 16: 486-491. 157060
Tung TCWi Chao CYHi Burnett J. (1999). A methodology to investigate the particulate penetration coefficient
through building shell. Atmos Environ, 33: 881-893. 049003
Turpin BJi Lim H-J. (2001). Species contributions to PM25 mass concentrations: revisiting common
assumptions for estimating organic mass. Aerosol Sci Technol, 35: 602-610. 017093
Turpin BJi Weisel CPi Morandi Mi Colome Si Stock Ti Eisenreich Si Buckley B. (2007). Relationships of Indoor,
Outdoor, and Personal Air (RIOPA): Part II. Analyses of concentrations of particulate matter species.
157062
U.S. EPA. (1982). Air quality criteria for particulate matter and sulfur oxides. U.S. Environmental Protection
Agency. Washington, D.C.. 017610
U.S. EPA. (1996). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA/600/P-95/001aF-cF. 079380
U.S. EPA. (1998). Particulate matter research needs for human health risk assessment to support future
reviews of the National Ambient Air Quality Standards for particulate matter. U.S. Environmental
Protection Agency. Durham, North Carolina. 083151
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA/600/P"99/002aF"bF. 056905
U.S. EPA. (2006). 2002 National Emissions Inventory Data and Documentation. 157070
U.S. EPA. (2006). Air quality criteria for ozone and related photochemical oxidants. EPA. DC. 088089
July 2009
3-290
DRAFT-DO NOT CITE OR QUOTE

-------
U.S. EPA. (2006). Provisional Assessment of Recent Studies on Health Effects of Particulate Matter Exposure.
U.S. Environmental Protection AGency. Research Triangle Park, NC. 157071
U.S. EPA. (2008). Analysis of particulate matter emissions from light-duty gasoline vehicles in Kansas City.
U.S. Environmental Protection Agency. Washington, D.C.. EPA420-R-08-010. 191767
U.S. EPA. (2008). Integrated Science Assessment for Oxides of Nitrogen - Health Criteria. U.S. Environmental
Protection Agency. Research Triangle Park, NC. 157073
U.S. EPA. (2008). Integrated science assessment for oxides of nitrogen and sulfur- Ecological criteria. EPA.
Research Triangle Park, NC. EPA/600/R-08/082F. 157074
U.S. EPA. (2008). Integrated Science Assessment for Sulfur Oxides?Health Criteria. U.S. Environmental
Protection Agency. Research Triangle Park, NC. 157075
U.S. EPA. (2008). National Air Quality Status and Trends Through 2007. US EPA. Research Triangle Park.
EPA-454/R-08-006 . 191190
U.S. EPA. (2008). Total Risk Integrated Methodology (TRIM) Air Pollutants Exposure Model Documentation
(TRIM.Expo/APEX, Version 4.3). Volume l: User's Guide. U.S. Environmental Protection Agency, Office
of Air Quality Planning and Standards. Research Triangle Park, NC. EPA-452/B"08"001b. 191775
U.S. EPA. (2008). U.S. EPA's 2008 Report on the Environment (Final Report). EPA. DC. 157076
U.S. EPA. (2009). Risk and Exposure Assessment for Review of the Secondary National Ambient Air Quality
Standards for Oxides of Nitrogen and Oxides of Sulfur Second Draft. U.S. Environmental Protection
Agency. Washington, D.C.. EPA-452/P-09-004a.
http'//www.epa.gov/ttnnaaqs/standards/no2so2sec/cr_rea.html. 191774
U.S. EPA. (2009). Verified Retrofit Technologies. [Online at httpV/www.epa.gov/otaq/retrofit/index.htm. 189885
UNCEC. (2007). Hemispheric Transport of Air Pollution 2007. 157078
Unsworth MH; Wilshaw JC. (1989). Wet, occult and dry deposition of pollutants on forests. , 47: 221-238.
046682
Van Aalst RM. (1982). Dry deposition of NOx. In Schneider, T.> Grant, L. (Ed.),Air pollution by nitrogen
oxidesAmsterdam, The Netherlands: Elsevier Scientific Publishing Company. 036481
van der Werf GR; Randerson JT; Giglio L; Collatz GJ; Kasibhatla PS; Arellano Jr AF. (2006). Interannual
variability in global biomass burning emissions from 1997 to 2004. , 6: 3423-3441. 157084
VanCuren RA; Cahill TA. (2002). Asian aerosols in North America: Frequency and concentration of fine dust. J
Geophys Res, 107 (2002) pArticle Number: 4804: Article. 157087
Vega E> Reyes Ei Wellens A; Sanchez G> Chow JC; Watson JG. (2003). Comparison of continuous and filter
based mass measurements in Mexico City. Atmos Environ, 37: 2783-2793. 105974
Venkataraman C; Lyons JM; Friedlander SK. (1994). Size distributions of polycyclic aromatic hydrocarbons and
elemental carbon 1 Sampling, measurement methods, and source characterization. Environ Sci Technol,
28: 555-562. 002475
Venn A; Lewis S; Cooper M; Hubbard R; Hill I; Boddy R; Bell M; Britton J. (2000). Local road traffic activity and
the prevalence, severity and persistence of wheeze in school children: combined cross sectional and
longitudinal study. Occup Environ Med, 57: 152-158. 007895
Venn AJ; Lewis SA; Cooper M; Hubbard R; Britton J. (2001). Living near a main road and the risk of wheezing
illness in children. Am J Respir Crit Care Med, 164: 2177-2180. 023644
Veranth JM; Moss TA; Chow JC; Labban R; Nichols WK; Walton JC; Walton JG; Yost GS. (2006). Correlation of
in vitro cytokine responses with the chemical composition of soil-derived particulate matter. Environ
Health Perspect, 114: 341-349. 087479
Viana M; Kuhlbusch TAJ; Querol X; Alastuey A; Harrison RM; Hopke PK; Winiwarter W', Vallius M; Szidat S;
Prevot ASH; Hueglin C; Bloemen H; Wahlin P; Vecchi R; Miranda AI; Kasper-Giebl A; Maenhaut W;
Hitzenberger R. (2008). Source apportionment of particulate matter in Europe: A review of methods and
results. J Aerosol Sci, 39(10): 827-849. 155269
Viana M; Querol X; Alastuey A. (2006). Chemical characterisation of PM episodes in NE Spain. Chemosphere,
62: 947-56. 096384
Viana M; Querol X; Alastuey A; Ballester F; Llop S; Esplugues A; FernAjndez-Patier R; GarcA-a dos Santos S;
Herce MD. (2008). Characterising exposure to PM aerosols for an epidemiological study. Atmos Environ,
42: 1552-1568. 156135
Violante FS; Barbieri A; Curti S; Sanguinetti G; Graziosi F; Mattioli S. (2006). Urban atmospheric pollution:
Personal exposure versus fixed monitoring station measurements. Chemosphere, 64: 1722-1729. 156140
Virkkula A; Makela T; Hillamo R; Yli-Tuomi T; Hirsikko A; Hameri K; Koponen IK. (2007). A simple procedure
for correcting loading effects of aethalometer data. J Air Waste Manag Assoc, 57: 1214-1222. 157098
Voisin D; Smith JN; Sakurai H; McMurry PH; Eisele FL. (2003). Thermal Desorption Chemical Ionization Mass
Spectrometer for Ultrafine Particle Chemical Composition. Aerosol Sci Technol, 37: 471-475. 156141
July 2009
3-291
DRAFT-DO NOT CITE OR QUOTE

-------
Volckens J; Olson DA; Hays MD. (2008). Carbonaceous species emitted from handheld two-stroke engines.
Atmos Environ, 42: 1239-1248. 105465
Vyas VM; Christakos G. (1997). Spatiotemporal analysis and mapping of sulfate deposition data over Eastern
USA. Atmos Environ, 31: 3623-3633. 156142
Wagner J; Leith D. (2001). Passive aerosol sampler. Part I: Principle of operation. Aerosol Sci Technol, 34: 186-
192. 190153
Wagner J; Leith D. (2001). Passive aerosol sampler. Part II: Wind tunnel experiments. Aerosol Sci Technol, 34:
193-201. 190154
Wallace L. (2000). Real-time monitoring of particles, PAH, and CO in an occupied townhouse. J Occup Environ
Hyg, 15: 39-47. 000803
Wallace Li Williams R. (2005). Use of personal-indoor-outdoor sulfur concentrations to estimate the infiltration
factor and outdoor exposure factor for individual homes and persons. Environ Sci Technol, 39: 1707-1714.
057485
Wallace Li Williams Ri Rea Ai Croghan C. (2006). Continuous weeklong measurements of personal exposures
and indoor concentrations of fine particles for 37 health-impaired North Carolina residents for up to four
seasons. Atmos Environ, 40: 399-414. 088211
Wallace Li Williams Ri Suggs Ji Jones P. (2006). Estimating contributions of outdoor fine particles to indoor
concentrations and personal exposures: effects of household characteristics and personal activities.
089190
Wang Ji Christopher SAi Nair US; Reid JS; Prins EM; Szykman J; Hand JL. (2006). Mesoscale modeling of
Central American smoke transport to the United States: l.??Top-down??assessment of emission strength
and diurnal variation impacts. J Geophys Res, 111: D05S17. 157109
Wang LH; Milford JB; Carter WPL. (2000). Reactivity estimates for aromatic compounds Part 2 uncertainty in
incremental reactivities. Atmos Environ4349"4360. 048365
Wang LH; Milford JB; Carter WPL. (2000). Reactivity estimates for aromatic compounds Part I: uncertainty in
chamber-derived parameters. Atmos Environ, 34: 4337-4348. 048357
Watson JG; Chen LW; Chow JC; Doraiswamy P; Lowenthal DH. (2008). Source apportionment: findings from
the U.S. Supersites Program. J Air Waste Manag Assoc, 58: 265-288. 157128
Watson JG; Chow JC. (2002). Comparison and evaluation of in situ and filter carbon measurements at the
Fresno Supersite. J Geophys Res, 107. 037873
Watson JG; Chow JC. (2007). Receptor models for source apportionment of suspended particles. In Introduction
to Environmental Forensics (pp. 273-298). San Diego, CA: Academic Press. 157127
Watson JG; Chow JC; Chen LWA. (2005). Summary of organic and elemental carbon/black carbon analysis
methods and intercomparisons. Aerosol Air Qual Res, 5- 69—102. 157125
Watson JG; Fujita EM; Chow JC; Zielinska B; Richards LW; Neff W; Dietrich D. (1998). Northern front range
air quality study Final report. 012257
Watson JG; Robinson NF; Chow JC; Henry RC; Kim BM; Pace TG; Meyer EL; Nguyen Q. (1990). The
USEPA/DRI chemical mass balance receptor model, CMB 70. Environ Software, 5: 38"49. 004848
Weber RJ; Orsini D; Daun Y; Lee Y-NJ Kotz PJ; Brechtel F. (2001). A particle-into-liquid collector for rapid
measurement of aerosol bulk chemical composition. Aerosol Sci Technol, 35: 718-727. 024640
Weber RJ; Orsini D; Duan Y; Baumann K; Kiang CS; Chameides W', Lee YN; Brechtel F; KLotz P; Jongejan P.
(2003). Intercomparison of near real time monitors of PM2. 5 nitrate and sulfate at the US
Environmental Protection Agency Atlanta Supersite. J Geophys Res, 108: 8421. 157129
Weijers EP; Khlystov AY; Kos GPA; Erisman JW. (2004). Variability of particulate matter concentrations along
roads and motorways determined by a moving measurement unit. Atmos Environ, 38: 2993-3002. 104186
Weingartner E; Saathoff H; Schnaiter M; Streit N; Bitnar B; Baltensperger U. (2003). Absorption of light by soot
particles: determination of the absorption coefficient by means of aethalometers. J Aerosol Sci, 34: 1445"
1463. 156149
Wenzel RJ; Liu DY; Edgerton ES; Prather KA. (2003). Aerosol time-of-flight mass spectrometry during the
Atlanta Supersite Experiment: 2. Scaling procedures. J Geophys Res, 108: 8426. 157139
Wesely ML; Eastman JA; Stedman DH; Yalvac ED. (1982). An eddy-correlation measurement of N02 flux to
vegetation and comparison to 03 flux. Atmos Environ, 16: 815-820. 036564
Wesely ML; Hicks BB. (2000). A review of the current status of knowledge on dry deposition. Atmos Environ, 34:
2261-2282. 025018
Westerdahl D; Fruin S; Sax T; Fine PM; Sioutas C. (2005). Mobile platform measurements of ultrafine particles
and associated pollutant concentrations on freeways and residential streets in Los Angeles. Atmos
Environ, 39: 3597-3610. 086502
Whitby KT. (1978). The physical characteristics of sulfur aerosols. Atmos Environ, 12: 135-159. 071181
July 2009
3-292
DRAFT-DO NOT CITE OR QUOTE

-------
Williams B; Goldstein A; Kreisberg Ni Hering S. (2006). An In-Situ Instrument for Speciated Organic
Composition of Atmospheric Aerosols' Thermal Desorption Aerosol GC/MS-FID (TAG). Aerosol Sci
Technol, 40: 627-638. 156157
Williams R; Rea A; Vette A; Croghan C> Whitaker D; Stevens C> McDow Si Fortmann Ri Sheldon Li Wilson Hi
Thornburg Ji Phillips Mi Lawless Pi Rodes Ci Daughtrey H. (2008). The design and field implementation
of the Detroit Exposure and Aerosol Research Study. J Expo Sci Environ Epidemiol, in press: in press.
191198
Williams TCi Shaddix CRi Jensen KAi Suo-Anttila JM. (2006). Measurements of the Dimensionless Extinction
Coefficient of Soot Within Laminar Diffusion Flames. 157148
Wilson JGi Zawar-Reza P. (2006). Intraurban-scale dispersion modelling of particulate matter concentrations:
applications for exposure estimates in cohort studies. Atmos Environ, 40: 1053-1063. 088292
Wilson WE; Brauer M. (2006). Estimation of ambient and non-ambient components of particulate matter
exposure from a personal monitoring panel study. J Expo Sci Environ Epidemiol, 16: 264-274. 088933
Wilson WE; Grover BD; Long RW; Eatough NL; Eatough DJ. (2006). The measurement of fine particulate
semivolatile material in urban aerosols. J Air Waste Manag Assoc, 56: 384-387. 091142
Wilson WE; Mage DT; Grant LD. (2000). Estimating separately personal exposure to ambient and nonambient
particulate matter for epidemiology and risk assessment: why and how. J Air Waste Manag Assoc, 50:
1167-1183. 010288
Wilson WE; Mar TF; Koenig JQ. (2007). Influence of exposure error and effect modification by socioeconomic
status on the association of acute cardiovascular mortality with particulate matter in Phoenix. J Expo Sci
Environ Epidemiol, 17: Sll. 157149
Wilson WE; Suh HH. (1997). Fine particles and coarse particles: concentration relationships relevant to
epidemiologic studies. J Air Waste Manag Assoc, 47: 1238-1249. 077408
Winkler PM; Steiner G; Vrtala A; Vehkamaki H; Noppel M; Lehtinen KEJ; Reischl GP; Wagner PE; Kulmala M.
(2008). Heterogeneous Nucleation Experiments Bridging the Scale from Molecular Ion Clusters to
Nanoparticles. , 319: 1374. 156160
Womiloju TO; Miller JD; Mayer PM; Brook JR. (2003). Methods to determine the biological composition of
particulate matter collected from outdoor air. Atmos Environ, 37: 4335-4344. 179954
Wongphatarakul V; Friedlander SK; Pinto JP. (1998). A comparative study of PM2.5 ambient aerosol chemical
databases. Environ Sci Technol, 32: 3926-3934. 049281
Woo K-S; Chen D-R; Pui DYH; McMurry PH. (2001). Measurement of Atlanta aerosol size distributions:
observations of ultrafine particle events. Aerosol Sci Technol, 34: 75"87. 011702
Wu C-F; Delfino RJ; Floro JN; Quintanai. (2005). Exposure assessment and modeling of particulate matter for
asthmatic children using personal nephelometers. Atmos Environ, 39: 3457-3469. 086397
Wu CF; Delfino RJ; Floro JN; Samimi BS; Quintana PJ; KLeinman MT; Liu LJ. (2005). Evaluation and quality
control of personal nephelometers in indoor, outdoor and personal environments. , 15: 99-110. 157155
Wu J; Houston D; Lurmann F; Ong P; Winer A. (2009). Exposure of PM2. 5 and EC from diesel and gasoline
vehicles in communities near the Ports of Los Angeles and Long Beach, California. Atmos Environ, 43:
1962-1971. 191773
Wu J; M Winer A; J Delfino R. (2006). Exposure assessment of particulate matter air pollution before, during,
and after the 2003 Southern California wildfires. Atmos Environ, 40: 3333-3348. 157156
Wu S; Mickley L; Jacob D; Rind D; Streets D. (2008). Effects of 2000-2050 changes in climate and emissions on
global tropospheric ozone and the policy-relevant background surface ozone in the United States. J
Geophys Res, 113: D18312. 190039
Xia X; Hopke PK. (2006). Seasonal Variation of 2"Methyltetrols in Ambient Air Samples. Environ Sci Technol,
40: 6934-6937. 179947
Xiaomin X; Zhen H; Jiasong W. (2006). The impact of urban street layout on local atmospheric environment. ,
41: 1352-1363. 156165
Yang H; Li Q; Yu JZ. (2003). Comparison of two methods for the determination of water-soluble organic carbon
in atmospheric particles. Atmos Environ, 37: 865-870. 156167
Yanosky JD; Paciorek CJ; Schwartz J; Laden F; Puett R; Suh HH. (2008). Spatio-temporal modeling of chronic
PM10 exposure for the Nurses' Health Study. Atmos Environ, 42(18): 4047-4062. 099467
Yanosky JD; Paciorek CJ; Suh HH. (2009). Predicting chronic fine and coarse particulate exposures using
spatiotemporal models for the northeastern and midwestern United States. Environ Health Perspect,
117: 522-529. 190114
Yi SM; Ambs JL; Patashnick H; Rupprecht G; Hopke PK. (2004). Particle Collection Characteristics of a
Prototype Electrostatic Precipitator (ESP) for a Differential TEOM System. Aerosol Sci Technol, 38: 46-
51. 156169
July 2009
3-293
DRAFT-DO NOT CITE OR QUOTE

-------
Yu H; Remer LA; Chin Mi Bian H; Kleidman RG; Diehl T. (2008). A satellite-based assessment of transpacific
transport of pollution aerosol. J Geophys Res, 113- 1"15. 157168
Yu JZ; Yang H; Zhang H; Lau AKH. (2004). Size distributions of water-soluble organic carbon in ambient
aerosols and its size-resolved thermal characteristics. Atmos Environ, 38- 1061-1071. 156172
Zanis P; Trickl T> Stohl A; Wernli H; Cooper Oi Zerefos Ci Gaeggeler Hi Schnabel Ci Tobler Li Kubik PWi Priller
Ai Scheel HE; Kanter HJi Cristofanelli Pi Forster Ci James Pi Gerasopoulos E> Delcloo A> Papayannis A>
Claude H. (2003). Forecast, observation and modelling of a deep stratospheric intrusion event over
Europe. , 3: 763-777. 053423
Zanobetti A; Wand MP; Schwartz J; Ryan LM. (2000). Generalized additive distributed lag models' quantifying
mortality displacement. , 1- 279-292. 004133
Zeger SL; Thomas D; Dominici F; Samet JM; Schwartz J; Dockery D; Cohen A. (2000). Exposure measurement
error in time-series studies of air pollution- concepts and consequences. Environ Health Perspect, 108:
419-426. 001949
Zeng Y. (2006). A comprehensive particulate matter monitoring system and dosimetry "based ambient
particulate matter standards. J Air Waste Manag Assoc, 56- 518-29. 098375
Zhang K; Wexler A. (2008). Modeling urban and regional aerosols—Development of the UCD Aerosol Module
and implementation in CMAQ model. Atmos Environ, 42- 3166-3178 . 191770
Zhang Q; Jimenez JL; Canagaratna MR; Jayne JT; Worsnop DR. (2005). Time- and size-resolved chemical
composition of submicron particles in Pittsburgh- Implications for aerosol sources and processes. J
Geophys Res, 110: 1-19. 157185
Zhang X; Kondragunta S; Schmidt C; Kogan F. (2008). Near Real-time Monitoring of Biomass Burning
Particulate Emissions (PM25) across Contiguous United States Using Multiple Satellite Instruments.
Atmos Environ, 42(29): 6959-6972. 155144
Zhang X; Zhuang G; Guo J; Yin K; Zhang P. (2007). Characterization of aerosol over the Northern South China
Sea during two cruises in 2003. Atmos Environ, 41: 7821-7836. 101119
Zhang Z; Che W', Liang Y; Wu M; Li N; Shu Y; Liu F; Wu D. (2007). Comparison of cytotoxicity and genotoxicity
induced by the extracts of methanol and gasoline engine exhausts. Toxicol In Vitro, 21: 1058-1065.
157186
Zhao W; Hopke PK; Norris G; Williams R; Paatero P. (2006). Source apportionment and analysis on ambient
and personal exposure samples with a combined receptor model and an adaptive blank estimation
strategy. Atmos Environ. 156181
Zhao W; Rabinovitch N; Hopke PK; Gelfand EW. (2007). Use of an expanded receptor model for personal
exposure analysis in schoolchildren with asthma. Atmos Environ, 41: 4084-4096. 156182
Zheng M; Cass GR; Schauer JJ; Edgerton ES. (2002). Source apportionment of PM2.5 in the southeastern
United States using solvent-extractable organic compounds as tracers. Environ Sci Technol, 36: 2361"
2371. 026100
Zheng M; Ke L; Edgerton ES; Schauer JJ; Dong M; Russell AG. (2006). Spatial distribution of carbonaceous
aerosol in the southeastern United States using molecular markers and carbon isotope data. J Geophys
Res, 111: D10S06. 157189
Zhou Y; Levy JI. (2007). Factors influencing the spatial extent of mobile source air pollution impacts: a meta-
analysis. BMC Public Health, 7: 89. 098633
Zhu K; Zhang J; Lioy PJ. (2007). Evaluation and comparison of continuous fine particulate matter monitors for
measurement of ambient aerosols. J Air Waste Manag Assoc, 57: 1499-506. 098367
Zhu Y; Eiguren-Fernandez A; Hinds WC; Miguel AH. (2007). In-Cabin Commuter Exposure to Ultrafine
Particles on Los Angeles Freeways. Environ Sci Technol, 41: 2138-2145. 179919
Zhu Y; Hinds WC. (2002). Concentration and size distribution of ultrafine particles near a major highway. J Air
Waste Manag Assoc, 52: 1032-1042. 041552
Zhu Y; Hinds WC; Shen S; Sioutas C. (2004). Seasonal trends of concentration and size distribution of ultrafine
particles near major highways in Los Angeles. Aerosol Sci Technol, 38: 5-13. 156184
Zhu Y; Kuhn T; Mayo P; Hinds WC. (2005). Comparison of Daytime and Nighttime Concentration Profiles and
Size Distributions of Ultrafine Particles near a Major Highway. Environ Sci Technol, 39: 2531-2536.
157191
Zidek JV; Shaddick G; Meloche J; Chatfield C; White R. (2007). A framework for predicting personal exposures
to environmental hazards. , 14: 411-431. 190076
Zidek JV; Wong H; Le ND; Burnett R. (1996). Causality, measurement error and multicollinearity in
epidemiology. Environmetrics, T- 441-451. 051879
July 2009
3-294
DRAFT-DO NOT CITE OR QUOTE

-------
July 2009	295	DRAFT - DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Chapter 4. Dosimetry
4.1. Introduction
Particle dosimetry refers to the characterization of deposition, translocation,
clearance, and retention of particles and their constituents within the respiratory tract and
extrapulmonary tissues. This chapter summarizes basic concepts presented in dosimetry
chapters of the 1996 and 2004 PM AQCDs (U.S. EPA, 1996, 079380; U.S. EPA, 2004,
056905), and updates the state of the science based upon new literature appearing since
publication of these PM AQCDS. Although the basic understanding of the mechanisms
governing deposition and clearance of inhaled particles has not changed, there has been
significant additional information on the role of certain biological determinants such as
gender, age and lung disease on deposition and clearance. Additionally, new studies have
further characterized the retention and translocation of ultrafine particles (also commonly
referred to as nanoparticles) following deposition in the respiratory tract.
The dose from inhaled particles deposited and retained in the respiratory tract is
governed by a number of factors. These include exposure concentration and duration,
activity and ventilatory parameters, and particle properties (e.g., particle size,
hygroscopicity, and solubility in airway fluids and cellular components). The basic
characteristics of particles as they relate to deposition and retention, as well as anatomical
and physiological factors influencing particle deposition and retention, were discussed in
depth in Chapter 10 of 1996 PM AQCD and updated in Chapter 6 of the 2004 PM AQCD.
Species differences between humans and rats in particle exposures, deposition patterns,
and pulmonary retention were also reviewed in Brown et al. (2005, 089308). The current
review of PM dosimetry focuses mainly on issues that may affect the susceptibility of an
individual to adverse effects as well as issues that affect our ability to extrapolate findings
between studies (e.g., in vitro to in vivo) and between species. Other than a brief overview
in this introductory section, the disposition (i.e., deposition, absorption, distribution,
metabolism, and elimination) of fibers and unique nano-objects (viz., dots, hollow spheres,
rods, fibers, tubes) is not reviewed herein. Exposures to fibers and unique nano-objects
generally occur in the occupational settings rather than the ambient environment.
The deposition by interception of micro-sized fibers was briefly discussed in the 1996
and 2004 PM AQCD, but fiber retention in the respiratory tract was not addressed.
Airborne fibers (length/diameter ratio > 3), frequently exceed 150 pm in length and appear
Note- Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health
and Environmental Research Online) at http• //epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in
the process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009
4-1
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
to be relatively stable in air. This is because their aerodynamic size is determined
predominantly by their diameter, not their length. Fibers, therefore, deposit largely by
interception and when aligned with the direction of airflow may penetrate deep into the
respiratory tract. Once deposited, macrophage mediated clearance is the primary
mechanism of removing micro-sized particles from the pulmonary region. The length of
fibers can, however, prevent their phagocytosis and clearance. For example, fibers of
greater than 17 um in length are too long to be fully engulfed by rat alveolar macrophages
and can protrude from macrophages (i.e., macrophage frustration) (Zeidler-Erdely et al.,
2006, 190967). Further discussion of the fiber disposition in the respiratory tract is beyond
the scope of this chapter.
The term "ultrafine particle" has traditionally been used by the aerosol research and
occupational and environmental health communities to describe airborne particles or other
laboratory generated aerosols used in toxicological studies that are smaller than 100 nm in
size (based on physical size, diffusivity, or electrical mobility). Generally consistent with the
definition of an ultrafine particle, the International Organization for Standardization (ISO)
recently defined a nanoparticle as an object with all three external dimensions in the
nanoscale, i.e., from approximately 1 and 100 nm (ISO, 2008, 190066). The ISO also defined
a nano-object as a material with one or more external dimensions in the nanoscale. The
terms, nanoparticle and ultrafine particle, have been used rather synonymously in the
recent literature. However, the terms nanoparticle and nano-object are more commonly
associated with engineered materials that are created for consumer products and industrial
applications. With the current interest in nanotechnologies, many nano-objects have been
created by manipulating materials at the atomic or molecular scale for the purpose of
forming new materials, structures, and devices that exploit the unique physical and
chemical properties associated with their nanoscale. Toxicological studies are becoming
available that evaluate in vivo translocation and heath effects unique of nano-objects (viz.,
dots, hollow spheres, rods, tubes). The in vivo disposition of these unique nano-objects is
not, however, necessarily relevant to the behavior of ultrafine aerosols in the urban
environment that are created by combustion sources and photochemical formation of
secondary organic aerosols. Therefore, the disposition of unique nano-objects (viz., dots,
hollow spheres, rods, fibers, tubes) is not considered in this chapter.
4.1.1. Size Characterization of Inhaled Particles
Particle size is a major determinant of the fraction of inhaled particles depositing in
and cleared from various regions of the respiratory tract. The distribution of particle sizes
in an aerosol is typically described by the lognormal distribution (i.e., the situation in which
the logarithms of particle diameter are distributed normally). The geometric mean is the
July 2009
4-2
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
median of the distribution, and the variability around the median is the geometric standard
deviation (GSD or og) and is given by:
Q SD = C = ^84% = ^50%
^50% ^16%
Equation 4-1
where: di6%, d5o%, ds4% are the particle diameters associated with the 16th, 50th (i.e., the
median), and the 84th percentiles from the cumulative frequency distribution of particle
sizes. By definition, GSD must be greater than one. The particle size associated with any
percentile of the distribution, di, is given by:
d = d 
-------
1
2
3
4
5
6
7
8
9
10
11
the tracheobronchial (TB) region which is the conducting airways and the alveolar region
which is the functional part or parenchyma of the lung. A recent review of interspecies
similarities and differences in the structure and function of the respiratory tract is provided
by Phalen et al. (2008, 156865). Although the structure varies, the illustrated anatomic
regions are common to all mammalian species with the exception of the respiratory
bronchioles. Respiratory bronchioles, the transition region between ciliated and fully
alveolated airways, are found in humans, dogs, ferrets, cats, and monkeys. Respiratory
bronchioles are absent in rats and mice and abbreviated in hamsters, guinea pigs, oxen,
sheep, and pigs. The branching structure of the ciliated bronchi and bronchioles also differs
between species from being a rather symmetric and dichotomous branching network of
airways in humans to a more monopodial branching network in other mammals.
Posterior
Nasal Passage
Nasal Part	
Oral Part	
Extrathoracic
Region
Pharynx
Larynx
Trachea
Main Bronchi
Tracheobronchial
Region
Bronchi
Bronchioles
Bronchiolar Region
Alveolar Interstitial
AI
Bronchioles
— Terminal Bronchioles
Respiratory Bronchioles
Alveolar
Region
AI
Alveolar Duct +
Alveoli
Source: Based on ICRP (1994,006988).
Figure 4-1. Diagrammatic representation of respiratory tract regions in humans. Structures are
anterior nasal passages, ETi; oral airway and posterior nasal passages, ET2; bronchial
airways, BB; bronchioles, bb; and alveolar interstitial, AI.
July 2009
4-4
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
Another species difference relevant to particle dosimetry is the route of breathing. For
instance, rats are obligate nose breathers, whereas most humans are oronasal breathers
who breathe through the nose when at rest and increasingly through the mouth with
increasing activity level. There is inter-individual variability in the route by which people
breathe. Most people, 87% (26 of 30) in the Niinimaa et al. (1981, 071758) study, breathed
through their nose until an activity level was reached where they switched to oronasal
breathing. Thirteen percent (4 of 30) of the subjects, however, were oronasal breathers even
at rest. These two subject groups are commonly referred to in the literature (e.g. ICRP,
1994, 006988) as "normal augmenters" and "mouth breathers," respectively. In contrast to
healthy subjects, Chadha et al. (1987, 037365) found that the majority (11 of 12) of patients
with asthma or allergic rhinitis breathe oronasally even at rest.
Bronchus
Aveolus
Bronchiolus
i
Tissue
Air
Tissue
Air
Tissue
Blood
Tissue
Air
Source: Panel (a) reproduced with permission (ER Weibel, Design and structure of the human lung,
In: Pulmonary Diseases and Disorders, ed. AP Fishman, McGraw-Hill, New York, 1980, p. 231 KFishman and Elias, 1980,156436)
Figure 4-2. Structure of lower airways with progression from the large airways to the alveolus.
Panel (a) illustrates basic airway anatomy. Structures are epithelial cells, EP; basement
membrane, BM; smooth muscle cells, SM; and fibrocartilaginous coat, FC. Panel (b)
illustrates the relative amounts of liquid, tissue, and blood with distal progression.
July 2009
4-5
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
The site of particle deposition within the respiratory tract has implications related to
lung retention and surface dose of particles as well as potential systemic distribution of
particles or their constituents. Figure 4-2 illustrates the progressive change in airway
anatomy with distal progression into the lower respiratory tract. In the bronchi there is a
thick liquid lining and mucociliary clearance rapidly moves deposited particles toward the
mouth. In general, in the bronchi, only highly soluble materials moving from the air into
the liquid layer will have systemic access via the blood. With distal progression, the
protective liquid lining diminishes and clearance rates slow. Soluble compounds and some
poorly soluble ultrafine particles may cross the air-liquid interface to enter the tissues and
the blood especially in the alveolar region.
4.2. Particle Deposition
Inhaled particles may be either exhaled or deposited in the ET, TB, or alveolar region.
A particle becomes deposited when it moves from the airway lumen to the wall of an airway.
The deposition of particles in the respiratory tract depends primarily on inhaled particle
size, route of breathing (through the nasal or oronasal), tidal volume (Vt), breathing
frequency (B, and respiratory tract morphology. The distinction between air passing
through the nose versus the mouth is important since the nasal passages more effectively
remove inhaled particles than the oral passage. Respiratory tract morphology, which affects
particle transport and deposition, varies between species, the size of an animal or human,
and health status.
The fraction of inhaled aerosol becoming deposited in the human respiratory tract has
been measured experimentally. Studies, using light scattering or particle counting
techniques to quantify the amount of aerosol in inspired and expired breaths, have
characterized total particle deposition for varied breathing conditions and particle sizes.
The vast majority of in vivo data on the regional particle deposition has been obtained by
scintigraphic methods. These data have shown highly variable regional deposition with
sites of highly localized deposition or "hot spots" in the obstructed lung relative to the
healthy lung. Even in the healthy lung, "hot spots" occur in the region of airway
bifurcations. Mathematical models aid in predicting the mixed effects of particle size,
breathing conditions, and lung volume on total and regional deposition. Experimentally,
however, there is considerable inter-individual variability in total and regional deposition
even when inhaled particle size and breathing conditions are strictly controlled.
Section 4.2.4 on Biological Factors Modulating Deposition provides more detailed
information on factors affecting deposition among individuals.
In order to potentially become deposited in the respiratory tract, particles must first
be inhaled. The inspirable particulate mass fraction of an aerosol is that fraction of the
ambient airborne particles that can enter the uppermost respiratory tract compartment,
July 2009
4-6
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
the head (Soderholm, 1985, 156992). The American Conference of Governmental Industrial
Hygienists (ACGIH) and the International Commission on Radiological Protection (ICRP)
have established inhalability criteria for humans (ACGIH, 2005, 156188; ICRP, 1994,
006988). These criteria are indifferent to route of breathing and assume random orientation
with respect to wind direction. They are based on experimental inhalability data for dae <
100 pm at wind speeds of between 1 and 8 m/s. For ACGIH criterion, inhalability is 97% for
an dae = 1 pm, 87% for an dae = 5 pm, 77% for an dae = 10 pm, and plateaus at 50% dae above
-40 pm. The ICRP criterion, which also plateaus at 50% for very large dae, does not become
of real importance until an dae = 5 pm where inhalability is 97%. Dai et al. (2006, 156377)
reported slightly lower nasal particle inhalability in humans during moderate exercise than
rest (e.g., 89.2 vs. 98.1% for 13 pm particles, respectively). Nasal particle inhalability is
similar between an adult and 7-year-old child (Hsu and Swift, 1999, 155855). Inhalability
into the mouth from calm air in humans also becomes important for dae >10 pm (Anthony
and Flynn, 2006, 155659; Brown, 2005, 156299). Unlike the inhalability from high wind
speeds which plateaus at 50% for dae greater than -40 pm, particle inhalability from calm
air continues to decrease toward zero with increasing dae.
Inhalability data in laboratory animals, such as rats, are only available for breathing
from relatively calm air (velocity < 0.3 m/s). For nasal breathing, inhalability becomes an
important consideration for dae of above 1 pm in rodents and 10 pm in humans (Menache et
al., 1995, 006533). The inhalability of particles having dae of 2.5, 5, and 10 pm is 80, 65, and
44% in rats, respectively, whereas it only decreases to 96% for a dae of 10 pm in humans
during nasal breathing (Menache et al., 1995, 006533). Asgharian et al. (2003, 153068)
suggested that an even more rapid decrease in inhalability with increasing dae may occur in
rats. Inhalability is a particularly important consideration for rodent exposures.
Section 4.2.3 provides additional discussion of interspecies patterns of particle deposition.
4.2.1. Mechanisms of Deposition
Particle deposition in the lung is predominantly governed by diffusion, impaction, and
sedimentation. Most discussion herein focuses on these three dominant mechanisms of
deposition. Simple interception, which is an important mechanism of fiber deposition, is not
discussed in this chapter. Electrostatic and thermophoretic forces as mechanisms of
deposition have not been thoroughly evaluated and receive limited discussion. Some
generalizations with regard to deposition by these mechanisms follows, but should not be
viewed as absolute rules. Both experimental studies and mathematical models have
demonstrated that breathing patterns can dramatically alter regional and total deposition
for all sized particles. The combined processes of aerodynamic and diffusive (or
thermodynamic) deposition are important for particles in the range of 0.1 pm to 1 pm.
July 2009
4-7
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Aerodynamic processes predominate above and thermodynamic processes predominate
below this range.
Diffusive deposition, by the process of Brownian diffusion, is the primary mechanism
of deposition for particles having physical diameters of less than 0.1 pm. For particles
having physical diameters of roughly between 0.05 and 0.1 pm, diffusive deposition occurs
mainly in the small distal bronchioles and the pulmonary region of the lung. However, with
further decreases in particle diameter below -0.05 pm, increases in particle diffusivity shift
more deposition proximally to the bronchi and ET regions.
Governed by inertial or aerodynamic properties, impaction and sedimentation
increase with dae. When a particle has sufficient inertia, it is unable to follow changes in
flow direction and strikes a surface thus depositing by the process of impaction. Impaction
occurs predominantly at bifurcations in the proximal airways, where linear velocities and
secondary eddies are at their highest. Sedimentation, caused by the gravitational settling of
a particle, is most important in the distal airways and pulmonary region of the lung. In
these regions, residence time is the greatest and the distances that a particle must travel to
reach the wall of an airway are minimal.
The electrical charge on some particles may result in an enhanced deposition over
what would be expected from size alone. With an estimated charge of 10-50 negative ions
per 0.5 pm particle, Scheuch et al. (1990, 006948) found deposition in humans (Vt = 500
mL, /= 15 min"1) to increase from 13.4% (no charge) to 17.8% (charged). This increase in
deposition is thought to result from image charges induced on the surface of the airway by
charged particles. Yu (1985, 006963) estimated a charge threshold level above which
deposition fractions would be increased of about 12, 30, and 54% for 0.3, 0.6, and 1.0 pm
diameter particles, respectively. Electrostatic deposition is generally considered negligible
for particles below 0.01 pm because so few of these particles carry a charge at Boltzmann
equilibrium. This mechanism is also thought to be a minor contributor to overall particle
deposition, but it may be important in some laboratory studies due to specific aerosol
generation techniques such as nebulization. Laboratory methods such as passage of the
aerosols through a Rr85 charge neutralizer prior to inhalation are commonly used to
mitigate this effect.
The National Radiological Protection Board (NRPB) recently evaluated the potential
for corona discharges from high voltage power lines to charge particles and enhance
particulate doses (NRPB, 2004, 156815). They concluded that electrostatic effects would be
the most important for particles in the size range from about 0.1" 1 pm, where deposition
may theoretically increase by a factor of three to ten. However, given that the small fraction
of ambient particles would pass through the corona to become charged, the small range of
relevant particle sizes (0.1-1 pm), and the subsequent required transport of charged
particles to expose individuals! the NRPB concluded that effects, if any, of electric fields on
particle deposition in the human respiratory tract would likely be minimal.
July 2009
4-8
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Thermophoretic forces on particles occur due to temperature differences between
respired air and respiratory tract surfaces. Temperature gradients of around 20°C are
thought to produce sufficient thermophoretic force to oppose diffusive and electrostatic
deposition during inspiration and to perhaps augment deposition by these mechanisms
during expiration (Jeffers, 2005, 156608). Thermophoresis is only relevant in the
extrathoracic and large bronchi airways and reduces to zero as the temperature gradient
decreases deeper in the lung. Theoretical analysis of thermophoresis has been done for
smooth walled tubes and is important over distances that are several orders of magnitude
smaller than the diameter of the trachea. The alteration of the flow patterns by airway
surface features such as cartilaginous rings may affect particle transport and deposition
over far greater distances than thermophoretic force.
4.2.2. Deposition Patterns
Knowledge of sites where particles of different sizes deposit in the respiratory tract
and the amount of deposition therein is necessary for understanding and interpreting the
health effects associated with exposure to particles. Particles deposited in the various
respiratory tract regions are subjected to large differences in clearance mechanisms and
pathways and, consequently, retention times. Deposition patterns in the human respiratory
tract were described in considerable detail in dosimetry chapters of prior PM AQCD
(U.S. EPA, 1996, 079380; U.S. EPA, 2004, 056905); as such, they are only briefly described
here.
Predicted total and regional deposition for an adult male during rest and light
exercise are illustrated in Figures 4-3 and 4-4, respectively. Note that a large proportion of
inhaled coarse particles in the 3-6 micron (dae) range can reach and deposit in the lower
respiratory tract, particularly the TB airways. Although these figures were provided in
Chapter 6 of the 2004 PM AQCD, they are reproduced here to illustrate changes in
deposition as a function of particle size and breathing conditions. The predictions were
based on two publicly available particle deposition models, the ICRP (1994, 006988) and the
Multi-Path Particle Dosimetry model (MPPD; Version 1.0, ©2002). The ICRP (1994,
006988) model was implemented by Lung Dose Evaluation Program (LUDEP; Version 2.07,
June 2000). The MPPD1 model was developed by the CUT Centers for Health Research
with support from the Dutch National Institute of Public Health and the Environment.
1 For more information about this model, the reader is referred to- http-//www.ara.com/products/mDDd capabilities.htm.
July 2009	4-9	DRAFT - DO NOT CITE OR QUOTE

-------
1.0
Total y
Breathing
Breathing
<£> and + Total
0.8
Total
0.6
C
0.4
S
0.2
ft' "fa
0.0
(
d
0.6
Mouth
Breathing
0.5
0.4
0.3
I
0.2-
&
0.1
TB
TB
0.04
0.'
Figure 4-3. Comparison of total and regional deposition results from the ICRP and MPPD models for
a resting breathing pattern (Vt = 625 mL, f = 12 min1) and corrected for particle
inhalahility. Regions are extrathoracic, ET; tracheobronchial, TB; and alveolar, A. Panels
a-b are for nose breathing; panels c-d are for mouth breathing.
Nose Breathing
Mouth Breathing
Figure 4-4. Comparison of total and regional deposition results from the ICRP and MPPD models for
a light exercise breathing pattern (Vt = 1250 mL, f= 20 min1) and corrected for particle
inhalahility. Regions are extrathoracic, ET; tracheobronchial, TB; and alveolar, A. Panels
a-b are for nose breathing; panels c-d are for mouth breathing.
July 2009
4-10
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
4.2.2.1. Total Respiratory Tract Deposition
The efficiency of deposition in the respiratory tract may generally be described as a
"U-shaped" curve on a plot of deposition efficiency versus the of log particle diameter. Total
deposition shows a minimum for particle diameters in the range of 0.1 to 1.0 pm, where
particles are small enough to have minimal sedimentation or impaction and sufficiently
large so as to have minimal diffusive deposition. Total deposition does not decrease to zero
for any sized particle because of mixing between particle laden tidal air and residual lung
air. The particles mixed into residual air remain in the lung following a breath and are
removed on subsequent breaths or gradually deposited. Total deposition approaches 100%
for particles of roughly 0.01 pm (physical diameter) due to diffusive deposition and for
particles of around 10 pm (dae) due to the efficiency of sedimentation and impaction.
Total human lung deposition, as a function of particle size, is depicted in Figure 4-5.
These experimental data were obtained by using monodisperse spherical test particles in
healthy adults during controlled breathing on a mouthpiece. Despite the control of inhaled
particle size and breathing conditions, these figures illustrate variability in deposition
efficiencies due to inter-individual differences in lung size and anatomical variability in
airway dimensions and branching patterns.
c
o
o
c
o
'
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
have become available, but are largely derived from computational fluid dynamics (CFD)
modeling and experimental measurements in casts. As most of these studies do not
substantially improve our understanding of deposition in the ET region they are not
reviewed here.
For particles >1 pm dae, deposition efficiency in the oral and nasal passages has been
generally described as a function of an impaction parameter (Stokes number) with the
addition of a flow regime parameter (Reynolds number) for the oral passages (Finlay and
Martin, 2008, 155776; Grgic et al., 2004, 155810; Kelly et al., 2005, 155894; Schroeter et al.,
2006, 156076). For an adult male, the CFD simulations of Schroeter et al. (2006, 156076)
predicted nasal deposition of 10 pm dae particles was 90%, and 100% for a Ve of 7.5 L/min
(rest) and 15 L/min (light activity), respectively. Thus, relatively few large coarse particles
will pass through the nasal passages into the lungs.
Since the nasal passages are more efficient at removing inhaled particles than the
oral passage, an individual's mode of breathing (i.e., oral vs. nasal) influences the quantity
of particles penetrating to the lung. In limited studies, it has been shown that children tend
to have more oral breathing both at rest and during exercise and also displayed more
variability than adults (Becquemin et al., 1999, 155679; Bennett et al., 2008, 156269;
James et al., 1997, 042422). In contrast to adults, there is little data on the uptake of
particles for oral or nasal breathing in children. Theoretical calculations by Xu and Yu
(1986, 072697) predict enhanced deposition of particles (greater than 2 pm) in the head
region for children when compared to adults. Studies of fine particle deposition in physical
models of the nose, scaled to adult vs. children sizes, predict that deposition efficiency in
the nose is a function of pressure drop across the nose (Phalen et al., 1989, 156023).
Consequently, these model analyses suggest that, when properly scaled physiological flows
are used in the calculation of nasal deposition, children, who have higher nasal resistance
than adults, should have higher nasal deposition compared to adults. Surprisingly, the few
studies reporting measures of nasal deposition in children, found lower nasal deposition
efficiencies for fine particles (1-3 pm dae) as compared to adults, despite their higher nasal
resistances (Becquemin et al., 1991, 009187; Bennett et al., 2008, 156269). These findings of
lesser nasal vs oral breathing and less efficient nasal deposition suggest that children's
lower respiratory tract (i.e., the TB and alveolar regions) may receive a higher dose of
ambient PM compared to adults. Normalized to lung surface area, the dose rate to the lower
airways of children vs. adults is increased further because children breathe at higher
minute ventilations relative to their lung volumes (see Section 4.2.4.2 on Age as a factor
modulating deposition).
4.2.2.3. Tracheobronchial and Alveolar Region
Inhaled particles passing the ET region enter and may become deposited in the lungs.
For any given particle size, the pattern of particle deposition influences clearance by
July 2009
4-12
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
partitioning deposited material between lung regions. Deposition in the tracheobronchial
airways and alveolar region cannot be directly measured in vivo. Much of the available
deposition data for the TB and A regions have been obtained from experiments with
radioactively labeled, poorly soluble particles (U.S. EPA, 1996, 079380) or by use of aerosol
bolus techniques (U.S. EPA, 2004, 056905). In general, the ability of these experimental
data to define specific sites of particle deposition is limited to anatomically large regions of
the respiratory tract such as the head, larynx, bronchi, bronchioles, and alveolar region.
Mathematical modeling can provide more refined predictions of deposition sites.
Comparisons of the modeling results obtained with two publicly available models were
provided in Figures 4-3 and 4-4. Highly localized sites of deposition within the bronchi are
described in Section 4.2.2.4. Both experimental and modeling techniques are based on many
assumptions that may be relatively good for the healthy lung but not for the diseased lung.
For discussion of these issues, the reader is referred to Sections 4.2.4.4 and 4.2.4.5.
4.2.2.4. Localized Deposition Sites
From a toxicological perspective, it is important to realize that not all epithelial cells
in an airway will receive the same dose of deposited particles. Localized deposition in the
vicinity of airway bifurcations has been analyzed using experimental and mathematical
modeling techniques. In the 1996 PM AQCD, experimental data were available illustrating
the peak deposition of coarse particles (3, 5, and 7 pm dae) in daughter airways during
inspiration and the parent airway during expiration, but always near the carinal ridge
(Kim and Iglesias, 1989, 078539; Kim et al., 1989, 078538). In the 2004 PM AQCD,
mathematical models predicted distinct "hot spots" of deposition in the vicinity of the
carinal ridge for both coarse (10 pm) and ultrafine (0.01 pm) particles (Heistracher and
Hofmann, 1997, 047514; Hofmann et al., 1996, 047515). In a model of lung generations 4-5
during inspiration, hot spots occurred at the carinal ridge for 10 pm dae particles due to
inertial impaction and for 0.01 pm particles due to secondary flow patterns formed at the
bifurcation. During expiration, preferential sites of deposition for both particle sizes
occurred l) approaching the juncture of daughter airways on the walls forming and across
the lumen from the carinal ridge and 2) the top and bottom (visualizing the Y-shaped
geometry laying horizontal) of the parent airway downstream of the bifurcation.
Recent studies further support these findings (Balashazy et al., 2003, 155671; Farkas
et al., 2006, 155771; Farkas and Balashazy, 2008, 157358; Isaacs et al., 2006, 155861). Most
of these studies quantified localized deposition in terms of an enhancement factor. Typically,
the enhancement factor is the ratio of the deposition in a pre-specified surface area (e.g.,
100 x 100 pm which corresponds to ~ 10 x 10 epithelial cells) to the average deposition
density for the whole airway geometry. These enhancement factors are very sensitive to the
size of the surface considered (Balashazy et al., 1999, 043201). The studies by Farkas et al.
(2006, 155771) and Farkas and Balashazy (2008, 157358) investigated the phenomena of
July 2009
4-13
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
localized deposition down to 0.001 pm particles. The deposition of 0.001 pm was rather
uniform, however, the deposition pattern became increasingly less uniform with increasing
particle size. These studies indicate that, for particles greater than -0.01 pm, some cells
located near the carinal ridge of bronchial bifurcations may receive hundreds to thousands
times the average dose (particles per unit surface area) of the parent and daughter airways.
Furthermore, the inertial impaction of particles > 1 pm dae at the carinal ridge of large
bronchi will increase with increasing inspiratory flows. In a comparison of constricted
versus healthy airways, Farkas et al. (2006, 155771) also reported that the overall
deposition efficiency of 10 pm dae particles at bifurcations downstream of a constriction may
be increased by 18 times. Given these considerations, Phalen and Oldham (2006, 156024)
noted that substantial doses of particles (> 1 pm dae) may be justified for in vitro studies
using tracheobronchial epithelial cell cultures.
4.2.3. Interspecies Patterns of Deposition
The primary purpose of this document is to assess the health effects of particles in
humans. As such, human dosimetry studies have been stressed in this chapter. Such
studies avoid the uncertainties associated with the extrapolation of dosimetric data from
laboratory animals to humans. However, animal models have been and continue to be used
in evaluating PM health effects because of ethical considerations regarding the types of
studies that can be performed with human subjects. Thus, there is a considerable need to
understand dosimetry in animals and dosimetric differences between animals and humans.
Limited new data are becoming available. Similar deposition efficiencies have been
reported in nasal casts of human and rhesus monkey for 1 to 10 pm dae for inspiratory flows
mimicking resting breathing patterns (Kelly et al., 2005, 155894). Oldham and Robinson
(2007, 156003) recently provided morphological data and predicted particle deposition in an
asthma mouse model.
Interspecies similarities and differences in deposition were described in detail in the
last two PM AQCDs (U.S. EPA, 1996, 079380; U.S. EPA, 2004, 056905). It was concluded
that the general pattern of total particle deposition efficiency was similar between
laboratory animals and humans^ deposition increases on both sides of a minimum that
occurs for particles of 0.2 to 1 pm. There are, however, marked interspecies differences in
uptake into the respiratory tract and regional deposition. For instance, the nasal
inhalability of 10 pm dae particles is predicted to be 96% in humans, whereas it is only 44%
in rats (Menache et al., 1995, 006533). In most laboratory animal species (rat, mouse,
hamster, guinea pig, and dogs), deposition in the ET region is near 100% for particles
greater than 5 pm dae (Raabe et al., 1988, 001439), indicating greater efficiency than that
seen in humans. Detailed presentation of dosimetric difference between rats and humans
are available elsewhere (Brown et al., 2005, 089308; Jarabek et al., 2005, 056756).
July 2009
4-14
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Brown et al. (2005, 089308) conducted a thorough evaluation of extrapolations
between rats and humans in relation to PM exposures. One of many factors they considered
was the choice of a dose metric appropriate for comparison between species. For example,
deposited mass may be an appropriate PM indicator for health effects associated with
soluble PM constituents. For health effects associated with insoluble PM, the particle
number, surface area, or mass may be appropriate indicators. Given interspecies differences
in deposition patterns and clearance rates, the question of retained versus deposited dose
was also discussed. It was concluded that for acute effects, the maximum deposited
incremental dose may be the appropriate type of dose metric. For chronic effects, long-term
burden maybe more appropriate. For various dose metrics, estimates of particle
concentration and exposure duration required for a rat to receive the same dose as received
by a human were obtained with consideration of activity levels and particle size
distributions. It was noted that high PM exposures over the period of months can lead to
particle overload in rats (see Section 4.3.4.4). Exposure regimes were derived as a function
of particle size and exposure duration that should avoid overwhelming macrophage
mediated clearance achieving particle overload in rats (see Table 12 in Brown et al., 2005,
089308). The dosimetric calculations indicated that to achieve nominally similar acute
doses per surface area in rats, relative to humans undergoing moderate to high exertion,
PM exposure concentrations for rats would need to be somewhat higher than for humans.
Since particle clearance from the lungs of rats is faster than humans, much higher exposure
concentrations are required for the rat to simulate retained burdens of humans. Illustrating
the complexity of such analyses, in some cases, rats were found to require lower exposures
than humans to have comparable doses (generally when considering a scenario of humans
at rest).
4.2.4. Biological Factors Modulating Deposition
Evaluation of factors affecting particle deposition is important to help understand
potentially susceptible subpopulations. Differences in biological response following
pollutant exposure may be caused by dosimetry differences as well as by differences in
innate sensitivity. The effects of different biological factors on deposition were discussed in
the 2004 PM AQCD (U.S. EPA, 2004, 056905) and are summarized briefly here.
4.2.4.1. Physical Activity
The activity level of an individual is well recognized to affect their minute ventilation
and route of breathing. Changes in minute ventilation during exercise are accomplished by
increasing both Vt and /(Table 4-1). Humans are oronasal breathers tending to breathe
through the nose when at rest and increasingly through the mouth with increasing activity
July 2009
4-15
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
level. There is considerable inter-individual variability in both the route by which people
breathe the way breathing pattern changes occur.
Table 4-1. Breathing patterns with activity level in adult human male.
Activity
Awake
Rest"
Slow
Walk"
Light
Exertion"
Moderate
Exertion"
Heavy
Exertion b
Breaths/min
12
16
19
28
26
Tidal volume, mi-
625
813
1000
1429
1923
Minute ventilation, L/min
7.5
13
19
40
50
Sources: "Winter-Sorkina and Cassee (2002, 043670); bNCRP (1994, 006988)
Individuals typically breathe through their nose while at rest, switching to oronasal
breathing as ventilation increases (Bennett et al., 2003, 191977; Niinimaa et al., 1981,
071758). The role of the nose in filtering particles is diminished as airflow is diverted from
the nose to the mouth during exercise, bringing more particles to the lower respiratory
tract. A recent study in adults (Bennett et al., 2003, 191977) found that nasal ventilation
during exercise varied as a function of both race and gender. African-Americans possessed a
greater nasal contribution to breathing during exercise than Caucasians. At similar
exercise efforts (i.e., normalized to a % maximum work capacity) the females also had a
greater nasal contribution to breathing during exercise than males.
In addition, when individuals increase their ventilation with activity the total number
of particles inhaled per unit time (i.e., exposure rate) increases, but the fractional
deposition of particles in each breath also changes with breathing pattern. Figures 4-3 and
4-4 illustrate predicted deposition fractions in the respiratory tract during rest vs. light
exercise, respectively. During exercise, both Vt and /increase. Fractional deposition for all
particles increases with increased Vt. Increasing the / however, decreases the fractional
deposition of fine and ultrafine particles due to decreased time for gravitational and
diffusive deposition. For particles of larger than a dae of roughly 3 pm, increasing /can
increase the deposition fraction due to increased impaction in the extrathoracic and TB
airways. Thus, it should be expected that the change in deposition fraction with activity
will vary among individuals depending on the relative influences of these two variables (i.e.,
Vt and f) in a given subject and the particle size to which they are exposed. Experimentally,
the lung deposition fractions of fine particles during moderate exercise and mouth
breathing are unchanged between rest and exercise (Bennett et al., 1985, 190034; Morgan
et al., 1984, 190035). Kim (2000, 013112) evaluated differences in deposition of 1, 3, and 5
pm (MMAD) particles under varying breathing patterns (simulating breathing conditions of
sleep, resting, and mild exercise). Total lung deposition increased with increasing Vt at a
given flow rate and with increasing flow rate at a given breathing period. These
experimental studies suggest that the total deposited dose rate (i.e., deposition per unit
July 2009
4-16
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
time) of particles will generally increase in direct proportion to the increase in minute
ventilation associated with exercise.
The changes in ventilation, i.e., breathing pattern and flow rate, may also alter the
regional deposition of particles. Coarse particle deposition increases in the TB and ET
regions during exercise due to the increased flow rates and associated impaction. A rapid-
shallow breathing pattern during exercise may result in more bronchial airway vs. alveolar
deposition, while a slow deep pattern will shift deposition to deeper lung regions (Valberg et
al., 1982, 190019). Bennett et al. (1985, 190034) showed for 2.6 pm particles that moderate
exercise shifted deposition from the lung periphery towards ET and larger, bronchial
airways. Similarly, Morgan et al. (1984, 190035) showed that even for fine particles (0.7 pm)
TB deposition was enhanced with exercise. This shift in deposition toward the bronchial
airways results in a much greater dose per unit surface area of tissue in those regions.
Morgan et al. (1984, 190035) also found that the apical-to-basal distribution of fine particles
increased with exercise, i.e., a shift towards increased deposition in the lung apices. This
shift may be less likely for larger particles, however, whose deposition in large airway
bifurcations may preclude their transport to these more apical regions (Bennett et al., 1985,
190034).
4.2.4.2. Age
Airway structure and respiratory conditions vary with age, and these variations may
alter the amount and site of particle deposition in the respiratory tract. It was concluded in
the 2004 PM AQCD (U.S. EPA, 2004, 056905) that significant differences between adults
and children had been predicted by mathematical models and observed in experimental
studies. Studies generally indicated that ET and TB deposition was greater in children and
that children received greater doses of particles per lung surface area than adults.
Deposition studies in the elderly are still quite limited.
A few studies have attempted to measure oronasal breathing in children as compared
to adults (Becquemin et al., 1999, 155679; Bennett et al., 2008, 156269; James et al., 1997,
042422). This is important since particles deposit with greater efficiency in the nose
relative to the mouth, thereby affecting exposure of the lower respiratory tract. James et al.
(1997, 012122) found that children (age 7-16, n = 10) displayed more variability than adults
with respect to their oronasal pattern of breathing with exercise. However, it was not
possible to predict the pattern of the partitioning of ventilation during exercise based on
age, gender, or nasal airway resistance. Further, in a limited number of children (age 8-16,
n = 10), Becquemin et al. (1999, 155679) found that the children tended to display more oral
breathing both at rest and during exercise than the adults. The highest oral fractions were
also found in the youngest children. None of these studies, however, was able to show a
relationship between nasal resistance and the relative contribution of nasal breathing in
their children. Bennett et al. (2008, 156269) made preliminary measurements of the
July 2009
4-17
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
relative contributions of oral versus nasal breathing at rest and during incrementally
graded submaximal exercise on the cycle ergometer for children (age 6-10, n = 12) adults
(age 18-27, n = 11). There was a trend for children to have a lesser nasal contribution to
breathing at rest and during exercise, but the differences from adults were not statistically
significant.
Breathing patterns are well recognized to change with increasing age, i.e., Vt increase
and respiratory rates decrease (Tabachnik et al., 1981, 157036; Tobin et al., 1983, 156122).
Bennett and Zeman (Bennett and Zeman, 1998, 076182) measured deposition fraction of
inhaled, fine particles (2 pm dae) in children as they breathed the aerosol with their natural,
resting breathing pattern. Among the children, variation in deposition fraction, measured
by photometry at the mouth, was highly dependent on intersubject variation in Vt. On the
other hand, they found no difference in deposition fraction for the children vs. adults for
these fine particles. This finding and the modeling predictions (Hofmann et al., 1989,
006922) are explained in part by the smaller Vt and faster breathing rate of children
relative to adults for natural breathing conditions. Bennett et al. (2008, 156269) also
recently reported measures of fine particle (l and 2 pm dae) deposition fraction for
ventilation associated with light exercise in children and adults and showed that, like with
resting breathing, deposition fraction was predicted by breathing pattern and did not differ
or tended to be less in children compared to adults. On the other hand, because children
breathe at higher minute ventilations relative to their lung volumes, the rate of deposition
of fine particles normalized to lung surface area may be greater in children vs. adults
(Bennett and Zeman, 1998, 076182).
Bennett and Zeman (2004, 155686) expanded their measures of fine particle
deposition during resting breathing to a larger group of healthy children (6" 13 yr! 20 boys,
16 girls) and found again that the variation in total deposition, was best predicted by Vt
(r = 0.79, p < 0.001). But both Vt and resting minute ventilation increased with both height
and body mass index (BMI) of the children. Interestingly, these data suggest that for a
given height and age, children with higher BMI have larger minute ventilations and Vt at
rest than those with lower BMI. These differences in breathing patterns as a function of
BMI translated into increased deposition of fine particles in the heaviest children. The rate
of deposition (i.e., particles depositing/time) in the overweight children was 2.8 times that of
the leanest children (p < 0.02). Among all children, the rate of deposition was significantly
correlated with BMI (r = 0.46, p < 0.004). Some of the increased deposition fraction in
heavier children may be due to their elevated Vt, which was well correlated with BMI
(r = 0.72, p < 0.001).
In 62 healthy adults with normal lung function aged 18-80, Bennett et al. (1996,
083284) showed there was no effect of age on the whole lung deposition fraction of 2-pm
particles under natural breathing conditions. Across all subjects, the deposition fraction
was found to be independent of age, depending on breathing period (r = 0.58, p < 0.001) and
July 2009
4-18
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
airway resistance (r = 0.46, p < 0.001). In the same adults breathing with a fixed pattern
(360 mL Vt, 3.4 sec breathing period), there was a mild decrease in deposition with
increasing age, which could be attributed to increased peripheral airspace dimensions in
the elderly.
4.2.4.3.	Gender
Males and females differ in body size, conductive airway size, and ventilatory
parameters; therefore, gender differences in deposition might be expected. In some of the
controlled studies, however, the men and women were constrained to breathe at the same
Vt and f. Since women are generally smaller than men, the increased minute ventilation of
women compared to their normal ventilation could affect deposition patterns. This may
help explain why gender related effects on deposition have been observed in some studies.
Kim and Hu (1998, 086066) assessed the regional deposition patterns of 1-, 3", and 5-
pm MMAD particles in healthy adult males and females using controlled breathing. The
total fractional deposition in the lungs was similar for both genders with the 1-pm particle
size, but was greater in women for the 3- and 5-pm particles. Deposition also appeared to be
more localized in the lungs of females compared to those of males. Kim and Jaques (2000,
012811) measured deposition in healthy adults using sizes in the ultrafine mode
(0.04-0.1 pm). Total fractional lung deposition was greater in females than in males for
0.04- and 0.06-pm particles. The region of peak fractional deposition was shifted closer to
the mouth and peak height was slightly greater for women than for men for all exposure
conditions. The total and regional deposition data from these studies are illustrated in
Figure 4-5. These differences were generally attributed to the smaller size of the upper
airways, particularly of the laryngeal structure, in females.
In another study (Bennett et al., 1996, 083284), the total respiratory tract deposition
of 2-pm particles was examined in adult males and females aged 18-80 years who breathed
with a normal resting pattern. There was a tendency for a greater deposition fraction in
females compared to males. However, since males had greater minute ventilation, the
deposition rate (i.e., deposition per unit time) was greater in males than in females. More
recently, Bennett and Zeman (2004, 155686) found no difference in the deposition of 2-pm
particles in boys versus girls aged 6-13 yr (n = 36).
4.2.4.4.	Anatomical Variability
Anatomical variability, even in the absence of respiratory disease, can affect
deposition throughout the respiratory tract. The ET region is the first exposed to inhaled
particles and, therefore, deposition within this region would reduce the amount of particles
available for deposition in the lungs. Variations in relative deposition within the ET region
will, therefore, propagate through the rest of the respiratory tract, creating differences in
calculated doses from individual to individual.
July 2009
4-19
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
The influence of variations in nasal airway geometry on particle deposition has been
investigated. Cheng et al. (1996, 047520) examined nasal airway deposition in healthy
adults using particles ranging in size from 0.004 to 0.15 pm and at two constant inspiratory
flow rates, 167 and 333 mL/s. Inter-individual variability in deposition was correlated with
the wide variation of nasal dimensions, in that greater surface area, smaller cross-sectional
area, and increasing complexity of airway shape were all associated with enhanced
deposition. Bennett and Zeman (2005, 155687) have also shown that nasal anatomy
influences the efficiency of particle uptake in the noses of adults. For light exercise
breathing conditions in adults, their study demonstrated that nasal deposition efficiencies
for both 1- and 2-pm monodisperse particles were significantly less in African Americans
versus Caucasians. The lesser nasal efficiencies in African-Americans were associated with
both lower nasal resistance and less elliptical nostrils compared to Caucasians.
Within the lungs, the branching structure of the airways may also differ between
individuals. Zhao et al. (2009, 157187) recently examined the bronchial anatomy of the left
lung in patients (132 M, 84 W; mean age 47 years) that under went conventional thoracic
computed tomography scans for various reasons. At the level of the segmental bronchus in
the upper and lower lobes, a bifurcation occurred in the majority of patients. A trifurcation,
however, was observed in 23% of the upper and 18% of the lower lobes. Other more unusual
findings were also reported such as four bronchi arising from the left upper lobe bronchus.
As described in Section 4.2.2.4, deposition can be highly localized near the carinal ridge of
bifurcations. The effect of a bifurcation versus other branching patterns on airflow patterns
and particle deposition has not been described in the literature. Martonen et al. (1994,
000847) showed that a wide blunt carinal ridge shape dramatically affected the flow stream
lines relative to a narrower and more rounded ridge shape. Specifically, there were high
flow velocities across the entire area of the blunt carinal ridge versus a smoother division of
the airstream in the case of the narrow rounded ridge shape. The implication may be that
localized particle deposition on the carinal ridge would increase with ridge width. A similar
situation might be expected for a trifurcation versus a bifurcation. These differences in
branching patterns provide a clear example of anatomical variability between individuals
that might affect both air flow patterns and sites of particle deposition.
4.2.4.5. Respiratory Tract Disease
The presence of respiratory tract disease can affect airway structure and ventilatory
parameters, thus altering deposition compared to that occurring in healthy individuals. The
effect of airway diseases on deposition has been studied extensively, as described in the
1996 and 2004 PM AQCD (U.S. EPA, 2004, 056905; U.S. EPA, 1996, 079380). Studies
described therein showed that people with chronic obstructive pulmonary disease (COPD)
had very heterogeneous deposition patterns and differences in regional deposition
compared to healthy individuals. People with obstructive pulmonary diseases tended to
July 2009
4-20
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
have greater deposition in the TB region than did healthy people. Furthermore, there
tended to be an inverse relationship between bronchoconstriction and the extent of
deposition in the A region, whereas total respiratory tract deposition generally increased
with increasing degrees of airway obstruction.
The vast majority of deposition studies in individuals with respiratory disease have
been performed during controlled breathing, i.e., all subjects breathed with the same Vt
and £ However, although resting Vt is similar or elevated in people with COPD compared to
healthy individuals, the former tend to breathe at a faster rate, resulting in higher than
normal tidal peak flow and resting minute ventilation. Thus, given that breathing patterns
differ between healthy and obstructed individuals, particle deposition data for controlled
breathing may not be appropriate for estimating respiratory doses from ambient PM
exposures.
Bennett et al. (1997, 078839) measured the fractional deposition of insoluble 2-pm
particles in moderate-to-severe COPD patients (n = 13; mean age 62 years) and healthy
older adults (n = 11; mean age 67 years) during natural resting breathing. COPD patients
had about a 50% greater deposition fraction and a 50% increase in resting minute
ventilation relative to the healthy adults. As a result, the patients had an average
deposition rate of about 2.5 times that of healthy adults. Similar to previously reviewed
studies (U.S. EPA, 1996, 079380; U.S. EPA, 2004, 056905), these investigators observed an
increase in deposition with an increase in airway resistance, suggesting that deposition
increased with the severity of airway disease.
Brown et al. (2002, 043216) measured the deposition of an ultrafine aerosol
(CMD = 0.033 jim) during natural resting breathing in 10 patients with moderate-to-severe
COPD (mean age 61 years) and 9 healthy adults (mean age 53 years). The COPD group
consisted of 7 patients with chronic bronchitis and 3 patients with emphysema. The total
deposition fraction in the bronchitic patients (0.67) was significantly (p < 0.02) greater than
in either the patients with emphysema (0.48) or the healthy subjects (0.54). Minute
ventilation increased with disease severity (healthy, 5.8 L/min; chronic bronchitic, 6.9
L/min! emphysema, 11 L/min). Relative to the healthy subjects, the average dose rate was
significantly (p < 0.05) increased by 1.5 times in the COPD patients, whereas the average
deposition fraction only tended to be increased by 1.1 times. These data further
demonstrate the need to consider dose rates (which depend on minute ventilation) rather
than just deposition fractions when evaluating the effect of respiratory disease on particle
deposition and dose.
July 2009
4-21
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Hygroscopic
c
o
w
o
Q.
CD
Q
Hydrophobic
0.0
0.03
0.1
0.4
Inhaled particle diameter (jjm)
Source: Adapted from Tu and Knutson (1984, 072870).
Figure 4-6. Total deposition of hygroscopic sodium chloride and hydrophobic aluminosilicate
aerosols during oral breathing (Vt = 1.0 L; f = 15 min1).
4.2.4.6. Hygroscopicity of Aerosols
Experimental and modeling studies of hygroscopic aerosol growth and deposition in
the lung were extensively discussed in Section 10.4.3.1 of the 1996 PM AQCD (U.S. EPA,
1996, 079380). Hygroscopic ambient aerosols include sulfates, nitrates, some organics, and
aerosols laden with sodium or potassium. The high relative humidity in the lungs
contributes to rapid growth of hygroscopic particles and dramatically alters the deposition
characteristics of ambient hygroscopic aerosols relative to nonhygroscopic aerosols.
Nonhygroscopic particles in the range of 0.3 pm have minimal intrinsic mobility and low
total deposition in the lungs. However, a 0.3 pm salt particle (dry) will grow in vivo to
nearly 2 pm and deposit to a far greater extent (Anselm et al., 1990, 156217). The
hygroscopic growth of particles in the respiratory tract decreases diffusive deposition and
increases aerodynamic deposition as illustrated in Figure 4-6.
4.2.5. Summary
Particle deposition in the respiratory tract occurs predominantly by diffusion,
impaction, and sedimentation. Deposition is minimal for particle diameters in the range of
0.1 to 1.0 pm, where particles are small enough to have minimal sedimentation or
impaction and sufficiently large so as to have minimal diffusive deposition. In humans,
total respiratory tract deposition approaches 100% for particles of roughly 0.01 pm
(physical diameter) due to diffusive deposition and for particles of around 10 pm dae due to
the efficiency of sedimentation and impaction.
The first line of defense for protecting the lower respiratory tract from inhaled
particles is the nose and mouth. Nasal deposition approaches 100% in humans for 10 pm dae
July 2009
4-22
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
particles. Experimental studies show lower nasal particle deposition in children than
adults. Relative to adults, children also tend to breathe more through their mouth which is
less efficient for removing inhaled particles than the nose. These findings suggest that the
lower respiratory tract of children may receive a higher dose of ambient PM compared to
adults. Since children breathe at higher minute ventilations relative to their lung volumes,
the rate of particle deposition normalized to lung surface area may be further increased
relative to adults.
People with COPD generally have greater total deposition and more heterogeneous
deposition patterns compared to healthy individuals. The observed increase in deposition
correlates with increases in airway resistance, suggesting that deposition increases with
the severity of airway disease. COPD patients also have an increased resting minute
ventilation relative to the healthy adults. This demonstrates the need to consider dose rates
(which depend on minute ventilation) rather than just deposition fractions when evaluating
the effect of respiratory disease on particle deposition and dose.
Modeling studies indicate that, for particles greater than -0.01 pm, some cells located
near the carinal ridge of bronchial bifurcations may receive hundreds to thousands times
the average dose (particles per unit surface area) of the parent and daughter airways. The
inertial impaction of particles > 1 pm dae at the carinal ridge of large bronchi increases with
increasing inspiratory flows. Airway constriction can further augment the overall
deposition efficiency of coarse particles at downstream bifurcations. These findings suggest
that substantial doses of particles (> 1 pm dae) may be justified for in vitro studies using
tracheobronchial epithelial cell cultures.
Our ability to extrapolate between species has not generally changed since the 2004
PM AQCD (U.S. EPA, 2004, 056905). However, some considerations related to coarse
particles warrant comment. The inhalability of particles having dae of 2.5, 5, and 10 pm is
80, 65, and 44% in rats, respectively, whereas it remains near 100% for a dae of 10 pm in
humans. In most laboratory animal species (rat, mouse, hamster, guinea pig, and dogs),
deposition in the extrathoracic region is near 100% for particles greater than 5 pm d ae. By
contrast, in humans nasal deposition approaches 100% for 10 pm dae. Oronasal breathing
versus obligate nasal breathing further contributes to greater penetration of coarse
particles into the lower respiratory tract of humans than rodents.
4.3. Clearance of Poorly Soluble Particles
This section discusses the clearance and translocation of poorly soluble particles that
have deposited in the respiratory tract. The term "clearance" is used here to refer to the
processes by which deposited particles are removed by mucociliary action or phagocytosis
from the respiratory tract. "Translocation" is used here mainly to refer to the movement of
free particles across cell membranes and to extrapulmonary sites. In the literature,
July 2009
4-23
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
translocation may also refer to the extra- and intracellular dissolution of particles and the
subsequent transfer of dissociated material to the blood through extra- and intracellular
fluids and across the various cell membranes and lung tissues. The clearance and
distribution of soluble particles and soluble constituents of particles are discussed in
Section 4.4.
Abasic overview of biological mechanisms and clearance pathways from various
regions of the respiratory tract are presented in the following sections. Then regional
kinetics of particle clearance are addressed. Subsequently an update on interspecies
patterns and rates of particle clearance is provided. The translocation of ultrafine particles
is also discussed. Finally information on biological factors that may modulate clearance is
presented.
4.3.1. Clearance Mechanisms and Kinetics
For any given particle size, the deposition pattern of poorly soluble particles
influences clearance by partitioning deposited material between lung regions.
Tracheobronchial clearance of poorly soluble particles in humans, with some exceptions, is
thought (in general) to be complete within 24 to 48 hours through the action of the
mucociliary escalator. Clearance of poorly soluble particles from the alveolar region is a
much slower process which may continue from months to years.
4.3.1.1.	Extrathoracic Region
Particles deposited in either the nasal or oral passages are cleared by several
mechanisms. Particles depositing in the mouth may generally be assumed to be swallowed
or removed by expectoration. Particles deposited in the posterior portions of the nasal
passages are moved via mucociliary transport towards the nasopharynx and swallowed.
Mucus flow in the most anterior portion of the nasal passages is forward, toward the
vestibular region where removal occurs by sneezing, wiping, or nose blowing.
4.3.1.2.	Tracheobronchial Region
Mucociliary clearance in the TB region has generally been considered to be a rapid
process that is relatively complete by 24-48 hours post-inhalation in humans. Mucociliary
clearance is frequently modeled as a series of "escalators" moving material proximally from
one generation to the next. As such, the removal rate of particles from an airway generation
increases with increasing tracheal mucus velocity. Assuming continuity in the amount of
mucus between airway generations, mucus velocities decrease and transit times within an
airway generation increase with distal progression. Although clearance from the TB region
is generally rapid, experimental evidence discussed in the 1996 and 2004 PM AQCD
July 2009
4-24
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
(U.S. EPA, 1996, 079380; U.S. EPA, 2004, 056905), showed that a fraction of material
deposited in the TB region is retained much longer.
The slowcleared TB fraction (i.e., the fraction of particles deposited in the TB region
that are subject to slow clearance) was thought to increase with decreasing particle size.
For instance, Roth et al. (1993, 156928) showed approximately 93% retention of ultrafine
particles (30 nm median diameter) thought to be deposited in the TB region at 24-hpost
inhalation. The slow phase clearance of these ultrafine particles continued with an
estimated half-time (ti/2) of around 40 days. Using a technique to target inhaled particles
(monodisperse 4.2 um MMAD) to the conducting airways, Moller et al. (2004, 155987)
observed that 49 ± 9% of particles cleared rapidly (ti/2 of 3.0 ± 1.6 h), whereas the remaining
fraction cleared considerably slower (ti/2 of 109 ± 78 days). The ICRP (1994, 006988) human
respiratory tract model assumes particles < 2.5 um (physical diameter) to have a slow-
cleared TB fraction of 50%. The slow-cleared fraction assumed by the ICRP (1994, 006988)
decreases with increasing particle size to <1% for 9 um particles. Considering the ultrafine
data of Roth et al. (1993, 156928) in addition to data considered by the ICRP (1994,
006988). Bailey et al. (1995, 190057) estimated a slow-cleared TB fraction of 75% for
ultrafine particles. At that time, they (Bailey et al. 1995) also estimated the slow-cleared
fraction to decrease with increasing particle size to 0% for particles > 6 um. Recent
experimental evidence from the same group (Smith et al., 2008, 190037) showed no
difference in TB clearance among humans for particles with geometric sizes of 1.2 vs. 5 um,
but the same dae (5 |um) so as to deposit similarly in the TB airways. For at least micron-
sized particles, these recent findings do not support the particle size dependence of a slow-
cleared TB fraction. As discussed further below, much of the apparent slow-cleared TB
fraction may be accounted for by differences in deposition patterns, i.e., greater deposition
in the alveolar region than expected.
A portion of the slow cleared fraction from the TB region appears to be associated with
small bronchioles. For large particles (dae = 6.2 pm) inhaled at very slow rate to
theoretically deposit mainly in small ciliated airways, 50% had cleared by 24-h post-
inhalation. Of the remaining particles, 20% cleared with a ti/2 of 2.0 days and 80% with a
11/2 of 50 days (Falk et al., 1997, 086080). Using the same techniques, Svartengren et al.
(2005, 157034) also reported the existence of long-term clearance in humans from the small
airways. It should be noted that the clearance rates for the slow-cleared TB fraction still
exceeds the clearance rate of the alveolar region in humans. Kreyling et al. (1999, 039175)
targeted inhaled particle (dae = 2.2 and 2.5 pm) deposition to the TB airways of adult beagle
dogs and subsequently quantified particle retention using scintigraphic and morphometric
analyses. Despite the use of shallow aerosol bolus inhalation to a volumetric lung depth of
less than the anatomic dead space, 2.5 to 25% of inhaled particles deposited in alveoli. At 24
and 96 hours post inhalation, more than 50% of the retained particles were in alveoli.
However, 40% of particles present at 24 and 96 hours were localized to small TB airways of
July 2009
4-25
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
between 0.3 and 1 mm in diameter. Collectively, these studies suggest that although
mucociliary clearance is fast and effective in healthy large airways, it is less effective and
sites of longer retention exist in the smaller TB airways.
The underlying sites and mechanisms of long-term TB retention in the smaller
airways remain largely unknown. Several factors may contribute to the existence or
experimental artifact of slow clearance from the smaller TB airways. Even when inhaled to
very shallow lung volumes, some particles reach the alveolar region (Kreyling et al., 1999,
039175). Therefore, experiments utilizing bolus techniques to target inhaled particle
deposition to the TB airways may have had some deposition in the alveolar region. This
may occur due to variability in path length and the number of generations to the alveoli
(Asgharian et al., 2001, 017025) and/or differences in regional ventilation (Brown and
Bennett, 2004, 190032). Nonetheless, the experimentally measured clearance rates
measured for the slow cleared TB fraction are faster than that of the alveolar region in both
humans and canines. Thus, although experimental artifacts likely occur, they do not
discount the existence of a slow cleared TB fraction. To some extent, it is possible that the
slow cleared TB fraction may be due to bronchioles that do not have a continuous ciliated
epithelium as in the larger bronchi. Neither path length, ventilation distribution, nor a
discontinuous ciliated epithelium explains an apparently slow cleared TB fraction with
decreasing particle size below 0.1 pm. As discussed in Section 4.3.3 on Particle
Translocation, ultrafine particles cross cell membranes by mechanisms different from larger
(~1 pm) particles. Based on that body of literature, particles smaller than a micron may
enter epithelial cells resulting in their prolonged retention, particularly in the bronchioles
where the residence time is longer and distances necessary to reach the epithelium are
shorter compared to that in the bronchi.
4.3.1.3. Alveolar Region
The primary alveolar clearance mechanism is macrophage phagocytosis and
migration to terminal bronchioles where the cells are cleared by the mucociliary escalator.
Alveolar macrophages originate from bone marrow, circulate briefly as monocytes in the
blood, and then become pulmonary interstitial macrophages before migrating to the
luminal surfaces. Under normal conditions, a small fraction of ingested particles may also
be cleared through the lymphatic system. This may occur by transepithelial migration of
alveolar macrophage following particle ingestion or free particle translocation with
subsequent uptake by interstitial macrophages. Snipes et al. (1997, 156092) have also
demonstrated the importance of neutrophil phagocytosis in clearance of particles from the
alveolar region. Rates of alveolar clearance of poorly soluble particles vary between species
and are briefly discussed in Section 4.3.2. The translocation of particles from their site of
deposition is discussed in Section 4.3.3.
July 2009
4-26
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
The efficiency of macrophage phagocytosis is thought to be greatest for particles
between 1.5 and 3 pm (Oberdorster, 1988, 006857). The decreased efficiency of alveolar
macrophage for engulfing ultrafines increases the time available for these particles to be
taken up by epithelial cells and moved into the interstitium (Ferin et al., 1992, 011101).
Consistent with this supposition (i.e., translocation increases with time), an increase in
titanium dioxide (TiCh) particle transport to lymph nodes has been reported following
inhalation of a cytotoxin to macrophages (Greenspan et al., 1988, 045031). Interestingly, the
long-term clearance kinetics of the poorly soluble ultrafine (15-20 nm CMD) iridium (Ir)
particles were found to be similar to the kinetics reported in the literature for micrometer-
sized particles (Semmler et al., 2004, 055641; Semmler-Behnke et al., 2007, 156080).
Semmler-Behnke et al. (2007, 156080) concluded Sthat ultrafine Ir particles are less
phagocytized by alveolar macrophage than larger particles, but are effectively removed
from the airway surface into the interstitium. Particles are then engulfed by interstitial
macrophages which then migrate to the airway lumen and are removed by mucociliary
clearance to the larynx. The major role of macrophage-mediated clearance was supported
by lavage of relatively few free particles versus predominantly phagocytized particles at
time-points of up to 6-months. It is also possible that some free particles as well as particle-
laden macrophage were carried from interstitial sites via the lymph flow to bronchial and
bronchiolar sites, including bronchial-associated lymphatic tissue, where they were
excreted again into the airway lumen.
4.3.2. Interspecies Patterns of Clearance and Retention
There are differences between species in both the rates of particle clearance from the
lung and manner in which particles are retained in the lung. For instance, based of models
of mucociliary clearance from un-diseased airways, >95% of particles deposited in the
tracheobronchial airways of rats are predicted to be cleared by 5-h post deposition, whereas
it takes nearly 40 hours for comparable clearance in humans (Hofmann and Asgharian,
2003, 055579). As noted in Section 4.3.1.2, however, there is considerable evidence that a
sizeable fraction of particles deposited at the bronchiolar level of the ciliated airways in
humans (as well as canines) are cleared at a far slower rate. The slow cleared TB fraction
appears to increase with decreasing particle size. In contrast to humans and canines,
studies of mice and rats show negligible long-term retention of even ultrafine particles in
the ciliated airways .
July 2009
4-27
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Human
Guinea Pig
Hamster
.00
Dog
0.10
0
100
200
300
400
Days after inhalation
Source: Adapted from Kreyling and Scheuch (2000,056281).
Figure 4-7. Retention of poorly soluble particles (0.5-5 /jm) in the alveolar region of the lung over
various mammalian species.
Figure 4-7 illustrates rates of alveolar clearance for 0.5-5 um particles in various
mammalian species. The alveolar clearance rate of particles smaller than 0.1 um and larger
than 5 um is slower than that of particles in the 0.5 to 5 um range. From interspecies
comparisons of alveolar clearance, the path length from alveoli to ciliated terminal
bronchioles may affect the particle transport rate (Kreyling and Scheuch, 2000, 056281).
The path length from alveoli to ciliated terminal bronchioli is longer in humans, monkeys,
and dogs, than in sheep, rats, hamsters, and mice. Transport time and hence retention
times may increase with path length. This hypothesis fits with all species in this
comparison, except guinea pigs, which have a short path length yet particle retention that
is nearly as long as in humans, monkeys, and dogs. However, sheep have a short path
length and particle transport as fast as rodents. In general, alveolar clearance rates appear
to increase with increasing path length from the alveoli to ciliated airways.
There are also distinct differences in the sites of particle retention between species.
Large mammals retain particles in interstitial tissues under normal conditions, whereas
rats retain particles in alveolar macrophages (Snipes, 1996, 076041). In rats, with chronic
high doses there is a shift in the pattern of dust accumulation and response from that
observed at lower doses in the lungs (Snipes, 1996, 076041; Vincent and Donaldson, 1990,
002162). Rats chronically exposed to high concentrations of insoluble particles experience a
reduction in their alveolar clearance rates and an accumulation of interstitial particle
burden (Bermudez et al., 2002, 055578; Bermudez et al., 2004, 056707; Ferin et al., 1992,
044401; Oberdorster et al., 1994, 046203; Oberdorster et al., 1994, 056285; Warheit et al.,
1997, 086055). The influence of exposure concentration on the pattern of particle retention
in rats (exposed to diesel soot) and humans (exposed to coal dust) was examined by Nikula
et al. (2001, 016641). In rats, the DE particles were found to be primarily in the lumens of
July 2009
4-28
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
the alveolar duct and alveoli! whereas in humans, retained dust was found primarily in the
interstitial tissue within the respiratory acini.
4.3.3. Particle Translocation
Mucociliary and macrophage mediated clearance of poorly soluble particles from the
respiratory tract was discussed in Section 4.3.1. There is evidence that particles may cross
cell membranes and move from their site of deposition by other mechanisms. The following
subsections discuss the movement of particles from the luminal surfaces of the alveolar
region and from the olfactory mucosa. The clearance and distribution of soluble particles
and soluble constituents of particles are discussed in Section 4.4.
4.3.3.1. Alveolar Region
Numerous studies have examined the translocation of ultrafine particles from their
site of deposition in the lung. Traditionally viewed as a relatively inert particle type,
ultrafine TiC>2 has received the most study. At the time the 2004 PM AQCD was released,
there were conflicting results regarding the rate and magnitude of ultrafine carbon
translocation from the human lung. Since that time, it has become well-established that the
transport of ultrafine carbon particles from the human lung is far slower than that of
soluble materials. However, it has also been shown in animal studies (primarily of rats)
that ultrafine particles cross cell membranes by mechanisms different from larger (~1 pm)
particles and that a small fraction of these particles enter capillaries and may distribute
systemically. Described in brief below, details of selected new studies investigating the
disposition of poorly soluble particles are provided in Annex B.
There has been some contention regarding ability of ultrafine carbon particles to
rapidly diffuse from the lungs into the systemic circulation. Based on their study of 5
healthy volunteers, Nemmar et al. (2002, 02 1!) 11) suggested that ultrafine carbon particles
(<100 nm) pass rapidly into the systemic circulation. However, Brown et al. (2002, 043216)
found that the majority of ultrafine carbon particles (CMD, 33 ± 2 nm) were still in the
lungs of healthy human adult volunteers (n = 9; aged 40 to 67 years) and COPD patients
(n = 10; 45 to 70 years) at 24-h post inhalation. Brown et al. (2002, 043216) and Burch
(2002, 056754) contended that the findings reported by Nemmar et al. (2002, 021!) 11) were
consistent with soluble pertechnetate clearance, but not insoluble ultrafine carbon
particles. Highly soluble in normal saline, pertechnetate clears rapidly from the lung with a
half-time of ~10 mins and accumulates most notably in the bladder, stomach, thyroid, and
salivary glands. Three recent studies have confirmed that the majority (>95%) of ultrafine
carbon particles deposited in the lungs of human volunteers are retained at 24-h post
inhalation (Mills et al., 2006, 088770; Wiebert et al., 2006, 157146; Wiebert et al., 2006,
156154). Wiebert et al. (2006, 157146) modified their aerosol generation system to reduce
July 2009
4-29
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
leaching of the 99mTc radiolabel from carbon particles. Except for a small amount of
radiotracer leaching from particles (1.0 ± 0.6% of initially deposited activity in urine by 24
h), these investigators found negligible radiolabel and associated particle clearance from
the lungs by 70 h. The available data show that there is not a rapid or significant amount of
ultrafine carbon particle migration into circulation (Mills et al., 2006, 088770; Moller et al.,
2008, 156771; Wiebert et al., 2006, 157146; Wiebert et al., 2006, 156154; Brown et al., 2002,
043216; Burch, 2002, 056754).
Although human studies show that the vast majority of ultrafine carbon particles are
retained in the lungs until at least 24-h post inhalation, both in vitro and in vivo studies
support the rapid [< 1 h] translocation of free ultrafine Ti02 particles across pulmonary cell
membranes (Churg et al., 1998, 085815; Ferin et al., 1992, 044401; Geiser et al., 2005,
087362). Peculiar to TiCh aerosols, there is evidence that particle aggregates may
disassociate once deposited in the lungs. This disassociation makes inhaled aggregate size
the determinant of deposition amount and site, but primary particle size the determinant of
subsequent clearance (Bermudez et al., 2002, 055578; Ferin et al., 1992, 044401; Takenaka
et al., 1986, 046210). Following disaggregation, the ultrafine TiC>2 particles are cleared
more slowly and cause a greater inflammatory response (neutrophil influx) than fine TiC>2
particles (Bermudez et al., 2002, 055578; Ferin et al., 1992, 044401; Landry et al., 1983,
003904; Oberdorster et al., 1994, 046203; Oberdorster et al., 1994, 056285; Putz-Anderson
et al., 1981, 003914). The differences in inflammatory effects and possibly lymph burdens
between fine and ultrafine TiC>2 in many studies appear related to lung burden in terms of
particle surface area and not particle mass or number (Oberdorster, 1996, 039852;
Oberdorster et al., 1992, 045110; Oberdorster et al., 2000, 039014; Tran et al., 2000,
013071). More recently, others have noted that particle surface area is not an appropriate
metric across all particle types (Warheit et al., 2006, 088436). Surface characteristics such
as roughness can also affect protein binding and potentially clearance kinetics, with
smoother Ti02 surfaces being more hydrophobic (Sousa et al., 2004, 089866).
Geiser et al. (2005, 087362) conducted a detailed examination of the disposition of
inhaled ultrafine Ti02 in 20 healthy adult rats. They found that distributions of particles
among lung tissue compartments appeared to follow the volume fraction of the tissues and
did not significantly differ between 1- and 24-h post-inhalation. Averaging 1- and 24-h data,
79.3 ± 7.6% of particles were on the luminal side of the airway surfaces, 4.6 ± 2.6% were in
epithelial or endothelial cells, 4.8 ± 4.5% were in connective tissues, and 11.3 ± 3.9% were
within capillaries. Particles within cells were not membrane bound. It is not clear why the
fraction of particles identified in compartments such as the capillaries did not differ
between 1- and 24-h post-inhalation. These findings were consistent with the smaller study
of 5 rats by Kapp et al. (2004, 156624) who reported identifying Ti02 aggregates in a type II
pneumocyte! a capillary close to the endothelial cells! and within the surface-lining layer
close to the alveolar epithelium immediately following a 1-h exposure. These studies
July 2009
4-30
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
effectively demonstrate that some inhaled ultrafine Ti02 particles once deposited on the
pulmonary surfaces can rapidly [< 1 h] translocate beyond the epithelium and potentially
into the vasculature.
Extrapulmonary translocation has also been described for poorly soluble ultrafine
gold and Ir particles. In male Wistar-Kyoto rats exposed to ultrafine gold particles (5-8 nm),
Takenaka et al. (2006, 156110) reported a low but significant fraction (0.03 to 0.06% of lung
concentration) of gold in the blood from 1 to 7 days post inhalation. Semmler et al. (2004,
055641) also found small but detectable amounts of poorly soluble Ir particle (15 and 20 nm
CMD) translocation from the lungs of male Wistar-Kyoto rats to secondary target organs
like the liver, spleen, brain, and kidneys. Each of these organs contained about 0.2% of
deposited Ir. The peak levels in these organs were found 7 days post inhalation. The
translocated particles were largely cleared from extrapulmonary organs by 20 days and Ir
levels were near background at 60 days post inhalation. Particles may have been
distributed systemically via the gastrointestinal tract. Immediately after the 6-h inhalation
exposure, 18 ± 5% of the deposited Ir particles had already cleared into the gastrointestinal
tract. After 3 weeks, 31 ± 5% of the deposited particles were retained in the lung. By 2 and
6 months post inhalation, lung retention was 17 ± 3 and 7 ± 1%, respectively. The particles
appeared to be cleared predominantly from the peripheral lung via the mucociliary
escalator into the GI tract and were found in feces.
A few recent studies have characterized differences in the behavior of fine and
ultrafine particles in vitro. Geiser et al. (2005, 087362) found that both ultrafine and fine
(0.025 |um gold, 0.078 |um TiC>2, and 0.2 jum TiCh) particles cross cellular membranes by non-
endocytic (i.e., involving vesicle formation) mechanisms such as adhesive interactions and
diffusion, whereas the phagocytosis of larger 1 urn TiC>2 particles is ligand-receptor
mediated. Edetsberger et al. (2005, 155759) found that ultrafine particles (0.020 pm
polystyrene) translocated into cells by first measurement (~1 min after particle
application). Intracellular agglomerates of 88-117 nm were seen by 15-20 mins and of 253-
675 nm by 50-60 mins after particle application. These intracellular aggregates were
thought to result from particle incorporation into endosomes or similar structures since
Genistein or Cytochalasin treatment generally blocked aggregate formation. Interestingly,
particles did not translocate into dead cells, rather they attached to the outside of the cell
membrane. Amine- or carboxyl-modified surfaces (46 nm polystyrene) did not affect
translocation across cultures of human bronchial epithelial cells with about 6% regardless
of the surface characteristics (Geys et al., 2006, 155789).
4.3.3.2. Olfactory Region
Numerous studies have demonstrated the translocation of soluble and poorly soluble
particles from the olfactory mucosa via the axons to the olfactory bulb of the brain. The vast
majority of these studies were conducted in rodents. However, DeLorenzo (DeLorenzo, 1970,
July 2009
4-31
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
156391) observed the rapid (within 30-60 mins) movement of 50 nm silver-coated colloidal
gold particles instilled on the olfactory mucosa into the olfactory bulb of squirrel monkeys.
The specifics of this and other key studies that have investigated the translocation of
particles to the olfactory bulb are provided in Annex B.
Two recent studies reported the movement of ultrafine particles deposited in the
olfactory region of the nose along the olfactory nerve and into the olfactory bulb of the brain
in rats. Oberdorster et al. (2004, 055639) exposed rats to ultrafine carbon particles (36 nm
CMD, 1.7 og) containing 13C in a whole-body chamber for 6 h. The distribution of 13C was
followed for 7 days postexposure. There was a significant increase in 13C in the olfactory
bulb on Day 1 with persistent and continued increased through Day 7. Elder et al. (2006,
089253) exposed rats to manganese (Mn) oxide (~30 nm equivalent sphere with 3-8 nm
primary particles) via whole-body inhalation exposure for 12 days (6 h/day, 5 days/wk) with
both nares open or Mn oxide for 2 days (6 h/day) with right nostril blocked. After the 12
days exposure via both nostrils, Mn in the olfactory bulb increased 3.5-fold. After the 2-day
exposure with the right nostril blocked, Mn was found mainly in the left olfactory bulb (2.4-
fold increase). These studies suggest the neuronal uptake and translocation of ultrafine
particles without particle dissolution and in the absence of mucosal injury.
Elder et al. (2006, 089253) also addressed the issue of whether solubilization of
particles was requisite for translocation along the olfactory nerve and into the brain.
Similar amounts of soluble manganese chloride (MnCk) and poorly soluble Mn oxide were
instilled onto the left naris of anesthetized rats. At 24-h post instillation, similar amounts of
Mn were found in the left olfactory bulb of rats instilled with MnCk (8.2 ± 3.6% of instilled)
and Mn oxide (8.2 ± 0.7% of instilled). If solubilization were required for translocation, then
a lower amount of Mn oxide than MnCb should have reached the olfactory bulb. Following
14 consecutive days of aerosol exposure, Dorman et al. (2001, 055433) demonstrated that
more soluble Mn sulfate reaches the olfactory bulb and striatum of rat brains than the
poorly soluble form of Mn tetroxide. Nonetheless, the Mn levels were statistically increased
in both the olfactory bulb and striatum following exposure to Mn tetroxide relative to filter
air. In a subsequent 13-wk exposure study, Dorman et al. (2004, 155752) also demonstrated
that more soluble manganese sulfate (MnSCh) reached the olfactory bulb than was observed
for the less soluble Mn form (hureaulite). Both the soluble and less soluble forms of Mn
resulted in statistically increased levels of Mn in the olfactory bulb relative to air exposed
controls. The soluble MnSCk was also observed to reach the striatum and cerebellum. In
addition, Yu et al. (2003, 156171) demonstrated increased Mn levels in the brains of rats
exposed to welding-fumes for 60 days, however, the role of transport via the blood is less
clear in this study.
The translocation of zinc (Zn) and TiCh to the olfactory bulb has also been reported in
the literature. Persson et al. (2003, 051846) observed the translocation of Zn to the olfactory
bulbs following instillation in both rats and freshwater pike. Wang et al. (2007, 156146)
July 2009
4-32
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
reported the translocation of both fine (155 nm) and ultrafine (21 and 71 nm) TiCh particles
in mice. Interestingly, a qualitative analysis of the data showed that more of the fine TiCh
than ultrafine Ti02 reached the olfactory bulb. Wang et al. (2007, 156146) suggested that a
strong hydrophilic character and propensity for aggregation reduced the translocation of
the ultrafine Ti02.
The importance of particle translocation to the brain is not yet understood.
Translocation via the axon to the olfactory bulb has been observed for numerous compounds
of varying composition, particle size, and solubility. Although the rate of translocation is
rapid, perhaps less than an hour, the magnitude of transport remains poorly characterized.
With regard to the magnitude of transport, Elder et al. (2006, 089253) found that as much
as 8% of both soluble and insoluble forms of Mn were translocated to the olfactory bulb in
rats following intranasal instillation. It is also still unclear to what extent translocation to
the olfactory bulb and other brain regions may vary between species. The olfactory mucosa
covers approximately 50% of the nasal epithelium in rodents versus only about 5% in
primates (Aschner et al., 2005, 155663). Additionally, a greater portion of inhaled air passes
through the olfactory region of rats relative to primates (Kimbell, 2006, 155902). These
differences may predispose rats, more so than humans, to deposition of particles in the
olfactory region with subsequent particle translocation to the olfactory bulb.
4.3.4. Factors Modulating Clearance
4.3.4.1. Age
It was previously concluded that there appeared to be no clear evidence for any age-
related differences in clearance from the lung or total respiratory tract, either from child to
adult, or young adult to elderly (U.S. EPA, 2004, 056905! U.S. EPA, 1996, 079380). Studies
showed either no change or some slowing in mucus clearance with age after maturity.
Although some differences in alveolar macrophage function were reported between mature
and senescent mice, no age-related decline in macrophage function had been observed in
humans. A comprehensive review of the recent and older literature supports a decrease in
mucociliary clearance with increasing age beyond adulthood in humans and animals.
Limited animal data also suggest macrophage-mediated alveolar clearance may also
decrease with age.
Studies addressing the effects of age on respiratory tract clearance are provided in
Annex B. Ho et al. (2001, 156549) demonstrated that nasal mucociliary clearance rates
were about 40% lower in old (age >40-90) versus young (age 11-40) men and women.
Tracheal mucus velocities in elderly (or aged) humans and beagle dogs are about 50% that
of young adults (Goodman et al., 1978, 071130; Whaley et al., 1987, 156153). Several
human studies have demonstrated decreasing rates of mucociliary particle clearance from
July 2009
4-33
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
the large and small bronchial airways with increasing age (Puchelle et al., 1979, 006863;
Svartengren et al., 2005, 157034; Vastag et al., 1985, 157088). Linear fits to the data show
that rapid clearance (within 1 h) from large bronchi and prolonged clearance (between 1-21
days) from the small bronchioles in an 80-year-old is only about 50% of that in 20-year-old
(Svartengren et al., 2005, 157034; Vastag et al., 1985, 157088). One study reported that
alveolar particle clearance rates decreased by nearly 40% in old versus young rats (Muhle
et al., 1990, 006853). Another study has reported that older rats have an increased
susceptibility to pulmonary infection due to altered alveolar macrophage function and
slowed bacterial clearance (Antonini et al., 2001, 156219). Although data are somewhat
limited, they consistently show a depression of clearance throughout the respiratory tract
with increasing age from young adulthood in humans and laboratory animals.
4.3.4.2.	Gender
Gender was not found to affect clearance rates in prior reviews (U.S. EPA, 1996,
079380; U.S. EPA, 2004, 056905). Studies not included in those reviews also show that
human males and females have similar nasal mucus clearance rates (Ho et al., 2001,
156549), tracheal mucus velocities (Yeates et al., 1981, 095391), and large bronchial airway
clearance rates (Vastag et al., 1985, 157088).
4.3.4.3.	Respiratory T ract Disease
At the time of the last two reviews (U.S. EPA, 1996, 079380; U.S. EPA, 2004, 056905),
it was well recognized that obstructive airways disease may influence both the site of initial
deposition and the rate of mucociliary clearance from the airways. When deposition
patterns are matched, mucociliary clearance rates are reduced in patients with COPD
relative to healthy controls. The effects of acute bacterial/viral infections and cough on
mucociliary clearance were briefly summarized in Section 10.4.2.5 (EPA, 1996, 079380) and
Section 6.3.4.4 (EPA, 2004, 056905) of past reviews. While cough is generally a reaction to
some inhaled stimulus, in some cases, especially respiratory disease, it can also serve to
clear the upper bronchial airways of deposited substances by dislodging mucus from the
airway surface. One of the difficulties in assessing effects on infection on mucociliary
clearance is that spontaneous coughing increases during acute infections. Cough has been
shown to supplement mucociliary clearance of secretions, especially in patients with
obstructive lung disease and primary ciliary dyskinesia.
Using a bolus technique to target specific lung regions, Moller et al. (2008, 156771)
examined particle clearance from the ciliated airways and alveolar region of healthy
subjects, smokers, and patients with COPD. Airway retention after 1.5 hours was
significantly lower in healthy subjects (89 ± 6%) than smokers (97 ± 3%) or COPD patients
(96 ± 6%). At 24 and 48 h, retention remained significantly higher in COPD patients
(86 ± 6% and 82 ± 6%, respectively) than healthy subjects (75 ± 10% and 70 ± 9%,
July 2009
4-34
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
respectively). However, these findings are confounded by the more central pattern of
deposition in the healthy subjects than in the smokers and COPD patients. Alveolar
retention of particles was similar between the groups at 48-h post-inhalation.
The effect of asthma on lung clearance of particles may depend on disease status. Lay
et al. (2009, 190060) found significantly (p < 0.01) more rapid particle (0.22 urn) mucociliary
clearance over a 2-h period post inhalation in mild asthmatics than in healthy volunteers.
Although the pattern of deposition tended to be more central in the asthmatics, there was
not a statistically significant difference from healthy controls. In vivo uptake by airway
macrophages in mild asthmatics was also enhanced relative to healthy volunteers (p <
0.01). In an ex vivo study, airway macrophages from individuals with more severe asthma
had impaired phagocytic capacity relative to less severely affect asthmatics and healthy
volunteers (Alexis et al., 2001, 190013). Lay et al. (2009, 190060) concluded that enhanced
uptake and processing of particulate antigens could contribute to the pathogenesis and
progression of allergic airways disease in asthmatics and may contribute to an increased
risk of exacerbations with particulate exposure.
Chen et al. (2006, 147267) investigated the effect of endotoxin on the disposition of
particles. Healthy rats and those pretreated with endotoxin (12 hours before particle
instillation) were instilled with ultrafine (56.4 nm) or fine (202 nm) particles. In healthy
rats, there were no marked differences in lung retention or systemic distribution between
the ultrafine and fine particles. In healthy animals, ultrafine particles were primarily
retained in lungs (72 ± 10% at 0.5-2 h; 65 ± 1% at 1 day! 62 ± 5% at 5 days). Particles were
also detected in the blood (2 ± 1% at 0.5-2 h; 0.1 ± 0.1% at 5 days) and liver (3 ± 2% at 0.5-2
h; 1 ± 0.1% at 5 days) of the healthy animals. At 1 day post-instillation, about 13% of the
particles were excreted in the urine or feces of the healthy animals. In rats pretreated with
endotoxin, by 2-h post-instillation, the ultrafine particles accessed the blood (5 vs. 2%) and
liver (11 vs. 4%) to a significantly greater extent than fine particles. The endotoxin-treated
rats also had significantly greater amounts of ultrafine particles in the blood (5% vs. 2%)
and liver (11% vs. 3%) relative to the healthy control rats. This study demonstrates that
acute pulmonary inflammation caused by endotoxin increases the migration of ultrafine
particles into systemic circulation.
Adamson and Preditis (1995, 189982) investigated the possibility that particle
deposition into an already injured lung might affect particle retention and enhance the
toxicity of "inert" particles. Bleomycin was instilled into mice to induce epithelial necrosis
and subsequent pulmonary fibrosis. Instilled 3 days following bleomycin treatment, while
epithelial permeability was compromised, carbon black particles in treated mice were
translocated to the interstitium and showed increased pulmonary retention relative to
untreated mice. When instilled 4 weeks post bleomycin treatment, after epithelial integrity
was restored, carbon black particle retention was similar between treated and untreated
mice with minimal translocation to the interstitium. The instillation of carbon particles did
July 2009
4-35
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
not appear to increase lung injury in the bleomycin treated mice at either time point. This
study shows that integrity of the epithelium affects particle retention and translocation into
interstitial tissues.
4.3.4.4. Particle Overload
Unlike other laboratory animals, rats appear susceptible to "particle overload" effects
due to impaired macrophage-mediated alveolar clearance. Numerous reviews have
discussed this phenomenon and the difficulties it poses for the extrapolation of chronic
effects in rats to humans (Miller, 2000, 011822; ILSI, 2000, 002892; Oberdorster, 1995,
046596; Oberdorster, 2002, 021111; Morrow, 1994, 006850). Large mammals have slow
pulmonary particle clearance and retain particles in interstitial tissues under normal
conditions, whereas rats have rapid pulmonary clearance and retain particles in alveolar
macrophages (Snipes, 1996, 076041). With chronic high doses of PM there is a shift in the
pattern of dust accumulation and response from that observed at lower doses in rat lungs
(Snipes, 1996, 076041; Vincent and Donaldson, 1990, 002462). Rats chronically exposed to
high concentrations of insoluble particles experience a reduction in their alveolar clearance
rates and an accumulation of interstitial particle burden (Bermudez et al., 2002, 055578;
Bermudez et al., 2004, 056707; Ferin et al., 1992, 044401; Oberdorster et al., 1994, 046203;
Oberdorster et al., 1994, 056285; Warheit et al., 1997, 086055). With continued exposure,
some rats eventually develop pulmonary fibrosis and both benign and malignant tumors
(Lee et al., 1985, 003847; Lee et al., 1985, 067628; Lee et al., 1986, 067629; Warheit et al.,
1997, 086055). Oberdorster (1996, 039852; 2002, 021111) proposed that high-dose effects
observed in rats may be associated with two thresholds. The first threshold is the
pulmonary dose that results in a reduction in macrophage-mediated clearance. The second
threshold, occurring at a higher dose than the first, is the dose at which antioxidant
defenses are overwhelmed and pulmonary tumors develop. Intrapulmonary tumors
following Ti02 exposures are exclusive to rats and are not found in mice or hamsters
(Mauderly, 1997, 084631). Moreover, Lee et al. (1985, 003847) noted that the squamous cell
carcinomas observed with prolonged high concentration Ti02 exposures developed from the
alveolar lining cells adjacent to the alveolar ducts, whereas squamous cell carcinomas in
humans are generally linked with cigarette smoking are thought to arise from basal cells of
the bronchial epithelium. Quoting Lee et al. (1986, 067629), "Since the lung tumors were a
unique type of experimentally induced tumor under exaggerated exposure conditions and
have not usually been seen in man or animals, their relevance to man in questionable."
4.3.5. Summary
For any given particle size, the pattern of poorly soluble particle deposition influences
clearance by partitioning deposited material between regions of the respiratory tract.
July 2009
4-36
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Particles depositing in the mouth may generally be assumed to be swallowed or removed by
expectoration. Particles deposited in the posterior portions of the nasal passages or the TB
airways are moved via mucociliary transport towards the nasopharynx and swallowed.
Although clearance from the TB region is generally rapid, there appears to be fraction of
material deposited in the TB region of humans that is retained much longer. The
underlying sites and mechanisms of long-term TB retention are not known. In contrast to
humans, mice and rats appear to have negligible long-term retention of particles in TB
airways. The primary alveolar clearance mechanism is macrophage phagocytosis and
migration to terminal bronchioles where the cells are cleared by the mucociliary escalator.
Clearance from both the TB and alveolar region is more rapid in rodents than humans.
Mucociliary and macrophage-mediated clearance decreases with age beyond adulthood.
Human data show that there is not a rapid or significant amount of ultrafine carbon
particle migration into circulation. However, both in vitro and in vivo animal studies
support the rapid [< 1 h] translocation of free ultrafine TiC>2 particles across pulmonary cell
membranes. Extrapulmonary translocation has also been described in rats for poorly
soluble ultrafine gold and Ir particles. Alow, but statistically significant, fraction (0.03 to
0.06% of lung concentration) of ultrafine gold particles has been observed in the blood of
rats from 1 to 7 days post inhalation. The translocation in detectable amounts (<1% of
deposited material) of poorly soluble Ir particles (15 and 20 nm CMD) from the lungs of rats
to secondary target organs like the liver, spleen, brain, and kidneys has also been reported.
However, the systemic distribution of particles may have occurred via normal clearance
from the lungs to the gastrointestinal tract.
Although the importance of particle translocation to the brain is not yet understood,
translocation from the olfactory mucosa via the axon to the olfactory bulb has been reported
in primates, rodents, and freshwater pike for numerous compounds of varying composition,
particle size, and solubility. The rate of translocation is rapid, perhaps less than an hour. In
rats, as much as 8% of material may become translocated to the olfactory bulb following
intranasal instillation. It is unclear to what extent translocation to the olfactory bulb and
other brain regions may vary between species. Interspecies differences may predispose rats,
more so than humans, to the deposition of particles in the olfactory region with subsequent
translocation to the olfactory bulb.
4.4. Clearance of Soluble Materials
Soluble particles and soluble constituents of particles may be absorbed through the
epithelium and distributed systemically or retained in the lung. The rate of dissolution
depends on a number of factors, including particle surface area and chemical structure.
Some dissolved materials bind to proteins or other components in the airway surface liquid
layer. In the ciliated airways, solutes are cleared by mucociliary transport and diffuse into
July 2009
4-37
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
underlying tissues and the blood. In the alveolar regions, the thin barrier between the air
and blood allows for rapid transport of solutes into the blood. The movement of soluble
materials depends on the site of deposition in the lung, the rate of material dissolution from
particles, and the molecular weight of the solute. The rate of soluble material clearance
from the lungs depends on epithelial permeability which may be affected by age,
respiratory disease, and concurrent exposures. While enhanced clearance of insoluble
particles acts to reduce dose to airway tissue, increased transport of soluble matter into the
blood stream may enhance effects on extra-pulmonary organs.
4.4.1. Clearance Mechanisms and Kinetics
The rate of absorption across the epithelium for materials that dissolve in the airway
or alveolar lining fluid is fairly rapid (minutes to hours) and is a function of their molecular
size and their water or lipid solubility (Oberdorster, 1988, 006857! Enna and Schanker,
1972, 155767; Huchon et al., 1987, 024923; Schanker et al., 1986, 005100). Huchon et al.
(1987, 024923) studied the clearance of a variety of aerosolized solutes from the lungs of
dogs. Solute clearance was inversely related to molecular weight. Negligible clearance of
the largest molecular weight solute (transferrin! mol wt -76,000 daltons) in their study was
found over a 30-min observation period. At the other extreme, free pertechnetate (mol wt
-163 daltons) had a clearance rate of 6% per min. Clearance of hydrophilic solutes is
diffusion limited by pore sizes associated with intercellular tight junctions (estimated at
0.6-1.5 nm). Absorption of lipophilic compounds that pass easily through cell membranes is
perfusion limited and thus generally occurs very rapidly. However, if lipophilic materials
are adsorbed onto insoluble particles their retention in the lung may be prolonged (Creasia
et al., 1976, 059713). In addition to diffusion through intercellular junctions, transcellular
transport of large solutes by pinocytosis in epithelial cells has also been observed (Chinard,
1980, 156341).
A portion of poorly soluble particles may become dissolved with subsequent solute
clearance. More rapid dissolution of poorly soluble nano- or ultrafine particles relative to
micro-sized particles occurs due to an increasing surface-to-volume ratio with decreasing
particle size. Kreyling et al. (2002, 037332) examined the dissolution of poorly soluble
ultrafine Ir particle agglomerates (15-80 nm CMD) composed of 5 nm primary particles.
After 7 days, less than 1% of the particles were dissolved in buffered saline, whereas 6%
dissolved in 1 N hydrochloric acid after 1 day. Thus, the high surface-to-volume ratio of
ultrafine particles should not be misconstrued to imply rapid dissolution of poorly soluble
particles following deposition in the respiratory tract. However, poorly soluble particles that
become phagocytosed may slowly dissolve in the acidic (pH of 4.3-5.3) environment of the
phagolysosome to be released in their solubilized form from the cell and potential move
across the epithelium into the bloodstream. The dissolution rate is inversely related to
July 2009
4-38
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
particle size and directly related to specific surface area (Kreyling and Scheuch, 2000,
056281) and facilitated by the acidic environment of the macrophage (Kreyling, 1992,
067243).
There is considerable evidence as well that soluble particles depositing in the
bronchial airways are also cleared by mucociliary transport (Bennett and Ilowite, 1989,
000835; Wagner and Foster, 2001, 156143; Matsui et al., 1998, 040405; Sakagami et al.,
2002, 156936; Lay et al., 2003, 155920). The relative contribution of their removal by
transepithelial absorption vs. mucociliary clearance is likely a function of both the
molecular size and water or lipid solubility of the material (Oberdorster, 1988, 006857;
Enna and Schanker, 1972, 155767; Huchon et al., 1987, 024923; Sakagami et al., 2002,
156936). Furthermore, the rate of mucociliary transport for soluble particles may be less
than that of insoluble particles (Lay et al., 2003, 155920). Consequently, non-permeating
hydrophilic solutes may remain in contact with the airway epithelium for a longer period
than insoluble particles. This may be due to diffusion of a greater portion of the solute into
the periciliary sol layer which may be transported less efficiently than the mucus layer
during mucociliary clearance. Bronchial blood flow has also been shown to modulate airway
retention of soluble particles (Wagner and Foster, 2001, 156143), i.e., decreasing blood flow
increases airway retention of soluble particles.
As an example of how transport of soluble components of PM may clear the lung by
transepithelial absorption, Wallenborn et al. (2007, 156144) measured elemental content of
lungs, plasma, heart, and liver of healthy male WKY rats (12-15 weeks old) 4 or 24 hours
following a single intratracheal (IT) instillation of saline or 8.33 mg/kg of oil combustion PM
containing a variety of transition metals with differing water and acid solubility. Metals
with high water solubility and relatively high concentration in oil combustion PM were
increased in extrapulmonary organs. Elements with low water or acid solubility, like silicon
and aluminum, were not detected in extrapulmonary tissues despite decreased levels in the
lung suggesting they cleared the lung primarily by mucociliary clearance. Thus, PM-
associated metals deposited in the lung may be released into systemic circulation at
different rates depending on their water/acid solubility.
The amount and type of water soluble or leachable metals associated with PM varies
with location and by source. Furthermore some metals such as Zn, copper, and iron are
essential to body function while others such as vanadium and nickel are nonessential
metals. Consequently the body and the lung have different ways of dealing with excesses in
inhaled soluble metals associated with PM. Bioavailability, and potentially the toxicity, of
leachable metals may be altered by protein binding within the lung and blood as well as the
affinities of these binding sites. For example Zn is tightly regulated by a variety of metal
binding proteins, including metallothionein and a family of Zn specific transporters. In the
plasma, Zn binds to many proteins, including a2-macroglobulin and albumin. Zn is an
example of a common abundant water soluble metal in ambient air that may contribute to
July 2009
4-39
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
increased respiratory and cardiovascular disease risk associated with PM exposure.
Wallenborn et al. (2009, 191172) recently showed that soluble Zn sulfate (in the form of
70Zn, a rare isotope of Zn) introduced into the lungs by instillation not only reaches, but
accumulates in extrapulmonary organs, including the heart and liver, following pulmonary
exposure. However, the retention of greater than 50% of 70Zn in the lung 4 hours post
instillation suggested that the transepithelial absorption of soluble Zn was indeed slowed
by binding to proteins in the lungs. While it could not be ascertained if 70Zn measured in
the heart was replacing endogenous Zn pools, any accumulation in cardiac Zn levels could
lead to mitochondrial dysfunction and ion channel disruption, possibly explaining adverse
cardiac effects from inhalation of Zn-rich PM. Effects of Zn instillation on epithelial
integrity were not evaluated.
4.4.2. Factors Modulating Clearance
A number of studies have evaluated the epithelial permeability by measuring the
clearance of 99mTc-diethylenetriaminepentaacetic acid (99mTc-DTPA), a small hydrophilic
solute (492 daltons, 0.57 nm). These studies are the basis for much of the discussion in this
section.
4.4.2.1.	Age
In humans, the clearance of water-soluble particles (99mTc-DTPA) from the alveolar
epithelium generally slows with increasing age (Pigorini et al., 1988, 156027; Braga et al.,
1996, 156289). However, Tankersley et al. (2003, 096363) recently showed enhanced
permeability of soluble particles (99mTc-DTPA) in terminally senescent mice just before
death, suggesting that a disintegration of the epithelial barrier may be a feature of lung
homeostatic loss during this period of terminal senescence.
4.4.2.2.	Physical Activity
The transepithelial transport rates of soluble particles, 99mTc-DTPA, have also been
found to increase during exercise (Hanel et al., 2003, 155826; Lorino et al., 1989, 155946;
Meignan et al., 1986, 156752). This enhancement was linked to increases in Vt associated
with exercise (Lorino et al., 1989, 155946). Regionally, this effect was dominated by
increased apical lung clearance and attributed to an increase in apical blood flow (Meignan
et al., 1986, 156752). The increased permeability with exercise appears to resolve to
baseline after a short period post exercise, i.e., within a couple hours (Hanel et al., 2003,
155826).
July 2009
4-40
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
4.4.2.3. Disease
Because the integrity of the epithelial surface lining of the lungs may be damaged
from lung disease, particles (either insoluble or soluble) may gain greater access to the
interstitium, lymph, and blood stream. Damage to the epithelial barrier is most likely to
acutely affect transepithelial transport rates of soluble particles. From bronchial biopsies,
Laitinen et al. (1985, 037521) found various degrees of epithelial damage, from loosening of
tight junctions to complete denudation of the airway epithelium, in asthmatics. Consistent
with these findings, Ilowite et al. (1989, 156584) found that asthmatics had increased
permeability of the bronchial mucosa to the hydrophilic solute 99mTc-DTPA. On the other
hand, a more recent study in a sheep model showed that the presence of bronchial edema
could slow the uptake of soluble DTPAinto the blood and enhanced retention in the
airways, likely within the expanded interstitial barrier (Foster and Wagner, 2001, 155778).
Both a leaky epithelial barrier and expanded interstitial barrier associated with asthma
may result in enhanced exposure of submucosal immune and smooth muscle cells to
xenobiotic substances.
Alveolar epithelial permeability was also shown to be affected by the presence of lung
inflammation. The most common finding has been a clear increase in alveolar permeability
induced by cigarette smoking (Jones et al., 1980, 155883). This effect appears to be
dependent the recent cigarette smoke exposure as indexed by carboxyhaemoglobin (Jones et
al., 1983, 155884) and is rapidly reversible within a week of smoking cessation (Mason et
al., 1983, 013169). In fact, Huchon et al. (1984, 156576) demonstrated that COPD patients
who have stopped smoking have normal clearance of 99mTc-DTPA.
In general, increased alveolar permeability to 99mTc-DTPAhas been found to be
associated with any lung syndrome characterized by pulmonary edema. While the trans-
alveolar transport of a small solute like DTPA is very sensitive to even mild acute lung
injury (such as that associated with even mild cigarette smoking), increased transport rates
of larger molecules (>100K daltons) across the alveolar epithelium require more severe
damage like that seen in adult respiratory distress syndrome (ARDS) (Peterson et al., 1989,
024922; Braude et al., 1986, 155701). Interstitial lung disease and pulmonary fibrosis are
also characterized by increased alveolar permeability (Bodolay et al., 2005, 156280;
Antoniou et al., 2006, 156220; Watanabe et al., 2007, 157115). Interestingly, these recent
studies have also shown that the increased permeability in these patients could be
corrected with immunosuppressive/steroid treatments (Bodolay et al., 2005, 156280;
Watanabe et al., 2007, 157115). Furthermore, studies of DTPA clearance in bleomycin
injured dogs, a model of pulmonary fibrosis, suggest that the enhanced permeability is
associated with the initial acute phase of the lung damage, with clearance rates returning
to normal as chronic fibrosis developed over time (Suga et al., 2003, 157021).
Finally, as evidence of lung complications associated with non-insulin dependent
diabetes (type 2) patients, Lin et al. (2002, 155932) found impairment of alveolar integrity
July 2009
4-41
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
as shown by increased transport rates of both hydophilic and lipophilic solutes from the
lungs in these patients. By contrast, a number of other studies have found epithelial
permeability reduced, i.e., slower transport rates, in diabetes (Ozsahin et al., 2006, 156833;
Caner, 1994, 156320; Mousa et al., 2000, 156786) that may be related to disease duration
and metabolic control (Ozsahin et al., 2006, 156833). These findings are consistent with
thickening of alveolar basement membrane detected in autopsies of diabetes patients
(Weynand et al., 1999, 157110).
4.4.2.4. Concurrent Exposures
The integrity of the alveolar epithelium may be disrupted by co-pollutants such that
soluble components of inhaled particles can more easily enter the interstitium and blood
stream. Like active cigarette smoking discussed previously, Beadsmoore et al. (2007,
156259) showed clearance half-times in healthy passive smokers to be shorter compared
with healthy non-smokers but still longer than in healthy smokers. These findings show a
progressive increase in epithelial permeability with exposure to cigarette smoke. Similarly,
acute exposure of humans to 0.4 ppm ozone for 2 hours with intermittent exercise has been
shown to alter epithelial integrity and increase clearance of soluble hydrophilic particles
from the alveolar surfaces of the lung (Kehrl et al., 1987, 040824). This effect persists to at
least 24 hours post-exposure to even low concentrations (0.24 ppm average for 130 minutes)
of ozone (Foster and Stetkiewicz, 1996, 079920). Similarly, 0.8 ppm O3 exposure for 2 hours
in rats shows increased permeability to macromolecules at all levels of the respiratory tract
(Bhalla et al., 1986, 040407) that persisted in the alveolar region beyond 24 hours post-
exposure. Cohen et al. (1997, 009213) may have best illustrated the competing effects of
mucociliary and transepithelial transport by showing that coexposure to ozone affected the
retention of inhaled chromium in rats differently depending on its solubility. In its soluble
potassium chromate form, ozone decreased the retention of chromium, but when chromium
was inhaled as insoluble barium chromate, its retention in the lung was increased by ozone
coexposure. Similarly, a study that showed decreased clearance of insoluble cesium oxide
particles following influenza infection also showed a virus-induced enhancement of
clearance for a soluble cesium chloride (Lundgren et al., 1978, 155950). Chang et al. (2005,
097776) also recently showed that ultrafine carbon black acts through a reactive oxygen
species (ROS) dependent pathway to increase epithelial permeability in mice.
But chronic exposure to other particulate or gaseous pollutants has not always led to
increased epithelial permeability. Studying subjects with a variety of occupational
exposures, Kaya et al. (2006, 156632) showed that nonsmoking welders actually have
decreased epithelial permeability relative to nonsmoking control subjects, and occupational
exposure of painters to isocyanates has no effect on bronchoalveolar epithelial permeability
(Kaya et al., 2003, 156631).
July 2009
4-42
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
4.4.3. Summary
The healthy airway and alveolar epithelium is generally impermeable to very large
insoluble macromolecules and particles. Water and acid soluble particles may more rapidly
move through the epithelium as they dissolve on the airway surface or within the
phagolysomes of macrophages. The presence of airway inflammation in a variety of airway
diseases (e.g., asthma, fibrosis, ARDS, pulmonary edema, inflammation from smoking)
alters epithelial integrity to allow more rapid movement of these solutes into the
bloodstream. While diabetics are another group recently shown to have increased
susceptibility to particulate air pollution (Zanobetti and Schwartz, 2002, 034821), it is
unclear whether transport of soluble particles across the epithelium is affected in these
patients. In general, it appears that coexposure to irritant pollutants results in a disruption
of epithelial integrity and macrophage function which, on the one hand, retards mucociliary
and alveolar clearance, but also allows for a more rapid movement of soluble constituents
across the epithelial surface into the interstitium and blood stream. Alterations in
epithelial permeability by disease, pollutant exposure, or infection may partially explain
increased susceptibility to PM associated with these co-conditions.
July 2009
4-43
DRAFT-DO NOT CITE OR QUOTE

-------
Chapter 4 References
ACGIH. (2005). TLVs and BEIs: Based on the documentation of the threshold limit values for chemical
substances and physical agents and biological exposure indices. 156188
Adamson I; Prieditis H. (1995). Response of mouse lung to carbon deposition during injury and repair. Environ
Health Perspect, 103: 72-76. 189982
Alexis N; Soukup J; Nierkens S; Becker S. (2001). Association between airway hyperreactivity and bronchial
macrophage dysfunction in individuals with mild asthma. , 280: L369-L375. 190013
Anselm A; Heibel T; Gebhart J; Ferron GA. (1990). In vivo studies of growth factors of sodium chloride particles
in the human respiratory tract. J Aerosol Sci, 21: S427-430. 156217
Anthony TR; Flynn MR. (2006). Computational fluid dynamics investigation of particle inhalability. J Aerosol
Sci, 37: 750-765. 155659
Antonini JM; Roberts JR; Clarke RW; Yang HM; Barger MWi Ma JYC; Weissman DN. (2001). Effect of age on
respiratory defense mechanisms: pulmonary bacterial clearance in Fischer 344 rats after intratracheal
instillation of Listeria monocytogenes. , 120: 240-249. 156219
Antoniou KM; Malagari E Tzanakis Ni Perisinakis E Symvoulakis EE Karkavitsas N> Siafakas NM; Bouros
D. (2006). Clearance of technetium-99m-DTPA and HRCT findings in the evaluation of patients with
Idiopathic Pulmonary Fibrosis. , 6: 4. 156220
Aschner M> Erikson EM; Dorman DC. (2005). Manganese Dosimetry: Species Differences and Implications for
Neurotoxicity. Crit Rev Toxicol, 35: 1-32. 155663
Asgharain B; Kelly JT; Tewksbury EW. (2003). Respiratory deposition and inhalability of monodisperse aerosols
in Long-Evans rats. Toxicol Sci, 71: 104-111. 153068
Asgharian B; Hofmann W; Miller FJ. (2001). Mucociliary clearance of insoluble particles from the
tracheobronchial airways of the human lung. J Aerosol Sci, 32: 817-832. 017025
Bailey M; Dorrian M; Birchall A. (1995). Implications of airway retention for radiation doses from inhaled
radionuclides. , 8: 373-390. 190057
Balashazy I; Hofmann W; Heistracher T. (1999). Computation of local enhancement factors for the
quantification of particle deposition patterns in airway bifurcations. J Aerosol Sci, 30: 185-203. 043201
Balashazy I; Hofmann W; Heistracher T. (2003). Local particle deposition patterns may play a key role in the
development of lung cancer. , 94: 1719-1725. 155671
Beadsmoore C; Cheow HE Szczepura E Ruparelia P; Peters AM. (2007). Healthy passive cigarette smokers
have increased pulmonary alveolar permeability. , 28: 75. 156259
Becquemin MH! Swift DL; Bouchikhi A; Roy M> Teillac A. (1991). Particle deposition and resistance in the noses
of adults and children. Eur Respir J, 4- 694-702. 009187
Becquemin MM; Bertholon JF; Bouchikhi A; Malarbet JL; Roy M. (1999). Oronasal Ventilation Partitioning in
Adults and Children: Effect on Aerosol Deposition in Airways. Radiat Prot Dosimetry, 81: 221-228.
155679
Bennett W', Messina M; Smaldone G. (1985). Effect of exercise on deposition and subsequent retention of inhaled
particles. J Appl Physiol, 59: 1046-1054. 190034
Bennett W', Zeman E Jarabek A. (2003). Nasal contribution to breathing with exercise: effect of race and
gender. J Appl Physiol, 95: 497-503. 191977
Bennett WD; Ilowite JS. (1989). Dual pathway clearance of 99mTc-DTPA from the bronchial mucosa. Am Rev
Respir Dis, 139: 1132-1138. 000835
Bennett WD; Zeman KL. (1998). Deposition of fine particles in children spontaneously breathing at rest. Inhal
Toxicol, 10: 831-842. 076182
Bennett WD; Zeman KL. (2004). Effect of body size on breathing pattern and fine-particle deposition in children.
J Appl Physiol, 97: 821-826. 155686
Bennett WD; Zeman KL. (2005). Effect of Race on Fine Particle Deposition for Oral and Nasal Breathing. Inhal
Toxicol, 17: 641-648. 155687
Bennett WD; Zeman KL; Jarabek AM. (2008). Nasal contribution to breathing and fine particle deposition in
children versus adults. J Toxicol Environ Health A, 71: 227-237. 156269
Bennett WD; Zeman KL; Em C. (1996). Variability of fine particle deposition in healthy adults: effect of age and
gender. Am J Respir Crit Care Med, 153: 1641-1647. 083284
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa. go v/her o. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009
4-44
DRAFT-DO NOT CITE OR QUOTE

-------
Bennett WD; Zeman KL; Kim C> Mascarella J. (1997). Enhanced deposition of fine particles in COPD patients
spontaneously breathing at rest. Inhal Toxicol, 9- 1"14. 078839
Bermudez E; Mangum JB; Asgharian B; Wong BA; Reverdy EE; Janszen DB; Hext PM; Warheit DB; Everitt JI.
(2002). Long-term pulmonary responses of three laboratory rodent species to subchronic inhalation of
pigmentary titanium dioxide particles. Toxicol Sci, 70- 86"97. 055578
Bermudez E; Mangum JB; Wong BA; Asgharian B; Hext PM; Warheit DB; Everitt JI. (2004). Pulmonary
responses of mice, rats, and hamsters to subchronic inhalation of ultrafine titanium dioxide particles.
Toxicol Sci, 77: 347-357. 056707
Bhalla DK; Mannix RC; KLeinman MT; Crocker TT. (1986). Relative permeability of nasal, tracheal, and
bronchoalveolar mucosa to macromolecules in rats exposed to ozone. J Toxicol Environ Health, 17: 269"
283. 040407
Bodolay E; Szekanecz Z; Devenyi K; Galuska L; Csipo I; Vegh J; Garai I; Szegedi G. (2005). Evaluation of
interstitial lung disease in mixed connective tissue disease (MCTD). , 44: 656-661. 156280
Braga F; Mango JC; Souza JF; Ferrioli E; De Andrade J; Iazigi N. (1996). Age-related reduction in 99TcnrDTPA
alveolar-capillary clearance in normal humans. , 17: 971-974. 156289
Braude S; Nolop KB; Hughes JMB; Barnes PJ; Royston D. (1986). Comparison of lung vascular and epithelial
permeability indices in the adult respiratory distress syndrome. , 133: 1002-1005. 155701
Brown J; Bennett W. (2004). Deposition of coarse particles in cystic fibrosis: model predictions versus
experimental results. , 17: 239-248. 190032
Brown JS. (2005). Particle inhalability at low wind speeds. Inhal Toxicol, 17: 831-837. 156299
Brown JS; Wilson WE; Grant LD. (2005). Dosimetric comparisons of particle deposition and retention in rats
and humans. Inhal Toxicol, 17: 355-385. 089308
Brown JS; Zeman KL; Bennett WD. (2002). Ultrafine particle deposition and clearance in the healthy and
obstructed lung. Am J Respir Crit Care Med, 166: 1240-1247. 043216
Burch WM. (2002). Comment on "Passage of inhaled particles into the blood circulation in humans". Circulation,
106: e!41-el42. 056754
Caner B. (1994). Impaired lung epithelial permeability in diabetics detected by technetium-99m-DTPA aerosol
scintigraphy. J Nucl Med, 35: 204-206. 156320
Chadha TS; Birch S; Sackner MA. (1987). Oronasal distribution of ventilation during exercise in normal
subjects and patients with asthma and rhinitis. , 92: 1037-1041. 037365
Chang C-C; Chiu H"F; Wu Y-S; Li Y-C; Tsai M-L; Shen C-K; Yang C-Y. (2005). The induction of vascular
endothelial growth factor by ultrafine carbon black contributes to the increase of alveolar-capillary
permeability. , 113: 454-460. 097776
Chen J; Tan M; Nemmar A; Song W', Dong M; Zhang G; Li Y. (2006). Quantification of extrapulmonary
translocation of intratracheal-instilled particles in vivo in rats: effect of lipopolysaccharide. Toxicology,
222: 195-201. 147267
Cheng K-H; Cheng Y"S; Yeh H"C; Guilmette RA; Simpson SQ; Yang Y-H; Swift DL. (1996). In vivo
measurements of nasal airway dimensions and ultrafine aerosol deposition in the human nasal and oral
airways. J Aerosol Sci, 27: 785-801. 047520
Chinard FP. (1980). The alveolar-capillary barrier: some data and speculations. Microvasc Res, 19: 1-17. 156341
Churg A; Stevens B; Wright JL. (1998). Comparison of the uptake of fine and ultrafine Ti02 in a tracheal
explant system. Am J Physiol, 274: L81-L86. 085815
Cohen MD; Zelikoff JT; Chen L-C; Schlesinger RB. (1997). Pulmonary retention and distribution of inhaled
chromium: effects of particle solubility and coexposure to ozone. Inhal Toxicol, 9: 843-865. 009213
Creasia DA; Poggenburg JK Jr; Nettesheim P. (1976). Elution of benzo[a]pyrene from carbon particles in the
respiratory tract of mice. J Toxicol Environ Health, l: 967-975. 059713
Dai YT; Juang YJ; Wu Y; Breysse PN; Hsu DJ. (2006). In vivo measurements of inhalability of ultralarge
aerosol particles in calm air by humans. J Aerosol Sci, 37: 967-973. 156377
DeLorenzo AJD. (1970). The olfactory neuron and the blood-brain barrier. In Taste and Smell in Vertebrates
(pp. 151-175). London: Churchill Livingstone. 156391
Dorman DC; McManus BE; Parkinson CU; Manuel CA; McElveen AM; Everitt JI. (2004). Nasal Toxicity of
Manganese Sulfate and Manganese Phosphate in Young Male Rats Following Subchronic (13-Week)
Inhalation Exposure. Inhal Toxicol, 16: 481-488. 155752
Dorman DC; Struve MF; James RA; Marshall MW; Parkinson CU; Wong BA. (2001). Influence of particle
solubility on the delivery of inhaled manganese to the rat brain: manganese sulfate and manganese
tetroxide pharmacokinetics following repeated (l4"day) exposure. Toxicol Appl Pharmacol, 170: 79"87.
055433
July 2009
4-45
DRAFT-DO NOT CITE OR QUOTE

-------
Edetsberger Mi Gaubitzer Ei Valic Ei Waigmann Ei Kohler G. (2005). Detection of nanometer-sized particles in
living cells using modern fluorescence fluctuation methods. Biochem Biophys Res Commun, 332- 109-116.
155759
Elder Ai Gelein Ri Silva Vi Feikert Ti Opanashuk Li Carter Ji Potter Ri Maynard Ai Ito Yi Finkelstein Ji
Oberdorster G. (2006). Translocation of inhaled ultrafine manganese oxide particles to the central
nervous system. Environ Health Perspect, 114: 1172-1178. 089253
Enna SJi Schanker LS. (1972). Absorption of drugs from the rat lung. Am J Physiol, 223: 1227-1231. 155767
Falk Ri Philipson Ki Svartengren Mi Jarvis Ni Bailey Mi Camner P. (1997). Clearance of particles from small
ciliated airways. , 23: 495-515. 086080
Farkas Ai Balashazy Ii SzQcs K. (2006). Characterization of Regional and Local Deposition of Inhaled Aerosol
Drugs in the Respiratory System by Computational Fluid and Particle Dynamics Methods. , 19: 329-343.
155771
Farkas Ai Balashazy I. (2008). Quantification of particle deposition in asymmetrical tracheobronchial model
geometry. , 38: 508-18. 157358
Ferin Ji Oberdorster Gi Penney DP. (1992). Pulmonary retention of ultrafine and fine particles in rats. Am J
Respir Cell Mol Biol, 6: 535-542. 044401
Finlay WHi Martin AR. (2008). Recent advances in predictive understanding of respiratory tract deposition. ,
21: 189-206. 155776
Fishman APi Elias JA. (1980). Fishman's Pulmonary diseases and disorders. 156436
Foster WMi Stetkiewicz PT. (1996). Regional clearance of solute from the respiratory epithelia: 18--20 h
postexposure to ozone. J Appl Physiol, 81: 1143-1149. 079920
Foster WMi Wagner EM. (2001). Bronchial edema alters 99mTc-DTPA clearance from the airway surface in
sheep. J Appl Physiol, 91: 2567-2573. 155778
Geiser Mi Rothen-Rutishauser Bi Kapp Ni Schurch Si Kreyling Wi Schulz Hi Semmler Mi Im Hof Vi Heyder Ji
Gehr P. (2005). Ultrafine particles cross cellular membranes by nonphagocytic mechanisms in lungs and
in cultured cells. Environ Health Perspect, 113: 1555-1560. 087362
Geys Ji Coenegrachts Li Vercammen Ji Engelborghs Yi Nemmar Ai Nemery Bi Hoet PHM. (2006). In vitro study
of the pulmonary translocation of nanoparticles A preliminary study. Toxicol Lett, 160: 218-226. 155789
Goodman RMi Yergin BMi Landa JF; Golinvaux MHi Sackner MA. (1978). Relationship of smoking history and
pulmonary function tests to tracheal mucous velocity in nonsmokers, young smokers, ex-smokers, and
patients with chronic bronchitis. Am Rev Respir Dis, 117: 205-214. 071130
Greenspan BJi Morrow PEi Ferin J. (1988). Effects of aerosol exposures to cadmium chloride on the clearance of
titanium dioxide from the lungs of rats. Exp Lung Res, 14: 491-499. 045031
Grgic Bi Finlay WHi Burnell PKPi Heenan AF. (2004). In vitro intersubject and intrasubject deposition
measurements in realistic mouth—throat geometries. J Aerosol Sci, 35: 1025-1040. 155810
Hanel Bi Law Ii Mortensen J. (2003). Maximal rowing has an acute effect on the blood-gas barrier in elite
athletes. , 95: 1076-1082. 155826
Heistracher Ti Hofmann W. (1997). Flow and deposition patterns in successive airway bifurcations. Presented
at In: Cherry, N.i Ogden, T., eds. Inhaled Particles VIII: proceedings of an international symposium on
inhaled particles organised by the British Occupational Hygiene Society; August 1996; Cambridge, UK.
Ann. Occup. Hyg. 4l(suppl.): 537-542. 047514
Ho JCi Chan KN; Hu WH! Lam WE Zheng L; Tipoe GL; Sun J; Leung R; Tsang KW. (2001). The effect of aging
on nasal mucociliary clearance, beat frequency, and ultrastructure of respiratory cilia. Am J Respir Crit
Care Med, 163: 983-988. 156549
Hofmann W> Asgharian B. (2003). The effect of lung structure on mucociliary clearance and particle retention in
human and rat lungs. Toxicol Sci, 73: 448-456. 055579
Hofmann W> Balashazy L Heistracher T; Koblinger L. (1996). The significance of particle deposition patterns in
bronchial airway bifurcations for extrapolation modeling. Aerosol Sci Technol, 25: 305-327. 047515
Hofmann Wi Martonen TBi Graham RC. (1989). Predicted deposition of nonhygroscopic aerosols in the human
lung as a function of subject age. : J. Aerosol Med. 2. 006922
Hsu DJi Swift DL. (1999). The measurements of human inhalability of ultralarge aerosols in calm air using
mannikins. J Aerosol Sci, 30: 1331-1343. 155855
Huchon GJi Montgomery ABi Lipavsky Ai Hoeffel JMi Murray JF. (1987). Respiratory clearance of aerosolized
radioactive solutes of varying molecular weight. J Nucl Med, 28: 894-902. 024923
Huchon GJi Russell JA; Barritault LG; Lipavsky A; Murray JF. (1984). Chronic air-flow limitation does not
increase respiratory epithelial permeability assessed by aerosolized solute, but smoking does. Am Rev
Respir Dis, 130: 457-460. 156576
ICRP. (1994). Human respiratory tract model for radiological protection: a report of a task group of the
International Commission on Radiological Protection. Ann ICRP, 24: 1-482. 006988
July 2009
4-46
DRAFT-DO NOT CITE OR QUOTE

-------
Ilowite JS; Bennett WD; Sheetz MS; Groth ML; Nierman DM. (1989). Permeability of the bronchial mucosa to
99mTc-DTPAin asthma. Am Rev Respir Dis, 139- 1139-1143. 156584
ILSI. (2000). The relevance of the rat lung response to particle overload for human risk assessment- a workshop
consensus report. Inhal Toxicol, 12: 1-17. 002892
Isaacs KK; Schlesinger RB; Martonen TB. (2006). Three-dimensional computational fluid dynamics simulations
of particle deposition in the tracheobronchial tree. , 19- 344-352. 155861
ISO. (2008). Nanotechnologies - Terminology and definitions for nano-objects -- Nanoparticle, nanofibre and
nanoplate. International Organization for Standardization. Geneva. 190066
James DS; Stidley CA; Lambert WE; Chick TW; Mermier CM; Samet JM. (1997). Oronasal distribution of
ventilation at different ages. Arch Environ Occup Health, 52: 118-123. 042422
Jarabek AM; Asgharian B; Miller FJ. (2005). Dosimetric adjustments for interspecies extrapolation of inhaled
poorly soluble particles (PSP). Inhal Toxicol, 17: 317-334. 056756
Jeffers DE. (2005). Relative magnitudes of the effects of electrostatic image and thermophoretic forces on
particles in the respiratory tract. Radiat Prot Dosimetry, 113: 189-194. 156608
Jones JG; Minty BD; Lawler P; Hulands G; Crawley JC; Veall N. (1980). Increased alveolar epithelial
permeability in cigarette smokers. Lancet, L 66"68. 155883
Jones JG; Minty BD; Royston D; Royston JP. (1983). Carboxyhaemoglobin and pulmonary epithelial
permeability in man. Thorax, 38: 129-133. 155884
Kapp N; Kreyling W', Schulz H; Im Hof V; Gehr P; Semmler M; Geiser M. (2004). Electron energy loss
spectroscopy for analysis of inhaled ultrafine particles in rat lungs. Microsc Res Tech, 63: 298-305.
156624
Kaya E; Fidan F; Unlu M; Sezer M; Tetik L; Acar M. (2006). Evaluation of alveolar clearance by Tc-99m DTPA
radioaerosol inhalation scintigraphy in welders. , 20: 503. 156632
Kaya M; Salan A; Tabakoglu E; Aydogdu N; Berkarda S. (2003). The bronchoalveolar epithelial permeability in
house painters as determined by Tc_99m DTPA aerosol scintigraphy. , 17: 305-308. 156631
Kehrl HR; Vincent LM; Kowalsky RJ; Horstman DH; O'Neil JJ; McCartney WH; Bromberg PA. (1987). Ozone
exposure increases respiratory epithelial permeability in humans. Am J Respir Crit Care Med, 135: 1124"
1128. 040824
Kelly JT; Asgharian B; Wong BA. (2005). Inertial Particle Deposition in a Monkey Nasal Mold Compared with
that in Human Nasal Replicas. Inhal Toxicol, 17: 823-830. 155894
Kim CS. (2000). Methods of calculating lung delivery and deposition of aerosol particles. Respir Care, 45: 695"
711.013112
Kim CS; Hu SC. (1998). Regional deposition of inhaled particles in human lungs: comparison between men and
women. J Appl Physiol, 84: 1834-1844. 086066
Kim CS; Iglesias AJ. (1989). Deposition of inhaled particles in bifurcating airway models: I inspiratory
deposition. , 2: 1-14. 078539
Kim CS; Iglesias AJ; Garcia L. (1989). Deposition of inhaled particles in bifurcating airway models: H expiratory
deposition. , 2: 15-27. 078538
Kim CS; Jaques PA. (2000). Respiratory dose of inhaled ultrafine particles in healthy adults. , 358: 2693-2705.
012811
Kimbell JS. (2006). Nasal Dosimetry of Inhaled Gases and Particles: Where Do Inhaled Agents Go in the Nose?.
Toxicol Pathol, 34: 270-273. 155902
Kreyling WG. (1992). Intracellular particle dissolution in alveolar macrophages. Environ Health Perspect, 97:
121-126. 067243
Kreyling WG; Blanchard JD; Godleski JJ; Haeussermann S; Heyder J; Hutzler P; Schulz H; Sweeney TD;
Takenaka S; Ziesenis A. (1999). Anatomic localization of 24" and 96"h particle retention in canine
airways. J Appl Physiol, 87: 269-284. 039175
Kreyling WG; Scheuch G. (2000). Clearance of particles deposited in the lungs. In Gehr, P.; Heyder, J.
(Ed.),Particle-lung interactionsNew York, NY: Marcel Dekker, Inc. 056281
Kreyling WG; Semmler M; Erbe F; Mayer P; Takenaka S; Schulz H; Oberdorster G; Ziesenis A. (2002).
Translocation of ultrafine insoluble iridium particles from lung epithelium to extrapulmonary organs is
size dependent but very low. J Toxicol Environ Health A, 65: 1513-1530. 037332
Laitinen LA; Heino M; Laitinen A; Kava T; Haahtela T. (1985). Damage of the airway epithelium and bronchial
reactivity in patients with asthma. Am Rev Respir Dis, 131: 599-606. 037521
Landry TD; Gushow TS; Langvardt PW; Wall JM; McKenna MJ. (1983). Pharmacokinetics and metabolism of
inhaled methyl chloride in the rat and dog. Toxicol Appl Pharmacol, 68: 473-486. 003904
Lay J; Alexis N; Zeman K; Peden D; Bennett W. (2009). In vivo uptake of inhaled particles by airway phagocytes
is enhanced in patients with mild asthma compared with normal volunteers. Thorax, 64: 313-320. 190060
July 2009
4-47
DRAFT-DO NOT CITE OR QUOTE

-------
Lay JC; Stang MR; Fisher PE; Yankaskas JR; Bennett WD. (2003). Airway Retention of Materials of Different
Solubility following Local Intrabronchial Deposition in Dogs. , 16- 153-166. 155920
Lee KP; Henry NW III; Trochimowicz HJ; Reinhardt CF. (1986). Pulmonary response to impaired lung
clearance in rats following excessive Ti02 dust deposition. Environ Res, 41' 144-167. 067629
Lee KP; Trochimowicz HJ; Reinhardt CF. (1985). Pulmonary response of rats exposed to titanium dioxide (Ti02)
by inhalation for two years. Toxicol Appl Pharmacol, 79: 179-192. 003847
Lee KP; Trochimowicz HJ; Reinhardt CF. (1985). Transmigration of titanium dioxide (Ti02) particles in rats
after inhalation exposure. Exp Mol Pathol, 42: 331-343. 067628
Lin CC; Chang CT; Li TC; Kao A. (2002). Objective Evidence of Impairment of Alveolar Integrity in Patients
with Non-Insulin-Dependent Diabetes Mellitus Using Radionuclide Inhalation Lung Scan. Lung, 180:
181-186. 155932
Lorino AM; Meignan M; Bouissou P; Atlan G. (1989). Effects of sustained exercise on pulmonary clearance of
aerosolized 99mTc-DTPA. , 67: 2055-2059. 155946
Lundgren DL; Hahn FF; Crain CR; Sanchez A. (1978). Effect of Influenza Virus Infection on the Pulmonary
Retention of Inhaled 144Ce and Subsequent Survival of Mice. , 34: 557. 155950
Martonen TB; Yang Y; Xue ZQ. (1994). Effects of carinal ridge shapes on lung airstreams. Aerosol Sci Technol,
21: 119-136. 000847
Mason GR; Uszler JM; Effros RM; Reid E. (1983). Rapid reversible alterations of pulmonary epithelial
permeability induced by smoking. , 83: 6-11. 013169
Matsui H; Randell SH; Peretti SW; Davis CW; Boucher RC. (1998). Coordinated clearance of periciliary liquid
and mucus from airway surfaces. J Clin Invest, 102: 1125-1131. 040405
Mauderly JL. (1997). Relevance of particle-induced rat lung tumors for assessing lung carcinogenic hazard and
human lung cancer risk. Presented at In: Driscoll, K. E.; Oberdorster, G., eds. Proceedings of the sixth
international meeting on the toxicology of natural and man-made fibrous and non-fibrous particles;
September 1996; Lake Placid, NY. Environ. Health Perspect. Suppl. 105(5): 1337-1346. 084631
Meignan M; Rosso J; Leveau J; Katz A; Cinotti L; Madelaine G; Galle P. (1986). Exercise increases the lung
clearance of inhaled technetium-99m DTPA. J Nucl Med, 27' 274-280. 156752
Miller FJ. (2000). Dosimetry of particles in laboratory animals and humans in relationship to issues
surrounding lung overload and human health risk assessment: a critical review. Inhal Toxicol, 12: 19-57.
011822
Mills NL; Amin N; Robinson SD; Anand A; Davies J; Patel D; de la Fuente JM; Cassee FR; Boon NA; Macnee W;
Millar AM; Donaldson K; Newby DE. (2006). Do inhaled carbon nanoparticles translocate directly into
the circulation in humans?. Am J Respir Crit Care Med, 173: 426-431. 088770
Morgan W', Ahmad D; Chamberlain M; Clague H; Pearson M; Vinitski S. (1984). The effect of exercise on the
deposition of an inhaled aerosol. , 56: 327-338. 190035
Morrow PE. (1994). Mechanisms and significance of "particle overload". Presented at . 006850
Mousa K; Onadeko BO; Mustafa HT; Mohamed M; Nabilla A; Omar A; Al"Bunni A; Elgazzar A. (2000).
Technetium 99mTc_DTPA clearance in the evaluation of pulmonary involvement in patients with
diabetes mellitus. Respir Med, 94: 1053-1056. 156786
Muhle H; Creutzenberg O; Bellmann B; Heinrich U; Mermelstein R. (1990). Dust overloading of lungs:
investigations ofvarious materials, species differences, and irreversibility of effects. , l: S111-S128.
006853
Menache MG; Miller FJ; Raabe OG. (1995). Particle inhalability curves for humans and small laboratory
animals. Ann Occup Hyg, 39: 317-328. 006533
Moller W; Felten K; Sommerer K; Scheuch G; Meyer G; Meyer P; Haussinger K; Kreyling WG. (2008).
Deposition, retention, and translocation of ultrafine particles from the central airways and lung
periphery. Am J Respir Crit Care Med, 177: 426-432. 156771
Moller W; Haussinger K; Winkler-Heil R; Stahlhofen W; Meyer T; Hofmann W', Heyder J. (2004). Mucociliary
and long-term particle clearance in the airways of healthy nonsmoker subjects. , 97: 2200-2206. 155987
Nemmar A; Hoet PHM; Vanquickenborne B; Dinsdale D; Thomeer M; Hoylaerts MF; Vanbilloen H; Mortelmans
L; Nemery B. (2002). Passage of inhaled particles into the blood circulation in humans. Circulation, 105:
411-414. 024914
Niinimaa V; Cole P; Mintz S; Shephard RJ. (1981). Oronasal distribution of respiratory airflow. Respir Physiol
Neurobiol, 43: 69-75. 071758
Nikula KJ; Vallyathan V; Green FHY; Hahn FF. (2001). Influence of exposure concentration or dose on the
distribution of particulate material in rat and human lungs. Environ Health Perspect, 109: 311-318.
016641
NRPB. (2004). Particle deposition in the vicinity of power lines and possible effects on health: report of an
independent advisory group on non-ionising radiation and its ad hoc group on corona ions. 156815
July 2009
4-48
DRAFT-DO NOT CITE OR QUOTE

-------
Oberdorster G. (1995). Lung particle overload- implications for occupational exposures to particles. Regul
Toxicol Pharmacol, 27: 123-135. 046596
Oberdorster G. (1996). Significance of particle parameters in the evaluation of exposure-dose-response
relationships of inhaled particles. Inhal Toxicol, 8 Supplement- 73"89. 039852
Oberdorster G. (2002). Toxicokinetics and effects of fibrous and nonfibrous particles. Inhal Toxicol, 14- 29"56.
021111
Oberdorster Gi Ferin J; Gelein R; Soderholm SC; Finkelstein J. (1992). Role of the alveolar macrophage in lung
injury- studies with ultrafine particles. Environ Health Perspect, 97- 193-199. 045110
Oberdorster Gi Finkelstein JNi Johnston Ci Gelein Ri Cox Ci Baggs Ri Elder ACP. (2000). Acute pulmonary
effects of ultrafine particles in rats and mice. 039014
Oberdorster Gi Sharp Zi Atudorei Vi Elder Ai Gelein Ri Kreyling Wi Cox C. (2004). Translocation of inhaled
ultrafine particles to the brain. Inhal Toxicol, 16- 437-445. 055639
Oberdorster G. (1988). Lung clearance of inhaled insoluble and soluble particles. , 1- 289-330. 006857
Oberdorster Gi Ferin Ji Lehnert BE. (1994). Correlation between particle size, in vivo particle persistence, and
lung injury. Environ Health Perspect, 102 : 173-179. 046203
Oberdorster Gi Ferin Ji Soderholm Si Gelein Ri Cox Ci Baggs Ri Morrow PE. (1994). Increased pulmonary
toxicity of inhaled ultrafine particles' due to lung overload alone?. Ann Occup Hyg, 38- 295-302. 056285
Oldham MJi Robinson RJ. (2007). Predicted tracheobronchial and pulmonary deposition in a murine asthma
model. , 290: 1309-1314. 156003
Persson Ei Henriksson Ji Tallkvist Ji Rouleau Ci Tjalve H. (2003). Transport and subcellular distribution of
intranasally administered zinc in the olfactory system of rats and pikes. Toxicology, 191: 97-108. 051846
Peterson BTi Dickerson KDi James HLi Miller EJi McLarty JWi Holiday DB. (1989). Comparison of three
tracers for detecting lung epithelial injury in anesthetized sheep. J Appl Physiol, 66: 2374-2383. 024922
Phalen RFi Oldham MJ. (2006). Aerosol dosimetry considerations. Clin Occup Environ Med, 5: 773-784. 156024
Phalen RFi Oldham MJi Mautz WJ. (1989). Aerosol Deposition in the Nose As a Function of Body Size. , 57: 299"
305. 156023
Phalen RFi Oldham MJi Wolff RK. (2008). The relevance of animal models for aerosol studies. , 21: 113-124.
156865
Pigorini Fi Maini CLi Pau Fi Giosue S. (1988). The influence of age on the pulmonary clearance of 99Tcm_DTPA
radioaerosol. , 9: 965-971. 156027
Puchelle Ei Zahm J-Mi Bertrand A. (1979). Influence of age on bronchial mucociliary transport. , 60: 307-313.
006863
Putz-Anderson Vi Setzer JVi Croxton JSi Phipps FC. (1981). Methyl chloride and diazepam effects on
performance. Scand J Work Environ Health, 7: 8"13. 003914
Raabe OGi Al"Bayati MAi Teague SVi Rasolt A. (1988). Regional deposition of inhaled monodisperse, coarse,
and fine aerosol particles in small laboratory animals. Presented at In: Dodgson, J.i McCallum, R. I.i
Bailey, M. R.i Fischer, D. R., eds. Inhaled particles VI: proceedings of an international symposium and
workshop on lung dosimetry; September 1985; Cambridge, United Kingdom. Ann. Occup. Hyg. 32(suppl.
1): 53-63. 001439
Roth C; Scheuch G; Stahlhofen W. (1993). Clearance of the human lungs for ultrafine particles. J Aerosol Sci,
24: S95-S96 . 156928
Sakagami M; Byron PR; Venitz J; Rypacek F. (2002). Solute disposition in the rat lung in vivo and in vitro:
determining regional absorption kinetics in the presence of mucociliary escalator. J Pharmacol Sci, 91:
594-604. 156936
Schanker LS; Mitchell EW! Brown RA Jr. (1986). Species comparison of drug absorption from the lung after
aerosol inhalation or intratracheal injection. Drug Metab Dispos, 14: 79"88. 005100
Scheuch G; Gebhart J; Roth C. (1990). Uptake of electrical charges in the human respiratory tract during
exposure to air loaded with negative ions. J Aerosol Sci, 21: S439-S442. 006948
Schroeter JD; Kimbell JS; Asgharian B. (2006). Analysis of Particle Deposition in the Turbinate and Olfactory
Regions Using a Human Nasal Computational Fluid Dynamics Model. , 19: 301-313. 156076
Semmler M; Seitz J; Erbe F; Mayer P; Heyder J; Oberdorster G; Kreyling WG. (2004). Long-term clearance
kinetics of inhaled ultrafine insoluble iridium particles from the rat lung, including transient
translocation into secondary organs. Inhal Toxicol, 16: 453-459. 055641
Semmler-Behnke M; Takenaka S; Fertsch S; Wenk A; Seitz J; Mayer P; Oberdorster G; Kreyling WG. (2007).
Efficient elimination of inhaled nanoparticles from the alveolar region: evidence for interstitial uptake
and subsequent reentrainment onto airways epithelium. Environ Health Perspect, 115: 728. 156080
Smith JR; Bailey MR; Etherington G; Shutt AL; Youngman MJ. (2008). Effect of particle size on slow particle
clearance from the bronchial tree. Exp Lung Res, 34: 287 - 312. 190037
July 2009
4-49
DRAFT-DO NOT CITE OR QUOTE

-------
Snipes MB. (1996). Current information on lung overload in nonrodent mammals' contrast with rats. Inhal
Toxicol, 8: 91-109. 076041
Snipes MB; Harkema JR.; Hotchkiss JA; Bice DE. (1997). Neutrophil Involvement in the Retention and
Clearance of Dust Intratracheally Instilled into the LUNGS of F344/N Rats. , 23: 65"84. 156092
Soderholm SC. (1985). Size-selective sampling criterial for inspirable mass fraction. : . 156992
Sousa SR; Moradas-Ferreira P; Saramango B; Viseu ML; Barbosa MA. (2004). Human serum albumin
adsorption on Ti02 from single protein solutions and from plasma. Langmuir, 20: 9745-9754. 089866
Suga K; Yuan Y; Ogasawara N; Tsukuda T; Matsunaga N. (2003). Altered Clearance of Gadolinium
Diethylenetriaminepentaacetic Acid Aerosol from Bleomycin "injured Dog Lungs: Initial Observations.
Am J Respir Crit Care Med, 167: 1704-1710. 157024
Svartengren M; Falk R; Philipson K. (2005). Long-term clearance from small airways decreases with age. Eur
Respir J, 26: 609-615. 157034
Tabachnik E; Muller NJ Toye B; Levison H. (1981). Measurement of ventilation in children using the respiratory
inductive plethysmograph. J Pediatr, 99: 895-899. 157036
Takenaka S; Dornhofer-Takenaka H; Muhle H. (1986). Alveolar distribution of fly ash and of titanium dioxide
after long-term inhalation by Wistar rats. J Aerosol Sci, 17: 361-364. 046210
Takenaka S; Karg EJ Kreyling W', Lentner B; Moller W; Behnke-Semmler M; Jennen L; Walch A; Michalke B;
Schramel P. (2006). Distribution Pattern of Inhaled Ultrafine Gold Particles in the Rat Lung. Inhal
Toxicol, 18: 733-740. 156110
Tankersley CG; Shank JA; Flanders SE; Soutiere SE; Rabold R; Mitzner W; Wagner EM. (2003). Changes in
lung permeability and lung mechanics accompany homeostatic instability in senescent mice. J Appl
Physiol, 95: 1681-1687. 096363
Tobin MJ; Chadha TS; Jenouri G; Birch SJ; Gazeroglu HB; Sackner MA. (1983). Breathing patterns. 1. Normal
subjects. , 84: 202-205. 156122
Tran CL; Buchanan D; Cullen RT; Searl A; Jones AD; Donaldson K. (2000). Inhalation of poorly soluble
particles II Influence of particle surface area on inflammation and clearance. Inhal Toxicol, 12: 1113-
1126. 013071
Tu KW; Knutson EO. (1984). Total deposition of ultrafine hydrophobic and hygroscopic aerosols in the human
respiratory system. Aerosol Sci Technol, 3: 453-465. 072870
U.S. EPA. (1996). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA/600/P-95/001aF-cF. 079380
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA/600/P"99/002aF"bF. 056905
Valberg P; Brain J; Sneddon S; LeMott S. (1982). Breathing patterns influence aerosol deposition sites in
excised dog lungs. J Appl Physiol, 53: 824-837. 190019
Vastag E; Matthys H; Kohler D; Gronbeck L; Daikeler G. (1985). Mucociliary clearance and airways obstruction
in smokers, ex-smokers and normal subjects who never smoked. , 139: 93-100. 157088
Vincent JH; Donaldson K. (1990). A dosimetric approach for relating the biological response of the lung to the
accumulation of inhaled mineral dust. Br J Ind Med, 47: 302-307. 002462
Wagner EM; Foster WM. (2001). Interdependence of bronchial circulation and clearance of 99mTc_DTPA from
the airway surface. J Appl Physiol, 90: 1275-1281. 156143
Wallenborn JG; Kovalcik KD; McGee JK; Landis MS; Kodavanti UP. (2009). Systemic translocation of (70)zinc:
kinetics following intratracheal instillation in rats. Toxicol Appl Pharmacol, 234: 25"32. 191172
Wallenborn JG; McGee John K; Schladweiler Mette C; Ledbetter Allen D; Kodavanti Urmila P. (2007). Systemic
translocation of particulate matter-associated metals following a single intratracheal instillation in rats.
J Toxicol Sci, 98: 231-239. 156144
Wang JX; Chen CY; Yu HW; Sun J; Li B; Li YF; Gao YX; He W; Huang YY; Chai ZF. (2007). Distribution of TiO
2 particles in the olfactory bulb of mice after nasal inhalation using microbeam SRXRF mapping
techniques. , 272: 527-531. 156146
Warheit DB; Hansen JF; Yuen IS; Kelly DP; Snajdr SI; Hartsky MA. (1997). Inhalation of high concentrations
of low toxicity dusts in rats results in impaired pulmonary clearance mechanisms and persistent
inflammation. Toxicol Appl Pharmacol, 145: 10-22. 086055
Warheit DB; Webb TR; Sayes CM; Colvin VL; Reed KL. (2006). Pulmonary instillation studies with nanoscale
Ti02 rods and dots in rats: toxicity is not dependent upon particle size and surface area. Toxicol Sci, 91:
227-236. 088436
Watanabe N; Tanada S; Sasaki Y. (2007). Pulmonary clearance of aerosolized 99mTc-DTPAin sarcoidosis I
patients. , 51: 82-90. 157115
Weynand B; Jonckheere A; Frans A; Rahier J; Saint-Luc CU. (1999). Diabetes mellitus Induces a Thickening of
the Pulmonary Basal Lamina. Respiration, 66: . 157140
July 2009
4-50
DRAFT-DO NOT CITE OR QUOTE

-------
Whaley SL; Muggenburg BA; Seiler FA; Wolff RK. (1987). Effect of aging on tracheal mucociliary clearance in
beagle dogs. J Appl Physiol, 62: 1331-1334. 156153
Wiebert P; Sanchez-Crespo A; Falk R; Philipson Ki Lundin A; Larsson Si Moller Wi Kreyling Wi Svartengren M.
(2006). No Significant Translocation of Inhaled 35 ¦nm Carbon Particles to the Circulation in Humans.
Inhal Toxicol, 18: 741-747. 156154
Wiebert Pi Sanchez-Crespo Ai Seitz Ji Falk Ri Philipson Ki Kreyling WGi Moller Wi Sommerer Ki Larsson Si
Svartengren M. (2006). Negligible clearance of ultrafine particles retained in healthy and affected human
lungs. Eur Respir J, 28: 286-290. 157146
Winter-Sorkina Rdei Cassee FR. (2002). From concentration to dose: factors influencing airborne particulate
matter deposition in humans and rats. 043670
Xu GBi Yu CP. (1986). Effects of age on deposition of inhaled aerosols in the human lung. Aerosol Sci Technol, 5'
349-357. 072697
Yeates DBi Gerrity TRi Garrard CS. (1981). Particle deposition and clearance in the bronchial tree. Ann Biomed
Eng, 9: 577-592. 095391
Yu CP. (1985). Theories of electrostatic lung deposition of inhaled aerosols. Ann Occup Hyg, 29: 219-227. 006963
Yu IJ; Park JD; Park ES; Song KS; Han KT; Han JH; Chung YH; Choi BS; Chung KH; Cho MH. (2003).
Manganese Distribution in Brains of Sprague—Dawley Rats After 60 Days of Stainless Steel Welding-
Fume Exposure. , 24: 777-785. 156171
Zanobetti A; Schwartz J. (2002). Cardiovascular damage by airborne particles: are diabetics more susceptible?. ,
13: 588-592. 034821
Zeidler-Erdely PC; Calhoun WJ; Ameredes BT; Clark MP; Deye GJ; Baron P; Jones W', Blake T; Castranova V.
(2006). In vitro cytotoxicity of Manville Code 100 gl ass fibers: Effect of fiber length on human alveolar
macrophages. Part Fibre Toxicol, 28: 3:5. 190967
Zhao X; Ju Y; Liu C; Li J; Huang M; Sun J; Wang T. (2009). Bronchial anatomy of left lung: a study of multi-
detector row CT. , 31: 85-91. 157187
Ozsahin K; Tugrul A; Mert SJ Yiiksel M; Tugrul G. (2006). Evaluation of pulmonary alveolo-capillary
permeability in Type 2 diabetes mellitus Using technetium 99mTc_DTPA aerosol scintigraphy and
carbon monoxide diffusion capacity. , 20: 205-209. 156833
July 2009
4-51
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Chapter 5. Possible Pathways/
Modes of Action
The mechanisms underlying pulmonary effects of inhaled PM have been well-studied
in the laboratory and there is general agreement regarding the key roles played by cellular
injury and inflammation. These pathways are initiated following the deposition of inhaled
particles on respiratory tract surfaces. Since most of these studies were conducted at
concentrations of PM higher than ambient levels, there is some question regarding the
relevance of these responses and mechanisms to ambient exposures.
Interestingly, inhaled PM may also affect the cardiovascular, hematopoietic and other
systems. Mechanisms underlying these extra-pulmonary effects are incompletely
understood. However, pulmonary inflammation can lead to systemic inflammation and
pulmonary reflexes can activate the autonomic nervous system (ANS). These latter
responses may mediate cardiovascular and other systemic effects, as will be discussed
below. In addition, it has been proposed that PM or soluble components of PM reach the
circulation by translocating across the epithelial and endothelial barriers of the respiratory
tract. In this way PM or its components may interact directly with cells in the vasculature
and blood and be transported to the heart and other organs. At this time, evidence clearly
supports the translocation of small solutes following inhalation exposures and the
translocation of soluble components of PM following some high dose exposures involving
intratracheal instillation. However there is insufficient evidence to support translocation of
appreciable amounts of intact particles following inhalation exposures at lower
concentrations (see Section 4.3.3.1). Future studies will be required to resolve these issues.
The following sections discuss biological pathways which comprise proposed modes of
action for the pulmonary and extra-pulmonary effects of inhaled PM. Overall themes are
emphasized and supportive evidence from new in vitro and in vivo animal studies is cited.
The characterization of evidence here is for PM in general, since most of the potential
pathways or modes of action do not appear to be specific to a particular size class of PM.
However, characteristics of ultrafine PM may allow for unique modes of action or effects
disproportionate to their mass, as will be described below. Finally, a compilation of results
from new inhalation studies which are relevant to ambient PM exposures and which
confirm and extend these proposed mechanisms is found at the end of this chapter. Detailed
descriptions of these key new studies are found in Chapters 6 and 7.
Note- Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health
and Environmental Research Online) at http • I lev a. go v/her o. HERO is a database of scientific literature used by U.S. EPA in
the process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009
5-1
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
5.1. Pulmonary Effects
5.1.1. Reactive Oxygen Species
A great deal of research interest has focused on the role of reactive oxygen species
(ROS) in the initiation of pulmonary injury and inflammation following exposure to PM.
Numerous studies have demonstrated PM oxidative potential in in vitro assay systems
(Ayres et al., 2008, 155666; Cho et al., 2005, 087937; Shi et al., 2003, 088248; Tao et al.,
2003, 156111). Both redox active surface components such as metals and organic species
and the surface characteristics of crystal structures have been shown to contribute to
oxidative potential (Jiang et al., 2008, 156609; Tao et al., 2003, 156111; Warheit et al., 2007,
090482). In this way, PM may be a direct source of ROS in the respiratory tract
(Figure 5" 1).
Redox Active
Surface Components
Metals, Organics
Surface Characteristics
of Crystal Structures
PM Oxidative
Potential
7 Cell-free assay or
oxidation of components
\ in a cellular system ,
Figure 5-1. PM oxidative potential.
PM may also act as an indirect source of ROS in the respiratory tract by stimulating
cells to produce ROS (Ayres et al., 2008, 155666; Tao et al., 2003, 156111) (Figure 5-2). This
July 2009
5-2
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
may explain the observation that the oxidative potential of isolated PM does not always
correlate with cellular or tissue oxidative stress induced by PM exposure. Exposure to PM
increases intracellular production of ROS by a variety of mechanisms. For example, PM
interaction with cell surfaces results in stimulation of NADPH oxidase in macrophages (i.e.,
the respiratory burst) (Dostert et al., 2008, 155753) and in epithelial cells (Amara et al.,
2007, 156212; Becher et al., 2007, 097125; Tamaoki et al., 2004, 157040). Absorption of PM
soluble components (e.g., PAH, transition metals) by respiratory tract cells can occur (Penn
et al., 2005, 088257) and be followed by microsomal transformation of PAHs to quinones or
by redox cycling of transition metals with production of intracellular ROS (Molinelli et al.,
2002, 035347; Xia et al., 2004, 087486). Disruption of intracellular iron homeostasis with
the subsequent generation of ROS has also been demonstrated following PM exposure (Ghio
and Cohen, 2005, 088272). In some cases, mitochondria serve as the source of ROS in
response to PM (Huang et al., 2003, 156573; Risom et al., 2005, 189016; Soberanes et al.,
2006,	156991; Soberanes et al., 2009, 190483). Furthermore, PM interaction with cells can
lead to the induction of nitric oxide synthase (Becher et al., 2007, 097125; Lindbom et al.,
2007,	155934; Xiao et al., 2005, 156164; Zhao et al., 2006, 100996) and the production of
nitric oxide and other reactive nitrogen species (RNS).
Although all size fractions of PM may contribute to oxidative and nitrosative stress,
ultrafine PM may contribute disproportionately to their mass due to their large
surface/volume ratio. The relative enrichment of redox active surface components such as
metals and organics per unit mass may translate to a relatively greater oxidative potential
of ultrafine PM compared with larger particles with similar surface components. In
addition, the greater surface per unit volume could potentially deliver relatively more
adsorbed soluble components to cells. These components may undergo intracellular redox
cycling following cellular uptake. Furthermore, per unit mass, ultrafine PM may have more
opportunity to interact with cell surfaces due to their greater surface area and their greater
particle number compared with larger PM. These interactions with cell surfaces can lead to
ROS generation, as described above. Recent studies have also demonstrated that ultrafine
PM have the capacity to cross cellular membranes by non-endocytotic mechanisms
involving adhesive interactions and diffusion (Geiser et al., 2005, 087362), as described in
Chapter 4. This may allow ultrafine PM to interact with or penetrate intracellular
organelles.
July 2009
5-3
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
Cellular Sources
of ROS/RNS
ROS/RNS Assay
Oxidation of Cellular Components
HO-1 Induction
NADPH
Oxidase
Mitochondrial
Electron Transport
Soluble metals
Nitric Oxide
Synthase
Microsomal
Metabolism
(PAH/Quinones)
Binding to Cell Surfaces
Phagocytosis
Cytoskeletal Interactions
Iron Sequestration
Redox Cycling
Altered iron homeostasis
Figure 5-2. PM stimulates pulmonary cells to produce ROS/RNS.
In general, high levels of intracellular ROS/RNS can lead to irreversible protein
modifications, loss of cellular membrane integrity, DNA damage and cellular toxicity. Lower
levels of ROS/RNS may cause reversible protein modifications that trigger intracellular
signaling pathways and/or adaptive responses. Thus PM-dependent generation of ROS may
be responsible for a continuum of responses from cell signaling to cellular injury.
5.1.2. Activation of Cell Signaling Pathways
Activation of cell signaling pathways by ROS/RNS has received increasing attention
in recent years. An early example was provided by Kaul and Forman (1996, 155892) who
demonstrated that respiratory burst-derived H2O2 activates the transcription factor nuclear
factor k a | )| ia -1 i gh l"ch a i ire n h a nee r of activated B cells (NFkB). Numerous studies since then
have demonstrated that PM, which serves as both a direct and indirect source of ROS/RNS,
activates cell signaling pathways by this mechanism.
July 2009
5-4
DRAFT-DO NOT CITE OR QUOTE

-------
Cell Surface
Interactions
PM
Endotoxin
Zinc
PAH
Initiation
Initiation
Amplification
Cell Signaling/
Transcription
Factor Activation
Cytokines
Chemokines
Proteases
Mediators
ROS/RNS
Cell Injury
Influx of Leukocytes
Pulmonary Inflammation
Figure 5-3. PM activates cell signaling pathways leading to pulmonary inflammation.
1	PM also has the potential to activate cell signaling by mechanisms that are
2	independent of ROS/RNS. For example, PM delivers water-soluble components such as
3	endotoxin and zinc to cell surfaces. Endotoxin binds to toll-like receptors on alveolar
4	macrophages and other cells, resulting in the up regulation of cytokines (Becker et al., 2002,
5	052419). Zinc, a transition metal which does not redox cycle, inhibits protein tyrosine
6	phophatases in airway epithelial cells resulting in a cascade of cell signaling events (Tal et
7	al., 2006, 108588). Similarly, PM-mediated delivery of lipid soluble components such as
8	PAH results in binding and activation of the arylhydrocarbon receptor (AhR). AhR is a
9	transcription factor responsible for the upregulation of CYP1A1, a cytochrome oxidase
July 2009	5-5	DRAFT - DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
involved in PAH biotransformation to metabolites capable of forming DNA adducts or
eliciting oxidative stress responses (Rouse et al., 2008, 156930). In addition, interaction of
PM with cell surfaces may activate cell signaling by perturbation of the cytoskeleton,
adherence, internalization, or receptor-mediated pathways.
Ultrafine PM may activate cell signaling by ROS-dependent and -independent
mechanisms disproportionately to their mass compared with larger sized particles for
reasons described in Section 5.1.1.
Recent studies involving PM exposures have focused on intracellular pathways
involving protein kinases such as mitogen-activated protein kinase (MAPK) (Bayram et al.,
2006, 088439; Lee et al., 2005, 156682; Roberts et al., 2003, 156051; Soberanes et al., 2009,
190483), AKT (Ahsan et al., 2005, 156200), src (Cao et al., 2007, 156322) and epidermal
growth factor receptor (Blanchet et al., 2004, 087982; Cao et al., 2007, 156322; Tamaoki et
al., 2004, 157040); as well as ras (Tamaoki et al., 2004, 157010), toll-like receptors (Becker
et al., 2002, 052419; Becker et al., 2005, 088590; U.S. EPA, 2000, 052149), protein tyrosine
phosphatases (Tal et al., 2006, 108588) phospholipases A2 (Lee et al., 2003, 156678),
calcium (Agopyan et al., 2003, 155649; Brown et al., 2004, 155705; Brown et al., 2004,
088663; Geng et al., 2005, 096689; 2006, 097026; Sakamoto et al., 2007, 096282), caspases
(Soberanes et al., 2006, 156991; Zhang et al., 2007, 156179), poly (ADP-ribose) polymerase
family member 1 (PARP-l) (Zhang et al., 2007, 156179) and histone acetylation (Gilmour et
al., 2003, 096959). The transcription factors regulated by these pathways, including NFkB
(Bayram et al., 2006, 088439; Lee et al., 2005, 156682; Takizawa et al., 2003, 157039),
activator protein 1 (AP-l) (Donaldson et al., 2003, 156408), signal transducers and
activators of transcription protein (STAT) (Cao et al., 2007, 156322), antioxidant response
element (ARE) (Li and Nel, 2006, 156694), and AhR (Rouse et al., 2008, 156930) have also
been studied following PM exposures. Activation of these intracellular pathways and
transcription factors leads to the upregulation of genes responsible for inflammatory,
immune and acute phase responses as well as genes responsible for antioxidant defense and
xenobiotic metabolism.
5.1.3. Pulmonary Inflammation
Following PM exposure, transcription factor activation in macrophages and epithelial
cells stimulates the synthesis and release of proteins involved in inflammatory and immune
responses including cytokines, chemokines, proteases and eicosanoids (Figure 5-3). These
soluble mediators play a role in recruiting inflammatory cells such as neutrophils,
monocytes, mast cells and eosinophils to the lung. Interactions between macrophages and
epithelial cells enhance these responses (Tao and Kobzik, 2002, 157044).
Inflammatory cells can serve as a source of extracellular ROS which, along with
soluble mediators derived from the inflammatory cells, amplify the inflammatory response.
July 2009
5-6
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Unchecked inflammation may cause cellular and tissue injury through the generation of
ROS and soluble mediators. In some cases the oxidative potential of PM is well-correlated
with the degree of inflammation (Dick et al., 2003, 036605), suggesting that the
inflammation is a direct consequence of PM-generated ROS. However, in other cases the
oxidative potential of PM is not well-correlated with the degree of inflammation (Beck-
Speier et al., 2005, 156262), suggesting that the inflammation is a consequence of the other
mechanisms by which PM activates intracellular signaling pathways.
Particle surface area has been identified as a key determinant of the extent of
inflammation in the case of low-toxicity, low-solubility particles (Donaldson et al., 2008,
190217). Given their large surface/volume ratio compared with other PM fractions,
ultrafine PM may cause inflammation disproportionately to their mass compared with PM
of larger sizes.
PM exposure often results in neutrophilic inflammation in laboratory studies (Tao et
al., 2003, 156111). Neutrophilic inflammation is also associated with acute lung injury in
humans as well as chronic lung diseases such as COPD and certain forms of asthma
(Barnes, 2007, 191139; Cowburn et al., 2008, 1911 12). Circulating neutrophils respond to
chemotactic factors in the lung such as leukotriene B4 and IL-8 (Barnes, 2007, 191139).
They migrate into the lung parenchyma across the pulmonary capillary network and into
the airways from the bronchial circulation (Cowburn et al., 2008, 1911 12). As a consequence
of priming by inflammatory mediators or contact with extracellular matrix components,
neutrophils become hyperresponsive to activating signals and insensitive to chemotactic
signals (Cowburn et al., 2008, 1911 12). Activation results in neutrophil degranulation,
respiratory burst responses and soluble mediator release (Cowburn et al., 2008, 191112).
Neutrophils eventually undergo apoptosis and are phagocytized by inflammatory
macrophages (Cowburn et al., 2008, 1911 12). This is accompanied by the release of anti-
inflammatory mediators such as IL-10 and transforming growth factor-B (TGF- 6) (Cowburn
et al., 2008, 191112). These steps are key to the resolution of inflammation and prevent
unregulated release of toxic neutrophil products such as neutrophil elastase (Cowburn et
al., 2008, 1911 12). Thus, circumstances leading to the decreased ingestion of apoptotic
neutrophils by macrophages may lead to tissue injury. Impairment of macrophage function
may serve as an important mechanism by which PM contributes to disease.
5.1.4. Respiratory Tract Barrier Function
Epithelial injury can lead to an increase in permeability of the airway epithelial and
alveolar-capillary barriers (Braude et al., 1986, 155701). Enhanced transport of soluble and
possibly of insoluble PM components into the circulation may occur under these conditions.
Increased epithelial permeability is also associated with enhanced immune responses to
proteins, including allergens, on the epithelial surface, presumably due to the greater
July 2009
5-7
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
availability of antigens to underlying immune cells (Wan et al., 1999, 191903).
Furthermore, endothelial injury can compromise the integrity of the the alveolar-capillary
barrier resulting in transvascular fluid and solute flux (Braude et al., 1986, 155701).
Soluble mediators derived from inflammatory and lung cells (Chang et al., 2005, 097776)
and peptides released by some nerve cells (Widdicombe and Lee, 2001, 019049) can increase
the permeability of the alveolar-capillary barrier and result in alveolar edema.
Compromised barrier function in the airways may lead to airway edema. Edema occurring
secondarily to nerve cell stimulation is one component of the process termed neurogenic
inflammation.
Given the small size of ultrafine PM, modest changes in epithelial permeability may
particularly affect the disposition of this fraction. Enhanced translocation to interstitial
compartments or to the circulation may be important sequelae. A recent study described in
Section 4.3.4.1 demonstrated greater translocation of ultrafine compared with PIVL.r, into
the circulation of rodents treated with endotoxin to induce acute lung injury prior to
intratracheal instillation of PM (Chen et al., 2006, 147267). Furthermore, epithelial injury
in another model resulted in greater translocation of ultrafine PM into the interstitial
compartment (Adamson and Prieditis, 1995, 189982).
5.1.5. Antioxidant Defenses and Adaptive Responses
Antioxidant defenses and adaptive responses are important modulators of oxidative
stress and other cellular stresses resulting from PM exposure. Antioxidants are present in
the epithelial lining fluid in all regions of the respiratory tract. In addition, they are present
in cells of the lung parenchyma and inflammatory cells found in airways and alveoli. Some
antioxidants act directly against oxidant species (e.g., glutathione, ascorbate, superoxide
dismutase) while others act indirectly (e.g., gamma-glutamylcysteine sythetase [yGCS],
glutathione reductase). Furthermore, some antioxidants (e.g., Phase 2 enzymes heme
oxygenase-1 [HO-l], NADPH quinone oxidoreductase 1 [NQOl], glutathione-S-transferase
[GST]) are inducible via activation of the nuclear factor (erythroid-derived 2)- related factor
2 (Nrf2)-ARE pathway, which occurs as an adaptive response to stress (Cho et al., 2006,
156345; Li and Nel, 2006, 156694). Antioxidants play an important role in reducing the
oxidative potential of those PM species that directly generate ROS. They also inhibit
responses due to generation of intracellular ROS.
Recently a three-tier response to oxidative stress was proposed (Li and Nel, 2006,
156694). In this scheme, mild oxidative stress enhances antioxidant defenses by
upregulating Phase 2 and other antioxidant enzymes (Tier l). Further increase in oxidative
stress induces inflammation (Tier 2) and cell death (Tier 3). Experimental evidence is
supportive of this scheme. Numerous studies have demonstrated that enhancement of lung
and cellular antioxidant defenses inhibits inflammation, cytotoxicity and other responses
July 2009
5-8
DRAFT-DO NOT CITE OR QUOTE

-------
1	following exposure to PM (Ahsan et al., 2005, 156200; Bachoual et al., 2007, 155667;
2	Bayram et al., 2006, 088439; Chang et al., 2005, 097776; Imrich et al., 2007, 155859; Koike
3	and Kobayashi, 2005, 088303; Koike et al., 2004, 058555; Li et al., 2007, 155929; Ramage
4	and Guy, 2004, 055640; Rhoden et al., 2004, 087969; Steerenberg et al., 2004, 087981;
5	Takizawa et al., 2003, 157039; Tao et al., 2003, 156111; Upadhyay et al., 2003, 097370; Wan
6	and Diaz-Sanchez, 2006, 097399; Wan and Diaz-Sanchez, 2007, 156145; Yin et al., 2004,
7	087983).
PM Core and
Soluble
Components
Irritant
Receptors
ROS/RNS
Pulmonary
Inflammation
and Injury
Altered
Lung Function
Allergic Asthma
And Other
Allergic
Disorders
AHR and
Airway
Remodeling
Impaired
Host Defense
and Infections
DNA Damage
and
Lung
Cancer
Progression of
Pre-existing
Lung
Disease
Death or Hospitalization for Asthma, Pneumonia, COPD and Lung Cancer
Figure 5-4. Potential pathways for the effects of PM on the respiratory system.
8	Cellular and tissue exposure to xenobiotics carried by PM can lead to induction of
9	Phase 1 and Phase 2 detoxifying enzymes following the activation of cell signaling
10	pathways and transcription factors AhR and ARE, respectively (Rengasamy et al., 2003,
11	156907; Rouse et al., 2008, 156930; Zhao et al., 2006, 100996).
July 2009
5-9
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
5.1.6.	Pulmonary Function
PM exposure may alter pulmonary function by a variety of different mechanisms
(Figure 5-4). In the short term, airway hyperresponsiveness (AHR) may ensue due to the
influence of inflammatory mediators. In the long-term, morphological changes may occur, in
some cases leading to mucus hypersecretion and airway remodeling. Activation of irritant
receptors and stimulation of the ANS in the respiratory tract is another mechanism by
which PM exposure may alter pulmonary function (Section 5.3).
5.1.7.	Allergic Disorders
PM exposure sometimes leads to the development of allergic immune responses
(Figure 5-4). These responses are predominately mediated by T helper 2 cells (Th2). Thl
responses, characterized by IFN-y and classical macrophage activation, are inflammatory!
in excess they can lead to tissue damage. Alternatively, Th2 responses are associated with
allergy and asthma and are characterized by IL-4, IL-5, IL-13, influx of eosinophils, 6—
lymphocyte production of IgE, and alternative macrophage activation. PM exposure can
also lead to the exacerbation of airway allergic responses, such as antigen-specific IgE
production and AHR.
Due to soluble mediators and immune cell trafficking, pulmonary exposure may result
in systemic immune alterations. Not only do macrophages ingest PM, but they are also
antigen presenting cells whose level of activation dictates costimulation and thus
subsequent T cell responses. These cells are highly mobile, and can transport PM to other
sites such as lymph nodes. Dendritic cells (DC) also play a key role as antigen presenting
cells and in modulating T and B cell activity. A cell culture model of the human epithelial
airway wall was used to demonstrate that DC extended processes between epithelial cells
through the tight junctions, collected particles in the lumenal space and transported them
across the epithelium (Blank et al., 2007, 096521). DC also transmigrated through the
epithelium to take up particles on the epithelial surface (Blank et al., 2007, 096521).
Furthermore, DC interacted with particle-loaded macrophages on top of the epithelium and
with other DC within or beneath the epithelium to transfer particles (Blank et al., 2007,
096521). In vitro studies also demonstrated that the adjuvant activity of diesel exhaust
particles (DEPs) involved stimulation of immature monocyte-derived dendritic cells
(iMDDC) to undergo maturation in response to an altered airway epithelial cell-derived
microenvironment (Bleck et al., 2006, 096560). Additionally, DEP directly influenced the
profile of cytokines secreted by DC and caused a predisposition toward Th2-mediated or
allergic responses (Chan et al., 2006, 097468). Thus PM can negatively affect both innate
immunity through effects on macrophage pathogen handling (see Section 5.1.8) as well as
adaptive immunity by altering macrophage or DC antigen presenting activity and
subsequent T cell responses.
July 2009
5-10
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Given their larger surface area and particle number per unit mass as well as their
propensity for trans-epithelial movement, ultrafine PM may have a disproportionate ability
to enhance allergic sensitization compared with particles of larger size.
5.1.8.	Impaired Lung Defense Mechanisms
PM exposure may impair lung defense mechanisms and result in frequent or
persistent infections (Figure 5-4). Potential targets include mucociliary transport,
surfactant function and pathogen clearance. Pathogen clearance is dependent on the
integrity of macrophages and their migration, phagocytosis and respiratory burst functions.
PM-mediated cytotoxicity of macrophages with the concomitant release of lysosomal
contents may affect pathogen clearance and cause damage to nearby cells and tissues.
Intratracheal instillation and cell culture experiments have demonstrated PM-dependent
impairment of lung defense mechanisms (Jaspers et al., 2005, 088115! Kaan and Hegele,
2003, 095753; Long et al., 2005, 087454; Moller et al., 2005, 156770; Monn et al., 2003,
052418; Roberts et al., 2007, 097623; Yin et al., 2004, 087983).
5.1.9.	Resolution of InflammationlProgression or
Exacerbation of Disease
Resolution of pulmonary inflammation and injury has been demonstrated in many
experimental models using higher than ambient concentrations of PM. Factors contributing
to this complex process are likely to include the uptake and clearance of PM by
macrophages; the retention of PM in parenchymal cells and tissues! the balance of
pro/anti-inflammatory soluble mediators, oxidants/antioxidants and proteases/anti-
proteases! and the presence of pre-existing disease. These factors may also influence the
resolution of pulmonary responses to ambient PM exposures (Figure 5-4). The long-term
consequences of prolonged inflammation are not likely to be beneficial and may lead to
remodeling of the respiratory tract and to the progression or exacerbation of disease.
5.1.9.1. Factors Affecting the Retention of PM
Clearance of poorly soluble particles from ET, TB and alveolar regions is extensively
discussed in Section 4.3. While clearance from ET and TB regions generally occurs over
hours to days, clearance from the alveolar compartment is much slower, occurring over
months to years depending on the species. Phagocytosis by alveolar macrophages and
transport by the mucociliary escalator is the primary mechanism of clearance from the
alveolar compartment although neutrophil phagocytosis also plays a role (Snipes et al.,
1997, 156092). Pre-existing disease can alter the extent and localization of PM deposition as
discussed in Section 4.2.4.5 and 4.2.5. In addition, mechanisms of clearance may be altered
July 2009
5-11
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
in cases of pre-existing disease as discussed in Section 4.3.4.3. While mild asthma was
associated with enhanced mucociliary clearance, acute lung injury was associated with
enhanced translocation to the circulation and the interstitial compartment. Whether
retained particles are localized in alveolar macrophages or parenchymal tissue also differs
according to species (Snipes, 1996, 076041).
Ultrafine PM may have a special propensity for retention given the decreased
efficiency of alveolar macrophages for phagocytosis of this size particle (Oberdorster, 1988,
006857) and the demonstration that ultrafine PM can readily cross cellular membranes
(Geiser et al., 2005, 087362). Some studies suggest that ultrafine PM are taken up by
epithelial cells and move to the interstitum where they are cleared by other pathways
(Semmler-Behnke et al., 2007, 156080); however clearance mechanisms are not entirely
understood.
Enhanced deposition of particles in "hot spots" may influence retention. For example,
deposition in the centriacinar region or proximal alveolar region, where clearance is slow,
may result in accumulated particle dose in this region and the potential for prolonged
inflammation at the site leading to the development of pulmonary fibrosis or emphysema
(Donaldson et al., 2008, 190217). A recent study suggests an important role for retained
particles in the progression of disease. Complexation of endogenous iron by retained
particles resulted in retained particles growing larger over time. The authors suggested
that redox cycling of complexed iron may be responsible for disease progression (Ghio and
Cohen, 2005, 088272; Ghio et al., 2004, 155790).
5.1.9.2. Factors Affecting the Balance of Pro/Anti-Inflammatory Mediators,
Oxidants/Anti-Oxidants and Proteases/Anti-Proteases
Inflammation can be enhanced by pro-inflammatory mediators or dampened by anti-
inflammatory mediators. Production of anti-inflammatory mediators normally occurs at
several steps of the inflammation pathway, such as the release of IL-8 and TGF- 6 during
phagocytosis of apoptotic neutrophils by macrophages (Cowburn et al., 2008, 191112).
Dsyregulation of the inflammatory process may prevent the resolution of inflammation. PM
exposure may result in the production of pro-inflammatory mediators as well as decrease
the production of anti*inflammatory mediators by impairing macrophages function.
An unfavorable balance of oxidants to antioxidants in the lung is associated with
inflammatory lung diseases including asthma and COPD (Rahman et al., 2006, 191165).
PM is likely to contribute to an unfavorable balance through its oxidative potential and
capacity to promote cellular production of ROS. Exacerbations of asthma and COPD
resulting from bacterial and viral infections are also associated with increased oxidative
stress (Barnes, 2007, 191139). In addition, antioxidants may reduce neutrophilic
inflammation associated with oxidative stress (Barnes, 2007, 191139).
July 2009
5-12
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Protease/anti-protease balance has long been tied to the pathogenesis of emphysema
and other forms of COPD (Owen, 2008, 191162). Key steps include the release of
proteinases by inflammatory cells which degrade the extracellular matrix components of
alveolar walls. Destruction of alveolar walls and airspace enlargement ensues. Endogenous
anti-protease defenses in the lung modulate this response but may be insufficient to
prevent it during prolonged inflammation. Proteases also play a role in airways pathologies
which lead to small airway fibrosis (Owen, 2008, 191162). Oxidative stress has been linked
to both activation of proteases and inactivation of anti-proteases (Owen, 2008, 191162). PM
may contribute to an unfavorable protease/anti-protease balance through the generation of
ROS.
Although there are numerous inflammatory cell-derived proteases and lung anti-
proteases, many recent studies have focused on matrix metalloproteinases (MMPs). MMPs
are a family of zinc-containing enzymes normally found in an inactive pro-enzyme form.
Activation involves proteolytic cleavage or oxidation of the "cysteine switch" (Pardo and
Selman, 2005, 191163). Inhibitors include tissue inhibitors of metalloproteinases (TIMPs)
(Pardo and Selman, 2005, 191163). In particular, MMP-1 is well-studied and found to play
an important role in physiological processes such as development and wound repair as well
as in diseases such as pulmonary emphysema, fibrosis, asthma and bronchial carcinoma (Li
et al., 2009, 190424; Pardo and Selman, 2005, 191163). In addition to its activity in
degrading collagenase, MMP-1 also acts on non-matrix substrates and cell surface
molecules suggesting that it may influence cell signaling (Pardo and Selman, 2005, 191163).
MMP-2 and MMP-9 are thought to be involved in the pathogenesis of disease through their
gelatinase activity. Interestingly, recent in vitro studies have demonstrated upregulation of
MMP-1 by hydrogen peroxide, cigarette smoke and DEPs (Amara et al., 2007, 156212; Li et
al., 2009, 190424; Mercer et al., 2004, 191180), and up-regulation of MMP-12 following
instillation of PM collected from the Paris subway (Bachoual et al., 2007, 155667). These
considerations suggest that particulate air pollution may act via MMP-mediated
progression or exacerbation of lung disease.
5.1.9.3. Pre-Existing Disease
In addition to its effects on deposition, retention and clearance of PM described above,
pre-existing disease may also affect the balance of the afore-mentioned factors. For
example, acute exacerbations of COPD are characterized by a rapid influx of neutrophils
into the airways (Owen, 2008, 191162). However, clearance of apoptotic neutrophils by
macrophages is impaired in COPD leading to greater release of neutrophil-derived
inflammatory mediators, oxidants and proteases (Owen, 2008, 191162). Thus exacerbation
of disease may occur as a result of unchecked inflammation.
July 2009
5-13
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
5.1.10.	Pulmonary DNA Damage
Pulmonary DNA damage can occur primarily or secondarily to PM exposure. Primary
effects include oxidative DNA injury or DNA adduct formation due directly to PM while
secondary effects occur due to PM-mediated inflammation (Gallagher et al., 2003, 140171;
Gabelova et al., 2007, 156457; Schins and Knaapen, 2007, 156074; de Kok et al., 2005,
189835). These responses may lead to chromosomal aberrations or DNA strand breaks. PM
effects on cell cycle arrest, proliferation, apoptosis, and DNA repair mechanisms may also
influence the genotoxic, mutagenic or carcinogenic potential of DNA damage as reviewed by
Schins et al. (2007, 156074).
5.1.11.Epigenetic	Changes
Epigenetic mechanisms regulate the transcription of genes without altering the
nucleotide sequence of DNA. These mechanisms generally involve DNA methylation and
histone modifications, leading to alterations which may have long-term consequences or are
heritable (Jones et al., 2007, 156615; Keverne and Curley, 2008, 1!) 1151) . DNA methylation
and histone modifications, which include methylation, acetylation, phosphorylation,
ubiquitylation and sumoylation, are known to be linked (Hitchler and Domann, 2007,
191151; Jones and Baylin, 2007, 191153). Numerous studies have identified epigenetic
processes in the control of cancer (Foley et al., 2009, 191144; Gopalakrishnan et al., 2008,
1911 17• Jones and Baylin, 2007, 191153; Valinluck et al., 2004, 191170). embryonic
development (Foley et al., 2009, 191144; Gopalakrishnan et al., 2008, 191147; Keverne and
Curley, 2008, 191154) and inflammation and other immune system functions (Adcock et al.,
2007, 191178).
Epigenetic modifications resulting in decreased expression of tumor suppressor genes
and increased expression of transforming genes have been observed in human tumors
(Valinluck et al., 2004, 191170). In general, transcription repression is associated with DNA
methylation in promoter regions of genes. Cytosine methylation in CpG dinucleotides has
emerged as an important, heritable epigenetic modification which can result in chromatin
remodeling and decreased gene expression (Valinluck et al., 2004, 191170). Global changes
in DNA methylation are also seen in cancer and hypomethylation is associated with
genomic instability (Gopalakrishnan et al., 2008, 191147).
Embryonic development is characterized by several phases of epigenetic
modifications. DNA methylation is very dynamic following fertilization, with demethylation
and re-methylation of egg and sperm genomes occurring immediately (Foley et al., 2009,
191144). Imprinted genes, however, retain the methylation profile of the parent of origin
(Foley et al., 2009, 191144). Epigenetic changes accumulated through a life course may be
passed from parent to offspring in the germline (i.e germline transmission of epimutation) if
they survive the epigenetic remodeling that occurs during gametogenesis and early
July 2009
5-14
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
embryogenesis (Foley et al., 2009, 1i)1 1 1 1). Early development is characterized by the
process of cell differentiation, which produces different cell types and involves the selective
activation of some sets of genes and the silencing of others in a temporal pattern (Foley et
al., 2009, 191144; Gopalakrishnan et al., 2008, 191117). DNA methylation is postulated to
provide a basis for cell differentiation (Gopalakrishnan et al., 2008, 191117).
In the lung, histone acetylation and methylation have been linked to inflammatory
gene expression, T cell differentiation, and the regulation of macrophage function following
pathogen challenge (Adcock et al., 2007, 191178). Furthermore, altered patterns of
methylation and acetylation have been reported in inflammatory diseases (Adcock et al.,
2007, 191178). Histone deacetylase has been identified as a potential therapeutic target for
epigenetic therapy (Adcock et al., 2007, 191178; Jones and Baylin, 2007, 191153). Reduced
expression and activity of histone deacetylase have been demonstrated in lung and
inflammatory cells in COPD and asthma (Adcock et al., 2007, 191178; Barnes, 2007,
191139).
Epigenetic mechanisms have been identified as potential targets for gene-
environment interactions and recent studies have demonstrated that diet, cigarette
smoking, endocrine disrupters, heavy metals and bacterial infection can alter the epigenetic
profile in animals and humans (Foley et al., 2009, 191 111). A role for PM in promoting
epigenetic changes has been proposed and new studies, discussed in later chapters, provide
some evidence for this pathway (Baccarelli et al., 2009, 188183; Liu et al., 2008, 156709;
Reed et al., 2008, 156903; Tarantini et al., 2009, 192010; Tarantini et al., 2009, 192153;
Yauk et al., 2008, 157164). Early life exposures may be especially important in this regard
since periods of rapid cell division and epigenetic remodeling are likely to occur at this time
(Foley et al., 2009, 191144; Keverne and Curley, 2008, 191958; Wright and Baccarelli, 2007,
191173). This may provide a basis for fetal origins of adult disease.
It has been suggested that DNA methylation is regulated by oxygen gradients and
redox status (Hitchler and Domann, 2007, 191151). While this is of particular importance
during development where oxygen gradients and redox status are linked to cellular
differentiation, these processes are also important for cell signaling during all stages of life.
A common metabolic precursor for both methylation reactions and glutathione availability
(involved in redox status) is methionine (Hitchler and Domann, 2007, 191151). Methionine
availability regulates the cell's ability to generate S-adenosyl methionine which is directly
involved in DNA and histone methylation and the cell's ability to generate
homocysteine/cysteine which is involved in glutathione biosynthesis (Hitchler and Domann,
2007, 191151). Furthermore the folate cycle is a key determinant of methionine
bioavailability (Hitchler and Domann, 2007, 191151). In this way cellular intermediary
metabolism is linked to epigenetic processes, with oxidative stress necessitating a metabolic
shift resulting in decreased DNA methylation and increased glutathione production.
July 2009
5-15
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
5.1.12. Lung Development
Lung development is a multi-step process which begins in embryogenesis and
continues to adult life (Pinkerton and Joad, 2006, 091237). This allows for a long period of
potential vulnerability to environmental and other stressors. Furthermore, enzymatic
systems responsible for detoxification of xenobiotic compounds are not fully developed until
the postnatal period (Pinkerton and Joad, 2006, 091237). Disruption of cell signaling during
development could affect cellular differentiation, branching morphogenesis and overall lung
growth, possibly leading to life-long consequences. Although very little is known about the
effects of maternal exposure to PM on the fetus or the effects of exposure during childhood,
recent animal studies demonstrate respiratory and immune system effects of perinatal
exposure to sidestream cigarette smoke (Pinkerton and Joad, 2006, 091237; Wang and
Pinkerton, 2007, 179975).
5.2. Systemic Inflammation
Pulmonary inflammation resulting from PM exposure may trigger systemic
inflammation through the action of cytokines and other soluble mediators which leave the
lung and enter the circulation (Figure 5-5). Epithelial permeability may exert an important
influence on this process (see Section 5.1.4.) Cytokines released by alveolar macrophages
can stimulate bone marrow production of leukocytes resulting in an increased number of
total and immature leukocytes in the circulation (Van Eeden and Hogg, 2002, 088111; Van
Eeden et al., 2001, 019018). They also activate neutrophils and promote their sequestration
in microvascular beds (Van Eeden et al., 2001, 019018). The time course of these responses
varies according to the acute or chronic nature of the PM exposure (van Eeden et al., 2005,
157086).
July 2009
5-16
DRAFT-DO NOT CITE OR QUOTE

-------
Ambient PM
Stroke
Thrombosis
Atherosclerosis
Arrhythmia
Myocardial
Infarction
Myocardial
Ischemia
Pro-
Coagulation
Effects
Autonomic
Nervous System
Death or Hospitalization
for Stroke
Liver
Acute
Phase
Response
Altered
Conduction/
Repolarization
Pulmonary Reflexes
Other Reflexes ?
Endothelial
Cell
Activation/
Dysfunction
Plaque Destabilization
Or Rupture
Systemic
Inflammation/
Oxidative
Stress
Altered
Vasoreactivity
of Coronary
Vessels
Death or
Hospitalization
for Thrombo-
Embolic Disease
Pulmonary
Oxidative Stress
and Inflammation
Altered Sympathetic/
Parasympathetic
Tone
Death or Hospitalization for Coronary
Heart Disease or Congestive Heart Failure
Direct Effects ?
T ranslocation/Absorption
of Soluble Components
Figure 5-5. Potential pathways for the effects of PM on the cardiovascular system.
1	Systemic inflammation is seen under conditions of mild pulmonary inflammation -
2	and sometimes under conditions of no measurable pulmonary inflammation - following PM
3	exposure. The time-dependent nature of pulmonary and systemic inflammatory responses
July 2009
5-17
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
may in part explain these findings since biomarkers of inflammation are frequently
measured only at one time point. Furthermore, chronic exposures may lead to adaptive
responses. In general, systemic inflammation is associated with changes in circulating
white blood cells, the acute phase response, pro-coagulation effects, endothelial dysfunction
and the development of atherosclerosis (Figure 5-5). Adverse effects on the cardiovascular
and cerebrovascular systems such as thrombosis, plaque rupture, MI and stroke may
result. Systemic inflammation may affect other organ systems such as the liver or the CNS.
One recent study demonstrated that alveolar macrophage-derived IL-6 mediated
pro-coagulation effects in mice exposed by intratracheal instillation to PMio (Mutlu et al.,
2007, 121441). This study provides a clear link between lung cytokines and systemic
responses in one model system. Whether this mechanism or others account for the majority
of extra-pulmonary effects following PM exposure is not yet known.
5.2.1. Endothelial Dysfunction and Altered Vasoreactivity
The lumenal surface of blood vessels is lined by endothelial cells which, in addition to
providing a barrier function, are key regulators of vascular homeostasis. Endothelial cells
synthesize and release vasodilators such as nitric oxide (NO) and prostacyclin and
vasoconstrictors such as endothelin (ET), which act on neighboring smooth muscle cells. ET
also stimulates endothelial NO synthesis through a feedback mechanism. Inhalation of
high concentrations of PM has been reported to increase ET levels in the circulation
(Thomson et al., 2005, 087554). ET has also been proposed to play a role in hypoxia-induced
MI (Caligiuri et al., 1999, 156318). However the role of ET in mediating cardiovascular
effects following ambient PM exposures is unclear.
Endothelial dysfunction can lead to or follow endothelial activation under conditions
of systemic inflammation and/or vascular oxidative stress. Both systemic inflammation and
vascular oxidative stress have been associated with PM exposure (Nurkiewicz et al., 2006,
088611; Van Eeden et al., 2001, 019018). Cytokines activate endothelial cells and
upregulate endothelial cell adhesion molecules. They also promote the sequestration of
neutrophils in microvascular beds. Neutrophil sequestration is sometimes associated with
the deposition of myeloperoxidase (MPO) on endothelial cell surfaces (Nurkiewicz et al.,
2006, 088611). ROS"derived from neutrophils, MPO, other adhered inflammatory cells
and/or other sources can perturb the balance of vasodilator and vasoconstrictor species
produced by endothelial cells. Oxidative stress can result in decreased synthesis of NO due
to limitation of the redox-sensitive essential cofactor tetrahydrobiopterin and in decreased
bioavailability of NO due to reaction with superoxide. Prostacyclin synthesis is also
decreased by oxidative stress. Importantly, these processes can affect vasoreactivity such
that blood vessels may be unable to respond to vasoconstrictor stimuli with compensatory
vasodilation.
July 2009
5-18
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Loss of NO and prostacyclin synthesis due to PM-dependent vascular oxidative stress
may have other consequences since both exert negative influences on platelet and
neutrophil activation. While endothelial surfaces normally are antithrombotic, endothelial
dysfunction can contribute to thrombus formation. Furthermore inflammation and
oxidative stress associated with endothelial dysfunction can contribute to the development
or progression of atherosclerosis (van Eeden et al., 2005, 157086).
5.2.2. Activation of Coagulation and Acute Phase Response
The primary function of the coagulation cascade is to stop the loss of blood after
vascular injury by forming a fibrin clot. However in some cases, activation of coagulation
can promote intravascular thrombosis (Karoly et al., 2007, 155890). It has been proposed
that air pollution-associated PM can activate clotting pathways and enhance the likelihood
of an obstructive cardiac ischemic event (e.g., MI) or cerebral event (e.g., stroke) (Seaton et
al., 1995, 045721).
Coagulation is regulated by intrinsic and extrinsic pathways. The intrinsic pathway
occurs following activation of Factor XII and does not require the addition of an exogenous
agent (Mackman, 2005, 156722). On the other hand, the extrinsic pathway is an inducible
signaling cascade that can be activated by tissue factor (TF) produced in response to
inflammation or endothelial injury (Karoly et al., 2007, 155890).
In general, platelets, red blood cells (RBCs) and endothelial cells are effector cells for
inducing a pro-coagulant state in the vasculature. Circulating factors may enhance
coagulation or promote activation of platelets. Cytokines formed during tissue damage and
inflammation lead to TF induction. TF is the initiating stimulus for coagulation following
vascular injury or plaque erosion. Complexes of TF^Factor Vila form on endothelial cell
surfaces and play a key role in thrombin generation by initiating the extrinsic blood
coagulation pathway (Gilmour et al., 2005, 087410). Thrombin generates fibrin from
fibrinogen and amplifies the intrinsic pathway (Karoly et al., 2007, 155890). TF and
thrombin also have pro-inflammatory actions independent of coagulation functions (Chu,
2005, 155730); thus activation of coagulation may lead to or potentiate inflammation, von
Willebrand factor, produced by endothelial cells, also plays an important role in
coagulation.
The fibrinolytic system opposes these processes by facilitating the removal of a clot.
The fibrinolytic pathway is regulated by the ratio of tissue plasminogen activator (tPA) and
plasminogen activator inhibitor (PAI). Furthermore, the endothelial cell surface has
antithrombotic properties due to the expression of tissue factor pathway inhibitor (TFPI)
and thrombomodulin (Mackman, 2005, 156722).
Inhibition of the fibrinolytic pathway, along with increased plasma viscosity, plasma
fibrinogen and Factor VII concentrations, contributes to a pro-thrombotic state (Gilmour et
July 2009
5-19
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
al., 2005, 087410). In acute lung injury, vascular cells have enhanced procoagulant activity
and impaired fibrinolytic activity (Gilmour et al., 2005, 087410). In arterial atherosclerosis,
TF expression is increased within plaques. As a result, spontaneous plaque rupture may
trigger intravascular clotting (Karoly et al., 2007, 155890).
Acute phase responses also play a role in hemostasis by exerting procoagulant effects.
Cytokines such as IL-6 stimulate the liver to produce acute phase proteins including O
reactive protein (CRP), fibrinogen and antiproteases (van Eeden et al., 2005, 157086). To
date, there is limited evidence supporting a role for ambient PM in stimulating acute phase
responses (Ruckerl et al., 2007, 156931 and reviewed therein! see also Sections 6.2.7.3 and
6.2.8.3).
5.2.3. Atherosclerosis
Atherosclerosis is a chronic progressive disease which contributes greatly to
cardiovascular disease (Libby, 2002, 192009). Mainly a disease of the large arteries, it is
characterized by the accumulation of lipid and fibrous tissue in atheromas, or swellings of
the vessel wall (Libby, 2002, 192009). Although a strong link is known to exists between
hypercholesterolemia and atherogenesis, recent studies demonstrate a key role for
inflammation in the initiation and progression of atherosclerosis (Libby, 2002, 192009).
Furthermore, inflammation has the potential to promote thrombosis which can complicate
this disorder and lead to MI and stroke (Libby, 2002, 192009). As discussed above, PM
exposure is associated with systemic inflammation, suggesting a potential role in the
development of atherosclerosis.
Atheroma formation in experimental animals fed a high fat diet begins with the
accumulation of modified lipoprotein particles in the arterial intima, as reviewed by Libby
(2002, 192009). The modification of lipoprotein particles often involves oxidation. As
discussed above, PM exposure is associated with oxidative stress, suggesting a potential
role for PM in the modification of lipoprotein particles. Endothelial dysfunction may also be
key to these early events (Halvorsen et al., 2008, 191149). Recent studies, which are
discussed in later chapters, demonstrate PM-dependent endothelial dysfunction
(Nurkiewicz et al., 2009, 191961). As described by Libby (2002, 191157), oxidative stress
leads to enhance lipid modification and uptake by endothelial cells and initiates an
inflammatory response by activating NFkB. As a result, cell adhesion molecules such as
vascular cell adhesion molecule-1 (VCAM-l) are upregulated and expressed in endothelial
cells. Pro-inflammatory cytokines associated with systemic inflammation may also
contribute to adhesion molecule upregulation. Subsequently, monocytes and T cell
lymphocytes adhere to the activated endothelium, then migrate into the tunica intima
directed by chemokines such as monocyte chemoattractant protein-1 (MCP-1) and IL-8.
Monocytes undergo transformation to tissue macrophages and later to foam cells. As a part
July 2009
5-20
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
of this transformation, monocyte/macrophages express scavenger receptors and bind to and
internalize the modified lipoprotein particles. These cells secrete growth factors and
cytokines, produce ROS, replicate within the lesion and contribute to further lesion
progression. Macrophage colony stimulating factor (M-CSF) and granulocyte-macrophage
colony stimulating factor (GM-CSF) are thought to be involved in these latter steps which
eventually lead to the formation of a fatty streak. T cell lymphocyes also become activated
in the atheroma and secrete pro- or anti-inflammatory cytokines. Degranulation of mast
cells found in the atheroma may also play an important role in the lesion.
Atheroma progression involves the proliferation of smooth muscle cells, which is
stimulated by macrophage-derived growth factors (Libby, 2002, 192009). This results in
smooth muscle accumulation in the lesion, the elaboration of extracellular matrix material
and the formation of a more bulky lesion which can occlude the arterial lumen. Matrix
proteins contribute to the evolution of the lesion from a fatty streak to a fibrous plaque.
Mechanisms which trigger plaque disruption, including endothelial erosions and plaque
rupture, can result in thrombosis as well as in further expansion of the lesion. Resident T
cells, mast cells and circulating platelets may also play a role in destabilizing plaques
(Halvorsen et al., 2008, 191149). It is thought that most atheromatous lesions progress in a
discontinuous manner due to cycles of disruption, expansion and repair.
A major factor regulating plaque disruption is the thickness of the fibrous cap, with
more stable plaques characterized by a thick fibrous cap (Libby, 2002, 192009). Collagen in
the fibrous cap can be degraded by proteases, especially MMPs. Inflammation in the intima
reduces collagen production by smooth muscle cells and promotes the expression and
activation of MMPs. ROS may mediate effects on MMP upregulation (Lund et al., 2009,
180257). Both macrophages and smooth muscle produce MMPs in the lesion areas
(Halvorsen et al., 2008, 191149); fully differentiated macrophages selectively upregulate
certain MMPs with more destructive potential (Newby, 2008, 191161). Rupture of the
fibrous cap allows pro-thrombotic factors within the plaque (e.g., TF) to come into contact
with coagulation factors in the blood possibly resulting in the formation of an occlusive
blood clot. However in some cases, blood fibrinolytic mechanisms minimize clot formation
and repair processes ensue. The result is a more fibrous plaque and/or an expanded lesion.
The proposed role of PM in activating coagulation pathways, which was discussed above,
may influence the outcome of plaque disruption.
In this manner oxidative stress, inflammation and pro-coagulant activity in the blood
are involved in the initiation and progression of atheromatous lesions as well as plaque
disruption and occlusive blood clot formation. PM exposure may contribute to these
pathways and in fact, studies described in later chapters demonstrate PM-dependent effects
on atherosclerosis progression (Araujo et al., 2008, 156222; Chen and Nadziejko, 2005,
087219; Sun et al., 2005, 186814; Ying et al., 2009, 190111).
July 2009
5-21
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
5.2.4. Activation of the Autonomic Nervous System by
Pulmonary Reflexes
Chemosensitive receptors, including rapidly activating receptors (RARs) and sensory
Ofiber receptors, are found at all levels of the respiratory tract and are sensitive to irritant
particles as well as to irritant gases (Alarie, 1973, 070967; Coleridge and Coleridge, 1994,
156362; Widdicombe, 2006, 155519). Activation of trigeminal afferents in the nose causes
CNS reflexes resulting in decreases in respiratory rate through a lengthened expiratory
phase, closure of the glottis, closure of the nares with increased nasal airflow resistance and
effects on the cardiovascular system such as bradycardia, peripheral vasoconstriction and a
rise in systolic arterial blood pressure (Alarie, 1973, 070967). Sneezing, rhinorrhea and
vasodilation with subsequent nasal vascular congestion are also nasal reflex responses
involving the trigeminal nerve (Sarin et al., 2006, 191166). Activation of vagal afferents in
the tracheobronchial and alveolar regions of the respiratory tract causes CNS reflexes
resulting in bronchoconstriction, mucus secretion, mucosal vasodilation, cough, apnea
followed by rapid shallow breathing and effects on the cardiovascular system such as
bradycardia and hypotension or hypertension (Coleridge and Coleridge, 1994, 156362;
Widdicombe, 2003, 157145; 2006, 155519; Widdicombe and Lee, 2001, 019049). Some
evidence suggests that cardiovascular responses may be mediated primarily by the C-fiber
receptors (Coleridge and Coleridge, 1994, 156362) and that irritants in the lower
respiratory tract cause more pronounced cardiovascular responses than irritants in the
upper respiratory tract (Widdicombe and Lee, 2001, 019049).
Early experiments demonstrated that sectioning of the trigeminal nerve abrogated
irritant effects on respiratory rate, heart rate and systolic arterial blood pressure (Alarie,
1973, 070967). These nasal reflexes were attributed to the ophthalmic branch of the
trigeminal nerve since they were identical to reflex responses following diving or immersion
of the face in water (Alarie, 1973, 070967). Early experiments also demonstrated that non-
nasal reflexes were mediated by cholinergic parasympathetic pathways involving the vagus
nerve and inhibited by atropine (Grunstein et al., 1977, 071445; Nadel et al., 1965, 014846).
More recent experiments have shown that noncholinergic mechanisms may also be
involved. For example, stimulation of C-fiber receptors can activate local axon reflexes.
These local axon pathways are responsible for secretion of neuropeptides and the
development of neurogenic inflammation (Widdicombe and Lee, 2001, 019049). It has been
proposed that, in some cases, neurogenic pulmonary responses can switch from their
normally protective function to one that perpetuates pulmonary inflammation (Wong et al.,
2003, 097707). Differences in respiratory tract innervation between rodents and humans
suggest that C-fiber mediated neurogenic inflammation may be more important in rodents
than in humans (Groneberg et al., 2004, 138134; 2003, 157145; Widdicombe and Lee, 2001,
July 2009
5-22
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
019049). However the role of neurogenic inflammation in mediating pulmonary responses
in humans is an active area of investigation.
VR1 receptors represent a subset of neuropeptide and acid-sensitive irritant receptors
which belong to the transient receptor potential (TRP) family. They are located on the
sensory Ofibers which lie underneath and between lung epithelial cells and on immune
and non-immune airway cells. Recent research interest has focused on the role played by
these receptors in mediating inflammation following exposure to PM (Veronesi and
Oortgiesen, 2001, 015977). Exposure to PM has been shown to result in an immediate
increase in intracellular calcium followed by the release of neuropeptides and inflammatory
cytokines (Veronesi et al., 1999, 048764; Veronesi et al., 2000, 017062). In one study this
response was found to be due to an intrinsic property of the particle core and was not
metal-dependent (Oortgiesen et al., 2000, 013998), while in another study electrostatic
charge was found to activate VR1 receptors (Veronesi et al., 2003, 094384). PM-mediated
activation of VR1 receptors which results in increases in intracellular calcium and
apoptosis in epithelial cells has also been demonstrated (Agopyan et al., 2003, 155649).
Recent studies, discussed in later chapters, provide evidence for the involvement of TRPV1
irritant receptor involvement in PM-dependent responses (Ghelfi et al., 2008, 156468;
Rhoden et al., 2005, 087878).
Recently it has been proposed that pulmonary reflex responses may be modulated by
CNS plasticity (Bonham et al., 2006, 191140; Sekizawa et al., 2008, 191167). Plasticity is a
property of neurons or synapses which allows change in response to previous events. In the
case of the respiratory tract, visceral afferent inputs to the CNS are integrated primarily in
the nucleus tractus soltarius (NTS) region of the brain (Bonham et al., 2006, 191110).
Inputs from local networks, higher brain regions and circulating mediators contribute to
the reflex output (Bonham et al., 2006, 1911 10). This integration allows for plasticity in
that repeated or prolonged exposure to a particular stimuli may lead to altered reflex
responses to the same or subsequent stimuli. An exaggerated reflex response was recently
observed in guinea pigs exposed to ETS for an extended period of time (Sekizawa et al.,
2008, 191167). It is not known whether CNS plasticity influences responses following acute
and chronic exposure to ambient PM, but it is a mechanism that may possibly explain
hyperresponsiveness and/or adaptation of reflex-related responses.
At this time, it is not clear how activation of the ANS by pulmonary reflexes
contributes to the kinds of altered conduction and/or repolarization properties of the heart
which may be linked to arrhythmias (Figure 5-5). Pulmonary reflexes, as they are currently
understood, initially lead to increases in parasympathetic tone. However decreased heart
rate variability appears to be reflective of decreased parasympathetic tone and/or increased
sympathetic tone. APM-dependent sympathetic stress response mediated by cytokines has
been postulated (Godleski et al., 2000, 000738) but there is little new information to
support this mechanism. Thus activation of the autonomic nervous system by mechanisms
July 2009
5-23
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
other than pulmonary reflexes seems likely in response to PM. Very little is known about
putative alternative mechanisms leading to sympathoexcitatory responses although one
study demonstrated a role for the olfactory bulb-NTS pathway in regulating cardiovascular
functions following smoke exposure (Moffitt et al., 2002, 191160). In addition,
sympathoexcitatory responses occur during myocardial ischemia, mediated by the release of
adenosine or the production of ROS by the myocardium (Longhurst et al., 2001, 191158).
Hence, PM-dependent effects leading to myocardial ischemia may stimulate
sympathoexcitatory responses. Furthermore, the effects of pre-existing alterations in the
ANS due to disease processes (e.g., increased sympathetic tone observed in cardiac
diseases) on PM responses are not understood. Possibly, integration of neural signals
resulting from pulmonary and cardiac reflexes at the level of the NTS may have an
influence on ANS responses to PM. Further investigation will be required to clarify these
mechanisms.
5.3. Translocation of Ultrafine PM or Soluble PM
Components
Ultrafine PM translocate into cells through non-endocytotic mechanisms involving
adhesive interactions and diffusion (Geiser et al., 2005, 087362), as described in Chapter
4.3.3.1. In this study, inhaled ultrafine particles rapidly crossed cell membranes of alveolar
epithelial cells and capillary endothelial cells but there was no measurable loss of PM from
the lung over 24 h. In another study, ultrafine PM were localized in macrophage
mitochondria as demonstrated by electron microscopy (Li et al., 2003, 042082). Other
studies demonstrated extrapulmonary translocation of poorly soluble ultrafine PM but the
process was slow and resulted in only a small amount leaving the lung. It is possible that in
these studies PM gained access to the circulation after initial transport to the lymph nodes
or the gastrointestinal system. To date, there is limited evidence that ultrafine or other PM
size fractions access the circulation by traversing the epithelial barrier of the respiratory
tract.
However, soluble components from all size fractions of PM have the potential to
translocate across the airway epithelium into the bronchial circulation or across the
alveolar epithelium into the systemic circulation as depicted in Figure 5"5. Absorption
across nasal epithelium may also occur (Ilium, 2006, 191205). Factors affecting this process
include the rate of dissolution of the solute from the particle and the molecular weight of
the solute (see Section 4.4.1). It should be noted that the high surface/volume ratio
characteristic of ultrafine PM may lead to more rapid dissolution of soluble components
associated with this size fraction compared with larger particles
Several interesting studies investigated the translocation of water-soluble metals
from the lung. Gilmour et al. (2006, 156472) demonstrated the rapid appearance of zinc in
July 2009
5-24
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
the plasma of rats intratracheally instilled with zinc sulfate. Similarly, Wallenborn et al.
(2007, 156144) demonstrated the rapid appearance of water-soluble metals in the blood,
heart and liver following intratracheal instillation of oil combustion PM in rats. Using a
more sensitive technique, these same investigators demonstrated the accumulation of 70Zn,
a rare isotope of zinc, in blood, heart and liver following intratracheal instillation of zinc
sulfate (Wallenborn et al., 2009, 191172). In three other studies, soluble zinc and copper
were associated with cardiac effects following intratracheal instillation of rats with
different forms of zinc and copper-containing PM (Gilmour et al., 2006, 088489; Gottipolu et
al., 2008, 191148; Kodavanti et al., 2008, 155907). These results suggest the possibility that
PM-derived soluble zinc and copper translocated across the alveolar-capillary barrier into
the circulation and exerted effects on the heart. However, in the two studies in which
barrier function was measured it was found to be compromised (Gilmour et al., 2006,
088489; Gottipolu et al., 2008, 191148) suggesting an acute lung injury response to
instillation of high concentrations of metal. Acute lung injury is not likely to occur in
healthy individuals exposed to PM at concentrations relevant to ambient levels. Cardiac
effects were also observed following subchronic inhalation exposure to low concentrations of
aerosolized zinc sulfate (Wallenborn et al., 2008, 191171). Since it was not possible to
measure extrapulmonary zinc in this study, it remains unclear whether cardiac effects were
a direct effect of translocated zinc or an indirect effect of exposure to zinc-containing PM.
Nonetheless, translocation of soluble components derived from inhaled PM remains a viable
hypothesis to explain some extra-pulmonary effects.
Epithelial permeability is a key determinant of translocation and is discussed in
detail in Section 4.4.2. In brief, a number of studies have measured clearance of 99mTc-DTPA
as an index of alveolar epithelial membrane integrity and permeability of alveolar-capillary
barrier (Braude et al., 1986, 155701). Endothelial integrity also contributes to the alveolar-
capillary barrier and is measured by transvascular protein flux but is not discussed here
(Braude et al., 1986, 155701). Interestingly, epithelial permeability was transiently
increased in human volunteers following 3 h of moderate exercise but not following 24-h
exposure to particle-rich urban air (Brauner et al., 2009, 190211). A previous study found
that the exercise-induced increase in epithelial permeability was transient and suggested
that it was due to increased ventilation and elevated vascular pressure which altered the
properties of tight junctions (Hanel et al., 2003, 155826). Smokers (Jones et al., 1983,
155884) and individuals with acute respiratory distress syndrome (Braude et al., 1986,
155701) or interstitial lung disease (Rinderknecht et al., 1980, 191965) also exhibited
increased alveolar epithelial permeability. The changes in smokers were reversible upon
cessation of smoking. In laboratory animals, increased alveolar permeability was
demonstrated in terminally senescent mice (Tankersley et al., 2003, 096363). Clearance of
99mTc-DTPA was also used to measure permeability of the bronchial mucosa in a study by
Ilowite (Ilowite et al., 1989, 156584) which found increased airway epithelial permeability
July 2009
5-25
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
in asthmatics. These studies demonstrate that epithelial permeability is increased following
moderate exercise and in lung syndromes associated with inflammation and suggest that
compromised epithelial barrier functions in the lung may contribute to PM-mediated
effects.
Interaction of circulating PM or soluble PM components with vascular endothelial
cells, platelets, and other leukocytes is a potential mechanism underlying the
cardiovascular and systemic effects of inhaled PM. A role for PM-derived ROS and/or
cellular-derived ROS has been proposed. Furthermore, soluble metals that do not
redox-cycle may activate cell signaling pathways without the generation of ROS. In this
way PM may promote adverse cardiovascular effects such as endothelial dysfunction,
atherosclerosis and thrombosis. Circulating PM or soluble PM components also have the
potential to impact other organ systems. However convincing evidence that this occurs to
an appreciable extent in healthy individuals following inhalation of PM at concentrations
relevant to ambient exposures is lacking.
5.4. Disease of the Cardiovascular and Other Organ
Systems
As discussed above, deposition of PM in the lung may lead not only to pulmonary
disease but also to diseases of other systems (Figure 5-5). In the cardiovascular system,
myocardial ischemia and MI may occur as a result of the above proposed effects of PM on
atherosclerosis, plaque instability, thrombosis, plaque rupture and/or altered vasoreactivity
of coronary vessels. Myocardial ischemia and MI may alter the conduction and
depolarization properties of the heart and lead to arrhythmic events. In addition,
thrombosis may lead to stroke and/or thromboembolic disease. Many of these processes may
be interlinked and responses to ambient PM exposures may involve multiple mechanisms
simultaneously with some variability depending on PM composition. Furthermore, it is not
clear at this time whether PM initiates cardiovascular disease or whether it perturbs
existing disease.
In addition, recent studies which are discussed in later chapters have demonstrated
PM-dependent effects on the CNS (Campbell et al., 2005, 087217; Kleinman et al., 2008,
190074; Sirivelu et al., 2006, 111151: Veronesi et al., 2005, 087481; WiirShwe et al., 2008,
190146). At this time, it is not known whether this is a direct or indirect consequence of PM
exposure. Translocation of soluble and poorly soluble particles from the olfactory mucosa
via the axons to the olfactory bulb of the brain has been proposed as a possible mechanism
by which PM or its components may directly access the CNS. Evidence for this pathway is
discussed in Section 4.3.2.2. Given their small size, ultrafine PM deposited onto nasal
epithelium may be more efficiently translocated by this mechanism compared with larger
size PM. Alternative mechanisms proposed for PM-mediated CNS effects involve systemic
July 2009
5-26
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
inflammation and autonomic responses. These are new and intriguing possibilities which
warrant further investigation.
PM-dependent effects on the reproductive system, reproductive outcomes and
perinatal development have also been identified and are discussed in a later chapter.
Mechanisms involved in these responses have not been determined. However it seems
possible that systemic inflammation and/or oxidative stress may play a role. Developmental
windows of susceptibility may also be a key consideration. Furthermore it has been
hypothesized that oxygen gradients and redox status are key to cell differentiation and
epigenetic processes occurring during development (see Section 5.1.11) (Hitchler and
Domann, 2007, 191151).
5.5. Acute and Chronic Responses
In general, repeated acute responses may lead to cumulative effects which manifest as
chronic disease. Several examples relevant to the modes of action discussed in this chapter
are that of allergic responses, atherosclerosis and lung development. Allergic responses
require repeated exposures to antigen over time. Co-exposure to an adjuvant, possibly DEP
or ultrafine concentrated ambient particles (CAPs), can enhance this response.
Furthermore, the presence of oxidative stress, as may occur in response to PM, can
contribute to allergic responses. Allergic sensitization often underlies allergic asthma,
characterized by inflammation and AHR. In this way, repeated or chronic exposures
involving multifactorial responses (immune system activation, oxidative stress,
inflammation) can lead to irreversible outcomes. Similarly, the development of
atherosclerosis involves inflammation and remodeling of the blood vessel wall. Factors
contributing to this process include systemic inflammation, endothelial dysfunction,
oxidative stress and high levels of circulating lipids. PM exposure is associated with three
out of four of these processes. The role of PM in initiating, promoting or complicating this
disease or its outcomes has yet to be determined. Critical windows of susceptibility during
development also provide an opportunity for repeated exposures to injurious agents to lead
to irreversible changes in organ structure and function. The extended period of postnatal
lung development in humans and other species heightens this vulnerability. PM may serve
as such an injurious agent as has been demonstrated previously for hyperoxia (Randell et
al„ 1990, 191956).
Furthermore, adverse outcomes may be precipitated by acute events imposed on
chronic disease states. In the case of allergic asthma, acute PM exposure may provoke
asthmatic responses through oxidative stress and inflammatory pathways. Additionally, PM
can act as a carrier of aeroallergens and other biological materials which can potentially
trigger asthma attacks. Similarly, PM exposure may provoke inflammatory or thrombotic
responses leading to rupture of an atherosclerotic plaque which subsequently leads to acute
July 2009
5-27
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
myocardial infarction. In this way, the outcome of an acute exposure to PM may be
drastically worsened by the underlying chronic disease.
5.6. Results of New Inhalation Studies which Contribute to
Modes of Action
Prior to this review, much of the evidence for the proposed modes of action was
obtained from animal studies involving intratracheal instillation or inhalation of high
concentrations of PM and from cell culture experiments. In many cases, the types of PM
used were of questionable relevance to ambient exposures (i.e., high concentrations of
ROFA, metals and ambient PM collected on filters). Since then, many inhalation studies
have been conducted using CAPs, combustion-derived PM, urban air and carbon black,
generally using concentrations of PM lower than 1 mg/m3. Much of this research has been
conducted in animal models of disease. These key new studies, described in detail in
Chapters 6 and 7, add to the understanding of modes of action which are relevant to
ambient PM exposure. A compilation of pertinent results is found below.
¦	Altered lung function including changes in respiratory frequency and AHR following
short-term exposures to CAPs and combustion-derived PM (Section 6.3.2.3)
¦	Mild pulmonary inflammation in response to short-term exposures to CAPs, urban
air, combustion-derived PM and carbon black (Section 6.3.3.3)
¦	Mild pulmonary injury in response to short-term exposure to CAPs and combustion-
derived PM (Section 6.3.5.3)
¦	Inhibition of cell proliferation in the proximal alveolar region of neonatal animals
following short-term exposure to iron-soot (Section 6.3.5.3)
¦	Pulmonary oxidative stress in response to short-term exposure to CAPs, urban air,
combustion-derived PM, carbon black and iron-soot! pulmonary nitrosative stress in
response to titanium dioxide (TiOa) (Section 6.3.4.2)
¦	Antioxidant intervention which ameliorates PM effects on oxidative stress, allergic
responses, and AHR (Sections 6.3.4.2 )
¦	Allergic sensitization and exacerbation of allergic responses in response to CAPs and
combustion-derived PM (Section 6.3.6.3)
¦	Altered methylation of promoter regions of IFN-y and IL-4 genes suggestive of pro-
allergic Th2 gene activation following short-term exposure to combustion-derived
PM in an allergy model (Section 6.3.6.3)
July 2009
5-28
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Increased susceptibility to respiratory infection following exposure to combustion-
derived PM (Section 6.3.7.2)
Effects on nasal epithelial mucosubstances, airway morphology and airway
mucosubstances following chronic exposure to urban air-derived PM and woodsmoke
(Section 7.3.5.1)
Worsening of papain-induced emphysema following chronic exposure to urban air-
derived PM (Section 7.3.5.1)
Effects on lung development following chronic exposure to urban air-derived PM
(Section 7.3.2.2 and 7.3.5.1)
Prolonged exposure to CAPs and combustion-derived PM leading sometimes to mild
pulmonary inflammation, oxidative stress and injury and sometimes to loss of
inflammatory, oxidative stress and AHR responses which were observed after short-
term exposures (Sections 7.3.2.2, 7.3.3.2. 7.3.5.1 and 7.3.6.2)
Hypermethylation of lung DNA following chronic exposure to combustion-derived
PM (Section 7.3.5.1)
A role for TRPV1 irritant receptors in activating local axon and CNS reflexes
following short term exposure to CAPs and combustion-derived PM (see
Section 6.2.9.3)
A role for TRPV1 irritant receptors in mediating lung and heart oxidative stress
through increased parasympathetic and sympathetic activity in response to CAPs.
(see Section 6.2.9.3 and 6.3.2.3)
Altered heart rate variability in response to CAPs, combustion-derived PM and
carbon black (Section 6.2.1.3)
Arrhymthmic events in response to CAPs and combustion-derived PM
(Section 6.2.2.2)
Decreased cardiac contractility following short-term exposure to CAPs and carbon
black (Section 6.2.6.1)
Enhanced myocardial ischemia following short-term exposure to CAPs
(Section 6.2.3.3).
July 2009
5-29
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
Endothelial dysfunction and altered vascular reactivity following short-term
exposure to CAPs, combustion-derived PM and Ti02 (Section 6.2.4.3)
3	¦ Increases in blood pressure following short-term exposure to CAPs and carbon black
4	(Section 6.2.5.3)
5	¦ Changes in blood leukocyte counts following short-term exposure to CAPs and
6	carbon black (Section 6.2.7.3)
7	¦ Increased levels of blood coagulation factors following short-term exposure to CAPs
8	and on-road highway aerosols (Section 6.2.8.3)
9	¦ Systemic and cardiovascular oxidative stress in response to short-term exposure to
10	CAPs, road dust and combustion-derived PM (Section 6.2.9.3)
11	¦ Progression of atherosclerosis, induction of TF in aortic plaques, vascular oxidative
12	stress and altered vasomotor function following long-term exposure to CAPs in a
13	susceptible animal model (Section 7.2.1.2)
14	¦ Vascular remodeling following chronic exposure to urban air-derived PM
15	(Section 7.2.1.2).
16	¦ Enhanced angiotensin 11-induced hypertension accompanied by vascular oxidative
17	stress and altered vasoreactivity in response to chronic exposure to CAPs
18	(Section 7.2.5.2)
19	¦ Exaggerated insulin resistance, visceral adiposity and systemic inflammation in
20	response to chronic exposure to CAPs and a high-fat diet (Section 7.2.3.1)
21	¦ CNS responses following short- and long-term exposures to CAPs and combustion-
22	derived PM (Section 6.4.3)
23	¦ Effects on the reproductive system, reproductive outcomes and perinatal
24	development following chronic exposure to urban-air derived PM (Section 7.4.2)
25	¦ DNA adducts in nose, lung and liver following chronic exposure to urban air
26	(Section 7. 5.2)
27	¦ Germ line mutations, DNA strand breaks and global hypermethylation in sperm
28	following chronic exposure to urban air-derived PM (Section 7.5.3)
29	These new studies confirm and extend findings from older studies. However this
30	increasing body of evidence does not provide a complete picture of the biological pathways
July 2009
5-30
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
involved in mediating PM effects. For example, a lack of information regarding the time-
dependence of many responses makes it difficult to understand the underlying biological
mechanisms. Existing gaps in knowledge include^
¦	The spatial distribution of retained particles in the lung and its impact
¦	The deposition, uptake and clearance of ultrafine particles in the lung
¦	Effects of ambient PM exposures on epithelial barrier function in the lung
¦	Time dependence of responses
¦	The putative modulation of neural reflexes by pre-existing disease or other factors.
¦	The putative role of neural reflexes besides those involving pulmonary irritant
receptors
¦	The putative role of ET in altering vasomotor tone following PM exposure
¦	The putative translocation of PM or soluble components across the epithelial barrier
of the lung into the circulation
¦	The putative translocation of PM from olfactory epithelium to the olfactory bulb and
other brain regions.
Additional studies will be required to clarify the biological mechanisms underlying
the health effects of PM.
July 2009
5-31
DRAFT-DO NOT CITE OR QUOTE

-------
Chapter 5 References
Adamson I; Prieditis H. (1995). Response of mouse lung to carbon deposition during injury and repair. Environ
Health Perspect, 103: 72-76. 189982
Adcock IMi Tsaprouni Li Bhavsar Pi Ito K. (2007). Epigenetic regulation of airway inflammation. Curr Opin
Allergy Clin Immunol, 19: 694-700. 191178
Agopyan Ni Bhatti Ti Yu Si Simon SA. (2003). Vanilloid receptor activation by 2_and lO-jim particles induces
responses leading to apoptosis in human airway epithelial cells. Toxicol Appl Pharmacol, 192: 21-35.
155649
Ahsan MKi Nakamura Hi Tanito Mi Yamada Ki Utsumi Hi Yodoi J. (2005). Thioredoxin-1 suppresses lung
injury and apoptosis induced by diesel exhaust particles (DEP) by scavenging reactive oxygen species and
by inhibiting DEP-induced downregulation of Akt. Free Radic Biol Med, 39: 1549-1559. 156200
Alarie Y. (1973). Sensory irritation by airborne chemicals. Crit Rev Toxicol, 2: 299-363. 070967
Amara Ni Bachoual Ri Desmard Mi Golda Si Guichard Ci Lanone Si Aubier Mi Ogier-Denis Ei Boczkowski J.
(2007). Diesel exhaust particles induce matrix metalloprotease-1 in human lung epithelial cells via a
NADP(H) oxidase/NOX4 redox-dependent mechanism. Am J Physiol Lung Cell Mol Physiol, 293: L170-
181. 156212
Araujo JAi Barajas Bi Kleinman Mi Wang Xi Bennett BJi Gong KWi Navab Mi Harkema Ji Sioutas Ci Lusis AJi
Nel AE. (2008). Ambient particulate pollutants in the ultrafine range promote early atherosclerosis and
systemic oxidative stress. Circ Res, 102: 589-596. 156222
Ayres JGi Borm Pi Cassee FRi Castranova Vi Donaldson Ki Ghio Ai Harrison RMi Hider Ri Kelly Fi Kooter IMi
Marano Fi Maynard RLi Mudway Ii Nel Ai Sioutas Ci Smith Si Baeza-Squiban Ai Cho Ai Duggan Si
Froines J. (2008). Evaluating the Toxicity of Airborne Particulate Matter and Nanoparticles by
Measuring Oxidative Stress Potential - A Workshop Report and Consensus Statement. Inhal Toxicol, 20:
75 - 99.155666
Baccarelli Ai Martinelli Ii Pegoraro Vi Melly Si Grillo Pi Zanobetti Ai Hou Li Bertazzi PAi Mannucci PMi
Schwartz J. (2009). Living near Major Traffic Roads and Risk of Deep Vein Thrombosis. Circulation, x: x.
188183
Bachoual Ri Boczkowski Ji Goven Di Amara Ni Tabet Li On Di Lecon-Malas Vi Aubier Mi Lanone S. (2007).
Biological effects of particles from the paris subway system. , 20: 1426-1433. 155667
Barnes PJ. (2007). New molecular targets for the treatment of neutrophilic diseases. , 119: 1055"62i quiz 1063-4.
191139
Bayram Hi Ito Ki Issa Ri Ito Mi Sukkar Mi Chung KF. (2006). Regulation of human lung epithelial cell numbers
by diesel exhaust particles. Eur Respir J, 27: 705-713. 088439
Becher Ri Bucht Ai Ovrevik Ji Hongslo Jan Ki Dahlman Hans Ji Samuelsen Jan Ti Schwarze Per E. (2007).
Involvement of NADPH Oxidase and iNOS in Rodent Pulmonary Cytokine Responses to Urban Air and
Mineral Particles. , 19: 645-55. 097125
Beck-Speier Ii Dayal Ni Karg Ei Maier KLi Schumann Gi Schulz Hi Semmler Mi Takenaka Si Stettmaier Ki
Bors Wi Ghio Ai Samet JMi Heyder J. (2005). Oxidative stress and lipid mediators induced in alveolar
macrophages by ultrafine particles. Free Radic Biol Med, 38: 1080-1092. 156262
Becker Si Fenton MJi Soukup JM. (2002). Involvement of microbial components and toll-like receptors 2 and 4
in cytokine responses to air pollution particles. Am J Respir Cell Mol Biol, 27: 611-618. 052419
Becker Si Mundandhara Si Devlin RBi Madden M. (2005). Regulation of cytokine production in human alveolar
macrophages and airway epithelial cells in response to ambient air pollution particles: further
mechanistic studies. Toxicol Appl Pharmacol, 207: 269-275. 088590
Blanchet Si Ramgolam Ki Baulig Ai Marano Fi Baeza-Squiban A. (2004). Fine particulate matter induces
amphiregulin secretion by bronchial epithelial cells. Am J Respir Cell Mol Biol, 230: 421-427. 087982
Blank Fi Rothen-Rutishauser Bi Gehr P. (2007). Dendritic cells and macrophages form a transepithelial
network against foreign particulate antigens. , 36: 669-77. 096521
Bleck Bi Tse Doris Bi Jaspers Ii de Lafaille Maria AC; Reibman J. (2006). Diesel exhaust particle-exposed
human bronchial epithelial cells induce dendritic cell maturation. , 176: 7431-7437. 096560
Bonham AC; Chen CY; Sekizawa SJ Joad JP. (2006). Plasticity in the nucleus tractus solitarius and its influence
on lung and airway reflexes. J Appl Physiol, 101: 322"7. 191140
Braude Si Nolop KBi Hughes JMBi Barnes PJi Royston D. (1986). Comparison of lung vascular and epithelial
permeability indices in the adult respiratory distress syndrome. , 133: 1002-1005. 155701
Brauner EVi Mortensen Ji Moller Pi Bernard Ai Vinzents Pi Wahlin Pi Glasius Mi Loft S. (2009). Effects of
ambient air particulate exposure on blood-gas barrier permeability and lung function. , 21: 38"47. 190244
July 2009
5-32
DRAFT-DO NOT CITE OR QUOTE

-------
Brown DM; Donaldson K> Borm PJ; Schins RP; Dehnhardt Mi Gilmour Pi Jimenez LAi Stone V. (2004). Calcium
and ROS-mediated activation of transcription factors and TNF-alpha cytokine gene expression in
macrophages exposed to ultrafine particles. Am J Physiol, 286- L344-L353. 155705
Brown DMi Donaldson Ki Stone V. (2004). Effects of PM10 in human peripheral blood monocytes and J774
macrophages. Respir Res, 5- 29. 088663
Caligiuri Gi Levy Bi Pernow Ji Thoren Pi Hansson GK. (1999). Myocardial infarction mediated by endothelin
receptor signaling in hypercholesterolemia mice. Presented at . 156318
Campbell Ai Oldham Mi Becaria Ai Bondy SCi Meacher Di Sioutas Ci Misra Ci Mendez LBi Kleinman M. (2005).
Particulate matter in polluted air may increase biomarkers of inflammation in mouse brain.
Neurotoxicology, 26: 133-140. 087217
Cao Di Tal TLi Graves LMi Gilmour Ii Linak Wi Reed Wi Bromberg PAi Samet JM. (2007). Diesel exhaust
particulate-induced activation of Stat3 requires activities of EGFR and Src in airway epithelial cells. Am
J Physiol Lung Cell Mol Physiol, 292: L422-L429. 156322
Chan RC-Fi Wang Mi Li Ni Yanagawa Yi Onoe Ki Lee JJi Nel AE. (2006). Pro-oxidative diesel exhaust particle
chemicals inhibit LPS"induced dendritic cell responses involved in T-helper differentiation. J Allergy Clin
Immunol, 118: 455-465. 097468
Chang C-Ci Chiu H"Fi Wu Y-Si Li Y-Ci Tsai M-Li Shen C-Ki Yang C-Y. (2005). The induction of vascular
endothelial growth factor by ultrafine carbon black contributes to the increase of alveolar-capillary
permeability. Environ Health Perspect, 113: 454-460. 097776
Chen Ji Tan Mi Nemmar A; Song Wi Dong Mi Zhang Gi Li Y. (2006). Quantification of extrapulmonary
translocation of intratracheal-instilled particles in vivo in rats: effect of lipopolysaccharide. Toxicology,
222: 195-201. 147267
Chen LCi Nadziejko C. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice
V CAPs exacerbate aortic plaque development in hyperlipidemic mice. Inhal Toxicol, 17: 217-224. 087219
Cho AE Sioutas C; Miguel AH; Kumagai Y; Schmitz DA; Singh M; Eiguren-Fernandez A; Froines JR. (2005).
Redox activity of airborne particulate matter at different sites in the Los Angeles Basin. Environ Res, 99:
40-47. 087937
Cho H-Y; Reddy SP; Kleeberger SR. (2006). Nrf2 defends the lung from oxidative stress. Antioxid Redox Signal,
8: 76-87. 156345
Chu AJ. (2005). Tissue factor mediates inflammation. Arch Biochem Biophys, 440: 123-132. 155730
Coleridge HM! Coleridge JCG. (1994). Pulmonary Reflexes: Neural Mechanisms of Pulmonary Defense. Annu
Rev Physiol, 56: 69-91. 156362
Cowburn AS; Condliffe AM; Farahi N; Summers C; Chilvers ER. (2008). Advances in neutrophil biology: clinical
implications. , 134: 606-12. 191142
de Kok TM; Hogervorst JG; Briede JJ; van Herwijnen MH! Maas LM; Moonen EJ; Driece HA; Kleinjans JC.
(2005). Genotoxicity and physicochemical characteristics of traffic-related ambient particulate matter. ,
46: 71-80. 189835
Dick CAJ; Brown DM; Donaldson K; Stone V. (2003). The role of free radicals in the toxic and inflammatory
effects of four different ultrafine particle types. Inhal Toxicol, 15: 39-52. 036605
Donaldson K; Borm PJ; Oberdorster G; Pinkerton KE; Stone V; Tran CL. (2008). Concordance between in vitro
and in vivo dosimetry in the proinflammatory effects of low-toxicity, low-solubility particles: the key role
of the proximal alveolar region. Inhal Toxicol, 20: 53"62. 190217
Donaldson K; Stone V; Borm PJ; Jimenez LA; Gilmour PS; Schins RP; Knaapen AM; Rahman I; Faux SP;
Brown DM; MacNee W. (2003). Oxidative stress and calcium signaling in the adverse effects of
environmental particles (PM10). Free Radic Biol Med, 34: 1369-1382. 156408
Dostert C; Petrilli V; Van Bruggen R; Steele C; Mossman BT; Tschopp J. (2008). Innate Immune Activation
Through Nalp3 Inflammasome Sensing of Asbestos and Silica. , 320: 674. 155753
Foley DL; Craig JM; Morley R; Olsson CA; Dwyer T; Smith K; Saffery R. (2009). Prospects for epigenetic
epidemiology. Am J Epidemiol, 169: 389-400. 191144
Gallagher J; Sams R; Inmon J; Gelein R; Elder A; Oberdorster G; Prahalad AK. (2003). Formation of 8-oxo"7,8"
dihydro-2'-deoxyguanosine in rat lung DNA following subchronic inhalation of carbon black. Toxicol Appl
Pharmacol, 190: 224-231. 140171
Geiser M; Rothen-Rutishauser B; Kapp N; Schurch S; Kreyling W; Schulz H; Semmler M; Im Hof V; Heyder J;
Gehr P. (2005). Ultrafine particles cross cellular membranes by nonphagocytic mechanisms in lungs and
in cultured cells. Environ Health Perspect, 113: 1555-1560. 087362
Geng H; Meng Z; Zhang Q. (2005). Effects of blowing sand fine particles on plasma membrane permeability and
fluidity, and intracellular calcium levels of rat alveolar macrophages. , 157: 129-137. 096689
Geng H; Meng Z; Zhang Q. (2006). In vitro responses of rat alveolar macrophages to particle suspensions and
water-soluble components of dust storm PM(2.5). Toxicol In Vitro, 20: 575-584. 097026
July 2009
5-33
DRAFT-DO NOT CITE OR QUOTE

-------
Ghelfi E; Rhoden CR; Wellenius GA; Lawrence J; Gonzalez-Flecha B. (2008). Cardiac oxidative stress and
electrophysiological changes in rats exposed to concentrated air particles are mediated by TRP"
dependent pulmonary reflexes. Toxicol Sci, 102- 328-336. 156468
Ghio AJ; Churg A; Roggli VL. (2004). Review- Ferruginous Bodies' Implications in the Mechanism of Fiber and
Particle Toxicity. , 32: 643. 155790
Ghio AJ; Cohen MD. (2005). Disruption of iron homeostasis as a mechanism of biologic effect by ambient air
pollution particles. Inhal Toxicol, 17: 709-716. 088272
Gilmour PS; Morrison ER; Vickers MA; Ford I; Ludlam CA; Greaves M; Donaldson K; MacNee W. (2005). The
procoagulant potential of environmental particles (PM10). Occup Environ Med, 62: 164-171. 087410
Gilmour PS; Nyska A; Schladweiler MC; McGee JK; Wallenborn JG; Richards JH; Kodavanti UP. (2006).
Cardiovascular and blood coagulative effects of pulmonary zinc exposure. Toxicol Appl Pharmacol, 21l:
41-52. 088489
Gilmour PS; Rahman I; Donaldson K; MacNee W. (2003). Histone acetylation regulates epithelial IL"8 release
mediated by oxidative stress from environmental particles. Retrieved , from . 096959
Gilmour PS; Schladweiler MC; Nyska A; McGee JK; Thomas R; Jaskot RH; Schmid J; Kodavanti UP. (2006).
Systemic imbalance of essential metals and cardiac gene expression in rats following acute pulmonary
zinc exposure. J Toxicol Environ Health A, 69: 2011-2032. 156472
Godleski JJ; Verrier RL; Koutrakis P; Catalano P; Coull B; Reinisch U; Lovett EG; Lawrence J; Murthy GG;
Wolfson JM; Clarke RW; Nearing BD; Killingsworth C. (2000). Mechanisms of morbidity and mortality
from exposure to ambient air particles. , 91; 5"88; discussion 89-103. 000738
Gopalakrishnan S; Van Emburgh BO; Robertson KD. (2008). DNA methylation in development and human
disease. , 647: 30-8. 191147
Gottipolu RR; Landa ER; Schladweiler MC; McGee JK; Ledbetter AD; Richards JH; Wallenborn GJ; Kodavanti
UP. (2008). Cardiopulmonary responses of intratr ache ally instilled tire particles and constituent metal
components. Inhal Toxicol, 20: 473-84. 191148
Groneberg DA; Quarcoo D; Frossard N; Fischer A. (2004). Neurogenic mechanisms in bronchial inflammatory
diseases. Allergy, 59: 1139-52. 138134
Grunstein MM; Hazucha M; Sorli J; Milic-Emili J. (1977). Effect of S02 on control of breathing in anesthetized
cats. J Appl Physiol, 43: 844-851. 071445
Gabelova A; Valovicova Z; Labaj J; Bacova G; Binkova B; Farmer Peter B. (2007). Assessment of oxidative DNA
damage formation by organic complex mixtures from airborne particles PM(10). , 620: 135-144. 156457
Halvorsen B; Otterdal K; Dahl TB; Skjelland M; Gullestad L; Oie E; Aukrust P. (2008). Atherosclerotic plaque
stability--what determines the fate of a plaque?. , 51: 183-94. 191149
Hanel B; Law I; Mortensen J. (2003). Maximal rowing has an acute effect on the blood-gas barrier in elite
athletes. , 95: 1076-1082. 155826
Hitchler MJ; Domann FE. (2007). An epigenetic perspective on the free radical theory of development. Free
Radic Biol Med, 43: 1023-36. 191151
Huang Y-CT; Soukup J; Harder S; Becker S. (2003). Mitochondrial oxidant production by a pollutant dust and
NO-mediated apoptosis in human alveolar macrophage. Am J Physiol Lung Cell Mol Physiol, 284: C24"
32. 156573
Ilium L. (2006). Nasal clearance in health and disease. , 19: 92-99. 191205
Ilowite JS; Bennett WD; Sheetz MS; Groth ML; Nierman DM. (1989). Permeability of the bronchial mucosa to
99mTc-DTPA in asthma. , 139: 1139-1143. 156584
Imrich A; Ning Y; Lawrence J; Coull B; Gitin E; Knutson M; Kobzik L. (2007). Alveolar macrophage cytokine
response to air pollution particles: oxidant mechanisms. Toxicol Appl Pharmacol, 218: 256-264. 155859
Jaspers I; Ciencewicki JM; Zhang W', Brighton LE; Carson JL; Beck MA; Madden MC. (2005). Diesel exhaust
enhances influenza virus infections in respiratory epithelial cells. Toxicol Sci, 85: 990-1002. 088115
Jiang J; Oberdrster G; Elder A; Gelein R; Mercer P; Biswas P. (2008). Does nanoparticle activity depend upon
size and crystal phase?. , 2: 33"42. 156609
Jones J; Stick S; Dingle P; Franklin P. (2007). Spatial variability of particulates in homes: implications for
infant exposure. Sci Total Environ, 376: 317-323. 156615
Jones JG; Minty BD; Royston D; Royston JP. (1983). Carboxyhaemoglobin and pulmonary epithelial
permeability in man. Thorax, 38: 129-133. 155884
Jones PA; Baylin SB. (2007). The epigenomics of cancer. Cell, 128: 683-92. 191153
Kaan PM; Hegele RG. (2003). Interaction between respiratory syncytial virus and particulate matter in guinea
pig alveolar macrophages. Am J Respir Cell Mol Biol, 28: 697-704. 095753
Karoly ED; Li Z; Dailey LA; Hyseni X; Huang YCT. (2007). Up-regulation of Tissue Factor in Human
Pulmonary Artery Endothelial Cells after Ultrafine Particle Exposure. , 115: 535. 155890
July 2009
5-34
DRAFT-DO NOT CITE OR QUOTE

-------
Kaul N> Forman HJ. (1996). Activation of NF?B by the respiratory burst of macrophages. Free Radic Biol Med,
21: 401-405. 155892
Keverne EB; Curley JP. (2008). Epigenetics, brain evolution and behaviour. , 29: 398-412. 191154
Keverne EB; Curley JP. (2008). Epigenetics, brain evolution and behaviour. Neuroendocrinology, 29: 398-412.
191958
KLeinman MT; Araujo JA; Nel A; Sioutas C> Campbell A; Cong PQ; Li H; Bondy SC. (2008). Inhaled ultrafine
particulate matter affects CNS inflammatory processes and may act via MAP kinase signaling pathways.
Toxicol Lett, 178: 127-130. 190074
Kodavanti UP; Schladweiler MC; Gilmour PS; Wallenborn JG; Mandavilli BS; Ledbetter AD; Christiani DC;
Runge MS; Karoly ED; Costa DL; Peddada S; Jaskot R; Richards JH; Thomas R; Madamanchi NR;
Nyska A. (2008). The role of particulate matter-associated zinc in cardiac injury in rats. , 116: 13-20.
155907
Koike E; Hirano S; Furuyama A; Kobayashi T. (2004). cDNA microarray analysis of rat alveolar epithelial cells
following exposure to organic extract of diesel exhaust particles. Toxicol Appl Pharmacol, 201: 178-185.
058555
Koike E; Kobayashi T. (2005). Organic extract of diesel exhaust particles stimulates expression of la and
costimulatory molecules associated with antigen presentation in rat peripheral blood monocytes but not
in alveolar macrophages. Toxicol Appl Pharmacol, 209: 277-285. 088303
Lee CC; Cheng YW; Kang JJ. (2005). Motorcycle exhaust particles induce IL-8 production through NF-kappaB
activation in human airway epithelial cells. J Toxicol Environ Health A, 68: 1537-1555. 156682
Lee SS; Woo CH; Chang JD; Kim JH. (2003). Roles of Rac and cytosolic phospholipase A2 in the intracellular
signalling in response to titanium particles. Cell Signal, 15: 339-345. 156678
Li J; Ghio AJ; Cho SH; Brinckerhoff CE; Simon SA; Liedtke W. (2009). Diesel exhaust particles activate the
matrix-metalloproteinase-1 gene in human bronchial epithelia in a beta-arrestin-dependent manner via
activation of RAS. Environ Health Perspect, 117: 400-409. 190424
Li N; Nel AE. (2006). Role of the Nrf2 -mediated signaling pathway as a negative regulator of inflammation:
Implications for the impact of particulate pollutants on asthma. Antioxid Redox Signal, 8: 88"98. 156694
Li N; Sioutas C; Froines JR; Cho A; Misr. (2003). Ultrafine particulate pollutants induce oxidative stress and
mitochondrial damage. Environ Health Perspect, 111: 455-460. 042082
Li Y-J; Kawada T; Matsumoto A; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2007). Airway inflammatory
responses to oxidative stress induced by low-dose diesel exhaust particle exposure differ between mouse
strains. , 33: 227-244. 155929
Libby P. (2002). Inflammation in atherosclerosis. Nature, 420: 868-874 . 192009
Libby P; Ridker PM; Maseri A. (2002). Inflammation and atherosclerosis. Circulation, 105: 1135-43. 191157
Lindbom J; Gustafsson M; Blomqvist G; Dahl A; Gudmundsson A; Swietlicki E; Ljungman AG. (2007). Wear
particles generated from studded tires and pavement induces inflammatory reactions in mouse
macrophage cells. , 20: 937-946. 155934
Liu J; Ballaney M; Al-Alem U; Quan C; Jin X; Perera F; Chen LC; Miller RL. (2008). Combined Inhaled Diesel
Exhaust Particles and Allergen Exposure Alter Methylation of T Helper Genes and IgE Production In
Vivo. Toxicol Sci, 102: 76-81. 156709
Liu X; Meng Z. (2005). Effects of airborne fine particulate matter on antioxidant capacity and lipid peroxidation
in multiple organs of rats. Inhal Toxicol, 17: 467-473. 088650
Long JF; Waldman WJ; Kristovich R; Williams M; Knight D; Dutta PK. (2005). Comparison of ultrastructural
cytotoxic effects of carbon and carbon/iron particulates on human monocyte-derived macrophages.
Environ Health Perspect, 113: 170-174. 087454
Longhurst JC; Tjen ALSC; Fu LW. (2001). Cardiac sympathetic afferent activation provoked by myocardial
ischemia and reperfusion. Mechanisms and reflexes. , 940: 74"95. 191158
Lund AK; Lucero A; Lucas S; Madden MC; McDonald JD; Seagrave JC; Knuckles TL; Campen MJ. (2009).
Vehicular emissions induce vascular MMP-9 expression and activity associated with endothelial—
mediated pathways. Arterioscler Thromb Vase Biol, 29: 511-517. 180257
Mackman N. (2005). Tissue "Specific Hemostasis in Mice. Arterioscler Thromb Vase Biol, 25: 2273-2281. 156722
Mercer BA; Kolesnikova N; Sonett J; DArmiento J. (2004). Extracellular regulated kinase/mitogen activated
protein kinase is up-regulated in pulmonary emphysema and mediates matrix metalloproteinase-1
induction by cigarette smoke. J Biol Chem, 279: 17690-17696. 191180
Moffitt JA; Grippo AJ; Holmes PV; Johnson AK. (2002). Olfactory bulbectomy attenuates cardiovascular
sympathoexcitatory reflexes in rats. , 283: H2575-83. 191160
Molinelli AR; Madden MC; McGee JK; Stonebuerner JG; Ghio AJ. (2002). Effect of metal removal on the toxicity
of airborne particulate matter from the Utah Valley. Inhal Toxicol, 14: 1069-1086. 035347
July 2009
5-35
DRAFT-DO NOT CITE OR QUOTE

-------
Moller W; Brown DM; Kreyling WG; Stone V. (2005). Ultrafine particles cause cytoskeletal dysfunctions in
macrophages' role of intracellular calcium. Part Fibre Toxicol, 2- 7. 156770
Monn C> Naef R; Koller T. (2003). Reactions of macrophages exposed to particles <10 "micron"m. Environ Res,
91: 35-44. 052418
Mutlu GM; Green D; Bellmeyer A; Baker CM; Burgess Z; Rajamannan N; Christman JW; Foiles N; Kamp DW;
Ghio AJ; Chandel NS; Dean DA; Sznajder JI; Budinger GR. (2007). Ambient particulate matter
accelerates coagulation via an IL"6"dependent pathway. , 117- 2952-61. 121441
Nadel JA; Salem H; Tamplin B; Tokiwa Y. (1965). Mechanism of bronchoconstriction during inhalation of sulfur
dioxide. J Appl Physiol, 20: 164-167. 014846
Newby AC. (2008). Metalloproteinase expression in monocytes and macrophages and its relationship to
atherosclerotic plaque instability. , 28: 2108-14. 191161
Nurkiewicz TR; Porter DW; Barger M; Millecchia L; Rao KM; Marvar PJ; Hubbs AF; Castranova V; Boegehold
MA. (2006). Systemic microvascular dysfunction and inflammation after pulmonary particulate matter
exposure. Environ Health Perspect, 114: 412-419. 088611
Nurkiewicz TR; Porter DW; Hubbs AF; Stone S; Chen BT; Frazer DG; Boegehold MA; Castranova V. (2009).
Pulmonary nanoparticle exposure disrupts systemic microvascular nitric oxide signaling. , 110: 191-203.
191961
Oberdorster G. (1988). Lung clearance of inhaled insoluble and soluble particles. , l: 289-330. 006857
Oortgiesen M; Veronesi B; Eichenbaum G; Kiser PF; Simon SA. (2000). Residual oil fly ash and charged
polymers activate epithelial cells and nociceptive sensory neurons. Am J Physiol, 278: L683-L695. 013998
Owen CA. (2008). Roles for proteinases in the pathogenesis of chronic obstructive pulmonary disease. , 3: 253"
68. 191162
Pardo A; Selman M. (2005). MMP*i: the elder of the family. , 37: 283-8. 191163
Penn A; Murphy G; Barker S; Henk W', Penn L. (2005). Combustion-derived ultrafine particles transport organic
toxicants to target respiratory cells. Environ Health Perspect, 113: 956-963. 088257
Pinkerton KE; Joad JP. (2006). Influence of air pollution on respiratory health during perinatal development.
Clin Exp Pharmacol Physiol, 33: 269-272. 091237
Rahman I; Biswas SK; Kode A. (2006). Oxidant and antioxidant balance in the airways and airway diseases. ,
533: 222-39. 191165
Ramage L; Guy K. (2004). Expression of C-reactive protein and heat-shock protein-70 in the lung epithelial cell
line A549, in response to PM10 exposure. Inhal Toxicol, 16: 447-452. 055640
Randell S; Mercer R; Young S. (1990). Neonatal hyperoxia alters the pulmonary alveolar and capillary structure
of 40-day-old rats. , 136: 1259. 191956
Reed MD; Barrett EG; Campen MJ; Divine KK; Gigliotti AP; McDonald JD; Seagrave JC; Mauderly JL; Seilkop
SK; Swenberg JA. (2008). Health effects of subchronic inhalation exposure to gasoline engine exhaust.
Inhal Toxicol, 20: 1125-1143. 156903
Rengasamy A; Barger MW! Kane E; Ma JKH; Castranova V; Ma JYC. (2003). Diesel exhaust particle-induced
alterations of pulmonary phase I and phase II enzymes of rats. J Toxicol Environ Health A, 66: 153-167.
156907
Rhoden CR; Lawrence J; Godleski JJ; Gonzalez-Flecha B. (2004). N-acetylcysteine prevents lung inflammation
after short-term inhalation exposure to concentrated ambient particles. Toxicol Sci, 79: 296-303. 087969
Rhoden CR; Wellenius GA; Ghelfi E; Lawrence J; Gonzalez-Flecha B. (2005). PM"induced cardiac oxidative
stress and dysfunction are mediated by autonomic stimulation. Biochim Biophys Acta, 1725: 305-313.
087878
Rinderknecht J; Shapiro L; Krauthammer M; Taplin G; Wasserman K; Uszler JM; Effros RM. (1980).
Accelerated clearance of small solutes from the lungs in interstitial lung disease. , 121: 105-117. 191965
Risom L; Mrller P; Loft S. (2005). Oxidative stress-induced DNA damage by particulate air pollution. , 592: 119-
137. 189016
Roberts ES; Richards JH; Jaskot R; Dreher KL. (2003). Oxidative stress mediates air pollution particle-induced
acute lung injury and molecular pathology. Inhal Toxicol, 15: 1327-1346. 156051
Roberts JR; Young S"H; Castranova V; Antonini JM. (2007). Soluble metals in residual oil fly ash alter innate
and adaptive pulmonary immune responses to bacterial infection in rats. Toxicol Appl Pharmacol, 221:
306-319. 097623
Rouse RL; Murphy G; Boudreaux MJ; Paulsen DB; Penn AL. (2008). Soot nanoparticles promote
biotransformation, oxidative stress, and inflammation in murine lungs. Am J Respir Cell Mol Biol, 39:
198-207. 156930
July 2009
5-36
DRAFT-DO NOT CITE OR QUOTE

-------
Ruckerl Ri Greven Si Ljungman Pi Aalto Pi Antoniades Ci Bellander Ti Berglind Ni Chrysohoou Ci Forastiere Fi
Jacquemin Bi von KLot Si Koenig Wi Kuchenhoff Hi Lanki Ti Pekkanen Ji Perucci CAi Schneider Ai
Sunyer Ji Peters A. (2007). Air pollution and inflammation (interleukin-6, Oreactive protein, fibrinogen)
in myocardial infarction survivors. Environ Health Perspect, 115- 1072-1080. 156931
Sakamoto Ni Hayashi Si Gosselink Ji Ishii Hi Ishimatsu Yi Mukae Hi Hogg JCi van Eeden SF. (2007). Calcium
dependent and independent cytokine synthesis by air pollution particle-exposed human bronchial
epithelial cells. Toxicol Appl Pharmacol, 225: 134-141. 096282
Sarin Si Undem Bi Sanico Ai Togias A. (2006). The role of the nervous system in rhinitis. , 118: 999-1016.
191166
Schins RPFi Knaapen AM. (2007). Genotoxicity of poorly soluble particles. Inhal Toxicol, 19 Suppl l: 189-198.
156074
Seaton Ai MacNee Wi Donaldson Ki Godden D. (1995). Particulate air pollution and acute health effects. Lancet,
8943: 176-178. 045721
Sekizawa Si Chen CYi Bechtold AGi Tabor JMi Brie JMi Pinkerton KEi Joad JPi Bonham AC. (2008). Extended
secondhand tobacco smoke exposure induces plasticity in nucleus tractus solitarius second-order lung
afferent neurons in young guinea pigs. , 28: 771-81. 191167
Semmler-Behnke Mi Takenaka Si Fertsch Si Wenk Ai Seitz Ji Mayer Pi Oberdorster Gi Kreyling WG. (2007).
Efficient Elimination of Inhaled Nanoparticles from the Alveolar Region: Evidence for Interstitial Uptake
and Subsequent Reentrainment onto Airways Epithelium. , 115: 728. 156080
Shi Ti Knaapen AMi Begerow Ji Birmili Wi Borm PJi Schins RP. (2003). Temporal variation of hydroxyl radical
generation and 8-hydroxy-2'-deoxyguanosine formation by coarse and fine particulate matter. Occup
Environ Med, 60: 315-321. 088248
Sirivelu MPi MohanKumar SMJi Wagner JGi Harkema JRi MohanKumar PS. (2006). Activation of the stress
axis and neurochemical alterations in specific brain areas by concentrated ambient particle exposure
with concomitant allergic airway disease. Environ Health Perspect, 114: 870-874. 111151
Snipes MB. (1996). Current information on lung overload in nonrodent mammals: contrast with rats. Inhal
Toxicol, 8: 91-109. 076041
Snipes MBi Harkema JRi Hotchkiss JAi Bice DE. (1997). Neutrophil Involvement in the Retention and
Clearance of Dust Intratracheal^ Instilled into the LUNGS of F344/N Rats. , 23: 65"84. 156092
Soberanes Si Panduri Vi Mutlu GMi Ghio Ai Bundinger GRSi Kamp DW. (2006). p53 mediates particulate
matter-induced alveolar epithelial cell mitochondria-regulated apoptosis. Am J Respir Crit Care Med,
174: 1229-1238. 156991
Soberanes Si Urich Di Baker CMi Burgess Zi Chiarella SEi Bell ELi Ghio AJi De Vizcaya-Ruiz Ai Liu Ji Ridge
KMi Kamp DW; Chandel NS; Schumacker PT; Mutlu GM; Budinger GRS. (2009). Mitochondrial complex
Ill-generated oxidants activate ASK1 and JNK to induce alveolar epithelial cell death following exposure
to particulate matter air pollution. J Biol Chem, 284: 2176-2186. 190483
Steerenberg P; Withagen CJ Dalen Wi Dormans Ji Loveren H. (2004). Adjuvant Activity of Ambient Particulate
Matter in Macrophage Activity-Suppressed, N"Acetylcysteine-Treated, iNOS" and IL"4"Deficient Mice.
Inhal Toxicol, 16: 835-843. 087981
Sun Qi Wang Ai Jin Xi Natanzon Ai Duquaine Di Brook RDi Aguinaldo J-GSi Fayad ZAi Fuster Vi Lippmann Mi
Chen Lung Ci Rajagopalan S. (2005). Long-term air pollution exposure and acceleration of
atherosclerosis and vascular inflammation in an animal model. , 294: 3003-3010. 186814
Takizawa Hi Abe Si Okazaki Hi Kohyama Ti Sugawara Ii Saito Yi Ohtoshi Ti Kawasaki Si Desaki Mi Nakahara
Ki Yamamoto Ki Matsushima Ki Tanaka Mi Sagai Mi Kudoh S. (2003). Diesel exhaust particles
upregulate eotaxin gene expression in human bronchial epithelial cells via nuclear factor-kappa B-
dependent pathway. Am J Physiol Lung Cell Mol Physiol, 284: L1055-1062. 157039
Tal TLi Graves LMi Silbajoris Ri Bromberg PAi Wu W', Samet JM. (2006). Inhibition of protein tyrosine
phosphatase activity mediates epidermal growth factor receptor signaling in human airway epithelial
cells exposed to Zn2+. Toxicol Appl Pharmacol, 214: 16-23. 108588
Tamaoki Ji Isono Ki Takeyama Ki Tagaya Ei Nakata Ji Nagai A. (2004). Ultrafine carbon black particles
stimulate proliferation of human airway epithelium via EGF receptor-mediated signaling pathway. Am J
Physiol Lung Cell Mol Physiol, 287: L1127-1133. 157040
Tankersley CGi Shank JAi Flanders SEi Soutiere SEi Rabold Ri Mitzner Wi Wagner EM. (2003). Changes in
lung permeability and lung mechanics accompany homeostatic instability in senescent mice. J Appl
Physiol, 95: 1681-1687. 096363
Tao F> Gonzalez-Flecha Bi Kobzik L. (2003). Reactive oxygen species in pulmonary inflammation by ambient
particulates. Free Radic Biol Med, 35: 327-340. 156111
Tao F> Kobzik L. (2002). Lung macrophage-epithelial cell interactions amplify particle-mediated cytokine
release. Am J Respir Cell Mol Biol, 26: 499-505. 157044
July 2009
5-37
DRAFT-DO NOT CITE OR QUOTE

-------
Tarantini L; Bonzini Mi Apostoli Pi Pegoraro Vi Bollati Vi Marinelli Bi Cantone Li Rizzo Gi Hou Li Schwartz Ji
Bertazzi PAi Baccarelli A. (2009). Effects of particulate matter on genomic DNA methylation content and
iNOS promoter methylation. Environ Health Perspect, 117: 217-222. 192010
Tarantini Li Bonzini Mi Apostoli Pi Pegoraro Vi Bollati Vi Marinelli Bi Cantone Li Rizzo Gi Hou Li Schwartz Ji
Bertazzi PAi Baccarelli A. (2009). Errata to Effects of particulate matter on genomic DNA methylation
content and iNOS promoter methylation . Environ Health Perspect, 117: A143. 192153
Thomson Ei Kumarathasan Pi Goegan Pi Aubin RAi Vincent R. (2005). Differential regulation of the lung
endothelin system by urban particulate matter and ozone. Toxicol Sci, 88: 103-113. 087554
U.S. EPA. (2000). Science Policy Council handbook: risk characterization. 052149
Upadhyay Di Panduri Vi Ghio Ai Kamp DW. (2003). Particulate matter induces alveolar epithelial cell DNA
damage and apoptosis: role of free radicals and the mitochondria. , 29: 180-187. 097370
Valinluck Vi Tsai HHi Rogstad DKi Burdzy Ai Bird Ai Sowers LC. (2004). Oxidative damage to methyl-CpG
sequences inhibits the binding of the methyl-CpG binding domain (MBD) of methyl-CpG binding protein
2 (MeCP2). , 32: 4100-8. 191170
Van Eeden SFi Hogg JC. (2002). Systemic inflammatory response induced by particulate matter air pollution:
the importance of bone-marrow stimulation. J Toxicol Environ Health A, 65: 65:1597-1613. 088111
Van Eeden SFi Tan WCi Suwa Ti Mukae Hi Terashima Ti Fujii Ti Qui Di Vincent Ri Hogg JC. (2001). Cytokines
involved in the systemic inflammatory response induced by exposure to particulate matter air pollutants
(PM10). Am J Respir Crit Care Med, 164: 826-830. 019018
van Eeden SFi Yeung Ai Quinlam Ki Hogg JC. (2005). Systemic response to ambient particulate matter:
relevance to chronic obstructive pulmonary disease. Presented at . 157086
Veronesi Bi Makwana Oi Pooler Mi Chen LC. (2005). Effects of subchronic exposures to concentrated ambient
particles VII Degeneration of dopaminergic neurons in Apo E7- mice. Inhal Toxicol, 17: 235-241. 087481
Veronesi Bi Oortgiesen M. (2001). Neurogenic inflammation and particulate matter (PM) air pollutants.
Neurotoxicology, 22: 795-810. 015977
Veronesi Bi Oortgiesen Mi Carter JDi Devlin RB. (1999). Particulate matter initiates inflammatory cytokine
release by activation of capsaicin and acid receptors in a human bronchial epithelial cell line. Toxicol
Appl Pharmacol, 154: 106-115. 048764
Veronesi Bi Oortgiesen Mi Roy Ji Carter JDi Simon SAi Gavett SH. (2000). Vanilloid (capsaicin) receptors
influence inflammatory sensitivity in response to particulate matter. Toxicol Appl Pharmacol, 169: 66"76.
017062
Veronesi Bi Wei Gi Zeng JQi Oortgiesen M. (2003). Electrostatic charge activates inflammatory vanilloid (VRl)
receptors. Neurotoxicology, 24: 463-473. 094384
Wallenborn JGi Evansky Pi Shannahan JHi Vallanat Bi Ledbetter ADi Schladweiler MCi Richards JHi
Gottipolu RRi Nyska Ai Kodavanti UP. (2008). Subchronic inhalation of zinc sulfate induces cardiac
changes in healthy rats. Toxicol Appl Pharmacol, 232: 69"77. 191171
Wallenborn JGi Kovalcik KDi McGee JKi Landis MSi Kodavanti UP. (2009). Systemic translocation of (70)zinc:
kinetics following intratracheal instillation in rats. Toxicol Appl Pharmacol, 234: 25"32. 191172
Wallenborn JG; McGee John Ki Schladweiler Mette Ci Ledbetter Allen Di Kodavanti Urmila P. (2007). Systemic
translocation of particulate matter-associated metals following a single intratracheal instillation in rats.
J Toxicol Sci, 98: 231-239. 156144
Wan Hi Winton HLi Soeller Ci Tovey ERi Gruenert DCi Thompson PJi Stewart GAi Taylor GWi Garrod DRi
Cannell MBi Robinson C. (1999). Der p 1 facilitates transepithelial allergen delivery by disruption of
tight junctions. J Clin Invest, 104: 123-133. 191903
Wan Ji Diaz-Sanchez D. (2006). Phase II enzymes induction blocks the enhanced IgE production in B cells by
diesel exhaust particles. , 177: 3477-3483. 097399
Wan Ji Diaz-Sanchez D. (2007). Antioxidant enzyme induction: a new protective approach against the adverse
effects of diesel exhaust particles. Inhal Toxicol, 19 Suppl l: 177-182. 156145
Wang Li Pinkerton KE. (2007). Air Pollutant Effects on Fetal and Early Postnatal Development. , 81: 144-154.
179975
Warheit DB; Webb TR; Colvin VL; Reed KL; Sayes CM. (2007). Pulmonary bioassay studies with nanoscale and
fine-quartz particles in rats: toxicity is not dependent upon particle size but on surface characteristics.
Toxicol Sci, 95: 270-280. 090482
Widdicombe J. (2006). Reflexes from the lungs and airways: historical perspective. J Appl Physiol, 101: 628.
155519
Widdicombe J; Lee L"Y. (2001). Airway reflexes, autonomic function, and cardiovascular responses. Environ
Health Perspect, 4- 579-584. 019049
Widdicombe JG. (2003). Overview of neural pathways in allergy and asthma. Pulm Pharmacol Ther, 16: 23-30.
157145
July 2009
5-38
DRAFT-DO NOT CITE OR QUOTE

-------
Win-Shwe TT; Yamamoto Si Fujitani Yi Hirano Si Fujimaki H. (2008). Spatial learning and memory function-
related gene expression in the hippocampus of mouse exposed to nanoparticle-rich diesel exhaust.
Neurotoxicology, 29: 940-947. 190146
Wong SSi Sun NNi Keith L Kweon OBi Foster DEi Schauer James J; Witten ML. (2003). Tachykinin substance
P signaling involved in diesel exhaust-induced bronchopulmonary neurogenic inflammation in rats. , 77-
638-650. 097707
Wright RO; Baccarelli A. (2007). Metals and neurotoxicology. J Nutr, 137: 2809-13. 191173
Xia Ti Korge P; Weiss JN; Li N> Venkatesen ML Sioutas C> Nel A. (2004). Quinones and aromatic chemical
compounds in particulate matter induce mitochondrial dysfunction: implications for ultrafine particle
toxicity. Environ Health Perspect, 112: 1347-1358. 087486
Xiao GG; Nel AE; Loo JA. (2005). Nitrotyrosine-modified proteins and oxidative stress induced by diesel exhaust
particles. Electrophoresis, 26: 280-292. 156164
Yauk C; Polyzos A; Rowan-Carroll A; Somers CM; Godschalk RW! Van Schooten FJ; Berndt ML; Pogribny IP;
Koturbash I; Williams A; Douglas GR; Kovalchuk O. (2008). Germ-line mutations, DNA damage, and
global hypermethylation in mice exposed to particulate air pollution in an urban/industrial location. ,
105: 605-610. 157164
Yin XJ; Ma JYC; Antonini JM; Castranova V; Ma JKH. (2004). Roles of reactive oxygen species and heme
oxygenase-1 in modulation of alveolar macrophage-mediated pulmonary immune responses to listeria
monocytogenes by diesel exhaust particles. Toxicol Sci, 82: 143-153. 087983
Ying Z; Kampfrath T; Thurston G; Farrar B; Lippmann M; Wang A; Sun Q; Chen LC; Rajagopalan S. (2009).
Ambient particulates alter vascular function through induction of reactive oxygen and nitrogen species.
Toxicol Sci, l: 1-36. 190111
Zhang J; Ghio AJ; Chang W', Kamdar 0; Rosen GD; Upadhyay D. (2007). Bim mediates mitochondria-regulated
particulate matter-induced apoptosis in alveolar epithelial cells. FEBS Lett, 581: 4148-4152. 156179
Zhao H; Barger MW! Ma JKH; Castranova V; Ma JYC. (2006). Cooperation of the inducible nitric oxide synthase
and cytochrome P450 1A1 in mediating lung inflammation and mutagenicity induced by diesel exhaust
particles. Environ Health Perspect, 114: 1253-1258. 100996
July 2009
5-39
DRAFT-DO NOT CITE OR QUOTE

-------
Chapter 6. Integrated Health Effects of
Short-Term PM Exposure
6.1. Introduction
This chapter summarizes, reviews, and integrates the evidence of relationships
between short-term exposures to PM and a variety of health-related endpoints.
Cardiovascular and respiratory health effects of short-term exposure to various size
fractions and sources of PM have been examined in an expansive number of epidemiologic,
controlled human exposure and toxicological studies. In addition, there is a large body of
literature evaluating the relationship between mortality and short-term exposure to PM.
The association between PM exposure and central nervous system function has also been
assessed, although far fewer studies are available. The research approaches used to
evaluate health effects of PM exposure are described in Section 1.5 along with advantages
and limitations of the various study types. Chapter 5 provides an overview of the potential
pathophysiological pathways and modes of action underlying the PM-induced health effects
observed in animal and human studies. Evidence from the scientific literature of specific
cardiovascular and systemic effects, respiratory effects, and central nervous system effects
associated with exposure to PM are presented in Sections 6.2, 6.3, and 6.4, respectively.
Evidence of associations between short-term exposure to PM and mortality are described in
Section 6.5. The chapter concludes with a preliminary evaluation of PM-induced health
effects attributable to specific constituents or sources (Section 6.6). More detailed
descriptions of each study evaluated for this assessment are presented in Annexes C, D,
and E.
Findings for cardiovascular and respiratory effects are presented by specific endpoint
or measure of effect, leading from more subtle health outcome measures (e.g., heart rate
variability [HRV]) to the more severe, such as hospitalization for cardiovascular disease.
Conclusions from the 2004 PM AQCD (U.S. EPA, 2004, 056905) are briefly summarized at
the beginning of each section, and the evaluation of evidence from recent studies builds
Note- Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health
and Environmental Research Online) at http • I leu a. gov/her o. HERO is a database of scientific literature used by U.S. EPA in
the process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009
6-1
DRAFT-DO NOT CITE OR QUOTE

-------
upon what was available during the previous review. For each health outcome, results are
summarized for studies from the specific scientific discipline, i.e., epidemiologic, controlled
human exposure, and toxicological studies. The sections conclude with summaries of the
evidence on the various health outcomes and integration of the findings that leads to
conclusions regarding causality based upon the framework described in Chapter 1.
Determination of causality is made for the overall health effect category, such as
cardiovascular effects, with coherence, consistency and plausibility being based upon the
evidence from across disciplines and also across the suite of related health outcomes
ranging from the more subtle health outcomes to cause-specific mortality. In the summary
sections for cardiovascular and respiratory effects, the evidence is summarized and
independent conclusions drawn for relationships with PM2.5, PM10-2.5, and UFPs (Sections
6.2.11 and 6.3.9, respectively). Evidence of central nervous system effects is also divided by
scientific discipline (Section 6.4); however, the lack of data does not allow for informative
summaries of effect by PM metric in discussing CNS effects.
6.2. Cardiovascular and Systemic Effects
6.2.1. Heart Rate and Heart Rate Variability
Heart rate (HR), HRV, and BP are all regulated, in part, by the sympathetic and
parasympathetic nervous systems. Changes in one or more may increase the risk of
cardiovascular events (e.g., arrhythmias, MI, etc.). Decreases in HRV have been associated
with cardiovascular mortality/morbidity in older adults and those with significant heart
disease (TFESC, 1996, 003061).
HRV is measured using electrocardiograms (ECG) and can be analyzed in the time
domain (e.g., standard deviation of all NN intervals [SDNN], square root of the mean
squared successive NN interval differences [rMSSD]), and/or the frequency domain
measured by power spectral analysis (e.g., high frequency [HF], low frequency [LF], ratio of
LF to HF [LF/HF]). SDNN generally reflects the overall modulation of HR by the autonomic
nervous system, whereas rMSSD generally reflects parasympathetic activity and high
frequency variations in HR. Thus, rMSSD is generally well correlated with HF, which also
reflects the parasympathetic modulation of HR. LF is predominately dictated by
sympathetic tone and increased LF/HF indicates sympathoexcitation, which correlates
July 2009
6-2
DRAFT-DO NOT CITE OR QUOTE

-------
overall with decreased overall HRV (SDNN, rMSSD). Thus LF/HF is thought to estimate
the ratio of sympathetic influences on HR to parasympathetic influences.
While HRV is commonly described as being a reflection of vagal and adrenergic input
to the heart, there is clearly a more complex phenomenon reflected in HRV parameters.
Rowan et al. (2007, 191911) provide a review of HRV and its use and interpretation with
respect to air pollution studies. To summarize, HRV indices are excellent measures of
extrapulmonary effects from inhaled pollutants, but the characterization of the acute,
reversible responses to air pollution as being either parasympathetic or sympathetic in
origin, much less predictive of some adverse outcomes such as ventricular arrhythmia, is
relatively unsupported by the clinical literature. This is consistent with the 2004 PM AQCD
(U.S. EPA, 2004, 056905) which stated that there is inherent variability in the
minute "to-minute spectral measurements, but long-term HRV measures demonstrate
excellent day-to-day reproducibility.
The 2004 PM AQCD (U.S. EPA, 2004, 056905) presented limited evidence of PM-
induced changes in HRV. However, findings from epidemiologic, controlled human exposure
and toxicological studies were seemingly contradictory, with reports of both decreases and
increases in HRV following PM exposure. Recent epidemiologic studies have demonstrated
a more consistent decrease in HRV (SDNN and rMSSD), which is supported by several
controlled human exposure studies published since 2003. In these studies, decreases in
HRV were observed among healthy adults following short-term exposures to PM2.5 and
PM10-2.5 CAPs. It is interesting to note that these effects were not observed in adults with
asthma or COPD. The effect of PM on HRV observed in animal toxicological studies
continues to vary greatly, which may be due in part to strain differences in baseline HRV.
6.2.1.1. Epidemiologic Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) reviewed several studies of PM
exposure and HR or HRV and described discrepant findings across studies. Several studies
have investigated the association between acute changes in multiple HRV parameters and
ambient air pollutant concentrations in the U.S., Canada, Europe, Mexico, and Asia.
Features and results of these studies are presented in Table 6-1, and are summarized
below.
In a multicity study, Liao and colleagues (2004, 056590) used data from the fourth
cohort evaluation of the Atherosclerosis Risk in Communities (ARIC) Study (1996-1998).
July 2009
6-3
DRAFT-DO NOT CITE OR QUOTE

-------
The 6,784 subjects were 45-64 yr of age and lived in Washington County, MD, Forsyth
County, NC, or the suburbs of Minneapolis, MN. At each HRV measurement session, each
subject rested comfortably for 15 min in the supine position in a quiet, semi-dark room,
with a constant temperature of 24°C. Then, resting, supine, 5-min beat-to-beat RR- interval
data were collected. All measurements were made between 8:30 a.m. and 12:30 p.m. Linear
regression models, adjusting for multiple covariates (i.e., age, ethnicity, gender, education,
smoking, BMI, cardiovascular medications, presence of coronary heart disease, diabetes,
hypertension, HR, humidity, temperature, and season), were used to examine the change in
HRV associated with PMio, O3, SO2, CO, and NO2 concentrations in the 1-3 days prior to
ECG measurement. Among all subjects, each 11.5 jug/m3 increase in mean daily PM10
concentration 1 day before the ECG measurement was associated with a 0.06 ms2 decrease
in log-transformed HF (95% CI: -0.10 to -0.02) and a 1.03 m decrease in SDNN (95%
CI: -1.64 to -0.42). A smaller non-significant decrease was also observed for log transformed
LF. This reduction in cardiac autonomic control was larger among hypertensive subjects,
suggesting that this group may be susceptible to the effects of PM.
In a study of 4,295 randomly selected participants in the Women's Health Initiative
(WHI), a multicity U.S. study, Whitsel et al. (2009, 191980) found decreases in rMMSD and
SDNN in association with PM10 concentration. The associations were stronger among
participants with diabetes. For example, among those with impaired fasting glucose, the
reduction in rMSSD was 8.3% (-13.9, 2.4) among those with high levels of insulin and 0.6%
(-2.1, 1.6) among those with low levels of insulin. Similar results were observed comparing
high and low levels of insulin resistance.
Timonen et al. (2006, 088747) conducted a multicity panel study of n = 131 elderly
subjects with stable coronary heart disease who lived in 3 European cities (Amsterdam,
Netherlands; Erfurt, Germany! or Helsinki, Finland). They collected ECGs biweekly for six
months in each subject. This analysis, done as part of the ULTRA Study, examined changes
in HRV (resting, paced breathing, supine, and five min beat-to-beat NN intervals)
associated with changes in fixed monitor particulate concentrations (PM2.5, PM10-2.5) with an
emphasis on ultrafine particle counts (UFP; 0.01-0.1 urn particles) and counts of
accumulation mode particles (ACP; 0.1-1.0 um particles). Mixed models adjusting for time
trend, weekday, humidity, barometric pressure, and temperature were first fit to estimate
the change in HRV associated with PM (UFP, ACP, PM2.5, and PM10-2.5) concentrations on
the same and previous four days in each city. Then, in pooled analyses, the most consistent
July 2009
6-4
DRAFT-DO NOT CITE OR QUOTE

-------
results were for LF/HF. During paced breathing, each 10,000 particles/cm3 increase in two
day lagged UFP was associated with a 13.5% decrease in LF/HF (95% CI: -20.1 to -7.1).
Each 1,000 particles/cm3 increase in 1 day lagged ACP was associated with a 7.8% decrease
in LF/HF (95% CI: -13.0 to -0.2). Although not statistically significant, each 10 ug/m3
increase in 2 day lagged mean PM10-2.5 concentration was associated with a 3.3% decrease in
LF/HF (95% CI: -12.7 to 6.1). For PM2.5, however, results were not consistent across cities,
and thus a pooled estimate was not appropriate. PM2.5 was associated with decreased HF
power and increased LF/HF in Helsinki, increased HF power and decreased LF/HF in
Erfurt, and not associated with any HRV metric in Amsterdam. The authors state that
these contrasting city-specific PM2.5 findings do not clearly support a PM/HRV association,
but suggest that effects may be dependent on PM sources and subject characteristics in
each city. In a subsequent analysis, de Hartog et al. (2009, 191904) investigated whether
exposure misclassification, effect modification by medication use, or particle composition
differences across the three cities could explain the result observed. They reported that
1 |ug/m3 increases in traffic related and long range transported PM2.5 were associated with -
0.28 ms (95% CI: -0.57 to 0.01) and -0.05 ms (95% CI: -0.17 to 0.07) changes in SDNN,
respectively, with smaller sized reductions at lags 0, 1, and 3. Further, these source
apportioned particles as well as outdoor PM2.5 were associated with reduced HRV only
among those not taking beta-blockers. There were no consistent patterns for other
medication classes. Thus, effect modification by medication use and particle composition
differences across the three cities may, in part, explain the heterogeneous PM2.5 findings in
the previous analysis (Timonen et al., 2006, 088747).
The association between HRV and short-term increases in PM2.5, PM10-2.5, PM10, other
size fractions and components was also examined in single-city studies (Table 6-1). Among
U.S. and Canadian cities, increases in PM2.5 were generally associated with decreased
SDNN (Adar et al., 2007, 001458; Ebelt et al., 2005, 056907; Fan et al., 2008, 191979;
Luttmann-Gibson et al., 2006, 089794; Park et al., 2005, 057331: Pope et al., 2004, 055238:
Schwartz et al., 2005, 074317) and/or decreased HF power (Adar et al., 2007, 001458;
Baccarelli et al., 2008, 157984; Ebelt et al., 2005, 056907; Luttmann-Gibson et al., 2006,
089794; Park et al., 2005, 057331: Schwartz et al., 2005, 074317). but not in all studies. Two
studies reported increased SDNN associated with PM2.5 concentrations (Riediker et al.,
2004, 056992; Wheeler et al., 2006, 088453). Yeatts et al. (2007, 091266) also reported
increased rMSSD, SDANN5 (standard deviation of the average of normal to normal
July 2009
6-5
DRAFT-DO NOT CITE OR QUOTE

-------
intervals in all 5-min intervals in a 24-h period), and SDNN24HR (standard deviation of
the average of all normal to normal intervals in a 24-h period), and HF power associated
with increased PM2.5 concentrations. Lipsett et al. (2006, 088753) reported significantly
decreased SDNN associated with increases in 2- and 6-h mean PM10 and PM10-2.5
concentrations. Similarly, Yeatts et al. (2007, 091266) reported decreased rMSSD,
SDNN24HR, SDANN5, ASDNN5 (mean of the standard deviation in all 5-min segments of
a 24-h recording), proportion of NN intervals <50 ms apart (pNN50) (7-min and 24-h), and
HF power associated with increased PM10-2.5 concentration. Of those studies examining
HRV associations with particle counts (Adar et al., 2007, 001458; Park et al., 2005, 057331),
only Adar et al. (2007, 001458) found clear evidence of such effects (e.g., decreased SDNN,
LF, HF). Decreased HRV was also associated with increases in ambient mean SC>42~
concentration (LuttmamrGibson et al., 2006, 089794), ambient mean BC concentration,
(Park et al., 2005, 057331; Schwartz et al., 2005, 074317), and traffic generated
particles/pollution (Adar et al., 2007, 001458; Riediker et al., 2004, 056992).
Studies in Asia, Europe, and Mexico have reported decreases in one or several HRV
metrics (see Figure 6"l) associated with increases in PM2.5 concentration or other size
fractions (Chan et al., 2004, 087398; Chuang et al., 2005, 087989; Chuang et al., 2007,
091063; Cardenas et al., 2008, 191900; Folino et al., 2009, 191902; Holguin et al., 2003,
057326; Min et al., 2008, 191901; Riojas-Rodriguez et al., 2006, 156913; Romieu et al., 2005,
086297). However, one study reported no PM-HRV associations (Barclay et al., 2009,
179935). Langrish et al. (2009, 191908) reported that study subjects, who were residents of
Beijing, had decreased HRV when they wore a face mask while walking for 2 h on a
predetermined route. HRV measures were higher when they did not wear the mask. In
contrast, Riojas-Rodriguez et al. (2006, 156913) reported significantly decreased LF and HF
power associated with each 1 ppm increase in CO concentration, but only small
non-significant decreases associated with PM2.5.
HRV studies investigated lagged pollutant concentrations from 2 h to 5 days before
ECG measurement, reporting effects associated with mean pollutant concentrations lagged
as short as 1-2 h, and more consistently lagged 24-48 h. Taken together, these international
and U.S./Canadian studies show decreases in HRV associated with PM2.5 in most studies
that use SDNN, rMSSD or HF power. The effects of PM10-2.5, UFP, and components were
evaluated in fewer studies but associations with decreased HRV (e.g., both time and
frequency measures) were observed. The proportion of studies reporting decreases in HRV
July 2009
6-6
DRAFT-DO NOT CITE OR QUOTE

-------
may be inflated by publication bias (i.e., studies showing little or no effects are not
submitted for publication).
HRV Studies Investigating Specific Mechanisms
Panel studies investigating PM-HRV associations have also been useful in
investigating potential mechanistic pathways by which PM may elicit a cardiovascular
response. A series of analyses using data from the Normative Aging Study (NAS), a cohort
of older men living in the Boston metropolitan area, has also provided mechanistic insights
into the PM/HRV association (Baccarelli et al., 2008, 191959; Chahine et al., 2007, 156327;
Park et al., 2005, 057331; Park et al., 2006, 091245; Park et al., 2008, 156845; Schwartz et
al., 2005, 086296).
Park et al. (2005, 057331) studied the association between short-term increases in
ambient air pollution and changes in HRV using n = 497 males enrolled in the NAS.
Subjects had ECG measurements made during a 4-min rest period between 8^00 a.m. and
1:00 p.m. Using linear regression models, adjusted for age, BMI, fasting blood glucose,
smoking, cardiac medications, room temperature, and season, the association between HRV
metrics and PM2.5, O3, NO2, SO2, CO, BC, and particle number count moving averages (ma)
in the previous 4, 24, and 48 h was examined. The modifying effects of hypertension,
diabetes, ischemic heart disease, and use of hypertensive medications were also estimated.
Of the pollutants examined, only PM2.5 and O3 were associated with reductions in HRV, and
each pollutant's effect appeared independent of the other. Each 8 |ug/m3 increase in mean
PM2.5 concentration in the previous 48 h was associated with a 20.8% decrease in the
component of HRV HF (95% CP -34.2% to -4.6%), with larger effects among subjects with
hypertension, ischemic heart disease (IHD), and diabetes. O3 effects were strongest with
the 4-h ma. The authors state that since BC concentrations were also associated with
adverse changes in HRV, this suggests that traffic pollution may be particularly toxic.
Schwartz et al. (2005, 086296) examined the hypothesis that adverse changes in HRV
due to PM2.5 are mediated by an oxidative stress response among participants in the NAS.
They examined whether the change in the HF component of HRV associated with each
10 jug/m3 increase in 48-h mean PM2.5 was modified by the presence or absence of the allele
for glutathione S-transferase Ml (GSTMl), use of statins, obesity, high neutrophil counts,
higher BP, and/or older age. In subjects without the GSTMl allele and its protection against
oxidative stress, each 10 |ug/m3 increase in 48-h mean PM2.5 concentration was associated
July 2009
6-7
DRAFT-DO NOT CITE OR QUOTE

-------
with a 34% decrease in HF (95% CP -52% to -9%). There was no association among those
with at least one copy of the allele. Obesity and high neutrophil counts also worsened the
effect of PM on HRV regardless of allele.
Park et al. (2006, 091245) investigated whether transition metals may be responsible
for cardiorespiratory effects that are observed in association with PM2.5. Again using the
NAS cohort, they investigated whether subjects with two hemochromatosis (HFE)
polymorphisms associated with increased iron uptake had a smaller decrease in HF HRV
associated with PM than those subjects without either variant. Each 10 ug/m3 increase in
48-h mean PM2.5 was associated with a 31.7% decrease in HF (95% CP -48.1 to -10.3%)
among subjects without either polymorphism, but not among those with the two protective
HFE alleles.
Chahine et al. (2007, 156327) reported a 10.5% reduction in SDNN (95% CP -18.2% to
¦2.2%) associated with each 10 ug/m3 increase in the mean 48-h PM2.5 concentration among
NAS participants without the GSTM1 allele, but only a 2.0% SDNN decrease
(95% CP -11.3%, 8.3%) in those with the allele. This confirmed the PM/HF HRV findings of
Schwartz et al. (2005, 074317). Further, subjects with the long repeat polymorphism in the
HOI promoter had a greater decline in SDNN associated with each 10 jug/m3 increase in
the mean 48-h PM2.5 concentration (-8.5% [95% CP -14.8% to -1.8%) than those with the
short repeat polymorphism in HOI (7.4 % increase [95% CP -8.7% to 26.2%). Again, this
suggests that PM-HRV changes are mediated, in part, by oxidative stress.
Baccarelli et al. (2008, 191959) investigated whether the HRVPM2.5 association was
modified by dietary intakes of methyl nutrients (folate, vitamins B6 and B12, and
methionine) and related gene polymorphisms thought to either confer increased or
decreased risk of CVD among men enrolled in the NAS. Each 10 jug/m3 increase in PM2.5 in
the previous 48 h was associated with -8.8% (95% CP -16.7 to -0.2) and -11.8% (95% CP -
20.8 to -1.8) decreases in SDNN, among those with CC/TT genotypes of the C677T
methylenetetrahydrofolate reductase (MTHFR) polymorphism, and the CC genotype of the
C1420T cytoplasmic serine hydroxymethyltransferase (cSHMT) polymorphism, respectively.
There were no changes among those with CC MTHFR and CC/TT cSHMT. Further, there
were similar HRV reductions in those subjects with lower intakes of B6, B12, and/or
methionine, but no decreases in those with high intakes. Thus these genetic and nutritional
variations in the methionine cycle may modify the PM/HRV association.
July 2009
6-8
DRAFT-DO NOT CITE OR QUOTE

-------
Finally, among those NAS subjects with high chronic lead exposure as measured
using X-ray fluorescence of the tibia, each 7 ug/ma increase in mean PM2.5 concentration in
the previous 48 h was associated with a 22% decrease in HF HRV (95% CP -37.4% to -1.7%)
(Park et al., 2008, 093027). Decreases in HF HRV were also associated with each 2.5 |ug/m3
increase in mean SC>42~ concentration in the previous 48 h (22% decrease [95% CP -40.4% to
1.6%) and each 16 ppb increase in mean O3 concentration in the previous 48 h (38%
decrease [95% CP -54.6% to -14.9%). The authors suggest that these findings are consistent
with an oxidative stress response. Although this series of studies suggest a role of oxidative
stress and perhaps methyl nutrients and related polymorphisms in these short-term
associations of PM2.5 with HRV, replication by other investigators in other cities and in
other populations will aid interpretations of these findings.
Using data from a randomized controlled trial in Mexico City, Romieu et al. (2005,
086297) investigated whether omega-3 fatty acids in fish oil supplements would mitigate
the adverse effects of acute PM exposure on HRV. Residents of a Mexico City nursing home
(N=50, aged >60 yr), subjects were randomized to either 2 g/day of fish oil or 2 g/day of soy
oil. They used random-effects regression models to estimate the change in HRV associated
with mean PM2.5 concentration in the pre-supplementation and supplementation phases. In
the group receiving the fish oil supplement, each 8 ug/m3 increase in 24-h mean total PM2.5
exposure (weighted average of indoor and outdoor PM2.5 based on time activity diaries) was
associated with a 54% reduction (95% CP -72% to -24%) in log transformed HF in the
pre-supplementation phase. However, in the supplementation phase of the trial, each
8 |ug/m3 increase in 24-h mean total PM2.5 concentration was associated with only a 7%
reduction in log transformed HF (95% CP -20% to 7%). Decreases in other HRV parameters
associated with PM2.5 were also muted in the supplementation phase. In the group
receiving the soy oil supplement, the reduction in HF was also smaller in magnitude during
the supplementation phase. However, among those receiving the soy oil supplement, the
differences between the pre-supplementation PM2.5-HF change and the supplementation
PM2.5-HF change were smaller compared to those receiving the fish oil, and were not
statistically significant. Romieu et al. (2008, 156922)also report that omega-3
polyunsaturated fatty acids appear to modulate the adverse effect of PM2.5 based on
measured biomarkers of oxidative response (see Section 6.2.9.1.).
In summary, several analyses of data from the Normative Aging Study have provided
evidence that HRV is modulated by genetic polymorphisms related to oxidative stress
July 2009
6-9
DRAFT-DO NOT CITE OR QUOTE

-------
(Chahine et al., 2007, 156327; Park et al., 2006, 091245; Schwartz et al., 2005, 086296)
preexisting conditions such as diabetes, IHD, and hypertension (Park et al., 2005, 057331;
Whitsel et al., 2009, 191980), or dietary methyl nutrients or related genetic polymorphisms
(Baccarelli et al., 2008, 191959). Another analysis reported that the HRV-PM association
was more pronounced among those with chronic lead exposure (Park et al., 2008, 093027).
In addition, omega-3 fatty acid was found to modulate the effect of PM2.5 on HRV in a
randomized trial conducted in Mexico City (Romieu et al., 2005, 086297) and beta-blocker
use (Folino et al., 2009, 191902) was found to modify the PM-HRV association in a study in
Padua, Italy.
Table 6-1. Characteristics of epidemiologic/panel studies investigating associations between PM and
changes in HRV.

PM Type,
Exposure Lag
Study Subjects
Ambient
Concentration Mean
(SD) **
Recording
Length
SDNN
LF
HF,
rMSSD
LF/HF
U.S. AND CANADIAN STUDIES
Park et al.
(2005, 057331)
PM2.6, 48-h avg
497 men (mean age - 73[7]
yr), Normative Aging Study
Boston MA
24-h: 11.4 (8.0) //g/m3
98th 30.58
4-min
4
1
1
t

PN, 48-h avg

24-h: 28,942 (13,527)
particles/cm3


1
1
1

BC, 48-h avg

24-h: 0.92 (0.47) //g/m3

1
1
1
t
Liao et al.
(2004, 056590)
24-h PM10, lag
1 -day
6784 (mean age - 62[6]
yr), ARIC study: MD, NC,
MN
24.3 //g/m3
5-min
1
1
1

Riedikeret al.
(2004, 056992)
ln-vehicle PM2.6
(mass) 9-h avg
9 state troopers
9-h in-vehicle avg 23 //g/m3
10-min
t

t
1
Schwartz et al.
(2005, 074317)
BC, 24-h
28 (61-89 yr), 12wk
24-h Median: 1.0//g/m3
23-min
1

1
t
PM2.6, 24-h

24-h Median: 10//g/m3

1

1
t

Secondary PM
(estimated), 24-h

1-h Median: -1.7 //g/m3

1

1
t
Yeatts et al.
(2007, 091266)
24-h PM10-2.B
12 adult asthmatics, Chapel
5.3 (2.8)//g/m3
5-min
1
1
1

24-h PM2.6

12.5 (6.0)//g/m3

t
1
t

Wheeler et al.
(2006, 088453)
PM2.6, 4-h avg
18 C0PD, Atlanta, GA
17.8 //g/m3
20-min
t
t
t
t
PM2.6, 4-h avg
12 Ml, Atlanta, GA


1
t
1
1

EC, 4-h avg
18 C0PD, Atlanta, GA
2.3//g/m3

t
Not
presentee
Not
1 presente
Not
id presented
EC, 4-h avg
12 Ml, Atlanta, GA
Not
Not
Not
presented
Dales 2004 PM2.6 24-h avg
(2004, 099036) (personal)
36 CAD patients, Toronto, 19.9 (13.8) //g/m3
Canada
Not described
Luttmann- PM2.6, lag 1-day
Gibson et al. 	
(2006, 089794) Sulfate, lag 1-day
32 (65+ yr), Steubenville
¦ OH
24-h: 19.7 //g/m3
24-h: 6.9 //g/m3
Nonsulfate PM, lag
1 -day
24-h: 10.0 //g/m3
-30-min. I
_L
4
July 2009	6-10	DRAFT - DO NOT CITE OR QUOTE

-------

PM Type,
Exposure Lag
Study Subjects
Ambient
Concentration Mean
(SD) **
R"ordi"B SDNN
Length
HF
LF rMSSD LF/HF

EC, lag 1 -day

24-h: 1.1 /yglm3
t
1 -
Adaret al.
(2007, 001458)
PM2.6, 24-h avg
44 (60+ yr), diesel bus
riders, St. Louis, MO
10.17/yg|m3
98th: 22.43
5-min. 1
4 4 t

BC, 24-h avg

330 ng/m3
4
4 4 t

PC fine

42 particles/cm3
4
4 4 t

PC course

0.02 particles/cm3
t
t t 1
Pope et al.
(2004, 055238)
24-h PM2.6 (FRM),
lag 1-day
88 subjects (65+ yr; 250
p-days), Utah Valley
23.7 (20.2) /yglm3
24-h 4
1
Sullivan et al.
(2005,109418)
PM2.6,1,2, 24-
h avg
21 subjects (65+ yr) with
CVD, Seattle WA
Median:10.7 /yglm3
20-min —>
—


13 subjects (65+ yr) w/out
CVD, Seattle WA

—
—
Lipsett et al.
(2006, 088753)
PM10, lag *
19 CAD patients (65+ yr),
¦ 12 wk fu, Coachella Valley,
CA
31.0 and 46.1 /yglm3
5-min F domain: 4
4 4
PM102.B, lag*
None given
domain 1
1 -

PM2.6, lag*

14 & 23.2/yg/m3
4
1 t
Ebelt et al.
(2005, 056907)
PM10 - 24 h
16 subjects with COPD in
¦ Vancouver, Canada
17 (6)//g/m3
24-h 4
1
PM10-2.E (calculated
from PM10 and
PM2.6 values)
5.6 (3.0) //g/m3
t


PM2.6 - 24-h

11.4 (4.6) //g/m3
98th: 23
1
1

PM2.6 Sulfate - 24-
h outdoor

2.0 (1.1) //g/m3
1
—
Baccarelli et al.
(2008,191959)
PM2.6 - 48 h
549 subjects in Normative
Aging Study and residents
of Boston metropolitan area
Geometric mean (95%
confidence interval)
10.5(10.0,10.9) /yglm3
7 min J,

Fan et al. (2008, PM2.6 personal-1 h 11 crossing guards in New Only change in 1-h PM2.6 24 h	4
191979)	Jersey	reported
Morning shift: 35.2 ± 25.9
Z^g/m3
Afternoon shift:
24.1 ± 22.1 /yglm3
Whitsel et al. PM10	4295 randomly selected Mean ± se	10 second 4	4
(2009,191980)	participants in the WHI Trial 28 ± 0.2 visit 1
24 h	27 ± 0.2 visit 2
27 ± 0.3 visit 3
INTERNATIONAL STUDIES
Timonen et al. UF, lags 0-2 days
(2006, 088747)
AC, lags 0-2 days
Stable CHD patients (65 +
Amsterdam:
5-min
V)
17,300 particles/cm3
(Pooled
n - 37: Amsterdam
Erfurt:
estimates during
n - 47: Erfurt
21,100 particles/cm3
paced breathing
n - 47: Helsinki

presented to the

Helsinki:
right)

17,000 particles/cm3

Amsterdam:


2100 particles/cm3


Erfurt:


1800 particles/cm3


Helsinki:


1400 particles/cm3

July 2009
6-11
DRAFT-DO NOT CITE OR QUOTE

-------

PM Type,
Exposure Lag
Study Subjects
Ambient
Concentration Mean
(SD) **
R"ordi"B SDNN
Length
LF HF'
rMSSD
LF/HF

PM2.6, lags 0-2
days

Amsterdam: 20.0//g/m3
4
t
1


Erfurt: 23.1 //g/m3
Helsinki: 12.7 //g/m3




PMio-2.6, 2-day lag

Amsterdam: 15.3 //g/m3
Erfurt: 3.7 //g/m3
Helsinki: 6.7//g/m3


1
Chan et al.
(2003, 089398)
NCo.02-1 1-4 h
9 adults (19-29 yr) with
lung function impairment
Taipei, Taiwan
23,407(19,836)
particles/cm3
5 min I
1 1
1


10 adults (42-79 yr) with
lung function impairment
25,529 (20,783)
particles/cm3
4
1 1
1


Taipei, Taiwan




Chuang et al.
(2005, 087989)
PM 1.0-0.31-4 h
16 CHD hypertensive
' patients, Taipei Taiwan
37.2 (25.8) //g/m3
5-min I
1 1
t
PM2.E-1.01-4 h
12.6 (7.8)//g/m3
1
1 1
t

PM102.B 1-4 h

14.0 (11.1) //g/m3
1
1 1
t

PM 1.0-0.31-4 h
10 CHD patients, Taipei
¦ Taiwan
26.8 (25.9) //g/m3
1
1 1
-

PM2.E-1.01-4 h
10.9 (8.5)//g/m3
1
1 1
1

PM 10-2.51-4 h

16.4 (10.7) //g/m3
1
1 1
t
Holguin et al.
(2003, 057326)
24-h PM2.6
21 without hypertension
(60-96 yr), Mexico City
37.2 (13.5)//g/m3
5-min
1 1
t


13 with hypertension
(60-88 yr), Mexico City


1 1
t
Romieu et al.
(1995, 089297)
24-h PM2.6 (outdoor
and indoor)
50 nursing home residents
65+ yr, Mexico City
Outdoor: 19.4 (5.7) //g/m3
Indoor: 18.3 (5.8)//g/m3
6-min I
(Indoor PM2.6,
p re-supplement
phase
presented)
1 1

Riojas-Rodrigues
et al. (2006,
156913)
Personal PM2.6
30 IHD patients, Mexico
City
Geometric mean:
46.8 //g/m3
5-min
1 1

Barclay et al. PMio - daily
(2009,179935)
Particle number
count (PNC)-daily
Estimated PM2.6
and PNC
132 subjects with stable
coronary heart failure
Aberdeen, Scotland
Range of daily means: 7.4 24 hs
to 68 //g/m3
Cardenas et al. PM2.6-outdoor
(2008,191900)
PM2.B-indoor
52 subjects (31 women, 21 Median PM2.6 outdoor: 28.3 15 min
men; 20-40 yr), southeast //g/m3
of Mexico City
Median PM2.6 indoor: 10.8
//g/m3
De Hartog et al. PM2.6 outdoor- 24
(2009,191904) h
Amsterdam, Netherlands
(37 subject)
Erfurt, Germany (47
subjects)
Helsinki, Finland (47
subjects)
Median //g/m3
Amsterdam: 16.7
Erfurt: 16.3
Helsinki: 10.6
5 min
July 2009
6-12
DRAFT-DO NOT CITE OR QUOTE

-------
PM Type,
Exposure Lag
Study Subjects
Ambient
Concentration Mean
(SD) **
Recording
Length
SDNN
LF
HF,
rMSSD
LF/HF
Folino et al. PMio24 h
(2009,191902)
	 PM2.6 24 h
PM0.26 24 h
39 subjects (36 male, 3
female; mean age - 60 yr)
Padua Italy
PM10
Mean ± std //g|m3
Summer46.4 ±
16.1 Winter73.0 ±
30.9Spring38.3 ±
15.4PM2.BMean ± std
/yg/m3summer33.9 ±
12.7Winter62.1 ±
27.9Spring30.8 ±
14.0PMo.2BMean ± std
/yg/m3Summer17.6 ±
7.5Winter30.5 ±
17.4Spring18.8 ± 10.8
24 h
Langrish et al.
(2009,191908)
PM2.b- personal
2 h
PNCpersonal
2 h
15 young healthy volunteers PM2.6 with mask: Mean -
(median age - 28 yr)	86 /yg/m3Without mask:
Beijing	Mean - 140/yg/m3PNC
with mask:Mean - 23,379
particles/cm3PNC without
mask:Mean - 24,184
particles/cm3
24 h
4	4
withoutmask withoutmask
Min et al.
(2008,191901)
PM10
12 h
1349 subjects (596 males;
mean age - 44 yr)
Korea
Mean ± std:
1-h avg
33.244 ± 19.017
5 min
Notes: Increases (|), decreases (J and no effects ( >) in HRV associated with PM concentration are indicated. Statistical significance was not necessary to categorize an effect as an increase or
decrease.
For time domain measures moving average lags up to 24-h were explored. For frequency domain measures lags of 2-h, 4-h and 24-h were explored.
** All concentrations are means, unless otherwise noted.
6.2.1.2. Controlled Human Exposure Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) cited one study in which HRV
indicators of parasympathetic activity increased relative to filtered air control following a
2-h exposure with intermittent exercise to PM2.5 CAPs (average concentration 174 (Jg/m3) in
both healthy and asthmatic volunteers (Gong et al., 2003, 042106). This effect was observed
immediately following exposure and at 1 day post-exposure, but not at 4 h post-exposure.
Although not statistically significant, HRV (total power) increased following exposure to
filtered air and decreased following exposure to CAPs. More recent controlled human
exposure studies are described below.
CAPs
Two new studies have evaluated the effect of PM2.5 CAPs (2-h exposures to
concentrations of20"200 |jg/m3) on HRV in elderly subjects (Devlin et al., 2003, 087348;
Gong et al., 2004, 087964). In both studies, subjects experienced significant decreases in
HRV following exposure to CAPs relative to filtered air exposures. Interestingly, Gong et al.
(2004, 087964) found that decreases in HRV were more pronounced in healthy older adults
July 2009
6-13
DRAFT-DO NOT CITE OR QUOTE

-------
than in those with COPD. In another study, healthy and asthmatic adults were exposed to
thoracic coarse CAPs (average concentration 157 (Jg/m3) for 2 h with intermittent exercise
(Gong et al., 2004, 055628). HRV was not affected immediately following the exposure, but
decreased in both groups at 4 and 22 h after the end of the exposure, with greater responses
observed in non-asthmatics. In a recent study among healthy adults exposed for 2 h with
intermittent exercise to thoracic coarse CAPs (average concentration 89 |Jg/m3, MMAD 3.59
|jm, Chapel Hill, NC), Graff et al. (2009, 191981) observed a significant decrease in overall
HRV (SDNN) at 20 h post-exposure, although no other measures of HRV were affected.
Using a similar study design, the same laboratory also evaluated the effect of ultrafine
CAPs (average concentration 49.8 jJg/m3, <0.16 |Jm in diameter) on various HRV parameters
(Samet et al., 2009, 191913) Relative to filtered air, HF and LF power were both observed to
increase 18 h following exposure to UF CAPs (36-42% increase per 105 particles/cm3).
Exposure to UF CAPs, expressed as mass concentration, was not associated with changes in
HF power, and time domain parameters of HRV did not differ between CAPs and filtered
air in the 24 h following exposure. Gong et al. (2008, 156483) also recently evaluated
changes in HRV following controlled human exposures to UF CAPs and reported a small
and transient decrease in LF power (p < 0.05) among healthy (n = 17) and asthmatic (n =
14) adults 4 h after the completion of a 2-h exposure with intermittent exercise in Los
Angeles (average concentration 100 jJg/m3, average particle count 145,000/cm3). No other
measure of HRV was shown to be significantly affected by exposure to UF CAPs. In one of
the largest studies of controlled human exposures to CAPs conducted to date, Fakhri et al.
(2009, 191914) evaluated changes in HRV among 50 adult volunteers during 2-h exposures
to fine CAPs (127 [Jg/m31 and O3 (114 ppb), alone and in combination. Neither exposure to
CAPs nor O3 resulted in any significant changes in HRV relative to filtered air. However,
trends were observed suggesting a negative concentration response relationship between
CAPs concentration and SDNN, rMSSD, HF power and LF power when subjects were
concomitantly exposed to O3.
Diesel Exhaust
In a double-blind, crossover, controlled-exposure study, Peretz et al. (2008, 156855)
exposed three healthy adult volunteers and 13 adults with metabolic syndrome while at
rest to filtered air and two levels of diluted DE (reported PM2.5 concentrations of 100 and
200 |jg/m3) in 2-h sessions. HRV parameters were assessed prior to exposure, as well as at
July 2009
6-14
DRAFT-DO NOT CITE OR QUOTE

-------
1, 3, 6 and 22 h following the start of exposure, and included both time domain (SDNN and
rMSSD) and frequency domain parameters (HF power, LF power, and the LF/HF ratio). The
authors observed an increase in HF power and a decrease in LF/HF 3 h after the start of
exposure to 200 (Jg/m3 relative to filtered air. Although these changes were statistically
significant (p < 0.05) the effects were not consistent among the study subjects. No other
significant effect of DE on HRV was observed at either concentration or time point. The
authors attributed the lack of consistent effects to the small and non-homogeneous
population and the timing of measurement. There was no difference in either baseline or
diesel-induced changes in HRV parameters between normal individuals and patients with
metabolic syndrome, although the number of normal individuals was quite small (n = 3). It
is unclear if patients with metabolic syndrome were taking any medications.
Model Particles
Several additional recent controlled human exposure studies have evaluated the effect
of laboratory generated particles on HRV in healthy and health-compromised individuals.
In a random order crossover controlled human exposure study, Routledge et al. (2006,
088674) examined the effects of ultrafine elemental carbon particles (50 (Jg/m3) alone and in
combination with 200 ppb SO2 on HRV among 20 healthy older adults (age 56-75), as well
as 20 older adults with coronary artery disease (age 52-74). Five minute recordings of HRV
data were obtained prior to and immediately following the 1-h exposure, as well as 3 h
post-exposure. In healthy subjects, exposure to carbon particles resulted in small increases
in RR-interval, SDNN, rMSSD, and LF power immediately following exposure compared to
filtered air control. At 3 h post-exposure, there were no significant differences in HRV
measures between carbon particle and filtered air exposures. Conversely, S02"induced
decreases in HRV were observed at 3 h, but not immediately following exposure.
Concomitant exposure to carbon particles and SO2 followed a pattern similar to that
observed with SO2 alone, but did not reach statistical significance. Subjects with coronary
artery disease did not experience any significant changes in HRV following exposure to
either pollutant. The authors postulated that this lack of effect may be due to differences in
medication between the two groups, as 70% of subjects with stable angina reported using 6
blockers, which are known to increase cardiac vagal control. The lack of any significant
effects on HRV following exposure to carbon particles is an important finding, as it provides
evidence to suggest that the health effects observed following exposure to PM may be due to
July 2009
6-15
DRAFT-DO NOT CITE OR QUOTE

-------
particle constituents other than carbon, or to reactive species found on the surface of the
particle. These findings are in agreement with those of Zareba et al. (2009, 190101) who
reported small and variable changes in HRV among a group of healthy adults following
exposure to UF elemental carbon. While exposure both at rest and during exercise to 10
(jg/m3 UF carbon resulted in an increase in time domain parameters (rMSSD and SNDD),
no such effect was observed following exposure to a higher concentration of UF carbon (25
(jg/m3) in the same subjects. A recent pilot study reported no effect of exposure to elemental
carbon and ammonium nitrate particles (250-300 (Jg/m3) on HRV parameters in 5 adults
with allergic asthma (Power et al., 2008, 191982). However, when the exposure occurred
concomitantly with O3 (0.2 ppm), subjects were observed to experience significant changes
in both time and frequency HRV parameters (decreases in SDNN and normalized HF and
LF power). These observations should be considered very preliminary as the study was
limited by a small sample size (n = 5) and did not evaluate the effect of exposure to O3
without particles. However, these findings are in agreement with the previously described
study of CAPs and O3 conducted by Fakhri et al. (2009, 191914). In addition to the studies
of laboratory generated carbon described above, Beckett et al. (2005, 156261) used ZnO as a
model particle and exposed twelve resting, healthy adults for 2 h to filtered air and
500 (Jg/m3 in the ultrafine (40.4 ±2.7 11m) and fine (291.2 ± 20.2 11m) modes. Neither
ultrafine nor fine ZnO produced a significant change in any time or frequency domain
parameter of HRV.
Summary of Controlled Human Exposure Study Findings for Heart Rate Variability
The results of several new controlled human exposure studies provide limited
evidence to suggest that acute exposure to near ambient levels of PM maybe associated
with small changes in HRV. Changes in HRV parameters, however, are variable with some
showing increased parasympathetic activity relative to sympathetic activity and others
showing the opposite. Although a direct comparison between younger and older adults has
not been made, PM exposure appears to result in a decrease in HRV more consistently in
healthy older adults (Devlin et al., 2003, 087348; Gong et al., 2004, 087964).
6.2.1.3. Toxicological Studies
Toxicological studies that examined HR and HRV are presented in the 2004 PM
AQCD (U.S. EPA, 2004, 056905) and overall demonstrated differing responses, which were
July 2009
6-16
DRAFT-DO NOT CITE OR QUOTE

-------
collectively characterized as providing limited evidence for PM-related cardiovascular
effects. The studies described that reported HR or HRV effects following PM exposure were
conducted with a variety of particle types (CAPs, diesel, ROFA, metals), exposure methods
(inhalation and IT instillation), and doses (100-3,000 |Jg/m3 for inhalation! up to 8.3 mg/kg
for IT instillation).
CAPs
Two groups of SH rats exposed to CAPs in Tuxedo, NY for 4 h (single-day mean PM2.5
concentrations 80 and 66 |Jg/m3; 2/2001 and 5/2001, respectively) demonstrated decreased
HR when exposure groups were combined that returned to baseline values when exposure
ceased (Nadziejko et al., 2002, 087460). Fine or ultrafine H2SO4 exposure (mean
concentration 225 and 468 (Jg/m3) did not induce any HR effects. Another study
demonstrated a trend toward increased HR in WKY rats following a 1- or 4-day PM2.5 CAPs
exposure in Yokohama City, Japan (4.5 h/day! 5/2004, 11/2004, 9/2005), but the correlation
between change in HR and cumulative PM mass collected was not significant (Ito et al.,
2008, 096823). Increased HR was observed in SH rats exposed to PM2.5 CAPs for two 5-h
periods during the spring (mean mass concentration 202 (Jg/m3) in a suburb of Taipei,
Taiwan (Chang et al., 2004, 055637). The response was less prominent in the summer
(mean mass concentration 141 (Jg/m3), despite the number concentrations being similar for
the two seasons (2.30 x 105 and 2.78 x 105particles/cm3, respectively).
For HRV, decreased SDNN was observed in SH rats exposed to PM2.5 CAPs (mean
mass concentration 202 (Jg/m3! mean number concentration 2.30 x 105 particles/cm3) for two
5-h periods separated by 24 h (Chang et al., 2005, 088662). Each of the four animals served
as their own control and the estimated mean PM effects for the SDNN decreases during
exposure were 85-60% of baseline. CAPs effects on rMSSD were less remarkable. In a study
of Tuxedo, NY, PM2.5 CAPs, no acute changes in rMSSD or SDNN were observed in either
ApoE '" or C57 mice when the 48-h time period postexposure was evaluated (6 h/day x 5
day/wk! mean mass concentration over 5-month period 110 (Jg/m3) (Chen and Hwang, 2005,
087218).
Diesel Exhaust
Anselme et al. (2007, 097084) used a MI model of CHF where the left anterior
descending coronary artery of WKY rats was occluded to induce ischemia. After 3 months of
recovery, rats were exposed to diesel emissions for 3 h (PM concentration 500 |Jg/m3; mass
July 2009
6-17
DRAFT-DO NOT CITE OR QUOTE

-------
mobility diameter 85 nm! NO2 1.1 ppm! CO 4.3 ppm) and decreases in rMSSD were
observed during the first 2 h of the exposure, which returned to baseline values for the last
hour of exposure. Healthy rats also demonstrated decreased rMSSD when measured over
the entire exposure period.
Model Particles
In WKY rats exposed to ultrafine carbon particles (mass concentration 180 |Jg/m3;
mean number concentration 1.6 x 107 particles/cm3) for 24 h, HR increased and SDNN
decreased during particle inhalation (Harder et al., 2005, 087371). These measures
returned to baseline values during the recovery period. This study provides evidence that
ultrafine carbon exerts its effects through changes in autonomic nervous system mediation,
as the HR and HRV responses occurred quickly after exposure started and pulmonary
inflammation was only observed at the 24-h time point (and not at 4 h). SH rats exposed to
ultrafine carbon particles under the same conditions (mass concentration 172 (Jg/m3; mean
number concentration 9.0 x 106 particles/cm3) demonstrated similar responses, albeit not
until recovery days 2 and 3 (Upadhyay et al., 2008, 159345).
A model of premature senescence has been developed by Tankersley et al. (2003,
053919), using aged AKR mice whose body weight abruptly declines ~5 wk prior to death
and is accompanied by deficiencies in other vital physiological function including HR and
temperature regulation. When exposed to CB (CB; mean concentration 160 |Jg/m3; 3 h/day x
3 day), terminal senescent mice responded with robust cardiovascular effects, including
bradycardia and increased rMSSD and SDNN (Tankersley et al., 2004, 094378). SDNN and
LF/HF were also increased in healthy senescent mice exposed to CB. These studies indicate
that HR regulatory mechanisms are altered in susceptible mice exposed to PM (sympathetic
and parasympathetic changes in healthy senescent mice and increased parasympathetic
influence in terminally senescent mice), which may translate into lowered homeostatic
competence in these animals. Results from the near-terminal group should be interpreted
with caution, as only 3 mice were in this group.
Subsequent research with a similar exposure protocol (mean CB concentration
159 (Jg/m3) used C57BL/6J and C3H/HeJ mice to determine whether an acute PM challenge
can modify HR regulation in two mice strains with differing baseline HR (Tankersley et al.,
2007, 097910) There were no CB-specific effects on HR or HRV in C3H/HeJ compared to
C57BL/6J mice (average HR -80 bpm lower than C3H/HeJ at baseline). Administration of a
July 2009
6-18
DRAFT-DO NOT CITE OR QUOTE

-------
sympathetic antagonist (propanolol) to C57BL/6J mice prior to CB exposure resulted in
elevated HR and decreased rMSSD compared to air during the last 2-h of exposure,
indicating withdrawal of parasympathetic tone. There may be differences in regional
particle deposition based on strain-specific breathing patterns that may affect HR and HRV
responses. However, this study revealed that inherent autonomic tone, which is genetically
varied between these mouse strains, may affect cardiovascular responses following PM
exposure. In extrapolating these results to humans, individual variation in genetic factors
likely plays some role in PM-induced adjustments in HR control via the ANS.
A recent study in mice (C3H/HeJ, C57BL/6J, and C3H/HeOuJ) examined the effects of
a 2-h O3 (mean concentration 0.584 ppm) pretreatment followed by a 3-h exposure to CB
(mean concentration 536 |jg/m3) on HR and HRV measures (Hamade et al., 2008, 156515)
HR decreased to the greatest extent during O3 pre-exposure for all strains that were then
exposed to CB. The percent change in SDNN and rMSSD were increased in C3H mice
during O3 pre-exposure and CB exposure compared to the filtered air group! however, these
HRV parameters gradually decreased over the duration of the experiment and appeared to
be O3 dependent. Together, these findings indicate that increases in parasympathetic tone
and/or decreases in sympathetic input may explain the observed bradycardia. In a subset of
all mice pre-exposed to O3, rMSSD remained significantly elevated during the CB exposure
compared to filtered air. The results from this study confirm what was observed in
Tankersly et al. (2007, 097910) in that genetic determinants affect HR regulation in mice
with exposure to air pollutants.
Summary of Toxicological Study Findings for Heart Rate and Heart Rate Variability
Both increases and decreases in HR have been observed in rats or mice following PM
exposure. Fine or ultrafine H2SO4 did not result in HR changes in SH rats. Similarly,
decreased SDNN was reported for ultrafine CAPs exposure and lowered rMSSD was
observed with diesel exposure. In near-terminal senescent mice, HRV responses were
robust following CB exposure and represented increased parasympathetic influence. Strain
differences in baseline HR and HRV likely contribute to PM responses. HRV changes with
preexposure to O3 and CB appeared to be O3 dependent, although rMSSD remained
elevated during PM exposure.
Source Apportionment and PM Components
July 2009
6-19
DRAFT-DO NOT CITE OR QUOTE

-------
An additional analysis of CAPs data (Chen and Hwang, 2005, 087218; Hwang et al.,
2005, 087957) was conducted to link short-term HR and HRV effects to major PM source
categories using source apportionment methodology (Lippmann et al., 2005, 087453). The
source categories were: (l) regional secondary SC>42~ comprised of high S, Si, and OC (mean
63.41 |jg/m3); (2) resuspended soil characterized by high concentrations of Ca, Fe, Al, and Si
(mean concentration 5.88 |Jg/m3); (3) fly ash emissions from power plants burning residual
oil in the eastern U.S. and containing high levels of V, Ni, and Se (mean concentration
1.53 |Jg/m3); and (4) motor vehicle traffic and other unknown sources (34.92 (Jg/m3)
(Lippmann et al., 2005, 087453). Exposures occurred from 9:00 a.m. to 3:00 p.m., 5 days/wk
for 5 months. PM2.5 mass was associated with a daily interquartile change of -4.1 beat/min
HR during exposure in ApoE"'" mice1 and a similar magnitude of effect was observed with
resuspended soil (-4.5 beat/min). Resuspended soil was also associated with a HR increase
in the afternoon post-exposure (2.6 beat/min); the secondary SO r factor was linked to
lowered HR in the same period (-2.5 beat/min). A 6.2% increase in rMSSD collected in the
afternoon post-exposure was associated with the residual oil factor, compared to a 5.6% and
2.4% decrease in rMSSD at night for secondary SC>42~ and PM2.5 mass, respectively.
Resuspended soil was associated with a 4.3% increase in rMSSD the night following CAPs
exposure. The residual oil and secondary SO r categories showed similar statistically
significant parameter estimates for SDNN as rMSSD.
Recent studies of ECG alterations in mice have indicated a role for PM-associated Ni
in driving the cardiovascular effects. Lippman et al. (2006, 091165) presented a posthoc
analysis of daily variations in PM2.5 CAPs (mean concentration: 85.6 |Jg/m3; 7/21/ 2004-
1/12/2005; Tuxedo, NY) and changes in cardiac dynamics in ApoE"'" mice. On the 14 days
that the exposed mice had unusually elevated HR, Ni, Cr, and Fe comprised 12.4% of the
PM mass, compared to only 1.5% on the other 89 days. Back trajectory analyses indicated
high-altitude winds from the northwest that did not traverse population centers and
industrial areas except the Sudbury Ni smelter in Ontario, Canada. On the 14 days that
1 Atherosclerosis and related pathways has been studied primarily in the Apolipoprotein E (ApoE) knockout mouse. Developed
by Nobuyo Maeda's group in 1992 (Piedrahita et al., 1992, 156868; Zhang et al., 1992, 157180). the ApoE7" mouse and related
models have become the workhorse of atherosclerosis research over the past 15 years. The ApoE molecule is involved in the
clearance of fats and cholesterol. When ApoE (or the LDL receptor) is deleted from the genome, mice develop severely
elevated lipid and cholesterol profiles; ApoE7' mice on a high-fat ("Western") diet exhibit cholesterol levels exceeding 1000
mgdL (normal is -150 mgdL) (Huber et al., 1999, 156575; Moore et al., 2005, 156780). As a result, the lipid uptake into the
vasculature is increased and the atherosclerotic process is dramatically hastened. Furthermore, the LDLs in ApoE'/' mice are
highly susceptible to oxidation (Hayek et al., 1994, 156527). which may be a crucial event in the air pollution-mediated
vascular changes. However it should be noted that this model is primarily one of peripheral vascular disease rather than
coronary artery disease.
July 2009
6-20
DRAFT-DO NOT CITE OR QUOTE

-------
high HR was observed, the HR elevation lasted for two days, but only the current day CAPs
concentration was statistically significant. SDNN decreases were statistically significant for
all 3 lags (0, 1, 2 days). The GAM regression analysis showed that only Ni produced a
statistically significant effect for HR and SDNN.
6.2.2. Arrhythmia
Epidemiologic and toxicological studies presented in the 2004 PM AQCD (U.S. EPA,
2004, 056905) provided some evidence of arrhythmia following exposure to PM. However, a
positive association between PM and ventricular arrhythmias among patients with
implantable cardioverter defibrillators was only observed in one study conducted in Boston,
MA, while toxicological studies reported arrhythmogenesis in rodents following exposure to
ROFA, DE (DE), or metals. Recent epidemiologic studies have confirmed the findings of
PM-induced ventricular arrhythmias in Boston, MA, and have also reported increases in
ectopic beats in studies conducted in the Midwest and Pacific Northwest regions of the U.S.
In addition, two studies from Germany have demonstrated positive associations between
traffic and combustion particles and changes in repolarization parameters among patients
with ischemic heart disease. Findings of recent toxicological studies are mixed, with both
demonstrated decreases and increases in frequency of arrhythmia following exposure to
CAPs.
6.2.2.1. Epidemiologic Studies
Studies of Arrhythmias Using Implantable Cardioverter Defibrillators
One study reviewed in the 2004 PM AQCD assessed the effect of short-term
fluctuations in PM2.5 on ventricular arrhythmias and several recent studies examining this
relationship have been conducted. Ventricular ectopy and arrhythmia include ventricular
premature beats (VPBs), ventricular tachycardia (VT), and ventricular fibrillation (VF). As
their name implies, VPBs are early-occurring depolarizations of the right or left ventricle.
VT refers to the succession of 3 or more VPBs at a rate of at least 100 beats/min while VF is
characterized by the total absence of properly formed QRS complexes and P waves.
Ventricular arrhythmia is commonly associated with myocardial infarction, heart failure,
cardiomyopathy, and other forms of structural (e.g., valvular) heart disease.
Pathophysiologic mechanisms underlying this established cause of sudden cardiac death
July 2009
6-21
DRAFT-DO NOT CITE OR QUOTE

-------
include dysfunction of the cardiac ion channels, gap junctions, and autonomic nervous
system.
Previously, Peters et al. (2000, 011347) conducted a pilot study in Boston, MA to
examine the association between short-term changes in ambient air pollutant
concentrations and increased risk of ventricular arrhythmias, among a cohort of patients
with implantable cardioverter defibrillators (ICD). ICDs continuously monitor subject's HR
and rhythm and upon detection of an abnormal rhythm (i.e., rapid HR), they can be
programmed to deliver pacing and/or shock therapy to restore normal sinus rhythm. Those
abnormal rhythms that are most severe or rapid are assumed to be due to ventricular
tachycardia or ventricular fibrillation (i.e., life-threatening arrhythmias), and are thus
treated with electric shock. These ICD devices also store information on each abnormal
rhythm detected, including the date, time, and therapy given. Thus, using the date and
time of those arrhythmias resulting in electric shock, Peters et al. (2000, 011347) reported
an increased risk of ICD shock associated with mean NO2 concentration in the previous two
days. Among subjects with frequent events (10 or more during 3 yr of followup) an
increased risk of ICD shock was also associated with interquartile range increases in CO,
NO2, PM2.5, and BC in the previous 2 days. Several studies were conducted to confirm these
findings. The study characteristics, as well as the reported effect estimates and 95% CI
associated with each PM metric, are shown in Table 6-2.
Dockery et al. (2005, 078995; 2005, 090743) conducted a follow-up study of n = 203
ICD patients living in eastern Massachusetts and followed subjects for a longer period of
time (up to 7 yr). They were the first to review the ECG, classify each ICD-detected
arrhythmia (e.g., ventricular arrhythmia, ventricular fibrillation, atrial tachycardia, sinus
tachycardia, etc.), and include only ventricular arrhythmias (ventricular fibrillation or
ventricular tachycardia! excluding supraventricular arrhythmias). In single pollutant
models using generalized estimating equations increased risks of confirmed ventricular
arrhythmias were associated with IQR increases in every pollutant (PM2.5, BC, SOr , NO2,
SO2, O3, and particle number count). Among those with a prior ventricular arrhythmia in
the past three days, interquartile range increases in 2 calendar day mean PM2.5, NO2, SO2,
CO, O3, SO r , and BC concentrations were all associated with significant and markedly
higher risks of ventricular arrhythmia than among those without a prior arrhythmia. The
pollutants associated with increased risk of ventricular arrhythmia implicate traffic
pollution.
July 2009
6-22
DRAFT-DO NOT CITE OR QUOTE

-------
Rich et al. (2005, 079620) conducted a case-crossover analysis of these same data to
investigate moving average pollutant concentrations lagged <48 h. They reported an
increased risk of ventricular arrhythmia associated with mean PM2.5 and O3 concentrations
in the 24 h before the arrhythmia. Each pollutant effect appeared independent in two
pollutant models. In single pollutant models, NO2 and SO2 were associated with increased
risk, but when included in two pollutant models with PM2.5, only PM2.5 remained associated
with increased risk. They did not, however, find evidence of a more acute arrhythmic
response to pollution (i.e., larger risk estimates associated with moving averages <24-h
before arrhythmia detection).
Rich et al. (2006, 089814) conducted another case-crossover study in the St. Louis,
MO metropolitan area. Using the same methods as in Boston, they reported increased risk
of ventricular arrhythmia associated mean SO2 concentration in the 24-h before the
arrhythmia, but not PM2.5 (in single-pollutant models). Again, they found no evidence of an
arrhythmic response with moving average pollutant concentrations <24-h before the
arrhythmia.
In Vancouver, Canada, Vedal et al. (2004, 055630) did not find increased risk of ICD
shocks associated with increases in any pollutant concentration (PM10, O3, SO2, NO2, and
CO), after adjusting for temporal trends, temperature, relative humidity, wind speed,
barometric pressure, and proportion of hours with rain. Secondary analyses among those
subjects with two or more discharges per year, and analyses stratified by season were also
null for PM10, although an association with SO2 (lag 2 days) was observed. A case crossover
analysis of these same data examining additional particle pollutant concentrations
available for a shorter time frame (e.g., PM2.5, SO42-, EC, and OC) also found no increased
risk of ICD shock associated with any pollutant. Rich et al. (2004, 055631) did not use the
time-stratified control selection procedure. They used an ambi-directional approach, taking
control periods 7 days before and after the case day when pollution data was available.
The largest ICD study to date examined the risk of ventricular arrhythmias
associated with increases in the daily concentration of numerous particulate and gaseous
pollutants in Atlanta, GA(Metzger et al., 2007, 092856) (see Table 6-2 for specific pollutants
evaluated). However, they also did not find significant or consistently increased risk of a
ventricular arrhythmia associated with any IQR increase in mean daily particulate or
gaseous pollutant concentration at any lag examined.
July 2009
6-23
DRAFT-DO NOT CITE OR QUOTE

-------
Ljungman et al. (2008, 180266) conducted a similar study, using case-crossover
methods, on 211 ICD patients in Gotheburg and Stockholm, Sweden from 2001-2006. They
investigated the triggering of confirmed ventricular arrhythmias by ambient PMio and NO2
concentrations, and reported increased relative odds of ventricular arrhythmia associated
with each 10 ug/ma increase in the 2-h ma PM10 concentration (OR = 1.22 [95% CP
1.00-1.51]), with a smaller non-significant risk associated with each 10.3 jug/m3 increase in
the 24-h ma PM10 concentration (OR = 1.23 [95% CP 0.87-1.73]). The NO2 and PM2.5 effect
estimates were much smaller and not statistically significant. Effect estimates were larger
for events occurring near the air pollution monitors, and larger in Gothenberg than
Stockholm.
Albert et al. (2007, 156201), although not investigating associations with ambient
pollution, conducted a case-crossover study of the association between ventricular
arrhythmia and traffic exposure in the hours before the arrhythmia. They reported an
increased risk of ventricular arrhythmia associated with traffic exposure or driving in the
previous hour. They hypothesized that this increased risk was due to either a stress
response from being in a car in heavy traffic, or from traffic-generated air pollution, or a
combination of both.
Although ICDs detect and treat potentially life-threatening ventricular arrhythmias,
other arrhythmias including episodes of paroxysmal atrial fibrillation (AF) may also be
detected. AF is the most common type of arrhythmia. In this condition, ectopic electrical
impulses arising in the atria or pulmonary veins, i.e., outside their normal anatomic origin
(the sinoatrial node), can result in atrioventricular dilatation, dysfunction, and/or
thromboembolism. Despite being common, clinical and subclinical forms of AF are
associated with reduced functional status and quality of life. Moreover, the arrhythmia
accounts for a large proportion of ischemic stroke (Laupacis et al., 1994, 190901;
Prystowsky et al., 1996, 156031) and is a strong risk factor for congestive heart failure (Roy
et al., 2009, 190902), contributing to both cardiovascular disease (CVD) and all-cause
mortality (Kannel et al., 1983, 156623).
In an ancillary case-crossover analysis of data from the Boston ICD study described
above, Rich et al. (2006, 088427) identified 91 confirmed episodes of paroxysmal AF among
29 subjects. In single pollutant models, they reported a significantly increased risk of AF
associated with mean O3 and PM2.5 concentrations in the hour before the arrhythmia and
BC concentration in the 24 h before the arrhythmia (Table 6-2).
July 2009
6-24
DRAFT-DO NOT CITE OR QUOTE

-------
Table 6-2. Studies of ventricular arrhythmia and ambient PM concentration, in patients with
implantable cardioverter defibrillators.
Reference
Outcome and
Sample Size
Study Design and
Analytic Method
Copollutants
PM
Metric
Ambient
Concentration
Lag and its
Increment
Units
OR
95%
Confidence
Interval
Dockery et al.
(2005, 078995;
2005, 090743)
Eastern MA
N - 670 days with >
1 confirmed ventricular
arrhythmias among
n - 84 subjects
Generalized estimating
equations
NO?, CO, SO?,
Os
PM2.6
Daily Median:
10.3 /yglm3
2 day
6.9/yg/m3
1.08
0.96,1.22
Lags Evaluated:
2 calendar day means

BC
Daily Median:
0.98/yg/m3
2 day
0.74/yg/m3
1.11
0.95,1.28



Sulfate
Daily Median:
2.55/yg/m3
2 day
2.04 /yglm3
1.05
0.92,1.20




Particle
Number
Daily Median:
29,300 particles/cm3
2 day
19,120
particles/cm3
1.14
0.87,1.50
Rich et al. (2005,
079620)
N - 798 confirmed
ventricular
arrhythmias among
n - 84 subjects
Time-stratified case -
crossover study.Conditional
logistic regression. Lags
evaluated: 3, 6, 24,48-h ma
NO?, CO, SO2,
Os
PM2.6
Daily Median:
9.8/yg/m3
24-h ma
7.8/yg/m3
1.19
1.02,1.38
Eastern MA

BC
Daily Median:
0.94/yg/m3
24-h ma
0.83 /yglm3
0.93
0.74,1.18
Rich et al. (2006,
089814)
N - 139 confirmed
ventricular
arrhythmias among
n - 56 subjects
Time-stratified
case-crossover study.
NO?, CO, SO2,
Os
PM2.6
Daily Median:
16.2/yg/m3
24-h ma
9.7/yg/m3
0.95
0.72,1.27
St. Louis metro
area
Conditional logistic
regression. Lags Evaluated:

EC
Daily Median:
0.6 /yglm3
24-h ma
0.5/yg/m3
1.18
0.93,1.50


6,12, 24, 48-h ma

Organic
Carbon
Daily Median:
4.0/yg/m3
24-h ma
2.3/yg/m3
1.08
0.81,1.43
Vedal et al.
(2004, 055630)
Vancouver, BC,
Canada
N - 257 days with >
1 ICD shock among
n - 50 subjects
Generalized estimating
equations
Lags Evaluated:
0,1, 2, 3 daily ma
NO2, CO, SO2,
Os
PM10
Daily Median:
11.6/yg/m3
Lag Day 0
5.6/yg/m3
1.00*
0.82,1.19*
Ljungman et al.
(2008,180266)
Gothenberg and
Stockholm,
Sweden
N - 114 ventricular
arrhythmias among 73
subjects. 211 total
subjects were
followed.
Conditional logistic
regression
Lags evaluated
2 h, 24 h
NO2
PM10,
Median Gothenberg
2 h: 18.95 /yglm3
24 h: 19.92/yglm3
Stockholm
2 h: 14.62/yg|m3
24 h: 15.23/yglm3
2-h ma:
14.16/yg/m3
24-h ma: 11:49
z^g/m3
2 h:
1.31
24 h:
1.24
1.00,1.72
0.87,1.76




PM2.6
Median
Stockholm/yg/m3
2 h: 9.17
24 h: 9.49 /yglm3
2-h ma: 6.69
z^g/m3
24-h ma: 5.27
/yg/m3
2 h:
1.23
24 h:
1.28
0.84,1.80
0.90,1.84
Rich et al. (2004,
055631)
N - 77 to 98 days
with with > 1 ICD
shock among n - 34
subjects
Ambi-directional
case-crossover study.
NO2, CO, SO2,
Os
PM2.6
Daily Mean:
8.2/yg/m3
Lag Day 0
5.2/yg/m3
1.0t
0.9, 1.1 t
Vancouver, BC,
Canada
Conditional logistic
regression

PM10
Daily Mean:
13.3/yg/m3
Lag Day 0
7.4/yg/m3
0.9t
0.5,1.5t


Lags Evaluated:
0,1, 2, and 3 day ma

EC
Daily Mean:
0.8/yg/m3
Lag Day 0
0.4/yg/m3
1.11
0.9,1.3t



Organic
Carbon
Daily Mean:
4.5 /yglm3
Lag Day 0
2.2/yg/m3
1.11
0.9,1.3t




Sulfate
Daily Mean:
1.3/yg/m3
Lag Day 0
0.9/yg/m3
0.9t
0.7,1.2t
Metzgeret al.
(2007, 092856)
Atlanta, GA
N - 6287 confirmed
ventricular
arrhythmias among
Generalized estimating
equations
FLags Evaluated:
0,1, and 2 day ma
NO2, CO, SO2,
Os
PM2.6
Daily Median:
16.2 /yglm3
24-h ma
10/yg/m3
1.00
0.95,1.0
n - 518 subjects

PM10
Daily Median:
26.4/yg/m3
24-h ma
10/yg/m3
1.00
0.97,1.03
July 2009
6-25
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Outcome and Study Design and
Sample Size Analytic Method
_ M . . PM
Copollutants |y|etrjc
Lag and its
Increment
Units
Ambient
Concentration
Increment OR Confidence
Interval
95%
elements
PM10.2.B Daily Median:
8.7 /yg/m3
PM2.6 EC Daily Median:
1.4/yg/m3
PM2.E OC Daily Median:
3.9/yg/m3
Daily Median:
4.1 /yg/m3

0.03 /yg|m:
Estimated from Figure 3 Vedal et al. (2004, 055630).t Estimated from Figure 3 Rich et al. (2004, 055631)
Since 2004, only two studies (in Boston and Sweden), reported adverse associations of
PM2.5, other size fractions and components with ICD detected ventricular arrhythmias
(Dockery et al., 2005, 078995; Dockery et al., 2005, 090743; Ljungman et al., 2008, 180266;
Rich et al., 2005, 079620), while other studies done elsewhere did not (Metzger et al., 2007,
092856; Rich et al., 2004, 055631; Vedal et al., 2004, 055630; Dusek et al., 2006, 155756). A
range in exposure lags was evaluated in the Boston study (3 h to 3 days) (Dockery et al.,
2005, 078995; Dockery et al., 2005, 090743; Rich et al., 2005, 079620) and Sweden study
(2 h and 24 h) (Ljungman et al., 2008, 180266). Given the unique and homogenous nature of
the study populations, it is not clear why there is not more consistency across these studies.
Rich et al. (2005, 079620) reported that use of the mean pollutant concentration from the
specific 24 h before the arrhythmia rather than just the day of the arrhythmia, resulted in
less exposure misclassification and less bias towards the null, possibly explaining the lack
of association when using just the day of ICD discharge and daily PM concentrations. Other
reasons for the inconsistent findings may include differing degrees of exposure
misclassification within each study or city due to differences in PM composition and
pollutant mixes (e.g., less transition metals and sulfates in the Pacific Northwest than the
Northeast U.S.), and differences in the size of study areas (Boston: within 40 km of PM2.5
monitoring site! Vancouver: Lower Mainland of British Columbia 90 km east of Vancouver).
Studies of ventricular arrhythmia and PM concentration in patients with ICDs are
summarized in Table 6-2.
July 2009
6-26
DRAFT-DO NOT CITE OR QUOTE

-------
Ectopy Studies Using ECG Measurements
A few panel studies have used ECG recordings to evaluate associations between
ectopic beats (ventricular or supraventricular) and mean PM concentrations in the
previous hours and/or days (Berger et al., 2006, 098702; Ebelt et al., 2005, 056907; Sarnat
et al., 2006, 090489). Ectopic beats are defined as heart beats that originate at a location in
the heart outside of the sinus node. They are the most common disturbance in heart
rhythm. Ectopic beats are usually benign, and may present with or without symptoms, such
as palpitations or dizziness. Such beats can arise in the atria or ventricles. When the origin
is in the atria the beat is called an atrial or supraventricular ectopic beat. When such a beat
occurs earlier than expected it is referred to as a premature supraventricular or atrial
premature beat. Likewise, when the origin is in the ventricle the beat is defined as a
ventricular ectopic beat, or when early a premature ventricular beat. When three or more
occur ectopic beats occur in succession, this is called a non-sustained run of either
supraventricular (atrial) or ventricular origin. When the rate of the run is greater than 100
beats per minute it is defined as a tachycardia. Sustained ventricular tachycardias are the
arrhythmias investigated in the ICD studies described above.
Sarnat et al. (2006, 090489) conducted a panel study among 32 nonsmoking older
adults residing in Steubenville, OH. In this study, the median daily PM2.5 concentration was
17.7 |ug/m3. The median daily SO r concentration was 5.7 ug/m3, and the median daily EC
concentration was 1.0 ug/m3. They used logistic regression models to examine lagged effects
of 1-10-day ma concentrations of PM2.5, SO42-, EC, O3, NO2, and SO2. Supraventricular
ectopy and ventricular ectopy were measured using Holter monitors during a 30-minute
protocol of alternating rest in the supine position, standing, walking and paced breathing.
In single pollutant models, each 10.0 jug/m3 increase in 5"day mean PM2.5 concentration was
associated with increased risk of supraventricular ectopy (OR = 1.42 [95% CP 0.99-2.04]),
but not ventricular ectopy (OR = 1.02 [95% CP 0.63-1.65]). Similarly, increased risk of
supraventricular ectopy, but not ventricular ectopy, was associated with each interquartile
range increase in 5"day mean SO42- and O3 concentration.
Ebelt et al. (2005, 056907) conducted a repeated measures panel study of 16 patients
with COPD in the summer of 1998 in Vancouver, British Columbia. Their goal was to
evaluate the relative impact of ambient and non-ambient exposures to PM2.5, PM10, and
PM10-2.5 on several health measures. Subjects wore an ambulatory ECG monitor for 24 h to
record heart rhythm data and ascertain supraventricular ectopic beats. The mean PM2.5
July 2009
6-27
DRAFT-DO NOT CITE OR QUOTE

-------
concentration during this study was 11.4 ug/m::. Using mixed models with random subjects
effects to investigate only same day PM concentrations, an increase in supraventricular
ectopic beats was associated with same day ambient exposures to each PM size fraction.
Berger and colleagues (2006, 098702) conducted a panel study of 57 men with
coronary heart disease living in Erfurt, Germany. Using 24-h ECG measurements made
once every 4 wk, they studied associations between runs of supraventricular and
ventricular tachycardia and lagged concentrations of PM2.5, UFP (0.01-0.1 um), ACP
(0.1-1.0 |um), SO2, NO2, CO, and NO. Using GAMs, as well as Poisson and linear regression
models, they reported increases in supraventricular tachycardia and the number of runs of
ventricular tachycardia associated with 5"day mean PM2.5, UFP counts (0.01-0.1 um), and
ACP counts (0.1-1.0 um). They found these associations at all lags evaluated (during ECG
recording, 0-23 h before, 24-47 h before, 48-71-h before, 72-95 h before, and 5-day mean),
but the largest effect estimates were generally associated with the 24-47-h mean and the
5-day mean.
Using data from the WHI done in 59 U.S. exam sites in 24 cities, Liao et al. (2009,
157456) estimated mean PM2.5 and PM10 concentrations at the address of 57,422 study
subjects undergoing ECG monitoring. They then estimated the risks of ventricular and
supraventricular ectopy during that 10 second ECG recording associated with increases in
mean PM10 and PM2.5 concentrations on the same day and previous 2 days, as well as over
the previous 30 days. Mean PM2.5 and PM10 concentrations during the study period were
13.8 and 27.5 |ug/m3, respectively. Using a 2-stage random effects model, they reported that
among smoking subjects, each 10 |u,g/m3 increase in PM2.5 concentration on lag day 1 was
associated with a significantly increased risk of ventricular ectopy (OR = 2.0 [95% CP
1.32-3.3]). Similarly, each 10 |ug/m3 increase in lag 1 PM10 concentration was associated
with an increased risk of ventricular ectopy (OR = 1.32 [95% CP 1.07-1.65]). The lag day 2
PM2.5 risk estimate was similar in size but not statistically significant. There were no
associations between PM10, PM2.5 and supraventricular ectopy among smokers or non-
smokers, and no association with any PM metric and ventricular ectopy among non-
smokers.
Four studies of ectopic beats and runs of supraventricular and ventricular
tachycardia, captured using ECG measurements, all report at least one positive association.
Further, they report findings in regions other than Boston and Sweden (i.e., Midwest U.S.,
Pacific Northwest, 24 U.S. cities, and Erfurt, Germany). A range of lags and/or moving
July 2009
6-28
DRAFT-DO NOT CITE OR QUOTE

-------
averages were investigated (0-30 days) with the strongest effects observed for either the 5-
day mean, same day, or 1-day lagged PM concentrations. Taken together, these ICD studies
and ectopy studies provide evidence of an arrhythmic response to PM, although further
study is needed to understand the variable ICD study findings.
ECG Abnormalities Indicating Arrhythmia
No reported investigations of the relationship of PM concentration and ECG
abnormalities indicating arrhythmia were conducted prior to 2002 and thus were not
included in the 2004 PM AQCD (U.S. EPA, 2004, 056905). Abnormalities in the myocardial
substrate, myocardial vulnerability, and resulting repolarization abnormalities are believed
to be key factors contributing to the development of arrhythmogenic conditions such as
those discussed above. These abnormalities include ECG measures of repolarization such as
QT duration (time for depolarization and repolarization of the ventricles), T-wave
complexity (a measure of repolarization morphology), and T-wave amplitude (height of the
T-wave). Abnormalities in repolarization may also identify subjects potentially at risk of
more serious events such as sudden cardiac death (Atiga et al., 1998, 156231; Berger et al.,
1997, 155688; Chevalier et al., 2003, 156338; Okin et al., 2000, 156002; Zabel et al., 1998,
156176). Recent studies of changes in these measures following acute increases in air
pollution are described below.
Two studies conducted in Erfurt, Germany (Henneberger et al., 2005, 087960; Yue et
al., 2007, 097968) examined the association between measures of repolarization (QT
duration, T-wave complexity, T-wave amplitude, T-wave amplitude variability) and
particulate air pollution. Henneberger et al. (2005, 087960) conducted a panel study of 56
males with ischemic heart disease. Each subject was measured every 2 wk for 6 months.
During the study, the median daily PM2.5 concentration was 14.9 ug/m3. The median EC
concentration was 1.8 |ug/m3, while the median OC concentration was 1.4 ug/m3. The
median count of UFP counts (0.01-0.1 um) was 11,444 particles/cm3, while the median count
of ACP (0.1-1.0 jum) was 1238 particles/cm3. They examined the change in these ECG
parameters associated with the mean pollutant (UFP, ACP, PM2.5, OC, and EC)
concentrations 0-5, 6-11, 12-17, 18-23, and 0-23 h before, and 2-5 days before the ECG
measurement. Significant decreases in T-wave amplitude were associated with PM2.5 mass,
UFP, and ACP. Each 16.4 ug/m3 increase in the mean PM2.5 concentration in the previous
5 h was associated with a 6.46 |uV decrease in T-wave amplitude (95% CP -10.88 to -2.04).
July 2009
6-29
DRAFT-DO NOT CITE OR QUOTE

-------
Each 0.7 jug/m3 increase in the mean OC concentration in the previous 5 h was associated
with a 4.15 ms increase in QT interval (95% CI: 0.22-8.09). There was a similar sized effect
for 24-h mean OC concentration. Significant increases in the variability of T-wave
complexity were also associated with acute increases in EC and OC concentration.
Yue et al. (2007, 097968) then used positive matrix factorization to identify 5 sources
of ambient PM (airborne soil, local traffic-related UFP, combustion generated aerosols,
diesel traffic-related particles, and secondary aerosols). Using similar statistical models,
they examined the association between these same repolarization changes and incremental
increases in the mean concentration of each particle source in the 24 h before the ECG
measurement. They also examined associations with CRP and vWF concentrations in the
blood. Both UFP from local traffic and diesel particles from traffic had the strongest
associations with repolarization parameters.
These two analyses demonstrate associations between PM pollution and
repolarization changes, at lags of 5 h to 2 days. Moreover, the findings from the Yue et al.
(2007, 097968) study demonstrate a potential role of traffic particles/pollution.
6.2.2.2. Toxicological Studies
The ECG of animal research models frequently exhibit different characteristics than
that of humans. Mice and rats are notable in this regard, as they do not have an isoelectric
ST-segment typical of larger species, likely owing to their rapid heart rates (-600 and -350
bpm, respectively) and repolarizing currents. However, the ultimate function of the
pumping heart is conserved and reflected by the ECG in a remarkably consistent manner
across species. Thus, atrial depolarization causes an electrical inflection represented by the
P-wave, ventricular depolarization elicits the QRS complex, and the T-wave represents
repolarization of the ventricles.
The earliest indication that there may be cardiovascular system effects of PM came
from ECG studies in susceptible animal models (rats with pulmonary hyptertension and
dogs with coronary occlusion), which were summarized in the 2004 PM AQCD (U.S. EPA,
2004, 056905). However, a study of dogs exposed to ROFA did not demonstrate ECG
changes, perhaps due to differences in disease state, as these were the oldest dogs in the
colony with clinical signs of preexisting, naturally occurring heart disease (Muggenburg et
al., 2000, 010279). Much of the research conducted since the release of the last PM AQCD
July 2009
6-30
DRAFT-DO NOT CITE OR QUOTE

-------
has been focused on exploring susceptibility or varying exposure methodologies, with little
new evidence into the mechanisms for ECG changes of inhaled PM.
CAPs
Wellenius et al. (2004, 087874) used a susceptible model that was previously shown to
produce significant results with exposures to ROFA (Wellenius et al., 2002, 025405) to
examine ECG-related PM2.5 effects. Using an anesthetized model of post-infarction
myocardium sensitivity, Wellenius and colleagues tested the effects of Boston, MA CAPs on
the induction of spontaneous arrhythmias in SD rats (l-h exposure! mean mass
concentration 523.11 |Jg/m3; range of mass concentration 60.3-2202 (Jg/m3). Decreased
(67.1%) ventricular premature beat (VPB) frequency was observed during the post-exposure
period in rats with a high number of pre-exposure VPB. No interaction was observed with
coexposure to CO (35 ppm). CAPs number concentration or the mass concentration of any
single element did not predict VPB frequency. In a followup publication, decreased number
of supraventricular ectopic beats (SVEB) was reported with CAPs (mean mass
concentration 645.7 (Jg/m3) (Wellenius et al., 2006, 156152). Furthermore, an increase in
CAPs number concentration of 1,000 particles/cm3 was associated with a 3.3% decrease in
SVEB frequency. The findings of decreased ventricular arrhythmia differ from those
observed following ROFA exposure in the same animal model in that an increased
frequency of premature ventricular complexes was observed with ROFA, albeit the ROFA
exposure concentration was >3,000 |jg/m3 (Wellenius et al., 2002, 025405). It is difficult to
directly compare the result of these studies due to differences in exposure concentrations
and particle type, but collectively they may suggest an important role for the soluble
components of PM, including transition metals, as only ROFA induced increases in
ventricular arrhythmia occurrence.
In older rats (Fisher 344; -18 months) exposed to PM2.5 CAPs in Tuxedo, NY (4 h;
mean concentration 180 |Jg/m3; 8/2,000), the frequency of delayed beats was greater than in
rats exposed to air (Nadziejko et al., 2004, 055632). The majority of these beats were
characterized as pauses (a delay of 2.5 times the adjacent interbeat intervals) rather than
premature beats. When the same animals were exposed to generated ultrafine carbon
particles (single-day concentrations 500 and 1280 |Jg/m3) or SO2 (1.2 ppm), no significant
differences were observed in arrhythmia frequency between air controls and exposed
animals. The authors also report using the same protocol for young WKY rats
July 2009
6-31
DRAFT-DO NOT CITE OR QUOTE

-------
(concentration 215 (Jg/m3) and very few arrhythmias were observed, thus precluding
statistical analysis. The results of this study indicate l) involvement of the sino-atrial node,
as the observed arrhythmias were mostly of a delayed nature and 2) particle size and PM2.5
constituents may play a role in these effects.
Diesel and Gasoline Exhaust
Anselme and colleagues (2007, 097084) exposed rats with and without induced CHF
to DE for 3 h (PM concentration 500 |Jg/m3; mass mobility diameter 85 nm! NO2 1.1 ppm!
CO 4.3 ppm). While no dramatic change was noted in HR, prominent increases in the
incidence of VPB were observed in CHF rats, which lasted at least 4-5 h after exposure
ceased. The duration of VPB attributable to diesel exposure in CHF rats lasted much longer
than the rMSSD change (>5 h post-exposure), indicating that the HRV response was not
driving the increased arrhythmia incidence. It is interesting to contrast the work of
Anselme with the studies by Wellenius et al. (2002, 025405; 2004, 087874; 2006, 156152), as
the arrhythmia incidence in the acute infarction model was greatest with ROFA, while the
CHF model demonstrated sensitivity to diesel exposure. However, several differences in the
research designs preclude strong comparisons.
Using ApoE"'" mice on a high-fat diet as a model of pre-existing coronary insufficiency
(Caligiuri et al., 1999, 156318), Campen and colleagues studied the impact of inhaled diesel
and gasoline exhaust and road dust (6 h/day x 3 day) on ECG morphology (Campen et al.,
2005, 083977; Campen et al., 2006, 096879). Moreover, a high efficiency particle filter was
used to compare the whole exhaust with an atmosphere containing only the gaseous
components. For gasoline exhaust, the PM-containing atmosphere (PM mean concentration
61 |Jg/m3; PNMD 15 nm! NOxmean concentration 18.8 ppm! CO mean concentration
80 ppm) induced T-wave morphological alterations, while the PM-filtered atmosphere did
not (Campen et al., 2006, 096879). Resuspended road dust (PM2.5), at up to 3500 |jg/m3 had
no impact on ECG. For DE (PM mean concentration 512, 770, or 3634 (Jg/m3; MMD 100 nm,
CMD 80 nm! NOx mean concentration 19, 105, 102 ppm for low whole exhaust, high PM
filtered, and high whole exhaust, respectively), dramatic bradycardia, decreased T-wave
area, and arrhythmia (atrioventricular-node block and VPB) were only observed in mice
exposed to high filtered and high whole exhaust (Campen et al., 2005, 083977). These
effects remained after filtration of PM, suggesting that the gaseous components of the
whole DE drove the cardiovascular findings. The diesel- and gasoline-induced ECG changes
July 2009
6-32
DRAFT-DO NOT CITE OR QUOTE

-------
contrast, in that the gasoline exhaust required particles to induce T-wave changes, whereas
the DE did not require PM to cause an effect on ECG. However, the differing responses
could be attributable to higher PM concentrations in the whole DE.
The above toxicological studies demonstrate mixed results for arrhythmias, which
may be somewhat attributable to the different disease models used. Wellenius et al. (2004,
087874; 2006, 156152) showed decreased frequency of VPB and SVEB following PM2.5 CAPs
exposure in rats with induced MI (>12-h prior to exposure). One study reported increased
frequency of premature beats in older rats exposed to CAPs, which were not observed with
ultrafine carbon particles (Nadziejko et al., 2004, 055632). Rats with a MI model of chronic
heart failure (3-month recovery) had increased incidence of VPB with DE exposure
(Anselme et al., 2007, 097084). As for ECG morphology changes, T-wave alterations were
reported for gasoline exhaust that were absent when the PM was filtered (Campen et al.,
2006, 096879). However, for DE, increased atrioventricular-node block, VPB, and decreased
T-wave area were observed with whole exhaust and remained after filtration of PM,
indicating that the gases were responsible for the effects (Campen et al., 2005, 083977).
6.2.3. Ischemia
Although no evidence from epidemiologic or controlled human exposure studies of
PM-induced myocardial ischemia was included in the 2004 PM AQCD (U.S. EPA, 2004,
056905). one toxicological study was cited that observed ST-segment changes in dogs
following a 3 day exposure to CAPs. In epidemiologic studies published since the 2004 PM
AQCD (U.S. EPA, 2004, 056905), associations have been demonstrated between PM and ST-
segment depression, and one new controlled human exposure study reported significant
increases in exercise-induced ST-segment depression among men with prior MI following a
controlled exposure to DE. Results from recent toxicological studies confirm the findings
presented in the 2004 PM AQCD (U.S. EPA, 2004, 056905) and provide coherence and
biological plausibility for the effects observed in epidemiologic and controlled human
exposure studies.
July 2009
6-33
DRAFT-DO NOT CITE OR QUOTE

-------
6.2.3.1. Epidemiologic Studies
ECG Abnormalities Indicating Ischemia
The ST-segment duration is typically in the range of 0.08-0.12 s (80-120 ms). The
direction of the ST change is influenced by the extent of the acute myocardial injury. If the
ischemia or infarction is transmural, i.e., penetrates the entire thickness of the ventricular
wall, it usually causes ST elevation, while ischemia confined primarily to the ventricular
endocardium often causes ST-segment depression. Several short-term exposure studies of
air pollution investigated the association of ST-segment depression with PM concentration.
Gold et al. (2005, 087558) studied 24 elderly residents of Boston, MA (aged 61-88 yr)
residing at or near an apartment complex that was ~ 0.5 km from an air pollution
monitoring station. A protocol of continuous Holter monitoring including 5 min of rest,
5 min of standing, 5 min of outdoor exercise, 5 min of rest, and then 20 cycles of paced
breathing was done up to 12 times for each subject (n = 269 ECG measurements for
analysis). From these ECG measurements, they identified occurrences of ST-segment
depression and examined whether mean BC, CO, and PM2.5 concentrations in the previous
5 and 12 h were associated with ST-segment depression. The median 5-h mean BC
concentration was 1.28 ug/rrrand the median 12-h mean BC concentration was 1.14 ug/m3.
The median 5-h mean PM2.5 concentration was 9.5 |ug/m3, and the median 12-h mean PM2.5
concentration was 9.8 ug/m3. The mean BC concentrations in the 5 and 12 h before testing
predicted ST-segment depression in most portions of the protocol. However, these effects
were strongest in the post-exercise periods. For example, during the post-exercise rest
period, each 10th-90th percentile increase (1.59 ug/m3) in the mean 5-h BC concentration
was associated with a -0.11 mm ST-segment depression (95% CP -0.18 to -0.05). In two
pollutant models, CO did not appear to confound this association. PM2.5 was not associated
with ST-segment depression in this study. These findings suggest traffic-generated
particulate pollution may be associated with ST-segment depression.
Previously, Pekkanen et al. (2002, 035050) conducted a panel study of 45 subjects with
stable coronary heart disease living in Helsinki, Finland. Each subject had biweekly
sub-maximal exercise testing for 6 months (n = 342 exercise tests with 72 exercise-induced
ST-segment depressions). During this study, the median daily PM2.5 concentration was
10.6 |ug/m3, and the median daily count of ACP (ACP: 0.1-1.0 |um) was 1,200 particles/cm3.
They examined the risk of ST-segment depression associated with mean pollutant
July 2009
6-34
DRAFT-DO NOT CITE OR QUOTE

-------
concentrations (UFP, ACP, PMi, PM2.5, PM10-2.5, NO2, CO) in the previous 24 h, and the 3
previous lagged 24-h periods. Each 7.9 |u,g/m3 increase in mean PM2.5 concentration, lagged
2 days, was associated with significantly increased risk of ST-segment depression >0.1 mV
(OR: 2.84 [95% CP 1.42-5.66]). Each 760 particles/cm3 increase in the count of ACP, lagged 2
days, was also associated with significantly increased risk of ST-segment depression >0.1
mV (OR: 3.29 [95% CP 1.57-6.92]). Similarly sized increased risks of ST-segment depression
were also found for other particulate pollutants, including PM10-2.5, PMi, and the counts of
UFP (0.01-0.1 nm).
This same research group, then conducted a principal components analysis to identify
five PM2.5 sources (crustal, long range transport, oil combustion, salt, and local traffic)
(Lanki et al., 2006, 088412). Using similar statistical models, each 1 jug/m3 increase in "local
traffic" particle concentration, lagged 2 days, was associated with increased risk of ST-
segment depression (OR: 1.53 [95% CP 1.19-1.97]). Similarly, each 1 jug/m3 increase in "long
range transport" particle concentration was also associated with increased risk of ST-
segment depression (OR: 1.11 [95% CP 1.02-1.20]). No significant associations for other
sources were reported for any lag time.
In Boston, Chuang et al. (2008, 155731) studied 48 patients with a prior percutaneous
intervention following MI, acute coronary syndrome (ACS) without MI, or stable coronary
artery disease without ACS, no earlier than 1-yr previous. Each patient had a 24-h ECG
measurement up to 4 times during study followup. Using logistic regression, they
estimated the risk of ST-segment depression of > 0.1 mm, during half hour segments,
associated with increases in the mean PM2.5, BC, CO, NO2, O3, and SO2 concentration in the
previous 24 h. Each 6.93 |u,g/m3 increase in mean PM2.5 concentration was associated with a
significantly increased risk of ST-segment depression (OR = 1.50 [95% CP 1.19-1.89]). Using
linear additive models to estimate the change in ST level associated with the same PM2.5
change, they observed a significant -0.031 mm change (95% CP -0.042 to -0.019). In single
pollutant models, risk estimates were of similar magnitude and statistically significant for
BC, NO2, and SO2. In two pollutant models, however, PM2.5 risk estimates were reduced to
1.0 in all models with BC, NO2, and SO2. In contrast, the risk estimates for BC, NO2, and
SO2 remained elevated and statistically significant when modeled with PM2.5.
In a panel study of 14 Helsinki resident, non-smoking, elderly subjects with coronary
artery disease, Lanki et al. (2008, 191984) used logistic regression to report that each 10
ug/m3 increase in personal PM2.5 concentration in the previous hour was associated with a
July 2009
6-35
DRAFT-DO NOT CITE OR QUOTE

-------
significantly increased risk of ST-segment depression (OR = 3.26 [95% CP 1.07-9.98]). In
addition, each 10 ug/m3 increase in outdoor mean PM2.5 concentration in the previous 4 h
was also associated with an increased risk (OR = 2.47 [95% CP 1.05-5.85]). Last, the risk
estimates for all time lags examined (l, 4, 8, 12, and 22 or 24 h) for all PM size fractions
were greater than 1.0, but none other than those described above were statistically
significant.
In a study of 57,908 participants in the WHI Trial, Zhang et al. (2009, 191970)
examined the change and risk of ST-segment abnormalities, T wave abnormalities, and T
wave amplitude associated with ambient PM2.5 concentrations on the same and previous 6
days. Using logistic regression, each 10 |ug/m3 increase in the mean PM2.5, on lag days 0 to
2, was associated with a 4% (95% CP -3% to 10%) increase in the relative odds of a
ST-segment abnormality, and a 5% (95% CP 0-9%) increase in the relative odds of a T wave
abnormality.
Summary of Epidemiologic Study Findings for Ischemia
These studies demonstrate associations between PM2.5 pollution and ST-segment
depression at lags of 1 h to 2 days. Morever, these findings demonstrate a potential role for
traffic (Chuang et al., 2008, 155731; Gold et al., 2005, 087558) and long-range transported
PM2.5 (Lanki et al., 2006, 089788).
6.2.3.2. Controlled Human Exposure Studies
Diesel Exhaust
Among a group of 20 men with prior MI, Mills et al. (2007, 091206) found that DE
(300 (Jg/m3 particle concentration, median particle diameter 54 nm) significantly increased
exercise-induced ischemic burden during exposure, calculated as the product of exercise
duration and change in ST-segment amplitude. The mechanism by which DE induced the
exacerbation of ischemic burden remains unclear, and appears to be unrelated to impaired
vasodilation. However, the authors suggest that this discrepancy may be due to the timing
of the vascular assessment, as measures of blood-flow were taken 5 h after the observed
increase in ischemic burden. Although it is reasonable to assume that the observed increase
in ST-segment depression during exercise represents an increased magnitude of ischemia, it
is important to note that there are other potential explanations for the ST change. For
example, it is possible that the ST-segment depression could be secondary to heterogeneity
July 2009
6-36
DRAFT-DO NOT CITE OR QUOTE

-------
of electrophysiological responses of particle exposure on the myocardium that is enhanced
by the metabolic and ionic conditions associated with ischemia or increased heart rate. It is
also important to note that the effects observed in this study cannot be conclusively
attributed to the particles per se, as subjects were also exposed relatively high levels of NO
(3.45 ppm), NO2 (1.01 ppm), CO (2.9 ppm), and total hydrocarbons (2.8 ppm).
6.2.3.3. Toxicological Studies
There were no toxicological studies cited in the 2004 PM AQCD (U.S. EPA, 2004,
056905) that directly examined myocardial ischemia. One study was reviewed in the 2004
PM AQCD that evaluated ST-segment changes in dogs during occlusion and following a 3-
day exposure to Boston CAPs reported increased magnitude and decreased time to
ST-segment elevation (Godleski et al., 2000, 000738).
CAPs
A study that examined ECG changes in dogs (female! retired mongrel breeder dogs)
following PM2.5 CAPs exposure in Boston, MA (mean mass concentration 345 |Jg/m3;
9/2000-3/2001) and left anterior descending coronary artery occlusion as an indicator of
myocardial ischemia reported changes in ST-segment (Wellenius et al., 2003, 055691). The
experimental protocol was a 6-h exposure to CAPs via tracheostomy, followed by a
preconditioning occlusion (5 min), rest interval (20 min), and the experimental occlusion (5
min). Increased ST-segment elevation was observed following PM2.5 during the
experimental occlusion period compared to filtered air. Furthermore, peak ST-segment
elevation attributable to CAPs was reported with the experimental occlusion, which
remained elevated 24 h post-exposure. Ventricular arrhythmias were rarely observed
during occlusion and when observed, were unrelated to CAPs exposure. The results from
this study support those previously observed (Godleski et al., 2000, 000738) and provides
greater support that enhanced myocardial ischemia occurs relatively quickly (within h)
following PM exposure.
A recent study in dogs (female mixed-breed canines) evaluated myocardial blood flow
during myocardial ischemia following 5"h PM2.5 Boston CAPs exposures (daily mean mass
concentration 94.1-1556.8 |Jg/m3; particle number concentration 3"69.3 x 103 particles/cm3;
BC concentration 1.3-32.0 (Jg/m3) (Bartoli et al., 2009, 179904). Similar methods were used
for the coronary occlusion and exposure method as Wellenius et al. (2003, 055691).
July 2009
6-37
DRAFT-DO NOT CITE OR QUOTE

-------
Immediately following exposure, microspheres were injected (15 |Jm diameter) into the left
atrium after 3 min of ischemia during the second occlusion. Post-mortem analysis of cardiac
tissue and blood samples allowed for quantification of microspheres. CAPs-exposed dogs
had decreased total myocardial blood flow and increased coronary vascular resistance
during occlusion that was greatest in tissue within or near the ischemic zone. The
rate-pressure product (product of HR and SBP) during occlusion was unchanged in animals
exposed to CAPs, indicating that cardiac metabolic demand was not altered. The multilevel
linear mixed models demonstrated that myocardial blood flow and coronary vascular
resistance during occlusion were inversely and significantly associated with CAPs mass
concentration, particle number concentration, and BC concentration, with the strongest
effects observed with particle number concentration. The results of this study provide
evidence that exacerbation of myocardial ischemia following PM exposure is due to reduced
myocardial blood flow, perhaps via dysfunctional collateral vessels.
Intratracheal Instillation
Cozzi et al. (2006, 091380) exposed ICR mice to ultrafine PM (100 |ug IT instillation),
and followed 24 h later by ischemia/reperfusion injury to the left anterior coronary artery.
The area-at-risk (the region of tissue perfused by the left anterior descending coronary
artery) and the infarct size were measured 2 h following reperfusion, and while the
area-at-risk was not affected by PM exposure, the infarct size was nearly doubled in mice
who received ultrafine PM. Increases in infarct size were associated with increased
myocardial neutrophil density in the infarct zone and lipid peroxidation in the myocardium.
Summary of Toxicological Study Findings for Ischemia
The studies described above provide evidence that PM can induce greater myocardial
responses following ischemic events, as demonstrated by, enhanced ischemia, decreased
myocardial blood flow and increased coronary vascular resistance, and increased infarct
size.
PM Components
The Wellenius et al. (2003, 055691) study employing dogs also attempted to link
ST-segment changes with four CAPs elements (Si, Ni, S, and BC) as tracers of PM2.5 sources
in Boston. In the multivariate regression analyses, peak ST-segment elevation and
integrated ST-segment change were significantly associated with only the mass
July 2009
6-38
DRAFT-DO NOT CITE OR QUOTE

-------
concentration of Si (Si mean concentration 8.17 |Jg/m3; Si concentration 2.31-13.93 |Jg/m3).
In the univariate regression analyses, Pb also demonstrated a significant association for
both ST-segment measures, although the p-value was greater than that observed with Si.
6.2.4. Vasomotor Function
The most noteworthy new health-related revelation in the past six yr with regards to
PM exposure is that the systemic vasculature may be a target organ. The vasculature of all
tissues is lined with endothelial cells that will naturally encounter any systemically
absorbed toxin. The endothelium (l) maintains barrier integrity to ensure fluid
compartmentalization, (2) communicates dilatory and constrictive stimuli to vascular
smooth muscle cells, and (3) recruits inflammatory cells to injured regions. Smooth muscle
cells lie within the layer of endothelium and are crucial to the regulation of blood flow and
pressure. In states of injury and disease, both cell types can exhibit dysfunction and even
pathological responses.
Endothelial dysfunction is a factor in many diseases and may contribute to the origin
and/or exacerbation of perfusion-limited diseases, such as MI or IHD, as well as
hypertension. Endothelial dysfunction is also a characteristic feature of early and advanced
atherosclerosis. A primary outcome of endothelial dysfunction is impaired vasodilatation,
frequently due to uncoupling of NOS. It is this uncoupling that appears central to impaired
vasodilation and thus endothelial dysfunction.
One controlled human exposure study cited in the 2004 PM AQCD (U.S. EPA, 2004,
056905) reported a decrease in bronchial artery diameter (BAD) among healthy adults
following exposure to CAPs in combination with O3. Conclusions based on this finding were
limited due to the concomitant exposure to O3 as well as a lack of published results from
epidemiologic and toxicological studies. Recent controlled human exposure studies have
provided support to the findings described in the 2004 PM AQCD (U.S. EPA, 2004, 056905),
with changes in vasomotor function observed following controlled exposures to DE and EC
particles. In addition, epidemiologic studies have observed associations between PM and
decreases in BAD and flow mediated dilatation in healthy adults and diabetics. These
findings are further supported by a large body of new toxicological evidence of impaired
vasodilation following exposure to PM.
July 2009
6-39
DRAFT-DO NOT CITE OR QUOTE

-------
6.2.4.1. Epidemiologic Studies
O'Neill et al. (2005, 088423) examined the association between 2 measures of vascular
reactivity (nitroglycerin mediated reactivity and flowmediated reactivity) and ambient
mean particulate pollutant concentration (PM2.5, SO r , BC, particle number count [PNC])
on the same and previous few days. They studied a panel of 270 subjects with diabetes or at
risk for diabetes, who lived in the greater Boston metropolitan area. The mean PM2.5
concentration during this study (1998-2002) was 11.5 ug/m3. Using linear regression models
adjusted individually for age, gender, BMI, and race, the change in vascular reactivity
associated with moving average pollutant concentrations across the same and previous 5
days was estimated. Interquartile range (values not reported) increases in the mean PM2.5
concentration, BC concentration, and PNC over the previous 6 days were associated with
decreased vascular reactivity among diabetics, but not among subjects at risk for diabetes.
For sulfates, the mean concentration on lag day 0, lag day 1, and the 3-day, 4-day, and 5-day
ma all were associated with similarly sized reductions in both metrics of vascular reactivity.
Among diabetics, each interquartile range increase in the mean S( )r concentration over
the previous 6 days was associated with a 5.4% decrease in nitroglycerin-mediated
reactivity (95% CP -10.5 to -0.1) and flow-mediated reactivity (-10.7% [95% CP -17.3 to
-3.5]). Also among diabetics, each interquartile range increase in the mean PM2.5
concentration over the previous 6 days was associated with a significant 7.6% decrease in
nitroglycerin-mediated reactivity (95% CP -12.8 to -2.1) and a non-significant 7.6% decrease
in flow-mediated reactivity (95% CP -14.9 to 0.4). Each interquartile range increase in the
mean BC concentration over the previous 6 days was associated with a 12.6% decrease in
flow mediated reactivity (95% CP -21.7 to -2.4), but not nitroglycerin-mediated reactivity.
PNC was associated with non-significant decreases in both. Effect estimates were larger for
type II diabetics than type 1 diabetics.
Dales et al. (2007, 155743) conducted a panel study of 39 healthy volunteers who sat
at 1 of 2 bus stops in Ottawa, Canada for 2 h. Flow-mediated vasodilation of the brachial
artery was measured immediately after the bus stop exposure, but not before. The mean
PM2.5 concentrations, measured at the 2 bus stops, were 40 and 10 |u,g/m3. They examined
the association between flow-mediated vasodilation and 2-h mean PM2.5, PMi, NO2, and
traffic density at the bus stop (vehicles/h). The authors report that each 30 ug/m3 increase
in 2-h mean PM2.5 concentration was associated with a significant 0.48% reduction in flow-
July 2009
6-40
DRAFT-DO NOT CITE OR QUOTE

-------
mediated dilation (FMD) (p = 0.05). This represented a 5% relative change in the maximum
ability to dilate.
This same research group conducted a panel study of 25 type I or II diabetic subjects
living in Windsor, Ontario (aged 18-65) (Liu et al., 2007, 156705). For each subject, personal
PMio concentrations were measured for 24 h before measurements of BAD, FMD, and other
biomarkers. The mean 24 -h mean PMio concentration, measured with personal monitors,
was 25.5 jug/m3. Each 10 jug/m3 increase in personal 24-h mean PMio concentration was
associated with a 0.20% increase in end-diastolic FMD (95% CP 0.04 to 0.36) and a 0.38%
increase in end-systolic FMD (95% CP 0.03-0.73), but decreases in end-diastolic basal
diameter (-2.52 jum [95% CP -8.93 to 3.89]) and end-systolic basal diameter (-9.02 jum [95%
CP -16.04 to -2.00]).
Rundell et al. (2007, 156060) examined the change in FMD associated with high and
low PMi (0.02-1.0 um) pollution in a panel of 16 young intercollegiate athletes (mean
age = 20.5 ± 2.4 yr) in Scranton, PA, who were non-smokers, non-asthmatics, and free of
cardiovascular disease (Rundell et al., 2007, 156060). Each subject had FMD of the brachial
artery measured 10-20 min before and 20-30 min after each of two 30-min exercise tests
(85-90% of maximal HR). The exercise tests were done outside either on an inner campus
location free of automobile and truck traffic (low PMi! mean = 5309 ± 1942 particles/cm3) or
on a soccer field adjacent to a major highway (high PMi! mean = 143,501 ± 58,565
particles/cm3). The order of the exercise test locations was chosen randomly. Using paired
t-tests for analysis, they reported FMD was impaired after high PMi exposure (pre-exercise:
6.8% ± 3.58%; post-exercise: 0.30% ± 2.74%; p = 0.0001 for the change) but not low PMi
exposure (pre-exercise: 6.6% ± 4.04%; post-exercise: 4.89% ± 4.42%; p-value for the change
not given). Further, they found basal brachial artery vasoconstriction (4%; pre-exercise
BAD: 4.66 ± 0.61 mm; post-exercise BAD: 4.47 ± 0.63 mm; p = 0.0002 for the change) after
the 'high PMi' exposure, but not the 'low PMi' exposure (-0.3% pre-exercise BAD: 4.66 ± 0.63
mm! post-exercise BAD: 4.68 ± 0.61 mm; p-value for the change not given).
In a prospective panel study of n = 22 Type II diabetics (aged 61 ± 8 yr), Schneider
et al. (2008, 191985) examined the change in FMD, BAD, small artery elasticity index,
larger artery elasticity index, and systemic vascular resistance associated with ambient
PM2.5 as measured in Chapel Hill, NC. The mean and standard deviation PM2.5
concentrations during the study period (November 2004 to December 2005) were 13.6 ± 7.0
ug/m3. Using additive mixed models with a random subject effect, each 10 ug/m3 increase in
July 2009
6-41
DRAFT-DO NOT CITE OR QUOTE

-------
PM2.5 in the previous 24 h was associated with a significant decrease in FMD (-17.3% [95%
CL -34.6 to 0.0]). Similarly, each 10 |u,g/m3 increases in PM2.5 was associated with a
significant decrease in small artery elasticity index lagged 1 day (-15.1% [95% CI: -29.3 to -
0.9]), and lagged 3 days (-25.4% [95% CI: -45.4 to -5.3]). Significant decreases in larger
artery elasticity index and significant increases in systemic vascular resistance lagged 2
and 4 days were also observed. Further, effects were greatest among those with high BMI,
high glycosylated hemoglobin Ale, low adiponectin, or the null GSTM1 polymorphism.
However, high myeloperoxidase levels were associated with greater PM2.5 effects on these
measures.
In a similar study done in Paris, France, Briet (2007, 093049) similarly reported that
each increase in PM2.5 was associated with a -0.32% (95% CI: -1.10 to 0.46) decrease in
FMD, although this estimate was not statistically significant. Significant FMD reductions
were associated with increased SO2, NO2, and CO concentrations. Each 1 standard
deviation increase (units not given) in PM2.5 in the previous 2 wk was associated with a
15.68% (95% CI: 7.11-23.30) increase in small artery reactive hyperemia. Each 1 standard
deviation increase (units not given) in PM10 in the previous 2 wk was associated with a
15.91% (95% CI: 7.74-24.0) increase in small artery reactive hyperemia.
Summary of Epidemiologic Study Findings for Vasomotor Function
Vasomotor function has been evaluated using several metrics in the studies described
above, including FMD, small artery elasticity index, larger artery index, systemic
vascular resistance, BAD, end diastolic basal diameter, and nitroglycerin-mediated
reactivity. The most common measures evaluated were BAD, a measure of the
relatively static, anatomic/physiological baseline vasomotor function, and FMD, the
dynamic measure of post- minus pre-occlusion BAD. Each study demonstrated an acute
association between these measures of vascular function and ambient PM concentrations
(Briet et al., 2007, 093049; Dales et al., 2007, 155743; Liu et al., 2007, 156705; O'Neill et al.,
2005, 088423; Rundell et al., 2007, 156060; Schneider et al., 2008, 191985). Three studied a
panel of diabetics (Liu et al., 2007, 156705; O'Neill et al., 2005, 088423; Schneider et al.,
2008, 191985), and three a panel of young healthy subjects (Briet et al., 2007, 093049; Dales
et al., 2007, 155743; Rundell et al., 2007, 156060). Only two studies investigated multiple
lags (lag days 0 to 6) (O'Neill et al., 2005, 088423; Schneider et al., 2008, 191985), with one
reporting the strongest association with the 6"day mean PM concentration (O'Neill et al.,
July 2009
6-42
DRAFT-DO NOT CITE OR QUOTE

-------
2005, 088423), and the other with lag day 0. In other studies, responses were observed in as
short as 30 min after the exposure (Rundell et al., 2007, 156060). The Rundell et al. (2007,
156060) findings are consistent with other studies showing an adverse response to ambient
particulate pollution emitted from vehicular traffic (Adar et al., 2007, 098635; Adar et al.,
2007, 001458; Riediker et al., 2004, 056992; Riediker et al., 2004, 091261).
6.2.4.2. Controlled Human Exposure Studies
Some evidence of a PM-induced increase in brachial artery vasoconstriction is
presented in the 2004 PM AQCD (U.S. EPA, 2004, 056905). Brook et al. (2002, 024987)
exposed 24 healthy adults to PM2.5 CAPs (150 |jg/m3) along with 120 ppb O3 for a period of
2 h. A significant decrease in BAD was observed immediately following exposure compared
with filtered air control. No significant changes were observed in either
endothelial-dependent or endotheliaHndependent vasomotor function, as determined by
FMD and nitroglycerin-mediated dilatation, respectively. As described below, many more
recent studies have evaluated the effects of various types of particles on vasomotor function
following controlled exposures among healthy and health-compromised individuals.
CAPs
A subsequent analysis of the CAPs constituents from the Brook et al. study revealed a
significant negative association between the post-exposure change in BAD and both the OC
and EC concentrations of CAPs (Urch et al., 2004, 055629). However, the observed
vasomotor effects cannot conclusively be attributed to PM2.5, as subjects were exposed
concurrently to PM2.5 and O3. Mills et al. (2008, 156766) evaluated the effect of fine and
ultrafine CAPs on vasomotor function in a group of 12 males with stable coronary heart
disease (average age 59 yr), as well as in 12 healthy males (average age 54 yr). Relative to
filtered air exposure, exposure to PM (average concentration 190 (Jg/m3) did not
significantly affect vascular function in either group. The authors attributed the lack of
response in endothelial function to the composition of the CAPs used in the study, which
were low in combustion-derived particles and consisted largely of sea salt.
Urban Traffic Particles
The effect of exposure to urban traffic particles on vasomotor function has recently
been evaluated among a group of adult volunteers (Brauner et al., 2008, 191966). In this
study, healthy young adults (average age 27 yr) exposed for 24 h to urban traffic particles
July 2009
6-43
DRAFT-DO NOT CITE OR QUOTE

-------
(average PM2.5 concentration 9.7 |jg/m3) were not observed to experience any change in
microvascular function after 6 or 24 h of exposure relative to filtered air.
Diesel Exhaust
Mills et al. (2005, 095757) exposed 30 healthy men (20-38 years old) to both diluted
DE (300 (Jg/m3) and filtered air control for 1 h with intermittent exercise. Half of the
subjects underwent vascular assessments at 6 to 8 h following exposure to DE or filtered
air, while in the other 15 subjects, vascular assessments were performed at 2-4 h
post-exposure. DE attenuated forearm blood flow increase induced by bradykinin,
acetylcholine, and sodium nitroprusside infusion measured 2 and 6 h after exposure. The
authors postulated that the effect of diesel on vasomotor function may be the result of
reduced NO bioavailability in the vasculature stemming from oxidative stress induced by
the nanoparticulate fraction of DE. A diesel-induced decrease in the release of tPA was also
observed at 6 h post-exposure, which may provide additional mechanistic evidence
supporting the observed association between air pollution and MI. As presented in
Tornqvist et al. (2007, 091279), changes in vascular function were also evaluated 24 h
following exposure in 15 of the 30 subjects. Compared with filtered air, exposure to DE
significantly reduced endothelium-dependent (acetylcholine) vasodilation (p = 0.01) at 24 h
post exposure. Bradykinin-induced vasodilation was marginally attenuated by DE
(p = 0.08), while no effects of diesel on endothelium-independent vasodilation (sodium
nitroprusside) were observed. Although the release of tPA was not affected by DE 24 h
following exposure, the authors suggest that the persistent association between diesel
exposure and vasomotor function observed in this study provides supporting mechanistic
evidence of increases in cardiovascular events occurring 24 h after a peak in PM
concentration.
To further investigate the effects of DE on vasomotor function, Mills et al. (2007,
091206) exposed 20 men (average age 60 yr) with previous MI on two separate occasions to
dilute DE (300 |Jg/m3; mean particle size 54 nm) or filtered air for 1 h with intermittent
exercise. Contrary to previous findings in younger, healthy adults (Mills et al., 2005,
095757), DE was found not to affect vasomotor function in peripheral resistance vessels at
6 h post-exposure as measured by endothelium-dependent (acetylcholine) and
endothelium-independent (sodium nitroprusside) vasodilation (forearm blood flow).
However, vascular assessments were not performed at 2 h post-exposure in this study. The
July 2009
6-44
DRAFT-DO NOT CITE OR QUOTE

-------
same laboratory evaluated the effect of exposure to DE with slightly higher particle
concentrations (330 |Jg/m3, particle number 1.26 x 106/cm3) on arterial stiffness among
healthy adults (Lundback et al., 2009, 191967). Using radial artery pulse wave analysis,
significant increases in augmentation pressure and augmentation index, as well as a
significant reduction in the time to wave reflection were observed 10 and 20 min following
exposure to DE relative to filtered air. This finding of a DE-induced reduction in arterial
compliance provides additional evidence to suggest that exposure to particles may
adversely affect vasomotor function.
Peretz et al. (2008, 156854) exposed both healthy adults (n = 10) and adults with
metabolic syndrome (n = 17) for 2 h to filtered air and two concentrations of diluted DE
(fine PM concentrations of 100 and 200 |Jg/m3). Compared with filtered air, DE at 200 (Jg/m3
elicited a statistically significant decrease in BAD (0.11 mm [95% CP 0.02-0.18 mm])
immediately following exposure. A smaller diesel-induced decrease in BAD (0.05 mm) was
observed following exposure to 100 |Jg/m3. Although this decrease was not statistically
significant, the average decrease was approximately one half the decrease at the higher
particle concentration, which provides suggestive evidence of a linear concentration
response in this range of concentrations. Exposure to DE was not shown to significantly
affect endothelium-dependent flow-mediated dilatation. Plasma levels of endothelin-1
(ET-l) were observed to increase relative to filtered air exposure approximately 1-h after
exposure to 200 (Jg/m3 DE (p = 0.01). The results of this study provide evidence of an acute
endothelial response and arterial vasoconstriction resulting from short-term exposure to
DE. Diesel-induced changes in vasoconstriction and ET-l release were more pronounced in
the healthy subjects than in the subjects with metabolic syndrome. The authors postulated
that subjects with metabolic syndrome may have stiffer vessels that are not as responsive
to vasoconstrictor stimuli. In a study utilizing a similar exposure protocol, Lund et al.
(2009, 180257) observed a significant increase in ET-l in healthy adults following a 1-h
exposure to DE with a particle concentration of 100 jJg/m3, a concentration that was not
shown to affect ET-1 levels in the Peretz et al. (2008, 156854) study.
In the previously described studies by Mills et al. (2005, 095757; 2007, 091206),
Peretz et al. (2008, 156854), Tornqvist et al. (2007, 091279) and Lund et al. (2009, 180257),
subjects were exposed to DE, which, in addition to PM, includes DE gases such as NOx, CO,
and hydrocarbons. Therefore, it is possible that the observed effects may be due in part to
exposure to non-particle components of DE. While the majority of these DE exposures have
July 2009
6-45
DRAFT-DO NOT CITE OR QUOTE

-------
contained relatively high levels of gaseous emissions including NO2 concentrations >2 ppm,
the concentrations of these gases were much lower in the Peretz et al. (2008, 156854) study
(NO2 concentrations ~ 20 ppb) which used a newer diesel engine (2002 Cummins B-series)
operating under load at 75% of rated capacity. In this study an apparent linear
concentration response relationship was observed between increasing DE exposure and
decreases in BAD at particle concentrations between 100 and 200 |Jg/m3.
Gasoline Emissions
Rundell and Caviston (2008, 191986) exposed 15 college athletes to particles
generated using a 2.5 hp gasoline engine, as well as a clean air control during 6 min periods
of maximal exercise on a cycle ergometer. Subjects were exposed twice under each
condition, with the two clean air exposures occurring first, separated by 3 days. The two
exposures to gasoline emissions were also separated by 3 days, with the first exposure
occurring 7 days after the second clean air exposure. During exposures to gasoline
emissions, average particle counts of particulate matter <1.0 |Jm were reported as
336,730/cm3 and396,200/cm3 during the first and second exposures, respectively, with an
average CO concentration of 6.3 ppm. There were no differences observed in total work
done (kJ) over the 6-minute exercise periods between the two clean air exposures or
between the clean air exposures and the first exposure to gasoline exhaust. However, the
second gasoline exhaust exposure was demonstrated to significantly decrease work
accumulated over the 6-minute exercise period compared with either of the other exposure
conditions (p < 0.01). The results of this study provide limited evidence to suggest that a
very short term exposure to gasoline emissions may affect exercise performance in healthy
adults. The authors speculated that the observed effect of exposure on work accumulated
during maximal exercise could be due to vasoconstriction and decrease in blood flow in the
skeletal muscle microcirculation. However, the effect of exposure on vasoreactivity was not
explicitly assessed.
Model Particles
The results of a recent study by Shah et al. (2008, 156970) provides evidence that
exposure to ultrafine EC particles (50 |Jg/m3) without coexposure to organics, metals, or
gaseous copollutants may alter vasomotor function in healthy adults. In this study, venous
occlusion plethysmography was used to measure reactive hyperemia of the forearm prior to
exposure, immediately following exposure, and 3.5 h, 21 h, and 45 h following a 2-h
July 2009
6-46
DRAFT-DO NOT CITE OR QUOTE

-------
exposure with intermittent exercise. Peak forearm blood flow was observed to increase after
exposure to filtered air, but not following exposure to ultrafine EC at 3.5 h post-exposure
(p = 0.03).
Summary of Controlled Human Exposure Study Findings for Vasomotor Function
Taken together, the two studies by Mills et al. (2005, 095757; 2007, 091206) along
with the studies by Peretz et al. (2008, 156854), Lund et al. (2009, 180257) and Tornqvist
et al. (2007, 091279) suggest that, in healthy subjects, DE exposure inhibits
endothelium-dependent and endothelium-independent vasodilation acutely (within 2-6 h),
and that the suppression of endothelium-dependent vasodilation may remain up to 24 h
following exposure. In patients with coronary artery disease, vasodilator function does not
appear to be affected 6-8 h following exposure! however, vascular assessments were not
performed at earlier time points. In addition, the use of medications in these patients may
have blunted the response to PM. The findings of Shah et al. (2008, 156970) suggest that
UFP carbon core may be sufficient to produce small changes in systemic vascular function,
but the mechanisms remain obscure. The authors demonstrated a decrease in nitrate levels
following exposure to EC UFP; however, venous nitrite level, which more closely reflects NO
production, was unchanged. Exposure to urban traffic particles was not demonstrated to
alter vasomotor function among healthy adults.
6.2.4.3. Toxicological Studies
Vascular dysfunction is a function of altered production of vasoconstrictors and
vasodilators. In the 2004 PM AQCD (U.S. EPA, 2004, 056905), studies examining ET as an
activator of vasoconstriction were limited to those conducted by Bouthiller et al. (1998,
087110) and Vincent et al. (2001, 021181), in which increased plasma endothelin (ET) levels
were observed in rats exposed to high concentrations (40 or 5 mg/m3) of resuspended
Ottawa (EHC-93) or diesel PM, respectively. The authors postulated that PM altered
vasoconstriction via elevated ET. No studies were cited in the 2004 PM AQCD (U.S. EPA,
2004, 056905) that looked at direct measures of vasoreactivity.
As this area is newly emerging, numerous studies are included below that utilize IT
exposure or high concentrations! the studies that exposed vessels directly to particles ex
vivo are included in Annex D only, as their relevance is questionable. There is clearly a need
for more toxicological research examining the relationship between vascular measurements
July 2009
6-47
DRAFT-DO NOT CITE OR QUOTE

-------
and PM exposures using ambient particles at lower concentrations. Furthermore, no new
studies have advanced the knowledge in regards to ET as abiomarker of PM-induced
vasoconstriction since the last PM review.
CAPs
SD rats were exposed to PM2.5 CAPs (5 h/day x 3 days! daily mean mass concentration
73.5-733 |Jg/m3; Boston, MA; 3/1997-6/1998) then the pulmonary arterial vasculature was
evaluated (Batalha et al., 2002, 088109). Some animals were repeatedly exposed to SO2
(5 h/day x 5 days/wk x 6 wks) to induce chronic bronchitis. Morphometric measurements
indicated that the pulmonary artery lumen-to-wall (L/W) ratio (an indicator of arterial
narrowing) was decreased for the both CAPs groups compared to the normal/air group.
Furthermore, decreased L/W ratio in CAPs-exposed animals (regardless of pre-treatment)
was significantly associated with particle mass and composition when the mean
concentrations from the second and third exposure days were used in a univariate linear
regression. These results indicate a change in vascular tone following acute exposure to
PM.
Diesel Exhaust
The venous circulation plays a prominent role in heart failure exacerbation (Gehlbach
and Geppert, 2004, 155784). In heart failure, patients are often volume overloaded and are
subsequently placed on diuretics to alleviate symptoms of pulmonary congestion and chest
pain. Knuckles et al. (2008, 191987) hypothesized that if veins constrict in a manner similar
to arteries, then patients with severe heart failure may have temporary shunting of fluid to
the pulmonary circulation, which may elicit signs and symptoms of heart failure. Using
mesenteric vessels from mice (C57BL/6) exposed to DE (350 (Jg/m3 x 4 h; MMD 100 nm,
CMD 80 nm), the authors reported a significant enhancement of ET-l-induced
vasoconstriction in veins with much weaker responses in arteries. In an ex vivo experiment,
venous constriction was blocked by the arginine analog, L-NAME, which eliminates the
feedback NOS activation via endothelial ETb receptors! this is indicative of impaired or
uncoupled eNOS. The authors hypothesized that volatile organic compounds might be
responsible these effects, but no significant effects were observed for acetaldehyde,
formaldehyde, acetone, hexadecane, or pristane.
July 2009
6-48
DRAFT-DO NOT CITE OR QUOTE

-------
Model Particles
A study by Nurkiewicz et al. (2008, 156816) compared the arteriole dilation responses
in the spinotrapezius muscle with inhalation exposure to fine or ultrafine Ti02 (l |Jm and
21 nm, respectively! mean mass concentration 3" 16 and 1.5-12 mg/m3, respectively) for
durations of 4-12 h in SD rats. Both size fractions of Ti02 induced impaired dilation with a
NOdependent Ca2+ ionophore in a dose-dependent manner. When ultrafine and fine TiCh
were compared at similar mass doses, the systemic microvascular dysfunction was greater
with the UFPs. Furthermore, three exposures of differing durations and concentrations
that produced equal calculated pulmonary deposition of ultrafine TiCh (30 |Jg) demonstrated
similar dilation responses, indicating that impairment is dependent upon the time x
concentration product. No effects on dilation were observed with a dose of 4 |jg ultrafine
Ti02 (1.5 mg/m3 for 4 h) or 8 |Jg fine TiCh (3 mg/m3 for 4 h).
In a followup study, Nurkiewicz et al. (2009, 191961) examined the effect of
pulmonary fine and ultrafine Ti02 exposure on endogenous microvasculature NO
production in SD rats. The exposure concentrations and durations were selected to produce
-50% impairment of microvascular reactivity (67 and 10 |Jg for fine1 and ultrafine2 Ti02,
respectively). Similar to the study above (Nurkiewicz et al., 2008, 156816), impaired
endothelium-dependent arteriolar dilation was observed 24 h post-exposure with infusion of
a Ca2+ ionophore. Earlier studies that used residual oil fly ash (ROFA) or TiCh IT
instillation reported similar findings, regardless of particle type (Nurkiewicz et al., 2004,
087968; Nurkiewicz et al., 2006, 088611). There was no difference in arteriolar dilation
between sham and TiC>2 exposed groups with direct administration of the NO donor sodium
nitroprusside (SNP) to the exterior arteriolar wall and this response was consistent with
that observed following ROFA administered IT (Nurkiewicz et al., 2004, 087968). The lack
of response to SNP indicates that vascular smooth muscle sensitivity to NO is not altered
after particle exposure. The amount of ROS in the microvascular wall was increased
following exposure to either Ti02 sizes. Local ROS may consume endothelial-derived NO
and generate peroxynitrite radicals, as microvascular nitrotyrosine (NT) formation (the end
product of peroxynitrite reactions) was demonstrated after Ti02 exposure. NO production
was compromised in a dose-dependent manner following particle exposure (8-90 |Jg for fine
1	Produced by a 300-min exposure to 16 mg/m3 of fine Ti02
2	Produced by a 240-min exposure to 6 mg/m3 of ultrafine TiC>2
July 2009
6-49
DRAFT-DO NOT CITE OR QUOTE

-------
and 4-38 |Jg for ultrafine TiCb), and was partially restored with agents for radical
scavenging or enzyme inhibition for NADPH oxidase and myeloperoxidase (MPO).
Intratracheal Instillation
Nurkiewicz et al. (2004, 087968; 2006, 088611) have shown impairment of
endothelium-dependent dilation in the systemic microvasculature of SD rats following
ROFAor Ti02 exposure (0.1 or 0.25 mg/rat). NOindependent arteriolar dilation was also
impaired by ROFA arteriole adrenergic sensitivity to phenylephrine (PHE) was not affected
by 0.25 mg ROFA, indicating that contractile activity was unchanged. In addition, increased
venular leukocyte rolling and adhesion in the spinotrapezius muscle was also observed
following ROFA exposure (Nurkiewicz et al., 2004, 087968).
Further characterization of the leukocyte adherence and "rolling" effects for both
ROFA and Ti02 were indicative of an activated endothelium (Nurkiewicz et al., 2006,
088611). Vascular deposition of MPO was observed in the spinotrapezius muscle 24 h post-
exposure and the authors suggested that the adherent leukocytes may have deposited the
MPO to be taken up by endothelial cells (Nurkiewicz et al., 2006, 088611). However, this is
in contrast to another study (Cozzi et al., 2006, 091380) that did not find changes in blood
neutrophil MPO release in ICR mice exposed to ultrafine PM (100 |Jg from Chapel Hill, NC;
assessed 24 h post-exposure), although this finding may be a reflection of differing
protocols. Increased oxidative stress in the arteriolar wall was also reported with exposure
to 0.25 mg ROFA. Ti02 and ROFA induced varying degrees of pulmonary inflammation in
these animals, but elicited very similar vascular effects, indicating that the vascular
responses may be due to PM presence in the lung rather than its physiochemical properties
or intrinsic pulmonary toxicity.
PM10
Tamagawa et al. (2008, 191988) reported reduced acetylcholine (ACh)-stimulated
relaxation in carotid arteries from rabbits (New Zealand White) exposed to PMio (EHC-93)
via intrapharyngeal instillation for 5 days or 4 wk (total doses 8 and 16 mg/kg,
respectively). Endothelium-dependent NO-mediated vasorelaxation correlated with
increased serum IL-6 levels in the acute study and during weeks 1 and 2 of the 4-wk
exposure, which may indicate a role for systemic inflammation in the response. Maximal
SNP-induced dilation was not affected by PM exposure, indicating that the dilatory
response was not acting via endothelium-independent NO-mediated mechanisms. This
July 2009
6-50
DRAFT-DO NOT CITE OR QUOTE

-------
finding is consistent with that by Nurkiewicz et al. (2004, 087968) and suggests that the
arteriolar smooth muscle is not involved in the PM-impaired dilatation response.
Vasoreactivity of aortic rings was measured in SH rats following exposure to 10 mg/kg
PMio (EHC-93), with an increase in ACh-induced vasorelaxation observed (Bagate et al.,
2004, 087945). This endothelium-dependent response was greatest at 4 h and was still
present at 24 h. Similarly, vasorelaxation induced by SNP 4-h post-PM exposure was
enhanced. The vasorelaxation response was attenuated after denudation of the aortic rings,
suggesting that the effect was endothelium-dependent. The findings of enhanced dilation
with PM exposure contrast with those reported by Nurkiewicz et al. (2004, 087968; 2006,
088611), Tamagawa et al. (2008, 191988), and Cozzi et al. (2006, 091380) and may be
attributable to differences in PM type, animal species, or disease models. The authors
attribute their findings to the SH rat as a well-documented model of sympathetic
hyperactivity (increased affinity of aortic smooth muscle cradrenergic receptors) that
demonstrates upregulation of NO formation and/or release (Safar et al., 2001, 156068). No
change in vasoconstriction was observed with PM with PHE or potassium chloride.
Consistent with the impaired vasodilatory responses observed in the
microvasculature and aortic rings following PM exposure, Courtois et al. (2008, 156369)
demonstrated less relaxation to ACh in intrapulmonary arteries of Wistar rats exposed to a
high dose (5 mg) of ambient PM (SRM1648). This response was only observed 12 h after PM
exposure and not at shorter (6 h) or longer (24 or 72 h) time points. Fine TiCh did not alter
ACh-induced relaxation.
Ultrafine PM
Cozzi et al. (2006, 091380) used ICR mice to examine the effects of ultrafine PM
exposure (100 |Jg collected from Chapel Hill, NC) on vascular reactivity following PM
exposure and ischemia/reperfusion injury. Aortic rings were evaluated for their contractile
and dilatory responses 24 h post-exposure and following the ischemia/reperfusion protocol.
Maximum ACh-induced relaxation was impaired in ultrafine PM-exposed vessels, as well as
a rightward shift in sensitivity to ACh. There was no difference in constriction to PHE
between aortic rings from control and PM-exposed mice. The reduced ACh-induced
relaxation may be important for reperfusion of critical vascular beds following occlusion,
potentially leading to a greater area of infarction (as in this study). Anew study in dogs
supports the results observed in the above study and provides evidence of reduced
July 2009
6-51
DRAFT-DO NOT CITE OR QUOTE

-------
myocardial blood flow following PM exposure (Bartoli et al., 2009, 179904), and is discussed
in more detail in Section 6.2.3.3.
Summary of Toxicological Study Findings for Vasoreactivity
The toxicological findings with respect to vascular reactivity are generally in
agreement and demonstrate impaired dilation following PM exposure that is likely
endothelium dependent. These effects have been demonstrated in varying vessels (right
spinotrapezius muscle, carotid arteries, and aortic rings) and in response to different PM
types (ROFA, TiCh, EH093, ultrafine ambient PM). The work by Nurkiewicz et al. (2004,
087968; 2006, 088611; 2008, 156816; 2009, 191961) supports a role for increased ROS and
RNS production in the microvascular wall that leads to altered NO bioavailability and
dysfunction following particle exposure. Only one study showed enhanced dilation with PM
exposure, but the authors attributed the conflicting results to the SH rat. No constriction
changes in response to PHE were observed following PM exposure. The responses observed
in the pulmonary circulation after PM exposure include pulmonary vasoconstriction,
decreased LAV ratio, and impaired vasodilation in intrapulmonary arteries. These results
are consistent and indicate altered vascular tone. Enhancement of vasoconstriction in
mesenteric veins following DE is the first study of its kind to report on venous circulatory
effects.
Endothelin
In addition to studies that look at vascular reactivity, three recent studies have
examined plasma ET levels following exposure to vehicle emissions and a few studies
examined the mRNA expression of ET-1 and the ET receptors in the hearts of rodents
following PM exposure.
CAPs
The upregulation of mRNA expressions of ET-1 and the ETa receptor in WKY rats
exposed to CAPs (l or 4 days! 4.5 h/day! mean mass concentration range 1,000-1,900 |Jg/m3;
5/2004, 11/2004, 9/2005; Yokohoma City, Japan) was correlated with increasing PM
cumulative mass collected on chamber filters (Ito et al., 2008, 096823). Furthermore,
relative cardiac mRNA expressions of ET-1 and ETa receptor were significantly correlated
with CYP1B1 and HO-1 expression, indicating a possible relationship between ET-1
metabolism or oxidative stress.
July 2009
6-52
DRAFT-DO NOT CITE OR QUOTE

-------
Another plasma mediator of vasomotor tone is asymmetric dimethylarginine (ADMA),
which is an endogenous inhibitor of NOS that is associated with impaired vascular function
and increased cardiovascular events. Dvonch et al. (2004, 055741) assessed levels of ADMA
in Brown Norway rats 24-h following a 3"day PM2.5 CAPs exposure in southwest Detroit
(8 h/day! 7/2002). CAPs (mean mass concentration 354 (Jg/m3) resulted in increased plasma
ADMA compared to air controls, although the levels reported were well below the 2 (jM/L
range associated with increased CVD risk in humans in chronic studies. Therefore, the
preliminary results identified a new potential biomarker of vascular tone that had not
previously been used in air pollution toxicological studies.
Traffic-Related Particles
A study of old rats (21 month; Fischer 344) exposed to on-road highway aerosols
(number concentration range 0.95-3.13 x 105 particles/cm3; Interstate 90 between Rochester
and Buffalo, NY) for 6 h demonstrated decreased plasma ET-2 (18 h post-exposure) and
unchanged levels of ET-1 and ET-3 (Elder et al., 2004, 087354).
Gasoline Exhaust
In contrast to the study above, circulating levels of ET-1 (measured 18 h
post-exposure) were elevated in animals exposed to gasoline exhaust and filtration of
particles did not reduce this effect (see study details in Section 6.2.2.2) (Campen et al.,
2006, 096879). The results of Campen et al. (2006, 096879) are consistent with those
observed by Bouthillier et al. (1998, 087110) following a very high exposure to EHC-93, but
it is difficult to attribute the effects to PM alone, as Campen et al. (2006, 096879) showed
that the gaseous components of the gasoline mixture were required for the ET-1 increase.
Aorta ET-1 mRNA expression was increased with a 7 day gasoline exhaust exposure
(60 |Jg/m3) in ApoE"'" mice, but was not changed following a single day exposure (Lund et al.,
2009, 180257). The expression and activity of MMP-2 and -9 and oxidative stress in aortas
of exposed mice were also elevated. The ET-1 and MMP-9 mRNA expressions were
attenuated with the addition of an ETa receptor antagonist (but not a radical scavenger),
indicating that ET-1 may mediate the expression of MMP-9 through the ETa receptor.
Model Particles
Another study examined the effects of ultrafine carbon particles (mass concentration
172 (Jg/m3; mean number concentration 9.0 x 106 particles/cm3) and there was no difference
in ET-1, ETa or ETb receptor mRNA expression between air- and particle-exposed SH rats 1
July 2009
6-53
DRAFT-DO NOT CITE OR QUOTE

-------
or 3 days post-exposure (Upadhyay et al., 2008, 159345). In lung homogenates, ET-1, ETa
and ETb receptor mRNA expressions were elevated 3 days after exposure to ultrafine
carbon particles (Upadhyay et al., 2008, 159345).
Summary of Toxicological Study Findings for Endothelin
The ET responses were mixed, with one study demonstrating ET-1 increases after
exposure to gasoline emissions that were particle independent and another reported
decreased ET-2, but no change in ET-1 or ET-3 with on-road highway exposure. Elevated
levels of ET-1 and ETa receptor mRNA expression were noted in hearts of rats exposed to
CAPs, but not in rats exposed to ultrafine carbon particles. However, ET-1, ETa and ETb
receptor mRNA expressions were increased in lung homogenates of rats following ultrafine
carbon exposure. The ETa receptor was found to be involved in the ET-1 and MMP-9
responses in the aortas of mice exposed to gasoline exhaust. A relatively novel marker,
AD MA, was used to evaluate vasomotor tone in rats and was found to be elevated following
exposure to CAPs, although the results are preliminary and have not been confirmed.
PM Components
In the Batalha et al. (2002, 088109) study described above, univariate analyses were
conducted that regressed log LAV on differential exposure concentrations of tracer elements
determined using principal components analysis. For CAPs exposure (regardless of
pre-treatment), CAPs mass, Si, Pb, SO42-, EC, and OC were all negatively correlated with
LAV ratio. Si and SO r were negatively correlated with LAV ratio in normal rats and Si and
OC were negatively correlated with LAV ratio in bronchitic rats. When a multivariate
analysis was conducted using normal and bronchitic animals, only the association with Si
remained significant. V was not associated with LAV ratio in any analysis.
6.2.5. Blood Pressure
One of the potential outcomes of air pollution-mediated alterations in vascular tone is
its impact on variable BP or hypertension. BP is tightly regulated by autonomic (central
and local), cardiac, renal, and regional vascular homeostatic mechanisms with changes in
arterial tone being countered by changes in cardiac contractility, HR, or fluid volume. The
evidence of PM-induced changes in BP presented in the 2004 PM AQCD (U.S. EPA, 2004,
056905) is limited and inconsistent. Recent epidemiologic, controlled human exposure, and
July 2009
6-54
DRAFT-DO NOT CITE OR QUOTE

-------
toxicological studies have similarly reported conflicting results regarding the effect of PM
on BP. However, the majority of these studies have evaluated changes in BP at some point
following exposure to PM. A significant increase in DBP was observed in the only controlled
human exposure study that evaluated BP during exposure (concomitant exposure to CAPs
and O3). In addition, evidence from toxicological studies suggests that the effect of PM on
BP may be modified by health status, as PM-induced increases in BP have been more
consistently observed in SH rats.
6.2.5.1. Epidemiologic Studies
Increased BP was associated with PM concentration in two of three studies reviewed
in the 2004 PM AQCD (U.S. EPA, 2004, 056905). Increases in left ventricular BP (systolic
and diastolic) are well established risk factors for cardiovascular mortality/morbidity (Welin
et al., 1993, 156151). Changes in HR and BP both reflect changes in autonomic tone, and
have been examined following short-term increases in PM pollution in several recent
studies.
Ibald-Mulli et al. (2004, 087415) examined associations between BP (systolic [SBP],
diastolic [DBP]) and ambient PM2.5 concentrations, UFP counts, and ACP counts in a
multicity panel study (Amsterdam, Netherlands! Helsinki, Finland! Erfurt, Germany) of
131 adults with coronary heart disease. Although based on the same ULTRA Study
(Tim on en et al., 2006, 088747) with study methods as described previously in
Section 6.2.1.1, the study period was different. They investigated changes in BP (SBP and
DBP) associated with mean PM2.5, UFP, and ACP concentration/counts (lag days 0, 1, and 2,
as well as the 5"day mean) in each city and then generated a pooled estimate across the
cities. The median PM2.5 concentration, median UFP count, and median ACP count for each
city are given below in Table 6-3. Pooled analyses across all 3 cities showed small, but
statistically significant decreases in SBP and DBP associated with different lagged
concentrations/counts of each particulate pollutant. Each 10 ug/m3 increase in the mean
PM2.5 concentration over the previous 5 days was associated with a 0.36 mmHg decrease in
SBP (95% CI: -0.99, 0.27) and a 0.39 mmHg decrease in DBP (95% CI: -0.75 to -0.03). Each
10,000 particles/cm3 increase in UFP was associated with a 0.72 mmHg decrease in SBP
(95% CP -1.92 to 0.49), and a 0.70 mmHg decrease in DBP (95% CP -1.38 to -0.02). Each
1,000 particles/cm3 increase in 5 day avg ACP was associated with a 1.11 mmHg decrease in
SBP (95% CI: -2.12 to -0.09) and a 0.95 mmHg decrease in DBP (95% CI: -1.53 to -0.37). The
July 2009
6-55
DRAFT-DO NOT CITE OR QUOTE

-------
authors concluded that these findings do not support previous findings of an increase in BP
associated with increases in particulate pollutant concentrations.
Table 6-3.
Median particle concentrations.

City
Median daily PM2.5 Median daily UFP concentration
concentration (ug/m3) (0.01-0.1 |jm; particles/cm3)
Median daily ACP concentration
(0.1-1.0 |jm; particles/cm3)
Amsterdam,
Netherlands
16.9 17,147
1,874
Erfurt, Germany
16.3 19,198
1,492
Helsinki, Finland
10.6 14,886
1,200
Single-city studies examining the association between BP and particulate air
pollution have been done in several U.S. and Canadian cities. Dales et al. (2007, 155743)
conducted a panel study of 39 healthy volunteers who sat outside at two different bus stops
for 2-h in Ottawa, Canada. The median PM2.5 concentrations measured at the bus stops
during each 2-h exposure session were 40 and 10 jug/m3. Post-exposure SBP and DBP were
not associated with the mean PM2.5 concentration measured at the bus stops during the 2-h
exposure session. The change in BP (post-exposure - pre-exposure) could not be evaluated,
as health measurements were only made after the 2-h exposure session.
Jansen et al. (2005, 082236) studied changes in BP among 16 older subjects (aged
60-86 yr) with asthma or COPD in Seattle, Washington, associated with indoor, outdoor,
and personal PM10, PM2.5, and BC measurements (levels within the health measurement
session) on 12 consecutive days. The mean daily outdoor PM10 and PM2.5 concentrations
were 13.47 and 10.47 |ug/m3, respectively. The mean daily outdoor BC concentration was
2.01 |ug/m3. The study authors reported that no associations were observed between BP and
daily mean PM10, PM2.5, or BC concentrations, but did not present any of these results in
the paper.
Zanobetti et al. (2004, 087489) examined the association between BP (SBP, DBP, and
mean arterial BP) and mean PM2.5 concentrations in the previous 24, 48, 72, 96, and 120 h
in 62 elderly, cardiac rehabilitation patients in Boston, MA (Zanobetti et al., 2004, 087489).
The median PM2.5 concentration during the study was 8.8 ug/ma. Each 10.4 ug/m3 increase
in mean PM2.5 concentration in the previous 120 h was associated with significant increases
in resting DBP (2.82 mmHg [95% CI: 1.26-4.41]), SBP (2.68 mmHg [95% CI: 0.04-5.38]), and
mean arterial BP (2.76 mmHg [95% CI: 1.07-4.48]).
July 2009
6-56
DRAFT-DO NOT CITE OR QUOTE

-------
Mar et al. (2005, 087566) studied this same PM2.5-BP association in 88 subjects aged
>57 yr in Seattle, WA. Among healthy subjects taking medications (bronchodilators, inhale
corticosteroids, anti-hypertensives, frblockers, calcium channel blockers, and/or cardiac
glycosides), each 10 |ug/m3 increase in mean outdoor PM2.5 concentration on the same day as
the BP measurement was made was associated with small increases in SBP and DBP (the
results for these analyses were presented in figures only). However, among all subjects,
each 10 |ug/m3 increase in same day mean PM2.5 concentration was associated with
non-significant decreases in SBP (-0.81 mmHg [95% CP -2.34 to 0.73]) and DBP
(-0.46 mmHg [95% CI: -1.49 to 0.57]).
As described earlier, Ebelt et al. (2005, 056907) conducted a repeated measures panel
study of 16 patients with COPD in the summer of 1998 in Vancouver, British Columbia to
evaluate the relative impact of ambient and non-ambient exposures to PM2.5, PM10, and
PM10-2.5 on multiple health outcomes including ectopy and BP. Using the same analytic
methods, pollutant concentrations, and lags, they reported decreased SBP associated with
same day ambient exposures to each PM size fraction (results were presented in figures
only).
Two similar studies were done in Incheon, South Korea (Choi et al., 2007, 093196) and
Taipei, Taiwan (Chuang et al., 2005, 156356). Both reported significant increases in BP
associated with acute increases in ambient PM. Choi et al. (2007, 093196) reported
significantly increased SBP and DBP associated with the mean PM10 concentration over the
same and previous 2 days in the warm season only (July to September). Chuang et al.
(2005, 156356) reported significant increases in SBP and DBP associated with the mean
UFP count (0.01 to 0.1 jum particles) 1*3 h before the BP measurement.
These studies (Choi et al., 2007, 093196; Chuang et al., 2005, 156356; Dales et al.,
2007, 155743; Ibald-Mulli et al., 2004, 087415; Mar et al., 2005, 087566; Zanobetti et al.,
2004,	087489) are not entirely consistent with regard to their BP-PM associations. Most
have reported increases in SBP and DBP associated with increases in either PM2.5, PM10, or
UFP (Choi et al., 2007, 093196; Chuang et al., 2005, 156356; Mar et al., 2005, 087566;
Zanobetti et al., 2004, 087489). However, two studies reported small decreases in BP
associated with multiple particulate pollutants (Ibald-Mulli et al., 2004, 087415; Mar et al.,
2005,	087566), Dales et al. (2007, 155743) reported no change in BP associated with a 2-h
exposure to bus stop PM2.5 and Jansen at al. (2005, 082236) reported null findings among
older adults in Seattle, WA. Exposure lags ranging from 1-3 h (Chuang et al., 2005,
July 2009
6-57
DRAFT-DO NOT CITE OR QUOTE

-------
156356), to the same day (Ebelt et al., 2005, 056907; Mar et al., 2005, 087566; Rich et al.,
2008, 156910), to the mean across the previous 5 days (Zanobetti et al., 2004, 087489) were
reported as having the strongest associations with BP.
Right Ventricular Pressure
Several recent studies, summarized in the section on hospital admissions and ED
visits for CVD causes, have reported increased risk of hospital admissions for congestive
heart failure associated with increased PM concentration on the same day (Wellenius et al.,
2005, 087483; Wellenius et al., 2006, 088748). As a possible mechanism for these reported
associations, Rich et al. (2008, 156910) hypothesized that these hospital admissions for
decompensation of heart failure would be preceded by more subtle increases in pulmonary
arterial (PA) and right ventricular (RV) diastolic pressures. They used passively monitored
PA and RV pressures on 5807 person-days, among 11 subjects implanted with the Chronicle
Implantable Hemodynamic Monitor [Medtronic, Inc. Medtronic, MN]). Using a two-stage
modeling process (generalized additive model and mixed effects model adjusted for time
trend, weekday, calendar month, apparent temperature, and barometric pressure), they
examined the change in daily mean right heart pressures associated with mean PM2.5
concentration on the same and previous 6 days. Each 11.62 ug/m3 increase in same day
mean PM2.5 concentration was associated with small but statistically significant increases
in estimated PA diastolic pressure (0.19 mmHg [95% CP 0.05-0.33]) and RV diastolic
pressure (0.23 mmHg [95% CP 0.11-0.34]). These effects were not attenuated when
controlling for all lags simultaneously. Thus, PM induced right heart pressure increases
may mark another potential pathway between PM exposure and incidence of cardiovascular
events, but further studies on this same hypothesis are needed for confirmation.
Wellenius et al. (2007, 092830) conducted a panel study of 28 subjects living in the
greater Boston metropolitan area, each with chronic stable heart failure and impaired
systolic function. They hypothesized that circulating levels of B-type natriuretic peptide
(BNP), measured in whole blood at 0, 6, and 12 wk, were associated with acute changes in
ambient air pollution, as a possible mechanistic explanation for the observed association
between hospital admissions for heart failure and ambient PM concentration (Wellenius et
al., 2005, 087483; Wellenius et al., 2006, 088748). During the study, the mean PM2.5
concentration was 10.9 |ug/m3, while the mean BC concentration was 0.73 ug/m3. Using
linear mixed models, they reported no association between any pollutant (PM2.5, CO, SO2,
July 2009
6-58
DRAFT-DO NOT CITE OR QUOTE

-------
N02, O3, and BC) and BNP at any lag (e.g., each 10 |u,g/m3 increase in mean daily PM2.5
concentration [0.8% increase in BNP (95% CP -16.4 to 21.5)]) (Wellenius et al., 2007,
092830). Yet, BNP the active peptide has a very short half-life and might not be the best
biomarker for such a study. Thus the absence of a correlation between PM and BNP may
not suggest that PM does not have an impact on RV or LV function in individuals with
impaired cardiac mechanics.
6.2.5.2. Controlled Human Exposure Studies
Only one controlled human exposure study cited in the 2004 PM AQCD (U.S. EPA,
2004, 056905) reported any PM-induced changes in BP. Gong et al. (2003, 042106) found
that exposure to PM2.5 (174 |jg/m3) decreased SBP in asthmatics, but increased SBP in
healthy subjects. Among healthy adults, BP was not affected following 2-h exposures to
200 |jg/m3 diesel PM (Nightingale et al., 2000, 011659), 150 |jg/m3 PM2.5 CAPs with 120 ppb
O3 (Brook et al., 2002, 024987), or 10 |jg/m3 ultrafine carbon particles (Frampton, 2001,
019051). The effect of PM on BP has been further investigated in several recent controlled
human exposure studies, which are described below.
CAPs
One recent study demonstrated a significant increase (9.3%) in DBP among healthy
adults immediately prior to the end of a 2-h exposure to 150 (Jg/m3 PM2.5 CAPs in
combination with 120 ppb O3 (Urch et al., 2005, 081080). The authors also found that the
magnitude of change in BP was significantly associated with PM2.5 carbon content, but not
total PM2.5 mass. It was postulated that the disparity between these findings and those of a
similar study by the same group (Brook et al., 2002, 024987) could be due to differences in
experimental methods. The Brook et al. (2002, 024987) study measured post-exposure BP
approximately 10 min following exposure, while the study by Urch et al. (2005, 081080)
measured BP during exposure. In a follow up study that evaluated changes in blood
pressure during a 2-h exposure to PM2.5 CAPs, Fakhri et al. (2009, 19191 1) reported a
significant increase in DBP with exposure to CAPs with, but not without, coexposure to O3.
Diesel Exhaust
Several recent studies have assessed BP changes following a 1-h exposure to DE with
a particle concentration of 300 |Jg/m3. Mills et al. (2005, 095757) evaluated changes in BP
2 h following exposure to DE and found a 6 mmHg increase in DBP of marginal statistical
July 2009
6-59
DRAFT-DO NOT CITE OR QUOTE

-------
significance (p = 0.08) compared to filtered air control. In this same group of subjects,
Tornqvist et al. (2007, 091279) did not observe any such changes in BP 24 h following DE
exposure. At lower particle concentrations in diluted DE (100-200 |Jg/m3 fine PM), Peretz
et al. (2008, 156854) did not observe any changes in systolic or DBP in either healthy adults
or adults with metabolic syndrome immediately following a 2-h exposure. Further, although
Lundback et al. (2009, 191967) reported an increase in arterial stiffness following exposure
to DE with a particle concentration of 330 (Jg/m3 among healthy young adults, no changes in
systolic or diastolic BP were observed during or following exposure relative to filtered air.
Model Particles
Routledge et al. (2006, 088674) did not observe any changes in BP among healthy
older adults and older adults with stable angina following a 1-h exposure to UF EC
(50 |Jg/m3), with or without coexposure to 200 ppb SO2. Similarly, neither Shah et al. (2008,
156970), nor Beckett et al. (2005, 156261) reported any changes in BP among healthy
adults following exposure to UF EC (50 [Jg/m31 or ZnO (500 (Jg/m3 fine and ultrafine),
respectively.
Summary of Controlled Human Exposure Study Findings for Blood Pressure
The findings of these new studies do not provide convincing evidence of an association
between PM exposure and an increase in BP; however, they do suggest that there is a need
for additional investigations of PM-induced changes in BP at various time points following
exposure.
6.2.5.3. Toxicological Studies
In healthy animal models, little evidence exists for significant BP changes following
inhalation exposure to environmentally-relevant concentrations of PM. Only one animal
toxicological study is mentioned in the 2004 PM AQCD (U.S. EPA, 2004, 056905) that
examined BP with PM exposure and no effect was observed (Vincent et al., 2001, 021181).
CAPs
In a recent study of dogs, exposure to PM2.5 CAPs from Boston (mean mass
concentration 358.1 |Jg/m3; mass concentration 94.1-1557 (Jg/m3) for 5 h resulted in
increased SBP (2.7 mmHg), DBP (4.1 mmHg), mean arterial pressure (3.7 mmHg), and
lowered pulse pressure (1.7 mmHg) when measured upstream of the femoral artery (Bartoli
et al., 2009, 156256). Administration of an a-adrenergic antagonist (prazosin) prior to CAPs
July 2009
6-60
DRAFT-DO NOT CITE OR QUOTE

-------
attenuated the BP responses. These findings indicate that CAPs exposure may have
activated a-adrenergic receptors and increased peripheral vascular resistance. Baroreflex
sensitivity was measured immediately before and after exposure during a transient
elevation of arterial pressure that was induced by PHE; increased baroreflex sensitivity
was observed in subgroup of dogs exposed to CAPs, which is consistent with an
upregulation of vagal reflexes.
Chang et al. (2004, 055637) noted slight increases in SH rat BP (5-10 mmHg) when
exposed to PM2.5 CAPs (mean mass concentration 202 ug/ma) during spring months.
However, during summer months, when the CAPs exposure level was less (140 ug/m3), this
effect was not observed. It was unclear, therefore, whether the effects were seasonal or
dose-related. In a preliminary study of SH rats exposed to CAPs during a dust storm event,
mean BP was elevated the third and fourth hour of a 6-h exposure, although interpretation
of this finding is difficult due to few animals in the exposure group (n = 2; details provided
above) (Chang et al., 2007, 155719). In another study, the increased change in mean BP
measured using the tail cuff method following CAPs exposure weakly correlated with PM
mass accumulated on chamber filters over the entire exposure duration (see Section 6.2.4.3
for details) (Ito et al., 2008, 096823). Furthermore, ETa receptor mRNA expression in
cardiac tissue was positively correlated with the change in mean BP.
Model Particles
In WKY rats, 24-h exposure to ultrafine carbon particles (mass concentration 180
(jg/m3; mean number concentration 1.6 x 107 particles/cm3) did not alter mean BP during
exposure or the recovery periods (Harder et al., 2005, 087371). SH rats exposed to ultrafine
carbon particles for 24 h (mass concentration 172 (Jg/m3; mean number concentration (9.0 x
106 particles/cm3) resulted in elevated mean BP (by 6 mmHg) on the first and second days of
recovery following exposure that was attributable to increases in both SBP and DBP
(Upadhyay et al., 2008, 159345). Increased plasma renin concentrations were observed in
CB-exposed rats on the first and second days of recovery, although renin activity and
angiotensin (Ang) I and II concentrations were not affected by particle exposure.
Summary of Toxicological Study Findings for Blood Pressure
Limited toxicological evidence provides support for elevated BP in dogs or
compromised rats with CAPs, ultrafine CAPs, CAPs during a dust storm event, or ultrafine
July 2009
6-61
DRAFT-DO NOT CITE OR QUOTE

-------
carbon particle exposure. However, most of the CAPs studies were conducted outside of the
U.S.
6.2.6. Cardiac Contractility
The 2004 PM AQCD (U.S. EPA, 2004, 056905) did not include any toxicological
studies that evaluated cardiac contractility either directly or indirectly following exposure
to PM. Two recent animal toxicological studies have demonstrated reductions in cardiac
fractional shortening, diminished ejection shortening, or changes in the QA interval
following PM exposure. The results of these studies provide some evidence of PM-induced
changes in cardiac contractility in animal models.
6.2.6.1. Toxicological Studies
The strength of the contracting heart is reflected by its contractility. In heart failure,
contractility wanes significantly and the heart cannot compensate during periods of
increased physical activity. Measuring true contractility in a whole animal is difficult,
requiring extensive surgical instrumentation and monitoring. There were no toxicological
studies that examined cardiac contractility in the last PM AQCD.
CAPs
Using radiotelemetry to indirectly measure cardiac contractility through the QA
interval, SH rats were repeatedly and alternately exposed to ultrafine CAPs in Taiwan on
separate days in spring or summer (details provided in Section 6.2.5.3) (Chang et al., 2004,
055637). The QA interval was calculated as the time duration between the Q wave in the
ECG and point A (upstroke in aortic pressure) in the pressure trace and is not as reliable as
other measures, such as echocardiography. During the spring exposure, QA interval
decreased by 1.6 ms as demonstrated by fixed effects in linear mixed-effects modeling,
which indicates an increase in cardiac contractility. There were no changes in QA interval
observed for the summer months, which may be attributable to lower ultrafine PM
concentrations (mean mass concentration 140 (Jg/m3) or differing PM compositions.
Model Particles
A recent study using old (18 to 28-months-old) mice (C57BL/6, C3H/HeJ, and B6C3F1)
demonstrated significant reductions in cardiac fractional shortening (due to increased left
July 2009
6-62
DRAFT-DO NOT CITE OR QUOTE

-------
ventricular end-diastolic and end-systolic diameters) following a 4-day (3 h/day) exposure to
CB (PM2.5 mean concentration 401 |Jg/m3; PM10 mean concentration 553 (Jg/m3) using
echocardiography (Tankersley et al., 2008, 157043). Hemodynamic measurements of
diminished ejection fraction and maximum change in pressure over time further supported
lowered myocardial contractility Furthermore, increased right ventricular pressure
associated with elevated right atrial and pulmonary vascular pressures and resistance, was
indicative of pulmonary vasoconstriction in CB exposed mice. Heart tissue and isolated
cardiomyocytes from exposed animals demonstrated enhanced ROS that was partially
attributable to NOS3-uncoupling and elevated MMP2 and MMP9 levels, which may
implicate myocardial remodeling. The combined results from this study suggest that
cellular mechanisms involving NOS-uncoupled ROS generation likely mediate PM-induced
cardiac effects. Furthermore, mRNA expression for atrial and brain natriuretic peptides
(ANP and BNP, respectively) was increased in hearts from exposed mice compared to
control, which is consistent with pulmonary congestion. There were no reported
strain-related differences in any response.
Intratracheal Instillation
Similar to the responses observed by Tankersley et al. (2008, 157043), decreases in
fractional shortening and increases in left ventricular end diastolic diameter measured by
echocardiography were also reported for SD rats at 24 h post-IT exposure to DEP (250 |Jg)
(Yan et al., 2008, 098625). A subset of rats received isoproterenol to induce myocardial
injury prior to IT instillation of DEP and these animals demonstrated lowered fractional
shortening at baseline, which was decreased to an even greater extent with DEP exposure!
left ventricular end diastolic diameter was not affected by DEP in these rats.
Summary of Toxicological Study Findings for Cardiac Contractility
The studies above provide some evidence that cardiac contractility may be altered
immediately following PM exposure in animal models. Results from the Tanksersley (2008,
157043) and Yan (2008, 098625) studies provide the strongest support for PM-induced
contractility changes, as echocardiography and hemodynamic measurements are
well-established for examining cardiac function.
July 2009
6-63
DRAFT-DO NOT CITE OR QUOTE

-------
6.2.7. Systemic Inflammation
The evidence presented in the 2004 PM AQCD (U.S. EPA, 2004, 056905) of increases
in markers of systemic inflammation associated with PM was limited and not sufficient to
formulate a definitive conclusion. Recent controlled human exposure and toxicological
studies continue to provide mixed results for an effect of PM on markers of systemic
inflammation including cytokine levels, Oreactive protein (CRP), and white blood cell
count. While results from recent epidemiologic studies have also been inconsistent across
studies, there is some evidence to suggest that PM levels may have a greater effect on
inflammatory markers among populations with preexisting diseases.
6.2.7.1. Epidemiologic Studies
Several studies reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905)
investigated the association of short-term fluctuations in PM concentration with markers of
inflammation. These studies were found to offer limited support for mechanistic
explanations of the associations between PM concentration and heart disease outcomes.
Recent studies, published since 2002, are reviewed below. CRP was measured in multiple
studies, allowing the consistency of findings across epidemiologic studies to be evaluated.
Several other markers were examined in only a few studies, in relation to a wide range PM
size fractions and components. These markers included IL-6, TNF-a, vascular cell adhesion
molecule-1 (VCAM-l), intercellular adhesion molecule-1 (ICAM-l), soluble CD40 ligand
(sCD40L), white blood cells (WBC) and soluble adhesion molecules (sP-selectin and e-
selectin).
Diez-Roux et al. (2006, 156400) examined whether CRP increased in response to
changes in the mean ambient PM2.5 concentrations in the prior day, prior 2 days, prior
week, prior 30 days, and prior 60 days among participants in the Multi-Ethnic Study of
Atherosclerosis (MESA) cohort. Subjects (n = 5,634) lived in either Baltimore City or
County, MD, Chicago, IL, Forsyth County, NC, Los Angeles County, CA, Northern
Manhattan and the Bronx, NY, or St. Paul, MN. The authors report finding no evidence of a
short-term effect of PM2.5 on CRP in their population based sample. Of the five exposure
measures examined, only the 30-day and 60-day mean exposures showed positive
associations with PM2.5 (3% [95% CP -2% to 10%] and 4% [95% CP -3% to 11%] per
10 |ug/m3, respectively).
July 2009
6-64
DRAFT-DO NOT CITE OR QUOTE

-------
Ruckerl et al. (2007, 156931) conducted a multicity longitudinal study to examine
whether changes in markers of inflammation were associated with short-term increases in
particulate concentrations (PMio, PM2.5, particle number concentration [PNC]) and gaseous
pollutant (NO2, SO2, CO, O3). Study subjects were MI survivors (n= 1003) living in either
Athens, Greece! Augsburg, Germany! Barcelona, Spain! Helsinki, Finland! Rome, Italy! or
Stockholm, Sweden. Repeated measurements of IL-6 and CRP were made during the study
Fibrinogen was also measured in this study and results are discussed in Section 6.2.8.1.
The mean city-specific pollutant concentrations during the study are shown below in
Table 6-4.
Table 6-4. Ambient concentrations in six European cities.
Pollutant
Helsinki
Stockholm
Augsburg
Rome
Barcelona
Athens
PNC (particles/cm3)
8,534
9,748
11,876
34,450
18,133
20,589
PM2.6 l/yg/m3'
8.2
8.8
17.4
24.5
24.2
23.0
PM10 (^/g/m31
17.1
17.8
33.1
42.1
40.7
38.5
Source: Ruckerl et al. (2007,156931)
In pooled analyses, each interquartile range (not provided) increase in PNC in the 12
to 17 h before the health measurement was associated with a 2.7% increase in the
geometric mean IL-6 (95% CP 1.0-4.6). None of the pollutants, at any lag, were associated
with CRP levels in these subjects. There did not appear to be effect modification of these
results by smoking, diabetes, or heart failure. Ljungman et al. (2009, 191983) studied the
modification of the IL-6 association with several PM size fractions (PM10, PM2.5, PNC) by
three IL-6 SNPs, one fibrinogen alpha chain (FGA) SNP and one fibrinogen beta chain
(FGB) SNP. The associations of PM2.5 and PM10 with plasma level of IL-6 were stronger
among those with the homozygous minor allele genotype of FGB rsl800790 and among
those homozygous for the major allele genotype of IL-6 rs2069832. Gene-environment
interactions were most pronounced for CO, modification the PNC-IL-6 association by
genotype was not apparent in these data nor was modification of the PM-IL-6 associations
by FBA.
Single-city studies of systemic inflammation have been conducted in the U.S. and
Canada since 2002. Delfino et al. (2008, 156390) measured CRP, IL-6, TNF-a, sP-selectin,
sVCAM-1 and sICAM-1 in blood during a period of 12 wk. Associations of these markers
with average PM concentration (PM0.25, PM0.25-2.5, PM10-2.5, particle number concentration
July 2009
6-65
DRAFT-DO NOT CITE OR QUOTE

-------
(PNC), EC, OC, BC, primary OC, secondary OC) 24-h to 9 days prior to the blood draw were
examined. Subjects included residents of two downtown Los Angeles nursing homes who
were 65+ years old with a history of coronary artery disease. Both 24-h avg and multiday
average concentrations of PM0.25, EC, primary OC, BC, PNC and gaseous pollutants were
associated with CRP, IL-6 and sP-selectin in this study.
Pope et al. (2004, 055238) conducted a panel study of 88 non-smoking, elderly subjects
residing in the Salt Lake City, Ogden, and Provo metropolitan area of Utah. The mean
PM2.5 concentration during the study was 18.9 ug/m3. Each 100 ug/m3 increase in same day
mean PM2.5 concentration was associated with a 0.81 mg/dL increase in CRP (95%
CP 0.48-1.14), but not WBCs. However, when excluding 1 influential subject, each
100 |ug/m3 increase in same day mean PM2.5 concentration was associated with only a
0.19 mg/dL increase in CRP (95% CP -0.01 to 0.39). Several markers of coagulation were
examined in this study and are discussed in Section 6.2.8.1.
Zeka et al. (2006, 157177) studied 710 elderly members of the VA Normative Aging
Study to examine changes in CRP, sediment rate and WBC with acute changes in PM
concentrations in the previous 48 h, 1 wk, and 2 wk. Results for fibrinogen will be discussed
in Section 6.2.8.1. The median 48"h PNC was 24,200 particles/cm3, while the median 48"h
PM2.5 concentration was 9.39 |ug/m3. The median 48-h BC concentration was 0.61 |ug/m3,
while the median 48"h SO42" concentration was 1.84 jug/m3. They did not find consistent or
significant associations with any pollutant and CRP. The authors state that the largest
effects were observed for the mean PNC and BC concentration in the previous 4 wk and
their data suggests that effect PM may be modified by obesity, GSTM1 genotype and statin
use may be present.
O'Neill et al. (2007, 091362) conducted a cross-sectional study of 92 Boston residents
with type 2 diabetes, to examine the association between plasma levels of ICAM-1, VCAM-1
and PM concentrations. Results for markers of coagulation measured in this study are
discussed in Section 6.2.8.1. PM2.5, BC, and SO42" concentrations were measured 0.5 km
from the patient exam site. The mean PM2.5 concentration during the study was 11.4 ug/m3.
The mean BC concentration was 1.1 |ug/m3, and the mean SO42" concentration was
3.0 |ug/m3. For all moving averages examined (1-6 days), increases in mean PM2.5 and BC
concentration were associated with increased ICAM-1 and VCAM-1 concentrations. Each
7.6 |ug/m3 increase in the mean PM2.5 concentration over the previous 6 days was associated
with a 11.76 ng/mL increase in VCAM-1 (95% CP 3.48-20.70), and each 0.6 jug/m3 increase
July 2009
6-66
DRAFT-DO NOT CITE OR QUOTE

-------
in the mean BC concentration over the previous 6 days was associated with a 27.51 ng/mL
increase in VCAM-1 (95% CI: 11.96-45.21). However, there were no consistent associations
between the mean SO42" concentration at any lag and any marker.
Sullivan et al. (2007, 100083) conducted a panel study of n = 47 subjects (aged >55 yr)
either with COPD (n = 23) or without COPD (n = 24) in Seattle, WA. They examined the
association between levels of CRP and mean daily PM2.5 concentration. Most values for IL-6
and TNF-a were below the limit of detection so these markers were not included in the
analyses. Results for fibrinogen and d-Dimer are discussed in Section 6.2.8.1. The median
PM2.5 concentration during the study was 7.7 ug/ma. They did not find any associations
between 24-h mean PM2.5 concentrations and levels of CRP in individuals with or without
COPD. In the study by Liu et al. (2006, 192002). conducted in Toronto, Ontario, neither
CRP (0.11 |ug/mL [95% CI: -0.03 to 0.25]) nor TNF-a (0.03 pg/mL [95% CI: -0.07 to 0.13]) was
associated with 24-h mean PM10 concentration. However, significant positive associations
with markers of oxidative stress, FMD and BP were found and are discussed in Sections
6.2.9.1, 6.2.4.1, and 6.2.5.1, respectively.
In the St. Louis Bus Study, each 5.4 ug/ma increase in the mean PM2.5 concentration
over the previous week was associated with 5.5% increase in WBC (95% CP 0.10-11)
(Dubowsky et al., 2006, 088750). Each 6.1 |ug/m3 increase in the mean PM2.5 concentration
over the previous 5 days was associated with a 14% increase in CRP among all subjects
(95% CP -5.4 to 37), but an 81% increase in CRP (95% CP 21-172) among subjects with
diabetes, obesity, and/or hypertension. Associations between PM2.5 and IL-6 were only
observed among those with diabetes, obesity, and/or or hypertension. In another study of
in-vehicle PM2.5, each 10 ug/m3 increase during a work-shift was associated with decreased
lymphocytes, increased mean corpuscular volume, neutrophils, and CRP over the next
10-14 h among 9 healthy North Carolina state troopers (Riediker et al., 2004, 056992).
Associations of roadside and ambient PM2.5 were weaker and non-significant in this
population. Personal exposure to PM10 (24-h averaging time) was not associated with CRP,
IL-6, ET-1 or TNF-a in as study 25 diabetic patients in Windsor, Ontario (Liu et al., 2006,
192002).
International studies of the effect of air pollution on markers of inflammation have
been conducted with mixed results. Two studies conducted among 57 male patients with
coronary heart disease in Efurt, Germany found associations of UFP, ACP and PM10 with
CRP (Ruckerl et al., 2006, 088754) and UFP and ACP with sCD40L, a marker for platelet
July 2009
6-67
DRAFT-DO NOT CITE OR QUOTE

-------
activation (Ruckerl et al., 2007, 156931). In a large cross-sectional study of healthy subjects
in Tel Aviv, Steinvil et al. (2008, 188893) examined biological markers of inflammation (CRP
and WBC) collected as part of routine health examinations for 3,659 individuals.
Associations with air pollutants (including PMio) measured at local monitoring sites for the
day of the examination and up to 7 days prior were examined. No significant associations
were found between pollutant levels and indications of enhanced inflammation. By
contrast, both PMio and PM2.5 SC>42~ and nitrate (3-day avg concentrations) were associated
with increases in hs-CRP in healthy students in Taiwan (Chuang et al., 2007, 091063).
PMio, PM2.5 and PM0.25 were not associated with CRP in a study of MI patients in italy,
although associations with autonomic dysregulation and more severe arrhythmias were
observed (Folino et al., 2009, 191902). Kelishadi et al. (2009, 191960) reports that CRP, as
well as markers of insulin resistance and oxidative stress (discussed in Section 6.2.9.1),
were associated in a cross sectional study, with PMio among a population based sample of
children 10-18 years old in Iran (mean PMio concentration was 122.08 jug/m3).
Summary of Epidemiologic Study Findings for Systemic Inflammation
The most commonly measured marker of inflammation in the studies reviewed was
CRP. CRP was not consistently associated with short-term PM concentrations (PM2.5, PMio,
SO42-, EC, OC, PNC). Amulticity study of MI survivors in Europe (Ruckerl et al., 2007,
156931) failed to provide evidence of an effect of PM (e.g., PMio, PM2.5, PNC) on CRP and no
effect was observed by Diez-Roux et al. (2006, 156400) in a population based study when
concentrations were averaged over periods less than 30 days. Several other markers of
inflammation have been examined in relation to several PM size fractions and components
but the number of studies examining the same marker/PM metric combination is too few to
allow results to be compared across epidemiologic studies.
6.2.7.2. Controlled Human Exposure Studies
Several controlled human exposure studies were included in the 2004 PM AQCD
(U.S. EPA, 2004, 056905) which evaluated markers of systemic inflammation following
exposure to PM. Salvi et al. (1999, 058637) exposed 15 healthy volunteers (21-28 years old)
for 1 h to DE (300 (Jg/m3 particle concentration) and observed a significant increase in
neutrophils in peripheral blood 6 h post-exposure compared with filtered air control.
However, Ghio et al. (2003, 087363) reported no changes in plasma cytokine levels (e.g.,
July 2009
6-68
DRAFT-DO NOT CITE OR QUOTE

-------
IL-6 and TNF-a), WBC count, or CRP 0 or 24 h following a 2-h exposure to PM2.5 CAPs
(120 jjg/m3). Gong et al. (2003, 042106) did not observe any effect of PM2.5 CAPs (174 |jg/m3)
on serum amyloid A, while Frampton (2001, 019051) reported no change in leukocyte
activation following exposure to a low concentration (10 (Jg/m3) of UF carbon. The results of
studies published since the completion of the 2004 PM AQCD (U.S. EPA, 2004, 056905) are
discussed below.
CAPs
In a study of exposures to PM2.5 CAPs (200 )Jg/m3), Gong et al. (2004, 087964)
observed increased peripheral basophils 4 h following a 2-h exposure in a group of healthy
older adults. While this provides some evidence of a CAPs-induced systemic inflammatory
response, several other studies have reported no change in plasma CRP levels 0-24 h after
exposure to ultrafine (average concentration 50-100 |Jg/m3), fine (average concentration 190
¦ 200 |jg/m3), or coarse (average concentration 89 |jg/m3) CAPs (Gong et al., 2004, 055628;
Gong et al., 2008, 156483; Graff et al., 2009, 191981; Mills et al., 2008, 156766; Samet et al.,
2009, 191913).
Urban Traffic Particles
In a recent investigation of controlled exposures (24 h) to urban traffic particles,
Brauner et al. (2008, 191966) observed no effect of particulate matter concentration (PM2.5
concentration 9" 10 (Jg/m3)
on markers of inflammation including CRP, IL-6 and TNF- a in peripheral venous
blood.
Diesel Exhaust
Recent controlled human exposure studies have observed no effect of DE on plasma
CRP concentrations or peripheral blood cell counts (Blomberg et al., 2005, 191991; Carlsten
et al., 2007, 155714; Mills et al., 2005, 095757; Mills et al., 2007, 091206; Tornqvist et al.,
2007, 091279). Mills et al. (2005, 095757) found no effect of DE (300 |jg/m3) on serum IL-6 or
TNF-a among healthy adult volunteers 6 h after exposure. However, as reported by
Tornqvist et al. (2007, 091279), a significant increase in these cytokines was observed 24 h
after exposure. Although the physiological significance of this finding is unclear, this study
does provide evidence of a mild systemic inflammatory response induced by exposure to DE.
In an effort to better understand the inflammatory response of exposure to PM, Peretz et al.
July 2009
6-69
DRAFT-DO NOT CITE OR QUOTE

-------
(2007, 156853) conducted a pilot study in which gene expression in peripheral blood
mononuclear cells (PBMCs) of healthy human volunteers was evaluated following a 2-h
controlled exposure to DE (200 (Jg/m3 PM2.5). Adequate RNA samples for microarray
analysis (Affymetrix U133 Plus 2.0) from both pre- and 4 h post-exposure to filtered air and
DE were available in 4 of the 11 subjects enrolled. The authors found differential expression
of 10 genes involved in the inflammatory response when comparing DE exposure (8
upregulated, 2 downregulated) and exposure to filtered air. Two participants had paired
samples from 20 h post-exposure which were adequate for analysis. At this time point, DE
was associated with 4 differentially expressed genes (l upregulated, 3 downregulated).
However, this study is limited by a small sample size with limited statistical power.
Wood Smoke
Barregard et al. (2006, 091381) recently reported an increase in serum amyloid A at 0,
3, and 20 h following a 4-h exposure to WS (PM2.5 concentrations of 240-280 jjg/m3' among a
group of 13 healthy adults (20-56 years old).
Model Particles
Frampton et al. (2006, 088665) evaluated the effect of varying concentrations
(10-50 (Jg/m3) of UF elemental carbon on blood leukocyte expression of adhesion molecules
in healthy and asthmatic adults. Healthy subjects (n = 40) were exposed for 2 h to filtered
air and UF carbon under three separate protocols: 10 |jg/m3 at rest (n = 12), 10 and
25 (Jg/m3 with intermittent exercise (n = 12), and 50 (Jg/m3 with intermittent exercise
(n = 16). Asthmatics (n = 16) were exposed at a single concentration (10 (Jg/m3) for 2 h with
intermittent exercise. Leukocyte expression of surface markers were quantified using flow
cytometry on peripheral venous blood samples collected prior to and immediately following
exposure, as well as at 3.5 and 21 h post-exposure. Among healthy resting adults, UFP
exposure at a concentration of 10 (Jg/m3 had no effect on blood leukocytes. The expression of
adhesion molecules CD54 and CD18 on monocytes, and CD18 on PMNs was shown to
decrease with UFP exposure in healthy exercising adults. In exercising asthmatics,
expression of CDllb on monocytes and eosinophils, as well as CD54 on PMNs were reduced
following exposure to UFP. In both asthmatics and healthy adults, a UFP-induced decrease
in eosinophils and basophils was observed 0-21 h following exposure. Although the clinical
significance of these findings is unclear, the authors concluded that their findings of
UFP-induced changes in leukocyte distribution and expression were consistent with
July 2009
6-70
DRAFT-DO NOT CITE OR QUOTE

-------
increased retention of leukocytes in the pulmonary vasculature, which may be due to an
increase in pulmonary vasoconstriction. Other studies have reported no changes in plasma
cytokine levels, peripheral blood counts, or CRP following exposure to ZnO or ultrafine EC
(Beckett et al., 2005, 156261; Routledge et al., 2006, 088674).
Summary of Controlled Human Exposure Study Findings for Systemic Inflammation
New studies involving controlled exposures to various particle types have provided
limited and inconsistent evidence of a PM-induced increase in markers of systemic
inflammation.
6.2.7.3. Toxicological Studies
There has been limited evidence that enhanced hematopoiesis may occur in animals
exposed to PM. Two studies in the 2004 PM AQCD (U.S. EPA, 2004, 056905) provided
support for this effect, with one study measured stimulated release of PMNs from bone
marrow and another examined peripheral blood PMN and blood cell counts! however, one
study did not find associations between CAPs and peripheral blood counts. Thus, it was
concluded that consistent evidence of PM-induced hematopoiesis remained to be
demonstrated. However, in a study of humans exposed to biomass burning during the 1997
Southeast Asian smoke-haze episodes, PMio demonstrated the best relationship with blood
PMN band cell counts expressed as a percentage of total PMN at lag 0 and 1, indicating a
relatively quick response (Tan et al., 2000, 002304).
CAPs
A2-day CAPs study employing SH rats did not report increased WBC 18-20 h
post-exposure (Kodavanti et al., 2005, 087946). A study utilizing fine and/or ultrafine CAPs
demonstrated decreased WBC in SH rats 18 h after a 2-day (6 h/day) exposure (Kooter et
al., 2006, 097547). The decrease was largely attributable to lowered neutrophils in the fine
CAPs-exposed rats and reduced lymphocytes in the ultrafine+fine CAPs animals.
Model Particles
In a study of fine and ultrafine CB particles (WKY rats! 7 h! mean mass concentration
1400 and 1660 (Jg/m3 for fine and ultrafine CB, respectively! mean number concentration
3.8 x 103 and 5.2 x 104 particles/cm3, respectively), only ultrafine CB induced elevated blood
leukocytes at 0 and 48 h post-exposure compared to the control rats and no effect was
July 2009
6-71
DRAFT-DO NOT CITE OR QUOTE

-------
observed at 16 h (Gilmour et al., 2004, 054175). In another study of SH rats exposed to
ultrafine carbon particles for 24 h (mass concentration 172 (Jg/m3! mean number
concentration 9.0 x 106 particles/cm3), the percent neutrophils and lymphocytes were
increased on the first recovery day, but not the third day (Upadhyay et al., 2008, 159345);
CRP was unchanged. In another study, blood neutrophils were decreased in SH rats
exposed to ultrafine CB for 6 h and no effects were observed in old Fischer 344 rats (Elder
et al., 2004, 055642). Plasma IL-6 levels were unchanged (Elder et al., 2004, 055642).
Coal Fly Ash
Smith et al. (2006, 110864) examined the hematology parameters in SD rats following
a 3"day inhalation exposure (4 h/day) to coal fly ash (mean mass concentration 1400 (Jg/m3)
and reported increased blood neutrophils and reduced blood lymphocytes at 36 h but not
18 h post-exposure.
Intratracheal Instillation
Elevated systemic IL-6 and TNF-a cytokine levels were observed following PMio
instillation in mice (details provided in Section 6.2.8.3) (Mutlu et al., 2007, 121111). IL-6
was decreased with PM exposure in macrophage-depleted mice, indicating that some of the
IL-6 release originated in macrophages. For mice (male C57B1/6J) exposed to PM10-2.5
derived from coal fly ash (200 |Jg), increased plasma IL-6 levels were only observed in
animals that also received 100 |Jg of LPS (Finnerty et al., 2007, 156434) and this response
was not observed with LPS alone, indicating a role for PM10-2.5.
Overall, these studies provide evidence of time-dependent responses of systemic
inflammation induced by PM exposure. Alterations in WBC have been reported generally as
elevations immediately (0 h) or <36 h post-exposure and no change or reductions are noted
from 18-24 h.
6.2.8. Hemostasis, Thrombosis and Coagulation Factors
The 2004 PM AQCD (U.S. EPA, 2004, 056905) presented limited and inconsistent
evidence from epidemiologic, controlled human exposure, and toxicological studies of PM-
induced changes in blood coagulation markers. The body of scientific literature
investigating hemostatic effects of PM has grown significantly since the publication of the
2004 PM AQCD (U.S. EPA, 2004, 056905), with a limited number of epidemiologic studies
July 2009
6-72
DRAFT-DO NOT CITE OR QUOTE

-------
demonstrating consistent increases in von Willebrand factor (vWf) associated with PM and
less consistent associations with fibrinogen. Recent controlled human exposure and
toxicological studies have also observed changes in blood coagulation markers (e.g.,
fibrinogen, vWf, factor VII, t'PA) following exposure to PM. However, the findings of these
studies are somewhat inconsistent, which may be due in part to differences in the post-
exposure timing of the assessment.
6.2.8.1. Epidemiologic Studies
Several studies investigating the association of short-term fluctuations in PM
concentration with markers of coagulation (e.g., blood viscosity and fibrinogen) were
included in the 2004 PM AQCD (U.S. EPA, 2004, 056905). These preliminary studies
offered limited support for mechanistic explanations of the associations of PM concentration
with heart disease outcomes. New studies, published since 2002, are reviewed in this
section. Only vWF and fibrinogen were measured in enough comparable studies to allow
the consistency of findings to be evaluated across epidemiologic studies. Other markers of
coagulation studied included D-dimer, prothombin time, Factor VII/VIII and tPA.
Liao et al. (2005, 088677) used a cross-sectional study to examine the association
between short-term increases in air pollutant concentrations (mean PMio, NO2, CO, SO2,
and O3 over the previous 3 days) and several plasma hemostatic markers (fibrinogen, factor
VIII-C, vWF, albumin). Study subjects were middle aged participants in the ARIC
(Atherosclerosis Risk in Communities) study (n = 10,208), and were residents of
Washington County, MD, Forsyth County, NC, selected suburbs of Minneapolis, MN, or
Jackson, MS. The mean PM10 concentration during the study was 29.9 ug/ma. Each
12.8 ug/m3 increase in the mean PM10 concentration 1 day before the health measurements
were made was associated with a 3.93% increase in vWF (95% CP 0.40-7.46) among
diabetics, but not among non-diabetics (-0.54% [95% CP -1.68 to 0.60]). Each 12.8 |ug/m3
increase in the mean PM10 concentration 1 day before the health measurements were made
was also associated with a 0.006 g/dL decrease in serum albumin (95% CP -0.012 to 0.000)
among those with CVD, but not among those without CVD (0.029 g/dL increase [95%
CP -0.004 to 0.062]). The mean CO concentration on the previous day was also associated
with a significant decrease in serum albumin. The authors reported significant curvilinear
associatins between PM10 and factor VIII-C, which may indicate a threshold effect. Similar
curvilinear associations were observed between O3 with fibrinogen, and vWF, and SO2 with
July 2009
6-73
DRAFT-DO NOT CITE OR QUOTE

-------
factor VIII-C, WBC, and serum albumin (Liao et al., 2005, 088677). A significant association
with fibrinogen was not observed.
In the European multicity study described in Section 6.2.7.1, Ruckerl et al. (2007,
156931) found that each 13.5 ug/m3 increase in the mean PMio concentration over the
previous 5 days was associated with a 0.6% increase in the arithmetic mean fibrinogen level
(95% CL 0.1-l.l). Further these investigators found that promoter polymorphisms within
FGA and FGB modified the association of 5-day avg PMio concentration with plasma
fibrinogen levels (Peters et al., 2009, 191992). This association was 8-fold higher among
those homozygous for the minor allele genotype of FGB rs 1800790 compared with those
homozygous for the major allele.
Several smaller studies have been conducted in the U.S. and Canada. Delfino et al.
(2008, 156390) measured fibrinogen and D-dimer in blood of subjects who resided at two
downtown Los Angeles nursing homes. As described in Section 6.2.7.1, measurements were
made over a period of 12 wk and subjects were 65+ years old with a history of coronary
artery disease. These markers were not associated with the broad array PM metrics studied
(e.g., PMo.25, PMo.25-2.5, PMio-2.5, EC, OC, primary OC, BC). In the study of 92 Boston
residents with type 2 diabetes described previously, O'Neill et al. (2007, 091362) found that
increases in mean PM2.5 and BC concentration were associated with vWF concentrations for
all moving averages examined (1-6 days). Reidiker et al. (2004, 056992) reported that
in-vehicle PM2.5 was associated with increased vWF over the next 10-14 h among 9 police
troopers. Sullivan et al. (2007, 100083) did not observe associations with fibrinogen, or
D-dimer in individuals with or without COPD. Neither RBC, platelets nor blood viscosity
were associated with PM2.5 concentration in a panel study of 88 non-smoking elderly
subjects residing in the Salt Lake City, Ogden and Provo metropolitan area of Utah (Pope et
al., 2004, 055238). Although Zeka et al. (2006, 157177) did not observe an association with
CRP in the analysis of the NAS population in Boston (Section 6.3.7.1), increased fibrinogen
level was associated with increases in the the number of particles/cm3 over the previous
48 h and 1 wk, and an incremental increase in BC concentration over the previous 4 wk.
There were no consistent findings for lagged PM2.5 or sulfates (Zeka et al., 2006, 157177).
Several studies of coagulation markers were conducted outside the U.S. and Canada.
In a study of healthy individuals in Taiwan, adverse associations were observed for PM2.5,
PMio, nitrate, and SO r concentrations with fibrinogen and plasminogen activator
fibrinogen inhibitor-1 (PALI) (Chuang et al., 2007, 091063). In a large cross-sectional study
July 2009
6-74
DRAFT-DO NOT CITE OR QUOTE

-------
of healthy subjects in Tel-Aviv, Steinvil et al. (2008, 188893) examined fibrinogen collected
as part of routine health examinations for 3,659 individuals. No significant associations
were found between pollutant levels (lagged 1-7 days) and fibrinogen. Finally, Baccarelli
and colleagues reported adverse associations between PMio and prothrombin time among
normal subjects (Baccarelli et al., 2007, 090733).
Summary of Epidemiologic Study Findings for Hemostasis, Thrombosis and Coagulation
The most commonly measured markers of coagulation in the studies reviewed were
fibrinogen and vWF. Associations of PMio (Liao et al., 2005, 088677) and PM2.5 (O'Neill et
al., 2007, 091362; Riediker et al., 2004, 056992) with increased vWF were observed across
the limited number of studies examining this association among both diabetics and healthy
state troopers (Liao et al., 2005, 088677; Liao et al., 2007, 180272; Riediker et al., 2004,
056992). Results for fibrinogen were not consistent across epidemiologic studies. Positive
associations with fibrinogen were reported in older adults residing in Boston (Zeka et al.,
2006, 157177) and in the multicity European study of MI survivors. Liao et al. (2005,
088677) in a population based multicity study and Sullivan et al. (2007, 100083) did not
observe associations of PMio or PM2.5 with fibrinogen. Several other markers have been
examined (e.g., D-dimer, prothrombin time) but not in adequate numbers of studies to allow
comparisons across epidemiologic studies.
6.2.8.2. Controlled Human Exposure Studies
In two separate studies conducted by Ghio and colleagues, controlled exposures (2 h)
to fine CAPs (Chapel Hill, NC) at concentrations between 15 and 350 |jg/m3 were shown to
increase blood fibrinogen 18-24 h following exposure among healthy adults (Ghio et al.,
2000, 012140; Ghio et al., 2003, 087363). Increases in blood fibrinogen or factor VII would
suggest an increase in blood coagulability, which could result in an increased risk of
coronary thrombosis. However, a similar study conducted in Los Angeles observed a PM2.5
CAPs-induced decrease in factor VII blood levels in healthy subjects and found no
association between PM2.5 CAPs and blood fibrinogen among healthy and asthmatic
volunteers (Gong et al., 2003, 042106). Since the publication of the 2004 PM AQCD
(U.S. EPA, 2004, 056905), several new controlled human exposure studies have evaluated
the effects of PM on blood coagulation markers.
July 2009
6-75
DRAFT-DO NOT CITE OR QUOTE

-------
CAPs
Two studies of controlled human exposures to Los Angeles CAPs among older adults
with COPD (fine CAPs) and adults with and without asthma (ultrafine CAPs) reported no
significant association between exposure and blood coagulation markers at 0, 4, or 22 h
post-exposure (Gong et al., 2004, 087964; Gong et al., 2008, 156483). Graff et al. (2009,
191981) observed a decrease in the concentration of D-dimer of marginal statistical
significance in healthy adults (11.3% decrease per 10 (Jg/m3- p = 0.07) following exposure to
thoracic coarse CAPs (89 jjg/m3' At 20 h post-exposure, levels of tPAin plasma were shown
to decrease by 32.9% from baseline per 10 (Jg/m3 increase in CAPs concentration. No other
markers of hemostasis or thrombosis were affected by exposure to thoracic coarse CAPs.
However, in a similar study from the same laboratory, Samet et al. (2009, 191913) reported
a statistically significant increase in D-dimer immediately following, as well as 18 h after a
2-h exposure to UF CAPs (49.8 |Jg/m3; 120,662 particles/cm3) in a group of healthy adults
(18-35 years old). Plasma concentrations of PAI-1 were also reported to increase 18 h after
exposure to UF CAPs, although this increase was not statistically significant (p = 0.1). No
changes in fibrinogen, tPA, vWF, plasminogen, or factor VII were observed. The finding of
an increase in D-dimer following exposure to UF CAPs provides potentially important
information in elucidating the relationship between elevated concentrations of PM and
cardiovascular morbidity and mortality observed in epidemiologic studies. Whereas many
coagulation markers provide evidence of an increased potential to form clots (e.g., an
increase in fibrinogen or a decrease in tPA), D-dimer is a degradation product of a clot that
has formed.
Urban Traffic Particles
In a study of controlled 24-h exposures to urban traffic particles (average PM2.5
concentration 9.7 |jg/m3) among 29 healthy adults, Brauner et al. (2008, 191966) did not
observe any particle-induced change in plasma fibrinogen, factor VII, or platelet count after
6 or 24 h of exposure. Similarly, Larsson et al. (2007, 091375) observed no change in PAI-1
or fibrinogen in peripheral blood of healthy adult volunteers 14 h after a 2-h exposure to
road tunnel traffic with a fine particle concentration of 46-81 |Jg/m3.
Diesel Exhaust
Mills and colleagues have recently demonstrated a significant effect of DE (particle
concentration 300 (Jg/m3) on fibrinolytic function both in healthy men (n = 30) and in men
July 2009
6-76
DRAFT-DO NOT CITE OR QUOTE

-------
with coronary heart disease (n = 20) (Mills et al., 2005, 095757; Mills et al., 2007, 091206).
In both groups of volunteers, bradykinin-induced release of tPA was observed to decrease
6 h following exposure to DE compared to filtered air exposure. The same laboratory did not
observe an attenuation of tPA release 24 h after a 1-h exposure to DE (300 (Jg/m3) in a group
of health adults (Tornqvist et al., 2007, 091279), or observe any change in markers of
hemostasis or thrombosis 6 or 24 h following DE exposure at the same particle
concentration among a group of older adults with COPD (Blomberg et al., 2005, 191991).
Carlsten et al. (2007, 155711) conducted a similar study involving exposure of healthy
adults to DE with a fine particle concentration of 200 |Jg/m3. Although the authors observed
an increase in D-dimer, vWF, and platelet count 6 h following exposure to DE, these
increases did not reach statistical significance. In a subsequent study with a similar study
design, the same laboratory found no effect of a 2-h exposure to DE (100 and 200 (Jg/m3
PM2.5) on prothrombotic markers in a group (n = 16) of adults with metabolic syndrome
(Carlsten et al., 2008, 156323). The authors postulated that the lack of significant findings
could be due to a relatively small sample size. In addition, Carlsten et al. (2007,
155714;(2008, 156323) exposed subjects at rest while Mills et al. (2005, 095757) exposed
subjects to a higher concentration (300 (Jg/m3) with intermittent exercise. A more recent
study of DE which exposed healthy adults to a slightly higher particle concentration (330
(jg/m3) evaluated the effect of DE on thrombus formation using an ex vivo perfusion
chamber (Lucking et al., 2008, 191993). Thrombus formation, as well as in vivo platelet
activation, was observed to significantly increase 2 h following exposure to DE relative to
filtered air, thus providing some evidence of a potential physiological mechanism which
may explain in part the associations between PM and cardiovascular events observed in
epidemiologic studies.
Wood Smoke
Barregard et al. (2006, 091381) recently evaluated the effect of WS on markers of
coagulation, inflammation, and lipid peroxidation. Subjects (n = 13) were healthy males and
females (20-56 years old) and were exposed for 4 h to PM2.5 concentrations of 240-280 |Jg/m3.
The authors reported an increase in the ratio of factor VHI/von Willebrand factor, which is
an indicator of an increased risk of venous thromboembolism, at 0, 3, and 20 h following
exposure to WS.
July 2009
6-77
DRAFT-DO NOT CITE OR QUOTE

-------
Model Particles
Routledge et al. (2006, 088674) did not observe any changes in fibrinogen or D-dimer
following a 1-h exposure to ultrafine carbon among a group of resting healthy older adults
and older adults with stable angina. Similarly, Beckett et al. (2005, 156261) found no
changes in hemostatic markers (e.g., factor VII, fibrinogen, and vWF) following exposure to
ultrafine and fine ZnO (500 (Jg/m3).
Summary of Controlled Human Exposure Study Findings for Hemostasis, Thrombosis and
Coagulation
Taken together, these new studies have provided some additional evidence that
short-term exposure to PM at near ambient levels may have small, yet statistically
significant effects on hemostatic markers in healthy subjects or patients with coronary
artery disease.
6.2.8.3. Toxicological Studies
In general, the limited toxicological studies reviewed in the 2004 PM AQCD
(U.S. EPA, 2004, 056905) reported positive and negative findings for plasma fibrinogen
levels or other factors involved in the coagulation cascade. Rats exposed to New York City
CAPs did not have any exposure-related effects on any measured coagulation markers
(Nadziejko et al., 2002, 050587), whereas rats exposed to a high concentration of ROFA
demonstrated increased plasma fibrinogen (Kodavanti et al., 2002, 025236).
CAPs
APM2.5 CAPs exposure conducted for 2 days (4 h/day! mean mass concentration
144-2758 |Jg/m3; 8-10/2001! RTP, NC) in SH rats induced plasma fibrinogen increases
(measured 18-20 h post-exposure) in 5 of 7 separate studies (Kodavanti et al., 2005,
087946). Fibrinogen was not different from the air control group on the two days with the
highest CAPs concentrations (1129 and 2758 |Jg/m3), indicating that the response was likely
not attributable to mass.
In SH rats exposed to PM2.5 CAPs for 6 h in one of three locations in the Netherlands
(mean mass concentration range 270-2400, 335-3720, and 655-3660 |Jg/m3), plasma
fibrinogen was increased 48-h post-exposure when all CAP-exposed animals were combined
in the analysis (Cassee et al., 2005, 087962) (Cassee et al., 2005, 087962). In WKY rats
pre-exposed to O3 (8 h; 1600 |jg/m3) and CAPs for 6 h, increases in RBC, hemoglobin, and
July 2009
6-78
DRAFT-DO NOT CITE OR QUOTE

-------
hematocrit were observed 2 days after CAPs exposure. For SH rats exposed to CAPs only,
decreased mean corpuscular hemoglobin concentration were reported.
A similar study conducted by the same group (Kooter et al., 2006, 097547) reported no
changes in plasma fibrinogen measured 18 h after a 2-day exposure (6 h/day) to fine or
fine+ultrafine CAPs (mean mass concentration range 399.0-1067.5 and 269.0-555.8 |Jg/m3,
respectively! 1/2003-4/2004). However, elevated vWF was observed in SH rats exposed to
the highest concentration of fine CAPs. Decreases in mean corpuscular volume (MCV), and
elevations in mean platelet volume (MPV) and mean platelet component (MPC) were
reported in SH rats 18 h following a 2-day exposure to ultrafine+fine CAPs in a freeway
tunnel.
Traffic-Related Particles
Plasma fibrinogen levels were elevated 18 h following a single 6-h exposure to on-road
highway aerosols when groups of rats pretreated with saline or influenza virus were
combined (i.e., there was a significant effect of particles) (Elder et al., 2004, 087354).
Model Particles
The coagulation effects of inhaled ultrafine CB at a concentration of 150 (Jg/m3
(number count not provided) for 6 h were evaluated 24-h post-exposure in two aged rat
models (11-14 months SH and 23 months Fischer 344), some of which received LPS via
intraperitoneal injection prior to particle exposure (Elder et al., 2004, 055642). LPS has
been shown to induce the expression of molecules involved in coagulation, inflammation,
oxidative stress, and the acute-phase response. In those animals only exposed to CB, SH
rats demonstrated increased thrombin-anti-thrombin complexes (TAT) and decreased
fibrinogen. For F344 rats, TAT complexes and fibrinogen were elevated only in those that
received LPS and CB. Whole-blood viscosity was not altered in either rat strain with
particle exposure.
In another study of SH rats exposed to ultrafine carbon particles for 24 h (mass
concentration 172 (Jg/m3! mean number concentration 9.0 x 106 particles/cm3), the number
of RBC and platelets and hematocrit percent, were unchanged 1 and 3 days following
exposure (Upadhyay et al., 2008, 159345). Fibrinogen levels were similar in both air and
ultrafine carbon-exposed groups. However, mRNA expression of PAI-1 and TF in lung
homogenates (but not in heart) was increased on recovery day 3 after exposure. A study of
similar design that employed SH rats did not report any effect on plasma fibrinogen 4 or
July 2009
6-79
DRAFT-DO NOT CITE OR QUOTE

-------
24 h following ultrafine carbon exposure (mass concentration 180 |Jg/m3; mean number
concentration 1.6 x 107 particles/cm3) (Harder et al., 2005, 087371). Similarly, clotting factor
Vila and thrombomodulin, PAT1, and tPA mRNA expression were not affected by ultrafine
carbon exposure at 24-h post-exposure.
Coal Fly Ash
One study that employed coal fly ash (mean mass concentration 1,400 |Jg/m3; 4 h/day
x 3 days) demonstrated increases in hematocrit and MCV in SD rats at 36 h but not 16 h
post-exposure (Smith et al., 2006, 110864).
Intratracheal Instillation
Mutlu et al. (2007, 121111) used a PMio sample collected from Dusseldorf, Germany
in mice (C57BL/6) with and without the gene coding for IL-6. The authors report using a
moderate IT dose (10 |Jg/mouse; roughly equivalent to 400-500 jJg/kg); the PM sample had
previously been characterized as having significant Fe, Ni, and V content (Upadhyay et al.,
2003, 097370). In C57BL/6 mice, the Dusseldorf PM shortened bleeding (32%), prothrombin
(13%), and activated partial thromboplastin (16%) times and increased platelet count,
fibrinogen, and Factors II, VIII, and X activities 24 h following exposure. The authors
further demonstrated accelerated coagulation by a reduction in the left carotid artery
occlusion time (experimentally-derived by direct application of FeCh). Additional
experiments demonstrated that IL-6"'" or macrophage-depleted mice showed dramatically
attenuated effects of PMio on hemostatic indices, thrombin generation, and occlusion time.
In IL-6"'" mice, there was no change in total cell counts or differentials in BALF compared to
the wild-type mice, despite the lack of IL-6. In contrast, the model of macrophage depletion
had reduced levels of macrophages and IL-6 in BALF, following PM exposure. These studies
suggest that instillation of Dusseldorf PMio activates clotting through an alveolar
macrophage-dependent release of IL-6; however, other factors may also be involved in the
prothrombotic response (i.e., activation of neutrophils, other inflammatory cells, or
alterations in the levels of other cytokines).
In a study employing PM10-2.5 collected from six European locations with contrasting
traffic profiles, fibrinogen increases were observed in SH rats exposed to 10 mg/kg IT at
24-h post-exposure and similar responses were observed with PM2.5 (Gerlofs-Nijland et al.,
2007, 097840). PM10-2.5 and PM2.5 samples from Prague or Barcelona administered IT to SH
rats (7 mg/kg) resulted in elevated plasma fibrinogen levels 24-h post-exposure compared to
July 2009
6-80
DRAFT-DO NOT CITE OR QUOTE

-------
rats instilled with water (Gerlofs-Nijland et al., 2009, 190353). No changes were observed in
vWF for whole particle suspensions, but Barcelona PM10-2.5 organic extract induced greater
levels of vWF than Barcelona PM10-2.5.
Summary of Toxicological Study Findings for Hemostasis, Thrombosis and Coagulation
Increases in coagulation and thrombotic markers were observed in some studies of
rats or mice exposed to PM. Plasma TAT complexes were increased in CB-exposed SH rats
and shortened bleeding, prothrombin, and activated partial thromboplastin times were
observed in mice exposed IT to PM10. Furthermore, the latter study also reported increased
levels of Factors II, VIII, and X activities in mice. Another study demonstrated increased
vWF in response to PM2.5 CAPs. As for plasma fibrinogen, these studies provide some
evidence that increased levels are observed 18 h to 48 h post-exposure to PM, although one
study reported no change and another reported a decrease in this biomarker. Alterations in
platelet measurements have also been observed with PM exposure, including increased
platelet number, mean platelet volume, and mean platelet component. The toxicological
results of RBOrelated measurements are limited and inconsistent following PM exposure,
which may be attributable to different exposure protocols, time of analysis, or rat strain.
6.2.9. Systemic and Cardiovascular Oxidative Stress
Very little information on systemic oxidative stress associated with PM was available
for inclusion in the 2004 PM AQCD (U.S. EPA, 2004, 056905). However, recent
epidemiologic studies have provided consistent evidence of PM-induced increases in
markers of systemic oxidative stress including plasma TBARS, CuZn-SOD, 8-oxodG, and
total homocysteine. This is supported by a limited number of controlled human exposure
studies that observed PM-induced increases in free-radical mediated lipid peroxidation as
well as upregulation of the DNA repair gene hOGGl. In addition, recent toxicological
studies have demonstrated an increase in cardiovascular oxidative stress following PM
exposure in rats.
6.2.9.1. Epidemiologic Studies
No studies of markers of oxidative stress were reviewed in the 2004 PM AQCD
(U.S. EPA, 2004, 056905). Since 2002, numerous studies have examined whether short-
July 2009
6-81
DRAFT-DO NOT CITE OR QUOTE

-------
term increases in mean PM concentrations are associated with adverse changes in systemic
markers of oxidative stress.
In an analysis of the randomized trial of omega-3 fatty acid supplementation in
Mexico City nursing home residents described previously (Section 6.6.1.1), Romieu et al.
(2008, 156922) investigated the effect of this intervention on markers of systemic oxidative
stress (Cu/Zn SOD activity, LPO in plasma and GSH in plasma). A significant decrease of
Cu/Zn SOD was associated with a 10 |jg/m3 increase of PM2.5 in both groups (Fish oil:
6 = -0.17 [SE = 0.05], p = 0.002; Soy oil: 6 = -0.06 [SE = 0.02] p = < 0.001). A decrease in
GSH was associated with a 10 |jg/m3 increase in PM2.5 in the fish oil group (6 = -0.09
(SE = 0.04, p = 0.017).
Two studies evaluated plasma homocysteine levels in relation to PM. Baccarelli et al.
(2007, 091310) investigated fasting and postmethionine-load total homocysteine (tHcy)
among 1,213 normal subjects in Lombardia, Italy. Plasma homocysteine is a risk factor for
CVD and a marker for oxidative stress. Among smokers, average PM10 level during the 24-h
preceding the measurement was associated with 6.3% (95% CI: 1.3-11.6) and 4.9% (95% CI:
0.5-9.6) increases in fasting and postmethionine-load tHcy, respectively. No associations
were observed among non-smokers. Park et al. (2008, 156845) investigated the association
of BC, OC, S042- and PM2.5 with tHcy among 960 male participants of the NAS. Effect
modification by folate and vitamins B6 and B12 was also examined. BC and OC were
associated with increases in tHcy and associations were more pronounced in those with
lower plasma folate and vitamin B12.
In smaller studies with 25 to 50 healthy or diseased participants, several markers of
oxidative stress have been associated with PM size fractions or components. These
associations include thiobarbituric acid reactive substances (TBARS) with 24-h PMioCLmetai.,
2006.192002)? Cu/Zn-SOD with several PM metrics (e.g., ultrafine, coarse, EC, OC, BC and PN)
(Delfino et al., 2008, 156390): PM2.5, BC, vanadium and chromium with plasma proteins
(Sorensen et al., 2003, 157000): DNA damage assessed by 7-hydro-8-oxo-2-deoxyguanosine
(8-oxodG) in lymphocytes (Sorensen et al., 2003, 157000) and, 8-OHdG with sulfates
(Chuang et al., 2007, 091063). In addition, a cross-sectional study of children (10-18 yr) in
Iran showed an association of PM10 with oxidized LDL (oxLDL), malondialdehyde (MDA)
and conjugated diene (CDE) (Kelishadi et al., 2009, 191960).
July 2009
6-82
DRAFT-DO NOT CITE OR QUOTE

-------
Summary of Epidemiologic Study Findings for Systemic and Cardiovascular Oxidative
Stress
Oxidative stress responses measured by plasma tHcy, CuZn-SOD, TBARS, 8-oxodG,
oxLDL and MDAhave been consistently observed (Baccarelli et al., 2007, 091310; Chuang
et al., 2007, 091063; Delfino et al., 2008, 156390; Kelishadi et al., 2009, 191960; Liu et al.,
2007, 156705; Romieu et al., 2008, 156922; So re n sen et al., 2003, 157000). In addition, a
series of analyses examining the modification the PM-HRV association by genetic
polymorphisms related to oxidative stress has provided insight into the possible
mechanisms of cardiovascular diseases observed in association with PM concentrations (see
Section 6.2.1.1).
6.2.9.2. Controlled Human Exposure Studies
Urban Traffic Particles
Brauner et al. (2007, 188507) recently investigated the effect of urban traffic particles
on oxidative stress-induced damage to DNA. Healthy adults (20-40 years old) were exposed
to low concentrations of urban traffic particles as well as filtered air for periods of 24 h,
with and without two 90-min periods of exercise. Exposures took place in an exposure
chamber above a busy road with high traffic density in Copenhagen. Non-filtered air was
pumped into the chamber from above the street, with average PM2.5 and PM10-2.5 mass
concentrations of 9.7 |Jg/m3 and 12.6 |Jg/m3, respectively. The ultrafine/fine (6-700 nm)
particle number concentration was continuously monitored throughout the exposure
(average particle count 10,067/cm3). The PM2.5 fraction was rich in sulfur, vanadium,
chromium, iron, and copper. PBMCs were isolated from blood samples collected at 6 and 24-
h. DNA damage, as measured by strand breaks (SB) and formamidopyrimidine-DNA
glycosylase (FPG) sites, was evaluated using the Comet assay. The activity and mRNA
levels of the DNArepair enzyme 7,8"dihydro-8-oxoguanine-DNA glycosylase (OGGl) were
also measured. The authors observed increased levels of DNA strand breaks and FPG sites
following 6 and 24 h of exposure to PM. Using a mixed-effects regression model, the particle
concentration at the 57 nm mode was found to be the major contributor of these measures
of DNA damage. The results of this study suggest that short-term (6-24 h) exposure to
ambient levels of UFPs cause systemic oxidative stress resulting in damage to DNA.
July 2009
6-83
DRAFT-DO NOT CITE OR QUOTE

-------
Diesel Exhaust
Tornqvist et al. (2007, 091279) reported an increase in plasma antioxidant capacity in
a group of healthy volunteers 24 h after a 1-h exposure to DE with a particle concentration
of 300 |Jg/m3. The investigators suggested that systemic oxidative stress occurring following
exposure may have caused this up-regulation in antioxidant defense. Peretz et al. (2007,
156853) observed some significant differences in expression of genes involved in oxidative
stress pathways between exposure to DE (200 |jg/m3 PM2.5) and filtered air. However, the
conclusions of this investigation are limited by a small number of subjects (n = 4).
Wood Smoke
In a controlled human exposure study of controlled exposure to WS, Barregard et al.
(2006, 091381) found an increase in urinary excretion of free 8"iso-prostaglandin2a among
healthy adults (n = 9) approximately 20 h following a 4-h exposure to PM2.5 (mass
concentration of 240-280 (Jg/m3). This finding provides evidence of a PM-induced increase in
free-radical mediated lipid peroxidation. From the same study, Danielsen et al. (2008,
156382) reported an increase in the mRNA levels of the DNA repair gene hOGGl in
peripheral mononuclear cells 20 h after exposure to WS relative to filtered air.
Summary of Controlled Human Exposure Study Findings for Systemic and Cardiovascular
Oxidative Stress
Based on the results of these studies, it appears that exposure to PM at or near
ambient levels may increase systemic oxidative stress in human subjects.
6.2.9.3. Toxicological Studies
Very little information was available for inclusion in the 2004 PM AQCD (U.S. EPA,
2004, 056905) on oxidative stress in the cardiovascular system. A few new studies have
evaluated ROS in blood or the heart following PM exposure. Some studies have used
chemiluminescence (CL), which is measured using the decay of excited states of molecular
oxygen, and may also be prone to artifact.
CAPs
Gurgueira et al. (2002, 036535) measured oxidative stress in SD rats immediately
following a 5"h CAPs exposure (PM2.5 mean mass concentration 99.6-957.5 |Jg/m3; Boston,
MA; 7/2000 to 2/2001) and reported increased in situ CL in hearts of CAPs-exposed
July 2009
6-84
DRAFT-DO NOT CITE OR QUOTE

-------
animals. CL evaluated after 1- and .'Hi CAPs exposure did not demonstrate changes from
the filtered air group, although a Frh exposure resulted in increased CL in hearts. When
animals were allowed to recover for 24 h, oxidative stress returned to control values. To
compare potential particle-induced differences in CL, rats were exposed to ROFA (1.7 mg/m3
for 30 min) or CB (170 (Jg/m3 for 5 h) and only the ROFA-treated animals exhibited
increased CL in cardiac tissue. Additionally, levels of antioxidant enzymes in the heart
(Cu/Zn-SOD and MnSOD) were increased in CAPs-exposed rats.
Recently, Rhoden et al. (2005, 087878) tested the role of the ANS in driving
CAPs-induced cardiac oxidative stress in heart tissues of SD rats. At PM2.5 mass
concentrations of 700 |Jg/m3 (Boston, MA), pretreatment with an antioxidant, a fir receptor
antagonist, or a muscarinic receptor antagonist attenuated the CL and TBARS effects
observed in the heart following a 5-h PM2.5 exposure. The wet/dry ratio (edema) of cardiac
tissue also returned to control values in animals treated with the antioxidant prior to
CAPs. These combined results indicate involvement of both the sympathetic and
parasympathetic pathways in the cardiac oxidative stress response observed following PM
exposure.
More recently, a type of irritant receptor, the transient receptor potential vanilloid
receptor 1 (TRPVl), was identified as central to the inhaled CAPS-mediated induction of
cardiac tissue CL and TBARS in SD rats (Ghelfi et al., 2008, 156468). In these studies
(PM2.5 mean mass concentration 218 |Jg/m3; Boston, MA), capsazapine (a TRPVl inhibitor)
abrogated cardiac CL, TBARS, edema, and QT-interval shortening when measured at the
end of the 5-h exposure. These studies provide some evidence that the ANS may be involved
in producing cardiac oxidative stress following exposure to CAPs. Furthermore, this
response could be acting, at least in part, via TRPV receptors.
In WKY rats exposed to PM2.5 CAPs in Japan, relative mRNA expression of HOI was
increased in cardiac tissue and was also significantly correlated with the cumulative mass
of PM collected on chamber filters throughout the exposure (Ito et al., 2008, 096823).
Road Dust
A composite of PM2.5 road dust samples obtained from New York City, Los Angeles,
and Atlanta induced cardiac ROS as measured by CL in the low exposure group (306 |jg/m3)
and TBARS in the high exposure group (954 (Jg/m3); thus, the CL and TBARS methods
provided different results for the various source types (Seagrave et al., 2008, 191990).
July 2009
6-85
DRAFT-DO NOT CITE OR QUOTE

-------
Gasoline and Diesel Exhaust
Gasoline exhaust exposure also resulted in increased ROS (measured by TBARS) in
aortas of ApoE'" mice, as discussed in Section 6.2.4.3 (Lund et al., 2009, 180257). Similarly,
a 6-h exposure to gasoline exhaust (PM mass concentration 60 |Jg/m3, CMD 15-20 nm! MMD
150 nm! CO concentration 104 ppm, NO concentration 16.7 ppm, NO2 concentration
1.1 ppm, SO2 concentration 1.0 ppm) in SD rats demonstrated increased CL in the heart,
but no change in TBARS and the CL response was not duplicated when the particles were
filtered (Seagrave et al., 2008, 191990). Increased lipid peroxides in the serum of male SH
rats exposed to gasoline exhaust (PM mass concentration 59.1 |Jg/m3; NO concentration
18.4 ppm! NO2 concentration 0.9 ppm! CO concentration 107.3 ppm! SO2 concentration
0.62 ppm) was observed following a 1-wk exposure to gasoline exhaust and this effect was
attenuated with particle filtration (Reed et al., 2008, 156903). An IT instillation study of
diesel particles in mice demonstrated increased myocardial MPO activity 12 and 24 h post-
exposure to the residual particle component that remained after extraction with
dichloromethane (Yokota et al., 2008, 190109).
Model Particles
Other studies presented in earlier sections also demonstrated ROS (via CL) and NT
expression (via ELISA) in the left ventricle with CB exposure (Tankersley et al., 2008,
157043) and oxidative stress in the systemic microvasculature following Ti02 inhalation
exposure (Nurkiewicz et al., 2009, 191961) or ROFAIT exposure (Nurkiewicz et al., 2006,
088611). Decreased HO-1 mRNA expression in hearts of SH rats exposed to ultrafine carbon
particles was observed 3 days following exposure (Upadhyay et al., 2008, 159345) and there
was a trend toward increased HO-1 mRNA expression 1 day post-exposure.
Summary of Toxicological Study Findings for Systemic and Cardiovascular Oxidative
Stress
When considered together, the above studies provide evidence that PM exposure
results in oxidative stress as measured in cardiac tissue by CL, TBARS, HO-1 mRNA
expression, and NT expression. However, the PM concentration/dose and method of ROS
measurement could also affect the response. Cardiac oxidative stress may have resulted
from PM stimulation of the ANS, although these studies have only been conducted in one
laboratory. Multiple studies from two different laboratories provide support for vascular
July 2009
6-86
DRAFT-DO NOT CITE OR QUOTE

-------
oxidative stress as demonstrated in aortas following gasoline exhaust exposure and in the
microvasculature after Ti02 inhalation or ROFAIT exposure.
PM Components
Individual PM component concentrations were linked to CL levels in rat heart tissue
using separate univariate linear regression models, with total PM mass, Al, Si, Ti, and Fe
having p-values < 0.007 (Gurgueira et al., 2002, 036535). The highest R2 value in the
regression analyses was for Al (0.67) and its concentration ranged from 0.000 to
8.938 |Jg/m3.
6.2.10. Hospital Admissions and ED Visits
The 1996 PM AQCD (U.S. EPA, 1996, 079380) considered just two time-series studies
regarding the association between daily variations in PM levels and the risk of
cardiovascular disease (CVD) morbidity as measured by the number of daily
hospitalizations with primary discharge diagnoses related to CVD (Burnett et al., 1995,
077226; Schwartz and Morris, 1995, 046186). In contrast, the 2004 PM AQCD (U.S. EPA,
2004, 056905) reviewed more than 25 publications relating PM and risk of CVD
hospitalizations. Results from a handful of larger multicity studies were emphasized, with
the greatest emphasis placed on findings from the U.S. National Morbidity, Mortality, and
Air Pollution Study (NMMAPS) (Samet et al., 2000, 010269) and a subsequent reanalysis
(Zanobetti and Schwartz, 2003, 157171). The NMMAPS study evaluated the effect of daily
changes in ambient PM levels on total CVD hospitalizations among elderly Medicare
beneficiaries in 14 U.S. cities and found a ~1% excess risk per 10 jug/m3 increase in PMio.
The 2004 PM AQCD concluded that these results, along with those of the other single- and
multicity studies reviewed "generally appear to confirm likely excess risk of CVD-related
hospital admissions for U.S. cities in the range of [0.6 to 1.7% per 10 |ug/m3] PMio, especially
among the elderly" (U.S. EPA, 2004, 056905). The 2004 PM AQCD (U.S. EPA, 2004, 056905)
also concluded that there was some evidence from single-city studies suggesting an excess
risk specifically for hospitalizations related to ischemic heart disease and heart failure.
Furthermore, the 2004 PM AQCD (U.S. EPA, 2004, 056905) found that "insufficient data
exist from the time-series CVD admissions studies [...] to provide clear guidance as to
which ambient PM components, defined on the basis of size or composition, determine
ambient PM CVD effect potency" (U.S. EPA, 2004, 056905). The key studies reviewed in the
July 2009
6-87
DRAFT-DO NOT CITE OR QUOTE

-------
2004 PM AQCD (U.S. EPA, 2004, 056905) on this topic included those by Burnett and
colleagues (1997, 084194; 1999, 017269), Lippman and colleagues (2000, 024579), Ito (2003,
042856), and Peters et al. (2001, 016546).
Recent large studies conducted in the U.S., Europe, and Australia and New Zealand
have confirmed these findings for PMio, and have also observed consistent associations
between PM2.5 and cardiovascular hospitalizations. However, findings from single-city
studies have demonstrated regional heterogeneity in effect estimates. It is apparent from
these recent studies that the observed increases in cardiovascular hospitalizations are
largely due to admissions for ischemic heart disease and congestive heart failure rather
than CBVDs (such as stroke). The new literature on hospitalizations and ED visits for
cardiovascular causes published since 2002 is reviewed in the following sections. First, the
specific CVD outcomes captured using ICD codes from hospital admissions databases are
discussed. Second, the methods used in the large and multicity studies are described. For
each outcome considered, evidence from large/multicity studies is emphasized and results
from U.S. and Canadian single-city studies are also discussed. Although the single-city
studies may lack statistical power needed to evaluate interactions and detect some of the
subtle effects of air pollution, they inform the interpretation of the heterogeneous effect
estimates that have been observed across North America.
Cardiovascular Disease ICD Codes
When the 2004 PM AQCD (U.S. EPA, 2004, 056905) was written, few studies had
evaluated the link between ambient PM and specific CVD outcomes such as congestive
heart failure, ischemic heart disease or ischemic stroke. In contrast, the majority of recent
studies have focused on specific CVD outcomes. This trend is justified by the fact that the
short-term exposure effects of PM may be very different for different cardiovascular
outcomes. For example, given the current putative biological pathways involved in the
acute response to PM exposure, there is no a priori reason why short-term fluctuations in
PM levels would have similar effects on the risk of acute MI, chronic atherosclerosis of the
coronary arteries, and hemorrhagic stroke.
Almost all of the published time-series studies of cardiovascular hospitalizations and
ED visits identified cases based on administrative discharge diagnosis codes as defined by
the International Classification of Disease 9th revision (ICD-9) or 10th revision (ICD-10)
(NCHS, 2007, 157194). A complicating factor in interpreting the results of these studies is
July 2009
6-88
DRAFT-DO NOT CITE OR QUOTE

-------
the lack of consistency in both defining specific health outcomes and in the nomenclature
used.
Table 6-5. Description of ICD-9 and ICD-10 codes for diseases of the circulatory system.
Description
ICD-9 Codes
ICD-10 Codes
All Cardiovascular Disease
390-459
I00-I99
Ischemic Heart Disease
410-414
I20-I25
Acute Ml
410
121
Diseases Of Pulmonary Circulation
415-417
I26-I28
Heart Failure
428
150
Arrhythmia
427
I47,148,149
CBVD
430-438
I60-I69
Ischemic Stroke And Transient Ischemic Attack (TIA)
430-432
I63
Hemorrhagic Stroke
433-435
I60-I62
Peripheral Vascular Disease (PVD)
440-448
I70-I79
Table 6-5 shows major groups of diagnostic codes used in air pollution studies for
diseases of the circulatory system. The codes ICD-9: 390-459 are frequently used to identify
all CVD morbidity. Note that this definition of CVD includes diseases of the heart and
coronary circulation, CBVD, and peripheral vascular disease. In contrast, the term cardiac
disease specifically excludes diseases not involving the heart or coronary circulation. While
this distinction is conceptually straightforward, the implementation of the definition of
cardiac disease in terms of ICD-9 or ICD-10 codes varies among authors. Even greater
heterogeneity can be found among studies in the implementation of definitions related to
CBVD.
Design and Methods of Large and Multicity Hospital Admission and ED Visit Studies
Recently, multiple research groups in the U.S., Europe, and Australia have created
large datasets to evaluate specific CVD and respiratory endpoints using more detailed and
relevant measures of PM concentration. In the U.S., the MCAPS analyses of Dominici et al.
(2006, 088398), Bell et al. (2008, 156266) and Peng et al. (2008, 156850) are large,
comprehensive and informative studies based on Medicare hospitalization data. Likewise,
the Atlanta-based SOPHIA study (Metzger et al., 2004, 044222; Peel et al., 2005, 056305;
Tolbert et al., 2007, 090316) is the largest and most comprehensive study of U.S.
cardiovascular and respiratory ED visits. In Europe, the APHEA initiative (Le Tertre et al.,
July 2009
6-89
DRAFT-DO NOT CITE OR QUOTE

-------
2002, 023746; Le Tertre et al., 2003, 042820) the more recent HEAPSS study (Von Klot et
al., 2005, 088070), and the French PSAS program (Host et al., 2008, 155852; Larrieu et al.,
2007, 093031) are similarly noteworthy for their large sample size, geographic diversity,
and consideration of specific CVD and/or respiratory endpoints. These studies contain
adequate data to examine interactions by season and region! the effects of different size
fractions, components and sources of PM; or the effect of PM on susceptible subpopulations.
The following section provides a detailed review of the study design and methods used by
each of the large studies. A discussion of the results of each study can be found in later
sections of the ISA.
MCAPS: Medicare Air Pollution Study
Dominici et al. (2006, 088398) created a database of daily time-series of hospital
admission rates (1999-2002) for a range of cardiovascular and respiratory outcomes among
Medicare beneficiaries aged > 65 yr, ambient PM2.5 levels, and meteorological variables for
204 U.S. urban counties. The specific CVD outcomes considered were: CBVD (ICD-9:
430-438), peripheral vascular disease (440-448), ischemic heart diseases (410-414, 429),
heart rhythm disturbances (426, 427), and heart failure (428). Injuries (800-849) were
evaluated as a control outcome. Gaseous and other particulate pollutant size fractions were
not considered.
Data on PM2.5 were obtained from the AQS database of the U.S. EPA. Within each
county, associations between cause-specific hospitalization rates and same-day PM2.5 levels
were evaluated using Poisson regression models controlling for long-term temporal trends
and meteorologic conditions with natural cubic splines. County-specific results were
subsequently averaged using Bayesian hierarchical models. In addition to evaluating single
day lags, 3-day distributed lag models (lags 0, 1, and 2 days) were also considered in a
subset of 90 U.S. counties with daily PM2.5 data available during the study time period.
Subsequently, Peng et al. (2008, 156850) and Bell et al. (2008, 156266) extended the
database of daily time-series of hospital admissions, PM2.5, and other covariates for 202
U.S. counties through 2005. Importantly, Peng et al. (2008, 156850) added data on PM10-2.5
to this database for 108 U.S. counties with one or more co-located PM2.5 and PM10 monitors.
Analyses with PM10-2.5 were carried out using similar methods to those of Dominici et al.
(2006, 088398). Peng et al. (2008, 156850) evaluated the robustness of PM2.5 associations to
adjustment for thoracic coarse PM (Peng et al., 2008, 156850). Gaseous pollutants were not
considered in these analyses.
July 2009
6-90
DRAFT-DO NOT CITE OR QUOTE

-------
SOPHIA: Study of Particulates and Health in Atlanta
SOPHIA investigators (Metzger et al., 2004, 044222; Peel et al., 2005, 056305; Tolbert
et al., 2000, 010320) compiled data on 4,407,535 ED visits between 1993 and 2000 to 31
hospitals in the Atlanta metropolitan statistical area (20 counties). Specific cardiovascular
outcomes considered were: ischemic heart disease (ICD-9: 410-414), acute MI (410), cardiac
dysrhythmias (427), cardiac arrest (427.5), congestive heart failure (428), peripheral
vascular and CBVD (433-437, 440, 443-444, 451-453), atherosclerosis (440), and stroke
(436). Finger wounds (883.0) were evaluated as a control outcome.
The air quality data included measurements of criteria pollutants (PM and gaseous
pollutants) for the entire study period, as well as detailed measurements of mass
concentrations for the fine (PM2.5) and thoracic coarse fractions (PM10-2.5) of PM and several
physical and chemical characteristics of PM2.5 for the final 25 months of the study using
data from the ARIES monitoring station. Rates of ED visits for specific causes were
assessed in relation to the 3-day ma (lags 0-2 days) of daily measures of air pollutants using
Poisson generalized linear models controlling for long-term temporal trends and
meteorologic conditions with cubic splines. Tolbert et al. (2007, 090316) published interim
results of this study in relation to both cardiovascular and respiratory disease visits,
Metzger et al. (2004, 011222) published the main results for CVD visits, and Peel et al.
(2005, 056305) published the main results for respiratory conditions. An analysis of co-
morbid conditions that may make individuals more susceptible to PM-related
cardiovascular risk was carried out by Peel et al. (2007, 090112). Tolbert et al. (2007,
090316) extended the available data through 2002 and compared results from single and
multipollutant models while Sarnat et al. (2008, 097972) evaluated the risk of ED visits for
cardiovascular and respiratory diseases in relation to specific sources of ambient PM using
the extended dataset.
APHEA and APHEA-2: Air pollution and Health: a European Approach
APHEA-2 investigators compiled daily data on cardiovascular (Le Tertre et al., 2002,
023746; Le Tertre et al., 2003, 042820) and respiratory (Atkinson et al., 2001, 021959;
Atkinson et al., 2003, 042797) disease hospital admissions in the following 8 European
cities^ Barcelona, Birmingham, London, Milan, the Netherlands, Paris, Rome, and
Stockholm. (The publications on respiratory diseases were reviewed in the 2004 PM AQCD).
The specific CVD outcomes considered in each city were: cardiac diseases (ICD-9: 390-429),
ischemic heart disease (410-413) and CBVDs (430-438). Routine registers in all cities
July 2009
6-91
DRAFT-DO NOT CITE OR QUOTE

-------
provided daily data on hospitalizations. Only emergency hospitalizations were considered,
except in Milan, Paris, and Rome where only general admissions data were available.
Ambient PMio levels were available in all cities except Paris (PM13 used), and Milan
and Rome (TSP used). Data on gaseous pollutants (NO2, SO2, CO, and O3) were also
available in most cities. Five of the eight cities provided data on black smoke (BS). The
length of the available time-series varied by city but generally spanned from the early to
mid 1990s.
Within each city, associations between cause-specific hospitalization rates and
same-day PM2.5 levels were evaluated using Poisson GAMs controlling for long-term
temporal trends and meteorologic conditions. City-specific results were subsequently
averaged using standard meta-analytic methods. The original analyses (Atkinson et al.,
2001, 021959; Le Tertre et al., 2002, 023746) were carried out using GAMs and LOESS
smoothers. Following reports of problems associated with using the default convergence
criteria in the standard S-plus GAM procedure (Dominici et al., 2002, 030458), study
authors reanalyzed the data on cardiac admissions using GAMs and stricter convergence
criteria, and generalized linear models (GLMs) with natural splines and penalized splines
(Atkinson et al., 2003, 042797; Le Tertre et al., 2003, 042820). The authors found that the
results of the original analyses were insensitive to the choice of convergence criteria and
that the use of GLMs with penalized splines yielded very similar results.
HEAPSS: Health Effects of Air Pollution among Susceptible Subpopulations
HEAPSS investigators collected data on patients hospitalized for a first MI in five
European cities between 1992 and 2000. Patients were identified from MI registers in
Augsburg and Barcelona, and from hospital discharge registers in Helsinki, Rome and
Stockholm. Data on daily levels of PM10, were measured at central monitoring sites in each
city. Particle number concentration, a proxy for UFPs, was measured for a year in each city
and then modeled retrospectively for the whole study period. Associations of outcomes with
gaseous criteria pollutants were also evaluated.
Von Klot et al. (2005, 088070) identified 22,006 survivors of a first MI in the five
participating European cities and collected data on subsequent first cardiac
re-hospitalizations between 1992 and 2001. Readmissions of interest were those with
primary diagnoses of acute MI, angina pectoris, or cardiac disease (which additionally
includes dysrhythmias and heart failure). Within each city, associations between
cause-specific hospitalization rates and same-day levels of PM10 were evaluated using
July 2009
6-92
DRAFT-DO NOT CITE OR QUOTE

-------
Poisson GAMs controlling for long-term temporal trends and meteorologic conditions using
penalized splines. City-specific results were combined using standard meta-analytic
methods. Subsequently Lanki et al. (2006, 089788) used HEAPSS data from 26,854
patients to evaluate the association between daily PMio and particle number concentrations
and the risk of hospitalization for first MI.
PSAS: The French National Program on Air Pollution Health Effects
Larrieu et al. (2007, 093031) evaluated the association between PMio and the risk of
hospitalization in 8 French cities between 1998 and 2003. The cities examined were:
Bordeaux, Le Havre, Lille, Lyon, Marseille, Paris, Rouen and Toulouse. The specific CVD
outcomes considered in each city included: total CVD (ICD-10:100-199), cardiac disease
(100-152), ischemic heart diseases (120-125) and stroke (160-164, G45-G46). The available
data did not differentiate between emergency and non-emergency hospitalizations. Daily
mean PMio and NO2 levels as well as 8-h max O3 levels were obtained from a network of
monitors in each city.
Within each city, associations between cause-specific hospitalization rates and 2-day
ma (lag 0-1 days) levels of PMio were evaluated using Poisson GAMs controlling for
long-term temporal trends and meteorologic conditions using penalized splines.
City-specific results were combined using standard meta-analytic methods. Host et al.
(2008, 155852) used a subset of these data (6 cities, 2000-2003) to compare the effects of the
fine (PM2.5) and coarse fractions (PM10-2.5) of ambient particles on the risk of cardiovascular
and respiratory admissions. CVD outcomes assessed in this analysis were all CVD (ICD-10
100-199), cardiac (100-152) and IHD (120-125). PM2.5 levels were obtained from the same
network of background monitors described above. PM10-2.5 was calculated by subtracting
PM2.5 levels from PMio levels. Gaseous pollutants and hospital admissions for stroke were
not considered in this analysis.
Multicity Studies in Australia and New Zealand
Barnett et al. (2006, 089770) collected data on daily CVD emergency hospital
admissions and pollution data between 1998 and 2001 in five Australian cities (Brisbane,
Canberra, Melbourne, Perth, Sydney) and two cities in New Zealand (Auckland,
Christchurch). In 2001, these cities covered 53% of the Australian population and 44% of
the New Zealand population. The specific outcomes considered in each city were: all
circulatory diseases (ICD-9 390-429, ICD-10 100-199 with exclusions); heart failure (ICD-9
428, ICD-10 150); arrhythmia (ICD-9 427 ICD-10 146-49); cardiac disease (ICD-9 390-429,
July 2009
6-93
DRAFT-DO NOT CITE OR QUOTE

-------
ICD-10 100-152, 197.0, 197.1, 198.1); ischemic heart disease (ICD-9 410-413, ICD-10 120-24,
125.2); acute MI (ICD-9 410, ICD-10 121-22); and stroke (ICD-9 430-438, ICD-10 160-66, 167,
168, 169, G45-46 with exclusions).
Air pollutants considered were 24-h avg PMio, 24-h avg PM2.5, BSP and gaseous
pollutants. Within each city, associations between cause-specific hospitalization rates and
2-day ma (lags 0-1 days) of PM10 were evaluated using the time-stratified case-crossover
approach which controls for long-term and seasonal time trends by design rather than
analytically. City-specific results were combined using random effects meta-analytic
methods.
EMECAS: Spanish Multicentric Study on the Relation between Air Pollution and Health
Ballester et al. (2006, 088746) collected data on daily cardiovascular emergency
hospital admission and air pollution data between approximately 1995 and 1999 in 14 cities
in Spain. The specific outcomes considered in each city were: total CVD (ICD-9: 390-459)
and heart diseases (410-414, 427, 428). Air pollutants considered were PM10, TSP, BS, SO2,
NO2 (24-h avg), CO and O3 (8-h max).
Within each city, associations between cause-specific hospitalization rates and daily
levels of each pollutant metric were evaluated using Poisson GAMs with strict convergence
criteria. In all models, pollutants were entered as linear continuous variables and included
control for confounding by meteorological variables, influenza rates, long-term time trends,
and unusual events. The authors considered both distributed lag models (lags 0-3 days) and
the 2-day ma of pollution (lags 0-1 days). City-specific results were combined using
standard meta-analytic methods.
6.2.10.1. All Cardiovascular Disease
The 2004 PM AQCD (U.S. EPA, 2004, 056905) incorporated the results of a large
number of time-series studies in the U.S. and elsewhere relating ambient PM levels and
risk of hospitalization for CVD. The 2004 PM AQCD (U.S. EPA, 2004, 056905) noted that
the strongest evidence for this association came from the NMMAPS study (Samet et al.,
2000, 010269) and the subsequent reanalysis by Zanobetti and Schwartz (2003, 042812).
Since then, the U.S. MCAPS study evaluated the association between PM2.5 and risk
of CVD hospitalization in 202 U.S. counties between 1999 and 2005 and found a 0.7% (95%
posterior interval (PI): 0.5, 1.0) increase in risk per 10 jug/m3 increase in PM2.5 on the same
day (Peng et al., 2008, 156850). In 108 U.S. counties with co-located PM10 and PM10-2.5
July 2009
6-94
DRAFT-DO NOT CITE OR QUOTE

-------
monitors, the authors found a 0.4% (95% PI, 0.1 to 0.7, lag 0) increase in risk per 10 jug/m3
PMio-2.5 and no associations at lags of 1 and 2 days (Peng et al., 2008, 156850). In a
2-pollutant model adjusted for PM2.5, the association between PM10-2.5 and CVD
hospitalization lost precision (0.3% [95% PL -0.1 to 0.6, lag 0]). Bell et al. (2008, 156266)
found evidence of substantial and statistically significant variability in the effects of PM2.5
on cardiovascular hospitalizations by season and region, with the highest national average
estimates occurring in the winter and the highest regional estimates in the northeastern
U.S. (1.08% [95% PL 0.79-1.37, lag 0, per 10 |u,g/m3 increase in pm2.b]). Estimates for the
nation (1.49% [95% PL 1.09-1.89, lag 0]) and northeast (2.01% [95% PL 1.39-2.63, lag 0])
were highest in the winter.
Bell et al. (2009, 191997) and Peng et al. (2009, 191998) used data from the MCAPS
study and the EPA's Speciation Trends Network (STN) to identify the components of PM2.5
that are most strongly associated with hospitalizations for cardiovascular disease. Peng et
al. (2009, 191998) focused on the components that make up the majority of PM2.5 mass
(SO42_, NOa , Si, EC, OC, NA+ and NH44") and found that in multipollutant models only EC
and OC were significantly associated with risk of hospitalization for cardiovascular disease.
Bell et al. (2009, 191997) used data from 20 PM2.5 components and found that EC, Ni, and V
were most positively and significantly associated with the risk of cardiovascular
hospitalizations. These results suggest that the observed associations between PM2.5 and
cardiovascular disease hospitalizations may be primarily due to particles from oil
combustion and traffic.
Additional evidence is provided by several large multicity studies conducted outside of
the U.S. The European APHEA2 study (Le Tertre et al., 2002, 023746) looked at admissions
for CVD (defined as ICD-9 390-429) among those aged > 65 and found a 0.7% (95% CP
0.4-1.0, lag 0-1 day avg) increase in risk per 10 |ug/m3 PM10. The Spanish EMECAS study
(Ballester et al., 2006, 088746) looked at admissions for CVD (defined as ICD-9 390-459)
and found a 0.9% (95% CP 0.4-1.5, lag 0-1 day avg) increase in risk per 10 |ug/m3 PM10. The
French PSAS program looked at CVD hospitalizations (defined as ICD-10 100-199) among
the elderly and found a 1.1% (0.5, 1.7%) increase in risk with PM10 and a 1.9% (95% CP
0.9-3.0, lag 0-1 day avg) increase in risk with a 10 |ug/m3 increase in PM2.5 (Host et al., 2008,
155852; Larrieu et al., 2007, 093031). Non-significant increases in association with PM10-2.5
were reported (1.0% [95% CP -1.0 to 3.0]) (Host et al., 2008, 155852). In multiple cities
across New Zealand and Australia, Barnett et al. (2006, 089770) looked among the elderly
July 2009
6-95
DRAFT-DO NOT CITE OR QUOTE

-------
(CVD defined as ICD-9 390-459 and found a 1.3% (95% CI: 0.6-2.0, lag 0-1 day avg) increase
in risk per 10 jug/m3 increase in PM2.5.
The Atlanta-based SOPHIA study found a 0.9% (95% CI: -0.2 to 1.9, lag 0-2 day avg)
and a 3.3% (95% CI: 1.0-5.6, lag 0-2 day avg) increase in risk with a 10 |ug/m3 increase in
PM10 and PM2.5, respectively (Metzger et al., 2004, 044222). In a more recent analysis from
this study with an additional 4 yr of data, ED visits for CVD were not significantly
associated with PM10 or PM2.5, but were significantly associated with total carbon (1.6%
[95% CP 0.5-2.6, per IQR increase]), EC (1.5% [95% CI: 0.5-2.5, per IQR increase]) and
organic carbon (1.5% [95% CI: 0.5-2.6, per IQR increase]) components of PM2.5 (2007,
090316). A weak non-significant association PM10-2.5 was observed in these data (Tolbert et
al., 2007, 090316) More recently, Sarnat et al. (2008, 097972) used multiple
source-apportionment methods to evaluate the association between all CVD ED visits and
specific PM2.5 sources and found consistent positive associations with sources related to
motor vehicles and biomass combustion. These results were insensitive to the
source-apportionment technique used. It is noteworthy that other traffic-related gaseous
pollutants were associated with CVD ED visits in the SOPHIA study (Metzger et al., 2004,
044222).
Using meta-regression techniques and the reported association between PM10 and
CVD hospitalizations from the 14 cities included in the NMMAPS analysis, Janssen et al.
(2002, 016743) examined whether the between-city variability in relative risk estimates
were related to the local contribution of a number of PM sources. The authors found that in
multivariate analyses PM10 coefficients increased significantly with increasing percentage
of PM10 emissions from highway vehicles/diesels and oil combustion.
A small number of additional single-city studies have been published showing positive
associations between hospital admissions and ambient PM in Copenhagen, Denmark
(Andersen et al., 2007, 093201), and weak nonsignificant associations in Spokane, WA
(Schreuder et al., 2006, 097959; Slaughter et al., 2005, 073854) and two small counties in
Idaho (Ulirsch et al., 2007, 091332). Schreuder et al. (2006, 097959) performed a source
apportionment analysis using seven yr of daily speciation data from the same residential
monitor in Spokane, WA used by Slaughter et al. (2005, 073854). These authors related
daily levels of four sources (woodsmoke, an As-rich source, motor vehicle emissions, and
airborne soil) to the excess risk of cardiovascular ED visits. During the heating season the
only notable association for CVD-related ED visits was with WS, while in the non-heating
July 2009
6-96
DRAFT-DO NOT CITE OR QUOTE

-------
season the only notable association was with airborne soil. While neither of these
associations reached statistical significance, the study likely lacked the statistical power to
find effects of the expected magnitude. In fact, it is doubtful that studies conducted outside
of large metropolitan areas have sufficient statistical power to detect associations of the
expected magnitude. Delfino et al. (2009, 191994) evaluated the effects of the 2003
California wildfires and observed a slightly larger excess risk of total cardiovascular disease
admissions during the wildfire period compared to the period prior to the wildfire although
excess risk estimates were generally weak and non-significant.
Studies in several cities in Australia have investigated the association of
cardiovascular disease admissions with PM concentration and sources. A study from
Sydney, Australia found a 0.3% (95% CP -0.8 to 1.4) and 1.8% (95% CP 0.4-3.2) excess risk
per 10 |ug/m3 increase in the 2-day ma (lags 0-1 days) in PMio and PM2.5, respectively
(Jalaludin et al., 2006, 189416). Johnston et al. (2007, 155882) and Hanigan et al. (2008,
156518) studied the association between PM10 and cardiovascular and respiratory
hospitalizations in Darwin, Australia, where the predominant source of PM is from biomass
combustion. The authors found little or no evidence of an association between PM10 and
cardiovascular disease hospital admissions in the general population.
Crustal material has also been investigated in an effort to explain associations of PM
concentration with cardiovascular disease admissions. Studies of a dust storm in the Gobi
desert that transported PM across the Pacific Ocean reaching the western U.S. in the
spring of 1998 have been conducted. An analysis of the health impacts of this event on the
population of British Columbia's (Canada) Lower Fraser Valley found no excess risk of
cardiac or respiratory hospital admissions despite hourly PM10 levels >100 ug/ma (Bennett
et al., 2006, 088061). On the other hand, a number of studies in Asia and eastern Europe
have identified some adverse health effects associated with dust storm events. Middleton
et al. (Middleton et al., 2008, 156760) found that dust storms in Cyprus were associated
with a 4.7% (95% CP 0.7-9.0) and 10.4% (95% CP -4.7 to 27.9) increase in risk of
hospitalization for all causes and cardiovascular diseases, respectively. Chan et al. (2008,
093297) studied the effects of Asian dust storms on cardiovascular hospital admissions in
Taipei, Taiwan and also found significant adverse effects during 39 Asian dust events with
high PM10 levels (daily PM10 >90 |ug/m3). Bell et al. (2008, 091268) analyzed these data
independently and concluded that Asian dust storms were positively associated with risk of
hospitalization for ischemic heart disease.
July 2009
6-97
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Location
Lag
Bell etal. (2008,156266)
202 US Counties
0

202 US Counties
0

202 US Counties
0

202 US Counties
0
Host et al. (2008,155852)
6 French Cities
0-1
Barnett et al. (2006, 089770)
7 Cities Aus/NZ
0-1
Metzqer et al. (2004, 044222)
Atlanta
0-2
Tolbert etal. (2007,090316)
Atlanta
0-2
Slauqhteret al. (2005, 073854)
Spokane
1
Delfino et al. (2009,191994)
CA
0-1
Excess Risk Estimate
65+ Overall ¦
Penq et al. (2008,156850)
108 US Counties
0
_^65+
PMio-

108 US Counties
1 h
«- 65+


108 US Counties
2 —
-65+

Host etal. (2008,155852)
6 Cities France
0-1 	
	•	 65+

Tolbert etal. (2007,090316)
Atlanta
0-2 	
—*	All Ages

Zanobetti & Schwartz (2003, 043119)
10 US Cities
0-1
~ 65+
PM10
Metzqer et al. (2004, 044222)
Atlanta
0-2 -i
	•	All ages

Tolbert etal. (2007,090316)
Atlanta
0-2
—¦	All Ages

Ballester et al. (2006, 088746)
14 Cities Spain
0-1
—.— 65+

Ulirsch et al. (2007, 091332)
Slauqhteret al. (2005, 073854)

n .
501

Spokane
1
	All ages

Le Tertre et al. (2002, 023746)
8 Cities Europe
0-1
65+

65+
-	All Ages
-	All Ages
-	All Ages, Wildfires
PM2.5
65+ NE
•— 65+, Winter
—•— 65+NE Winter
—«	 65+
All Ages
-4
-2
Figure 6-1.
Excess risk estimates per 10//g/m3 increase in PMio, PM2.5 and PM10 2.5 for studies of
CVD ED visits* and hospitalizations. Studies represented in the figure include all
multicity studies. Single-city studies conducted in the U.S. or Canada are also included.
July 2009
6-98
DRAFT-DO NOT CITE OR QUOTE

-------
Table 6-6. Characterization of ambient PM concentrations in studies of hospital admission and ED
visits for cardiovascular diseases.
Pollutant Author Location Mean Concentration Upper Percentile
	(|jg/m3)	Concentration (Mg/m3)
PMw

Balelster et al. (2006, 088746)
14 cities in Spain
NR
NR

Barnett et al. (2006, 089770)
5 cities in Australia and
New Zealand
NR
NR

Burnett et al. (1999, 017269)
Toronto, Canada
30.2


Ito et al. (2003, 042856); Lippman (2000, 024579)
Detroit, Ml
31


Jalaludin et al. (2006,189416)
Sydney, Australia
16.8
75th: 19.9
Max: 103.9

Larrieu et al. (Larrieu et al., 2007, 093031)
8 cities in France



Le Tertre et al. (2002, 023746)
8 cities in Europe
Ranqe: 15.5-55.7
Ranqe 75th: 19.9-66

Linn et al. (2000, 002839)
Los Anqeles, California
45
78 (summer) -132 (fall)

Metzqer et al. (2004, 044222)
Atlanta, GA
26.3
90th: 44.7

Morris et al. (1998, 024924)
Chicago, Illinois
41
75th: 51
Max: 117

Peters et al. (2001, 016546)
Boston, MA



Schwartz et al. (1995, 046186)
Detroit, Ml
48
90th: 82

Slauqhteret al. (2005, 073854)
Spokane, WA
NR
90th: 41.9

Tolbert et al. (2007,090316)
Atlanta, GA



Ulirsch et al. (2007, 091332)*
2 cities in southeast Idaho
24.2123.2
90th: 40.7137.4

Wellenius et al. (2005, 087483)
Pittsburqh, PA
31.1
95th: 70.5

Wellenius et al. (2005, 087483)
9 cities in the U.S.
28.4 (median)
90th: 57.9

Wellenius et al. (2005, 087483)
7 cities in the U.S.
28.3 (median)
90th: 57

Zanobetti and Schwartz (2005, 088069)
Boston, MA
28.4 (median)
90th: 53.6
PM2.5

Barnett et al. (2006, 089770)
7 cities in Australia



Bell et al. (2008,156266)
202 counties in the U.S.
12.92
34.16

Burnett et al. (1999, 017269)
Toronto Canada
18


Dominici et al. (2006, 088398)
204 counties in the U.S.
NR
NR

Delfino et al. (2009,191994)
6 counties CA
18.4-32.7
45.3-76.1 (wildfire period)

Host et al. (2008,155852)
6 cities in France
13.8-18.8
NR

Ito et al. (2003, 042856); Lippman (2000, 024579)
Detroit, Ml
18
98th: 55.2

Lisabeth et al. (2008,155939)

7
75th: 10

Metzper et al. (2004, 044222)
Atlanta, GA
17.8
90th: 32.3
98th: 39.8

Pope et al. (2006, 091246)
Wasatch Front, Utah



Slauqhteret al. (2005, 073854)
Spokane, WA
NR
90th: 20.2

Sullivan et al. (2005, 050854)
Kinq County, WA
12.8
90th 27.3, Max: 147

Symons et al. (2006, 091258)
Baltimore, MD
16
Max: 69.2

Tolbert et al. (2007,090316)
Atlanta, GA
17.1
98th: 38.7

Villenueve et al. (2006, 090191)
Edmonton, Canada
8.5
75th: 11

Zanobetti and Schwartz (2005, 088069)
Boston, MA
11.1 (median)
95th: 26.31
98th: 55.2
PM 18-2.5

Burnett et al. (1999, 017269)
Toronto, Canada
12.2
Max: 68

Host et al. (2008,155852)
6 cities in France
7-11
NR

Ito et al. (2003, 042856); Lippman (2000, 024579)
Detroit, Ml
13
Max: 50

Le Tertre et al. (2002, 023746)
8 cities in Europe
NR
NR

Metzper et al. (2004, 044222)
Atlanta, GA
9.1
90th: 16.2

Penpet al. (2008,156850)
204 cities in the U.S.
9.8 (Median)
NR

Peters et al. (2001, 016546)
Boston, MA
7.4


Slauqhteret al. (2005, 073854)
Spokane, WA
NR
NR

Tolbert et al. (2007,090316)
Atlanta, GA
9

"Results presented separately for 2 separate time series
The effect estimates from multicity studies and single-city studies conducted in the
U.S. and Canada are included in Figure 6-l. Information on PM concentrations during the
relevant study period is presented in Table 6-6. In summary, large studies from the U.S.,
Europe, and Australia/New Zealand published since the 2004 PM AQCD (U.S. EPA, 2004,
056905) provide support for an association between short-term increases in ambient levels
July 2009
6-99
DRAFT-DO NOT CITE OR QUOTE

-------
of PMio and PM2.5 and increased risk of hospitalization for total CVD. The evidence for an
association of CVD hospitalization with PM10-2.5 is relatively limited. Peng et al. (2008,
156850) reported that their PM10-2.5 estimate was not robust to adjustment for PM2.5 and
estimates from the other studies are imprecise. The average excess risk among the U.S.
elderly is likely in the range of 0.5 to 1.0% per 10 ug/ma increase in PM2.5, although
substantial variability by region of the country and season has been demonstrated. An
excess risk of ED visits for CVD of a similar magnitude appears likely. The excess risk of
CVD hospitalization maybe somewhat greater in Europe and Australia/New Zealand than
in the U.S. Sources including wood burning, oil burning, traffic and crustal material have
been associated with increases in cardiovascular hospitalization or ED visits, but the best
evidence suggests that in the U.S., oil combustion, wood burning, and traffic are likely the
sources of PM2.5 most strongly associated with cardiovascular hospitalizations or ED visits.
6.2.10.2. Cardiac Diseases
Cardiac disease represents a subset of CVD which specifically excludes
hospitalizations for CBVD, peripheral vascular disease, and other circulatory diseases not
involving the heart or coronary circulation. Only a small number of studies published since
the 2004 PM AQCD (U.S. EPA, 2004, 056905) have evaluated the association between
ambient PM and hospitalizations for cardiac diseases, as most investigators have focused
instead on more narrowly defined outcomes.
The French PSAS program found a 1.5% (95% CP 0.5-2.2, lag 0-1) and 2.4% (95% CP
1.2-3.7, lag 0-1) excess risk among the elderly per 10 |u.g/m3 increase in PM10 and PM2.5,
respectively (Host et al., 2007, 155851; Larrieu et al., 2007, 093031). Host et al. (2008,
155852) also found a positive less precise association with PM10-2.5, (excess relative risk per
10 jug/m3: 1.6% [95% CP -0.8 to 4.1%]). The European HEAPSS study looked at cardiac
readmissions among survivors of a first MI and found a 2.1% (95% CP 0.4-3.9, lag 0) excess
risk per 10 |ug/m3 increase in PM10 (Von Klot et al., 2005, 088070). A 1.9% (95% CP 1.0-2.7,
lag 0-1) excess risk per 10 |ug/m3 increase in PM2.5 was observed in several cities in
Australia and New Zealand (Barnett et al., 2006, 089770). Single-city studies of hospital
admissions from Kaohsiung and Taipei, Taiwan, and an ED visit study from Sydney,
Australia also reported statistically significant positive associations (Chang et al., 2005,
088662; Jalaludin et al., 2006, 189416; Yang et al., 2004, 055603). On the other hand,
July 2009
6-100
DRAFT-DO NOT CITE OR QUOTE

-------
Slaughter et al. (2005, 073854) found no association between either PMio or PM2.5 and risk
of cardiac hospitalization in Spokane, Washington.
In summary, although relatively few studies have focused on all cardiac diseases,
large studies from Europe and Australia/New Zealand published since the 2004 PM AQCD
(U.S. EPA, 2004, 056905) report positive associations between short-term increases in
ambient levels of PM10, PM2.5 and PM10-2.5 and increased risk of hospitalization for cardiac
disease. The results from small single-city studies are less consistent. The excess risk for
cardiac hospitalizations may be somewhat larger than for total CVD hospitalizations.
6.2.10.3. Ischemic Heart Disease
IHD represents a subset of all cardiac disease hospitalizations and typically includes
acute MI (ICD 9: 410), other acute and subacute forms of IHD (411), old MI (412), angina
pectoris (413), and other forms of chronic ischemic heart disease (414). Some authors term
this category coronary heart disease. Published studies evaluating IHD as a single outcome
are considered first, followed by consideration of studies looking at acute MI, a specific form
of IHD.
In 1995 Schwartz and Morris (1995, 046186) reported a 0.6% (95% CP 0.2-1.0%)
excess risk of hospitalization for IHD per 10 jug/m3 increase in mean PM10 levels over the
previous two days amo elderly Medicare beneficiaries living in Detroit between 1986 and
1989. As reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905), similar associations
were subsequently observed in many single-city studies including: London, England
(Atkinson et al., 1999, 007882), Toronto, Canada (Burnett et al., 1999, 017269), and Seoul,
Korea (Lee et al., 2003, 095552). Studies in Hong Kong (Wong et al., 1999, 009172; Wong et
al., 2002, 023232), Birmingham, England (Anderson et al., 2001, 017033), and London,
England (Wong et al., 2002, 025436) yielded positive point estimates of a similar
magnitude, but did not reach statistical significance.
The positive associations between short-term changes in PM and IHD hospitalizations
observed in the early single-city studies have been confirmed in several large multicity
studies. The U.S. MCAPS study (Dominici et al., 2006, 088398) found a 0.4% (95% CP
0.0-0.8) excess risk of hospitalization for IHD per 10 ug/m3 increase in PM2.5 two days
earlier. The European APHEA-2 study (Le Tertre et al., 2002, 023746) considered PM10 and
found a 0.8% (95% CP 0.3-1.2, lag 0-1) excess risk among those aged > 65 yr. Among the
elderly in 5 cities in Australia and New Zealand (Barnett et al., 2006, 089770) there was a
July 2009
6-101
DRAFT-DO NOT CITE OR QUOTE

-------
4.3% (95% CI: 1.9-6.4, lag O-l) excess risk per 10 jug/m3 increase in PM2.5. Among the elderly
in several French cities there was a 4.5% (95% CP 2.3-6.8, lag O-l), 6.4% (95% CI:
1.6-11.4%, lag 0"l) and 2.9% (95% CP 1.5-4.3, lag O-l) excess risk per 10 jug/m3 increase in
PM2.5, PM10-2.5, (Host et al., 2008, 155852) and PM10, respectively (Larrieu et al., 2007,
093031).
With regard to ED visits, the Atlanta-based SOPHIA study (Metzger et al., 2004,
011222) found positive associations with PM10 and PM2.5 (ranging from 1.1 to 2.3%), but the
effect estimates did not reach statistical significance. Similarly, associations of EC and OC
with IHD were increased but not significant. In 6 cities across Canada, Szyszkowicz (2009,
191996) observed a 1.4% (95% CP 0.7-2.0%) and 2.4% (95% CP 1.2-3.6%) excess risk of ED
visits for angina per 10 |ug/m3 increase in same-day PM10 and PM2.5, respectively. Although
excess risks were generally weak and non-significant, Delfino et al. (2009, 191994) observed
a slightly larger excess risk of IHD during wildfires compared to the pre-wildfire period. In
Sydney, Australia, Jalaludin et al. (2006, 189416) found a 0.8% (95% CP -1.2 to 2.8%) and
2.6% (95% CP 0.1-5.2) excess risk of ED visits for IHD per 10 jug/m3 increase in 2-day ma of
PM10 and PM2.5, respectively. A recent study in Helsinki, Finland found no evidence of an
association with UFP, ACP, PM2.5, PM10-2.5, or source-specific PM2.5 (Halonen et al., 2009,
180379).
July 2009
6-102
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Location
Lao

Excess Risk Estimate

ISCHEMIC HEART DISEASE
Ito (2003, 042856)
Detroit
1

	•	 65 +
PM2.5

Wasatch Front
0

	*	 All Ages

Host (2007,155851)
France (6 Cities)
0-1

—«	 All Ages

Metzqer et al. (2004, 044222)
Atlanta
0-3

	*	 All Ages

Barnett (2006, 089770)
Au/NZ 5 Cities
0-1

	.	 15-64

Dominici et al. (2006, 088398)
204 US Counties
0

• 65 +


204 US Counties
1

• 65 +


204 US Counties
2

* 65 +


204 US Counties
0-2 DL

65 +

Barnett (2006, 089770)
Au/NZ 5 Cities
0-1

	»	 65 +

Host (2007,155851)
France (6 Cities)
0-1

	»	 65 +

Burnett (1999,017269)
Toronto
0,1

—*— All Ages

~elfino et al. (2009,191994)
CA wildfires
0,1

—*— All ages

Ito (2003, 042856)
Detroit
1

	.	 65 +
PMio 2.5
Metzqer et al. (2004, 044222)
Atlanta
0-3
All Ages
;	»	

Host (2007,155851)
France 6 Cities
0-1

All Ages	•	


France 6 Cities
0-1

			 65 +

Burnett et al. (1999, 017269)
Toronto
0

—All Ages

Ito (2003, 042856)
Detroit
1

—«— 65 +
PM10
Le Tertre et al. (2002, 023746)
Europe 8 Cities
0-1

•- <65

Metzqer et al. (2004, 044222)
Atlanta
0-2

—•—All Ages

Larrieu (2007, 093031)
France 8 Cities
0-1

» All Ages

Burnett (1999,017269)
Toronto
0-1

All Ages

Le Tertre et al. (2002, 023746)
Europe 8 Cities
0-1

* 65 +

Jalaludin (2006,189416)
Sydney
0-1

—•— 65 +

Larrieu (2007, 093031)
France 8 Cities
0-1

65 +

Myocardial Infarction
Peters et al. (2001, 016546)
Boston
2 h

mean age 61.6 ¦
	*

Boston
24 h

mean age 61.6
	•—£
Sullivan (2005, 050854)
King County WA
1 h

21-98 yrs	¦	


King County WA
2 h

21-98 yrs	•	


King County WA
4 h

21-98 yrs	•—


King County WA
24 h

21-98 yrs	•	

Zanobetti & Schwartz (2006, 090195)
Boston
0

65+	«	






PMio
Peters et al. (2001, 016546)
Boston
2 h
avg age 61.6
•s	¦	


Boston
24 h
avg age 61.6
*	

Linn et al. (2000, 002839)
Los Angeles
0

*>30 yrs
PM10
Peters et al. (2001, 016546)
Boston
2 h

avg age 61.6 ¦	


Boston
24 h

avg age 61.6 *
	*
Zanobetti & Schwartz (2005, 088069)
21 US Cities
0

•* 65 +

-5 -2 2 4 6 8 12 16 20 24 28
Figure 6-2. Excess risk estimates per 10 yug/m3 increase in PMio, PM2.5, PM10 2.5 for studies of
EDvisits * and hospitalizations for IHD and Ml. Studies represented in the figure inclue
all multicity studies. Single-city studies conducted in the U.S. and Canada are also
included.
To explore this link further, Pope et al. (2006, 091246) used data from an ongoing
registry of patients undergoing coronary angiography at a single referral center in Salt
July 2009
6-103
DRAFT-DO NOT CITE OR QUOTE

-------
Lake City, UT, between 1994 and 2004. The authors found a 4.8% (95% CI: 1.0-8.8, lag 0)
excess risk of acute MI or unstable angina per 10 jug/m3 increase in PM2.5 among 4,818
patients. These results were robust to changes in the definition of the outcome. The results
of this study are particularly noteworthy given the high specificity of the outcome
definition.
In summary, large studies from the U.S., Europe, and Australia/New Zealand
published since the 2004 PM AQCD (U.S. EPA, 2004, 056905) provide support for an
association between short-term increases in ambient levels of PM10 and PM2.5 and increased
risk of hospitalization or ED visits for ischemic heart diseases. Although estimates are less
precise for PM10-2.5, most results from single pollutant models provide evidence of a positive
association between PM10-2.5 and IHD. Moreover, Host et al. (2008, 155852) found that the
effect estimates for the association of PM2.5 and PM10-2.5 with IHD were very similar when
scaled to the IQR of each metric. Estimates of the excess risk vary considerably between
studies, but as was the case for total CVD hospitalizations, the excess risk appears to
somewhat greater in Europe and Australia/New Zealand. Results from multicity studies
and U.S. and Canadian single-city studies are presented in Figure 6-2.
6.2.10.4. Acute Ml
Because even IHD refers to a heterogeneous collection of diseases and syndromes,
several authors have evaluated the association between short-term fluctuations in ambient
PM and acute MI, a specific form of ischemic heart disease.
In 2001, Peters et al. (2001, 016546) published their study evaluating the effects of
PM on the risk of MI among 772 Boston-area participants in the Determinants of MI Onset
Study. The authors found that a 10 (Jg/m3 increase in the 2-h or 24-h avg levels of PM2.5 was
associated with a 17% (95% CP 4-32) and 27% (95% CP 6-53) excess risk of MI, respectively.
An imprecise, non-significant association between PM10-2 .5 and onset of MI was observed in Boston. In
contrast, a study among 5793 patients in King County, WAthat used similar methods,
found no association with PM2.5 with lag times of 1, 2, 4, or 24 h (Sullivan et al., 2005,
050854). Among 852 hospitalized patients in Augsburg, Germany, Peters et al. (2005,
087759) also found no association between PM2.5 and MI risk within this time frame,
although they did find a positive and statistically significant association with time spent in
traffic (Peters et al., 2004, 087464).
July 2009
6-104
DRAFT-DO NOT CITE OR QUOTE

-------
These three studies are particularly important because in each one: (l) cases were
prospectively identified based on clinical criteria rather than retrospectively based on
discharge diagnoses! and (2) time of MI symptom onset was used for exposure assessment
rather than date of hospital admission. Whether the discrepant results among these studies
are due to regional differences in population characteristics and/or air pollution sources
remains unclear. The King County study suggests that differences in statistical approaches
are unlikely to account for the discrepant results (Sullivan et al., 2005, 050854). Analyses
from the U.S. MCAPS study suggest that substantial heterogeneity of effects are to be
expected across regions of the country (Bell et al., 2008, 156266)
Several studies have assessed the association between acute exposure to ambient PM
and MI using administrative databases. In the U.S., MI was not one of the specific
endpoints evaluated in the MCAPS study (Dominici et al., 2006, 088398) or in the
Atlanta-based SOPHIA study of ED visits (Metzger et al., 2004, 044222). However,
Zanobetti and Schwartz (2005, 088069) found a 0.7% (95% CP 0.3-1.0) excess risk of MI per
10 ug/ma increase in same-day PMio among elderly Medicare beneficiaries in 21 cities.
Subsequently, the same authors found that among elderly Medicare beneficiaries living in
the Boston metropolitan region a 10 |jg/m3 increase in PM2.5 was associated with a 4.9%
(95% CP 1.1-8.2) excess risk on the same day (Zanobetti and Schwartz, 2006, 090195).
This body of evidence may implicate traffic-related pollution generally as a risk factor
for MI. In the study described above, Peters et al. (2001, 016546) found positive associations
between risk of hospitalization for MI and potential markers of traffic-related pollution
measured at a central monitor including BC, CO and NO2. However, none of these
associations were statistically significant in models adjusting for season, meteorological
variables, and day of week. Zanobetti and Schwartz (2006, 090195)
examined the association between traffic-related pollution and risk of hospitalization
for MI among Medicare beneficiaries in the Boston area and found that MI risk was
positively and significantly associated with measures of PM2.5, BC, NO2, and CO, but not
with levels of non-traffic-related PM2.5. Peters et al. (2004, 087464) interviewed 691 subjects
with MI who survived at least 24-h after the event and found a strong positive association
between self-reported exposure to traffic and the onset of MI within 1 h (OR: 2.9 [95% CP
2.2-3.8] p < 0.001). The association was somewhat stronger among subjects traveling by
bicycle or public transportation in the hour prior to the event. Of note, however, this study
did not directly measure traffic-related pollution.
July 2009
6-105
DRAFT-DO NOT CITE OR QUOTE

-------
Similar studies with administrative databases have been conducted in Europe,
Australia, and New Zealand. In Rome, D'Ippoliti et al. (2003, 074311) carried out a
case-crossover study and found a statistically significant positive association between TSP
and the risk of hospitalization for MI. Barnett et al. (2006, 089770) observed that in 5 cities
in Australia and New Zealand, a 10 |jg/m3 increase in PM2.5 was associated with a 7.3%
(95% CP 3.5-11.4, lag 0-1 day) excess risk. In contrast, the HEAPSS study found no
evidence of an association between PM10 and risk of hospitalization for a first MI in 5
European cities (Lanki et al., 2006, 089788), although there is some indication that among
survivors of a first MI, risk of re-hospitalization for MI may be related to transient
elevations in PM10 (Von Klot et al., 2005, 088070).
In summary, large studies from the U.S., Europe, and Australia/New Zealand
published since the 2004 PM AQCD (U.S. EPA, 2004, 056905) provide support for an
association between short-term increases in ambient levels of PM10 and PM2.5 and increased
risk of hospitalization for MI, but not all studies have found statistically significant
associations. Some of the heterogeneity of results is likely explained by regional differences
in pollution sources, components, and measurement error. One study of the effect of 2- and
24-h avg PM10-2.5 concentration on admissions for MI produced effect estimates that were
positive but imprecise (Peters et al., 2001, 016546). These results need to be interpreted
together with those studies evaluating hospitalization for IHD since Mis make up the
majority of hospitalizations for ischemic heart diseases. U.S. studies of MI are included in
Figure 6-2.
6.2.10.5. Congestive Heart Failure
Perhaps the first suggestion of an association between ambient PM and
hospitalization for congestive heart failure (CHF) was provided by the study of Schwartz
and Morris (1990, 073222). These authors reported that among elderly Medicare
beneficiaries living in Detroit between 1986 and 1989, a 10 (Jg/m3 increase in mean PM10
levels over the previous two days was associated with a 1.0% (95% CP 0.4-1.6%) increase in
risk of hospitalization for CHF. As reviewed in the 2004 PM AQCD (U.S. EPA, 2004,
056905), using similar approaches, statistically significant positive associations with PM10
or PM2.5 were subsequently reported in single-city studies looking at hospitalizations for
CHF in Toronto (Burnett et al., 1999, 017269), Hong Kong (Wong et al., 1999, 009172). and
Detroit (Lippmann, 2000, 024579), but not Los Angeles (Linn et al., 2000, 002839) or
July 2009
6-106
DRAFT-DO NOT CITE OR QUOTE

-------
Denver (Koken et al., 2003, 049466). Burnett et al. (1999, 017269) reports a significantly
increased risk of CHF hospitalization with PM10-2.5 while Metzger et al. (2004, 011222) and
(Ito, 2003, 042856) less precise associations.
Subsequent multicity studies support the presence of a positive association. In the
U.S., Wellenius et al. (2006, 088748) reported a 0.7% (95% CI: 0.4-1.1) excess risk of
hospitalization for CHF per 10 ug/m3 increase in same-day PM10 among elderly Medicare
beneficiaries in 7 cities. The larger MCAPS study found a 1.3% (95%: 0.8-1.8) excess risk
per 10 (Jg/m3 increase in same-day PM2.5 (Dominici et al., 2006, 088398). In Australia and
New Zealand, Barnett et al. (2006, 089770) found a 9.8% (95% CP 4.8-14.8, lag 0-1 day) and
4.6% (95% CP 2.8-6.3, lag 0-1 days) excess risk of hospitalization for CHF associated with a
10 (Jg/m3 increase in PM2.5 and PM10, respectively. Results from more recent single-city
studies in Pittsburgh (Wellenius et al., 2005, 087483), Utah's Wasatch Front (Pope et al.,
2008, 191969), Kaohsiung, Taiwan (Lee et al., 2007, 093042) and Taipei, Taiwan (Yang,
2008, 157160) have also reported positive associations. In addition, Yang et al. (2009,
190341) found that hospitalizations for heart failure were elevated during or immediately
following 54 Asian dust storm events (while single day lags 0 to 3 were evaluated,
maximum excess risk occurred at lag V- 11.4% [95% CP -0.7% to 25.0%]). Delfino et al.
(2009, 191994) observed a slightly larger excess risk of total CHF during the wildfire
occurring in California compared to the period before the wildfires (risks for CVD reported
were generally weak and non-significant).
While most studies suggest an association at very short lags (0-1 days), the study by
Pope et al. (2008, 191969) failed to find such short term associations and instead suggested
that PM2.5 levels averaged over the past 2-3 wk may be more important. Pope et al. (2008,
191969) observed a 13.1% (1.3%, 26.2%) increase in CHF hospitalization per 10 |u,g/m3
increase in PM2.5 (imputed values used in analysis). Whether findings at longer lags in this
population represent true cumulative effects of PM or are due to misclassification of
symptom onset times remains to be determined.
Findings from the Atlanta-based SOPHIA study (Metzger et al., 2004, 011222) also
support the presence of a positive association. Specifically, the SOPHIA study found a 5.5%
(95% CP 0.6-10.5, lag 0-2 days) excess risk of ED visits for CHF per 10 (Jg/m3 increase in the
3-day ma of PM2.5. Positive associations were also observed for CHF and EC and organic
carbon components of PM2.5. No associations were observed with PM10 and a weak imprecise
increase was observed in association with PM10-2.5.
July 2009
6-107
DRAFT-DO NOT CITE OR QUOTE

-------
Only one published study has attempted to evaluate the effects of ambient particles
on CHF symptom exacerbation using data which was not derived from administrative
databases. Symons et al. (2006, 091258) interviewed 135 patients with prevalent CHF
hospitalized for symptom exacerbation in Baltimore, MD. The authors found a 7.4% (95%
CP "7.5 to 24.2) excess risk of hospitalization per 10 (Jg/m3 increase in PM2.5 two days prior
to symptom onset. Although the authors' findings did not reach statistical significance, the
study lacked the statistical power needed to find an effect of the expected magnitude.
In summary, large studies from the U.S., Europe, and Australia/New Zealand
published since the 2004 PM AQCD (U.S. EPA, 2004, 056905) provide support for an
association between short-term increases in ambient levels of PM10 and PM2.5 and increased
risk of hospitalization and ED visits for heart failure. Although the number of studies is
fewer (and only Metzger et al., 2004, 011222 is new since the 2005 AQCD), elevated risks of
hospitalization or ED visits for CHF in association with PM10-2.5 have been observed. The
excess risk associate with heart failure hospitalizations and ED visits are consistently
greater than those observed for other CVD endpoints. The results of multicity studies and
U.S. and Canadian single-city studies are summarized in Figure 6"3.
July 2009
6-108
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Location
Lag
Burnett (1999, 017269), All Ages
Toronto
0-2
Ito et al. (2003, 042856); Lippman (2000, 024579), 65 +
Detroit
1
Metzger et al. (2004, 044222)*, All Ages
Atlanta
0-2
Barnett (2006, 0897701,15-64
AustraliafNZ
0-1
Svmons (2006, 091258), All Ages
Baltimore
2
Dominici et al. (2006, 088398)
204 US Counties
0
Barnett (2006, 089770), 65 +
AustraliafNZ
0-1
Delfino et al. (2009,191994), all ages
CA wildfires
0-1
Pope et al. (2008,191969)

14 d DL
Excess Risk Estimates
PMio
Burnett (1999,017269), All Ages	Toronto	0-2
Ito et al. (2003, 042856); Lippman (2000, 024579), 65+ Detroit	0
Metzger et al. (2004,044222)", All Ages	Atlanta	0-2		
Burnett (1999,017269)

0-2
Linnet al. (2000, 002839), >30
Los Angeles
0
Ito et al. (2003, 042856); Lippman (2000, 024579), 65 +
Detroit
0
Metzger et al. (2004, 044222)*, All Ages
Atlanta
0-2
Barnett (2006, 0897701,15-64
AustraliafNZ
0-1
Schwartz et al. (1995, 046186)

0-1
Morris (1998, 086857), 65 +
Chicago
0
Wellenius (2005, 087483), 65 +
Pittsburgh
0
Wellenius (2006, 088748), 65 +
US 7 Cities
0
Barnett (2006, 089770), 65 +
AustraliafNZ
0-1
I	1	1	1	1	1	1	1	1	1	1	1	1
-6-8-4-2 0 2 4 « 8 tO 12 14 16
Figure 6-3. Excess risk estimates per 10 yug/m3 increase in PMio, PM102.5 and PM10 2.5 for studies of
CHF ED visits* and hospitalizations. Studies represented in the figure include all
multicity studies. Single-city studies conducted in the U.S. and Canada are also included.
6.2.10.6. Cardiac Arrhythmias
A number of studies based on administrative databases have sought to evaluate the
association between short-term fluctuations in ambient PM levels and the risk of
hospitalization for cardiac arrhythmias (also known as dysrhythmias). Typically in these
studies a primary discharge diagnosis of ICD-9 427 has been used to identify hospitalized
patients. However, ICD-9 427 includes a heterogeneous group of arrhythmias including
paroxysmal ventricular or supraventricular tachycardia, atrial fibrillation and flutter,
July 2009	6-109	DRAFT - DO NOT CITE OR QUOTE

-------
ventricular fibrillation and flutter, cardiac arrest, premature beats, and sinoarterial node
dysfunction. One study in the Netherlands found that the positive predictive value of ICD-9
codes related to ventricular arrhythmias and sudden cardiac death was 82% (De Bruin et
al., 2005, 155746). The positive predictive value of other codes related to cardiac
arrhythmias is unknown but likely to be lower.
The results from early studies of arrhythmia-related hospitalizations have been
inconsistent with positive findings in Toronto (Burnett et al., 1999, 017269) and null
findings in Detroit (Schwartz and Morris, 1995, 046186), Los Angeles (Linn et al., 2000,
002839), and Denver (Koken et al., 2003, 049466). The U.S. MCAPS study found a
statistically significant 0.6% (95% CP 0.0-1.2%) excess risk of hospitalization for the
combined end point of cardiac arrhythmias and conduction disorders (ICD-9: 426, 427) per
10 jug/m3 increase in same-day PM2.5 (Dominici et al., 2006, 088398). Amulticity study in
Australia and New Zealand found no evidence of an association between arrhythmia
hospitalizations and either PM10 or PM2.5 (Barnett et al., 2006, 089770). A study in
Helsinki, Finland found no evidence of an association between either PM2.5 or PM10-2.5 and
risk of hospitalization for arrhythmias (Halonen et al., 2009, 180379), although there was
an association with smaller particles (0.03-0.1 |um).
With regard to ED visits, the Atlanta-based SOPHIA study found no evidence of an
association between any measure of ambient PM and the rate of ED visits for cardiac
arrhythmias (Metzger et al., 2004, 044222). However, In Sao Paulo, Brazil, Santos et al.
(2008, 192004) found a 3.0% (95% CP 0.5-5.4%) excess risk of ED visits for arrhythmias per
10 ug/ma increase in PM10 on the same day.
A number of studies in patients with ICD have been able to evaluate in more detail
the potential link between ambient air pollution and the risk of atrial and ventricular
arrhythmias (as opposed to hospital admissions or ED visits for these events) (Berger et al.,
2006, 098702; Dockery et al., 2005, 078995; Dockery et al., 2005, 090743; Ljungman et al.,
2009, 191983; Metzger et al., 2007, 092856; Peters et al., 2000, 011347; Rich et al., 2004,
055631; Vedal et al., 2004, 055630). An important strength of these studies is the ability to
examine recordings of arrhythmic episodes, thereby reducing misclassification of the
outcome. These studies are reviewed in detail in Section 6.1.2.1.
In summary, the current evidence does not support the presence of a consistent
association between short-term increases in ambient levels of PM10, PM2.5 or PM10-2.5 and
increased risk of hospitalization for cardiac arrhythmias. However, it should be noted that
July 2009
6-110
DRAFT-DO NOT CITE OR QUOTE

-------
studies of hospital admissions or ED visits are ill-suited to the study of cardiac arrhythmias
since most arrhythmias do not lead to hospitalization. Studies in patients with implanted
defibrillators, human panel studies with ambulatory ECG recordings, and animal
toxicological studies provide a more appropriate setting for evaluating this endpoint.
Results of these studies are described in Section 6.1.2.
6.2.10.7. Cerebrovascular Disease
Time-series studies evaluating the hypothesis that short-term increases in ambient
PMio or PM2.5 levels are associated with increased risk of hospitalization for CBVD have
been inconsistent with a minority of studies reporting statistically significant positive
associations (Chan et al., 2006, 090193; Dominici et al., 2006, 088398; Metzger et al., 2004,
044222; Wordley et al., 1997, 082745), and several studies reporting null or negative
associations (Anderson et al., 2001, 017033; Barnett et al., 2006, 089770; Burnett et al.,
1999, 017269; Halonen et al., 2009, 180379; Jalaludin et al., 2006, 189416; Larrieu et al.,
2007, 093031; Le Tertre et al., 2002, 023746; Peel et al., 2007, 090442; Villeneuve et al.,
2006, 090191; Wong et al., 1999, 009172).
The U.S. MCAPS study found a 0.8% (95% CP 0.3-1.4) excess risk of hospitalization
for CBVD per 10 |ug/m3 increase in same-day PM2.5 (Dominici et al., 2006, 088398). The
association showed regional variability with the strongest associations observed in the
eastern U.S. The Atlanta-based SOPHIA study found a 2.0% (95% CP -0.1 to 4.3, lag 0-2
days) excess risk of ED visits for a combined endpoint of cerebrovascular and peripheral
vascular disease excluding hemorrhagic strokes per 10 jug/m3 increase in PM10 and a 5.0%
(95% CP 0.8-9.3, lag 0-2 days) excess risk for PM2.5 (Metzger et al., 2004, 044222). Delfino
et al. (2009, 191994) observed a similar weak excess risks of CBVD admissions before and
during the wildfire occurring in California and slightly higher risks after the wildfire period
(risks for CVD reported were generally weak and non-significant).
In contrast, large multicity studies outside of North America have failed to observe an
association. The APHEA study, found no excess risk (0.0% [95% CP -0.3 to 0.3]) of
hospitalization for CBVD per 10 (Jg/m3 increase in the 2-day ma of PM10 in 8 European
cities (Le Tertre et al., 2002, 023746). Investigators from the French PSAS program
reported a 0.8% (95% CP -0.9 to 2.5, lag 0-1 days) excess risk per 10 |ug/m3 increase in PM10
among patients aged > 65 yr and a 0.2% (95% CP -1.6 to 1.9, lag 0-1 days) excess risk
among all patients (Larrieu et al., 2007, 093031). Although neither estimate was
July 2009
6-111
DRAFT-DO NOT CITE OR QUOTE

-------
statistically significant, the estimated excess risk among the elderly is very similar to that
observed in the U.S. MCAPS study. Barnett et al. (2006, 089770) examined this hypothesis
in New Zealand and Australia and found no association, but the authors did not report
point estimates or confidence intervals.
All of the above studies have identified cases of CBVD based on ICD-9 or ICD-10
codes (most commonly ICD-9 430-438). However, the range of ICD codes commonly used in
these studies includes ischemic strokes, hemorrhagic strokes, transient ischemic attacks
(TIAs) and several poorly defined forms of acute neurological events (e.g., seizures from a
vascular cause) (see Table 6-5). It is plausible that ambient PM has different effects on each
of these disparate outcomes.
Ischemic Strokes and Transient Ischemic Attacks
An increasing number of studies have specifically evaluated the association between
PMio and PM2.5 and the risk of ischemic stroke (Chan et al., 2006, 090193; Henrotin et al.,
2007, 093270; Linn et al., 2000, 002839; Lisabeth et al., 2008, 155939; Tsai et al., 2003,
080133; Villeneuve et al., 2006, 090191; Wellenius et al., 2005, 087483; Low et al., 2006,
090441; Szyszkowicz, 2008, 192128). Linn et al. (2000, 002839) found a 1.3% (95% CP
1.0-1.6 per 10 |Jg/m3, lag 0) excess risk of hospitalization for ischemic stroke in the Los
Angeles metropolitan area. Wellenius et al. (2005, 087483) reported a statistically
significant 0.4% (95% CP 0.0-0.9) excess risk per 10 jug/m3 increase in same-day PM10
among elderly Medicare beneficiaries in 9 U.S. cities. Low et al. (2006, 090111) reported an
absolute increase of 0.08 (95% CI; 0.002-0.16) ischemic stroke hospitalizations per 10 |u,g/m3
increase in PM10 in New York City. In Kaohsiung, Taiwan, Tsai et al. (2003, 080133) found a
5.9% (95% CP 4.3-7.4, lag 0-2 days) excess risk of hospitalization for ischemic stroke per
10 ug/m3 increase in PM10 after excluding days with mean daily temperature <20°C.
Meanwhile, in Taipei, Taiwan, Chan et al. (2006, 090193) found a 1.6% (95% CP -0.8 to 3.9,
lag 3) and 3.0% (95% CP -0.8 to 6.6, lag 3) excess risk per 10 |jg/m3 increase in PM10 and
PM2.5, respectively. Villeneuve et al. (2006, 090191) and Szyszkowicz et al. (2008, 192128)
found no association between either PM2.5 or PM10 and ED visits for acute ischemic stroke
in Edmonton, Canada.
Two recent studies are particularly noteworthy given the high specificity of the
outcome definition. Henrotin et al. (2007, 093270) used data on 1432 confirmed cases of
ischemic stroke from the French Dijon Stroke Register and found 0.9% (-7.0, 9.4%) excess
July 2009
6-112
DRAFT-DO NOT CITE OR QUOTE

-------
risk of ischemic stroke per 10 (Jg/m3 increase in PMio on the same day and a 1.1% ("0.2,
9.4%) excess risk on the previous day (lag 1 day). Lisabeth et al. (2008, 155939) used data
on 2,350 confirmed cases of ischemic stroke and 1,158 cases of TIAfrom the Brain Attack
Surveillance in Corpus Christi Project (BASIC), a population-based stroke surveillance
project designed to capture all strokes in Nueces County, Texas. The authors found a 6.0%
(95% CP -0.8 to 13.2) and 6.0% (95% CP -1.8 to 14.4) excess risk of ischemic stroke/TIA per
10 |ug/m3 increase in PM2.5 on the previous day and the same day, respectively.
Only the study by Villeneuve et al. (2006, 090191) specifically evaluated the
association between ambient PM and the risk of TIAs. This study failed to find any
evidence of an association with either PM2.5 or PM10.
A limitation of all of these studies is that they have assessed exposure based on the
date of hospital admission or ED presentation rather than the date and time of stroke
symptom onset. It has been shown that this can bias health effect estimates towards the
null by up to 60% (Lokken et al., 2009, 186774). Therefore, if there is a causal link between
PM and the risk of stroke, it is likely that the existing studies underestimate the true
effects. Moreover, most of these studies have evaluated only very short-term effects (lags of
0-2 days) and none have considered lags longer than 5 days. It is possible that the lag
structure of the association between PM and stroke differs from that of other
cardiovascular diseases and it might even differ by stroke type.
Hemorrhagic Strokes
Most of the studies in the preceding section also evaluated the association between
ambient PM and the risk of hemorrhagic stroke (Chan et al., 2006, 090193; Henrotin et al.,
2007, 093270; Tsai et al., 2003, 080133; Villeneuve et al., 2006, 090191; Wellenius et al.,
2005, 087483). In Kaohsiung, Taiwan, Tsai et al. (2003, 080133) noted a 6.7% (95% CP
4.2-9.4, lag 0"2 days) excess risk of hospitalization for hemorrhagic stroke per 10 (Jg/m3
increase in PM10, after excluding days where the mean temperature was <20°C. However,
in the U.S., Wellenius et al. (2005, 087483) failed to find any association between ambient
PM10 levels and risk of hemorrhagic stroke among Medicare beneficiaries in 9 U.S. cities.
Similarly, Villeneuve et al. (2006, 090191) found no evidence of an association between ED
visits for hemorrhagic stroke and either PM10 or PM2.5 levels in Edmonton, Canada.
Henrotin et al. (2007, 093270) found no evidence of an association between risk of
hospitalization and PM10 levels in Dijon, France, and Chan et al. (2006, 090193) found no
July 2009
6-113
DRAFT-DO NOT CITE OR QUOTE

-------
evidence of an association between risk of hospitalization and either PMio or PM2.5 levels in
Taipei, Taiwan.
In summary, large studies from the U.S., Europe, and Australia/New Zealand
published since the 2004 PM AQCD (U.S. EPA, 2004, 056905) provide inconsistent support
for an association between short-term increases in ambient levels of PM10 and PM2.5 and
risk of hospitalization and ED visits for CBVD (Figure 6-4). Studies of PM10-2.5 and CBVD or
stroke have not been conducted. The heterogeneity in results is likely partly attributed to:
l) differences in the sensitivity and specificity of the various outcome definitions used in the
relevant studies! 2) lag structures between PM exposure and stroke onset which may vary
by stroke type and patient characteristics! and 3) exposure misclassification due to the use
of hospital admission date rather than stroke onset time, which may vary by region,
population characteristics, and stroke type. Effect estimates from multicity studies and
single-city U.S. and Canadian studies are included in Figure 6-4.
6.2.10.8. Peripheral Vascular Disease
In the U.S., the large MCAPS study Dominici et al. (2006, 088398) evaluated the
association between mean daily PM2.5 levels and the risk of hospitalization among elderly
Medicare beneficiaries in 204 urban counties and found that a 10 (Jg/ni3 increase in PM2 5 was not
significantly associated with risk of hospitalization for peripheral vascular disease 0-2 days later. An
earlier study in Toronto (Burnett et al., 1999, 017269) found a negative association with PM2 5 (point
estimate and confidence intervals not reported), a positive non-significant association with PM10 (0.5%
[95% CI: -0.5 to 1.6%]), and a positive statistically significant association with PMi0-2.5 (2.2% [95% CI:
0.1-4.3%]). The Atlanta-based SOPHIA study (Metzger et al., 2004, 044222) of ED visits grouped visits
for PVD with those for CBVD, making interpretation of these results challenging.
July 2009
6-114
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Location
Lag

Excess Risk Estimates
Dominici et al. (2006, 088398), 65 +
204 US Counties
0

+ CBVD PM2.5

Northeast
0

r«- CBVD

Southeast
0

-1*- CBVD

Midwest
0

-4 CBVD

South
0

-p»— CBVD
Metzqer et al. (2004, 044222), All Aqes
Atlanta
0-2

\	«	 CBVDIPVD
Delfino et al. (2009,191994), all aqes
California
0-1

-1*— California Wildfire, CBVD
Metzqer et al. (2004, 044222), All Aqes
Atlanta, GA
0-2

[
[
-M	 CBVDIPVD PM10
Le Tertre et al. (2002, 0237461/2003 reanalvsis)
8 Cities Europe
0-1

• lCBVD
Larrieu et al. (2007, 093031), 65 +
8 Cities France
0-1

—CBVD
Larrieu et al. (2007, 093031), All Aqes
8 Cities France
0-1

—*t- CBVD
1
i
Villeneuve et al. (2006, 090191), 65 +
Edmonton, Canada
0-2 IS, Cool
«	
[
[
	«	¦	 PM2.5


0-2




0-2
<	
	*	1	 TIA, Cool


0-2
«	
—		1	 TIA, Warm
[
Wellenius (2005, 0886851,65 +
9 U.S. Cities
0

l
[
#1 IS PM10
Lisabeth et al. (2008,155939), All Aqes
Nueces County, TX
1

—I » IS/TIA


0

	1	» IS/TIA
Villeneuve et al. (2006, 090191), 65 +
Edmonton, Canada
0-2

	•—1	 IS, Cool


0-2

	[—•	 IS, Warm


0-2

	•	1— TIA, Cool


0-2

	•	;— TIA, Warm
Linnet al. (2000, 002839), >30
LA, California
0

t IS
[


[

PM2.5
Villeneuve et al. (2006, 090191), 65 +
Edmonton, Canada 0-2 HS, Cool ^—
	1	
	«	
	>

0-2
HS, Warm	1—

—*	>
Wellenius (2005,088685), 65+	9 U.S. Cities 0	—*-[— HS	PMio
Villeneuve et al. (2006, 090191), 65+	Edmonton, Canada 0-2 HS, Cool ^	[	»	J
0-2	H S, Warm			»	>
l
I I I I I I I l'l I I I I I I I I I I
-14 -10 -6 -2 2 4 6 8 12 18 20
Figure 6-4. Excess risk estimates per 10 (Jg/m3 increase in PM10 and PM2.5 for studies of ED
visits* and hospitalizations for CBVDs. Studies represented in the figure included all
multicity studies. Single-city studies conducted in the U.S. and Canada are also included.
July 2009
6-115
DRAFT-DO NOT CITE OR QUOTE

-------
In summary, there is insufficient published data to determine whether or not there
may be an association between short-term increases in ambient levels of PMio, PM2.5, or
PM10-2.5 and increased risk of hospitalization and ED visits for PVD.
6.2.10.9. Copollutant Models
Relatively few studies have evaluated the effects of PM2.5 and PM10-2.5 on the risk of
hospital admissions and ED visits in the context of 2 pollutant models. Generally, results
for health effects of both size fractions are similar even after controlling for SO2 or O3 levels
(Figure 6"5). However, controlling for NO2 or CO has yielded conflicting results. Among the
large multicity studies, the Atlanta-based SOPHIA study found that the association
between PM2.5 (total carbon) and risk of cardiovascular ED visits was somewhat attenuated
in 2-pollutant models additionally controlling for either CO or NO2 (Tolbert et al., 2007,
090316). Barnett et al. (2006, 089770) found that the associations they observed between
PM2.5 and cardiac hospitalizations in Australia and New Zealand were attenuated after
control for 24-h NO2, but not after control for CO.
Only a few studies have attempted to evaluate the effects of one PM size fraction
while controlling for another PM size fraction. The large U.S. MCAPS study evaluating the
effects of PM10-2.5 on cardiovascular hospital admissions lost precision after controlling for
PM2.5, but did not consider gaseous pollutants (Peng et al., 2008, 156850). Andersen et al.
(2008, 189651) found that associations between both PM10 and PM2.5 and cardiovascular
hospitalizations in Copenhagen were not attenuated by control for particle number
concentration, a measure of UFPs.
A number of studies have also evaluated PM10 effects in the context of 2-pollutant
models with inconsistent results. The multicity Spanish EMECAS study (Ballester et al.,
2006, 088746) found that the statistically significant positive associations observed between
PM10 and cardiac hospitalizations were robust to control for other pollutants in 2-pollutant
models. Jalaludin et al. (2006, 189416) found that the effects of PM10 as well as PM2.5 on
cardiovascular ED visits in Sydney Australia were attenuated by additional control for
either NO2 or CO. Wellenius et al. (2005, 087483) found that the PMicrrelated risk of
hospitalization for CHF in Allegheny County, PA, was attenuated in 2-pollutant models
controlling for either CO or NO2. In contrast, Chang et al. (2004, 055637) examined CHF
hospitalizations in Taipei and found attenuation of PM10 effects by control for NO2 or CO,
but only during warm days. In Kaohsiung, Taiwan, Tsai et al. (2003, 080133) found that the
July 2009
6-116
DRAFT-DO NOT CITE OR QUOTE

-------
association between PMio and ischemic stroke hospitalizations was not materially
attenuated in 2-pollutant models controlling for either NO2 or CO.
The inconsistent findings after controlling for gaseous pollutants or other size
fractions are likely due to differences in the correlation structure among pollutants as well
as differing degrees of exposure measurement error.
July 2009
6-117
DRAFT-DO NOT CITE OR QUOTE

-------
Authors
Outcome
Pollutant

Excess Risk Estimate
Tolbert, Klein (2007, 090316)
CVD
PM2.6

—f*— PM2.5 adjusted for gases


PM2.6TC

1 	»	


PM2.BTC+CO

	1—§


PM2.BTC+NO2

1 .


PM2.BTC+CO+NO2

	1 t
Barnett, Williams (2006, 089770)
HRT
PM2.6

' 	1	


PM2.6+NO2




PM2.6+CO

1 *	
Villeneuve, Chen (2006, 090191)
HS
PM2.6
<	
	1	•	>


PM2.6+NO2
<	
	!-•	>
Burnett, Cakmak (1997, 084194)
HRT
PM2.6

1 •	


PM2.6+O3

1 ¦


PM2.6+NO2




PM2.6+SO2




PM2.6+CO

1
Ito (2003, 042856); Lippman et al. (2000, 011938)
CHF
PM2.6

1 *


PM2.6+O3




PM2.B+SO2

1 |


PM2.B+NO2




PM2.6+CO

1	1	

IHD
PM2.B

1 g


PM2.B+O3




PM2.B+SO2

—¦ 1


PM2.B+NO2

—j—1	


PM2.B+CO

	1 i
Moolgavkar (2003, 051316); (2000, 010305)
CVD
PM2.B

> —§—


PM2.B+CO


Jalaludin, Morgan (2006,189416)
CVD
PM2.B
<	
	1	1	>


PM2.B+NO2
<	
	1	1	
1
Ito (2003, 042856); Lippman et al. (2000, 011938)
IHD
PM10-2.B

1 >	 PM 10-2.5 adjusted for gases


PM10-2.B+O3

1 ,


PM10-2.B+SO2

,	1


PMio-2.b+N02

t	1	


PMio-2.b+CO

1 .

CHF
PM10-2.B

	1	~


PM10-2.B+O3

	1	•	


PM10-2.B+SO2

—:	~	


PMio-2.b+N02




PMio-2.b+CO

	1—«	
Burnett, Cakmak (1997, 084194)
HRT
PM10-2.B

1


PM10-2.B+O3




PMio-2.b+N02

	'—•	


PM10-2.B+SO2

1


PMio-2.b+CO

1 •
t
Peng, Chang (2008,186787)
CVD
PM10-2.B

V PM adjusted for other size fractions


PM10-2.B+PM2.B

»-


PM2.6




PM2.B + PM10-2.B


Andersen, Whalin (2008,189651)
CVD
PM2.B

" 	•	


Number Cone.

1 |
1
1
	.	1	.	1	.	.	1	.	.	«
I I I I I I I I I I I I
-6 -4 -2 0 2 4 6 fl 10 12 14 16
Figure 6-5. Excess risk estimates per 10 (Jg/m3 increase in PM2.5, and PM102.5 for studies of ED
visits* and hospitalizations for CVDs. Results presented for multipollutant models.
July 2009
6-118
DRAFT-DO NOT CITE OR QUOTE

-------
6.2.10.10. Concentration Response
The concentration-response relationship has been extensively analyzed primarily
through studies that examined the relationship between PM and mortality. These studies,
which have focused on both short- and long-term exposures to PM have consistently found
no evidence for deviations from linearity or a safe threshold (Daniels et al. 2004; Schwartz
et al. 2004; Samoli et al. 2005; Schwartz et al. 2008) (see Section 6.5.2.7 and 7.1.4).
Although on a more limited basis, studies that have examined PM effects on cardiovascular
hospital admissions and ED visits have also analyzed the PM concentration-response
relationship, and contributed to the overall body of evidence which suggests a log-linear, no-
threshold PM concentration-response relationship.
The evaluation of cardiovascular hospital admission and ED visit studies in 2004 PM
AQCD (U.S. EPA, 2004, 056905) found no evidence for a threshold in the dose-response
relationship between short-term exposure to PMio and IHD hospital admissions (Schwartz
and Morris, 1995, 046186). An evaluation of both recent single and multicity studies of
hospital admission and ED visits for cardiovascular diseases further supports this finding.
Ballester et al. (2006, 088746) examined the linearity of the relationship between air
pollutants (including PMio) and cardiovascular hospital admissions in 14 Spanish cities
within the EMECAM project. In this exploratory analysis, the authors examined the models
used when pollutants were added in either a linear or non-linear way (i.e., with a spline
smoothing function) to the model. Although the study does not present the results for each
of the pollutants evaluated individually, overall Ballester et al. (2006, 088746) found that
the shape of the pollutant-cardiovascular hospital admission relationship was most
compatible with a linear curve. Wellenius et al. (2005, 087483) conducted a similar analysis
when examining the relationship between PMio and CHF hospital admissions among
Medicare beneficiaries. The authors examined the assumption of linearity using fractional
polynomials and linear splines. The results of both approaches contributed to Wellenius
et al. (2005, 087483) concluding that the assumption of linearity between the log relative
risk of cardiovascular hospital admissions and PM concentration was reasonable.
Unlike the aforementioned studies that examined the linearity in the concentration-
response curve as part of the model selection process (i.e., to determine the most
appropriate model to use to examine the relationship between PM and cardiovascular
hospital admissions and ED visits), Zanobetti and Schwartz (2005, 088069) conducted an
extensive analysis of the shape of the concentration-response curve and the potential
July 2009
6-119
DRAFT-DO NOT CITE OR QUOTE

-------
presence of a threshold when examining the association between PMio and MI hospital
admissions among older adults in 21 U.S. cities. The authors examined the concentration-
response curve by fitting a piecewise linear spline with slope changes at 20 |jg/m3 and 50
|jg/m3. This approach resulted in an almost linear concentration-response relationship
between PMio and MI hospital admissions wit h a steeper slope occurring below 50 |Jg/m3
(see Figure 6-6). Additionally, Zanobetti and Schwartz (2005, 088069) found no evidence for
a threshold.
o
.S 3.0
S 2.0
«
a. U
t	1	r
30 40 50 60
PMLJiifl/m3)
Source: Zanobetti and Schwartz (2005, 088069).
Figure 6-6. Combined random-effect estimate of the dose-response relationship between Ml
emergency hospital admissions and PMio, computed by fitting a piecewise linear spline,
with slope changes at 20 jjgfm3 and 50 pg/m3.
Overall, the limited evidence from the studies that examined the concentration-
response relationship between PM and cardiovascular hospital admissions and ED visits
supports a no-threshold, log-linear model, which is consistent with the observations made
in studies that examined the PM-mortality relationship (see Section 6.5.2.7).
6.2.10.11.Out of Hospital Cardiac Arrest
One study of out of hospital cardiac death conducted in Seattle, WA (Checkoway et al.,
2000, 015527), which reported no association with PM was included in the 2004 PM AQCD
(U.S. EPA, 2004, 056905). In the U.S., the survival rate of sudden cardiac arrest is less than
July 2009
6-120
DRAFT-DO NOT CITE OR QUOTE

-------
5%. In addition, as discussed above, Zeka et al. (2006, 088749) found that the estimated
mortality risk due to short-term exposure to PMio was much higher for out-of-hospital
cardiovascular deaths than for in-hospital cardiovascular deaths. The analysis of studies
that examine the association between PM and cardiac arrest could provide evidence for an
important link between the morbidity and mortality effects attributed to PM.
Sullivan et al. (2003, 043156) examined the association between the incidence of
primary cardiac arrest and daily measures of PM2.5 (measured by nephelometry) using a
case-crossover analysis of 1,206 Washington State out-of-hospital cardiac arrests
(1985-1994) among persons with (n = 774) and without (n = 432) clinically recognized heart
disease. The authors examined PM associations at 0- through 2-day lags using the
time-stratified referent sampling scheme (i.e., the same day of the week and month of the
same year). The estimated relative risk for a 13.8-|Jg/m3 increase in 1-day lag PM2.5
(nephelometry: IQR = 0.54 km1 bsp) was 0.94 (95% CI: 0.88-1.02), or 0.96 (95% CI: 0.91-1.0)
per 10 (Jg/m3 increase. Similar estimates were reported for 0- or 2-day lags. The presence or
absence of clinically recognized heart disease did not alter the result. This finding is
consistent with the previous study of cardiac arrest in Seattle (Levy et al., 2001, 017171)
that reported no PM association. It is also consistent with the Sullivan et al. (2005, 050854)
analysis of PM and onset of MI, and the Sullivan et al. (2007, 100083) analysis of PM and
blood markers of inflammation in the elderly population, both of which were conducted in
Seattle. Note also that the analysis of the NMMAPS data for the year 1987-1994 also found
no PM10 association for all-cause mortality in Seattle. Overall, the results of studies
conducted in Seattle consistently found no association between PM and cardiovascular
outcomes or all-cause mortality.
Rosenthal et al. (2008, 156925) examined associations between PM2.5 and
out-of-hospital cardiac arrests in Indianapolis, Indiana for the years 2002-2006 using a
case-crossover design with time-stratified referent sampling. Using all the cases (n = 1,374),
they found no associations between PM2.5 and cardiac arrest in any of the 0 through 3-day
lags or multiday averages thereof (e.g., for 0-day lag, OR = 1.02 [CI: 0.94-1.11] per 10 |ug/m3
increase in PM2.5). However, for cardiac arrests witnessed by bystanders (n = 511), they
found a significant association with PM2.5 exposure (by TEOM, corrected with FRM
measurements) during the hour of the arrest (OR = 1.12 [CI: 1.01-1.25] per 10 jug/m3
increase in PM2.5), and even larger risk estimates for older adults (age 60-75) or those that
July 2009
6-121
DRAFT-DO NOT CITE OR QUOTE

-------
presented with asystole. There have been very few PM studies that used hourly PM
measurements, and further studies are needed to confirm associations at such time scales.
In Rome, Forastiere et al. (2005, 086323) examined associations between air pollution
(particle number concentration, or PNC, PMio, CO, NO2, and O3) and oul,-of-hospital
coronary deaths (n = 5,144) for the study period of 1998-2000. A case-crossover design with
the time-stratified referent sampling was used to examine the pollution indices at lag 0
through 3 days and the average of 0-1 lags. They found associations between deaths and
PNC (lag 0 and 0-1), PM10 (lag 0, 1, and 0-1), and CO (lag 0 and 0-1) but not with NO2 or O3.
The risk estimate for 0-day lag PM10 was 1.59% (CP 0.03-3.18) per 10 |ug/m3 increase. The
older adults (65-74 and 75+ age groups) showed higher risk estimates than the younger
(35-64) age group. Because PNC is considered to be associated with UFPs, and CO was also
associated with out-of-hospital cardiac arrests, combustion sources were implicated.
In summary, only a few studies have examined out-of-hospital cardiac arrest or
deaths. The two studies from Seattle, WA consistently found no association (also consistent
with other cardiac effects and mortality studies conducted in that locale); a study in
Indianapolis, IN found an association with hourly PM2.5 but not daily PM2.5; and a study in
Rome found an association with PM10 but also with particle numbers and CO. Because
multicity mortality studies examining this association found heterogeneity in PM risk
estimates across regions, future studies of out-of-hospital cardiac arrest will need to
consider location and the air pollution mixture during their design.
6.2.11. Short-term Exposure to PM and Cardiovascular
Mortality
An evaluation of studies that examined the association between short-term exposure
to PM2.5 and PM10-2.5 and mortality provides additional evidence for PM-related
cardiovascular health effects. Although the primary analysis in the majority of mortality
studies evaluated consists of an examination of the relationship between PM2.5or PM10-2.5
and all-cause (non-accidental) mortality, some studies have examined associations with
cause-specific mortality including cardiovascular-related mortality.
Multicity mortality studies that examined the PM2.5"cardiovascular mortality
relationship on a national scale: Franklin et al. (2007, 091257)- 27 U.S. cities! Franklin et
al. (2008, 155779) - 25 U.S. cities! and Zanobetti and Schwartz (2009, 188462) - 112 U.S.
cities) have found consistent positive associations between short-term exposure to PM2.5 and
July 2009
6-122
DRAFT-DO NOT CITE OR QUOTE

-------
cardiovascular mortality ranging from 0.47 to 0.85% per 10|ug/m3 at lag 0-1 (see Section
6.5). The associations observed on a national scale are consistent with those presented by
Ostro et al. (2006, 087991) in a study that examined the I'M2 -mortality relationship in 9
California counties (0.6% [95% CP 0-1. l] per 10 |u,g/m3). Of the multicity studies evaluated,
one examined single day lags and found evidence for slightly larger effects at lag 1
compared to the average of lag days 0 and 1 for cardiovascular mortality (94% [95% CP -
0.14 to 2.02] per 10 jug/m3) (Franklin et al., 2007, 091257). Although the overall effect
estimates reported in the multicity studies evaluated are consistently positive, it should be
noted that a large degree of variability exists between cities when examining city-specific
effect estimates potentially due to differences between cities and regional differences in
PM2.5 composition (see Figure 6-24). A limited number of studies that examined the PM2.5-
cardiovascular mortality association have not conducted extensive analyses of potential
confounders, as a result, PMio-mortality studies provide evidence which suggests that PM2.5
risk estimates are fairly robust to the inclusion of gaseous copollutants in models. Overall,
the cardiovascular PM2.5 effects observed were similar to those reported for all-cause (non-
accidental) mortality (see Section 6.5), and are consistent with the effect estimates observed
in the single- and multicity studies evaluated in the 2004 PM AQCD.
Zanobetti and Schwartz (2009, 188462) also examined PM10-2.5 mortality associations
in 47 U.S. cities and found evidence for cardiovascular mortality effects (0.32% [95% CP
0.00-0.64] per 10 |ug/m at lag 0-1) similar to those reported for all-cause (non-accidental)
mortality (0.46% [95% CP 0.21-0.67] per 10 |ug/m). In addition, Zanobetti and Schwartz
(2009, 188462) reported seasonal (i.e., larger in spring and summer) and regional
differences in PM10-2.5 cardiovascular mortality risk estimates. A few single-city studies
evaluated also reported associations, albeit somewhat larger than the multicity study,
between PM10-2.5 and cardiovascular mortality in Phoenix, AZ (Wilson et al., 2007, 157149)
(3.4-6.6% at lag l) and Vancouver, CAN (Villeneuve et al., 2003, 055051) (5.4% at lag 0).
The difference in the PM10-2.5 risk estimates observed between the multi- and single-city
studies could be due to a variety of factors including differences between cities and
compositional differences in PM10-2.5 across regions (see Figure 6-29). Only a small number
of studies have examined potential confounding by gaseous copollutants or the influence of
model specification on PM10-2.5 mortality risk estimates, but the effects are relatively
consistent with those studies evaluated in the 2004 PM AQCD.
July 2009
6-123
DRAFT-DO NOT CITE OR QUOTE

-------
6.2.12. Summary and Causal Determinations
6.2.12.1. PM2.5
Several studies cited in the 2004 AQCD reported positive associations between short-
term PM2.5 concentrations and hospital admissions or ED visits for cardiovascular diseases,
although few were statistically significant. In addition, U.S. and Canadian-based studies
(both multi- and single-city) that examined the PM2.5"mortality relationship reported
associations for cardiovascular mortality consistent with those observed for all-cause
(nonaccidental) mortality and relatively stronger than those for respiratory mortality.
Significant associations were also observed between MI and short-term PM2.5 concentration
(averaged over 2 or 24 h) as well as decreased HRV in association with PM2.5. Several
controlled human exposure and animal toxicological studies demonstrated HRV effects from
exposure to PM2.5 CAPs, as well as changes in blood coagulation markers. However, the
effects in these studies were variable. Arrhythmogenesis was reported for toxicological
studies and generally these results were observed in animal models of disease (SH rat, MI,
pulmonary hypertension) exposed to combustion-derived PM2.5 (i.e., ROFA, DE, metals).
One study demonstrated significant vasoconstriction in healthy adults following controlled
exposures to CAPs, although this response could not be conclusively attributed to the
particles as subjects were concomitantly exposed to relatively high levels of O3. The results
reported for systemic inflammation in toxicological studies were mixed.
A large body of evidence from studies of the effect of PM2.5 on hospital admissions and
ED visits for cardiovascular diseases has been published since the 2004 PM AQCD.
Associations with PM2.5 are consistently positive with the majority of studies reporting
increases in hospital admissions or ED visits ranging from a 0.5 to 3.4% per 10 (Jg/m3
increase in PM2.5 (section 6.2.10). The largest U.S-based multicity study, MCAPs, reported excess
risks in the range of approximately 0.7% with the largest excess risks in the North East
(1.08%) and in the winter (1.49%), providing evidence of regional and seasonal
heterogeneity (Bell et al., 2008, 156266; Dominici et al., 2006, 088398). Weak or null
findings for PM2.5 have been observed in two single-city studies both conducted in
Washington state (Slaughter et al., 2003, 086294; Sullivan et al., 2007, 100083) and may be
explained by this heterogeneity. Weak associations were also reported in Atlanta for PM2.5
and CVD ED visits, with PM2.5 traffic components being more strongly associated with CVD
July 2009
6-124
DRAFT-DO NOT CITE OR QUOTE

-------
ED visits (Tolbert et al., 2007, 090316). Multicity studies conducted outside the U.S. and
Canada have shown positive associations with PM2.5. Studies of specific CVD outcomes
indicate that IHD and CHF may be driving the observed associations (Sections 6.2.10.3 and
6.2.10.5, respectively). Although estimates from studies of cerebrovascular diseases are less
precise and consistent, ischemic diseases appear to be more strongly associated with PM2.5
compared to hemorrhagic stroke (Section 6.2.10.7). The available evidence suggests that
these effects occur at very short lags (0-1 days), although effects at longer lags have rarely
been evaluated. Overall, the results of these studies provide support for associations
between short-term PM2.5 exposure and increased risk of cardiovascular hospital
admissions in areas with mean concentrations ranging from 7 to 18 |Jg/m3.
Epidemiologic studies that examined the association between PM2.5 and mortality
provide additional evidence for PlV^.r,-related cardiovascular effects (Section 6.2.11). The
multicity studies evaluated found consistent, precise positive associations between short-
term exposure to PM2.5 and cardiovascular mortality ranging from 0.47 to 0.85% at mean
24-h avg PM2.5 concentrations above 13 ug/m3. These associations were reported at short
lags (0-1 days), which is consistent with the associations observed in the HA and ED visit
studies discussed above. Although examinations of potential confounders of the PM2.5-
cardiovascular mortality relationship are limited, the observed associations are supported
by PMicrmortality studies, which found that PM risk estimates remained robust to the
inclusion of copollutants in models.
Recent studies that apportion ambient PM2.5 into sources and components suggest
that cardiovascular hospital admissions associated with PM2.5 may be attributable to
traffic-related pollution, and in some cases biomass burning (Section 6.2.10). Further
supporting evidence is provided by studies that have used PM10 collection filters (median
diameter generally <2.5 |Jm) to identify combustion- or traffic-related sources associated
with cardiovascular HA. Metals have also been implicated in these effects (Bell et al., 2009,
191997). A limited number of older publications have reported that particle acidity of PM2.5
is not more strongly associated with CVD hospitalizations or ED visits than other PM
metrics.
Changes in various measures of cardiovascular function have been demonstrated by
multiple independent laboratories following controlled human exposures to different types
of PM2.5. The most consistent effect is changes in vasomotor function, which has been
demonstrated following exposure to CAPs and DE. The majority of the new evidence of
July 2009
6-125
DRAFT-DO NOT CITE OR QUOTE

-------
particle-induced changes in vasomotor function comes from studies of exposures to DE
(Section 6.2.4.2). None of these studies have evaluated the effects of DE with and without a
particle trap. Therefore, the changes in vasomotor function cannot be conclusively
attributed to the particles in DE as subjects are also concomitantly exposed to relatively
high levels of NO2, NO, CO, and hydrocarbons. However, it is important to note that a study
by Peretz et al. (2008, 156854) used a newer diesel engine with lower gaseous emissions
and reported significant DE-induced decreases in BAD. In addition, increasing the particle
exposure concentration from 100 to 200 |ug/m3, without proportional increases in NO, NO2,
or CO, resulted in an approximate 100% increase in response. An additional consideration
is that while fresh DE used in these studies contains relatively high concentrations of fine
particles, the MMAD is typically < 100 nm, which makes it difficult to determine whether
the observed effects are due to the fine particles, or more specifically due to the ultrafine
fraction. Further evidence of a particle effect on vasomotor function is provided by
significant changes in BAD demonstrated in healthy adults following controlled exposure to
CAPs with O3 (Brook et al., 2002, 024987). These findings are consistent with epidemiologic
studies of various measures of vasomotor function (e.g., FMD and BAD were the most
common), which have demonstrated an association with short-term PM2.5 concentration in
healthy and diabetic populations (Section 6.2.4.1). A limited number of epidemiologic
studies examined multiple lags and the strongest associations were with either the 6 day
mean concentration (O'Neill et al., 2005, 098094) or the concurrent day (Schneider et al.,
2008, 191985) either the 6 day mean concentration (O'Neill et al., 2005, 098094) or the
concurrent day (Schneider et al., 2008, 191985).
The toxicological findings with respect to vascular reactivity are generally in
agreement and demonstrate impaired dilation following PM2.5 exposure that is likely
endothelium dependent (Section 6.2.4.3). These effects have been demonstrated in varying
vessels and in response to different PIVk.stypes, albeit using IT exposure in most studies.
Further support is provided by IT studies of ambient PM10 that also demonstrate impaired
vasodilation and a PM2.5 CAPs study that reported decreased LAV ratio of the pulmonary
artery. An inhalation study of Boston PM2.5 CAPs reported increases in coronary vascular
resistance during ischemia, which indicated a possible role for PM-induced coronary
vasoconstriction. The mechanism behind impaired dilation following PM exposure may
include increased ROS and RNS production in the microvascular wall that leads to altered
NO bioavailability and endothelial dysfunction. Despite the limited number of inhalation
July 2009
6-126
DRAFT-DO NOT CITE OR QUOTE

-------
studies conducted with concentrations near ambient levels, the toxicological studies
collectively provide coherence and biological plausibility for the myocardial ischemia
observed in controlled human exposure and epidemiologic studies.
Consistent with the observed effects on vasomotor function, one recent controlled
human exposure study reported an increase in exercise-induced ST-segment depression (a
potential indicator of ischemia) during exposure to DE in a group of subjects with prior MI
(Mills et al., 2007, 091206). In addition, toxicological studies from Boston that employed
CAPs provide further evidence for PM2.5 effects on ischemia, with changes in ST-segment
and decreases in total myocardial blood flow reported (Section 6.2.3.3). These findings from
toxicological and controlled human exposure studies provide coherence and biological
plausibility for the associations observed in epidemiologic studies, particularly those of
increases in ED visits and hospital admissions for IHD. Several epidemiology studies have
reported associations between short-term PM2.5 concentration (including traffic sources or
components such as BC) and ST-segment depression or abnormality (Section 6.2.3.1).
Toxicological studies provide biological plausibility for the PM2.5 associations with
CHF hospital admissions by demonstrating increased right ventricular pressure and
diminished cardiac contractility in rodents exposed to CB and DE (Section 6.2.6.1).
Similarly, increased coronary vascular resistance was observed following PM2.5 CAPs
exposure in dogs with experimentally-induced ischemia. Further, a recent epidemiology
study reported small but statistically significant decreases in passively monitored diastolic
pressure and right ventricular diastolic pressure (Rich et al., 2008, 156910).
In addition to the effects of PM on vasomotor response, there is a growing body of
evidence that demonstrates changes in markers of systemic oxidative stress following
controlled human exposures to DE, WS, and urban traffic particles. However, these effects
may be driven in part by the ultrafine fraction of PM2.5. Toxicological studies provide
evidence of increased cardiovascular ROS following PM2.5 exposure to CAPs, road dust, CB,
and TiC>2, as well as increased systemic ROS in rats exposed to gasoline exhaust (Section
6.2.9.3). Epidemiologic studies of markers of oxidative stress (e.g., tHcy, CuZn-SOD,
TBARS, 8-oxodG, oxLDL and MDA) are consistent with these toxicological findings (Section
6.2.9.1).
A few epidemiologic studies of ventricular arrhythmias recorded on ICDs that were
conducted in Boston and Sweden (Table 6-2) found associations with short-term PM2.5
concentration (also BC and sulfate). While Canadian and U.S. studies conducted outside of
July 2009
6-127
DRAFT-DO NOT CITE OR QUOTE

-------
Boston did not find positive associations between PM2.5 and ICD recorded ventricular
arrhythmias, several such studies observed associations with ectopic beats and runs of
supraventricular or ventricular tachycardias (Section 6.2.2.1). Toxicological studies also
provide limited evidence of arrhythmia, mainly in susceptible animal models (i.e., older
rats, rats with chronic heart failure) (Section 6.2.2.2).
Most epidemiologic studies of HRVhave reported decreases in SDNN, LF, HF, and
rMSSD (Section 6.2.1.1). While there are also a significant number of controlled human
exposure studies reporting PM-induced changes in HRV, these changes are often variable
and difficult to interpret (Section 6.2.1.2). Similarly, HRV increases and decreases have
been observed in animal toxicological studies that employed CAPs or CB (Section 6.2.1.3).
In a study in mice, resuspended soil, secondary sulfate, residual oil, and motor vehicle/other
sources, as well as Ni were implicated in HRV effects (Lippmann et al., 2006, 091165).
Further, cardiac oxidative stress has been implicated as a consequence of ANS stimulation
in response to CAPs. Modification of the PM-HRV association by genetic polymorphisms
related to oxidative stress has been observed in a series of analyses of the population
enrolled in the NAS. Changes in HRV measures (whether increased or decreased) are likely
to be more useful as indicators of PM exposure rather than predictive of some adverse
outcome. Furthermore, the HRV result may be reflecting a fundamental response of an
individual that is determined in part by a number of factors including age and pre-existing
conditions.
Although not consistently observed across studies, some investigators have reported
PM2.5-induced changes in BP, blood coagulation markers, and markers of systemic
inflammation in controlled human exposure studies (Sections 6.2.5.2, 6.2.8.2, and 6.2.9.2,
respectively). Findings from epidemiologic studies, which are largely cross-sectional and
measure a wide array markers of inflammation and coagulation, are not consistent!
however, a limited number of recent studies of gene environment interactions offer insight
into potential individual susceptibility to these effects (Ljungman et al., 2009, 191983;
Peters et al., 2009, 191992). Similarly, toxicological studies demonstrate mixed results for
systemic inflammatory markers and generally indicate relatively little change at 16-20 h
post-exposure (Section 6.2.7.3). Increases in BP have been observed in toxicological studies
(Section 6.2.5.3), with the strongest evidence coming from dogs exposed to PM2.5 CAPs. For
blood coagulation parameters, the most commonly reported change in animal toxicological
July 2009
6-128
DRAFT-DO NOT CITE OR QUOTE

-------
studies is elevated plasma fibrinogen levels following PM2.5 exposure, but this response is
not consistently observed (Section 6.2.8.3).
In summary, associations of hospital admissions or ED visits with PM2.5 for
cardiovascular diseases (predominantly IHD and CHF) are consistently positive with the
majority of studies reporting increases ranging from a 0.5 to 3.4% per 10 (Jg/m3 increase in
PM2.5. Seasonal and regional variation observed in the large multicity study of Medicare
recipients is consistent with null findings reported in several single city studies conducted
in the Western U.S. The results from the HA and ED visit studies are supported by the
associations observed between PM2.5 and cardiovascular mortality, which also provide
additional evidence for regional and seasonal variability in PM2.5 risk estimates. Changes in
various measures of cardiovascular function that may explain these epidemiologic findings
have been demonstrated by multiple independent laboratories following controlled human
exposures to different types of PM2.5. The most consistent PM2.5 effect is on vasomotor
function, which has been demonstrated following exposure to CAPs and DE. Toxicological
studies finding reduced myocardial blood flow during ischemia and altered vascular
reactivity provide coherence and biological plausibility for the myocardial ischemia that has
been observed in both controlled human exposure and epidemiologic studies. Further, PM2.5
effects on ST-segment depression have been observed across disciplines. In addition to
ischemia, PM2.5 may act through several other pathways. Plausible biological mechanisms
(e.g., increased right ventricular pressure and diminished cardiac contractility) for the
associations of PM2.5 with CHF have also been proposed based on toxicological findings.
There is a growing body of evidence from controlled human exposure, toxicological and
epidemiologic studies demonstrating changes in markers of systemic oxidative stress with
PM2.5 exposure. Inconsistent effects of PM on BP, blood coagulation markers and markers of
systemic inflammation have been reported across the disciplines. Together, the collective
evidence is sufficient to conclude that a causal relationship exists between short-term PM2.5
exposures and cardiovascular effects.
6.2.12.2. PM102.5
There was little evidence in the 2004 AQCD regarding PM10-2.5 cardiovascular health
effects. Two single-city epidemiologic studies found positive associations of PM10-2.5 with
cardiovascular HA in Toronto (Burnett et al., 1999, 017269) and Detroit, MI (Ito, 2003, 042856;
Lippmann, 2000, 024579) and the effect estimates were of the same general magnitude as for PMi0
July 2009
6-129
DRAFT-DO NOT CITE OR QUOTE

-------
and PM2 5. Both studies reported positive associations and estimates appeared robust to
adjustment for gaseous copollutants in 2-pollutant models. An imprecise, non-significant
association between PM10-2.5 and onset of MI was observed in Boston (Peters et al., 2001, 016546). No
controlled human exposure or toxicological studies of PM10-2.5 were presented in the 2004
AQCD.
Several recent epidemiologic studies of the effect of ambient PM10-2.5 concentration on
hospital admissions or ED visits for cardiovascular diseases were conducted (Section
6.2.10).	In a study of Medicare patients in 108 U.S. counties, Peng et al. (2008, 156850)
reported a significant association between PM10-2.5 and cardiovascular disease
hospitalizations in their single pollutant model. In a study of six French cities, Host et al.
(2008, 155852) reported a significant increase in IHD hospital admissions in association
with PM10-2.5. In contrast, associations of cardiovascular outcomes with PM10-2.5 were weak
for CHF and null for IHD in the Atlanta-based SOPHIA study (Metzger et al., 2004,
044222). Results from single city studies are generally positive but effect sizes are
heterogeneous and estimates are imprecise (Section 6.2.10). Crustal material from a dust
storm in the Gobi desert that was largely coarse PM (generally indicated using PM10) was
associated with hospitalizations for cardiovascular diseases including IHD and CHF in
most studies (Section 6.2.10). Mean PM10-2.5 concentrations in the hospital admission and
ED visit studies ranged from 7.4-13 |Jg/m3. A few epidemiologic studies that examined the
association between short-term exposure to PM10-2.5 and cardiovascular mortality (Section
6.2.11)	provide supporting evidence for the hospital admission and ED visit studies at
similar 24~h avg PM10-2.5 concentrations (i.e., 6.1-16.4 (Jg/m3). Amulticity study reported risk
estimates for cardiovascular mortality of similar magnitude to those for all-cause
(nonaccidental) mortality (Zanobetti and Schwartz, 2009, 188462). However, the single-city
studies evaluated (Wilson et al., 2007, 157149; Villeneuve et al., 2003, 055051) reported
substantially larger effect estimates, but this could be due to differences between cities and
compositional differences in PM10-2.5 across regions. Of note is the lack of analyses within
the studies evaluated that examined potential confounders of the PMio-2 5-cardiovascular
mortality relationship.
The U.S. study of Medicare patients (Peng et al., 2008, 156850) and the multicity
study that examined the association between PM10-2.5 and mortality (Zanobetti and
Schwartz, 2009, 188462) were the only studies to adjust PM10-2.5 for PM2.5. Peng, et al.
(2008, 156850) found that the PM10-2.5 association with cardiovascular disease
July 2009
6-130
DRAFT-DO NOT CITE OR QUOTE

-------
hospitalizations remained, but diminished slightly after adjustment for PM2.5. These results
are consistent with those reported by Zanobetti and Schwartz (2009, 188462), which found
PMio-2.5-cardiovascular mortality risk estimates remained relatively robust to the inclusion
of PM2.5 in the model. Because of the greater spatial heterogeneity of PM10-2.5, exposure
measurement error is more likely to bias health effect estimates towards the null for
epidemiologic studies of PM10-2.5 versus PM10 or PM2.5, making it more difficult to detect an
effect of the coarse size fraction. In addition, models that include both PM10-2.5 and PM2.5
may suffer from instability due to collinearity. Further, the lag structure of PM10-2.5 effects
on risk of cardiovascular hospital admissions and ED visits, as well as mortality, has not
been examined in detail.
Several epidemiologic studies of cardiovascular endpoints including HRV, BP,
ventricular arrhythmia, and ECG changes indicating ectopy or ischemia were conducted
since publication of the 2004 PM AQCD. Supraventricular ectopy and ST-segment
depression were associated with PM10-2.5 (Section 6.2.3.1) and the only study to examine the
effect of PM10-2.5 on BP reported a decrease in SBP (Ebelt et al., 2005, 056907) (Section
6.2.5.1). HRV findings were mixed across the epidemiologic studies (Section 6.2.1.1). A
limited number of studies have evaluated the effect of controlled exposures to PM10-2.5 CAPs
on cardiovascular endpoints in human subjects. These studies have provided some evidence
of decreases in HRV (SDNN) and tPA concentration among healthy adults approximately
20 hours following exposure (Section 6.2.1.2). However, it is important to note that no other
measures of HRV (e.g., LF, HF, or LF/HF), nor other hemostatic or thrombotic markers (e.g.,
fibrinogen) were significantly affected by particle exposure in these studies.
There are very few toxicological studies that examined the effect of exposure to
PM10-2.5 on cardiovascular endpoints or biomarkers in animals. The only studies that
evaluated cardiovascular responses were comparative studies of various size fractions and
only blood or plasma parameters were measured (Sections 6.2.7.3 and 6.2.8.3). These
studies used IT instillation methodologies, as there are challenges to exposing rodents via
inhalation to PMio-2,5, due to near 100% deposition in the ET region for particles >5 |Jm
(Raabe et al., 1988, 001439) and only 44% nasal inhalability of a 10 |Jm particle in the rat
(Menache et al., 1995, 006533). These studies also employed relatively high doses of
PM10-2.5. Despite these shortcomings, increased plasma fibrinogen was observed and the
response was similar to that observed with PM2.5. At this time, evidence of biological
plausibility for cardiovascular morbidity effects following PM10-2.5 exposure is sparse, due to
July 2009
6-131
DRAFT-DO NOT CITE OR QUOTE

-------
the small number of studies, few endpoints examined, and the limitations related to the
interpretation of IT exposures.
In summary, several epidemiologic studies report associations with cardiovascular
endpoints including IHD hospitalizations, supraventricular ectopy, and changes in HRV. In
studies examining multiple size fractions, most of the effect estimates observed were of a
similar magnitude (e.g., the PM2.5 effect was not larger in all studies). Further, dust storm
events resulting in high concentrations of crustal material are linked to increases in
cardiovascular disease hospital admissions or ED visits for cardiovascular diseases. A large
proportion of inhaled coarse particles in the 3-6 |um (dae) range can reach and deposit in the
lower respiratory tract, particularly the TB airways (see Figures 4-3 and 4-4). The few
toxicological and controlled human exposure studies examining the effects of PM10-2.5
provide limited evidence of cardiovascular effects and biological plausibility to support the
epidemiologic findings. Therefore the available evidence is suggestive of a causal relationship
between PM10-2.5 exposures and cardiovascular morbidity.
6.2.12.3. Ultrafine PM
There was very little evidence available in the 2004 PM AQCD (U.S. EPA, 2004,
056905) on the cardiovascular effects of ultrafine PM. Findings from one study presented in
the 2004 PM AQCD (U.S. EPA, 2004, 056905) of controlled exposures to ultrafine elemental
carbon suggested no particle-related effects on various cardiovascular endpoints including
blood coagulation, HRV, and systemic inflammation. No epidemiologic studies of short-term
ultrafine particle concentration and cardiovascular endpoints were included in the 2004
AQCD and there were no relevant toxicological studies reviewed in the 2004 PM AQCD
(U.S. EPA, 2004, 056905) that exposed animals to ultrafine PM. A small number of new
epidemiologic studies, as well as several controlled human exposure and toxicological
studies have been conducted in recent years, but substantial uncertainties remain as to the
cardiovascular effects of ultrafine PM. For a given mass, the enormous number and large
surface area of UFPs highlight the importance of considering the size of the particle in
assessing response. For example, UFPs with a diameter of 20 nm, when inhaled at the
same mass concentration, have a number concentration that is approximately 6 orders of
magnitude higher than for a 2.5*|Jm diameter particle. Particle surface area is also greatly
increased with ultrafine PM. Many studies suggest that the surface of particles or
substances released from the surface (e.g., transition metals, organics) interact with
July 2009
6-132
DRAFT-DO NOT CITE OR QUOTE

-------
biological substrates, and that surface-associated free radicals or free radical-generating
systems may be responsible for toxicity, resulting in greater toxicity of ultrafine PM per
particle surface area than larger particles. Additionally, smaller particles may have greater
potential to cross cell membranes and epithelial barriers.
Controlled human exposure studies are increasingly being utilized to evaluate the
effect of ultrafine PM on cardiovascular function. While the number of studies of exposure
to UFPs per se is still limited, there is a large body of evidence from exposure to fresh DE,
which is typically dominated by UFPs. As described under the summary for PM2.5, studies
of controlled exposures to DE (100-300 jug/m3) have consistently demonstrated effects on
vasomotor function among adult volunteers (Section 6.2.4.2). In addition, exposure to
ultrafine EC (50 |ug/m3, particle count 10.8 x 106/cm3) was recently shown to attenuate FMD
(Shah et al., 2008, 156970). Changes in vasomotor function have been observed in animal
toxicological studies of ultrafine PM, although very few studies have been conducted
(Section 6.2.4.3). Inhaled ultrafine TiCh impaired arteriolar dilation when compared to fine
Ti02 at similar mass doses (Nurkiewicz et al., 2008, 156816). This response may have been
due to ROS in the microvascular wall, which may have led to consumption of endothelial -
derived NO and generation of peroxynitrite radicals. Support for an ultrafine PM effect on
altered vascular reactivity is also provided by studies of DE and IT exposure to ambient
PM. The response to DE did not appear to be due to VOCs. One epidemiologic study showed
that particle number count was associated with a nonsignificant decrease in flow- and
nitroglycerine-mediated reactivity as measures of vasomotor function in diabetics living in
Boston (O'Neill et al., 2005, 088423).
New studies have reported increases in markers of systemic oxidative stress in
humans following controlled exposures to different types of PM consisting of relatively high
concentrations of UFPs from sources including WS, urban traffic particles, and DE (Section
6.2.9.2).	Increased cardiac oxidative stress has been observed in mice and rats following
gasoline exhaust exposure and it appeared the effect was particle-dependent (Section
6.2.9.3).
The associations between ultrafine PM and HRV measures in epidemiologic studies
include increases and decreases (Section 6.2.1.1), providing some evidence for an effect.
Exposure to ultrafine CAPs has been observed to alter parameters of HRV in controlled
human exposure studies, although this effect has been variable between studies (Section
6.2.1.2). Alterations in HR, HRV, and BP were reported in rats exposed to <200 (Jg/m3
July 2009
6-133
DRAFT-DO NOT CITE OR QUOTE

-------
ultrafine CB (<1.6 x 107 particles/cm3) (Sections 6.2.1.3 and 6.2.5.3). The effects of ultrafine
PM on BP have been mixed in epidemiologic studies (Section 6.2.5.1).
There is some evidence of changes in markers of blood coagulation in humans
following controlled exposure to ultrafine CAPs, as well as WS and DE; however, these
effects have not been consistently observed across studies (Section 6.2.8.2). Toxicological
studies demonstrate mixed results for systemic inflammation and blood coagulation as well
(Sections 6.2.7.3 and 6.2.8.3).
Few time-series studies of cardiovascular hospital admissions have evaluated
ultrafine PM. The SOPHIA study found no association between any outcome studied (all
CVD, dysrhythmia, CHF, IHD, peripheral vascular and cerebrovascular disease) and 24-h
mean levels of UFP (Metzger et al. 2004). The median UF particle count in Atlanta during
the study period was 25,900 particles/cm3. UFP were not associated with CVD
hospitalizations in the elderly in Copenhagen, Denmark, but were associated with cardiac
readmission or fatal MI in the European HEAPSS study (Section 6.2.10). In the
Copenhagen study, the mean count of particles with a 100 nm mean diameter was 0.68 x
104 particles/cm3, whereas the PNC range was approximately 1.2-7.6 x 104 particles/cm3 in
HEAPSS study. Spatial variation in UFP concentration, which diminishes within a short
distance from the roadway, may introduce exposure measurement error, making it more
difficult to observe an association if one exists.
A limited number of epidemiologic studies have evaluated subclinical cardiovascular
measures and a number of these were conducted in Boston. Ultrafine PM has been linked to
ICD-recorded arrhythmias in Boston and supraventricular ectopic beats in Erfurt, Germany
(Section 6.2.2.1). One study reported no UFP association with ectopy (Barclay et al., 2009,
179935). ST-segment depression in subjects with stable coronary heart disease was
associated with ultrafine PM in Helsinki (Section 6.2.3.1). The limited number of studies
that examine this size fraction makes it difficult to draw conclusions about these
cardiovascular measures.
In summary, there is a large body of evidence from controlled human exposure studies
of fresh DE, which is typically dominated by UFPs, demonstrating effects of UFP on the
cardiovascular system. In addition, cardiovascular effects have been demonstrated by a
limited number of laboratories in response to ultrafine CB, urban traffic particles and
CAPs. Responses include altered vasomotor function, increased systemic oxidative stress
and altered HRV parameters. Studies using ultrafine CAPs, as well as WS and DE, provide
July 2009
6-134
DRAFT-DO NOT CITE OR QUOTE

-------
some evidence of changes in markers of blood coagulation, but findings are not consistent.
Toxicological studies conducted with ultrafine TiCh, CB, and DE demonstrate changes in
vasomotor function as well as in HRV. Effects on systemic inflammation and blood
coagulation are less consistent. PM-dependent cardiac oxidative stress was noted following
exposure to gasoline exhaust. The few epidemiologic studies of ultrafine conducted do not
provide strong support for an association of UFPs with effects on the cardiovascular system.
Based on the above findings, the evidence is suggestive of a causal relationship between ultrafine PM
exposure and cardiovascular effects.
6.3. Respiratory Effects
6.3.1. Respiratory Symptoms and Medication Use
The 2004 PM AQCD (U.S. EPA, 2004, 056905) presented evidence from epidemiologic
studies of increases in respiratory symptoms associated with PM, although this was not
supported by the findings of a limited number of controlled human exposure studies. Recent
epidemiologic studies have provided evidence of an increase in respiratory symptoms and
medication use associated with PM among asthmatic children, with less evidence of an
effect in asthmatic adults. The lack of an observed effect of PM exposure on respiratory
symptoms in controlled human exposure studies does not necessarily contradict these
findings, as very few studies of controlled exposures to PM have been conducted among
groups of asthmatic or healthy children.
6.3.1.1. Epidemiologic Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) concluded that the effects of PMio on
respiratory symptoms in asthmatics tended to be positive, although they were somewhat
less consistent than PMio effects on lung function. Most studies showed increases in cough,
phlegm, difficulty breathing, and bronchodilator use, although these increases were
generally not statistically significant for PMio. The results from one study of respiratory
symptoms and thoracic coarse particles (Schwartz and Neas, 2000, 007625) found a
statistically significant association with cough with PM10-2.5. The results of two studies
examining respiratory symptoms and PM2.5 revealed slightly larger effects for PM2.5 than
for PMio.
July 2009
6-135
DRAFT-DO NOT CITE OR QUOTE

-------
Asthmatic Children
Two large, longitudinal studies in urban areas of the U.S. investigated the effects of
ambient PM on respiratory symptoms and/or asthma medication use with similar analytic
techniques (i.e., multistaged modeling and generalized estimating equations [GEE]): the
Childhood Asthma Management Program (CAMP) (Schildcrout et al., 2006, 089812) and the
National Cooperative Inner-City Asthma Study (NCICAS) (Mortimer et al., 2002, 030281).
A number of smaller panel studies conducted in the U.S. evaluated the effects of ambient
PM concentrations on respiratory symptoms and medication use among asthmatic children
(Delfino et al., 2002, 093740; Delfino et al., 2003, 090941; Delfino et al., 2003, 050460; Gent
et al., 2003, 052885; Gent et al., 2009, 180399; 2006, 088031; Slaughter et al., 2003,
086294).
July 2009
6-136
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Location
Mar et al. (2004,057309)	Spokane, WA	0
Mar et al. (2004,057309)	Spokane, WA	0
Rodriguez et al. (2007, 092842)	Perth, Australia	0-5
Ranzi et al. (2004,089500)	EmiliaRomagna,	0-3
Aekplakorn et al. (2003,089908)	Thailand	0
Gent et al. (2009,180399)	New Haven, CT	0-2
Mar et al. (2004,057309)	Spokane, WA	0
Rodriguez et al. (2007, 092842)	Perth, Australia	0-5
Gent et al. (2009,180399)	New Haven, CT	0-2
Gent et al. (2009,180399)	New Haven, CT	0-2
OConnor et al. (2008,156818)	Multicity ¦ US	0-4
Mar et al. (2004,057309)	Spokane, WA	0
Mar et al. (2004,057309)	Spokane, WA	0
Rodriguez et al. (2007, 092842)	Perth, Australia	0-5
Gent et al. (2009,180399)	New Haven, CT	0-2
Mar et al. (2004,057309)	Spokane, WA	0
Aekplakorn et al. (2003,089908)	Thailand	0
Aekplakorn et al. (2003,089908)	Thailand	0
Mar et al. (2004,057309)	Spokane, WA	0
Gent et al. (2009,180399)	New Haven, CT	0-2
Rabinovitch et al. .(2006,088031)	Denver, CO	1
Slaughter et al. (2003, 086294)	Seattle, WA	1
Rabinovitch et al. (2004,096753)	Denver, CO	0-2
Slaughter et al. (2003, 086294)	Seattle, WA	1
Effect Estimates
p Cough
-A— Wheeze
Shortness of Breath
j
~l	
Shortness of Breath
-i—~— Chest Tightness
l 9	 Any Symptoms
LRS
-f- Medication Use
Asthma Exacerbation
TM7
Maret al. (2004,
Maret al. (2004,
Aekplakorn et al.
Maret al. (2004,
Maret al. (2004,
Maret al. (2004,
Maret al. (2004,
Aekplakorn et al.
Aekplakorn et al.
Maret al. (2004,
057309)
057309)
(2003, 089908)
057309)
057309)
057309)
057309)
(2003, 089908)
(2003, 089908)
057309)
Spokane, WA
Spokane, WA
Thailand
Spokane, WA
Spokane, WA
Spokane, WA
Spokane, WA
Thailand
Thailand
Spokane, WA
Runny Nose, Phlegm
Cough
Wheeze
I* Shortness of Breath
-r*— Any Symptoms
URS
LRS
LRS
TM7
Maret al. (2004,057309)	Spokane, WA	0
Mar et al. (2004,057309)	Spokane, WA	0
Aekplakorn et al. (2003,089908)	Thailand	0
Just et al. (2002,035429)	Paris, France	0-4
Jalaludin et al. (2004,056595)	Australia	0-5
Mar et al. (2004,057309)	Spokane, WA	0
Jalaludin et al. (2004,056595)	Australia	0-5
Mar et al. (2004,057309)	Spokane, WA	0
Mar et al. (2004,057309)	Spokane, WA	0
Delfino et al. (2003, 050460)	California	0
Mortimer et al. (2002, 030281)	Multicity, US	1-2
Rabinovitch et al. (2004,096753)	Denver, CO	0-3
Rabinovitch et al. (2004,096753)	Denver, CO	0-3
Mar et al. (2004,057309)	Spokane, WA	0
Delfino et al. (2002, 093740)	California	0-2
Delfino et al. (2002, 093740)	California	0-2
Delfino et al. (2002, 093740)	California	0-2
Schildcrout et al. (2006, 089812)	US and Canada	0-2
Aekplakorn et al. (2003,089908)	Thailand	0
Aekplakorn et al. (2003,089908)	Thailand	0
Just et al. (2002,035429)	Paris, France	0-4
Mar et al. (2004,057309)	Spokane, WA	0
Rabinovitch et al. (2004,096753)	Denver, CO	0-3
Jalaludin et al. (2004,056595)	Australia	0-5
Slaughter et al. (2003, 086294)	Seattle, WA	0
Schildcrout et al. (2006, 089812)	US and Canada	0-2
Rabinovitch et al. (2004,096753)	Denver, CO	0-3
Just et al. (2002,035429)	Paris, France	0-4
Slaughter et al. (2003, 086294)	Seattle, WA	0
—Runny Nose, Phlegm
Cough
Wheeze

Shortness of Breath
- Score > 1, Any Symptom
Current Dar-
Previous Night
1-h Max
8-h Max
24-h Avg
URS
- LRS
j-»- LRS
Medication Use
1 Asthma Exacerbation
0.1
~~r~
0.6
~I
1.1
I
1.6
~~r~
2.1
I
2.6
Figure 6-7. Respiratory symptoms and/or medication use among asthmatic children following acute
exposure to PM2.5. ORs and 95% CIs standardized to increments of 10 |jg/m3
July 2009
6-137
DRAFT-DO NOT CITE OR QUOTE

-------
In the CAMP study, the association between ambient air pollution and asthma
exacerbations in children (n = 990) from eight North American cities was investigated
(Schildcrout et al., 2006, 089812). In contrast to several past studies (Delfino et al., 1996,
080788; 1998, 051406). no associations were observed between PMio and asthma
exacerbations or medication use. PMio concentrations were measured on less than 50% of
study days in all cities except Seattle and Albuquerque. While PMio effects were not
observed for the entire panel of children, they were observed in recent reports on the
children participating at the Seattle center (Slaughter et al., 2003, 086294; Yu et al., 2000,
013254). In a smaller panel study of asthmatic children (n = 133) enrolled in the CAMP
study, daily particle concentrations averaged over 3 central sites in Seattle was used as the
exposure metric (Slaughter et al., 2003, 086294). Children were followed for 2 months, on
average. Daily health outcomes included both a 3-category measure of asthma severity
based on symptom duration and frequency, and inhaled albuterol use. In single-pollutant
models, an increased risk of asthma severity was associated with a 10 |u,g/m3 increase in lag
1 PM2.5 (OR 1.20 [95% CP 1.05-1.37]) and with a 10 |ug/m3 increase in lag 0 PMio (OR 1.12
[95% CP 1.05-1.22]). In copollutant models with CO, the associations remained (OR for
PM2.5 1.16 [95% CP 1.03-1.30]; OR for PMio 1.11 [95% CP 1.03-1.19]). Associations between
inhaler use and PM were positive in single-pollutant models (RR lag 1 PM2.5 1.08 [95% CP
1.01-1.15]; RR lag 0 PMio 1.05 [95% CP 1.00-1.09]), but attenuated and no longer
statistically significant in copollutant models.
The eight cities included in the NCICAS (Mortimer et al., 2002, 030281) were all in
the East or Midwest: New York City (Bronx, E. Harlem), Baltimore, Washington DC,
Cleveland, Detroit, St. Louis, and Chicago. In this study, 864 asthmatic children, aged 4-9
yr, were followed daily for four 2~wk periods over the course of nine months. Morning and
evening asthma symptoms (analyzed as none vs. any) and peak flow were recorded. For the
three urban areas with data, each 10 |u,g/m3 increase in the mean of the previous 2 days (lag
1-2) PMio, increased the risk for morning asthma symptoms (OR 1.12 [95% CP 1.00-1.26]).
This effect was robust to the inclusion of O3 (OR 1.12 [95% CP 0.98-1.27]), though
attenuated in models including O3, SO2, and NO2 (OR 1.07 [95% CP 0.89-1.22]). In a related
study, O'Connor et al. (2008, 156818) examined the relationship between short-term
fluctuations in outdoor air pollutant concentrations and changes in pulmonary function and
respiratory symptoms among children with asthma in 7 U.S. inner-city communities. PM2.5
concentration was not significantly associated with respiratory symptoms in this study.
July 2009
6-138
DRAFT-DO NOT CITE OR QUOTE

-------
Table 6-7. Characterization of ambient PM concentrations from studies of respiratory morbidity and
short-term exposures in asthmatic children and adults. All concentrations are for the 24-
h avg unless otherwise noted.
Reference
Location
Mean Concentration
(|jg/m3>
Upper Percentile
Concentrations (|jg/m3)
PMzs
Adamkiewicz et al. (2004, 087925)
Steubenville, OH
20.43
75th: 23
98th: 51.79
Max: 51.79
Adaretal. (2007, 001458)
St. Louis, MO
10.13
98th: 22.43
Max: 23.24
Aekplakorn et al. (2003, 089908)
North Thailand

Max: 24.8-26.3
Allen et al. (2008,156208)
Seattle, WA
11.2

Barraza-Villarreal et al. (2008,156254)
Mexico City
8-h max: 28.9
Max: 102.8
Bourotte et al. (2007,150040)
Sao Paulo, Brazil
11.9
Max: 26.6
de Hartog et al. (2003, 001061)
Multicity, Europe
12.8-23.4
Max: 39.8-118.1
Delfino et al. (2006,090745)
Southern CA
3.9-6.9
Max: 8.8-11.6
DeMeo et al. (2004, 087346)
Boston, MA
10.8

Ebelt et al. (2005,056907)
Vancouver, Canada
11.4
98th: 23
Max: 28.7
Ferdinands et al. (2008,156433)
Atlanta, GA
27.2
Max: 34.7
Fischer et al. (2007,156435)
The Netherlands
56
75th: 187
Gent et al. (2003, 052885)
CT&MA
13.1
60th: 12.1
80th: 19.0
Gent et al. (2009,180399)
New Haven, CT
17.0

Girardot et al. (2006, 088271)
Smoky Mountains
13.9
Max: 38.4
Hogervorst et al. (2006,189460)
The Netherlands
19.0

Hong et al. (2007, 091347)
Incheon City, Korea
20.27
Max: 36.28
Jansen et al. (2005, 082236)
Seattle, WA
14.0
Max: 44

Johnston et al. (2006, 091386)
Darwin, Australia
11.1
Max: 36.5
Koenig et al. (2003,156653)
Seattle, WA
13.3
Max: 40.4
Laporio et al. (2006, 089800)
Rome, Italy
27.2
Max: 100
Lee et al. (2007, 093042)
Seoul, South Korea
51.15
75th: 87.54
Max: 92.71
Lewis et al. (2004, 097498)
Detroit, Ml
15.7-17.5
Max: 56.1
Liu et al. (2009,192003)
Windsor, Ontario
7.1
95th: 19.0
98th: 19.0.
Maretal. (2004,057309)
Spokane, WA
8.1-11.0

Maretal. (2005, 088759)
Seattle, WA
5-26

McCreanor et al. (2007, 092841)
London, England
1-h avg: 11.9-28.3
1-h max: 55.9-76.1
Moshammer et al. (2006, 090771)
Linz, Austria
8-h avg: 15.70
Max 24-h avg: 76.39
Murata et al. (2007,156787)
Tokyo, Japan
39.0
Max 1-h avg: 120
July 2009
6-139
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Location
Mean Concentration
(|jg/m3>
Upper Percentile
Concentrations (|jg/m3)
O'Connor et al. (2008,156818)
Multicity, U.S.
14
Max: 35
Peledetal. (2005,156015)
Multicity, Israel
23.9-29.2

Penttinen et al. (2006, 087988)
Helsinki, Finland
8.37
75th: 11.15
Max: 33.53
Rabinovitch et al. (2004, 096753)
Denver, CO
10.8
98th: 29.3
Max: 53.5
Rabinovitch et al. (2006, 088031)
Denver, CO
10.8
98th: 23.4
Ranzi et al. (2004, 089500)
Emilia-Romagna, Italy
Urban: 53.07
Rural: 29.11

Rodriguez et al. (2007, 092842)
Perth, Australia
1-h avg: 20.8
24-h avg: 8.5
Max 1-h avg: 93.4
Max 24-h avg: 39.4
Slaughter et al. (2003, 086294)
Seattle, WA
7.3"
75th: 11.3
Strand et al. (2006,157017)
Denver, CO
12.7

Timonen et al. (2004, 087915)
Multicity, Europe
12.7-23.1
Max: 39.8-118.1
Trenqa et al. (2006,155209)
Seattle, WA
8.6-9.6"
75th: 13.1-14.8
Max: 40.4-41.5
von Klot et al. (2002, 034706)
Erfurt, Germany
30.3b
75th: 41.3b
Max: 133.8b
Ward et al. (2002, 025839)
Birmingham and Sandwell, U.K.
12.3-12.7
Max: 28-37
PM 102.5
Aekplakorn et al. (2003, 089908)
North Thailand
NR
NR
Bourotte et al. (2007,150040)
Sao Paulo, Brazil
21.7
Max: 62.0
Ebelt et al. (2005,056907)
Vancouver, Canada
5.6
Max: 11.9
Laqorio et al. (2006, 089800)
Rome, Italy
15.6
Max: 39.6
Maretal. (2004,057309)
Spokane, WA
8.7-13.5

von Klot et al. (2002, 034706)
Erfurt, Germany
10.3
75th: 14.6
Max: 64.3
PM10
Aekplakorn et al. (2003, 089908)
North Thailand
31.9-37.5
Max: 113.3-153.3
Boezen et al. (2005, 087396)
The Netherlands
26.6-44.1
Max: 89.9-242.2
de Hartoq et al. (2003, 001061)
Multicity, Europe
19.6-36.5
Max: 67.4-112.0
Delfino et al. (2002,093740)
Alpine, CA
20
90th: 32
Max: 42
Delfino et al. (2003,050460)
Los Angeles, CA
59.9
90th: 86/0/Max: 126
Delfino etal. (2004,056897)
Alpine, CA
29.7
90th: 40.9
Max: 50.7
Delfino etal. (2006,090745)
Southern CA
35.7-70.8
Max: 105.5-154.1
Desquevroux et al. (2002, 026052)
Paris, France
23-28
Max: 63-84
Ebelt etal. (2005,056907)
Vancouver, Canada
17
Max: 36
Honq et al. (2007, 091347)
Incheon City, Korea
35.3
Max: 124.87
Jalaludin et al. (2004, 056595)
Sydney, Australia
22.8
75th: 122.8
Jansen et al. (2005, 082236)
Seattle, WA
18.0
Max: 51
Johnston et al. (2006, 091386)
Darwin, Australia
20
Max: 43.3
July 2009
6-140

DRAFT-DO NOT CITE OR QUOTE

-------
Reference
, .. Mean Concentration
Location . , ,,
Upper Percentile
Concentrations (|jg/m3)
Just et al. (2002, 035429)
Paris, France
23.5
Max: 44.0
Laporio et al. (2006, 089800)
Rome, Italy
42.8
Max: 123
Laurent et al. (2008,156672)
Strasbourg, France
20.8
Max: 106.3
Lee et al. (2007, 093042)
Seoul, South Korea
71.40
75th: 87.54
Max: 148.34
Maretal. (2004,057309)
Spokane, WA
16.8-24.5

Mortimer et al. (2002, 030281)
Multicity, UC
53

Moshammer et al. (2006, 090771)
Linz, Austria
8-h avg: 24.85
Max 24-h: 76.39
Odajima et al. (2008,192005)
Fukuoka, Japan
3-havg: 32.6-41.5
Max 3-havg: 126.0-191.3
Peacock et al. (2003, 042026)
Southern England
21.2
Max: 87.9
Peledetal. (2005,156015)
Multicity, Israel
31.0-67.1

Preutthipan et al. (2004, 055598)
Bangkok, Thailand
111.0
Max: 201
Rabinovitch et al. (2004, 096753)
Denver, CO
28.1
Max: 102.0
Seqala et al. (2004, 090449)
Paris, France
24.2
Max: 97.4
Schildcrout et al. (2006, 089812)
Multicity, U.S.
17.7-32.4"
75th: 26.2-42.7
90th: 32.5-53.9
Slaughter et al.(2003, 086294)
Seattle, WA
21.0"
75th: 29.3
Steinvil et al. (2008,188893)
Tel Aviv, Israel
64.5
75th: 60.7
von Klot et al. (2002, 034706)
Erfurt, Germany
45.4
75th: 59.7
Max: 172.4
8Median
Includes UFP, for complete information on number concentration from this study, please see corresponding table in Annex E.
Mar et al. (2004, 057309) studied asthmatic children (n = 9) in Spokane, WA.
Increases in 0, 1 or 2 day lags of each of the PM size classes studied were associated with
cough. When all lower respiratory tract symptoms (wheeze, cough, shortness of breath,
sputum production) were grouped together, positive associations were reported for each
10 |ug/m3 increase in same-day PMio (OR 1.07 [95% CI: 1.00-1.14]), or lag 0 or lag 1 PM2.5
(OR 1.18 [95% CI: 1.00-1.38]; OR 1.21 [95% CI: 1.00-1.46], respectively), and 10 |ug/m3
increase in lag 0 and lag 1 PM1.0 (OR 1.21 [95% CI: 1.01-1.44]; OR 1.25 [95% CI: 1.01-1.55],
respectively). No associations were reported for PM10-2.5 and grouped lower respiratory tract
symptoms (Mar et al., 2004, 057309).
Gent et al. (2003, 052885) reported on daily symptom and medication use during one
summer for 271 asthmatic children living in southern New England. In single-pollutant
models for users of maintenance medication (n = 130), PM2.5 > 19 |ug/m3 lagged by 1 day was
associated with a 10 to 25% increase in risk of symptoms compared to PM2.5 <6.9 |u.g/m3: OR
for persistent cough 1.12 (95% CI: 1.02-1.24); OR for chest tightness 1.21 (95% CI:
1.00-1.46); OR for shortness of breath 1.26 (95% CI: 1.02-1.54). Effects were attenuated in
July 2009
6-141
DRAFT-DO NOT CITE OR QUOTE

-------
models including O3 (OR for persistent cough 1.00 95% CP 0.88-1.15]; OR for chest
tightness 0.91 [95% CP 0.71-1.17]; OR for shortness ofbreath 1.20 [95% CI: 0.94-1.52]). No
statistical associations between ambient particle exposure and respiratory health were
found for asthmatic children not on maintenance medication.
Annual PM2.5 levels at monitoring sites in New Haven, CT exceed the annual
standard of 15 |ug/m3. Gent et al. (2009, 180399) conducted a study here to examine the
associations between daily exposure to PM2.5 components and sources identified through
source apportionment, and daily symptoms and medication use in asthmatic children.
Asthmatic children (n = 149) aged 4-12 yr were enrolled in the study between 2000 and
2003. Factor analysis was used to identify six sources of PM2.5 (motor vehicle, road dust,
sulfur, biomass burning, oil, and sea salt). Total PM2.5 was not associated with any
symptoms or medication use, however trace elements originating from motor vehicle, road
dust, biomass burning and oil sources were associated with symptoms and/or medication
use. For example, an increased risk of wheeze, shortness ofbreath, chest tightness or short-
acting inhaler use was associated with increasing EC mass concentration. Risks remain in
models that include all six PM2.5 sources as well as NO2, which may be considered a marker
for traffic. NO2 was found to be an independent risk factor for increased wheeze.
Two panel studies were conducted over the course of three winters at a school in
Denver (Rabinovitch et al., 2004, 096753; Rabinovitch et al., 2006, 088031). In the first
report, approximately 86 different children contributed data on asthma symptoms and
medication use over three consecutive winters (Rabinovitch et al., 2004, 096753). The
exposure metric was a 3-day ma of PM2.5 measured at a site located next to the school for
the first 2 winters and from a central site located 4.8 km (3 miles) away for the third. A
strong correlation was observed during the first two winters between PM2.5 values
measured locally and at a downtown monitoring station (Pearson product-moment
correlation = 0.93) and between PM10 values measured locally and at a downtown
monitoring station (correlation = 0.84). Therefore, in year 3, all ambient data were collected
from nearby community monitoring stations. No statistical associations were found
between asthma symptoms or medication use and PM. Rabinovitch et al. (2006, 088031)
enrolled a panel of 73 children and evaluated associations with morning maximum PM2.5
measured at the central site. PM measurements were available hourly from 2 co-located
monitors, an FRM and a TEOM monitor. Each 10 ug/ma increase in morning maximum 1-h
PM2.5 concentration was associated with an increased likelihood of rescue medication use
July 2009
6-142
DRAFT-DO NOT CITE OR QUOTE

-------
(OR for FRM exposure data 1.02 [95% CI: 1.01-1.03]; OR for TEOM 1.03 [95% CI: 1.00-1.6]).
Interestingly, the association between inhaler use and particle exposure was not evident
when the 24-h avg PM2.5 was used in the model.
Two smaller panel studies enrolling asthmatic children conducted by Pel fino et al.
(2002, 093740; 2003, 050460) in southern California examined the health effects of different
averaging times for PM10 (l-h, 8-h, 24-h) (Delfino et al., 2002, 093740), and 24-h avg of two
PM10 components (EC and OC) (Delfino et al., 2003, 050460). In the first study, 22 children
living in a "lower" pollution area were followed daily for two months in spring. In contrast
with Gent et al. (2003, 052885), positive statistical associations with asthma symptoms
(measured on a 6-point severity scale) were found only for the children not taking
anti-inflammatory medication. For these 12 children, in single-pollutant models each
10 ug/m3 increase in lag 0 l*h max PM10 nearly doubled the risk of clinically meaningful
symptoms (i.e., an asthma symptom score > 3) (OR 1.14 [95% CP 1.04-1.24]) and each
10 jug/m3 increase in 3"day avg 24-h PM10 increased the risk by 1.25 (95% CP 1.06-1.48). No
statistical associations were found between exposure to ambient particles and symptoms in
the 10 children who were taking anti-inflammatory medication. No multipollutant models
were reported. The second study enrolled 22 asthmatic children living in an area of higher
pollution. For children living in this community, each 10 ug/m3 increase in lag 0, 24-h PM10
was associated with an increased risk of asthma symptom score >P OR 1.10, (95% CP
1.03-1.19) (Delfino et al., 2003, 050460). The correlation among PM10, EC and OC was
substantial: 0.80 between PM10 and either EC or OC, and 0.94 between EC and OC.
Associations between EC or OC and asthma symptoms were very similar to those for PMio:
each 3 |u,g/m3 increase in lag 0, 24-h EC or 5 |ug/m3 increase in lag 0, 24-h OC was associated
with an increased risk of asthma symptoms (OR 1.85 [95% CP 1.11-3.08] or OR 1.88 [95%
CP 1.12-3.17], respectively) (Delfino et al., 2003, 050460).
Studies from Australia (Rodriguez et al., 2007, 092842), Europe (Laurent et al., 2008,
156672; Ranzi et al., 2004, 089500; Laurent et al., 2009, 192129), and Asia (Aekplakorn et
al., 2003, 089908) provide additional evidence of an association between ambient PM and
respiratory symptoms and/or medication use among asthmatic children. Two studies
(Jalaludin et al., 2004, 056595; Just et al., 2002, 035429) found no association between
ambient PM levels and these health endpoints.
July 2009
6-143
DRAFT-DO NOT CITE OR QUOTE

-------
Asthmatic Adults
Since the 2004 PM AQCD (U.S. EPA, 2004, 056905), one U.S. and several European
studies have investigated the effects of ambient PM levels on respiratory symptoms and
medication use among asthmatic adults. The respiratory symptom and medication use
results from these studies are summarized by particle size and displayed in Table 6.7 and
Figure 6.8. Relatively few studies examined these effects in healthy adults, and they did not
identify a relationship between ambient PM levels and respiratory symptoms or medication
use. These studies of healthy adults are summarized in Annex E, but will not be described
in detail in this section.
Mar et al. (2004, 057309) studied asthmatic adults (N = 16) in Spokane, WAover a
3-yr time period. No associations were found between PM and respiratory symptoms among
the adults.
Several panel studies conducted in Europe have examined effects of daily exposures to
air pollution on adults with asthma, including studies in the Pollution Effects on Asthmatic
Children in Europe (PEACE) study (Boezen et al., 2005, 087396), Exposure and Risk
Assessment for Fine and UFPs in Ambient Air (ULTRA) study (De Hartog et al., 2003,
001061). in Germany (Von Klot et al., 2002, 031706). and in Paris (Desqueyroux et al., 2002,
026052; 2004, 090449). Boezen et al. (2005, 087396) enrolled 327 elderly adults in the
Netherlands to examine the role of AHR and IgE levels in susceptibility to air pollution. For
subjects with both AHR (defined as > 20% FEVi decline at < 2 mg cumulative methacholine)
and high total IgE (>20 kU/L), each 10 |ug/m3 increase in lag 2 PMio concentration was
associated with an increased risk of upper respiratory symptoms (URS) among males (OR
1.06 [95% CP 1.02-1.10]), and at lag 0 with increased cough among females (OR 1.04 [95%
CP 1.00-1.08]). Each 10 |ug/m3 increase in BS at lag 0, lag 1, and the 5-day mean was
associated with URS and cough among males. The strongest association in both cases was
for the 5"day mean (OR for URS 1.43 [95% CI: 1.20-1.69]; OR for cough 1.16 [95% CI:
1.05-1.29]). The authors suggest that the sex differences observed maybe explained by
differential daily exposure to traffic exhaust experienced by men compared to women
(Boezen et al., 2005, 087396).
As part of the multicenter ULTRA study, de Hartog et al. (2003, 001061) enrolled 131
older adults with coronary artery disease in three cities (Amsterdam, Erfurt [Germany],
and Helsinki). Pooling data from all three cities, associations were observed between PM2.5
and shortness of breath and phlegm: each 10 |jg/m3 increase in the 5-day avg PM2.5 was
July 2009
6-144
DRAFT-DO NOT CITE OR QUOTE

-------
associated with an increased risk of symptoms (OR for shortness of breath 1.12 [95% CI:
1.02-1.24]; OR for phlegm 1.16 [95% CI: 1.03-1.32]). Unlike fine particles, UFPs were not
consistently associated with symptoms.
In a study that took place in Erfurt, Germany, von Klot et al. (2002, 034706) examined
daily, winter time exposure to ambient PM10-2.5, PM2.5-0.01 and PM0.1-0.01 particles and
respiratory health effects in 53 P2"agonists or inhaled corticosteroids and exposure to
particles in single and multipollutant models. Particle exposure metrics examined included
same-day, 5-day and 14-day avg. No effects were observed for wheeze and exposure to
PM10-2.5 for any averaging time. The strongest association between wheeze and exposure to
UFPs was for a 14-day avg: each 7,700 increase in the NCo.01-0.1 increased the risk of wheeze
by 27% (OR 1.27 [95% CP 1.13-1.43]). The effect was attenuated in copollutant models that
also included PM2.5-0.01 (OR 1.12 [95% CI: 1.01-1.24]), NO2 (OR 1.12 [95% CI: 0.99-1.26]), CO
(OR 1.05 [95% CI: 0.92-1.19]) or SO2 (OR 1.14 [95% CI: 1.04-1.26]). The correlations
between UFPs and two gaseous pollutants, NO2 and CO, were high: 0.66 for each.
July 2009
6-145
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Location
Lao
Effect Estimate
de Hartoq et al. (2003, 001061)
Netherlands
0-4
Phlfir)m 1 T PM?s
Maretal. (2004,057309)
Spokane, WA
0

Maretal. (2004,057309)
Spokane, WA
0

Maretal. (2004,057309)
Spokane, WA
0
„! . nniir|h
de Hartoq et al. (2003, 001061)
Netherlands
0-4
1
Maretal. (2004,057309)
Spokane, WA
0
1 _
Maretal. (2004,057309)
Spokane, WA
0
1
Johnston et al. (2006, 091386)
Australia
0-5
Asthma Symptoms
Maretal. (2004,057309)
Spokane, WA
0

Maretal. (2004,057309)
Spokane, WA
0
1 1 IRS
1
PM102.5 [
Maretal. (2004,057309)
Spokane, WA
0
m\ Runny Nose, Phlegm
Maretal. (2004,057309)
Spokane, WA
0

von Klot et al. (2002, 034706)
Germany
0-4
¦
Maretal. (2004,057309)
Spokane, WA
0
—'r—
Maretal. (2004,057309)
Spokane, WA
0

Maretal. (2004,057309)
Spokane, WA
0

Maretal. (2004,057309)
Spokane, WA
0
T ¦ Any Symptom
von Klot et al. (2002, 034706)
Germany
0-4
1
Maretal. (2004,057309)
Spokane, WA
0

PM10 ;
Maretal. (2004,057309)
Spokane, WA
0
ml Runny Nose, Phlegm
Maretal. (2004,057309)
Spokane, WA
0

Boezen et al. (2005, 087396)
Netherlands
0-4
_*L Cough
Seqala et al. (2004, 090449)
Paris, France
0
1
Maretal. (2004,057309)
Spokane, WA
0

Maretal. (2004,057309)
Spokane, WA
0

Maretal. (2004,057309)
Spokane, WA
0
¦B Shortness of breath
1
Desquevroux et al. (2002, 026052)
Paris, France
3-5

Johnston et al. (2006, 091386)
Australia
0-5
1
-f»-
Maretal. (2004,057309)
Spokane, WA
0

Boezen et al. (2005, 087396)
Netherlands
0-4
URS
¦
Maretal. (2004,057309)
Spokane, WA
0
-*-J-LRS
1	1	1	1	1	1
0,50 0.73 1.00 1.21 1.50 1.78
Figure 6-8. Respiratory symptoms and/or medication use among asthmatic adults following
acute exposure to particles. Summary of studies using 24-h avg of PMio, PM2.5, PM102.5.
ORs and 95% CIs were standardized to increments of 10 (Jg/m3.
In the same study, no association was found between exposure to thoracic coarse, fine
or UFPs and use of short-acting inhalers, though there was an association with
maintenance medication. Increased likelihood of maintenance medication was significantly
associated with PM of all sizes and all averaging times (same day, 5- and 14-day avg) and
gaseous copollutants in single or multipollutant models. The strongest effects were seen for
14-day avg of PM10-2.5 (for each 10 |u,g/m3 increase OR 1.43 [95% CP 1.28-1.60]), PM2.5-0.01 (for
July 2009
6-146
DRAFT-DO NOT CITE OR QUOTE

-------
each 20 |ug/m3 increase OR 1.54 [95% CI: 1.43-1.66]), NCo.oi-o.i (for each 7,700 increase OR
1.45 [95% CL 1.29-1.63]). For PM2.B-0.01, effects were unchanged in copollutant models,
including a model with UFPs. The authors conclude that this is evidence for independent
effects of fine and UFPs (Von Klot et al., 2002, 034706).
In Paris, Segala et al. (2004, 090449) recruited 78 adults from an otolaryngology clinic
and followed them for three months. Both PM10 and BS (which were very highly correlated
[r =.88]) were associated with cough: OR 1.24 (95% CI: 1.01-1.52) for a 10 |ug/m3 increase in
mean 0-4 day PM10 and OR 1.18 (95% CI: 1.02-1.39) for a 10 |ug/m3 increase in BS.
Also in Paris, 60 severe asthmatics were followed for 13 months and the relationship
between daily air quality (including 24-h PM10 as measured at the site nearest to the
subject's home) and asthma attack (defined as the need to increase rescue medication use
and one or more positive signs on clinical exam of wheezing, expiratory brake, thoracic
distention, hypertension with tachycardia, polypnea) were examined with GEE models
(Desqueyroux et al., 2002, 026052). Each 10 jug/m3 increase in PM10 increased the risk of
asthma attack, but only after lags of 3 to 5 days. The strongest effect was seen for the mean
lag of days 3 to 5 (OR 1.21 [95% CI: 1.04-1.40]). Effect sizes were larger among patients not
on regular oral steroid therapy: for PM10 lag 3-5 (OR 1.41 [95% CI: 1.15-1.73]). This effect
persisted in multipollutant models for winter time levels of PM10 and SO2 (OR 1.51 [95% CI:
1.20-1.90]) or NO2 (OR 1.43 [95% CI: 1.16-1.76]), but not in summer time models with O3
(OR 1.09 [95% CI: 0.71-1.67]).
Copollutant Models
A limited number of respiratory symptoms studies reported results of copollutant
models. Generally, the associations between respiratory symptoms and PM were robust to
the inclusion of copollutants (Figure 6"9), though Desqueyroux et al. (2002, 026052) indicate
the effects of PM may be potentiated by NO2 and SO2 during the winter months. Gent et al.
(2003, 052885) also reported the results of copollutant models, though the categorical
exposure groups used in the analysis did not allow these results to be included in Figure 6-
9. As reported above, the investigators found that effects were attenuated in models
including O3.
July 2009
6-147
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Outcome
Pollutant
Effect Estimate
Slauqhter et al. (2003, 08B294\
Asthma
severity
PM2.6
PM2.6+CO
PM2.5 Asthmatic Children
1 *
¦
1 a
Aekplakorn et al. (2003,0899081
Cough
PM2.6
PM2.6+SO2
1
1
1
1
1
1
1
1
Aekplakorn et al. (2003,0899081
Cough
PM 10-2.6
PM10.2.B+SO2
PM10 2.5 Asthmatiij Children
	1	m	
1
	1 U
1
1
1
1
Slauqhter et al. (2003, 08B294)
Asthma
severity
PM10
PM10+CO
PM10 Asthmatic Ohildren
1
1
1
Mortimer et al. (2002, 030281)
AM asthma
symptoms
PM10
PM10+O3
1
¦
1 •
1
Aekplakorn et al. (2003, 089908I
Cough
PM10
PM10+SO2
1 ¦
1
1
1
1*
1
1
1
Desqueyroux et al. (2002, Q2GQ52\
Asthma
PM10
PM10 Asthmatic Ajdults

attack

« „


PM10 + NO2
¦


PM10+O3
1


PM10+SO2
1
i	i	1	1
0.5	1.0	1.5	2.0
Figure 6-9. Respiratory symptoms following acute exposure to particles and additional criteria
pollutants. Circles represent single pollutant effect estimates and squares represent
copollutant effect estimates.
6.3.1.2. Controlled Human Exposure Studies
CAPs
Neither new controlled human exposure studies, nor studies cited in the 2004 PM
AQCD (U.S. EPA, 2004, 056905) have found significant effects of CAPs on respiratory
symptoms among healthy or asthmatic adults, or among older adults with COPD (Gong et
July 2009
6-148
DRAFT-DO NOT CITE OR QUOTE

-------
al., 2000, 155799; Gong et al., 2003, 042106; Gong et al., 2004, 087964; Gong et al., 2004,
055628; Gong et al., 2005, 087921; Gong et al., 2008, 156483; Petrovic et al., 2000, 004638).
Urban Traffic Particles
One new study reported an increase in respiratory symptoms (upper and lower
airways) among healthy volunteers (19"59 years old) during a 2-h exposure to road tunnel
traffic (PM2.5 concentration 46-81 |jg/m3) (Larsson et al., 2007, 091375). However,
information on specific respiratory symptoms (e.g., throat irritation, wheeze or chest
tightness) was not provided. In addition, this study only evaluated respiratory symptoms
pre- versus post-exposure, and did not compare response with a filtered air control
exposure.
Diesel Exhaust
Respiratory symptoms including mild nose and throat irritation have been reported
following controlled exposure to DE; however, other symptoms such as cough, wheeze and
chest tightness have not been observed (Mudway et al., 2004, 180208).
Model Particles
Pietropaoli et al. (2004, 156025) found no association between exposure to ultrafine
carbon particles and respiratory symptoms in healthy adults at concentrations between 10
and 50 |Jg/m3, or asthmatics at a concentration of 10 |Jg/m3. Beckett et al. (2005, 156261)
exposed healthy subjects to ultrafine and fine ZnO (500 (Jg/m3) and observed no difference
in respiratory symptoms compared to filtered air control 24 h following exposure. In a study
evaluating respiratory effects of exposure to ammonium bisulfate or aerosolized H2SO4 (200
and 2,000 |jg/m3) among healthy and asthmatic adults, Tunnicliffe et al. (2003, 088744)
observed no change in respiratory symptoms with either particle type or concentration
relative to filtered air. This finding is in agreement with many similar older studies which
have generally reported no increase in respiratory symptoms following exposure to acid
aerosols at concentrations <1,000 |jg/m3 (U.S. EPA, 1996, 079380; U.S. EPA, 2004, 056905)
Summary of Controlled Human Exposure Study Findings for Respiratory Symtoms
These new studies confirm previous reports that have found no association between
PM exposure and respiratory symptoms.
July 2009
6-149
DRAFT-DO NOT CITE OR QUOTE

-------
6.3.2. Pulmonary Function
Epidemiologic studies cited in the 2004 PM AQCD (U.S. EPA, 2004, 056905) observed
small decrements in pulmonary function associated with both PMio and PM2.5 (U.S. EPA,
2004, 056905). The majority of controlled human exposure studies reported no effect of PM
on pulmonary function, while the results from toxicological studies were mixed, with some
evidence of changes in tidal volume and respiratory rate following exposure to CAPs.
Epidemiologic studies published since the 2004 PM AQCD (U.S. EPA, 2004, 056905) have
reported an association between PM2.5 concentration and decrements in FEVi, particularly
among asthmatic children. These findings are coherent with results from a number of
recent toxicological studies which have observed increases in airway hyperresponsiveness
(AHR) following CAPs exposure. Results from recent controlled human exposure studies
have been inconsistent, with some studies demonstrating small decreases in arterial oxygen
saturation or maximal mid-expiratory flow following exposure to CAPs or EC. It is
interesting to note that these effects appear to be more pronounced among healthy adults
than adults with asthma or COPD.
6.3.2.1. Epidemiologic Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) concluded that both PM10 and PM2.5
appeared to affect lung function in asthmatics. UFPs did not appear to have any notably
stronger effect than other larger-diameter fine particles (U.S. EPA, 2004, 056905). Few
analyses were able to clearly distinguish the effects of PM10 and PM2.5 from other
pollutants. Results for PM10 peak flow analyses in non-asthmatic studies were inconsistent,
with fewer studies reporting strong associations.
Asthmatic Children
Seven recent panel studies have been conducted in the U.S. examining the association
of exposure to ambient PM and lung function in asthmatic children (Allen et al., 2008,
156208 in Seattle! Lewis et al., 2003, 088413 in Southern California; 2004, 097498; Lewis et
al., 2005, 081079 in Detroit; O'Connor et al., 2008, 156818; Rabinovitch et al., 2004, 096753
in Denver). Mean concentration data from these studies are summarized in Table 6-7.
In the Inner-City Asthma Study (ICAS), FEVi and PEFT were significantly related to
the 5-day avg of PM2.5, but not to the 1-day avg concentration (O'Connor et al., 2008,
156818). The risk of experiencing a percent-predicted FEVi more than 10% below personal
July 2009
6-150
DRAFT-DO NOT CITE OR QUOTE

-------
best was significantly related to the 5-day avg concentration of PM2.5 (1.14 [95% CI:
1.01-1.29]). The risk of experiencing a percent-predicted PEFR more than 10% below
personal best was significantly related to PM2.5 (1.18 [95% CI: 1.03-1.35]). This effect
remained robust in multipollutant models with O3 and NO2 for the FEVi effect, but not the
PEFR effect.
The Denver study (Rabinovitch et al., 2004, 096753), described in Section 6.3.1.1, also
examined daily forced expiratory volume in 1 sec (FEVi) and peak expiratory flow (PEF) in
86 asthmatic children over the course of 3 winters (some subjects participated in more than
one winter). Lung function measurements were performed under supervision daily at the
elementary school where all subjects attended, and without supervision every evening and
on nonschool days. As described above, the authors chose to use a 3-day ma of 24-h PM2.5 or
PM10 as the exposure metric. No statistical associations were observed between morning or
afternoon FEVi or PEF and particle exposure. The same group of researchers (Strand et al.,
2006, 157017) used regression calibration to estimate personal exposures to ambient PM2.5
and found that a 10 |ug/m3 increase in PM2.5 was associated with a 2.2% (95% CI: 0.0-4.3)
decrease in FEVi at a 1-day lag as compared with the estimate of a 1.0% decrease in FEVi
using ambient PM2.5 concentrations from fixed monitors. These results underscore the
effects of exposure error on epidemiologic study results! the effect estimate using an
estimate of personal exposure to ambient PM2.5 was twice that for central site PM2.5.
From winter 2001 to the spring of 2002, the same number (n = 86) of primary
school-age asthmatic children participated in six, 2-wk seasonal assessments of lung
function in Detroit (Lewis et al., 2005, 081079). Using a protocol similar to that used in
Rabinovitch et al. (2004, 096753), morning lung function measurements (FEVi, PEF) were
self-administered at school under supervision by research staff. Evening and weekend
measurements were made without supervision by research staff at home. Community-level
exposure was assessed using monitors placed on a school roof top of both of the two
communities. Most of the subjects (82 of 86) lived within 5 km of their respective
community monitors. In single-pollutant models using GEE and only among children
reporting the use of maintenance medication (corticosteroids), each 10 ug/m3 increase in lag
2 PM10 was associated with a decrease in the lowest daily percent predicted FEVi (a
reduction of 1.15%, [95% CI: -2.1 to -0.25]). Among children reporting presence of URI on
the day of lung function measurement, increases in the average of lag 3-5 of either PM2.5 or
PM10 resulted in a decrease in the lowest daily FEVi (for a 10 ug/m3 increase in PM2.5 the
July 2009
6-151
DRAFT-DO NOT CITE OR QUOTE

-------
reduction was 2.24% [95% CI; -4.4 to -0.25]; and for a 10 jug/m3 increase in PMio the
reduction was 2.4% [95% CI: -4.5 to -0.3]). In copollutant models that included one particle
pollutant and O3, and among children using maintenance medication, lag 3"5 PM2.5
continued to be associated with lowest daily FEVi as well as diurnal FEVi variability: each
10 |ug/m3 increase was associated with a 2.23% decrease in FEVi (95% CP -3.92 to -0.57)
and a 2.22% increase in FEVi variability (95% CP 1.0 to 3.50). Increases in lag 1 or lag 2 of
PM10 were associated with FEVi and FEVi diurnal variability in copollutant models. The
strongest association was with lag 2 for diurnal variability (for each 10 |ug/m3 increase
variability increased by 7.0% [95% CF 4.2 to 9.6). It is unclear what role the lack of
supervision during the evening and weekend measures may have had on these diurnal
results.
Two panel studies in southern California examined the association of PM exposure on
lung function in asthmatic children (Delfino et al., 2003, 050460; Delfino et al., 2004,
056897). In Delfino et al. (2003, 050460), described above, no association between exposure
to particles and PEF was found for 22 Hispanic, asthmatic children living in an area of
relatively high pollution. In Delfino et al. (2004, 056897) 19 asthmatic children aged 9-17 yr
were followed for 2 wk and daily, self-administered FEVi measurements were taken.
Particle exposures studied included central-site PM10 in addition to personal PM (in the
range of 0.1-10 jum range, with the highest response in the fine PM range), and home
stationary measurements of both PM10 and PM2.5. The authors report inverse associations
between percent expected FEVi and PM indicators. The strongest association for exposure
to personal PM was for a 5-day ma of 12-h daytime PM: for each 10 jug/m3 increase, FEVi
decreased by 7.1% (95% CF -9.9 to -2.9). Effects for all stationary sites (inside and outside of
residence, central site) for PM2.5 were on the order of 1 to 2% reductions in FEVi, with the
strongest associations for the 5-day ma (given in figures only). Likewise for PM10 measured
at stationary sites, the strongest effects were for 5-day ma and ranged from approximately
3.8% reduction associated with indoor monitors to about 1.5% for both the outdoor and
central site monitors (given in figures only). A helpful comparison among all 24-h measures
is given for 10 jug/m3 increases in personal PM and PM2.5 associated with decreases in
percent predicted FEVi: an increase of 10 |ug/m3 personal PM is associated with a decrease
in FEVi of 3.0% (95% CF -5.6 to -0.5); 10 jug/m3 increase in indoor PM with 2.4% decrease
(95% CF -4.2 to -0.6); 10 jug/m3 increase in outdoor PM with 1.5% decrease (95% CF -3.4 to
0.1); 10 jug/m3 increase in central site PM with 0.9% decrease (95% CF -2.6 to 0.5).
July 2009
6-152
DRAFT-DO NOT CITE OR QUOTE

-------
Trenga et al. (2006, 155209) reported associations among personal, residential, and
central site PM2.5 and lung function in 17 asthmatic children in Seattle. The only statistical
association with decline in FEVi was with indoor measurements of PM2.5: each 10 ug/m3
increase in lag 1 indoor PM2.5 was associated with a decline in FEVi of 64.8 mL (95% CP
¦111.3 to 18.3) (a 3.4% decline from the mean of 1.9 L). Indoor PM2.5 (lag l) was also
associated with declines in PEF (by 9.2 L/min [95% CP -17.5 to -0.9], a 3.6% decline from
the 254 L/min avg) and in MMEF for the 6 subjects not taking anti-inflammatory
medication (by 12.6 L/min [95% CP -20.7 to -4.6], a 13.7% decline from the 92 L/min avg).
Personal PM2.5 (lag l) was only statistically associated with PEF for the 6 subjects not on
anti-inflammatory medication: each 10 ug/m3 increase resulted in a 10.5 L/min ([95% CP
-18.7 to -2.3], a 4.5% decline from the 233 L/min avg) reduction in PEF. Anti-inflammatory
medication use significantly attenuated associations with PM2.5.
Also in Seattle, Allen et al. (2008, 156208) evaluated the effect of different PM2.5
exposure metrics in relation to lung function among children in WS'impacted areas. The
authors found that the ambient-generated component of PM2.5 exposure was associated
with decrements in lung function only among children not using inhaled corticosteroids,
whereas no association was reported with the nonambient exposure component. All of the
ambient concentrations were associated with decrements in both PEF and MEF. There were
no associations between any exposure metrics and FVC. The authors suggest that lung
function may be especially sensitive to the combustion-generated component of ambient
PM2.5, whereas airway inflammation may be more closely related to some other constituent.
In a longitudinal study, Liu et al. (2009, 192003) examined the association between
acute increases in ambient air pollutants and pulmonary function among children (ages 9*
14 yr) with asthma. FEVi and FEF25-75% exhibited a consistent trend of negative
associations with PM2.5 across lag days 0, 1, 0-1, and 0-2, with the strongest effects for
FEF25-75% on lag day 0 (-1.12% [95% CP -2.06 to -0.18]) and lag days 0-1 (-1.18% [95% CP -
2.24 to -0.12]). Two pollutant models including O3, SO2 or NO2 did not result in marked
changes in the PM2.5 risk estimates for FEVi or FEF25-75%.
Moshammer and Neuberger (2003, 041956) used a novel technique for assessing
exposure to PM in a study they conducted in Austria. They employed a diffusion charging
particle sensor (model LQ 1-DC, Matter Engineering AG, Wohlen, Switzerland) and a
photoelectric aerosol sensor (model PAS 2000 CE, EcoChem Analytics, League City, TX) to
relate the spirometry scores of Upper Austrian children aged 7-10 to particle surface area
July 2009
6-153
DRAFT-DO NOT CITE OR QUOTE

-------
and particle-bound PAH concentration, respectively. Details on these methods for
measuring surface area and PAH can be found in Shi et al. (2001, 078292) and Burtscher
(2005, 155710), respectively. By measuring the surface area distribution, it was possible to
understand potential for contact area with the respiratory tract cells. The authors found
that acute decrements of pulmonary function (FVC, FEVi, MEF50) were related to the active
surface of particles after adjustment for PM10. For short-term lung impairments, this
indicates that active particle surface is a better index of exposure than PM mass.
A number of additional panel studies conducted outside of the U.S. and Canada also
examined lung function using more traditional exposure metrics. Several European and
Asian studies reported associations with decrements in pulmonary function (FEVi, FVC,
FEF, MEF, PEFR) (Hogervorst et al., 2006, 189460; Hong et al., 2007, 091347; Moshammer
et al., 2006, 090771; Odajima et al., 2008, 192005; Peacock et al., 2003, 042026; Peled et al.,
2005, 156015). Others found little evidence for a relationship between PM and daily
changes in PEF after correction for the confounding effects of weather, trends in the data,
and autocorrelation (Fischer et al., 2002, 025731; Holguin et al., 2007, 099000; Just et al.,
2002, 035429; Preutthipan et al., 2004, 055598; Ranzi et al., 2004, 089500; Ward, 2003,
157111).
Adults
Trenga et al. (2006, 155209) examined personal, residential, and central site
monitoring of particles and the relationship with lung function in Seattle. In models
controlling for gaseous copollutants (CO, NO2), adults, regardless of COPD status,
experienced a significant decline in FEVi associated only with measurements of PM2.5 at
the central site: each 10 |jg/m3 increase in lag 0 PM2.5 was associated with a 35.3 mL (95%
CP -70 to -1.0) decrease in FEVi. This represents a 2.2% decline in mean FEVi (mean 1.6 L
during the study). Results for personal, indoor and outdoor measures of PM2.5 were
inconsistent. No significant associations were reported with outdoor PM10-2.5.
Girardot et al. (2006, 088271) assessed the effects of PM2.5 on the pulmonary function
of adult day hikers in the Great Smoky Mountains National Park. Hikers performed
spirometry both before their hike and when they returned from their hike. The authors
reported no statistically significant responses in pulmonary function with an average of 5 h
of outdoor exercise at ambient PM2.5 levels that were below the current NAAQS.
July 2009
6-154
DRAFT-DO NOT CITE OR QUOTE

-------
Specifically, posthike percentage changes in FVC, FEVi, FEVi/FVC, FEF25-75, and PEF were
not associated with PM2.5 exposure.
Ebelt et al. (2005, 056907) developed an approach to separately estimate exposures to
PM of ambient and nonambient origin based on a mass balance model. These exposures
were linked with respiratory and cardiovascular health endpoints for 16 patients with
COPD in Vancouver, Canada (mean age 74 yr). Effect estimates for estimated ambient
exposure were generally equal to or larger than those for the respective ambient
concentration levels for post-FEV and AFEVi, and were statistically significant for all
AFEVi comparisons (estimated from figure).
Several studies outside of the U.S. and Canada examined the relationship between
PM concentrations and lung function and all reported a decrease in lung function in adults
(FEVi, FVC, PEFR) associated with PM exposure (Boezen et al., 2005, 087396; Bourotte et
al., 2007, 150040; Lagorio et al., 2006, 089800; Lee et al., 2007, 093042; McCreanor et al.,
2007, 092841; Penttinen et al., 2006, 087988).
Measures of Oxygen Saturation
Oxygen saturation measures the percentage of hemoglobin binding sites in the
bloodstream occupied by oxygen. DeMeo et al. (2004, 087346) estimated the change in
oxygen saturation and mean PM2.5 concentration in the previous 24 h in a panel of elderly
subjects. They used the same panel of elderly Boston residents (n = 28) and study protocol
and analytic methods (12 wk of repeated oxygen saturation measurements) as Gold et al.
(2005, 087558) and Schwartz et al. (2005, 074317) in studies of ST-segment depression and
HRV, respectively. At each clinic visit, subjects had 5 min each of rest, standing,
post-exercise rest, and 20 cycles of paced breathing. The median PM2.5 concentration during
the study period was 10.0 |ug/m3 (Schwartz et al., 2005, 074317). Each 13.4 jug/m3 increase
in the mean PM2.5 concentration in the previous 6 h was associated with a 0.2% decrease in
oxygen saturation (95% CP -0.3 to 0.0) during the baseline rest period. Each 13.4 jug/m3
increase in mean 6"h PM2.5 concentration was also associated with a decline in oxygen
saturation during the post-exercise period (-0.2% [95% CP -0.3 to 0.0]), and post-exercise
paced breathing period (-0.1% [95% CP -0.3 to 0.0]), but not during the exercise period. The
authors suggest that these oxygen saturation reductions may result from pulmonary
vascular and inflammatory changes.
In a similar study, Goldberg et al. (2008, 180380) examined the association between
oxygen saturation, pulse rate, and ambient PM2.5, NO2, and SO2 concentrations in a panel
July 2009
6-155
DRAFT-DO NOT CITE OR QUOTE

-------
of 31 subjects in Montreal, with NYHA Class II or III heart failure who were aged 50-85 yr.
Although each 7.3 |ug/m3 increase in PM2.5 on lag day 0 was associated with a -0.087 (95%
CI = -0.143 to -0.031) change in oxygen saturation in unadjusted models, once adjusted for
temperature and barometric pressure, the estimated change was smaller and no longer
significant (-0.056 [95% CI = -0.117 to 0.005). Only SO2 was significantly associated with
reduced oxygen saturation in multivariate adjusted models. None of the pollutants
examined, including PM2.5, were associated with a change in pulse rate.
6.3.2.2. Controlled Human Exposure Studies
As with respiratory symptoms, there is little evidence from controlled human
exposure studies of PM-induced changes in pulmonary function. One study cited in the
2004 PM AQCD (U.S. EPA, 2004, 056905) noted a significant decrement in thoracic gas
volume in healthy adults following a 2-h exposure to PM2.5 CAPs (92 (Jg/m3); however, no
significant changes were observed in spirometric measurements, diffusing capacity (DLCO),
total lung capacity, or airways resistance (Petrovic et al., 2000, 004638). Other studies
found no significant changes in pulmonary function in healthy adults following exposure to
inhaled iron oxide particles (Lay et al., 2001, 020613), or in healthy and asthmatic adults
following exposure to CAPs (Ghio et al., 2000, 012140; 2003, 087365; Gong et al., 2000,
155799). Rudell et al. (1996, 056577) reported a significant increase in specific airways
resistance following exposure to DE, an effect that was not attenuated by reducing the
particle number by 46% (2.6 x 106/cm3 compared with 1.4 x 106/cm3) using a particle trap.
The particle trap did not affect the concentrations of other measured diesel emissions
including NO2, NO, CO, or total hydrocarbons. As described below, more recent controlled
human exposure studies provide limited and inconsistent evidence of changes in lung
function following exposure to particles from various sources.
CAPs
Among a group of healthy and asthmatic adults exposed to UFPs (Los Angeles, mean
concentration 100 |Jg/m3), Gong et al. (2008, 156483) observed small, yet statistically
significant decrements in arterial oxygen saturation immediately following exposure, 4 h
post-exposure, and 22 h post-exposure (0.5% mean decrease relative to filtered air across all
time points, p < 0.05). A statistically significant decrease in FEVi was also observed, but
only at 22 h post-exposure (2% decrease relative to filtered air, p < 0.05). The responses
July 2009
6-156
DRAFT-DO NOT CITE OR QUOTE

-------
demonstrated in this study were not significantly affected by health status. No such effects
were observed in a similar study conducted in Chapel Hill, NC which exposed healthy
adults to a lower concentration of ultrafine CAPs (49.8 |jg/m3) (Samet et al., 2009, 191913).
In addition, two studies evaluating effects of exposure to thoracic coarse CAPs (average
concentration 89-157 (Jg/m3) on lung function observed no changes in spirometric
measurements, diffusing capacity or arterial oxygen saturation 0-22 h post-exposure in
asthmatic or healthy adults (Gong et al., 2004, 055628; Graff et al., 2009, 191981). While
Gong et al. (2004, 087964) did not observe a significant association between exposure to
PM2.5 CAPs and spirometry in older subjects (60-80 yr), the investigators did report a
decrease in oxygen saturation immediately following CAPs exposure. This effect was
observed more consistently in healthy older adults than in older adults with COPD. These
findings were confirmed by a subsequent study conducted by the same laboratory (Gong et
al., 2005, 087921). The authors also observed a small decrease in maximal mid-expiratory
flow (MMEF) following a 2-h exposure to PM2.5 CAPs (200 (Jg/m3) which was more
pronounced in healthy subjects.
Urban Traffic Particles
Neither short term exposure to relatively high levels of urban traffic particle nor
longer exposures to lower concentrations of urban particles have been observed to alter
pulmonary function in controlled exposures among healthy adults. Larsson et al. (2007,
091375) exposed 16 adults for 2 h to fine particle concentrations of 46-81 (Jg/m3 in a room
adjacent to a busy road tunnel, with concomitant exposure to NO2 (0.12 ppm), NO
(0.71 ppm), and CO (5 ppm). Although respiratory effects in this study were not compared
to filtered air control, no difference in lung function was observed 14 h after exposure to
traffic particles relative to lung function measured on a day following typical activities that
did not include transit though a road tunnel. In a study of 24-h exposures to urban traffic
particles (PM2.5 9.7 |Jg/m3), no change in lung function was reported at 2.5 h after the start
of exposure relative to filtered air (Brauner et al., 2009, 191179).
Diesel Exhaust
Mudway et al. (2004, 180208) exposed 25 healthy adults to DE with an average
particle concentration of 100 (Jg/m3 and observed mild bronchoconstriction (airways
resistance) immediately following exposure relative to filtered air. No changes were
observed in FEV1 or FVC following DE exposure in these subjects, or in a group of 15
July 2009
6-157
DRAFT-DO NOT CITE OR QUOTE

-------
asthmatics exposed using the same protocol (Mudway et al., 2004, 180208; Stenfors et al.,
2004, 157009)
Model Particles
Pietropaoli et al. (2004, 156025) observed a reduction in MMEF and DLCO in healthy
adults 21 h after a 2-h exposure to ultrafine carbon particles (50 |Jg/m3). This reduction in
DLCO may reflect a PM-induced vasoconstrictive effect on the pulmonary vasculature.
Tunnicliffe et al. (2003, 088744) did not observe any significant change in lung function
following exposure to ammonium bisulfate or aerosolized H2SO4 (200 and 2,000 (Jg/m3) in
healthyor asthmatic adults, which is consistent with findings of the majority of studies of
controlled exposures to acid aerosols presented in the last two PM AQCDs (U.S. EPA, 1996,
079380; 2004, 056905).
Summary of Controlled Human Exposure Study Findings for Pulmonary Function
Taken together, the majority of controlled human exposure studies do not provide
evidence of PM-induced changes in pulmonary function! however, some investigators have
observed slight decreases in DLCO, MMEF, FEVi, oxygen saturation, or increases in
airways resistance following exposure to CAPs, DE, or ultrafine EC.
6.3.2.3. Toxicological Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) included three animal toxicological
studies which measured pulmonary function following multiday short-term inhalation
exposure to CAPs. A decreased respiratory rate was noted in one study involving dogs.
Increased tidal volume was observed in one study involving rats while no changes were
observed in the other rat study. AHR was found in 4 studies of mice, healthy rats or SH rats
exposed to ROFAby IT instillation or inhalation. Studies conducted since the last review
are discussed below.
CAPs
SH rats exposed to Tuxedo, NY CAPs via nose-only inhalation for 4 h (mean
concentration 73 |ug/m3; single-day concentrations 80 and 66 ug/m3; 2/2001 and 5/2001
respectively) had a statistically significant decreased respiratory rate compared with
air-exposed controls (Nadziejko et al., 2002, 087460). This measure was obtained from BP
fluctuations using radiotelemetry. The decrease in respiratory rate of 25-30 breaths/min
July 2009
6-158
DRAFT-DO NOT CITE OR QUOTE

-------
was an immediate response to CAPs, beginning shortly after the exposure began and
ceasing with the end of exposure. It was accompanied by a decrease in HR (see
Section 6.2.1.3). Rats were also exposed to fine (MMAD 160 nm! 49-299 |Jg/m3) and
ultrafine HaSO-^MMAD 50-75 nm! 140-750 |jg/m3) (Nadziejko et al., 2002, 087460) because
H2SO4 aerosols have the potential to activate irritant receptors. Irritant receptors, found at
all levels of the respiratory tract, include rapidly-adapting receptors and sensory Ofiber
receptors (Alarie, 1973, 191136; Coleridge and Coleridge, 1994, 156362; 2003, 157145; 2006,
155519; Widdicombe and Lee, 2001, 019049) Activation of trigeminal afferents in the nose
causes CNS reflexes resulting in decreases in respiratory rate through a lengthened
expiratory phase, closure of the glottis, closure of the nares with increased nasal airflow
resistance and effects on the cardiovascular system such as bradycardia, peripheral
vasoconstriction and a rise in systolic arterial blood pressure. Sneezing, rhinorrhea and
vasodilation with subsequent nasal vascular congestion are also nasal reflex responses
involving the trigeminal nerve (Sarin et al., 2006, 191166). Activation of vagal afferents in
the tracheobronchial and alveolar regions of the respiratory tract causes CNS reflexes
resulting in bronchoconstriction, mucus secretion, mucosal vasodilation, cough, and apnea
followed by rapid shallow breathing. Besides effects on the respiratory system, effects on
the cardiovascular system can also occur including bradycardia and hypotension or
hypertension. Fine H2SO4 induced an overall decrease in respiratory rate, with ultrafine
H2SO4 resulting in elevated respiratory rate compared to control (Nadziejko et al., 2002,
087460). The authors suggested that both CAPs and fine H2SO4 aerosols activated sensory
irritant receptors in the upper airways, resulting in a decreased respiratory rate. The
response to ultrafine H2SO4 aerosols differed from the other responses and was thought to
be due to deposition of UFPs deeper into the lung with the subsequent activation of
pulmonary irritant receptors which trigger an increase in respiratory rate. Since irritant
receptors in nasal, tracheobronchial and alveolar regions act via trigeminal- and
vagal-mediated pathways, this study indicates a role for neural reflexes in respiratory
responses to CAPs.
Kodavanti et al. (2005, 087946) measured respiratory frequency 1 day after a 2-day
exposure of SH and WKY rats to CAPs from RTP, NC (mean mass concentration range
144-2,758 |ug/m3; less than 2.5 |um in size! 8/27-10/24/2001) for 4 h/day. Increases in
inspiratory and expiratory times were seen in SH, but not WKY rats, exposed to CAPs
compared with filtered air controls.
July 2009
6-159
DRAFT-DO NOT CITE OR QUOTE

-------
Effects of CAPs on pulmonary function were also investigated in a rat model of
pulmonary hypertension using SD rats pre-treated with monocrotaline (Lei et al., 2004,
087999). In this study, rats were exposed to CAPs from an urban high traffic area in Taiwan
(mean mass concentration 371 ug/ma) for 6 h/day on 3 consecutive days and pulmonary
function was evaluated 5 h post-exposure using whole-body plethysmography A statistically
significant decrease in respiratory frequency and an increase in tidal volume were observed
following CAPs exposure, along with an increase in airway responsiveness (measured as
Penh) following methacholine challenge.
In many animal studies changes in ventilatory patterns are assessed using whole
body plethysmography, for which measurements are reported as enhanced pause (Penh).
Some investigators report increased Penh as an indicator of AHR, but these are
inconsistently correlated and many investigators consider Penh solely an indicator of
altered ventilatory timing in the absence of other measurements to confirm AHR. Therefore
use of the terms AHR or airway responsiveness has been limited to instances in which the
terminology has been similarly applied by the study investigators.
Diesel Exhaust
Li et al. (2007, 155929) exposed BALB/c and C57BL/6 mice to clean air or to low dose
DE (containing 100 |ug/m3 DEP) for 7 h/day and 5 days/wk for 1, 4 and 8 wk. average gas
concentrations were reported to be 3.5 ppm CO, 2.2 ppm NO2, and less than 0.01 ppm SO2.
AHR was evaluated by whole body plethysmography at day 0 and after 1, 4 and 8 wk of
exposure. Exposure to DE for 1 wk resulted in an increased sensitivity of airways to
methacholine, measured as Penh, in C57BL/6, but not BALB/c, mice. Other short-term
responses of this study are discussed in Sections 6.3.3.3 and 6.3.4.2.
McQueen et al. (2007, 096266) investigated the role of vagally-mediated pathways in
respiratory responses to PM. Respiratory minute volume (RMV) was increased in
anesthetized Wistar rats 6 h after treatment with 500 |ug DEP (SRM2975) by IT instillation.
This response was blocked by severing the vagus nerve or pretreatment with atropine. The
absence of a respiratory response with vagotomy or atropine indicated that the increase in
RMV following DEP exposure involved a neural reflex acting via vagal afferents. No
statistically significant changes in mean BP, HR or HRV were observed in response to DEP
in this study. Vagally-mediated inflammatory responses to DEP were also observed in this
study and are discussed in Section 6.3.3.3.
July 2009
6-160
DRAFT-DO NOT CITE OR QUOTE

-------
Model Particles
In a study by Last et al. (2004, 097334), BALB/c mice were exposed to 250 |ug/m3
laboratory-generated iron-soot (size range 80-110 nm! about 200 |ug/m3 as soot) for 4 h/day
and 3 days/wk for 2 wk. Pulmonary function was measured by whole-body plethysmography
after challenge with methacholine. No increase in AHR, as measured by Penh, was
observed following 2-wk exposure to iron-soot. Other findings of this study are reported in
Sections 6.3.3.3, 6.3.5.2 and 7.3.2.2.
Summary of Toxicological Study Findings for Pulmonary Function
Several recent studies demonstrated alterations in respiratory frequency and AHR
following short-term exposure to CAPs and DE. Two studies provide evidence for the
involvement of irritant receptors and vagally-mediated neural reflexes in mediating
changes in respiratory functions.
6.3.3. Pulmonary Inflammation
The discussion of the effects of PM on pulmonary inflammation in the 2004 PM AQCD
(U.S. EPA, 2004, 056905) was limited by a relative lack of information from controlled
human exposure and toxicological studies. Although no epidemiologic studies of pulmonary
inflammation were described in the 2004 PM AQCD (U.S. EPA, 2004, 056905), several
recent studies have observed a positive association between PM concentration and exhaled
NO. New controlled human exposure and toxicological studies have also generally observed
an increase in markers of inflammation in the airways following exposure to PM.
6.3.3.1. Epidemiologic Studies
No epidemiologic studies of pulmonary inflammation were described in the 2004 PM
AQCD (U.S. EPA, 2004, 056905).
Exhaled Nitrogen Oxide - Asthmatic Children
Exhaled NO, a biomarker for airway inflammation, was the outcome studied in panels
of asthmatic children in southern California (Wu et al., 2006, 157156) and Seattle (Allen et
al., 2008, 156208; Koenig et al., 2003, 156653; 2005, 087384; Mar et al., 2005, 088759).
Mean concentration data from these studies are summarized in Table 6-7. Delfino et al.
(2006, 157156) followed 45 asthmatic children for 10 days with offline fractional eNO and
July 2009
6-161
DRAFT-DO NOT CITE OR QUOTE

-------
examined the associations with exposures to personal PM2.5 and 24-h PM2.5, EC and OC as
well as ambient PM2.5, EC and OC. The strongest associations were between eNO and
2-day avg pollutant concentrations: for a 10 |ug/m3 increase in personal PM2.5, eNO
increased by 0.46 ppb (95% CP 0.04-0.79); for 0.6 |ug/m3 personal EC, eNO increased by 0.7
ppb (95% CP 0.3-1.1). An association with exposure to ambient PM2.5 was only statistically
significant in 19 subjects taking inhaled corticosteroids: for each 10 ug/m3 increase in PM2.5,
eNO increased by 0.77 ppb (95% CP 0.07-1.47).
In a panel of 19 asthmatic children in Seattle, effects were observed only among the
10 non-users of inhaled corticosteroids. For each 10 jug/m3 increase in personal, outdoor,
indoor, or central site PM2.5, eNO increased from 3.82 ppb (associated with central site, 95%
CP 1.22-6.43) to 4.48 ppb (with personal PM2.5, 95% CP 1.02-7.93) (Koenig et al., 2003,
156653). Further analysis examining the association between eNO and outdoor and
indoor-generated particles suggested that eNO was associated more strongly with ambient
particles, but only for non-users of medication: each 10 ug/m3 increase in estimated ambient
PM2.5 results in an increase in eNO of 4.98 ppb (95% CP 0.28-9.69) (Koenig et al., 2005,
087384).
Also in Seattle, WA, Mar et al. (2005, 088759) examined the association between eNO
and ambient PM2.5 concentration among children (aged 6-13 yr) recruited from an
asthma/allergy clinic. Fractional exhaled nitric oxide (FeNO) was associated with hourly
averages of PM2.5 up to 10-12 h after exposure. Each 10 jug/m3 increase in 1-h mean PM2.5
concentration was associated with a 6.99 ppb increase in eNO (95% CP 3.43-10.55) among
children not taking inhaled corticosteroids, but associated with only a 0.77 ppb decrease in
eNO (95% CP -4.58 to 3.04) among those taking inhaled corticosteroids.
Allen et al. (2008, 156208), in a reanalysis of data from Koenig et al. (2005, 087384),
evaluated the effect of different PM2.5 exposure metrics in relation to airway inflammation
among children in WS-impacted areas of Seattle. The authors found that for the 9 non-
users of inhaled corticosteroids, the ambient-generated component of PM2.5 exposure was
associated with respiratory responses, both airway inflammation and decrements in lung
function, whereas the nonambient PM2.5 exposure component was not. They did note,
however, different relationships for airway inflammation and decrements in lung function,
with the former significantly associated with total personal PM2.5, personal light-absorbing
carbon (LAC), and ambient generated personal PM2.5 and the latter related to ambient
PM2.5 and its combustion markers. The different results between FeNO and lung function
July 2009
6-162
DRAFT-DO NOT CITE OR QUOTE

-------
were not unexpected; epidemiologic data show that airway inflammation indicated by FeNO
does not correlate strongly with either respiratory symptoms or lung function (Smith and
Taylor, 2005, 192176). The authors conclude that lung function decrements may be
associated with the combustion-generated component of ambient PM2.5, whereas airway
inflammation may be related to some other constituent of the ambient PM2.5 mixture.
In a longitudinal study, Liu et al. (2009, 192003) examined the association between
acute increases in ambient air pollutants and FeNO among children (ages 9-14 yr) with
asthma. Median PM2.5 = 6.5, 95th percentile = 19.0. FeNO had a trend of positive
associations with PM2.5, with the strongest association on lag day 0 (3.12% [95% CP -2.12 to
8.82]). Two pollutant models including O3, SO2 or NO2 did not result in marked changes in
the PM2.5 risk estimates for FeNO.
Several studies outside of the U.S. examined eNO in relation to PM exposure among
children. Fischer et al. (2002, 025731) and Murata et al. (2007, 156787) found a significant
association between increases in PM and increases in the percent of eNO. Holguin et al.
(2007, 099000) found no association between exposure to PM and eNO. However, they did
see significant associations between increases in eNO for the 95 asthmatic subjects and
measures of road density of roads 50- and 75-m from the home.
Exhaled Nitrogen Oxide - Adults
Three recent panel studies examined the effects of particle exposure on eNO
measured in older adults (Adamkiewicz et al., 2004, 087925 in Steubenville, OH; Adar et
al., 2007, 001458; Jansen et al., 2005, 082236 in Seattle). Mean concentration data from
these studies are characterized in Table 6-7. Breath samples were collected weekly for 12
wk from a group of 29 elderly adults in Steubenville, OH (Adamkiewicz et al., 2004,
087925). In single-pollutant models, each 10 jug/m3 increase in 24-h ambient PM2.5
increased eNO by 0.82 ppb (95% CP 0.19-1.45), a change of 15% compared to mean eNO (9.9
ppb). Effects were essentially unchanged in multipollutant models that included ambient
and/or indoor NO. The effect estimates for the 7 COPD subjects were significantly higher
than for normal subjects (2.20 vs. 0.45 ppb, p = 0.03) (Adamkiewicz et al., 2004, 087925).
In the Seattle panel of older adults (aged 60-86 yr), 7 subjects were asthmatic and 9
had a diagnosis of COPD (5 with asthma and 4 without) (Jansen et al., 2005, 082236).
Exhaled NO was measured daily for 12 days, along with personal, indoor, outdoor and
central site PM10, PM2.5 and BC. Significant associations between 24-h avg PM and eNO
July 2009
6-163
DRAFT-DO NOT CITE OR QUOTE

-------
were found only for the asthmatic subjects: 10 jug/m3 increases in outdoor levels (measured
outside the subjects' homes) of PM2.5 or PM10 were associated with increases in eNO of 4.23
ppb (95% CL 1.33-7.13), an increase of 22% above the group mean of 19.2 ppb, and 5.87 ppb
(95% CP 2.87-8.88), an increase of 31%, respectively. BC measured indoors, outdoors or
personally was also associated with significant increases in eNO (of 3.97, 2.32, and 1.20
ppb, respectively) (Jansen et al., 2005, 082236).
Adar et al. (2007, 001458) conducted a panel study of 44 non-smoking seniors residing
in St. Louis, MO (60 yr). As part of the study, subjects were taken on group trips to a
theater performance, Omni movie, outdoor band concert, and a Mississippi River boat
cruise. Subjects were driven to and from each event aboard a diesel bus. Before and after
each bus trip, eNO was measured on each subject. Two carts containing continuous air
pollution monitors were used to measure group-level micro-environmental exposures to
PM2.5, BC, and size-specific particle counts (0.3-2.5 |um and 2.5-10 |um) on the day of each
trip. Each 10 |ug/m3 increase in 24-h mean PM2.5 concentration was associated with a 36%
increase in eNO pre-trip (95% CP 5"7l). Each 10 jug/m3 increase in micro-environmental
PM2.5 concentration (i.e., during the bus ride) was associated with a 27% increase in eNO
post-trip (95% CP 17-38).
These studies all demonstrated an association between increased levels of eNO and
increases in PM in the previous 4-24 h. Further, three studies demonstrated effects in
elderly populations (Adamkiewicz et al., 2004, 087925; Adar et al., 2007, 001458; Jansen et
al., 2005, 082236) while four others reported a similar acute increase in eNO among
children (Dellino et al., 2006, 090745; Koenig et al., 2003, 156653; 2005, 087384; 2005,
088999).
Outside of the U.S., one study examined eNO in a panel of 60 adult asthmatic subjects
in London. McCreanor et al. (2007, 092841) reported that 1 jug/m3 increase in personal
exposure to EC was associated with significant increases of approximately 1.75-2.25% in
eNO (results were presented graphically only) for up to 22 h post-exposure.
Other Biomarkers of Pulmonary Inflammation and Oxidative Stress
Other biomarkers of respiratory distress that have been examined in recent panel
studies include urinary leukotriene E4 (LTE4) in asthmatic children (Rabinovitch et al.,
2006, 088031); two oxidative stress markers: thiobarbituric acid reactive substances
(TBARS) and 8-isoprostane in asthmatic children (Liu et al., 2009, 192003) and breath
July 2009
6-164
DRAFT-DO NOT CITE OR QUOTE

-------
acidification in adolescent athletes (Ferdinands et al., 2008, 156433). Mean concentration
data from these studies are characterized in Table 6-7.
In Rabinovitch et al. (2006, 088031), LTE4, an asthma-related biological mediator, was
used to study the response to short-term particle exposure. In the second winter of their
2-yr study of asthmatic children (described above in Section 6.3.1.1, under respiratory
symptom and medication use outcomes), urine samples were collected at approximately the
same time of day from 57 subjects for 8 consecutive days. Controlling for days with URI
symptoms, each 10 jug/m3 increase in morning maximum PM2.5 (measured by TEOM), was
associated with an increase in LTE4 levels by 5.1% (95% CP 1.6-8.7). No significant effects
were observed on the same day or up to 3 days later based on 24-h averaged concentrations
from the TEOM monitor or from the FRM central site monitor.
In a longitudinal study conducted in Windsor, Ontario, Liu et al. (2009, 192003)
examined the association between acute increases in ambient air pollutants and TBARS
and 8-isoprostane among children (ages 9-14 yr) with asthma. TBARS, but not 8-
isoprostane, was positively associated with PM2.5 (percent change in TBARS 40.6% [95% CP
11.8-81.3], lag 0-2 days). The association with TBARS persisted for at least 3 days. Adverse
changes in pulmonary function (see above) were consistent with those of TBARS in
response to PM2.5 with a similar lag structure, suggesting a coherent outcome for small
airway function and oxidative stress.
The effects of vigorous outdoor exercise during peak smog season in Atlanta, GA on
breath pH, a biomarker of airway inflammation, in adolescent athletes (n = 16, mean
age = 14.9 yr) were examined by Ferdinands et al. (2008, 156433). Median pre-exercise
breath pH was 7.58 (range 4.39-8.09) and median post-exercise breath pH was 7.68 (range
3.78-8.17). The authors observed no significant association between ambient PM and
post-exercise breath pH. However both pre- and post-exercise breath pH were strikingly low
in these athletes when compared to 14 relatively sedentary healthy adults and to published
values of breath pH in healthy subjects. The authors speculate that repetitive vigorous
exercise may induce airway acidification.
Effect of Measurement Location on Studies of Pulmonary Function and Inflammation
A number of studies examining exposure to PM2.5 and pulmonary function and
inflammation have compared the results of exposure assessment based on concentrations
recorded from personal, indoor, outdoor, and/or ambient monitors (Allen et al., 2008,
July 2009
6-165
DRAFT-DO NOT CITE OR QUOTE

-------
156208; Delfino et al., 2004, 056897; Delfino et al., 2006, 090745; Koenig et al., 2005,
087384; Trenga et al., 2006, 155209). Two investigations evaluated PM2.5 concentrations
from indoor, outdoor, personal and central site monitors and the relationship with FEVi.
Delfino et al. (2004, 056897) reported that personal exposure estimates showed a stronger
association with FEVi than any of the stationary exposures, and that indoor exposure
estimates were associated with a stronger effect than either outdoor or central site exposure
estimates. However, Trenga et al. (2006, 155209) reported the largest declines in FEVi
associated with central site exposure estimates, though the most consistent association
with declines in FEVi came from the exposure estimates measured by indoor monitors.
Delfino et al. (2006, 090745) used personal and ambient exposure estimates in a study of
FeNO among asthmatic children and found that the personal exposure estimates were more
robust than the ambient exposure estimates. Two studies conducted in Seattle, WA
partitioned personal exposure to into its ambient-generated and indoor-generated
components. Koenig et al. (2005, 087384) reported that ambient-generated PM2.5 was
consistently associated with an increase in FeNO, while the indoor-generated component of
PM2.5 was less strongly associated with FeNO. This could reflect the difference in
composition of indoor-generated PM2.5 as compared to ambient-generated PM2.5. Similarly,
Allen et al. (2008, 156208) found that FeNO was associated with the ambient-generated
component of personal PM2.5 exposure, but not with ambient PM2.5 concentrations
measured by central site monitors. Overall, these studies provide a unique perspective on
how measurement location influences the findings of epidemiologic studies. This small
group of studies indicates that effects are associated with all types of PM measurement,
suggesting health effects of both ambient-generated and indoor-generated particles. It is
likely that variability in season, meteorology, topography, geography, behavior and exposure
patterns contribute to the observed differences.
6.3.3.2. Controlled Human Exposure Studies
Studies of controlled human exposures presented in the 2004 PM AQCD (U.S. EPA,
2004, 056905) provided evidence of pulmonary inflammation induced by exposure to PM.
Lay et al. (1998, 007683) found that instillation of iron oxide particles (2.6 |Jm) produced an
increase in alveolar macrophages and neutrophils in BALF collected 24 h post-instillation.
Ghio and Devlin (2001, 017122) evaluated the inflammatory response following instillation
of particles extracted from filters collected in the Utah Valley both prior to and after the
July 2009
6-166
DRAFT-DO NOT CITE OR QUOTE

-------
closure of an area steel mill. Subjects who underwent pulmonary instillation of particles
(500 |Jg) collected while the steel mill was operating (n = 16) had significantly higher levels
of neutrophils 24 h post-instillation compared with either saline instillation or with subjects
(n = 8) who were instilled with the same mass of PM collected during the mill's closure. This
finding indicates that metals may be an important PM component for this health outcome.
In an inhalation study of exposure to fine CAPs (23-311 (Jg/m3) from Chapel Hill, NC, Ghio
et al. (2000, 012110) observed an increase in airway and alveolar neutrophils 18 h after the
2-h exposure. A similar finding was reported by Rudell et al. (1999, 001964) following
exposure to DE among healthy adults. In this study, reducing the particle number from 2.6
x 106/cm3 to 1.3 x 106/cm3 while maintaining the concentration of gaseous diesel emissions,
was not observed to attenuate the response. As summarized below, several recent studies of
controlled exposures have provided some additional evidence of pulmonary inflammation
associated with PM.
CAPS
A series of exposures to ultrafine, fine, and thoracic coarse CAPs from Los Angeles
with average particle concentrations between 100 and 200 |jg/m3 have not been shown to
have a significant effect on markers of airway inflammation in healthy or health-
compromised adults (Gong et al., 2004, 087964; 2004, 055628; 2005, 087921; 2008, 156483)
However, two recent studies conducted in Chapel Hill, NC reported significant increases in
percent PMNs and concentration of IL-8 in BALF among healthy adults 18-20 h following
controlled exposures to coarse (89 [Jg/m3' and ultrafine (49.8 [Jg/m3' CAPs, respectively
(Graff et al., 2009, 191981; Samet et al., 2009, 191913). As discussed above, the same
laboratory previously reported a mild inflammatory response in the lower respiratory tract
following exposure to fine CAPs (Ghio et al., 2000, 012110). In a follow-up analysis, Huang
et al. (2003, 087377) found the increase in BAL neutrophils demonstrated by Ghio et al.
(2000, 012110) to be positively associated with the Fe, Se, and SO42- content of the particles.
Alexis et al. (2006, 154323) recently evaluated the effect of PM10-2.5 on markers of
airway inflammation, specifically focusing on the impact of biological components of
PM10-2.5. Healthy men and women (n = 9) between the ages of 18 and 35 inhaled nebulized
saline (0.9%) as well as aerosolized PM10-2.5 collected from ambient air. Subjects were
exposed to PM10-2.5 on two separate occasions, once using PM10-2.5 that had been heated to
inactivate biological material and once using non-heated PM10-2.5. Approximately 0.65 mg
July 2009
6-167
DRAFT-DO NOT CITE OR QUOTE

-------
PMio-2.5 was deposited in the respiratory tract of subjects during the exposures. Markers of
inflammation and immune function were analyzed in induced sputum collected 2-3 h after
inhalation of saline or PM10-2.5. Both heated and non-heated PM10-2.5 were observed to
increase the neutrophil response compared with saline. Exposure to non-heated PM10-2.5
was found to increase levels of monocytes, eotaxin, macrophage TNF-a mRNA, and was also
associated with an upregulation of macrophage cell surface markers. No such effects were
observed following exposure to biologically inactive PM10-2.5. These results suggest that
while thoracic coarse fraction PM-induction of neutrophil response is not dependent on
biological components, heat sensitive components of coarse PM (e.g., endotoxin) may be
responsible for PM-induced alveolar macrophage activation.
Traffic Particles
Larsson et al. (2007, 091375) exposed 16 healthy adults to air pollution in a road
tunnel for 2 h during the afternoon rush hour in Stockholm, Sweden. The median PM2.5 and
PM10 concentrations during the road tunnel exposures were 64 (Jg/m3 and 176 |Jg/m3,
respectively. Bronchial biopsies were obtained and broncoscopy and bronchoalveolar lavage
were performed 14 h after the exposure. The results were compared with a control exposure
which consisted of exposure to urban air during normal activity. The authors reported
significant BALF increases in percentage of lymphocytes, total cell number, and alveolar
macrophages following exposure to road tunnel exposure versus control. These results
provide evidence of a significant association between exposure to road tunnel air pollution
and airway inflammation. However, unlike other controlled exposure studies, the control
exposure was not a true clean air control, but only a lower dose exposure group with no
characterization of personal exposure. In addition, it is not possible to separate out the
contributions of each air pollutant, including PM, on the observed inflammatory response.
Diesel Exhaust
In a recent study evaluating the effect of DE exposure on markers of airway
inflammation, Behndig et al. (2006, 088286) exposed healthy adults (n = 15) for 2 h with
intermittent exercise to filtered air or DE with a reported PM10 concentration of 100 |Jg/m3.
Eighteen hours after exposure to DE, the authors found significant increases in neutrophil
and mast cell numbers in bronchial tissue, as well as significant increases in neutrophil
numbers and IL-8 in bronchial lavage fluid compared with filtered air control. Similarly,
Stenfors et al. (2004, 157009) observed an increase in pulmonary inflammation (e.g.,
July 2009
6-168
DRAFT-DO NOT CITE OR QUOTE

-------
airways neutrophilia and an increase in IL-8 in BALF) among healthy adults 6 h following
exposure to DE (PMio average concentration 108 |Jg/m3). It is interesting to note, however,
that no such inflammatory effects were observed in a group of mild asthmatic subjects in
the same study The DE-induced neutrophil response in the airways of healthy subjects
observed in these two studies (Behndig et al., 2006, 088286; Stenfors et al., 2004, 157009) is
qualitatively consistent with the findings of Ghio et al. (2000, 0121 10) who exposed healthy
subjects to Chapel Hill fine CAPs. In a group of healthy volunteers, Bosson et al. (2007,
156286) demonstrated that exposure to O3 (2 h at 0.2 ppm) may enhance the airway
inflammatory response of DE relative to clean air (l-h exposure to 300 |Jg/m3). Exposure to
O3 was conducted 5 h after exposure to DE, and resulted in an increase in the percentage of
neutrophils in induced sputum collected 18 h after exposure to O3. In a subsequent study
using a similar protocol at the same concentrations, prior exposure to DE was shown to
increase the inflammatory effects of O3 exposure, demonstrated as an increase in neutrophil
and macrophage numbers in bronchial wash (Bosson et al., 2008, 156287).
Wood Smoke
Barregard et al. (2008, 155675) examined the effect of a short-term exposure (4 h) to
WS (240-280 (Jg/m3) on markers of pulmonary inflammation in a group of healthy adults.
Exposure to WS increased alveolar NO compared to filtered air (2.0 ppb versus 1.3 ppb) 3 h
after exposure. Although these results provide some evidence of a PM-induced increase in
pulmonary inflammation, the physiological significance of the relatively small increase in
alveolar NO is unclear.
Model Particles
Pietropaoli et al. (2004, 156025) observed a lack of airway inflammatory response 21 h
after exposure to ultrafine EC particles (10-50 |jg/m3) among healthy and asthmatic adults.
The same laboratory reported no effect of exposure to ultrafine or fine ZnO (500 (Jg/m3) on
total or differential sputum cell counts 24 h after exposure in a group of healthy adults
(Beckett et al., 2005, 156261). Tunnicliffe et al. (2003, 088744) measured levels of eNO as a
marker of airway inflammation following l-h controlled exposures to ammonium bisulfate
or aerosolized H2SO4 (200 and 2,000 |jg/m3) in a group of healthy and asthmatic adults.
While exposure to ammonium bisulfate increased the concentration of eNO immediately
following exposure in asthmatics, no such effect was observed in healthy adults, or in either
healthy or asthmatic adults following exposure to aerosolized H2SO4.
July 2009
6-169
DRAFT-DO NOT CITE OR QUOTE

-------
Instillation
Schaumann et al. (2004, 087966) investigated the inflammatory response of human
subjects instilled with PM2.5 (100 |Jg) collected from two different cities in Germany,
Hettstedt and Zerbst. Although instillation of PM from both cities were shown to induce
airway inflammation, instillation of PM from the more industrial area (Hettstedt) resulted
in greater influxes of BALF monocytes compared to PM collected from Zerbst. The authors
postulated that the difference in response between PM from the two cities may be due to
the higher concentration of transition metals observed in the samples collected from
Hettstedt. Another study reported no change in inflammatory markers in nasal lavage fluid
4 and 96 h following intranasal instillation of DEP (300 (Jg/nostril) in asthmatics and
healthy adults (Kongerud et al., 2006, 156656). Pre-exposure of DEP to O3 was not shown to
have any effect on the response. Although not a cross-over design, these findings suggest
that exposure to DEP without the gaseous component of DE may have little effect on
inflammatory responses in human subjects.
Summary of Controlled Human Exposure Study Findings for Pulmonary Inflammation
These new studies strengthen the evidence of PM-induced pulmonary inflammation,
with the majority of the evidence associated with fine and thoracic coarse fractions. The
response appears to vary significantly depending on the source and composition of the
particles.
6.3.3.3. Toxicological Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) discussed numerous studies
investigating pulmonary inflammation in response to CAPs, ROFA, DEPs, metals and acid
aerosols. A wide variety of responses was reported depending on the type of PM and route of
administration. In general, IT exposure to fly ash and metal PM resulted in notable
pulmonary inflammation. In contrast, inhalation of sulfates and acid aerosols had minimal
if any effect on pulmonary inflammation. More recent animal toxicological studies using
CAPs, DE and other relevant PM types are summarized below.
CAPs
The 2004 PM AQCD (U.S. EPA, 2004, 056905) found that exposure to fine CAPs at
concentrations of 100-1,000 jug/m3 for 1-6 h/day and 1-3 days generally resulted in minimal
to mild inflammation in rats and dogs. Somewhat enhanced inflammation was observed in
July 2009
6-170
DRAFT-DO NOT CITE OR QUOTE

-------
a model of chronic bronchitis. Since the last review, numerous studies have investigated
inflammatory responses to fine and ultrafine CAPs in both healthy and compromised
animal models.
In one study of healthy animals, SD rats were exposed to CAPs for 4 h/day on 3
consecutive days in Fresno, CA, in fall 2000 and winter 2001 (PM2.5; mean mass
concentration 190-847 |u,g/m3) (Smith et al., 2003, 042107). The particle concentrator used in
these studies was capable of enhancing the concentration of ultrafine as well as fine
particles. Immediately after exposure on the third day, BALF was collected and analyzed
for total cells and neutrophils. Statistically significant increases were observed in numbers
of neutrophils during the first week of the fall exposure period and in numbers of total cells,
neutrophils and macrophages during the first week of the winter exposure period. CAPs
concentrations were >800 (Jg/m3 during both of those weeks.
Two studies were conducted using CAPs in Boston. In a study by Godleski et al. (2002,
156478), healthy SD rats were exposed for 5 h/day for 3 consecutive days to CAPs ranging
in concentration from 73.5-733.0 ug/m3. BALF and lung tissue were collected for analysis 1
day later. Neutrophilic inflammation was indicated by a statistically significant increase in
percent neutrophils in BALF. Microarray analysis of RNA from lung tissue and BALF cells
demonstrated increased gene expression of pro-inflammatory mediators, markers of
vascular activation and enzymes involved in organic chemical detoxification. This study
overlapped in part with previously described studies by Saldiva et al. (2002, 025988) and
Batalha et al. (2002, 088109) (see Section 6.2.4.3). In another study, healthy SD rats were
exposed for 5 h to CAPs (mean mass concentration 1228 jug/m3; 6/20-8/16/2002; (Rhoden et
al., 2004, 087969). A statistically significant increase in BALF neutrophils was observed 24-
h following CAPs exposure. Histological analysis confirmed the influx of inflammatory cells
(Section 6.3.5.3). Inflammation was accompanied by injury which is discussed in
Section 6.3.5.3.
Kodavanti et al. (2005, 087946) reported two sets of studies involving fine CAPs
exposure during fall months in RTP, NC. In the first study, SH rats were exposed to filtered
air or CAPs (mean mass concentration range 1,138-1,765 jug/m3; less than 2.5 jum in size) for
4 h and analyzed 1-3 h later. No increase in BALF inflammatory cells or other measured
parameter was observed. In the second study, SH and WKY rats were exposed to filtered air
or CAPs (mean mass concentration range 144-2,758 jug/m3; less than 2.5 jum in size) for
4 h/day on 2 consecutive days and analyzed 1 day afterward. Differences in baseline
July 2009
6-171
DRAFT-DO NOT CITE OR QUOTE

-------
parameters were noted for the two rat strains since SH rats had greater numbers of BALF
neutrophils than WKY rats. Following the 2-day CAPs exposure, increased BALF
neutrophils were observed in the WKY rats but not in the SH rats compared with filtered
air controls. Inflammation was not accompanied by increases in BALF markers of injury
(see Section 6.3.5.3).
Two CAPs studies involving SH rats were conducted in the Netherlands. In the first,
SH rats were exposed by nose-only inhalation to CAPs (ranging in concentration from
270-3,660 jug/m3 and in size from 0.15-2.5 jum) from 3 different sites in the Netherlands
(suburban, industrial and near-freeway) for 6 h (Cassee et al., 2005, 087962). Increased
numbers of neutrophils were observed in BALF 2 days post-exposure compared to air
controls. When CAPs exposure was used as a binary term, the relationship between CAPs
concentration and number of PMN in BALF was statistically significant. In contrast,
Kooter et al. (2006, 097547) reported no changes in markers of pulmonary inflammation
measured 18 h after a 2-day exposure (6 h/day) of SH rats to fine or fine+ultrafine CAPs
from sites in the Netherlands (mean mass concentration range 399-3613 and
269-556 |Jg/m3, respectively! fine CAPs site in Bilthoven and fine+ultrafine CAPs site in
freeway tunnel in Hendrik Ido Ambacht).
Pulmonary inflammation was investigated in 2 studies using a rat model of
pulmonary hypertension (i.e., SD rats pre-treated with monocrotaline). In the first study,
rats were exposed to fine CAPs from an urban high traffic area in Taiwan (mean mass
concentration of 371 |ug/m3) (Lei et al., 2004, 087999) for 6 h/day on 3 consecutive days and
BALF was collected 2 days later. A statistically significant increase in total cells and
neutrophils was observed in BALF. Levels of TNFa and IL-6 in the BALF were not altered
by CAPs exposure. In the second study, rats were exposed to fine CAPs (mean mass
concentration 315.6 and 684.5 |Jg/m3 for 6 and 4.5 h, respectively! Chung-Li area, Taiwan)
during a dust storm event occurring 3/18-3/19/2002 (Lei et al., 2004, 087884). Only one
animal served as control during the 6"h exposure (from 2100-300 on the first exposure day)
so results were combined with that of3 control animals from the 4.5"h exposure (from
300-730) on the second exposure day. A statistically significant increase in total cells and
neutrophils in BALF occurred in both CAPs-exposed groups. In addition, increases in BALF
IL-6 and markers of injury (see Section 6.3.5.3) were observed as a function of CAPs
exposure.
July 2009
6-172
DRAFT-DO NOT CITE OR QUOTE

-------
In summary, pulmonary inflammation was noted in all 3 studies involving multiday
exposure of healthy rats to CAPs from different locations. No pulmonary inflammation was
seen in one study of SH rats exposed to CAPs for 4 h and analyzed 1-3 h later. In studies
involving multiday exposure of SH rats, one demonstrated pulmonary inflammation while
two did not. In the rat monocrotaline model of pulmonary hypertension, both single "day and
multiday exposures to CAPs resulted in mild pulmonary inflammation.
On-Road Exposures
In a study by Elder et al. (2004, 087354) (21 months) were exposed to on-road
highway aerosols (particle concentration range 0.95-3.13 x 10B particles/cm3; mass
concentration estimated to be 37-106 |u,g/m3; Interstate 90 between Rochester and Buffalo,
NY) for 6 h on 1 or 3 consecutive days. No increase in BALF inflammatory cells was
observed 18 h post-exposure in any of the treatment groups.
Urban Air
To evaluate inflammatory responses to ambient particles from vehicles, Wistar rats
were exposed to ambient urban air from a high traffic site (concentration range
22-225 ug/m3 PMio! Porto Alegre, Brazil) or to the same air which was filtered to remove the
PM (Pereira et al., 2007, 156019). Concentrations of gases were not reported. Compared
with filtered air controls, a significant increase in total number of BALF cells was observed
24 h following the 20 h continuous exposure but not following the 6 h of exposure to
unfiltered urban air.
Diesel Exhaust
The 2004 PM AQCD (U.S. EPA, 2004, 056905) summarized findings of the 2002 EPA
Diesel Document regarding the health effects of DE. Short-term inhalation exposure to low
levels of DE results in the accumulation of DPM in lung tissue, pulmonary inflammation
and alveolar macrophage aggregation and accumulation near the terminal bronchioles.
More recent studies are summarized below.
Pulmonary inflammatory responses were investigated in C57BL/6 mice exposed to
diesel engine emissions 7 h/day for 6 consecutive days (Harrod et al., 2003, 097046).
Compared with controls, inflammatory cell counts in BALF were increased in mice exposed
to the higher concentration of DE (1,000 ug/m3 DEP) but not in mice exposed to the lower
July 2009
6-173
DRAFT-DO NOT CITE OR QUOTE

-------
concentration of DE (30 Lig/m3 DEP). Concentrations of gases present in the higher dose DE
were reported to be 43 ppm NOx, 20 ppm CO and 364 ppb SO2.
In a second study evaluating DE effects on BALF inflammatory cells, no increases in
numbers of neutrophils, lymphocytes or eosinophils were observed in BALB/c mice exposed
by inhalation to 500 or 2000 |ug/m3 DEP for 4 h/day on 5 consecutive days (Stevens et al.,
2008, 157010). Concentrations of gases reported in this study were 4.2 ppm CO, 9.2 ppm
NO, 1.1 ppm NO2, and 0.2 ppm SO2 for the higher concentration of DE. Transcriptional
microarray analysis demonstrated upregulation of chemokine and inflammatory cytokine
genes, as well as genes involved in growth and differentiation pathways, in response to the
higher concentration of DE. No gene expression results were reported for the lower
concentration of DE. Sensitization and challenge with ovalbumin significantly altered these
findings (see Section 6.3.6.2). These results demonstrate that changes in gene expression
can occur in the absence of measurable pulmonary inflammation or injury markers (see
Section 6.3.5.3).
Li et al. (2007, 155929) exposed mice to clean air or to low dose DE (DEP 100 |u,g/m3)
for 7 h/day and 5 days/wk for 1, 4 and 8 wk as described in Section 6.3.2.3. Analysis of
BALF and histology of lung tissues was carried out at day 0 and after 1, 4 and 8 wk of
exposure. Total numbers of cells and macrophages in BALF were significantly increased in
C57BL/6 mice but not in BALB/c mice after 1-wk exposure to DE compared with 0 day
controls. Neutrophils and lymphocytes were increased after 1-wk exposure to DE in both
strains compared with 0 day controls. Differences in BALF cytokines were also noted
between the 2 strains after 1-wk exposure to DE. No changes were observed by histological
analysis. Pulmonary function and oxidative responses were also evaluated (Sections 6.3.2.3
and 6.3.4.2) Long-term exposure responses are discussed in Sections 7.3.2.2, 7.3.3.2 and
7.3.4.1.
Healthy Fisher 344 rats and A/J mice were exposed to DE containing 30, 100, 300 and
1000 ug/m3 PM by whole body inhalation for 6 h/day, 7 days/wk for either 1 wk or 6 months
in a study by Reed et al. (2004, 055625). Concentrations of gases were reported to be from
2.0-45.3 ppm NO, 0.2-4.0 ppm NO2, 1.5-29.8 ppm CO and 8-365 ppb for SO2 in these
exposures. One week of exposure resulted in no measurable effects on pulmonary
inflammation. Long-term exposure responses are discussed in Section 7.3.3.2.
In a study by Wong et al. (2003, 097707) and Witten et al. (2005, 087485), Fisher
344/NH rats were exposed nose-only to filtered room air or to DE at concentrations of
July 2009
6-174
DRAFT-DO NOT CITE OR QUOTE

-------
35.3 ju,g/m3and 669.3 jug/m3 DEP (particle size range 7.2-294.3 rim) for 4 h/day and 5
days/wk for 3 wk. Gases associated with the high dose exposure were reported to be
3.59 ppm NO, 3.69 ppm NOx, 0.1 ppm NO2, 2.95 ppm CO, 518.96 ppm CO2 and 0.031 ppm
total hydrocarbon. The focus of this study was on the possible role of neurogenic
inflammation in mediating responses to DE. Neurogenic inflammation is characterized by
both the influx of inflammatory cells and plasma extravasation into the lungs following the
release of neuropeptides from bronchopulmonary Ofibers. Pulmonary inflammation was
evaluated by histological analysis of lung tissue at the end of the 3-wk exposure period.
Following high, but not low, dose-exposure to DE, a large number of alveolar macrophages
was found in the lungs. Small black particles, presumably DEP, were found in the
cytoplasm of these alveolar macrophages. Perivascular cuffing consisting of mononuclear
cells was also observed in high dose-exposed animals. Influx of neutrophils or eosinophils
was not seen although mast cell number was increased in high-dose exposed animals.
Pulmonary plasma extravasation was measured by the 99mTechnecium-albumin technique
and found to be dose-dependently increased in the bronchi and lung parenchyma. Alveolar
edema was also observed by histopathology in high dose-exposed animals. A significant
decrease in substance P content in lung tissue was reported in DE-exposed rats. These
responses initially suggested that DE resulted in stimulation of Ofibers and activation of a
local neuron reflex resulting in the repeated release of the stored neuropeptide substance P.
Subsequent experiments were conducted using capsaicin pretreatment, which inhibits
neurogenic inflammation by activating Ofibers and causing the depletion of neuropeptide
stores. Pretreatment with capsaicin was found to reduce the influx of inflammatory cells
but not plasma extravasation in response to DE. Hence, DE is unlikely to act through
bronchopulmonary Ofibers to cause neurogenic inflammation in this model, although there
may be a different role for bronchopulmonary Ofibers in mediating the inflammatory cell
influx.
Stimulation of bronchopulmonary Ofibers can result in activation of both local and
CNS reflexes through vagal parasympathetic pathways. McQueen et al. (2007, 096266)
investigated the role of vagally-mediated pathways in acute inflammatory responses to
DEP. A statistically significant increase in BAL neutrophils was observed 6 h after IT
treatment of anesthetized Wistar rats with 500 |ug DEP (SRM2975). This response was
blocked by severing the vagus nerve or pretreatment with atropine (McQueen et al., 2007,
096266). Similarly, atropine treatment blocked the increase in BAL neutrophils seen 6 h
July 2009
6-175
DRAFT-DO NOT CITE OR QUOTE

-------
after DEP exposure in conscious Wistar rats. These results provide evidence for the
involvement of a pulmonary vagal reflex in the inflammatory response to DEP.
In summary, several studies demonstrate that short-term inhalation exposure to DE
(100-1,000 jug/m3 DEP) causes pulmonary inflammation in rodents. No attempt was made
in these studies to determine whether the responses were due to PM components or to
gaseous components. However, PM from DE was found to be capable of inducing an
inflammatory response, as demonstrated by the one IT instillation study described above.
Evidence was presented suggesting that DEP may act through bronchopulmonary Ofibers
to stimulate pulmonary inflammation.
Gasoline Emissions and Road Dust
Healthy male Swiss mice were exposed to gasoline exhaust (635 |ug/m3 PM and
associated gases) or filtered air for 15 min/day for 7, 14, and 21 days (Sureshkumar et al.,
2005, 088306). BALF was collected for analysis 1-h after the last exposure. Histological
analysis was also carried out at 7, 14, and 21 days. The number of leukocytes in BALF was
increased after exposure to gasoline exhaust but this increase did not achieve statistical
significance. However, levels of the pro-inflammatory cytokines TNFa and IL-6 were
significantly increased in BALF following 14 and 21 days of exposure. Furthermore,
inflammatory cell infiltrate in the peribronchiolar and alveolar regions were observed by
histology. Evidence of lung injury was also found (see Section 6.3.5.3). In this study, BALF
analysis of inflammatory cells was a less sensitive indicator of pulmonary inflammation
than BALF analysis of cytokines and histological analysis of lung tissue. Unfortunately
results of this study cannot entirely be attributed to the presence of PM in the gasoline
exhaust since 0.11 mg/m3 SOx, 0.49 mg/m3 of NOx and 18.7 ppm of CO were also present
during exposure.
Using ApoE '" mice on a high-fat diet, Campen et al. (2006, 096879) studied the impact
of inhaled gasoline emissions and road dust (6 h/day x 3 day) on pulmonary inflammation.
Moreover, the investigators used a high efficiency particle filter to compare the whole
exhaust with an atmosphere containing only the gaseous components. For gasoline
emissions, the PM-containing atmosphere (PM mean concentration 61 |Jg/m3; NOx mean
concentration 18.8 ppm! CO mean concentration 80 ppm) failed to increase numbers of
inflammatory cells in BALF collected 18 h after the last exposure. However, a statistically
July 2009
6-176
DRAFT-DO NOT CITE OR QUOTE

-------
significant increase in total cells and macrophages was observed in response to resuspended
road dust (PM2.5) at 3500 |Jg/m3, but not at 500 |u,g/m3.
Model Particles
In a study by Elder et al. (2004, 055642), pulmonary inflammation was investigated
in two compromised, aged animal models (11-14 mo old SH and 23 mo old Fischer 344)
exposed by inhalation to ultrafine CB (count median diameter = 36 nm) at a relevant
concentration (150 |Jg/m3). No changes in BALF cells were seen 24 h post-exposure in either
model.
An increase in BALF neutrophils was observed at 24 h but not at 4 h in WKY rats
exposed to ultrafine carbon particles (median particle size 38 nm! mass concentration 180
(jg/m3; mean number concentration 1.6 x 107 particles/cm3) for up to 24 h (Harder et al.,
2005, 087371). Changes in HR and HRV demonstrated in this study (see Section 6.2.1.3)
occurred much more rapidly than the inflammatory response.
No evidence of pulmonary inflammation was found by analysis of BALF or pulmonary
histopathology 1 or 3 days following 24-h exposure of SH rats to ultrafine carbon particles
under similar conditions (median particle size 31 nm! mass concentration 172 (Jg/m3; mean
number concentration 9.0 x 106 particles/cm3) (Upadhyay et al., 2008, 159345). However
increased expression of HOI, ET-1, ETa and ETb, tPA and, plasminogen activator-1 was
found in lung tissue 3 days following exposure.
In a study by Gilmour et al. (2004, 054175), adult Wistar rats were exposed for 7 h to
fine and ultrafine CB particles (mean mass concentration 1400 and 1660 (Jg/m3 for fine and
ultrafine CB, respectively! mean number concentration 3.8 x 103 and 5.2 x 104
particles/cm3, respectively! count median aerodynamic diameter 114 nm and 268 nm,
respectively). Both treatments resulted in increased BALF neutrophils 16 h post-exposure,
with the UFPs having the greater response. UFPs also increased total BALF leukocytes and
macrophage inflammatory protein-2 mRNAin BALF cells. Although these exposures may
not be relevant to ambient exposures, this study demonstrated the greater propensity of
ultrafine CB particles to cause a pro-inflammatory response compared with fine CB
particles.
In a study by Last et al. (2004, 097334), mice were exposed to 250 |u,g/m3
laboratory-generated iron-soot over a 2-wk period as described in Section 6.3.2.3. BALF was
collected 1-h after the last exposure and analyzed for total cells. No increase in total cell
July 2009
6-177
DRAFT-DO NOT CITE OR QUOTE

-------
number was observed following iron-soot exposure. Other findings of this study are
described in Sections 6.3.2.3 and 6.3.5.3.
Pinkerton et al. (2008, 190471) exposed young adult male SD rats to filtered air, iron,
soot or iron-soot. Increased levels of the pro-inflammatory cytokine IL-16 were observed in
lung tissue of rats exposed for 6 h/day for 3 days to 90 |ug/m3, but not 57 |u.g/m3, iron. No
change in BALF inflammatory cells was observed after exposure to to 57 ug/ma or 90 ug/m3
iron. Exposures to 250 jug/m3 soot in combination with 45 jug/m3 iron also resulted in
increased levels of lung IL-16 and activation of the transcription factor NFkB. Levels of
lung IL-16 were increased in neonatal rats exposed to 250 jug/m3 soot in combination with
100, but not 30, jug/m3 iron. This study is described in greater detail in Sections 6.1.4.2.
6.3.4. Oxidative Responses
The results of a small number of controlled human exposure and toxicological studies
presented in the 2004 PM AQCD (U.S. EPA, 2004, 056905) provided some initial evidence of
an association between exposure to PM and pulmonary oxidative stress. Recent controlled
human exposure studies have provided support to previous findings of an increase in
markers of pulmonary oxidative stress following exposure to DE, and one new study has
observed a similar effect following controlled exposure to WS. New findings from
toxicological studies provide further evidence that oxidative species are involved in PM-
mediated effects. No epidemiologic studies have evaluated the association between PM
concentration and pulmonary oxidative response.
6.3.4.1. Controlled Human Exposure Studies
Two studies cited in the 2004 PM AQCD (U.S. EPA, 2004, 056905) observed effects on
markers of airway oxidative response in healthy adults following controlled exposures to
fresh DE or resuspended DEP (Blomberg et al., 1998, 051246; Nightingale et al., 2000,
011659). Several recent studies are described below which have further evaluated the
oxidative response following exposure to particles in human volunteers.
Diesel Exhaust
Pourazar et al. (Pourazar et al., 2005, 088305) exposed 15 adults (11 males and 4
females) for 1 h to air or DE (PMio concentration 300 (Jg/m3) in a controlled cross-over study.
Bronchoscopy with airway biopsy was performed 6 h after exposure. The expression of
July 2009
6-178
DRAFT-DO NOT CITE OR QUOTE

-------
NF-kB, AP-1 (c-jun and c-fos), p38, and JN K in bronchial epithelium was quantified using
immunohistochemical staining. DE was observed to significantly increase nuclear
translocation of NF-kB, AP-1, phosphorylated p38, and phosphorylated JN K; however, the
findings of this study require confirmation with more quantitative methods such as
Western blot analysis. The observed activation of redox-sensitive transcription factors by
DE may result in the induction of pro-inflammatory cytokines. There is some evidence to
suggest that this bronchial response to DE is mediated through the epidermal growth factor
receptor signaling pathway (Pourazar et al., 2008, 156884). Behndig et al. (Behndig et al.,
2006, 088286) evaluated the upregulation of endogenous antioxidant defenses following
exposure to DE (100 (Jg/m3 PMio) in a group of 15 healthy adults. Increases in urate and
reduced GSH were observed in alveolar lavage, but not bronchial wash, 18 h after exposure.
In a study utilizing the same exposure protocol, Mudway et al. (2004, 180208) observed an
increase in GSH and ascorbate in nasal lavage fluid 6 h following exposure to DE in a group
of 25 healthy adults.
Wood Smoke
Barregard et al. (2008, 155675) observed a significant increase in malondialdehyde
levels in breath condensate of healthy volunteers (n = 13) immediately following and 20 h
after a 4-h exposure to WS (240-280 )Jg/m3).
Instillation
Schaumann et al. (2004, 087966) demonstrated an increased oxidant radical
generation of BAL cells following instillation of urban particles compared with instillation
of particles collected in a rural area. The authors suggested that this difference was likely
due to the greater concentration of transition metals found in the urban particles.
Summary of Controlled Human Exposure Study Findings for Blood Pulmonary Oxidative
Responses
Taken together, these studies suggest that short-term exposure to PM at near
ambient levels may produce mild oxidative stress in the lung. Limited data suggest that
proximal and distal lung regions may be subject to different degrees of oxidative stress
during exposures to different pollutant particles.
July 2009
6-179
DRAFT-DO NOT CITE OR QUOTE

-------
6.3.4.2. Toxicological Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) reported one study which provided
evidence that ROS were involved in PM-mediated responses. This particular study used
pre-treatment with the antioxidant DMTU to block the neutrophilic response to ROFA.
More recently, several studies evaluated the effects of PM exposure on pulmonary oxidative
stress. Oxidative stress can be directly determined by measuring ROS or oxidation products
of lipids and proteins. An indirect assay involves measurement of the enzyme HO-1 or of
the antioxidant enzymes SOD or catalase, all of which can be induced by oxidative stress.
Antioxidant interventions which inhibit or prevent responses are a further indirect
measure of oxidative stress playing a role in the pathway of interest.
CAPs
Gurgueira et al. (2002, 036535) measured oxidative stress as in situ
chemiluminescence (CL). Immediately following a 5-h CAPs exposure (PM2.5; mean mass
concentration range 99.6-957.5 |Jg/m3; Boston, MA) increased CL was observed in lungs of
CAPs-exposed SD rats. CL evaluated after CAPs exposure durations of 3 h was also
increased but did not achieve statistical significance compared to the filtered air group.
When animals were allowed to recover for 24 h following the 5-h CAPs exposure, CL levels
returned to control values. Interestingly, a decrease in lung CL was observed in rats
breathing filtered air for 3 days compared with rats breathing room air for the same
duration. Exposure to CAPs for 3 and 5 h also increased lung wet/dry ratios, indicating the
presence of mild edema. To compare potential particle-induced differences in in situ CL,
rats were exposed to ROFA (1.7 mg/m3 for 30 min) or CB (170 (Jg/m3 for 5 h). Only the
ROFA-treated animals exhibited increased CL in lung tissue. Additionally, levels of
antioxidant enzymes in the lung (MnSOD and catalase) were increased in CAPs-exposed
rats. A CAPs-associated increase in CL was also seen in the heart (see Section 6.2.9.3) but
not the liver.
In a similar study, Rhoden et al. (2004, 087969) exposed SD rats for 5 h to CAPs from
Boston (mean mass concentration 1228 ug/m3) or to filtered air. Significant increases in
TBARS (a measure of lipid peroxidation) and protein carbonyl content (a measure of protein
oxidation) were observed 24 h post-exposure to CAPs. Pretreatment with the thiol
antioxidant NAC (50 mg/kg i.p.) 1-h prior to exposure prevented not only the lipid and
protein oxidation observed in response to CAPs, but also the increase in BALF neutrophils
July 2009
6-180
DRAFT-DO NOT CITE OR QUOTE

-------
and and pulmonary edema in this model (see Sections 6.3.3.3 and 6.3.5.3). Results of this
study demonstrate the key role played by oxidative stress in these CAPs-mediated effects.
A later study by Rhoden et al. (2008, 190475) investigated the role of superoxide in
mediating pulmonary inflammation following exposure to ambient air particles. In this
study, adult SD rats were IT exposed to 1 mg of SRM1649. Two h prior to exposure, half of
the rats were pretreated with the membrane-permeable SOD mimetic MnTBAP (10 mg/kg,
i.p.). MnTBAP abrogated the inflammatory response, measured by increased BALF
inflammatory cells, and the increase in lung superoxide, measured by CL observed 4 h
following exposure to urban air particles.
Kooter et al. (2006, 097547) reported an increase in HOI in BALF and lung tissue
measured 18 h after a 2-day exposure (6 h/day) of SH rats to fine or fine+ultrafine CAPs
(mean mass concentration range 399-3613 and 269-556 |Jg/m3, respectively! fine CAPs site
in Bilthoven and ultrafine+fine site in freeway tunnel in Hendrik Ido Ambacht, the
Netherlands). This occurred in the absence of any measurable pulmonary inflammation
(see Section 6.3.3.3).
Urban Air
To evaluate oxidative stress responses to ambient particles from vehicles, Wistar rats
were exposed to ambient urban air from a high traffic site (concentration range
22-225 ug/m3 PMio! Porto Alegre, Brazil) or to the same air which was filtered to remove the
PM (Pereira et al., 2007, 156019). Several exposures regimens were carried out: 6 and 20-h
continuous exposures or to intermittent exposures of 5 h/day for 4 consecutive days. A
significant increase in lipid peroxidation (measured as malondialdehyde) was seen in lung
tissue immediately following the 20"h continuous exposure but not following the 6"h
exposure or the intermittent exposures.
Diesel Exhaust
Li et al. (2007, 155929) exposed mice to clean air or to low dose DE (DEP 100 |u,g/m3)
for 7 h/day and 5 days/wk for 1, 4 and 8 wk as described in Section 6.3.2.3. HOI mRNA and
protein were increased in lung tissues of both mouse strains after 1 wk of DE exposure. In
addition, AHR and changes in BAL cells and cytokines were observed (see Sections 6.3.2.3
and 6.3.3.3). Pretreatment with the thiol antioxidant NAC (320 mg/kg, i.p.) on days 1-5 of
DE exposure greatly attenuated the AHR and inflammatory response seen after 1 wk of DE
exposure. Long-term responses are discussed in Sections 7.3.2.2, 7.3.3.2 and 7.3.4.1.
July 2009
6-181
DRAFT-DO NOT CITE OR QUOTE

-------
A study by Whitekus et al. (2002, 157142) investigated the adjuvant effects of DEP in
an allergic animal model and is discussed in detail below (see Section 6.3.6.3). Intervention
with the thiol antioxidants bucillamine and NAC inhibited the increases in allergen-specific
IgE and IgGi as well as the increases in protein carbonyl and lipid hydroperoxides in the
lung following DE exposure.
Gasoline Exhaust
Pulmonary oxidative stress was evaluated by measurement of CL and TBARS
following exposure of SD rats to gasoline engine exhaust (Seagrave et al., 2008, 191990).
Animals were exposed for 6 h in a nose-only inhalation exposure system. PM mass
concentration was reported to be 60 |Jg/m3; count median diameter 20 nm! mass median
diameterl50 nm! while the concentrations of gaseous copollutants were 104 ppm CO,
16.7 ppm NO, 1.1 ppm NO2 and 1.0 ppm SO2. A statistically significant increase in lung CL
was observed without a concomitant increase in lung TBARS. Discordant results were also
observed for road dust exposures in the heart (see Section 6.2.9.3). The discrepancy between
oxidative stress indicators suggests that the responses may follow different time courses.
Furthermore, no CL was seen when the gasoline exhaust was filtered to remove the
particulate fraction.
Model Particles
Increased expression of HO-1 was observed in lung tissue 3 days following 24-h
exposure of SH rats to ultrafine carbon particles (median particle size 31 nm! mass
concentration 172 (Jg/m3; mean number concentration 9.0 x 106 particles/cm3) despite no
evidence of pulmonary inflammation (see Section 6.3.3.3) (Upadhyay et al., 2008, 159345)
In a study conducted by Pinkerton et al. (2008, 190171) young adult male SD rats
were exposed to filtered air, soot, iron or iron-soot for 6 h/day for 3 days. The iron
particulates were mainly less than 100 nm aerodynamic diameter, while the soot
particulates were initially 20-40 nm in diameter but formed clusters of 100-200 nm in
diameter. The size-distribution of iron-soot particulates was bimodal over 10-250 nm and
averaged 70-80 nm in diameter. Rats were exposed 6 h/day for 3 days to 45, 57 and 90 |ug/m3
iron or to 250 ug/m3 soot alone or in combination with 45 ug/m3 iron. A statistically
significant decrease in total antioxidant power and a statistically significant increase in
glutathione-S-transferase activity were observed in lung tissue from rats exposed to 90
ug/m3 iron. This high concentration iron exposure also resulted in increased levels of
July 2009
6-182
DRAFT-DO NOT CITE OR QUOTE

-------
ferritin protein in lung tissue, indicating the presence of free iron which has the potential to
redox cycle and cause oxidative stress. Lung tissue total antioxidant power was decreased
and glutathione redox ratio was increased by the combined exposure to 250 ug/m3 soot and
45 jug/m3 iron. The iron-soot exposure also increased oxidized glutathione in BALF and lung
tissue. These results demonstrate that co-exposure to soot enhanced iron-mediated
oxidative stress. Furthermore, co-exposure to soot and iron resulted in increased expression
of cytochrome P450 isozymes CYP1A1 and CYP2E1 in lung tissue, an effect not observed in
response to either agent alone. Other endpoints of this study are described in Sections
6.1.3.3.
In a parallel study, Pinkerton et al. (2008, 190471) exposed neonatal male SD rats to
iron-soot or filtered air 6 h/day for 3 days during the second and fourth week of life. Both 30
ug/m3 and 100 ug/m3 iron in combination with 250 ug/m3 soot resulted in increased BALF
oxidized glutathione, glutathione redox ratio and glutathione-S-transferase activity and
decreased total antioxidant power. The higher concentration exposure resulted in increased
ferritin expression in lung tissue. Effects on cellular proliferation in specific regions of the
lung were also noted as described in Section 6.1.5.3.
Nurkiewicz et al. (2009, 191961) exposed SD rats to fine (count median diameter
710 nm) and ultrafine (count median diameter 100 nm) TiC>2 particles via aerosol inhalation
at concentrations of 1.5-16 mg/m3 for 240-720 min. These exposures were chosen in order to
produce deposition of 4-90 |Jg/rat, which was demonstrated in a previous study to result in
different degrees of impaired microvascular function (Nurkiewicz et al., 2008, 156816).
Histopathological analysis of lung tissue did not find any significant inflammation although
particle accumulation in alveolar macrophages and a frequent association of alveolar
macrophage with the alveolar wall was observed 24 h following exposure (Nurkiewicz et al.,
2008, 156816). Although the main focus of the more recent study was on effects of TiC>2 on
NO production and microvascular reactivity in the spinotrapezius muscle (see
Section 6.2.4.3), the presence of nitrotyrosine was determined in both lung tissue and
spinotrapezius muscle as a measure of peroxynitrite formation. Peroxynitrite formation
occurs mainly as a result of the rapid reaction of NO with superoxide and suggests an
increase in local superoxide production. The area of lung tissue containing nitrotyrosine
immunoreactivity increased 3-fold 24 h following exposure to 10 |Jg ultrafine Ti02.
Nitrotyrosine immunore activity was localized in inflammatory cells found in the alveolar
region of the lung.
July 2009
6-183
DRAFT-DO NOT CITE OR QUOTE

-------
6.3.5. Pulmonary Injury
The 2004 PM AQCD (U.S. EPA, 2004, 056905) presented evidence from several
toxicological studies of small PM-induced increases in markers of pulmonary injury
including thickening of alveolar walls and increases in BALF protein. These findings are
consistent with the results of recent toxicological studies demonstrating mild pulmonary
injury accompanying inflammatory responses to CAPs. One recent epidemiologic study has
also observed a positive association between PM and urinary concentrations of lung Clara
cell protein.
6.3.5.1.	Epidemiologic Studies
One epidemiologic study examined biomarkers of pulmonary injury. The mean
concentration data from this study are characterized in Table 6" 7. Tim on en et al. (2004,
087915) enrolled subjects with coronary heart disease in Amsterdam (n = 37), Erfurt,
Germany (n = 47) and Helsinki (n = 47) to study daily variation in PM and urinary
concentrations of lung Clara cell protein (CC16). No associations were seen between the
particle number concentration of the smallest particles (NCo.oi-o.i) and CC16. Significant
associations with NCo.i-i and PM2.5 (which were strongly correlated with each other
[r = 0.8]) were seen only for Helsinki subjects: same day, lag 3 and 5-day mean NC0.1-1
increases of 1000/cm3 were associated with increases in In (CC16/creatinine) of 15.5% (95%
CI: 0.001-30.9), 17.4% (95% CI: 3.4-31.4), and 43.2% (95% CI: 17.4-69.0), respectively.
Similar associations were seen for 10 |ug/m3 increases in PM2.5: lag 0 and 5-day mean PM2.5
were associated with increases in In (CC16/creatinine) of 23.3% (95% CI: 6.3-40.3) and
38.8% (95% CI: 15.8-61.8), respectively.
6.3.5.2.	Controlled Human Exposure Studies
No studies of controlled human exposures presented in the 2004 PM AQCD
(U.S. EPA, 2004, 056905) specifically examined the effect of PM on pulmonary injury.
However, several recent studies have evaluated changes in markers of injury and increased
alveolar permeability following exposures to various types of particles.
Urban Traffic Particles
Brauner et al. (2009, 191179) evaluated the effect of exposure to urban traffic
particles (24-h exposure, PM2.5 9.7 (Jg/m3) on the integrity of the alveolar epithelial
July 2009
6-184
DRAFT-DO NOT CITE OR QUOTE

-------
membrane in a group of 29 healthy adults, with and without exercise. Following 2.5 h of
exposure, alveolar epithelial permeability was assessed by measuring the pulmonary
clearance of 99mTc-DTPA, which was administered as an aerosol during 3 min of tidal
breathing. While pulmonary clearance of 99mTc-DTPA was observed to increase following
exercise, there was no significant difference in clearance between exposure to urban traffic
particles and filtered air. In addition, PM exposure was not observed to affect the level of
CC16 in plasma or urine at 6 or 24 h after the start of exposure.
Diesel Exhaust
Relative to filtered air, exposure for 1 h to DE (300 |jg/m3 particle concentration) was
not observed to affect the plasma CC16 concentration at 6 or 24 h post exposure in a group
of 15 former smokers with COPD (Blomberg et al., 2005, 191991).
Wood Smoke
In a study examining the respiratory effects of WS, Barregard et al. (2008, 155675)
exposed two groups of healthy adults in separate 4-h sessions to WS with median particle
concentrations of 243 and 279 |Jg/m3. At 20 h post-exposure, the mean serum CC16
concentration was significantly higher after exposure to WS when compared with filtered
air. However, when the analysis was stratified by exposure session, a statistically
significant effect of WS on serum CC16 was observed in the subjects in session 1, but not
those in session 2. It is interesting to note that while the mean particle concentration was
only slightly higher in session 1, the mean particle number in session 1 was almost 90%
higher than the particle number in session 2, with geometric mean particle diameters of 42
and 112 nm, respectively.
Summary of Controlled Human Exposure Study Findings for Pulmonary Injury
The findings from these studies provide limited evidence to suggest that exposures to
particles may increase markers of pulmonary injury in healthy adults.
6.3.5.3. Toxicological Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) reported mild increases in BALF
protein, a marker of pulmonary injury, in several studies involving inhalation exposure to
CAPs. In addition, histopathological analysis demonstrated that the bronchoalveolar
junction was the site of the greatest inflammation following CAPs exposure. Low level
July 2009
6-185
DRAFT-DO NOT CITE OR QUOTE

-------
exposure to DE was associated with Type 2 cell proliferation and thickening of alveolar
walls near alveolar macrophages according to the 2002 EPA Diesel Document (U.S. EPA,
2002, 042866). In addition, IT instillation of fly ash and metal-containing PM generally
caused pulmonary injury as measured by increases in BALF protein, LDH and albumin.
Proliferation of bronchiolar epithelium was also noted. More recent studies of BALF
markers of pulmonary injury and histopathological analysis of lung tissue are summarized
below.
BALF Markers of Pulmonary Injury and Increased Permeability
CAPs
Kodavanti et al. (2005, 087946) exposed SH and WKY rats to filtered air or CAPs
from RTP, NC as described in Section 6.3.3.3. Differences in baseline parameters were
noted for the two rat strains since SH rats had greater levels of protein and lower levels of
levels of LDH, NAG, ascorbate and uric acid in the BALF than WKY rats. One day after the
2-day CAPs exposure, increased levels of GGT were observed in BALF (a marker of
epithelial injury) of SH rats but not WKY rats compared with filtered air controls. Injury
was not accompanied by inflammation (see Section 6.3.3.3).
In a study by Cassee et al. (2005, 087962), SH rats were exposed for 6 h by nose-only
inhalation to CAPs from 3 different sites in the Netherlands as described in Section 6.3.3.3.
The pulmonary injury marker CC16 was increased in BALF 2 days following CAPs
exposure. Inflammation was also observed (see Section 6.3.3.3).
Gurgueira et al. (2002, 036535) exposed to Boston, MA, CAPs as described in
Section 6.3.4.2 and reported a small but statistically significant increase in lung wet/dry
ratios after 3 and 5 h of exposure, indicating the presence of mild edema. This response was
accompanied by increased oxidative stress as measured by in situ chemiluminescence (CL)
(see Section 6.3.4.2). In a similar study, Rhoden et al. (2004, 087969) reported an increase
in lung wet/dry ratio 24 h following a 5"h exposure to Boston CAPs which was diminished
by pre-treatment of the antioxidant NAC (see Section 6.3.4.2).
Pulmonary injury was investigated in 2 studies using a rat model of pulmonary
hypertension (SD rats pre-treated with monocrotaline which is described in greater detail
in Section 6.3.3.3 (Lei et al., 2004, 087999). Significant increases in BALF LDH and protein
were observed in response to CAPs. Pulmonary inflammation was observed in both of these
studies (see Section 6.3.3.3).
July 2009
6-186
DRAFT-DO NOT CITE OR QUOTE

-------
Diesel Exhaust
In a study evaluating the effects of DE, no changes were observed in BALF protein
and LDH in mice exposed by inhalation to concentrations of 50 and 2000 jug/m3 DEP for
4 h/day on 5 consecutive days as described in Section 6.3.3.3 (Stevens et al., 2008, 157010).
Changes in gene expression were observed in the higher exposure group. This study
demonstrates that changes in gene expression can occur in the absence of measurable
markers of injury or pulmonary inflammation.
In a study by Wong et al. (2003, 097707) and Witten et al. (Witten et al., 2005,
087485), rats were exposed nose-only to filtered room air or to DE over a 3-wk period. This
study, focusing on neurogenic inflammation, is described in greater detail in Section 6.3.3.3.
Pulmonary plasma extravasation was measured by the 99mTechnecium-albumin technique
and found to be dose-dependently increased in the bronchi and lung. Pretreatment with
capsaicin, which inhibits neurogenic inflammation by activating Ofibers and causing the
depletion of neuropeptide stores, did not reduce plasma extravasation following DE
exposure. Hence, DE is unlikely to act through bronchopulmonary Ofibers to cause
neurogenic inflammation in this model. Inflammatory responses measured in this study are
discussed in Section 6.3.3.3.
Gasoline Exhaust
Healthy male Swiss mice were exposed to gasoline exhaust (635 |ug/m3 PM and
associated gases) or filtered air for 15 min/day for 7, 14, and 21 days as described in
Section 6.3.3.3 (Sureshkumar et al., 2005, 088306). BALF was collected for analysis 1-h
after the last exposure. Statistically significant increases in BALF markers of lung injury,
alkaline phosphatase, gamma-glutamyl transferase and LDH, were observed at all time
points studied. Alveolar edema was noted following 14 and 21 days of exposure. Other
findings of this study including inflammation and histopathology are discussed in
Section 6.3.3.3.
Histopathology
CAPs
Histopathological changes were demonstrated in rats exposed for 5 h to Boston CAPs
as described in Section 6.3.3.3 (Rhoden et al., 2004, 087969). Slight bronchiolar
inflammation and thickened vessels at the bronchiole were observed 24 h post-exposure,
July 2009
6-187
DRAFT-DO NOT CITE OR QUOTE

-------
consistent with the influx of polymorphonuclear leukocytes observed in BALF (see
Section 6.3.3.3).
Diesel Exhaust
In a study by Wong et al. (2003, 097707) and Witten et al. (2005, 087485), rats were
exposed nose-only to filtered room air or to DE over a 3~wk period. This study, focusing on
neurogenic inflammation, is described in greater detail in Section 6.3.3.3. Pulmonary
inflammation was evaluated by histopathological analysis of lung tissue. Following high,
but not low, dose-exposure to DE, a large number of alveolar macrophages was found in the
lungs. Small black particles, presumably DEP, were found in the cytoplasm of these
alveolar macrophages. Perivascular cuffing consisting of mononuclear cells was also
observed in high dose-exposed animals. Influx of neutrophils or eosinophils was not seen
although mast cell number was increased. Other indices of injury demonstrated in this
study are described above.
Gasoline Exhaust
Another study, which is described in greater detail in Section 6.3.3.3, demonstrated
histopathological responses to gasoline exhaust in mice exposed to gasoline exhaust or
filtered air for 15 min/day for 7, 14, and 21 days (Sureshkumar et al., 2005, 088306).
Histological observations showed inflammatory cell infiltrate in the peribronchiolar and
alveolar region, alveolar edema and thickened alveolar septa at 14 and 21 days post-
exposure. Levels of pro-inflammatory cytokines and marker enzymes of lung damage were
also increased in BALF. The numbers of inflammatory cells in BALF was increased but not
significantly, demonstrating that BALF analysis of inflammatory cells was a less sensitive
indicator of pulmonary inflammation in this study than histophathological analysis. Other
indices of injury found in this study are described above.
Model Particles
In a study investigating the effects of iron-soot, mice were exposed to 250 |u,g/m3
laboratory-generated iron-soot as described in Sections 6.3.2.3 and 6.3.3.3 (Last et al., 2004,
097334). Analysis of airway collagen content was conducted by histology and by biochemical
analysis of microdissected airways. No increases in airway collagen content were found by
either method in mice exposed to iron-soot for 2 wk. Furthermore, no goblet cells were
observed in airways of air or iron-soot exposed animals. Other findings of this study are
described in Sections 6.3.2.3 and 6.3.3.3.
July 2009
6-188
DRAFT-DO NOT CITE OR QUOTE

-------
An interesting study demonstrating histopathological responses to PM in neonatal
rats was reported by Pinkerton et al. (2004, 087465). Rat pups (10 days old) were exposed to
soot and iron particles (mean mass concentration of 243 ug/m3; iron concentration 96 ug/rn3:
size range 10-50 rim) for 6 h/day on 3 consecutive days. Cell proliferation in different lung
regions was evaluated following bromodeoxyuridine injection 2-h prior to necropsy. The rate
of cell proliferation in the proximal alveolar region (immediately beyond the terminal
bronchioles) was significantly reduced in iron-soot exposed animals compared to controls.
This was a region-specific response since the rate of cell proliferation was not altered in the
terminal bronchioles or the general lung parenchyma. However alveolar septation, the
process by which alveoli are formed during development, and alveolar growth were not
altered by iron-soot exposure. Decreased cell viability and increased LDH was also noted in
BALF of neonatal rats (Pinkerton et al., 2008, 190471). The authors suggest the possibility
of greater susceptibility to air pollution during the critical postnatal lung development
period which occurs in animals and humans and that neonatal exposure to PM may
contribute to impaired lung growth seen in children.
Relative Toxicity of PM Size Fractions
Ambient PM Studies
A recently undertaken multinational project entitled "Chemical and biological
characterization of ambient thoracic coarse (PM10-2.5), fine (PM2.5-0.2), and UFPs (PM0.2) for
human health risk assessment in Europe" (PAMCHAR) takes a systematic approach to
expanding the present knowledge about the physiochemical and toxicological effects of
these three PM size fractions. Six European cities were selected that represented
contrasting ambient PM profiles: Helsinki, Duisburg, Prague, Amsterdam, Barcelona, and
Athens. For PM collected at all sites, PM10-2.5 induced the greatest pulmonary effects in
C57B1/6J mice intratracheally instilled with 1, 3, or 10 mg/kg of particles (Happo et al.,
2007, 096630). Dose-response relationships in BALF parameters measured 24-h post-IT
exposure, including BALF cell number and protein, were observed for all sites following
PM10-2.5 and neutrophils were the predominant cell type (Happo et al., 2007, 096630).
Prague PM10-2.5 exposure resulted in decreased macrophages in BALF at 12 h and
Amsterdam, Barcelona, and Athens PM10-2.5 induced lymphoplasmacytic cells in BALF
(Happo et al., 2007, 096630). No inflammatory responses were observed for ultrafine PM
measured 12-h after exposure. Protein was elevated for PM10-2.5 for all locations with the 10
July 2009
6-189
DRAFT-DO NOT CITE OR QUOTE

-------
mg/kg dose! Athens ultrafine PM induced protein release only at the two lowest doses 12 h
post-exposure. For TNF-a and IL-6, the greatest response was observed with PM10-2.5 4 h
following exposure (Happo et al., 2007, 096630). Ultrafine Duisburg PM exposure resulted
in elevated TNF-a for the 1 and 3 mg/kg doses. Only the Helsinki sample appeared to
induce the same level of IL-6 release for PM10-2.5 and PM0.2 at 10 mg/kg, albeit the collection
times differed. In vitro TNF-a and IL-6 responses did not always reflect in vivo effects
(Table 6-8), as the Duisburg PM10-2.5 sample was the most potent in vivo compared to the
other sites and elicited much lower cytokine release compared to other cities (except
Helsinki) in vitro (Happo et al., 2007, 096630;(Jalava et al., 2008, 098968).
Helsinki PM was collected in the spring and generally had the lowest in vivo and in
vitro activity for PM10-2.5 compared to the other cities (Happo et al., 2007, 096630;(Jalava et
al., 2008, 098968) . Spring-time samples were collected because episodes of resuspended
road dust occur frequently during this season (Pennanen et al., 2007, 155357). There was a
high correlation between EC content in PM2.5 and PM10-2.5, indicating that traffic impacted
both size fractions (Sillanpaa et al., 2005, 156980). Duisburg PM collected in fall had the
greatest amounts of Mn and Zn compared to PM samples from other locations (Pennanen et
al., 2007, 155357). Metals industries in Duisburg are likely contributors to the observed PM
metals concentrations. For the Prague winter PM samples, the As content was higher than
any other location (Pennanen et al., 2007, 155357). Prague also had the highest PAH levels
in all three size fractions, possibly attributable to stable atmosphere conditions and
incomplete combustion of coal and biomass in residential heating (Pennanen et al., 2007,
155357). High levels of ammonium and nitrate in PM samples from Amsterdam suggest
traffic as a large source of air pollution (Pennanen et al., 2007, 155357). Approximately
one-third of PM10-2.5 mass from Amsterdam was comprised of sea salt (Sillanpaa et al., 2005,
156980), double that of any other city. In Barcelona and Athens, high calcium or Ca2+
contents in spring and summer PM2.5 and PM10-2.5 are indicative of resuspended soil-derived
particles (Pennanen et al., 2007, 155357).
July 2009
6-190
DRAFT-DO NOT CITE OR QUOTE

-------
Table 6-8.
PAMCHAR PM
102.5 inflammation results with ambient PM.



City and Season

In vivo3 (mg/kg)

In vitrob (|jg/mL)

BALF protein
BALF TNF-a
BALF IL-6
BALF KC
BALFPMN BALF AM
TNF-a
IL-6
MIP-2
Helsinki spring
+10
+10
+10
[+310]
+10
+150,300
+150,300
+150,300
Duisburg fall
+10
+10
+10
+10
+10
+150,300
+150,300
+300
Prague winter
+10
[+310]
+10
[+310]
+10 +10
+150,300
+150,300
+150,300
Amsterdam winter
+10
+10
+10
+10
+10
+150
+150,300
+150,300
Barcelona spring
+10
+10
[+310]
+10
+10
+150,300
+150,300
+150,300
Athens summer
+10
[+310]
[+310]
[+310]
+10
+150,300
+150,300
+150,300
aSource: Happo et al. (2007,096630); 2 cell lines used for in vitro study were RAW264.7
bSource: Jalava et al. (2006,155872); + indicates increased response and numbers that follow indicate at which dose the response was observed
Schins et al. (2004, 054173) employed PM from two cities in Germany, Duisburg and
Borken, in another study In contrast to the PAMCHAR study where animals were
administered PM suspended in pathogen-free water (Happo et al., 2007, 096630), animals
received PM via IT instillation suspended in saline at a dose of 320 |Jg (Schins et al., 2004,
054173). In female Wistar rats, neutrophils in BALF were significantly elevated for PM10-2.5
from Duisburg and Borken (Table 6"9), albeit the percent of neutrophils with the PM10-2.5
from Borken was nearly double that of Duisburg. The responses with PM2.5 were much
smaller. When these PM10-2.5particles were introduced into whole blood to determine overall
inflammogenic capacity, IL-8 and TNF-a were released in greater quantities than in
response to fine PM. Furthermore, PM10-2.5 from Borken induced higher cytokine responses
than Duisburg PM10-2.5.
An in vivo study involving SH rats was conducted using PM10-2.5 and PM2.5 from six
different European locations with varying traffic densities (3 or 10 mg/kg IT; ultrafine PM
was not collected) (Gerlofs-Nijland et al., 2007, 097840). It was reported that PM10-2.5
generally induced greater responses than PM2.5. IT instillation of PM10-2.5 from a location
with high traffic influence in Munich, Germany demonstrated the greatest response in
terms of LDH activity, BALF protein, total cells, neutrophils, and lymphocytes 24 -h
post-exposure. PM10-2.5 collected from a low traffic site in Munich induced the greatest
cytokine response for TNF-a and MIP-2. Some correlations were observed between PM10-2.5
components (Ba and Cu) and BALF parameters, but were largely driven by one location
(Gerlofs-Nijland et al., 2007, 097840).
July 2009
6-191
DRAFT-DO NOT CITE OR QUOTE

-------
Table 6-9. Other ambient PM - in vivo PM102.5 studies - BALF results, 18-24 h post-IT exposure
Location
Endotoxin
(~ Values)
Dose
(mg/kg)
Ceil
Differentials
Cytokines
Injury
Biomarkers
Reference
Germany, Borken; rural Feb-May
2000
6.6 EU/mg
0.58-0.91
f%PMN
t TNF-a

Schins et al. (2004,054173)
Germany, Duisburg; heavy industry
Feb-May 2000
5.0 EU/mg
0.58-0.91
t % PMN
t MIP-2

Schins et al. (2004,054173)
USA, Seattle, WA
6.0 EU/mg
1.25, 5.0



Gilmour, et al. (2007, 096433)
Feb-March 2004






USA, Salt Lake City, UT
6.3 EU/mg
1.25, 5.0


T protein
Gilmour, et al. (2007, 096433)
Apr-May 2004






USA, South Bronx, NY
2.8 EU/mg
1.25, 5.0
t PMN
t MIP-2

Gilmour, et al. (2007, 096433)
Dec 2003-Jan 2004






USA, Sterling Forest, NY
2.9 EU/mg
1.25, 5.0



Gilmour, et al. (2007, 096433)
Dec 2003-Jan 2004






USA, RTP, NC
0.96 EU/mg
0.5, 2.5, 5.0
ft PMN
t IL-6

Dick (2003, 088776)
Oct-Nov 1996






Germany, Munich Ost Bahnof; high
traffic A
Aug 2002
2.9 EU/mg
3,10
ft* total cells
ft AM
tt'PMN
ft* Lymph
ft MIP-2
ft TNF-a
tfLDH
T* protein
Gerlofs-Nijland, et al. (2007, 097840)
Netherlands, Hendrik-ldo-Ambacht;
high traffic
Sept 2002
6.5 EU/mg
3,10
ft total cells
tfAM
ft PMN
ft Lymph
t MIP-2
ft TNF-a
ft LDH
T protein
Gerlofs-Nijland, et al. (2007, 097840)
Italy, Rome; high traffic
1.5 EU/mg
3,10
f total cells
ft MIP-2
ft LDH
Gerlofs-Nijland, et al. (2007, 097840)
Apr 2002


tfAM
ft PMN
ft Lymph
ft TNF-a


Netherlands, Dordrecht; moderate
traffic
Apr 2002
0.6 EU/mg
3,10
ft total cells
t AM
ft PMN
f Lymph

ft LDH
T protein
Gerlofs-Nijland, et al. (2007, 097840)
Germany, Munich Grosshadern
Hospital; low traffic
Jun-Jul 2002
2.9 EU/mg
3,10
f total cells
tfAM
ft PMN
ft Lymph
tf MIP-2
ft* TNF-a
tfLDH
T protein
Gerlofs-Nijland, et al. (2007, 097840)
Sweden, Lycksele; low traffic
0.9 EU/mg
3,10
ft total cells

ft LDH
Gerlofs-Nijland, et al. (2007, 097840)
Feb-March 2002


t AM
ft PMN
f Lymph

T protein

For Gerlofs-Nijland study, composition data were averaged across seasons, t significant only at highest dose, ft Significant at lowest and highest dose. * Greatest potency for that endpoint and
study. Gilmour et al. (2007, Q96433)exposure was via aspiration.
July 2009
6-192
DRAFT-DO NOT CITE OR QUOTE

-------
A more recent study by these investigators (Gerlofs-Nijland et al., 2009, 190353)
compared responses to PM from 3 different European cities based on size fraction and
content of metals and PAH. SH rats were intratracheally instilled with 7 mg/kg PM, and
markers of toxicity and inflammation were measured in BALF 24 h later. Blood markers of
systemic inflammation and coagulation were also measured and are described in
Section 6.2.7.3 and 6.2.8.3. In the first part of the study, both fine and coarse fractions of
PM from Duisburg were found to have dramatic effects on inflammatory cell influx and
activation as well as on the injury markers LDH, protein and albumin in the BALF. The
antioxidant species uric acid was increased in BALF from rats exposed to both size fractions
and was interpreted as an adaptive response to oxidative stress. Statistical analysis
demonstrated that coarse PM was more potent in eliciting these responses than fine PM. In
the second part of the study, responses to metal-rich PM from Duisburg and metal-poor PM
from Prague were determined. A statistically significant greater enhancement of BALF
markers of inflammation and injury was observed for the Duisburg PM compared with the
Prague PM. Furthermore, responses to PAH-rich coarse PM from Prague and PAH-poor
coarse PM from Barcelona were determined. PM10-2.5 from Prague was found to have
statistically significant greater effects compared with PM10-2.5 from Barcelona. However,
organic extracts of these PM10-2.5 fractions had very little capacity to produce inflammation
or toxicity in this model. These findings suggest an important role for specific components
associated with the coarse fraction in mediating the pro-inflammatory effects.
In another study investigating specific components of PM10-2.5, BALB/c mice were
intratracheally-instilled with 25 and 50 |Jg PM10-2.5 from a rural area of the San Joaquin
Valley, California (Wegesser and Last, 2008, 190506). Inflammatory cell influx into BALF
began at 6 h and peaked at 24 h following IT instillation with 50 |Jg PM, with the increase
in neutrophils preceding the increase in macrophages. Pro-inflammatory effects were found
to be mainly due to insoluble components of PM. Furthermore, he at-treatment, which was
capable of inactivating endotoxin, had no effect on inflammation. Numbers of neutrophils in
the BALF were found to correlate with the content of macrophage inflammatory protein-2,
a known neutrophil chemoattractant released from macrophages and epithelial cells. Taken
together, these results demonstrate that the pro-inflammatory effect of this PM10-2.5 was
associated with insoluble components and not with endotoxin.
In an in vivo study that employed ambient PM collected in fall 1996 from Research
Triangle Park (RTP), NC, neutrophilic influx was observed in BALF of female CD1 mice 18-
July 2009
6-193
DRAFT-DO NOT CITE OR QUOTE

-------
h post-IT exposure (10, 50 or 100 |Jg) of coarse PM (3.5-20 |Jm), although a dose-response
relationship was not evident (Dick et al., 2003, 088776). Only the two highest doses of PM
for the smaller size fractions induced elevated neutrophils. Significant responses in
albumin and TNF-a were only observed for the fine PM (1.7-3.5 |Jm) exposure group. Total
protein, LDH and NAG responses were absent for all PM size fractions. Levels of IL-6 were
elevated in mice exposed to 100 |Jg for coarse, fine, and fine/ultrafine (<1.7 |Jm) PM. When
dimethylthiourea (DMTU) was administered intravenously prior to exposure, the
neutrophil response was attenuated in all groups to levels below control.
Another study compared thoracic coarse, fine, and ultrafine PM collected in Seattle,
WA, Salt Lake City, UT, South Bronx, NY, and Sterling Forest, NY (Gilmour et al., 2007,
096433). In female BALB/c mice, the 100 |Jg dose of PM10-2.5 (approximately 5 mg/kg) from
Salt Lake City induced a significant increase in protein in BALF and the level released was
almost as high as that observed after LPS exposure. PM10-2.5 from the South Bronx resulted
in dose-related increases in neutrophil number and MIP-2 levels in BALF. In contrast, no
effects were observed with PM10-2.5 from Sterling Forest. The greatest amount of LPS was
observed in the Salt Lake City and Seattle PM10-2.5 samples. There was a less discernable
pattern of response with fine and ultrafine PM.
Coal Fly Ash
Coal fly ash of differing size fractions and composition was administered to female
CD1 mice via oropharyngeal aspiration (25 or 100 |Jg) to assess lung inflammation and
injury 18 h following exposure (Gilmour et al., 2004, 057120). Montana (lowsulfur
subbituminous! 0.83% sulfur, 11.72% ash content) or western Kentucky (high-sulfur
bituminous! 3.11% sulfur, 8.07% ash content) coal was combusted using a laboratory-scale
down-fired furnace. Interestingly, no significant PM10-2.5 effects for either coal fly ash were
observed for BALF neutrophils, TNF-a, MIP-2, albumin, total protein, LDH activity, or
NAG activity 18 h post-exposure. However, the ultrafine fraction (PM0.2) of combusted
Montana coal induced greater numbers of neutrophils than PM10-2.5 or PM2.5 at both doses.
TNF-a was only elevated in animals exposed to 100 |Jg of the Montana ultrafine PM; MIP-2
was also increased at both doses. The PM2.5 western Kentucky coal fly ash caused increased
BALF neutrophils, MIP-2, albumin, and protein (Gilmour et al., 2004, 057120).
In a similar study employing Montana subbituminous coal fly ash particles >2.5 |Jm,
C57B1/6J mice were intratr ache ally instilled with PM alone or PM+LPS and BALF was
obtained 18 h post-exposure (Finnerty et al., 2007, 156434). TNF-a and IL-6 in lung
July 2009
6-194
DRAFT-DO NOT CITE OR QUOTE

-------
homogenates were only elevated in the animals exposed to PM+100 |Jg LPS, although it
appeared that there was a greater-than additive effect. Total cells and cell differentials
were not measured.
Summary of Toxicological Study Findings for Relative Toxicity of PM Size Fractions
Biomarkers of injury and inflammation were measured in in vivo and in vitro studies
comparing the toxicity of different size fractions of ambient PM from various locations.
Responses were measured in BALF of rodents following IT instillation or aspiration of PM.
In general, the PM10-2.5 size fraction was more potent than fine or ultrafine PM and
endotoxin levels did not appear responsible. In one study, rural PM10-2.5 from Germany
induced a greater inflammatory and cytokine response than PM10-2.5 from an industrial
location. In contrast, PM10-2.5 from Sterling Forest, NY did not lead to any change in BALF
inflammation or injury markers. A study that employed coal fly ash indicated that the
ultrafine PM fraction was the most inflammogenic. All of these studies were conducted
using high doses of PM (0.58-10 mg/kg) and it is unclear if similar effects would be observed
at lower doses.
6.3.6. Allergic Responses
A large number of toxicological and controlled human exposure studies cited in the
2004 PM AQCD (U.S. EPA, 2004, 056905) reported an exacerbation of existing allergic
airway disease following exposure to laboratory-generated and ambient particles. In
addition, numerous studies have demonstrated that PM can alter the immune response to
challenge with specific antigens and suggest that PM may act as an adjuvant to promote
allergic sensitization. Recent toxicological studies have provided evidence of enhanced
allergic responses and allergic sensitization following exposure to CAPs and diesel that is
consistent with the findings presented in the 2004 PM AQCD. PM can enhance allergic
responses by facilitating delivery of allergenic material and promoting subsequent immune
reactivity. The initiation or exacerbation of allergic responses has important implications
for allergic asthma, the most common form of asthma. Additionally, PM has been shown to
alter ventilatory measures in non-allergic animal models, suggesting a possible role in
other forms of asthma.
July 2009
6-195
DRAFT-DO NOT CITE OR QUOTE

-------
6.3.6.1.	Epidemiologic Studies
Allergy contributes to a number of respiratory morbidity outcomes, including asthma.
However, relatively few epidemiologic studies of PM have specifically examined indicators
of allergy. The 2004 PM AQCD (U.S. EPA, 2004, 056905) presented one study (Hajat et al.,
2001, 016693) showing an association between doctor visits for allergic rhinitis and PMio
among children in London. This association was strongest at a lag of 3 or 4 days. Similar
results were obtained in a new study by Tecer et al. (2008, 180030). which found significant
associations between PM2.5, PM10, and PM10-2.5 with hospital admissions for allergic rhinitis
in Turkish children, particularly at lag day 4. While exacerbation of allergic symptoms may
occur relatively rapidly, repeated or longer exposures may be required for allergic
sensitization to develop! a number of studies associating long-term exposure to PM with
specific indicators of allergic sensitization are described in Chapter 7.
6.3.6.2.	Controlled Human Exposure Studies
Exacerbation of Allergic Responses
Diesel Exhaust and Diesel Exhaust Particles
Exposure to DEP was shown to increase the allergic response among atopic
individuals in several controlled human exposure studies cited in the 2004 PM AQCD
(U.S. EPA, 2004, 056905). Nordenhall et al. (2001, 025185) found that exposure to DE
significantly decreased the concentration of methacholine required to induce a 20%
decrease in FEVi in a group of atopic asthmatics 24 h post-exposure. In addition,
Diaz-Sanchez et al. (1997, 051247) demonstrated an increase in allergen-specific IgE
following exposure via intranasal spray to ragweed plus DEP (0.3 mg) relative to ragweed
allergen alone. Decreases in IFN-y and IL-2, as well as increases in IL-4, IL-5, IL-6, IL-10,
and IL-13 were also observed when ragweed allergen was administered with DEP. It should
be noted that the DEP used in this study were collected during a cold start of a light-duty
Isuzu diesel engine, and thus contained relatively high levels of incomplete combustion
materials and semi-volatiles organics (e.g., PAHs). One new study using the same source of
DEP (Bastain et al., 2003, 098690) also observed an increase in IL-4 and allergen specific
IgE, as well as a decrease in IFN-y following intranasal administration of ragweed allergen
with DEP (0.3 mg) in atopic adults. The protocol was repeated in this study for all subjects,
and the enhancement of allergic response by coexposure to DEP was observed to be highly
July 2009
6-196
DRAFT-DO NOT CITE OR QUOTE

-------
reproducible within individuals. In addition, Gilliland et al. (2004, 156471) demonstrated
that GST polymorphisms may alter the adjuvant effects of DEP on allergic response, with
individuals with GSTM1 null or GSTP11105 wild type genotypes showing the largest
effects.
Allergic Sensitization
Diesel Exhaust and Diesel Exhaust Particles
One controlled human exposure study has demonstrated that de novo sensitization to
a neoantigen can be induced by exposure to DEP. In this study, Diaz-Sanchez et al. (1999,
011346) dosed 25 atopic adults intranasally with 1 mg keyhole limpet hemocyanin (KLH),
followed by 2 biweekly challenges with 100 |Jg KLH. In 15 of the 25 subjects, cold-start DEP
(0.3 mg) were administered intranasally 24-h prior to each KLH exposure, while in the
other 10 subjects, no diesel particles were administered. No KLH-specific IgE was observed
in the nasal lavage fluid of any of the subjects exposed to KLH without exposure to diesel
particles. However, KLH-specific IgE was present in the nasal lavage fluid of 9 out of 15
subjects 28"32 days after the initial KLH immunization when exposures were preceded by
administration of DEP.
6.3.6.3. Toxicological Studies
Exacerbation of Allergic Responses
Increased use of actual ambient air particle mixes in toxicological studies since the
2004 CD has greatly expanded evidence relevant to assessing these and other immunotoxic
effects. A number of studies have also included ambient level doses, although many still
include relatively high doses of questionable relevance compared to the doses inhaled by
humans. Recent dosimetric models reveal that a small fraction of epithelial cells located at
the carinal ridges of airway bifurcations can receive massive doses that may be even a few
hundred times higher than the average dose for the whole airway (see Chapter 4). These
areas, coincidentally, are locations of bronchus associated lymphoid tissues (BALT) which
are sites at which interaction of T and B lymphocytes with antigen presenting cells (APC)
occurs. Hence the deposited particles are in near-ideal proximity to immunologically active
tissues. Doses used for assessing PM immunotoxicity should be viewed with this
perspective. In many animal studies changes in ventilatory patterns are assessed using
July 2009
6-197
DRAFT-DO NOT CITE OR QUOTE

-------
whole body plethysmography, for which measurements are reported as enhanced pause
(Penh). Some investigators report increased Penh as an indicator of airway
hyperresponsiveness (AHR), but these are inconsistently correlated and many investigators
consider Penh solely an indicator of altered ventilatory timing in the absence of other
measurements to confirm AHR. Therefore use of the terms AHR or airway responsiveness
has been limited to instances in which the terminology has been similarly applied by the
study investigators.
CAPs
Existing allergic sensitization confers susceptibility to the effects of PM in rodent
models. For example, studies in allergic rats (Harkema et al., 2004, 056842; Morishita et
al., 2004, 087979) suggest that allergic sensitization enhances the retention of PM in the
airways. Recovery of anthropogenic trace elements (La, V, Mn, S) from lung tissue was
greater for Detroit CAPs PM2.5 exposed ovalbumin (OVA) sensitized/challenged BN rats
than for air exposed or non-allergic CAPs exposed controls (24-h post-exposure for 4 or 5
consecutive 10"h days during July or September! time weighted average mass concentration
of 676 ± 288 or 313 ± 119 |ug/m3, respectively) (Harkema et al., 2004, 056842). Interestingly,
despite lower average mass concentration, increases in these elements were observed in
September, when the average number concentration of UFPs was nearly double that of July
(10,879 ± 5126 vs. 5753 ± 2566 particles/cm3). September CAPs was associated with
eosinophil influx and BALF protein content, as well as significantly increased airway
mucosubstances, and the authors speculated that the high concentration of UFPs facilitated
particle penetration into the alveolar region of the lungs. IT instillation of fractionated
insoluble PM2.5 collected from this period resulted in a mild pulmonary neutrophilic
inflammation in healthy BN rats, but no differential effects were obtained after IT
instillation of total, soluble, or insoluble PM2.5 in allergic rats.
Research has also been conducted to determine the effect of proximity to the roadway
on exacerbation of existing allergic disease. OVA-allergic BALB/c mice were exposed to
CAPs (fine, F, < 2.5 or ultrafine, UF, < 0.15, avg total concentration 400 ug/ma) for five 4-h
days a week over 2 wk at 50 or 150 m downwind of a heavily trafficked road (Kleinman et
al., 2005, 189364). Markers of allergy (serum OVA-specific IgE and IgGl, lung IL-5 and
eosinophils) were significantly higher in mice exposed to CAPs (UF or F) than in
air-exposed mice after OVA challenge. IL-5, IgGl, and eosinophils were higher in mice
closer to the roadway (50 m) than in mice 150 m downwind. The authors suggest that the
July 2009
6-198
DRAFT-DO NOT CITE OR QUOTE

-------
enhanced responses closer to the roadway may reflect a greater proportion of UF particles
in this vicinity, given that the concentrations of sutr25-nm particles decrease rapidly with
distance from the roadway and the F CAPs nearer the roadway contained a greater number
of particles for a similar mass, a portion of which are UF. Animal "to-animal variability
among the biomarkers tested made it necessary to combine values from two exposures
spanning two years for statistical power (determined prior to the start of the experiment). A
subsequent publication (Kleinman et al., 2007, 097082) included a third exposure regimen
as well as compositional analysis. Fine CAPs mass concentration was intentionally
adjusted to an average concentration of approximately 400 |Jg/m3, ranging from 163 to
500 |Jg/m3, with an estimated particle number of 2.1 x 105 particles/cm3 at 50 m and 1.6 x
105 particles/cm3 at 150 m. UF ranged from 146 to 430 jJg/m3, with particle counts of 4.9 ±
1.4 x 105 particles/cm3 at 50 m, and 4.4 ±2.1 x 10B particles/cm3 at 150 m. Analysis of results
from the three exposures indicated that OVA-sensitized mice exposed 50 m downwind of the
roadway exhibited increased levels of IL-5 and IgGi compared to mice exposed 150 m
downwind or exposed to air. No markers of allergy-related responses were observed in the
150 m exposure groups, and very little difference was seen between fine and ultrafine CAPs
responses, perhaps because fine material contained 20-32% ultrafine components. The
strongest associations between component concentrations and biological markers of allergy
(IL-5 and IgGl) were with EC and OC. These studies demonstrate that CAPs can enhance
allergic responses, and that proximity to a source may be an important factor.
In a BN rat model for allergic asthma (Heidenfelder et al., 2009, 190026), thirteen 8-h
days of exposure to Grand Rapids, MI CAPs (PM2.5) alone did not result in differential gene
expression or indicators of asthmatic pathology in the lung, but the combination of CAPs
and OVA resulted in differential expression of genes predominantly related to inflammation
and airway remodeling, along with significant increases in IgE, mucin, and total protein in
BALF. Consistent with these changes in gene expression and BALF markers, OVA with
CAPs also induced a more severe allergic bronchopneumonia (distribution and severity of
bronchiolitis and alveolitis), and increased mucus cell metaplasia/hyperplasia and
mucosubstances, indicating exacerbation of allergic or asthmatic disease. CAPs was
collected in July and characterized as having an average mass of 493 ± 391, OC 244 ± 144,
EC 10 ± 4, S042- 79 ± 131 (13 day avg was only about 10% of the CAPs, but a spike occurred
during the first week), nitrate 39 ± 67, ammonium 39 ± 59, and urban dust (estimated from
Fe, Al, Ca, and Si) 18 ± 6 (mean ± SD in jug/m3).
July 2009
6-199
DRAFT-DO NOT CITE OR QUOTE

-------
Diesel Exhaust Particles
Resuspended DEP influences airway responses in mice with existing allergic
sensitization. A single 5-h nose-only exposure to 870 jug/m3 aerosolized filter-collected DEP
(PM2.5) increased Mch-induced increases in ventilatory timing (Penh, a parameter that has
been correlated with airways resistance in some studies) in OVA sensitized/challenged
C57BL/6J mice (Farraj et al., 2006, 088469). Intranasal pretreatment with an antibody
against the pan neurotrophin receptor p75 attenuated the DEP-induced increase in airflow
obstruction, indicating a role for neurotrophins. Neurotrophins are expressed by various
structural, nerve and immune cells within the respiratory tract and are linked to the
etiology of asthma in both humans and animal models. DEP alone in unsensitized mice
caused a significant increase in lung macrophages! this response was also inhibited by
anti-p75, which may suggest mediation of macrophage influx by neurotrophin or
alternatively may reflect anti-p75 dependent depletion of macrophages due to expression of
the p75 receptor. Aside from increased macrophages, the single exposure to DEP had little
effect on other markers of airway inflammation. In a similar subsequent study, these
authors demonstrate neurotrophin-mediated DEP-induced airflow obstruction in OVA
sensitized and challenged BALB/c mice (Farraj et al., 2006, 141730), in this case using a
higher 2000 jug/m3 single 5"h exposure to aerosolized filter-collected PM2.5. Differences
between whole body plethysmography and tracheal ventilation measurements indicated
that airflow obstruction may have originated in the nasal passages. Again, very few indices
of inflammation were increased! however, similar neurotrophin-dependent increases in lung
macrophages were observed after DEP exposure alone, and BALF IL-4 protein levels were
increased 5-fold in sensitized, challenged, DEP-exposed mice. This neurotrophin-dependent
IL-4 response was not evident in the first study, and may be related to the higher dose used
in the second study or the characteristic allergic/Th2 bias of the BALB/c strain. Airflow
obstruction in the absence of airway inflammation in OVA-sensitized animals seen in both
studies by Farraj et al. (2006, 088469! 2006, 141730) may reflect DEP-induced acute
enhancement of neurogenic as opposed to immunologic inflammation.
Diesel Exhaust
Exposure to relatively low doses of DE has been shown to exacerbate asthmatic
responses in OVA sensitize d/challenged BALB/c mice (Matsumoto et al., 2006, 098017).
Mice were intranasally challenged one day prior to chamber exposure to DE (DEP
100 jug/m3! CO, 3.5 ppm! NO2, 2.2 ppm! SO2 <0.01 ppm) for 1 day or 1, 4, or 8 wk (7h/day, 5
July 2009
6-200
DRAFT-DO NOT CITE OR QUOTE

-------
days/wk, endpoints 12-h post-DE exposure). Results from the 8 wk study are described in
Section 7.3.6.2. It should be noted that control mice were left in a clean room as opposed to
undergoing chamber exposure to filtered air. Significant AHR upon Mch challenge was
observed after 1 and 4 wk of exposure, and airway sensitivity (provocative concentration of
Mch causing a 200% increase in Penh) was significantly increased after 1 wk of exposure
but not 4 wk. DE had no effect on total cells in BALF, but transiently increased expression
of IL-4, IL-5, and IL-13 after 1 day of exposure, MDC after 1 wk, and RANTES after 2 and 3
wk. Eotaxin, TARC, and MCP-1 were elevated without statistical significance after
short-term (l day or wk) exposure. Statistical power may have been lacking due to an n of
3. Protein levels of IL-4 and RANTES were significantly elevated after 1 day of DE
exposure, respectively. DE had no effect on OVA challenge-induced peribronchial
inflammatory or mucin positive cells. Therefore DE-induced airway hyperresponsiveness
was observed in the absence of cellular inflammation, similar to the responses described for
aerosolized or nebulized DEP by Farraj et al. (2006, 088469; 2006, 111730) and Hao et al.
(2003, 096565).
Gasoline Exhaust
Acute exposure to fresh gasoline engine exhaust (GEE) PM does not appear to
exacerbate allergic responses (Day et al., 2008, 190204). BALB/c mice were exposed to
whole exhaust diluted 1:10 (H), l: 15 (M), or 1:90 (L), filtered exhaust at the 1:10 (HF), or
clean air for 6 h/day over three days. Analytes for the high (H) and high filtered (HF)
concentrations were: PM mass (|Jg/m3) 59.1 ± 28.3 (H) and 2.3 ± 2.6 (HF); PM number
(#/cm3) 5.0 x 105 and 1.1 x 104; CO (mg/m3) 102.8 ± 33.0 and 99.5 ± 1.6; NO (mg/m3) 18.4 ±
2.8 and 17.2 ± 1.9; NO2 (mg/m3) 1.4 ± 0.3 and 1.7 ± 0.2; SO2 (|Jg/m3) 1366.8 ± 56.0 and 1051.1
± 43.0; NH3 (|Jg/m3) 1957.7 ± 8.1 and 1241.5 ± 6.1; NMHC (mg/m3) 15.9 and 25.9. Particles
represented only 0.04% of the total exposure mass and particle size in the H exposure
ranged from 5.5 - 150 nm with the majority between 5-20 nm (MMD 150 nm) (McDonald et
al., 2008, 191978). Although particles were filtered out, it should be noted that NMHC (non-
methane volatile organics) increased by 62%. Mice were exposed with or without prior
sensitization to OVA, after one aerosol challenge and with or without secondary challenge.
Acute GEE exposure had variable effects on inflammatory and allergic markers depending
on the exposure protocol, but there were no statistically significant differences between the
H and HF exposure results, suggesting that the PM fraction of GEE does not appear to
contribute significantly to observed health effects.
July 2009
6-201
DRAFT-DO NOT CITE OR QUOTE

-------
Hardwood Smoke
One study indicated that hardWS (HWS) exposure only minimally exacerbated
indices of allergic airway inflammation in an OVA-sensitized BALB/c mouse model and did
not alter Thl/Th2 cytokine levels (Barrett et al., 2006, 155677). Trend analysis indicated
increasing BALF eosinophils with increasing dose of HWS, becoming significantly elevated
at 300 |ug/m3 (CO, 1.6 ± 0.3 ppm! total vapor hydrocarbon, 0.6 ± 0.2 ppm! NOx, below limit of
quantitation, PM MMAD 0.35 ± 2.0 urn), and increasing but not statistically significant
OVA-specific IgE levels with HWS up to 1000 |ug/m3.
Model Particles
Exposures to an aerosol of soot and iron oxide generated from ethylene (PM2.5, 0.235
mg/m3) were conducted to test whether the sequence of exposure to OVA aerosol challenge
and PM affected the observed response of OVA sensitized BALB/c mice (Last et al., 2004,
097334). Though called PM2.5, the authors characterized the PM material as ultrafine,
80-110 nm, with the iron oxide crystals often spatially segregated from the soot (200 |u,g/m3
soot, remainder iron oxide, CO <0.8 ppm, NOx <0.4 ppm, PAH below detection). Mice were
exposed to PM via chamber inhalation for 2 wk (4h/day, 3 days/wk) before or after 4 wk of
OVA inhalation, or simultaneously to PM and OVA for 6 wk. Among endpoints (BALF cells,
Penh, airway collagen, and goblet cells) only goblet cell counts were significantly increased
with PM exposure in any combination with OVA. There was a trend toward increased Penh
responses with exposure to PM alone or with OVA, particularly when PM exposure
immediately preceded Mch challenge (after or during OVA challenge). Results from this
study are difficult to interpret due to the varied elapsed times between cessation of PM or
OVA treatment and endpoint determination. The mild responses to PM may be related to
the intraperitoneal sensitization protocol used, reputed to generate a highly allergic mouse
in which any additive effects of PM may be obscured by maximal responses to antigen
challenge (Deurloo et al., 2001, 156396; Hao et al., 2003, 096565).
Residual Oil Fly Ash
Arantes-Costa and colleagues (2008, 187137) estimated that 60 |ug of ROFA would be
inhaled by a mouse during one day of exposure to Sao Paulo air. This dose, given
intranasally every other day for four days, increased AHR in both nonsensitized and OVA
sensitized/challenged BALB/c mice upon Mch challenge two days after the last exposure.
ROFA had no significant impact on eosinophil or macrophage numbers in the lung, nor did
it increase the chronic lung inflammation or thickening induced by OVA. In many studies
July 2009
6-202
DRAFT-DO NOT CITE OR QUOTE

-------
particular effects such as airway obstruction are only evident when allergic sensitization
precedes exposure, but this study and a few others demonstrate allergen-independent AHR
after exposure to PM including CAPs (Lei et al., 2004, 087999) and DEP or DE (Hao et al.,
2003, 096565; Li et al., 2007, 155929).
Allergy in pregnancy or early life
Pregnancy or in utero exposure may confer susceptibility to PM-induced asthmatic
responses. Exposure of pregnant BALB/c mice to aerosolized ROFAleachate by inhalation
or to DEP intranasally increased asthma susceptibility in their offspring (Fedulov et al.,
2008, 097482; Hamada et al., 2007, 091235). The offspring from dams exposed for 30 min to
50 mg/mL ROFA 1, 3, or 5 days prior to delivery responded to OVA immunization and
aerosol challenge with AHR and increased antigen-specific IgE and IgGl antibodies. AHR
was also observed in the offspring of dams intranasally instilled with 50 |ug of DEP or Ti02,
or 250 ug CB, indicating that the same effect could be demonstrated using relatively "inert"
particles. Pregnant mice were particularly sensitive to exposure to DEP or Ti02 particles,
and genetic analysis indicated differential expression of 80 genes in response to Ti02 on the
pregnant background. Thus pregnancy may enhance responses to PM, and exposure to even
relatively inert particles may result in offspring predisposed to asthma.
Allergic Sensitization
A large number of in vivo animal studies and in vitro studies have demonstrated that
particles can alter the immune response to challenge with specific antigens and suggest
that PM acts as an adjuvant to promote allergic sensitization. This phenomenon was
introduced in the 2002 Diesel Document, and has been noted in multiple animal and
human studies by the 2004 PM AQCD. Adjuvants enhance the immune response to
antigens through various means, including chemoattraction, cytokines, or enhanced antigen
presentation and costimulation, and may originate via effects on a number of cell types.
Importantly, adjuvants may be major contributors to the development of inappropriate
immune responses. These immune responses, mediated by T helper cells, fall along a
continuum from T helper type 1 (Thl) to T helper type 2 (Th2). Thl responses,
characterized by IFN-y, are inflammatory and in excess can lead to tissue damage.
Alternatively, Th2 responses are characterized by IL-4, IL-5, IL-13, eosinophils, and IgE,
and are associated with allergy and asthma. Autoimmune diseases may be driven by Thl,
July 2009
6-203
DRAFT-DO NOT CITE OR QUOTE

-------
Th2, or mixed responses, but allergic diseases are predominantly Th2 mediated, and many
of the immunologic effects observed for PM fall into the Th2 category
It has been suggested that the capacity of particles to enhance allergic sensitization is
associated more strongly with particle number and surface area than particle mass, and
several studies comparing size fractions of the same material show greater adjuvant
activity for an equivalent mass dose of smaller particles (Inoue et al., 2005, 088625;
Nygaard et al., 2004, 058558; de Haar et al., 2006, 144746). This is particularly true of inert
or homogeneous materials, such as carbon, polystyrene, and Ti02, which vary little in
composition with size fraction. Studies using CAPs have also observed that adjuvancy and
allergic exacerbation are more strongly associated with the ultrafine fraction, possibly due
to greater oxidative potential (Kleinman et al., 2005, 189364; 2007, 097082; Li et al., 2009,
190457). In some studies of ambient PM, however, coarse particles have demonstrated
equal or greater adjuvancy compared to fine particles (Nygaard et al., 2004, 058558;
Steerenberg et al., 2004, 096024; Steerenberg et al., 2005, 088649). More inhalation studies
to compare size fractions are needed in order to elucidate the role of particle size in
mediating adjuvancy, but this may prove difficult given the influence of composition, e.g.,
combustion related materials (Steerenberg et al., 2006, 088249) and metal content (Gavett
et al., 2003, 053153), which differs among various size fractions and sources.
CAPs
As little as 0.1 ug of ultrafine Los Angeles CAPs administered intranasally with OVA
was able to significantly boost allergic antibody responses in BALB/c mice (Li et al., 2009,
190457). Acomparison of UFPs (UFP, aerodynamic diameter <0.15 |Jm) with a mix of sub-
2.5 |Jm particles (F/UFP) collected 200 m from a major freeway delivered intranasally five
times over the course of nine days showed that UFP but not F/UFP were associated with
significant adjuvant effects. 0.5 jig of UFP with OVA (but not alone) led to an increase in
BALF eosinophils, allergic cytokines, inflammatory mediators, and serum OVA-specific
IgE/IgGl, as well as allergic tissue inflammation in the upper and lower airways. Adjuvant
effects of UFP were observed with two independently collected samples (Jan 07 and Sep 06)
and could not be replicated by administering the same amount of endotoxin measured in
the particles, indicating that the effects were not unique to the sampling period nor
mediated by contaminating endotoxin. UFP had a greater OC and PAH content than F/UFP,
and induced greater oxidative stress in vitro. Partial blocking of the adjuvant effects by
antioxidant administration implicates redox potential as a key factor in mediating these
July 2009
6-204
DRAFT-DO NOT CITE OR QUOTE

-------
effects. The authors suggest that the lack of adjuvancy for ultrafine carbon particles (being
mostly EC) is due to a lack of redox cycling compounds, but this was not tested. In contrast,
ultrafine (30-50 nm) CB particles have demonstrated intranasal adjuvant activity in other
studies (de Haar et al., 2005, 097872) when administered with OVA over three consecutive
days. A 200 ug dose increased serum OVA-specific IgE, local lymph node dendritic cells and
OVA-specific Th2 lymphocytes in the lung draining lymph nodes and lung, as well as post-
challenge airway eosinophilia. Doses as low as 20 ug were able to activate adoptively
transferred OVA-specific T cells.
Diesel Exhaust Particles
Resuspended DEP have been shown to enhance OVA-specific IgGl and IgE in BALB/c
mice exposed via inhalation to doses as low as 200 and 600 |ug/m3, respectively (Whitekus et
al., 2002, 157142). Mice were exposed to DEP (200, 600 and 2000 |ug/m3) for 1-h daily for 10
days prior to aerosol OVA challenge. Compared with responses to OVA alone, antibody
levels were increased by all OVA+DEP exposures. Statistical significance was reached for
IgGl at all DEP exposure levels, whereas OVA specific IgE was significantly increased at
the 600 and 2000 |ug/m3 doses and total IgE was significantly elevated at 2000 |ug/m3.
Although strong adjuvant effects were observed, no general markers of inflammation such
as eosinophils, IL-5, GM-CSF, mucin, morphological changes, or eosinophilic major basic
protein (MBP) deposition in the airways were observed in exposed mice. In vitro
experiments using the RAW 264.7 macrophage-like cell line indicated a DEP-induced lipid
peroxidation and protein oxidation, which could be inhibited by a variety of antioxidants.
Also observed was a decrease in the GSH: GSSG ratio and an increase in HO-1 expression,
both of which were inhibited only by the thiol antioxidants NAC and BUC. These same thiol
antioxidants were able to completely block DEP-related increases in IgE and IgGl, as well
as lipid peroxides and oxidized proteins recovered from lung lavage fluids at the 2000 jug/m3
dose. Thus solid correlations between in vivo and in vitro antioxidant activities were found,
and the reversal of adjuvant effects by antioxidants in vivo clearly indicates a link between
oxidative stress and DEP adjuvancy. However, the intranasal adjuvant activity of Ottawa,
Canada, dust (EHC-93) in the same strain of mice was not inhibited by NAC pretreatment
(Steerenberg et al., 2004, 087981), suggesting that disparate pathways may be utilized by
different materials to exert immune stimulation.
July 2009
6-205
DRAFT-DO NOT CITE OR QUOTE

-------
Diesel Exhaust
DEP inhalation during allergen exposure has been shown to augment IgE production
and alter methylation of T helper genes in BALB/c mice (Liu et al., 2008, 156709) Animals
were exposed to 1280 |u,g/m3 DEP over a 3-wk period, 5h per day, concurrent with periodic
intranasal sensitization to the common fungus Aspergillus fumigatus. Gas concentrations
were not reported. Total IgE and BALF eosinophils were elevated with A. fumigatus
sensitization and further increased by concomitant DEP exposure. Greater methylation of
the IFN-y promoter was observed following DEP and A. fumigatus exposure (but not DEP
alone) compared to A. fumigatus alone, indicating that combined DEP and allergen
exposure might induce methylation and thus suppress expression of Thl genes.
Furthermore, hypomethylation of the IL-4 promoter was detected after exposure to A.
fumigatus and DEP compared with exposure to A. fumigatus or DEP alone, suggesting
pro-allergic Th2 gene activation upon combined exposure to allergen and DEP. The changes
in methylation status of these genes were associated with alterations in IgE levels in
individual animals, indicating that modifications at the genetic level could result in
predicted downstream effects. This study shows for the first time that DEP exposure can
exert pro-allergic in vivo effects on the mouse immune system at the epigenetic level.
A toxicogenomic approach to investigate early response mechanisms of DEP
adjuvancy was taken by Stevens et al. (2008, 157010). BALB/c mice were chamber exposed
to filtered air, 500 or 2000 jug/m3 PM in DE for 4 h/day over 5 consecutive days and
intranasally exposed to OVA on each of the first 3 days. In the low vs. high PM exposures,
CO, NO, NO2, and SO2 were <0.1 vs. 4.3, <2.5 vs. 9.2, <0.25 vs. 1.1 and <0.06 vs. 0.2 ppm;
particle number median diameters were 80 and 86 nm, and volume median diameters were
184 and 195 nm, respectively. Lung tissues were assessed for alterations in global gene
expression (n = 4) 4 h after the last DE exposure on day 4. Mice were intranasally
challenged with OVA or saline on day 18 and then with OVA on day 28. Post-challenge
results demonstrated mild adjuvancy with antigen and DE exposure as evidenced by
significant increases in eosinophils, neutrophils, lymphocytes, and IL-6 in the BALF.
Antibody responses were not significantly affected by DE exposure, although a slight
increase in IgE after high dose exposure was observed. DE alone only increased
neutrophils, indicating the need for combined exposure to DE and antigen in the
development of allergic outcomes. Comparison of low DE (500 jug/m3)/OVA versus air/OVA
resulted in no significant changes in gene sets associated with this treatment. Comparison
July 2009
6-206
DRAFT-DO NOT CITE OR QUOTE

-------
of the high (2000 jug/m3) DE/OVA versus air/OVA, however, showed significant changes in 23
gene sets, including neutrophil homing and other chemokines, inflammatory cytokines,
numerous interleukins and TNF subtypes, and growth/differentiation pathways.
Summary of Findings for Allergic Responses
Studies conducted since the last review confirm and extend the 2004 PM AQCD's
(U.S. EPA, 2004, 056905) finding that PM can modulate immune reactivity in both humans
and animals to promote allergic sensitization and exacerbate allergic responses. Numerous
forms of PM, including inert materials, have been shown to function as adjuvants, and
although toxicological studies of relatively homogeneous materials demonstrate greater
adjuvancy for smaller particles, some analyses of ambient PM do not. Recent toxicological
studies comparing size fractions of well-characterized ambient PM for adjuvant activity in a
direct, controlled fashion via inhalation exposure suggest a role for oxidative potential, but
thus far the relative contributions of size and composition are not entirely clear. Although
epidemiologic studies examining specific allergic outcomes and short-term exposure PM are
relatively rare, the available studies, conducted primarily in Europe, positively associate
various PM size fractions with allergic rhinitis or hay fever and skin prick reactivity to
allergens. Similar findings from a number of long term studies are described in Chapter 7.
6.3.7. Host Defense
The normal, and very important, role of respiratory immune defense is the detection
and/or destruction of pathogens that enter the lung via inhalation and removal of damaged,
transformed (cancerous), or infected cells. Innate immune defenses of the respiratory tract
include mucociliary clearance, release of toxic antimicrobial proteins into airway surface
liquid, and activation of alveolar macrophages. The innate immune system is the earliest
responder to irritation or infection, initiating the normal inflammatory response including
the majority of detrimental inflammatory processes discussed. Activated macrophages and
epithelial cells release cytokines and chemokines that can bring into play the adaptive
immune system, which in turn can produce long-lasting pathogen-specific immune
responses critical for resolving and preventing infections.
July 2009
6-207
DRAFT-DO NOT CITE OR QUOTE

-------
6.3.7.1.	Epidemiologic Studies
Collectively, results from multicity studies of hospital admissions and ED visits for
respiratory infection as well as single-city studies conducted in the U.S. and Canada
(summarized in Figure 6" 14) show a positive association between PM and respiratory
infections. Lag structure was not investigated in most studies and effects have been
observed in association with current day concentration (Zanobetti and Schwartz, 2006,
090195) as well as with concentrations modeled using a 14 day distributed lag function
(Peel et al., 2005, 056305). Of studies examining multiple lag times, associations with
increasing lag times were observed (Dominici et al., 2006, 088398; Peel et al., 2005, 056305;
Peng et al., 2008, 156850). Although no significantly positive associations were reported,
Slaughter et al. (2005, 073854) observed a trend of increasing association with increasing
lag for acute respiratory infection ED visits with PMi, PM2.5, PM10 and PM10-2.5. This delay
in the onset of disease may reflect the time necessary for an infection to become established
and symptomatic. The majority of toxicological evidence, described below and in the 2004
PM AQCD (U.S. EPA, 2004, 056905), suggests that PM impairs innate immunity, the first
line of defense in preventing infection.
6.3.7.2.	Toxicological Studies
Several toxicological studies were cited in the 2004 PM AQCD (U.S. EPA, 2004,
056905) that demonstrated increased susceptibility to infectious agents following exposure
to PM. A limited number of new studies have evaluated the effect of PM on host defense in
rodents. Two recent studies have observed an increase in susceptibility to influenza
infection and respiratory syncytial virus in mice. However, one new study found that
woodsmoke had no effect on bacterial clearance in rodents.
Bacterial Infection
Several studies included in the 2004 PM AQCD demonstrated increased susceptibility
to infectious agents following exposure to various forms of PM. CAPs exposed aged rats
demonstrated increased S. pneumoniae burdens when a 24-h exposure (65 ug/m3) followed
infection (Zelikoff et al., 2003, 039009). In another study, IT exposure to ROFA was found to
affect bacterial clearance (Antonini et al., 2002, 035342). Examinations of mechanisms
related to PM interference with host defenses have demonstrated impaired mucociliary
clearance and modified macrophage phagocytosis and chemotaxis. Prolonged exposure to
July 2009
6-208
DRAFT-DO NOT CITE OR QUOTE

-------
inhaled particles at sufficiently high concentrations can lead to diminished clearance of PM
from the alveolar region of the lung, resulting in the accumulation of retained particles and
an accompanying chronic alveolar inflammation. Diminished clearance of PM may also
increase susceptibility to pulmonary infection by impeding clearance of pathogens.
Impaired phagocytosis by alveolar macrophages may contribute to a decrease in the lung's
capacity to deal with increased particle loads (as occurs during high-pollution episodes) or
infections and affect the local and systemic responses through the release of biologically
active compounds (cytokines, ROS, NO, isoprostanes).
Diesel Exhaust
Since the last review, several additional studies have reported impairment of
pathogen clearance following exposure to various sources of PM. All levels of DE (30, 100,
300 or 1,000 |ug/m3) decreased lung bacterial clearance in C57BL/6 mice exposed for 1 wk (7
days/wk, 6 h/day) prior to infection with Pseudomonas aeruginosa (Harrod et al., 2005,
088144). This effect appeared dose dependent up to 100 |ug/m3 and was not enhanced at
higher doses. Lung inflammation was not induced by DE in the absence of infection, but
infection-induced inflammation was exacerbated by DE at all concentrations without
apparent dose dependency. Measures of histopathology in infected animals were dose
dependently increased by DE exposure, peaking at 100 |ug/m3 and leveling off or decreasing
with higher doses. Particle deposition was readily apparent in the lungs after exposure to
the lowest concentration of 30 |ug/m3. A loss of ciliated cells was observed at 30 ug/m3 and
100 ug/m3 in large airways and in small airways at the higher dose. Alterations in Clara cell
morphology and function were observed at both doses as well. Concentrations of gases were
reported to be from 2.0-45.3 ppm NO, 0.2-4.0 ppm NO2, 1.5-29.8 ppm CO and 8-365 ppb for
SO2 (McDonald et al., 2004, 055644). PM mass median diameter was -100-150 nm at all
exposure levels (>90% below 1 |Jm in aerodynamic diameter), with lower exposure
concentrations having a slightly smaller size distribution (Reed et al., 2004, 055625).
Gasoline Exhaust
In a study by Reed et al. (2008, 156903), short or long-term exposure to fresh gasoline
exhaust (6h/day, 7day/wk for 1 wk or 6 months) did not affect clearance of P. aeruginosa
from the lungs of C57BL/6 mice. [Atmospheric characterizations are described above for the
Day et al. (2008, 190204) and McDonald et al. (2008, 191978) studies in Section 6.3.6.3,
Exacerbation of Allergic Responses/Gasoline Exhaust.]
July 2009
6-209
DRAFT-DO NOT CITE OR QUOTE

-------
Hardwood Smoke
Similar to gasoline exhaust, HWS does not appear to have significant impact on
pathogen clearance. C57BL/6 mice were exposed to 30-1000 |ug/m3 HWS by whole body
inhalation for 1 wk and 6 months (Reed et al., 2006, 156043). Long-term responses are
discussed in Sections 7.3.3.2 and 7.3.7.2. Concentrations of gases ranged from
229.0-14887.6 mg/m3 for CO, 54.9-139.3 jug/m3 for ammonia, and 177.6-3455.0 jug/m3
nonmethane volatile organic carbon in these exposures. Bacterial clearance of instilled
Pseudomonas aeruginosa was unaffected by HWS.
Residual Oil Fly Ash
Antonini et al. (2004, 097199) compared sources of ROFAin SD rats. Precipitator
ROFAinduced an inflammatory response and diminished pulmonary clearance of L.
monocytogenes while air heater ROFA had no effect on lung bacterial clearance at the same
dose of 1 mg/lOOg body weight. Precipitator ROFA generated a metal-dependent hydroxyl
radical suggesting that differences in metal composition were a determinant of the
immunotoxicity of ROFA. ROFA has also been shown to result in ciliated cell loss in BALB/c
mice after intranasal administration of 60 ug every other day for four days (Arantes-Costa
et al., 2008, 187137).
Viral Infection
Diesel Exhaust
Viral respiratory infections in early life are associated with increased incidence of
childhood asthma and other pulmonary diseases. DE exposure can enhance the progression
of influenza infection. BALB/c mice that were chamber exposed to DE 4 h/day for 5 days
and subsequently instilled with influenza A/Bangkok/1/79 virus had increased
susceptibility to influenza infection (Ciencewicki et al., 2007, 096557). Exposures to two
doses of DEP were conducted: 500 |ug/m3 (0.9 ppm CO, <0.25 ppm NO2, <2.5 ppm NO, and
0.06 ppm SO2) and 2000 |ug/m3 (5.4 ppm CO, 1.13 ppm NO2, 10.8 ppm NO, and 0.32 ppm
SO2). Responses were greater for animals exposed to 500 |u,g/m3 DEP than to 2000 |ug/m3,
and were associated with a significant increase in IL-6 protein and mRNA expression and
IFN-6 expression. The authors present the possibility that damage to the epithelium at the
higher dose prevented viral infection and replication. After exposure to 500 ug/m3 DEP
alone or prior to infection, decreased expression of surfactant proteins (SP) A and D was
observed. These proteins are part of the IFN-independent defense against influenza.
July 2009
6-210
DRAFT-DO NOT CITE OR QUOTE

-------
Similarly, Harrod et al. (2003, 097046) demonstrated decreased SP-A expression in
the lungs following DE exposure and linked it to increased susceptibility to respiratory
syncytial virus (RSV), the most common cause of respiratory infection in young children.
C57BL/6 mice, a relatively RSV-resistant strain, were exposed via inhalation to DE at a
concentration of 30 or 1000 ug/m3 DPM 6h/day for 7 consecutive days prior to intratracheal
viral inoculation. Gaseous copollutants ranged from 2.0-43.3 ppm for NOx (~ 90% NO),
0.94-29.0 ppm CO, and 8.3-364.9 ppb SO2. Exposure to 30 |ug/m3 DEP did not induce a
statistically significant increase in BALF cell numbers compared to air-treated, infected
animals. However, distinct consolidated inflammatory infiltrates were observed in the
peribronchial regions of RSV-infected animals exposed to this dose, along with alterations
in Clara cell morphology, decreased CCSP production by these cells, and occasional regional
myofibril layer thickening. These changes were more pronounced in RSV-infected animals
exposed to 1000 |ug/m3, and the higher dose also resulted in significant increases in
inflammatory cells, predominantly macrophages, in both uninfected and infected mice
compared to air-exposed controls. Both doses elicited significant levels of TNF-a and IFN-y
in the lungs of infected animals, but decreased levels of SP-A. Consistent with this study's
finding of decreased SP-A and increased viral gene and inflammatory cytokine expression
after DE exposure, SP-A"'" mice demonstrate decreased clearance of RSV concordant with
increased lung inflammation (LeVine et al., 1999, 156687). Thus, DEP may enhance
susceptibility to respiratory viral infections by reducing the expression and production of
SP (Ciencewicki et al., 2007, 096557; Harrod et al., 2003, 097046), although the
contribution of gaseous copollutants, in some instances concentrated lOOOx, should be
considered for both studies. SP are also essential for clearance of other pathogens, including
group B Streptococcus (GBS), Haemophilus influenzae, and P. aeruginosa (LeVine and
Whitsett, 2001, 155928).
A reduction in host defense molecules and an increase in viral entry sites was
observed by Gowdy et al. (Gowdy et al., 2008, 097226) after BALB/c mice were exposed to
HEPA filtered room air or DE at 0.5 or 2.0 mg/m3 for 4hr/day for one or five consecutive
days [O2 (%) 21.0 ± 0.10 or 20.7 ± 0.09, CO (ppm) 1.7 ± 0.15 or 5.4 ± 0.07, NOx (ppm) 2.0 ±
0.36 or 7.4 ± 0.61, SO2 (ppm) 0.0 ± 0.0 or 0.4 ± 0.3, number median (nm) 96.2 ± 2 or 97 ± 2,
volume median (nm) 238 ± 2 or 249 ± 2, OC/EC (wt ratio) 0.4 ± 0.04 or 0.4 ± 0.07 for the 0.5
or 2.0 mg/m3 exposures, respectively]. One of the more notable features of this study was
the observation that effects of extended exposure to the lower dose (0.5 mg/m3 for 5 days)
July 2009
6-211
DRAFT-DO NOT CITE OR QUOTE

-------
tended to persist beyond 18 h post-exposure. Exposure to DE significantly increased BAL
neutrophils in the higher dose group, and this response persisted beyond 18 h only after the
five day exposure. An increase in ICAM-1 expression (a viral entry site) was observed in
both dose groups, and was persistent in the lower dose group after a 5 day exposure.
Persistently elevated expression of pro-inflammatory cytokines IL-6 and TNF-a and pro-
allergic cytokine IL-13 was observed after 5 days of low dose exposure. Non-statistically
significant effects of either dose or dosing regimen included increased IFN-y and MIP-2.
Host defense molecules CCSP, SP-A and SP-D were decreased after either exposure
regimen, persisting beyond 18 h in the low dose group.
Taken together, these data suggest that exposure to DE can weaken host defenses, in
some cases persistently. A role for the PM component is supported by studies demonstrating
changes in host defense molecules and viral entry sites in vitro consistent with those
observed in vivo. In lung epithelial cells, DEP increased the mRNA expression of
intercellular adhesion molecule-1 (ICAM-l), low-density lipoprotein (LDL) and
platelet-activating factor (PAF) receptors, which can act as receptors for viruses or bacteria
(Ito et al., 2006, 096648). DEP may therefore enhance the susceptibility to infection by the
upregulation of bacterial and viral invasion sites in the lungs. Expression of the
6"defensin-2 gene, which is one antimicrobial mechanism of host defense in the airway, was
significantly inhibited by V and not Ni or Fe in airway epithelial cells incubated with
aqueous leachate of ROFA (Klein-Patel et al., 2006, 097092).
Immunosuppressive Effects of PM
Diesel Exhaust
DEP may affect systemic immunity. Decreased thymus weight was observed in female
F344 rats exposed to 300 |u,g/m3 DEP for 1 wk by Reed et al. (2004, 055625). Concentrations
of gases for this dose were reported to be approximately 16.1 ppm for NO, 0.8 ppm for NO2,
9.8 ppm for CO, and 115 ppb for SO2. Long-term responses are discussed in Section 7.3.8.
Summary of Findings for Host Defense
Toxicological studies demonstrate that short-term inhalation exposures to CAPs and
DE, but not gasoline exhaust or woodsmoke, can increase susceptibility to infection by
bacterial and viral pathogens. While gaseous copollutants may be contributing factors, a
role for particulate components is demonstrated by studies utilizing IT exposure and in
July 2009
6-212
DRAFT-DO NOT CITE OR QUOTE

-------
vitro studies of particles where biomarkers parallel those observed in vivo. Although ethical
considerations limit controlled exposure studies in humans, epidemiologic evidence reflects
an association between most PM size fractions and hospital admissions for respiratory
infections. Importantly, toxicological studies demonstrate impaired host defense against the
etiological agents of influenza, pneumonia (S. pneumoniae), and bronchiolitis (RSV), which
are commonly reported respiratory morbidities associated with PM.
6.3.8. Respiratory ED Visits, Hospital Admissions and
Physician Visits
The epidemiologic evidence presented in the 2004 PM AQCD (U.S. EPA, 2004,
056905) linking short-term increases in PM concentration with respiratory hospitalizations
and ED visits was consistent across studies. Recent investigations provide further support
for this relationship, with larger effect estimates observed among children and older adults.
However, effect estimates are clearly heterogeneous, with evidence of both regional and
seasonal differences at play.
Excess risk estimates for hospitalizations or ED visits for all respiratory diseases
combined, reported in studies reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905) fell
within the range of approximately 1 to 4% per 10 (Jg/m3 increase in PMio. On average,
excess risks for asthma were higher than excess risks for COPD and pneumonia.
Associations with ambient fine particles (PM2.5, PMi) and coarse thoracic particles (PM10-2.5)
were also reported in the limited body of evidence reviewed in the 2004 AQCD. Excess risk
estimates fell within the range of approximately 2.0 to 6.0% per 10 (Jg/m3 increases in PM2.5
or PM10-2.5 for all respiratory diseases combined as well as COPD admissions. Larger
estimates were reported for asthma admissions. Many of the associations of respiratory
admissions and ED visits with short-term PM2.5 concentration were statistically significant.
The associations with PM10-2.5 were less precise with fewer reaching statistical significance
(U.S. EPA, 2004, 056905). Finally, several studies reviewed in the 2004 AQCD reported
associations of PM with outpatient physician visits suggesting that the population impacted
by short-term increases in PM is not be restricted to those admitted to the hospital or
seeking medical attention through an ED.
July 2009
6-213
DRAFT-DO NOT CITE OR QUOTE

-------
Table 6-10. Description of ICD-9 and ICD-10 codes for diseases of the respiratory system.
Description
ICD 9 Codes
ICD 10 Codes
Diseases of the Respiratory System
460-519
J00-J99
Asthma
493
J45
C0PD and allied conditions
490-496 (asthma, chronic bronchitis, emphysema,
bronchiectasis, extrinsic allergic alveolitis)

Chronic lower respiratory diseases

J40-J47 (bronchitis, emphysema, other C0PD, asthma, status
asthmaticus, bronchiactasis)
Acute Respiratory Infections
460-466 (common cold, sinusitis, pharyngitis, tonsillitis,
laryngitis & tracheitis, bronchitis & bronchiolitis)

Acute Upper Respiratory Infections

J00-J06 (common cold, sinusitis, pharyngitis, tonsillitis, laryngitis
& tracheitis, croup & epiglottitis)
Acute bronchitis and bronchiolitis
466
J20-J22
Allergic Rhinitis
477
J30.1
Pneumonia
480-486
J13-J18
Wheezing
786.09

Hospital admissions or ED visits for respiratory diseases and ambient concentrations
of PM have been the subject of more than 90 peer-reviewed research publications since
2002 (see Annex E). Included among these new publications are several large single-city
and multicity studies. These new studies complement those reviewed in the 2004 AQCD by
examining the effect of several PM size fractions and components on increasingly specific
disease endpoints as well as evaluating the presence of effect modification by factors such
as season and region.
Specific design and methodological considerations of the large and multicity studies
included in this review were discussed previously (Section 4.4.10.2). Like the CVD
endpoints discussed, the respiratory endpoints examined in these studies were
heterogeneous and approaches to selecting cases for inclusion in the studies were varied.
ICD codes commonly used in hospital admission and ED visits studies are found in Table 6*
10.
6.3.8.1. All Respiratory Diseases
Findings from new studies of PM and respiratory hospitalization and ED visits among
children are summarized in Figure 6" 10. Results from new studies of adults are
summarized in Figure 6-11. Information on the PM concentrations during the relevant
study periods is found in Table 6-11.
July 2009
6-214
DRAFT-DO NOT CITE OR QUOTE

-------
Children
Barnett et al. (2005, 087394) used a case-crossover design to study respiratory
hospital admissions (ICD-9 460-519) of children (age groups 0, 1 to 4, 5 to 14 yr) in seven
cities in Australia and New Zealand from 1998 to 2001. All respiratory diseases (ICD10
J00-J99) except Mendelson's Syndrome, post-procedural disorders, asphyxia and certain
other symptoms (ICD10 codes J95.4-J95.9, R09.1, R09.8) were included in the study. In
addition, scheduled admissions and transfers from other hospitals were excluded. Using an
a priori lag (0"1 day avg), increases in respiratory hospital admissions of 2.0% (95% CI:
-0.13 to 4.3) among infants less than 1 year old, 2.3% (95% CI: 1.9-7.3) among children 1-4
years old and 2.5% (95% CI: 0.1-5.1) among children 5-14 years old and, per 10 (Jg/m3
increase in 24-h avg PMio were observed. Increases of 6.4% (95% CI: 2.7-10.3) among
infants less than 1 year and 4.5% (95% CI: 1.9-7.3) among children 1-4 yr per 10 (Jg/m3
increase in PM2.5 were observed.
Ostro et al. (2009, 191971) studied the effect of PM2.5 and components on respiratory
disease (ICD9 460-519) hospitalizations among children less than 19 yr from 2000 to 2003
in six counties in California. The full-year and cool season mean PM2.5 levels were 19.4
ug/m3 and 24.8 ug/m3, respectively. The nine components examined (EC, OC, nitrates,
sulfates, Cu, Fe, K, Si and Zn), were chosen because they made up relatively large
proportion of PM2.5, had a signal to noise ratio greater than 2, or the majority of their
values were greater than the LOD. Single day lags of 0-3 days were evaluated. The largest
risks were observed at lag 3 days for PM2.5 (2.8% 95%CP 1.2-4.3 per 10 |ug/m3), EC (5.4 95%
CP 0.8-10.3 per IQR) and Fe (4.7% (95% CI: 2.2-7.2 per IQR increase). Although not as
great, positive associations were also observed for OC, SO42 , nitrate, Cu and Zn.
In a study of PM2.5 from wildfires in California during 2003, Delfino et al. (2009,
191994) evaluated conducted stratified analyses comparing PM2.5 associations pre-, post-
and during the wildfires. Four age groups (0-4, 5-19, 20-64 and 65+) were considered in
these analyses. Authors found increased respiratory disease admissions in the periods
before (2.6% 95%CI: -5.4 to 11.3) and during (2.7% 95%CI: -1.6 to 7.6) the wildfires among
children 5-19 years old, but not after the wildfire period. Among younger children (0"4 yr),
hospital admission were increased during fire periods (4.5% 95% CI: 1*8.2) but not before or
after the wildfire period. Estimated zip code level PM2.5 concentrations were 90 ug/m3 and
75 |ug/m3 during heavy and light smoke conditions, respectively, compared to 20 ug/m3
during non-fire periods.
July 2009
6-215
DRAFT-DO NOT CITE OR QUOTE

-------
In the study of six cities in France described previously (PSAS), investigators report a
change of 0.4% (95%CP -1.2 to 2) per 10 (Jg/m3 increases in PM2.5 for all respiratory diseases
combined (ICD-10: J00-J99) among children from 0-14 years old (Host et al., 2008, 155852).
The same study reported a larger increase associated with PM10-2.5 of 6.2% (95% CP
0.4-12.3, 0-1 day avg) per 10 (Jg/m3 increase among children. A relatively large effect for
PM10-2.5 (31% 95% CP -4.7 to 80) was also observed in a single-city study of children less
than 3 years old in Vancouver (Yang et al., 2004, 087488). The non-significant PM2.5 effect
estimates were not presented in the publication. Luginaah et al. (2005, 057327) did not
observe significant increases in respiratory hospitalizations with increasing PM10
concentrations among male or female children in Ontario Canada, while Ulirsch et al.
(2007, 091332) reported increased admissions for respiratory hospitalizations, ED and
urgent care visits combined among children <17 years old in association with PM10.
Adults and All Ages Combined
In the study of 4 million ED visits from 31 hospitals in Atlanta described previously,
SOPHIA investigators reported an excess risk of 1.3% (95% CP 0.4-2.1, lag 0-2) per 10 (Jg/m3
increase in 24-h avg PM10 for ED visits for respiratory causes combined (ICD-9: 460-466,
477, 480-486, 491-493, 496, 786.09) among all ages in Atlanta during the period January,
1993 to August, 2000 (Peel et al., 2005, 056305). PM2.5, PM10-2.5, ultrafine number count and
PM2.5 components (SO42-, acidity, EC, OC, and an index of water-soluble transition metals)
were available for inclusion in analyses beginning August 1, 1998. Larger increases in ED
visits for respiratory diseases were associated with PM2.5 compared to PM10-2.5. Excess risks
of 1.6% (95% CP -0.003 to 3.5) per 10 (Jg/m3 increase in 24-h avg PM2.5 and 0.6% (95% CP
-3.6 to 5.1) per 10 |jg/m3 increase in PM10-2.5 were reported. Weaker, less precise associations
with components were reported and no increase with ultrafine particle count was observed.
Analyses with four additional years of data were conducted and more recently
reported by SOPHIA investigators (Tolbert et al., 2007, 090316). Single pollutant results
are included in Figure 6-11. As shown, effect of PM10 remains with the additional years of
data, while the effect of PM2.5 is diminished and a decrease in ED visits with PM10-2.5 was
observed. The association of PM10 with respiratory disease ED visits was robust to
adjustment for O3, CO and NO2. In another recent analysis using SOPHIA data from 1998
through 2002 to compare source apportionment methods, Sarnat et al. (2008, 097972)
reported that PM2.5 from mobile sources, PM2.5 from biomass burning and SO r -rich
July 2009
6-216
DRAFT-DO NOT CITE OR QUOTE

-------
secondary PM2.5 were associated with respiratory ED visits and associations were robust to
the choice of the method. Excess risks were statistically significant, ranging from
approximately 2%-4%, depending on the method.
In the French multicity study larger increases were observed in association with 2,4-h
avg PM10-2.5 concentration compared to PM2.5 concentration among adults as well as
children. Among adults 15-64 yr, investigators reported increases in respiratory
hospitalizations of 0.8% (95%CP -0.7, 2.3) and 2.6% (95%CP -0.5 to 5.8) per 10 |jg/m3 for
PM2.5 and PM10-2.5, respectively (lag 0-1 days) (Host et al., 2008, 155852).
In a study of respiratory hospital admission and ED visits (ICD-9 Codes 460-519)
among all ages conducted in Spokane, Washington, no associations were observed with any
size fraction of PM considered (e.g., PMi, PM2.5, PM10-2.5, PM10) (Slaughter et al., 2005,
073854). However, authors observe that there was a suggestion of greater effect estimates
with PM2.5 compared to PM10-2.5. Furthermore, several of the same investigators conducted a
source apportionment analysis using daily fine PM filter samples from the same residential
monitor in Spokane (Schreuder et al., 2006, 097959). In this investigation, PlVk.sfrom
vegetative burning in the previous day (lag l) was associated with respiratory hospital
admissions (2.3% [95% CP 0.9-3.8] per interquartile range increase in the source marker).
In a study of PM2.5 from wildfires in California during 2003, associations with respiratory
hospitalizations were generally stronger relative to associations in the periods before and
after the fires (Delfino et al., 2009, 191994). Among adults 20-64 years, an increase of 2.4%
(95% CP 0.5-4.4 per 10 (Jg/m3) during the wildfire period compared to 0.9% (95%CP -0.1 to
1.8 per 10 (Jg/m3) for all periods combined (pre-, post- and during wildfires).
Luginaah et al. (2005, 057327) examined respiratory hospital admissions in relation
to PM10 concentration across strata for age and gender and compared time series to case
crossover approaches. The results for all ages combined, which were relatively precise,
stratified by gender and all lags are presented in the figure! the largest estimates for PM10
were for adult males (15-64 years old), however. Fung et al. (2005, 093262) did not report
evidence of an association between respiratory admissions and 24-h PM10 concentration
among adults less than 65 yr, in a study in Vancouver, while Ulirsch et al. (2007, 091332)
reported a significant positive association among all ages and adults (18-64 yr) in two
Southeast Idaho cities for hospitalizations, ED and urgent care visits combined. This
estimate was robust to adjustment for gaseous pollutants.
July 2009
6-217
DRAFT-DO NOT CITE OR QUOTE

-------
Older Adults
Among older adults, MCAPS investigators observed largely null findings for PM2.5
and respiratory hospitalizations (ICD-9: 490-492, 464-466, 480-487) for the U.S. as a whole
but reported heterogeneity in effect estimates across the country that were explained by
regional and seasonal factors (Bell et al., 2008, 156266). The nationwide excess risk of
respiratory admissions with PM2.5 was 0.22% (95% posterior interval (PI): -0.12 to 0.56, lag
0) (Bell et al., 2008, 156266). The largest increase was observed during the winter in the
Northeast (1.76% [95% PL 0.60-2.93], lag 0). Significant increases in respiratory admissions
were also observed at lag 2. In an analysis of PM10-2.5, MCAPS investigators observed small
imprecise increases in respiratory admissions with 24-h PM10-2.5 concentration (0.33% [95%
PI: -0.21 to 0.86, per 10 |Jg/m3, lag 0]) (Peng et al., 2008, 156850), which decreased after
adjustment for PM2.5 (0.26% [95% PI: -0.32 to 0.84 per 10 |jg/m3 lag 0]). Associations with
PM2.5 increased (0.7% [95% PI: 0-1.5, lag 0]) or persisted (0.6% [95% PI: -0.2 to 1.25, lag 2]),
after adjustment for PM10-2.5.
Two recent MCAPS analyses evaluate the effect of PM2.5 components on respiratory
hospital admissions. Bell et al. (2009, 191997) analyzed a subset of MCAPS data restricted
to 106 counties with data available for both long-term average concentrations of PM2.5
components (Bell et al., 2007, 155683) and PM2.5 total mass (1999-2005). The components
evaluated included 20 chemicals with demonstrated toxicity or that contribute a large
proportion of PM2.5 mass (Al, NH4+, As, Ca, CI, Cu, EC, OCM, Fe, Pb, Mg, Ni, NO3", K, Si,
Na+, Ti, V, Zn). Increases in effect estimates of 511% (95% PI: 80.7-941) for EC, 223% (95%
PI: 36.9-410) for nickel and 392% (95% CI: 46.3-738) for vanadium per IQR increases in
county-specific component fraction were observed. Associations were somewhat reduced and
non-significant in two pollutant models. When Queens or New York County were excluded
the association of vanadium with hospital admissions lost significance. Associations were
also diminished when alternative lag structures were considered.
Peng et al. (2009, 191998) linked data on hospital admissions for respiratory causes
among older adults from 2000-2006 to daily air levels from the STN in 119 counties in
which both sets of data were available. Chemical components evaluated were SO42-, nitrate,
silicon, EC, OCM, sodium and ammonium ions. Single day lags of 0-2 days were considered.
These investigators found a 0.82% increase (95% PI: 0.22-1.44) per IQR increase in same
day OCM. After adjustment for the other components, a 1.01% (95% PI: 0.04-1.98, lag 0)
July 2009
6-218
DRAFT-DO NOT CITE OR QUOTE

-------
increase in respiratory admissions was observed. A similar effect estimate was observed at
lag 2.
French PSAS investigators reported a non-significant increase in hospitalizations for
respiratory diseases (ICD-10 J00-J99) with 24-h avg PM10-2.5 among older adults. PM2.5
estimates were also not significant and were closer to the null that estimates for PM10-2.5
(Host et al., 2008, 155852). Adjusted estimates from 2-pollutant models were not presented.
Positive associations of first hospitalization, overall hospitalizations and readmission for
respiratory diseases and PM10-2.5 were also reported among older adults in Vancouver (Chen
et al., 2005, 087555). PM10-2.5 was associated with an increase of 15% (95% CP 4.8-22.8) in
overall admissions per 10 |Jg/m3. Increases associated with PM10-2.5 were larger for
readmissions compared to overall admissions. The association for PM2.5 with overall
admissions was 5.1% (95% CP -4.9 to 13) and the association with readmissions was not
larger. In this study, effect estimates for PM10 and PM10-2.5 lost precision but were robust to
adjustment for gaseous pollutants while the estimate for PM2.5 was null after adjustment
for gaseous pollutants. In Vancouver, Fung et al. (2006, 089789) also report larger effect
estimates for PM10-2.5 than PM2.5 among adults 65 yr and older. These authors report
increased admissions of 1.8% (95% CP -2.5 to 5.8) per 10 (Jg/m3 increase in PM2.5 and 3.8%
(95% CP 0-7.6) per 10 |jg/m3 increase in PM10-2.5 (lag 0-1 day avg).
In a multicity Australian study, Simpson et al. (2005, 087438) examined the
association between fine particles measured by nephelometry and respiratory hospital
admissions (ICD-9 460-519) among older adults (65+ yr) and reported significant
associations (1.055 [95% CP 1.008-1.1045], lag 0-1 day avg) from a meta-analysis combining
effect estimates from all cities. Results from three statistical models were considered,
including standard GAM, which produced similar results.
Delfino et al. (2009, 191994) reported that PM2.5 from wildfire in California was
associated with respiratory hospital admissions among older adults (3% 95% CP 1.1-4.9 per
10 |ug/m3). In two analyses of data collected in Copenhagen Denmark between 1999 and
2004, several size fractions including ultrafine and accumulation mode (Andersen et al.,
2008, 189651) and PM10 sources (Andersen et al., 2007, 093201) were investigated in
relation to respiratory hospitalizations (J41-42, J43, J44-46) among adults greater than
65 yr of age. Of the size fractions examined (NC total, NC median diameter of 12 nm
[NCai.2], NCa23, NCa57, NCaioo, NCa2i2, PMio, PM2.5) NCa2i2, typically aged secondary
long-range transported, NC57 and PM10 were significantly associated with respiratory
July 2009
6-219
DRAFT-DO NOT CITE OR QUOTE

-------
hospitalizations (Andersen et al., 2008, 189651). PMio sources including biomass
combustion, secondary inorganic compounds, oil combustion and crustal were associated
with respiratory hospitalizations (Excess risks ranged from 3.5% to 5.4% per interquartile
range respectively) (Andersen et al., 2007, 093201). PMio associations were diminished
somewhat in 2 pollutant models (2007, 093201; Andersen et al., 2008, 189651); the authors
note that it was difficult to separate the effects of PMio and NCa2i2, which were highly
correlated in these data. PM2.5 was not associated with respiratory hospitalizations in these
data.
Results from other single-city studies offer somewhat consistent evidence for the
effect of PMio on respiratory admissions among older age groups. Ulirsch et al. (2007,
091332) found increases in hospitalizations, ED and urgent care visits combined among this
age group in 2 cities of Southeast Idaho. Two studies in Vancouver report increased
admissions for respiratory causes with the largest effects observed for a 3 day ma (0-2 days)
(Chen et al., 2005, 087555; Fung et al., 2006, 089789). Fung et al. (2006, 089789) observed
non-significant increases in admissions with PMio among older adults in Ontario, Canada
while the study by Luginaah et al. (2005, 057327). which was also conducted in Ontario, did
not provide compelling evidence for an effect that was robust to method selection, although
some increases among males were observed. Finally, a study of hospital admissions for
cardiopulmonary conditions combined among older adults (65+ yr) in Allegheny County, PA
found a positive association with PMio at lag 0 (Arena et al., 2006, 088631).
July 2009
6-220
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Location
Lag
Excess Risk (%)
Barnett et al. (2005, 0873941 Australia/NZ
0-1
PM2.5
< 1 y ¦
Australia/IMZ
0-1
1-4 y
Australia/IMZ
0-1
5-14 y
Host et al. (2008.1558521
6 Cities France
0-1 d avg
0-14 y
Ostro et al. (2009,1919711
6 counties CA
19 y
Delfino et al. (2009.190254)
CA wildfires
0-1
4 y
CA wildfires
0-1
5-19 y
Host et al. (2008.155852)
6 Cities France
0-1 d avg
PM102.5
0-14 y
Yang et al. (2004. 087488)
Vancouver
< 3 y
(31% (-4.7 to 80)
Barnett et al. (2005, 087394)
Au/NZ
0-1 d avg
PM10
<1 y.
Au/NZ
0-1 d avg
1-4 y
Au/NZ
0-1 d avg
5-14 y
Ulirsch et al. (2007.091332)
Idaho
0
0-17 y
Luginaah et al. (2005, 057327) Windsor Ontario
Females, 0-14 y
Windsor Ontario
Females, 0-14 y
Windsor Ontario
Females, 0-14 y
Windsor Ontario
Males, 0-14 y
Windsor Ontario
Males, 0-14 y
Windsor Ontario
Males, 0-14 y
Figure 6-10.
Excess risk estimates per 10 (Jg/m3 24-h avg PM concentration for studies of ED
visits and hospitalizations for respiratory diseases in children. Studies represented in
the figure include all multicity studies. Single-city studies conducted in the U.S. or
Canada are also included.
As shown in Figure 6" 10, studies of respiratory hospitalizations or ED visits reported
increased risks to children in association with all size fractions. However, increased risk
among boys was not observed in Ontario (Luginaah et al., 2005, 057327). Estimates are
imprecise and it is not clear if associations with fine, coarse, or both are driving
associations observed with PMio. Effect estimates for adults (and combined age groups) as
well as older adults are found in Figure 6-11. Effects observed in single-city studies are
generally imprecise but most studies report positive associations. Regional and seasonal
variation was observed with the largest effect estimate reported by Bell et al. (2008,
156266) in the Northeast during the winter. Although the number of studies examining
components or sources was limited, EC, OC, nickel, vanadium, and PM2.5 from mobile
sources were associated with increased respiratory admissions. Several additional studies
conducted outside the U.S. and Canada reported positive associations of respiratory
July 2009
6-221
DRAFT-DO NOT CITE OR QUOTE

-------
hospitalizations with PMio for different age groups and lags (Bedeschi et al., 2007, 090712;
Chen et al., 2006, 087947; Hanigan et al., 2008, 156518; Middleton et al., 2008, 156760;
Oftedal et al., 2003, 055623; Lai and Cheng, 2008, 180301; Larrieu et al., 2009, 180294),
PM2.5 (Hinwood et al., 2006, 088976; Neuberger et al., 2004, 093249; Vigotti et al., 2007,
090711), BS (Bartzokas et al., 2004, 093252; Tecer et al., 2008, 180030) and with PM10-2.5
(Tecer et al., 2008, 180030). Other studies reported no associations with PM10 (Vegni and
Ros, 2004, 087448) or TSP (Llorca et al., 2005, 087825).
6.3.8.2. Asthma
Results from multicity studies of hospital admissions and ED visits for asthma as well
as single-city studies conducted in the U.S. and Canada are summarized in Figure 6-12.
Studies reviewed in the 2004 AQCD are included for continuity. Concentrations of PM for
the relevant study period are found in Table 6-11.
Children
SOPHIA investigators (Peel et al., 2005, 056305) reported that, of the PM mass
indicators examined, the largest effect estimate observed using the a priori lag (0-2 day avg)
was the association of PM10 with pediatric (2-18 yr) asthma ED visits (1.6% [95% CP -0.2 to
3.4]). ED visits for both asthma (ICD-9: 493) and wheezing (ICD-9: 786.09) were included in
their study. Asthma hospital admissions (ICD-10 J45, J46, J44.8) in children less than
14 years old were examined in the Australia / New Zealand multicity study (Barnett et al.,
2005, 087394). In this study, associations for asthma hospital admissions with PM10 and
PM2.5 were increased but imprecise.
Lin et al. (2002, 026067) used both time series and case crossover approaches to
investigate the influence of PM on asthma hospitalization in children, 6" 12 years old, in
Toronto from 1981 to 1993. These authors report relatively small differences in results
obtained through bi-directional case crossover and time series approaches but indicate that
unidirectional case-crossover methods may overestimate the relative risks. Single to 7-
day avg lags were investigated and estimates appeared to increase and then level off at the
longer lags (0-2 day and 0-5 day lags are shown in figure). Effect estimates for PM2.5 are not
easily distinguished from the null but PM10-2.5 is significantly associated with asthma
admissions among boys and among girls. These associations were imprecise but robust to
adjustment for gaseous pollutants among all children combined.
July 2009
6-222
DRAFT-DO NOT CITE OR QUOTE

-------
Although Ostro et al. (2009, 191971) presented estimates for all respiratory diseases
combined, these authors note that PM2.5 and its components were associated with asthma
hospitalizations among the children in 6 counties of Los Angeles studied. Delfino et al.
(2009, 191994) examined the association of PM2.5 before, during, and after wildfires in
California with asthma hospitalizations among age and gender subgroups. Significant
associations were observed for children 0-4 yr among children during the wildfire period
(8.3% 95% CI: 2.1-14.9, per 10 jug/m3) but not before or after the wildfire period. For older
children, 5" 19 yr, non-significant increases in asthma hospitalizations were found before the
wildfire period but not during or after the fires.
July 2009
6-223
DRAFT-DO NOT CITE OR QUOTE

-------
Rpfprpnpp


Atlanta
R-7 H
Atlanta
R-7
Snnkann
1
fi Ritifis Franr.fi
R-1
Ralifnrnia Wilrlfims
R-1
Atlanta
R-7 rl
Atlanta
R-7
Snnkann
1
fi Ritifis Frannfi
R-1
Atlanta
R-7
Atlanta
R-7
Winrlsnr
1

7

3

7

3
Snnkanfi
1

7

3
Irlahn
R
Irlahn
R
flntarin
R

R-1

0-7
AHnlts (All Anes RnmhinpHI
Pfifilfital f7RRR. RRR3RR). All Anns*
Tnlhfirt fit al f7RR7. flPflHIfil. All anns*
Slaunhtfir fit al f7RRR. R73RR4}. All Anns*
Hnst fit al f7RRR. 1RRRR71
Dfilfinn fit al. 17Rfifl.inR7R4)
Pfifil fit al. (7RRR. RRR3RR). All Anns*
Tnlhfirt fit al. (7RR7. R9R31fi). All anns*
Slaunhtfir fit al. (7RRR. R73RR4). All Anns*
Hnst fit al. (7RRR. 1RRRR7)
Pfifilfital f7RRR. RRR3RR). All Anns*
Tnlhfirt fit al f7RR7. RRfilVI fi). All anns*
I nninaah fit al f7RRR. RR7377). All Anns
Slannhtfir fit al f7RRR. R73RR4). All Anns* Snnkanfi
IJIirsnhfit al f7RR7. (191337). All Anns*
lllirsrh fit al f7RR7. RRm7) **
Fnnn fit al OflflR. R937R7)
1R-R4 vrs
7R-R4\f
i
1R-R4 vrs
i _
j
Ffimalfi 		"y"
Ffimalfi p1
Ffimalfi r
Malfi !™
Malfi i
PM?
PMim 5
PM in
Malfi 1-
-H
1R-R4 vfiars
<" RR vfiars
< RR vfiars	* .
< RR vfiars
i
fllHpr flHultc IRR-i- ward
Roll pt al f7RRR 1RR7RR)
Hnst fit al f7RRR. 1RRRR7)
Fnnn fit al f7RRR. RRR7R9)
Rhfin fit al f7RRR. RR7RRR)
Dfilfinn fit al. (7RR9. 19R7R4)
Hnst fit al. (7RRR. 1RRRR7)
Pfinn fit al. (7RRR. 1RRRRR)
Fnnn fit al. (7RRR. RR97R9)
Rhfin fit al. (7RRR. RR7RRR)
Fnnn fit al f^nnfi RRG7RG1
lllirsnhfital /7RR7. R9133?r*
Rhfin fit al f7RRR. RR7RRR)
Fnnn fit al f7RRR. RR37R7)
7R7 IIS flnnntipc
R Ritifis Franr.fi
Vannnnvfir
Vannnnvfir
Ralifnrnia Wildfires
R Ritifis Frannfi
1 RR US Rnuntifis
Vannnin/fir
Vannnnvfir
Vannnnvfir
Vannnnvfir
Vannnnvfir
Vannnnvfir
Vannnnvfir
Vannnnvfir
Irlahn
Vannnnvfir
flntarin
*FD Visits
*"*Hnsnitali7atinn. FD. and Urnent Ram Visits
R
1
7
R
R
R-1
R
R-7H
R-4 rl
R.Rrl
1 rl
R-1 rl
R-7
R-1
R-1
R
1
7
R
R-7 rl
R-4
R-R
1
R-1
R-7
R
R-7 rl
R-4
R.R
R
R
R-1
R-7
R
R-1
R-7
PlWIn
Wintfir.IIS »"
Wintfir. Nnrthfiast 1
PMim
PM in
I	1	1	1	1	1-
¦4-2 0 2 4 6
~I	1	1	1
10 12 14 16
Figure 6-11. Excess risks estimates per 10 (Jg/m3 increase in 24-h avg PM for studies of ED visits
and hospitalizations for respiratory diseases among adults. Studies represented in the
figure include all multicity studies. Single-city studies conducted in the U.S. or Canada
are also included.
July 2009
6-224
DRAFT-DO NOT CITE OR QUOTE

-------
Hirshon et al. (2008, 180375) studied hospital admissions and ED visits by children 0-
17 years old in Baltimore, MD from June 2002-November 2002 in relation to zinc as a
component in PM2.5. Single day lags from 0 to 2 days were tested with the highest estimates
observed for the previous day A 23% (95% CP 7-41%) increase in admissions was observed
comparing medium (8.63-20.76 ng/m3) concentrations on the previous day to low
concentrations (<8.63 ng/m3) on the previous day. Previous day high concentration (>20.76
ng/m3) was associated with an increase in admissions of 16% (95% CI: -3 to 39) compared to
previous day low concentration. Zinc associations were robust to adjustment for EC, CO,
NO2, nickel, and chromium. However, evidence of effect modification by EC and NO2 at lags
1 and 2 was observed.
Mohr et al. (2008, 180215) used measurements of EC, O3, SO2, and total NOx from the
EPAsupersite in St. Louis for June 2001 through May 2003 to examine the association of
EC, temperature and season with asthma ED visits among children 2-17 years old. The
association of EC with asthma ED visits varied by age, season and weekday versus
weekend. The largest associations were observed for 2-5 year olds during the fall weekends
(3% 95% CP 1-5 per 0.1 jug/m3) and 11-17 year olds during winter weekdays (3% 95% CP
0-5, per 0.1 jug/m3) and summer weekends (9% 95% CP 2-17, per 0.1 jug/m3). Investigators
also report that temperature modified the effect of EC after adjusting for gaseous
copollutants such that the association of ED visits with EC increased with increasing
temperature during the summer and increased with decreasing temperature during the
winter. Authors attribute the pattern of associations to time activity patterns among this
age group.
In Copenhagen, Anderson et al. (2007, 093201) found an association between PM10
attributed to vehicle emissions and asthma hospitalizations among children 5-18 yr (5.4%
95% CP 0.57, 22.9 per 10 jJg/m3, 0-5 day avg) in Copenhagen, Denmark. In an analysis of
size distribution and number concentration, accumulation mode particles were most
strongly associated with asthma admissions (8% [95% CP 0-17] per 495 particles per cm3,
lag 0-5). (Andersen et al., 2008, 189651). In Helsinki, Halonen et al. (2008, 189507)
examined the association of various size fractions of PM (e.g., Aitken, accumulation mode,
PM2.5, PM10-2.5) with ED visits for asthma among children <15 yr. These authors evaluated
lags 0 to 5 and noted a different lag structure depending on age with children experiencing
greater effects at lags 3 to 5 days compared to adults at lag 0. Aitken, accumulation mode
particles and traffic-related PM were significantly and most strongly associated with
July 2009
6-225
DRAFT-DO NOT CITE OR QUOTE

-------
asthma visits among children, while no association with PM10-2.5 was observed in this age
group.
Sinclair and Tolsma (2004, 088696) investigated respiratory ambulatory care visits
using ARIES data in Atlanta, GA(also used by SOPHIA investigators) and health insurance
records. These authors evaluated three 3-day ma lags (0-2, 2-5 and 6-8 days) and reported
relative risks, with no confidence intervals, for significant results only (not included in
Figure 6-12). For childhood asthma outpatient visits, OHC, PM10-2.5, PM10, EC and OC were
significantly associated with ambulatory care visits at lags 0-2 or 2-5 days.
A study in Anchorage used medical records to examine effects of particle exposure on
pediatric asthma outpatient visits, inpatient visits and prescriptions for short-acting
inhalers (Chimonas and Gessner, 2007, 093261). Authors examined Medicaid claims for
asthma-related and lower respiratory infection visits among children less than 20 yr of age
for five years (approximately 25,000 children were enrolled in Medicaid each year between
1999 and 2003). Citing work done in the mid-1980s, the authors describe their city's
particles as arising primarily from natural, geologic sources (PM10), and to a lesser extent
from local automotive emissions (PM2.5) (Pritchett and Cooper, 1985, 156886). Using GEE in
a time-series analysis of daily and weekly effects of particle exposure on health outcomes,
the authors found that each 10 |ug/m3 increase in 24-h avg PM10 was associated with a 0.6%
increase (95% CP 0.1-1.3)) in outpatient visits for asthma. The same increase in weekly
PM10 concentration resulted in a 2.1% increase (95% CP 0.4-3.8)) in asthma visits, after
adjustment for gaseous pollutants. No meaningful associations were observed for PM2.5.
Adults and All Ages Combined
Results from the Atlanta SOPHIA study based on the a priori models examining a
3-day ma (lag 0-2 days) revealed no statistically significant associations with asthma
(ICD-9 493, 786.09) among all ages for any of the PM metrics studied (e.g., PM10, PM2.5,
PM10-2.5, ultrafine Particle Count, PM components) (Peel et al., 2005, 056305). However, the
14-day unconstrained distributed lag model produced an excess risk of 9.9% (95% CP
6.5-13.5). The authors note that associations of PM2.5 and OC with asthma tended to be
stronger during the warmer months. Sinclair and Tolsma (2004, 088696) report a
significant association between adult outpatient visits for asthma and UFPs, but not other
PM size fractions (not included in Figure 6-12 because only significant results were
presented).
July 2009
6-226
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Lag
Excess Risk Estimate
Children
Rarnfitt fit al f?00fi. (18977(11.Ails tralia/W7
0-1
0-1
_i	» 5-14 y
Lin et al. (2002, 026067), Toronto
Lin et al. (2002, 026067), Toronto
0-2
0-5 < ,	
0-?
i • Girls, 6-12 y

0-5
1 * Girls, 6-1? y
Delfino et al. (2006, 090745), 6 Counties CA, wildfire
0-1
J 	a	0-4 y

0-1
a 5-19 v
Chimonas and Gessner (2007, 093261), Anchoraqe
0 «	.	
o ¦
I Inpatient, 0-19 y
Lin et al. (2002, 026067), Toronto
0-2


0-5


Lin et al. (2002, 026067), Toronto
0-2
Rirk fi-1 ? y | ,

0-5
Rirk fi-1? y
>¦ >
i
Barnett et al. (2006, 089770), Australia/NZ
0-1
1-4 y-i	«	
5-14 v
PMio
Peel et al. (2005, 056305)*, Atlanta
0-2
"J—*	2-18 v

Lin et al. (2002, 026067), Toronto
Lin et al. (2002, 026067), Toronto
0-2
0-5
0-2
0-5
	J	•	 Girls, 6-12 y
l • Girls, 6-12 v

Chimonas and Gessner (2007, 093261), Anchoraqe
0
0
• Inpatient, 0-19 y
Outpatient, 0-19 y

i
Adults and All Ages Combined ¦
Peel et al. (2005, 056305)*, Atlanta
0-2

PM2.5
Itoet al. (2007,091262)*, NYC
0-1
0-1
0-1
! —•	 All Year
i 	•	 Warm Season
1 *	 Cool Season

Slauqhteret al. (2005, 073854)*, Spokane
1
2
3
	1	»	

Delfino et al. (2006, 090745), 6 Counties CA, wildfire
0-1
0-1
-j	*	 20-64 y

Sheppard et al. (2003, 042826) Seattle
0


i
Peel et al. (2005, 056305)*, Atlanta
0-2
	'i	
PMio 2.5
Slauqhteret al. (2005, 073854)*, Spokane
1
2
3
—i—~	
—¦*	
	*i	

Sheppard et al. (2003, 042826) Seattle
0


i
Peel et al. (2005, 056305)*, Atlanta
0-2
0-13

PMio
Jaffe etal. (2003, 041957)*, Ohio
3
3
2
3
—*-•	 All Cities, 5-34 y


i	»	 Cleveland, 5-34 y

Slauqhteret al. (2005, 073854)*, Spokane
1
2
3
i '
	I*—

Sheppard et al. (2003, 042826) Seattle
0
1—m—

-8 -4 0 2 4 6 8 10 14 18 22
Figure 6-12. Excess risk estimates per 10 (Jg/m3 increase in 24-h avg PMio, PM2.5 and PM10 2.5 for
studies of asthma ED visits* and hospitalizations. Studies represented in the
figure include all multicity studies. Single-city studies conducted in the U.S. or Canada
are also included.
Jaffe et al. (2003, 041957) examined the effects of ambient pollutants (PM10, O3, NO2
and SO2) during the summer months (June through August) on the daily number of ED
July 2009
6-227
DRAFT-DO NOT CITE OR QUOTE

-------
visits for asthma among Medicaid recipients aged 5 to 34 yr from 1991 to 1996 in
Cincinnati, Columbus and Cleveland. Lags 1 to 3 were tested and only statistically
significant lags were presented. For all cities combined, the overall effect estimate for 24-
h avg PMio was 1.0% (95% CI: -1.44 to 3.54 per 10 (Jg/m3 increase). The effect estimate for
Cleveland was the only significantly elevated estimate (2.3 95% CI: 0.0-4.9 per 10 |jg/m3
increase) when the cities were examined independently. The authors report results from
analyses indicating a possible concentration response for O3 but no consistent effects for
PM10.
In New York City, Ito et al. (2007, 091262) examined numbers of ED visits for asthma
among all ages (ICD-9 493) in relation to pollution levels from 1999 to 2002; several
weather models were evaluated. Although the association with NO2 was the strongest,
PM2.5 was significantly associated with asthma ED visits in each weather model (strongest
during the warm months) and remained significant after adjustment for O3, NO2, CO and
SO2. Slaughter et al. (2005, 073854) reported no associations with ED visits or
hospitalizations for asthma, among all ages, in Spokane Washington for the PM size
fractions studied (PMi, PM2.5, PM10, PM10-2.5). An association with CO, which the authors
attribute to combustion related pollution in general, was observed. The effect of 24-h avg
and l"h maximum PM2.5, PM10-2.5, EC and OC on ED visits for asthma comparing two
communities in New York City was investigated (New York State Department of, 2006,
090132). In the Bronx, an increase in visits of 3.1% (95% CP 0.6-6.2, per 10 jug/m3) was
observed in relation to 24-h avg PM2.5 with similar findings for 1-h maximum PM2.5. For
PM10-2.5, an increase of 2.7% (95% CP 0.0-5.4) was observed in the Bronx. Smaller, less
precise estimates were observed for Manhattan. Increased asthma visits were observed
with OC, EC and total metals. The association of 1-h maximum PM2.5 with ED visits was
robust to adjustment for copollutants.
Delfino et al. (2009, 191994) examined the association of PM2.5 before, during and
after wildfires in California with asthma hospitalizations among age and gender subgroups.
The increase among older adults greater than 65 yr of 10% (95% CP 3-17.8, per 10 jug/m3)
was larger than the increase among adults 20-64 yr of 4.1% (95% CP -0.5 to 9, per 10
ug/m3). For older adults, the association was stronger during the wildfire period compared
to the pre-wildfire period and did not diminish during the post-wildfire period.
Effect estimates from studies of hospital admissions and ED visits for asthma are
summarized in Figure 6-12. Associations with PM2.5 concentration among children are
July 2009
6-228
DRAFT-DO NOT CITE OR QUOTE

-------
imprecise and not consistently positive across different age groups and lags. Findings from
two studies of PM10-2.5 (Sinclair and Tolsma, 2004, 088696) as well as PM10 studies both
show positive associations, although estimates lack precision. Among adults and adults and
children combined, associations of asthma hospital admissions and ED visits with PM2.5
concentration were observed in most studies. Positive, non-significant associations of
PM10-2.5 concentration with asthma admissions and ED visits were observed in some studies
of adults. Again, PM10 estimates are more consistently positive and precise compared to
other size fractions. Associations were observed with several PM2.5 components (e.g., EC,
OC and zinc) and sources (e.g., traffic, wildfires). Many factors (e.g., the underlying
distribution of individual sensitivity and severity, medication use and other personal
behaviors) can influence the lag time observed in observational studies (Forastiere et al.,
2008,	186937). Excess risk estimates for asthma were generally sensitive to choice of lag
and increase with longer or cumulative lags times. Most additional single-city studies
conducted in Europe, South America and Asia, have investigated the associations of asthma
hospitalizations, ED visits or doctor visits and most have reported evidence of an
association with TSP (Arbex et al., 2007, 091637; Migliaretti and Cavallo, 2004, 087425;
2005,	088689), PM10 (Bell et al., 2008, 156266; Chardon et al., 2007, 091308; Chen et al.,
2006,	087947; Erbas et al., 2005, 073849; Galan et al., 2003, 087408; Jalaludin et al., 2004,
056595; Kuo et al., 2002, 036310; Lee et al., 2002, 034826; Lee et al., 2006, 090176; Ko et
al., 2007, 091639; Kim et al., 2007, 092837) and PM2.5 (Chardon et al., 2007, 091308; Ko et
al., 2007, 091639), while a few have not shown an association with PM10 (Larrieu et al.,
2009,	180294; Masjedi et al., 2003, 052100; Tsai et al., 2006, 089768; Yang and Chen, 2007,
092847).
6.3.8.3. C0PD
Results from multicity studies of hospital admissions and ED visits for COPD as well
as single-city studies conducted in the U.S. and Canada are summarized in Figure 6-13.
Studies reviewed in the AQCD are included in the figure for continuity. Concentrations of
PM for the relevant study period are found in Table 6-11.
In a study of Medicare recipients in 204 U.S. counties, Dominici et al. (2006, 088398)
reported an overall increase of about 1% in COPD hospitalizations (ICD-9 490-492)
associated with 24-h avg PM2.5, with the largest effects at lags 0 and 1. In this study effect
July 2009
6-229
DRAFT-DO NOT CITE OR QUOTE

-------
estimates were heterogeneous aross the U.S. with a significant increase of about 4%
observed in the Southeast at lag 0. In another study using Medicare data in 36 U.S. cities
(1986-1999) short-term exposure to PMio was associated with an increase in COPD hospital
admissions (ICD-9 490-496, excluding 493) of 1.47% (95% CI: 0.93-2.01, lag l) during the
warm season (Medina-Ramon et al., 2006, 087721). A smaller effect was observed during
the cold season.
In Atlanta, SOPHIA investigators reported a comparably sized effect estimate for
COPD (ICD-9 491, 492, 496) and 24-h avg PM2.5 (1.5% [95% CI: -3.1 to 6.3, 0-2 day avg]).
The association of PMio with COPD reported by Peel et al. (2005, 056305) was 1.8%
(95% CI: -0.6 to 4.3). No associations were observed for PM10-2.5, ultrafine or PM2.5
components. Slaughter et al. (2005, 073854) reported no associations between any size
fraction of PM in Spokane, Washington (PMio, PM2.5, PM10-2.5) and COPD (ICD-9 491, 492,
494, 496). In contrast, Chen et al. (2004, 087262) reported increases in COPD admissions
(ICD-9 490-492, 494, 496) for PMio (16.5% [95% CI: 6.88-27.02, 0-3 day avg]), PM2.5 (17.1%
[95% CI: 4.6-31.0) and PM10-2.5 (10.0% [95% CI: -1.2 to 22.8, 0-3 day avg]). However, the
estimates for PM metrics were diminished after adjustment for NO2, however.
Delfino et al. (2009, 190254) examined the association of PM2.5 from the wildfires of
2003 in California with COPD hospitalizations among age and gender subgroups. Among
older adults (65+ years), associations were similar across pre-, post- and wildfire periods
with none reaching significance. The increase for all periods combined in this age group was
1.9% (95% CI: -0.6 to 4.4, per 10 |ug/m3). Michaud et al. (2004, 089900) reported an
association for asthma and COPD ED visits combined with PMi (lag l) in Hilo, Hawaii in a
study designed to investigate the effect of volcanic fog.
Halonen et al. (2008, 189507) conducted a study of ED visits for COPD and asthma
combined (J41, J44-J46) among adults 15-64 yr and older adults greater than 65 yr. These
authors examined the effects of aitken mode particles, accumulation mode particles, PM2.5
and PM10-2.5 as well as several sources of PM2.5 (traffic, long range transported particles,
road dust and coal/oil combustion). Concentrations, lagged from 0-5 days, were examined
and the largest effects among older adults were observed in association with concurrent day
PM2.5, PM10-2.5 and accumulation mode particles, NO2 and CO concentrations. The PM2.5
association was diminished with adjustment for UFPs, NO2 and CO. A similar
diminishment was observed when PM10-2.5 was adjusted for PM2.5, NO2 and CO. However,
traffic related particles and long range transported particles (e.g., accumulation mode
July 2009
6-230
DRAFT-DO NOT CITE OR QUOTE

-------
particles such as carbon compounds, sulftates and nitrates from central Europe and Russia)
were associated with COPD and asthma among older adults.
This same research group conducted additional analyses of hospital admissions using
the same PM metrics focusing on older adults (65+ yr) (Halonen et al., 2009, 180379). The
PM2.5 results and lag structure were similar to the earlier ED visit study. The strongest
effect was for accumulation mode particles with COPD/asthma admissions. Traffic related
PM2.5 was associated with COPD/asthma admissions at lag 1 while no effect was observed
with concurrent day concentration. Long range transported particles and road dust were
also associated with admissions for asthma and COPD. With the exception of one study
conducted in Spokane Washington (Slaughter et al., 2005, 073854), associations have been
consistently observed for PM2.5 and PM10 with COPD in multicity and single-city studies
conducted in the U.S. and Canada. Associations with PM10-2.5 are fewer and less consistent.
A study that examined 7 single day lags in association with pooled COPD and asthma ED
visits in Finland reports that PM2.5, PM10-2.5, traffic sources as well as gaseous pollutants
had a more immediate effect in older adults (lags 0 and l) compared to the children
experiencing asthma (3-5 day lags) (Halonen et al., 2008, 189507). Larger estimates at
shorter lags were not observed consistently across other studies. Most single-city studies
conducted outside of the U.S. or Canada focused on PM10 (Chiu et al., 2008, 191989;
Hapcioglu et al., 2006, 093263; Ko et al., 2007, 091639; Ko et al., 2007, 092844; Martins et
al., 2002, 035059; Masjedi et al., 2003, 052100; Sauerzapf et al., 2009, 180082; Yang and
Chen, 2007, 092847).
6.3.8.4. Pneumonia and Respiratory Infections
Results from multicity studies of hospital admissions and ED visits for respiratory
infection as well as single-city studies conducted in the U.S. and Canada are summarized in
Figure 6-14. The figure includes studies of respiratory infection reviewed in the 2004
AQCD. Concentrations of PM for the relevant study period are found in Table 6-11.
July 2009
6-231
DRAFT-DO NOT CITE OR QUOTE

-------
Reference	Location	Lag	Excess Risk Estimates
Dominici et al. (2006, 088398)
204 Urban US
0

i


PM2.5


1

|-#-





2







0-2 d DL

r*-



Peel (2005, 056305)*, All Ages
Atlanta
0-2 d avg

—i-i—



Slaughter et al. (2005, 073854)*
Spokane
1
	1—
	1	





2

	{"¦	





3

	¦	



Chen (2004, 087262), 65 +
Vancouver
1

1

	•—



2

1

-9	



3

1


—•	


3 d avg

1
1


—•	
Delfino et al. (2009,190254), 65 +
Southern CA
0-1

' •



Ito (2003, 042856), 65 +
Detroit
0

	•	



Moolqavkar (2003, 0513161,65 +
LA
0

1 f




Peel (2005, 056305)*
Atlanta
3 d ma
	¦	
1


PM102.5
Slaughter et al. (2005, 073854)*
Spokane
1

	¦	





2

-1-1	





3

	Lft	



Chen (2004, 087262)
Vancouver
1

1—

	•—



2

1
1
—•—




3

1


	»


3 d avg



	•	

Ito (2003, 042856)
Detroit
0

1
-•	



Zanobetti & Schwartz (2003, 043119)
14 US Cities
0-1

1 -§-


PM10
Medina-Ramon (2006, 087721)
36 US Cities
0

l-t-
Warm




1

| +
Warm




0

~
Cold




1

*
Cold


Peel (2005, 056305)*
Atlanta
3 d ma

| ¦





0-13 d DL

i -

—I	

Slaughter et al. (2005, 073854)*
Spokane
1






2







3

	hi	



Chen (2004, 087262)
Vancouver
1

1
1

	~	



2

1

	•	



3

i

-•	



0-2 d avg

1
1


—•	
Ito (2003, 042856)
Detroit
0

	>-•	



Moolqavkar (2003, 051316)
LA
0

r*-



Moolqavkar (2003, 051316)
Cook County
0








I I 1 1 I
-12 -8 -4
r+T
0 2
1 1
4 6
1 1 1
8 12
1 1 1 1 1
16 20 24
Figure 6-13. Excess risks estimates per 10 (Jg/m3 increase in 24-h avg PMio, PM2.5 and PM102.5 for
studies of COPD ED visits* and hospitalizations among older adults (65+ yr, unlessother
age group is noted). Studies represented in the figure include all multicity studies.
Single-city studies conducted in the U.S. or Canada are also included.
July 2009
6-232
DRAFT-DO NOT CITE OR QUOTE

-------
Children
In the study of 7 cities in Australia and New Zealand, associations of PM2.5 with
pneumonia and acute bronchitis (ICD-10 J12-J17, J18.0, J18.1, J18.8, J18.9, J20, J21) were
observed among infants less than 1-years old (4.54% [95% CI: 0.00-9.20]) and children
1-4 years old (6.44% [95% CI: 0.26-12.85]) (Barnett et al., 2005, 087394). Although
quantitative results are only presented for all respiratory diseases combined, Ostro et al.
(2009, 191971) examined several specific respiratory diseases including acute bronchitis
and pneumonia. They reported that PM2.5 and its components were more strongly
associated with these endpoints compared to other respiratory diseases. Delfino et al. (2009,
190254) reports imprecise increases in admissions among children during wildfire periods
for acute bronchitis and bronchiolitis as well as pneumonia.
Inpatient and outpatient visits for lower respiratory tract infections among children
in Anchorage, Alaska, were not associated with PM2.5 or PM10 (Chimonas and Gessner,
2007, 093261). Lin et al. (2005, 087828) observed associations of respiratory infections
(ICD-9 464, 466, 480-487) with PM10-2.5 and PM10 that persisted after adjustment for
gaseous pollutants among subjects less than 15 years old. Analyses were stratified by
gender and both single and multiple day lags were examined (4 and 6 day avg were
presented). The largest significant effect estimates were for PM10-2.5. The size of the PM2.5
estimate varied by gender and was sensitive to the choice of lag. PM2.5 results were not
generally robust to adjustment for gases.
All Ages and Older Adults
SOPHIA investigators examined ED visits for upper respiratory tract infections (URI)
(ICD-9 460-466, 477) and pneumonia (ICD-9 480-486) among all ages. An excess risk of
1.4% (95% CP 0.4-2.5 per 10 |ug/m3, lag 0-2 day avg) for PM10 was associated with URI
visits. With the exception of a small increase in risk for OC of 2.8% (95% CI: 0.4-5.3, per
2 |ug/m3, 0-2 day avg) with pneumonia visits, Peel et al. (2005, 056305) reported no
association with other PM size fractions or components evaluated. However, Sinclair and
Tolsma (2004, 088696), who also used ARIES data in their analysis, reported significant
associations with outpatient visits for LRI. These associations were generally observed for
3-5-day ma lags, in association with PM10-2.5, PM10, EC, OC, and PM2.5 water soluble metals
(not pictured in figure because only significant lags were reported). No associations with
pneumonia for any size fractions were observed among all ages in a study conducted in
July 2009
6-233
DRAFT-DO NOT CITE OR QUOTE

-------
Spokane, Washington (not pictured because effect estimates were not reported) (Slaughter
et al., 2005, 073854).
French PSAS investigators examined the effect of fine and coarse PM on hospital
admissions for respiratory infection (ICD-10: J10-22) among all ages. Increases of 2.5%
(95% CP 0.1-4.8) and 4.4% (95%CP 0.9-8.0) per 10 |u,g/m3 were observed in association with
PM2.5 and PM10-2.5, respectively (Host et al., 2008, 155852). In a multicity study of older
adults (65+ yr) Medina-Ramon et al. (2006, 087721) examined hospital admissions for
pneumonia (ICD-9 480-487) in 36 U.S. cities in relation to 24-h avg PM10 concentration. An
increase in pneumonia admissions of 0.84% (95% CP 0.50-1.19, per 10 |ug/m3, lag 0) was
reported by these investigators during the warm season. Cold season associations were
weaker (0.30% 95% CP 0.07-0.53, per 10 |ug/m3, lag 0) as were lag 1 associations. Dominici
et al. (2006, 088398) investigated hospital admissions for all respiratory infections
including pneumonia (ICD-9 464-466, 480-487) among older adults in 204 urban U.S.
counties in relation to PM2.5 and reported a significant increased risk only at lag 2.
Heterogeneity in effect estimates was observed across the U.S. with the largest associations
reported for the South and Southeast.
In Boston, excess risks of pneumonia hospitalization in association with PM2.5, BC,
and CO were observed among older adults (Zanobetti and Schwartz, 2006, 090195). A
measure of non-traffic PM, e.g., the residuals from the regression of PM2.5 on BC, was not
associated with pneumonia hospitalization in these data. In a California study (Delfino et
al., 2009, 190254) effect estimates of similar magnitude for pneumonia admissions
associated with PM2.5 from wildfires among all ages combined and older adults (2.8% 95%
CP 0.7-5.0, per 10 |ug/m3, all ages combined). The PM2.5 association with acute bronchitis
and bronchiolitis admissions during the wildfire period for all age groups showed an
approximately 10% increase (9.6% 95% CP 1.8-17.9, per 10 |ug/m3). The increase was not
larger during the wildfire period compared to the pre-fire period for either endpoint.
In a study of 4 cities in Australia, statistically significant associations of pneumonia
and acute bronchitis with NO2 and particles measured by nephelometry (but not PM2.5
mass) were observed among older adults (Simpson et al., 2005, 087438). Halonen et al.
(2009, 180379) examined pneumonia among older adults (ICD10 J12-J15) in their most
recent analysis. Associations of PM2.5 (5.0% 95% CP 1.0-9.3, per 10 |u,g/m3, lag 5-day mean)
as well as accumulation mode particles with pneumonia admissions were observed.
July 2009
6-234
DRAFT-DO NOT CITE OR QUOTE

-------
Although the body of literature is small, several studies of children reported
associations of PM2.5, PM10-2.5 and PM10 with respiratory infections but endpoints studied
are heterogeneous and effect estimates are imprecise. Studies of adults show a similar
pattern of increased risk for each of these size fractions. Several other single-city studies
conducted outside the U.S. and Canada reported associations for PM10 (Cheng et al., 2007,
093034; Hwang and Chan, 2002, 023222; Nascimento et al., 2006, 093247) and PM2.5
(Hinwood et al., 2006, 088976) with hospitalization or ED visits for respiratory infections.
July 2009
6-235
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Location	Lao	Pollutant
Excess Risk Estimate
Rarnptt et al f?RRR RRQ77R1
Rhimnnas ft Rpssnpr f?RR7 RQ39R11
riplfinn Pt al f?RRR 1PR9R41
:= I in pt al f?RRR. RR7R?R)
o
I in pt al f?RRR. RR7R?R)
7 Ritips	fl-1
n.1
flnnhnranp Al fl
n
R nnnntips S R-1
R.1
R-1
Pnpnmnnia 
-------
Authors
Pollutant Outcome
Peng, Chang (2008,156850)
PM2.6
RESP

PM2.6+PM 10-2.6

Lin, Stieb (2005,087828)
PM2.5
RTI

PM2.6+CO, SO2,

Chen, Yang (2005, 087555)
PM2.6
RESP

PM2.6+CO+O3+N

Chen, Yang (2005,087362)
PM2.6
C0PD

PM2.5 "t-PM 10-2.5


PM2.6+CO


PM2.5+O3


PM2.6+NO2


PM2.6+SO2

Ito (2003, 042856)
PM2.6
PNEU

PM2.6+O3

Lippman et al. (2000, 011938)
PM2.B+SO2


PM2.6+NO2


PM2.B+C0

Moolgavkar (2003,042864) (2000,
PM2.S
C0PD

PM2.6+NO2

Sheppard (2003, 042826) (1999, 086921)
PM2.B
ASTH

PM2.6+CO

Lin, Chen (2002,026067)
PM2.B
ASTH

PM2.5


PM2.B+CO, S02,

Delfino (1998, 093624)
PM2.6
RESP

PM2.B+O3

Burnett, Cakmak (1997,084194)
PM2.B
RESP

PM2.G + O3


PM2.B+NO2


PM2.B+SO2


PM2.B

Thurston (1994, 043921)
PM2.B


PM2.B+O3

Peng, Chang (2008,156850)
PM102.B
RESP

PMlO-2.B + PM2.6

Lin, Stieb (2005, 087828)
PM10-2.B
RTI

PM102.B + C0, SO2,

Chen, Yang(2005,087555)
PMlO-2.5
RESP

PMlO-

Chen, Yang (2005, 087362)
PM10-2.B
C0PD

PM10.2.B + PM2.B


PM 10-2.5 + CO


PM10-2.B+03


PM102.B+N02


PM10-2.B+S02

Yang, Chen (2004,087488)
PMlO-2.5
RESP

PM10-2.S + C0, S02,

Ito (2003, 042856)
PM 10-2.B
C0PD

PM10-2.B + 03

Lippman et al. (2000, 011938)
PM 10-2.B + S02


PM102.B + NO2


PM 10-2.5 + CO


PM 10-2.5
PNEU

PM 10-2.5 + 03


PM 10-2.5 + S02


PMlO-2.5 + NO2


PM 10-2.5 + CO

Lin, Chen (2002,026067)
PM 10-2.5
ASTH

PMlO-2.5


PMlO-2.5 + C0, S02,

Burnett, Cakmak (1997, 084194)
PMlO-2.5
RESP

PM 10-2.5 + 03


PMlO-2.5 + NO2


PM 10-2.5 + S02


PM 10-2.5 + CO

PM2 5 adjusted for gases or other size fractions
«-
20.94(4.05-40.6)
22.34(5.08-42.75)
<	
<	*-
25.65(10.97-42.68)
20.69(4.65-37.5)
f
PMh>2 5 adjusted for gases or other size fractions
	>
23.31(6.55-42.68)
<	
30.97(-4.7-80)
60.55(4.83-154.322)
20.55(3.58-41.68)

i—r
1—1—1—1—1
Figure 6-15.
1—1—1—1—1—1—r
-8 -4 0 2 4 6 8 12 16
Excess risk estimates per 10//g/m3 increase in PM2.5, and PM102.5 for studies of ED
visits* and HAs for respiratory diseases. Results for multipollutant models.
July 2009
6-237
DRAFT-DO NOT CITE OR QUOTE

-------
Table 6-11. PM concentrations in studies of respiratory diseases published since 2002.
Pollutant Author
Location
Mean Concentration (/vg/m3)
Upper Percentile
concentrations (/vg/m3)
Puffin
Andersen et al. (2007, 093201)
Cooenhaqen, Denmark
25124
75th: 301 99th: 72
Barnett et al. (2005, 087394)
7 Cities, Australia, NZ
16.5-20.6
Max: 50.2-156.3
Chardon et al. (2007, 091308)
Paris
23
Max: 97.3
Chen et al. (2004, 087262; 2005, 087555)
Vancouver, Canada
13.3
Max: 52.2
Chimonas and Gessner (2007, 093261)
Anchoraqe, Alaska
27.6
Max: 421
Funa et al. (2005, 093262)
Ontario, Canada
38
Max: 248
Fund et al. (2006, 089789)
Vancouver, Canada
13.3
Max: 52.17
Gordian and Choudhurv (2003, 054842)
Anchoraqe, AK
36.11
Max: 210.0
Jaffe et al. (2003, 041957)
Cincinnati
43
Max: 90
Jalaludin et al. (2004, 056595)
Svdnev. Australia
22.8
Max: 44.9
Lin et al. (2002,026067)
Toronto, Canada
30.16
Max: 116.20
Lin et al. (2005,087828)
Ontario, Canada
20.41
Max: 73
Luainaah et al. (2005, 057327)
Ontario, Canada
50.6
Max: 349
Medina-Ramon et al. (2006, 087721)
36 U.S. Cities
15.9-44.0
NR
Moolqavkar (2003, 051316)
Los Anqeles, CA
22 (median)
Max: 86
Moolqavkar (2003, 051316)
Cook Countv. IL
35 (median)
Max: 365
Peel et al. (2005, 056305)
Atlanta, GA
27.9
Max: 44.7
Sinclair and Tolsma (2004, 088696)
Atlanta, GA
29.03
NR
Slaughter et al. (2005, 073854)
Sookane, WA
NR
Max: 41.9 (usinq 90% of concentrations)
Tolbert et al. (2007,090316)
Atlanta, GA
26.6
90th: 42.8
Ulirsch et al. (2007,091332)
Idaho
23.2
Max: 183.0
Yanq et al. (2004, 087488)
Vancouver, Canada
13.3
Max: 52.2
Zanobetti (2003, 042812); Samet et al. (2000, 010269)
14 U.S. Cities
24.4-45.3
Max 94.8-605.8
PM,S
Andersen et al. (2007, 093201)
Cooenhaqen, Denmark
10
99'": 28
Barnett et al. (2005, 087394)
7 Cities Australia, NZ
8.1-11
Max: 29.3-122.8
Bell et al. (2008,156266)
202 U.S. counties
12.92
98th: 34.16
Chardon et al. (2007, 091308)
Paris, France
14.7
75th: 18.2
Chen et al. (2004, 087262; 2005, 087555)
Vancouver, Canada
7.7
Max: 32
Chimonas and Gessner (2007, 093261)
Anchoraqe, AK
6.1
Max: 69.8
Delfino et al. (2009,190254)
6 counties, CA
18.4-32.7
45.3-76.1 (mean durinq wildfire oeriod)
Dominici et al. (2006, 088398)
204 U.S. counties
13.4
75th: 15.2
Funa et al. (2006, 089789)
Vancouver, Canada
7.72
Max: 32
Halonen et al. (2008,189507)
Helsinki, Finland
NR; Median - 9.5
Max: 69.5
Host et al. (2008,155852)
6 Cities France
13.8-18.8
95th: 25.0-33.0
Itoetal. (2007,091262)
New York, NY
All vr: 15.1
All vr: 95th: 32
Lin et al. (2002,026067)
Toronto Canada
17.99
Max: 89.59
Lin et al. (2005,087828)
Ontario, Canada
9.59
Max: 73
Moolqavkar (2003, 051316)
Los Anqeles, CA
22 (median)
Max: 86
Peel et al. (2005, 056305)
Atlanta, GA
19.2
90th: 32.3; 98th: 39.8
Sinclair and Tolsma (2004, 088696)
Atlanta, GA
17.62
NR
Sheooard et al. (2003, 042826)
Seattle, WA
16.7
98th: 46.6
Slauqhter et al. (2005, 073854)
Sookane, WA
NR
Max: 20.2 (usinq 90% of concentrations)
Tolbert et al. (2007,090316)
Atlanta, GA
17.1
90th: 28.8; 98th: 38.7
Yanq et al. (2004, 087488)
Vancouver, Canada
7.7
Max: 32.0
Zanobetti and Schwartz (2009,188462)
112 U.S. cities



Chen et al. (2004, 087262; 2005, 087555)
Vancouver, Canada
5.6
Max: 24.6
Funq et al. (2006, 089789)
Vancouver, Canada
5.6
Max: 27.07
Halonen et al. (2008,189507)
Helsinki, Finland
NR; Median: 9.9
Max: 101.4
Host et al. (2008,155852)
6 Cities France
7.0-11.0
95th: 12.5-21.0
Lin et al. (2002,026067)
Toronto, Canada
12.17
Max: 68.00
Lin et al. (2005,087828)
Ontario, Canada
10.86
Max: 45
Peel et al. (2005, 056305)
Atlanta, GA
9.7
90th: 16.2
Penq et al. (2008,156850)
108 U.S. counties
NR; Median: 9.8
75th: 15.0
Sinclair and Tolsma (2004, 088696)
Atlanta, GA
9.67
NR
Sheooard et al. (2003, 042826)
Seattle, WA
16.2
Max: 88
Slauqhter et al. (2005, 073854)
Sookane, WA
NR
NR
Tolbert et al. (2007,090316)
Atlanta, GA
9
90th: 15.1; Max: 50.3
Yanq et al. (2004, 087488)
Vancouver, Canada
7.7
Max: 24.6
III TRAFINF
Andersen et al. (2008,189651)
Cooenhaqen, Denmark
Mean oarticles/cmJ: 6847
99th: 19,895 oarticles/cmJ
Halonen et al. (2008,189507)

NR: Median oarticles/cmJ: 8,203
Max: 50,990 oarticles/cmJ
July 2009
6-238
DRAFT-
DO NOT CITE OR QUOTE

-------
6.3.8.5. Copollutant Models
Some studies have investigated potential confounding by copollutants through the
application of multipollutant models. Several Canadian studies of respiratory hospital
admissions reported larger effects for PM10-2.5 compared to PM2.5 that were robust to
adjustment for gaseous pollutants (Chen et al., 2005, 087555; Lin et al., 2002, 026067; Yang
et al., 2004, 087488). The COPD associations between PM2.5 and PM10-2.5 reported by Chen
et al. 2004 remained positive but were diminished slightly after adjustment for NO2. The
associations reported by Ito et al. 2003 of PM2.5 and PM10-2.5 with pneumonia hospital
admissions remained after adjustment for gases while the association of PM10-2.5 with
COPD admissions was not robust to adjustment for O3. Associations reported by Burnett et
al. (1986, 084184), Moolgavkar et al. (2003, 042864) and Delfino et al. (1998, 09362 Dwere
not consistently robust to adjustment for gaseous copollutants. In the MCAPS study the
effect of PM2.5 was robust to adjustment for PM10-2.5 while the PM10-2.5 effect on respiratory
admissions was diminished after adjustment for PM2.5 (Peng et al., 2008, 156850).
Multiple pollutant analyses for other size fractions and components have been
conducted in a some additional studies. Effect estimates for PM10 were robust to adjustment
for gases in several recent studies (Tolbert, 2007, 090316; Ulirsch et al., 2007, 091332;
Anderson and Bogdan, 2007, 156214). PM10 associations with respiratory disease did not
change in models also containing total number concentration, nor did the association of
ACP diminish after adjustment for UFP concentration (Anderson and Bogdan, 2007,
156214). Finally, Peng et al. (2009, 191998) reports an OCM effect that was robust to
adjustment for other components while the associations with nickel, vanadium and EC
were somewhat diminished in models containing multiple components.
Inconsistency across these study findings is likely due to differences in the correlation
structure among pollutants as well as differing degrees of exposure measurement error.
6.3.9. Short-term Exposure to PM and Respiratory Mortality
An evaluation of studies that examined the association between short-term exposure
to PM2.5 and PM10-2.5 and mortality provides additional evidence for PM-related respiratory
health effects. Although the primary analysis in the majority of mortality studies evaluated
consists of an examination of the relationship between PM2.5 or PM10-2.5 and all-cause (non-
July 2009
6-239
DRAFT-DO NOT CITE OR QUOTE

-------
accidental) mortality, some studies have examined associations with cause-specific
mortality including respiratory-related mortality.
Multicity mortality studies that examine the PM-respiratory mortality relationship on
a national scale - Franklin et al. (2007, 091257): 27 U.S. cities and Zanobetti and Schwartz
(2009, 188462): 112 U.S. cities - have found consistent positive associations between short-
term exposure to PM2.5 and respiratory mortality of approximately 1.68% per 10 |ug/m3 at
lag 0-1 (see Section 6.5). The associations observed on a national scale are consistent with
those presented by Ostro et al. (2006, 087991) in a study that examined the PM2.5-mortality
relationship in 9 California counties (2.2% [95% CP 0.6-3.9] per 10 jug/m3). An evaluation of
studies that examined additional lag structures of associations found smaller respiratory
mortality effect estimates when using the average of lag days 1 and 2 (1.01% [95% CP -0.03
to 2.05] per 10 |u,g/m3) (Franklin et al., 2008, 155779), and associations consistent with those
observed at lag 0-1 when examining single day lags, specifically lag 1 (1.78% [95% CP
0.2-3.36]). Although the overall effect estimates reported in the multicity studies evaluated
are consistently positive, it should be noted that a large degree of variability exists between
cities when examining city-specific effect estimates potentially due to differences between
cities and regional differences in PM2.5 composition (see Figure 6-25). A limited number of
studies that examined the PM2.5-respiratory mortality association have conducted extensive
analyses of potential confounders, as a result, PMio-mortality studies provide evidence
which suggests that PM2.5 risk estimates are fairly robust to the inclusion of gaseous
copollutants in models. Overall, the respiratory PM2.5 effects observed were larger, but less
precise than those reported for all-cause (nonaccidental) mortality (see Section 6.5), and are
consistent with the effect estimates observed in the single- and multicity studies evaluated
in the 2004 PMAQCD.
Zanobetti and Schwartz (2009, 188462) also examined PM10-2.5 mortality associations
in 47 U.S. cities and found evidence for respiratory mortality effects (1.16% [95% CP 0.43,
1.89] per 10 |ug/m at lag 0-1), which are somewhat larger than those reported for all-cause
(non-accidental) mortality (0.46% [95% CP 0.21, 0.671] per 10 |u,g/m). In addition, Zanobetti
and Schwartz (2009, 188462) reported seasonal (i.e., larger in spring) and regional
differences in PM10-2.5 respiratory mortality risk estimates. However, single-city studies
conducted in Atlanta, GA (Klemm et al., 2004, 056585) and Vancouver, CAN ((Villeneuve et
al., 2003, 055051) reported no associations between short-term exposure to PM10-2.5 and
respiratory mortality. The difference in the results observed between the multi- and single-
July 2009
6-240
DRAFT-DO NOT CITE OR QUOTE

-------
city studies could be due to a variety of factors including differences between cities and
compositional differences in PM10-2.5 across regions (see Figure 6"30). Only a small number
of studies have examined potential confounding by gaseous copollutants or the influence of
model specification on PM10-2.5 mortality risk estimates, but the effects are relatively
consistent with those studies evaluated in the 2004 PM AQCD.
6.3.10. Summary and Causal Determinations
6.3.10.1. PM2.5
Several studies of the effect of PM2.5 on hospital admissions for respiratory diseases
reviewed in the 2004 AQCD reported positive associations for several diseases. The 2004
AQCD presented limited epidemiologic evidence of PM2.5 being associated with respiratory
symptoms (including cough, phlegm, difficulty breathing, and bronchodilator use);
observations for PM2.5were positive, with slightly larger effects for PM2.5 than for PM10. In
addition, mortality studies reported relatively higher PM2.5 risk estimates for respiratory-
related mortality compared to all-cause (non-accidental) mortality. Controlled human
exposure studies did not provide support for effects of CAPs on respiratory symptoms.
Small decrements in peak flow for both PM2.5 and PM10 in asthmatics and nonasthmatics
were reported in epidemiologic studies included in the 2004 PM AQCD, whereas controlled
human exposure and animal toxicological studies reported few or no effects on pulmonary
function with inhalation to CAPs. In addition, the 2004 PM AQCD presented a number of
controlled human exposure and toxicological studies that reported mild pulmonary
inflammation following exposure to PM2.5 CAPs and DE or DEP, as well as ROFA or other
metal-containing PM in animals. The 2004 PM AQCD described controlled human exposure
studies showing increases in allergic responses among previously sensitized atopic subjects
after short term exposure to DEP. These observations were supported by many toxicological
studies that added to existing evidence demonstrating that various forms of PM could
promote allergic disease and exacerbate allergic asthma in animal models. Toxicological
studies also indicated that PM2.5 increased susceptibility to respiratory infection.
Overall, in recent studies PM2.5 effects on respiratory hospitalizations and ED visits
have been consistently observed. Most effect estimates were in the range of-1-4% and were
observed in areas with mean 24-h PM2.5 concentrations between 6.1 and 22 ug/m3. Further,
recent studies have focused on increasingly specific disease endpoints such as asthma,
July 2009
6-241
DRAFT-DO NOT CITE OR QUOTE

-------
COPD and respiratory infection. The strongest evidence of an association comes from large
multicity studies of COPD, respiratory tract infection and all respiratory diseases among
Medicare recipients (65+ years old) published since 2004 (Dominici et al., 2006, 088398;
Bell et al., 2008, 156266). Studies of children have also found evidence of an effect of PM2.5
on hospitalization for all respiratory diseases, including asthma and respiratory infection.
However, many of these effect estimates are imprecise, their magnitude and statistical
significance are sensitive to choice of lag, and some null associations were observed.
Although the association of PM2.5 with pediatric asthma was not examined specifically, it is
noteworthy that one of the strongest associations observed in the Atlanta-based SOPHIA
study was between PM10 and pediatric asthma visits! PM2.5 makes up a large proportion of
PM10 in Atlanta (Peel et al., 2005, 056305); Positive associations between PM2.5 (or PM10)
and hospital admissions for respiratory infection (Figure 6" 14) are supported by animal
toxicological studies which add to previous findings of increased susceptibility to infection
following exposure to PM2.5. These include studies demonstrating reduced clearance of
bacteria (Pseudomonas, Listeria) or enhanced pathogenesis of viruses (influenza, RSV)
after exposure to DE or ROFA.
Epidemiologic studies that examined the association between PM2.5 and mortality
provide additional evidence for PlV^.r,-related respiratory effects (Section 6.3.9). The
multicity studies evaluated found consistent, precise positive associations between short-
term exposure to PM2.5 and respiratory mortality ranging from 1.67 to 2.20% at mean 24-h
PM2.5 avg concentrations above 13 ug/m3. Although examinations of potential confounders of
the PM2.5-respiratory mortality relationship are limited, the observed associations are
supported by PMicrmortality studies, which found that PM risk estimates remained robust
to the inclusion of copollutants in models.
Epidemiologic studies of asthmatic children have observed increases in respiratory
symptoms and asthma medication use associated with higher PM2.5 or PM10 concentrations.
Associations with respiratory symptoms and medication use are less consistent among
asthmatic adults, and there is no evidence to suggest an association between respiratory
symptoms with PM2.5 among healthy individuals. In addition, respiratory symptoms have
not been reported following controlled exposures to PM2.5 among healthy or health-
compromised adults (Section 6.3.1.2).
Although more recent epidemiologic studies of pulmonary function and PM2.5 have
yielded somewhat inconsistent results, the majority of studies have found an association
July 2009
6-242
DRAFT-DO NOT CITE OR QUOTE

-------
between PM2.5 concentration and FEVi, PEF, and/or MMEF. In asthmatic children, a
10 (Jg/m3 increase in PM2.5 is associated with a decrease in FEVi ranging from 1-3.4%
(Section 6.3.2.1). A limited number of controlled human exposure studies have reported
small decreases in arterial oxygen saturation and MMEF following exposure to PM2.5 CAPs
with more pronounced effects observed in healthy adults than in asthmatics or older adults
with COPD (Section 6.3.2.2). In toxicological studies, changes in pulmonary function have
been observed in healthy and compromised rodents after inhalation exposures to CAPs
from a variety of locations or DE. A role for the PM component of DE is supported by altered
pulmonary function in healthy rats after IT instillation of DEP (Section 6.3.2.3).
Several lines of evidence suggest that fine PM promotes and exacerbates allergic
disease, which often underlies asthma (Section 6.3.6). Although epidemiologic studies
examining specific allergic outcomes and short-term exposure to PM are relatively rare, the
available studies, conducted primarily in Europe, positively associate PM2.5 and PM10 with
allergic rhinitis or hay fever and skin prick reactivity to allergens. Short term exposure to
DEP in controlled human exposure studies has been shown to increase the allergic response
among previously sensitized atopic subjects, as well as induce de novo sensitization to an
antigen. Toxicological studies continue to provide evidence that PM2.5, in the form of CAPs,
resuspended DEP, or DE, but not woodsmoke, spurs and intensifies allergic responses in
rodents. Proposed mechanisms for these effects include mediation by neurotrophins and
oxidative stress, and one study demonstrated that effects were mediated at the epigenetic
level (Liu et al., 2008, 156709).
A large body of evidence, primarily from toxicological studies, indicates that various
forms of PM induce oxidative stress, pulmonary injury, and inflammation. Notably, CAPs
from a variety of locations induce inflammatory responses in rodent models, although this
generally requires multiday exposures. The toxicology findings are consistent with several
recent epidemiologic studies of PM2.5 and the inflammatory marker eNO, which reported
statistically significant, positive effect estimates with some inconsistency in the lag times
and use of medication. In asthmatic children, a 10 (Jg/m3 increase in PM2.5 is associated
with an increase in eNO ranging from 0.46 to 6.99 ppb. Several new controlled human
exposure studies report traffic or dieseHnduced increases in markers of inflammation (e.g.,
neutrophils and IL-8) in airway lavage fluid from healthy adults. Recent studies have
provided additional evidence in support of a pulmonary oxidative response to DE in
humans, including induction of redox-sensitive transcription factors and increased urate
July 2009
6-243
DRAFT-DO NOT CITE OR QUOTE

-------
and GSH concentrations in nasal lavage. In addition, exposure to WS has recently been
demonstrated to increase the levels of eNO and malondialdehyde in breath condensate of
healthy adults (Barregard et al., 2008, 155675). Preliminary findings indicate little to no
pulmonary injury in humans following controlled exposures to fine urban traffic particles or
DE, in contrast to a number of toxicological studies demonstrating injury with CAPs or DE
(Sections 6.3.5.2 and 6.3.5.3, respectively).
Recent studies have reported associations of hospital admissions, ED or urgent care
visits for several respiratory diseases with PM2.5 components and sources including Ni, V,
OC and EC, woodsmoke and traffic emissions, in studies of both children and adults.
Delfino et al. (2003, 090941; 2006, 090745) found positive associations between EC and OC
components of PM and asthma symptoms and between EC and eNO. Particle composition
and/or source also appears to heavily influence the increase in markers of pulmonary
inflammation demonstrated in studies of controlled human exposures to PM2.5. For
example, whereas exposures to fine CAPs from Chapel Hill, NC have been shown to
increase BAL neutrophils in healthy adults, no such effects have been observed in similar
studies conducted in Los Angeles. In addition, differential inflammatory responses have
been observed following bronchial instillation of particles collected at different times or
from different areas (Section 6.3.3.2). One new study found that the increased airway
neutrophils previously observed by Ghio et al. (2000, 012110) in human volunteers after
Chapel Hill CAPs exposure could be largely attributed to the content of sulfate, Fe, and Se
in the soluble fraction (Huang et al., 2003, 087377).
In summary, new evidence of ED visits and hospital admissions builds upon the
positive and statistically significant evidence presented in the 2004 PM AQCD to support a
consistent association with ambient concentrations of PM2.5. Most effect estimates with
respiratory hospitalizations and ED visits were in the range of ~1% to 4% and were
observed in areas with mean 24-h PM2.5 concentrations between 6.1 and 22 jug/m3. The
evidence for PM2.5 induced respiratory effects is strengthened by similar HA and ED visit
associations for PM10, along with the consistent positive associations observed between
PM2.5 and respiratory mortality in multicity studies. Panel studies also indicate associations
with PM2.5 and respiratory symptoms, pulmonary function, and pulmonary inflammation
among asthmatic children. Further support for these observations is provided by recent
controlled human exposure studies in adults demonstrating increased markers of
pulmonary inflammation following DE and other traffic-related exposures, oxidative
July 2009
6-244
DRAFT-DO NOT CITE OR QUOTE

-------
responses to DE and woodsmoke, and exacerbations of allergic responses and allergic
sensitization following exposure to DEP. Although not consistent across studies, some
controlled human exposure studies have reported small decrements in various measures of
pulmonary function following controlled exposures to PM2.5. Numerous toxicological studies
demonstrating a wide range of responses provide biological plausibility for the associations
between PM2.5 and respiratory morbidity observed in epidemiologic studies. Altered
pulmonary function, mild pulmonary inflammation and injury, oxidative responses, AHR in
allergic and non-allergic animals, exacerbations of allergic responses and increased
susceptibility to infections were observed in a large number of studies involving exposure to
CAPs, DE, other traffic-related PM and woodsmoke. The evidence for an effect of PM2.5 on
respiratory outcomes is somewhat restricted by limited coherence between some of the
findings from epidemiologic and controlled human exposure studies for the specific health
outcomes reported and the sub-populations in which those health outcomes occur. Although
there is evidence for respiratory symptoms among asthmatic children in epidemiologic
panel studies, the studies of hospital admissions and ED visits provide more evidence for
effects from COPD and respiratory infections than for asthma. Additionally, controlled
human exposure studies report greater effects in healthy adults when compared to
asthmatics or those suffering from COPD. Finally, there is limited information which could
explain the relationship between the clinical and subclinical respiratory outcomes observed
and the magnitude of the PM2.5"respiratory mortality associations reported.Therefore, the
evidence is sufficient to conclude that a causal relationship is likely to exist between
short-term PM2.5 exposures and respiratory effects.
6.3.10.2. PM10-2.5
The 2004 PM AQCD presented the results from several epidemiologic studies of
respiratory symptoms and thoracic coarse particles, which provided limited evidence for
cough and effects on morning PEF. Toxicology data for PM10-2.5 were extremely limited, and
there were no controlled human exposure studies presented in the 2004 PM AQCD that
evaluated the effect of PM10-2.5 on respiratory symptoms, pulmonary function, or
inflammation. Epidemiologic studies of the effect of PM10-2.5 on hospitalizations or ED visits
for respiratory diseases (i.e., pneumonia, COPD and respiratory diseases combined)
reviewed in the 2004 AQCD reported positive associations. Additionally, the few mortality
July 2009
6-245
DRAFT-DO NOT CITE OR QUOTE

-------
studies that examined cause-specific mortality suggested somewhat larger risk estimates
for respiratory mortality compared to all-cause (non-accidental) mortality
Several new studies report associations between PM10-2.5 and respiratory
hospitalizations with the most consistent evidence among children (Figures 6-10 through 6-
14), however, effect estimates are imprecise. Although a number of studies provide evidence
of respiratory effects in older adults, a recent analysis of MCAPS data reports that weak
associations of PM10-2.5 with respiratory hospitalizations are further diminished after
adjustment for PM2.5. It is not clear that PM10-2.5 estimates across all populations and
regions are confounded by PM2.5. An examination of PM10-2.5 mortality associations on a
national scale found a strong association between PM10-2.5 and respiratory mortality, but
this association varied when examining city-specific risk estimates (Zanobetti and
Schwartz, 2009, 188462). Additionally, copollutant analyses were not conducted in this
study, and the associations observed are inconsistent with those reported in single-city
studies. There is greater spatial heterogeneity in PM10-2.5 compared to PM2.5 and
consequently greater potential for exposure measurement error in epidemiologic studies
relying on central site monitors. This exposure measurement error may bias effect
estimates toward the null.
Mar et al. (2004, 057309) provide evidence for an association with increased
respiratory symptoms in asthmatic children but not asthmatic adults. Consistent with this,
controlled human exposures to PM10-2.5 have not been observed to affect lung function or
respiratory symptoms in healthy or asthmatic adults. However, increases in markers of
pulmonary inflammation have been demonstrated in healthy volunteers. In these studies,
an increase in neutrophils in BAL fluid or induced sputum was observed, with additional
evidence of alveolar macrophage activation associated with biological components of PM10-2.5
(i.e., endotoxin). Toxicological studies using inhalation exposures are still lacking, but
pulmonary injury and inflammation have been observed in animals after IT exposure and
both rural and urban PM10-2.5 have induced these responses. In some cases, PM10-2.5 from
urban air was more potent than PM2.5 (Section 6.3.3.3). PM10-2.5 respiratory effects may be
due to components other than endotoxin (Wegesser and Last, 2008, 190506).
Overall, the most compelling new evidence comes from a number of recent
epidemiology studies conducted in Canada and France showing significant associations
between respiratory ED visits or hospitalization and short-term exposure to PM10-2.5. Effects
have been observed in areas where the mean 24-h avg PM10-2.5 concentrations ranged from
July 2009
6-246
DRAFT-DO NOT CITE OR QUOTE

-------
7.4 to 13.0 |Jg/m3. The strongest relationships were observed among children, whereas
studies of adults and older adults show less consistent evidence of an association. While
controlled human exposure studies have not observed an effect on lung function or
respiratory symptoms in healthy or asthmatic adults in response to exposure to PM10-2.5,
healthy volunteers have exhibited increases in markers of pulmonary inflammation.
Toxicological studies using inhalation exposures are still lacking, but pulmonary injury has
been observed in animals after IT exposure to both rural and urban PM10-2.5, which may not
be entirely attributed to endotoxin. Overall, epidemiologic studies, along with the limited
number of controlled human exposure and toxicological studies that examined PM10-2.5 and
respiratory outcomes, provide evidence that is suggestive of a causal relationship between
short-term PM10 2.5 exposures and respiratory effects.
6.3.10.3. Ultrafine PM
The 2004 PM AQCD included a few epidemiologic or controlled human exposure
studies which provided limited evidence of an association between ultrafine PM and
respiratory symptoms, medication use, or decreased pulmonary function! none assessed
inflammation and no studies of controlled human exposure to ultrafine PM were available.
Evidence from toxicological studies presented in the 2004 AQCD, although limited,
suggested that exposure via inhalation to high concentrations of ultrafine TiC>2 may
increase pulmonary inflammation in healthy rodents. Since the publication of the 2004
AQCD there has been an increased focus among the scientific community on gaining a
better understanding of the potential health effects associated with exposure to ultrafine
particles (UFPs). A number of recent controlled human exposure and toxicological studies
have evaluated respiratory responses following exposures to fresh DE. While these
atmospheres contain both fine and UFPs, the MMAD is typically < 100 nm, and therefore
the results of these studies may be used to support findings from studies utilizing other
sources of UFP
Recent epidemiologic studies conducted in Copenhagen, Denmark and Helsinki,
Finland reported associations between UFPs and HA or ED visits for respiratory diseases
including childhood asthma and pneumonia in adults (Halonen et al., 2008, 189507!
Andersen et al., 2008, 189651). The median UFP number concentrations in Copenhagen
and Helsinki were 6,243 particles/cm3 and 8,203 particles/cm3, respectively. Associations
July 2009
6-247
DRAFT-DO NOT CITE OR QUOTE

-------
between UFP and ED visits for respiratory diseases were not observed in the Atlanta-based
SOPHIA study, where the mean UFP number concentration was 38,000 particles/cm3.
A single recent epidemiologic study has examined associations between UFP and
pulmonary function, and observed that asthmatic adults exhibited decreased lung function
after exposure to diesel traffic pollution in London (McCreanor et al., 2007, 092841). Two
new controlled human exposure studies have reported small decreases in pulmonary
function among healthy adults approximately 20 hours following exposure to Los Angeles
UF CAPs (100 |ug/m3, particle count 145,000/cm3) or UF EC (50 |ug/m3, particle count 10.8 x
106/cm3) (Pietropaoli et al., 2004, 156025; Gong, 2008, 156483). Exposures to lower
concentrations of UF CAPs (~50 ug/m3, particle count 120,662/cm3) from Chapel Hill, NC
did not result in any changes in pulmonary function between 0 and 18 hours after exposure
(Samet et al., 2009, 191913). However, while Gong et al. (2008, 156483) did not observe any
effect of exposure to UF CAPs on markers of pulmonary inflammation, Samet et al. (2009,
191913)reported an UF CAPs-induced increase in IL-8 in BAL fluid at 18 hours post-
exposure. A limited number of studies have also demonstrated increases in the pulmonary
inflammatory response following exposure to ultrafine and fine particles from DE, which
may be enhanced by exposure to O3 (Section 6.3.3.2).
Altered pulmonary function and inflammation have also been observed in
toxicological studies of DE and ultrafine model particles (see Sections 6.3.2.3 and 6.3.3.3).
Although the contributions of gaseous components of DE to changes in respiratory function
are unknown, IT instillation of DEP can result in similar effects. In one rat model,
pulmonary inflammation is observed after exposure to ultrafine carbon black at
concentrations as low as 180 |ug/m3 (Harder et al., 2005, 087371). However, inflammatory
responses vary considerably depending on the animal model, dose, test material, and
exposure duration. In cases where pulmonary inflammation is not observed, oxidative
stress is often evident (Section 6.3.4.2). For example, although gasoline exhaust does not
appear to induce inflammation, the exhaust particles have been shown to increase ROS.
Oxidative stress is a major mechanism by which PM may exert effects (see Chapter 5), and
some toxicological studies suggest that UFPs are more potent than fine particles, possibly
due to a higher proportion of pro-oxidative OC and PAH content and greater surface area
with which to deliver these components.
The relationship between exposure to UFP and pulmonary injury has not been widely
examined. No association with pulmonary injury biomarkers was found for UFP in a
July 2009
6-248
DRAFT-DO NOT CITE OR QUOTE

-------
European multicity epidemiologic study (Timonen et al., 2004, 087915). In controlled
human exposure studies, UFP from woodsmoke resulted in significantly increased markers
of injury in healthy adults, but this effect was not evident in COPD sufferers exposed to DE
(Section 6.3.5.2). Exposure of neonatal rats to ultrafine iron-soot particles resulted in a
significantly reduced rate of cell proliferation in the proximal alveolar region, which
suggests that postnatal lung development may be susceptible to air pollution, consistent
with impaired lung function growth observed in children (Pinkerton et al., 2004, 087465).
In contrast, no histopathological responses were evident in adult mice exposed to ultrafine
iron-soot particles (Last et al., 2004, 097334). Some toxicological studies have observed
pulmonary injury after inhalation of DE or gasoline exhaust (Section 6.3.5.3). In studies
that evaluated ambient PM size fractions from a variety of European and U.S. cities for
relative toxicity in rodents following IT exposure, ultrafine PM was generally less injurious
than the larger size fractions. However, the ultrafine fraction of Montana coal fly ash
induced greater injury and inflammation than the PM10-2.5 fraction (Gilmour et al., 2004,
057420).
In rodent studies, ultrafine CAPs appeared to be more potent than fine CAPs in
inducing and exacerbating allergic responses (Section 6.3.6.3). In addition to CAPs,
ultrafine carbon black or iron-soot particles, but not particles from fresh gasoline exhaust,
have been shown to induce or exacerbate allergic responses in mice. Bacterial clearance
appears unaffected by HWS or gasoline engine exhaust. Diesel exhaust, however, has been
shown to reduce bacterial clearance, impair defenses against viral infection, and reduce
thymus weight, which may indicate systemic immunosuppression.
Taken together, a limited number of epidemiologic studies have provided some
evidence of an association between short-term exposure to UFPs and respiratory symptoms
as well as asthma hospitalizations. However, these findings have been inconsistent across
studies. The ultrafine number concentrations reported in the hospital admissions studies
ranged from a median of 6243 particles/cm3 to a mean of 38,000 particles/cm3. Although the
effect of controlled exposures to UFPs has not been extensively examined in humans, two
controlled human exposure studies have observed small ultrafine particle-induced
decreases in pulmonary function. However, no increases in respiratory symptoms have been
reported and effects on pulmonary inflammation are not consistent. The results from
animal toxicological studies examining the respiratory effects of UFPs are mixed, but
several studies demonstrate oxidative, inflammatory, and allergic responses. Some effects,
July 2009
6-249
DRAFT-DO NOT CITE OR QUOTE

-------
such as inflammation or pulmonary histopathology, may be observed only in particular
animal models (e.g., immature or compromised). Additionally, although a number of
controlled human exposure and toxicological studies using controlled exposures to fresh DE
report respiratory effects, the relative contributions of gaseous copollutants remain
unresolved. Thus, the current collective evidence is suggestive of a causal relationship
between short-term UFP exposure and respiratory effects.
6.4. Central Nervous System Effects
While evidence of an effect of PM on the central nervous system was not presented in
the 2004 PM AQCD (U.S. EPA, 2004, 056905), a limited number of recent epidemiologic,
controlled human exposure and toxicological studies provide some evidence that exposure to
PM may be associated with changes in neurological function. The majority of studies
included in this section are of short-term exposure, however, there are also a few studies of
long-term exposure. As CNS effects of PM are a newly emerging area, and since there are so
few studies, all studies that evaluate CNS responses are included in this section.
6.4.1. Epidemiologic Studies
Chen and Schwartz (2009, 179945) used extant data on CNS function from the Third
National Health and Nutrition Examination Survey (NHANES III) to characterize the
association between cognitive function in adults (ages 20-59) and exposure to ambient air
pollution. Three computerized neurobehavioral tests were used: a simple reaction time test
(SRTT), a basic measure of visuomotor speed; a symbol digit substitution test (SDST) on
coding ability! and a serial digit learning test (SDLT) on attention and short-term memory.
The authors used annual PMio concentrations to approximate the long-term exposure to
ambient air pollution prior to the NHANES-III examination. Increased PMio levels were
associated with reduced performance in all three neurobehavioral tests, and were
particularly strong for SDST and SDLT scores in models adjusted for age and sex. However,
after additional adjustment for race/ethnicity or SES, the magnitudes of these associations
were greatly diminished and largely null. It is possible that the observed associations
disappeared after adjustment for race/ethnicity and SES due to the potential confounding
by residential segregation of ethnic minorities and poorer people in areas with high levels of
ambient PMio concentrations.
July 2009
6-250
DRAFT-DO NOT CITE OR QUOTE

-------
Two additional epidemiologic studies evaluated the effect of ambient PM on the CNS
(Calderon-Garciduenas et al., 2008, 156317; Suglia et al., 2008, 157027). These studies
examined long-term exposure to non-specific PM indicators and are detailed in Annex E.
6.4.2.	Controlled Human Exposure Studies
In a recent controlled human exposure study, Cruts et al. (2008, 156374) exposed 10
healthy males (18-39 years old) to filtered air and dilute DE (300 |jg/m3 particulate
concentration) for 1-h using a randomized crossover study design. Changes in brain activity
were measured during and following exposure using quantitative electroencephalography
(QEEG). Exposure to DE was observed to significantly increase the median power
frequency (MPF) in the frontal cortex during exposure, as well as in the hour following the
completion of the exposure. While this study does provide some evidence of an acute cortical
stress response to DE, it is important to note that the QEEG findings are very nonspecific,
and could have been caused by factors other than diesel PM such DE gases (e.g., CO, NO
and NO2) or the odor of the DE.
6.4.3.	Toxicological Studies
Evidence is mounting that the CNS may be a critical target of PM and that adverse
health effects may result from PM exposure. Whether these health effects are a direct or
indirect effect of PM has not yet been established. One hypothesis suggests that ultrafine
PM which deposits onto nasal olfactory epithelium enters the CNS by axonal olfactory
transport to the olfactory bulb and leads to a cascade of effects involving inflammatory
cytokines and ROS. An increased potential for neurodegenerative processes may ensue.
Evidence for translocation of ultrafine PM to the olfactory bulb via olfactory neurons is
discussed in Chapter 4, but its relevance to CNS health effects is unknown. Another
hypothesis suggests that brain inflammation occurs secondarily to PM-mediated systemic
inflammation. Finally, it has been suggested that PM-stimulation of the autonomic nervous
system via respiratory tract receptors results in inflammatory or other effects in the CNS.
This is an emerging field with many unknowns.
July 2009
6-251
DRAFT-DO NOT CITE OR QUOTE

-------
6.4.3.1. Urban Air
Calderon-Garciduenas et al. (2003, 156316) conducted a long-term observational
study in mongrel dogs from Mexico City and Tlaxcala. DNA damage and inflammation in
the brain and respiratory tract were evaluated in dogs living in Mexico City (exposed group)
and dogs living in Tlaxcala (control group). These cities are similar in altitude but differ in
air pollutant levels. Measurements of air pollutant levels were presented only for Mexico
City, the more polluted city. Statistically significant greater levels of apurinic/apyrimidinic
sites (an indicator of DNA damage) were observed in the olfactory bulbs and hippocampus
of Mexico City dogs compared with controls. These differences were not seen in other brain
regions examined or in nasal respiratory epithelium. In addition, Mexico City dogs
demonstrated greater histopathological changes in the respiratory and olfactory epithelium
of the nasal cavity compared with controls. Immunohistochemical staining of brain tissue
from the Mexico City dogs demonstrated greater immunoreactivity for NFkB, iNOS,
cyclooxygenase-2, glial fibrillatory acidic protein (GFAP), ApoE, amyloid precursor product
and 6-amyloid compared with controls. These results are indicative of inflammation and
stress protein responses. This study has several limitations given that the dogs were of
mixed breeds and of variable ages and that there was no standardization of exposures or
diets. However results suggest a possible relationship between air pollution and brain
inflammation.
6.4.3.2. CAPS
Several new inhalation studies have provided evidence of CNS effects due to ambient
PM exposures. In one study, Campbell et al. (2005, 087217) exposed ovalbumin-sensitized
BALB/c mice to filtered air or near-highway Los Angeles CAPs (a 20-fold concentration of
fine+ultrafine or ultrafine only! mean exposure concentration ultrafine 282.5 ug/m3 and fine
441.7 ug/ma) for 4 h/day and 5 days/wk over a 2-wk period. The animals were subsequently
challenged with ovalbumin to elicit an allergic response in the lungs! brain tissue was
obtained 1 day later. Exposure to CAPs, but not filtered air, resulted in activation of the
immune-related transcription factor NF-kB and upregulation of the cytokines TNFa, and
IL-la in the brain, demonstrating pro-inflammatory responses that could contribute to
neurodegenerative disease. While this study demonstrates CAPs effects in an allergic
animal model, it is not known whether these responses also occur in non-allergic animals.
July 2009
6-252
DRAFT-DO NOT CITE OR QUOTE

-------
In a second study, control or ovalbumin-sensitized and challenged Brown Norway rats
were exposed for 8-h to filtered air or fine CAPs (Grand Rapids, MI; 500 jug/m3fine PM)
(Sirivelu et al., 2006, 111151). Brain tissue was obtained 1 day later. CAPs exposure
resulted in brain region-specific modulation of neurotransmitters. In animals which were
not pretreated with ovalbumin, statistically significant increases in norepinephrine were
observed in the paraventricular nucleus and olfactory bulb of CAPs-exposed rats compared
with filtered air controls. In animals which were pretreated with ovalbumin, a statistically
significant increase in dopamine was observed in the medial preoptic area in CAPs-exposed
rats compared with controls. Furthermore, exposure to CAPs resulted in a statistically
significant increase in serum corticosterone. These data suggest that the
hypothalamo-pituitary-adrenal axis (i.e., stress axis) maybe activated by PM exposure,
causing aggravation of allergic airway disease. The authors discuss the possible role of the
olfactory bulb in mediating neuroendocrine control of autonomic activities involved in
respiratory and cardiovascular functions! however these relationships require clarification.
Pro-inflammatory responses were examined in a subchronic CAPs study involving
normal (C57BL/6J) and ApoE"'" (Kleinman et al., 2008, 190074). Mice were exposed to
filtered air or to two concentrations of ultrafine CAPs from a near-highway area of central
Los Angeles (average of 30.4 and 114.2 |jg/m3) for 5 h/day and 3 days/wk over a 6-wk period.
Brain tissue was harvested one day after the last exposure and cortical samples prepared.
CAPs exposure resulted in activation of transcription factors, with a dose-dependent
increase observed for AP-1 and an increase in NFkB observed at the higher concentration.
Increased levels of GFAP (representing activation of astrocytes) and phosphorylated JNK
(representing MAP kinase activation) were observed at the lower but not higher
concentration of CAPs. No changes were observed in levels of or activation of the other MAP
kinases p38 and ERK or of IkB. These findings provide evidence that inhalation of CAPs
can lead to activation of cell signaling pathways involved in upregulation of pro-
inflammatory cytokine genes in the cortical region of the mouse brain.
In another study utilizing normal (C57BL/6) and ApoE"'" mice, brain histopathology
was examined following a 4-mo chronic exposure to fine CAPs from Tuxedo, NY (March,
April or May through September 2003) (Veronesi et al., 2005, 087481). The average PM2.5
exposure concentration was 110 |Jg/m3. CAPs exposure resulted in a statistically significant
decrease in dopaminergic neurons, measured by tyrosine hydroxylase immunoreactivity, in
the substantia nigra of ApoE"'" mice but not in control mice. This population of neurons is
July 2009
6-253
DRAFT-DO NOT CITE OR QUOTE

-------
targeted in neurodegenerative diseases such as Parkinson's. Furthermore, a statistically
significant increase in GFAP immunoreactivity, a marker for astrocytes, was observed in
the nucleus compacta of CAPs-exposed ApoE"'" mice compared to air-exposed ApoE"'" mice.
These results suggest that the ApoE"'" mice, a genetic model involving increased oxidative
stress, are susceptible to PM-induced neurodegeneration. Evidence for brain oxidative
stress has also been found in normal animals following IT instillation of high
concentrations of PM2.5 from Taiyuan, China (Liu and Meng, 2005, 088650) and of gasoline
exhaust (Che et al., 2007, 096460) and following chronic exposure to ROFAby intranasal
instillation (Zanchi et al., 2008, 157173).
6.4.3.3. Diesel Exhaust
A recent study tested the effects of DE inhalation on spatial learning and memory
function-related gene expression in the hippocampus (Win-Shwe et al., 2008, 190516). Male
BALB/c mice were exposed to DE (148.86 [Jg/m3 particulate concentration) for 5 h/day and 5
day/wk over a 4-wk period. Particle size was 26.21 ± 1.50 nm and particle number count
was 1.92 x 106 ± 6.18 x 104. Concentrations of gases were 3.27 ppm CO, 0.01 ppm SO2,
0.53 ppm NO2, 0.98 ppm NO and 0.07 ppm CO2. Half of the animals were injected i.p. once
per week with lipoteichoic acid (LTA), a bacterial cell wall component used to induce
systemic inflammation. The ability of the mice to perform spatial learning tasks was
examined the day after the final exposure to DE and on two subsequent days. Impaired
acquisition of spatial learning was observed in DE-exposed mice on the first day and on all
three days in DE-exposed mice which had also been treated with LTA. LTA by itself had no
effect. Since the NMDA (a type of neurotransmitter) receptors in the hippocampus play an
important role in spatial learning ability, mice were sacrificed and total RNAfrom
hippocampus was extracted and analyzed for expression of NMDA receptor subunits. DE
exposure resulted in a statistically significant increase in the expression of one subunit
while the combined exposure to DE and LTA resulted in statistically significant increases in
the expression of three subunits compared with controls. The expression of pro-
inflammatory cytokines was also examined in the hippocampus. DE exposure resulted in a
statistically significant increase in TNFa mRNA while LTA exposure resulted in a
statistically significant increase IL-16 mRNA compared with controls. Neither exposure
altered the expression of HO-1. These results demonstrated that subchronic exposure to
ultrafine-rich DE resulted in impaired spatial learning and altered expression of
July 2009
6-254
DRAFT-DO NOT CITE OR QUOTE

-------
hippocampal genes involved in memory function and inflammation. These responses were
modulated by systemic inflammation.
In summary, PM may produce adverse effects in the CNS by direct or indirect
mechanisms which are at present incompletely understood. Two recent short-term fine
CAPs inhalation studies demonstrated pro-inflammatory responses in the brain and brain
region-specific modulation of neurotransmitters and suggest the involvement of
neuroimmunological pathways. One recent chronic fine CAPs inhalation study
demonstrated loss of dopaminergic neurons in the substantia nigra and suggested that
oxidative stress contributes to neurodegeneration. Veronesi et al. (2005, 087481) have noted
that the brain is very vulnerable to the oxidative stress induced by PM due to the brain's
high energy demands, low levels of endogenous free radical scavengers, and high content of
lipids and proteins. PM-mediated upregulation of inflammatory cytokines and mediators
may also contribute to neurodegeneration. In fact, a recent subchronic study involving
ultrafine CAPs demonstrated the activation of cell signaling pathways associated with
upregulation of pro-inflammatory cytokines in brain cortical regions. Furthermore, a
subchronic study involving ultrafine-rich DE demonstrated impaired spatial learning and
altered expression of pro-inflammatory and neurotransmitter receptor genes in the
hippocampus. Further investigations are required to delineate mechanisms involved in
these responses.
6.4.4. Summary and Causal Determination
Recent animal toxicological studies involving acute or chronic CAPs exposure have
demonstrated pro-inflammatory responses in the brain, brain region-specific modulation of
neurotransmitters and loss of dopaminergic neurons in the substantia nigra (Campbell et
al., 2005, 087217; Kleinman et al., 2008, 190074; Sirivelu et al., 2006, 111151; Veronesi et
al., 2005, 087481). However, the mechanisms underlying these effects need to be delineated.
A single controlled human exposure study provides some evidence of an acute cortical stress
response to DE, though these findings are nonspecific and could have been caused by DE
gases rather than diesel PM (Cruts et al., 2008, 156374). Similar consideration is
warranted for the single animal toxicological study involving DE which demonstrated
impaired spatial learning and altered expression of pro-inflammatory and neurotransmitter
genes in the hippocampus following subchronic exposure (Win-Shwe et al., 2008, 190516).
The single epidemiology study that examined CNS outcomes did not find associations
July 2009
6-255
DRAFT-DO NOT CITE OR QUOTE

-------
between long-term exposure to PMio and cognitive function in adults after adjustment for
race/ethnicity or SES (Chen and Schwartz, 2009, 179945). Though the effect of ambient air
pollution on CNS outcomes has recently begun to draw more attention, the evidence for a
PM-induced CNS effect is limited. While most available studies have evaluated the effects
of fine particle exposures, there is insufficient evidence to draw conclusions regarding
effects of specific PM size fractions. Overall, the evidence is inadequate to determine if a
causal relationship exists between short-term exposures to PM2.5, PIVWs, UFPs, or specific PM
components and CNS effects.
6.5. Mortality Associated with Short-Term Exposure
The relationship between short-term exposure to PM and mortality has been
extensively addressed in previous PM assessments (Burnett et al., 2004, 086247; U.S. EPA,
1982, 017610; U.S. EPA, 1996, 079380; 2004, 056905). Apositive association between PM
concentration and mortality was consistently demonstrated across studies cited in the 2004
PM AQCD (U.S. EPA, 2004, 056905); these results are summarized below in Section 6.5.1.
Numerous studies have been published since the previous review, including a number of
multicity analyses and many single-city studies. The current body of evidence examines the
association between short-term exposure to PM of various size fractions (i.e., PMio, PM10-2.5,
PM2.5, and UFPs [0.01-0.1 jum]) and mortality through the use of time-series and/or
case-crossover studies. Both study designs aim to disentangle the PM-mortality effect
through either complex modeling (i.e., time-series) or matching strategies (i.e., case
crossover). Overall, the results of the more recent studies build upon the conclusions from
the previous review, showing consistent positive associations between mortality and short-
term exposure to PM2.5 and PM10-2.5.
Section 6.5.2 reviews and summarizes the results of recent studies that examined
mortality associations with the four PM size classes listed above. Each section integrates
the results of recent studies with those available in previous PM reviews. This assessment
first focuses on multicity studies that examined mortality associations with PMio because
this is a large body of literature that provides important information on potential effect
modifiers, potential confounding by copollutants, evaluation of concentration-response
relationships, and the influence of different modeling approaches on the PM-mortality
relationship (Section 6.5.2.1). The PMio studies have provided the most data among the PM
July 2009
6-256
DRAFT-DO NOT CITE OR QUOTE

-------
indices thus far! therefore this evaluation begins with the consideration of those findings as
they relate to the general association between PM and mortality. It is difficult to interpret
the extent to which these studies inform an evaluation of the effects of PM2.5 or PM10-2.5,
since data are combined from multiple cities with different PM composition. Interpretations
of the PM size fraction that contributes the most to the PM10 effects observed are provided
when appropriate in the following review. The multicity studies that examine the
association between PM10 and mortality also offer new evidence on regional and seasonal
differences in effect estimates, building upon observations made in the 2004 PM AQCD
(U.S. EPA, 2004, 056905).
Recent study findings on associations with PM2.5, PM10-2.5, and UFPs are evaluated in
Sections 6.5.2.2, 6.5.2.3, and 6.5.2.4, respectively. For PM2.5, the focus of the assessment
remains on multicity study findings! however, for PM10-2.5 and UFPs, some additional
emphasis is placed on single-city studies, due to the relative sparseness of the evidence base
for these size fractions. Some studies have also evaluated relationships between mortality
and specific components and sources of PM, and the results are summarized in Sections
6.5.2.4 and 6.5.2.5. Finally, Section 6.5.2.6 assesses evidence on the concentration-response
relationship between short-term PM exposure and mortality.
6.5.1. Summary of Findings from 2004 PM AQCD
The 2004 PM AQCD (U.S. EPA, 2004, 056905) found strong evidence that PMio and
PM2.5, or one or more PM2.5 components, acting alone and/or in combination with gaseous
copollutants, are associated with total (non-accidental) mortality and various cause-specific
mortality outcomes. For PM10, several multicity studies in the U.S., Canada, and Europe
provided strong support for this conclusion, reporting associations with total mortality
highlighted by effect estimates ranging from ~0.2 to 0.7% (per 10 (Jg/m3 increase in PM10)
(U.S. EPA, 2004, 056905). Numerous studies also reported PM10 associations with
cause-specific mortality, specifically cardiovascular- and respiratory-related mortality. For
PM2.5, the strength of the evidence varied across endpoints, with relatively stronger
evidence for associations with cardiovascular compared to respiratory endpoints. The
resulting effect estimates reported from the U.S.- and Canadian-based studies (both multi-
and single-city) analyzed for these two endpoints ranged from 1.2 to 2.7% for
cardiovascular-related mortality and 0.8 to 2.7% for respiratory-related mortality, per
10 |jg/m3 increase in PM2.5 (U.S. EPA, 2004, 056905). In regards to thoracic coarse particles
July 2009
6-257
DRAFT-DO NOT CITE OR QUOTE

-------
(PM10-2.5), the PM AQCD found a limited body of evidence that was suggestive of
associations between short-term exposure to ambient PM10-2.5 and various mortality
outcomes (e.g., 0.08-2.4% increase in total [non-accidental] mortality per 10 (Jg/m3 increase
in PM10-2.5). The positive effect estimates obtained from studies that analyzed the
association between PM10-2.5 and mortality resulted in the conclusion that PM10-2.5, or some
constituent component(s) (including those on the surface) of PM10-2.5, may contribute, in
certain circumstances, to increased human health risks.
Some additional studies examined the association between specific PM2.5 chemical
components and mortality. These studies observed associations for SO42-, nitrate, and CoH,
but not crustal particles. The strength of the association for each component varied from
city to city (U.S. EPA, 2004, 056905). Source-oriented analyses were also conducted to
identify specific source-types associated with mortality. These studies implicate fine
particles from anthropogenic origin, such as motor vehicle emissions, coal combustion, oil
burning, and vegetative burning, as being important in contributing to increased mortality
(U.S. EPA, 2004, 056905).
6.5.2. Associations of Mortality and Short-Term Exposure to
PM
The recent literature examines the association between short-term exposure to
various PM size fractions (i.e., PM10, PM10-2.5, PM2.5, UFPs [UFP], or species [e.g., OC, EC,
transition metals, etc.]) and mortality. This ISA, similar to previous AQCDs, focuses more
heavily on multicity studies, and especially those conducted in the U.S. and Canada (see
Table 6-12). By using this approach it is possible to: (1) obtain a more representative
sample of or insight into the PM-mortality relationship observed across the U.S.!
(2) analyze the association between mortality and short-term exposure to PM at or near
ambient conditions observed in the U.S.; (3) examine the potential heterogeneity in effect
estimates between cities and regions! and (4) analyze the confounders and/or effect
modifiers that may explain the PM-mortality relationship in the U.S. Although this
section focuses on mortality outcomes in response to short-term exposure to PM, it does not
evaluate studies that examine the association between PM and infant mortality. These
studies are evaluated in Section 7.5, Reproductive, Developmental, Prenatal and Neonatal
Outcomes, although it is possible that short- and long-term in utero exposures may
contribute to infant mortality. In addition, the exposure windows of interest for this unique
July 2009
6-258
DRAFT-DO NOT CITE OR QUOTE

-------
health outcome can be difficult to characterize and may span both short- and long-term
periods.
Table 6-12. Overview of U.S. and Canadian multicity PM studies of mortality analyzed in the 2004 PM
AQCD and the PM ISA"
Reference
Location
Mean Concentration
M/m3)
98th; 99th Percentiles
(/uglm31
Upper Percentile:
Concentrations (/uglm31
PMn
Dominici et al. (2003,156407)°
90 U.S. cities
15.3-53.2

NR
Burnett and Goldberp (2003, 042798)"
8 Canadian cities
25.9

95th: 54
Maximum: 121
Penqet al. (2005, 087463)
100 U.S. cities
13-49

50th: 27.1; 75th: 32.0
Maximum: 48.7
Dominici et al. (2006, 088398)'
100 U.S. cities
13-49

50th: 27.1; 75th: 32.0
Maximum: 48.7
Weltv and Zeqer (2005, 087484)'
100 U.S. cities
13-49

50th: 27.1
75th: 32.0
Maximum: 48.7
Bell et al. (2009,191997)
84 U.S. urban communities
NR

NR
Burnett et al. (2004, 086247)
12 Canadian cities
NR

NR
Samoli et al. (2008,188455)
12 Canadian cities
90 U.S. cities8
22 European cities
NR

NR
Schwartz (2004, 078998)
14 U.S. cities
23-36d

75th: 31-57
Schwartz (2004, 053506)
14 U.S. cities
23-36d

75th: 31-57
Zeka et al. (2005, 088068)
20 U.S. cities
15-37.5

NR
Zeka et al. (2006, 088749)
20 U.S. cities
15.9-37.5

NR
PMzs
Burnett and Goldberp (2003, 042798)8
8 Canadian cities
13.3

95th: 32
Maximum: 86
Dominici et al. (2007, 099135)
100 U.S. cities
NR

NR
Zanobetti and Schwartz (2009,188462)
112 U.S. cities
13.2
34.3; 38.6
Maximum: 57.4
Franklin et al. (2007, 091257)
27 U.S. cities
15.6
45.8; 54.7
Maximum: 239
Franklin et al. (2008, 097426)°
25 U.S. cities
Winter: 9.6-34.4
Spring: 6.7-27.6
Summer: 7.6-26.0
Fall: 9.5-32.1

NR
Ostro et al. (2006, 087991)
9 California counties
19.9
68.2:82.0
95th: 61.3
Maximum: 160.0
Ostro et al. (2007, 091354)
6 California counties
18.4
61.2:70.1
Maximum: 116.1
Burnett et al. (2004, 086247)
12 Canadian cities
12.8

NR
July 2009
6-259
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Location
Mean Concentration
l/ug|m3)
98th; 99th Percentiles
Ouglm31
Upper Percentile:
Concentrations (/uglm31
PM10-2.5
Burnett and Goldberp (2003, 042798)8
8 Canadian cities
12.6

95th: 30
Maximum: 99
Zanobetti and Schwartz (2009,188462)
47 U.S. cities
11.8
40.2; 47.2
Maximum: 88.3
Burnett et al. (2004, 086247)
12 Canadian cities
11.4

NR
Villeneuve et al. (2003, 055051)
Vancouver, Canada
6.1

90th: 13.0
Maximum: 72.0
Klemm et al. (2004, 056585)
Atlanta, Georgia
9.7

50th: 9.34:75th: 11.94
Maximum: 25.17
Slaughter et al. (2005, 073854)
Spokane, Washington
NR

NR
Wilson etal. (2007,157149)
Phoenix, Arizona
NR

NR
Kettunen et al. (2007, 091242)
Helsinki, Finland
Cold season: 6.7d
Warm season: 8.4d

Cold season
50th: 6.7:75th: 12.5
Maximum: 101.4
Warm season
50th: 8.4:75th: 11.8
Maximum: 42.0
Perez et al. (2008,156020)
Barcelona, Spain
Saharan Dust Days: 16.4
NonSaharan Dust Days: 14.9

Saharan Dust Days
50th: 14.8:75th: 21.8
Maximum: 36.7
Non Saharan Dust Days
50th: 12.6:75th: 18.9
Maximum: 93.1
8	Multicity studies examined in the 2004 PM AQCD (U.S. EPA, 2004, 056905)
b Because only two multicity study was identified that examined PM102.5, single-city and international studies that examined PM102.5 were analyzed in this ISA and are included in this table.
e The majority of multicity studies examined in the PM ISA provide the mean PM concentration of each individual city, not an overall PM concentration across all cities. As a result, the range of PM
concentrations for a particular study are presented, which represents the lowest and highest mean PM concentrations reported across cities, if an overall mean is not provided within the study.
d Median PM concentration.
"The study included 90 U.S. cities in the 1 -day lag analysis, but only 15 U.S. cities in the analysis of the average of lag days 0-1.
'The concentrations reported for these studies were estimated from Peng et al. (2005, 087463) because they used the same number of cities and years of data from NMMAPS.
9	This study did not present an overall mean 24-h ave PM2.5 concentration across all cities for each season. The range of mean 24-h avg concentrations reported in this table for each season
represents the lowest mean 24-h avg PM2.5 concentration and the highest 24-h avg PM2.5 concentration reported across all cities included in the study.
6.5.2.1. PM10
The majority of studies that examined the association between short-term exposure to
PM and mortality focused on effects attributed to PM10. Although these studies do not
characterize the compositional differences in PM10 across the cities examined in each of the
studies evaluated, they can provide an underlying basis for the overall pattern of
associations observed when examining the relationship between PM10-2.5 and PM2.5 and
July 2009
6-260
DRAFT-DO NOT CITE OR QUOTE

-------
mortality. The studies evaluated in this review analyzed the PMicrmortality relationship
through either a time-series or case-crossover design.1
Time-Series Analyses
Mortality associations with short-term exposure to PMio in the U.S. have been
examined in several updated time-series analyses of the National Morbidity and Mortality
Air Pollution Study (NMMAPS). In the previous NMMAPS analysis (Dominici et al., 2003,
156407; Samet et al., 2000, 005809; Samet et al., 2000, 010269) of the 1987-1994 data,
which was reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905), the strongest
association was found for non-accidental mortality for 1-day lag, with a combined estimate
across 90 cities of 0.21% (95% posterior interval [PI]: 0.09, 0.33) per 10 |ug/m3 increase in
PMio. The association was found to be robust to the inclusion of other gaseous copollutants
in the regression models, but the investigators found heterogeneity across regions, with the
strongest associations in northeastern cities. In the new updated analyses, the
investigators examined additional issues surrounding the association between PM and
mortality including: seasonal effect modification; change in risk estimates over time!
sensitivity of results to alternative weather models! and effect modification by air
conditioning use. The NMMAPS data has also been used to examine the PM concentration-
response relationship using PMio data from 20 cities (see Section 6.5.2.7). A few multicity
studies conducted in Canada and Europe provide additional information, which further
clarifies and supports the association between PM and mortality presented in the
NMMAPS analyses.
1 Schwartz (2004, 188945) used a case-crossover study design, but also conducted a time-series analysis to validate the results
obtained using the case-crossover approach.
July 2009
6-261
DRAFT-DO NOT CITE OR QUOTE

-------
1.5-
Q.
c
®
1 1-0'
8
c
,E
cb
a.
o
0.5 •
00
-0.5 •
>,	1 -5 ¦
*-»
I	10
1	05
1 0 0 •
-0.5 ¦
I
Industrial Midwest
Northeast
I Northwest
[Soiitierri California 1
—-
' I r 1
E !
^J | j

jthwcst
Upper Midwosr
All Regions 1



! i
\ j
1 I
j I


J
K# * ' " '1
[ !
I . . . !
10) too 300	100 200 300	100 200 300
Day in year
100 200 300
Source: Peng et al. (2005,087463)
Figure 6-16.
National and regional estimates of smooth seasonal effects for PM10 at a 1-day lag and
their sensitivity to the degrees of freedom assigned to the smooth function of time in the
updated NMMAPS data 1987-2000. Note: The degrees of freedom chosen were 3 df
(short-dashed line), 5 df (dotted line), 7 df (solid line), 9 df (dotted-and-dashed line), and
11 df (long-dashed line) per year of data.
Seasonal Analyses of PMio-Mortality Associations
Using the updated NMMAPS data, which consisted of 100 U.S. cities for the period
1987-2000, Peng et al. (2005, 087463) examined the effect of season on PMicrmortality
associations. In their first stage regression model, for each city, the PMio effect was modeled
to have a sinusoidal shape that completes a cycle in a year, but was constrained to be
periodic across years using sine/cosine terms. The authors also considered a model that
consisted of PMicrseason interactions using season indicators. Both of these models also
included covariates that were used in their earlier NMMAPS analyses. In the second stage
model, the seasonal patterns of PMio mortality coefficients were estimated for seven
geographic regions and on average for the entire U.S. Peng et al. (2005, 087463) found for
1-day lag, at the national level, season specific increases in non-accidental mortality per
10 jug/m3 increase in PMio of 0.15% (PL -0.08 to 0.39), 0.14% (PL -0.14 to 0.42), 0.36% (PL
0.11-0.61), and 0.14% (PL -0.06 to 0.34) for winter, spring, summer, and fall, respectively.
July 2009
6-262
DRAFT-DO NOT CITE OR QUOTE

-------
The corresponding all-season estimate was 0.19% (PL 0.10-0.28). After the inclusion of SO2,
O3, or NO2 in the model with PM10 in a subset of cities (i.e., 45 cities) for which data existed
for all pollutants resulted in fairly robust PM10 risk estimates. An analysis by geographic
region found a strong seasonal pattern in the Northeast. Figure 6-16 presents the estimated
seasonal pattern of PM10 risk estimates by region from Peng et al. (2005, 087463), which
includes a sensitivity analysis aimed to determine the appropriate number of degrees of
freedom for temporal adjustment. It is clear from Figure 6-16 that the Northeast has the
strongest association, which peaks in the summer and is robust to the extent of temporal
adjustment. The industrial Midwest also shows the summer peak, but with smaller risk
estimates. Other regions have either no seasonal pattern (Southeast) or a suggestion of a
spring peak that appears to be sensitive to the extent of temporal adjustment. On a
nationwide basis, the PM10 risk estimates appear to peak between spring and summer.
Overall, this study identified an effect modifier that may be useful in identifying the
specific chemical component(s) of PM that are related to specific regions and times of the
year.
Change in PMio-Mortality Associations over Time
Dominici et al. (2007, 099135) conducted an analysis of the extended NMMAPS data
set (i.e., 1987-2000) to examine if short-term PMicrmortality risk estimates changed during
the course of the study period. The investigators estimated the average PM10 mortality risk
coefficient for 1-day lag, using essentially the same model specification as in their 2003
analysis, separately for three time periods: 1987-1994, 1995-2000, and 1987-2000, for the
"eastern U.S." (62 counties), the "western U.S." (38 counties), and all 100 U.S. counties. To
produce national and regional estimates, two-stage hierarchical models were used as in the
previous NMMAPS studies. As shown in Table 6-13, the authors found a continuation of the
PMicrmortality association in the nationwide data for the entire study period. A comparison
of the relative risk estimates for 1987-1994 vs. 1995-2000 suggests weak evidence (not a
significant difference) that short-term effects declined. Most of the decline in the national
estimate appears to be attributable to the eastern U.S. counties. However, the decline in the
risk estimate for all-cause mortality in the eastern U.S. appears to be disproportionately
influenced by the reduction in the risk estimate for the "other" mortality category
(i.e., all-cause minus cardiorespiratory category, which may be 40 to 50% of all-cause
deaths in U.S. cities). Likewise, the apparent increase in the risk estimate for all-cause
mortality in the western U.S. appears to be affected by the increase in the risk estimate for
July 2009
6-263
DRAFT-DO NOT CITE OR QUOTE

-------
the "other" mortality category. Because the study does not clearly identify the specific
cause(s) in the "other" mortality category that are affected by PM, interpreting the
reduction in risk estimates for all-cause mortality requires caution. In contrast, the
apparent reductions (-23%) in PMio risk estimates for cardiorespiratory deaths were more
comparable between the two regions.
In addition, the investigators estimated time-varying PMio mortality risk as a linear
function of calendar time for the period 1987-2000, producing the percentage rate change in
the PMio risk estimate with a change in time of 1 year. The estimated rate of decline in
slope for all-cause mortality and the combination of cardiovascular and respiratory
mortality were -0.012 (PI: -0.037, 0.014) and -0.016 (PI: -0.058, 0.027), respectively. The
authors also estimated a PM2.5 mortality risk for the period 1999-2000 (discussed in
Section 6.5.2.2.).
Table 6-13. NMMAPS national and regional percentage increase in all-cause, cardio-respiratory, and
other-cause mortality associated with a 10yL/g/m3 increase in PM10 at lag 1 day for the
periods 1987-1994,1995-2000, and 1987-2000.

1987-1994
95% PI
1996-2000
95% PI
1987-2000
95% PI
ALL CAUSE
East
0.29
0.12, 0.46
0.13
¦0.19, 0.44
0.25
0.11,0.39
West
0.12
¦0.07,0.30
0.18
¦0.07, 0.44
0.12
¦0.02, 026
National
0.21
0.10,0.32
0.18
0.00, 0.35
0.19
0.10, 0.28
CARDIORESPIRATORY
East
0.39
0.16, 0.63
0.30
¦0.13, 0.73
0.34
0.15, 0.54
West
0.17
¦0.07,0.40
0.13
¦0.23, 0.50
0.14
¦0.05, 0.33
National
0.28
0.14,0.43
0.21
¦0.03, 0.44
0.24
0.13, 0.36
OTHER
East
0.21
¦0.03, 0.44
0.00
¦0.49, 0.50
0.15
¦0.09, 0.39
West
0.09
¦0.21,0.38
0.23
¦0.15, 0.62
0.11
¦0.10, 0.33
National
0.15
¦0.02, 0.32
0.17
¦0.07, 0.41
0.15
0.00, 0.29
Source: Dominici et al. (2007, 099135)
The objective of the Dominici et al. (2007, 099135) study described above was
motivated by accountability research, the idea of measuring the impact of policy
interventions. However, unlike the intervention studies conducted in Hong Kong (Hedley et
al., 2002, 040284) and Dublin, Ireland (Clancy et al., 2002, 035270) that were reviewed in
the 2004 PM AQCD (U.S. EPA, 2004, 056905), this study was not designed to estimate a
reduction in mortality in response to a sudden change in air pollution. In fact, the figure of
July 2009
6-264
DRAFT-DO NOT CITE OR QUOTE

-------
observed trend in PMio levels presented in the Dominici et al. (2007, 099135) study
indicates that the decline in PMio levels during the study period was very gradual, with
much of the decline appearing in the first few years (median values of ~33 ug/m3 in 1987 to
~25 jug/m3 in 1992, then down to ~23 jug/m3 in 2000). A flaw in the use of the time-series
study design for this type of analysis is that it adjusts for long-term trends, and, therefore,
does not estimate the change in mortality in response to the gradual change in PMio. The
apparent change, though weak, in the PMio risk estimates may also reflect a potential
change in the composition of PMio (i.e., PM10-2.5 or PM2.5). The study listed a number of
PMio-related air pollution control programs that were implemented between 1987 and
2000. Some of these programs, such as the Acid Rain Control Program, did result in major
reductions in emissions, and, therefore, could have contributed to the results observed, but
the analytic approached used in the study does not allow for a systematic analysis of the
effect of air pollution policies on the risk of mortality.
Sensitivity of PM-Mortality Associations to Alternative Weather Models
To examine the sensitivity of PMio mortality risk estimates to alternative weather
models that consider longer lags, Welty and Zeger (2005, 087484) analyzed the updated
NMMAPS 100 U.S. cities data. All of the previous NMMAPS analyses only considered
temperature and dew point up to 3-day lags. In this analysis, the authors considered
various forms of a constrained distributed lag model: (l) containing a step function of
temperature with steps at lag 0, 2, 7 and extended to 14 days! (2) similar to (l) but with
time-varying coefficients to change over season and study period! and, (3) containing a
smooth function to account for non-linearity in the temperature-mortality relationship.
With the combination of degrees of freedom for temporal trends and the number of
distributed lags, more than 20 models were applied to each of the three lag days (0, 1, and
2) of PMio. These city-specific risk estimates were then combined across the 100 cities in the
second stage Bayesian model. The combined PMio risk estimates were generally consistent
within the lag. In particular, the risk estimates for non-accidental mortality for lag 1-day
ranged between 0.15% and 0.25% per 10 |ug/m3 increase in PMio, and were always
statistically significant regardless of the model used. In addition, the range of these point
estimates across the models was found to be much narrower than the regression posterior
intervals. Thus, the PMio risk estimates at lag 1 day were robust to alternative
temperature models that considered temperature effects lasting up to a 2-wk period.
July 2009
6-265
DRAFT-DO NOT CITE OR QUOTE

-------
In summary, the above three analyses of the updated NMMAPS data provided useful
information on PM mortality risks, resulting in the following conclusions^ (l) estimated
PMio risk is particularly high in the northeast and in the summer! (2) there remains an
overall PMicrmortality association in the 1987-2000 time period as well as the 1995-2000
time period! (3) there is a weak indication that PMio mortality risk estimates are declining!
and (4) PMio risk estimates were not sensitive to alternative temperature models.
Effect Modification of PMio-Mortality Associations by Air Conditioning Use
It has been hypothesized that air conditioning use reduces an individual's exposure to
PM and subsequently modifies the PM-mortality association. Bell et al. (2009, 191997)
investigated the role of air conditioning (AC) use on the relationship between PMio and all-
cause mortality using the NMMAPS PMio risk estimates from 84 U.S. urban communities
from 1987-2000.1 Bayesian hierarchical modeling was used to examine if AC prevalence
(i.e., fraction of households with central or any AC) explained city-to-city variation in PMio
risk estimates. The authors calculated yearly, summer-only, and winter-only effect
estimates stratified by housing stock that had either central AC or any AC, which includes
window units. Risk estimates for lag 1 (previous day) were used in the analysis because this
lag has been found to show the strongest association with mortality in the original
NMMAPS analyses. Community-specific AC prevalence was calculated from national
survey U.S. Census American Housing Survey (AHS) data, which is available every two
years. The investigators computed percent change in PMio effect estimates per an
additional 20% of the population acquiring AC.
The AC variables were not strongly correlated with socio-economic variables (poverty
rate, unemployment, and education) from the U.S. Census (correlation ranged from -0.27 to
0.29). Bell et al. (2009, 191997) found that communities with higher AC prevalence had
lower PMio mortality risk estimates for all-cause mortality (-30.4% [95% P.I.: -80.4 to 19.6]
per an additional 20% of the population acquiring any AC; -39.0% [95% P.I.: -81.4 to 3.3] for
central AC), but results were not statistically significant. When restricting the analysis to
the summer months and focusing on the 45 cities with summer-peaking PMio
concentrations, the authors reported positive (i.e., lower PMio risks in cities with lower AC
use) non-significant risk estimates. A similar analysis was conducted for winter months
1 This study also examined risk estimates for cardiovascular and respiratory hospital admissions in older adults (> 65).
July 2009	6-266	DRAFT - DO NOT CITE OR QUOTE

-------
using data from 6 cities with winter peaking PMio concentrations, but the confidence bands
were too wide (due to small sample size) for meaningful interpretation.
Although the estimated reductions in PMio all-cause mortality risks from AC use
reported in the Bell et al. (2009, 191997) study were not statistically significant, their large
magnitude suggests that AC use may reduce an individual's exposure to PM. Given the
expected additional increase in AC use in the future, and the results from recent multicity
studies, which have reported stronger PM-mortality associations during the warm season,
AC use may play a larger role in determining an individuals exposure to PM. Studies that
have examined the effect of AC use on the PM2.r,-mortal ity association have reported similar
results. For example, Franklin et al. (2007, 091257) (discussed in detail in Section 6.1.2.2)
found that AC use non-significantly modified PM2.5 mortality risk estimates, but the result
was suggestive of higher PM2.5 effects in cities with lower AC use, especially in cities with
summer-peaking PM2.5 concentrations. Overall, further investigation is needed to fully
understand the relationship between AC use and mortality attributed to short-term
exposure to PM.
PMio-Mortality Associations in Canada and Europe
Burnett et al. (2004, 086247) examined the association between mortality and various
air pollutants in 12 Canadian cities, and reported that the most consistent association was
found for NO2. For this analysis, PM was measured every 6th -day for the majority of the
study period, and the PMio concentrations used in the study represent the sum of the PM2.5
and PM10-2.5, which were directly measured by dichotomous samplers. The authors found
that the simultaneous inclusion of NO2 and PMio in a model, on those days with PM data,
greatly reduced the PMio association with non-accidental mortality, from 0.47% (95% CP
0.04-0.89) to 0.07% (95% CP -0.44 to 0.58) per 10 |ug/m3 increase. The previous Canadian
multicity analysis (Burnett and Goldberg, 2003, 042798), a re-analysis of Burnett et al.
(2000, 010273) reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905). did not consider
gaseous pollutants. Thus, PMio risk estimates in the Canadian data appear to be more
sensitive to NO2 than those estimates reported in U.S. studies.
The association between PMio and mortality in Europe was also reviewed in the 2004
PM AQCD (U.S. EPA, 2004, 056905) through Katsouyanni et al. (2003, 042807). which
presented results from the Air Pollution and Health: a European Approach (APHEA2)
study, a multicity study that examined PMio effects on total mortality in 29 European cities.
Analitis et al. (2006, 088177) published a brief report on effect estimates for cardiovascular
July 2009
6-267
DRAFT-DO NOT CITE OR QUOTE

-------
and respiratory deaths also based on the 29 European cities, within the APHEA2 study.
They reported for the average of 0- and 1-day lags, PMio risk estimates per 10 |u,g/m3 of
0.76% (95% CL 0.47-1.05) for cardiovascular deaths and 0.71% (95% CI: 0.22-1.20) for
respiratory deaths in random effects models.
Comparison of PM-Mortality Associations in Europe, Canada, and the U.S.
The APHENA study (Samoli et al., 2008, 188455) was a collaborative effort by the
APHEA, NMMAPS, and the Canadian multicity study investigators to evaluate the
coherence of PMio mortality risk estimates across locations and possible effect modifiers of
the PM-mortality relationship using a common protocol. In this study, to adjust for
temporal trends Samoli et al. (2008, 188455) used 3, 8, and 12 df with natural splines and
penalized splines, as well as the minimization of the sum of the absolute values of the
partial auto-correlation function (PACF). The investigators also included a smooth function
of temperature on the same day of death and the day before death. The study reported risk
estimates for a 1-day lag (from all three data sets), the average of lag day 0 and 1 (all but
for the Canadian data because PM data was collected every 6th day), and an unconstrained
distributed lag model using lags of 0, 1, and 2 days (all but for the Canadian data). The
second-stage regression included: (a) the average pollution level and mix in each city! (b) air
pollution exposure characterization (e.g., number of monitors, density of monitors); (c) the
health status of the population (e.g., cardio-respiratory deaths as a percentage of total
mortality, crude mortality rate, etc.); and (d) climatic conditions (e.g., mean and variance of
temperature). In addition, unemployment rate was examined for 14 European cities and all
U.S. cities. Effect modification patterns were examined only for cities with complete time-
series data and using the average of lags 0 and 1 day, resulting in the exclusion of the
Canadian data.
Generally, the risk estimates from Europe and the U.S. were similar, but those from
Canada were substantially higher.1 For example, the percent excess risks per 10 (Jg/m3
increase in PMio for all ages using 8 df/yr and penalized splines were 0.84% (0.30, 1.40),
0.33% (0.22, 0.44), and 0.29% (0.18, 0.40) for the Canadian, European, and U.S. data,
respectively. Note that the risk estimate for the 90 U.S. cities is slightly larger than that
reported in the original NMMAPS study (0.21%, using natural splines, and more
1 The risk estimate reported for the 12 Canadian cities examined in the APHENA study is higher than that reported by
Burnett et al. (2004, 086247). This is because the APHENA study did not use the 12 cities data from Burnett et al. (2004,
086247). but instead used a composite of the data from three previous studies conducted by the same group (Burnett et al.,
2000, 010273; Burnett et al., 1998, 029505; Burnett and Goldberg, 2003, 042798)
July 2009
6-268
DRAFT-DO NOT CITE OR QUOTE

-------
temperature variables). In the all ages model, the average of lag days 0 and 1, and the
distributed lag model with lags 0, 1, and 2 did not result in larger risk estimates compared
to those for a 1 day lag. In copollutant models, PMio risk estimates did not change when
controlling for O3. Figure 6" 17 shows the risk estimates from the three data sets for
alternative extent of temporal smoothing and smoothing methods. The Canadian data
appear less sensitive to the extent of temporal smoothing or smoothing methods (see Panel
A of Figure 6-17). When stratifying by age the risk estimates for the older age group (age >
75) were consistently larger than those for the younger age group (age <75) (e.g., 0.47% vs.
0.12% for the U.S. data) for all the three data sets. Although the study did not
quantitatively present the results from the effect modification analyses, some evidence of
effect modification across the study regions was observed. The investigators reported that,
in the European data, higher levels of NO2 and a larger NO2/PM10 ratio were associated
with greater PM10 risk estimates, and that while this pattern was also present in the U.S.
data, it was less pronounced. Additionally, in the U.S. data, smaller PM10 risk estimates
were observed among older adults in cities with higher O3 levels. Effect modification by
temperature was also observed, but only in the European data.
a
£ 1.5
0
e
"i 10
e
1	0-5
a.
0.0
-C.5
ir ir •
1 1 :
3 8 12
Canada (n= 12)
¦ PS
A NS
K
I I I
3 8 12
Europe in =22}
df/year
i i i
3 8 12
USA (n = 90}
a
£ 1-5
u
""I 1,0
CD
£ 0.5
Q.
0.0
-0.5

~1 I I I
3 8 12PACF

H I I I
3 8 12PA0F
Europe |n = 22) USA (n = 15)
If/fear
n I I i
3 8 12 PACF
Europe-USA
Source: Samoli et al. (2008,188455)
Figure 6-17. Percent increase in the daily number of deaths, for all ages, associated with a 10-//g/m3
increase in PM10: lag 1 (A) and lags 0 and 1 (B) for all three centers. PACF indicates df
based on minimization of PACF.
In this study, the underlying basis for the larger PM10 risk estimates (by 2-fold) in the
Canadian data compared to the European and U.S. data could not be identified, even when
July 2009
6-269
DRAFT-DO NOT CITE OR QUOTE

-------
consistent statistical methods were applied across each of the data sets. Because the effect
modification of PMio risk estimates were not examined in the Canadian data, the potential
influence of air pollution type or mixture could not be ruled out as a potential source of
heterogeneity across the three data sets. It should be noted that both the original U.S. and
European studies reported regional heterogeneity in PM risk estimates while the U.S. data
also reported seasonal heterogeneity. In both of these cases the specific characteristics
associated with the regions that contributed to the heterogeneity observed were not
identified. Thus, further investigation is needed to identify factors that influence the
heterogeneity in PM risk estimates observed between different countries and across
regions.
Case-Crossover Analyses
Since the 2004 PM AQCD (U.S. EPA, 2004, 056905) investigators have used the
case-crossover study design more frequently as an alternative to time-series analyses to
examine the association between short-term exposure to PM and mortality. This study
design allows for the control of seasonal variation, time trends, and slow time varying
confounders without the use of complex models. However, similar to any study design,
biases can be introduced into the study depending on the control (i.e., referent) period
selected (Janes et al., 2005, 087535) . The multicity case-crossover analyses discussed below
match cases (i.e., days in which a death occurred) to controls (i.e., days in which a death did
not occur), to control for (l) seasonal patterns and gaseous pollutants, or (2) temperature.
In addition, the studies attempted to examine the heterogeneity of effect estimates through
the analysis of individual-level and city-specific effect modification.
Controlling for Temperature
Schwartz (2004, 078998) investigated the PMicrmortality association in 14 U.S. cities
for the years 1986-1993 (some cities started in later years because of PMio data availability)
using a case-crossover study design. Note that in this analysis, four more cities (Boulder,
CO; Cincinnati, OH; Columbus, OH; and Provo-Orem, UT) were added to the cities
Schwartz (2003, 042800) previously analyzed using a time-series study design. These cities
were chosen for this analysis because they collected daily PMio data, unlike most U.S.
cities, which only monitor PMio every six days. Lag 1-day PMio risk estimates were
computed using several methods. Models (l) (i.e., the main model) and (2) were constructed
from a case-crossover analysis with bidirectional control days (7-15 days before and after
July 2009
6-270
DRAFT-DO NOT CITE OR QUOTE

-------
the case). Model (l) obtained city-specific estimates in the first stage analysis, followed by a
second stage random-effects model to obtain a combined estimate. Model (2) is the same as
model (l), but consisted of a single stage model, which included data from all 14 cities.
Models (3) and (4) were also constructed from a case-crossover analysis, but used
time-stratified control days (i.e., matched on season and temperature within the same
degree in Celsius). Model (3) obtained single-city estimates in the first stage analysis,
followed by a second stage random-effects model to obtain combined estimates. Model (4)
used the same approach as model (3), but consisted of a single stage model including data
from all 14 cities. The final model, (5), consisted of a two-stage Poisson time-series model,
which produced city-specific estimates in the first stage, and combined estimates across
cities in the second stage. In the main model, (l) above, the estimated excess risk for
non-accidental mortality was 0.36% (CI: 0.22, 0.50) per 10 |ug/m3 increase in PMio. The
other models yielded a similar magnitude of effect estimates, ranging from 0.32% (model 2)
to 0.53% (model 4). Thus, the methods used to select control days and adjust for weather in
the case-crossover design did not result in major differences in effect estimates, and in
addition, were comparable to the estimates obtained from the time-series analysis, 0.40%
(model 5).
Controlling for Gaseous Pollutants
In a subsequent analysis, Schwartz (2004, 053506) analyzed the same 14 cities data
described above, using a case-crossover design, to investigate the potential confounding
effects of gaseous pollutants. For each case day, control days were selected from all other
days of the same month of the same year. In addition, case days were matched to control
days that had gaseous pollutant levels that were within a defined concentration: 1 ppb, 1
ppb, 2 ppb, and 0.03 ppm for SO2, NO2, 1-h max O3, and CO, respectively. Unlike the study
described above (Schwartz, 2004, 078998) in this analysis, the excess risk was estimated for
the average of 0- and 1-day lag PM10 (rather than 1-day lag). In addition, apparent
temperature (a composite index of temperature and humidity) was used rather than
temperature and humidity individually. The case-crossover analysis was conducted in each
city, and a combined estimate was computed in a second-stage random effects model. The
number of cities analyzed varied across pollutants depending on the availability of
monitors. The study reported PM10 risk estimates for non-accidental mortality of 0.81% (CI:
0.47-1.15), 0.78% (CI: 0.42-1.15), 0.45% (CI: 0.12-0.78), and 0.53% (CI: 0.04-1.02) per
July 2009
6-271
DRAFT-DO NOT CITE OR QUOTE

-------
10 |ug/m3 increase, for the analysis matched by SO2 (10 cities), NO2 (8 cities), O3 (13 cities),
and CO (13 cities), respectively.
Schwartz (2004, 053506) only presented PM10 risk estimates matched by gaseous
pollutants, therefore, it is unclear in this analysis how matching by gaseous pollutants
affected (i.e., reduced or increased) unmatched PM10 risk estimates. The estimates reported
were computed using the average of 0- and 1-day lagged PM10 and, therefore, cannot be
directly compared to the 1-day lag PM10 risk estimates obtained in the Schwartz (2004,
078998) 14-city study described above. The estimates reported in Schwartz (2004, 053506)
are generally larger than those obtained in the Schwartz (2004, 078998) analysis, which
was expected since the Schwartz (2004, 053506) analysis used two-day avg PM10. However,
the estimates reported in Schwartz (2004, 053506) are comparable to the average of 0- and
1-day lagged PM10 risk estimate for non-accidental mortality (0.55% [CP 0.39-0.70]) per
10 |ug/m3 increase from the 10-city study (Schwartz, 2003, 042800), which was reviewed in
the 2004 PM AQCD (U.S. EPA, 2004, 056905). Overall, Schwartz (2004, 053506) provided
an alternative method to assess the influence of gaseous copollutants. The results suggest
that PM10 is significantly associated with all-cause mortality after controlling for each of
the gaseous copollutants.
City-Level Effect Modification
Zeka et al. (2006, 088749) expanded the 14 cities analyses conducted by Schwartz
(2004, 078998; 2004, 053506) to 20 cities, added more years of data (1989-2000), and
investigated PM10 effects on total and cause-specific mortality using a case-crossover
design. Individual 0-, 1-, and 2-day lags as well as an unconstrained distributed lag model
with 0, 1, and 2 lag days were examined. For each case day, control days were defined as
every third day in the same month of the same year, to eliminate serial correlation. The
authors also investigated potential effect modifiers in the second stage regression using
city-specific variables including percent using air conditioning, population density,
standardized mortality rates, the proportion of elderly in each city, daily minimum
apparent temperature in summer, daily maximum apparent temperature in winter, and the
estimated percentage of primary PM10 from traffic sources.
July 2009
6-272
DRAFT-DO NOT CITE OR QUOTE

-------
a£
€
0
e
,E
1
£
U
0.8
0.6
0,4
0,2
0
-0,2
~Q, <4
1000 3000
(count/mile2)
25% 75%
25% 75% 25% 75%
Population
density
Variance of
summer AT
Mean of
winter AT
% Primary
PM10 from
traffic
City effect modifiers
Source: Zeka et al. (2005, 088068)
Figure 6-18. Effect modification by city characteristics in 20 U.S. cities. Note: The two estimates
and their CI for each of the modifying factors represent the percentage increase in
mortality for a 10yL/g/m3 increase in PMio, for the 25th percentile, and 75th percentile of
the modifier distribution across the 20 cities.
The investigators found that, for all-cause (non-accidental) mortality, lag 1-day
showed the largest risk estimate (0.35% [CP 0.21-0.49] per 10 jug/m3) among the individual
lags. Respiratory mortality exhibited associations at lag 0, 1, and 2 days (0.34-0.52, and
0.51%, respectively), whereas cardiovascular mortality was most strongly associated with
PMio at lag day 2 (0.37%). The sum of the distributed lag risk estimates (e.g., 0.45% [CP
0.25-0.65] for all-cause mortality) was generally larger than those for single-day lag
estimates. The excess risk estimates for single-day lags for specific respiratory and
cardiovascular causes had generally wider confidence intervals due to their smaller daily
mortality counts, but some of the categories showed markedly larger estimates when
included in the combined distributed lag model (e.g., pneumonia 1.24% [CP 0.46-2.02]). As
shown in Figure 6-18, Zeka et al. (2005, 088068) also found evidence indicative of several
PMio effect modifiers including higher population density and the estimated percentage of
primary PMio from traffic. When 25th vs. 75th percentiles of these city-specific variables
were evaluated, the estimated percent increase in mortality attributed to PMio appears to
July 2009
6-273
DRAFT-DO NOT CITE OR QUOTE

-------
contrast substantially (e.g., 0.09 vs. 0.52% for variance of summer time apparent
temperature).
The effect modifiers investigated by Zeka et al. (2005, 088068) consisted of
city-specific variables. Some of these variables are ecological in nature, and therefore,
interpreting the meaning of "effect modification" requires some caution. As the
investigators pointed out, the population density and the estimated percentage of primary
PMio from traffic were correlated in this data set (r = 0.65)1. These variables may also be a
surrogate for another or composite aspects of "urban" characteristics. Thus, the apparent
effect modification by traffic associated PMio needs further investigation. Interestingly, the
percent of homes with central air conditioning was not a significant effect modifier of PMio
risk estimates, which questions the impact of reduced ventilation rates on PM exposure.
Overall, this study presented PMio risk estimates that are consistent with those found in
other analyses, but also provided new information on the risk estimated for broad and
specific respiratory and cardiovascular mortality, along with possible effect modifying city-
specific characteristics.
1 The correlation coefficient was calculated based on the numbers provided in Table 1 of Zeka et al. (2005, 088068).
July 2009	6-274	DRAFT - DO NOT CITE OR QUOTE

-------
Gender
Race
to
o
CM

in
o
o
o
9
9
<
XL
<8
eg
O
SM"
s
to
o
O
o
iq
9
° White '¦
»• Black!
i
X
in
w
3
U1
«
O
c>
o
"1
9
Age group
f
1%
0 <65 !
A 6b-7b ,
• >7h I
3
o
CM

Education

UJ
r,S
q
o
in
9
O [_Qyy
:fi Medium
:• High
•I
13
?
2r
s

-------
Location of death
Season
14
€S1
o
CvJ
$
c
in
d
o
o
m
9
f
f
0 In-hospital 1
[• Qut-of-hospitalj
ui
cJ
o
CsJ

£ <=*

-------
was not statistically significant. The study also examined effect modification by location of
death ("out-of-hospital" vs. "in-hospital") and season (see Figure 6-20). The "out-of-hospital"
deaths showed larger PMio risk estimates than were found for "in-hospital deaths" with a
significant contrast per 10 |ug/m3 for all-cause (0.71% vs. 0.22%) and heart disease (0.93%
vs. 0.15%) deaths. Stroke deaths also showed a significant contrast (0.87% vs. 0.06%, not
shown in Figure 6-20).
Overall, Zeka et al. (2006, 088749) showed a consistent pattern of effect modification
by contributing causes of death (i.e., pneumonia, stroke, heart failure, and diabetes) on
PMio risk estimates for primary causes of death (Figure 6-21; not all results for
contributing cause are shown). However, because the contributing causes of death counts
were relatively small, as reflected by the wide confidence intervals in Figure 6-21, most of
the contrasts observed did not achieve statistical significance.
5
i4
£
«
h
All cause
\
\
\
±
i
~ • + • + •
If U //
Respiratory	Ml
Primary cause of death
Stroke
Source: Zeka et al. (2006, 088749)
Figure 6-21. PMio risk estimates (per 10 yug/m3) by contributing causes of deaths. The estimates with
'*' (added to the original figure) indicates a significant difference.
In addition, when examining the other effect modifiers, the results that show no
gender or race differences in PMio risk estimates for all-cause and cardiovascular deaths
July 2009
6-277
DRAFT-DO NOT CITE OR QUOTE

-------
are important, given the relatively narrow confidence bands of these estimates. The effect
modification by the location of death has been reported previously in smaller studies, but
the large contrast found for all-cause and cardiovascular mortality in this large multicity
analysis is noteworthy The elevated PMio risks reported by Zeka et al. (2006, 088749) for
all-cause, heart disease (and stroke) "out-of-hospital" deaths are also consistent with the
hypothesis of acute PMio effects on "sudden deaths" brought on by systemic inflammation
or dysregulation of the autonomic nervous system. The finding regarding the seasonal effect
modification, though significant only for respiratory deaths, is somewhat in contrast with
the Peng et al. (2005, 087463) analysis of the extended NMMAPS data, which observed the
greatest effects during the summer season. The apparent inconsistency may be due to the
difference in geographic coverage (i.e., 20 vs. 100 cities) or methodology (i.e., case-crossover
with referent days in the same month of the same year vs. time-series analysis with
adjustment for temporal trend in the regression model).
Summary of PMio Risk Estimates
Overall, the recent studies continue to show an association between short-term
exposure to PM and mortality. Although these studies do not examine mortality effects
attributed to PM size fractions that compose PMio, the regional, seasonal, and effect
modification analyses conducted contribute to the evidence for the PM2.5 and PM10-2.5
associations presented in Sections 6.5.2.2 and 6.5.2.3, respectively. Of the PMio studies
evaluated, depending on the lag/averaging time and the number of cities included, the
estimates for all-cause (non-accidental) mortality for all ages ranged from 0.12% (Dominici
et al., 2007, 099135) to 0.84% (Samoli et al., 2008, 188455) per 10 |u,g/m3 increase in PMio,
regardless of the study design used (i.e., time-series vs. case crossover) (see Figure 6-22).
Although this range of PM mortality risk estimates is smaller than those reported for
PM10-2.5 and PMa.sthey do support the association between PM and mortality. The majority
of studies examined present estimates for either a lag of 1 day or a 2-day avg (lag 0-1), both
of which have been found to be strongly associated with the risk of death (Schwartz, 2004,
078998; 2004, 053506) The use of a distributed lag model (using lag 0, 1, and 2 days) was
found to result in slightly larger (by -30%) estimates compared to those for single-day lags
in the 20 cities study (Zeka et al., 2005, 088068), but when using the 15 cities data from
NMMAPS analyzed in the APHENA study (Samoli et al., 2008, 188455), the 1-day lag
combined risk estimate was larger than the distributed lag (lag, 0, 1, and 2 days) estimate.
July 2009
6-278
DRAFT-DO NOT CITE OR QUOTE

-------
Overall, an examination of the PMio risk estimates stratified by cause-specific mortality
and age, for all U.S.- and Canadian-based studies, further supports the findings of the
multicity studies discussed in the 2004 PM AQCD (U.S. EPA, 2004, 056905) and this ISA,
however, it must be noted that a large degree of variability exists between cities when
examining city-specific risk estimates.
Multicity PMio Studies	% Increase
Pena et al. (2005. 087463). NMMAPS. 1
All seasons

Winter
Snrinn
Summer
Fall

Dominici et al. (2007, 097361), NMMAPS, 1

Nationwide, 1987-1994
1995-2000
1987-2000



East, 1987-1994
1995-2000
1987-2000



West, 1987-1994 —
1995-2000 —
1987-2000 H

Weltv & Zeger (2005, 087484), NMMAPS, 1
Schwartz (2004,188945a), 14 U.S. cities, 1
—i-nm—
Bidirectional, 2-stage
Bidirectional, 1 -stage
Matched bv temperature, 2-stage
Matched bv temperature, 1 -stage
Time-series

Schwartz (2004, 053506), 14 U.S. cities, 0-1

Matched bv CO (13 cities)
Matched bv 03 (13 cities)
Matched bv NO2 (8 cities)
Matched bv SO2 (10 cities)

Zeka et al. (2005, 088068), 20 U.S. cities

0	lag
1	lag
2	lag
Sum of distributed lag 0-2

Zeka et al. (2006, 088749), 20 U.S. cities, 1 -2

Winter
Summer 	
Spring/Fall



In-hospital
Out-of-hospital

Burnett et al. (2004, 086247), 12 Canadian cities, 1
Samoli et al. (2008,188455), APHENA, 1

Canada (12 cities)
U.S. (90 cities)


1 1 1 1 1 1 1 1
-0.4	0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Figure 6-22. Summary of PM10 risk estimates (per 10 yug/m3) for all-cause mortality from recent
multicity studies. The number after the study location indicates lag/average used for
PM10 (e.g., "01" indicates the average of lag 0 and 1 days). For Welty and Zeger (2005,
087484), the vertical lines represent point estimates for 23 different weather models,
and the horizontal band spans the 95% posterior intervals of these point estimates.
July 2009
6-279
DRAFT-DO NOT CITE OR QUOTE

-------
The variability in PMio mortality risk estimates reported within and between
multicity studies may be due to the difference in the cities analyzed and the potential
regional differences in PM composition. The NMMAPS studies have found that geographic
regions and seasons are the two most important factors that determine the variability in
risk estimates, with estimates being larger in the eastern U.S. and during the summer,
respectively. These findings were fairly consistent across studies, but Zeka et al. (2006,
088749) did observe the strongest association during the transition period (spring and fall);
however, this may be due to the difference in geographic coverage or the difference in the
model specification used compared to Peng et al. (2005, 087463).
Finally, examination of potential confounders showed that the size of PMio risk
estimates are fairly robust to the inclusion of gaseous copollutants in models (Peng et al.,
2005, 087463) or by matching days with similar gaseous pollutant concentrations
(Schwartz, 2004, 078998). These findings further confirmed that PMio risk estimates are
not, at least in a straightforward manner, confounded by gaseous copollutants.
6.5.2.2. PM2.5
A nationwide monitoring system for PM2.5 was not established until 1999. This in
conjunction with the unavailability of nationwide mortality data from the National Center
of Health Statistics (NCHS) starting in 20011, has contributed to the relatively small
literature base that has examined the association between short-term exposure to PM2.5
and mortality. To date, the studies that have been conducted examined national (i.e., in
multiple cities across the country) or regional (i.e., in one location of the country) PM2.5
associations with mortality.
PM2.5 - Mortality Associations on a National Scale
The NMMAPS study conducted by Dominici et al. (2007, 099135) (described in
Section 6.5.2.1.), also conducted a national analysis of PM2.5-mortality associations using
the same methodology and data for 1999-2000. The PM2.5 risk estimates at lag 1-day were
0.29% (PL 0.01-0.57) and 0.38% (PI: -0.07 to 0.82) per 10 |ug/m3 increase for all-cause and
cardio-respiratory mortality, respectively. The authors also conducted a sensitivity analysis
of the risk estimates based on the extent of adjustment for temporal trends in the model,
1 In 2008 the EPA facilitated the availability of the mortality data for EPA-funded researchers, which should eventually
increase the literature base of studies that examine the association between short-term exposure to PM2.5 and mortality.
July 2009
6-280
DRAFT-DO NOT CITE OR QUOTE

-------
changing the degrees of freedom (df) of temporal adjustment from 1 to 20/yr (the main
result used 7 df/yr). In comparison to the PMio results, the PM2.5 risk estimates appeared
more sensitive to the extent of temporal adjustment between 5 and 10 df/yr, but this may be
in part due to the much smaller sample size used for the PM2.5 analysis compared to the
PM10 analysis.
Franklin et al. (2007, 091257) analyzed 27 cities across the U.S. that had PM2.5
monitoring and daily mortality data for at least 2 yr of a 6-yr period 1997 to 2002. The
mortality data up to year 2000 were obtained from the NCHS, while the 2001-2002 data
were obtained from six states (CA, MI, MN, PA, TX, and WA), resulting in 12 out of the 27
cities having data up to 2002. The start year for each city included in the study was set at
1999, except for Milwaukee, WI (1997) and Boston, MA (1998), which is due to PM2.5 data
availability in these two cities. In the case-crossover analysis in each city, control days for
each death were chosen to be every third day within the same month and year that death
occurred in order to reduce auto-correlation. The first stage regression examined the
interaction of effects with age and gender, while the second stage random effects model
combined city-specific PM2.5 risk estimates and examined possible effect modifiers using
city-specific characteristics (e.g., prevalence of central air conditioning and geographic
region). For all of the mortality categories, the estimates for lag 1-day showed the largest
estimates. The combined estimates at lag 1 day were: 1.2% (CP 0.29-2.1), 0.94% (CP -0.14
to 2.0), 1.8% (CP 0.20-3.4), and 1.0% (CP 0.02-2.0) for all-cause, cardiovascular, respiratory,
and stroke deaths, respectively, per 10 ug/m3. When examining the city-specific risk
estimates most of the cities with negative estimates are also those with a high prevalence of
central air conditioning (Dallas, 89%; Houston, 84%; Las Vegas, 93%; Birmingham, 77%). It
is unclear why these cities exhibit negative (and significant) risk estimates rather than null
effects.
In the analysis of effect modifiers, Franklin et al. (2007, 091257) found that
individuals >75 showed significantly higher PM2.5 risk estimates. The estimated effects
were also found to vary by geographic location with larger estimates in the East than in the
West, which are consistent with the regional pattern found in the NMMAPS PM10 risk
estimates. In addition, a higher prevalence of central air conditioning was associated with
decreased PM2.5 risk estimates when comparing the lower (25th percentile) vs. the higher
(75th percentile) air conditioning use rates, especially in the cities where PM2.5
concentrations peak in the summer. Finally, the risk estimates were not found to be
July 2009
6-281
DRAFT-DO NOT CITE OR QUOTE

-------
different between communities with PM2.5 levels less than or equal to vs. higher than
15 jug/m3. The risk estimates for each effect modifier are presented in Figure 6-25. Note the
wide confidence intervals associated with each of the risk estimates, specifically for
Franklin et al. (2007, 091257) and Ostro et al. (2006, 087991), which suggests low
statistical power for testing the differences between effect modifiers.
Franklin et al. (2008, 097426) analyzed 25 cities that had PM2.5 monitoring and daily
mortality data between the years 2000 to 2005 (with the study period varying from city to
city). The choice of the 25 communities was based on the availability of PM2.5 mass
concentrations and daily mortality records for at least 4 yr, along with PM2.5 speciation data
for at least 2 yr between 2000 and 2005. Similar to Franklin et al. (2007, 091257), all-cause,
cardiovascular, respiratory, and stroke deaths were examined; however, of the 25 cities
included in the study, only 15 overlap with the 27 cities analyzed in Franklin et al. (2007,
091257). The authors obtained mortality data from the NCHS and various state health
departments (California, Massachusetts, Michigan, Minnesota, Missouri, Ohio,
Pennsylvania, Texas, and Washington). Although the main objective of the study was to
examine the role of PM2.5 chemical species in the second stage analysis, the first stage
analysis conducted a time-series regression of mortality on PM2.5. In addition, the first
stage regression performed a seasonal analysis in order to take advantage of seasonal
variation in PM2.5 chemical species across cities and to possibly explain the city-to-city
variation in PM2.5 mortality risk estimates. From this analysis a strong seasonal pattern
was observed with the greatest effects occurring in the spring and summer seasons (see
Figure 6-25).
Overall, the risk estimates for all-cause, cardiovascular, and respiratory deaths
reported by Franklin et al. (Franklin et al., 2008, 097426) are comparable to those
presented in the 27 cities study (2007, 091257), as shown in Figure 6-26. When comparing
the 2007 and 2008 studies conducted by Franklin et al. (2007, 091257; 2008, 097426),
although only 15 cities overlap between the two studies and each study was designed
differently (i.e., time-series vs. case-crossover), the magnitude of the PM2.5 risk estimates
reported were similar for the same averaging time, and both studies reported a regional
pattern (East > West) similar to that found in the NMMAPS studies previously discussed.
Zanobetti and Schwartz (2009, 188462) conducted a multicity time-series study to
examine associations between PM2.5 and mortality in 112 U.S. cities. The cities included in
this analysis encompass the majority of cities included in the Franklin et al. (2007, 091257)
July 2009
6-282
DRAFT-DO NOT CITE OR QUOTE

-------
and (2008, 097426) analyses. In this analysis a city represents a single county! however, 14
of the cities represent a composite of multiple counties. In addition to examining PM2.5, the
investigators also analyzed PM10-2.5; these results are discussed in Section 6.5.2.3. Zanobetti
and Schwartz (2009, 188462) analyzed PM2.5 associations with all-cause, cardiovascular
disease (CVD), MI, stroke, and respiratory mortality for the years 1999-2005. To be
included in the analysis, each of the cities selected had to have at least 265 days of PM2.5
data per year and at least 300 days of mortality data per year. The authors conducted a
city- and season-specific Poisson regression to estimate excess risk for PM2.5 lagged 0- and
1-days, adjusting for smooth functions (natural cubic splines) of days (1.5 df per season), the
same-day and previous day temperature (3 df each), and day-of-week. The city specific
estimates were then combined using a random effects model. Based on the assumption that
climate affects PM exposures (e.g., ventilation and particle characteristics), the
investigators combined city-specific estimates into six regions classified based on the
Koppen climate classification scheme (e.g., "Mediterranean climates" for CA, OR, WA, etc.).
The overall combined excess risk estimates were: 0.98 % (0.75, 1.22) for all-cause! 0.85
% (0.46, 1.24) for CVD, 1.18 % (0.48, 1.89) for Ml; 1.78 % (0.96, 2.62) for stroke, and 1.68 %
(1.04, 2.33) for respiratory mortality for a 10 (Jg/m3 increase in PM2.5 at lag 0-1. When the
risk estimates were combined by season, the spring estimates were the largest for all-cause
and for all of the cause-specific mortality outcomes examined. For example, the risk
estimate for all-cause mortality for the spring was 2.57% (1.96-3.19) with the estimates for
the other seasons ranging from 0.25% to 0.95%. When examining cities that had both PM2.5
and PM10-2.5 data (i.e., 47 cities), the addition of PM10-2.5 in the model did not alter the PM2.5
estimates substantially, only decreasing slightly from 0.94% in a single pollutant model to
0.77% in a copollutant model with PM10-2.5,. When the risk estimates were combined by
climatic regions, the estimated PM2.srisk for all-cause mortality were similar (all above 1%
per 10 (Jg/m3 increase) for all the regions except for the "Mediterranean" region (0.5%)
which include cities in CA, OR and WA, though the estimates in that region were
significantly heterogeneous (Figure 6-24).
July 2009
6-283
DRAFT-DO NOT CITE OR QUOTE

-------
Climatic Region
All-Cause
CVD
Respiratory
Humid subtropical and maritime
Warm summer continental
Hot summer continental
Dry
Dry continental
Mediterranean
Increase
Increase
Increase
Source: Zanobetti and Schwartz (2009,188462).
Figure 6-23. Percent increase in mortality for 10 (Jg/m3 increase in the average of 0- and 1-day lagged
PM2.5, combined by climatic regions.
The PM2.5 risk estimate for all-cause mortality reported by Zanobetti and Schwartz
(2009, 188462) for 112 cities (0.98% per 10 |u,g/m3 increase in the average of 0- and 1-day
July 2009
6-284
DRAFT-DO NOT CITE OR QUOTE

-------
lags) is generally consistent with that reported by Franklin et al. (2007, 091257) for 27
cities (0.82% [0.02-1.63]) and Franklin et al. (2008, 097426) for 25 cities (0.74% [0.41-1.07])
using the same 0" and 1-day avg exposure time. The seasonal pattern (i.e., higher risk
estimates in the spring) found in this study is also consistent with the result from Franklin
et al. (2008, 097426). Figure 6-23 highlights the risk estimates for all-cause, CVD, and
respiratory morality combined by region. The regional division based on climatic types used
in this study makes it difficult to directly compare the regional pattern of results from
previous studies. However, an examination of empirical Bayes-adjusted effect estimates for
each of the cities included in the analysis further confirms the heterogeneity observed
between some cities and regions of the country (see Figure 6-24). It is noteworthy that,
unlike NMMAPS, which focused on PMio and indicated larger risk estimates in the
northeast, Zanobetti and Schwartz (2009, 188162Hound that the all-cause mortality risk
estimates were fairly uniform across the climatic regions, except for the "Mediterranean"
region.
July 2009
6-285
DRAFT-DO NOT CITE OR QUOTE

-------
I—I—I—I—I—I—I I I—I—I—I—1 t I	I—[—1—I—I—I—I—I—[—J—I—I—I—I—I	I—I—I t * I—I—11)1—I—I—I—I
-3MI -1J -gJ| 8Jt t.§ 1M &§ 4jB	-as -fj| -fts. ftt	3UI 4M	CUS f.f 5ES 341 4S
Increase in Total Mortality	Increase in Cardiovascular Mortality	Increase in Respiratory Mortality
Figure 6-24. Empirical Bayes-adjusted city-specific effect estimates for total, cardiovascular, and
respiratory mortality for 10 (Jg/m3 increase in the average of 0- and 1-day lagged PM2.5
by decreasing mean 24-h avg PM2.5 concentrations. Based on estimates calculated from
Zanobetti and Schwartz (2009,188462) using the approach specified in Le Tertre et al.
(2005, 087560).
Table 6-14. Key to Figure 6-24
City
Mean
CO
CO
City
Mean
CO
CO
City
Mean
CO
CO
City
Mean
CO
CO
Rubidoux, CA
24.7
68.0
Taylors, SC
15.0
32.2
Waukesha, Wl
13.4
35.3
Phoenix, AZ
11.4
30.7
Bakersfield, CA
21.7
80.3
Toledo, OH
14.9
36.6
BatonRouge, LA
13.4
30.1
Tacoma, WA
11.4
38.1
Los Angeles, CA
19.7
51.1
Anaheim, CA
14.9
44.1
Memphis, TN
13.3
32.4
Port Arthur, TX
11.1
25.7
Fresno, CA
18.7
64.9
NewYork, NY
14.7
38.1
Erie, PA
12.9
36.1
Cedar Rapids, IA
11.0
31.0
Atlanta, GA
17.6
38.2
Washington, PA
14.7
37.0
Dallas, TX
12.8
28.7
Dodge, Wl
10.9
32.9
Steubenville, OH
17.1
41.4
Winston, NC
14.7
34.1
Houston, TX
12.8
27.5
Oklahoma, OK
10.8
26.1
Cincinnati, OH
17.1
39.9
Elizabeth, NJ
14.6
38.2
Chesapeake, VA
12.8
29.8
Des Moines, IA
10.5
27.9
Birmingham, AL
16.5
38.8
Philadelphia, PA
14.6
36.6
Wilkesbarre, PA
12.8
32.5
Jacksonville, FL
10.5
25.3
Middletown, OH
16.5
38.4
St. Louis, M0
14.5
33.7
Norfolk, VA
12.7
29.6
Omaha, NE
10.5
28.0
Indianapolis, IN
16.4
38.2
Allentown, PA
14.4
38.9
Sacramento, CA
12.6
45.0
Denver, CO
10.5
26.4
Cleveland, OH
16.3
40.5
Richmond, VA
14.3
33.0
Springfield, MA
12.5
35.1
Pinellas, FL
10.4
23.1
Dayton, OH
16.3
38.3
Spartanburg, SC
14.2
31.4
NewOrleans, LA
12.5
29.0
Austin, TX
10.4
24.5
July 2009
6-286
DRAFT-DO NOT CITE OR QUOTE

-------
City
Mean
CO
CO
City
Mean
CO
CO
City
Mean
CO
CO
City
Mean
CO
CO
Columbus, OH
16.2
38.3
Durham, NC
14.2
32.9
Ft. Worth, TX
12.4
27.7
Orlando, FL
10.3
24.3
Detroit, Ml
16.2
41.0
LittleRock, AR
14.2
31.8
Pensacola, FL
12.3
31.2
Klamath, OR
10.2
40.7
Akron, OH
16.0
39.0
Easton, PA
14.2
39.7
Davenport, IA
12.3
32.1
Seattle, WA
10.1
27.9
Louisville, KY
15.9
38.0
Raleigh, NC
14.1
31.8
Avondale, LA
12.3
28.6
Medford, OR
10.0
37.3
Chicago, IL
15.8
39.1
Greensboro, NC
14.1
31.0
Boston, MA
12.3
30.2
Bath, NY
9.6
29.3
Pittsburgh, PA
15.7
43.1
Mercer, PA
14.1
36.4
Holland, Ml
12.1
35.0
Provo, UT
9.5
38.5
Harrisburg, PA
15.6
40.2
Annandale, VA
14.0
34.6
Charleston, SC
12.1
27.9
Miami, FL
9.4
20.5
Baltimore, MD
15.6
38.8
Nashville, TN
13.9
31.0
Tampa, FL
12.1
25.8
El Paso, TX
9.0
24.4
Youngstown, OH
15.6
38.1
Dumbarton, VA
13.8
31.9
Tulsa, OK
12.1
32.3
Spokane, WA
8.9
30.6
Knoxville, TN
15.5
32.9
Columbia, SC
13.7
30.7
Kansas, M0
12.0
28.6
SanAntonio, TX
8.9
21.9
Gary, IN
15.5
37.5
Milwaukee, Wl
13.7
36.3
Scranton, PA
11.9
33.0
Portland, OR
8.9
25.4
Charlotte, NC
15.3
32.7
NewHaven, CT
13.6
36.8
Hartford, CT
11.8
33.5
Davie, FL
8.4
19.1
Warren, OH
15.2
37.4
Grand Rapids, Ml
13.6
36.4
Minneapolis, MN
11.6
31.6
Eugene, OR
8.1
29.9
Washington, DC
15.2
37.2
El Cajon, CA
13.5
34.9
Worcester, MA
11.5
30.2
Palm Beach, FL
7.8
18.4
Wilmington, DE
15.1
37.6
Gettysburg, PA
13.4
36.5
Salt Lake, UT
11.5
52.4
Bend, OR
7.7
23.5
Carlisle, PA
15.1
40.0
State College, PA
13.4
38.5
Providence, Rl
11.5
30.5
Albuquerque, NM
6.6
17.9
Note: The top effect estimate in the figures represents the overall effect estimate for that mortality outcome across all cities. The remaining effect estimates are ordered by the highest (i.e., Rubidoux, CA) to
lowest (i.e., Albuquerque, NM) mean 24-h PM2.5 concentrations across the cities examined. In the key the cities are reported in this order, which represents the policy relevant concentrations for the annual
standard, but the policy relevant PM2.5 concentrations for the daily standard (i.e., 98th percentile of the 24-h average) are also listed for each city (from Zanobetti and Schwartz (2009))
PM2.5 - Mortality Associations on a Regional Scale: California
Ostro et al. (2006, 087991) examined associations between PM2.5 and daily mortality
in nine heavily populated California counties (Contra Costa, Fresno, Kern, Los Angeles,
Orange, Riverside, Sacramento, San Diego, and Santa Clara) using data from 1999 through
2002. The authors used a two-stage model to examine all-cause, respiratory, cardiovascular,
ischemic heart disease, and diabetes mortality individually and by potential effect modifier
(i.e., age, gender, race, ethnicity, and education level). The a priori exposure periods
examined included the average of 0- and 1-day lags (lag 0-1) and the 2-day lag (lag 2). The
authors selected these non-overlapping lags (i.e., rather than selecting lag 1 as the
single-day lag) because previous studies have reported stronger associations at lags of 1 or
2 days or with cumulative exposure over three days. It is unclear why the investigators
chose these non-overlapping lags (i.e., single-day lag of 2 instead of l) even though they
state they based the selection of their lag days on results presented in previous studies,
which found the strongest association for PM lagged 1 or 2 days. Using the average of 0-
and 1-day lags Ostro et al. (2006, 087991) reported combined estimates of 0.6%(CI: 0.2-1.0),
0.6% (CI: 0.0-1.1), 0.3% (CI: -0.5 to 1.0), 2.2% (CI: 0.6-3.9), and 2.4% (CI: 0.6-4.2) for
all-cause, cardiovascular, ischemic heart disease, respiratory, and diabetes deaths,
respectively, per 10 |u.g/m3. The authors also conducted a sensitivity analysis of risk
July 2009
6-287
DRAFT-DO NOT CITE OR QUOTE

-------
estimates based on the extent of temporal adjustment, which showed monotonic reductions
for all of the death categories examined when 4, 8, and 12 degrees of freedom per year were
used.
Five of the nine counties examined in the Ostro et al. (2006, 087991) analysis contain
cities that are among the 27 cities examined in the Franklin et al. (2007, 091257) analysis
for the same period, 1999-2002. While the lags used were different between these two
studies, both presented PM2.5 risk estimates in individual cities or counties (graphically in
the Franklin et al. study (2007, 091257); in a table in the Ostro et al. study (2006, 087991),
which allowed for a cursory evaluation of consistency between the two analyses. In
Franklin et al. (2007, 091257), PM2.5 risk estimates at lag 1 day for the cities Los Angeles
and Riverside were slightly negative, whereas Fresno, Sacramento, and San Diego showed
positive values above 1% per 10 |ug/m3 increase in PM2.5. The 2-day lag result presented in
Ostro et al. (2006, 087991) is qualitatively consistent, with Los Angeles and Riverside, both
of which show slightly negative estimates, while the other 3 locations all show positive, but
somewhat smaller estimates, than those reported by Franklin et al. (2007, 091257). The
estimates for the average of 0- and 1-day lags for these five counties in Ostro et al. (2006,
087991), which contain cities examined in Franklin et al. (2007, 091257), were all positive.
Thus, these two PM2.5 studies showed some consistencies in risk estimates even though
they used different lag periods and a different definition for the study areas of interest (i.e.,
counties vs. cities). The risk estimates for Franklin et al. (2007, 091257) and Ostro et al.
(2006, 087991), stratified by various effect modifiers (gender, race, etc.), are summarized in
Figure 6-25. Of note is the contrast in the results presented for the effect modification
analysis for "in-hospital" vs. "out-of-hospital" deaths for Ostro et al. (2006, 087991), which
differs from the results presented in the PM10 study conducted by Zeka et al. (2006,
088749). Ostro et al. (2006, 087991) observed comparable risk estimates for "in-hospital" vs.
"out-of-hospital" deaths, whereas Zeka et al. (2006, 088749) observed a large difference
between the two in the 20 cities study discussed earlier. This difference in effects observed
between the two studies is more than likely due to the compositional differences in PM10 in
the cities examined in Zeka et al. (2006, 088749) (i.e., PM10 more or less dominated by PM2.5
and the subsequent composition of PM2.5).
July 2009
6-288
DRAFT-DO NOT CITE OR QUOTE

-------
Multicitv PM7 5 Studies
% Increase
Ostro et al. (2006, 087991), 9 CA Counties, 0-1
All-cause
4ge > 65
Male
Female
White
Black
Hispanic
1
1
1
1
1
T*~
ln-hospital
Out-of-hospital
> High School
< High School
Franklin et al. (2007, 091257), 27 U.S. cities, 1
All-cause
Age > 75
Age < 75
Male
Female
East
West
PM2.6 > 15//g/m3
PM2.6 < 15/ig/m3
25th percentile air conditioning
75th percentile air conditioning
Summer peaking PM2.6 cities:
25th percentile air conditioning
75th percentile air conditioning
Franklin et al. (2008, 097426), 25 U.S. cities, 0-1
All-cause
Winter
Spring
Summer
Fall
West
East & Central
Zanobetti and Schwartz (2009,188462), 112 U.S. cities, 0-1
All-cause
Winter
Spring
Summer
Autumn
Figure 6-25. Summary of all-cause mortality PM2.5 risk estimates per 10 yug/m3 by various effect
modifiers.
July 2009
6-289
DRAFT-DO NOT CITE OR QUOTE

-------
PM2.5-Mortality Associations in Canada
An analysis of multiple pollutants, including PM2.5, in 12 Canadian cities found the
most consistent associations for NO2 (Burnett et al., 2004, 086247). In this analysis, PM2.5
was only measured every 6th-day in much of the study period, and the simultaneous
inclusion of NO2 and PM2.5 in a model on the days when PM2.5 data were available
eliminated the PM2.5 association (from 0.60% to -0.10% per 10 jug/m3 increase in PM2.5).
However, the investigators noted that during the later study period of 1998-2000 when
daily TEOM PM2.5 data were available for 11 of the 12 cities, a simultaneous inclusion of
NO2 and PM2.5 resulted in considerable reduction of the NO2 risk estimate, while the PM2.5
risk estimate was only slightly reduced from 1.1% to 0.98% (CP "0.16 to 2.14). Thus, the
relative importance of NO2 and PM2.5 on mortality effect estimates has not been resolved
when using the Canadian data sets.
Summary of PM2.5 Risk Estimates
The risk estimates for all-cause mortality for all ages ranged from 0.29% Dominici
et al. (2007, 099135) to 1.21% Franklin et al. (2007, 091257) per 10 |ug/m3 increase in PM2.5
(see Figure 6-26). An examination of cause-specific risk estimates found that PM2.5 risk
estimates for cardiovascular deaths are similar to those for all-cause deaths (0.30-1.03%),
while the effect estimates for respiratory deaths were consistently larger (1.01-2.2%), albeit
with larger confidence intervals, than those for all-cause or cardiovascular deaths using the
same lag/averaging indices. Figure 6-27 summarizes the PM2.5 risk estimates for all U.S.-
and Canadian-based studies by cause-specific mortality.
An examination of lag structure observed results similar to those reported for PM10
with most studies reporting either single day lags or two-day avg lags with the strongest
effects observed on lag 1 or lag 0-1. In addition, seasonal patterns of PM2.5 risk estimates
were found to be similar to those reported for PM10, with the warmer season showing the
strongest association. An evaluation of regional associations found that in most cases the
eastern U.S. had the highest PM2.5 mortality risk estimates, but this was dependent on the
geographic designations made in the study. When grouping cities by climatic regions,
similar PM2.5 mortality risk estimates were observed across the country except in the
Mediterranean region, which included CA, OR, and WA. Unlike the examination of PM10
risk estimates, the recently evaluated U.S.-based multicity studies did not analyze
potential confounding of PM2.5 risk estimates by gaseous pollutants. Burnett et al. (2004,
July 2009
6-290
DRAFT-DO NOT CITE OR QUOTE

-------
086247) in a Canadian multicity study did analyze gaseous pollutants and found mixed
results, with possible confounding by NO2. Therefore, it is unclear if gaseous pollutants
confound the PM2.5 mortality association.
Multicity PM2.5 Studies	% Increase
Dominici et al. (2007, 099135), NMMAPS
¦
All-cause, lag 1	1	»	
¦
Cardiorespiratory, lag 1	¦ i
Franklin et al. (2007, 091257), 27 U.S. cities
All-cause, lag 1
Cardiovascular, lag 1
Respiratory, lag 1
All-cause, lag 0-1
Cardiovascular, lag 0-1
Respiratory, lag 0-1
Franklin et al. (2008, 097426), 25 U.S. cities
All-cause, lag 0-1
Cardiovascular, lag 0-1
Respiratory, lag 1-2
Zanobetti and Schwartz (2009,188462), 112 U.S. cities
All-cause, lag 0-1
Cardiovascular, lag 0-1
Respiratory, lag 0-1
Ostro et al. (2006, 087991), 9 CA Counties
All-cause, lag 0-1
Cardiovascular, lag 0-1
Respiratory, lag 0-1
I	1	1	1	1	1	1	1	1	1
-0.5 ao as i.o t.s 2.0 2.5 10 as 4.0
Figure 6-26. Summary of PM2.5 risk estimates per 10 |jg/m3 for major underlying causes of death.
July 2009
6-291
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Location
Lan
% Innrnacn
Schwartz et al. (2003, 042800)
Burnett and Goldberp (2003, 042798)
Klemm and Mason (2003, 042801)
Moolgavkar (2003, 051316)
Ito (2003, 042856)
Klemm and Mason (2000, 010389)
Fairlev (2003, 042850)
Tsaiet al. (2000, 006251)
Tsaiet al. (2000, 006251)
Tsaiet al. (2000, 006251)
Chock et al. (2000, 010407)
Chock et al. (2000, 010407)
Dominici et al. (2007, 097361 a)
Zanobetti and Schwartz (2009,188462)
Franklin et al. (2007, 091257)
Franklin et al. (2008, 097426)
Burnett et al. (2004, 086247)
Ostro et al. (2006, 087991)
Slaughter et al. (2005, 073854)
Klemm et al. (2004, 056585)
Villeneuve et al. (2003, 055051)
Tsaiet al. (2000, 006251)
Tsaiet al. (2000, 006251)
Tsai et al. (2000, 006251)
Dominici et al. (2007, 097361 a)
Klemm and Mason (2003, 042801)
Ostro et al. (1995, 079197)
Lipfert et al. (2000, 004088)
Moolgavkar (2003, 051316)
Ito (2003, 042856)
Maretal. (2003, 042841)
Fairley (2003, 042850)
Zanobetti and Schwartz (2009,188462)
Franklin et al. (2007, 091257)
Franklin et al. (2008, 097426)
Ostro et al. (2007, 091354)
Ostro et al. (2006, 087991)
Holloman et al. (2004, 087375)
Wilson etal. (2007,157149)
Villeneuve et al. (2003, 055051)
Klemm and Mason (2003, 042801)
Ostro et al. (1995, 079197)
Moolqavkar (2003, 051316)
Ito (2003, 042856)
Fairley (2003, 042850)
Zanobetti and Schwartz (2009,188462)
Franklin et al. (2007, 091257)
Franklin et al. (2008, 097426)
Ostro et al. (2006, 087991)
Villeneuve et al. (2003, 055051)
"Studies shaded in grey represent the collective evidence
from the 2004 PM AQCD (U.S. EPA, 2004, 056905).
10 U.S. Cities
8	Canadian Cities
6 U.S. Cities
Los Angeles, CA
Detroit, Ml
Atlanta, GA
Santa Clara County,
Newark, NJ
Elizabeth, NJ
Camden, NJ
Pittsburgh, PA
Pittsburgh, PA
100 U.S. cities
112 U.S. cities
27 U.S. cities
25 U.S. cities
12 Canadian cities
9	California counties
Spokane, WA
Atlanta, GA
Vancouver, CAN
Newark, NJ
Elizabeth, NJ
Camden, NJ
100 U.S. cities
6 U.S. Cities
Southern California
Philadelphia, PA
Los Angeles, CA
Detroit, Ml
Phoenix, AZ
Santa Clara County,
112 U.S. cities
27 U.S. cities
25 U.S. cities
6	California counties
9 California counties
7	NC counties
Phoenix, AZ
Vancouver, CAN
6 U.S. Cities
Southern California
Los Angeles, CA
Detroit, Ml
Santa Clara County,
112 U.S. cities
27 U.S. cities
25 U.S. cities
9 California counties
Vancouver, CAN
0-1
1
0-1
1
3
0
0
0
0
0
0
0
1
0-1
1
0-1
1
0-1
1
0-1
0
0
0
0
0-1
0
1
1
1
1
0
0-1
1
0-1
3
0-1
0
0-5
1
0-1
0
1
0
0
0-1
0-1
1-2
0-1
0
(t-
65-t
Non-Accidental
-<75
¦754
65-t
Cardiorespiratory
Cardiovascular
>16
65#"
"654
-125
Respiratory
-1 f
1	1	1	1	1
7 8 11 13 15
Figure 6-27. Summary of PM2.5 risk estimates (per 10 yug/m3) for cause-specific mortality for all
U.S.- and Canadian-based studies. The three vertical lines for the Wilson et al. (2007,
157149) estimate represent the central, middle, and outer Phoenix estimates.
July 2009
6-292
DRAFT-DO NOT CITE OR QUOTE

-------
6.5.2.3. Thoracic Coarse Particles (PM102.5)
In the 2004 PM AQCD (U.S. EPA, 2004, 056905), a limited number of studies, mostly
single-city analyses, were evaluated that examined thoracic coarse (PM10-2.5) PM for its
association with mortality. Of these studies a small number examined both PM2.5 and
PM10-2.5 effects, and found some evidence for PM10-2.5 effects of the same magnitude as PM2.5.
However, multiple limitations in these studies were identified including measurement and
exposure issues for PM10-2.5 and the correlation between PM2.5 and PM10-2.5. These
limitations increased the uncertainty surrounding the concentrations at which PM10-2.5-
mortality associations are observed.
A thorough analysis of PM10-2.5 mortality associations requires information on the
speciation of PM10-2.5. This is because, while a large percent of the composition of coarse
particles may consist of crustal materials by mass, depending on available sources, the
surface chemical characteristics of PM10-2.5 may also vary from city to city. Thus, without
information on the chemical speciation of PM10-2.5, the apparent variability in observed
associations between PM10-2.5 and mortality across cities is difficult to characterize.
Although this type of information is not available in the current literature, the relative
importance of the associations observed between PM10-2.5 and mortality in the following
studies is of interest.
PM10 2.5 Concentrations Estimated Using the Difference Method
The Zanobetti and Schwartz (2009, 188462) multicity analysis, described for PM2.5
section (Section 6.5.2.2), also examined the association between computed PM10-2.5 and all-
causes, cardiovascular disease (CVD), MI, stroke, and respiratory mortality for the years
1999-2005. Of the 112 cities included in the PM2.5 analysis only 47 cities had both PM2.5 and
PM10 data available. PM10-2.5 was estimated in these cities by differencing the countywide
averages of PM10 and PM2.5. In addition to examining the association between PM10-2.5 and
mortality for the average of lags 0- and 1-day, the investigators also considered a
distributed lag of 0-3 days. The risk estimates for PM10-2.5 were presented for both a single
pollutant model and a copollutant model with PM2.5, and were also combined by season and
climatic regions as was done in the PM2.5 analysis.
The study found a significant association between the computed PM10-2.5 and all-
cause, CVD, stroke, and respiratory mortality. The combined estimate for the 47 cities
using the average of 0- and 1-day lag PM10-2.5 for all-cause mortality was 0.46% (95% CP
July 2009
6-293
DRAFT-DO NOT CITE OR QUOTE

-------
0.21-0.71) per 10 |ug/m3 increase with the estimate obtained using the distributed lag model
being smaller (0.31% [95% CP 0.00-0.63]). The seasonal analysis showed larger risk
estimates in the spring for all-cause (1.01%) and respiratory mortality (2.56%) (i.e., the
same pattern observed in the PM2.5 analysis); however, for CVD mortality, the estimates for
spring (0.95%) and summer (1.00%) were comparable. When the risk estimates were
combined by climatic region (Figure 6-28), for all-cause mortality, the "dry, continental"
region (which included Salt Lake City, Provo, and Denver, all of which had relatively high
estimated PM10-2.5 concentrations) showed the largest risk estimate (l.ll%[95% CP
0.11-2.11]), but the "dry" region (which included Phoenix and Albuquerque, the two cities with
high PM10-2.5 concentrations) and the "Mediterranean" region (which included cities in CA, OR, and WA)
did not show positive associations. The other three regions (i.e., "hot summer, continental," "warm
summer, continental," and "humid, subtropical and maritime"), which included cities that correspond to
the mid-west, southeast, and northeast geographic regions as defined in previous NMMAPS analyses, all
showed significantly positive associations. Similar regional patterns of associations were found for CVD
and respiratory mortality, which are further confirmed when examining the empirical Bayes-adjusted city-
specific estimates in Figure 6-29. The regional pattern of associations for MI and stroke are less clear,
because of the wider confidence intervals due to the smaller number of deaths in these specific categories.
The lack of a PMio-2 5-mortality association in the "dry" region reported in this study is in contrast to the
result from three studies that analyzed Phoenix data and found associations, as reviewed in the 2004 PM
AQCD (U.S. EPA, 2004, 056905). and Wilson et al. (2007, 157149) (discussed below).
Although the results from this analysis are informative because it is the first
multicity U.S.-based study that examined the association between short-term exposure to
PM10-2.5 and mortality on a large scale, some limitations do exist. Specifically, it is not clear
how the computed PM10-2.5 measurements used by Zanobetti and Schwartz (2009, 188462)
compare with the PM10-2.5 concentrations obtained by directly measuring PM10-2.5 using a
dichotomous sampler, or the PM10-2.5 concentrations computed using the difference of PM10
and PM2.5 measured at co-located samplers.
Additional studies evaluated the association between short-term exposure to PM10-2.5
and mortality using PM10-2.5 concentrations estimated by subtracting PM10 from PM2.5
concentrations at co-located monitors. Although PM10-2.5 concentrations estimated using this
approach are not ideal, the results from these studies are informative in evaluating the
PM10-2.5 mortality association.
July 2009
6-294
DRAFT-DO NOT CITE OR QUOTE

-------
Climatic Region
All-Cause
CVD
Respiratory
HumiH snhtrnninal and
Warm summer continental
Hot summer continental
Dry
H*-

Dry continental
Mediterranean
-1-
_1
Source: Zanobetti and Schwartz (2009,188462).
Figure 6-28. Percent increase in mortality for 10 (Jg/m3 increase in the average of 0- and 1-day lagged
PMio2.5, combined by climatic regions.
July 2009	6-295	DRAFT - DO NOT CITE OR QUOTE

-------
-2.5 -1.5 -0.5 0.5 1.0 f.5 2.0 2.5 3.0 3.5 -2.5 -1.5 -0.5 0.5 1.0 1..S 2.0 2.5 3.0 3.5 -M -I* -0.5 0.5 5 0 IS 2 0 fcj 3.0 3.8
Increase in Total Mortality	Increase in Cardiovascular Mortality	Increase in Respiratory Mortality
Figure 6-29. Empirical Bayes-adjusted city-specific effect estimates for total, cardiovascular, and
respiratory mortality for 10 (Jg/m3 increase in the average of 0- and 1-day lagged PM10 2.5
by decreasing 98th percentile of mean 24-h avg PM2.5 concentrations. Based on
estimates calculated from Zanohetti and Schwartz (2009,188462) using the approach
specified in Le Tertre et al. (2005, 087560).
Slaughter et al. (2005, 073854) examined the association of various PM size fractions
(PMi, PM2.5, PM10, PM10-2.5) and CO with ED visits, HAs, and mortality in Spokane, WA for
the period 1995-2001. Although the authors did not report mortality risk estimates for
PM10-2.5, they did not find an association between any PM size fraction (or CO) and
mortality or cardiac HAs at lags of 0- to 3-days.
Wilson et al. (2007, 157149) examined the association between size-fractionated PM
(PM2.5 and PM10-2.5) and cardiovascular mortality in Phoenix for the study period 1995-1997,
using mortality data aggregated for three geographic regions^ "Central Phoenix," "Middle
Ring," and "Outer Phoenix," which were constructed as a composite of ZIP codes of
residence in order to compare population size among the three areas. The authors reported
apparently different patterns of associations between PM2.5 and PM10-2.5 in terms of the size
July 2009
6-296
DRAFT-DO NOT CITE OR QUOTE

-------
of the risk estimate across the three areas and temporal patterns of associations. In the
"Middle Ring" where PM10-2.5 showed the strongest association, the estimated risk per
10 |ug/m3 increase for a 1 day lag was 3.4% (95% CP 1.0-5.8). The estimated risk for PM2.5
found for "Central Phoenix" was 6.6% (95% CP 1.1-12.5) for lag 1. The authors speculated
that the apparent difference in estimated risks across the areas might be due to the lower
SES in "Central Phoenix" or the lower exposure error, but the relatively wide confidence
bands of these estimates make it difficult to establish such relationships (see Section 8.1.7
for a detailed discussion on SES and susceptibility to PM exposure).
Table 6-15. Key for Figure 6-29
City
98th
Mean
City
98th
Mean
City
98th
Mean
City
98th
Mean
El Paso, TX
105.1
25.4
Cleveland, OH
51.2
15.2
Sacramento, CA
31.5
10.2
Louisville, KY
23.3
8.3
St. Louis, M0
81.9
15.2
Davenport, IA
49.9
15.3
Tampa, FL
29.1
12.9
Wilkesbarre, PA
22.2
6.2
Phoenix, AZ
80.1
33.3
Birmingham, AL
49.6
14.2
Toledo, OH
28.8
7.6
New York, NY
22.0
6.4
Detroit, Ml
77.5
17.3
Provo, UT
49.3
18.2
Washington, PA
27.8
6.5
Wilmington, DE
21.8
7.0
Gary, IN
71.3
6.9
Chicago, IL
46.1
12.4
Allentown, PA
27.8
4.5
Raleigh, NC
20.9
6.9
Omaha, NE
65.6
24.7
Easton, PA
43.9
12.0
Atlanta, GA
27.4
8.6
Scranton, PA
19.2
6.1
Albuquerque, NM
64.3
22.9
Steubenville, OH
43.5
12.1
Davie, FL
25.5
9.4
Harrisburg, PA
18.6
5.4
New Haven, CT
58.4
11.9
Columbia, SC
42.9
8.4
Taylors, SC
25.4
8.0
Akron, OH
17.7
5.3
Bakersfield, CA
55.9
16.1
Los Angeles, CA
42.5
13.5
Memphis, TN
24.3
9.3
Charleston, SC
17.6
6.6
Des Moines, IA
55.0
16.2
Spokane, WA
41.8
13.8
Seattle, WA
23.7
9.0
Winston, NC
16.5
7.4
Denver, CO
53.8
18.1
Columbus, OH
40.0
11.2
Baltimore, MD
23.5
8.9
Erie, PA
14.9
3.1
Salt Lake, UT
52.6
19.2
Pittsburgh, PA
32.0
9.4
Cincinnati, OH
23.3
7.8



Note: The top effect estimate in the figures represents the overall effect estimate for that mortality outcome across all cities. The remaining effect estimates are ordered by the highest (i.e.,
Rubidoux, CA) to lowest (i.e., Albuquerque, NM) 98th percentile of the mean 24-h PM102.5 concentrations across the cities examined, which is the policy relevant concentration for the daily standard
(from Zanobetti and Schwartz (2009)).
Kettunen et al. (2007, 091242) analyzed UFPs, PM2.5, PM10, PM10-2.5, and gaseous
pollutants for their associations with stroke mortality in Helsinki during the study period of
1998-2004. The authors did not observe an association between air pollution and mortality
for the whole year or cold season, but they did find associations for PM2.5 (13.3% [95% CP
2.3-25.5] per 10 |ug/m3), PM10, and CO during the warm season, most strongly at lag 1 day.
An association was also observed for PM10-2.5 during the warm season (7.8% [95% CP -7.4 to
25.5] per 10 |u,g/m3 at lag 1 day); however, it was weaker than PM2.5.
The Perez et al. (2008, 156020) analysis tested the hypothesis that outbreaks of
Saharan dust exacerbate the effects of PM2.5 and PM10-2.5 on daily mortality. Changes of
effects between Saharan and non-Saharan dust days were assessed using a time-stratified
case-crossover design involving 24,850 deaths between March 2003 and December 2004 in
July 2009
6-297
DRAFT-DO NOT CITE OR QUOTE

-------
Barcelona, Spain. Saharan dust days were identified from back-trajectory and satellite
images. Chemical speciation, but not an analysis for microbes or fungi, was conducted
approximately once a week during the study period. On Saharan dust days, mean
concentrations were 1.2 times higher for PM2.5 (29.9 ug/m3) and 1.1 times higher for PM10-2.5
(16.4 jug/m3) than on non-Saharan dust days. During Saharan dust days (90 days out of
602), the PM10-2.5 risk estimate was 8.4% (95% [CP 1.5-15.8]) per 10 |ug/m3 increase at lag 1
day, compared with 1.4% (95% CP -0.8% to 3.4%]) during non-Saharan dust days. In
contrast, there was not an additional increased risk of daily mortality for PM2.5 during
Saharan dust days (5.0% [95% CP 0.5-9.7]) compared with non-Saharan dust days (3.5%
[95% CP 1.6-5.5]). Plowever, differences in chemical composition (i.e., PM2.5 was primarily
composed of nonmineral carbon and secondary aerosols! whereas PM10-2.5 was dominated by
crustal elements) did not explain these observations. Note also when examining all days
combined, both size fractions were associated with mortality, but the PM2.5 association was
found to be stronger.
PM10 2.5 Concentrations Directly Measured
In Burnett et al. (2004, 086247), which analyzed the association of multiple pollutants
with mortality in 12 Canadian cities, described previously! the authors also examined
PM10-2.5. In this study the authors collected PM10-2.5 using dichotomous samplers with an
every-6th-day schedule. When both NO2 and PM10-2.5 were included in the regression model,
the PM10-2.5 effect estimate was reduced from 0.65% (CP -0.10 to 1.4) to 0.31% (95% CP -0.49
to l.l) per 10 |ug/m3 increase in 1-day lag PM10-2.5. These risk estimates are similar to those
reported for PM2.5, which were also reduced upon the inclusion of NO2 in the two-pollutant
model, but to a greater extent, from 0.60% (95% [CP -0.03 to 1.2]) to -0.1% (95% [CP -0.86 to
0.67]).
Villeneuve et al. (2003, 055051) analyzed the association between PM2.5, PM10-2.5, TSP,
PM10, SO42 , and gaseous copollutants in Vancouver, Canada, using a cohort of
approximately 550,000 whose vital status was ascertained between 1986 and 1999. In this
study PM2.5 and PM10-2.5 were directly measured using dichotomous samplers. The authors
examined the association of each air pollutant with all-cause, cardiovascular, and
respiratory mortality, but only observed significant results for cardiovascular mortality at
lag 0 for both PM10-2.5 and PM2.5. They found that PM10-2.5 (5.4% [95% CP 1.1-9.8] per
July 2009
6-298
DRAFT-DO NOT CITE OR QUOTE

-------
10 ug/m3', was more strongly associated with cardiovascular mortality than PM2.5 (4.8%
[95% CL -1.9 to 12.0] per 10 |u,g/m3).
Klemm et al. (2004, 056585) analyzed various components of PM and gaseous
pollutants for their associations with mortality in Fulton and DeKalb Counties, Georgia for
the two-yr period, 1998-2000. PM10-2.5 concentrations were obtained from the ARIES
database, which directly measured PM10-2.5 using dichotomous samplers. In this analysis
the authors adjusted for temporal trend using quarterly, monthly, and biweekly knots, and
reported estimates for all-cause, circulatory, respiratory, cancer, and other causes mortality
for each scenario. Overall, PM2.5 was, generally, more strongly associated with mortality
than PM10-2.5. For example, using the average of 0- and 1-day lags, the risk estimates for
PM2.5 and PM10-2.5 in the monthly knots model for all-cause mortality, ages > 65 were 5.6%
(95% [CP 1.9-9.5]) and 6.4% (95% [CI: -0.5 to 14.1]) per 10 |ug/m3 increase, respectively.1
1 The monthly knot model was selected for comparison because, overall, PM2.5 showed the strongest association with all-cause
mortality among the 15 air pollution indices examined when using this model.
July 2009
6-299
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Location
Lag
% Increase
Klemm et al. (2003, 042801)
Burnett and Goldberg (2003, 0427981
Ito (2003, 042856)
Klemm and Mason (2000, 010389)
Fairley (2003, 042850)
Chock et al. (2000, 010407)
Chock et al. (2000.010407)
Zanobetti and Schwartz (2009,188462)
Burnett et al. (2004,086247)
Klemm et al. (2004.056585)
Villeneuve et al. (2003, 055051)
Lipfert et al. (2000, 004088)
Ito (2003, 042856)
Maret al. (2003. 042841)
Fairley (2003, 042850)
Ostro et al. (2003, 042824)
Zanobetti and Schwartz (2009,188462)
Wilson et al. (2007,157149)
Villeneuve et al. (2003, 055051)
Ito (2003.042856)
Fairley (2003, 042850)
Zanobetti and Schwartz (2009,188462)
Villeneuve et al. (2003,055051)
6 U.S. Cities
8 Canadian Cities
Detroit, Ml
Atlanta, GA
Santa Clara County, CA
Pittsburgh, PA
Pittsburgh, PA
47 U.S. cities
12 Canadian cities
Atlanta, GA
Vancouver, CAN
Philadelphia, PA
Detroit, Ml
Phoenix, AZ
Santa Clara County, CA
Coachella Valley, CA
47 U.S. cities
Phoenix, AZ
Vancouver, CAN
Detroit, Ml
Santa Clara County, CA
47 U.S. cities
Vancouver, CAN
"Studies shaded in grey represent the collective evidence from the
2004 PM ADCD (U.S. EPA, 2004, 056905).
0-1
1
1
0
0
0
0
0-1
1
0-1
0
1
1
1
0
0
0
0-1
0-5
0
2
0
0-1
0
654
Nonaccidental
<75
- 75 +
. 65-t
Cardiovascular
++
>25
.654
Respiratory
654
-1
~l	1	1
11 13 15
Figure 6-30. Summary of PM10 2.5 risk estimates (per 10 yug/m3) for cause-specific mortality for all
U.S.-, Canadian-, and international-based studies. The three vertical lines for the Wilson
et al. (2007,157149) estimate represent the central, middle, and outer Phoenix
estimates.
Summary of PM10 2.5 Risk Estimates
The results from newly available studies that examined the association between
short-term exposure to PM10-2.5 primarily consisted of single-city studies. Collectively these
studies found inconsistent associations, which varied depending on the study location. The
evidence from those single-city studies conducted in the U.S. and Canada in combination
July 2009
6-300
DRAFT-DO NOT CITE OR QUOTE

-------
with the multicity studies evaluated (i.e., in the U.S. and Canada), provide evidence for
PMio-2.5 effects. Although there are limitations in the method used in the new U.S. multicity
study (Zanobetti and Schwartz, 2009, 188462) to estimate PM10-2.5 (i.e., PM10-2.5 was
estimated by the difference between the county average PM10 and PM2.5), associations
between PM10-2.5 and mortality were observed throughout regions of the country. However,
some of the findings of this new multicity study (e.g., no associations in "dry" region where
PM10-2.5 levels are high) are not consistent with the findings of the PM10-2.5 studies evaluated
in the 2004 PM AQCD (U.S. EPA, 2004, 056905), and suggest that the coarse fraction is
associated with mortality in areas of the U.S. where PM10-2.5 levels are not high. Limitations
also exist in the PM10-2.5 associations reported due to the small number of PM10-2.5 studies
that have investigated confounding by gaseous copollutants or the influence of model
specification on PM10-2.5 risk estimates. Additionally, more data is needed to characterize the
chemical and biological components that may modify the potential toxicity of PM10-2.5.
Figure 6-30 summarizes the PM10-2.5 risk estimates for all U.S.-, Canadian-, and
international-based studies by cause-specific mortality.
6.5.2.4. Ultrafine PM
The 2004 PM AQCD (U.S. EPA, 2004, 056905) reviewed Wichmann et al.'s (2000,
013912; reanalyzed by Stolzel et al., 2003, 042842) study of fine and ultrafine particles
(UFP) (diameter: 0.01-0.1 jum) in Erfurt, Germany for the study period 1995-1998. Stolzel
et al. (2007, 091374) extended the study period to include the years 1995-2001 and updated
the analysis. Number concentrations (NC) for four size ranges of UFP (0.01-0.1, 0.01-0.03,
0.03-0.05, and 0.05-0.1 |um) as well as mass concentration (MC) for three size ranges
(0.01-2.5, 0.1-0.5, and 10 |u,m) were analyzed. The authors found associations with UFP NC
and all-cause as well as cardio-respiratory mortality, each for a 4 day lag. The risk
estimates associated with a 9,748/cm3 increase in UFP NC was 2.9% (95% CP 0.3-5.5) for
all-cause mortality and 3.1% (95% CP 0.3-6.0) for cardio-respiratory mortality. The
UFP-mortality association, and the lag structure of association, is consistent with the
results from their earlier analysis, but the PM2.5 association found in the previous study
was not observed in the updated analysis. Both UFP and PM2.5 concentrations were higher
during the cold season in this locale.
Breitner et al. (2009) analyzed UFP data from Erfurt, Germany over a 10.5-yr period
(October 1991 through March 2002) after the German unification, when air quality
July 2009
6-301
DRAFT-DO NOT CITE OR QUOTE

-------
improved. In this analysis associations between all-cause mortality and UFP and PM2.5
were analyzed from September 1995 to March 2002, while PM10, NO2 and CO was analyzed
for the whole study duration. The exposure time window / averages used in this study were
different from those used by Stolzel (2003, 042842) and Stolzel et al. (2007, 091374).
Breitner et al. investigated the cumulative effect of air pollution on mortality at lags 0-5
and, 0-14 using (a) a semiparametric Poisson regression model and (b) a third degree
polynomial distributed lag (PDL) model. The authors estimated the mortality risk for the
entire study period as well as specific time periods to examine the effect of declining air
pollution levels on the air pollution-mortality association. Of the air pollutants examined,
UFP were found to be most consistently associated with mortality. NO2 and CO were also
found to be significantly associated with mortality using the 15-day PDL and 15-day avg
models, respectively. PM2.5 and PM10 also showed positive, but much weaker associations
with mortality. In this data set, UFP were only moderately correlated with PM2.5 (r = 0.48)
and PM10 (r = 0.57). Of the pollutants examined, NO2 showed the strongest (but overall a
moderate) correlation with UFP (r = 0.62). When the risk estimates were compared between
the two latter time periods of the study (September 1995 to February 1998; and March 1998
to March 2002), the estimates obtained using the 6-day avg for these pollutants generally
declined. For example, the all-cause mortality risk estimates associated with a 8,439/cm3
increase in UFP NC was 5.5% (95% CP 1.1-10.5) for the earlier period and -1.1% (95% CP -
6.8 to 4.9) for the later period. However, such patterns were less clear when using 15-
day avg pollutants concentrations. In summary, UFP appear to be the pollutant most
consistently associated with mortality in Erfurt, Germany, but combined with the results
for NO2 and CO, these associations may implicate the role of local combustion sources on
the mortality association observed.
Kettunen et al.'s (2007, 091242) study in Helsinki also examined the relationship
between UFP and stroke mortality. As described earlier, PM2.5, PM10, and CO was
associated with stroke mortality only during the warm season. The association with UFP
was borderline non-significant (8.5% [95% CP -1.2 tol9.l] per 4,979/cm3 increase in UFP at
lag 1 day), but its lag structure of association and the magnitude of the effect estimate per
interquartile-range are similar to those for PM2.5. Note that the UFP NC levels in Helsinki
(median equals 8,986/cm3 during the cold season and 7,587/cm3 during the warm season)
are lower than those in Erfurt (mean = 13,549/cm3), but clearly higher in the cold season.
July 2009
6-302
DRAFT-DO NOT CITE OR QUOTE

-------
Summary of Ultra-Fine Particle (UFP) Risk Estimates
Only a few new studies, all of them from Europe, examined and reported associations between
UFP and mortality. In Erfurt, UFP showed the strongest associations with mortality among all of the PM
indices, but its lag structure of association is either unique (strongest association at lag 4 days in Stolzel et
al., 2007, 091374). and not consistent with the lag structure of associations found in other mortality
studies, or the time-windows examined are longer (0-5 and 0-14 days in Breitner et al., 2009), making it
difficult to compare whether the associations observed are consistent with those reported in other studies.
In Helsinki, the association between UFP and stroke mortality was weaker than that for PM2.5, but its lag
structure of association was similar to that for PM2.5 (strongest at lag 1 day). However, Kettunen et al.
(2007, 091242) only examined lags 0 through 3 days. Overall, the results of these studies should be
viewed with caution because UFP were consistently found to be correlated with gaseous pollutants
derived from local combustion sources, and one or more of the gaseous pollutants were also found to be
associated with mortality. Clearly, more research is needed to further investigate the role of UFP on
PM-mortality associations.
6.5.2.5. Chemical Components of PM
A few recent studies have examined the association between mortality and
components of PM2.5. This endeavor has been undertaken by some investigators through
the use of the newly available PM2.5 chemical speciation network data. The PM2.5 chemical
speciation network consists of more than 250 monitors that have been collecting over 40
chemical species since 2000; however, most sites started collecting data in 2001. One caveat
to the new network is that because the sampling frequencies of the monitors are either
every third day or every sixth day, there have not been, generally, a sufficient number of
days to examine associations with mortality in single cities. To circumvent this issue, some
investigators (Bell et al., 2009, 191997; Dominici et al., 2007, 099135; Franklin et al., 2008,
097426; Lippmann et al., 2006, 091165) have used the PM2.5 chemical species data in a
second stage regression to explain the heterogeneity in PM10 or PM2.5 mortality risk
estimates across cities. However, it should be noted these studies assume that the relative
contributions of PIVb.shave remained the same over time. There have also been some
studies that directly analyzed speciated PM2.5 data (e.g., Klemm et al., 2004, 056585; Ostro
et al., 2007, 091354).
July 2009
6-303
DRAFT-DO NOT CITE OR QUOTE

-------
Explaining the Heterogeneity of PM10 Risk Estimates Using PM2.5 Chemical Speciation
Data
Lippmann et al. (2006, 091165), in addition to their primary analysis1, investigated
the consistency of the associations between specific elements and health outcomes by
examining the heterogeneity of published 1 "day lagged NMMAPS PM10 mortality risk
estimates for 1987-1994 across cities as a function of the average PM2.5 chemical
components across cities. They matched PM2.5 chemical species in 60 out of 90 cities.
Lippmann et al. (2006, 091165) noted that the concentrations of the 16 chemical species
examined averaged over the years 2000-2003 were highly skewed across cities. They
therefore regressed PM10 risk estimates on each of the PM2.5 components, raw and
log-transformed, with weights based on the standard error of the PM10 risk estimates. The
log-transformed values yielded better predictive power, and the authors subsequently
presented the results with log-transformed values. As shown in Figure 6"31, the 16 PM2.5
species showed varying extent of predictive power in explaining the PM10 risk estimates
across 60 cities, with Ni and V being the best predictors.
1 The main focus of the study was to examine the role of PM2.5 chemical components in a mouse model of atherosclerosis
(ApoK ) exposed to concentrated fine PM (CAPs) in Tuxedo, NY.
July 2009
6-304
DRAFT-DO NOT CITE OR QUOTE

-------
Ni-
V-
EC-
Zn-
50,-
Cy-
Pb-
OC-
pmk-
Se-
Cf-
Mn-
Fe-
As-
N03-
ai-
St-
-0.5	0.0	0.5	1.0
Percent per 10-^g/iii3 increase in PM„
Source: Lippmann et al. (2006,091165)
Figure 6-31. Percent increase in PM10 risk estimates (point estimates and 95% confidence intervals)
associated with a 5th-to-95th percentile: increase in PM2.5 and PM2.5 chemical
components. The PM2.5 chemical components were log-transformed in the regression.
The PM10 risk estimates were for 60 NMMAP cities for 1987-1994.
July 2009
6-305
DRAFT-DO NOT CITE OR QUOTE

-------
No communities removed
New York removed
-20 -10	0	10	20	30
Percent increase in PM10 risk estimates per IQR Ni
Source: Dominici et al. (2007,099135)
Figure 6-32. Sensitivity of the percent increase in PM10 risk estimates (point estimates and 95%
confidence intervals) associated with an interquartile increase in l\li. The l\li
concentration was not log-transformed in this regression model. The PM10 risk estimates
were for 72 NMMAP cities for 1987-2000. The top estimate is achieved by including
data for all the 69 communities. The other estimates are calculated by excluding one of
the 69 communities at a time.
Dominici et al. (2007, 099135) aNi and V on the updated NMMAPS PMio mortality
risk estimates for 1987-2000, using 72 counties in which Ni and V data were collected. A
Bayesian hierarchical model was used to estimate the role of Ni and V on the heterogeneity
of PMio risk estimates. While they found both Ni and V to be significant predictors of
variation in PMio mortality risk estimates across cities, they also noted that this result was
sensitive to the inclusion of the New York City data. Lippmann et al. (2006, 091165) and
Dominici et al. (2007, 099135) both reported that the Ni levels in New York City are
particularly high (~10 times the national average). Figure 6"32 shows the result of the
sensitivity analysis for Ni. Note that the Ni in this result was not log-transformed, as
clearly reflected in the change in the width of confidence bands when the New York data
were removed (i.e., a skewed distribution produces narrow bands). Dominici et al. (2007,
099135) further noted that they reached "the same conclusion" when log-transformed data
were used in the analysis, but the results were not presented.
Bell et al. (2009, 191997) presented a supplemental analysis similar to both
Lippmann et al. (2006, 091165) and Dominici et al. (2007, 099135) in their examination of
whether the variation in PM2.5 risks for cardiovascular and respiratory hospital admissions
July 2009
6-306
DRAFT-DO NOT CITE OR QUOTE

-------
is due to differences in PM2.5 chemical composition. The authors used the 100 U.S. cities
included in the Peng et al. (2005, 087463) analysis and PM10 data for the years 1987-2000
along with PM2.5 chemical component data for 2000-2005. Using a Bayesian hierarchical
model, Bell et al. (2009, 191997) found that PM10 relative risks for total mortality were
greater in counties and during seasons with higher PM2.5 Ni concentrations. However, in a
sensitivity analysis when selectively removing cities from the overall estimate, the
significant association between the PM10 mortality risk estimate and the PM2.5 Ni fraction
was diminished upon removing New York city from the analysis, which is consistent with
the results presented by Dominici et al. (2007, 099135).
Explaining the Heterogeneity of PM2.5 Risk Estimates Using PM2.5 Chemical Speciation
Data
The first stage of the Franklin et al. (2008, 097426) 25 cities study, described
previously, focused on a time-series regression of mortality on PM2.5 by season. In the
second stage random effects meta regression, the PM2.5 mortality risk estimates (25 cities x
4 seasons = 100 estimates) were regressed on the ratio of mean seasonal PM2.5 species to
the total PM2.5 mass. The authors included those species that had at least 25% of the
reported concentrations above the minimum detection limit, which resulted in 18 species
being included in the analysis. Their rationale for using species proportions as effect
modifiers, according to the investigators, was that "in the first stage of the analysis the
mortality risk was estimated per unit of the total PM2.5 mass, which encompassed all
measured species, and therefore it would not be meaningful to use the species
concentrations directly as the effect modifier" (Franklin et al., 2008, 097426). In the second
stage regression model, Franklin et al. (2008, 097426) also included a quadratic function of
seasonally averaged temperature to capture the inverted U-shape relationship between
PM2.5 penetration and temperature. They found that the fitted relationship between PM2.5
risk estimates across cities and seasonally averaged temperature substantiates the use of
temperature as a surrogate for ventilation (Franklin et al., 2008, 097426). Table 6-16 shows
the resulting effect modification by PM2.5 species. Al, As, Ni, Si, and SC>42~ were found to be
significant effect modifiers of PM2.5 risk estimates, and simultaneously including Al, Ni, and
SO42- together, or Al, Ni, and As together further increased explanatory power. Of all the
species examined, Al and Ni explained the most residual heterogeneity. Franklin et al.
(2008, 097426) also examined the effect of demographic variables on PM2.5 risk estimates
and found that only median household income was significantly associated with mortality.
July 2009
6-307
DRAFT-DO NOT CITE OR QUOTE

-------
Table 6-16. Effect modification of composition on the estimated percent increase in mortality with a
10 (Jg/m3 increase in PM2.5.
Cause
Species
p-value for effect
% increase in non-accidental mortality per 10 jjg/m3
Heterogeneity


modification by species to
increase in PM2.5 for an interquartile increase in
explained (%)f


PM2.5 mass proportion
species to PM2.5 mass proportion*

Non-accidental
Al
<0.001
0.58
45
Univariate
As

Br
Cr
0.02
0.55
35

EC
Fe
0.11
0.38
5

K
0.12
0.33
16

Mn




Na +
0.79
0.06
0

Ni




NOs
0.43
0.12
3

NH4




OC
0.10
0.41
28

Pb
Si
0.42
0.14
10

SO42-




V
0.22
0.20
14

Zn
0.01
0.37
41


0.07
-0.49
28


0.84
0.04
3


0.59
-0.02
4


0.31
0.17
11


0.03
0.41
25


0.01
0.51
33


0.28
0.30
3


0.19
0.23
15
Non-accidental
Al
<0.001
0.79

Multivariate
Ni

(1)
SO42-
0.01
0.34
100


<0.001
0.75

Non-accidental
Al
<0.001
0.61

Multivariate
Ni

(2)
As
0.01
0.35
100


<0.001
0.58

Adjusted for temperature
tlncludes heterogeneity explained by temperature
Source: Franklin et al. (2008, 097426)
Although Lippmann et al. (2006, 091165) used NMMAPS PM10 risk estimates and
Franklin et al. (2008, 097426) used PM2.5 risk estimates to examine effect modification due
to various PM species, 14 out of the 18 species analyzed in these two studies overlap (see
Figure 6-31 and Table 6-16). Both studies found that Ni explained the heterogeneity in PM
risk estimates. Note that New York City was not included in the 25 cities examined in
Franklin et al. (2008, 097426) and, thus, could not influence the result. Sulfate positively,
but not significantly, explained the PM10 risk estimates in the Lippmann et al. (2006,
091165) analysis. However, SO42- was a significant predictor of PM2.5 risk estimates in the
July 2009
6-308
DRAFT-DO NOT CITE OR QUOTE

-------
Franklin et al. (2008, 097426) analysis. A1 and Si were negative (i.e., less than the average
PMio risk estimates across cities), though not significant, predictors in the Lippmann et al.
(2006, 091165) analysis. Unlike the Franklin et al. (2008, 097426) analysis, arsenic (As)
showed no association with mortality in the Lippmann et al. (2006, 091165) analysis. The
source of these differences may be due to the difference in geographic coverage, PM size
(PM2.5 may represent more secondary aerosols than PM10), or the difference in the
analytical methods used in each study. Specifically, the analytical approach used by
Franklin et al. (2008, 097426) does have an advantage of delineating seasonal variations in
PM components and the associated potential seasonal mortality effects.
In light of the results presented in speciation studies it must be noted that second
stage analyses that use PM chemical species as effect modifiers have some limitations.
Unlike analyses that directly examine the associations between chemical species and
mortality, if an effect modification is observed it may be confounded if the variations of the
mean levels of the chemical species examined are correlated with other demographic factors
that vary across cities. Thus, more concrete conclusions could be formulated if direct
associations are found between mortality and PM chemical components in time-series
analyses.
Association between PM2.5 Chemical Components and Mortality
Ostro et al. (2007, 091354) examined associations between PM2.5 chemical components
and mortality in six California counties (Fresno, Kern, Riverside, Sacramento, San Diego,
and Santa Clara), which had at least 180 days of speciation data for the years 2000 to 2003.
The study examined all-cause, cardiovascular, and respiratory mortality for individual lags
of 19 specific PM2.5 chemical components. The second stage random-effects model combined
risk estimates at each lag across cities. The number of available days for chemical species
data ranged from 243 (San Diego County) to 395 (Sacramento County). The authors found
an association between mortality, especially cardiovascular mortality, and several chemical
components. For example, cardiovascular mortality was associated with EC, OC, nitrate,
Fe, K, and Ti at various lags.
Even though this was a multicity study, the relatively small number of available days
and the every-third-day (or every-sixth-day) sampling frequency for PM2.5 chemical species
made it difficult to interpret the results of the lag structure of associations observed for the
chemical species. To evaluate the impact of non-daily sampling frequency, Ostro et al. (2007,
July 2009
6-309
DRAFT-DO NOT CITE OR QUOTE

-------
091354) examined both the PM2.5 series that coincides with the speciation sampling days
(for the initial six counties [i.e., PM2.5J) and PM2.5 data that was available on all days for an
extended set of counties (the initial six counties plus Contra Costa, Los Angeles, and
Orange Counties [i.e., PM2.5ext]). Figure 6-33 shows the association between all-cause
mortality and selected PM2.5 chemical species as well as for PM2.sc and PM2.5ext. Note the
wide confidence bands for the risk estimates for each PM2.5 chemical species and PM2.sc,
apparently reflecting the low statistical power of the data. The lag structure of associations
is more clearly defined for PM2.5ext, and appears to be different from that for PM2.5.
3,00
2.
1
2.00
100
u
m 0
*
Q
a-
«
100
200
* *
• •
• »•
» •
• •
• *
C " ? 3 0 1 2 3 0)2 3 own e ' 2 'J 0 r ?
m,,,	EC	QC	»W,,	SO.	C.I
Species and lag day
0 1 2 3 0 12 3 It I :>
K	2n	,
Source: Ostro et al. (2007,091354)
Figure 6-33. Excess risk (CI) of total mortality per IQR of concentrations. Note: PM2.5 has the same
sampling days as chemical species. PM2.5 has all available PM2.5ext data for nine counties.
* p < 0.10;** p < 0.05
Ostro et al. (2008, 097971) used the speciation data from the six counties analyzed in
their 2007 analysis, described above, in an additional analysis to examine effect
modification of cardiovascular mortality effects, which showed the strongest association in
the 2007 analysis, attributed to PM2.5 and 13 chemical components by socio-economic and
demographic factors. The results of the analysis were combined using random effects
meta-analysis. The investigators tested statistical differences in risk estimates between
July 2009
6-310
DRAFT-DO NOT CITE OR QUOTE

-------
strata using a t-test, and reported that, for many of the PM2.5 chemical species! there were
significantly higher effect estimates among those with lower educational attainment and
Hispanics. While these patterns were apparent in their results table, interpretation of the
results is not straightforward because the table only presented the most significant (and
positive) lags, and they were often different between the strata (e.g., the most frequent
significant lag for the Hispanic group was 1 day, while it was 2 or 3 days for the White
group). As the investigators pointed out, the every third-day sampling frequency of the
speciation data also complicates the interpretation of the results for different lags.
Overall, the two studies by Ostro et al. (2007, 091354) were the first attempt to
directly analyze associations between the newly available chemical speciation data and
mortality. While suggestive associations between several chemical species and mortality
were reported, a longer length of observations is needed to more clearly determine the
associations.
6.5.2.6. Source-Apportioned PM Analyses
Chemically speciated PM data allow for the source apportionment of PM. The idea of
using source-apportioned PM for health effects analyses is appealing because, if such
source-apportionment could be reliably conducted, it will allow for an evaluation of PM2.5
mass concentrations by source types. However, the uncertainties associated with
source-apportionment methods have not been well characterized.
To address this issue, in 2003, several groups of EPA-funded researchers organized a
workshop and independently conducted source apportionment on two sets of data: Phoenix,
AZ, and Washington, DC, compared the results (Hopke et al., 2006, 088390), and then
conducted time-series mortality regression analyses using each group's source-apportioned
data (Thurston et al., 2005, 097949) Ito, 2006 #5669 (Mar et al., 2006, 086143). The various
research groups generally identified the same major source types, each with similar
elemental compositions. Inter-group correlation analyses indicated that soil", SO42-,
residual oil-, and salt-associated mass concentrations were most unambiguously identified
by various methods, whereas vegetative burning and traffic were less consistent. Aggregate
source-specific mortality relative risk (RR) estimate confidence intervals overlapped each
other, but the SOr -related PM2.5 component was most consistently significant across
analyses in these cities.
July 2009
6-311
DRAFT-DO NOT CITE OR QUOTE

-------
The results from the source-apportionment workshop quantitatively characterized the
uncertainties associated with the factor analysis-based methods, but they also raised new
issues. The mortality analyses conducted in Phoenix, AZ, and Washington, DC, both found
that different source-types showed varying lag structure of associations with mortality. For
example, Figure 6"34 shows cardiovascular mortality risk estimates for three of the PM2.5
sources from the Phoenix, AZ, analysis (Mar et al., 2006, 086143). The strongest
associations for "traffic" PM2.5 was found for lag 1-day, while for "secondary SO r " PM2.5, it
was lag 0, with a monotonic decline towards longer lags. These results are consistent with
those in the 2004 PM AQCD (U.S. EPA, 2004, 056905), in which associations were reported
with combustion-related PM2.5, but not crustal source PM2.5. It is conceivable that PM from
different source types produces different lagged effects, but it is also likely that different
PM species have varying lagged correlations with the covariates in the health effects
regression models (e.g., temperature, day-of-week) resulting in apparent differences in
lagged associations with mortality. Thus, interpretation of these source-apportioned PM
health effect estimates remains challenging.
Secondary sulfate
Traffic
A8CDEFGH I	A8CDEFGH I	ABCDEFQH I
Source: Mar et al. (2006, 086143)
Figure 6-34. Relative risk and CI of cardiovascular mortality associated with estimated PM2.5 source
contributions. Y-axis: relative risk per 5th-to-95th percentile increment of estimated
PM2.5 source contribution. X-axis: the Dbet denotes investigator/ method; lagged PM2.5
source contribution for lag 0 through 5 days, left to right, are shown for each
investigator/method.
July 2009
6-312
DRAFT-DO NOT CITE OR QUOTE

-------
6.5.2.7. Investigation of Concentration-Response Relationship
The results from large multicity studies reviewed in the 2004 PM AQCD (U.S. EPA,
2004, 056905) suggested that strong evidence did not exist for a clear threshold for PM
mortality effects. However, as discussed in the 2004 PM AQCD (U.S. EPA, 2004, 056905),
there are several challenges in determining and interpreting the shape of PM-mortality
concentration-response functions and the presence of a threshold, including: (l) limited
range of available concentration levels (i.e., sparse data at the low and high end); (2)
heterogeneity of susceptibility in at-risk populations! and (3) the influence of measurement
error. Regardless of these limitations, studies have continued to investigate the
PM-mortality concentration-response relationship.
Daniels et al. (2004, 087343) evaluated three concentration-response models: (l)
log-linear models (i.e., the most commonly used approach, from which the majority of risk
estimates are derived); (2) spline models that allow data to fit possibly non-linear
relationship; and (3) threshold models, using PMio data in 20 cities from the 1987-1994
NMMAPS data. They reported that the spline model, combined across the cities, showed a
linear relation without indicating a threshold for the relative risks of death for all-causes
and for cardiovascular-respiratory causes in relation to PMio, but "the other cause" deaths
(i.e., all cause minus cardiovascular-respiratory) showed an apparent threshold at around
50 |Jg/m3 PMio, as shown in Figure 6"35. For all-cause and cardkrrespiratory deaths, based
on the Akaike's Information Criterion (AIC), a log-linear model without threshold was
preferred to the threshold model and to the spline model.
The HEI review committee commented that interpretation of these results required
caution, because (l) the measurement error could obscure any threshold! (2) the
city-specific concentration-response curves exhibited a variety of shapes! and (3) the use of
AIC to choose among the models might not be appropriate due to the fact it was not
designed to assess scientific theories of etiology. Note, however, that there has been no
etiologically credible reason suggested thus far to choose one model over others for
aggregate outcomes. Thus, at least statistically, the result of Daniels et al. (2004, 087343)
suggests that the log-linear model is appropriate in describing the relationship between
PMio and mortality.
July 2009
6-313
DRAFT-DO NOT CITE OR QUOTE

-------
o
§
CVDRESP Mortality
Total Mortality
Other Cause Mortality
'•ssussiissiiee
PM,0 {pg/ffl3}
SSSS^SSSBES
PM„ Wnfl
9SS8g«S 888888 S
PM,jnm
Source: Daniels et al. (2004,087343)
Figure 6-35. Concentration-response curves (spline model) for all-cause, cardiovascularrespiratory
and other cause mortality from the 20 NMMAPS cities. Estimates are posterior means
under Bayesian random effects model. Solid line is mean lag, triangles are lag 0 (current
day), and squares are lag 1 (previous day).
The Schwartz (2004, 078998) analysis of PMio and mortality in 14 U.S. cities,
described in Section 6.5.2.1., also examined the shape of the concentration-response
relationship by including indicator variables for days when concentrations were between 15
and 25 |ug/m3, between 25 and 34 |u.g/m3, between 35 and 44 |u,g/m3, and 45 |ug/m3 and above.
In the model, days with concentrations below 15 ug/ma served as the reference level. This
model was fit using the single stage method, combining strata across all cities in the
case-crossover design. Figure 6"36 shows the resulting relationship, which does not provide
sufficient evidence to suggest that a threshold exists. The authors did not examine
cityto-city variation in the concentration-response relationship in this study.
July 2009
6-314
DRAFT-DO NOT CITE OR QUOTE

-------
2.5 r-
~
_« 2.0
= 1.5
8
~
~
O 1.0
~
0.0 h A
10
20	30	40	50	60
PM10 N/m3)
Source: Schwartz (2004,078998)
Figure 6-36. Percent increase in the risk death on days with PM10 concentrations in the ranges of
15-24, 25-34, 35-44, and 45 //g/m3 and greater, compared to a reference of days when
concentrations were below 15//g/m3. Risk is plotted against the mean PM10
concentration within each category.
Samoli et al. (2005, 087436) investigated the concentration-response relationship
between PMio and mortality in 22 European cities (and BS in 15 of the cities) participating
in the APHEA project. In nine of the 22 cities, PMio levels were estimated using a
regression model relating co-located PMio to BS or TSP They used regression spline models
with two knots (30 and 50 |ug/m3) and then combined the individual city estimates of the
splines across cities. The investigators concluded that the association between PM and
mortality in these cities could be adequately estimated using the log-linear model. However,
in an ancillary analysis of the concentration-response curves for the largest cities in each of
the three distinct geographic areas (western, southern, and eastern European cities):
London, England; Athens, Greece! and Cracow, Poland, Samoli et al. (2005, 087436)
observed a difference in the shape of the concentration-response curve across cities. Thus,
while the combined curves (Figure 6"37) appear to support no-threshold relationships
between PMio and mortality, the heterogeneity of the shapes across cities makes it difficult
to interpret the biological relevance of the shape of the combined curves.
July 2009	6-315	DRAFT - DO NOT CITE OR QUOTE

-------
t/i
ec
07
0
V)

-------
2.7% for respiratory-related mortality, per 10 (Jg/m3 increase in PM2.5 (U.S. EPA, 2004,
056905).
In the current review, PM2.5 risk estimates were found to be consistently positive, and
slightly larger than those reported for PM10 for all-cause, and respiratory- and
cardiovascular-related mortality. The risk estimates for all-cause (non-accidental) mortality
ranged from 0.29% (Dominici et al., 2007, 099135) to 1.21% (Franklin et al., 2007, 091257)
per 10 |ug/m3 increase in PM2.5. These associations were consistently observed at lag 1 and
lag 0*1, which have been confirmed through extensive analyses in PMicrmortality studies.
Cardiovascular-related mortality risk estimates were found to be similar to those for
all-cause mortality! whereas, the risk estimates for respiratory-related mortality were
consistently larger: 1.01% (Franklin et al., 2007, 091257) to 2.2% (Ostro et al., 2006,
087991) using the same lag (i.e., lag 1 and lag 0-1) and averaging indices. Regional and
seasonal patterns in PM2.5 risk estimates were observed with results similar to those
presented for PM10 (Dominici et al., 2007, 099135; Peng et al., 2005, 087463; Zeka et al.,
2006, 088749), with the greatest effects occurring in the eastern U.S. (Franklin et al., 2007,
091257; Franklin et al., 2008, 097426) and during the spring (Franklin et al., 2007, 091257;
Zanobetti and Schwartz, 2009, 188462). Of the studies evaluated, no U.S.-based multi-city
studies conducted a detailed analysis of the potential confounding of PM2.5 risk estimates by
gaseous pollutants, but Burnett et al. (2004, 086247) found mixed results, with possible
confounding by NO2 when analyzing gaseous pollutants in a multi-city Canadian-based
study. However, it should be noted that U.S.-based multi-city studies that focused on the
association between PM10 and mortality found that gaseous pollutants are not likely to
confound the PM-mortality relationship. An examination of effect modifiers (e.g.,
demographic and socioeconomic factors), specifically air conditioning use which is
sometimes used as a surrogate for decreased pollutant penetration indoors, has suggested
that PM2.5 risk estimates increase as the percent of the population with access to air
conditioning decreases (Franklin et al., 2007, 091257; Franklin et al., 2008, 097426).
Collectively, the epidemiologic evidence on the effect of short-term exposure to PM2.5 on
mortality is sufficient to conclude that a causal relationship is likely to exist at ambient
concentrations.
July 2009
6-317
DRAFT-DO NOT CITE OR QUOTE

-------
6.5.3.2. PMio-2.5
The 2004 PM AQCD (U.S. EPA, 2004, 056905) found a limited body of evidence that
was suggestive of associations between short-term exposure to ambient PM10-2.5 and various
mortality outcomes (e.g., 0.08 to 2.4% increase in total [non-accidental] mortality per
10 (Jg/m3 increase in PM10-2.5). As a result, the AQCD concluded that PM10-2.5, or some
constituent component(s) (including those on the surface) of PM10-2.5, may contribute, in
certain circumstances, to increased human health risks.
The majority of studies evaluated in this review that examined mortality associations
with PM10-2.5 reported mixed results in terms of the relative impact of PM10-2.5 on mortality.
These studies were conducted in areas with mean 24-h avg concentrations ranging from
6.1-16.4 (Jg/m3. This can primarily be attributed to the majority of studies being conducted
in individual cities and the cityto-city variation in the chemical composition of PM10-2.5. A
new study of 47 U.S. cities (Zanobetti and Schwartz, 2009, 188462), which estimated
PM10-2.5 by calculating the difference between the county-average PM10 and PM2.5, found
associations between PM10-2.5 and mortality across the U.S., including regions where
PM10-2.5 levels are not high. One well conducted multicity Canadian study (Burnett et al.,
2004, 086247) also provides evidence for an association between short-term exposure to
PM10-2.5 and mortality. However, unlike PM2.5 not many of the PM10-2.5 studies have
investigated confounding by gaseous copollutants or the influence of model specification on
PM10-2.5 risk estimates. Zanobetti and Schwartz (2009, 188462) did provide preliminary
evidence for greater effects during the warmer months (i.e., spring and summer), which is
consistent with the PMio-mortality studies (Peng et al., 2005, 087463; Zeka et al., 2006,
088749). Overall, although more data is needed to adequately characterize the chemical
and biological components that may modify the potential toxicity of PM10-2.5, consistent
positive associations between short-term exposure to PM10-2.5 and mortality were observed
in the U.S. and Canadian-based multicity studies, as well as in single-city studies
conducted in these locations. Therefore, the epidemiologic evidence on the effect of
short-term exposure to PM10 2.5 on mortality is suggestive of a causal relationship at ambient
concentrations.
July 2009
6-318
DRAFT-DO NOT CITE OR QUOTE

-------
6.5.3.3. Ultrafine PM
Limited evidence was available during the review of the 2004 PM AQCD (U.S. EPA,
2004, 056905) regarding the potential association between UFP and mortality. The lone
study evaluated was conducted in Germany and provided some evidence for an association,
but this association was reduced upon the inclusion of gaseous copollutants in a two-
pollutant model.
Only a few new studies, all of them from Europe, were identified during this review,
which examined the association between short-term exposure to UFP and mortality.
Inconsistencies were observed in the lag structure of association reported by each study in
terms of both the lag day with the greatest association and the number of lag days
considered in the study. Overall the studies consistently found that UFP were correlated
with gaseous pollutants derived from local combustion sources and that one or more of the
gaseous pollutants were also associated with mortality. The limited number of studies
available and the discrepancy in results between studies further confirms the need for
additional data to examine the UFP-mortality relationship. In conclusion, the epidemiologic
evidence on the effect of short-term exposure to UFP on mortality is inadequate to infer a
causal association at ambient concentrations.
6.6. Attribution of Ambient PM Health Effects to Specific
Constituents or Sources
From a mechanistic perspective, it is highly plausible that the chemical composition of
PM would be a better predictor of health effects than particle size. This would be consistent
with observed regional heterogeneity in PM-related health effects in some epidemiologic
studies. Also, data from the CSN demonstrate gradients in a number of PM2.5 components,
including EC, OC, nitrate, and S042~ (Section 3.5.1.1). Recent studies in epidemiology,
controlled human exposure, and toxicology have begun using apportionment of ambient PM
into chemical constituents and sources, in an attempt to link them to health outcomes and
endpoints, and thus examine their possible role in health effects.
This section focuses on studies that have assessed effects for a range of PM sources or
components. Some of the studies reviewed earlier in this chapter evaluated the relationship
between specific chemical constituents derived from ambient PM and health effects,
including mortality and cardiovascular and respiratory morbidity. However, many of these
July 2009
6-319
DRAFT-DO NOT CITE OR QUOTE

-------
studies, as well as others only included in the Annexes, only considered one or a small
number of PM constituents. Controlled human exposure and toxicological studies that use a
single compound found in PM rather than ambient PM are also included in this category.
Additionally, studies that presented ambient PM composition and health data without
systematically and explicitly investigating relationships are also excluded from this section.
In contrast, other epidemiologic, controlled human exposure, and toxicological studies took
into consideration a large set of PM constituents (typically minerals, metals, EC, OC, and
ions such as SOr ), aiming to segregate which constituents or groups of constituents may
be responsible for the PM-related health effects observed. This section compiles these latter
studies and provides a review of their findings. The focus of this section is limited to those
studies that use ambient measurements of chemically speciated PM to examine
associations with health outcomes and endpoints. One of the most salient characteristics of
ambient PM speciation data is the large number of constituents that compose PM and the
strength of the correlation between each constituents.
Prior to the 2004 PM AQCD (U.S. EPA, 2004, 056905), only a handful of epidemiologic
studies had attempted to relate specific constituents or sources of ambient PM to health
outcomes without selecting constituents a priori. In this review, approximately 40 new
epidemiologic, controlled human exposure, and toxicological studies explore the health
effects attributed to chemical constituents and sources of ambient PM. The following
summary (Section 6.6.3) provides a synthesis of the findings, including discussions on the
robustness, coherence, and consistency of the results.
6.6.1. Evaluation Approach
Relating a large number of ambient PM constituents with a large number of health
outcomes presents difficulties that are related to both the nature of PM, and the methods of
quantitative analysis. First, the number of constituents that comprise PM is not only large,
but the correlations between them are inherently high. Reducing the correlation between
constituents has been accomplished in most of the recent studies through various forms of
factor analysis, which limits the correlations between constituents by grouping the most
highly correlated ambient PM constituents into less correlated groups or factors. Some
studies identify the resulting groups or factors with named sources of ambient PM, but
many do not draw explicit links between factors and actual sources. The methods for
estimating source contributions to ambient PM are reviewed in Section 3.5.4.
July 2009
6-320
DRAFT-DO NOT CITE OR QUOTE

-------
Most studies, regardless of discipline, were based on data for between 7 and 20
ambient PM constituents, with EC, OC, SO4, and NO3 most commonly measured. Most, but
not all of the studies, reduced the number of ambient PM constituents by grouping them
with various factorization or source apportionment techniques before using a separate
analysis to examine the relationship between the grouped PM constituents and various
health effects. However, not all studies labeled the constituent groupings according to their
presumed source. A few performed these steps simultaneously using Partial Least Squares
(PLS) procedures or Structural Equation Modeling (SEM). A small number of controlled
human exposure and toxicological studies did not apply any kind of grouping to the ambient
PM speciation data.
There are some differences in the type of PM constituent data used in epidemiologic,
controlled human exposure and toxicological studies. In epidemiologic studies, ambient PM
speciation data is obtained from atmospheric monitors! for controlled human exposure and
toxicological studies, the experimental exposure technique used determines the type of PM
data. Thus, all epidemiologic studies relied on monitor data, while all of the controlled
human exposure and the majority of the toxicological studies used CAPs. The remaining
toxicological studies used ambient PM samples collected on filters at various U.S. sites.
Further details on the studies included can be found in Appendix F.
Some important limitations in interpreting these studies together is that few, if any of
the results are easily comparable, due to: differences in the sets of ambient PM constituents
that make up each of the factors! the subjectivity involved in labeling factors as sources! the
numerous potential health effects examined in these studies, including definitive outcomes
(e.g., HAs) as well as physiological alterations (e.g., increased inflammatory response); and
the various statistical methods and analytical approaches used in the studies. There are no
well-established, objective methods for conducting the various forms of factor analysis and
source apportionment, leaving much of the model operation and factor assignment open to
judgment by the individual investigator. For example, the Al/Si factor identified in one
study may differ from the Al/Ca/Fe/Si factor from another study, and the "Resuspended
Soil" factor from a third study.
After factorization or apportionment of the ambient PM data, various methods were
used for analyzing the potential association between ambient PM constituents or sources
and health effects. Except for the studies that used PLS or SEM, controlled human
exposure and toxicological studies all used univariate mixed model regression for every
July 2009
6-321
DRAFT-DO NOT CITE OR QUOTE

-------
identified PM factor or source. A number of toxicological studies followed the univariate
step with multivariate regression for all factors. Epidemiologic studies generally related
short-term exposure to sources with health outcomes through various forms of Poisson
regression; whereas, one long-term exposure study used time-to-event data analysis
(survival analysis) methodology.
6.6.2. Findings
The results that follow are organized by discipline, with epidemiologic studies
followed by controlled human exposure and toxicological studies. This section ends with a
summary table, Table 6-17. Table 6-17 is broken out by PM2.5 sources, and includes those
epidemiologic, controlled human exposure, and in vivo toxicological studies that either
grouped ambient PM2.5 constituents or used tracers for each source. The table does not
include all factors or sources examined in the studies listed: those factors or sources for
which no association with effects was found are excluded.
6.6.2.1. Epidemiologic Studies
Results from the 2004 PM AQCD
Three epidemiologic studies were evaluated in the 2004 PM AQCD (U.S. EPA, 2004,
056905) that examined the association between PM constituents or sources and specific
health effects. Of these studies, one study associated daily mortality with a mobile sources
PM factor in Knoxville, TN and St. Louis, MO and coal in Boston, MA; while the crustal
factor was not found to be significant for any of the six cities studied (Laden et al., 2000,
012102). Another study demonstrated an association between a regional SO42- factor and
total mortality at lag 0 in Phoenix and factors for regional SO42-, motor vehicles, and
vegetative burning with cardiovascular mortality at lags of 0, 1, and 3, respectively (Mar et
al., 2000, 001760; 2003, 042841). Negative associations were observed between total
mortality and regional SO42- at lag 3, along with local SO2 and soil factors (Mar et al., 2000,
001760: 2003, 042841). Finally, Tsai et al. (2000, 006251) identified significant associations
between PMis-derived industrial sources and total daily deaths in Newark and Camden,
NJ; SO r was also linked to cardiopulmonary deaths in both locations. Total mortality and
cardiopulmonary deaths were also significantly associated with PM from oil burning in
Camden (2000, 006251).
July 2009
6-322
DRAFT-DO NOT CITE OR QUOTE

-------
Comparative Analyses of Source Apportionment Methods
Hopke et al. (2006, 088390) conducted a comparative analysis of source
apportionment techniques used by investigators at multiple institutions, and subsequently
used in epidemiologic analyses (Ito et al., 2006, 088391; Mar et al., 2006, 086143). An
overarching conclusion of this set of analyses, reported in Thurston et al. (2005, 097949), is
that variation in the source apportionment methods was not a major source of uncertainty
in the epidemiologic effect estimates. In the primary analyses, mortality was associated
with secondary SOr in both Phoenix and Washington D.C., although lag times differed (0
and 3, respectively). The SO r effect was stronger for total mortality in Washington D.C.
and for cardiovascular mortality in Phoenix (Ito et al., 2006, 088391; Mar et al., 2006,
086143). In addition, Ito et al. (2006, 088391) found some evidence for associations with
primary coal and traffic with total mortality in Washington D.C. (Ito et al., 2006, 088391)
while copper smelter, traffic, and sea salt were associated with cardiovascular mortality in
Phoenix at various lag times (Mar et al., 2006, 086143). In contrast to Phoenix, sea salt and
traffic were not associated with mortality in Washington D.C. (Ito et al., 2006, 088391), but
in both locations no associations were observed between biomass/wood combustion and
mortality (Ito et al., 2006, 088391; Mar et al., 2006, 086143). In an additional study that
compared three source apportionment methods in Atlanta - PMF, modified CMB, and a
single-species tracer approach - found that the epidemiologic results were robust to the
analytic method used (Sarnat et al., 2008, 097972). There were consistent associations
between ED visits for cardiovascular diseases with PM2.5 from mobile sources (gasoline and
diesel) and biomass combustion (primarily prescribed forest burning and residential wood
combustion), whereas PM2.5 from secondary SO42- was associated with respiratory disease
ED visits (Sarnat et al., 2008, 097972). Sarnat et al. (2008, 097972) also found that the
primary power plant PM2.5 source identified by the CMB approach was negatively
associated with respiratory ED visits while no association was found for PM2.5 from soil and
secondary nitrates/ammonium nitrate. In these studies, effect estimates based on the
different source apportionment methods were generally in close agreement.
Source Apportionment Studies
A study that examined associations with mortality in Santiago, Chile identified a
motor vehicle source of PM2.5 as having the greatest association with total and cardiac
mortality at lag 1 (Cakmak et al., 2009, 191995). There was effect modification by age, with
July 2009
6-323
DRAFT-DO NOT CITE OR QUOTE

-------
the total mortality relative risks associated with PM2.5 from motor vehicles being greatest
for those >85 yr. Soil and combustion sources were also associated with cardiac mortality
Risk estimates for respiratory mortality were the greatest for the motor vehicle source, with
combustion and soil source factors also demonstrating positive associations for lag 1
(Cakmak et al., 2009, 191995).
An epidemiologic study that evaluated respiratory ED visits was conducted in
Spokane, WA and used tracers as indicators of ambient PM2.5 sources (Schreuder et al.,
2006, 097959). In this study, only PM2.5 from vegetative burning (total carbon) was
associated with increased respiratory ED visits for lag 1, while PM2.5 indicators for motor
vehicles (Zn) and soil (Si) were not associated with cardiac hospital or respiratory ED visits.
Andersen et al. (2007, 093201) conducted a source apportionment analysis to identify the
sources of ambient PM10 associated with cardiovascular and respiratory hospital
admissions in older adults and children (ages 5-18) in Copenhagen, including two-pollutant
models with various sources of PM10. Andersen et al. (2007, 093201 Hound that secondary
and crustal sources of PM10 were associated with cardiovascular hospital admissions!
biomass sources were associated with respiratory hospital admissions! and vehicle sources
were associated with asthma hospital admissions.
Several panel epidemiologic studies have examined the association between PM
sources and physiological alterations in cardiovascular function. Lanki et al. (2006, 089788)
reported associations between PM2.5 from local traffic and ST-segment depression in elderly
adults in a study conducted in Helsinki, Finland. In an additional study, Yue et al. (2007,
097968) found that adult males with coronary artery disease in Erfurt, Germany
demonstrated changes in repolarization parameters associated with traffic-related PM2.5,
with increased vWF linked to traffic and combustion-generated particles, although the
source apportionment analysis was based solely on particle size distribution. In addition,
elevated CRP levels were associated with all sources of PM2.5 (soil, local traffic, secondary
aerosols from local fuel combustion, diesel, and secondary aerosols from multiple sources)
(Yue et al., 2007, 097968). Reidiker et al. (2004, 056992), in a study of young highway patrol
officers, found that the most significant effects (HRV, supraventricular ectopic beats,
hematological markers, vWF) were associated with a "speed-change" factor for PM2.5 (2004,
056992). In addition, the authors observed an association between crustal factor and
cardiovascular effects, but no health-related associations with steel wear or gasoline PM2.5
source factors.
July 2009
6-324
DRAFT-DO NOT CITE OR QUOTE

-------
Two recent studies have examined the associations between ambient PM2.5 sources
and respiratory symptoms and lung function. Positive associations with PM2.5 motor vehicle
and road dust sources were reported for respiratory symptoms and inhaler use in asthmatic
children in New Haven, CT, and negative associations with wheeze or inhaler use for
biomass burning at lag 0-2 (Gent et al., 2009, 180399). These positive effects for motor
vehicle and road dust sources were robust to the inclusion of a gaseous copollutant (NO2,
CO, SO2, or O3) in the regression model. Penittinen et al. (2006, 087988) in a study
consisting of asthmatic adults living in Helsinki, Finland, found that decrements in PEF
were associated with ambient PM2.5 soil, long-range transport, and local combustion sources
at lags from 0-5 days. In addition, negative associations with asthma symptoms and
medication use were reported for PM2.5 from sea salt and long-range transport sources
(Penttinen et al., 2006, 087988).
PM Constituent Studies
Some studies considered large sets of ambient PM constituents and attempted to
identify which were associated with various health effects, but without grouping them into
factors, or identifying sources. The majority of these studies focused on health effects
associated with short-term exposure to PM2.5. Peng et al. (2009, 191998) examined the
association between PM2.5 constituents (i.e., EC, OC, SO42", NO3", Si, Na, NH4"1") and
cardiovascular and respiratory hospital admissions in 119 U.S. cities. When including each
constituent in a multipollutant model, Peng et al. (2009, 191998) found that EC and OC
were robust to the inclusion of the other constituents at lag 0 for cardiovascular and
respiratory hospital admissions, respectively. Although this study did not conduct analyses
to identify sources of the constituents examined, EC and OC are often attributed to motor
vehicle emissions, particularly diesel engines, and wood burning (Peng et al., 2009, 191998).
Ostro et al. (2007, 091354; 2008, 097971) conducted two studies in six California counties to
examine the association between ambient PM constituents and mortality. In the 2007
analysis, Ostro et al. (2006, 087991) found associations for Cu and all-cause mortality! EC,
K, and Zn and CVD mortality! and Cu and Ti and respiratory mortality at lags ranging
from 0 to 3 days. Associations during the summer were only observed for K for both CVD
and respiratory mortality! and Al, CI, Cu, Pb, Ti, and Zn for respiratory mortality. Overall,
the most consistent associations were observed during the cool season. In a subsequent
analysis, Ostro et al. (2008, 097971) examined the association between ambient PM
July 2009
6-325
DRAFT-DO NOT CITE OR QUOTE

-------
constituents and cardiovascular mortality in potentially susceptible subpopulations. The
authors found positive associations between EC, OC, NO3", SO42", K, Cu, Fe, and Zn and
cardiovascular mortality These associations were higher in individuals with lower
educational attainment and of Hispanic ethnicity In addition, similar to the 2007 analysis,
associations were observed at lags ranging from 0 to 3 days.
One study long-term exposure to PM2.5 has included evaluation of multiple PM2.5
components and an indicator of traffic density (Lipfert et al., 2006, 088756). Using health
data from a cohort of U.S. military veterans and PM2.5 data from EPA's CSN, Lipfert et al.
(2006, 088756) reported a positive association for mortality with sulfates. Positive
associations were found between mortality and long-term exposures to nitrates, EC, Ni and
V, as well as traffic density and peak O3 concentrations. In multipollutant models,
associations with traffic density remained significant, as did nitrates, Ni and V in some
models.
Evaluation of Effect Modification by PM Constituents
Several studies have conducted secondary analyses to examine the whether variation
in associations between PM2.5 and morbidity or PM10 and mortality reflects differences in
PM2.5 constituents. An assumption in these types of analyses, especially when examining
the effects on PMio-mortality risk estimates, is that the relative contributions of PM2.5 have
remained the same over time, since these studies used PM10 data for years prior to 2000,
while PM2.5 speciation data has only been routinely collected since about 2000. Bell et al.
(2009, 191997) found statistically significant associations between the county average
concentrations of V, Ni, and EC (106 counties) and effect estimates for both cardiovascular
and respiratory hospital admissions with short-term exposure to PM2.5. In this analysis the
authors only focused on those ambient PM2.5 constituents (i.e., NH4"1", EC, OC, NO3", and
SO42") that comprised the majority of PM2.5 total mass in the study locations. Bell et al.
(2009, 191997) also conducted a similar analysis for PMio-mortality risk estimates and
found that only Ni increased the risk estimate. However, in a sensitivity analysis, when
selectively dropping out the communities examined one-by-one, the Ni association was
diminished when removing New York City. Both Lippmann et al. (2006, 091165) and
Dominici et al. (2007, 099135) conducted similar analyses, albeit using a smaller subset of
cities and/or different years of PM10 data. In both studies, Ni and V were found to modify
the PMio-mortality risk estimates. In addition, Dominici et al. (2007, 099135) also found in
July 2009
6-326
DRAFT-DO NOT CITE OR QUOTE

-------
a sensitivity analysis that the association with Ni and V was diminished when excluding
New York City from the analysis.
6.6.2.2.	Controlled Human Exposure Studies
A few controlled human exposure studies employed PCA, although not all linked
groupings of PM constituents to the measured physiological parameters. Huang et al.
(2003, 087377) demonstrated associations between increased fibrinogen and Cu/Zn/V and
increased BALF neutrophils and Fe/Se/SCk in young, healthy adults exposed to RTP, NC
CAPs! however, only water-soluble constituents were analyzed. In the other study that
examined physiological cardiovascular effects, Fe and EC were associated with changes in
ST-segment, while SO r was associated with decreased SBP in asthmatic and healthy
human volunteers exposed to Los Angeles CAPs (2003, 087377). In Gong et al. (2003,
087365) the majority of the PM was in the thoracic coarse fraction. In the other study that
used Los Angeles CAPs, the only observed association was between SO42- content and
decreased lung function (FEVi and FVC) in elderly volunteers with and without COPD
(Gong et al., 2005, 087921). Two additional controlled human exposure studies that did not
perform grouping and employed Toronto CAPs plus O3 demonstrated increased DBP and
increased brachial artery vasoconstriction associated with carbon content (2004, 055629;
Urch et al., 2005, 081080).
6.6.2.3.	Toxicological Studies
The only toxicological in vivo study that characterized PM sources corresponding to
identified sources was conducted in Tuxedo, NY, over a 5-month period. This study reported
that all sources (regional SO42~, resuspended soil, residual oil, traffic and other unknown
sources) were linked to HR or HRV changes in mice at one time or another during or after
daily exposure (Lippmann et al., 2005, 087453). In a similar in vitro study using CAPs from
the same location, NF-kB in BEAS-2B cells were correlated with the oil combustion factor (r
= 0.289 and 0.302 for V and Ni, respectively) (Maciejczyk and Chen, 2005, 087456). The
other in vitro toxicological study (Duvall et al., 2008, 097969) that named sources employed
samples from five U.S. cities and found a good fit for the regression model with increased
IL-8 release in primary human airway epithelium cells and coal combustion (R2 = 0.79),
secondary nitrate (R2 = 0.63), and mobile sources (R2 = 0.39). In addition, soil (R2 = 0.48),
residual oil combustion (R2 = 0.38), and wood combustion (R2 = 0.33) were associated with
July 2009
6-327
DRAFT-DO NOT CITE OR QUOTE

-------
COX-2 effects! whereas, secondary SC>42~ (R2 = 0.51) was correlated with HOI. Wood
combustion and soil were negatively associated with HOI.
Several toxicological studies employed Boston CAPs and identified at least four
groupings of ambient PM2.5 constituents (V/Ni, S, Al/Si, and Br/Pb), but they named sources
only partially and tentatively (Batalha et al., 2002, 088109; Clarke et al., 2000, 011806;
Godleski et al., 2002, 156478; Nikolov et al., 2008, 156808; Saldiva et al., 2002, 025988;
Wellenius et al., 2003, 055691). When examining cardiovascular effects these studies
reported that Si was associated with changes in the ST-segment of dogs (Wellenius et al.,
2003, 055691) and decreased LAV ratio in rat pulmonary arteries (Batalha et al., 2002,
088109) in multivariate analyses. In addition, blood hematological results were associated
with V/Ni, Al/Si, Na/Cl, and S in dogs (Clarke et al., 2000, 011806). An examination of
respiratory effects in the latter study found that V/Ni and Br/Pb were associated with
increased inflammation in BALF for only the 3rd day of exposure (Clarke et al., 2000,
011806). Decreased respiratory rate and increased airway irritation (Penh) in dogs were
associated with road dust (Al) and motor vehicles (OC), respectively (Nikolov et al., 2008,
156808). Individual PM2.5 constituents associated with elevated neutrophils in BALF were
Br, EC, OC, Pb, and SC>42~ (Godleski et al., 2002, 156478), which is consistent with the
findings (Br, EC, OC, Pb, V, and CI) of Saldiva et al. (2002, 025988).
The toxicological studies that used PLS methodologies identified PM2.5 constituents
linked to respiratory parameters. Seagrave et al. (2006, 091291) demonstrated associations
between cytotoxic responses and a gasoline plus nitrates source factor (OC, Pb,
hopanes/steranes, nitrate, and As) along with inflammatory responses and a gasoline plus
diesel source factor (including major metal oxides) in rats exposed IT. In the other study,
Veranth et al. (2006, 087479) collected loose surface soil from 28 sites in the Western U.S.
and exposed BEAS-2B cells to PM2.5. In the multivariate redundancy analysis, OCi, OC3,
OC2, EC2, Br, ECi, and Ni correlated with IL-8 release, decreased IL-6 release, and
decreased viability at low and high doses (10 and 80 |Jg/cm2, respectively).
July 2009
6-328
DRAFT-DO NOT CITE OR QUOTE

-------
Table 6-17. Study-specific PM2.5 factor/source categories associated with health effects.




Typ


Source Category
Location
Health Effects
Time
eol
Stu
dy1
Species
Reference
CRUSTAL/SOIL/ROAD DUST
Al, Si, Fe
Phoenix, AZ
negative association with
total mortality
Lag 2
E
Human
Maretal. (2000,001760)
Not provided
Washington, D.C.
fCV mortality
Lag 4
E
Human
ltoetal.(2006, 088391)
Al, Ca, Fe, Si
Santiago, Chile
fCV mortality
f respiratory mortality
Lag 1
E
Human
Cakmak et al. (2009,191995)
Al, Si, Ca, K, Fe
Helsinki, Finland
ST-segment depression
Lag 3
E
Human
Lanki et al. (2006, 089788)
Al, Si, Ca, K, Fe
Los Angeles, CA
I ST-segment voltage
2 days post-exposure
H
Human
Gong et al. (2003, 042106)
Al, Si
Boston, MA
ST-segment change
Following exposure
T
Dog
Wellenius et al. (2003,055691)
Al, Si, Ca
Boston, MA
I lumen/wall ratio
24 h post-exposure
T
Rat
Batalha et al. (2002, 088109)
Al, Si, Ti, Fe
Wake County, NC
T uric acid
T mean cycle length
Lag 15 h
E
Human
Riediker et al. (2004, 056992)
Al, Si, Ca, Fe
Tuxedo, NY
4 HR
t HR
t SDNN, t RMSSD
During exposure
Afternoon post-exposur
e. Night post-exposure
T
Mouse
Lippmann et al. (2005, 087453)
Al, Si
Boston, MA
t blood PMN %
I blood lymphocytes %
t WBC
Following exposure
T
Dog
Clarke et al.(2000, 011806)
Si, Fe, A I, Ca, Ba, Ti
New Haven, CT
T respiratory symptoms and
inhaler use
Lag 0-2
E
Human
Gent et al. (2009,180399)
Si, Al, Ca, Fe, Mn
Helsinki, Finland
I mean PEF
Lag 3
E
Human
Penttinen et al. (2006, 087988)
Al
Boston, MA
I airway irritation (penh)
During exposure
T
Dog
Nikolovet al. (2008,156808)
SALT
Not provided
Phoenix, AZ
fCV mortality
ftotal mortality
negative association with
total mortality
Lag 5
LagO
E
Human
Maretal. (2006,086143)
Na, CI
Helsinki, Finland
ST-segment depression
Lag 3
E
Human
Lanki et al. (2006, 089788)
Na, CI
Boston, MA
f blood lymphocyte %
Following exposure
T
Dog
Clarke et al.(2000, 011806)
Na, CI
Helsinki, Finland
Negatively associated with
bronchodilator use and
corticosteroid use
Lag 0-5 avg
E
Human
Penttinen et al. (2006, 087988)
Na, CI
Boston, MA
f lung PMN density
24 h post-exposure
T
Rat
Saldiva et al. (2002, 025988)
SECONDARY SO/ /
LONG RANGE TRANSPORT
S
Phoenix, AZ
f total mortality
negative association with
total mortality
LagO
Lag 5
E
Human
Maretal. (2000,001760)
Not provided
Washington, D.C.
f total mortality
Lag 3
E
Human
Itoetal. (2006, 088391)
Not provided
Phoenix, AZ
fCV mortality
Lag 0
E
Human
Maretal. (2006, 086143)
S, K, Zn, Pb
Helsinki, Finland
ST-segment depression
Lag 2
E
Human
Lanki et al. (2006, 089788)
July 2009
6-329
DRAFT-DO NOT CITE OR QUOTE

-------




Typ


Source Category
Location
Health Effects
Time
eof
Stu
dy1
Species
Reference
SO42-
Los Angeles, CA
4 SBP
4 h post exposure
H
Human
Gongetal. (2003, 042106)
S, Si, oc
Tuxedo, NY
4 HR
| SDNN, | RMSSD
Afternoon post-
exposure
Night post-exposure
T
Mouse
Lippmann et al. (2005, 087453)
s
Boston, MA
| RBC
| hemoglobin
Following exposure
T
Dog
Clarke et al. (2000,011806)
SO42-, NH4+, oc
Atlanta, GA
f respiratory ED visits
Lag 0
E
Human
Sarnat et al. (2008, 097972)
S, K, Zn, PM mass
Helsinki, Finland
J, mean PEF. Negative
association with asthma
symptom prevalence
Lag 1
Lag 3
E
Human
Penttinen et al. (2006, 087988)
SO42 (+NO2)
Los Angeles, CA
| FEVi
| FVC
Following exposure
H
Human
Gongetal. (2005, 087921)
TRAFFIC
Pb, Br, Cu
Harvard Six Cities
f total mortality
Lag 0-1
E
Human
Laden et al. (2000,012102)
Not provided
Phoenix, AZ
fCV mortality
Lag 1
E
Human
Maretal. (2006, 086143)
Mn, Fe, Zn, Pb, OC, EC,
CO.NO?
Phoenix, AZ
f CV mortality
Lag 1
E
Human
Maret al. (2000, 001760)
CO, NO?, EC, OC
Santiago, Chile
fCV mortality
f respiratory mortality
Lag 1
E
Human
Cakmak et al. (2009,191995)
Gasoline (OC, NO3', NH4+)
Atlanta, GA
t CVD ED visits
Lag 0
E
Human
Sarnat et al. (2008, 097972)
Diesel (EC, OC, NO3)
Atlanta, GA
t CVD ED visits
Lag 0
E
Human
Sarnat et al. (2008, 097972)
NOx, EC, ultrafine count
Helsinki, Finland
ST-segment depression
Lag 2
E
Human
Lanki et al. (2006, 089788)
Speed-change factor (Cu,
S, aldehydes)
Wake County, NC
f blood urea nitrogen
f mean red cell volume
t blood PMN %
J, blood lymphocytes %
f von Willebrand factor
(vWF)
4 protein C
f mean cycle length
t SDNN
t PNN50
| supraventricular ectopic
beats
Lag 15 h
E
Human
Riediker et al. (2004, 056992)
Motor vehicle/other (Br,
Pb, Se, Zn, N03-)
Tuxedo, NY
4 RMSSD
Afternoon post-
exposure
T
Mouse
Lippmann et al. (2005, 087453)
EC, Zn, Pb, Cu, Se
New Haven, CT
f respiratory symptoms
Lag 0-2
E
Human
Gent et al. (2009,180399)
Local combustion (NOx,
ultrafine PM, Cu, Zn, Mn,
Fe)
Helsinki, Finland
4 mean PEF
Lag 0-5 avg
E
Human
Penttinen et al. (2006, 087988)
Gasoline+secondary
nitrate*
Birmingham, AL;
Atlanta, GA;
Pensacola, FL;
Centreville, AL
cytotoxic responses
(potency)
24 h post-exposure
T
Rat
Seagrave et al. (2006, 091291)
Gasoline-t-diesel*
Birmingham, AL;
Atlanta, GA;
Pensacola, FL;
Centreville, AL
inflammatory responses
(potency)
24 h post-exposure
T
Rat
Seagrave et al. (2006, 091291)
July 2009
6-330
DRAFT-DO NOT CITE OR QUOTE

-------




Typ


Source Category
Location
Health Effects
Time
eof
Stu
dy1
Species
Reference
OIL COMBUSTION
V, Ni
Boston, MA
t blood PMN %
J, blood lymphocytes %
f BALF AM %
Following exposure
Following exposure
24 h post exposure
T
Dog
Clarke et al. (2000,011806)
V, Ni, Se
Tuxedo, NY
4 SDNN
4 RMSSD
Afternoon post-
exposure
T
Mouse
Lippmann et al. (2005, 087453)
Ni
Boston, MA
J, respiratory rate
During exposure
T
Dog
Nikolov et al. (2008,156808)
V, Ni
Boston, MA
f lung PMN density
24 h post-exposure
T
Rat
Saldiva et al. (2002, 025988)
COAL COMBUSTION
Se, S042-
Harvard Six Cities
f total mortality
Lag 0-1
E
Human
Laden et al. (2000,012102)
Not provided
Washington, D.C.
f total mortality
Lag 3
E
Human
Itoetal. (2006, 088391)
OTHER METALS
Cu smelter (not provided)
Phoenix, AZ
fCV mortality
ftotal mortality
LagO
E
Human
Maretal. (2006,086143)
Incinerator
Washington, D.C.
Negative association with
total and CV mortality
LagO
E
Human
Itoetal. (2006,088391)
Metal processing (SO42~,
Fe, NH4+, EC, 0C)
Atlanta, GA
t CVD ED visits
Lag 0
E
Human
Sarnat et al. (2008, 097972)
Combustion (Cr, Cu, Fe,
Mn, Zn)
Santiago, Chile
fCV mortality
f respiratory mortality
Lag 1
E
Human
Cakmak et al. (2009,191995)
WOODSMOKE/
VEGETATIVE BURNING
0C, K
Phoenix, AZ
f CV mortality
Lag 3
E
Human
Maret al. (2000, 001760)
0C,EC,K,NH4+
Atlanta, GA
t CVD ED visits
Lag 0
E
Human
Sarnat et al. (2008, 097972)
Total C
Spokane, WA
f respiratory ED visits
Lag 1
E
Human
Schreuderet al. (2006, 097959)
UNNAMED FACTORS
Zn-Cu-V
Chapel Hill, NC
f blood fibrinogen
18 h post-exposure
H
Human
Huang et al. (2003, 087377)
Fe-Se-S042-
Chapel Hill, NC
t BALF PMN
18 h post-exposure
H
Human
Huang et al. (2003, 087377)
Br, CI, Pb
Santiago, Chile
fCV mortality
f respiratory mortality
Lag 1
E
Human
Cakmak et al. (2009,191995)
Br, Pb
Boston, MA
t BALF PMN %
24 h post-exposure
T
Dog
Clarke et al.(2000, 011806)
Br, Pb
Boston, MA
f lung PMN density
24 h post-exposure
T
Rat
Saldiva et al. (2002, 025988)
"Constituents not provided.
1 E = Epidemiologic study; h = Controlled human exposure study; T = Toxicological study
An in vitro toxicological study that employed Chapel Hill PMio used PCAbut did not
name specific PM sources (Becker et al., 2005, 088590). In this study, the release of IL-6
from human alveolar macrophages and 11/8 from normal human bronchial epithelial cells
was associated with a PMio factor comprised of Cr, Al, Si, Ti, Fe, and Cu. No statistically
significant effects were observed for a second PMio factor (Zn, As, V, Ni, Pb, and Se).
July 2009
6-331
DRAFT-DO NOT CITE OR QUOTE

-------
Those toxicological studies that did not apply groupings to the ambient PM2.5
speciation data demonstrated a variety of results. Two Boston CAPs studies demonstrated
lung oxidative stress correlated with a number of individual PM2.5 constituents including,
Mn, Zn, Fe, Cu, and Ca (Gurgueira et al., 2002, 036535) (Gurgueira et al., 2002, 036535)
and Al, Si, Fe, K, Pb, and Cu (Rhoden et al., 2004, 087969) in rats using univariate
regression. The remaining toxicological study that did not use ambient PM constituent
groupings reported a correlation between Zn and plasma fibrinogen in SH rats when
constituents were normalized per unit mass of CAPs (Kodavanti et al., 2002, 035344).
6.6.3. Summary by Health Effects
Recent epidemiologic, toxicological, and controlled human exposure studies have
evaluated the health effects associated with ambient PM constituents and sources, using a
variety of quantitative methods applied to a broad set of PM constituents, rather than
selecting a few constituents a priori. As shown in Table 6-17, numerous ambient PM2.5
source categories have been associated with health effects, including factors for PM from
crustal and soil, traffic, secondary SO42-, power plants, and oil combustion sources. There is
some evidence for trends and patterns that link particular ambient PM constituents or
sources with specific health outcomes, but there is insufficient evidence to determine
whether these patterns are consistent or robust.
For cardiovascular effects, multiple outcomes have been linked to a PM
crustal/soil/road dust source, including cardiovascular mortality in Washington D.C. (Ito et
al., 2006, 08839 Dand Santiago, Chile (Cakmak et al., 2009, 191995) and ST-segment
changes in Helsinki (Lanki et al., 2006, 089788), Los Angeles (Gong et al., 2003, 042106),
and Boston (Wellenius et al., 2003, 055691). Interestingly, the ST-segment changes have
been observed in an epidemiologic panel study, a controlled human exposure study, and a
toxicological study, although the majority of the CAPs in the controlled human exposure
study was PM10-2.5. Further support for a crustal/soil/road dust source associated with
cardiovascular health effects comes from a PM10 source apportionment study in
Copenhagen that reported increased cardiovascular hospital admissions (Andersen et al.,
2007, 093201).
PM2.5 traffic and WS/vegetative burning sources have also been linked to
cardiovascular effects. Cardiovascular mortality in Phoenix (Mar et al., 2000, 001760; Mar
et al., 2006, 086143) and Santiago, Chile (Cakmak et al., 2009, 191995) was associated with
July 2009
6-332
DRAFT-DO NOT CITE OR QUOTE

-------
traffic at lag 1. Gasoline and diesel sources were associated with ED visits in Atlanta for
cardiovascular disease at lag 0 (Sarnat et al., 2008, 097972). Cardiovascular mortality in
Phoenix (Mar et al., 2000, 001760) and ED visits in Atlanta (Sarnat et al., 2008, 097972)
were associated with WS/vegetative burning.
Studies that only examined the effects of individual PM2.5 constituents linked EC to
cardiovascular hospital admissions in a multicity analysis (Peng et al., 2009, 191998) and
cardiovascular mortality in California (Ostro et al., 2007, 091354; 2008, 097971).
These studies suggest that cardiovascular effects maybe associated with PIVL n from
motor vehicle emissions, wood or biomass burning, and PM (both PM2.5 and PM10-2.5) from
crustal or road dust sources. In addition, there are many studies that observed associations
between other sources (i.e., salt, secondary S042~/long-range transport, other metals) and
cardiovascular effects, but at this time, there does not appear to be a consistent trend or
pattern of effects for those factors.
There is less consistency in observed associations between PM sources and
respiratory health effects, which may be partially due to the fact that fewer studies have
been conducted that evaluated respiratory-related outcomes and measures. However, there
is some evidence for associations with secondary SC>42~ PM2.5. Sarnat et al. (2008, 097972)
found an increase in respiratory ED visits in Atlanta that was associated with a PM2.5
secondary SC>42~ factor. Decrements in lung function in Helsinki (Lanki et al., 2006, 089788)
and Los Angeles (Gong et al., 2005, 087921) in asthmatic and healthy adults, respectively,
were also linked to this factor. Health effects relating to the crustal/soil/road dust and
traffic sources of PM included increased respiratory symptoms in asthmatic children (Gent
et al., 2009, 180399) and decreased PEF in asthmatic adults (Penttinen et al., 2006,
087988). Inconsistent results were also observed in those PM2.5 studies that use individual
constituents to examine associations with respiratory morbidity and mortality, although
Cu, Pb, OC, and Zn were related to respiratory health effects in two or more studies.
A few studies have identified PM2.5 sources associated with total mortality. These
studies found an association between mortality and a PM2.5 coal combustion factor (Laden
et al., 2000, 012102). while others linked mortality to a secondary SO r /long-range
transport PM2.5 source (Ito et al., 2006, 088391; Mar et al., 2006, 086143). One long-term
exposure study reported an association between EC in PM2.5 and total mortality, and some
evidence for associations with Ni and V, in a cohort of U.S. veterans (Lipfert et al., 2006,
088756).
July 2009
6-333
DRAFT-DO NOT CITE OR QUOTE

-------
Recent studies have evaluated whether epidemiologic effect estimates from multicity
studies were related to the specific PM2.5 constituents in the cities (Bell et al., 2009, 191997;
Dominici et al., 2007, 099135; Lippmann et al., 2006, 091165). In all three studies, PMicr
mortality effect estimates were greater in areas with higher proportion of Ni in PM2.5, but
when New York City was excluded in a sensitivity analysis in two of the studies, the
association was diminished. A relationship was also found between V concentration in PM2.5
for PMicrmortality effect estimates as well as those for PM2.5 with respiratory and
cardiovascular hospital admissions (Bell et al., 2009, 191997).
Recent studies show that source apportionment methods have the potential to add
useful insights into which sources and/or PM constituents may contribute to different
health effects. Of particular interest are several epidemiologic studies that compared source
apportionment methods and the associated results. One set of studies compared
epidemiologic associations with PM2.5 source factors using several methods - PCA, PMF,
and UNMIX - independently analyzed by separate research groups (Hopke et al., 2006,
088390; Ito et al., 2006, 088391; Mar et al., 2006, 086143; Thurston et al., 2005, 097949).
Schreuder et al. (2006, 097959) compared UPM and two versions of UNMIX to derive
tracers and Sarnat et al. (2008, 097972) compared PMF, modified CMB, and a single-species
tracer approach. In all analyses, epidemiologic results based on the different methods were
generally in close agreement. The variation in risk estimates for daily mortality between
source categories was significantly larger than the variation between research groups (Ito
et al., 2006, 088391; Mar et al., 2006, 086143; Thurston et al., 2005, 097949). Additionally,
the variation in risk estimates based on the source apportionment model used had a much
smaller effect than the variation caused by the different source constituents. Further, the
most strongly associated source types were consistent across all groups. This supports the
general validity of such approaches, though integration of results would be simpler if the
methods employed for grouping PM constituents were more consistent across studies and
disciplines. Further research would aid understanding of the contribution of different
factors, sources, or source tracers of PM to health effects by increasing the number of
locations where similar health endpoints or outcomes are examined.
In summary, these findings are consistent with the conclusions of the 2004 PM AQCD
(U.S. EPA, 2004, 056905), that a number of source types, including motor vehicle emissions,
coal combustion, oil burning, and vegetative burning, are associated with health effects
(U.S. EPA, 2004, 056905). Although the crustal factor of fine particles was not associated
July 2009
6-334
DRAFT-DO NOT CITE OR QUOTE

-------
with mortality in the 2004 PM AQCD (U.S. EPA, 2004, 056905), recent studies have
suggested that PM (both PM2.5 and PM10-2.5) from crustal, soil or road dust sources or PM
tracers linked to these sources are associated with cardiovascular effects. In addition,
secondary SC>42~ PM2.5 has been associated with both cardiovascular and respiratory effects.
Overall, the results displayed in Table 6-17 indicate that many constituents of PM can be
linked with differing health effects and the evidence is not yet sufficient to allow
differentiation of those constituents or sources that are more closely related to specific
health outcomes.
July 2009
6-335
DRAFT-DO NOT CITE OR QUOTE

-------
Chapter 6 References
Adamkiewicz Gi Ebelt Si Syring Mi Slater Ji Speizer FEi Schwartz Ji Suh Hi Gold DR. (2004). Association
between air pollution exposure and exhaled nitric oxide in an elderly population. Thorax, 59- 204-209.
087925
Adar SDi Adamkiewicz Gi Gold DRi Schwartz Ji Coull BAi Suh H. (2007). Ambient and microenvironmental
particles and exhaled nitric oxide before and after a group bus trip. Environ Health Perspect, 115- 507"
12.098635
Adar SDi Gold DRi Coull BAi Schwartz Ji Stone PHi Suh H. (2007). Focused exposures to airborne traffic
particles and heart rate variability in the elderly. Epidemiology, 18- 95-103. 001458
Aekplakorn Wi Loomis Di Vichit-Vadakan Ni Shy Ci Plungchuchon S. (2003). Acute effects of S02 and particles
from a power plant on respiratory symptoms of children, Thailand. , 34- 906-914. 089908
Alarie Y. (1973). Sensory irritation by airborne chemicals. , 2- 299-363. 191136
Albert CMi Rosenthal Li Calkins Hi Steinberg JSi Ruskin JNi Wang Pi Muller JEi Mittleman MA. (2007).
Driving and Implantable Cardioverter-Defibrillator Shocks for Ventricular Arrhythmias' Results From
the TOVA Study. J Am Coll Cardiol, 50: 2233-2240. 156201
Alexis NEi Lay JCi Zeman Ki Bennett WEi Peden DBi Soukup JMi Devlin RBi Becker S. (2006). Biological
material on inhaled coarse fraction particulate matter activates airway phagocytes in vivo in healthy
volunteers. J Allergy Clin Immunol, 117: 1396-1403. 154323
Allen RWi Mar Ti Koenig J; Liu LJi Gould Ti Simpson Ci Larson T. (2008). Changes in lung function and airway
inflammation among asthmatic children residing in a woodsmoke-impacted urban area. Inhal Toxicol, 20:
423-433. 156208
Analitis Ai Katsouyanni Ki Dimakopoulou Ki Samoli Ei Nikoloulopoulos AKi Petasakis Yi Touloumi Gi
Schwartz Ji Anderson HRi Cambra Ki Forastiere Fi Zmirou Di Vonk JMi Clancy Li Kriz Bi Bobvos Ji
Pekkanen J. (2006). Short-term effects of ambient particles on cardiovascular and respiratory mortality. ,
17: 230-233. 088177
Andersen ZJi Wahlin Pi Raaschou-Nielsen Oi Ketzel Mi Scheike Ti Loft S. (2008). Size distribution and total
number concentration of ultrafine and accumulation mode particles and hospital admissions in children
and the elderly in Copenhagen, Denmark. , 65: 458-66. 189651
Andersen ZJi Wahlin Pi Raaschou-Nielsen Oi Scheike Ti Loft S. (2007). Ambient particle source apportionment
and daily hospital admissions among children and elderly in Copenhagen. 625-636. 093201
Anderson HRi Bremner SAi Atkinson RWi Harrison RMi Walters S. (2001). Particulate matter and daily
mortality and hospital admissions in the west midlands conurbation of the United Kingdom: associations
with fine and coarse particles, black smoke and sulphate. Occup Environ Med, 58: 504-510. 017033
Anderson MEi Bogdan GM. (2007). Environments, indoor air quality, and children. Pediatr Clin North Am, 54:
295-307, viii. 156214
Anselme Fi Loriot Si Henry J-Pi Dionnet Fi Napoleoni J-Gi Thnillez Ci Morin J-P. (2007). Inhalation of diluted
diesel engine emission impacts heart rate variability and arrhythmia occurrence in a rat model of chronic
ischemic heart failure. Arch Toxicol, 81: 299-307. 097084
Antonini JMi Roberts JRi Jernigan MRi Yang H"Mi Ma JYCi Clarke RW. (2002). Residual oil fly ash increases
the susceptibility to infection and severely damages the lungs after pulmonary challenge with a bacterial
pathogen. Toxicol Sci, 70: 110-119. 035342
Antonini JMi Taylor MDi Leonard SSi Lawryk NJi Shi Xi Clarke RWi Roberts JR. (2004). Metal composition
and solubility determine lung toxicity induced by residual oil fly ash collected from different sites within
a power plant. Mol Cell Biochem, 255: 257-265. 097199
Arantes-Costa Fi Lopes Fi Toledo Ai Magliarelli-Filho Pi Moriya Hi Carvalho-Oliveira Ri Mauad Ti Saldiva Pi
Martins M. (2008). Effects of residual oil fly ash (ROFA) in mice with chronic allergic pulmonary
inflammation. Toxicol Pathol, 36: 680. 187137
Arbex MAi Martins LCi De Oliveira RCi Pereira LAi Arbex FFi Cancado JEi Saldiva PHi Braga AL. (2007). Air
pollution from biomass burning and asthma hospital admissions in a sugar cane plantation area in
Brazil. J Epidemiol Community Health, 61: 395-400. 091637
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health
and Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in
the process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009	6-336	DRAFT - DO NOT CITE OR QUOTE

-------
Arena VC; Mazumdar Si Zborowski JVi Talbott EOi He Si Chuang YHi Schwerha JJ. (2006). A retrospective
investigation of PM10 in ambient air and cardiopulmonary hospital admissions in Allegheny County,
Pennsylvania^ 1995-2000. J Occup Environ Med, 48: 38-47. 088631
Atiga WLi Calkins Hi Lawrence JHi Tomaselli GFi Smith JM; Berger RD. (1998). Beat-to-beat repolarization
lability identifies patients at risk for sudden cardiac death. , 9: 899-908. 156231
Atkinson RW; Anderson HR; Sunyer J; Ayres J; Baccini Mi Vonk JM; Boumghar Ai Forastiere Fi Forsberg Bi
Touloumi Gi Schwartz J; Katsouyanni K. (2001). Acute effects of particulate air pollution on respiratory
admissions^ results from APHEA 2 project. Am J Respir Crit Care Med, 164: 1860-1866. 021959
Atkinson RW! Anderson HR; Sunyer J; Ayres J; Baccini M> Vonk JM; Boumghar A; Forastiere F> Forsberg Bi
Touloumi Gi Schwartz J; Katsouyanni K. (2003). Acute effects of particulate air pollution on respiratory
admissions. 042797
Atkinson RW; Bremner SA; Anderson HR; Strachan DP; Bland JM; Ponce de Leon A. (1999). Short-term
associations between emergency hospital admissions for respiratory and cardiovascular disease and
outdoor air pollution in London. Arch Environ Occup Health, 54: 398-411. 007882
Baccarelli A; Cassano P; Litonjua A; Park S; Suh H; Sparrow D; Vokonas P; Schwartz J. (2008). Cardiac
autonomic dysfunction: effects from particulate air pollution and protection by dietary methyl nutrients
and metabolic polymorphisms. , 117: 1802. 191959
Baccarelli A; Martinelli I; Zanobetti A; Grillo P; Hou LF; Bertazzi PA; Mannucci PM; Schwartz J. (2008).
Exposure to Particulate Air Pollution and Risk of Deep Vein Thrombosis. Arch Intern Med, 168: 920-927.
157984
Baccarelli A; Zanobetti A; Martinelli I; Grillo P; Hou L; Giacomini S; Bonzini M; Lanzani G; Mannucci PM;
Bertazzi PA; Schwartz J. (2007). Effects of exposure to air pollution on blood coagulation. J Thromb
Haemost, 5: 252-260. 090733
Baccarelli A; Zanobetti A; Martinelli I; Grillo P; Hou L; Lanzani G; Mannucci PM; Bertazzi PA; Schwartz J.
(2007). Air pollution, smoking, and plasma homocysteine. Environ Health Perspect, 115: 176-181. 091310
Bagate K; Meiring JJ; Gerlofs-Nijland ME; Vincent R; Cassee FR; Borm PJA. (2004). Vascular effects of
ambient particulate matter instillation in spontaneous hypertensive rats. Toxicol Appl Pharmacol, 197:
29-39. 087945
Ballester F; Rodriguez P; Iniguez C; Saez M; Daponte A; Galan I; Taracido M; Arribas F; Bellido J; Cirarda FB;
Canada A; Guillen JJ; Guillen-Grima F; Lopez E; Perez-Hoyos S; Lertxundi A; Toro S. (2006). Air
pollution and cardiovascular admisisons association in Spain: results within the EMECAS project. J
Epidemiol Community Health, 60: 328-336. 088746
Barclay JL; Miller BG; Dick S; Dennekamp M; Ford I; Hillis GS; Ayres JG; Seaton A. (2009). A panel study of
air pollution in subjects with heart failure: negative results in treated patients. Occup Environ Med, 66:
325-334. 179935
Barnett AG; Williams GM; Schwartz J; Best TL; Neller AH; Petroeschevsky AL; Simpson RW. (2006). The
effects of air pollution on hospitalizations for cardiovascular disease in elderly people in Australian and
New Zealand cities. Environ Health Perspect, 114: 1018-1023. 089770
Barnett AG; Williams GM; Schwartz J; Neller AH; Best TL; Petroeschevsky AL; Simpson RW. (2005). Air
pollution and child respiratory health: a case-crossover study in Australia and New Zealand. Am J Respir
Crit Care Med, 171: 1272-1278. 087394
Barraza-Villarreal A; Sunyer J; Hernandez-Cadena L; Escamilla-Nunez MC; Sienra-Monge JJ; Ramirez-Aguilar
M; Cortez-Lugo M; Holguin F; Diaz-Sanchez D; Olin AC; Romieu I. (2008). Air pollution, airway
inflammation, and lung function in a cohort study of Mexico City schoolchildren. Environ Health
Perspect, 116: 832-838. 156254
BarregardL; Sallsten G; Andersson L; Almstrand AC; Gustafson P; Andersson M; Olin AC. (2008).
Experimental exposure to wood smoke: effects on airway inflammation and oxidative stress. Occup
Environ Med, 65: 319-324. 155675
Barregard L; Sallsten G; Gustafson P; Andersson L; Johansson L; Basu S; Stigendal L. (2006). Experimental
exposure to wood-smoke particles in health humans: effects on markers of inflammation, coagulation,
and lipid peroxidation. Inhal Toxicol, 18: 845-853. 091381
Barrett EG; Henson RD; Seilkop SK; McDonald JD; ReedMD. (2006). Effects of hardwood smoke exposure on
allergic airway inflammation in mice. Inhal Toxicol, 18: 33"43. 155677
Bartoli CR; Wellenius GA; Coull BA; Akiyama I; Diaz EA; Lawrence J; Okabe K; Verrier RL; Godleski JJ.
(2009). Concentrated ambient particles alter myocardial blood flow during acute ischemia in conscious
canines. Environ Health Perspect, 117: 333-337. 179904
Bartoli CR; Wellenius GA; Diaz EA; Lawrence J; Coull BA; Akiyama I; Lee LM; Okabe K; Verrier RL; Godleski
JJ. (2009). Mechanisms of Inhaled Fine Particulate Air Pollution-Induced Arterial Blood Pressure
Changes. Environ Health Perspect, 117: 361-366. 156256
July 2009
6-337
DRAFT-DO NOT CITE OR QUOTE

-------
Bartzokas A; Kassomenos P; Petrakis M> Celessides C. (2004). The effect of meteorological and pollution
parameters on the frequency of hospital admissions for cardiovascular and respiratory problems in
Athens. Indoor Built Environ, 13- 271-275. 093252
Bastain TM; Gilliland FD; Li Y_F> Saxon A; Diaz-Sanchez D. (2003). Intraindividual reproducibility of nasal
allergic responses to diesel exhaust particles indicates a susceptible phenotype. , 109- 130-136. 098690
Batalha JR; Saldiva P H; Clarke RW! Coull BA; Stearns RC; Lawrence J; Murthy GG; Koutrakis P; Godleski JJ.
(2002). Concentrated ambient air particles induce vasoconstriction of small pulmonary arteries in rats.
Environ Health Perspect, 110: 1191-1197. 088109
Becker Si Mundandhara Si Devlin RBi Madden M. (2005). Regulation of cytokine production in human alveolar
macrophages and airway epithelial cells in response to ambient air pollution particles' further
mechanistic studies. Toxicol Appl Pharmacol, 207: 269-275. 088590
Beckett WSi Chalupa DFi Pauly-Brown Ai Speers DMi Stewart JCi Frampton MWi Utell MJi Huang LSi Cox Ci
Zareba Wi Oberdorster G. (2005). Comparing inhaled ultrafine versus fine zinc oxide particles in healthy
adults - A human inhalation study. Am J Respir Crit Care Med, 171: 1129-1135. 156261
Bedeschi Ei Campari Ci Candela Si Collini Gi Caranci Ni Frasca Gi Galassi Ci Francesca Gi Vigotti MA. (2007).
Urban air pollution and respiratory emergency visits at pediatric unit, Reggio Emilia, Italy. J Toxicol
Environ Health A, 70: 261-265. 090712
Behndig AFi Mudway ISi Brown JLi Stenfors N Helleday R Duggan STi Wilson SJi Boman C Cassee FRi Frew
AJi Kelly FJi Sandstrom T Blomberg A. (2006). Airway antioxidant and inflammatory responses to diesel
exhaust exposure in healthy humans. Eur Respir J, 27: 359-365. 088286
Bell Mi Ebisu Ki Peng Ri Samet J; Dominici F. (2009). Hospital Admissions and Chemical Composition of Fine
Particle Air Pollution. Am J Respir Crit Care Med, 179: 1115-1120. 191997
Bell ML; Dominici F; Ebisu Ki Zeger SL; Samet JM. (2007). Spatial and Temporal Variation in PM2. 5 Chemical
Composition in the United States for Health Effects Studies. , 115: 989. 155683
Bell ML; Ebisu K; Peng RD; Walker J; Samet JM; Zeger SL; Dominic F. (2008). Seasonal and regional short-
term effects of fine particles on hospital admissions in 202 U.S. counties, 1999-2005. Am J Epidemiol,
168: 1301-1310. 156266
Bell ML; Levy JK; Lin Z. (2008). The effect of sandstorms and air pollution on cause-specific hospital admissions
in Taipei, Taiwan. Occup Environ Med, 65: 104-111. 091268
Bennett CM; McKendry IG; Kelly S; Denike K; Koch T. (2006). Impact of the 1998 Gobi dust event on hospital
admissions in the Lower Fraser Valley, British Columbia. Sci Total Environ, 366: 918-925. 088061
Berger A; Zareba W; Schneider A; Ruckerl R; Ibald-Mulli A; Cyrys J; Wichmann HE; Peters A. (2006). Runs of
ventricular and supraventricular tachycardia triggered by air pollution in patients with coronary heart
disease. J Occup Environ Med, 48: 1149-58. 098702
Berger RD; Kasper EE Baughman KL; Marban E; Calkins H; Tomaselli GF. (1997). Beat"to"Beat QT Interval
Variability : Novel Evidence for Repolarization Lability in Ischemic and Nonischemic Dilated
Cardiomyopathy. , 96: 1557-1565. 155688
Blomberg A; Sainsbury C, Rudell B; Frew AJ; Holgate ST; Sandstrom T; Kelly FJ. (1998). Nasal cavity lining
fluid ascorbic acid concentration increases in healthy human volunteers following short term exposure to
diesel exhaust. Free Radic Res, 28: 59"67. 051246
Blomberg A; Tornqvist H; Desmyter L; Deneys V; Hermans C. (2005). Exposure to diesel exhaust nanoparticles
does not induce blood hypercoagulability in an at-risk population. J Thromb Haemost, 3: 2103-2105.
191991
Boezen HM; Vonk JM; Van Der Zee SC; Gerritsen J; Hoek G; Brunekreef B; Schouten JP; Postma DS. (2005).
Susceptibility to air pollution in elderly males and females. Eur Respir J, 25: 1018-1024. 087396
Bosson J; Barath S; Pourazar J; Behndig AF; Sandstrom T; Blomberg A; Adelroth E. (2008). Diesel exhaust
exposure enhances the ozone-induced airway inflammation in healthy humans. Eur Respir J, 31: 1234"
1240. 156287
Bosson J; Pourazar J; Forsberg B; Adelroth E; Sandstrom T; Blomberg A. (2007). Ozone enhances the airway
inflammation initiated by diesel exhaust. Respir Med, 101: 1140-1146. 156286
Bourotte C; Curi-Amarante A"P; Forti M"C; APereira LA; Braga AL; Lotufo PA. (2007). Association between
ionic composition of fine and coarse aerosol soluble fraction and peak expiratory flow of asthmatic
patients in Sao Paulo city (Brazil). Atmos Environ, 41: 2036-2048. 150040
Bouthillier L; Vincent R; Goegan P; Adamson IYR; Bjarnason S; Stewart M; Guenette J; Potvin M;
Kumarathasan P. (1998). Acute effects of inhaled urban particles and ozone: lung morphology,
macrophage activity, and plasma endothelin-1. Am J Pathol, 153: 1873-1884. 087110
Brauner EV; Forchhammer L; Moller P; Simonsen J; Glasius M; Wahlin P; Raaschou-Nielsen O; Loft S. (2007).
Exposure to ultrafine particles from ambient air and oxidative stress-induced DNA damage. , 115: 1177-
82.188507
July 2009
6-338
DRAFT-DO NOT CITE OR QUOTE

-------
Briet Mi Collin Ci Laurent Si Tan Ai Azizi Mi Agharazii Mi Jeunemaitre Xi Alhenc-Gelas Fi Boutouyrie P.
(2007). Endothelial function and chronic exposure to air pollution in normal male subjects. Hypertension,
50: 970-976. 093049
Brook RDi Brook JR; Urch Bi Vincent R; Rajagopalan Si Silverman F. (2002). Inhalation of fine particulate air
pollution and ozone causes acute arterial vasoconstriction in healthy adults. , 105- 1534-1536. 024987
Brauner EV; Mortensen J; Moller Pi Bernard Ai Vinzents Pi Wahlin Pi Glasius Mi Loft S. (2009). Effects of
ambient air particulate exposure on blood-gas barrier permeability andlung function. Inhal Toxicol, 21-
38-47. 191179
Brauner EV; Moller Pi Barregard Li Dragsted LCK Glasius Mi Wahlin Pi Vinzents Pi Raaschou-Nielsen Oi Loft
S. (2008). Exposure to ambient concentrations of particulate air pollution does not influence vascular
function or inflammatory pathways in young healthy individuals. Part Fibre Toxicol, 5- 13. 191966
Burnett RT; Brook J; Dann Ti Delocla Ci Philips Oi Cakmak Si Vincent Ri Goldberg MS; Krewski D. (2000).
Association between particulate- and gas-phase components of urban air pollution and daily mortality in
eight Canadian cities. Inhal Toxicol, 12- 15-39. 010273
Burnett RT; Cakmak Si Brook JR. (1998). The effect of the urban ambient air pollution mix on daily mortality
rates in 11 Canadian cities. Can J Public Health, 89- 152-156. 029505
Burnett RT; Cakmak Si Brook JR.; Krewski D. (1997). The role of particulate size and chemistry in the
association between summertime ambient air pollution and hospitalization for cardiorespiratory
diseases. Environ Health Perspect, 105: 614-620. 084194
Burnett RT; Dales R; Krewski D; Vincent R; Dann T; Brook JR. (1995). Associations between ambient
particulate sulfate and admissions to Ontario hospitals for cardiac and respiratory diseases. Am J
Epidemiol, 142: 15-22. 077226
Burnett RT; Goldberg MS. (2003). Size-fractionated particulate mass and daily mortality in eight Canadian
cities. 042798
Burnett RT; Smith-Doiron M; Stieb D; Cakmak S; Brook JR. (1999). Effects of particulate and gaseous air
pollution on cardiorespiratory hospitalizations. Arch Environ Occup Health, 54: 130-139. 017269
Burnett RT; Stieb D; Brook JR; Cakmak S; Dales R; Raizenne M; Vincent R; Dann T. (2004). Associations
between short-term changes in nitrogen dioxide and mortality in Canadian cities. Arch Environ Occup
Health, 59: 228-236. 086247
Burtscher H. (2005). Physical characterization of particulate emissions from diesel engines: a review. J Aerosol
Sci, 36: 896-932. 155710
Cakmak S; Dales R; Vida C. (2009). Components of particulate air pollution and mortality in Chile. Int J Occup
Environ Health, 15: 152. 191995
Calderon-Garciduenas L; Maronpot RR; Torres-Jardon R; Henriquez-Roldan C; Schoonhoven R; Acuna-Ayala H;
Villarreal-Calderon A; Nakamura J; Fernando R; Reed W; Azzarelli B; Swenberg JA. (2003). DNA
damage in nasal and brain tissues of canines exposed to air pollutants is associated with evidence of
chronic brain inflammation and neurodegeneration. Toxicol Pathol, 31: 524-538. 156316
Calderon-Garciduenas L; Mora-Tiscareno A; Ontiveros E; Gomez-Garza G; Barragan-Mejia G; Broadway J;
Chapman S; Valencia-Salazar G; Jewells V; Maronpot RR; Henriquez-Roldan C; Perez-Guille B; Torres-
Jardon R; Herrit L; Brooks D; Osnaya-Brizuela N; Monroy M. (2008). Air pollution, cognitive deficits and
brain abnormalities: A pilot study with children and dogs. Brain Cognit, 68: 117-127. 156317
Caligiuri G; Levy B; Pernow J; Thoren P; Hansson GK. (1999). Myocardial infarction mediated by endothelin
receptor signaling in hypercholesterolemia mice. Presented at . 156318
Campbell A; Oldham M; Becaria A; Bondy SC; Meacher D; Sioutas C; Misra C; Mendez LB; KLeinman M. (2005).
Particulate matter in polluted air may increase biomarkers of inflammation in mouse brain.
Neurotoxicology, 26: 133-140. 087217
Campen MJ; Babu NS; Helms GA; Pett S; Wernly J; Mehran R; McDonald JD. (2005). Nonparticulate
components of diesel exhaust promote constriction in coronary arteries from ApoE -/- mice. Toxicol Sci,
88: 95-102. 083977
Campen MJ; McDonald JD; Reed MD; Seagrave J. (2006). Fresh gasoline emissions, not paved road dust, alter
cardiac repolarization in ApoE-/- mice. Cardiovasc Toxicol, 6: 199-210. 096879
Carlsten C; Kaufman JD; Trenga CA; Allen J; Peretz A; Sullivan JH. (2008). Thrombotic markers in metabolic
syndrome subjects exposed to diesel exhaust. Inhal Toxicol, 20: 917-921. 156323
Carlsten C; Kaufman Joel D; Peretz A; Trenga Carol A; Sheppard L; Sullivan Jeffrey H. (2007). Coagulation
markers in healthy human subjects exposed to diesel exhaust. , 120: 849-855. 155714
Cassee FR; Boere AJF; Fokkens PHB; Leseman DLAC; Sioutas C; Kooter IM; Dormans JAMA. (2005).
Inhalation of concentrated particulate matter produces pulmonary inflammation and systemic biological
effects in compromised rats. J Toxicol Environ Health A, 68: 773-796. 087962
July 2009
6-339
DRAFT-DO NOT CITE OR QUOTE

-------
Chahine Ti Baccarelli A; Litonjua A; Wright RO; Suh H; Gold DR; Sparrow D; Vokonas P; Schwartz J. (2007).
Particulate air pollution, oxidative stress genes, and heart rate variability in an elderly cohort. Environ
Health Perspect, 115: 1617-1622. 156327
Chan C-C; Chuang K-J; Chien LrC> Chen W-J; Chang W"T. (2006). Urban air pollution and emergency
admissions for cerebrovascular diseases in Taipei, Taiwan. Eur Heart J, 27: 1238-1244. 090193
Chan CC; Chuang KJ; Chen WJ; Chang WT; Lee CT; Peng CM. (2008). Increasing cardiopulmonary emergency
visits by long-range transported Asian dust storms in Taiwan. Environ Res, 106: 393-400. 093297
Chan CC; Chuang KJ; Shiao GM; Lin LY. (2004). Personal exposure to submicrometer particles and heart rate
variability in human subjects. Environ Health Perspect, 112: 1063-1067. 087398
Chang C-C; Hwang J-S; Chan C"C; Wang P-Y; Cheng T-J. (2007). Effects of concentrated ambient particles on
heart rate, blood pressure, and cardiac contractility in spontaneously hypertensive rats during a dust
storm event. Inhal Toxicol, 19: 973-978. 155719
Chang C-C; Hwang J-S; Chan C"C; Wang P-Y; Hu T"H; Cheng T-J. (2004). Effects of concentrated ambient
particles on heart rate, blood pressure, and cardiac contractility in spontaneously hypertensive rats.
Inhal Toxicol, 16: 421-429. 055637
Chang CC; Hwang JS; Chan CC; Wang PY; Hu TH; Cheng TJ. (2005). Effects of concentrated ambient particles
on heart rate variability in spontaneously hypertensive rats. J Occup Health, 47: 471-480. 088662
Chardon B; Lefranc A; Granados D; Gremy I. (2007). Air pollution and doctors' house calls for respiratory
diseases in the greater Paris area (2000-3). Occup Environ Med, 64: 320-4. 091308
Che W; Zhang Z; Zhang H; Wu M; Liang Y; Liu F; Shu Y; Li N. (2007). Compositions and oxidative damage of
condensate, particulate and semivolatile organic compounds from gasoline exhausts. Environ Toxicol
Pharmacol, 24: 11-18. 096460
Checkoway H; Levy D; Sheppard L; Kaufman J; Koenig J; Siscovick D. (2000). A case-crossover analysis of fine
particulate matter air pollution and out-of-hospital sudden cardiac arrest. 015527
Chen C-H; Xirasagar S; Lin H"C. (2006). Seasonality in adult asthma admissions, air pollutant levels, and
climate: a population-based study. J Asthma, 43: 287-292. 087947
Chen J; Schwartz J. (2009). Neurobehavioral effects of ambient air pollution on cognitive performance in US
adults. Neurotoxicology, 30: 231-239. 179945
Chen LC; Hwang JS. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice
IV Characterization of acute and chronic effects of ambient air fine particulate matter exposures on
heart-rate variability. Inhal Toxicol, 17: 209-216. 087218
Chen Y; Yang Q; Krewski D; Burnett RT; Shi Y; McGrail KM. (2005). The effect of coarse ambient particulate
matter on first, second, and overall hospital admissions for respiratory disease among the elderly. Inhal
Toxicol, 17: 649-655. 087555
Chen Y; Yang Q; Krewski D; Shi Y; Burnett RT; McGrail K. (2004). Influence of relatively low level of
particulate air pollution on hospitalization for COPD in elderly people. Inhal Toxicol, 16: 21-25. 087262
Cheng M-F; Tsai S"S; Wu T-N; Chen P-S; Yang C-Y. (2007). Air pollution and hospital admissions for
pneumonia in a tropical city: Kaohsiung, Taiwan. J Toxicol Environ Health A, 70: 2021-6. 093034
Chevalier P; Burri H; Adeleine P; Kirkorian G; Lopez M; Leizorovicz A; Andre-Fouet X; Chapon P; Rubel P;
Touboul P. (2003). QT dynamicity and sudden death after myocardial infarction: results of a long-term
follow-up study. , 14: 227-233. 156338
Chimonas MA; Gessner BD. (2007). Airborne particulate matter from primarily geologic, non-industrial sources
at levels below National Ambient Air Quality Standards is associated with outpatient visits for asthma
and quick-relief medication prescriptions among children less than 20 years old enrolled in Medicaid in
Anchorage, Alaska. Environ Res, 70: 2021-2026. 093261
Chiu H; Tiao M; Ho S; Kuo H; Wu T; Yang C. (2008). Effects of Asian Dust Storm events on hospital admissions
for chronic obstructive pulmonary disease in Taipei, Taiwan. Inhal Toxicol, 20: 777-781. 191989
Chock DP; Winkler SL; Chen C. (2000). A study of the association between daily mortality and ambient air
pollutant concentrations in Pittsburgh, Pennsylvania. J Air Waste Manag Assoc, 50: 1481-1500. 010407
Choi JH; Xu QS; Park SY; Kim JH; Hwang SS; Lee KH; Lee HJ; Hong YC. (2007). Seasonal variation of effect of
air pollution on blood pressure. J Epidemiol Community Health, 61: 314-318. 093196
Chuang K-J; Chan C"C; Chen N"T; Su T-C; Lin L-Y. (2005). Effects of particle size fractions on reducing heart
rate variability in cardiac and hypertensive patients. Environ Health Perspect, 113: 1693-1697. 087989
Chuang K-J; Chan C"C; Su T"C; Lee C-T; Tang C-S. (2007). The effect of urban air pollution on inflammation,
oxidative stress, coagulation, and autonomic dysfunction in young adults. Am J Respir Crit Care Med,
176: 370-6. 091063
Chuang KJ; Chan CC; Shiao GM; Su TC. (2005). Associations between submicrometer particles exposures and
blood pressure and heart rate in patients with lung function impairments. J Occup Environ Med, 47:
1093-1098. 156356
July 2009
6-340
DRAFT-DO NOT CITE OR QUOTE

-------
Chuang KJi Coull BAi Zanobetti A; Suh Hi Schwartz J; Stone PHi Litonjua A; Speizer FEi Gold DR. (2008).
Particulate Air Pollution as a Risk Factor for ST-Segment Depression in Patients With Coronary Artery
Disease. 155731
Ciencewicki J; Gowdy Ki Krantz QTi Linak WPi Brighton Li Gilmour ML Jaspers I. (2007). Diesel exhaust
enhanced susceptibility to influenza infection is associated with decreased surfactant protein expression.
, 19: 1121-33. 096557
Clancy Li Goodman Pi Sinclair Hi Dockery DW. (2002). Effect of air pollution control on death rates in Dublin,
Ireland: an intervention study. Lancet, 360: 1210-1214. 035270
Clarke RW; Catalano Pi Coull B; Koutrakis Pi Krishna Murthy GG; Rice T; Godleski JJ. (2000). Age-related
responses in rats to concentrated urban air particles (CAPs). Presented at In: Phalen, R. F., ed.
Inhalation toxicology: proceedings of the third colloquium on particulate air pollution and human health
(first special issue); June, 1999; Durham, NC. Inhalation Toxicol. 12(suppl. l): 73"84. 011806
Coleridge HM! Coleridge JCG. (1994). Pulmonary Reflexes: Neural Mechanisms of Pulmonary Defense. Annu
Rev Physiol, 56: 69-91. 156362
Courtois A; Andujar Pi Ladeiro Yi Baudrimont L Delannoy Ei Leblais V; Begueret Hi Galland MABi Brochard Pi
Marano F. (2008). Impairment of NODependent Relaxation in Intralobar Pulmonary Arteries:
Comparison of Urban Particulate Matter and Manufactured Nanoparticles. Environ Health Perspect,
116: 1294. 156369
Cozzi Ei Hazarika Si Stallings HWi Cascio WEi Devlin RBi Lust RMi Wingard CJi Van Scott MR. (2006).
Ultrafine particulate matter exposure augments ischemia-reperfusion injury in mice. Am J Physiol, 291:
H894-H903.091380
Cruts Bi van Etten Li Tornqvist Hi Blomberg Ai Sandstrom Ti Mills NLi Borm PJ. (2008). Exposure to diesel
exhaust induces changes in EEG in human volunteers. Part Fibre Toxicol, 5- 4. 156374
Cardenas Mi Vallejo Mi Romano-Riquer Pi Ruiz-Velasco Si Ferreira-Vidal ADi Hermosillo AG. (2008). Personal
exposure to PM2.5 air pollution and heart rate variability in subjects with positive or negative head-up
tilt test. Environ Res, 108: 1-6. 191900
D'Ippoliti Di Forastiere Fi Ancona Ci Agabiti Ni Fusco Di Michelozzi Pi Perucci CA. (2003). Air pollution and
myocardial infarction in Rome: a case-crossover analysis. , 14: 528-535. 074311
Dales R. (2004). Ambient carbon monoxide may influence heart rate variability in subjects with coronary artery
disease. J Occup Environ Med, 46: 1217-21. 099036
Dales Ri Liu L; Szyszkowicz M; Dalipaj M; Willey J; Kulka R; Ruddy TD. (2007). Particulate air pollution and
vascular reactivity: the bus stop study. Int Arch Occup Environ Health, 81: 159-164. 155743
Daniels MJi Dominici Fi Zeger SLi Samet JM. (2004). The national morbidity, mortality, and air pollution study
Part III: PM10 concentration-response curves and thresholds for the 20 largest US cities. 087343
Danielsen PHi Brauner EVi Barregard Li Sallsten Gi Wallin Mi Olinski Ri Rozalski Ri Moller Pi Loft S. (2008).
Oxidatively damaged DNA and its repair after experimental exposure to wood smoke in healthy humans.
, 642: 37-42. 156382
Day KCi Reed MDi McDonald JDi Keilkop SKi Barrett EG. (2008). Effects of Gasoline Engine Emissions on
Preexisting Allergic Airway Responses in Mice. Inhal Toxicol, 20: 1145-1155. 190204
De Bruin MLi van Hemel NMi Leufkens HGi Hoes AW. (2005). Hospital discharge diagnoses of ventricular
arrhythmias and cardiac arrest were useful for epidemiologic research. , 58: 1325-1329. 155746
de Haar Ci Hassing Ii Bol Mi Bleumink Ri Pieters R. (2005). Ultrafine carbon black particles cause early airway
inflammation and have adjuvant activity in a mouse allergic airway disease model. , 87: 409-418. 097872
de Haar Ci Hassing Ii Bol Mi Bleumink Ri Pieters R. (2006). Ultrafine but not fine particulate matter causes
airway inflammation and allergic airway sensitization to co-administered antigen in mice. Clin Exp
Allergy, 36: 1469-1479. 144746
De Hartog JJi Hoek Gi Peters Ai Timonen KLi Ibald-Mulli Ai Brunekreef Bi Heinrich Ji Tiittanen Pi Van
Wijnen JHi Kreyling Wi Kulmala Mi Pekkanen J. (2003). Effects of fine and ultrafine particles on
cardiorespiratory symptoms in elderly subjects with coronary heart disease: the ULTRA study. Am J
Epidemiol, 157: 613-623. 001061
Delfino Ri Brummel Si Wu Ji Stern Hi Ostro Bi Lipsett Mi Winer Ai Street Di Zhang Li Tjoa T. (2009). The
relationship of respiratory and cardiovascular hospital admissions to the southern California wildfires of
2003. Occup Environ Med, 66: 189. 191994
Delfino RJi Chang Ji Wu Ji Ren Ci Tjoa Ti Nickerson Bi Cooper Di Gillen DL. (2009). Repeated hospital
encounters for asthma in children and exposure to traffic-related air pollution near the home. Ann
Allergy Asthma Immunol, 102: 138-44. 190254
Delfino RJi Coate BDi Zeiger RSi Seltzer JMi Street DHi Koutrakis P. (1996). Daily asthma severity in relation
to personal ozone exposure and outdoor fungal spores. Am J Respir Crit Care Med, 154: 633-641. 080788
July 2009
6-341
DRAFT-DO NOT CITE OR QUOTE

-------
Delfino RJ; Gone H; Linn WS; Pellizzari ED; Hu Y. (2003). Asthma symptoms in Hispanic children and daily
ambient exposures to toxic and criteria air pollutants. Environ Health Perspect, 111 - 647-656. 050460
Delfino RJ; Gong H; Linn WS; Hu Y; Pellizzari ED. (2003). Respiratory symptoms and peak expiratory flow in
children with asthma in relation to volatile organic compounds in exhaled breath and ambient air. J
Expo Sci Environ Epidemiol, 13: 348-363. 090941
Delfino RJ; Murphy-Moulton AM; Becklake MR. (1998). Emergency room visits for respiratory illnesses among
the elderly in Montreal: association with low level ozone exposure. Environ Res, 76' 61-11. 093624
Delfino RJ; Quintana PJE; Floro J; Gastanaga VM; Samimi BS; KLeinman MT; Liu L-JS; Bufalino C; Wu OF;
McLaren CE. (2004). Association of FEV1 in asthmatic children with personal and microenvironmental
exposure to airborne particulate matter. Environ Health Perspect, 112: 932-941. 056897
Delfino RJ; Staimer N; Gillen D; Tjoa T; Sioutas C; Fung K; George SC; Kleinman MT. (2006). Personal and
ambient air pollution is associated with increased exhaled nitric oxide in children with asthma. Environ
Health Perspect, 114: 1736-1743. 090745
Delfino RJ; Staimer N; Tjoa T; Polidori A; Arhami M; Gillen DL; Kleinman MT; Vaziri ND; Longhurst J;
Zaldivar F; Sioutas C. (2008). Circulating biomarkers of inflammation, antioxidant activity, and platelet
activation are associated with primary combustion aerosols in subjects with coronary artery disease.
Environ Health Perspect, 116: 898-906. 156390
Delfino RJ; Zeiger RS; Seltzer JM; Street DH. (1998). Symptoms in pediatric asthmatics and air pollution:
differences in effects by symptom severity, anti"inflammatory medication use and particulate averaging
time. Environ Health Perspect, 106: 751-761. 051406
Delfino RJ; Zeiger RS; Seltzer JM; Street DH; McLaren CE. (2002). Association of asthma symptoms with peak
particulate air pollution and effect modification by anti-inflammatory medication use. Environ Health
Perspect, 110: A607-A617. 093740
DeMeo DL; Zanobetti A; Litonjua AA; Coull BA; Schwartz J; Gold DR. (2004). Ambient air pollution and oxygen
saturation. Am J Respir Crit Care Med, 170: 383-387. 087346
Desqueyroux H; Pujet J-G Prosper M; Squinazi F; Momas I. (2002). Short-term effects of low-level air pollution
on respiratory health of adults suffering from moderate to severe asthma. Environ Res, 89: 29"37. 026052
Deurloo DT; van Esch BO Hofstra CL; Nijkamp FP; van Oosterhout AJ. (2001). CTLA4"IgG reverses asthma
manifestations in a mild but not in a more "severe" ongoing murine model. Am J Respir Cell Mol Biol, 25:
751-760. 156396
Devlin RB; Ghio AJ; Kehrl H; Sanders G; Cascio W. (2003). Elderly humans exposed to concentrated air
pollution particles have decreased heart rate variability. Eur Respir J, 40: 76S-80S. 087348
de Hartog JJ; Lanki T; Timonen KL; Hoek G; Janssen NA; Ibald-Mulli A; Peters A; Heinrich J; Tarkiainen TH;
van Grieken R; van Wijnen JH; Brunekreef B; Pekkanen J. (2009). Associations between PM2.5 and
heart rate variability are modified by particle composition and beta-blocker use in patients with coronary
heart disease. Environ Health Perspect, 117: 105-111. 191904
Diaz-Sanchez D; Garcia MP; Wang M; Jyrala M; Saxon A. (1999). Nasal challenge with diesel exhaust particles
can induce sensitization to a neoallergen in the human mucosa. J Allergy Clin Immunol, 104: 1183-1188.
011346
Diaz-Sanchez D; Tsien A; Fleming J; Saxon A. (1997). Combined diesel exhaust particulate and ragweed
allergen challenge markedly enhances human in vivo nasal ragweed-specific IgE and skews cytokine
production to a T helper cell 2"type pattern. J Immunol, 158: 2406-2413. 051247
Dick CA; Singh P; Daniels M; Evansky P; Becker S; Gilmour MI. (2003). Murine pulmonary inflammatory
responses following instillation of size-fractionated ambient particulate matter. J Toxicol Environ Health
A, 66: 2193-2207. 088776
Diez-Roux AV; Auchincloss AH; Astor B; Barr RG; Cushman M; Dvonch T; Jacobs DR Jr; Kaufman J; Lin X;
Samson Px. (2006). Recent exposure to particulate matter and Oreactive protein concentration in the
multi-ethnic study of atherosclerosis x. Am J Epidemiol, 164: 437-448. 156400
Dockery DW; Luttmann-Gibson H; Rich DQ; Link MS; Mittleman MA; Gold DR; Koutrakis P; Schwartz JD;
Verrier RL. (2005). Association of air pollution with increased incidence of ventricular tachyarrhythmias
recorded by implanted cardioverter defibrillators. Environ Health Perspect, 113: 670-674. 078995
Dockery DW; Luttmann-Gibson H; Rich DQ; Link MS; Schwartz JD; Gold DR; Koutrakis P; Verrier RL;
Mittleman MA. (2005). Particulate air pollution and nonfatal cardiac events Part II Association of air
pollution with confirmed arrhythmias recorded by implanted defibrillators. 090743
Dominici F; McDermott A; Daniels M; Zeger SL; Samet J. (2003). Revised Analyses of Time-Series Studies of Air
Pollution and Health: Mortality Among Residents of 90 Cities. 156407
Dominici F; McDermott A; Zeger SL; Samet JM. (2002). On the use of generalized additive models in time-series
studies of air pollution and health. Am J Epidemiol, 156: 193-203. 030458
July 2009
6-342
DRAFT-DO NOT CITE OR QUOTE

-------
Dominici Fi Peng RD; Bell ML; Pham L; McDermott A; Zeger SL; Samet JL. (2006). Fine particulate air
pollution and hospital admission for cardiovascular and respiratory diseases. JAMA, 295- 1127-1134.
088398
Dominici F> Peng RD; Ebisu K; Zeger SL; Samet JM; Bell ML. (2007). Does the effect of PM10 on mortality
depend on PM nickel and vanadium content? A reanalysis of the NMMAPS data. Environ Health
Perspect, 115: 1701-3. 099135
Dominici F; Peng RD; Zeger SL; White RH; Samet JM. (2007). Particulate air pollution and mortality in the
United States: did the risks change from 1987 to 2000?. Am J Epidemiol, 166: 880-8. 097361
Dubowsky SD; Suh H; Schwartz J; Coull BA; Gold DR. (2006). Diabetes, obesity, and hypertension may enhance
associations between air pollution and markers of systemic inflammation. Environ Health Perspect, 114:
992-998. 088750
Dusek R; Frank GP; Hildebrandt L; Curtius J; Schneider J; Walter S; Chand D; Drewnick F; Hings S; Jung D;
Borrmann S; Andreae MO. (2006). Size matters more than chemistry for cloud-nucleating ability of
aerosol particles. 1375-1378. 155756
Duvall RM; Norris GA; Dailey LA; Burke JM; McGee JK; Gilmour MI; Gordon T; Devlin RB. (2008). Source
apportionment of particulate matter in the US and associations with lung inflammatory markers. Inhal
Toxicol, 20: 671-83. 097969
Dvonch JT; Brook RD; Keeler GJ; Rajagopalan S; D'Alecy LG; Marsik FJ; Morishita M; Yip FY; Brook JR;
Timm EJ; Wagner JG; Harkema JR. (2004). Effects of concentrated fine ambient particles on rat plasma
levels of asymmetric dimethylarginine. Inhal Toxicol, 16: 473-480. 055741
Ebelt ST; Wilson WE; Brauer M. (2005). Exposure to ambient and nonambient components of particulate
matter: a comparison of health effects. Epidemiology, 16: 396-405. 056907
Elder A; Gelein R; Finkelstein J; Phipps R; Frampton M; Utell M; Kittelson DB; Watts WF; Hopke P; Jeong CH;
Kim E; Liu W; Zhao W; Zhuo L; Vincent R; Kumarathasan P; Oberdorster G. (2004). On-road exposure to
highway aerosols 2 Exposures of aged, compromised rats. Inhal Toxicol, 16 Suppl l: 41-53. 087354
Elder ACP; Gelein R; Azadniv M; Frampton M; Finkelstein J; Oberdorster G. (2004). Systemic effects of inhaled
ultrafine particles in two compromised, aged rat strains. Inhal Toxicol, 16: 461-471. 055642
Erbas B; Kelly A"M; Physick B; Code C; Edwards M. (2005). Air pollution and childhood asthma emergency
hospital admissions: estimating intra-city regional variations. Int J Environ Health Res, 15: 11-20.
073849
Fairley D. (2003). Mortality and air pollution for Santa Clara County, California, 1989-1996. 042850
Fakhri AA; Ilic LM; Wellenius GA; Urch B; Silverman F; Gold DR; Mittleman MA. (2009). Autonomic effects of
controlled fine particulate exposure in young healthy adults: Effect modification by ozone. Environ
Health Perspect, In Press: In Press. 191914
Fan Z; Meng Q; Weisel C; Laumbach R; Ohman-Strickland P; Shalat S; Hernandez M; Black K. (2008). Acute
exposure to elevated PM (2.5) generated by traffic and cardiopulmonary health effects in healthy older
adults. J Expo Sci Environ Epidemiol, 19: 525-533. 191979
Farraj AK; Haykal-Coates N; Ledbetter AD; Evansky PA; Gavett SH. (2006). Inhibition of pan neurotrophin
receptor p75 attenuates diesel particulate "induced enhancement of allergic airway responses in
C57/B16J mice. Inhal Toxicol, 18: 483-491. 088469
Farraj AK; Haykal-Coates N; Ledbetter AD; Evansky PA; Gavett SH. (2006). Neurotrophin mediation of allergic
airways responses to inhaled diesel particles in mice. , 94: 183-92. 141730
Fedulov AV; Leme A; Yang Z; Dahl M; Lim R; Mariani TJ; Kobzik L. (2008). Pulmonary exposure to particles
during pregnancy causes increased neonatal asthma susceptibility. , 38: 57"67. 097482
Ferdinands JM; Crawford CA; Greenwald R; Van Sickle D; Hunter E; Teague WG. (2008). Breath acidification
in adolescent runners exposed to atmospheric pollution: a prospective, repeated measures observational
study. Environ Health, 7: 10. 156433
Finnerty K; Choi J-E; Lau A; Davis-Gorman G; Diven C; Seaver N> Linak William P; Witten M; McDonagh Paul
F. (2007). Instillation of coarse ash particulate matter and lipopolysaccharide produces a systemic
inflammatory response in mice. J Toxicol Environ Health A, 70: 1957-1966. 156434
Fischer PH; Steerenberg PA; Snelder JD; Van Loveren H; Van Amsterdam JGC. (2002). Association between
exhaled nitric oxide, ambient air pollution and respiratory health in school children. Int Arch Occup
Environ Health, 75: 348-353. 025731
Fischer SL; Koshland CP. (2007). Daily and peak 1 h indoor air pollution and driving factors in a rural Chinese
village. Environ Sci Technol, 41: 3121-3126. 156435
Folino AF; Scapellato ML; Canova C; Maestrelli P; Bertorelli G; Simonato L; Iliceto S; Lotti M. (2009).
Individual exposure to particulate matter and the short-term arrhythmic and autonomic profiles in
patients with myocardial infarction. Eur Heart J, 30: 1614-1620. 191902
July 2009
6-343
DRAFT-DO NOT CITE OR QUOTE

-------
Forastiere F; Stafoggia M; Berti G; Bisanti L; Cernigliaro A; Chiusolo M; Mallone S; Miglio R; Pandolfi P;
Rognoni M; Serinelli M; Tessari R; Vigotti M; Perucci C. (2008). Particulate Matter and Daily Mortality:
A Case-Crossover Analysis of Individual Effect Modifiers. Epidemiology, 19: 571-580. 186937
Forastiere F; Stafoggia M; Picciotto S; Bellander T; D'Ippoliti D; Lanki T; Von Klot S; Nyberg F; Paatero P;
Peters A; Pekkanen J; Sunyer J; Perucci CA. (2005). A case-crossover analysis of out-of-hospital coronary
deaths and air pollution in Rome, Italy. Am J Respir Crit Care Med, 172: 1549-1555. 086323
Frampton MW. (2001). Systemic and cardiovascular effects of airway injury and inflammation: ultrafine particle
exposure in humans. Environ Health Perspect, 109: 529-532. 019051
Frampton MW; Stewart JC; Oberdorster G; Morrow PE; Chalupa D; Pietropaoli AP; Frasier LM; Speers DM;
Cox C; Huang LS; Utell MJ. (2006). Inhalation of ultrafine particles alters blood leukocyte expression of
adhesion molecules in humans. Environ Health Perspect, 114: 51-58. 088665
Franklin M; Koutrakis P; Schwartz J. (2008). PM25 composition & daily mortality in 25 US communities.
097426
Franklin M; Koutrakis P; Schwartz J. (2008). The role of particle composition on the association between PM2.5
and mortality. Epidemiology, 19: 680-689. 155779
Franklin M; Zeka A; Schwartz J. (2007). Association between PM2.5 and all-cause and specific-cause mortality
in 27 US communities. J Expo Sci Environ Epidemiol, 17: 279-287. 091257
Fung KY; Khan S; Krewski D; Chen Y. (2006). Association between air pollution and multiple respiratory
hospitalizations among the elderly in Vancouver, Canada. Inhal Toxicol, 18: 1005-1011. 089789
Fung KY; Luginaah IKMG; Webster G. (2005). Air pollution and daily hospitalization rates for cardiovascular
and respiratory diseases in London, Ontario. Int J Environ Stud, 62: 677-685. 093262
Galan I; Tobias A; Banegas JR; Aranguez E. (2003). Short-term effects of air pollution on daily asthma
emergency room admissions. Eur Respir J, 22: 802-808. 087408
Gavett SH; Haykal-Coates N> Copeland L B; Heinrich J; Gilmour MI. (2003). Metal composition of ambient
PM2.5 influences severity of allergic airways disease in mice. Environ Health Perspect, 111: 1471-1477.
053153
Gehlbach BE Geppert E. (2004). The pulmonary manifestations of left heart failure. , 125: 669-682. 155784
Geiser M; Rothen-Rutishauser B; Kapp N> Schurch S; Kreyling W', Schulz H; Semmler M; Im Hof V; Heyder J;
Gehr P. (2005). Ultrafine particles cross cellular membranes by nonphagocytic mechanisms in lungs and
in cultured cells. Environ Health Perspect, 113: 1555-1560. 087362
Gent JF; Koutrakis P; Belanger K; Triche E; Holford TR; Bracken MB; Leaderer BP. (2009). Symptoms and
medication use in children with asthma and traffic-related sources of fine particle pollution. Environ
Health Perspect, In Press: 1-41. 180399
Gent JF; Triche EW; Holford TR; Belanger K; Bracken MB; Beckett WS; Leaderer BP. (2003). Association of
low-level ozone and fine particles with respiratory symptoms in children with asthma. JAMA, 290: 1859"
1867. 052885
Gerlofs-Nijland ME; Rummelhard M; Boere AJF; Leseman DLAC; Duffin R; Schins RPF; Borm PJA; Sillanpaa
M; Salonen RO; Cassee FR. (2009). Particle induced toxicity in relation to transition metal and polycyclic
aromatic hydrocarbon contents . Environ Sci Technol, In Press: 1-8. 190353
Gerlofs-Nijland ME; Dormans JA; Bloemen HJ; Leseman DL; John A; Boere F; Kelly FJ; Mudway IS; Jimenez
AA; Donaldson K; Guastadisegni C; Janssen NA; Brunekreef B; Sandstrom T; van Bree L; Cassee FR.
(2007). Toxicity of coarse and fine particulate matter from sites with contrasting traffic profiles. , 19:
1055-69. 097840
Ghelfi E; Rhoden CR; Wellenius GA; Lawrence J; Gonzalez-Flecha B. (2008). Cardiac oxidative stress and
electrophysiological changes in rats exposed to concentrated air particles are mediated by TRP"
dependent pulmonary reflexes. Toxicol Sci, 102: 328-336. 156468
Ghio AJ; Devlin RB. (2001). Inflammatory lung injury after bronchial instillation of air pollution particles. Am J
Respir Crit Care Med, 164: 704-708. 017122
Ghio AJ; Hall A; Bassett MA; Cascio WE; Devlin RB. (2003). Exposure to concentrated ambient air particles
alters hematologic indices in humans. Inhal Toxicol, 15: 1465-1478. 087363
Ghio AJ; Kim C; Devlin RB. (2000). Concentrated ambient air particles induce mild pulmonary inflammation in
healthy human volunteers. Am J Respir Crit Care Med, 162: 981-988. 012140
Gilliland FD; Li YF; Saxon A; Diaz-Sanchez D. (2004). Effect of glutathione-S-transferase Ml and PI genotypes
on xenobiotic enhancement of allergic responses: randomised, placebo-controlled crossover study. Lancet,
363: 119-125. 156471
Gilmour MI; McGee J; Duvall Rachelle M; Dailey L; Daniels M; Boykin E; Cho S"H; Doerfler D; Gordon T;
Devlin Robert B. (2007). Comparative toxicity of size-fractionated airborne particulate matter obtained
from different cities in the United States. , 19 Suppl l: 7" 16. 096433
July 2009
6-344
DRAFT-DO NOT CITE OR QUOTE

-------
Gilmour ML O'Connor Si Dick CAJi Miller CAi Linak WP. (2004). Differential pulmonary inflammation and in
vitro cytotoxicity of size-fractionated fly ash particles from pulverized coal combustion. J Air Waste
Manag Assoc, 54: 286-295. 057420
Gilmour PSi Ziesenis Ai Morrison ERi Vickers MAi Drost EMi Ford Ii Karg Ei Mossa Ci Schroeppel Ai Ferron
GAi Heyder J; Greaves Mi MacNee W', Donaldson K. (2004). Pulmonary and systemic effects of short-
term inhalation exposure to ultrafine carbon black particles. Toxicol Appl Pharmacol, 195: 35"44. 054175
Girardot SPi Ryan PBi Smith SMi Davis WTi Hamilton CBi Obenour RAi Renfro JR; Tromatore KA; Reed GD.
(2006). Ozone and PM25 exposure and acute pulmonary health effects: a study of hikers in the Great
Smoky Mountains National Park. Environ Health Perspect, 113: 612-617. 088271
Godleski JJi Clarke RWi Coull BAi Saldiva PHNi Jiang NFi Lawrence J; Koutrakis P. (2002). Composition of
inhaled urban air particles determines acute pulmonary responses. Ann Occup Hyg, 46: 419-424. 156478
Godleski JJi Verrier RLi Koutrakis Pi Catalano Pi Coull Bi Reinisch U> Lovett EG; Lawrence J; Murthy GG;
Wolfson JM; Clarke RW! Nearing BD; Killingsworth C. (2000). Mechanisms of morbidity and mortality
from exposure to ambient air particles. , 91: 5"88; discussion 89-103. 000738
Gold DR; Litonjua AA; Zanobetti A; Coull BA; Schwartz J; MacCallum G; Verrier RL; Nearing BD; Canner MJ;
Suh H; Stone PH. (2005). Air pollution and ST-segment depression in elderly subjects. Environ Health
Perspect, 113: 883-887. 087558
Goldberg MS; Giannetti N; Burnett RT; Mayo NE; Valois MF; Brophy JM. (2008). A panel study in congestive
heart failure to estimate the short-term effects from personal factors and environmental conditions on
oxygen saturation and pulse rate. Occup Environ Med, 65: 659-666. 180380
Gong H Jri Sioutas C; Linn WS. (2003). Controlled exposures of healthy and asthmatic volunteers to
concentrated ambient particles in metropolitan Los Angeles. 087365
Gong H; Sioutas C; Linn WS; Clark KW; Terrell SL; Terrell LL; Anderson KR; Kim S; Chang MC. (2000).
Controlled human exposures to concentarted ambient fine particles in metropolitan Los Angeles:
Methodology and preliminary health-effect findings. Inhal Toxicol, 12: 107-119. 155799
Gong H Jri Linn WS; Clark KW; Anderson KR; Geller MD; Sioutas C. (2005). Respiratory responses to
exposures with fine particulates and nitrogen dioxide in the elderly with and without COPD. Inhal
Toxicol, 17: 123-132. 087921
Gong H Jr; Linn WS; Clark KW; Anderson KR; Sioutas C; Alexis NE; Cascio WE; Devlin RB. (2008). Exposures
of healthy and asthmatic volunteers to concentrated ambient ultrafine particles in Los Angeles. Inhal
Toxicol, 20: 533-545. 156483
Gong H Jri Linn WS; Sioutas C; Terrell SL; Clark KW; Anderson KR; Terrell LL. (2003). Controlled exposures
of healthy and asthmatic volunteers to concentrated ambient fine particles in Los Angeles. Inhal Toxicol,
15: 305-325. 042106
Gong H Jri Linn WS; Terrell SL; Anderson KR; Clark KW; Sioutas C; Cascio WE; Alexis N; Devlin RB. (2004).
Exposures of elderly volunteers with and without chronic obstructive pulmonary disease (COPD) to
concentrated ambient fine particulate pollution. Inhal Toxicol, 16: 731-744. 087964
Gong H Jr; Linn WS; Terrell SL; Clark KW; Geller MD; Anderson KR; Cascio WE; Sioutas C. (2004). Altered
heart-rate variability in asthmatic and healthy volunteers exposed to concentrated ambient coarse
particles. Inhal Toxicol, 16: 335-343. 055628
Gordian ME; Choudhury AH. (2003). PM10 and asthma medication in schoolchildren. Arch Environ Occup
Health, 58: 42-47. 054842
Gowdy K; Krantz QT; Daniels M; Linak WP; Jaspers I; Gilmour MI. (2008). Modulation of pulmonary
inflammatory responses and antimicrobial defenses in mice exposed to diesel exhaust. Toxicol Appl
Pharmacol, 229(3): 310-319. 097226
Graff D; Cascio W; Rappold A; Zhou H; Huang Y; Devlin R. (2009). Exposure to concentrated coarse air pollution
particles causes mild cardiopulmonary effects in healthy young adults . Environ Health Perspect, 117:
1089-1094. 191981
Gurgueira SA; Lawrence J; Coull B; Murthy GGK; Gonzalez-Flecha B. (2002). Rapid increases in the steady-
state concentration of reactive oxygen species in the lungs and heart after particulate air pollution
inhalation. Environ Health Perspect, 110: 749-755. 036535
Hajat S; Haines A; Atkinson RW! Bremner SA; Anderson HR; Emberlin J. (2001). Association between air
pollution and daily consultations with general practitioners for allergic rhinitis in London, United
Kingdom. Am J Epidemiol, 153: 704-714. 016693
Halonen JI; Lanki T; Yli-Tuomi T; Kulmala M; Tiittanen P; Pekkanen J. (2008). Urban air pollution, and
asthma and COPD hospital emergency room visits. Thorax, 63: 635-41. 189507
Halonen JI; Lanki T; Yli-Tuomi T; Tiittanen P; Kulmala M; Pekkanen. (2009). Particulate air pollution and
acute cardiorespiratory hospital admissions and mortality among the elderly. Epidemiology, 20: 143-153.
180379
July 2009
6-345
DRAFT-DO NOT CITE OR QUOTE

-------
Hamada Ki Suzaki Y> Leme A; Ito T> Miyamoto Ki Kobzik L> Kimura H. (2007). Exposure of pregnant mice to an
air pollutant aerosol increases asthma susceptibility in offspring. J Toxicol Environ Health A, 70- 688"
695. 091235
Hamade AK; Rabold R; Tankersley CG. (2008). Adverse cardiovascular effects with acute particulate matter and
ozone exposures' interstrain variation in mice. Environ Health Perspect, 116- 1033-1039. 156515
Hanigan ICi Johnston FH; Morgan GG. (2008). Vegetation fire smoke, indigenous status and cardio respiratory
hospital admissions in Darwin, Australia, 1996-2005- a time-series study. , 7- 42. 156518
Hao Mi Cornier Si Wang Mi Lee James Ji Nel A. (2003). Diesel exhaust particles exert acute effects on airway
inflammation and function in murine allergen provocation models. J Allergy Clin Immunol, 112- 905-914.
096565
Hapcioglu Bi Issever Hi Kocyigit E> Disci R; Vatansever Si Ozdilli K. (2006). The effect of air pollution and
meteorological parameters on chronic obstructive pulmonary disease at an Istanbul hospital. Indoor Built
Environ, 15: 147-153. 093263
Happo MS; Salonen RCK Halinen AI; Jalava PI; Pennanen AS; Kosma VM; Sillanpaa M; Hillamo R; Brunekreef
B; Katsouyanni K; Sunyer J; Hirvonen MR. (2007). Dose and time dependency of inflammatory responses
in the mouse lung to urban air coarse, fine, and ultrafine particles from six European cities. , 19: 227"
246. 096630
Harder V; Gilmour P; Lentner B; Karg E; Takenaka S; Ziesenis A; Stampfl A; Kodavanti U; Heyder J; Schulz H.
(2005). Cardiovascular responses in unrestrained WKY rats to inhaled ultrafine carbon particles. Inhal
Toxicol, 17: 29-42. 087371
Harkema JR; Keeler G; Wagner J; Morishita M; Timm E; Hotchkiss J; Marsik F; Dvonch T; Kaminski N; Barr
E. (2004). Effects of concentrated ambient particles on normal and hypersecretory airways in rats.
056842
Harrod KS; Jaramillo RJ; Berger JA; Gigliotti AP; Seilkop SK; Reed MD. (2005). Inhaled diesel engine
emissions reduce bacterial clearance and exacerbate lung disease to Pseudomonas aeruginosa infection in
vivo. Toxicol Sci, 83: 155-165. 088144
Harrod KS; Jaramillo RJ; Rosenberger CL; Wang S"Z; Berger JA; McDonald JD; Reed MD. (2003). Increased
susceptibility to RSV infection by exposure to inhaled diesel engine emissions. , 28: 451-463. 097046
Hayek T; Oiknine J; Brook JG; Aviram M. (1994). Increased plasma and lipoprotein lipid peroxidation in apo E-
deficient mice. Biochem Biophys Res Commun, 201: 1567-1574. 156527
Hedley AJ; Wong C"M; Thach TQ; Ma S; Lam T-H; Anderson HR. (2002). Cardiorespiratory and all-cause
mortality after restrictions on sulphur content of fuel in Hong Kong: an intervention study. Lancet, 360:
1646-1652. 040284
Heidenfelder BL; Reif DM; Harkema JR; Cohen Hubal EA; Hudgens EE; Bramble LA; Wagner JG; Morishita M;
Keeler GJ; Edwards SW; Gallagher JE. (2009). Comparative microarray analysis and pulmonary changes
in brown norway rats exposed to ovalbumin and concentrated air particulates. Toxicol Sci, 108: 207-221.
190026
Henneberger A; Zareba W', Ibald-Mulli A; Ruckerl R; Cyrys J; Couderc J-P; Mykins B; Woelke G; Wichmann H-
E; Peters A. (2005). Repolarization changes induced by air pollution in ischemic heart disease patients.
Environ Health Perspect, 113: 440-446. 087960
Henrotin JB; Besancenot JP; Bejot Y; Giroud M. (2007). Short-term effects of ozone air pollution on ischaemic
stroke occurrence: a case-crossover analysis from a 10-year population-based study in Dijon, France.
Occup Environ Med, 64: 439-445. 093270
Hinwood AL; De Klerk N> Rodriguez C; Jacoby P; Runnion T; Rye P; Landau L; Murray F; Feldwick M; Spickett
J. (2006). The relationship between changes in daily air pollution and hospitalizations in Perth, Australia
1992-1998: a case-crossover study. Int J Environ Health Res, 16: 27"46. 088976
Hirshon JM; Shardell M; Alles S; Powell JL; Squibb K; Ondov J; Blaisdell CJ. (2008). Elevated ambient air zinc
increases pediatric asthma morbidity. Environ Health Perspect, 116: 826-831. 180375
Hogervorst JG; de Kok TM; Briede JJ; Wesseling G; Kleinjans JC; van Schayck CP. (2006). Relationship
between radical generation by urban ambient particulate matter and pulmonary function of school
children. , 69: 245-62. 189460
Holguin F; Flores S; Ross Z; Cortez M; Molina M; Molina L; Rincon C; Jerrett M; Berhane K; Granados A;
Romieu I. (2007). Traffic-related exposures, airway function, inflammation, and respiratory symptoms in
children. Am J Respir Crit Care Med, 176: 1236-42. 099000
Holguin F; Tellez-Rojo MM; Hernandez M; Cortez M; Chow JC; Watson JG; Mannino D; Romieu I. (2003). Air
pollution and heart rate variability among the elderly in Mexico City. , 14: 521-527. 057326
Holloman CH; Bortnick SM; Morara M; Strauss WJ; Calder CA. (2004). A Bayesian hierarchical approach for
relating PM2.5 exposure to cardiovascular mortliaty in North Carolina. Environ Health Perspect, 112:
1282-1288. 087375
July 2009
6-346
DRAFT-DO NOT CITE OR QUOTE

-------
Hong Y-C; Hwang S-S; Kim JH; Lee K-H; Lee H-J; Lee K-H; Yu S-D; Kim D-S. (2007). Metals in particulate
pollutants affect peak expiratory flow of schoolchildren. Environ Health Perspect, 115- 430-434. 091347
Hopke PE Ito E Mar T; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Liu H;
Neas L; Pinto J; Stolzel M; Suh H; Paatero P; Thurston GD. (2006). PM source apportionment and health
effects' 1 Intercomparison of source apportionment results. J Expo Sci Environ Epidemiol, 16- 275-286.
088390
Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V;
Prouvost H; Lefranc A. (2007). Short-term Associations between Fine and Coarse Particles and
Cardiorespiratory Hospitalizations in Six French Cities. Occup Environ Med. 155851
Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V;
Prouvost H; Lefranc A. (2008). Short-term associations between fine and coarse particles and hospital
admissions for cardiorespiratory diseases in six French cities. Occup Environ Med, 65- 544-551. 155852
Huang Y-CT; Ghio AJ; Stonehuerner J; McGee J; Carter JD; Grambow SC; Devlin RB. (2003). The role of
soluble components in ambient fine particles-induced changes in human lungs and blood. Inhal Toxicol,
15: 327-342. 087377
Huber SA; Sakkinen P; Conze D; Hardin N; Tracy R. (1999). Interleukin-6 exacerbates early atherosclerosis in
mice. Arterioscler Thromb Vase Biol, 19: 2364-2367. 156575
Hwang J-S; Chan C"C. (2002). Effects of air pollution on daily clinic visits for lower respiratory tract illness. Am
J Epidemiol, 155: M0. 023222
Hwang J-S; Nadziejko C; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles
(CAPs) in mice: HI Acute and chronic effects of CAPs on heart rate, heart-rate fluctuation, and body
temperature. Inhal Toxicol, 17: 199-207. 087957
Ibald-Mulli A; Timonen KL; Peters A; Heinrich J; Wolke G; Lanki T; Buzorius G; Kreyling WG; De Hartog J;
Hoek G; Ten Brink HM; Pekkanen J. (2004). Effects of particulate air pollution on blood pressure and
heart rate in subjects with cardiovascular disease: a multicenter approach. Environ Health Perspect,
112: 369-377. 087415
Inoue K; Takano H; Yanagisawa R; Sakurai M; Ichinose T; Sadakane K; Yoshikawa T. (2005). Effects of nano
particles on antigen-related airway inflammation in mice. Respir Res, 16. 088625
Ito K. (2003). Associations of particulate matter components with daily mortality and morbidity in Detroit,
Michigan. 042856
Ito K; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Neas L; Hopke PK;
Thurston GD. (2006). PM source apportionment and health effects: 2 An investigation of intermethod
variability in associations between source-apportioned fine particle mass and daily mortality in
Washington, DC. J Expo Sci Environ Epidemiol, 16: 300-310. 088391
Ito K; Thurston GD; Silverman RA;. (2007). Association between coarse particles and asthma emergency
department (ED) visits in New York City. 091262
Ito T; Okumura H; Tsukue N; Kobayashi T; Honda K; Sekizawa K. (2006). Effect of diesel exhaust particles on
mRNA expression of viral and bacterial receptors in rat lung epithelial L2 cells. , 165: 66-70. 096648
Ito T; Suzuki T; Tamura K; Nezu T; Honda K; Kobayashi T. (2008). Examination of mRNA expression in rat
hearts and lungs for analysis of effects of exposure to concentrated ambient particles on cardiovascular
function. Toxicol Sci, 243: 271-283. 096823
Jaffe DH; Singer ME; Rimm AA. (2003). Air pollution and emergency department visits for asthma among Ohio
Medicaid recipients, 1991-1996. Environ Res, 91: 21-28. 041957
Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Simpson R; Corbett S. (2006). Associations between ambient
air pollution and daily emergency department attendances for cardiovascular disease in the elderly (65+
years), Sydney, Australia. , 16: 225-37. 189416
Jalaludin BB; O'Toole BI; Leeder SR. (2004). Acute effects of urban ambient air pollution on respiratory
symptoms, asthma medication use, and doctor visits for asthma in a cohort of Australian children.
Environ Res, 95: 32-42. 056595
Jalava PI; Salonen RO; Halinen AL Penttinen P; Pennanen AS; Sillanpaa M; Sandell E; Hillamo R; Hirvonen
M-R. (2006). In vitro inflammatory and cytotoxic effects of size-segregated particulate samples collected
during long-range transport of wildfire smoke to Helsinki. Toxicol Appl Pharmacol, 215: 341-353. 155872
Jalava PI; Salonen RO; Pennanen AS; Happo MS; Penttinen P; Halinen AL Sillanpaa M; Hillamo R; Hirvonen
MR. (2008). Effects of solubility of urban air fine and coarse particles on cytotoxic and inflammatory
responses in RAW 264.7 macrophage cell line. Toxicol Appl Pharmacol, 229: 146-60. 098968
Janes H; Sheppard L; Lumley T. (2005). Case-crossover analyses of air pollution exposure data: referent
selection strategies and their implications for bias. , 16: 717-726. 087535
July 2009
6-347
DRAFT-DO NOT CITE OR QUOTE

-------
Jansen KL; Larson TV; Koenig JQ; Mar TF; Fields Ci Stewart J; Lippmann M. (2005). Associations between
health effects and particulate matter and black carbon in subjects with respiratory disease. Environ
Health Perspect, 113: 1741-1746. 082236
Janssen NAH; Schwartz J; Zanobetti A; Suh HH. (2002). Air conditioning and source-specific particles as
modifiers of the effect of PM10 on hospital admissions for heart and lung disease. Environ Health
Perspect, 110: 43-49. 016743
Johnston FH; Bailie RS; Pilotto LS; Hanigan IC. (2007). Ambient biomass smoke and cardio-respiratory hospital
admissions in Darwin, Australia. , T- 240. 155882
Johnston FH; Webby RJ; Pilotto LS; Bailie RS; Parry DL; Halpin SJ. (2006). Vegetation fires, particulate air
pollution and asthma: a panel study in the Australian monsoon tropics. Int J Environ Health Res, 16:
391-404. 091386
Just J; Segala C; Sahraoui F; Priol G; Grimfeld A; Neukirch F. (2002). Short-term health effects of particulate
and photochemical air pollution in asthmatic children. Eur Respir J, 20: 899-906. 035429
Kannel WB; Abbott RD; Savage DD; McNamara PM. (1983). Coronary heart disease and atrial fibrillation: the
Framingham Study. Am Heart J, 106: 389-396. 156623
Katsouyanni K; Touloumi G; Samoli E; Petasakis Y; Analitis A; Le Tertre A; Rossi G; Zmirou D; Ballester F;
Boumghar A; Anderson HR; Wojtyniak B; Paldy A; Braunstein R; Pekkanen J; Schindler C; Schwartz J.
(2003). Sensitivity analysis of various models of short-term effects of ambient particles on total mortality
in 29 cities in APHEA2. 042807
Kelishadi R; Mirghaffari N; Poursafa P; Gidding S. (2009). Lifestyle and environmental factors associated with
inflammation, oxidative stress and insulin resistance in children. Atherosclerosis, 203: 311-319. 191960
Kettunen J; Lanki T; Tiittanen P; Aalto PP; Koskentalo T; Kulmala M; Salomaa V; Pekkanen J. (2007).
Associations of fine and ultrafine particulate air pollution with stroke mortality in an area of low air
pollution levels. Stroke, 38: 918-922. 091242
Kim SY; O'Neill MS; Lee JT; Cho Y; Kim J; Kim H. (2007). Air pollution, socioeconomic position, and emergency
hospital visits for asthma in Seoul, Korea. Int Arch Occup Environ Health, 80: 701-710. 092837
Klein-Patel ME; Diamond G; Boniotto M; Saad S; Ryan LK. (2006). Inhibition of beta-defensin gene expression
in airway epithelial cells by low doses of residual oil fly ash is mediated by vanadium. , 92: 115-125.
097092
Kleinman MT; Araujo JA; Nel A; Sioutas C, Campbell A; Cong PQ; Li H; Bondy SC. (2008). Inhaled ultrafine
particulate matter affects CNS inflammatory processes and may act via MAP kinase signaling pathways.
Toxicol Lett, 178: 127-130. 190074
Kleinman MT; Hamade A; Meacher D; Oldham M; Sioutas C; Chakrabarti B; Stram D; Froines JR; Cho AK.
(2005). Inhalation of concentrated ambient particulate matter near a heavily trafficked road stimulates
antigen-induced airway responses in mice. , 55: 1277-88. 189364
Kleinman MT; Sioutas C; Froines JR; Fanning E; Hamade A; Mendez L; Meacher D; Oldham M. (2007).
Inhalation of concentrated ambient particulate matter near a heavily trafficked road stimulates antigen-
induced airway responses in mice. , 19 Suppl l: 117-26. 097082
Klemm RJ; Lipfert FW; Wyzga RE; Gust C. (2004). Daily mortality and air pollution in Atlanta: two years of
data from ARIES. Inhal Toxicol, 16 Suppl l: 131-141. 056585
Klemm RJ; Mason R. (2003). Replication of reanalysis of Harvard Six-City mortality study. 042801
Klemm RJ; Mason RM Jr. (2000). Aerosol Research and Inhalation Epidemiological Study (ARIES): air quality
and daily mortality statistical modeling-"interim results. J Air Waste Manag Assoc, 50: 1433-1439.
010389
Knuckles T; Lund A; Lucas S; Campen M. (2008). Diesel exhaust exposure enhances venoconstriction via
uncoupling of eNOS. Toxicol Appl Pharmacol, 230: 346. 191987
Ko FWS; Tam W', Wong TW; Chan DPS. (2007). Temporal relationship between air pollutants and hospital
admissions for chronic obstructive pulmonary disease in Hong Kong. Thorax, 62: 780-785. 091639
Ko FWS; Tam W; Wong TW; Lai CKW;. (2007). Effects of air pollution on asthma hospitalization rates in
different age groups in Hong Kong. Clin Exp Allergy, 37: 1312-1319. 092844
Kodavanti UP; Schladweiler MC; Ledbetter AD; Hauser R; Christiani DC; McGee J; Richards JR; Costa DL.
(2002). Temporal association between pulmonary and systemic effects of particulate matter in healthy
and cardiovascular compromised rats. J Toxicol Environ Health A, 65: 1545-1569. 025236
Kodavanti UP; Schladweiler MC; Ledbetter AD; McGee JK; Walsh L; Gilmour PS; Highfill JW; Davies D;
Pinkerton KE; Richards JH; Crissman K; Andrews D; Costa DL. (2005). Consistent pulmonary and
systemic responses from inhalation of fine concentrated ambient particles: roles of rat strains used and
physicochemical properties. Environ Health Perspect, 113: 1561-1568. 087946
July 2009
6-348
DRAFT-DO NOT CITE OR QUOTE

-------
Kodavanti UP; Schladweiler MCJ; Ledbetter AD; Hauser R; Christiani DC; Samet JM; McGee J; Richards JH;
Costa DL. (2002). Pulmonary and systemic effects of zinc-containing emission particles in three rat
strains' multiple exposure scenarios. Toxicol Sci, 70- 73"85. 035344
Koenig J; Allen R; Larson T; Liu S. (2005). Response to "Indoor- and Outdoor-Generated Particles and Children
with asthma" [letter]. Environ Health Perspect, 113: A581. 088999
Koenig JQ; Jansen K; Mar TF; Lumley T; Kaufman J; Trenga CA; Sullivan J; Liu LJ; Shapiro GG; Larson TV.
(2003). Measurement of offline exhaled nitric oxide in a study of community exposure to air pollution.
Environ Health Perspect, 111: 1625-1629. 156653
Koenig JQ; Mar TF; Allen RW; Jansen K; Lumley T; Sullivan JH; Trenga CA; Larson T; Liu LJ. (2005).
Pulmonary effects of indoor- and outdoor-generated particles in children with asthma. Environ Health
Perspect, 113: 499-503. 087384
Koken PJM; Piver WT; Ye F; Elixhauser A; Olsen LM; Portier CJ. (2003). Temperature, air pollution, and
hospitalization for cardiovascular diseases among elderly people in Denver. Environ Health Perspect,
111: 1312-1317. 049466
Kongerud J; Madden MC; Hazucha M; Peden D. (2006). Nasal responses in asthmatic and nonasthmatic
subjects following exposure to diesel exhaust particles. Inhal Toxicol, 18: 589-594. 156656
Kooter IM; Boere AJ; Fokkens PH; Leseman DL; Dormans JA; Cassee FR. (2006). Response of spontaneously
hypertensive rats to inhalation of fine and ultrafine particles from traffic: experimental controlled study.
Part Fibre Toxicol, 15: 3-7. 097547
Kuo HW; Lai JS; Lee MC; Tai RC; Lee MC. (2002). Respiratory effects of air pollutants among asthmatics in
central Taiwan. Arch Environ Occup Health, 57: 194-200. 036310
Laden F; Neas LM; Dockery DW; Schwartz J. (2000). Association of fine particulate matter from different
sources with daily mortality in six US cities. Environ Health Perspect, 108: 941-947. 012102
Lagorio S; Forastiere F; Pistelli R; Iavarone I; Michelozzi P; Fano V; Marconi A; Ziemacki G; Ostro BD. (2006).
Air pollution and lung function among susceptible adult subjects: a panel study. , 5: 11. 089800
Lai LW; Cheng WL. (2008). The impact of air quality on respiratory admissions during Asian dust storm
periods. Int J Environ Health Res, 18: 429-450. 180301
Langrish JP; Mills NL; Chan JK; Leseman DL; Aitken RJ; Fokkens PH; Cassee FR; Li J; Donaldson K; Newby
DE; Jiang L. (2009). Beneficial cardiovascular effects of reducing exposure to particulate air pollution
with a simple facemask. Part Fibre Toxicol, 6: 8. 191908
Lanki T; De Hartog JJ; Heinrich J; Hoek G; Janssen NAH; Peters A; Stolzel M; Timonen KL; Vallius M;
Vanninen E; Pekkanen J. (2006). Can we identify sources of fine particles responsible for exercise-
induced ischemia on days with elevated air pollution? The ULTRA study. Environ Health Perspect, 114:
655-660. 088412
Lanki T; Hoek G; Timonen K; Peters A; Tiittanen P; Vanninen E; Pekkanen J. (2008). Hourly variation in fine
particle exposure is associated with transiently increased risk of ST segment depression. Br Med J, 65:
782. 191984
Lanki T; Pekkanen J; Aalto P; Elosua R; Berglind N; D'Ippoliti D; Kulmala M; Nyberg F; Peters A; Picciotto S;
Salomaa V; Sunyer J; Tiittanen P; Von Klot S; Forastiere F; for the HEAPSS Study Group. (2006).
Associations of traffic-related air pollutants with hospitalisation for first acute myocardial infarction: the
HEAPSS study. Occup Environ Med, 63: 844-851. 089788
Larrieu S; Jusot J-F; Blanchard M; Prouvost H; Declercq C; Fabre P; Pascal L; Le Tertre A; Wagner V; Riviere
S; Chardon B; Borelli D; Cassadou S; Eilstein D; Lefranc A. (2007). Short term effects of air pollution on
hospitalizations for cardiovascular diseases in eight French cities: The PSAS program. Sci Total Environ,
387: 105-112. 093031
Larrieu S; Lefranc A; Gault G; Chatignoux E; Couvy F; Jouves B; Filleul L. (2009). Are the Short-term Effects of
Air Pollution Restricted to Cardiorespiratory Diseases?. Am J Epidemiol, 169: 1-8. 180294
Larsson B"M; Sehistedt M; Grunewald J; Skold CM; Lundin A; Blomberg A; Sandstrom T; Eklund A;
Svartengren M. (2007). Road tunnel air pollution induces bronchoalveolar inflammation in healthy
subjects. Eur Respir J, 29: 699-705. 091375
Last JA; Ward R; Temple L; Pinkerton KE; Kenyon NJ. (2004). Ovalbumin "induced airway inflammation and
fibrosis in mice also exposed to ultrafine particles. , 16: 93-102. 097334
Laupacis A; Boysen G; Connolly S. (1994). Risk factors for stroke and efficacy of antithrombotic therapy in atrial
fibrillation: Analysis of pooled data from five randomized controlled trials. , 154: 1449-1457. 190901
Laurent O; Pedrono G; Filleul L; Segala C; Lefranc A; Schillinger C; Riviere E; Bard D. (2009). Influence of
socioeconomic deprivation on the relation between air pollution and beta-agonist sales for asthma. Chest,
135: 717-723. 192129
July 2009
6-349
DRAFT-DO NOT CITE OR QUOTE

-------
Laurent 0> Pedrono G> Segala Ci Filleul L; Havard Si Deguen Si Schillinger Ci Riviere Ei Bard D. (2008). Air
pollution, asthma attacks, and socioeconomic deprivation- a small-area case-crossover study. Am J
Epidemiol, 168: 58-65. 156672
Lay JCi Bennett WD; Kim CS; Devlin RB; Bromberg PA. (1998). Retention and intracellular distribution of
instilled iron oxide particles in human alveolar macrophages. Am J Respir Cell Mol Biol, 18: 687-695.
007683
Lay JCi Zeman KL; Ghio AJ; Bennett WD. (2001). Effects of inhaled iron oxide particles on alveolar epithelial
permeability in normal subjects. Inhal Toxicol, 13: 1065-1078. 020613
Le Tertre A; Medina S; Samoli E; Forsberg B; Michelozzi P; Boumghar A; Vonk JM; Bellini A; Atkinson R;
Ayres JG; Sunyer J; Schwartz J; Katsouyanni K. (2002). Short term effects of particulate air pollution on
cardiovascular diseases in eight European cities. J Epidemiol Community Health, 56: 773-779. 023746
Le Tertre A; Medina S; Samoli E; Forsberg B; Michelozzi P; Boumghar A; Vonk JM; Bellini A; Atkinson R;
Ayres JG; Sunyer J; Schwartz J; Katsouyanni K. (2003). Short-term effects of particulate air pollution on
cardiovascular diseases in eight European cities. 042820
Le Tertre A; Schwartz J; Touloumi G. (2005). Empirical Bayes and adjusted estimates approach to estimating
the relation of mortality to eposure of PM10. Risk Anal, 25: 711-718. 087560
Lee J-T; Kim H; Song H; Hong Y-C; Cho Y-S; Shin S-Y; Hyun Y-J; Kim Y-S. (2002). Air pollution and asthma
among children in Seoul, Korea. , 13: 481-484. 034826
Lee J-T; Son J-Y; Cho Y"S. (2007). A comparison of mortality related to urban air particles between periods with
Asian dust days and without Asian dust days in Seoul, Korea, 2000-2004. Environ Res, 105: 409-13.
093042
Lee JT; Kim H; Cho YS; Hong YC; Ha EH; Park H. (2003). Air pollution and hospital admissions for ischemic
heart diseases among individuals 64+ years of age residing in Seoul, Korea. , 58: 617-623. 095552
Lee SL; Wong WHS; Lau YL. (2006). Association between air pollution and asthma admission among children in
Hong Kong. Clin Exp Allergy, 36: 1138-1146. 090176
Lei Y-C; Chan C-C; Wang P-Y; Lee C-T; Cheng T-J. (2004). Effects of Asian dust event particles on
inflammation markers in peripheral blood and bronchoalveolar lavage in pulmonary hypertensive rats.
Environ Res, 95: 71-76. 087884
Lei YC; Chen MC; Chan CC; Wang PY; Lee CT; Cheng TJ. (2004). Effects of concentrated ambient particles on
airway responsiveness and pulmonary inflammation in pulmonary hypertensive rats. Inhal Toxicol, 16:
785-792. 087999
LeVine AM; Gwozdz J; Stark J; Bruno M; Whitsett J; Korfhagen T. (1999). Surfactant protein-A enhances
respiratory syncytial virus clearance in vivo. J Clin Invest, 103: 1015-1021. 156687
LeVine AM; Whitsett JA. (2001). Pulmonary collectins and innate host defense of the lung. , 3: 161-166. 155928
Levy D; Sheppard L; Checkoway H; Kaufman J; Lumley T; Koenig J; Siscovick D. (2001). A case-crossover
analysis of particulate matter air pollution and out-of-hospital primary cardiac arrest. , 12: 193-199.
017171
Lewis CW; Klouda GA; Ellenson WD. (2004). Radiocarbon measurement of the biogenic contribution to
summertime PM-2.5 ambient aerosol in Nashville, TN. Atmos Environ, 38: 6053-6061. 097498
Lewis CW; Norris GA; Conner TL; Henry RC. (2003). Source apportionment of Phoenix PM2.5 aerosol with the
Unmix Receptor model. J Air Waste Manag Assoc, 53: 325-338. 088413
Lewis TC; Robins TG; Dvonch JT; Keeler GJ; Yip FY; Mentz GB; Lin X; Parker EA; Israel BA; Gonzalez L; Hill
Y. (2005). Air pollution-associated changes in lung function among asthmatic children in Detroit.
Environ Health Perspect, 113: 1068-1075. 081079
Li N; Wang M; Bramble LA; Schmitz DA; Schauer JJ; Sioutas C; Harkema JR; Nel AE. (2009). The adjuvant
effect of ambient particulate matter is closely reflected by the particulate oxidant potential . Environ
Health Perspect, 117: 1116-1123. 190457
Li Y-J; Kawada T; Matsumoto A; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2007). Airway inflammatory
responses to oxidative stress induced by low-dose diesel exhaust particle exposure differ between mouse
strains. , 33: 227-244. 155929
Liao C-M; Chiang Y-H; Chio C-P. (2009). Assessing the airborne titanium dioxide nanoparticle-related exposure
hazard at workplace. J Hazard Mater, 162: 57"65. 157456
Liao D; Duan Y; Whitsel EA; Zheng Z-J; Heiss G; Chinchilli VM; Lin H-M. (2004). Association of higher levels of
ambient criteria pollutants with impaired cardiac autonomic control: a population-based study. Am J
Epidemiol, 159: 768-777. 056590
Liao D; Heiss G; Chinchilli VM; Duan Y; Folsom AR; Lin HM; Salomaa V. (2005). Association of criteria
pollutants with plasma hemostatic/inflammatory markers: a population-based study. J Expo Sci Environ
Epidemiol, 15: 319-328. 088677
July 2009
6-350
DRAFT-DO NOT CITE OR QUOTE

-------
Liao KJ; Tagaris E; Manomaiphiboon K; Napelenok SL; Woo JH; He S; Amar P; Russell AG. (2007).
Sensitivities of Ozone and Fine Particulate Matter Formation toEmissions under the Impact of Potential
Future Climate Change. Environ Sci Technol, 41- 8355-8361. 180272
Lin M; Chen Y; Burnett RT; Villeneuve PJ; Krewski D. (2002). The influence of ambient coarse particulate
matter on asthma hospitalization in children- case-crossover and time-series analyses. Environ Health
Perspect, 110: 575-581. 026067
Lin M; Stieb DM; Chen Y. (2005). Coarse particulate matter and hospitalization for respiratory infections in
children younger than 15 years in Toronto- a case-crossover analysis. , 116- 235-240. 087828
Linn WS; Szlachcic Y> Gong H Jr', Kinney PL; Berhane KT. (2000). Air pollution and daily hospital admissions
in metropolitan Los Angeles. Environ Health Perspect, 108: 427-434. 002839
Lipfert FW; Baty JD; Miller JP; Wyzga RE. (2006). PM2.5 constituents and related air quality variables as
predictors of survival in a cohort of U.S. military veterans. Inhal Toxicol, 18: 645-657. 088756
Lipfert FW; Morris SC; Wyzga RE. (2000). Daily mortality in the Philadelphia metropolitan area and size-
classified particulate matter. J Air Waste Manag Assoc, 50: 1501-1513. 004088
Lippmann M. (2000). Environmental toxicants: human exposures and their health effects. 024579
Lippmann M; Hwang J; Maclejczyk P; Chen L. (2005). PM source apportionment for short-term cardiac function
changes in ApoE-/- mice. Environ Health Perspect, 113: 1575-1579. 087453
Lippmann M; Ito K; Hwang JS; Maciejczyk P; Chen LC. (2006). Cardiovascular effects of nickel in ambient air.
Environ Health Perspect, 114: 1662-9. 091165
Lippmann M; Ito K; Nadas A; Burnett RT. (2000). Association of particulate matter components with daily
mortality and morbidity in urban populations. 011938
Lipsett MJ; Tsai FC; Roger L; Woo M; Ostro BD. (2006). Coarse particles and heart rate variability among older
adults with coronary artery disease in the Coachella Valley, California. Environ Health Perspect, 114:
1215-1220. 088753
Lisabeth LD; Escobar JD; Dvonch JT; Sanchez BN; Majersik JJ; Brown DL; Smith MA; Morgenstern LB.
(2008). Ambient air pollution and risk for ischemic stroke and transient ischemic attack. , 64: 53"59.
155939
Liu J; Ballaney M; Al-Alem U; Quan C; Jin X; Perera F; Chen LC; Miller RL. (2008). Combined Inhaled Diesel
Exhaust Particles and Allergen Exposure Alter Methylation of T Helper Genes and IgE Production In
Vivo. Toxicol Sci, 102: 76-81. 156709
Liu L; Poon R; Chen L; Frescura AM; Montuschi P; Ciabattoni G; Wheeler A; Dales R. (2009). Acute effects of
air pollution on pulmonary function, airway inflammation, and oxidative stress in asthmatic children.
Environ Health Perspect, 117: 668-674. 192003
Liu L; Ruddy TD; Dalipaj M; Szyszkowicz M; You H; Poon R; Wheeler A; Dales R. (2007). Influence of personal
exposure to particulate air pollution on cardiovascular physiology and biomarkers of inflammation and
oxidative stress in subjects with diabetes. J Occup Environ Med, 49: 258-265. 156705
Liu SX; Hou FF; Guo ZJ; Nagai R; Zhang WR; Liu ZQ; Zhou ZM; Zhou M; Xie D; Wang GB; Zhang X. (2006).
Advanced oxidation protein products accelerate atherosclerosis through promoting oxidative stress and
inflammation. , 26: 1156-1162. 192002
Liu X; Meng Z. (2005). Effects of airborne fine particulate matter on antioxidant capacity and lipid peroxidation
in multiple organs of rats. Inhal Toxicol, 17: 467-473. 088650
Ljungman P; Bellander T; Schneider A; Breitner S; Forastiere F; Hampel R; Illig T; Jacquemin B; Katsouyanni
K; von Klot S. (2009). Modification of the interleukin-6 response to air pollution by interleukin-6 and
fibrinogen polymorphisms. Environ Health Perspect, x: x. 191983
Ljungman PLS; Berglind N> Holmgren C; Gadler F; Edvardsson N; Pershagen G; Rosenqvist M; Sjogren B;
Bellander T. (2008). Rapid effects of air pollution on ventricular arrhythmias. Eur Heart J, 29: 2894"
2901.180266
Llorca J; Salas A; Prieto-Salceda D; Chinchon-Bengoechea V; Delgado-Rodriguez M. (2005). Nitrogen dioxide
increases cardiorespiratory admissions in Torrelavega (Spain). J Environ Health, 68: 30-35. 087825
Lokken PR; Wellenius GA; Coull BA; Burger MR; Schlaug G; Suh HH; Mittleman MA. (2009). Air Pollution and
Risk of Stroke: Underestimation of Effect Due to Misclassification of Time of Event Onset. Epidemiology,
20: 137. 186774
Low RB; Bielory L; Qureshi AI; Dunn V; Stuhlmiller DF; Dickey DA. (2006). The relation of stroke admissions
to recent weather, airborne allergens, air pollution, seasons, upper respiratory infections, and asthma
incidence, September 11, 2001, and day of the week. , 37: 951-957. 090441
Lucking A; Lundback M; Mills N; Faratian D; Barath S; Pourazar J; Cassee F; Donaldson K; Boon N; Badimon
J. (2008). Diesel exhaust inhalation increases thrombus formation in man. Eur Heart J, 29: 3043. 191993
July 2009
6-351
DRAFT-DO NOT CITE OR QUOTE

-------
Luginaah IN; Fung KY; Gorey KM; Webster GJ Wills C. (2005). Association of ambient air pollution with
respiratory hospitalization in a government designated "area of concern"' the case of Windsor, Ontario.
Environ Health Perspect, 113- 290-296. 057327
Lund AE Lucero A; Lucas Si Madden MCi McDonald JDi Seagrave JCi Knuckles TLi Campen MJ. (2009).
Vehicular emissions induce vascular MMP-9 expression and activity associated with endothelial—
mediated pathways. Arterioscler Thromb Vase Biol, 29- 511-517. 180257
Lundback Mi Mills NLi Lucking Ai Barath Si Donaldson Ki Newby DEi Sandstrom Ti Blomberg A. (2009).
Experimental exposure to diesel exhaust increases arterial stiffness in man. Part Fibre Toxicol, 6- 7.
191967
Luttmann-Gibson Hi Suh HHi Coull BAi Dockery DWi Sarnet SEi Schwartz Ji Stone PHi Gold DR. (2006).
Short-term effects of air pollution on heart rate variability in senior adults in Steubenville, Ohio. J Occup
Environ Med, 48: 780-788. 089794
Maciejczyk Pi Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in
mice: VIII source-related daily variations in in vitro responses to CAPs. Inhal Toxicol, 17: 243-253.
087456
Mar TFi Ito Ki Koenig JQ; Larson TV; Eatough DJ; Henry RC; Kim E> Laden F; Lall R; Neas Li Stolzel Mi
Paatero Pi Hopke PKi Thurston GD. (2006). PM source apportionment and health effects: 3 Investigation
of inter-method variations in associations between estimated source contributions of PM25 and daily
mortality in Phoenix, AZ. J Expo Sci Environ Epidemiol, Epidemiol.16: 311-320. 086143
Mar TF; Jansen K; Shepherd K; Lumley T; Larson TV; Koenig JQ. (2005). Exhaled nitric oxide in children with
asthma and short-term PM25 exposure in Seattle. Environ Health Perspect, 113: 1791-1794. 088759
Mar TF; Koenig JQ; Jansen K; Sullivan J; Kaufman J; Trenga CA; Siahpush SH; Liu L-JS; Neas L. (2005). Fine
particulate air pollution and cardiorespiratory effects in the elderly. , 16: 681-687. 087566
Mar TF; Larson TV; Stier RA; Claiborn C; Koenig JQ. (2004). An analysis of the association between respiratory
symptoms in subjects with asthma and daily air pollution in Spokane, Washington. Inhal Toxicol, 16:
809-815. 057309
Mar TF; Norris GA; Koenig JQ; Larson TV. (2000). Associations between air pollution and mortality in Phoenix,
1995-1997. Environ Health Perspect, 108: 347-353. 001760
Mar TF; Norris GA; Larson TV; Wilson WE; Koenig JQ. (2003). Air pollution and cardiovascular mortality in
Phoenix, 1995-1997. 042841
Martins LC; Latorre MRDO; Saldiva PHN; Braga ALF. (2002). Air pollution and emergency room visits due to
chronic lower respiratory diseases in the elderly: an ecological time-series study in Sao Paulo, Brazil. J
Occup Environ Med, 44: 622-627. 035059
Masjedi MR; Jamaati HR; Dokouhaki P; Ahmadzadeh Z; Taheri SA; Bigdeli M; Izadi S; Rostamian A; Aagin K;
Ghavam SM. (2003). The effects of air pollution on acute respiratory conditions. Respirology, 8: 213-230.
052100
Matsumoto A; Hiramatsu K; Li Y; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2006). Repeated exposure to
low-dose diesel exhaust after allergen challenge exaggerates asthmatic responses in mice. , 121: 227-235.
098017
McCreanor J; Cullinan P; Nieuwenhuijsen MJ; Stewart-Evans J; Malliarou E; Jarup L; Harrington R;
Svartengren M; Han I"K; Ohman-Strickland P; Chung KF; Zhang J. (2007). Respiratory effects of
exposure to diesel traffic in persons with asthma. , 357: 2348-2358. 092841
McDonald J; Barr E; White R; Kracko D; Chow J; Zielinska B; Grosjean E. (2008). Generation and
characterization of gasoline engine exhaust inhalation exposure atmospheres. Inhal Toxicol, 20: 1157-
1168.191978
McDonald JD; Barr EB; White RE Chow JC; Schauer JJ; Zielinska B; Grosjean E. (2004). Generation and
characterization of four dilutions of diesel engine exhaust for a subchronic inhalation study. Environ Sci
Technol, 38: 2513-2522. 055644
McQueen DS; Donaldson K; Bond SM; McNeilly JD; Newman S; Barton NJ; Duffin R. (2007). Bilateral
vagotomy or atropine pre-treatment reduces experimental diesel-soot induced lung inflammation. Toxicol
Appl Pharmacol, 219: 62-71. 096266
Medina-Ramon M; Zanobetti A; Schwartz J. (2006). The effect of ozone and PM10 on hospital admissions for
pneumonia and chronic obstructive pulmonary disease: a national multicity study. Am J Epidemiol, 163:
579-588. 087721
Metzger KB; Klein M; Flanders WD; Peel JL; Mulholland JA; Langberg JJ; Tolbert PE. (2007). Ambient air
pollution and cardiac arrhythmias in patients with implantable defibrillators. Epidemiology, 18: 585-592.
092856
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd KH; Mulholland JA; Ryan PB; Frumkin H.
(2004). Ambient air pollution and cardiovascular emergency department visits. , 15: 46"56. 044222
July 2009
6-352
DRAFT-DO NOT CITE OR QUOTE

-------
Michaud J-P; Grove JS; Krupitsky D. (2004). Emergency department visits and "vog"-related air quality in Hilo,
Hawai'i. Environ Res, 95: 11-19. 089900
Middleton N> Yiallouros P; KLeanthous S> Kolokotroni Oi Schwartz J; Dockery DWi Demokritou P; Koutrakis P.
(2008). A 10-year time-series analysis of respiratory and cardiovascular morbidity in Nicosia, Cyprus: the
effect of short-term changes in air pollution and dust storms. , T- 39. 156760
Migliaretti Gi Cadum Ei Migliore Ei Cavallo F. (2005). Traffic air pollution and hospital admission for asthma: a
case-control approach in a Turin (Italy) population. Int Arch Occup Environ Health, 78: 164-169. 088689
Migliaretti Gi Cavallo F. (2004). Urban air pollution and asthma in children. Pediatr Pulmonol, 38: 198-203.
087425
Mills NLi Robinson SD; Fokkens PHi Leseman DLi Miller MRi Anderson D> Freney EJ; Heal MR; Donovan RJ;
Blomberg A; Sandstrom T; MacNee W; Boon NA; Donaldson K; Newby DE; Cassee FR. (2008). Exposure
to concentrated ambient particles does not affect vascular function in patients with coronary heart
disease. Environ Health Perspect, 116; 709-715. 156766
Mills NL; Tornqvist H; Robinson SD; Gonzalez M; Darnley K; MacNee W; Boon NA; Donaldson K; Blomberg A;
Sandstrom T; Newby DE. (2005). Diesel exhaust inhalation causes vascular dysfunction and impaired
endogenous fibrinolysis. , 112: 3930-3936. 095757
Mills NL; Tornqvist H; Gonzalez MC; Vink E; Robinson SD; Soderberg S; Boon NA Donaldson K; Sandstrom T;
Blomberg A; Newby DE. (2007). Ischemic and thrombotic effects of dilute diesel-exhaust inhalation in
men with coronary heart disease. , 357: 1075-1082. 091206
Min KB; Min JY; Cho SI; Paek D. (2008). The relationship between air pollutants and heart-rate variability
among community residents in Korea. Inhal Toxicol, 4- 435-444. 191901
Mohr LB; Luo S; Mathias E; Tobing R; Homan S; Sterling D. (2008). Influence of season and temperature on the
relationship of elementalcarbon air pollution to pediatric asthma emergency room visits. J Asthma, 45:
936-943. 180215
Moolgavkar SH. (2000). Air pollution and daily mortality in three US counties. Environ Health Perspect, 108:
777-784. 012054
Moolgavkar SH. (2000). Air pollution and hospital admissions for diseases of the circulatory system in three US
metropolitan areas. J Air Waste Manag Assoc, 50: 1199-1206. 010305
Moolgavkar SH. (2003). Air pollution and daily deaths and hospital admissions in Los Angeles and Cook
counties. 042864
Moolgavkar SH. (2003). Air pollution and daily mortality in two US counties: season-specific analyses and
exposure-response relationships. Inhal Toxicol, 15: 877-907. 051316
Moore KJ; Kunjathoor W; Koehn SL; Manning JJ; Tseng AA; Silver JM; McKee M; Freeman MW. (2005). Loss
of receptor-mediated lipid uptake via scavenger receptor A or CD36 pathways does not ameliorate
atherosclerosis in hyperlipidemic mice. J Clin Invest, 115: 2192-2201. 156780
Morishita M; Keeler G; Wagner J; Marsik F; Timm E; Dvonch J; Harkema J. (2004). Pulmonary retention of
particulate matter is associated with airway inflammation in allergic rats exposed to air pollution in
urban Detroit. Inhal Toxicol, 16: 663-674. 087979
Morris RD; Naumova EN. (1998). Carbon monoxide and hospital admissions for congestive heart failure:
evidence of an increased effect at low temperatures. Environ Health Perspect, 106: 649-653. 086857
Morrison D; Skwarski D; Millar AM; Adams W', MacNee W. (1998). A comparison of three methods of measuring
99mTc-DTPA lung clearance and their repeatability. Eur Respir J, 11: 1141-1146. 024924
Mortimer KM; Neas LM; Dockery DW; Redline S; Tager IB. (2002). The effect of air pollution on inner-city
children with asthma. Eur Respir J, 19: 699-705. 030281
Moshammer H; Hutter H"P; Hauck H; Neuberger M. (2006). Low levels of air pollution induce changes of lung
function in a panel of schoolchildren. Eur Respir J, 27: 1138-1143. 090771
Moshammer H; Neuberger M. (2003). The active surface of suspended particles as a predictor of lung function
and pulmonary symptoms in Austrian school children. Atmos Environ, 37: 1737-1744. 041956
Mudway IS; Stenfors N; Duggan ST; Roxborough H; Zielinski H; Marklund SL; Blomberg A; Frew AJ;
Sandstrom T; Kelly FJ. (2004). An in vitro and in vivo investigation of the effects of diesel exhaust on
human airway lining fluid antioxidants. Arch Biochem Biophys, 423: 200-212. 180208
Mudzinski SP; Rudofsky UH; Mitchell DG; Lawrence DA. (1986). Analysis of lead effects on in vivo antibody-
mediated immunity in several mouse strains. Toxicol Appl Pharmacol, 83: 321-330. 084184
Muggenburg BA; Barr EB; Cheng YS; Seagrave JC; Tilley LP; Mauderly JL. (2000). Effect of inhaled residual
oil fly ash on the electrocardiogram of dogs. Presented at In: Grant, L. D., ed. PM2000: particulate matter
and health. Inhalation Toxicol. 12(suppl. 4): 189-208. 010279
Murata A; Kida K; Hasunuma H; Kanegae H; Ishimaru Y; Motegi T; Yamada K; Yoshioka H; Yamamoto K;
Kudoh S. (2007). Environmental influence on the measurement of exhaled nitric oxide concentration in
school children: special reference to methodology. , 74: 30-36. 156787
July 2009
6-353
DRAFT-DO NOT CITE OR QUOTE

-------
Mutlu GM; Green D; Bellmeyer A; Baker CM; Burgess Z; Rajamannan N; Christman JW; Foiles N; Kamp DW;
Ghio AJ; Chandel NS; Dean DA; Sznajder JI; Budinger GR. (2007). Ambient particulate matter
accelerates coagulation via an IL"6"dependent pathway. , 117- 2952-61. 121441
Menache MG; Miller FJ; Raabe OG. (1995). Particle inhalability curves for humans and small laboratory
animals. Ann Occup Hyg, 39- 317-328. 006533
Nadziejko CJ Fang K; Chen LC; Cohen B; Karpatkin M; Nadas A. (2002). Effect of concentrated ambient
particulate matter on blood coagulation parameters in rats. 050587
Nadziejko Ci Fang K> Nadziejko E> Narciso SP; Zhong Mi Chen LC. (2002). Immediate effects of particulate air
pollutants on heart rate and respiratory rate in hypertensive rats. , 2- 245-252. 087460
Nadziejko Ci Fang Ki Narciso Si Zhong Mi Su WCi Gordon Ti Nadas Ai Chen LC. (2004). Effect of particulate
and gaseous pollutants on spontaneous arrhythmias in aged rats. Inhal Toxicol, 16- 373-380. 055632
Nascimento LFi Pereira LAi Braga ALi Modolo MCi Carvalho JA Jr. (2006). Effects of air pollution on children's
health in a city in southeastern Brazil. Rev Saude Publica, 40: 77-82. 093247
NCHS. (2007). Classification of Diseases, Functioning, and Disability...Monitoring the Nation's Health.
Retrieved , from #http7/www.cdc.gov/nchs/icd9.htm#RTF#. 157194
Neuberger Mi Schimek MGi Horak F Jri Moshammer Hi Kundi Mi Frischer Ti Gomiscek Bi Puxbaum Hi Hauck
Hi AUPHEP-Team. (2004). Acute effects of particulate matter on respiratory diseases, symptoms and
functions: epidemiological results of the Austrian Projects on Health Effects of Particulate Matter
(AUPHEP). Atmos Environ, 38: 3971-3981. 093249
New York State Department of Health. (2006). Study of ambient air contaminants and asthma in New York
City, parts A and B. 090132
Nightingale JAi Maggs Ri Cullinan Pi Donnelly LEi Rogers DFi Kinnersley Ri Chung KFi Barnes PJi Ashmore
Mi Newman-Taylor A. (2000). Airway inflammation after controlled exposure to diesel exhaust
particulates. Am J Respir Crit Care Med, 162: 161-166. 011659
Nikolov MCi Coull BAi Catalano PJi Diaz Ei Godleski JJ. (2008). Statistical methods to evaluate health effects
associated with major sources of air pollution: a case-study of breathing patterns during exposure to
concentrated Boston air particles. J Roy Stat Soc C Appl Stat, 57: 357-378. 156808
Nordenhall Ci Pourazar Ji Ledin M"Ci Levin J-Oi Sandstrom Ti Adelroth E. (2001). Diesel exhaust enhances
airway responsiveness in asthmatic subjects. Eur Respir J, 17: 909-915. 025185
Nurkiewicz TRi Porter DWi Barger Mi Castranova Vi Boegehold MA. (2004). Particulate matter exposure
impairs systemic microvascular endothelium-dependent dilation. Environ Health Perspect, 112: 1299"
1306. 087968
Nurkiewicz TRi Porter DWi Barger Mi Millecchia Li Rao KMi Marvar PJi Hubbs AFi Castranova Vi Boegehold
MA. (2006). Systemic microvascular dysfunction and inflammation after pulmonary particulate matter
exposure. Environ Health Perspect, 114: 412-419. 088611
Nurkiewicz TRi Porter DWi Hubbs AFi Cumpston JLi Chen BTi Frazer DGi Castranova V. (2008). Nanoparticle
inhalation augments particle-dependent systemic microvascular dysfunction. Part Fibre Toxicol, 5- 1.
156816
Nurkiewicz TRi Porter DWi Hubbs AFi Stone Si Chen BTi Frazer DGi Boegehold MAi Castranova V. (2009).
Pulmonary nanoparticle exposure disrupts systemic microvascular nitric oxide signaling. , 110: 191-203.
191961
Nygaard UCi Samuelsen Mi Aase Ai Lovik M. (2004). The capacity of particles to increase allergic sensitization
is predicted by particle number and surface area, not by particle mass. Toxicol Sci, 82: 515-524. 058558
O'Connor GTi Neas Li Vaughn Bi Kattan Mi Mitchell Hi Crain EFi Evans R, 3rd; Gruchalla R; Morgan W', Stout
J; Adams GK; Lippmann M. (2008). Acute respiratory health effects of air pollution on children with
asthma in US inner cities. J Allergy Clin Immunol, 121: 1133-1139 e!131. 156818
O'Neill MS; Hajat S; Zanobetti A; Ramirez-Aguilar M; Schwartz J. (2005). Impact of control for air pollution and
respiratory epidemics on the estimated associations of temperature and daily mortality. Int J
Biometeorol, 50: 121-9. 098094
O'Neill MS; Veves A; Sarnat JA; Zanobetti A; Gold DR; Economides PA; Horton ES; Schwartz J. (2007). Air
pollution and inflammation in type 2 diabetes: a mechanism for susceptibility. Occup Environ Med, 64:
373-379. 091362
O'Neill MS; Veves A; Zanobetti A; Sarnat JA; Gold DR; Economides PA; Horton ES; Schwartz J. (2005).
Diabetes enhances vulnerability to particulate air pollution-associated impairment in vascular reactivity
and endothelial function. , 111: 2913-2920. 088423
Odajima H; Yamazaki S; Nitta H. (2008). Decline in peak expiratory flow according to hourly short-term
concentration of particulate matter in asthmatic children. Inhal Toxicol, 20: 1263. 192005
July 2009
6-354
DRAFT-DO NOT CITE OR QUOTE

-------
Oftedal B; Nafstad P; Magnus P; Bjorkly Si Skrondal A. (2003). Traffic related air pollution and acute hospital
admission for respiratory diseases in Drammen, Norway 1995-2000. Eur J Epidemiol, 18- 671-675.
055623
Okin PMi Devereux RB; Howard BVi Fabsitz RR; Lee ETi Welty TK. (2000). Assessment of QT interval and QT
dispersion for prediction of all-cause and cardiovascular mortality in American Indians' The Strong
Heart Study. , 101: 61-66. 156002
Ostro B. (1995). Fine particulate air pollution and mortality in two Southern California counties. Environ Res,
70' 98-104. 079197
Ostro Bi Broadwin R; Green Si Feng W"Yi Lipsett M. (2006). Fine particulate air pollution and mortality in nine
California counties: results from CALFINE. Environ Health Perspect, 114: 29"33. 087991
Ostro Bi Feng W"Yi Broadwin Ri Green Si Lipsett M. (2007). The effects of components of fine particulate air
pollution on mortality in California: results from CALFINE. Environ Health Perspect, 115: 13"9. 091354
Ostro Bi Roth Li Malig Bi Marty M. (2009). The effects of fine particle components on respiratory hospital
admissions in children. Environ Health Perspect, 117: 475. 191971
Ostro BDi Broadwin Ri Lipsett MJ. (2003). Coarse particles and daily mortality in Coachella Valley, California.
042824
Ostro BDi Feng WY; Broadwin R; Malig BJ; Green RS; Lipsett MJ. (2008). The impact of components of fine
particulate matter on cardiovascular mortality in susceptible subpopulations. Occup Environ Med, 65:
750-756. 097971
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Schwartz J. (2005). Effects of air pollution on heart rate
variability^ the VA normative aging study. Environ Health Perspect, 113: 304-309. 057331
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Spiro A, 3rd; Tucker KL; Suh H; Hu H; Schwartz J. (2008).
Traffic-related particles are associated with elevated homocysteine: the VA normative aging study. Am J
Respir Crit Care Med, 178: 283-289. 156845
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Wright RO; Coull B; Nie H; Hu H; Schwartz J. (2008). Air
pollution and heart rate variability: effect modification by chronic lead exposure. , 19: 111-120. 093027
Park SK; O'Neill MS; Wright RO; Hu H; Vokonas PS; Sparrow D; Suh H; Schwartz J. (2006). HFE genotype,
particulate air pollution, and heart rate variability A gene-environmental interaction. , 114: 2798-2805.
091245
Peacock JL; Symonds P; Jackson P; Bremner SA; Scarlett JF; Strachan DP; Anderson HR. (2003). Acute effects
of winter air pollution on respiratory function in schoolchildren in southern England. Occup Environ
Med, 60: 82-89. 042026
Peel JL; Metzger KB; Klein M; Flanders WD; Mulholland JA; Tolbert PE. (2007). Ambient air pollution and
cardiovascular emergency department visits in potentially sensitive groups. Am J Epidemiol, 165: 625"
633. 090442
Peel JL; Tolbert PE; Klein M; Metzger KB; Flanders WD; Knox T; Mulholland JA; Ryan PB; Frumkin H. (2005).
Ambient air pollution and respiratory emergency department visits. , 16: 164-174. 056305
Pekkanen J; Peters A; Hoek G; Tiittanen P; Brunekreef B; de Hartog J; Heinrich J; Ibald-Mulli A; Kreyling WG!
Lanki T; Timonen KL; Vanninen E. (2002). Particulate air pollution and risk of ST-segment depression
during repeated submaximal exercise tests among subjects with coronary heart disease: the exposure
and risk assessment for fine and ultrafine particles in ambient air (ULTRA) study. , 106: 933-938. 035050
Peled R; Friger M; Bolotin A; Bibi H; Epstein L; Pilpel D; Scharf S. (2005). Fine particles and meteorological
conditions are associated with lung function in children with asthma living near two power plants. Public
Health, 119: 418-425. 156015
Peng R; Bell M; Geyh A; McDermott A; Zeger S; Samet J; Dominici F. (2009). Emergency admissions for
cardiovascular and respiratory diseases and the chemical composition of fine particle air pollution.
Environ Health Perspect, 117: 957-963. 191998
Peng RD; Chang HH; Bell ML; McDermott A; Zeger SL; Samet JM; Dominici F. (2008). Coarse particulate
matter air pollution and hospital admissions for cardiovascular and respiratory diseases among Medicare
patients. JAMA, 299: 2172-2179. 156850
Peng RD; Dominici F; Pastor-Barriuso R; Zeger SL; Samet JM. (2005). Seasonal analyses of air pollution and
mortality in 100 US cities. Am J Epidemiol, 161: 585-594. 087463
Peng RD;Chang HH;Bell ML;McDermott A;Zeger SL;Samet JM;Dominici F. (2008). Coarse particulate matter
air pollution and hospital admissions for cardiovascular and respiratory diseases among Medicare
patients. , 299: 2172-9. 186787
Pennanen AS; Sillanpaa M; Hillamo R; Quass U; John AC; Branis MJ Hunova I; Meliefste K; Janssen NA;
Koskentalo T; Castano-Vinyals G; Bouso L; Chalbot MC; Kavouras IG; Salonen RO. (2007). Performance
of a high-volume cascade impactor in six European urban environments: mass measurement and
chemical characterization of size-segregated particulate samples. , 374: 297-310. 155357
July 2009
6-355
DRAFT-DO NOT CITE OR QUOTE

-------
Penttinen P; Vallius M> Tiittanen P; Ruuskanen J; Pekkanen J. (2006). Source-specific fine particles in urban
air and respiratory function among adult asthmatics. Inhal Toxicol, 18- 191-198. 087988
Pereira CEL; Heck TG; Saldiva PHN; Rhoden CR. (2007). Ambient particulate air pollution from vehicles
promotes lipid peroxidation and inflammatory responses in rat lung. , 40- 1353-1359. 156019
Peretz A; Kaufman JD; Trenga CA; Allen J; Carlsten C> Aulet MR; Adar SD; Sullivan JH. (2008). Effects of
diesel exhaust inhalation on heart rate variability in human volunteers. Environ Res, 107- 178-184.
156855
Peretz A; Peck EC; Bammler TK; Beyer RP; Sullivan JH; Trenga CA; Srinouanprachnah S; Farin FM; Kaufman
JD. (2007). Diesel exhaust inhalation and assessment of peripheral blood mononuclear cell gene
transcription effects' an exploratory study of healthy human volunteers. Inhal Toxicol, 19- 1107-1119.
156853
Peretz A; Sullivan JH; Leotta DF; Trenga CA; Sands FN; Allen J; Carlsten C; Wilkinson CW; Gill EA; Kaufman
JD. (2008). Diesel exhaust inhalation elicits acute vasoconstriction in vivo. Environ Health Perspect, 116:
937-942. 156854
Perez L; Tobias A; Querol X; Kunzli N; Pey J; Alastuey A; Viana M; Valero N; Gonzalez-Cabre M; Sunyer J.
(2008).	Coarse particles from Saharan dust and daily mortality. , 19: 800-807. 156020
Peters A. (2005). Particulate matter and heart disease: evidence from epidemiological studies. Toxicol Appl
Pharmacol, 207: S477-S482. 087759
Peters A; Dockery DW! Muller JE; Mittleman MA. (2001). Increased particulate air pollution and the triggering
of myocardial infarction. , 103: 2810-2815. 016546
Peters A; Greven S; Heid I; Baldari F; Breitner S; Bellander T; Chrysohoou C; Illig T; Jacquemin B; Koenig W.
(2009).	Fibrinogen Genes Modify the Fibrinogen Response to Ambient Particulate Matter. Am J Respir
Crit Care Med, 179: 484-491. 191992
Peters A; Liu E; Verrier RL; Schwartz J; Gold DR; Mittleman M; Baliff J; Oh JA; Allen G; Monahan K; Dockery
DW. (2000). Air pollution and incidence of cardiac arrhythmia. , 11: 11-17. 011347
Peters A; Von Klot S; Heier M; Trentinaglia I; Hormann A; Wichmann HE; Lowel H. (2004). Exposure to traffic
and the onset of myocardial infarction. , 351: 1721-1730. 087464
Petrovic S; Urch B; Brook J; Datema J; Purdham J; Liu L; Lukic Z; Zimmerman B; Tofler G; Downar E; Corey
P; Tarlo S; Broder I; Dales R; Silverman F. (2000). Cardiorespiratory effects of concentrated ambient
PM25: a pilot study using controlled human exposures. Inhal Toxicol, l: 173-188. 004638
Piedrahita JA; Zhang SH; Hagaman JR; Oliver PM; Maeda N. (1992). Generation of mice carrying a mutant
apolipoprotein E gene inactivated by gene targeting in embryonic stem cells. Presented at Proceedings of
the National Academy of Science, 5/15/1992. 156868
Pietropaoli AP; Frampton MWi Hyde RW; Morrow PE; Oberdorster G; Cox C; Speers DM; Frasier LM; Chalupa
DC; Huang LS; Utell MJ. (2004). Pulmonary function, diffusing capacity, and inflammation in healthy
and asthmatic subjects exposed to ultrafine particles. Inhal Toxicol, 16: 59"72. 156025
Pinkerton KE; Zhou Y; Teague SV; Peake JL; Walther RC; Kennedy IM; Leppert VJ; Aust AE. (2004). Reduced
lung cell proliferation following short-term exposure to ultrafine soot and iron particles in neonatal rats:
key to impaired lung growth?. Inhal Toxicol, l: 73-81. 087465
Pinkerton KE; Zhou Y; Zhong C; Smith KR; Teague SV; Kennedy IM; Menache MG. (2008). Mechanisms of
particulate matter toxicity in neonatal and young adult rat lungs. Health Effects Institute. Boston, MA.
135. 190471
Pirrone N; Keeler GJ; Warner PO. (1995). Trends of ambient concentrations and deposition fluxes of particulate
trace metals in Detroit from 1982 to 1992. Sci Total Environ, 162: 43-61. 089297
Pope C; Renlund D; Kfoury A; May H; Home B. (2008). Relation of heart failure hospitalization to exposure to
fine particulate air pollution. , 102: 1230-1234. 191969
Pope CA; Hansen ML; Long RW; Nielsen KR; Eatough NL; Wilson WE; Eatough DJ. (2004). Ambient
particulate air pollution, heart rate variability, and blood markers of inflammation in a panel of elderly
subjects. Environ Health Perspect, 112: 339-345. 055238
Pope CA III; Muhlestein JB; May HT; Renlund DG; Anderson JL; Home BD. (2006). Ischemic heart disease
events triggered by short-term exposure to fine particulate air pollution. Circulation, 114: 2443-2448.
091246
Pourazar J; Blomberg A; Kelly FJ; Davies DE; Wilson SJ; Holgate ST; Sandstrom T. (2008). Diesel exhaust
increases EGFR and phosphorylated C-terminal Tyr 1173 in the bronchial epithelium. Part Fibre
Toxicol, 5:8. 156884
Pourazar J; Mudway IS; Samet JM; Helleday R; Blomberg A; Wilson SJ; Frew AJ; Kelly FJ; Sandstrom T.
(2005). Diesel exhaust activates redox-sensitive transcription factors and kinases in human airways. Am
J Physiol, 289: L724-L730. 088305
July 2009
6-356
DRAFT-DO NOT CITE OR QUOTE

-------
Power E Balmes J; Solomon C. (2008). Controlled exposure to combined particles and ozone decreases heart
rate variability. J Occup Environ Med, 50: 1253. 191982
Preutthipan A; Udomsubpayakul U> Chaisupamongkollarp Ti Pentamwa P. (2004). Effect of PM10 pollution in
Bangkok on children with and without asthma. Pediatr Pulmonol, 37: 187-192. 055598
Pritchett LC; Cooper JA. (1985). Aerosol Characterization Study of Anchorage, Alaska: Chemical Analysis and
Source Apportionment. 156886
Prystowsky EN; Benson DW Jr', Fuster V; Hart RG; Kay GN; Myerburg RJ; Naccarelli GV; Wyse DG. (1996).
Management of patients with atrial fibrillation. A Statement for Healthcare Professionals. From the
Subcommittee on Electrocardiography and Electrophysiology, American Heart Association. , 93: 1262"
1277. 156031
Raabe OG; Al-Bayati MA; Teague SV; Rasolt A. (1988). Regional deposition of inhaled monodisperse, coarse,
and fine aerosol particles in small laboratory animals. In Inhaled particles VP Proceedings of an
international symposium and workshop on lung dosimetry (pp. 53"63). Cambridge, United Kingdom:
Pergamon Press. 001439
Rabinovitch N; Strand M; Gelfand EW. (2006). Particulate levels are associated with early asthma worsening in
children with persistent disease. Am J Respir Crit Care Med, 173: 1098-1105. 088031
Rabinovitch N; Zhang LN; Murphy JR; Vedal S; Dutton SJ; Gelfand EW. (2004). Effects of wintertime ambient
air pollutants on asthma exacerbations in urban minority children with moderate to severe disease. J
Allergy Clin Immunol, 114: 1131-1137. 096753
Ranzi A; Gambini M; Spattini A; Galassi C; Sesti D; Bedeschi M; Messori A; Baroni A; Cavagni G; Lauriola P.
(2004).	Air pollution and respiratory status in asthmatic children: hints for a locally based preventive
strategy AIRE study. Eur J Epidemiol, 19: 567-576. 089500
Reed MD; Barrett EG; Campen MJ; Divine KK; Gigliotti AP; McDonald JD; Seagrave JC; Mauderly JL; Seilkop
SK; Swenberg JA. (2008). Health effects of subchronic inhalation exposure to gasoline engine exhaust.
Inhal Toxicol, 20: 1125-1143. 156903
Reed MD; Campen MJ; Gigliotti AP; Harrod KS; McDonald JD; Seagrave JC; Mauderly JL; Seilkop SK. (2006).
Health effects of subchronic exposure to environmental levels of hardwood smoke. Inhal Toxicol, 18: 523"
539. 156043
Reed MD; Gigliotti AP; McDonald JD; Seagrave JC; Seilkop SK; Mauderly JL. (2004). Health effects of
subchronic exposure to environmental levels of diesel exhaust. Inhal Toxicol, 16: 177-193. 055625
Rhoden CR; Ghelfi E; Gonzalez-Flecha B. (2008). Pulmonary inflammation by ambient air particles is mediated
by superoxide anion. Inhal Toxicol, 20: 11-15. 190475
Rhoden CR; Lawrence J; Godleski JJ; Gonzalez-Flecha B. (2004). N-acetylcysteine prevents lung inflammation
after short-term inhalation exposure to concentrated ambient particles. Toxicol Sci, 79: 296-303. 087969
Rhoden CR; Wellenius GA; Ghelfi E; Lawrence J; Gonzalez-Flecha B. (2005). PM"induced cardiac oxidative
stress and dysfunction are mediated by autonomic stimulation. Biochim Biophys Acta, 1725: 305-313.
087878
Rich DQ; Freudenberger RS; Ohman-Strickland P; Cho Y; Kipen HM. (2008). Right heart pressure increases
after acute increases in ambient particulate concentration. Environ Health Perspect, 116: 1167-1171.
156910
Rich DQ; Kim MH; Turner JR; Mittleman MA; Schwartz J; Catalano PJ; Dockery DW. (2006). Association of
ventricular arrhythmias detected by implantable cardioverter defibrillator and ambient air pollutants in
the St Louis, Missouri metropolitan area. Occup Environ Med, 63: 591-596. 089814
Rich DQ; Mittleman MA; Link MS; Schwartz J; Luttmann-Gibson H; Catalano PJ; Speizer FE; Gold DR;
Dockery DW. (2006). Increased risk of paroxysmal atrial fibrillation episodes associated with acute
increases in ambient air pollution. Environ Health Perspect, 114: 120-123. 088427
Rich DQ; Schwartz J; Mittleman MA; Link M; Luttmann-Gibson H; Catalano PJ; Speizer FE; Dockery DW.
(2005).	Association of short-term ambient air pollution concentrations and ventricular arrhythmias. Am J
Epidemiol, 161: 1123-1132. 079620
Rich KE; Petkau J; Vedal S; Brauer M. (2004). A case-crossover analysis of particulate air pollution and cardiac
arrhythmia in patients with implantable cardioverter defibrillators. Inhal Toxicol, 16: 363-372. 055631
Riediker M; Cascio WE; Griggs TR; Herbst MC; Bromberg PA; Neas L; Williams RW; Devlin RB. (2004).
Particulate matter exposure in cars is associated with cardiovascular effects in healthy young men. Am J
Respir Crit Care Med, 169: 934-940. 056992
Riediker M; Devlin RB; Griggs TR; Herbst MC; Bromberg PA; Williams RW; Cascio WE. (2004). Cardiovascular
effects in patrol officers are associated with fine particulate matter from brake wear and engine
emissions. Part Fibre Toxicol, l: 2. 091261
July 2009
6-357
DRAFT-DO NOT CITE OR QUOTE

-------
Riojas-Rodriguez Hi Escamilla-Cejudo JA; Gonzalez-Hermosillo JA; Tellez-Rojo MM; Vallejo Mi Santos-Burgoa
Ci Rojas-Bracho L. (2006). Personal PM2.5 and CO exposures and heart rate variability in subjects with
known ischemic heart disease in Mexico City. J Expo Sci Environ Epidemiol, 16- 131-137. 156913
Rodriguez Ci Tonkin Ri Heyworth J; Kusel Mi De Klerk Ni Sly PDi Franklin Pi Runnion Ti Blockley A; Landau
Li Hinwood AL. (2007). The relationship between outdoor air quality and respiratory symptoms in young
children. Int J Environ Health Res. 092842
Romieu Ii Garcia-Esteban Ri Sunyer Ji Rios Ci Alcaraz-Zubeldia Mi Velasco SRi Holguin F. (2008). The effect of
supplementation with omega-3 polyunsaturated fatty acids on markers of oxidative stress in elderly
exposed to PM(2.5). Environ Health Perspect, 116: 1237-1242. 156922
Romieu Ii Tellez-Rojo MMi Lazo Mi Manzano-Patino Cortez-Lugo Mi Julien Pi Belanger MCi Hernandez-Avila
MHolguin F. (2005). Omega-3 fatty acid prevents heart rate variability reductions associated with
particulate matter. Am J Respir Crit Care Med, 172: 1534-1540. 086297
Rosenthal FSi Carney JPi Olinger ML. (2008). Out-of-hospital cardiac arrest and airborne fine particulate
matter: a case-crossover analysis of emergency medical services data in Indianapolis, Indiana. Environ
Health Perspect, 116: 631-636. 156925
Routledge HCi Manney Si Harrison RMi Ayres JGi Townend JN. (2006). Effect of inhaled sulphur dioxide and
carbon particles on heart rate variability and markers of inflammation and coagulation in human
subjects. Heart, 92: 220-227. 088674
Rowan 3rd WHi Campen MJi Wichers LBi Watkinson WP. (2007). Heart rate variability in rodents: uses and
caveats in toxicological studies. Cardiovasc Toxicol, T- 28-51. 191911
Roy Di Talajic Mi Dubuc Mi Thibault Bi Guerra Pi Macle Li Khairy P. (2009). Atrial fibrillation and congestive
heart failure. Curr Opin Cardiol, 24: 29. 190902
Ruckerl Ri Greven Si Ljungman Pi Aalto Pi Antoniades Ci Bellander Ti Berglind Ni Chrysohoou Ci Forastiere Fi
Jacquemin Bi von Klot Si Koenig Wi Kuchenhoff Hi Lanki Ti Pekkanen Ji Perucci CAi Schneider Ai
Sunyer Ji Peters A. (2007). Air pollution and inflammation (interleukin-6, C-reactive protein, fibrinogen)
in myocardial infarction survivors. Environ Health Perspect, 115: 1072-1080. 156931
Ruckerl Ri Ibald-Mulli Ai Koenig WSchneider Ai Woelke Gi Cyrys Ji Heinrich Ji Marder Vi Frampton Mi
Wichmann HEi Peters A. (2006). Air pollution and markers of inflammation and coagulation in patients
with coronary heart disease. Environ Health Perspect, 173: 432-441. 088754
Rudell Bi Blomberg Ai Helleday Ri Ledin M"Ci Lundback Bi Stjernberg Ni Horstedt Pi Sandstrom T. (1999).
Bronchoalveolar inflammation after exposure to diesel exhaust: comparison between unfiltered and
particle trap filtered exhaust. Occup Environ Med, 56: 527-534. 001964
Rudell Bi Ledin M"Ci Hammarstrom Ui Stjernberg Ni Lundback Bi Sandstrom T. (1996). Effects on symptoms
and lung function in humans experimentally exposed to diesel exhaust. Occup Environ Med, 53: 658-662.
056577
Rundell KWi Caviston R. (2008). Ultrafine and fine particulate matter inhalation decreases exercise
performance in healthy subjects. , 22: 2-5. 191986
Rundell KWi Hoffman JR.; Caviston Ri Bulbulian Ri Hollenbach AM. (2007). Inhalation of ultrafine and fine
particulate matter disrupts systemic vascular function. Inhal Toxicol, 19: 133-140. 156060
Safar Mi Chamiot-Clerc Pi Dagher Gi Renaud JF. (2001). Pulse pressure, endothelium function, and arterial
stiffness in spontaneously hypertensive rats. Hypertension, 38: 1416-1421. 156068
Saldiva PHNi Clarke RWi Coull BAi Stearns RCi Lawrence Ji Krishna-Murthy GGi Diaz Ei Koutrakis Pi Suh Hi
Tsuda Ai Godleski JJ. (2002). Lung inflammation induced by concentrated ambient air particles is
related to particle composition. Am J Respir Crit Care Med, 165: 1610-1617. 025988
Salvi Si Blomberg Ai Rudell Bi Kelly Fi Sandstrom Ti Holgate STi Frew A. (1999). Acute inflammatory
responses in the airways and peripheral blood after short-term exposure to diesel exhaust in healthy
human volunteers. Am J Respir Crit Care Med, 159: 702-709. 058637
Samet JMi Dominici Fi Zeger SLi Schwartz Ji Dockery DW. (2000). National morbidity, mortality, and air
pollution study Part I: methods and methodologic issues. 005809
Samet JMi Rappold Ai Graff Di Cascio WEi Berntsen JHi Huang YCi Herbst Mi Bassett Mi Montilla Ti Hazucha
MJi Bromberg PAi Devlin RB. (2009). Concentrated ambient ultrafine particle exposure induces cardiac
changes in young healthy volunteers. Am J Respir Crit Care Med, 179: 1034-1042. 191913
Samet JMi Zeger SLi Dominici Fi Curriero Fi Coursac Ii Dockery DWi Schwartz Ji Zanobetti A. (2000). The
national morbidity, mortality, and air pollution study Part II: morbidity, mortality, and air pollution in
the United States. 010269
Samoli Ei Analitis Ai Touloumi Gi Schwartz Ji Anderson HRi Sunyer Ji Bisanti Li Zmirou Di Vonk JMi
Pekkanen Ji Goodman Pi Paldy Ai Schindler Ci Katsouyanni K. (2005). Estimating the exposure-
response relationships between particulate matter and mortality within the APHEA multicity project.
Environ Health Perspect, 113: 88-95. 087436
July 2009
6-358
DRAFT-DO NOT CITE OR QUOTE

-------
Samoli E> Peng R; Ramsay Ti Pipikou Mi Touloumi Gi Dominici Fi Burnett Ri Cohen Ai Krewski Di Samet J.
(2008). Acute effects of ambient particulate matter on mortality in Europe and North America- results
from the APHENA study. Environ Health Perspect, 116: 1480. 188455
Santos Ui Terra-Filho Mi Lin C> Pereira Li Vieira Ti Saldiva Pi Braga A. (2008). Cardiac arrhythmia emergency
room visits and environmental air pollution in Sao Paulo, Brazil. J Epidemiol Community Health, 62:
267. 192004
Sarin Si Undem B> Sanico Ai Togias A. (2006). The role of the nervous system in rhinitis. , 118: 999-1016.
191166
Sarnat JAi Marmur A', Klein Mi Kim Ei Russell AG; Sarnat SE; Mulholland JA; Hopke PK; Tolbert PE. (2008).
Fine particle sources and cardiorespiratory morbidity: an application of chemical mass balance and factor
analytical source-apportionment methods. Environ Health Perspect, 116: 459-66. 097972
Sarnat SE; Suh HH; Coull BA; Schwartz J; Stone PH; Gold DR. (2006). Ambient particulate air pollution and
cardiac arrhythmia in a panel of older adults in Steubenville, Ohio. Occup Environ Med, 63: 700-706.
090489
Sauerzapf V; Jones AP; Cross J. (2009). Environmental factors and hospitalisation for chronic obstructive
pulmonary disease in a rural county of England. J Epidemiol Community Health, 63: 324-328. 180082
Schaumann F; Borm PJA; Herbrich A; Knoch J; Pitz M; Schins RPF; Luettig B; Hohlfeld JM; Heinrich J; Krug
N. (2004). Metal-rich ambient particles (particulate matter 25) cause airway inflammation in healthy
subjects. Am J Respir Crit Care Med, 170: 898-903. 087966
Schildcrout JS; Sheppard L; Lumley T; Slaughter JC; Koenig JQ; Shapiro GG. (2006). Ambient air pollution and
asthma exacerbations in children: an eight-city analysis. Am J Epidemiol, 164: 505-517. 089812
Schins RPF; Lightbody JH; Borm PJA; Shi T; Donaldson K; Stone V. (2004). Inflammatory effects of coarse and
fine particulate matter in relation to chemical and biological constituents. Toxicol Appl Pharmacol, 195:
1-11. 054173
Schmitz H; Hilgers U; Weidner M. (2000). Assimilation and metabolism of formaldehyde by leaves appear
unlikely to be of value for indoor air purification. New Phytol, 147: 307-315. 015623
Schneider A; Neas L; Herbst M; Case M; Williams R; Cascio W', Hinderliter A; Holguin F; Buse J; Dungan K.
(2008). Endothelial dysfunction: associations with exposure to ambient fine particles in diabetic
individuals. Environ Health Perspect, 116: 1666. 191985
Schreuder AB; Larson TV; Sheppard L; Claiborn CS. (2006). Ambient woodsmoke and associated respiratory
emergency department visits in Spokane, Washington. Int J Occup Environ Health, 12: 147-53. 097959
Schwartz J. (2003). Airborne particles and daily deaths in 10 US cities. 042800
Schwartz J. (2004). Is the association of airborne particles with daily deaths confounded by gaseous air
pollutants? An approach to control by matching. Environ Health Perspect, 112: 557-561. 053506
Schwartz J. (2004). The effects of particulate air pollution on daily deaths: a multi-city case crossover analysis. ,
61: 956-961. 188945
Schwartz J. (2004). The effects of particulate air pollution on daily deaths: a multi-city case crossover analysis.
Occup Environ Med, 61; 956-961. 078998
Schwartz J; Litonjua A; Suh H; Verrier M; Zanobetti A; Syring M; Nearing B; Verrier R; Stone P; MacCallum G;
Speizer FE; Gold DR. (2005). Traffic related pollution and heart rate variability in a panel of elderly
subjects. Thorax, 60: 455-461. 074317
Schwartz J; Marcus A. (1990). Mortality and air pollution in London: a time series analysis. Am J Epidemiol,
131: 185-194. 073222
Schwartz J; Morris R. (1995). Air pollution and hospital admissions for cardiovascular disease in Detroit,
Michigan. Am J Epidemiol, 142: 23-35. 046186
Schwartz J; Neas LM. (2000). Fine particles are more strongly associated than coarse particles with acute
respiratory health effects in schoolchildren. , 11: 6"10. 007625
Schwartz J; Park SK; O'Neill MS; Vokonas PS; Sparrow D; Weiss S; Kelsey K. (2005). Glutathione-S-transferase
Ml, obesity, statins, and autonomic effects of particles: gene-by-drug-by-environment interaction. Am J
Respir Crit Care Med, 172: 1529-1533. 086296
Seagrave J; Campen M; McDonald J; Mauderly J; Rohr A. (2008). Oxidative stress, inflammation, and
pulmonary function assessment in rats exposed to laboratory-generated pollutant mixtures. , 71: 1352.
191990
Seagrave JC; McDonald JD; Bedrick E; Edgerton ES; Gigliotti AP; Jansen JJ; Ke L; Naeher LP; Seilkop SK;
Zheng M; Mauderley JL. (2006). Lung toxicity of ambient particulate matter from southeastern US sites
with different contributing sources: relationships between composition and effects. Environ Health
Perspect, 114: 1387-93. 091291
Segala C; Poizeau D; Neukirch F; Aubier M; Samson J; Gehanno P. (2004). Air pollution, passive smoking, and
respiratory symptoms in adults. Arch Environ Occup Health, 59: 669-676. 090449
July 2009
6-359
DRAFT-DO NOT CITE OR QUOTE

-------
Shah AP; Pietropaoli AP; Frasier LM; Speers DM; Chalupa DC; Delehanty JM; Huang LS; Utell MJ; Frampton
MW. (2008). Effect of inhaled carbon ultrafine particles on reactive hyperemia in healthy human
subjects. Environ Health Perspect, 116- 375-380. 156970
Sheppard L. (2003). Ambient air pollution and nonelderly asthma hospital admissions in Seattle, Washington,
1987-1994. 042826
Sheppard L; Levy D; Norris G; Larson TV; Koenig JQ. (1999). Effects of ambient air pollution on nonelderly
asthma hospital admissions in Seattle, Washington, 1987-1994. , 10- 23-30. 086921
Shi JP; Harrison RM; Evans D. (2001). Comparison of ambient particle surface area measurement by
epiphaniometer and SMPS/APS. Atmos Environ, 35- 6193-6200. 078292
Sillanpaa M; Frey A; Hillamo R; Pennanen AS; Salonen RO. (2005). Organic, elemental and inorganic carbon in
particulate matter of six urban environments in Europe. , 5- 2869-2879. 156980
Simpson R; Williams G; Petroeschevsky A; Best T; Morgan G; Denison L; Hinwood A; Neville G. (2005). The
short-term effects of air pollution on hospital admissions in four Australian cities. Aust N Z J Public
Health, 29: 213-221. 087438
Sinclair AH; Tolsma D. (2004). Associations and lags between air pollution and acute respiratory visits in an
ambulatory care setting: 25-month results from the aerosol research and inhalation epidemiological
study. J Air Waste Manag Assoc, 54: 1212-1218. 088696
Sirivelu MP; MohanKumar SMJ; Wagner JG; Harkema JR; MohanKumar PS. (2006). Activation of the stress
axis and neurochemical alterations in specific brain areas by concentrated ambient particle exposure
with concomitant allergic airway disease. Environ Health Perspect, 114: 870-874. 111151
Slaughter JC; Kim E; Sheppard L; Sullivan JH; Larson TV; Claiborn C. (2005). Association between particulate
matter and emergency room visits, hospital admissions and mortality in Spokane, Washington. J Expo
Sci Environ Epidemiol, 15: 153-159. 073854
Slaughter JC; Lumley T; Sheppard L; Koenig JQ; Shapiro GG. (2003). Effects of ambient air pollution on
symptom severity and medication use in children with asthma. Ann Allergy Asthma Immunol, 91: 346"
353. 086294
Smith AD; Taylor DR. (2005). Is exhaled nitric oxide measurement a useful clinical test in asthma?. Curr Opin
Allergy Clin Immunol, 5: 49-56. 192176
Smith KR; Kim S; Recendez JJ; Teague SV; Menache MG! Grubbs DE; Sioutas C; Pinkerton KE. (2003).
Airborne particles of the California Central Valley alter the lungs of healthy adult rats. Environ Health
Perspect, 111: 902-908. 042107
Smith KR; Veranth JM; Kodavanti UP; Aust AE; Pinkerton KE. (2006). Acute pulmonary and systemic effects of
inhaled coal fly ash in rats: comparison to ambient environmental particles. , 93: 390-9. 110864
Steerenberg P; Withagen C; Dalen W', Dormans J; Loveren H. (2004). Adjuvant Activity of Ambient Particulate
Matter in Macrophage Activity-Suppressed, N"Acetylcysteine-Treated, iNOS" and IL"4"Deficient Mice.
Inhal Toxicol, 16: 835-843. 087981
Steerenberg PA; Van Amelsvoort L; Lovik M; Hetland RB; Alberg T; Halatek T; Bloemen HJT; Rydzynski K;
Swaen G; Schwarze P; Dybing E; Cassee FR. (2006). Relation between sources of particulate air pollution
and biological effect parameters in samples from four European cities: an exploratory study. Inhal
Toxicol, 18: 333-346. 088249
Steerenberg PA; Withagen CE; van Dalen WJ; Dormans JA; Heisterkamp SH; van Loveren H; Cassee FR.
(2005). Dose dependency of adjuvant activity of particulate matter from five European sites in three
seasons in an ovalbumin-mouse model. Inhal Toxicol, 17: 133-145. 088649
Steerenberg PA; Withagen CET; van Dalen WJ; Dormans JAMA; Cassee FR; Heisterkamp SH; van Loveren H.
(2004). Adjuvant activity of ambient particulate matter of different sites, sizes, and seasons in a
respiratory allergy mouse model. Toxicol Appl Pharmacol, 200: 186-200. 096024
Steinvil A; Kordova-Biezuner L; Shapira I; Berliner S; Rogowski O. (2008). Short-term exposure to air pollution
and inflammation-sensitive biomarkers. , 106: 51-61. 188893
Stenfors N; Nordenhall C; Salvi SS; Mudway I; Soderberg M; Blomberg A; Helleday R; Levin JO; Holgate ST;
Kelly FJ; Frew AJ; Sandstrom T. (2004). Different airway inflammatory responses in asthmatic and
healthy humans exposed to diesel. Eur Respir J, 23: 82"86. 157009
Stevens T; Krantz QT; Linak WP; Hester S; Gilmour MI. (2008). Increased transcription of immune and
metabolic pathways in naive and allergic mice exposed to diesel exhaust. Toxicol Sci, 102: 359-370.
157010
Stolzel M; Peters A; Wichmann H"E. (2003). Daily mortality and fine and ultrafine particles in Erfurt,
Germany. 042842
Strand M; Vedal S; Rodes C', Dutton SJ; Gelfand EW! Rabinovitch N. (2006). Estimating effects of ambient
PM2.5 exposure on health using PM2.5 component measurements and regression calibration (vol 16, pg
30, 2006). J Expo Sci Environ Epidemiol, 16: 471-471. 157017
July 2009
6-360
DRAFT-DO NOT CITE OR QUOTE

-------
Stolzel Mi Breitner Si Cyrys Ji Pitz Mi Wolke Gi Kreyling Wi Heinrich Ji Wichmann H"Ei Peters A. (2007).
Daily mortality and particulate matter in different size classes in Erfurt, Germany. J Expo Sci Environ
Epidemiol, 17: 458-467. 091374
Suglia SFi Gryparis A; Wright RCK Schwartz Ji Wright RJ. (2008). Association of black carbon with cognition
among children in a prospective birth cohort study. Am J Epidemiol, 167: 280-286. 157027
Sullivan Ji Ishikawa Ni Sheppard Li Siscovick Di Checkoway Hi Kaufman J. (2003). Exposure to ambient fine
particulate matter and primary cardiac arrest among persons with and without clinically recognized
heart disease. Am J Epidemiol, 157: 501-509. 043156
Sullivan Ji Sheppard Li Schreuder Ai Ishikawa Ni Siscovick Di Kaufman J. (2005). Relation between short-term
fine-particulate matter exposure and onset of myocardial infarction. , 16: 41-48. 050854
Sullivan JHi Hubbard Ri Liu SLi Shepherd Ki Trenga CAi Koenig JQi Chandler WLi Kaufman JD. (2007). A
community study of the effect of particulate matter on blood measures of inflammation and thrombosis in
an elderly population. Environ Health Perspect, 6: 3. 100083
Sullivan JHi Schreuder ABi Trenga CAi Liu SLi Larson TV; Koenig JQi Kaufman JD. (2005). Association
between short term exposure to fine particulate matter and heart rate variability in older subjects with
and without heart disease. Thorax, 60: 462"6. 109418
Sureshkumar V; Paul Bi Uthirappan Mi Pandey R; Sahu APi Lai K> Prasad AKi Srivastava Si Saxena Ai
Mathur Ni Gupta YK. (2005). Proinflammatory and anti-inflammatory cytokine balance in gasoline
exhaust induced pulmonary injury in mice. Inhal Toxicol, 17: 161-168. 088306
Symons JMi Wang Li Guallar E> Howell E> Dominici F> Schwab Mi Ange BAi Samet Ji Ondov Ji Harrison D>
Geyh A. (2006). A case-crossover study of fine particulate matter air pollution and onset of congestive
heart failure symptom exacerbation leading to hospitalization. Am J Epidemiol, 164: 421-33. 091258
Szyszkowicz M. (2008). Ambient air pollution and daily emergency department visits for ischemic stroke in
Edmonton, Canada. Int J Occup Med Environ Health, 21: 295-300. 192128
Szyszkowicz M. (2009). Air pollution and ED visits for chest pain. Am J Emerg Med, 27: 165-168. 191996
Sorensen Mi Daneshvar Bi Hansen Mi Dragsted LOi Hertel O; Knudsen L; Loft S. (2003). Personal PM2.5
exposure and markers of oxidative stress in blood. Environ Health Perspect, 111: 161-166. 157000
Tamagawa Ei Bai N> Morimoto K> Gray Ci Mui T> Yatera K; Zhang X; Xing L> Li Y> Laher I. (2008). Particulate
matter exposure induces persistent lung inflammation and endothelial dysfunction. , 295: L79. 191988
Tan WC; Qiu D; Liam BL; Ng TP; Lee SH; Van Eeden SF; D'Yachkova Y; Hogg JC. (2000). The human bone
marrow response to acute air pollution caused by forest fires. Am J Respir Crit Care Med, 161: 1213-
1217. 002304
Tankersley CG; Bierman A; Rabold R. (2007). Variation in heart rate regulation and the effects of particle
exposure in inbred mice. , 19: 621-9. 097910
Tankersley CG; Campen M; Bierman A; Flanders SE; Broman KW; Rabold R. (2004). Particle effects on heart-
rate regulation in senescent mice. Inhal Toxicol, 16: 381-390. 094378
Tankersley CG; Champion HC; Takimoto E; Gabrielson K; Bedja D; Misra V; El-Haddad H; Rabold R; Mitzner
W. (2008). Exposure to inhaled particulate matter impairs cardiac function in senescent mice. Am J
Physiol Regul Integr Comp Physiol, 295: R252-263. 157043
Tankersley CG; Irizarry R; Flanders SE; Rabold R; Frank R. (2003). Unstable heart rate and temperature
regulation predict mortality in AKR/J mice. Am J Physiol, 284: R742-R750. 053919
Tecer LH; Alagha O; Karaca F; Tuncel G; Eldes N. (2008). Particulate Matter (PM2.5, PM10-2.5, and PM10) and
Children's Hospital Admissions for Asthma and Respiratory Diseases: A Bidirectional Case-Crossover
Study. J Toxicol Environ Health A, 71: 512-520. 180030
TFESC. (1996). Heart rate variability: standards of measurement, physiological interpretation, and clinical use.
Task Force of the European Society of Cardiology. North American Society of Pacing and
Electrophysiology. Eur Heart J, 17: 354-381. 003061
Thurston G; Ito K; Mar T; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Liu H;
Neas L; Pinto J; Stolzel M; Suh H; Hopke PK. (2005). Results and implications of the workshop on the
source apportionment of PM health effects. , 16: S134-S135. 097949
Thurston GD; Ito K; Hayes CG; Bates DV; Lippmann M. (1994). Respiratory hospital admissions and
summertime haze air pollution in Toronto, Ontario: consideration of the role of acid aerosols. Environ
Res, 65: 271-290. 043921
Timonen KL; Hoek G; Heinrich J; Bernard A; Brunekreef B; De Hartog J; Hameri K; Ibald-Mulli A; Mirme A;
Peters A; Tiittanen P; Kreyling WG; Pekkanen J. (2004). Daily variation in fine and ultrafine particulate
air pollution and urinary concentrations of lung Clara cell protein CC16. Occup Environ Med, 61: 908"
914. 087915
July 2009
6-361
DRAFT-DO NOT CITE OR QUOTE

-------
Timonen KL; Vanninen Ei De Hartog J; Ibald-Mulli A; Brunekreef B; Gold DR; Henrich J; Hoek G> Lanki T>
Peters A; Tarkiainen Ti Tiittanen P; Kreyling Wi Pekkanen J. (2006). Effects of ultrafine and fine
particulate and gaseous air pollution on cardiac autonomic control in subjects with coronary artery
disease- the ULTRA study. J Expo Sci Environ Epidemiol, 16- 332-341. 088747
Tolbert PE; Klein M; Metzger KB; Peel J; Flanders WD; Todd K; Mulholland JA; Ryan PB; Frumkin H. (2000).
Interim results of the study of particulates and health in Atlanta (SOPHIA). J Expo Sci Environ
Epidemiol, 10: 446-460. 010320
Tolbert PE; KLeina MJ Peelb JL; Sarnata SE; Sarnata JA. (2007). Multipollutant modeling issues in a study of
ambient air quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 24: 938"
945. 090316
Tornqvist H; Mills NL; Gonzalez M; Miller MR; Robinson SD; Megson IL; MacNee W; Donaldson K; Soderberg
S; Newby DE; Sandstrom T; Blomberg A. (2007). Persistent endothelial dysfunction in humans after
diesel exhaust inhalation. Am J Respir Crit Care Med, 176: 395-400. 091279
Trenga CA; Sullivan JH; Schildcrout JS; Shepherd KP; Shapiro GG; Liu LJ; Kaufman JD; Koenig JQ. (2006).
Effect of particulate air pollution on lung function in adult and pediatric subjects in a Seattle panel
study. , 129: 1614-22. 155209
Tsai FC; Apte MG; Daisey JM. (2000). An exploratory analysis of the relationship between mortality and the
chemical composition of airborne particulate matter. Inhal Toxicol, 12: 121-135. 006251
Tsai S-S; Cheng M"H; Chiu H-F; Wu T-N; Yang C-Y. (2006). Air pollution and hospital admissions for asthma in
a tropical city: Kaohsiung, Taiwan. Inhal Toxicol, 18: 549-554. 089768
Tsai S-S; Goggins WB; Chiu H"F; Yang C"Y. (2003). Evidence for an association between air pollution and daily
stroke admissions in Kaohsiung, Taiwan. , 34: 2612-2616. 080133
Tunnicliffe WS; Harrison RM; Kelly FJ; Dunster C; Ayres JG. (2003). The effect of sulphurous air pollutant
exposures on symptoms, lung function, exhaled nitric oxide, and nasal epithelial lining fluid antioxidant
concentrations in normal and asthmatic adults. Occup Environ Med, 60. 088744
U.S. EPA. (1982). Air quality criteria for particulate matter and sulfur oxides. U.S. Environmental Protection
Agency. Washington, D.C.. 017610
U.S. EPA. (1996). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA/600/P-95/001aF-cF. 079380
U.S. EPA. (2002). Health assessment document for diesel engine exhaust. 042866
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA/600/P"99/002aF"bF. 056905
Ulirsch GV; Ball LM; Kaye W; Shy CM; Lee CV; Crawford-Brown D; Symons M; Holloway T. (2007). Effect of
particulate matter air pollution on hospital admissions and medical visits for lung and heart disease in
two southeast Idaho cities. J Expo Sci Environ Epidemiol, 17: 478-487. 091332
Upadhyay D; Panduri V; Ghio A; Kamp DW. (2003). Particulate matter induces alveolar epithelial cell DNA
damage and apoptosis: role of free radicals and the mitochondria. , 29: 180-187. 097370
Upadhyay S; Stoeger T; Harder V; Thomas RF; Schladweiler MC; Semmler-Behnke M; Takenaka S; Karg E;
Reitmeir P; Bader M; Stampfl A; Kodovanti U; Schulz H. (2008). Exposure to ultrafine carbon particles at
levels below detectablepulmonary inflammation affects cardiovascular performance inspontaneously
hypertensive rats. Part Fibre Toxicol, 5- 19. 159345
Urch B; Brook JR.; Wasserstein D; Brook RD; Rajagopalan S; Corey P; Silverman F. (2004). Relative
contributions of PM2.5 chemical constituents to acute arterial vasoconstriction in humans. Inhal Toxicol,
16: 345-352. 055629
Urch B; Silverman F; Corey P; Brook JR; Lukic KZ; Rajagopalan S; Brook RD. (2005). Acute blood pressure
responses in healthy adults during controlled air pollution exposures. Environ Health Perspect, 113:
1052-1055. 081080
Van der Werf GR; Randerson JT; Collatz J; Giglio L. (2003). Carbon emissions from fires in tropical and
subtropical ecosystems. , 9: 547-562. 089398
Vedal S; Rich K; Brauer M; White R; Petkau J. (2004). Air pollution and cardiac arrhythmias in patients with
implantable cardiovascular defibrillators. Inhal Toxicol, 16: 353-362. 055630
Vegni FE; Ros O. (2004). Hospital accident and emergency burden is unaffected by today's air pollution levels.
Eur J Emerg Med, 11: 86-88. 087448
Veranth JM; Moss TA; Chow JC; Labban R; Nichols WK Walton JC; Walton JG; Yost GS. (2006). Correlation of
in vitro cytokine responses with the chemical composition of soil-derived particulate matter. Environ
Health Perspect, 114: 341-349. 087479
Veronesi B; Makwana O; Pooler M; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient
particles VII Degeneration of dopaminergic neurons in Apo E7- mice. Inhal Toxicol, 17: 235-241. 087481
July 2009
6-362
DRAFT-DO NOT CITE OR QUOTE

-------
Vigotti MA; Chiaverini F> Biagiola P; Rossi G. (2007). Urban air pollution and emergency visits for respiratory
complaints in Pisa, Italy. J Toxicol Environ Health A, 70- 266-269. 090711
Villeneuve PJ; Burnett RT; Shi Y> Krewski D; Goldberg MS; Hertzman C; Chen Y; Brook J. (2003). A time-
series study of air pollution, socioeconomic status, and mortality in Vancouver, Canada. J Expo Sci
Environ Epidemiol, 13: 427-435. 055051
Villeneuve PJ; Chen L; Stieb D; Rowe BH. (2006). Associations between outdoor air pollution and emergency
department visits for stroke in Edmonton, Canada. Eur J Epidemiol, 21; 689-700. 090191
Vincent R; Kumarathasan P; Goegan P; Bjarnason SG; Guenette J; Berube D; Adamson IY; Desjardins S;
Burnett RT; Miller FJ; Battistini B. (2001). Inhalation toxicology of urban ambient particulate matter:
acute cardiovascular effect in rats. 021184
Von KLot S; Peters A; Aalto P; Bellander T; Berglind N; D'Ippoliti D; Elosua R; Hormann A; Kulmala M; Lanki
T; Lowel H; Pekkanen J; Picciotto S; Sunyer J; Forastiere F; Health Effects of Particles on Susceptible
Subpopulations (HEAPSS) Study Group. (2005). Ambient air pollution is associated with increased risk
of hospital cardiac readmissions of myocardial infarction survivors in five European cities. Circulation,
112: 3073-3079. 088070
Von Klot S; Wolke G; Tuch T; Heinrich J; Dockery DW; Schwartz J; Kreyling WG; Wichmann HE; Peters A.
(2002). Increased asthma medication use in association with ambient fine and ultrafine particles. Eur
Respir J, 20: 691-702. 034706
Ward DJ; Roberts KT; Jones N; Harrison RM; Ayres JG; Hussain S; Walters S. (2002). Effects of daily variation
in outdoor particulates and ambient acid species in normal and asthmatic children. Thorax, 57: 489-502.
025839
Ward PA. (2003). Acute lung injury: how the lung inflammatory response works. Eur Respir J, 44: 22s-23s.
157111
Wegesser TC; Last JA. (2008). Lung response to coarse PM: bioassay in mice. Toxicol Appl Pharmacol, 230: 159-
166. 190506
Welin L; Eriksson H; Larsson B; Svardsudd K; Wilhelmsen L; Tibblin G. (1993). Risk Factors for Coronary
Heart Disease during 25 Years of Follow-Up. Cardiology, 82: 223-228. 156151
Wellenius GA; Batalha JRF; Diaz EA; Lawrence J; Coull BA; Katz T; Verrier RL; Godleski JJ. (2004). Cardiac
effects of carbon monoxide and ambient particles in a rat model of myocardial infarction. Toxicol Sci, 80:
367-376. 087874
Wellenius GA; Bateson TF; Mittleman MA; Schwartz J. (2005). Particulate air pollution and the rate of
hospitalization for congestive heart failure among medicare beneficiaries in Pittsburgh, Pennsylvania.
Am J Epidemiol, 161: 1030-1036. 087483
Wellenius GA; Coull BA; Batalha JRF; Diaz EA; Lawrence J; Godleski JJ. (2006). Effects of ambient particles
and carbon monoxide on supraventricular arrhythmias in a rat model of myocardial infarction. Inhal
Toxicol, 18: 1077-1082. 156152
Wellenius GA; Coull BA; Godleski JJ; Koutrakis P; Okabe K; Savage ST. (2003). Inhalation of concentrated
ambient air particles exacerbates myocardial ischemia in conscious dogs. Environ Health Perspect, 111:
402-408. 055691
Wellenius GA; Saldiva PHN; Batalha JRF; Murthy GGK; Coull BA; Verrier RL; Godleski JJ. (2002).
Electrocardiographic changes during exposure to residual oil fly ash (ROFA) particles in a rat model of
myocardial infarction. Toxicol Sci, 66: 327-335. 025405
Wellenius GA; Schwartz J; Mittleman MA. (2005). Air pollution and hospital admissions for ischemic and
hemorrhagic stroke among medicare beneficiaries. , 36: 2549-2553. 088685
Wellenius GA; Schwartz J; Mittleman MA. (2006). Particulate air pollution and hospital admissions for
congestive heart failure in seven United States cities. Am J Cardiol, 97: 404-408. 088748
Wellenius GA; Yeh GY; Coull BA; Suh HH; Phillips RS; Mittleman MA. (2007). Effects of ambient air pollution
on functional status in patients with chronic congestive heart failure: a repeated-measures study. , 6: 26.
092830
Welty LJ; Zeger SL. (2005). Are the acute effects of particulate matter on mortality in the National Morbidity,
Mortality, and Air Pollution study the result of inadequate control for weather and season? A sensitivity
analysis using flexible distributed lag models. Am J Epidemiol, 162: 80-88. 087484
Wheeler A; Zanobetti A; Gold DR; Schwartz J; Stone P; Suh HH. (2006). The relationship between ambient air
pollution and heart rate variability differs for individuals with heart and pulmonary disease. Environ
Health Perspect, 114: 560-566. 088453
Whitekus MJ; Li N> Zhang M; Wang M; Horwitz MA; Nelson SK; Horwitz LD; Brechun N; Diaz-Sanchez D; Nel
AE. (2002). Thiol antioxidants inhibit the adjuvant effects of aerosolized diesel exhaust particles in a
murine model for ovalbumin sensitization. , 168: 2560-2567. 157142
July 2009
6-363
DRAFT-DO NOT CITE OR QUOTE

-------
Whitsel E; Quibrera P; Christ S; Liao D; Prineas R; Anderson G; Heiss G. (2009). Heart rate variability,
ambient particulate matter air pollution, and glucose homeostasis' the environmental epidemiology of
arrhythmogenesis in the Women's Health Initiative. Am J Epidemiol, x- x. 191980
Wichmann H"E; Spix C; Tuch T; Wolke G; Peters A; Heinrich J; Kreyling WG; Heyder J. (2000). Daily mortality
and fine and ultrafine particles in Erfurt, Germany Part I- role of particle number and particle mass.
013912
Widdicombe J. (2006). Reflexes from the lungs and airways' historical perspective. J Appl Physiol, 101' 628.
155519
Widdicombe J; Lee L"Y. (2001). Airway reflexes, autonomic function, and cardiovascular responses. Environ
Health Perspect, 4- 579-584. 019049
Widdicombe JG. (2003). Overview of neural pathways in allergy and asthma. Pulm Pharmacol Ther, 16' 23-30.
157145
Wilson WE; Mar TF; Koenig JQ. (2007). Influence of exposure error and effect modification by socioeconomic
status on the association of acute cardiovascular mortality with particulate matter in Phoenix. J Expo Sci
Environ Epidemiol, 17: Sll. 157149
Win-Shwe TT; Yamamoto S; Fujitani Y> Hirano S; Fujimaki H. (2008). Spatial learning and memory function-
related gene expression in thehippocampus of mouse exposed to nanoparticle-rich diesel exhaust.
Neurotoxicology, 29: 940-947. 190516
Witten ML; Wong SS ; Sun NN; Keith I; Kweon C; Foster DE; Schauer JJ; Sherrill DL. (2005). Neurogenic
responses in rat lungs after nose-only exposure to diesel exhaust. 087485
Wong OM; Atkinson RW! Anderson HR; Hedley AJ; Ma S; Chau PY-K; Lam T"H. (2002). A tale of two cities:
effects of air pollution on hospital admissions in Hong Kong and London compared. Environ Health
Perspect, 110: 67-77. 023232
Wong SS; Sun NN; Keith I; Kweon C"B; Foster DE; Schauer James J; Witten ML. (2003). Tachykinin substance
P signaling involved in diesel exhaust-induced bronchopulmonary neurogenic inflammation in rats. , 77:
638-650. 097707
Wong TW; Lau TS; Yu TS; Neller A; Wong SL; Tam W', Pang SW. (1999). Air pollution and hospital admissions
for respiratory and cardiovascular diseases in Hong Kong. Occup Environ Med, 56: 679-683. 009172
Wong TW; Tam WS; Yu TS; Wong AHS. (2002). Associations between daily mortalities from respiratory and
cardiovascular diseases and air pollution in Hong Kong, China. Occup Environ Med, 59: 30-35. 025436
Wordley J; Walters S; Ayres JG. (1997). Short term variations in hospital admissions and mortality and
particulate air pollution. Occup Environ Med, 54: 108-116. 082745
Wu J; M Winer A; J Delfino R. (2006). Exposure assessment of particulate matter air pollution before, during,
and after the 2003 Southern California wildfires. Atmos Environ, 40: 3333-3348. 157156
Yan YH; Huang CH; Chen WJ; Wu MF; Cheng TJ. (2008). Effects of diesel exhaust particles on left ventricular
function in isoproterenol-induced myocardial injury and healthy rats. Inhal Toxicol, 20: 199-203. 098625
Yang C-Y; Chang C"C; Chuang H"Y; Tsai S"S; Wu T"N; Ho C"K. (2004). Relationship between air pollution and
daily mortality in a subtropical city: Taipei, Taiwan. Environ Int, 30: 519-523. 055603
Yang CY. (2008). Air pollution and hospital admissions for congestive heart failure in a subtropical city: Taipei,
Taiwan. J Toxicol Environ Health A, 71: 1085-1090. 157160
Yang CY; Chen CJ. (2007). Air pollution and hospital admissions for chronic obstructive pulmonary disease in a
subtropical city: Taipei, Taiwan. J Toxicol Environ Health A, 70: 1214-9. 092847
Yang CY; Cheng MH; Chen CC. (2009). Effects of Asian Dust Storm Events on Hospital Admissions for
Congestive Heart Failure in Taipei, Taiwan. , 72: 324-328. 190341
Yang Q; Chen Y; Krewski D; Shi Y; Burnett RT; McGrail KM. (2004). Association between particulate air
pollution and first hospital admission for childhood respiratory illness in Vancouver, Canada. Arch
Environ Occup Health, 59: 14-21. 087488
Yeatts K; Svendsen E; Creason J; Alexis N; Herbst M; Scott J; Kupper L; Williams R; Neas L; Cascio W', Devlin
RB; Peden DB. (2007). Coarse particulate matter (PM25-10) affects heart rate variability, blood lipids,
and circulating eosinophils in adults with asthma. Environ Health Perspect, 115: 709-714. 091266
Yokota S; Seki T; Naito Y; Tachibana S; Hirabayashi N; Nakasaka T; Ohara N; Kobayashi H. (2008). Tracheal
instillation of diesel exhaust particles component causes blood and pulmonary neutrophilia and enhances
myocardial oxidative stress in mice. J Toxicol Sci, 33: 609-620. 190109
Yu O; Sheppard L; Lumley T; Koenig JQ; S. (2000). Effects of ambient air pollution on symptoms of asthma in
Seattle-area children enrolled in the CAMP study. Environ Health Perspect, 108: 1209-1214. 013254
Yue W; Schneider A; Stolzel M; Ruckerl R; Cyrys J; Pan X; Zareba W; Koenig W; Wichmann HE; Peters A.
(2007). Ambient source-specific particles are associated with prolonged repolarization and increased
levels of inflammation in male coronary artery disease patients. , 621: 50-60. 097968
July 2009
6-364
DRAFT-DO NOT CITE OR QUOTE

-------
Zabel Mi KLingenheben Ti Franz MRi Hohnloser SH. (1998). Assessment of QT Dispersion for Prediction of
Mortality or Arrhythmic Events After Myocardial Infarction • Results of a Prospective, Long-term
Follow-up Study. , 97: 2543-2550. 156176
Zanchi ACi Venturini CDi Saiki Mi Nascimento Saldiva PHi Tannhauser Barros HMi Rhoden CR. (2008).
Chronic nasal instillation of residual-oil fly ash (ROFA) induces brain lipid peroxidation and behavioral
changes in rats. Inhal Toxicol, 20: 795-800. 157173
Zanobetti Ai Canner MJi Stone PHi Schwartz Ji Sher Di Eagan-Bengston Ei Gates KAi Hartley LHi Suh Hi
Gold DR. (2004). Ambient pollution and blood pressure in cardiac rehabilitation patients. , 110: 2184"
2189. 087489
Zanobetti Ai Schwartz J. (2003). Airborne particles and hospital admissions for heart and lung disease. 043119
Zanobetti Ai Schwartz J. (2003). Airborne particles and hospital admissions for heart and lung disease. In:
Revised analyses of time-series studies of air pollution and health. Special Report. 157174
Zanobetti Ai Schwartz J. (2003). Multicity assessment of mortality displacement within the APHEA2 project.
042812
Zanobetti Ai Schwartz J. (2005). The effect of particulate air pollution on emergency admissions for myocardial
infarction: a multicity case-crossover analysis. Environ Health Perspect, 113: 978-982. 088069
Zanobetti Ai Schwartz J. (2006). Air pollution and emergency admissions in Boston, MA. J Epidemiol
Community Health, 60: 890-895. 090195
Zanobetti Ai Schwartz J. (2009). The effect of fine and coarse particulate air pollution on mortality: A national
analysis. Environ Health Perspect, 117: 1-40. 188462
Zareba Wi Couderc JPi Oberdorster Gi Chalupa Di Cox Ci Huang LSi Peters Ai Utell MJi Frampton MW.
(2009). ECG parameters and exposure to carbon ultrafine particles in young healthy subjects. Inhal
Toxicol, 21: 223-233. 190101
Zeka Ai Sullivan JRi Vokonas PS; Sparrow Di Schwartz J. (2006). Inflammatory markers and particulate air
pollution: characterizing the pathway to disease. Int J Epidemiol, 35: 1347-1354. 157177
Zeka Ai Zanobetti A; Schwartz J. (2005). Short term effects of particulate matter on cause specific mortality:
effects of lags and modification by city characteristics. Occup Environ Med, 62: 718-725. 088068
Zeka A; Zanobetti A; Schwartz J. (2006). Individual-level modifiers of the effects of particulate matter on daily
mortality. Am J Epidemiol, 163: 849-859. 088749
Zelikoff JT; Chen LC; Cohen MD! Fang K> Gordon T; Li Y; Nadziejko C; Schlesinger RB. (2003). Effects of
inhaled ambient particulate matter on pulmonary antimicrobial immune defense. Inhal Toxicol, 15: 131-
150. 039009
Zhang SH; Reddick RL; Piedrahita JA; Maeda N. (1992). Spontaneous hypercholesterolemia and arterial lesions
in mice lacking apolipoprotein E. , 258: 468-471. 157180
Zhang Z; Whitsel E; Quibrera P; Smith R; Liao D; Anderson G; Prineas R. (2009). Ambient Fine Particulate
Matter Exposure and Myocardial Ischemia in the Environmental Epidemiology of Arrhythmogenesis in
the Women?s Health Initiative (EEAWHI) Study. Environ Health Perspect, 117: 751. 191970
July 2009
6-365
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Chapter 7. Integrated Health Effects of
Long-Term PM Exposure
7.1.	Introduction
This chapter summarizes reviews and integrates the evidence on relationships
between health effects and long-term exposures to various size fractions and sources of PM.
Cardiopulmonary health effects of long-term exposure to PM have been examined in an
extensive body of epidemiologic, human clinical, and toxicological studies. Both
epidemiologic and toxicological studies provide a basis for examining reproductive and
developmental and cancer health outcomes with regard to long-term exposure to PM. In
addition, there is a large body of epidemiologic literature evaluating the relationship
between mortality and long-term exposure to PM.
Conclusions from the 2004 PM AQCD are summarized briefly at the beginning of each
section, and the evaluation of evidence from recent studies builds upon what was available
during the previous review. For each health outcome (e.g., respiratory infections, lung
function), results are summarized for studies from the specific scientific discipline,
i.e., epidemiologic and toxicological studies. The sections conclude with summaries of the
evidence on the various health outcomes and integration of the findings that leads to
conclusions regarding causality based upon the framework described in Chapter 1.
Determination of causality is made for the overall health effect category, such as
cardiovascular effects, with coherence and plausibility being based upon the evidence from
across disciplines and also across the suite of related health outcomes. In these summary
sections, the evidence is summarized and independent conclusions drawn for relationships
with PM2.5, PM10-2.5, and ultrafine particles.
7.2.	Cardiovascular and Systemic Effects
Studies examining associations between long-term exposure to ambient PM (over
months to years) and CVD morbidity had not been conducted and thus were not included in
the 1996 or 2004 PM Air Quality Criteria Documents (U.S. EPA, 1996, 079380; U.S. EPA,
2004, 056905). A number of studies were included in the 2004 PM AQCD that evaluated the
effect of long-term PM2.5 exposure on cardiovascular mortality and found consistent
associations. No toxicological studies examined chronic atherosclerotic effects of PM
Note- Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health
and Environmental Research Online) at http• //epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in
the process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009
7-1
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
exposure in animal models. However, a subchronic study that evaluated atherosclerosis
progression in hyperlipidemic rabbits was discussed and this study provided the foundation
for the subsequent work that has been conducted in this area (Suwa et al., 2002, 028588).
No previous toxicological studies evaluated effects of subchronic or chronic PM exposure on
diabetes measures, or HR or HRV changes, nor were there animal toxicological studies
included in the 2004 PM AQCD that evaluated systemic inflammatory or blood coagulation
markers following subchronic or chronic PM exposure.
Several new epidemiologic studies have examined the long-term PM-CVD association
among U.S. and European populations. The studies investigate the association of both
PM2.5 and PM10 exposures with a variety of clinical and subclinical CVD outcomes.
Epidemiologic and toxicological studies have provided evidence of the adverse effects of
long-term exposure to PM2.5 on cardiovascular outcomes, including atherosclerosis, clinical
and subclinical markers of cardiovascular morbidity, and cardiovascular mortality. The
evidence of these effects from long-term exposure to PM10-2.5 is weaker.
7.2.1. Atherosclerosis
Atherosclerosis is a progressive disease that contributes to several adverse outcomes,
including myocardial infarction and aortic aneurysm. It is multifaceted, beginning with an
early injury or inflammation that promotes the extravasation of inflammatory cells. Under
conditions of oxidative or nitrosative stress and high lipid or cholesterol concentrations, the
vessel wall undergoes a chronic remodeling that is characterized by the presence of foam
cells, migrated and differentiated smooth muscle cells, and ultimately a fibrous cap. The
advanced lesion that develops from this process can occlude perfusion to distal tissue,
causing ischemia, and erode, degrade, or even rupture, revealing coagulant initiators
(tissue factor) that promote thrombosis, stenosis, and infarction or stroke. Several detailed
reviews of atherosclerosis pathology have been published elsewhere (Ross, 1999, 156926;
Stocker and Keaney JF, 2004, 157013).
7.2.1.1. Epidemiologic Studies
Measures of Atherosclerosis
Although no study has examined the association between long-term PM exposure and
longitudinal change in subclinical markers of atherosclerosis, several cross sectional studies
have been conducted. Markers of atherosclerosis used in these studies include coronary
artery calcium (CAC), carotid intima-media thickness (CIMT), ankle-brachial index (ABI),
and abdominal aortic calcium (AAC). These measures are descried briefly below.
CAC is a measure of atherosclerosis assessed by non-contrast, cardiac-gated electron
beam computed tomography (EBCT) or multidetector computed tomography (MDCT) of the
July 2009
7-2
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
coronary arteries in the heart (Greenland and Kizilbash, 2005, 156496; Hoffmann et al.,
2005, 156556; Mollet et al., 2005, 155988). The prevalence of CAC is strongly related to age.
Few people have detectable CAC in their second decade of life but the prevalence of CAC
rises to approximately 100% by age 80 (Ardehali et al., 2007, 155662). Previous studies
suggest that while the absence of CAC does not rule out atherosclerosis, it does imply a
very low likelihood of significant arterial obstruction (Achenbach and Daniel, 2001, 156189;
A rail et al., 1996, 155661; Shaw et al., 2003, 156083; Shemesh et al., 1996, 156085).
Conversely, the presence of CAC confirms the existence of atherosclerotic plaque and the
amount of calcification varies directly with the likelihood of obstructive disease (Ardehali et
al., 2007, 155662). CAC is a quantified using the Agatston method (Agatston et al., 1990,
156197). Its repeatability depends on the laboratory and the method of calculation
(O'Rourke et al. 2000). Agatston scores are frequently used to classify individuals into one
of five groups (zero! mild; moderate! severe! extensive) or according to age- and sex-specific
percentiles of the CAC distribution (Erbel et al., 2007, 155768).
CIMT is a measure of atherosclerosis assessed by high-resolution, B-mode
ultrasonography of the carotid arteries in the neck, the walls of which have inner (intimal),
middle (medial) and outer (adventitial) layers (Craven et al., 1990, 155740! O'Leary et al.,
1999, 156826! Wendelhag et al., 1993, 157136). CIMT estimates the distance in mm or jum
between the innermost (blood-intima) and outermost (media-adventitia) interfaces, often by
averaging over three arterial segments in the common carotid, carotid bulb, and internal
carotid artery (Amato et al., 2007, 155656). CIMT has been associated with atherosclerosis
risk factors (Heiss et al., 1991, 156535; O'Leary et al., 1992, 156825; Salonen and Salonen,
1991, 156938), prevalent coronary heart disease (Chambless et al., 1997, 156329;
Geroulakos et al., 1994, 155788), and incident coronary and cerebral events (O'Leary et al.,
1999, 156826; van der Meer et al., 2004, 156129). Several studies have indicated that CIMT
measurements are accurate (Girerd et al., 1994, 156474; Pignoli et al., 1986, 156026;
Wendelhag et al., 1991, 157135) and reproducible (Montauban van Swijndregt et al., 1999,
156777; Smilde et al., 1997, 156988; Willekes et al., 1999, 157147), especially for the
common carotid artery (Montauban van Swijndregt et al., 1999, 156777).
ABI, which is also known as the ankle-arm or resting (blood) pressure index, is a
measure of lower extremity arterial occlusive disease commonly caused by advanced
atherosclerosis (Weitz et al., 1996, 156150). It is assessed by continuous wave Doppler and
manual or automated oscillometric sphygmomanometry, the latter having been shown to
have higher repeatability and validity (Weitz et al., 1996, 156150). ABI is defined as the
unitless ratio of ankle to brachial systolic blood pressures measured in mmHg. As ankle
pressure is normally equal to or slightly higher than arm pressure (resulting in an ABI >
1.0), epidemiologic studies typically define the normal ABI range as 0.90 to 1.50 (Resnick et
al., 2004, 156048). Low ABI has been associated with all-cause and CVD mortality
July 2009
7-3
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
(Newman et al., 1993, 156805; Vogt et al., 1993, 157100), as well as myocardial infarction
and stroke (Karthikeyan and Lip, 2007, 156626).
AAC is a measure of atherosclerosis assessed by non-contrast, EBCT or MDCT of the
abdominal aorta. It is scored much like CAC (Agatston et al., 1990, 156197), but the age-
specific prevalence and extent of AAC is greater, particularly among women and at ages
>50 years. Although AAC has not been studied as extensively as CAC, it is associated with
carotid and coronary atherosclerosis as well as cardiovascular morbidity and mortality
(Allison et al., 2004, 156210; Allison et al., 2006, 155653; Hollander et al., 2003, 156562;
Khoury et al., 1997, 156636; Oei et al., 2002, 156820; Walsh et al., 2002, 157103; Wilson et
al., 2001, 156159; Witteman et al., 1986, 156161) and measurements are sufficiently
reproducible to allow serial investigations over time (Budoff et al., 2005, 192105).
Study Findings
Diez et al. (2008, 156401) conducted cross-sectional analyses of the association of
three of these subclinical markers of atherosclerosis (CAC, CIMT and ABI), collected from
2000 to 2003 during baseline examinations of participants enrolled in the Multi-Ethnic
Study of Atherosclerosis (MESA), with long-term exposure to PM2.5 and PM10. The study
population included 5,172 ethnically diverse people (53% female) residing in Baltimore,
MD; Chicago, IL; Forsyth County, NC; Los Angeles, CA; New York, NY; and St. Paul, MN
ranging in age from 45 to 84 years old. Authors used spatio-temporal modeling of pollutant
concentrations, weather and demographic data to impute 20-yr avg exposures to PM2.5 and
PM10. They reported small increases in CIMT of 1% (95% CP 0%-1.4%) and 0.5% (95%
CP 0%-l%), which correspond to absolute changes of 8 (95% CP 0-12) and 7 (95%
CP 0-14) jum, per 10 pg/m3 increase in 20-yr avg PM10 and PM2.5 concentration, respectively.
Evidence of age-, gender-, lipid- and smoking-related susceptibility was lacking. They also
reported weak, non-significant increases in the relative prevalence of CAC of 1% (95%
CP -2% to 4%) and 0.5% (95% CP -2% to 3%) per 10 pg/m3 increase in PM10 and PM2.5,
respectively. Among the subset of 2,586 participants with EBCT-identified calcification,
similarly weak associations were observed. There was little evidence of modification of the
CAC associations by demographic, socioeconomic or clinical characteristics. Finally, the
authors report no differences in mean ABI with PM10 or PM2.5 concentrations. The null
findings for ABI exhibited little heterogeneity among participant subgroups and were
similarly null when ABI was modeled as a dichotomous outcome using a cutpoint of 0.9
units.
MESA investigators also examined the chronic PM2.5-AAC association in a
residentially stable subset of 1,147 participants (mean age = 66 yr; 50% female) randomly
selected from all MESA centers, except Baltimore, MD for enrollment in its Aortic Calcium
Ancillary Study (Allen et al., 2009, 189644). The authors used kriging and inverse
residence-to-monitor distance-weighted averaging of EPAAQS data to estimate 2-yr mean
exposures to PM2.5. In cross-sectional analyses, the authors found a 6% (95% CP -4% to
July 2009
7-4
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
16%) excess risk of a non-zero Agatston score and an 8% (95% CP -30% to 46%) increase in
AAC, i.e., approximately 50 (95% CI: -251 to 385) Agatston units, per 10 pg/m3 increase in
PM2.5 concentration. These associations were stronger among users than non-users of anti-
hyperlipidemics.
Kunzli et al. (2005, 087387) used baseline data collected between 1998-2003 from two
randomized placebo-controlled clinical trials, the Vitamin E Atherosclerosis Progression
Study (VEAPS) and the B-Vitamin Atherosclerosis Intervention Trial (BVAIT), for their
ancillary cross-sectional analyses of the effect of long-term PM2.5 exposure on CIMT. The
study population included 798 residents of the greater Los Angeles, CA area who were more
than 40 years old at baseline and 44% were female. The authors used universal kriging of
PM2.5 data from 23 state and local monitors operating in 2000 to estimate 1-yr avg exposure
to PM2.5 at each participant's geocoded U.S. Postal Service ZIP code. They found a 4.2%
(95% CP -0.2% to 8.9%) or approximately 32 (95% CP -2 to 68) pm increase in CIMT per
10 pg/m3 increase in PM2.5 concentration. In contrast to findings from the relatively large,
ethnically diverse, yet geographically overlapping MESA ancillary study described above,
PM-related increases in CIMT were two- to three-fold larger among older and female
participants taking anti-hyperlipidemics in this study. PM-related increases in CIMT were
also higher in never smokers when compared with current or former smokers.
Hoffmann et al. (2007, 091163) conducted a cross-sectional analysis of data collected
at baseline (2000 to 2003) for 4,494 residents of Essen, Mulheim and Bochum, Germany
enrolled in the Heinz Nixdorf Recall Study from 2000 to 2003. The age of particpants
ranged from 45-74 years and 51% were female. In this cross-sectional study the authors
used dispersion and chemistry transport modeling of emissions, climate and topography
data to estimate one-yr avg exposure to PM2 .5 in 2002 (the midpoint of the baseline exam. ) They reported
an imprecise 43% (95% CP -15% to 115%) or 102 (95% CP -77 to 273) Agatston unit increase
in CAC per 10 pg/m3 increase in PM2.5. Differences in strength of association between
subgroups defined by demographic and clinical characteristics were small. The authors
reported a more consistent association of CAC with traffic exposure (distance from a major
roadway) than with PM2.5 in this study.
In a subsequent analysis of these data, Hoffmann et al. (2009, 190376) examined the
PM-ABI association in this population. In this cross-sectional study, no changes in ABI were
observed in association with PM2.5 concentration nor was evidence of effect modification by
demographic and clinical characteristics apparent. As in the previous study, residing near a
major roadway was a stronger predictor of atherosclerotic changes. Absolute changes in ABI
of -0.024 (95% CP -0.047 to -0.001) were associated with living within 50 m of a major
roadway compared to living more than 200 m away.
Each of the studies described above relied on cross-sectional analyses examining
differences in long-term average PM2.5 concentrations across space (as well as time to the
extent baseline examinations were conducted over time). Such associations may reflect the
July 2009
7-5
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
effect of compositional differences in PM2.5 as well as the effect of higher PM2.5
concentrations. Most associations of PM2.5 with CAC (Diez Roux et al., 2008, 156401;
Hoffmann et al., 2007, 091163), CIMT (Diez Roux et al., 2008, 156401; Kunzli et al., 2005,
087387), ABI (Diez Roux et al., 2008, 156401; Hoffmann et al., 2009, 190376) and AAC
(Allen et al., 2009, 189644) reviewed in this section were weak and/or imprecise. However,
several factors including exposure measurement error, variation in baseline measures
atherosclerosis, as well as limited power may contribute to the insensitivity of these cross-
sectional studies to detect small differences in CAC, CIMT, ABI and AAC. The study by
Hoffmann et al. (2007, 091163), which reported large, imprecise and non-significant
increases in CAC in association with PM2.5, is not distinguished from the other studies
reviewed by a superior study design or larger sample size. The several fold difference in the
magnitude of CIMT associations reported by Kunzli et al. (2005, 087387) and Diez Roux et
al. (2008, 156401) may be related to differences between the study populations. The
ambient PM concentrations from these studies are characterized in Table 7-1.
7.2.1.2. Toxicological Studies
In the only study of this kind described in the 2004 PM AQCD, Suwa et al. (2002,
028588) demonstrated more advanced atherosclerotic lesions based on phenotype and
volume fraction in the left main and right coronary arteries of rabbits exposed to PM10 (5
mg/kg, 2 times/wk x 4 wks). Although this study was conducted using IT exposure
methodology at a relatively high dose, it provided the first experimental evidence that PM
exposure may result in progression of atherosclerosis. Recent toxicological studies
conducted using inhalation exposures have replicated these findings at relevant
concentrations and are discussed below.
CAPs
New studies have demonstrated increased atherosclerotic plaque area in aortas of
ApoE"'" mice exposed to PM2.5 CAPs for 4-6 months (6 h/day x 5 days/wk). Average CAPs
concentrations ranged from 85 to 138 pg/m3 and all were conducted in Tuxedo or
Manhattan, NY. Chen and Nadziejko (2005, 087219) reported that the percentage of aortic
intimal surface covered by atherosclerotic lesions in ApoE"'" mice was increased. In male
ApoE'/LDLR'" mice, both lesion area and cellularity in the aortic root were enhanced by
Tuxedo, NY CAPs exposure, although there was no change in lipid content. Genetic profiles
within plaques recovered from ApoE"'" mice included many of the molecular pathways
known to contribute to atherosclerosis, including inflammation (Floyd et al., 2009, 190350).
Sun (2005, 186814) similarly demonstrated an enhancement of atherosclerosis in ApoE"'"
mice exposed Tuxedo, NY CAPs. Plaque area in the aortic arch and abdominal aorta was
significantly increased in the PM-exposed, high fat-chow group compared to air-exposed,
high fat-chow group. Macrophage infiltration in the abdominal aorta was also observed in
July 2009
7-6
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
the groups exposed to CAPs. A study conducted in Manhattan for 4-months (5-9/2007)
showed that PM2.5 CAPs exposure increased atherosclerotic plaque area and led to higher
levels of macrophage infiltration, collagen deposition, and lipid composition in thoracic
aortas of ApoE"'" mice (Ying et al., 2009, 190111). which is consistent with the previous two
studies described that were conducted in Tuxedo, NY.
Alteration of vasomotor function has been observed in aortic rings of ApoE-'- mice on a
high fat diet with long-term exposure to CAPs (Sun et al., 2005, 186814; Ying et al., 2009,
190111). Sun (2005, 186814) reported that PM2.5-exposed animals exhibited increased
vasoconstrictor responsiveness to serotonin and PE. Increased ROS and elevated iNOS
protein expression in aortic sections of CAPs-exposed mice may have resulted alterations in
the NO pathway and generation of peroxynitirite that could have affected vascular
reactivity. In contrast, Ying, et al. (2009, 190111) demonstrated decreased maximum
constriction induced by PE following Manhattan CAPs exposure. Pretreatment with the
soluble guanylate cyclase (sGC) inhibitor ODQ attenuated the response, indicating that
CAPs exposure resulted in abnormal NO/sGC signaling. iNOS mRNA and protein
expression was increased in aortas of CAPs-exposed mice, further supporting a role for NO
production. In conjunction with increased NO, aortic superoxide production was
demonstrated that appeared to be partially driven by increased NADPH oxidase activity.
The difference in vasoconstrictor responses between these two studies may be attributable
to varying durations (6 vs. 4 months, respectively) or CAPs compositions.
Sun (2005, 186814) and Ying et al. (2001, 019011) reported similar relaxation
responses to ACh for air- and CAPs-exposed mice. However, Manhattan CAPs-exposed mice
had a markedly decreased response to A23187, indicating that NO release occurred via
Ca2+-dependent mechanisms (Ying et al., 2009, 190111). Abnormal eNOS function is likely
responsible for the decreased relaxation response, as activation of eNOS (but not iNOS) is
Ca2+-dependent.
A recent study (Sun et al., 2008, 157033) that was part of the research described
above (Sun et al., 2005, 186814) investigated tissue factor (TF) expression in aortas, which
is a major regulator of hemostasis and thrombosis following vascular injury or plaque
erosion. In PM2.5-exposed ApoE"'" mice on a high-fat diet, TF was significantly elevated in
the plaques of aortic sections compared to air-exposed mice on the high-fat diet. TF
expression was generally detected in (l) the extracellular matrix surrounding macrophages
and foam cell-rich areas and (2) around smooth muscle cells.
One new study of CAPs PM2.5 or ultrafine PM derived from traffic was conducted.
Araujo et al. (2008, 156222) compared the relative impact of ultrafine (0.01-0.18 pm) and
fine (0.01-2.5 pm) PM inhalation on aortic lesion development in ApoE"'" mice following a
40"day exposure (5 h/day x 3 days/wk for 75 total hours). Animals were on a normal chow
diet and exposed to CAPs in a mobile inhalation laboratory parked 300 m from a freeway in
downtown Los Angeles. Exposure concentrations were -440 pg/m3 for PM2.5 and -110 pg/m3
July 2009
7-7
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
for ultrafine PM, and the number concentrations were roughly equivalent (4.56 x 105 and
5.59 x 10B particles/cm3 for PM2.5 and ultrafine PM, respectively). Significant increases in
plaque size (estimated by lesions at the aortic root) were reported for mice exposed to
ultrafine PM only. The lesions were largely comprised of macrophages with intracellular
lipid accumulation. Increased total cholesterol measured at the end of the exposure protocol
was observed only in the PM2.5 group. HDL isolated from the ultrafine PM-exposed mice
demonstrated decreased anti-inflammatory protective capacity against LDL-induced
monocyte chemotactic activity in an in vitro assay. The livers from the ultrafine
PM-exposed mice demonstrated significant increases in lipid peroxidation and several
stress-related gene products (catalase, glutathione S-transferase Ya, NADPH-quinone
oxidoreductasel, superoxide dismutase 2). Thus, ultrafine PM in these exposures had a
substantially greater impact on the systemic response than did PM2.5.
PM10
A study employing young BALB/c mice examined the effects of a 4-month exposure (24
h/day x 7 days/wk) to ambient air on arterial histopathology (Lemos et al., 2006, 088594).
Outdoor exposure chambers were located in downtown Sao Paulo, Brazil next to streets of
high traffic density. In the control chamber, PM10 and NO2 were filtered with 50% and 75%
efficiency, respectively. The average pollutant concentrations were 2.06 ppm for CO (8-h
mean), 104.75 pg/m3 for NO2 (24-h mean), 11.07 pg/m3 for SO2 (24-h mean), and 35.52 pg/m3
for PM10 (24-h mean) at a monitoring site within 100 m of the inhalation chambers. The
pulmonary and coronary arteries demonstrated significant decreases in LAV ratio for
animals exposed to the entire ambient mixture compared to controls, indicating thicker
walls in these vessels. There was no difference reported for the LAV ratio in renal arteries.
Morphologic examination suggested that the increases in LAV ratio were due to muscular
hypertrophy rather than fibrosis. The results of this study indicate vascular remodeling of
the pulmonary and coronary arteries, as opposed to changes in tone.
To examine the role of systemic inflammation and recruitment of monocytes into
plaque tissue as a possible pathway for accelerated atherosclerosis, Yatera et al. (2008,
157162) exposed female Watanabe heritable hyperlipidemic rabbits (42 weeks old) to
Ottawa PM10 (EHC-93) via intratracheal instillation (5 mg/rabbit; approximately 1.56
mg/kg) twice a week for 4 wk. Transfusion of whole blood harvested to from exposed and
non-exposed animals to donor rabbits supplied labeled monocytes for assessment of
monocyte recruitment from the blood to the aortic wall. The fraction of aortic surface and
volume of aortic wall taken up by atherosclerotic plaque was increased and the number of
labeled monocytes in the atherosclerotic plaques was elevated in rabbits exposed to PM10.
In addition, labeled monocytes were attached onto the endothelium overlying
atherosclerotic plaques and the number that migrated into the smooth muscle underneath
plaques in aortic vessel walls was greater with PM10 exposure compared to control. These
July 2009
7-8
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
responses were not observed in normal vessel walls. ICAM-1 and VCAM-1 expression was
elevated in atherosclerotic lesions, likely indicating enhanced monocyte adhesion to
endothelium and migration into plaques. Monocytes in plaque tissue stained with
immunogold demonstrated foam cell characteristics, which were more numerous in the
rabbits exposed to PMio.
Gasoline Exhaust
Lund and colleagues (2007, 125741) used whole emissions from gasoline exhaust to
investigate changes in the transcriptional regulation of several gene products with known
roles in both the chronic promotion and acute degradation/destabilization of atheromatous
plaques. These 50-day exposures (6 h/day x 7 days/wk) employed ApoE"'" mice on high-fat
chow and the concentrations of the high exposure group were 61 pg/m3 for PM, 19 ppm for
NOx, 80 ppm for CO, and 12.0 ppm for total hydrocarbons. The average particle number
median diameter was approximately 15 nm (McDonald et al., 2007, 156746). Dilutions of
gasoline engine emissions induced a concentration-dependent increase in transcription of
matrix metalloproteinase (MMP) isoform 9, ET-1, and HOI in aortas! MMP-3 and -9 mRNA
levels were only increased in animals in the highest exposure group. Strong increases in
oxidative stress markers (nitrotyrosine and TBARS) in the aortas were also observed.
However, using a high-efficiency particle trap, they established that most of the effects were
caused by the gaseous portion of the emissions and not the particles. This study did not
directly address lesion area.
7.2.2. Venous Thromboembolism
One epidemiologic study examined the relationship between long term PMio
concentration, venous thromboembolism, and laboratory measures of hemostasis
(prothrombin and activated partial thomboplastin times [PT; PTT]). PT and PTT measure
the extrinsic and intrinsic blood coagulation pathways, the former activated in response to
blood vessel injury, the latter, key to subsequent amplification of the coagulation cascade
and propagation of thrombus (Mackman et al., 2007, 156723). Decreases in PT and PTT are
consistent with a hypercoagulable, prothrombotic state.
7.2.2.1. Epidemiologic Studies
Baccarelli et al. (2007, 090733) studied 2,081 residents (56% female) of the Lombardy
region of Italy whose ages ranged from 18 to 84 years old. In this case-control study of 871
patients with ultrasonographically or venographically diagnosed lower extremity deep vein
thrombosis (DVT) and 1,210 of their healthy friends or relatives (1995-2005), the authors
used arithmetic averaging of PMio data available at 53 monitors in nine geographic areas to
estimate one-yr avg residence-specific exposures. They found -0.09 (95% CP -0.16 to 0) and
July 2009
7-9
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
¦0.18 (95% CL -0.35 to 0) decreases in standardized correlation coefficients for PT as well as
0.02 (95% CL -0.05 to 0.06) and -0.11 (95% CI: -0.29 to 0.12) decreases in standardized
correlation coefficients for PTT among cases and controls, respectively, per 10 pg/m3
increase in PMio. Patients with DVT who were taking heparin or coumarin anticoagulants
were not asked to stop taking them before measurement of PT and aPTT. Of additional
note, PT was neither adjusted for differences in reagents used to determine it nor
conventionally reported as the International Normalized Ratio (INR). The ambient PM
concentrations from this study are characterized in Table 7-1.
7.2.3.	Diabetes
7.2.3.1. Toxicological Studies
Diabetics as a potentially susceptible subpopulation have only recently been
evaluated. A toxicological study of a diet-induced obesity mouse model (C57BL/6 fed high-
fat chow for 10 wk) examined the effects of a 128-day PM2.5 CAPs exposure (mean mass
concentration 72.7 pg/m3; Tuxedo, NY) on insulin resistance, adipose inflammation, and
visceral adiposity (Sun et al., 2009, 190487). Elevated fasting glucose and insulin levels
were observed in CAPs-exposed mice compared to air-exposed during the glucose tolerance
test. Aortic rings of mice exposed to CAPs demonstrated decreased peak relaxation to ACh
or insulin, which was associated with reduced NO bioavailability. Additionally, insulin
signaling was impaired in aortic tissue via lowered endothelial Akt phosphorylation.
Increases in adipokines and systemic inflammatory markers (i.e., TNF-a, IL-6, E-selectin,
ICAM-1, PAP1, resistin, leptin) were reported for CAPs-exposed mice. CAPs resulted in
increased visceral and mesenteric fat mass, as well as greater adipose tissue macrophages
in epididymal fat pads and larger adipocyte size compared to mice in the filtered air group.
The results of this study demonstrate that PM2.5 exposure can exaggerate insulin
resistance, visceral adiposity, and inflammation in mice fed high-fat chow.
7.2.4.	Systemic Inflammation, Immune Function, and Blood
Coagulation
7.2.4.1. Epidemiologic Studies
Chen and Schwartz (2008, 190106) studied 2,978 residentially stable participants in
33 U.S. communities (age range = 20-89 years! 49% female) who were examined during
phase 1 of the National Health and Nutrition Examination Survey III (1989-1991). In this
cross-sectional study, the authors used inverse-distance weighted averaging of U.S. EPA
AQS monitored data from participant and adjacent counties of residence to estimate one-
July 2009
7-10
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
yr avg exposures to PMio (median concentration within quartiles = 23.1, 31.2, 38.8 and 53.7
jug/m3). They found that after adjustment, residents of communities in quartile 1 had 138
(95% CP 2-273) fewer white blood cells (x 106/L) than residents of communities in quartiles
2-4. This difference increased with increasing number of metabolic abnormalities (insulin
resistance! hypertension! hypertriglyceridemia! low high-density lipoprotein cholesterol!
abdominal obesity) reported by the participant.
Forbes et al. (2009, 190351) studied approximately 25,000 adults (age > 16 yr! 53%
female) who were representatively sampled from 720 English postcode sectors and
participated in the Health Survey for England (1994, 1998 and 2003). In this fixed-effects
meta-analysis of year-specific cross-sectional findings, the authors used dispersion
modeling of emissions and weather data to estimate two-yr avg exposures to PMio at
participant postcode sector centroids (median in 1994, 1998 and 2003 = 19.5, 17.9 and 16.2
ug/ma). They found little evidence of a PMicrinflammatory marker association, i.e., only a -
0.08% (95% CP -0.25% to 0.10%) decease in fibrinogen concentration and a 0.14% (95%
CP -1.00% to 1.30%) increase in CRP concentration per 1 pg/m3 increase in PMio.
Calderon-Garciduenas et al. (2007, 091252) compared residentially stable, non-
smoking healthy children (age range: 6-13 yr) living and attending school between 2003-
2004 in Mexico City (historically high PM! altitude 2,250 m) and Polotitlan (historically low
PM! altitude 2,380 m). In this ecologic study, residents of Mexico City (n = 59! 93% female)
had fewer white blood cells and neutrophils (x 109/L) than residents of Polotitlan (n = 22!
69% female): unadjusted mean 6.2 (95% CP 5.7-6.6) versus 6.9 (95% CP 6.3-7.5) and 2.9
(95% CP 2.3-3.5) versus 3.8 (95% CP 3.2-4.4), respectively.
Calderon-Garciduenas et al. (2009) subsequently compared 37 unadjusted mean
measures of immune function and inflammation among an expanded number of these
participants. They found that under a two-sided type I error rate (a) = 0.05, sixteen (43%) of
the measures were significantly different in residents of southwest Mexico City (n = 66!
48% female) than those in Polotitlan (n = 93! 57% female). However, only eight measures
were significantly different after Bon Ferroni-correction (a = 0.05 / 37 = 0.001) and even
fewer would be after adjustment for reported correlation between the measures of immune
function and inflammation, e.g., CRP and lipopolysaccharide binding protein (Pearson's r =
0.71).
Only two cross-sectional analyses of PMio concentration and markers of immune
function or inflammation have been conducted with significant changes observed in the
NHANES population (stronger effects among those with metabolic disorders) but not in a
relative large survey of adults, which was conducted in England. Ecological analyses
comparing children in high versus low pollution regions in Mexico show differences in
unadjusted blood markers that may be related to PM concentration or other unmeasured
risk factors that differs across the communities studied.
July 2009
7-11
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
7.2.4.2. Toxicological Studies
In addition to the PM2.5 study mentioned previuosly that showed increased TF
expression (an important initiator of thrombosis) in aortas of ApoE"'" mice following
subchronic CAPs exposure (Sun et al., 2008, 157033). three recent studies examined
hematology and clotting parameters in rats and mice exposed to DE, gasoline exhaust, or
HWS for 1 week or 6 months (Reed et al., 2004, 055625; Reed et al., 2006, 156043; Reed et
al., 2008, 156903). In all studies, male and female F344 rats were exposed to the mixtures
by whole-body inhalation for 6 h/day, 7 day/wk. Respiratory effects for these studies are
presented in Section 7.3.3.
Diesel Exhaust
The target PM concentrations in the diesel exhaust study was 30, 100, 300, and
1000 pg/m3 and the MMAD was 0.10-0.15 pm (Reed et al., 2004, 055625). Male and female
rats exposed to DE at the highest concentration (NO concentration 45.3 ppm! NO2
concentration 4.0 ppm! CO concentration 29.8 ppm! SO2 concentration 365 ppb) for 6
months demonstrated decreased serum Factor VII, but no change in plasma fibrinogen or
TAT (Reed et al., 2004, 055625). White blood cells were decreased only in female rats in the
highest exposure group. Another DE study of shorter duration (4 wk, 4 h/day, 5day/wk; PM
mass concentration 507 or 2201 pg/m3, CO 1.3 and 4.8 ppm, NO <2.5 and 5.9 ppm, NO2
<0.25 and 1.2 ppm, SO2 0.2 and 0.3 ppm for low and high PM exposures, respectively) did
not demonstrate changes in hematologic parameters or those related to coagulation (i.e.,
plasma prothrombin time, activated partial thromboplastin time, plasma fibrinogen, D-
dimer) or inflammation (i.e., CRP) in SH or WKY rats (Gottipolu et al., 2009, 190360).
Together, these findings do not support a DE-related stimulation of blood coagulation
following 1 or 6 months of exposure.
Hardwood Smoke
The target PM concentrations in the HWS study was 30, 100, 300, and 1000 pg/m3
and the MMAD was 0.25-0.36 pm (Reed et al., 2006, 156043). In male rats exposed to HWS,
the mid-low group (PM concentration 113 pg/m3! NO, NO2, SO2 concentrations 0 ppm! CO
concentration 1832.3 ppm) had the greatest responses in hematology parameters, including
increased hematocrit, hemoglobin, lymphocytes, and decreased segmented neutrophils
(Reed et al., 2006, 156043). Platelets were elevated in male and female rats after 1 week of
exposure, but this response returned to control values following the 6 month exposure. No
changes were observed for any coagulation markers at 6 months.
Gasoline Exhaust
PM mass in the gasoline exhaust study ranged from 6.6 to 59.1 pg/m3, with the
corresponding number concentration between 2.6 x 104 and 5.0 x 105 particles/cm3; the
July 2009
7-12
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
dilutions for the gasoline exhaust were 1^10, 1^15 or 1:90 and filtered PM at the 1^10
dilution (Reed et al., 2008, 156903). Similar to the responses observed with HWS, male and
female rats in the mid and high gasoline exhaust exposure groups (NO concentrations 11.9
and 18.4 ppm! NO2 concentrations 0.5 and 0.9 ppm! CO concentration 73.2 and 107.3 ppm!
SO2 concentration 0.38 and 0.62 ppm, respectively) demonstrated elevated hematocrit and
hemoglobin; RBC count was also elevated in these groups (Reed et al., 2008, 156903). The
only response that appeared somewhat dependent on the presence of particles was
increased RBC in female rats at 6 months, although the authors attributed the observed
increases to the high concentration of CO.
Collectively, these studies do not indicate robust systemic inflammation or coagulation
responses in F344 rats following 6-month exposures to diesel, HWS, or gasoline exhaust.
The limited effects that were observed could possibly be due to the varying gas
concentrations in the exposure mixtures.
7.2.5. Renal and Vascular Function
Two recent epidemiologic studies have tested associations between PM exposure and
indicators of renal and vascular function (urinary albumin to creatinine ratio [UACR] and
blood pressure). UACR is a measure of urinary albumin excretion (National Kidney, 2008,
156796). When calculated as the ratio of albumin to creatinine concentrations in untimed
("spot") urine samples, UACR approximates 24-h urinary albumin excretion and can be
used to identify albuminuria, a marker of generalized vascular endothelial damage (Xu et
al., 2008, 157157). Values > 30 mg/g (3.5 mg/mmol) and > 300 mg/g (34 mg/mmol) usually
define micro- and macroalbuminuria, both of which are associated with increases in CVD
incidence and mortality (Bigazzi et al., 1998, 156272; Deckert et al., 1996, 156389; Dinneen
and Gerstein, 1997, 156403; Gerstein et al., 2001, 156466; Mogensen, 1984, 156769).
Several researchers have called the dichotomization of albuminuria into question, observing
that there is no threshold below which risk of cardiovascular and end-stage kidney disease
disappears (Forman and Brenner, 2006, 156439; Knight and Curhan, 2003, 179900;
Ruggenenti and Remuzzi, 2006, 156933).
Systolic, diastolic, pulse, and mean arterial blood pressures (SBP; DBP; PP; MAP) in
mmHg have also been used as measures of cardiovascular disease. Franklin et al. (1997,
156446) suggested that SBP and PP were the only two measures predictive of carotid
stenosis in a multivariable analysis considering all four measures, whereas Khattar et al.
(2001, 155896) suggested that their prognostic significance in hypertensive populations
may differ by age, with SBP and PP being most predictive among those > 60 and DBP
among those <60 years old (Khattar et al., 2001, 155896).
July 2009
7-13
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
7.2.5.1. Epidemiologic Studies
O'Neill et al. (2007, 156006) examined the association of UACR with PM2.5 and PM10
among members of the MESA population described previously (Diez Roux et al., 2008,
156401). For this study of UACR, which included cross-sectional and longitudinal analyses,
the study population was restricted to a subset of 3,901 participants (mean age = 63 yr! 52%
female) with complete covariate, outcome and exposure data at their first through third
exams (2000-2004). In cross-sectional analyses, the authors found that after adjustment for
demographic and clinical characteristics, 10 pg/m3 increases in 20-year imputed exposures
to PM2.5 and PM10 were associated with negligible 0.002 (95% CP -0.048 to 0.052) and -0.002
(95% CP -0.038 to 0.035) mean differences in baseline log UACR, respectively. Similarly,
small non-statistically significant decreases in the prevalence of microalbuminuria (defined
in this setting as > 25 mg/g) provided little evidence of an effect on renal function. These
largely null cross-sectional findings mirrored those based on the study's shorter-term (30-
and 60-day) PM2.5 and PM10 exposures. Moreover, longitudinal analyses revealed only a
weak association between three-year change in log UACR and 20-yr PM10 exposure.
Evidence of effect modification by demographic and geographic characteristics was not
apparent in either the cross-sectional or longitudinal analyses.
Auchincloss et al. (2008, 156234) focused on automated, oscillometric,
sphygmomanometric measures of blood pressures in mmHg (SBP; DBP; PP; MAP). Like
O'Neill (2007, 156006), Diez et al. (2008, 156401) and Allen et al. (2007, 156006),
Auchincloss et al. (2008, 156234) based their examination on the previously described
MESA population. The authors included 5,112 study participants (age range = 45-84 years!
52% female) who were free of clinically manifested CVD at their baseline exam in one of six
primarily urban U.S. locations (2000-2002). In this cross-sectional study, they used
arithmetic averaging of EPAAQS PM2.5 data available at the monitor nearest to each
participant's geocoded U.S. Postal Service ZIP code centroid to estimate 30- and 60-day avg
exposures to PM2.5. They found small nonsignificant increases of 1.5 (95% CP -0.2 to 3.2),
0.2 (95% CP -0.7 to 1.0), 1.3 (95% CP 0.1 to 2.6), and 0.6 (95% CP -0.4 to 1.7) mmHg
increases in SBP, DBP, PP and MAP, respectively, per 10 ug/m3 increase in 30-day avg PM2.5
exposure, Associations were slightly weaker for 60-day avg PM2.5 exposure and among
participants without hypertension, during cooler weather, in the presence of low NO2,
residing >300 m from a highway, or surrounded by lower road density.
Finally, the Calderon-Garciduenas et al. (2007, 091252) ecologic study introduced in
Section 7.2.3.1 also found that children residing in Mexico City had higher mean pulmonary
artery pressure as assessed by Doppler echocardiography and fasting plasma endothelin-1
(ET-l) than residents in Polotitlan: unadjusted mean 17.5 (95% CP 15.7-19.4) versus 14.6
(95% CP 13.8-15.4) mmHg and 2.23 (95% CP 1.93-2.53) versus 1.23 (95% CP 1.11-1.35)
pg/mL, respectively. Within Mexico City, ET-l was higher in residents of the Northeast
(historically higher PM2.5) than those of the Southwest (historically lower PM2.5).
July 2009
7-14
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
The MESA analyses of UACR (O'Neill et al., 2007, 156006) and the ecologic study of
children living in a highly polluted area of Mexico (Calderon-Garciduenas et al., 2007,
091252) provide little evidence that long-term exposure to PM2.5 had an effect on renal and
vascular function. Auchincloss et al. (2008, 156234) reports small nonsignificant
associations of blood pressure with 30 and 60 day avg PM2.5 concentrations. PM
concentrations from the analyses are characterized in Table 7-1.
Table 7-1. Characterization of ambient PM concentrations from studies of subclinical measures of
cardiovascular diseases.
Reference
Location
Mean Concentration (|JCj/m3)
Upper Percentile Concentrations
(|jg/m3)
PMia
Diez et al. (2008,156401)
MESA: 6 Cities U.S.
20 yr imputed mean: 34
NR
O'Neill etal. (2007,156006)
MESA: 6 Cities U.S.
Long-Term Exposure:
1982-2002: 34.7
1982-1987:40.5
1988-1992:38
1993-1997: 30.6
1998-2002: 29.7
Previous Month: 27.5
NR
Baccarelli et al. (2008,157984)
Lombardy Italy
Annual avg: 41
NR
Rosenlund et al. (2006, 089796)
Stockholm, Sweden
30-y avg PMio (traffic)
Cases: 2.6
Controls: 2.4
5th-95th %: 0.5-6
0.6-5.9
PMzs
Hoffmann et al.(2007, 091163)
HNRS, 3 Cities Germany
Annual avg: 22.8
NR
Allen etal. (2007,156006)
MESA: 5 Cities
Annual avg: 15.8
Min-Max: 10.6-24.7
Kunzlietal. (2005,087387)
VEAPS BVAIT
Annual avg: 20.3
Min-Max: 5.2-26.9
Auchincloss et al. (2008,156234)
MESA: 6 Cities
Prior 30 days: 16.8
Prior 60 days: 16.7
NR
O'Neill (2007,156006)
MESA: 6 Cities U.S.
Previous Month: 16.5
NR
Diez et al. (2008,156401)
MESA: 6 Cities U.S.
20-y imputed mean: 21.7
NR
Hoffmann et al. (2009,190376)
HNRS: 3 Cities Germany
Annual avg: 22.8
Min-max: 19.8-26.8
MESA: Multi-Ethnic Study of Atherosclerosis
HNRS: Heinz Nixdorf Recall Study
VEAPS: Vitamin E Atherosclerosis Progression Study
BVAIT: B-Vitamin Atherosclerosis Intervention Trial
7.2.5.2. Toxicological Studies
In a PM2.5 CAPs study of 10 wk (6 h/day x 5 days/wk) in Tuxedo, NY (mean mass
concentration 79.1 pg/m3), there was no difference in mean arterial pressure (MAP) in SD
rats between groups (Sun et al., 2008, 157032). When angiotensin II (Ang II) was infused
during the last week of exposure to induce systemic hyptertension, the MAP slope was
consistently greater in the CAPs-exposed rats compared to the filtered air group.
Furthermore, thoracic aortic rings were more responsive to phenylephrine-induced
constriction and less responsive to ACh-induced relaxation in the PM+Ang II vessels. In
July 2009
7-15
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
contrast to the latter findings, the relaxation response was exaggerated in the PM+Ang II
aortic segments with a Rho-kinase (ROCK) inhibitor. Superoxide production in aortic rings
increased in the PM+Ang II group compared to the filtered air group and the addition of
NAD(P oxidase inhibitor (apocymin) or a NOS inhibitor (L-NAME) attenuated the
superoxide generation. The levels of tetrahydrobiopterin (BH4) were decreased in
mesenteric vasculature and the heart by 46% and 41% in the PM+Ang II group compared to
controls, respectively! furthermore, levels of BH4 in the liver were similarly reduced, which
is consistent with a systemic effect of CAPs. Together, these findings indicate that CAPs
potentiate Ang 11-induced hypertension and alter vascular reactivity, perhaps through
activated NADPH oxidase and eNOS uncoupling that result in oxidative stress generation
and triggering of the Rho/ROCK signaling pathway.
7.2.6. Autonomic Function
7.2.6.1. Toxicological Studies
Hwang et al. (2005, 087957) and Chen and Hwang (2005, 087218) used
radiotelemetry to examine the chronic changes in HR and HRV resulting from the same
CAPs exposures described previously (Chen and Nadziejko, 2005, 087219). The overall
average CAPs exposure concentration was 133 pg/m3 and results indicate differing
responses to CAPs between ApoE"'" mice and their genetic background strain, C57BL/6J
mice (Hwang et al., 2005, 087957). Using the time period of 1^30 to 4:30 a.m., C57BL/6J
mice showed a HR increase only over the last month of exposure. In contrast, ApoE '" mice
had chronic decreases of 33.8 beat/min for HR. Changes in HRV (SDNN and rMSSD) were
somewhat more complicated, with biphasic responses in ApoE"'" mice over the 5 month
period (initial increase over first 6 wk, decrease over next 12 wk, and slight upward turn for
remainder of the study)(Chen and Hwang, 2005, 087218). Increasing linear trends were
observed in C57BL/6J mice for SDNN and rMSSD. The average CAPs concentration for the
HRV study was 110 pg/m3. However, only 3 C57BL/6J mice in the exposure group were
included in the analysis compared to 10 ApoE"'" animals, thus making it difficult to
interpret the C57BL/6J mice responses (Chen and Hwang, 2005, 087218; Hwang et al.,
2005, 087957).
July 2009
7-16
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
7.2.7. Cardiac changes
7.2.7.1. Toxicological studies
Two recent toxicological studies have evaluated the effects of PM on cardiac effects
including pathology and gene expression. Cardiac mitochondrial function has also been
evaluated following PM exposure in rats.
Diesel Exhaust
A recent study of DE exposure (PM mass concentration 507 or 2201 pg/m3, CO 1.3 or
4.8 ppm, NO <2.5 or 5.9 ppm, NO2 <0.25 or 1.2 ppm, SO2 0.2 or 0.3 ppm for low and high
PM exposures, respectively! geometric median number diameter 85 11m) indicated a
hypertensive "like cardiac gene expression in WKY rats that mimicked baseline patterns in
air-exposed SH rats (Gottipolu et al., 2009, 190360). Exposure to the high concentration of
DE for 4 wk (4 h/day, 5 day/wk) led to downregulation of genes involved in stress,
antioxidant compensatory response, growth and extracellular matrix regulation, membrane
transport of molecules, mitochondrial function, thrombosis regulation, and immune
function. No genes were affected by DE in SH rats. A dose-dependent inhibition of
mitochondrial aconitase activity in both rat starins was observed, indicating a DE effect on
oxidative stress. It should be noted that while DE-related cardiovascular effects were found
in WKY rats only, pulmonary inflammation and injury were observed in both strains (see
Sections 7.3.3.2. and 7.3.5.1.).
Model Particles
Wallenborn et al. (2008, 191171) examined the subchronic (5 h/day, 3 day/wk, 16 wk)
pulmonary, cardiac, and systemic effects of nose-only exposure to particulate ZnS04 (9, 35,
or 120 pg/m3) in WKY rats. Particle size was reported to be 31-44 nm measured as number
median diameter. Although changes in pulmonary inflammation or injury and cardiac
pathology were not observed, effects on cardiac mitochondrial protein and enzyme levels
were noted (i.e., increased ferritin levels, decrease in succinate dehydrogenase activity),
possibly indicating a small degree of mitochondrial dysfunction. Glutathione peroxidase, an
antioxidant enzyme, was also decreased in the cardiac cytosol. Gene expression analysis
identified alterations in cardiac genes involved in cell signaling events, ion channels
regulation, and coagulation in animals exposed to the highest ZnS04 concentration only.
This study demonstrates a possible direct effect of ZnS04 on extrapulmonary systems, as
suggested by the lack of pulmonary effects (see Section 7.3.3.2).
July 2009
7-17
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
7.2.8. Left Ventricular Mass and Function
Van Hee et al. (2009) studied 3,827 participants (age range = 45-84 yr! 53% female)
who underwent magnetic resonance imaging (MRI) of the heart at the baseline examination
of the MESA cohort (2000-2002). This cross-sectional study focused on two MRPbased
outcome measures^ left ventricular mass index (LVMI, g/m2) and ejection fraction (EF, %),
the former estimated using the DuBois formula for body surface area, the latter as the ratio
of stroke volume to end diastolic volume. The study also estimated annual mean exposures
to PM2.5 at participants' geocoded residential addresses in 2000 using ordinary kriging of
U.S. EPAAQS concentration data. In fully adjusted models, it found 3.8 (95% CP -6.1 to
13.7) g/m2 and -3.0% (-8.0% to 2.0%) differences in LVMI and EF per 10 jug/m3 increment in
PM2.5. The findings were small and imprecise, albeit suggestive of a slight, PM-associated
increase in the mass and decrease in the function of the left ventricle. The effect of living
within 50 m of a major roadway on LVMI was greater than the effect of PM2.5 (i.e., 1.4 g/m2
[95% CP 0.3-2.5] per 10 |ug/m3.)
Table 7-2. Characterization of ambient PM concentrations from studies of clinical cardiovascular
diseases.



Reference
Location
Mean Annual
Concentration ([jg/m3)
Upper Percentile
Concentrations (|jg/m3)
PMia
Puett et al. (2008,156891)
13 U.S. States
21.6

Zanobetti and Schwartz (2007, 091247)
21 U.S. Cities
28.8
Overall range NR
Rosenlund et al. (2006, 089796)
Stockholm, Sweden
30 y avg PMio (traffic)
5th-95th Percentile


Cases: 2.6
0.5-6.0


Controls: 2.4
0.6-5.9
Maheswaran et al. (2005, 090769)
Sheffield, U.K.
Range of means in each quintile: 16-23.3
NR
Baccarelli et al. (2007, 090733)
Lombardy, Italy
Sep-Nov: 51.2
148.9


Dec-Feb: 68.5
238.3


Mar-May:64.1
158.5


Jun-Aug: 44.3
94.7
PMzs
Miller et al. (2007, 090130)
WHI: 36 Metropolitan areas
Citywide avg (yr 2000): 13.5
Min-max: 4-19.3
Hoffmann et al. (2006, 091162)
HNRS: 2 Cities Germany
23.3
NR
WHI: Womens Health Initiative
HNRS: Hans Nixdorf Recall Study
7.2.9. Clinical Outcomes in Epidemiologic Studies
Several epidemiologic studies of U.S. and European populations have examined
associations between long-term PM exposures and clinical CVD events (Baccarelli et al.,
July 2009
7-18
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
2008, 157984; Hoffmann et al., 2006, 091162; Hoffmann et al., 2009, 190376; Maheswaran
et al., 2005, 088683; Maheswaran et al., 2005, 090769; Miller et al., 2007, 090130;
Rosenlund et al., 2006, 089796; Solomon et al., 2003, 156994; Zanobetti and Schwartz,
2007, 091247). Results from these studies are summarized in Figure 7-1. The ambient PM
concentrations from these studies are characterized in Table 7-2.
Coronary Heart Disease
Epidemiologic studies examining the association of coronary heart disease (CHD) with
long-term PM exposure are discussed below (Hoffmann et al., 2006, 091162; Maheswaran et
al., 2005, 090769; Miller et al., 2007, 090130; Rosenlund et al., 2006, 089796; Rosenlund et
al., 2009, 190309; Zanobetti and Schwartz, 2007, 091247). Cases of CHD were variably
defined in these studies to include history of angina pectoris, MI, coronary artery
revascularization (bypass graft; angioplasty! stent; atherectomy), and congestive heart
failure (CHF). Results pertaining to death from CHD are described in Section 7.6.
Miller et al. (2007, 090130) studied incident, validated MI, revascularization, and
CHD death, both separately and collectively, among 58,610 post-menopausal female
residents of 36 U.S. metropolitan areas (age range = 50-79 yr) enrolled in the Women's
Health Initiative Observational Study (WHI OS, 1994-1998). In this prospective cohort
study of participants free of CVD at baseline (median duration of follow-up = 6 yr), the
authors used arithmetic averaging of year 2000 EPAAQS PM2.5 data available at the
monitor nearest to each participant's geocoded U.S. Postal Service five-digit ZIP code
centroid to estimate one-yr avg exposures. They found 6% (95% CP -15% to 34%), 20% (95%
CP 0-43%) and 21% (95% CP 4-42%) increases in the overall risk of MI, revascularization,
and their combination with CHD death per 10 pg/m3 increase in PM2 .5, respectively. Hazards
were higher within than between cities and in the obese. For the combined CVD outcome
(MI, revascularization, stroke, CHD death, cerebrovascular disease), authors reported a
24% (95% CP 9-41%) increase in risk that was higher among participants at higher than
lower quintiles of body mass index, waist-to-hip ratio, and waist circumference. The
PM2.5-CVD association was stronger among non-diabetic than diabetic participants.
Puett et al. (2008, 156891) studied incident, validated CHD, CHD death, and non-
fatal MI among 66,250 female residents (mean age = 62 years) of metropolitan statistical
areas in thirteen northeastern U.S. states who were enrolled in the Nurses' Health Study
(NHS, 1992-2002). In this prospective cohort study of women without a history of non-fatal
MI at baseline (maximum duration of followup = 4 years), the authors used two-stage,
spatially smoothed, land use regression to estimate residence-specific, 1-yr ma PM10
exposures from U.S. EPAAQS and emissions, IMPROVE, and Harvard University monitor
data. They found a 10% (95% CP -6 to 29) increase in risk of first CHD event per 10 p/m3
increase in 1-yr avg PM10 exposure, while the association with MI was close to the null
value. The association with fatal CHD event of 30% (95% CP 0-71%) was stronger.
July 2009
7-19
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Furthermore, associations with CHD death were higher in the obese and in the never
smokers.
Rosenlund et al. (2006, 089796) studied 2,938 residents of Stockholm County, Sweden
(age range = 45-70 years! 34% female). In this case-control study of 1,085 patients with
their first, validated non-fatal MI and an age-, gender- and catchment-stratified random
sample of 1,853 controls without MI (1992-1994), the authors used street canyon-adjusted
dispersion modeling of emissions data to estimate 30-yr avg exposure to PMio
(median = 2.4 pg/m3). They found that the OR for prevalent MI per 10 pg/m3 increase in
PMio was 0.85 (95% CI: 0.50-1.42). The OR for fatal MI was non-significantly elevated.
In a more recent study, Rosenlund et al. (2009, 190309) evaluated 554,340 residents
(age range = 15-79 years! 49% female) of Stockholm County, Sweden (1984-1996). In this
population-based, case-control study of 43,275 cases of incident, validated MI, the authors
used dispersion modeling of traffic emissions and land use data to estimate 5-yr avg
exposure to PMio. They found that after adjustment for demographic, temporal, and
socioeconomic characteristics, the OR for MI per 5 pg/m3 increase in PMio was 1.04 (95% CP
1.00-1.09). ORs were higher after restriction to fatal cases, in- or out-of-hospital deaths, and
participants who did not move between population censuses. Authors state that control for
confounding was superior in their previous study (Rosenlund et al., 2006, 089796) although
the size of the population was larger in this recent study (Rosenlund et al., 2009, 190309).
Zanobetti and Schwartz (2007, 091247) studied ICD-coded recurrent MI (ICD-9 410)
and post-infarction CHF (ICD-9 428) among 196,131 Medicare recipients (age > 65 years!
50% female) discharged alive following MI hospitalization in 21 cities from 12 U.S. states
(1985-1999). In this ecologic, open cohort study of re-hospitalization among MI survivors
(mean duration of followup = 3.6 and 3.7 years for MI and CHF, respectively), the authors
used arithmetic averaging of EPAAQS PMio data available in the county of hospitalization
to estimate one-yr avg exposures. They found 17% (95% CP 5-31%) and 11% (95% CP 3-
21%) increases in the risk of recurrent MI and post-infarction CHF, respectively, per
10 pg/m3 increase in PMio exposure. Hazards were somewhat higher among persons aged
>75 years.
Hoffmann et al. (2006, 091162) studied self-reported CHD (MI or revascularization)
among 3,399 residents of Essen and Mulheim, Germany (age range = 45-75 years! 51%
female) at the baseline exam of the Heinz Nixdorf Recall Study (2000-2003) introduced
previously. In this cross-sectional ancillary study, the authors used dispersion modeling of
emissions, climate and topography data to estimate 1-yr avg exposure to PM2.5
(mean = 23.3 pg/m3). They found little evidence of an association between PM2.5 and CHD in
these data. After adjustment for geographic, demographic and clinical characteristics, the
OR for prevalent CHD per 10 pg/m3 increase in exposure was 0.55 (95% CP 0.14-2.11).
July 2009
7-20
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Avg
Time
Risk Estimate
PROSPECTIVE COHORT STUDIES
Miller et al.(2007, 090130)
1 yr
. Incident Ml
PM2.5
58,610 postmenopausal women enrolled in WHI,
incident, validated cases, 36 US cities
Revascularization
. Stroke
Cerebrovascular Disease
, All CVD
Puett et al. (2008,156891)
1 yr
1st CHD Event
PM11
75,809 women in the Nurses Health Study,
validated cases, 13 metro areas, NE states
CASE CONTROL STUDIES
Rosenlund et al. (2009,190309)
5 yr
. Fatal/Non-Fatal Ml
PM10
24,347 cases, l\l-276,926 randomly selected
population based controls, Stockholm, Sweden
Rosenlund et al. (2006, 089796)
30 yr
, Fatal/Non-Fatal Ml
2,246 cases, N-3,206 randomly selected
population controls, Stockholm, Sweden
Baccarelli et al. (2008,157984)
1 yr
, DVT
871 cases, N -1210 healthy friend controls,
Lombardy, Italy
OTHER STUDY DESIGNS
PM2.5
Hoffman et al. (2006, 091162)
1 yr
. Self-reported CHD
Hoffman et al. (2009,190376)
1 yr
_Self-reported PVD
Prevalent cases at baseline in 3,399 residents,
cross-sectional, 2 German cities
Zanobetti and Schwartz (2007, 091247)
1 yr
^Recurrent Ml Hospitalization
PM10
196,131 Ml survivors (ICD9 410), readmitted for
Ml or CHF, ecologic, open cohort, 21 US cities
Post-MI CHF Hospitalization
Maheswaran et al. (2005, 090769; 2005, 088683) 5 yr
_CHD Hospitalization
Hospital admissions among 199,682 residents,
ecologic design, Sheffield, UK
Stroke Hospitalization
I I I I I I II I I I I I I II I I I I I II
0.0 0.2 0.4 0.8 0.8 1.0 1.2 1.4 1.8 1.8 2.0 22 2.4
Risk estimates for the associations of clinical outcomes with long-term exposure to
ambient PM2.5 and PM10.
Figure 7-1.
1	In the study of 1030 census enumeration districts in Sheffield, U.K. described
2	previously, Maheswaran et al. (2005, 090769) studied 11,407 ICD-10-coded emergency
3	hospitalizations for CHD (ICD10 120-25) among 199,682 residents (age > 45 years! 45%
4	female). In this ecologic study, the authors used dispersion modeling of emissions and
5	climate data to estimate 5-yr avg exposure to PM10. They found that after adjusting for
July 2009
7-21
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
smoking prevalence, controlling for socioeconomic factors, and smoothing, the age- and
gender-standardized rate ratios for CHD admission were 1.01 (95% CI: 0.92-1.11), 1.04
(95% CI: 0.93-1.15), 0.97 (95% CI: 0.87-1.08), and 1.07 (95% CI: 0.95-1.20) across PMio
quintiles. The linear trend was somewhat stronger for CHD mortality (see Section 7.3).
The study of post-menopausal women enrolled in the WHI OS by Miller et al. (2007,
090130) was the only U.S. study to examine the effect of PM2.5 rather than PM10. This
study, which provides strong evidence of an association, was distinguished by its
prospective cohort design, validation of incident cases and large population. Puett et al.
(2008, 156891), the other U.S. study with comparable design features, provides evidence of
an association of incident CHD with long- term PM10 exposure. Findings from Swedish case
control studies of incident validated cases of MI were not consistent. A cross-sectional study
of self-reported CHD did not provide evidence of an association with PM2.5, while findings
from two ecologic studies of PM10 indicated positive associations of CHD hospitalizations
with PM10 (Maheswaran et al., 2005, 088683; Zanobetti and Schwartz, 2007, 091247).
Stroke
Miller et al. (2007, 090130) found 28% (95% CI: 2-61%) and 35% (95% CI: 8-68%)
increases in the overall risk of validated stroke and cerebrovascular disease, respectively,
per 10 pg/m3 increase in one-yr avg PM2.5 exposure. Risks were higher within than between
cities. In the study of 1030 Census of enumeration districts in Sheffield, U.K. described
previously, Maheswaran et al. (2005, 088683) studied 5,122 ICD-10-coded emergency
hospital admissions for stroke (160-69) among 199,682 residents (age > 45 yr! 45% female)
of 1,030 census enumeration districts in Sheffield, U.K. (1994-1999). In this ecologic study,
the authors used dispersion modeling of emissions and climate data to estimate five-yr avg
exposure to PM10. They found that the age- and gender-standardized rate ratios for stroke
admission were 1.05 (95% CI: 0.94-1.17), 1.07 (95% CI: 0.95-1.20), 1.06 (95% CI: 0.94-1.20),
and 1.15 (95% CI: 1.01-1.31) across PM10 quintiles. Linear trend was somewhat stronger for
stroke mortality (see Section 7.6).
These studies examining the long-term PM-stroke relationship provide evidence of
association. Maheswaran et al. (2005, 088683) examined emergency room HAs in Sheffield,
U.K. using an ecologic design while results reported by Miller et al. (2007, 090130) are
based on the prospective cohort study of the WHI OS population (both introduced
previously).
Peripheral Arterial Disease
The German Heinz Nixdorf Recall cross-sectional study described in Section 7.2.1.1
(Hoffmann et al., 2009, 190376) also evaluated the association between one-yr avg exposure
to PM2.5 and peripheral arterial disease (self-reported history of a surgical or procedural
intervention or an ABI <0.9 in one or both legs). The authors found no evidence of an
July 2009
7-22
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
increase in risk. The OR for peripheral arterial disease was 0.87 (95% CI: 0.57-1.34) per 3.9
pg/m3 increase in PM2.5. However, evidence of an association with traffic exposure was
present in these data. ORs of 1.77 (95% CI: 1.01-3.10), 1.02 (95% CI: 0.58-1.80), and 1.07
(95% CI: 0.68-1.68) for residing < 50, 50-100, and 100-200 m of a major road (reference
category: >200 m), respectively were observed. ORs were higher among participants with
CAC scores < 75th percentile, women, and smokers.
Deep Vein Thrombosis
The Italian case-control study (introduced in Section 7.2.1.2) also examined the
chronic PMicrDVT association (Baccarelli et al., 2008, 157984). The authors found a 70%
(95% CI: 30-223%) increase in the odds of deep vein thrombosis (DVT) per 10 pg/m3 increase
in one-yr avg PM10 exposure. This finding was consistent with the decreases in PT and PTT
also observed among controls in this context as well as the 47% (95% CI: 11*96%) increase
in the odds of DVT per inter-decile range (242 m) increase in the residence-to-major-
roadway distance observed among a subset of cases and controls (Baccarelli et al., 2009,
188183). The PMicrDVT and distance-DVT associations were both weaker among women
and among users of oral contraceptives or hormone therapy.
7.2.10. Cardiovascular Mortality
New epidemiologic evidence reports a consistent association between long-term
exposure to PM2.5 and increased risk of cardiovascular mortality. There is little evidence for
the long-term effects of PM10-2.5 on cardiovascular mortality. This section focuses on
cardiovascular mortality outcomes in response to long-term exposure to PM. The studies
that investigate long-term exposure and mortality due to any specific or all (non-accidental)
causes are evaluated in Section 7.6. A summary of the mean PM concentrations reported for
the studies characterized in this section is presented in Table 7-8, and the effect estimates
are presented in Figures 7-7 and 7-8.
A number of large, U.S. cohort studies have found consistent associations between
long-term exposure to PM2.5 and cardiovascular mortality. The American Cancer Society
(ACS) (Pope et al. 2004, 055880) reported positive associations with deaths from specific
cardiovascular diseases, particularly ischemic heart disease, and a group of cardiac
conditions including dysrhythmia, heart failure and cardiac arrest (RR for cardiovascular
mortality = 1.12 [95% CI: 1.08-1.15] per 10 |ug/m3 PM2.5). In an additional reanalysis that
extended the follow-up period for the ACS cohort to 18 years (1982-2000) (Krewski et al.,
2009, 191193), investigators found effect estimates that were similar, though generally
higher, than those reported in previous ACS analyses.
A follow-up to the Harvard Six Cities study (Laden et al., 2006, 087605) used updated
air pollution and mortality data and found positive associations between long-term
July 2009
7-23
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
exposure to PM2.5 and mortality. Of special note is a statistically significant reduction in
mortality risk reported with reduced long-term fine particle concentrations. This reduced
mortality risk was observed for deaths due to cardiovascular and respiratory causes, but
not for lung cancer deaths.
The WHI cohort study (Miller et al., 2007, 090130) (described previously) found that
each 10 |ug/m3 increase of PM2.5 was associated with a 76% increase in the risk of death
from cardiovascular disease (hazard ratio, 1.76 [95% CP 1.25-2.47]). The WHI study not
only confirms the ACS and Six City Study associations with cardiovascular mortality in yet
another well characterized cohort with detailed individual-level information, it also has
been able to consider the individual medical records of the thousands of WHI subjects over
the period of the study. This has allowed the researchers to examine not only mortality, but
also related morbidity in the form of heart problems (cardiovascular events) experienced by
the subjects during the study. These morbidity co*associations with PM2.5 in the same
population lend even greater support to the biological plausibility of the air pollution-
mortality associations found in this study.
In an analysis for the Seventh-Day Adventist cohort in California (AHSMOG), a
positive, association with coronary heart disease mortality was reported among females (92
deaths! RR = 1.42 [95% CP 1.06-1.90] per 10 |ug/m3 PM2.5), but not among males (53 deaths!
RR = 0.90 [95% CP 0.76-1.05] per 10 |ug/m3 PM2.5) (Chen et al., 2005, 087942). Associations
were strongest in the subset of postmenopausal women (80 deaths! RR = 1.49 [95% CP 1.17-
1.89] per 10 |ug/m3 PM2.5). The authors speculated that females maybe more sensitive to air
pollution-related effects, based on differences between males and females in dosimetry and
exposure. As was found with fine particles, a positive association with coronary heart
disease mortality was reported for PM10-25 and PM10 among females (RR = 1.38 [95% CP
0.97-1.95] per 10 |ug/m3 PM10-2&RR = 1.22 [95% CP 1.01-1.47] per 10 |u,g/m3 PM10), but not
for males (RR = 0.92 [95% CP 0.66-1.29] per 10 jug/m3 PMio-2&RR = 0.94 [95% CP 0.82-1.08]
per 10 |ug/m3 PM10); associations were strongest in the subset of postmenopausal women
(80 deaths) (Chen et al., 2005, 087942).
Two additional studies explored the effects of PM10 on cardiovascular mortality. The
Nurses' Health Study (Puett et al., 2008, 156891) is an ongoing prospective cohort study
examining the relation of chronic PM10 exposures with all-cause mortality and incident and
fatal coronary heart disease consisting of 66,250 female nurses in MSAs in the northeastern
region of the U.S. The association with fatal CHD occurred with the greatest magnitude
when compared with other specified causes of death (hazard ratio 1.42 [95% CP 1.11-1.81]).
The North Rhine-Westphalia State Environment Agency (LUANRW) initiated a cohort of
approximately 4,800 women, and assessed whether long-term exposure to air pollution
originating from motorized traffic and industrial sources was associated with total and
cause-specific mortality (Gehring et al., 2006, 089797). They found that cardiopulmonary
mortality was associated with PM10 (RR = 1.52 [95% CP 1.09-2.15] per 10 |u,g/m3 PM10).
July 2009
7-24
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
In summary, the 2004 PM AQCD concluded that there was strong evidence that long-
term exposure to PM2.5 was associated with increased cardiopulmonary mortality Recent
studies investigating cardiovascular mortality provide some of the strongest evidence for a
cardiovascular effect of PM. A number of large cohort studies have been conducted
throughout the U.S. and reported consistent increases in cardiovascular mortality related
to PM2.5 concentrations. The results of two of these studies have been replicated in
independent reanalyses. These effects are biologically plausible and coherent with
epidemiologic and toxicological studies of short-term exposure and CVD morbidity and
mortality and long-term exposure to PM2.5 and CVD morbidity.
7.2.11. Summary and Causal Determinations
7.2.11.1. PM2.5
Epidemiologic studies examining associations between long-term exposure to ambient
PM (over months to years) and CVD morbidity had not been conducted and thus were not
included in the 1996 or 2004 PM AQCDs (U.S. EPA, 1996, 079380; U.S. EPA, 2006, 157071).
A number of studies were included in the 2004 AQCD that evaluated the effect of long-term
PM2.5 exposure on cardiovascular mortality and found strong and consistent associations.
No toxicological studies had evaluated the effects of subchronic or chronic PM exposure on
CVD outcomes in the 2004 PM AQCD. Recently, epidemiologic and toxicological studies
have provided evidence of the adverse effects of long-term exposure to PM2.5 on
cardiovascular outcomes, including atherosclerosis and clinical and subclinical markers of
cardiovascular morbidity.
The strongest evidence for a CVD health effect related to long-term PM2.5 exposure
comes from epidemiologic studies of cardiovascular mortality. A number of large, multicity
U.S. studies (the ACS, Six Cities Study, WHI, and AHSMOG) provide consistent evidence of
an effect between long-term exposure to PM2.5 and cardiovascular mortality (Section 7.2.10).
These studies were conducted in urban areas across the U.S. where mean concentrations
ranged from 10.2-29.0 |ug/m3 (Table 7-8). An epidemiologic study investigating the
relationship between PM2.5 and clinical CVD morbidity among post-menopausal women
(Miller et al., 2007, 090130) provides evidence of an effect that is coherent with the
cardiovascular mortality studies. This large, prospective cohort study of incident, validated
cases found large increases in the adjusted risk of MI, revascularization, and stroke using a
1-yr avg PM2.5 concentration (mean = 13.5 pg/m3). Across-sectional analyses of self-reported
prevalence of CHD and peripheral arterial disease found no such increase in the odds of
CVD morbidity (Hoffmann et al., 2006. 091162); the inconsistency of these findings with
Miller et al. (2007, 090130) may be explained by differences in study design or location.
July 2009
7-25
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
The effect of long-term PM2.5 exposure on pre-clinical measures of atherosclerosis
(CIMT, CAC, AAC or ABI) has been studied in several populations using a cross-sectional
study design. The magnitude of the PM2.5 effects and their consistency across different
measures of atherosclerosis in these studies varies widely, and they may be limited in their
ability to discern small changes in these measures. Kunzli et al. (2005, 087387) observed a
non-significant 4.2% increase in CIMT associated with long-term PM2.5 exposure among
participants of a clinical trial in greater Los Angeles, which was several fold higher than
the 0.5% increase observed by Diez-Roux et al. (2008, 156401) in their analyses of MESA
baseline data. The associations in MESA of CAC and ABI with long-term PM2.5 exposure
were largely null (Diez Roux et al., 2008, 156401), while an increase in AAC with long-term
PM2.5 exposure was reported (Chang et al., 2008, 180393). By contrast, a 43% increase in
CAC was associated with long-term PM2.5 exposure in a German study but no similar
association with ABI was observed (Hoffmann et al., 2009, 190376). Although the number of
studies examining these relationships is limited, effect modification by use of
hyperlipidemics and smoking status was reported in more than one study of long-term PM
exposure.
Evidence of enhanced atherosclerosis development was demonstrated in new
toxicological studies that demonstrate increased plaque and lesion areas, lipid deposition,
and TF in aortas of ApoE '" mice exposed to CAPs (Section 7.2.1.2). In addition, alterations
in vasoreactivity were observed, suggesting an impaired NO pathway. Additional
toxicological studies of PM10 are consistent with these results. Further support is provided
by a study that reported decreased LAV ratio in the pulmonary and coronary arteries of
mice exposed to ambient air. However, PM2.5 CAPs derived from traffic in Los Angeles did
not affect plaque size (Araujo et al., 2008, 156222). Collectively, these toxicological studies
provide biological plausibility for the small effects observed in epidemiologic studies.
There is limited evidence for the effects of PM2.5 on renal or vascular function. Cross-
sectional and longitudinal epidemiologic analyses of PM2.5 and UACR revealed no evidence
of an effect (O'Neill et al., 2007, 156006) while small non-statistically significant increases
in BP with 30- and 60-day avg PM2.5 concentrations were reported (Auchincloss et al., 2008,
156234). A toxicological study did not show changes in MAP with CAPs, but indicated a
CAPs-related potentiation of experimentally-induced hypertension (Sun et al., 2008,
157032). In addition, CAPs has induced changes in insulin resistance, visceral adiposity,
and inflammation in a diet-induced obesity mouse model (Sun et al., 2009, 190487),
indicating that diabetics may be a potentially susceptible population to PM exposure.
In summary, a number of large U.S. cohort studies report associations of long-term
PM2.5 concentration with cardiovascular mortality. These studies provide the strongest
evidence for an effect of long-term PM2.5 exposure on CVD effects. Additional evidence
comes from a methodologically rigorous epidemiology study that demonstrates coherent
associations between long-term PM2.5 exposure and CVD morbidity among post-menopausal
July 2009
7-26
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
women. Toxicological studies demonstrate that this effect is biologically plausible and the
effect is coherent with studies of short-term PM2.5 exposure and CVD morbidity and
mortality and with long-term exposure to PM2.5 and CVD mortality. Associations between
PM2.5 and subclinical measures of atherosclerosis are inconsistent, but cross-sectional
studies may be limited in their ability to discern small changes in these measures. In
addition, potential modification of the PM2.5-CVD association by smoking status and the use
of anti-hyperlipidemics has been demonstrated in epidemiologic studies that used
individual-level data. Toxicological studies provide evidence for accelerated development of
atherosclerosis in ApoE"'" mice exposed to CAPs and show effects on coagulation factors,
experimentally-induced hypertension, and vascular reactivity. Available studies of clinical
cardiovascular disease outcomes report inconsistent results. Based on the above findings,
the epidemiologic and toxicological evidence is sufficient to infer a causal relationship
between long-term PM2.5 exposures and cardiovascular effects.
7.2.11.2.	PM10-2.5
One epidemiologic study evaluated the relationship between long-term exposure to
PM10-2.5 and cardiovascular mortality and found a positive association with coronary heart
disease mortality among females, but not for males! associations were strongest in the
subset of postmenopausal women (Chen et al., 2005, 087942). No toxicological studies of
long-term exposure to ambient PM10-2.5 and cardiovascular effects have been conducted to
date. Evidence is inadequate to infer the presence or absence of a causal relationship.
7.2.11.3.	Ultrafine PM
A few toxicological studies of long-term exposure to ultrafine PM have been
conducted. Increased plaque size was reported in mice exposed to ultrafine CAPs derived
from traffic (Araujo et al., 2008, 156222). Studies of diesel and gasoline exhaust reported
relatively few changes in hematologic or coagulation parameters (Section 7.2.4.2) and one
DE study demonstrated altered cardiac gene expression in normotensive rats that reflected
the development of hypertension (Gottipolu et al., 2009, 190360). Whole and filtered
gasoline exhaust induced increases in gene products involved in atheromatous plaque
formation and/or degradation, but these effects were largely due to the gaseous emissions
(Lund et al., 2007, 125741). Evidence from these studies alone is inadequate to infer the
presence or absence of a causal relationship due to a few studies being conducted without
gaseous co-pollutants.
July 2009
7-27
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
7.3. Respiratory Effects
Several cohort studies reviewed in the 2004 PM AQCD provided evidence for
relationships between long-term PM exposure and effects on the respiratory system, though
it did not rule out the possibility that the observed respiratory effects may have been
confounded by other pollutants. In 12 southern California communities in the Children's
Health Study (CHS), Gauderman et al. (2000, 012531; 2002, 026013) found that decreases
in lung function growth among schoolchildren were associated with long-term exposure to
PM. Declines in pulmonary function were reported with all three major PM size classes -
PMio, PMio-2.5 and PM2.5 - though the three PM measures were highly correlated. In
another analysis of data from the CHS cohort, McConnell et al. (1999, 007028). reported an
increased risk of bronchitis symptoms in children living in communities with higher PM10
and PM2.5 concentrations. These results were found to be consistent with results of cross-
sectional analyses of the 24-cities study by Dockery et al. (1996, 077269) and Raizenne et
al. (1996, 077268), that were assessed in the 1996 PM AQCD. These studies reported
associations between increased bronchitis rates and decreased peak flow with fine particle
sulfate and fine particle acidity. However, the high correlation of PM10, acid vapor and NO2
precluded clear attribution of the bronchitis effects reported by McConnell et al. (1999,
007028) to PM alone. In a prospective cohort study among a subset of children in the CHS
(n = 110) who moved to other locations during the study period, Avol et al. (2001, 020552)
reported that those subjects who moved to areas of lower PM10 showed increased growth in
lung function compared with subjects who moved to communities with higher PM10
concentrations. Finally, the 2004 PM AQCD concluded that there was strong epidemiologic
evidence for associations between long-term exposures to PM2.5 and cardiopulmonary
mortality, though the respiratory effects were not separated from the cardiovascular effects
in this conclusion.
The 2004 PM AQCD (U.S. EPA, 2004, 056905) concluded that the evidence for an
association between long-term exposure to PM and respiratory effects may be confounded
by other pollutants. Gauderman et al. (2002, 026013) reported declines for FEVi and
McConnell et al. (1999, 007028) reported increased ORs for bronchitic symptoms in
asthmatics for PM10 and PM2.5. Recent epidemiologic literature includes results from
several prospective cohort studies, which found consistent, positive associations between
long-term exposure to PM and respiratory morbidity. Associations were reported with PM2.5
and PM10, and the studies showing associations only with PM10 were conducted in locations
where the PM was predominantly fine particles, providing support for associations with
long-term exposure to fine particles. These results are summaraized below! further details
of these studies are summarized in Annex E.
Very few subchronic and chronic toxicological studies investigating respiratory effects
were available in the 2004 PM AQCD. However, the 2002 EPA Health Assessment
July 2009
7-28
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Document for DE reported that chronic exposure to DE was associated with histopathology
including alveolar histiocytosis, aggregation of alveolar macrophages, tissue inflammation,
increased polymorphonuclear leukocytes, hyperplasia of bronchiolar and Type 2 epithelial
cells, thickened alveolar septa, edema, fibrosis, emphysema and lesions of the trachea and
bronchi. Since then a number of animal toxicological studies have been conducted involving
inhalation exposure to CAPs, urban air, DE, gasoline exhaust, and woodsmoke. These
subchronic and chronic studies provide evidence of altered pulmonary function,
inflammation, histopathological changes and oxidative and allergic responses following
PM2.5 exposures. These results are summarized below! further details of these studies are
summarized in Annex D.
7.3.1. Respiratory Symptoms and Disease Incidence
7.3.1.1. Epidemiologic Studies
New longitudinal cohort studies provide the best evidence to evaluate the relationship
between long-term exposure to ambient PM and increased incidence of respiratory
symptoms or disease. A summary of the mean PM concentrations reported for the long-term
exposure studies characterized in this section is presented in Table 7-3.
Bayer-Oglesby et al. (2005, 086245) examined the decline of ambient pollution levels
and improved respiratory health demonstrated by a reduction in respiratory symptoms and
diseases in school children (n = 9,591) in Switzerland. Reduced air pollution exposure
resulted in improved respiratory health of children. Further, the average reduction of
symptom prevalence was more pronounced in areas with stronger reduction of air pollution
levels. The average decline of PM10 between 1993 and 2000 across the nine study regions
was 9.8 |ug/m3 (29%). Declining levels of PM10 were associated with declining prevalence of
chronic cough, bronchitis, common cold, nocturnal dry cough, and conjunctivitis symptoms,
but no significant associations were reported for wheezing, sneezing, asthma, and hay fever,
as shown in Figure 7*2. In Figure 7-2, Panel (B) illustrates that on an aggregate level
across region, the mean change in adjusted prevalence of chronic cough is associated with
the mean change in PM10 levels (r = 0.78; p = 0.02). Similar associations were seen for
nocturnal dry cough and conjunctivitis symptoms and PM10 levels. Roosli et al. (2000,
010296; 2001, 108738; 2005, 156923) have demonstrated that PM10 levels are
homogeneously distributed within regions of Basel, Switzerland and are not significantly
affected by local traffic, justifying the single-monitor approach for assignment of PM10
exposures. Based on parallel measurements of PM2.5 and PM10 at seven sites in
Switzerland, PM2.5 and PM10 at all sites are generally highly correlated (r2 ranging from
0.85 to 0.98 for the seven cities) (Gehrig and Buchmann, 2003, 139678), indicating that
PM10 is predominantly fine particles in these locations.
July 2009
7-29
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Schindler et al. (2009, 191950) reported that sustained reduction in ambient PMio
concentrations can lead to decreases in respiratory symptoms among Swiss adults in the
SAPALDLIA study. They compared baseline data in 1991 to a follow-up interview in 2002
after a substantial decline in PMio concentrations served as a natural experiment. Each
subject was assigned model-based estimates of PMio concentrations averaged over the 12
months preceding each health assessment with mean decline in PMio levels of 6.2 ug/m3
(SD = 3.9 |ug/m3). When the authors tested the joint hypothesis of no association between
the PMio difference and symptom incidence or persistence, positive results were obtained
for regular cough, chronic cough or phlegm and wheezing but not regular phlegm or
wheezing without a cold.
Pierse et al. (2006, 088757) studied the association between primary PMio (particles
directly emitted from local sources/traffic) and the prevalence and incidence of respiratory
symptoms in a randomly sampled cohort of 4,400 children (aged 1-5 years) in
Leicestershire, England surveyed in 1998 and again in 2001. Annual exposure to primary
PMio was calculated for the home address using the Airviro statistical dispersion model.
After adjusting for confounders, mean annual exposure to locally generated PMio was
associated with an increased prevalence of cough without a cold in both the 1998 (OR 1.21
[95% CI: 1.07-1.38], n = 2164) and 2001 surveys (OR 1.56 [95% CI: 1.32-1.84], n = 1756).
Nordling et al. (2008, 097998) examined the relationship between estimated PM
exposure levels and respiratory health effects in a Swedish birth cohort (n = 4089) of
preschool children. The spatial distributions of PM from traffic in the study area were
estimated with emission databases and statistical dispersion modeling. Children were
examined at 2 months and 1, 2, and 4 years of age. Using GIS methods, the average
contribution of traffic-generated PMio above regional background to the children's
residential outdoor air pollution levels was determined. To evaluate the exposure
assessment, the authors compared the estimated levels of traffic-generated PMio with PM2.5
measurements from 42 locations (Hoek et al., 2002, 042364) and reported modeled traffic-
generated PMio correlated reasonably well with measured PM2.5 (r = 0.61). Persistent
wheezing (cumulative incidence up to age 4) was associated with exposure to traffic-
generated PMio (OR 2.28 [95% CP 0.84-6.24] per 10 pg/m3 increase) while transient and
late onset wheezing was not associated. This study demonstrates that respiratory effects
may be present in preschool children.
July 2009
7-30
DRAFT-DO NOT CITE OR QUOTE

-------
Table 7-3. Characterization of ambient PM concentrations from studies of respiratory
symptoms/disease and long-term exposures.
Reference
Location
Mean Annual Concentration (jjg/m3)
Upper Percentile
Concentrations
(|jg/m3>
PM2.5
Annesi Maesano et al. (2007, 093180)
6 French Cities
Range of means across sites: 8.7-23.0
Avg of means across sites: 15.5

Braueret al. (2007, 090691)
The Netherlands
16.9
75th: 18.1
90th: 19.0
Max: 25.2
Goss et al. (2004, 055624)
U.S.
13.7
75th: 15.9
Islam et al. (2007, 090697)
12 CHS/CA communities

Max: 29.5
Janssen et al. (2003,133555)
The Netherlands
20.5
75th: 22.1
Max: 24.4
Kim et al. (2004,087383)
San Francisco, CA
Range of means across sites: 11-15
Avg of means across sites: 12

McConnell et al. (2003, 049490)
12 CHS/CA communities
13.8
Max: 28.5
Morgenstern et al. (2008,156782)
Munich, Germany
11.1

PM,o
BaverOglesbv et al. (2005, 086245)
Nine study regions in Switzerland

Max: 46
Kunzli et al. (2009,191949)
Switzerland
21.5

Nord ling et al. (2008, 097998)
Sweden
4*

Schin dler et al. (2009,191950)
Switzerland


McConnell et al. (2003, 049490)
12 CHS/CA communities
30.8
Max: 63.5
Pierse et al. (2006, 088757)
Leicestershire,U.K.
1.33
75th: 1.84
"Source specific; PM10 from traffics
July 2009
7-31
DRAFT-DO NOT CITE OR QUOTE

-------
0
t
1
f")
f
I
5
1.5
1.4 —|
1.3
1.2 H
1.1
1.0
0.9-
o.e
0.?
0 6 -
0.5 -
0.4 -
0.3-
0.2 —
0.1 -
0.0
_cF

*1
- / /
* * S J
{
0	g x
Is 1 S
¦c a w
m if?
1	1 o
ro O
Be.
-20 15 io ¦ r,
Mean change of annual
average PM10 (pg/m3)
T
0#	~	~
/ / ^
«/
T~


#•
~
Source: Bayer Oglesby et al. (2005, 086245)
Figure 7-2. Adjusted ORs and 95% CIs of symptoms and respiratory diseases associated with a
decline of 10//g/m3 PM10 levels in Swiss Surveillance Program of Childhood Allergy
and Respiratory Symptoms1. Inset: Mean change in adjusted prevalence (1998-2001 to
1992-1993) versus mean change in regional annual averages of PM10 (1997-2000 to
1993) for chronic cough, across nine SCARPOL regions (An: Anieres. Be: Bern. Bi: Biel.
Ge: Geneva. La:, Langnau. Lu: Lugano. Mo: Montana. Pa: Payerne. Zh:
Zurich).
1	McConnell et al. (2003, 049490) conducted a prospective study examining the
2	association between air pollution and bronchitic symptoms in 475 school children with
3	asthma in 12 Southern California communities as part of the CHS from 1996 to 1999. They
4	investigated both the differences between- communities with 4-yr avg and within-
5	communities yearly variation in PM (i.e., PMio, PM2.5, PM10-2.5, EC, and OC). Based on a
6	10 pg/m3 change in PM2.5, within-communities effects were larger (OR 1.90 [95% CP 1.10-
7	2.70]) than those for between-communities (OR 1.30 [95% CP 1.10-1.50]). The OR for the
8	10 pg/m3 range in 4-yr avg PM2.5 concentrations across the 12 communities was 1.29 (95%
L Adjusted for age, sex, nationality, parental education, number of siblings! farming status, low birth weight, breastfeeding,
child who smokes, family history of asthma, bronchitis, and/or atopy, mother who smokes, indoor humidity, mode of heating
and cooking, carpeting, pets allowed in bedroom, removal of carpet and/or pets for health reasons, person who completed
questionnaire, month when questionnaire was completed, number of days with the maximum temperature <0°C, and belief
of mother that there is an association between environmental exposures and children's respiratory health
July 2009
7-32
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
CL 1.06-1.58). Similar results were reported for PMio and PM10-2.5 but the effect estimates
were smaller in magnitude and generally not statistically significant. Within-community
associations were not confounded by any time-fixed personal covariates. In two-pollutant
models, the within-community effect estimates for PM2.5 and OC were significant in the
presence of several other pollutants. While the within-community single-pollutant effect of
PM2.5 (13 = 0.085/pg/m3) was only modestly attenuated after adjusting for some pollutants, it
was markedly reduced after adjusting for NO2 or OC. The between-community effect
estimates generally were not significant in the presence of other pollutants in two-pollutant
models.
In the CHS already discussed, Islam et al. (2007, 090697) examined the hypothesis
that ambient air pollution attenuates the reduced risk for childhood asthma that is
associated with higher lung function (n = 2057). At each age a distribution of pulmonary
functions exists. Haland et al. (2006, 156511) found evidence that children with high lung
function have a reduced risk for asthma. Islam et al. (2007, 090697) used the CHS data to
study how the association of asthma incidence with lung function is modified by long-term
PM exposure. The incidence rate (IR) of newly diagnosed asthma increased from 9.5/1000
person-years for children with percent-predicted FEF25-75 values > 120% to 20.4/1000
person-years for children with FEF25-75 value < 100%. Over the 10th"90th percentile range
for FEF25-75 (57.1%), the hazard ratio of new onset asthma was 0.50 (95% CP 0.35-0.71). The
IR of asthma for FEF25-75> 120% in the "high" PM2.5 (13.7-29.5 jug/m3) communities was
15.9/1000 person-years compared to 6.4/1000 person-years in "low" PM2.5 (5.7-8.5 |ug/m3)
communities. Loss of protection by high lung function against new onset asthma in the
"high" PM2.5 communities was observed for all the lung function measures. Figure 7.3
shows the effect of PM2.5 on the association of lung function with asthma. Of all the
pollutants examined (NO2, PM10, PM2.5, acid vapor, O3, EC, and OC), PM2.5 appeared to
have the strongest modifying effect on the association between lung function with asthma
as it had the highest R2 value (0.42). Over the 10th-90th percentile range of FEF25-75, the
hazard ratio of new onset asthma was 0.34 (95% CP 0.21-0.56) in a community with low
PM2.5 (<13.7 pg/m3) and 0.76 (95% CP 0.45-1.26) in a community with high PM2.5
(> 13.7 pg/m3). The data do not indicate that PM exposure increased rates of incident
asthma among children with poor lung function at study entry because rates among those
with poor lung function were similar in both low and high pollution communities.
July 2009
7-33
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
1.00
High PM2 5 - Communities
1.50-
UP
° 1.20-
Low PM25
Communities
ML
SD
0.90-
(8 	~
N
X 0.60 —J
¦o
w
LE
RV
LM
AT
BM
LB
r2 = 0.42
P = 0.01
AL
0.30 H
5	10	15	20	25	30
PM2,g(|jg/m3)
Source: Islam et al. (2007, 090697)
Figure 7-3. Effect of PM2.5 on the association of lung function with asthma. Community-specific
hazard ratio of newly diagnosed asthma over 10-90th percentile range (57.1%) of FEF25
75% by level of ambient PM2.5 ((Jg/m3). The 12 CHS communities are shown.
In a prospective birth cohort study (n = 4,000) in The Netherlands, Brauer et al.
(2007, 090691) assessed the development of asthma, allergic symptoms, and respiratory
infection during the first four years of life in relation to long-term PM2.5 concentration at
the home address with a validated model using GIS. PM2.5 was associated with doctor-
diagnosed asthma (OR = 1.32 [95% CP 1.04-1.69]) for a cumulative lifetime indicator. These
findings extend observations made at 2 years of age in the same cohort (Brauer et al., 2002,
035192) providing greater confidence in the association. No associations were observed for
bronchitis.
Kunzli et al. (2009, 191949) used the SAPALDIA cohort study discussed previously in
this section to evaluate the relationship between the 11-year change (1991-2002) in traffic-
related PM10 and asthma incidence-adult onset asthma. In a cohort of 2,725 never-smokers
without asthma at baseline (ages: 18-60 years in 1991), subjects reporting doctor-diagnosed
asthma at follow-up were considered incident cases. Modeled traffic-related PM10 levels
were used. Cox proportional hazard models for time to asthma onset were used with
adjustments for cofounders. The study findings suggest that PM contributes to asthma
development and that reductions in PM decrease asthma risk. A strong feature of
SAPALDIA is the ability to assign space, time, and source-specific pollution to each subject.
Further, Kunzli et al. (2008, 129258) discusses the impact of attributable health risk models
for exposures that are assumed to cause both chronic disease and its exacerbations. The
added impact of causing disease increases the risk compared to only exacerbations.
A matched case-control study of infant bronchiolitis (ICD 9 code 466.1) hospitalization
and two measures of long-term exposure - the month prior to hospitalization (subchronic)
and the lifetime average (chronic) - to PM2.5 and gaseous air pollutants in the South Coast
July 2009
7-34
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Air Basin of southern California was conducted by Karr et al. (2007, 090719) among 18,595
infants born between 1995-2000. For each case, 10 controls matched on date were randomly
selected from birth records. Exposure was based on PM2.5 measurements collected every
third day. The mean distance between the subjects' residential ZIP code and the assigned
monitor was generally 4-6 miles with a maximum distance of 30 miles. For 10 pg/m3
increases in both sub-chronic and chronic PM2.5 exposure, an adjusted OR of 1.09 (95% CP
I.04-1.14)	was observed. In multipollutant model analyses, the association with PM2.5 was
robust to the inclusion of gaseous pollutants. Also, in a cohort of children in Germany,
Morgenstern et al. (2008, 156782) modeled PM2.5 data at birth addresses found statistically
significant effects for asthmatic bronchitis, hay fever, and allergic senstation to pollen.
Goss et al. (2004, 055624) conducted a national study examining the relationship
between air pollutants and health effects in a cohort of cystic fibrosis (CF) patients (n =
II,484)	over the age of 6 years (mean age = 18.4, SD = 10) enrolled in the Cystic Fibrosis
Foundation National Patient Registry in 1999 and 2000. Exposure was assessed by linking
air pollution values from the closest population monitor from the Air Quality System (AQS)
with the centroid of the patient's home ZIP code that was within 30 miles. The mean
distance from the patient's ZIP code to monitors for PM2.5 and PM10 was 10.8 miles (SD 7.8)
and 11.5 miles (SD 7.9), respectively. PM2.5 and PM10 24-h avg were collected every 1 to 12
days. CF diagnosis involves genetic screening panels and a common severe mutation used is
the loss of phenylalanine at the 508th position. Genotyping was available in 74% of the
population and of those genotyped, 66% carried one or more delta F508 deletions. After
adjusting for confounders, a 10 pg/m3 increase in PM2.5 or PMiowas associated with a 21%
(95% CP 7-33) or 8% (95% CP 2-15), respectively, increase in the odds of two or more
exacerbations defined as a CF-related pulmonary condition requiring admission to the
hospital or use of home intravenous antibiotics. The estimate for the associations between
pulmonary exacerbations and PM2.5 and PM10 were attenuated when the models were
adjusted for lung function. Brown et al. (2001, 012307) found that particle deposition was
increased in CF and that particle distribution in the lungs was enhanced in poorly
ventilated tracheobronchial regions in CF patients. Such focal deposition may partially
explain the association of PM and CF exacerbation.
Annesi-Maesano (2007, 093180) from 5,338 school children (10.4 ±0.7 years)
attending 108 randomly chosen schools in 6 French cities to the concentration of PM2.5
monitored in school yards. Atopic asthma was related to PM2.5 (OR 1.43 [95% CP 1.07-1.91])
when high PM2.5 concentrations (20.7 pg/m3) were compared to low PM2.5 concentrations
(8.7 pg/m3). The report is consistent with the results in an earlier paper (Penard'Morand et
al., 2005, 087951) in the same sample of children that related the findings to PM10.
Kim et al. (2004, 087383) conducted a school-based cross-sectional study in the San
Francisco metropolitan area in 2001 comprised of 10 neighborhoods to examine the
relationship between traffic-related pollutants and current bronchitic symptoms and
July 2009
7-35
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
asthma obtained by parental questionnaire (n = 1109). They related traffic-related
pollutants (PM) and bronchitic and asthma symptoms in the past 12 months. No
multipollutant models were evaluated because of the high interpollutant correlations. PM2.5
levels ranged across the school sites from 11 to 15 ug/m3.
Schikowski et al. (2005, 088637) examined the relationship between both long-term
air pollution exposure and living close to busy roads and COPD in the Rhine-Ruhr Basin of
Germany from 1985 to 1994 using consecutive cross-sectional studies. Seven monitoring
stations that were <8 km to a woman's home address provided TSP data that PM10 was
estimated from using a conversion factor (obtained from parallel measurement of TSP and
PM10 conducted at 7 sites in the Ruhr area). Distance to a major road was determined using
GIS. The results of the study suggest that long-term exposure to air pollution from PM10
and living near a major road might increase the risk of developing COPD and can have a
detrimental effect on lung function. All ORs for 5-yr exposures were stronger than those for
1-yr exposures.
In summary, the 2004 PM AQCD evaluated the available studies which primarily
related effects to bronchitic symptoms in school-age children. The new evidence is more
informative, as it includes longitudinal cohort studies that have been conducted in different
countries by different researchers that use different study designs. Some new studies are
using several different methods to include individual estimates of exposure to ambient PM
that may reduce the impact of exposure error. The strength and consistency of the outcomes
is enhanced by results being reported by several different researchers in different countries
using different designs. Most recent studies have focused on children, but a few studies
have also reported associations in adults.
The CHS (McConnell et al., 2003, 049490) provides evidence in a prospective
longitudinal cohort study that relates PM2.5 and bronchitic symptoms and reports larger
associations for within-community effects that are less subject to confounding than
between-community effects. Several new studies report similar findings with long-term
exposure to PM10 in areas where fine particles predominate PM10. In England, in a cohort of
4,400 children (aged 1-5), an association is seen with an increased prevalence of cough
without a cold. Further evidence includes a reduction of respiratory symptoms
corresponding to decreasing PM levels in "natural experiments" in both a cohort of Swiss
school children (Bayer-Oglesby et al., 2005, 086245) and adults (Schindler et al., 2009,
191950).
In a separate analysis of the CHS, Islam et al. (2007, 090697) showed that PM2.5 had
the strongest modifying effect on the association between lung function with asthma such
that loss of protection by high lung function against new onset asthma in high PM2.5
communities was observed for all the lung function measures from 10 to 18 years of age.
This relates new onset asthma to long-term PM exposure. In the Netherlands, Brauer et al.
(2007, 090691) augments the literature with data examining the first 4 years of life in a
July 2009
7-36
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
birth cohort showing an association with doctor-diagnosed asthma. Further, in an adult
cohort in the SALPALDIA study, Kunzli et al. (2009, 191949) relate PM to asthma
incidence.
7.3.2. Pulmonary Function
Several cohort studies reviewed in the 2004 PM AQCD provided evidence for
relationships between long-term PM exposure and effects on the respiratory system. In 12
southern California communities in the Children's Health Study (CHS), Gauderman et al.
(2000, 012531; 2002, 026013) found that decreases in lung function growth among school
children were associated with long-term exposure to PM. Declines in pulmonary function
were reported with all three major PM size classes - PMio, PM10-2.5 and PM2.5 - though the
three PM measures were highly correlated. These results were found to be consistent with
results of cross-sectional analyses of Raizenne et al. (1996, 077268), that was assessed in
the 1996 PM AQCD. That study reported associations between decreased peak flow with
fine particle sulfate and fine particle acidity. Finally, in a prospective cohort study among a
subset of children in the CHS (n = 110) who moved to other locations during the study
period, Avol et al. (2001, 020552) reported that those subjects who moved to areas of lower
PM10 showed increased growth in lung function compared with subjects who moved to
communities with higher PM10 concentrations who showed decrease growth in lung
function.
7.3.2.1. Epidemiologic Studies
New longitudinal cohort studies have evaluated the relationship between long-term
exposure to PM and changes in measures of pulmonary function (FVC, FEVi, and measures
of expiratory flow). Cross-sectional studies also offer supportive information (see Annex E)
and may provide insights derived from within community analysis. Lung function increases
continue through early adulthood with growth and development, then declines with aging
(Stanojevic et al., 2008, 157007; Thurlbeck, 1982, 093260; Zeman and Bennett, 2006,
157178). A summary of the mean PM concentrations reported for the long-term exposure
studies characterized in this section is presented in Table 7-4.
Table 7-4. Characterization of ambient PM concentrations from studies of FEVi and long-term
exposures.


Reference
Location
Mean Annual Upper Percentile
Concentration (ng/m3) Concentrations (ng/m3)
PM2.5
Gauderman et al. (2002, 026013)
12 CHS/CA communities
5-30
Gauderman et al. (2004, 056569)
12 CHS/CA communities
6-27
Goss et al. (2004, 055624)
U.S.
13.7 75th: 15.9
July 2009
7-37
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Reference
Location
Mean Annual
Concentration (ng/m3)
Upper Percentile
Concentrations (ng/m3)
Gotschi et al. (2008,180364)
21 European cities
Range of means across sites: 3.7-44.7
Avg of mean across sites: 16.8

PMm
Downs et al. (2007, 092853)
8 cities in Switzerland
Range of means across sites: 9-46
Avg of mean across sites: 21.6

Gauderman et al. (2002, 026013)
12 CHS/CA communities
Range of means across sites: 13-78
Avg of mean across sites: NR

Gauderman et al. (2004, 056569)
12 CHS/CA communities
Range of means across sites: 18-68
Avg of mean across sites: NR

Nordling et al. (2008, 097998)
Sweden


A vol et al. (2001,020552)
Southern CA/CHS


Rojas-Martinez (2007,188508)
Mexico City, Mexico
75.6
75th: 92.2
90th: 112.7
The CHS prospectively examined the relationship between air pollutants and lung
function (FVC, FEVi, MMEF) in a cohort (n = 1,759) of children between the ages of 10 and
18 years, a period of rapid lung development (Gauderman et al., 2004, 056569). Air
pollution monitoring stations provided data in each of the 12 study communities from 1994-
2000. The results for O3, PM10, NO2, PM2.5, acid vapor, and EC are depicted in Figure 7-4. In
general, two-pollutant models for any pair of pollutants did not provide a substantially
better fit to the data than the corresponding single-pollutant models due to the strong
correlation between most pollutants. The pollution-related deficits in the average growth in
lung function over the eight-yr period resulted in clinically important deficits in attained
lung function at the age of 18 years.
Downs et al. (2007, 092853) prospectively examined 9,651 randomly selected adults
(18-60 years of age) in 8 cities in Switzerland (see alsoAckermann-Liebrich et al., 1997,
077537) to ascertain the relationship between reduced exposure to PM10 and age-related
decline in lung function (FVC, FEVi, and FEF25-50). An evaluated statistical dispersion
model (Liu et al., 2007, 093093) provided spatially resolved concentrations of PM10 that
enabled assignment to residential addresses for the participant examinations in 1991 and
2002 that yielded a median decline of 5.3 |ug/m3 (IQR 4.1-7.5). Decreasing PM10
concentrations attenuated the decline in lung function. Effects were greater in tests
reflecting small airway function. No other pollutant relationships were evaluated, though a
related study indicated that levels of NO2 also declined over the same period (Ackermann-
Liebrich et al., 2005, 087826). Generalized cross-validation essentially chose a linear fit for
the the concentration-response curve for age-related decline in lung function.
These data show that improvement in air quality may slow the annual rate of decline
in lung function in adulthood indicating positive consequences for public health. Further
evidence on improvement in respiratory health with reduction in air pollution levels is
provided from studies conducted in East Germany related to dramatic emissions reductions
after the reunification in 1990 (Fryer and Collins, 2003, 156454; Heinrich et al., 2002,
July 2009
7-38
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
034825; Sugiri et al., 2006, 088760). This type of "natural experiment" provides additional
support for epidemiologic findings that relatively low levels of airborne particles have
respiratory effects.
In a prospective cohort study consisting of school-age children (n = 3,170) who were 8
years of age at the beginning of the study, had not been diagnosed with asthma, and were
located in Mexico City (Rojas-Martinez et al., 2007, 188508), evaluated the association
between long-term exposure to PMio, O3 and NO2 and lung function growth every 6 months
from April 1996 through May 1999. Exposure data were provided by 10 air quality monitor
stations located within 2 km of each child's school. The multipollutant model effect of PM10
over the age of 8 to 10 years of life in this cohort on FVC, FEVi, and FEF25-75 showed an
association. Single pollutant models showed an association between ambient pollutants (O3,
PM10 and NO2) and deficits in lung function growth. The association between PM10 and
FEF25-75 was not statistically significant. While the estimates from two-pollutant models
were not substantially different than single pollutant models, independent effects for
pollutants could not be estimated accurately because the traffic-related pollutants were
correlated. Although no PM2.5 data were presented in this study, Chow et al. (2002) in a
separate study report that for Mexico City during the winter of 1997 that approximately
50% of PM10 was in the PM2.5 fraction.
Gotschi et al. (2008, 156485) examined the relationship between air pollution and
lung function in adults in the European Community Respiratory Health Survey (ECRHS).
FEVi and FVC were assessed at baseline and after 9 years of follow-up from 21 European
centers (followed-up sample n = 5,610). No significant associations were found between city-
specific annual mean PM2.5 and average lung function levels which is in contrast to the
results seen by Ackermann-Liebrich (1997, 077537) (SAPALDIA) and Schikowski et al.
(2005, 088637) (SALIA) which compared across far more homogenous populations than for
the population assessed in the ECRHS. Misclassification and confounding may partially
explain the discrepancy in findings.
July 2009
7-39
DRAFT-DO NOT CITE OR QUOTE

-------
H 10-1
I
g
R-0.04
P-0,89
~ LB
~ UP
SD
~ ML
~ RV
~ AT
~ SM ~ LM
~	AL
~	LE
<( LN
~ LA
25	35	45	55	65
Oj from 10 a.m. to 6 p.m. (PPb}
75
J
OS
>
ID-
S'
.y 6-
2 <
0	4'
1	2.
oq 1
V
i
~ UP
R-0.66
P-0,02
~ 1
V
£
R-0.69
P-0,01
~ UP
~ SD
~ LN
4	6
Acid Vapor (ppb)
10
12
£. 10-
S
"5
>
R-0.74
P-0.006
~ LM
~ LN
~ UP
0.0 0.2 0.4 0.6 0.S 1.0
Elemental Carbon (f»g/m3)
1.2
1.4
Source: Adapted from Gauderman et al. (2004, 056569)
Figure 7-4. Proportion of 18-year olds with a FEVi below 80% of the predicted value plotted against
the average levels of pollutants from 1994 through 2000 in the 12 southern California
communities of the Children's Health Study (AL = Alpine; AT = Atascadero; LA = Lake
Arrowhead; LB = Long Beach; LE = Lake Elsinore; LM = Lompoc; Ll\l = Lancaster;
ML = Mira Loma; RV = Riverside; SD = San Dimas; SM = Santa Maria; UP = Uplannd.)
1	In a birth cohort (n = 2,170) in Oslo, Norway, Oftedal et al. (2008, 093202) examined
2	effects of exposure to PM2.5 and PMioon lung function (FVC, FEVi, FEFscm). Spirometry was
3	performed in 2,307 children aged 9-10 years old in 2001-2002. Residential air pollution
4	levels over the time period 1992-2002 were calculated using EPISODE dispersion models to
5	provide three time scales of exposure^ l) first year of life, 2) lifetime exposure, and 3) just
6	before the lung function test. Only single pollutant models were evaluated because air
7	pollutants were highly correlated (r = 0.83-0.95). PM exposure was associated with changes
8	in adjusted peak respiratory flow, especially in girls. No effect was found for forced volumes.
July 2009
7-40
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Adjusting for contextual socioeconomic factors diminished associations. Results for PMio
were similar to those for PM2.5.
In an exploratory study, Mortimer et al. (2008, 187280) examined the association of
prenatal and lifetime exposure to air pollutants using geocoded monthly average PM10
levels with pulmonary function in a San Joaquin Valley, California cohort of 232 children
(aged 6" 11) with asthma. First and second trimester PM10 exposures (based on monthly
average concentrations) had a negative effect on pulmonary function and may relate to
prenatal exposures affecting the lungs as they begin to develop at 6 wk gestation.
Dales et al. (2008, 156378) in a cross-sectional prevalence study examined the
relationship of pulmonary function and PM measures, other pollutants, and indicators of
motor vehicle emissions in Windsor, Ontario, in a cohort of 2,402 school children. PM2.5 and
PM10 concentrations were estimated for each child's residence at the postal code level. Each
10 pg/m3 increase in PM2.5 was associated with a 7.0% decrease in FVC expressed in a
percentage of predicted.
In Leicester, England, investigators examined the carbon content of airway
macrophages in induced sputum in 64 of 114 healthy children 8 to 15 years of age (Grigg et
al., 2008, 156499; Kulkarni et al., 2006, 089257). The carbon content of airway
macrophages (Finch et al., 2002, 054603; Strom et al., 1990, 157020) was used as a marker
of individual exposure to PM10. Near each child's home, exposure to PM10 was estimated
using a statistical dispersion model (Pierse et al., 2006, 088757). The authors reported a
dose-dependent inverse association between the carbon content of airway macrophages and
lung function in children and found no evidence that reduced lung function itself causes an
increase in carbon content. Consistent results were obtained for both FVC and FEF25-75.
Caution should be used when interpreting these results as the accuracy of the estimates on
individual PM10 exposures were not validated! there is potential for confounding by ethnic
origin! and there is concern that the magnitude of the changes in pulmonary function
associated with increased particle area appear large (Boushey et al. 2008).
Nordling et al. (2008, 097998) discussed above in the respiratory symptoms section,
also reported that lower PEF at age 4 was associated with exposure to traffic-related PM10
(-8.93 L/min [95% CP -17.78 to -0.088]). Goss et al. (2004, 055624), discussed in the
respiratory symptoms section, found strong inverse relationships between FEVi and PM2.5
concentrations in both cross-sectional and longitudinal analyses.
In summary, recent studies have greatly expanded the evidence available for the 2004
PM. The earlier CHS studies followed young children for 2-4 years. New analyses have been
conducted that include longer follow-up periods of this cohort through 18 years of age
(considered early adulthood for lung development (Stanojevic et al., 2008, 157007) and
provide evidence that effects from exposure to PM2.5 persist into early adulthood.
Longitudinal studies follow effects over time and are considered to provide the best
evidence as opposed to studies across communities as in cross-sectional studies. The
July 2009
7-41
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
longitudinal cohort studies in the 2004 PM AQCD provided data for children in one location
in one study and new longitudinal studies have been conducted in other locations.
Gauderman et al. (2004, 056569) reported that PM2.5 exposure was associated with
clinically and statistically significant deficits in FEVi attained at the age of 18 years.
Clinical significance was defined as a FEVi below 80% of the predicted value, a criterion
commonly used in clinical settings to identify persons at increased risk for adverse
respiratory conditions. This clinical aspect is an important enhancement over the earlier
results reported in the 2004 PM AQCD. Further, the association reported in this study that
evaluated the 8-yr time period into early adulthood not only provided evidence for the
persistence of the effect, but in addition the strength and robustness of the outcomes were
more positive, larger, and more certain than previous CHS studies of shorter follow-up.
Supporting this result are new longitudinal cohort studies conducted by other
researchers in other locations with different methods. Though these studies report results
for PM10, available data discussed above indicate that the majority of PM10 is composed of
PM2.5U1 these areas. New studies provide positive results from Mexico City, Sweden, and a
national cystic fibrosis cohort in the U.S. One study reported null results in a European
cohort described as having potential misclassification and confounding concerns as well as
lacking a homogenous population potentially rendering the outcome as non-informative. A
natural experiment in Switzerland, where PM levels had decreased, reported that
improvement in air quality may slow the annual rate of decline in lung function in
adulthood, indicating positive consequences for public health. These natural experiments
are considered especially supportive.
The relationship between long-term PM exposure and decreased lung function is thus
seen during lung growth and lung development in school-age children into adulthood. At
adult ages studies continue to show a relationship between decreased lung function and
long-term PM exposure. Some newer studies attempting to study the relationship of long-
term PM exposure from birth through preschool are reporting a relationship. Thus, the
impact of long-term PM exposure is seen over the time period of lung function growth and
development and the decline of lung function with aging.
Overall, effect estimates from these studies are negative (i.e., indicating decreasing
lung function) and the pattern of effects are similar between the studies for FVC and FEVi.
Thus, the data are consistent and coherent across several designs, locations, and
researchers. With cautions noted, the results relating carbon content of airway
macrophages to decreased measures of pulmonary function add plausibility to the
epidemiologic findings. Some new studies are using individual estimates of exposure to
ambient PM to reduce the impact of exposure error (Downs et al., 2007, 092853; Jerrett et
al., 2005, 087381).
As was found in the 2004 PM AQCD, the studies report associations with PM2.5 and
PM10, while most did not evaluate PM10-2.5. Associations have been reported with fine
July 2009
7-42
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
particle components, particularly EC and OC. Source apportionment methods generally
have not been used in these long-term exposure studies. However, numerous studies have
evaluated exposures to PM related to traffic or motor vehicle sources. For example, Meng et
al. (2007, 093275) investigated the associations between traffic and outdoor pollution levels
and poorly controlled asthma among adults who were respondents to the California Health
Interview Survey and found associations for traffic density and PMio, but not PM2.5.
7.3.2.2. Toxicological Studies
Urban Air
An important new study evaluated the effects of chronic exposure to ambient levels of
urban particles on lung development in the mouse (Mauad et al., 2008, 156743). Both
functional and anatomical indices of lung development were measured. Male and female
BALB/c mice were continuously exposed to ambient or filtered Sao Paolo air for 8 months.
Concentrations in the "polluted chamber" vs. "clean chamber" were 16.8 vs. 2.9 ug/ma PM2.5.
Thus PM levels were reduced by filtration but not entirely eliminated. Ambient
concentrations of CO, NO2 and SO2 were 1.7 ppm, 89.4 jug/m3 and 8.1 |u,g/m3, respectively.
Concentrations of gaseous pollutants were assumed to be similar to ambient levels in both
chambers. After 4 months, the animals were mated and the offspring were divided into 4
groups to provide for a prenatal exposure group, a postnatal exposure group, a pre and
postnatal exposure group and a control group. Animals were sacrificed at 15 and 90 days of
age for histological analysis of lungs. Pulmonary pressure-volume measurements were also
conducted in the 90 day old offspring. Statistically significant reductions in inspiratory and
expiratory volumes were found in the pre and postnatal exposure group, but not in the
prenatal or postnatal exposure group, compared with controls. These changes in pulmonary
function correlated with anatomical changes which are discussed in Section 7.3.5.1.
Diesel Exhaust
Li et al. (2007, 155929) exposed BALB/c and C56BL/6 mice to clean air or to low dose
DE (at a PM concentration of 100 |u,g/m3) for 7 hs/day and 5 days/week for 1, 4 and 8 wk.
Average gas concentrations were reported to be 3.5 ppm CO, 2.2 ppm NO2, and less than
0.01 ppm SO2. AHR was evaluated by whole body plethysmography at Day 0 and after 1, 4
and 8 wk of exposure. Short-term responses are discussed in Section 6.3.2.3, 6.3.3.3 and
6.3.4.2. The increased sensitivity of airways to methacholine (measured as Penh) seen in
C57BL/6 but not BALB/c mice at 1 week was also seen at 4 wk but not at 8 wk. This study
suggests that adaptation occurs during prolonged DE exposure. Influx of inflammatory
cells, histopathology, markers of oxidative stress and effects of antioxidant intervention
were also evaluated (see Sections 7.3.3.2 and 7.3.4.1). Although no attempt was made in
this study to determine the effects of gaseous components of DE on the measured responses,
July 2009
7-43
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
concentrations of gases were very low suggesting that PM may have been responsible for
the observed effects.
In many animal studies changes in ventilatory patterns are assessed using whole
body plethysmography, for which measurements are reported as enhanced pause (Penh).
Some investigators report increased Penh as an indicator of AHR, but these are
inconsistently correlated and many investigators consider Penh solely an indicator of
altered ventilatory timing in the absence of other measurements to confirm AHR. Therefore
use of the terms AHR or airway responsiveness has been limited to instances in which the
terminology has been similarly applied by the study investigators.
Gottipolu et al. (2009, 190360) exposed WKY and SH rats to filtered air or DE
(particulate concentration 500 and 2000 |ug/m3) for 4 h/day and 5 days/week over a 4-week
period. Concentrations of gases were 1.3 and 4.8 ppm CO, NO <2.5 and 5.9 ppm NO, <0.25
and 1.2 ppm NO2, 0.2 and 0.3 ppm SO2 for low and high PM exposures, respectively. Particle
size, measured as geometric median number and volume diameters, was 85 and 220 nm,
respectively. No DE-related effects were found for breathing parameters measured by
whole-body plethysmography. Other pulmonary effects are described in Sections 7.3.3.2 and
7.3.5.1.
Woodsmoke
One study evaluated the effects of subchronic woodsmoke exposure on pulmonary
function in Brown Norway rats. Rats were exposed 3 h/day and 5 days/week for 4 and 12 wk
to air or to 1000-10,000 ug/ma concentrated wood smoke from the pinon pine which is native
to the U.S. Southwest (Tesfaigzi et al., 2002, 025575). PM concentrations in the woodsmoke
were 1,000 and 10,000 jug/m3 PM. The particles in this woodsmoke had a bimodal size
distribution with the smaller size fraction (74%) characterized by a MMAD of 0.405 um and
the larger size fraction (26%) characterized by a MMAD of 6.7-11.7 um. Many of these
larger particles would not be inhalable by the rat since 8 um MMAD particles are about
50% inhalable (Menache et al., 1995, 006533). Concentrations of gases were reported to be
15-106.4 ppm CO, 2.2-18.9 ppm NO, 2.4-19.7 ppm NOx and 3.5-13.8 ppm total hydrocarbon
in these exposures. Respiratory function measured by whole-body plethysmography
demonstrated a statistically significant increase in total pulmonary resistance in rats
exposed to 1000 |ug/m3 woodsmoke. Additional effects were found at 10,000 |u,g/m3.
Inflammatory and histopathological responses were also evaluated (see Sections 7.3.3.2 and
7.3.5.1).
July 2009
7-44
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
7.3.3. Pulmonary Inflammation
7.3.3.1.	Epidemiologic Studies
One epidemiologic study examined the relationship of airway inflammation (eNO) and
PM measures, other pollutants, and indicators of motor vehicle emissions in Windsor,
Ontario (Dales et al., 2008, 156378). This cohort of 2,402 school children estimated PM2.5
and PM10-2.5 for each child's residence at the postal code level with an evaluated statistical
model (Wheeler et al., 2006, 103905). Each 10 pg/m3 increase in 1-yr PM2.5 was associated
with a 39% increase in eNO (p = 0.058). Associations between eNO and PM10-2.5 were
positive but not statistically significant.
7.3.3.2.	Toxicological Studies
CAPs Studies
An important set of subchronic studies involved exposure of normal (C57BL1/6) mice,
ApoE"'" and the double-knockout ApoE'/LDLR"'" mice to Tuxedo, NY CAPs for 5-6 month
(March, April or May through September 2003 (Lippmann et al., 2005, 087452). The
average PM2.5 exposure concentration was 110 pg/m3. Animals were fed a normal chow diet
during the CAPs exposure period. No pulmonary inflammation was observed in response to
CAPs exposure as measured by BALF cell counts and histology. The lack of a persistent
pulmonary response may have been due to adaptation of the lung following repeated
exposures. In fact, a parallel study examined CAPs-related gene expression in the double-
knockout animals and found upregulation of numerous genes in lung tissue (Gunnison and
Chen, 2005, 087956). An in vitro study conducted simultaneously found daily variations in
CAPs-mediated NFkB activation in cultured human bronchial epithelial cells, suggesting
that transcription factor-mediated gene upregulation could occur in response to CAPs
(Maciejczyk and Chen, 2005, 087456). It should be noted that significant cardiovascular
effects were observed in these subchronic studies which are discussed in Section 7.2.1.2.
Araujo et al. (2008, 156222) compared the relative impact of ultrafine (0.01-0.18 pm)
versus fine (0.01-2.5 pm) PM inhalation in ApoE"'" mice following a 40 day exposure (5
hs/day x 3 days/week for 75 total hours). Animals were fed a normal chow diet and exposed
to PM from November 3-December 12, 2005 in a mobile inhalation laboratory that was
parked 300 m from the 110 Freeway in downtown Los Angeles. Particles were concentrated
to -440 pg/m3 for the fine exposures and -110 pg/m3 for the ultrafine exposures,
representing a roughly 15-fold increase in concentration from ambient levels! the number
concentration of PM in the fine and ultrafine chambers were roughly equivalent (4.56 x 105
and 5.59 x 10B particles/cm3, respectively). Over 50% of the ultrafine PM was comprised of
organic carbon compared to only 25% for PM2.5. No major increase in BALF inflammatory
July 2009
7-45
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
cells was found in response to PM. However ultrafine PM exposure resulted in significant
cardiovascular and systemic effects (see Section 7.3.5.2).
Diesel Exhaust
Gottipolu et al. (2009, 190360) exposed WKY and SH rats to filtered air or DE for
4 wk as described in Section 7.3.2.2. Previous studies from this laboratory have shown
enhanced effects of PM in SH compared with WKY rats. Although the main focus of this
recent study was on DE-induced mitochondrial oxidative stress and hypertensive gene
expression in the heart (see Section 7.2.7.1), some pulmonary effects were also found.
Subchronic exposure to DE resulted in a dose-dependent increase in BALF neutrophils in
both rat strains although levels of measured cytokines were not altered. Histopathological
analysis of lung tissue rats exposed to the higher concentration of DE demonstrated
accumulation of particle-laden macrophages as well as focal alveolar hyperplasia and
inflammation. Effect on indices of injury are discussed in Section 7.3.5.1.
Ishihara and Kagawa (2003, 096404) exposed Wistar rats to filtered air and DE
containing 200, 1,000 and 3,000 |ug/m3 PM for 16 h/day and 6 days/week for 6, 12, 18 or 24
months. The mass median particle diameter was reported to be between 0.3 and 0.5 |um.
Concentrations of gases ranged from 2.93-35.67 ppm NOx, 0.23-4.57 ppm SO2, 1.8-21.9 ppm
CO in the DE exposures. Statistically significant increases in total numbers of
inflammatory cells and neutrophils in BALF were observed beginning at 6-12 months of
exposure to DE containing 1,000 and 3,000 |u,g/m3 PM. When rats were exposed to DE
containing 1,000 ug/m3 PM, which was filtered to remove PM, the inflammatory cell
response was significantly diminished. These results implicate the PM fraction of DE as a
key determinant of the inflammation. The PM fraction was also found to mediate the
increase in protein levels, the decrease in PGE2 levels and alterations in mucus and
surfactant components observed in BALF (see Section 7.3.5.1).
Li et al. (2007, 155929) exposed BALB/c and C56BL/6 mice to low dose DE as
described in Section 7.3.2.2. for 1, 4 and 8 wk. Increases in numbers of BALF macrophages
and total inflammatory cells were observed in BALB/c mice at 8 wk but not 4 wk of DE
exposure. Persistent increases in numbers of BALF neutrophils and lymphocytes were
observed in both strains at 4 and 8 wk of DE exposure. Corresponding increases in BALF
cytokines differed between the two strains. These results should be interpreted with
caution since comparisons were made with Day 0 controls rather than age-matched
controls. No histopathological changes in the lungs were seen at any time point after DE
exposure. This study demonstrated differences in pulmonary responses to low dose DE
between 2 mouse strains. Altered AHR, pulmonary inflammation, markers of oxidative
stress and effects of antioxidant intervention were also evaluated (see Sections 7.3.2.2 and
7.3.4.1). Although no attempt was made in this study to determine the effects of gaseous
July 2009
7-46
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
components of DE on the measured responses, concentrations of gases were very low
suggesting that PM may have been responsible for the observed effects.
In a study by Hiramatsu et al. (2003, 155846), BALB/c and C57BL/6 mice were
exposed to DE (PM concentrations 100 and 3000 ug/m3) for 1 or 3 months. Concentrations of
gases were reported to be 3.5-9.5 ppm CO, 2.2-14.8 ppm NOx, and less than 0.01 ppm SO2.
Modest increases in BALF neutrophils and lymphocytes were observed in response to DE in
both mouse strains at 1 and 3 months. Histological analysis demonstrated DEP-laden
alveolar macrophages in alveoli and peribronchial tissues at both time points. Bronchus-
associated lymphoid tissue developed after 3 months exposure to the higher concentration
of DE in both mouse strains. Mac-1 positive cells (a marker of phagocytic activation of
alveolar macrophages) were also increased in BALF of BALB/c mice exposed to the higher
concentration of DE for 1 and 3 months. Increased expression of several cytokines and
decreased expression of iNOS mRNA was observed in DE-exposed mice at 1 and 3 months.
NFkB activation was also noted following 1 month exposure to the lower concentration of
DE. No attempt was made in this study to determine the responses to gaseous components
of the DE.
In a study by Reed et al. (2006, 156043), healthy Fisher 344 rats and A/J mice were
exposed to DE (PM concentration = 30, 100, 300 and 1000 |ug/m3) by whole body inhalation
for 6 h/day, 7 days/week for either 1 week or 6 months. Concentrations of gases were
reported to be 2.0-45.3 ppm NO, 0.2-4.0 ppm NO2, 1.5-29.8 ppm CO and 8-365 ppb SO2.
Short-term responses are discussed in Section 6.3.3.3 and 6.3.7.1, and sub-chronic systemic
effects are presented in Section 7.2.4.1. Six months of exposure resulted in no measurable
effects on pulmonary inflammation. However numerous black particles were observed
within alveolar macrophages after 6 months of exposure.
Seagrave et al. (2005, 088000) evaluated pulmonary responses in male and female
CDF (F-344)/CrlBR rats exposed 6 h/day for 6 months to filtered air or DE at concentrations
ranging from 30-1000 ug/m3 PM. Concentrations of gases were reported for the highest
exposure as 45.3 ppm NO, 4.0 ppm NO2, 29.8 ppm CO and 2.2 ppm total vapor
hydrocarbon. No changes in BALF cells were noted. A small decrease in TNFa was seen in
BALF of female rats exposed to the highest concentration of DE for 6 months. Pulmonary
injury also was evaluated (Section 7.3.5.1). Thus changes in BALF markers are modest and
gender specific.
Woodsmoke
Seagrave et al. (2005, 088000) also evaluated pulmonary responses in male and
female CDF (F344)/CrlBR rats exposed 6 h/day for 6 months to filtered air or HWS
concentrations ranging from 30-1000 ug/m3 PM. Concentrations of gases were reported for
the highest exposure as 3.0 ppm CO and 3.1 ppm total vapor hydrocarbon. A small increase
in BALF neutrophils was observed in male rats exposed to the lowest concentration of
July 2009
7-47
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
HWS. Female rats exhibited a decrease in BALF MIP-2 at the highest concentration of
HWS. Pulmonary injury also was evaluated (Section 7.3.5.1). In general, responses to HWS
were more remarkable than responses to DE seen in a parallel study. However these
gender-specific responses are modest and difficult to interpret.
In a study by Reed et al. (2006, 156043), Fisher 344 rats, SHR rats, A/J mice and
C57BL/6 mice were exposed to clean air or HWS (PM concentrations 30, 100, 300 and
1000 |ug/m3) by whole body inhalation for 6 h/day, 7 days/week for either 1 week or 6
months. Concentrations of gases ranged from 229.0-14887.6 mg/m3 for CO, 54.9-139.3 jug/m3
for ammonia, and 177.6- 3455.0 jug/m3 nonmethane VOC in these exposures. Short-term
responses are discussed in Section 6.3.7.2 and sub-chronic effects are presented in
Section 7.2.4.1. Histological analysis of lung tissue showed minimal increases in alveolar
macrophages. The effects of HWS on bacterial clearance are discussed below
(Section 7.3.7.2).
Another study evaluated the effects of subchronic woodsmoke exposure in Brown
Norway rats and is described in detail in Section 7.3.2.2 (Tesfaigzi et al., 2002, 025575).
Numbers of alveolar macrophages in BALF were significantly increased in rats exposed to
1000 jug/m3 woodsmoke for 12 wk, but no changes were seen in numbers of other
inflammatory cells. A large percent of BALF macrophages contained carbonaceous material.
Histological analysis of lung tissue showed minimal to mild inflammation in the epiglottis
of the larynx in rats exposed to both concentrations of woodsmoke.
Ramos et al. (2009, 190116) examined the effects of subchronic woodsmoke exposure
on the development of emphysema in guinea pigs. Inflammation is thought to be involved in
the pathogenesis of this form of COPD. Statistically significant increases in total numbers
of BALF cells were observed in guinea pigs exposed to smoke for 1-7 months, with numbers
of macrophages increased at 1-4 months and numbers of neutrophils increased at 4-7
months. At 4 months, alveolar mononuclear phagocytic and lymphocytic peribronchiolar
inflammation were observed by histopathological analysis of lung tissue. This study is
discussed in depth in Section 7.3.5.1.
Model Particles
Wallenborn et al. (2008, 191171) examined the pulmonary, cardiac and systemic
effects of subchronic exposure to particulate ZnSC>4. WKY rats were exposed nose-only to
10, 30, or 100 pg/m3 ultrafine particulate ZnSCk for 5 h/day and 3 day/wk over a 16-week
period. Particle size was reported to be 31-44 nm measured as number median diameter. No
changes in pulmonary inflammation or injury were observed although cardiac effects were
noted (see Section 7.2.7.1.). This study possibly demonstrates a direct effect of ZnSCk on
extrapulmonary systems, as suggested by the lack of pulmonary effects.
July 2009
7-48
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
7.3.4. Pulmonary Oxidative Response
7.3.4.1. Toxicological Studies
Urban Air
An interesting new study evaluated the effects of subchronic exposure to ambient
levels of urban particles on the development of emphysema in papain-treated mice (Lopes
et al., 2009, 190430). Since oxidative stress is thought to contribute to the development of
emphysema, 8-isoprostane levels were measured in lung tissue from the 4 groups of mice
used in this study A statistically significant increase in 8-isoprostane, a marker of oxidative
stress, was observed in lungs from mice treated with papain and exposed to ambient air
compared with the other groups of mice. This study is described in greater depth in
Section 7.3.5.1.
Diesel Exhaust
Li et al. (2007, 155929) exposed mice to low dose DE for 1, 4 and 8 wk as described in
Section 7.3.2.2. Markers of oxidative stress and effects of antioxidant intervention were
evaluated in this model. While HOI mRNA and protein were increased in lung tissues of
both mouse strains after 1 week of DE exposure (see Section 6.3.4.2), at 8 wk of DE
exposure, HOI protein levels remained high in C57BL/6 mice but returned to control
values in BALB/c mice. This study demonstrates differences in pulmonary responses to low
dose DE between two mouse strains. Furthermore, this study suggests that adaptation
occurs in BALB/c mice during prolonged DE exposure since the increase in HOI protein
seen in both strains at 1 week of exposure was only seen in C57BL/6 mice at 8 wk. Altered
AHR (Section 7.3.2.2) and pulmonary inflammation (Section 7.3.3.2) were also evaluated.
Although no attempt was made in this study to determine the effects of gaseous
components of DE on the measured responses, concentrations of gases were very low. This
suggests that PM may have been responsible for the observed effects.
7.3.5. Pulmonary Injury
7.3.5.1. Toxicological Studies
Urban Air
An important new study evaluated the effects of chronic exposure to ambient levels of
urban particles on lung development in the mouse (Mauad et al., 2008, 156743). Both
functional and anatomical indices of lung development were measured in mice exposed
July 2009
7-49
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
prenatally and/or postnatally as described in Section 7.3.2.2. Animals were sacrificed at 15
and 90 days of age for histological analysis of lungs. Histological analysis demonstrated the
presence of mild foci of macrophages containing black dots of carbon pigment in the
prenatal and postnatal exposure group at 90 days. In addition, the alveolar spaces of 15-day
old mice in the prenatal and postnatal exposure group were enlarged compared with
controls. Morphometric analysis demonstrated statistically significant decreases in surface
to volume ratio at 15 and 90 days in the pre and postnatal group compared with controls.
Since alveolarization is normally complete by 15 days of age, these results suggest
incomplete alveolarization in the 15-day old group and an enlargement of air spaces in the
90"day old group. These anatomical changes correlated with decrements in pulmonary
function which are discussed in Section 7.3.2.2.
Prolonged exposure to low levels of ambient air pollution beginning in early life has
been linked to secretory changes in the nasal cavity of mice, specifically increased
production of acidic mucosubstances (Pires-Neto et al., 2006, 096734). Six day-old Swiss
mice were continuously chamber exposed to ambient or filtered Sao Paulo air for 5 months.
Concentrations in the "polluted chamber" vs. "clean chamber" were (in ug/m3) 59.52 vs.
37.08 for NO2, 12.52 vs. 0 for BC, and 46.49 vs. 18.62 for PM2.5. Thus, pollutant levels were
reduced by filtration but not entirely eliminated. Compared to filtered air, exposure to
ambient air resulted in increased total mucus and acidic mucus in the epithelium lining the
nasal septum, but no statistically significant differences in other parameters (amount of
neutral mucus, volume proportions of neutral mucus, total mucus, or nonsecretory
epithelium, epithelial thickness, or ratio between neutral and acidic mucus). The
physicochemical properties of mucus glycoproteins are critical to the protective function of
the airway mucus layer. Acidified mucus is more viscous, and is associated with a decrease
in mucociliary transport. Thus acidic mucosubstances may represent impaired defense
mechanisms in the respiratory tract.
An interesting new study evaluated the effects of subchronic exposure to ambient
levels of urban particles on the development of emphysema in papain-treated mice (Lopes
et al., 2009, 190430). Emphysema is a form of COPD caused by the destruction of
extracellular matrix in the alveolar region of the lung which results in airspace
enlargement, airflow limitation and a reduction of the gas-exchange area of the lung.
Inflammation, oxidative stress, protease imbalance and apoptosis are thought to contribute
to the development of emphysema. In this study, male BALB/c mice were continuously
exposed to ambient or filtered Sao Paulo air for 2 months. Concentrations of PM2.5 in the
"polluted chamber" vs. "clean chamber" were 33.86 ± 2.09 vs. 2.68 ± 0.38 |u.g/m3. Thus
filtration reduced PM levels considerably. Ambient concentrations of CO and SO2 were 1.7
ppm and 16.2 ug/m3 respectively. No significant difference was observed in the
concentrations of NO2 in the "polluted chamber" vs "clean chamber" (60-80 ug/m3). Half of
the mice were pre-treated with papain by intranasal instillation in order to induce
July 2009
7-50
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
emphysema. Morphometric analysis of lung tissue demonstrated a statistically significant
increase in mean linear intercept, a measure of airspace enlargement, in papain-treated
mice compared with saline "treated controls exposed to filtered air. While exposure to
ambient air failed to increase mean linear intercept values in saline-treated mice, mean
linear intercept values were significantly increased in papain-treated mice exposed to
ambient air compaired with papain-treated mice exposed to filtered air. A similar pattern of
responses was observed for the volume proportion of collagen and elastin fibers in alveolar
tissue, which are markers of alveolar wall remodeling. Lung immunohistochemical analysis
demonstrated an effect of papain, but not ambient air, on macrophage cell density and
matrix metalloproteinase 12-positive cell density. No differences in caspase-3 positive cells,
a marker of apoptosis, were observed between the 4 groups of mice. Oxidative stress was
evaluated in this model as described in see Section 7.3.4.1. Taken together, results of this
study demonstrate that urban levels of PM, mainly from traffic sources, worsen protease-
induced emphysema in an animal model.
Pulmonary vascular remodeling, measured by a decrease in the lumen to wall ratio,
was observed in mice exposed to ambient Sao Paulo air for 4 months (Lemos et al., 2006,
088594). This study is described in greater detail in Section 7.2.1.2.
Kato and Kagawa (2003, 089563) exposed Wistar rats to roadside air contaminated
mainly with automobile emissions (55.7 to 65.2 ppb NO2 and 63 to 65 |ug/m3 suspended PM
[SPM]) and examined the effects on respiratory tissue after 24, 48, or 60 wk of exposure.
The surface of the lungs was light gray in color after all durations of exposure, and BC
particle deposits accumulated with prolonged exposure. These characteristics were not
evident in filtered air-exposed control animals, although filtered air contained low levels of
air pollutants (< 6.2 ppb NO2 and 15 jug/m3 SPM). The most common change observed using
transmission electron microscopy was the presence of particle laden (anthracotic) alveolar
macrophages, or anthracosis, in a wide range of pulmonary tissues, including the
submucosa, tracheal- and bronchiole-associated lymph nodes, alveolar wall and space,
pleura, and perivascular connective tissue. These changes were evident after 24 wk and
increased with duration of exposure. Other changes included increases in the number of
mucus granules in goblet cells, mast cell infiltration (but no degranulation) after 24 wk,
increased lysosomes in ciliated cells, some altered morphology of Clara cells, and
hypertrophy of the alveolar walls after 48 wk. No goblet cell proliferation was observed, but
slight, variable acidification of mucus granules appeared after 24 and 48 wk and
disappeared after 60 wk. Anthracotic macrophages were seen in contact with plasma cells
and lymphocytes in the lymphoid tissue, suggesting immune cell interaction in the
immediate vicinity of particles. Even after 60 wk, no lymph node anthracosis was observed
in the filtered air group.
In a post-mortem study of lung tissues from 20 female lifelong residents of Mexico
City, a high PM locale, histology demonstrated significantly greater amounts of fibrous
July 2009
7-51
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
tissue and muscle in the airway walls compared to subjects from Vancouver (Churg et al.,
2003, 087899), a city with relatively low PM levels. Electron microscopy showed
carbonaceous aggregates of ultrafine particles, which the authors conclude penetrate into
and are retained in the walls of small airways. The study shows an association between
retained particles and airway remodeling in the form of excess muscle and fibrotic walls.
The subjects were deemed suitable for examination based on never-smoker status, no use of
biomass fuels for cooking, no known occupational particle/dust exposure, death by cause
other than respiratory disease, and extended residence in each locale (lifelong for Mexico
City and >20 years for Vancouver). However, subjects from the two locales were not
matched with respect to ethnicity, sex (20 females from Mexico City vs. 13 females and 7
males from Vancouver), or mean age at death (66 ± 9 vs. 76 ± 11), and other possibly
influential factors such as exercise or diet were not considered.
Diesel Exhaust
Gottipolu et al. (2009, 190360) exposed WKY and SH rats to filtered air or DE as
described in Section 7.3.2.2. (Previous studies from this laboratory have shown enhanced
effects of PM in SH compared with WKY rats. Although the main focus of this recent study
was on DE-induced mitochondrial oxidative stress and hypertensive gene expression in the
heart (see Section 7.2.7.1), some pulmonary effects were found. Inflammatory effects are
described in Section 7.3.3.2. GGT activity in BALF was increased in both strains in
response to the higher concentration of DE. No DE-related changes were observed in BALF
protein or albumin. Histopathological analysis of lung tissue rats exposed to the higher
concentration of DE demonstrated accumulation of particle-laden macrophages as well as
focal alveolar hyperplasia and inflammation. No effects on indices of pulmonary function
were observed (see Section 7.3.2.2.)
Ishihara and Kagawa (2003, 096404) exposed rats to DE for up to 24 months as
described in Section 7.3.3.2. A statistically significant increase in BALF protein was
observed at 12 months of exposure to DE containing 1000 jug/m3 PM. This response was
attenuated when the DE was filtered to remove PM. Pulmonary inflammation was noted
and is described in Section 7.3.3.2.
Seagrave et al. (2005, 088000) evaluated pulmonary responses in rats exposed to DE
for up to 6 months as described in Section 7.3.3.2. A small increase in LDH was seen in
BALF of female rats exposed to the highest concentration of DE for 6 months. Pulmonary
inflammation was also evaluated (Section 7.3.3.2). The changes in BALF markers in this
study were modest and gender specific.
Gasoline Exhaust
Reed et al. (2008, 156903) examined a variety of health effects following subchronic
inhalation exposure to gasoline engine exhaust. Male and female CDF (F344)/CrlBR rats,
July 2009
7-52
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
SHR rats and male C57BL/6 mice were exposed for 6 h/day and 7 days/week for a period of
3 days to 6 months. The dilutions for the gasoline exhaust were L10, 1-15 and 1:90; filtered
PM was at the L10 dilution. PM mass ranged from 6.6 to 59.1 pg/m3, with the
corresponding number concentration between 2.6 x 104 and 5.0 x 105 particles/cm3.
Concentrations of gases ranged from 12.8-107.3 ppm CO, 2.0-17.9 ppm NO, 0.1-0.9 ppm
NO2, 0.09-0.62 ppm SO2 and 0.38-3.37 ppm NH3. Other effects for these studies are
presented in Sections 7.2.4.1.and 7.3.6.1. No pulmonary inflammation or histopathological
changes were noted in the F344 rats and A/J mice, except for a time-dependent increase in
the number of macrophages containing PM. However statistically significant increases of
47% and 29% in BALF LDH were observed in female and male F344 rats, respectively, after
6 months of exposure to the highest concentration of engine exhaust. This response was
absent when gasoline exhaust was filtered, implicating PM as a key determinant of this
response. In addition, exposure to the highest concentration of gasoline exhaust resulted in
statistically significant decreases in hydrogen peroxide and superoxide production in
unstimulated and stimulated BALF macrophages. Interestingly, hypermethylation of lung
DNA was observed in male F344 rats following 6 months of exposure to gasoline exhaust
containing 30 pg/m3 particulates. This response was PM-dependent since it was absent in
mice exposed to filtered gasoline exhaust. The significance of this epigenetic change in
terms of respiratory health effects is not known. However, alterered patterns of DNA
methylation can affect gene expression and are sometimes associated with altered immune
responses and/or the development of cancer.
Woodsmoke
Seagrave et al. (2005, 088000) also evaluated pulmonary responses in rats to HWS for
6 months as described in Section 7.3.3.2. Increases in BALF LDH and protein were seen in
male but not female rats Female rats exhibited a decrease in BALF glutathione at the
highest concentration of HWS. Decreases in BALF alkaline phosphatase were found in both
males and females exposed to 1000 |ug/m3 HWS. Male rats exposed to 100 and 300 |ug/m3
HWS exhibited a decrease in BALF 6-glucuronidase activity. Pulmonary inflammation was
also evaluated (Section 7.3.3.2). These changes in BALF markers in this study were modest
and gender specific.
Another study evaluated the effects of subchronic woodsmoke exposure in Brown
Norway rats as described in Section 7.3.2.2. (Tesfaigzi et al., 2002, 025575). Exposure to
1,000 ug/m3 woodsmoke for 12 wk resulted in a statistically significant increase in Alcian
Blue- (AB) and Periodic Acid Schiff- (PAS) positive airway epithelial cells compared to
controls, indicating an increase in mucous secretory cells containing neutral and acid
mucus, respectively. More significant histopathological responses were found following
exposure to 10,000 ug/m3 of DE. Pulmonary function and inflammation were evaluated also
July 2009
7-53
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
but are not discussed here due to the extremely high exposure level (Sections 7.3.2.2 and
7.3.3.2).
Ramos et al. (2009, 190116) examined the effects of subchronic woodsmoke exposure
on the development of emphysema in guinea pigs. In particular, the involvement of
macrophages and macrophage-derived matrix metalloproteinases (MMP) in woodsmoke-
related responses was investigated. Guinea pigs were exposed to ambient air or to whole
smoke from pine wood for 3 h/day and 5 days/week over a 7-month period. PMio and PM2.5
concentrations in the exposure chambers were reported to be 502 ± 34 and 363 ± 23 |u,g/m3,
respectively, while the concentration of CO was less than 80 ppm. COHb levels were
reported to be 6% in controls and 15-20% in smoke-exposed guinea pigs. Statistically
significant decreases in body weight were observed in guinea pigs exposed to smoke for 4 or
more months compared with controls. Statistically significant increases in total numbers of
BALF cells were observed in guinea pigs exposed to smoke for 1-7 months, with numbers of
macrophages increased at 1-4 months and numbers of neutrophils increased at 4-7 months.
At 4 months, alveolar mononuclear phagocytic and lymphocytic peribronchiolar
inflammation, as well as bronchiolar epithelial and smooth muscle hyperplasia, were
observed by histopathological analysis of lung tissue. Emphysematous lesions, smooth
muscle hyperplasia and pulmonary arterial hypertension were noted at 7 months.
Morphometric analysis of lung tissue demonstrated statistically significant increases in
mean linear intercept values, a measure of airspace enlargement, in guinea pigs at 6 and 7
months of exposure. Statistically significant increases in elastolytic activity was observed in
BAL macrophages and lung tissue homogenates at 1-7 months of exposure. Lung
collagenolytic activity was also increased at 4-7 months of exposure and corresponded in
time with the presence of active forms of MMP-2 and MMP-9 in lung tissue homogenates
and BALF. Furthermore, MMP-1 and MMP-9 immunoreactivity was detected in
macrophages, epithelial and interstitial cells in smoke-exposed animals at 7 months.
Increased levels of MMP-2 and MMP-9 mRNA were also found in smoke-exposed guinea
pigs after 3-7 months. Apoptosis was found in BAL macrophages (TUNEL assay) from
guinea pigs exposed to smoke for 3-7 months and in alveolar epithelial cells (caspase-3
immunore activity) after 7 months. Taken together, these results provide evidence that
subchronic exposure to woodsmoke leads to the development of emphysematous lesions
accompanied by the accumulation of alveolar macrophages, increased levels and activation
of MMP's, connective tissue remodeling and apoptosis. Unfortunately, the high levels of CO
and COHb reported in this study make it difficult to conclude that woodsmoke PM alone is
responsible for these dramatic effects.
July 2009
7-54
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
7.3.6. Allergic Responses
7.3.6.1. Epidemiologic Studies
A number of epidemiologic studies have found associations between PM and allergic
(or atopic) indicators. Allergy is a major driver of asthma, which has been associated with
PM in studies discussed in previous sections. In a study by Annesi-Maesano (2007, 093180)
(described in Section 7.3.1.1) atopic asthma was related to PM2.5 (OR 1.43 [95% CP
1.07-1.91]) and positive skin prick test to common allergens was also increased with higher
PM levels. This report is consistent with the results from an earlier study (Penard'Morand
et al., 2005, 087951) in the same sample of children that associated allergic rhinitis and
atopic dermatitis with PM10. Also, Morgenstern et al. (2008, 156782) found statistically
significant effects for asthmatic bronchitis, hay fever, and allergic sensitization to pollen in
a cohort of children in Germany examining modeled PM2.5 data at birth addresses. Distance
to a main road had a dose-response relationship with sensitization to outdoor allergens.
Nordling et al. (2008, 097998) (discussed above in Section 7.3.2.1) reported a positive
association of PM10 exposure during the first year of life with allergenic sensitization (igE
antibodies) to inhaled allergens, especially pollen. In a study by Brauer et al. (2007,
090691) (discussed above in Section 7.3.1.1) an interquartile range increase in PM2.5was
associated with an increased risk of sensitization to food allergens (OR 1.75 [95% CI
1.23-2.47]). A significant association was found for sensitization to any allergen, but none
was found for sensitization to specific indoor or outdoor aeroallergens or atopic dermatitis
(eczema). In a study by Janssen et al. (2003, 133555), PM2.5 was associated with allergic
indicators such as hay fever (ever), skin prick test reactivity to outdoor allergens, current
itchy rash, and conjunctivitis in Dutch children. These same outcomes were also associated
with proximity of the school to truck traffic but not car traffic, suggesting a role for diesel-
related pollution. Consistent with the aforementioned Dutch study by Brauer et al. (2007,
090691). PM2.5 was not associated with eczema.
Mortimer et al. (2008, 187280) examined the association between prenatal and early-
life exposures to air pollutants with allergic sensitization in a cohort of 170 children with
asthma, age 6-11, living in central California. Sensitization to at least one allergen was
associated with higher levels of PM10 and CO during the entire pregnancy and 2nd
trimester and higher PM10 during the first two years of life. Sensitization to at least one
indoor allergen was associated with higher exposures to PM10 and CO in during the entire
pregnancy and during the 2nd trimester. However, no significant associations remained for
PM10 after adjustment for co-pollutants, effect modifiers, or potential cofounders in addition
to year of birth. The authors advise that the large number of comparisons may be of concern
and this study should be viewed as an exploratory, hypothesis-generating undertaking. In
examining the National Health Interview Survey for the years 1997-2006, Bhattacharyya
July 2009
7-55
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
et al. (2009, 180154) found relationships between air quality and the prevalence of hay
fever and sinusitis. However, the air quality data were not clearly defined and as such
caution is required in interpretation of these results. In contrast, Bayer-Oglesby et al.
(2005, 086245) found no significant association between declining levels of PMio and hay
fever in Switzerland. In a study by Oftedal et al. conducted in Oslo, Norway (2007, 191948),
earlylife exposure to PMio or PM2.5 was generally not associated with sensitization to
allergens in 9-10-year-old children! lifetime exposures to PMio and PM2.5 were associated
with dust mite allergy, but the association was diminished by adjustment for socioeconomic
factors . In Norway, wood burning in the wintertime is thought to account for about half of
the PM2.5 levels. Although associations between PM and reactivity to specific allergens have
been reported in long-term studies, there is a consistent lack of correlation between PM and
total IgE levels, indicating a selective enhancement of allergic responses.
7.3.6.2. Toxicological Studies
Diesel Exhaust
Exposure to relatively low doses of DE has been shown to exacerbate asthmatic
responses in OVA sensitized and challenged BALB/c mice (Matsumoto et al., 2006, 098017).
Mice were intraperitoneally sensitized and intranasally challenged lday prior to inhalation
exposure to DE (PM concentration 100 jug/m3! CO, 3.5 ppm! NO2, 2.2 ppm! SO2 <0.01 ppm)
for 1 day or 1, 4, or 8 wk (7/h/day, 5 days/week, endpoints 12-h post DE exposure). Results
from the 1- and 4-week exposures are described in Section 6.3.6.3. It should be noted that
control mice were left in a clean room as opposed to undergoing chamber exposure to
filtered air. The significant increases in AHR and airway sensitivity observed following
shorter exposure periods did not persist at 8 wk. BALF cytokines were altered by DE
exposure with only RANTES significantly elevated after 8 wk. DE had no effect on OVA
challenge-induced peribronchial inflammatory or mucin positive cells. These results suggest
that adaptive processes may have occurred during prolonged exposure to DE.
Gasoline Exhaust
In a study by Reed et al. (2008, 156903), BALB/c mice were exposed to whole gasoline
exhaust diluted 1^10 (H), l: 15 (M), or 1:90 (L), filtered exhaust at the 1^10 (HF), or clean air
for 6 h/day (atmospheric characterization described in Section 6.3.6.3). GEE exposure from
conception through four weeks of age induced slight but non-significant increases in ova-
specific IgGl in offspring but had no significant effect on airway reactivity, BAL cytokine or
cell concentrations, although there were non-significant increases in lung neutrophils and
eosinophils. Significant increases in total serum IgE were observed, but this effect persisted
after filtration of particles and was thus attributed to gas phase components.
July 2009
7-56
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Woodsmoke
In a study by Tesfaigzi et al. (2005, 156116), Brown Norway rats were sensitized and
challenged with ovalbumin. Rats were exposed for 70 days to filtered air or to 1000 ug/m';
HWS. Particles were characterized by a MMAD of 0.36 |um. Concentrations of gases were
reported to be 13.0 ppm CO and 3.1 ppm total vapor hydrocarbon with negligible NOx.
Respiratory function was measured in anesthetized animals by whole-body
plethysmography and demonstrated a significant increase in functional residual capacity as
well as a significant increase in dynamic lung compliance in HWS-exposed animals
compared to controls. No change in total pulmonary resistance or airway responsiveness to
methacholine was observed. BALF inflammatory cells were not increased, although
histological analysis demonstrated focal inflammation including granulomatous lesion and
eosinophilic infiltrations in HWS-exposed rats. Alterations of several cytokines in BALF
and plasma were noted. Changes in airway epithelial mucus cells and intraepithelial stored
mucosubstances were modest and did not achieve statistical significance. Results of this
study demonstrate that subchronic exposure to HWS had minimal effects on pulmonary
responses in a rat model of allergen sensitization and challenge.
7.3.7. Host Defense
7.3.7.1. Epidemiologic Studies
Epidemiologic studies of respiratory infections indicate an association with PM. This
is more evident when considering short-term exposures (Chapter 6), but studies of long-
term exposures have observed associations with general respiratory symptoms often caused
by infection, such as bronchitis. In a birth cohort study of approximately 4,000 Dutch
children, Brauer et al. (2007, 09069lXdescribed in Section 7.3.1.1) found significant positive
associations for PM2.5 with ear/nose/throat infections and doctor-diagnosed flu/serious cold
in the first four years of life. These results are consistent with an earlier study by Brauer et
al. (2006, 090757), which found that an increase of 3 |ug/m3 PM2.5 was associated with
modestly increased risk for ear infections in the Netherlands [OR 1.13 (95% CI, 1.00-1.27)].
ASwiss study by Bayer-Oglesby et al. (2005, 086245), discussed in Section 7.3.1.1 above,
demonstrated that declining levels of PM10 were associated with declining prevalence of
common cold and conjunctivitis. Because traffic-related pollutants such as ultrafine
particles are high near major roadways and then decay exponentially over a short distance,
Williams, et al. (2009, 191945) assessed exposure according to residential proximity to
major roads in a Seattle area study of postmenopausal women. Proximity to major roads
was associated with a 21% decrease in natural killer cell function, which is an important
defense against viral infection and tumors. This finding was limited to women who reported
exercising near traffic! other markers of inflammation and lymphocyte proliferation did not
July 2009
7-57
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
consistently differ according to proximity to major roads. In the Puget Sound region of
Washington, Karr et al. (2009, 191946) reported that there may be a modest increased risk
of bronchiolitis related to PM2.5 exposure for infants born just before the peak respiratory
syncytial virus (RSV) season. Risk estimates were stronger when restricted to cases
specifically attributed to RSV and for infants residing closer to highways. Emerging
evidence suggests that respiratory infections, particularly infection by viruses such as RSV,
can cause asthma or trigger asthma attacks.
7.3.7.2. Toxicological Studies
Diesel Exhaust
DE may affect systemic immunity. The proliferative response of A/J mouse spleen
cells following stimulation with T cell mitogens was suppressed by 6 months of daily
exposure to DE at concentrations at or above 300 |u,g/m3 PM (Burchiel et al., 2004, 055557).
B cell proliferation was increased at 300 |ug/m3 but unaffected at higher concentrations (up
to 1000 |ug/m3). Concentrations of gases and were reported in the parallel study by Reed et
al. (2004, 055625), described in Section 7.3.3.2 Pulmonary Inflammation/Diesel exhaust.
The Reed study reported a decrease in spleen weight in male mice (27% reduction in the
300 jug/m3 exposure group). The immunosuppressive effects of DE were not due to PAHs or
benzo(ayrene (BaP)-quinones (BPQs) since there were little, if any, of these compounds
present in the chamber atmosphere. It should be noted that sentinel animals were negative
for mouse parvovirus at the start of the study, but seroconverted by the end of the study,
indicating possible infection. Parvovirus can interfere with the modulation of lymphocyte
mitogenic responses (Baker, 1998, 156245). A six month exposure (6h/day, 7d/week) to 30,
100, 300 or 1000 ug/m3 of PM in DE did not significantly affect bacterial clearance in
C57BL/6 mice infected with Pseudomonas aeruginosa, although all levels reduced bacterial
clearance when the exposure only lasted a week (Harrod et al., 2005, 088144).
Characterization of the exposure atmosphere was given by Reed et al. (2004, 055625)
(Section 7.3.3.2 Pulmonary Inflammation/Diesel exhaust).
Gasoline Exhaust
In a study by Reed et al. (2008, 156903) (described in Section 6.3.6.3) long-term
exposure to fresh gasoline exhaust (6h/day, 7d/week for 6 months) did not affect clearance of
P. aeruginosa from the lungs of C57BL/6 mice.
Hardwood Smoke
One study demonstrated immunosuppressive effects of HWS (Burchiel et al., 2005,
088090}. Exposure to HWS increased proliferation of T cells from A/J mice exposed daily to
100 jug/m3 PM for 6 months, but produced a concentration-dependent suppression of
July 2009
7-58
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
proliferation at PM concentrations >300 Lig/m:;. No effects on B cell proliferation were
observed. Concentrations of NO and NO2 were not detectable or <40 ppb for all exposure
levels. CO was reported to be 2, 4, and 13 ppm for the 100, 300 and 1000 |ug/m3 PM
concentrations, respectively. Exposure atmospheres contained significant levels of
naphthalene and methylated napthalenes, fluorene, phenanthrene, and anthracene, as well
as low concentrations of several metals (K, Ca, and Fe) (Burchiel et al., 2005, 088090). It
should be noted that serologic analysis of study sentinel animals indicated infection with
parvovirus , which can interfere with the modulation of lymphocyte mitogenic responses
(Baker, 1998, 156245). In another study by Reed et al. (2006, 156043) C57BL/6 mice were
exposed to 30-1000 |u.g/m3 HWS by whole body inhalation for 6 months prior to instillation
of P. aeruginosa. Exposure characterizations are described in Section 7.3.3.2. Although
there was a trend toward increased clearance with increasing exposure concentrations,
there was no statistically significant effect of HWS exposure on bacterial clearance.
7.3.8.	Respiratory Mortality
Two large U.S. cohort studies examined the effect of long-term exposure to PM2.5 on
respiratory mortality with mixed results. In the ACS, Pope et al. (2004, 055880) reported
positive associations with deaths from specific cardiovascular diseases, but no PM2.5
associations were found with respiratory mortality. A followup to the Harvard Six Cities
study (Laden et al., 2006, 087605) used updated air pollution and mortality data and found
positive associations between long-term exposure to PM2.5 and mortality. Of special note is a
statistically significant reduction in mortality risk reported with reduced long-term fine
particle concentrations observed for deaths due to cardiovascular and respiratory causes,
but not for lung cancer deaths. There is some evidence for an association between PM2.5 and
respiratory mortality among post-neonatal infants (ages 1 mo to 1 yr) (see Section 7.4.1). In
summary, when deaths due to respiratory causes are separated from all-cause (non-
accidental) and cardiopulmonary deaths, there is limited and inconsistent evidence for an
effect of PM2.5 on respiratory mortality, with one large cohort study finding a reduction in
deaths due to respiratory causes associated with reduced PM2.5 concentrations, and another
large cohort study finding no PM2.5 associations with respiratory mortality.
7.3.9.	Summary and Causal Determinations
7.3.9.1. PM2.5
The epidemiologic studies reviewed in the 2004 PM AQCD suggested relationships
between long-term PM10 and PM2.5 (or PM2.1) exposures and increased incidence of
respiratory symptoms and disease. One of these studies indicated associations with
July 2009
7-59
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
bronchitis in the 24-city cohort (Dockery et al., 1996, 077269). They also suggested
relationships between long-term exposure to PM2.5 and pulmonary function decrements in
the CHS (Gauderman et al., 2000, 012531; Gauderman et al., 2002, 026013). These findings
added to the database of the earlier 22-city study of PM2.1 (Raizenne et al., 1996, 077268)
that found an association between exposure to ambient particle strong acidity and
impairment of lung function in children. No long-term exposure toxicological studies were
reported in the 2004 PM AQCD.
Recent studies have greatly expanded the evidence available since the 2004 PM
AQCD. New analyses have been conducted that include longer follow-up periods of the CHS
cohort through 18 years of age and provide evidence that effects from exposure to PM2.5
persist into early adulthood. Gauderman et al. (2004, 056569) reported that PM2.5 exposure
was associated with clinically and statistically significant deficits in FEVi attained at the
age of 18 years. In addition the strength and robustness of the outcomes were larger in
magnitude, and more precise than previous CHS studies with shorter followup periods.
Supporting this result are new longitudinal cohort studies conducted by other researchers
in other locations with different methods. These studies report results for PM10 that is
dominated by PM2.5. New studies provide positive associations from Mexico City, Sweden,
and a national cystic fibrosis cohort in the U.S. A natural experiment in Switzerland, where
PM levels had decreased, reported that improvement in air quality may slow the annual
rate of decline in lung function in adulthood, indicating positive consequences for public
health. Thus, the data are consistent and coherent across several study designs, locations
and researchers. As was found in the 2004 PM AQCD, the studies report associations with
PM2.5 and PM10, while most did not evaluate PM10-2.5. Associations have been reported with
fine particle components, particularly EC and OC. Source apportionment methods generally
have not been used in these long-term exposure studies.
Coherence and biological plausibility for the observed associations with lung function
decrements is provided by recent toxicological studies (Section 7.3.2.2). A recent study
demonstrated that pre- and postnatal exposure to ambient levels of urban particles affected
mouse lung development, as measured by anatomical and functional indices (Mauad et al.,
2008, 156743). Another study suggested that the developing lung may be susceptible to PM
since acute exposure to ultrafine iron-soot decreased cell proliferation in the proximal
alveolar region of neonatal rats (Pinkerton et al., 2004, 087465) (Section 6.3.5.3). Impaired
lung development is a viable mechanism by which PM may reduce lung function growth in
children. Other animal toxicological studies have demonstrated alterations in pulmonary
function following exposure to DE and WS (Section 7.3.2.2).
An expanded body of epidemiologic evidence for the effect of PM2.5 on respiratory
symptoms and asthma incidence now includes prospective cohort studies conducted by
different researchers in different locations, both within and outside the U.S. with different
methods. The CHS provides evidence in a prospective longitudinal cohort study that relates
July 2009
7-60
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
PM2.5 and bronchitic symptoms and reports larger associations for within-community
effects that are less subject to confounding than between-community effects (McConnell et
al., 2003, 049490). Several new studies report similar findings with long-term exposure to
PMio in areas where fine particles predominate. In England, an association was seen with
an increased prevalence of cough without a cold. Further evidence includes a reduction of
respiratory symptoms corresponding to decreasing PM levels in natural experiments in
cohorts of Swiss school children (Bayer-Oglesby et al., 2005, 086245) and adults (Schindler
et al., 2009, 191950).
New studies examine the relationship between long-term PM2.5 exposure and asthma
incidence. PM2.5 had the strongest modifying effect on the association between lung function
with asthma in an analysis of the CHS (Islam et al., 2007, 090697). The loss of protection
by high lung function against new onset asthma in high PM2.5 communities was observed
for all the lung function measures. In the Netherlands, an association with doctor-
diagnosed asthma was found in a birth cohort examining the first 4 years of life (Brauer et
al., 2007, 090691) Further, findings from an adult cohort suggest that traffic-related PM10
contributes to asthma development and that reductions in PM decrease asthma risk
(Kunzli et al., 2009, 191949).
A large proportion of asthma is driven by allergy, and the majority of recent
epidemiologic studies examining allergic (or atopic) indicators found positive associations
with PM2.5 or PM10 (Section 7.3.6.1). Limited evidence for PM-mediated allergic responses is
provided by toxicological studies of DE and woodsmoke, while effects of gasoline exhaust
were attributed to gaseous components (Section 7.3.6.2).
Long-term PM2.5 exposure is associated with pulmonary inflammation and oxidative
responses. An epidemiologic study found a relationship between PM2.5 and increased
inflammatory marker eNO among school children (Dales et al., 2008, 156378). Toxicological
studies of pulmonary inflammation have demonstrated mixed results, with subchronic DE
exposures generating increases and CAPs and WS inducing little or no response (Section
7.3.3.2). The pulmonary inflammation observed with DE was attributable to the particle
fraction. Toxicological studies also reported evidence of oxidative responses (Section 7.3.4.1).
Adaptation to prolonged DE was observed for some oxidative responses in addition to some
allergic and pulmonary function responses (Section 7.3.2.2 and 7.3.6.2).
Additional support for the relationship between long-term PM2.5 exposures and
respiratory outcomes is provided by pulmonary injury responses observed in toxicological
studies (Section 7.3.5.1). Markers of pulmonary injury were increased in rats exposed to DE
and gasoline exhaust! and these changes were attributable to PM. Further, lung DNA
methylation was observed in the gasoline exhaust study. Histopathological changes have
also been reported following exposure to heavily-trafficked urban air and to woodsmoke.
Findings include nasal and airway mucous cell hyperplasia accompanied by alterations in
mucus production which can lead to a loss of mucus-mediated protective functions!
July 2009
7-61
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
exacerbation of protease-induced emphysema! and mast cell infiltration and hypertrophy of
alveolar walls. These results provide biological plausibility for adverse respiratory outcomes
following long-term PM exposure.
Limited information is available on host defense responses (Section 7.3.7) and
respiratory mortality (Section 7.3.8) resulting from PM2.5 exposure. Several recent
epidemiologic studies suggest a relationship between long-term exposure to PM2.5or PM10
and infection in children and infants (Section 7.3.7.1). A few toxicological studies suggest
that DE exposure affects systemic immunity and although impaired bacterial clearance is
associated with short-term exposures to DE, neither DE or gasoline exhaust seems to have
this effect after longer exposures (Section 7.3.7.2).
In summary the strongest evidence for a relationship between long-term exposure to
PM2.5 and respiratory morbidity is provided by epidemiologic studies demonstrating
associations with decrements in lung function growth in children and with respiratory
symptoms and disease incidence in adults. Mean PM2.5 concentrations in these study
locations ranged from 13.8 to 30 pg/m3 during the study periods. These studies provide
evidence for associations in areas where PM is predominantly fine particles. A major
challenge to interpreting the results of these studies is that the PM size fractions and
concentrations of other air pollutants are often correlated! however, the consistency of
findings across different locations supports an independent effect of PM2.5. Recent
toxicological studies provide support for the associations with PM2.5 and decreases in lung
function growth in children. Pre- and postnatal exposure to ambient levels of urban
particles was found to affect mouse lung development, which provides biological plausibility
for the epidemiologic findings. Recent subchronic and chronic toxicological studies also
demonstrate altered pulmonary function, mild inflammation, oxidative responses,
histopathological changes including mucus cell hyperplasia and enhanced allergic
responses in response to CAPs, DE, urban air and woodsmoke and provide further
coherence and biological plausibility. Exacerbation of emphysematous lesions was noted in
one study involving exposure to urban air in a heavily-trafficked area. Collectively, the
evidence is sufficient to conclude that the relationship between long-term PM2.5 exposure and
respiratory effects is likely to be causal.
7.3.9.2. PM10-2.5
The 2004 PM AQCD did not report long-term exposure studies for PM10-2.5. The only
recent study to evaluate long-term exposure to PM10-2.5 found positive, but not statistically
significant associations with eNO (Dales et al., 2008, 156378). The evidence is inadequate to
determine if a causal relationship exists between long-term PM10 2.5 exposures and respiratory
effects.
July 2009
7-62
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
7.3.9.3. Ultrafine PM
The 2004 PM AQCD did not report long-term exposure studies for ultrafine PM. The
current evidence for long-term ultrafine PM effects is limited to toxicological studies.
Generally, subchronic exposure to DE induced pulmonary inflammation, which was in
contrast to ultrafine CAPs exposure (Section 7.3.3.2) It appeared that the PM fraction was
responsible for the inflammatory response with DE exposure. Long-term exposure to DE
also resulted in oxidative and allergic responses, although lung injury was not remarkable
(Sections 7.3.4.1 and 7.3.6.2). The evidence is inadequate to determine if a causal relationship
exists between long-term UFP exposures and respiratory effects.
7.4. Reproductive, Developmental, Prenatal and Neonatal
Outcomes
7.4.1. Epidemiologic Studies
This section evaluates and summarizes the scientific evidence on PM and
developmental and pregnancy outcomes and infant mortality. Infants and fetal development
processes may be particularly vulnerable to PM exposure, and although the physical
mechanisms are not fully understood, several hypotheses have been proposed involving
direct effects on fetal health, altered placenta function, or indirect effects on the mother's
health (Bracken et al., 2003, 156288; Clifton et al., 2001, 156360; Maisonet et al., 2004,
156725; Schatz et al., 1990, 156073; Sram et al., 2005, 087442). Study of these outcomes
can be difficult given the need for detailed data and potential residential movement of
mothers during pregnancy. Two recent articles have reviewed methodological issues
relating to the study of outdoor air pollution and adverse birth outcomes (Ritz and Wilhelm,
2008, 156914; Slama et al., 2008, 156985). Some of the key challenges to interpretation of
these study results include the difficulty in assessing exposure as most studies use existing
monitoring networks to estimate individual exposure to ambient PM; the inability to
control for potential confounders such as other risk factors that affect birth outcomes (e.g.,
smoking); evaluating the exposure window (e.g., trimester) of importance! and limited
evidence on the physiological mechanism of these effects (Ritz and Wilhelm, 2008, 156914;
Slama et al., 2008, 156985). Another uncertainty is whether PM effects differ by the child's
sex. A review of pre-term birth and low birth weight studies found limited indication that
effects may differ by gender, however sample size was limited (Ghosh et al., 2007, 091233).
Previous summaries of the association between PM concentrations and pregnancy
outcomes and infant mortality were presented in previous PM AQCDs. The 1996 PM AQCD
concluded that although few studies had been conducted on the link between PM and infant
mortality, the research "suggested an association," particularly for post-neonates (U.S. EPA,
July 2009
7-63
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
1996, 079380). In the 2004 PM AQCD, additional evidence was available on PM's effect on
fetal and early postnatal development and mortality (U.S. EPA, 2004, 056905) and although
some studies indicated a relationship between PM and pregnancy outcomes, others did not.
Studies identifying associations found that exposure to PMio early during pregnancy (first
month of pregnancy) or late in the pregnancy (six weeks prior to birth) were linked with
higher risk of preterm birth, including models adjusted for other pollutants, and that PM2.5
during the first month of pregnancy was associated with interuterine growth restriction.
However, other work did not identify relationships between PM10 exposure and low birth
weight. The state of the science at that time, as indicated in the 2004 PM AQCD, was that
the research provided mixed results based on studies from multiple countries, and that
additional research was required to better understand the impact of PM on pregnancy
outcomes and infant mortality. Considering evidence from recent studies discussed below,
along with previous AQCD conclusions, epidemiologic studies consistently report
associations between PM10 and PM2.5 exposure and low birth weight and infant mortality,
especially during the post-neonatal period. Animal toxicological evidence supports these
associations with PM2.5, but provides little mechanistic information or biological
plausibility. Information on the ambient concentrations of PM10 and PM2.5 in these study
sites can be found in Table 7-5.
Low Birth Weight
A large number of studies have investigated exposure to ambient PM and low birth
weight at term, including a U.S. national study, as well as two studies in the northeast
U.S., and four in California have been conducted. Parker and Woodruff (2008, 156846)
linked U.S. birth records for singletons delivered at 40 wk gestation in 2001-2003 during
the months of March, June, September and December to quarterly estimates of PM
exposure by county of residence and month of birth. They found an association between
PM10-2.5 and birthweight (-13 g [95% CP -18.3 to -7.6]) per 10 pg/m3 increase), but no such
association for PM2.5.
Maisonet et al. (2001, 016624) analyzed 89,557 births (1994-96) in six northeastern
cities (Boston and Springfield MA, Hartford CT, Philadelphia and Pittsburgh PA,
Washington DC). Each city had three PM10 monitors measuring every sixth day. Results
from multiple monitors were averaged in each city. Exposure was determined for each
trimester of pregnancy and categorized by quartiles (<25, 25-30, 31-35, 36-43 pg/m3) and
95th percentile (>43pg/m3). There was no increased risk for low birth weight at term
associated with PM10 exposure during any trimester of pregnancy. When birth weight was
considered as a continuous outcome, exposure to PM10 was not associated with a reduction
in mean birth weight.
July 2009
7-64
DRAFT-DO NOT CITE OR QUOTE

-------
Table 7-5. Characterization of ambient PM concentrations from studies of reproductive,
developmental, prenatal and neonatal outcomes and long-term exposure.
Reference
Location
Mean Annual
Concentration (jjg/m3)
Upper Percentile
Concentrations (jjg/m3)
PM2.5
Basu et al. (2004, 087896)
CA
Range of means across sites: 14.5-18.2
Avg of means across sites: 16.2
Max: 26.3-34.1
Bell et al. (2007,091059)
CT&MA
22.3

Braueret al. (2008,156292)
Vancouver, Canada
5.3
Max: 37.0
Huvnhet al. (2006, 091240)
CA
Range of means across trimesters: 17.5-18.8
Avg of means across trimesters: 18.2

Jalaludin et al. (2007,156601)
Sydney, Australia
9.0

Liu (2007, 090429)
Multicity, Canada
12.2
75th: 15
Loomis et al. (1999, 087288)
Mexico City
27.4
Max: 85
Mannes et al. (2005, 087895)
Sydney, Australia
9.4
75th: 11.2; Max: 82.1
Parker et al. (2005, 087462)
CA
15.4

Ritz et al. (2007, 096146)
Los Angeles, CA
20.0

Wilhelm and Ritz (2005, 088668)
Los Angeles, CA
21.0
Max: 38.9-48.5
Woodruff et al. (2006, 088758)
CA
19.2"
75th: 22.7
Woodruff et al. (2008, 098386)
U.S.
Range of means across effects: 14.5-14.9°
Avg of means across effects: 14.8°
75th: 18.5-18.7
PM10-2.5
Parker etal. (2008.156013)
U.S.
13.2
75th: 17.5
PMm
Bell et al. (2007, 093256)
CT&MA
22.3

Braueret al. (2008,156292)
Vancouver, Canada
12.7
Max: 35.4
Chenet al. (2002, 024945)'
Washoe County, NV
31.53
75th: 39.35; Max: 157.32
Gilboaetal. (2005,087892)
TX
23.8"
75th: 29
Ha et al. (2003, 042552)
Seoul, South Korea
69.2
75th: 87.7; Max: 245.4
Hansen et al. (2006, 089818)
Brisbane, Australia
19.6
Max: 171.7
Hansen et al. (2007, 090703)
Brisbane, Australia
19.6
75th: 22.7; Max: 171.7
Jalaludin et al. (2007,156601)
Sydney, Australia
16.3

Kim etal. (2007,156642)
Seoul, Korea
Range of means across time: 88.7-89.7
Avg of means across time: 89.2

Lee et al. (2003, 043202)
Seoul, Korea
71.1
75th: 89.3; Max: 236.9
Leem et al. (2006, 089828)
Incheon, Korea
53.8"
75th: 64.6; Max: 106.39
Lipfert et al. (2000, 004103)
U.S.
33.1
Max: 59
Maisonet et al. (2001, 016624)
NE U.S.
31.0"
75th: 36.1; Max: 46.5
Mannes et al. (2005, 087895)
Sydney, Australia
16.8
75th: 19.9; Max: 104.0
Pereira et al. (1998, 007264)
Sao Paulo, Brazil
65.04
Max: 192.8
Ritz et al. (2000, 012068)
CA
49.3
Max: 178.8
Ritz et al. (2006, 089819)
CA
46.3
Max: 83.5
Rogers and Dunlop (2006, 091232)
GA
3.75
75th: 15.07
Romieu et al. (2004, 093074)
Ciudad Juarez, Mexico
33.0-45.9

July 2009

7-65
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Reference
Location
Mean Annual
Concentration (ng/m3)
Upper Percentile
Concentrations (ng/m3)
Sagivet al. (2005, 087468)
PA
Range of means across time: 25.3-27.1
Avg of means across time: 26.2
Max: 68.9-156.3
Salam et al. (2005, 087885)
CA
Range of means across trimesters: 45.4-46.6
Avg of means across trimesters: 45.8

Suh et al. (2008,192077)
Seoul, Korea
Range of means across trimesters: 54.6-61.1
Avg of means across trimesters: 58.27
75th: 62.8-67.8
Max: 85.1-107.36
Tsai et al. (2006, 090709)
Kaohsiung, Taiwan
81.5
75th: 111.5; Max: 232.0
Wilhelm and Ritz (2005, 088668)
Los Angeles, CA
38.1
Max: 74.6-103.7
Woodruff et al. (2008, 098386)
U.S.
Range of means across effects: 28.6-29.8°
Avg of means across effects: 29.1°
75th: 33.8-36.5
Yang et al. (2006, 090760)
Taipei, Taiwan
53.2
75th: 64.9; Max: 234.9
8Median concentration
In contrast, Bell et al. (2007, 093256) reported positive associations for both PM2.5 and
PM10 with birth weight in a study of births (n = 358,504) in Connecticut and Massachusetts
(1999-2002). Birth data indicated county, not street address or ZIP code, so women were
assigned exposure based on county residence at delivery. The difference in birth weight per
10 pg/m3 associated with PM2.5 was -66.8 (95% CP 77.7 to -55.9) gm. For PM10 it was -11.1
(95% CP -15.0 to -7.2) gm. The increased risk for low birth weight was OR = 1.054 (95% CP
1.022-1.087) for PM2.5 and OR = 1.027 (95% CP 0.991-1.064) for PM10, based on average
exposure during pregnancy. Reductions in birth weight were also associated with third
trimester exposure to PM10 and second and third trimester exposure to PM2.5. Comparing
this study to Maisonet et al. (2001, 016624), a larger sample size was able to detect a small
increase in risk. In addition, birth weight was reduced more by exposure to PM2.5 than by
exposure to PM10. Measured PM2.5 concentrations were not available in the earlier study.
The Children's Health Study is a population based cohort of children living in 12
southern California communities, selected on the basis of differing levels of air pollution
(Salam et al., 2005, 087885). as previously discussed in Section 7.3. The children in grades
4, 7 and 10 were recruited through schools. A subset of this cohort (n = 6,259) were born in
California from 1975-1987. Of these, birth certificates were located for 4,842, including
3,901 infants born at term and 72 cases of low birth weight at term. Using the mother's ZIP
code at the time of birth, exposure was determined by inverse distance weighting of up to 3
PM10 monitors within 50 km of the ZIP code centroid. If there was a PM10 monitor within 5
km of the ZIP code centroid (40% of data), exposure from that monitor was used. Exposure
was calculated for the entire pregnancy, and for each trimester of pregnancy. A 10 pg/m3
increase in PM10 during the third trimester reduced mean birth weight -10.9 g (95% CP
-21.1 to -0.6) in single pollutant models, but became non-significant in multipollutant
models controlling for the effects of O3. Increased risks of low birth weight (<2500 gm) were
not statistically significant (OR = 1.3 [95% CP 0.9-1.9]). A strength of this study was the
cohort data available included information on SES and smoking during pregnancy. A
July 2009
7-66
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
limitation is the assignment of exposure based on monitoring stations up to 50 km distant!
this may have introduced significant exposure misclassification obscuring some
associations.
Parker et al. examined births in California within 5 miles of a monitoring station
(n = 18,247) (Parker et al., 2005, 087462). Only infants born at 40 wk gestation were
included. Thus all infants were the same gestational age, and had been exposed in the same
year. Exposure to PM2.5 in quartiles (<11.9, 11.9-13.9, 14.0-18.4, >18.4) was associated with
decrements in birth weight. Infants exposed to >13.9 pg/m3 experienced reductions in birth
weight (third quartile -13.7 g (95% CP -34.2 to 6.9), fourth quartile -36.1 g (95% CP -55.8 to
-16.5). These are larger reductions than have been seen in some other studies. However,
this study reduced misclassification by including only women living within 5 miles of a
monitoring station, and only included births at 40 wk gestation. Reducing misclassification
should lead to a stronger association, if the association is causal.
The effects of spatial variation in exposure were also investigated by Wilhelm and
Ritz (2005, 088668). Their study included all women living in ZIP codes where 60% of the
ZIP code was within two miles of a monitoring station in the Southern California Basin,
and women with known addresses in Los Angeles County within 4 miles of a monitoring
station. Exposure to average PM10 in the third trimester was analyzed for increased risk of
low birth weight at term (> 37~wk gestation). Analysis at the ZIP code level did not detect
increased risk (per 10 pg/m3 PM10, OR = 1.03 [95% CP 0.97-1.09]). However the analysis
based on geocoded addresses indicated that increasing exposure to PM10 was associated
with increased risk of low birth weight for women living within 1 mile of the station where
PM10 was measured. For these women (n = 247 cases, 10,981 non-cases), each 10 pg/m3
increase in PM10 was associated with a 22% increase in risk of term low birth weight
(OR = 1.22 [95% CP 1.05-1.41]). In the categorical analysis, exposure to PM10 >44.4 pg/m3
was associated with a 48% increase in risk (OR = 1.48 [95% CP 1.00-2.19]). Increased risk
of low birth weight also was associated with exposure to CO in single pollutant models.
However, when multipollutant models were considered, the effects of CO were attenuated
but the effects of PM10 increased. Controlling for CO, NO2, and O3, each 10 pg/m3 increase
in exposure to PM10 increased risk of low birth weight 36% (OR = 1.36 [95% CP 1.12-1.65]).
Spatial variation in PM2.5 exposure was investigated by Basu et al. (2004, 087896).
They included only mothers who lived within 5 miles of a PM2.5 monitor and within a
California county with at least one monitor. To minimize potential confounding, they
included only white (n = 8597) or Hispanic (n = 8114) women, who were married, between
20-30 years of age, completed at least high school and were having their first child.
Consistently, PM2.5 exposure measured by the county monitor was more strongly associated
with reductions in birth weight than exposure measured by the neighborhood monitor. The
results were replicated in both the white and the Hispanic samples. Reductions in birth
weight ranged from 15.2 g to 43.5 g per 10 pg/m3 increase in PM2.5.
July 2009
7-67
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
In the remaining U.S. study, Chen et al. (2002, 024945) analyzed 33,859 birth
certificates of residents of Washoe County in northern Nevada (1991-1999). There were four
sites monitoring PMio during the study period, it appears (not stated) that exposure was
averaged over the county. A 10 pg/m3 increase in exposure to PMio during the third
trimester of pregnancy was associated with an 11 gm reduction in birth weight (95% CP
¦2.3 to -19.8). Effects on risk of low birth weight were not significant. For exposure in the
third trimester of 19.77 to 44.74 pg/m3 compared to <19.74 pg/m3 the odds ratio for low
birth weight was 1.05 (95% CP 0.81-1.36). Comparing exposure >44.74 to the same
reference category, the odds ratio was 1.10 (95% CP 0.71-1.71). Misclassification of exposure
may have occurred when exposure was averaged over a large geographic area (16,968 km2).
Recent international studies investigating effects of particles on low birth weight
include one in Munich (Slama et al., 2007, 093216), two in Canada (Brauer et al., 2008,
156292; Dugandzic RDodds et al., 2006, 088681), two in Australia (Hansen et al., 2007,
090703; Mannes et al., 2005, 087895), two in Taiwan (Lin et al., 2004, 089827; Yang et al.,
2003, 087886) one in Korea (Ha et al., 2003, 042552) and two in Sao Paulo, Brazil (Gouveia
et al., 2004, 055613; Medeiros and Gouveia, 2005, 089824). The majority of these studies
found that PM concentrations were associated with low birth weight, though two studies
(Hansen et al., 2007, 090703; Lin et al., 2004, 089827) found no associations. The effect
estimates were similar in magnitude to those reported in the U.S. studies.
Issues in Interpreting Results of Low Birth Weight Studies
Studies included subjects at distances from monitoring stations varying from as close
as 1 mile or 2 km, to as far as 50 km or the size of the county. However, studies that only
included subjects living within a short distance (l mile, 2 km) of the monitoring station
(thus likely reducing exposure measurement error) were more likely to find that PM
exposure was associated with increased risk of low birth weight. However, Basu et al.
(2004, 087896) reported a stronger association between PM2.5 exposure and birth weight
when exposure was estimated based on the county monitor, rather than the monitor within
5 miles of the residence. They suggest that county level exposure may be more
representative of where women spend their time, including not only home, but also other
time spent away from home. Other pollutants also appeared to influence the risk associated
with particle exposure. In one study, exposure to PMio in a single pollutant model reduced
birth weight by 11 g, but became non-significant in multipollutant models with O3 (Salam et
al., 2005, 087885). In another study the risk associated with PMio exposure increased from
22% to 36% when other pollutants were included in the model (Wilhelm and Ritz, 2005,
088668). All but one study in the U.S. found some association between particle exposure
and reduced birth weight (Maisonet et al., 2001, 016624). The results of international
studies were inconsistent. This might be related to the chemical composition of particles in
the U.S., or to differences in the pollutant mixture. Studies with null results must be
July 2009
7-68
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
interpreted with caution when the comparison groups have significant exposure. This was
certainly the situation in studies in Taiwan and Korea (Lee et al., 2003, 043202; Lin et al.,
2004, 089827; Yang et al., 2003, 087886). Differences in geographical locations, study
samples and linkage decisions may contribute to the diverse findings in the literature on
the association between PM and birthweight, even within the U.S. (Parker and Woodruff,
2008, 156846).
Preterm Birth
A potential association of exposure to airborne particles and preterm birth has been
investigated in numerous epidemiologic studies, including some conducted in the U.S. and
others in foreign countries. Three U.S. studies have been carried out by the same group of
investigators in California.
A natural experiment occurred when an open-hearth steel mill in Utah Valley was
closed from August 1986 through September 1987. Parker et al. (2008, 156013) compared
birth outcomes for Utah mothers within and outside of the Utah Valley, before, during, and
after the mill closure. They report that mothers who were pregnant around the time of the
closure of the mill were less likely to deliver prematurely than mothers who were pregnant
before or after. The strongest effect estimates were observed for exposure during the second
trimester (14% decrease in risk of preterm birth during mill closure). Preterm birth outside
of the Utah Valley did not change during the time of the mill closure.
In 2000, Ritz et al. (2000, 012068) published the first study investigating the
association of preterm birth with PM in the U.S. The study population was women living in
the southern California Basin. There were 8 monitoring stations measuring PMio every
sixth day during the study period. Birth certificates (1989-1993) were analyzed for women
living in zip codes within 2 miles of a monitoring station. Women with multiple gestations,
chronic disease prior to pregnancy and women who delivered by cesarean section were
excluded resulting in a study population of 48,904 women. The risk of preterm birth
increased by 4% (RR = 1.04 [95% CP 1.02-1.6]) per 10 pg/m3 increase in PMio averaged in
the 6 wk before birth. Exposure to PMio in the first month of pregnancy resulted in a 3%
increase in risk (RR = 1.03 [95% CP 1.01-1.05]). These results were robust in multipollutant
models.
Wilhelm and Ritz (2005, 088668) reinvestigated this association among women in the
same area in 2005, when air pollution had declined from a mean level near 50 pg/m3 to a
mean level near 40 pg/m3. Birth certificate data from 1994-2000 was analyzed for women
living in ZIP codes within 2 miles of a monitoring station, or with addresses within 5 miles
of the monitoring station. No significant effects of exposure to PMio were reported.
Exposure to PM2.5 6 wk before birth resulted in an increase in preterm birth (RR = 1.19
[95% CP 1.02-1.40]) for the highest quartile of exposure (PM2.5 >24.3 pg/m3). Using a
July 2009
7-69
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
continuous measure of PM2.5, there was a 10% increase in risk for each 10 pg/m3 increase in
PM2.5 (RR = 1.10 [95% CI: 1.00-1.21]).
There have been two major criticisms of air pollution studies using birth certificate
data. First, that birth certificates only indicate the address at birth and the exposure of
women who moved during pregnancy may be misclassified! second, that information about
some important confounders may not be available (e.g., smoking). To obtain more precise
information about these variables, Ritz et al. (2007, 096146) conducted a case control study
nested within a cohort of birth certificates (Jan 2003-Dec 2003) in Los Angeles County.
Births to women residing in ZIP codes (n = 24) close to monitoring stations or major
population centers or roadways (n = 87) were eligible (n = 58,316 births). All cases of low
birth weight or preterm birth and an equal number of randomly sampled controls in the 24
zip codes close to monitors were selected. In the other 87 ZIP codes, 30% of cases and an
equal number of controls were randomly sampled. Of 6,374 women selected for the case
control study, 2,543 (40%) were interviewed. The association of preterm birth with exposure
to PM2.5 differed between women responding to the survey and women who did not respond.
Among responders, exposure to each 10 pg/m3 increase in PM2.5 concentration in the first
trimester increased risk to preterm birth by 23% (RR = 1.23 [95% CP 1.02-1.48]). There was
no increase in risk among non-responders (RR = 0.95 [95% CP 0.82-1.10]), or in the entire
birth cohort (RR = 1.00 [95% CP 0.94-1.07]).
An additional case control study of preterm birth and PM2.5 exposure Huynh et al.
(2006, 091240) used California birth certificate data. Singleton preterm infants (24-36-wk
gestation) born in California (1999-2000) whose mothers lived within 5 miles of a PM2.5
monitor were eligible. Each of these 10,673 preterm infants were matched to three term
(39-44-wk gestation) controls (having a last menstrual period within 2 wk of the case
infant), resulting in a study population of 42,692. Controlling for maternal race/ethnicity,
education, marital status, parity and CO exposure, exposure to PM2.5 >17.7 pg/m3 increased
the risk of preterm birth by 14% (OR = 1.14 [95% CP 1.07-1.23]). Averaging PM2.5 exposure
over the first month of pregnancy, the last 2 wk before birth, or the entire pregnancy did not
substantially change the risk estimate.
Two additional studies of preterm birth and exposure to particulate air pollution have
been conducted in the U.S. Each has used a unique methodology. Sagiv et al. (2005, 087468)
used time series to analyze births in four Pennsylvania counties between January 1997 and
December 2001. In this analysis, exposure to PM10 is compared to the rate of preterm births
each day. Both acute exposure (on the day of birth) and longer term exposure (average
exposure for the preceding six weeks) were considered in the analysis. An advantage of this
analysis is that days, rather than individuals are compared, so confounding by individual
risk factors is minimized. For exposure averaged over the 6 wk prior to birth, there was a
non-significant increase in risk (RR = 1.07 [95% CP 0.98-1.18]); for acute exposure with a 2
July 2009
7-70
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
day lag (RR = 1.10 [95% CI: 1.00-1.21]) and 5 day lag (RR = 1.07 95% CI: 0.98-1.18]) results
were marginal.
Rogers and Dunlop (2006, 091232) examined exposure to particles and risk of delivery
of an infant weighing less than 1500 g (all of which were preterm) from 24 counties in
Georgia. The study included 69 preterm small for gestational age (SGA) infants, 59 preterm
appropriate for gestational age (AGA) infants and 197 term AGAcontrols. Exposure was
estimated using an environmental transport model that considered PMio emissions from 32
geographically located industrial point sources, meteorological factors, and geographic
location of the birth home. Exposure was categorized by quartiles. Comparing women who
delivered a preterm AGA infant to those who delivered a term AGA infant, exposure to
PMio>15.07 pg/m3 tripled the risk (OR = 3.68 [95% CI: 1.44-9.44]).
Brauer et al. (2008, 156292) evaluated the impacts of PM2.5 on preterm birth using
spatiotemporal exposure metrics in Vancouver, Canada. The authors found similar results
when they used a land-use regression model or inverse distance weighting as the exposure
metric. For preterm births <37 wk, they reported an OR of 1.06 (95% CI: 1.01-1.11), and for
preterm births <35 wk the OR increased to 1.12 (95% CI: 1.02-1.24). There were no
consistent trends for early or late gestational period to be more strongly associated with
preterm births.
Suh et al. (2008, 192077) conducted a study to determine if the effects of exposure to
PM10 during pregnancy on preterm delivery are modified by maternal polymorphisms in
metabolic genes. They analyzed the effects of the gene-environment interaction between the
GSTM1, GSTT1, CYPlal-T6235C and -1462V polymorphisms and exposure to PM10 during
pregnancy on preterm birth in a case-control study in Seoul, Korea. PM10 concentration >
75th percentile alone was significant in the third trimester fo pregnancy (OR = 2.33 [95%
CI: 1.33-4.80]), but not in the first or second trimester. The risk of preterm delivery
conferred by the GSTM1 null genotype was increased, and the highest risk was found
during the third trimester of pregnancy (OR = 2.58 [95% CI: 1.34-4.97]). There were no
statistical associations with the GSTT1 or CYP1A1 genotypes. When the gene-environment
interaction was analyzed, the risk for preterm birth was significantly higher for women who
carried the GSTM1 null genotype and were exposed to high levels of PM10 (> 75th
percentile) than for those who carried the GSTM1 positive genotype but were only exposed
to low levels of PM10 (<75th percentile) during the third trimester of pregnancy (OR = 6.22,
95% CI: 2.14-18.08).
In Incheon, Korea, Leem et al. (2006, 089828) estimated PM10 exposure spatially as
well as temporally. Exposure was based on 26 monitors and kriging was used to determine
exposure for 120 dongs (administrative districts, mean area 7.82 km2, median area 1.42
km3). The sample included 52,113 births, from 2001-2002. PM10 was very weakly correlated
with other pollutants. Exposure was compared in quartiles for the first and third trimester
of pregnancy. In the first trimester, relative risks for the second, third and fourth quartiles
July 2009
7-71
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
were RR = 1.14 (95% CI: 0.97-1.34), RR = 1.07 (95% CI: 0.94-1.37), and RR = 1.24 (95% CI:
1.09-1.41), respectively. Exposure to PMio in quartile one (reference group) was 26.9 -
45.9 pg/m3; fourth quartile exposure equaled 64.6-106.4 pg/m3. The p-value for trend was
0.02. Exposure in the third trimester was not related to preterm birth, however no
information was provided to determine how exposure in the third trimester was adjusted
for women who delivered preterm.
Two studies investigating risks of preterm birth related to particle exposure have
been reported from Australia. In Brisbane, Hansen et al. (2006, 089818) studied 28,200
births (2000-2003) in an area of low PMio concentrations. Exposure to an interquartile
range increase in PMio exposure in the first trimester resulted in a 15% increased risk of
preterm birth (OR = 1.15 [95% CI: 1.06-1.25]). This result was strongly influenced by the
effect of PMio exposure in the first month of pregnancy (OR = 1.19 [95% CI: 1.13-1.26]).
PMio was correlated with ozone r = (0.77) in this study and ozone also increased risk in the
first trimester. No effects were associated with exposure to PMio in the third trimester.
In Sydney, associations between exposure to particles and preterm birth varied by
season. Jalaludin et al. (2007, 156601) obtained information on all births in metropolitan
Sydney (1998-2000). Exposure to PM2.5 in the three months preceding birth was associated
with an increased risk of preterm birth (OR = 1.11 [95% CI: 1.04-1.19]). Additional effects
were dependent on season of conception. Both PMio (OR = 1.3 95% CI: 1.2-1.5]) and PM2.5
(OR = 1.4 [95% CI: 1.3-1.6]) were associated with increased risk for conceptions in the
winter. Conceptions in summer were associated with reductions in risk (PMio OR = 0.91
[95% CI: 0.88-0.93]) (PM2.5 OR = 0.87 [95% CI: 0.84-0.92]). Due to both positive and
negative findings, the authors recommend caution in interpreting their results.
Issues in Analyzing Environmental Exposures and Preterm Birth
A major issue in studying environmental exposures and preterm birth is selecting the
relevant exposure period, since the biological mechanisms leading to preterm birth and the
critical periods of vulnerability are poorly understood (Bobak, 2000, 011448). Exposures
proximate to the birth may be most relevant if exposure causes an acute effect. However,
exposure occurring in early gestation might affect placentation, with results observable
later in pregnancy, or cumulative exposure during pregnancy may be the most important
determinate. The studies reviewed have dealt with this issue in different ways. Many have
considered several exposure metrics based on different periods of exposure.
Often the time periods used are the first month (or first trimester) of pregnancy and
the last month (or six weeks) prior to delivery. Using a time interval prior to delivery
introduces an additional problem since cases and controls are not in the same stage of
development when they are compared. For example, a preterm infant delivered at 36 wk, is
a 32 wk fetus 4 wk prior to birth, while an infant born at term (40 wk) is a 36 wk fetus 4 wk
prior to birth. Only one study (Huynh et al., 2006, 091210) adjusted for this in the design.
July 2009
7-72
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Many of these studies compare exposure in quartiles, using the lowest quartile as the
reference (or control) group. No studies use a truly unexposed control group. If exposure in
the lowest quartile confers risk, than it may be difficult to demonstrate additional risk
associated with a higher quartile. Thus negative studies must be interpreted with caution.
Preterm birth occurs both naturally (idiopathic preterm), and as a result of medical
intervention (iatrogenic preterm). Ritz et al. (2000, 012068S 2007, 096146) excluded all
births by Cesarean section, to limit their studies to idiopathic preterm. No other studies
attempted to distinguish the type of preterm birth, although PM exposure maybe associated
with only one type. This is a source of potential effect misclassification.
Growth Restriction
Low birth weight has often been used as an outcome measure because it is easily
available and accurately recorded on birth certificates. However, low birth weight may
result from either short gestation, or inadequate growth in utero. Most of the studies
investigating air pollution exposure and low birth weight, limited their analysis to term
infants to focus on inadequate growth. A number of studies were identified that specifically
addressed growth restriction in utero by identifying infants who failed to meet specific
growth standards. Usually these infants had birth weights less than the 10th percentile for
gestational age, using an external standard. Many of these studies have been previously
discussed, since they also examined other reproductive outcomes (low birth weight or
preterm delivery).
Three studies in the U.S. examined intrauterine growth. A recent study (Rich et al.,
2009, 180122) investigated very small for gestational age (defined as a fetal growth ratio
<0.75), small for gestational age (defined as > 75 and <85) and "reference" births (> 85) to
women residing in New Jersey and mean air pollutant concentrations during the first,
second and third trimesters. They reported an increased risk of SGA associated with first
and third trimester PM2.5 concentrations (1.116 [95% CP 1.012, 1.232], and 1.106 [1.008-
1.212], per 10 |ug/m3 PM2.5, respectively). Parker et al. (2005, 087462) reported a positive
association between exposure to PM2.5. Since this study only included singleton live births
at 40-wk gestation, birth weights less than 2872 g for girls and 2986 g for boys were
designated SGA, based on births in California. Infants exposed to the highest quartile PM2.5
(>18.4 pg/m3) compared to the lowest quartile PM2.5 (<11.9 pg/m3) were 23% more likely to
be small for gestational age (OR = 1.23 [95% CP 1.03-1.50]). Very similar results were found
for exposure in each of the three trimesters respectively (OR = 1.26 [95% CP 1.04-1.51],
OR = 1.24 [95% CP 1.04-1.49], OR = 1.21 [95% CP 1.02-1.43]). These results controlled for
exposure to CO, which did not increase risk for SGA.
In contrast, Salam et al. (2005, 087885) found no association between exposure to
PM10 and intrauterine growth retardation (IUGR) in the California Children's Health
Study. IUGR was defined as less than the 15th percentile of predicted birth weight based on
July 2009
7-73
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
gestational age and sex in term infants. Apparently no external standard was used since
15% of infants in the study were designated as IUGR. An IQR increase in PMio exposure
was not significantly associated with IUGR for the whole pregnancy (OR =1.1 [95% CP 0.9-
1.3]) or for any specific trimester. Differences between this study and the study by Parker et
al. (2005, 087462) include measurement of PMio vs. PM2.5, a less stringent definition of
IUGR, and exposures determined by monitors located much farther away from the subjects'
residences (up to 50 km vs. within 5 miles). All of these factors could lead to
misclassification.
Two studies investigating particle exposure and SGA were conducted in Australia,
with differing results (Hansen et al., 2007, 090703; Mannes et al., 2005, 087895). Mannes et
al. (2005, 087895) defined SGA as birth weight less than two standard deviations below the
national mean birth weight for gestational age. In this study there was a statistically
significant effect of exposure to both PMio (OR = 1.10 [95% CP 1.00-1.48], per 10 pg/m3
increase) and PM2.5 (OR = 1.34 [95% CP 1.10-1.63], per 10 pg/m3 increase) for exposure
during the second trimester. When analysis was restricted to births within 5 km of the
monitoring station, the association for PMio became slightly stronger (OR = 1.22 [95% CP
1.10-1.34]). Exposure during other trimesters of pregnancy was not associated with IUGR.
In Brisbane, Hansen et al. (2007, 090703) examined head circumference (HC), crown
heel length (CHL) and risk of SGA, defined as less than the tenth percentile of weight for
gestational age and gender based on an Australian national standard. There was no
consistent relationship between PMio exposure and SGA, HC or CHL in any trimester of
pregnancy. PMio exposure was determined by averaging values from the five monitoring
stations. Due to the sample size and limited number of monitoring stations, it was not
possible to analyze the data for women living within 5 km of a monitoring station, as was
done in Sydney.
In Canada, Liu et al. (2007, 090429) investigated the effect of PM2.5 exposure on fetal
growth in three cities, Calgary, Edmonton and Montreal. Intrauterine growth retardation
(IUGR) was defined as birth weight below the tenth percentile, by sex and gestational week
(37-42) for all singleton live births in Canada between 1986 and 2000. Models were
adjusted for maternal age, parity, infant sex, season of birth, city of residence, and year of
birth. A 10 pg/m3 increase in PM2.5 was associated with an increased risk for IUGR
(OR = 1.07 [95% CP 1.03-1.10]) in the first trimester, and similar risks were associated with
exposure in the second or third trimesters. The effect of PM2.5 was reduced in
multipollutant models including CO and NO2.
Brauer et al. (2008, 156292) observed consistent increased risks of SGA for PM2.5,
PMio, NO2, NO and CO in Vancouver, Canada (20% increase in risk in PM2.5 and PMio per
10 pg/m3 increase). The effects were similar for exposure estimates based on nearest
monitor, inverse distance weighting, and land-use regression modeling. ORs for early or
July 2009
7-74
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
late pregnancy exposure windows were remarkably similar to those for the full duration of
pregnancy
Birth Defects
Four recent studies examined PM and birth defects. The Seoul, Korea study discussed
above also considered congenital anomalies, defined as a defect in the child's body structure
(Kim et al., 2007, 156642). PMio levels were associated with higher risk of birth defects for
the second trimester, with a 16% (95% CP 0-34) increase in risk per 10 jug/m3 in PMio.
Two U.S. studies examined air pollution and risk of birth defects. Data were collected
from the California Birth Defects Monitoring Program for four counties in Southern
California (Los Angeles, Riverside, San Bernardino, and Orange) for the period 1987 to
1993, although each county included a subset of this period (Ritz et al., 2002, 023227).
Cases (i.e., infants with birth defects) were identified as live birth infants and fetal deaths
from 2()-wk gestation to 1 year post birth, with isolated, multiple, syndrome, or
chromosomal cardiac or orofacial cleft defects. Cases were restricted to those with registry
data for gestational age and residence zip code, and those with residences <10 miles from
an air pollution monitor. Six types of categories were included: aortic defects! atrium and
atrium septum defects! endocrinal and mitral value defects! pulmonary artery and valve
defects! conotruncal defects! and ventricular septal defects not part of the conotruncal
category. PMio measurements were available every six days. While results indicated
increased risk of birth defects for higher levels of CO or O3, the authors determined that
results for PMio were inconclusive, finding no consistent trend of effect after adjustment for
CO and O3.
The other U.S. study examined birth defects through a case-control design in seven
Texas counties for the period 1997 to 2000 (Gilboa et al., 2005, 087892). Births were
excluded for parents <18 years and several non-air pollution risk factors known to be
associated with birth defects (e.g., maternal diabetes, holoprosencephaly in addition to oral
clef). Comparison of the highest (> 29.0 |u.g/m3) and lowest (<19.521 jug/m3) quartiles of PMio
for exposure defined as the third to eighth weeks of pregnancy generated an OR of 2.27
(95% CP 1.43-3.60) for risk of isolated artrial septal defects and 1.26 (95% CP 1.03-1.55) for
individual artrial septal defects. Including other pollutants (CO, NO2, O3, SO2) in the model
did not greatly alter results! numerical results for co-pollutant analysis were not provided.
Strong evidence was not observed for a relationship between PMio and the other birth
defect categories. Review articles have concluded that the scientific literature is not
sufficient to conclude a relationship between air pollution and birth defects (Sram et al.,
2005, 087442).
A recent study of oral clefts conducted in Taiwan found no association between this
birth defect and concentrations of PMio during the first or second gestational month
July 2009
7-75
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
(Hwang et al., 2006, 088971). This population-based case-control study included 653 cases
and a random sample of 6,530 controls born in Taiwan between 2001 and 2003.
Infant Mortality
Many studies have identified strong associations between exposure to particles and
increased risk of mortality in adults or the general population, including for short- and
long-term exposure (see Section 7.6). Less evidence is available for the potential impact on
infant mortality, although studies have been conducted in several countries. The results of
these infant mortality studies are presented here with the other reproductive and
developmental outcomes because it is likely that in vitro exposures contribute to this
outcome. Both long-term and short-term exposure studies of infant mortality are included
in this section. Results on PM and infant mortality includes a range of findings, with some
studies finding associations and many non-statistically significant or null effects. Yet, more
consistency is observed when results are divided into the type of health outcome based on
the age of infant and cause of death.
An important question regarding the association between PM and infant mortality is
the critical window of exposure during development for which infants are susceptible.
Several age structures have been explored: infants (<1 year); neonatal (<1 month); and
postneonatal (l month to 1 year). Within these various age categories, multiple causes of
deaths have been investigated, particularly total deaths and respiratory-related deaths.
The studies reflect a variety of study designs, particle size ranges, exposure periods,
regions, and adjustment for confounders.
Stillbirth
Only one study of stillbirths and PM was identified. A prospective cohort of pregnant
women in Seoul, Korea from 2001 to 2004 was examined with respect to exposure to PMio
(Kim et al., 2007, 156642). Gestational age was estimated by the last menstrual period or
by ultrasound. Whereas many of the previously discussed studies of PM and pregnancy
outcomes were based on national registries, this study examined medical records and
gathered individual information through interviews on socio-economic condition, medical
history, pregnancy complications, smoking, second-hand smoke exposure, and alcohol use.
Mother's exposure to PMio was based on residence for each month of pregnancy, each
trimester defined as a three month period, and the six weeks prior to death. Exposure was
assigned by the nearest monitor. A 10 ug/m3 increase in PMio in the third trimester was
associated with an 8% (95% CP 2-14) increase in risk of stillbirth.
In Sao Paulo, Brazil, Poisson regression of stillbirth counts for the period 1991 and
1992 found that a 10 |ug/m3 increase in PMio was associated with a 0.8% increase in
stillbirth rates (Pereira et al., 1998, 007264). When other pollutants (NO2, SO2, CO, O3)
were included simultaneously in the model, the association did not remain. Stillbirths were
defined as fetal loss at >28 wk of pregnancy age, weight >1000 g, or length of fetus >35 cm.
July 2009
7-76
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
As discussed below, there exists some limited evidence for a link between PM and
stillbirth.
Infant Mortality and Infant Respiratory Mortality, < lYear
A literature search did not reveal new studies on PM and infant mortality (<1 year)
since the previous PM AQCD. Previously conducted studies include a case-control study
that reported associations between infant mortality and TSP levels over the period between
birth and death for infants in the Czech Republic (Bobak and Leon, 1999, 007678). An
ecological study evaluated U.S. PMio data for the year 1990 using long-term pollution levels
in 180 U.S. counties (Lipfert et al., 2000, 004103). The authors found a 9.64% (95% CP 4.60-
14.9) increase in risk of infant mortality for non-low birth weight infants per 10 ug/m3
increase in PMio, a 13.4% (95% CP -10.3 to 43.5%) increase in non-low birth weight
respiratory-disease related deaths (ICD-9 460-519) and a 19.5% (95% CP 0.07-42.8)
increase in all non-low birth weight respiratory-related infant deaths (ICD 9 460-519, 769,
770).
Neonatal Mortality and Neonatal Respiratory Mortality, < 1 month
Studies on PM and neonatal mortality (<1 month) included a time-series analysis of
PMio for four years of data (1998-2000) for Sao Paulo, Brazil (Lin et al., 2004, 095787). The
analysis used daily counts of deaths from government registries and adjusted for temporal
trend, day of the week, weather, and holidays. Findings indicated that a 10 ug/m3 increase
in PMio was associated with a 1.71% (95% CP 0.31-3.32) increase in risk of neonatal death.
A case-crossover study of 11 years (1989-2000) in Southern California did not find an
association between PMio and neonatal deaths (Ritz et al., 2006, 089819). Quantitative
results were not provided. The authors considered adjustment for season, county, parity,
gender, prenatal care, and maternal age, education, and race/ethnicity. The overall levels of
PMio in these studies were similar.
These results add to previous work on PM and neonatal death, including studies
identifying higher risk of neonatal mortality with higher TSP in the Czech Republic in an
ecological analysis (Bobak and Leon, 1992, 044415) and case-crossover study (Bobak and
Leon, 1999, 007678), and a Poisson model study in Kagoshima City, Japan (Shinkura et al.,
1999, 156978). An ecological study evaluated U.S. PMio data for the year 1990 using long-
term pollution levels in 180 U.S. counties (Lipfert et al., 2000, 001103). Analysis considered
birth weight, sex, month of birth, location by state and county, prenatal care, and mother's
race, age, educational level, marital status, and smoking status. County-level variables
were included for socio-economic status, altitude, and climate. Results indicate a 13.1%
increase in neonatal mortality (95% CP 4.4-22.6) per 10 jug/m3 PMio for non-low birth
weight infants. Statistically significant associations were also observed considering all
infants or low birth weight infants. However, higher levels of SO2 were associated with
lower risk of infant mortality. When sulfate and an estimate of non-sulfate particles were
July 2009
7-77
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
included in the regression simultaneously, associations were observed with non-sulfate
particles and an inverse relationship with sulfate particles. Respiratory neonatal mortality
was not associated with higher TSP in the Czech Republic case-control study (Bobak and
Leon, 1999, 007678).
Postneonatal Mortality and Post-neonatal Respiratory Mortality, 1 Month-1 Year
Several studies have been conducted on PM and postneonatal mortality since the
previous PM AQCD, including three from the U.S., one from Mexico, and three from Asia.
Two case-control studies examined the risk of PM to postneonatal death in California.
Research focused on Southern California for the period 1989 to 2000 linked birth and death
certificates and considered PMio two months prior to death with adjustment for prenatal
care, gender, parity, county, season, and mother's age, race/ethnicity, and education (Ritz et
al., 2006, 089819). As previously noted, this study did not find an association between PMio
and neonatal mortality (<1 month), however an association was observed for postneonatal
mortality, with a 10 |ug/m3 increase in PMio associated with a 4% (95% CP 1-6) increase in
risk. The exposure period of 2 wk before death was also considered, producing effect
estimates of 5% (95% CP 1-10) for the same PMio increment. Even larger effect estimates
were observed for those who died at ages 4 to 12 months. When CO, NO2, and O3 were
simultaneously included with PMio in the model, the central estimate reduced to 2% for the
2-week exposure period and 4% for the 2-month exposure period, and both estimates lost
statistical significance. The other case-control study of California considered PM2.5 from
1999 to 2000 for infants born to mothers within five miles of a PM2.5 monitoring station
(Woodruff et al., 2006, 088758). Infants who died during the postneonatal period were
matched to infants with date of birth within two weeks and birth weight category. Exposure
was estimated from the time of birth to death. Models considered parity and maternal race,
education, age, and martial status. A 10 |ug/m3 increase in PM2.5 was associated with a 7%
(95% CP -7 to 24) increase in postneonatal death
County-level PMio and PM2.5 for the first two months of life for births in urban U.S.
counties (> 250,000 residents) from 1999 to 2002 were evaluated in relation to postneonatal
mortality with GEE models (Woodruff et al., 2008, 098386). Analyses were adjusted for
primiparity (first born), community-level poverty, region, month, year, and mother's race,
marital status, education, and age. Births were restricted to singleton births with
gestational age < 44 wk, same county of residence at birth and death, and non-missing data
on birth order, birth weight, and maternal race, education, and martial status. Higher
levels of either PM metric were associated with higher risk of postneonatal mortality, with
4% (95% CP -1 to 10) increase in mortality risk per 10 jug/m3 in PMio and 4% (95% CP -2
to 11) increase in mortality risk for the same increment of PM2.5. This work builds on a
previous study of 86 U.S. urban areas from 1989 to 1991, finding a 4% (95% CP 2-7)
increase in postneonatal mortality per 10 ug/m3 county-level PMio over the first two months
of life (Woodruff et al., 1997, 084271).
July 2009
7-78
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
In Ciudad Juarez, Mexico, a case-crossover approach was applied to data from 1997 to
2001 based on death certificates and the cumulative PMio for the day of death and previous
two days (Romieu et al., 2004, 093074). A case-crossover study of Kaohsiung, Taiwan from
1994-2000 compared the average of PMio on the day of death and two previous days to PMio
in control periods a week before and week after death (Tsai et al., 2006, 090709). A similar
approach was also applied to 1994 to 2000 data from Taipei, Taiwan, also using case-
crossover methods for the lag 0-2 PMio with referent periods the week before and after
death (Yang et al., 2006, 090760). In these case-crossover studies, season was addressed
through matching in the study design. A 10 ug/m3 increase in PMio was associated with a
2.0% (95% CP -2.8 to 7.0) increase in the Mexico study, a 0.59 (95% CP -15.0 to 18.8)
increase in postneonatal death in the Kaohsiung study, and a 1.02% (95% CP -13.2 to 17.6)
increase in the Taipei study. A study in Seoul, South Korea from 1995 to 1999 used time-
series approaches adjusted for temporal trend and weather, based on national death
registries excluding accidental deaths (Ha et al., 2003, 042552). A 10 |ug/m3 increase in PMio
was associated with a 3.14% (95% CP 2.16-4.14) increase in risk of death for postneonates.
A subset of the studies examining postneonatal mortality also considered the subset of
postneonatal deaths from respiratory causes. These include the time-series study in South
Korea, finding a 17.8% (95% CP 14.4 to 21.2) increase in respiratory-mortality per 10 jug/m3
increase in PMio (Ha et al., 2003, 042552) and the case-crossover study in Mexico, for which
the same increment in PMio was associated with a 1.5% (95% CP -14.1 to 13.0) decrease in
risk (Romieu et al., 2004, 093074). Both case-control California studies identified
associations, with a 5% (l, 10%) increase in risk in Southern California (Ritz et al., 2006,
089819) and 57.4% (95% CP 7.0-132) increase in California per 10 |ug/m3 PMio (Woodruff et
al., 2006, 088758). The U.S. study found this increment in PMio to be linked with a 16%
(95% CP 6.0-28.0) increase in respiratory postneonatal mortality, although effect estimates
for PM2.5 were not statistically significant (Woodruff et al., 2008, 098386). Earlier studies on
respiratory-related postneonatal mortality include the study of 86 U.S. urban areas, finding
statistically significant effects (Woodruff et al., 1997, 084271).
Sudden Infant Death Syndrome
Three studies examining the relationship between PM and sudden infant death
syndrome have been published from 2002 onward. These studies examined infant mortality
and were thereby discussed in this section previously. A case-control study over a 12-yr
period (1989 to 2000) matched 10 controls to deaths (cases) in Southern California (Ritz et
al., 2006, 089819). A 10 |u.g/m3 increase in PMio the two months prior to death was
associated with a 3% (95% CP -1 to 8) increased in SIDS. Adjusted for other pollutants (CO,
NO2, and O3), the effect estimate reduced to 1% (95% CP -5 to 7).
A case-control study, also based in California, found an OR of 1.008 (95% CP 1.006-
1.012) per 10 |u,g/m3 increase in PM2.5, considering a SIDS definition of ICD 10 R95
(Woodruff et al., 2006, 088758). Due to changes in SIDS diagnosis, another SIDS definition
July 2009
7-79
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
was explored for ICD 10 R99 in addition to ICD 10 R95. Under this SIDS definition, the
effect estimate changed to 1.03 (95% CI: 0.79-1.35). The authors also examined whether the
relationship between PM2.5 and SIDS differed by season, finding no significant difference.
PM10 and PM10-2.5 were not associated with risk of SIDS; numerical results were not
provided for these PM metrics. The third recent study of PM and SIDS examined U.S.
urban counties from 1999 to 2002 (Woodruff et al., 2008, 098386). Non-statistically
significant relationships were observed between SIDS and PM10 or PM2.5 in the first two
months of life.
These studies add to earlier work, such as a U.S. study that found higher risk of SIDS
with higher annual PM2.5 levels, including in a separate analysis of normal birth weight
infants (Lipfert et al., 2000, 001103). and a U.S. study identifying a 12% (95% CP 7-17)
increase in SIDS risk per 10 ug/ma in PM10 for the first two months of life for normal weight
births (Woodruff et al., 1997, 084271). A study based on Taiwan found higher SIDS risk
with lower visibility (Knobel et al., 1995, 155905), whereas a 12 city Canadian time-series
study identified no significant associations (Dales et al., 2004, 087342).
Deaths by SIDS were identified by different methods in the studies, partly due to
transition from ICD 9 to ICD 10, but also due to different choices within the research
design. Two studies examined multiple approaches (ICD 10 R95, ICD 10 R95 and R99)
(Woodruff et al., 2006, 088758; Woodruff et al., 2008, 098386), and other studies
investigated ICD 9 798.0 and ICD 10 R95 (Ritz and Wilhelm, 2008, 156914), ICD 9 798.0
(Woodruff et al., 1997, 084271). ICD 9 798.0 and 799.0 (Knobel et al., 1995, 155905), as well
as a sudden unexplained death of infant <1 year for which an autopsy did not identify a
specific cause of death (Dales et al., 2004, 087342). These variations in the definition of
health outcomes add to differences in populations and study designs.
Although some findings indicate a potential effect of PM on risk of SIDS, with the
strongest evidence perhaps from the case-control study in California (Woodruff et al., 2006,
088758), others do not find an effect or observe an uncertain association. For the
relationship between PM and SIDS, a 2004 review article concluded consistent evidence
exists compared to evidence for other infant mortality effects (Glinianaia et al., 2004,
087898), whereas other reviews found weaker or insufficient evidence (Heinrich and Slama,
2007, 156534). Another review concluded that the scientific literature on air pollution and
SIDS suggests an effect, but that further research is needed to draw a conclusion (Tong and
Colditz, 2004, 087883).
Comparisons across Studies and Key Issues
Comparison of results across studies can be challenging due to several issues,
including differences in methodologies, populations and study areas, pollution levels, and
the exposure timeframes used. Given the large variation in study designs, the methods to
address potential confounders vary. For example, weather and season were addressed in
July 2009
7-80
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
the case-control studies by matching, in the time-series study through non-linear functions
of temperature and temporal trend, and in the ecological study through county-level
variables. All studies included consideration of seasonality and weather. Researchers used
different definitions of respiratory-related deaths, including ICD 9 460-519 (Bobak and
Leon, 1999, 007678; Lipfert et al., 2000, 004103); ICD 9 460-519, 769-770 (Lipfert et al.,
2000, 004103); ICD 9 codes 460-519, 769, 770.4, 770.7, 770.8, 770.9, and ICD 10 J00-J98,
P22.0, P22.9, P27.1, P27.9, P28.0, P28.4, P28.5, and P28.9 (Ritz et al., 2006, 089819); and
ICD 9 460-519 and ICD 10 J00-J99 for any cause on death certificate (Romieu et al., 2004,
093074); ICD 10 J00-99 and P27.1 excluding J69.0 (Woodruff et al., 2006, 088758; Woodruff
et al., 2008, 098386); and ICD 9 460-519 (Woodruff et al., 1997, 084271).
Socioeconomic conditions were included at the individual level, typically maternal
education, in many studies (e.g.,Bobak and Leon, 1999, 007678; Ritz and Wilhelm, 2008,
156914; Ritz et al., 2006, 089819; Woodruff et al., 1997, 084271; Woodruff et al., 2006,
088758) and at the community-level in others (e.g.,Bobak and Leon, 1992, 044415; Penna
and Duchiade, 1991, 073325) or for both individual and community-level data (e.g.,Lipfert
et al., 2000, 001103). The time-series approach is unlikely to be confounded by
socioeconomic and other variables that do not exhibit day-to-day variation. Similarly, case-
crossover methods use each case as his/her own control, thereby negating the need for
individual-level confounders such as socioeconomic status (e.g.,Romieu et al., 2004, 093074;
Tsai et al., 2006, 090709; Yang et al., 2006, 090760). All studies published after 2001
incorporated individual-level socioeconomic data or were of case-crossover or time-series
design. One study specifically examined whether socioeconomic status modified the PM and
mortality relationship, dividing subjects into three socioeconomic strata based on the zip
code of residence at death (Romieu et al., 2004, 093074). This work, based in Mexico, found
that at lower socio-economic levels the association between PMio and postneonatal
mortality increased. Although the overall association showed higher risk of death with
higher PMio with statistical uncertainty, for the lowest socio-economic group, a 10 ug/ma
increment in cumulative PMio over the two days before death was associated with a 60%
(95% CP 3-149) increase in postneonatal death. A trend of higher effect for lower socio-
economic condition is observed in all three lag structures.
Studies differ in terms of the timeframe of pregnancy that was used to estimate
exposure. Exposure to PM for infant mortality (<1 year) was estimated as the levels
between birth and death (Bobak and Leon, 1999, 007678), annual community levels (Lipfert
et al., 2000, 004103; Penna and Duchiade, 1991, 073325) and the 3 to 5 days prior to death
(Loomis et al., 1999, 087288). For neonatal deaths, exposure timeframes considered were
the time between birth and death (Bobak and Leon, 1992, 044415; Bobak and Leon, 1999,
007678), annual levels (Bobak and Leon, 1999, 007678; Lipfert et al., 2000, 004103),
monthly levels (Shinkura et al., 1999, 156978), the same day concentrations (Lin et al.,
2004, 095787), and the two months or two weeks prior to death (Ritz et al., 2006, 089819).
July 2009
7-81
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Postneonatal mortality was associated with PM concentrations based on annual levels
(Bobak and Leon, 1992, 044415; Lipfert et al., 2000, 004103), between birth and death
(Bobak and Leon, 1999, 007678; Woodruff et al., 2006, 088758), two months before death
(Ritz et al., 2006, 089819), the first two months of life (Woodruff et al., 1997, 084271;
Woodruff et al., 2006, 088758), the day of death (Ha et al., 2003, 042552), and the average of
the same day as death and previous two days (Romieu et al., 2004, 093074; Tsai et al., 2006,
090709; Yang et al., 2006, 090760). Thus, no consistent window of exposure was identified
across the studies.
PMio concentrations were highest in South Korea (69.2 jug/m3) (Ha et al., 2003,
042552) and Taiwan (81.45 |ug/m3) (Tsai et al., 2006, 090709), and lowest in the U.S.
(29.1 |ug/m3) (Woodruff et al., 2008, 098386) and Japan (21.6 |u,g/m3) (Shinkura et al., 1999,
156978). All studies used community level exposure information based on ambient
monitors, as opposed to exposure measured at the individual level (e.g., subject's home) or
personal monitoring.
Given similar sources for multiple pollutants (e.g., traffic), disentangling the health
responses of co-pollutants is a challenge in the study of ambient air pollution. Several
studies examined multiple pollutants, most by estimating the effect of different pollutants
through several univariate models. Some studies noted the difficulty of separating particle
impacts from those of other pollutants, but noted stronger evidence for particles than other
pollutants (Bobak and Leon, 1999, 007678). A few studies applied co-pollutant models by
including multiple pollutants simultaneously in the same model. Effect estimates for the
relationship between PMio and neonatal deaths in Sao Paulo were reduced to a null effect
when SO2 was incorporated (Lin et al., 2004, 095787). Associations between PMio and
postneonatal mortality or respiratory postneonatal mortality remained but lost statistical
significance in a multiple pollutant model with CO, NO2, and O3 (Ritz et al., 2006, 089819).
July 2009
7-82
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Neonatal
Neonatal
Figure 7-5. Percent increase in postneonatal mortality per 10 yug/m3 in PMio, comparing risk for
total and respiratory mortality.
Panel a (left) provides central estimates; panel b (right) also adds the 95% intervals. The points reflect
central estimates and the lines the 95% intervals. Solid lines represent statistically
significant effect estimates; dashed lines represent non-statistically significant
estimates.1
Several review articles in recent years have examined whether exposure to PM affects
risk of infant mortality, generally concluding that more consistent evidence has been
observed for postneonatal mortality particularly from respiratory causes (Bobak and Leon,
1999, 007678; Heinrich and Slama, 2007, 156534; Lacasana et al., 2005, 155914; Sram et
al., 2005, 087442). In one review authors identified 14 studies on infant mortality and air
pollution and determined that studies on PM and infant mortality do not provide consistent
results, although more evidence was present for an association for some subsets of infant
mortality such as postneonatal respiratory-related mortality (Bobak and Leon, 1999,
007678). The relationship between PM and postneonatal respiratory mortality was
concluded to be causal in one review (Sram et al., 2005, 087442). and strong and consistent
in another (Heinrich and Slama, 2007, 156534). Meta-analysis using inverse-variance
weighting of PMio studies found that a 10 |ug/m3 increase in acute PMio exposure was
associated with 3.3% (95% CP 2.4-4.3) increase in risk of postneonatal mortality, whereas
the same increment of chronic PMio exposure was linked with a 4.8% (95% CP 2.2-7.2)
1 Studies included are Bobak and Leon (1999, 007678). Ha et al. (2003, 042552). Ritz et al. (2006, 089819). Romieu et al. (2004,
093074). Romieu et al. (2008, 156922). Woodruff et al. (1997, 084271). Woodruff et al. (2006, 088758). Findings from Bobak
and Leon (1999, 007678) were based on TSP and were converted to PMio estimates assuming PMio/TSP = 0.8 as per
summary data in the original article (Bobak and Leon, 1999, 007678). Findings from Woodruff et al. (1997, 084271) for
respiratory-related mortality were based on non-low birth weight infants. Results for Woodruff et al. (2006, 088758) were
based on PM2.5 and were converted to PMio assuming PM2.5/PM10 = 0.6.
July 2009
7-83
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
increase in postneonatal mortality and a 21.6% (95% CP 10.2-34.2) increase for respiratory
postneonatal mortality (Lacasana et al., 2005, 155914).
Studies that examined multiple outcomes and ages of death allow a direct comparison
based on the same study population and methodologies, thereby negating the concern that
inconsistent results are due to underlying variation in population, approaches, etc. In this
review, one study finds higher effect estimates for postneonatal mortality, for both total and
respiratory-related mortality, and the other study found higher effects for risk in the
neonatal period. Another study, based in Southern California identified no association for
neonatal effects (numerical results not provided) but statistically significant results for
postneonatal mortaliy (Ritz et al., 2006, 089819). Figure 7.6 compares risk for the
postneonatal period for respiratory and total mortality. In six of the seven studies, higher
effect estimates were observed for respiratory-related mortality. Results from the neonatal
period found higher effects for total mortality compared to respiratory mortality (Bobak and
Leon, 1999, 007678) and the reverse for a study examining infant mortality (Lipfert et al.,
2000, 004103). Thus, there exists evidence for a stronger effect at the postneonatal period
and for respiratory-related mortality, although this trend is not consistent across all
studies.
Decrements in Sperm Quality
Limited research conducted in the Czech Republic on the effect of ambient air
pollution on sperm production has found associations between elevated air pollution and
decrements in proportionately fewer motile sperm, proportionately fewer sperm with
normal morphology or normal head shape, proportionately more sperm with abnormal
chromatin (Selevan et al., 2000, 012578), and an increase in the percentage of sperm with
DNA fragmentation (Rubes et al., 2005, 078091). These results were not specific to PM, but
for exposure to a high-, medium- or low-polluted air mixture. Similarly, in Salt Lake City,
Utah, fine particulate matter (PM2.5) was associated with decreased sperm motility and
morphology (Hammoud et al. 2009). Research in Los Angeles, California examined 5134
semen samples from 48 donors in relation to ambient air pollution measured 0"9, 10-14, 70-
90 days before semen collection over a 2-yr period (1996-1998). Ambient ozone during all
exposure periods had a significant negative correlation with average sperm concentration,
and no other pollutant measures were significantly associated with sperm quality
parameters, or presented quantitatively (Sokol et al., 2006, 098539).
7.4.2. Toxicological Studies
This section summarizes recent evidence on reproductive health effects reported with
exposure to ambient I'M! no evidence was presented in this area in the 2004 PM AQCD.
Studies from different toxicological rodent models allow for investigation of specific
mechanisms and modes of action for reproductive changes. Emphasis is placed here on
July 2009
7-84
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
results from different windows of development, i.e., if exposure in utero, neonatally or as an
adult can affect reproductive outcomes as an adult. In addition, studies evaluating whether
fertility is affected in female and/or male animals equally by a similar exposure, and how
exposures are transmitted to the fertility of the F1 offspring, are summarized. Hormonal
changes which can lead to decreased sperm count or changes in the estrous cycle are also of
interest. Pregnancy losses and placental sufficiency are also followed. Most recently, the
role of environmental chemicals in shifting sex ratios (also seen in epidemiologic studies)
and in affecting heritable DNA changes have become endpoints of interest.
7.4.2.1. Female Reproductive Effects
Significant work has been done in male rodent models to determine the effect of PM
exposure on reproductive success! fewer studies have been done on female rodents. Tsukue
et al. (2004, 096643) exposed pregnant C57-BL mice to DE 0.1 mg DEP/m3 diluted in
charcoal and HEPA-filtered clean air or to for 8 h/day GD 2-13 and at GD 14 collected the
female fetuses for analysis of mRNAfor Ad4BP-l/SF-l and MIS, and found no significant
changes. The concentration of the gaseous materials including NO, NOx, NO2, CO and SO2
are 2.2 + 0.34 ppm, 2.5 + 0.34 ppm, 0.0 ppm, 9.8 ± 0.69 ppm, and <0.1ppm (not detectable),
respectively. Work by Yoshida et al. (2006, 097015) showed changes in these two transcripts
in male ICR fetuses exposed to similar doses of DEP, albeit with different daily durations of
exposure. Further work by Yoshida et al. showed that of three mouse strains tested, ICR
male fetuses were the most sensitive to DE-dependent changes in these two genes.
Nonetheless, strain sensitivity to DEP may also differ by sex. Thus, it appears that female
mice exposed in utero to DE show a lack of response at the mRNA level of MIS or Ad4bP-
1/SF-l, important genes in male sexual differentiation that showed DE-dependent changes
in male pups from dams exposed in utero. Female fetuses do however show a decrease in
BMP-15, which is related to oocyte development. Possible ingestion exposure by the
animals during grooming cannot be ruled out in this study.
Windows of exposure are important in determining reproductive success as an adult.
Exposure as a neonate may have a drastically more profound impact than does a similar
adult exposure in females. To test this, female BALB/C mice were exposed to ambient air in
Sao Paulo as neonates or as adults and then were bred to non-exposed males (Mohallem et
al., 2005, 088657). Concentrations of pollutants in this ambient air including CO, NO2,
PM10, and SO2 as measured locally were 2.2 ± 1.0, 107.8 ± 42.3, 35.5 ± 12.8, and 11.2 ± 5.3,
respectively. They reported decreased fertility in animals exposed as newborn but not in
adult-exposed female BALB/c mice. There were a significantly higher number of liveborn
pups from dams housed in filtered chambers (PM removed as well as chemical substances)
versus animals exposed to ambient air as newborns. There was also a higher incidence of
implantation failures in dams reared as newborns in polluted chambers. Sex ratio (unlike
in epidemiologic studies), number of pregnancies per group, resorptions, fetal deaths, and
July 2009
7-85
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
fetal placental weights did not differ significantly by treatment group. Thus, in these
studies, exposure to ambient air pollution and its associated PM affected future
reproductive success of females if they were exposed as neonates and not if they are
exposed as adults.
Environmental chemicals have been shown to act as endocrine disruptors by acting on
the androgen pathway, including the phthalates, which have manifested their anti-
androgenic activity in numerous ways including decreased anogenital distance in male
rodents (Foster et al., 1980, 094701; Foster et al., 2001, 156442). To access the role of DE
exposure on reproductive success and anti-androgenic effects on offspring, Tsukue et al.
(2002, 030593) exposed 6-weeks-old female C57-BL mice to 4 months of DE (0.3, 1.0, or 3.0
mg/m3) or filtered air (controls). Some animals were euthanized at the end of this exposure!
DE-exposed estrous females from this group were found to have significantly decreased
uterine weight (1.0 mg/m3). Some of these DE-exposed females were bred to unexposed
males. It was determined that DE-exposure led to increased but not significant rates of
pregnancy loss in mated females (up to 25%). The rate of good nest construction of the
pregnant exposed dams at the highest dose group was significantly lower than control (3.0
mg DEP/m3). Offspring were weighed after birth with significant decreases in body weight
seen at 6 and 8 wk (males and females 1.0 and 3.0 mg DEP/m3) and in female offspring
(9 wk of age, 1.0 and 3.0 mg/m3). Anogenital distance, a sensitive marker of anti-androgen
activity in males, was significantly decreased in 30-day old DEP exposed male offspring (0.3
mg DEP/m3) v. controls. Thymus weight was significantly decreased in 30-day old female
offspring (3.0 mg DEP/m3) and remained decreased at 70 days (0.3 and 1.0 mg DEP/m3).
Ovary weight of female offspring was significantly decreased (3.0 mg DEP/m3) at 30 days,
but no longer significantly different at 70 days. In males at 70 days of age, body weights
were significantly decreased and AGD was significantly shorter (3.0 mg DEP/m3). In
females at 70 days of age, the 1.0 mg DEP/m3 group showed significantly lower organ
weights (adrenals, liver, and thymus) and the 3.0 mg DEP/m3 group had decreased body
weight. Thymus weight of the 0.3 mg/m3 females was significantly lower at 70 days. Also,
crown to rump length in females from dams exposed to DEP (1.0 and 3.0 mg DEP/m3) was
also significantly lower. In conclusion, adult exposure to DEP led to maternal-dependent
reproductive changes that affected outcomes in offspring manifesting as decreased pup body
weight, anti-androgenic effects like decreased AGD and decreased organ weight (which may
be confounded by changes in body weight).
7.4.2.2. Male Reproductive Effects
Rodent strains differ in their sensitivity and response to various environmental
chemicals. Studies were performed to determine PM-dependent strain sensitivity using
male steroidogenic enzymes as the model pathway. In utero exposure of 3 strains of
pregnant mice (ICR, C57B1/6J or ddY mice) via inhalation exposure of DE at 0.1 mg DE
July 2009
7-86
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
particles (DEP)/m3 in HEPA-filtered clean air occurred or clean air as controls continuously
over gestational days 2-13 (Yoshida et al., 2006, 156170). At gestational day 14, dams were
euthanized and fetuses were collected from the uteri. Male fetuses were collected from each
dam for mRNA analysis of genes related to male gonad development including mullerian
inhibiting substance (MIS), steroid transgenic factor (Ad4BP/SF-l), an enzyme in the
testosterone synthesis pathway, cytochrome P450 cholesterol side chain cleavage enzyme
(P450scc), and other steroidogenic enzymes [l76-hydroxysteroid dehydrogenase (HSD),
cytochrome P450 17-(rhydroxylase (P450cl7), and 3-6hydroxysteroid dehydrogenase
(3BHSD)]. There were significant decreases in MIS (ICR, and C57BL/6 mice) and
Ad4BP/SF-l (ICR mice) versus control at gestational day (GD) 14. SF-1 transcriptionally
regulates T secretion. MIS is crucial in for sexual differentiation including mullerian duct
regression in males. The ddY strain showed no significant changes in Ad4BP/SF-l or MIS
which the authors hypothesized may be due to changes in 3frh I ), which had marked
changes in expression in the ddY strain when compared to non-DE exposed controls. From
these studies, it appears that mouse strains with in utero exposure to DE show differential
sensitivity in gonadal differentiation genes (mRNA) expression in male offspring! ICR are
the most sensitive, followed by C57BL/6 with ddY mice the least sensitive.
Yoshida et al. (2006, 097015) also monitored changes in the male reproductive tract
after in utero exposure to DE. Timed-pregnant ICR dams were exposed during gestation
(2dpc to 16dpc) to continuous DE generated to concentrations of 0.3, 1.0 or 3.0 mg DEP/m3
in HEPA-filtered air or clean air as controls. PM deposition on the fur and ingestion of the
dams by grooming is another possible exposure route in this study. The reproductive tracts
of male offspring were monitored at 4 wk postnatally. These pups received possible
continued exposure through lactation as dams exposed to DE during gestation nursed pups.
There was a threshold effect! 0.3 mg/m3 had no effect on male reproductive organ weight or
serum testosterone (T). Exposure to the higher doses (1.0 and 3.0 mg/m3) of DEP led to
significant increases in reproductive gland weight [testis, prostate, seminal vesicle (3.0 mg
DEP/m3 only) and coagulating gland]. The intermediate dose of 1.0 mg DEP/m3 induced
significant increases in serum T. The organ weights are presented as absolute numbers and
not adjusted for body weight, which is sometimes problematic for complete representation
of hormonal changes as body weight may confound absolute organ weight changes.
Nonetheless, there were also significant decreases in mRNA for the steroidogenesis related
enzymes 36HSD (3.0 mg DEP/m3) and aromatase (3.0 mg DEP/m3). Transcripts relating to
male sexual differentiation [Mullerian inhibitory substance (mis) and steroid transgenic
factor (AD4BP/SF-1), 1.0 and 3.0 mg DEP/m3] were also significantly decreased. Sexual
differentiation is a tightly regulated process. For example, SF-1 missense mutations result
in XY individuals with external female genitalia. Thus the effect of environmental DE-
exposure should not be underscored.
July 2009
7-87
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
This study demonstrated effects of DE exposure on male spermatogenesis. Exposure
of pregnant ICR mice to DE (2 dpc-16 dpc continuous inhalation exposure to 1.0 mg DEP/m3
in filtered air or to filtered clean air) led to impaired spermatogenesis in offspring (Ono et
al., 2007, 156007). Male offspring were followed at PND 8, 16, 21 (3 wk), 35 (5 wk) and 84
(12 wk). After 16dpc but before termination of the study, all of the animals were transferred
to a regular animal care facility. No cross fostering was performed in this experiment, so
pups that were born to DE-exposed dams were also nursed on these dams and may have
received lactational exposure to DE from milk. The gaseous components of the diluted DE
included nitric oxide (NO), NO2, sulfur dioxide (SO2), and CO2 at concentrations of 11.75 ±
1.18, 4.62 ± 0.36, 0.21 ± 0.01, and 4922 ± 244 ppm, respectively. The average proportion of
sulfur in the fuel during this study was 0.043%. Body weight was significantly depressed at
PNDs 8 and 35. Accessory gland relative weight was significantly increased at PND 8 and
16 only. Serum testosterone was significantly decreased at 3 wk and at 12 wk was
significantly increased. At 5 and 12 wk, daily sperm production (DSP) was significantly
decreased. FSH receptor and star mRNA levels were significantly increased at 5 and 12 wk,
respectively. Relative testis weight and relative epididymal weight were unchanged at all
time points. All endpoints were measured at each time point and if not mentioned above,
those data reported no significant changes. Histological changes showed Sertoli cells with
partial vacuolization and a significant increase in testicular multinucleated giant cells in
the seminiferous tubules of DE exposed animals compared to control. This study indicates
that in utero exposure to DE had effects on spermatogenesis in offspring at the
histologically, hormonally and functionally.
In utero exposure to DE and its effect on adult body weight, sex ratio, and male
reproductive gland weight was measured by Yoshida et al. (2006, 097015). Pregnant ICR
mice were exposed by inhalation to DE (0.3, 1.0 or 3.0 mg DEP/m3 or to clean air) from 2dpc
to 16dpc. Pups were allowed to nurse in clean air on exposed dams until weaning and at
PND 28, male pups were sacrificed. At this time, serum testosterone and pup reproductive
gland weight was determined. Significant increases in relative reproductive organ weights
were reported at 1.0 and 3.0 mg DEP/m3 for the seminal vesicle, testis, epididymis,
coagulating gland, prostate and liver. Male pup serum testosterone was significantly
increased at 1.0 mg DEP/m3. Mean testosterone positively correlated with testis weight,
daily sperm production, aromatase and steroidogenic enzyme message level (P450cc, cl7
lyase, and P450 aromatase). Sex ratio did not differ in DE-exposed animals versus control.
Male pup body weight of DE-exposed animals was significantly increased at PND 28 (1.0
and 3.0 mg DEP/m3). These studies show that in utero DE-exposure led to increased serum
testosterone and increased reproductive gland weight in male offspring early in life.
The effects of DE on murine adult male reproductive function were studied by
exposing ICR male mice (6 wk of age) to DE (clean air control, 0.3, 1.0 or 3.0 mg DECP/m3)
for 12 h/day for 6 months with another group receiving a one month recovery of clean air
July 2009
7-88
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
exposure post-exposure (Yoshida and Takedab, 2004, 097760). After six months exposure,
there was a dose-dependent significant increase in degeneration of seminiferous tubules of
mice exposed to DEP. After six months, there was a significant decrease in daily sperm
production (DSP)/g of testis tissue in DEP exposed animals. After six months exposure to
DEP plus one month recovery with clean air exposure, significant decreases remained in
DSP at the higher doses! the effect was lost at 0.3 mg/m3. This adult exposure and other
work with in utero exposure to DE showed similar outcomes. The effect of ingestion of
deposited PM from the fur during grooming post-inhalation exposure cannot be ruled out as
a possible mechanism of exposure in this experiment.
Earlier studies showed an inverse correlation between environmental levels of PM
and sperm count in adult men (Mehta and Anad Kumar, 1997, 157197). To expand on PM-
dependent changes in spermatogenesis, an eloquent DE-exposure model was designed to
determine if PM or the gaseous phase of DE was responsible for changes in sperm
production in rodents (Watanabe, 2005, 087985). Pregnant dams (F344/DuCrj rats) exposed
to DE (6 hs/day exposure to 0.17 or 1.71 mg DEP/m3) or filtered air (removing PM only, high
dose filtered air and low dose filtered air) from GD7 to parturition produced adult offspring
with a decreased number of Sertoli cells and decreased daily sperm production (PND 96)
when compared to control mice exposed to clean air (Watanabe, 2005, 087985). The
concentrations of NO2 for the high filtered and low filtered exposure groups were 0.8 and
0.1 ppm, respectively. Because both filtered and DE-exposure groups showed the same
outcomes, the effects are likely due to gaseous components of DE.
Another source of PM emissions that is common around the world is motorcycle
exhaust. Adult male (8 weeks old) Wistar rats were exposed to motorcycle exhaust (ME) for
one hour in the morning and one hour in the afternoon Monday through Friday at L50
dilution in filtered clean air for 4 wk (group A) or 1:10 for 2 (group B) or 4 wk (group C) or to
clean air (Huang et al., 2008, 156574) via a head and nose inhalation chamber. After 4 wk of
exposure, both exposed groups had significantly decreased body weight v. control. All three
groups showed a decreased number of spermatids in the testis after ME exposure. Both
1:10 exposure groups also showed a decrease in caudal epididymal sperm counts. Group C
showed significant decreases in testicular weight. Group C had decreased mRNAfor the
cytochrome P450 substrate 7-ehtoxycoumarin Ode-ethylase, and increased IL-6, IL-1B,
and cox-2 mRNA control. Decreased protein levels of antioxidant superoxide dismutase and
increased IL-6 protein were reported for group C when compared to control. Serum
testosterone was significantly decreased in group C and co-treatment of group C with the
antioxidant vitamin E resulted in partial rescue of serum T levels and caudal epididymal
sperm counts (albeit still significantly decreased versus control), and returned IL-6, IL-16,
and COX-2 ME-exposure-dependent message levels to baseline. The glutathione
antioxidant system and lipid peroxidation were unchanged after these ME exposures at the
time points measured. Male animals exposed to ME in this experiment showed significant
July 2009
7-89
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
decrements in body weight, spermatid number, and serum testosterone with an increase in
inflammatory cytokines. Vitamin E co-treatment with ME-exposure led to an attenuation of
inflammation and a partial rescue of testosterone levels and sperm numbers. No filtration
was done in this experiment to determine if the effects were gas- or particle-dependent.
In summary, laboratory animals exposed to DE in utero or as adults manifest with
abnormal effects on the male reproductive system. In utero exposure to DE induced
increased reproductive gland weight and increased serum testosterone in early life
(PND28), which may lead to early puberty (albeit not measured in this study). With similar
in utero DE exposures, later life outcomes include decreased DSP, abberant sperm
morphology, and hormonal changes (T and FSHr decrements). Chronic exposure of adult
mice to DE also induced decreased DSP and seminiferous tubule degeneration. DE-
dependent effects on male reproductive function have been reported in multiple animal
models with only one model separating exposure based on particulate versus gaseous
components. DE and filtered air (gaseous phase only) exposure in utero induced Sertoli cell
and DSP decrements in both groups, indicating that the gaseous phase of DE was
causative. Adult male rats exposed to ME manifested with decreased spermatid number,
serum T, and an increase in inflammatory cytokines. Significant effects on the male
reproductive system have been demonstrated after exposure to ambient PM sources (DE,
ME, or PAHs). Nonetheless, these models often include a complex mixture of gaseous
component and PM exposure, which makes interpreting the contribution from PM alone
difficult.
7.4.2.3. Multiple Generation Effects
Multi-generational, chronic exposure to traffic-generated fine PM affects reproductive
and fetal outcomes in female mice including estrous cyclicity, follicle development, mating,
fertility, and pregnancy success. In this study, Veras et al. (2009, 190496) investigated
pregnancy and female reproductive outcomes in BalbC female mice exposed to ambient air
(NF) or filtered air (F) at one of two different time periods (before conception and during
pregnancy) at the University of Sao Paulo School of Medicine near an area of high traffic
density. The exposure system is described by Mohallem et al., earlier in this ISA. Two
groups of 2nd generation (G2) nulliparous female mice were continuously exposed to
filtered air (F) or non-filtered ambient air (NF) since birth. Estrous cyclicity and ovarian
follicle classification were followed at PND60 (reproductive maturation) in one group. A
further group was subdivided into 4 groups by exposures during pregnancy following
reproductive capability and pregnancy outcomes of the G2 mice. Exposure were 27.5 and
6.5 ug/m3PM2.5, respectively for NF and F chambers with 101 ug/m3 NO2, 1.81 ug/m3 CO,
and 7.66 ppm SO2 in both chambers (Veras et al., 2009, 190496).
The results of this study showed that animals exposed to NF air versus F air had an
extended time in estrous and thus a reduction in the number of cycles during the
July 2009
7-90
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
experiment study period. The number of antral follicles was significantly decreased in the
NF versus the F animals. Other follicular quantification (number of small, growing or
preovulatory follicles) showed no significant differences between the two chambers. There
was a significant increase in the time necessary for mating, a significant decrease in the
fertility index, and a significant increase in the pregnancy index in the NF group versus the
F group. Specifically, in the NF1 and NF2 groups, there was a significant increase in rate of
the post-implantation loss. However, there was no statistically significant change in
number of pups in the litter. Fetal weight was significantly decreased in all treatment
groups (F2, NF1, and NF2) when compared to F1 or animals raised entirely in filtered air,
showing that fetal weight was affected by both pre-gestational and gestational PM exposure
(Veras et al., 2009, 190496).
PM exposure prior to conception is associated with increased time in estrous, which in
other animal models can be related to ovarian hormone dysfunction and ovulatory
problems. These estrous issues can contribute to fecundity issues. There was no significant
difference in number of preovulatory follicles in this model, but there was a statistically
significant decrease in the number of antral follicles. Antral follicles are the last stage in
follicle development prior to ovulation, and a decrease in antral follicle number can be
related to premature reproductive senescence, premature ovarian failure, or early
menopause, which was not followed in this model (Veras et al., 2009, 190496).
From the experimental design, one sees that the males that were used to generate the
G1 and G2 generations were also exposed to NF or F air, and thus the reproductive
contribution of these males to the overall fertility and mating changes in the females
mentioned above cannot be totally eliminated as a possible confounder as the literature has
characterized PM exposure as being associated with adverse male reproductive outcomes.
Thus these effects are hard to differentiate as male or female-dependent and can thus
indicate a general loss of reproductive fitness. Interestingly, both pre and gestational
exposure to NF induced a significant loss in post-implantation of fetuses and this may be
related to placental insufficiency as has been described in other work by this lab (Veras et
al., 2008, 190493).
7.4.2.4. Receptor Mediated Effects
Arylhydrocarbon Receptor (AhR) and DEP
The AhR is often activated by chemicals classified as endocrine disrupting compounds
(EDCs), exogenous chemicals that behave as hormonally active agents, disrupting the
physiological function of endogenous hormones. DEPs are known to activate the AhR. A
recent study by Izawa et al. (2007, 190387) showed that certain polyphenols (quercetin from
the onion) and food extracts (Ginkgo biloba extract (GBE)) are able to attenuate DEP-
July 2009
7-91
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
dependent AhR activation when measured with the Ah-Immunoassay, thus possibly
attenuating the EDC activity of DEP.
7.4.2.5. Developmental Effects
Sex Ratio
A direct correlation between air pollution (PMio) exposure and a decrease in
standardized sex ratios (SSRs) has been reported in humans exposed to air pollution
(Lichtenfels et al., 2007, 097041; Wilson et al., 2000, 010288) pollution, with fewer numbers
of male births reported. To understand this shift, two groups (control and exposed) of male
Swiss mice were housed concurrently in Sao Paulo and received either ambient air
exposure or filtered air (chemical and particulate filtering) from PND10 for four months.
Filtration efficiency for PM2.5, carbon black, and NO2 inside the chamber was found to be
55%, 100%, and 35%, respectively. After this exposure, non-exposed females were placed in
either chamber to mate. After mating, the males were sacrificed and testes collected; males
exposed to ambient air showed decreased testicular and epididymal sperm counts,
decreased total number of germ cells, and decreased elongated spermatids, but no
significant change in litter size. Females were housed in the chambers and sacrificed on
GD19 when the number of pups born alive and the sex ratio were obtained. There was a
significant decrease in the SSR for pups born after living in the ambient air-exposed
chamber compared to the filtered chamber. In this study, a shift in SSR has been shown for
both humans and rodents exposed to air pollution, but other studies with DE exposure
(Yoshida et al., 2006, 156170) or ambient air in Sao Paulo (Mohallem et al., 2005, 088657)
showed no changes in rodent sex ratio. Possible exposure to PM and other components of
ambient air via ingestion during grooming cannot be ruled out in this rodent model.
Immunological Effect-Placenta
Placental insufficiency can lead to the loss of a pregnancy or to adverse fetal
outcomes. DE-exposure has been shown to induce inflammation in various models.
Fujimoto et al. (2005, 096556) assessed cytokine/immunological changes of DE-dependent
inhalation exposure on the placenta during pregnancy. Pregnant Sl<> CR mice were exposed
to DE (0.3, 1.0, or 3.0 mg DEP/m3 in HEPA and charcoal-filtered clean air from 2dpc to 13
dpc) or clean air in inhalation chambers! dams, placenta, and pups were collected at 14dpc.
There was a significant increase in the number of absorbed placentas in DE exposed
animals (0.3 and 3.0 mg DEP/m3) with a significant decrease in the number of absorbed
placentas in DE exposed animals at the middle dose (1.0 mg DEP/m3). Absorbed placentas
from DE exposed mice had undetectable levels of CYP1A1 and two fold increases in TNF-a!
CYP1A1 placental mRNAfrom healthy placentas of DE-exposed mice was unchanged
versus control. Interleukin (!L)-2, IL-5, IL-12a, IL-12-6 and granulocyte macrophage colony-
July 2009
7-92
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
stimulating factor (GM-CSF) mRNA significantly increased in placentas of DE-exposed
animals (0.3 and 3 mg DEP/m3). Placental IL-6 mRNA was increased ten-fold in DE-
exposed mice (3.0 mg DEP/m3). Fujimoto et al. reported DE-induced significant increases in
multiple inflammatory markers in the placenta with significant increases in the number of
absorbed placentas.
Immunological Effects: Asthma
In utero exposure may confer susceptibility to PM-induced asthmatic responses in
offspring. Exposure of pregnant BALB/c mice to aerosolized ROFAleachate by inhalation or
to DEP intranasally increases asthma susceptibility to their offspring (Fedulov et al., 2008,
097482; Hamada et al., 2007, 091235). The offspring from dams exposed for 30 min to 50
mg/mL ROFA 1, 3, or 5 days prior to delivery responded to OVA immunization and aerosol
challenge with airway hyperreactivity and increased antigen-specific IgE and IgGl
antibodies. Airway hyperreactivity was also observed in the offspring of dams intra-nasally
instilled with 50 |ug of DEP or Ti02, or 250 |ug CB, indicating that the same effect could be
demonstrated using relatively "inert" particles. Pregnant mice were particularly sensitive
to exposure to DEP or TiCh particles, and genetic analysis indicated differential expression
of 80 genes in response to Ti02 in pregnant dams. Thus pregnancy and in utero exposure
may enhance responses to PM, and exposure to even relatively inert particles may result in
offspring predisposed to asthma.
Placental Morphology and Urban PM Exposure
Exposure to ambient air pollution during pregnancy is associated with reduced fetal
weight in both human and animal models. The effect of particular urban air pollution on
the functional morphology of the mouse placenta was explored by exposing second
generation mice in one of four groups to urban Sao Paulo air (PM was 67% PM2.5, mainly of
vehicular origin) or filtered air (Veras et al., 2008, 190493). Experimental design follows
with group F-F being comprised of mice that were raised in filtered air chambers and
completed pregnancy in filtered air chambers! group F-nF was raised in filtered air and
pregnant in ambient air! group nF-nF was raised and completed pregnancy in non-filtered
air chambers! group nF-F mice were raised in ambient air and received filtered air during
pregnancy. Exposure was from PND20-PND60. At this time, the animals were mated and
then maintained in their respective chambers during pregnancy! pregnancy was terminated
at GD18 (near term) with placentas and fetuses collected then for analysis.
Exposure to ambient PM pregestationally or gestationally led to significantly smaller
fetal weight (total litter weight). Progestational exposure to ambient air induced significant
increases in fetal capillary surface area and in total mass-specific conductance. But
maternal/dam blood space and diameters were reduced. Gestational exposure to non-
filtered air was associated with reduced volume, diameter (caliber) and surface area of
July 2009
7-93
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
maternal blood space with compensatory greater fetal capillary surface and oxygen
diffusion conduction rates. Intravascular barrier thickness, a quantitative relationship
between trophoblast volume and the combined surfaces of maternal blood spaces and fetal
capillaries, was not reduced with ambient air exposure. Fetal/placental circulatory
adaptation to maternal blood deficits after ambient PM exposure seems to be insufficient to
overcome PM-dependent birth weight deficits in mice exposed to ambient air with the
magnitude of this effect greater in the gestationally exposed groups.
Placental Weights and Birth Outcomes
Pregnant female Swiss mice were exposed to ambient air (Sao Paulo) or filtered air
over various portions of gestation to determine if there was an association between fetal or
placental weight or birth outcomes with exposure to air pollution. The concentration of
various components of the ambient air as measured by a State Environmental Sanitation
Agency 100 m away from the rodent exposure chambers reported PMio (42 ± 17 pm/m3),
NO2 (97 ± 39 pg/m3), and SO2 (9 ± 4 pg/m3) concentrations. By using six windows of
exposure that covered one to three weeks of gestation, which is all of gestation in a mouse,
these authors (Silva et al., 2008, 156981) determined that a significant decrease in near-
term fetal weight(GD19) could be induced by ambient air-exposure at least during the first
week of gestation. Decreased placental weight could be induced by ambient air exposure
during any of the three weeks of gestation. These studies point to possible windows of
exposure that may be important in evaluating epidemiologic study results.
Neurodevelopmental Effects
The diagnosis of autism is on the rise in the Western world with its etiology mostly
unknown. Autism is associated cell loss in specific brain regions that is hypothesized to be
developmental in origin. Sugamata et al. (2006, 097166) exposed pregnant ICR mice to DE
(0.3 mg DEP/m3) continuously from 2 days post-coitus (dPC) to 16 dPC. Pups with in utero
exposure to DE were nursed in clean air chambers but may have received gastro-intestinal
exposure via lactational transfer of various components of DE. At 11 wk of age, cerebullar
brain tissue was collected. Twenty animals were in each group (10 females, 10 males) with
one group receiving clean air exposure and one receiving DE; no filtration was used to
compare PM v. gaseous DE exposure. Earlier work has shown that DEP (<0.1 pm) have
been detected in the brains (cerebral cortex and hippocampus) of newborn pups who were
born to dams who were exposed to DE during pregnancy (Sugamata et al., 2006, 097166).
Histological analysis of DE-exposed pup cerebella revealed significant increases in caspase-
3 (c-3) positive cells compared to control and significant decreases in cerebella Purkinje cell
numbers in DE-exposed animals versus control. The ratio of cells positive for apoptosis (c*3
positive) showed a nearly significant sex difference with males displaying increased
apoptosis versus females (p = 0.09). In humans with autism, the cerebellum has a decreased
July 2009
7-94
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
number of Purkinje cells, which is thought to be fetal and developmental in origin! further,
these authors speculate that humans maybe more sensitive to DE-dependent neuronal
brain changes as the human placenta is 2 layers thick whereas the mouse placenta is 4
layers thick.
Behavioral Effects
Body weight decrements at birth have recently been associated through the Barker
hypothesis with adverse adult outcomes. Thus, many publications have begun to focus on
decreased birth weight for gestational age and associated adult changes. Hougaard et al.
(2008, 156570) exposed 40 timed-pregnant C57BL/6 dams to DEP reference materials (aged
DE particulate extract) via inhalation chamber over GD7-GD19 of pregnancy. They found
significantly decreased pup weight at weaning, albeit not at birth. PM-dependent liver
changes were monitored by following various inflammatory and gentoxicityrelated mRNA
transcripts! there were no significant differences in pups at PND2. The comet assay from
PND2 pup livers showed no significant differences between DEP-exposed and control
animals. Thyroxine was unchanged in control and DEP-exposed dams and offspring at
weaning. At two months, female DEP-exposed pups required less time than controls to
locate the platform in its new location during the first trial of the spatial reversal learning
task in the Morris water maze (p < 0.05). DEP extract exposure during in utero
development led to decreased body weight at weaning and no changes in inflammatory
markers, or thyroid hormone levels.
The effect of in utero DE exposure on CNS motor function was evaluated in male pups
(ICR mice); dams received DE exposure 8hxd/5d/wk via total body inhalation GD2-GD17
(Yokota et al., 2009, 190518). Spontaneous motor activity was significantly decreased in
pups (PND35) as was the dopamine metabolite homovanillic acid (HVA) as measured in the
striatum and nucleus accumbens indicating decreased dopamine (DA) turnover. DA levels
were unchanged in the same areas of the brain. The authors conclude that these data
demonstrate that maternal exposure to DE induced hypolocomotion, similar to earlier
studies with adult and neonatal DEP exposure (Peters et al., 2000, 001756), with decreased
extracellular DA release. Concentrations of DE constituents in the Yokota study for DE
particle mass, CO, NO2, and SO2 are 1.0 mg/m3, 2.67 ppm, 0.23 ppm, and <0.01ppm,
respectively.
Lactation
Lactational exposure to various environmental compounds is an area of research that
is often overlooked. Breast milk is a complex matrix, which is essential for the survival of
many species. Compounds that are especially lipophilic are commonly found in breast milk
of exposed dams and this maternal load can be transferred to the developing neonate. PM
in DE adsorbs many chemicals, including polycyclic aromatic hydrocarbons (PAHs), which
July 2009
7-95
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
have been shown to have to be mutagenic and to be estrogenic/antiestrogenic and
antiandrogenic (Hirose et al., 2001, 156548; Kizu et al., 2003, 096196). Thus, Tozuka et al.
(2004, 090864) monitored the transfer of aromatic hydrocarbons to fetuses and breast milk
of Fisher 344 rats exposed to DE for 2 wk from GD7-GD 20 (minus four days of the weekend
with no exposures) for 6h/day with PMio concentration of 1.73 mg/m3. Concentrations of
individual PAHs were monitored in the inhalation chambers including Ace, Fie, Phe, Ant,
Flu, Pyr, BaA, Chr, BbF, BkF, BaP, DBA, BghiP, and IDP at 150 ± 34, 3160 ± 401,
2280 ± 291, 70.3 ± 10.9, 148 ± 19, 133 ± 5, 17.2 ± 2.7, 39.9 ± 6.8, 9.9 ± 2.1, 4.9 ± 1.1, 3.7 ± 0.5,
<1.4, <6.0, and 4.2 ±0.1 ng/m3, respectively. At PND 14, milk was collected from exposed
and control rats. Fifteen PAHs were monitored in DE-generated air. Seven of these were
quantified in dam blood with levels of phenanthrene (Phe), anthracene (Ant) and
benz[a]anthracene (BaA) in the DE group being significantly higher than control group. In
breast milk, acenaphthene (Ace), fluorene (Fie), Phe, Ant, fluoranthene (Flu), pyrene (Pyr),
BaA and chrysene (Chr) were quantified. Ant, Flu, Pyr and Chr showed significant
increases in the DE group compared to control milk. BaA tended to be about four fold
higher than the control group in breast milk, but the increase was not significant. PAHs in
dam livers of DE versus control were not significantly different. PAHs are transferred
across the placenta from the DE-exposed dam to the fetus. Lactational transfer through the
breast milk is also likely as PAHs are detected in dam breast milk, but this should be
confirmed in future studies that cross foster control and exposed dams and pups. The
lipophilicity of the PAH based on its structure affected its uptake to the dam from the air as
PAHs with 3 or 4 rings were found in maternal blood and PAHs with 5 or 6 rings were not
detected in dam blood.
Heritable DNA Changes and Heritable Epigenetic Changes
To address the role of ambient air exposure on heritable changes, Somers et al. (2004,
078098) exposed mice to ambient air in at a rural Canadian site or at an urban site near a
steel mill. They showed that offspring of mice exposed to ambient air in urban regions
inherited paternal-origin expanded simple tandem repeat (ESTR) mutations 1.9 to 2.1
times more frequently than offspring of mice exposed to HEPA filtered air or those exposed
to rural ambient air. Mouse expanded simple tandem repeat (ESTR) DNA is composed of
short base pair repeats which are unstable in the germline and tend to mutate by insertion
or deletion of repeat units. In vivo and in situ studies have shown that murine ESTR loci
are susceptible to ionizing radiation, and other environmental mutatgen-dependent
germline mutations, and are thus good markers of exposure to environmental
contaminants.
Expanding upon the above work and to determine if PM or the gaseous phase of the
urban air was responsible for heritable mutations, Yauk et al. (2008, 157164) exposed
mature male C57BlxCBAFl hybrid mice to either HEPA-filtered air or to ambient air in
July 2009
7-96
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Hamilton, Ontario, Canada for three, ten, or 10 wk plus 6 wk of clean air exposure (16 wk).
Sperm DNA was monitored for expanded simple tandem repeat (ESTR) mutations. In
addition, male-germ line (spermatogonial stem cell) DNA methylation was monitored post-
exposure. This area in Hamilton is near 2 steel mills and a major highway. Air composition
provided by the Ontario Ministry of the Environment showed TSP concentration of
9.38 ± 17 |ug/m3, PAH concentration of 8.3 ± 1.7 ng/m3, and metal at 3.6 ± 0.7 mg/m3.
Mutation frequency at ESTR Ms6"hm locus in sperm DNA from mice exposed 3 or 10 wk
did not show elevated ESTR mutation frequencies, but there was a significant increase in
ESTR mutation frequency at 16 wk in ambient air exposed males versus HEPA-filter
exposed animals, pointing to a PM-dependent mechanism of action. When compared to
HEPA-filter air exposed males, ambient air exposed males manifested with
hypermethylation of germ-line DNA at 10 and 16 wk. These PM-dependent epigenetic
modifications (hypermethylation) were not seen in the halploid stage (3 wk) of
spermatogenesis, but were nonetheless seen in early stages of spermatogenesis (10 wk) and
remained significantly elevated in mature sperm even after removal of the mouse from the
environmental exposure (16 wk). Thus, these studies indicate that the ambient PM phase
and not the gaseous phase is responsible for the increased frequency of heritable DNA
mutations and epigenetic modifications.
7.4.3. Summary and Causal Determinations
7.4.3.1. PM2.5
The 1996 PM AQCD concluded that while few studies had been conducted on the link
between PM and infant mortality, the research "suggested an association," particularly for
post-neonates (U.S. EPA, 1996, 079380). In the 2004 PM AQCD, additional evidence was
available on PM's effect on fetal and early postnatal development and mortality and while
some studies indicated a relationship between PM and pregnancy outcomes, others did not
(U.S. EPA, 2004, 056905). Studies identifying associations found that exposure to PM10
early during pregnancy (first month of pregnancy) or late in the pregnancy (six weeks prior
to birth) were linked with higher risk of preterm birth, including models adjusted for other
pollutants, and that PM2.5 during the first month of pregnancy was associated with IUGR.
However, other work did not identify relationships between PM10 exposure and LBW. The
state of the science at that time, as indicated in the 2004 PM AQCD, was that the research
provided mixed results based on studies from multiple countries.
Building on the evidence characterized in the previous AQCDs, recent epidemiologic
studies conducted in the U.S. and Europe were able to examine the effects of PM2.5, and all
found an increased risk of LBW (Section 7.4.1). Exposure to PM2.5 was usually associated
with greater reductions in birth weight than exposure to PM10. All of the studies that
July 2009
7-97
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
examined the relationship between PM2.5 and preterm birth report positive associations,
and most were statistically significant. The studies evaluating the association between
PM2.5 and growth restriction all found positive associations, with the strongest evidence
coming when exposure was assessed during the first or second trimester (Section 7.4.1). For
infant mortality (<1 year), several studies examined PM2.5 and found positive associations
(Section 7.4.1).
Animal toxicological studies reported effects including decreased uterine weight,
limited evidence of male reproductive effects, and conflicting reports of reproductive
outcomes in male offspring, particularly in studies of DE (Section 7.4.2). Toxicological
studies also reported effects for several development outcomes, including immunological
effects in the placenta and related to asthma and neurodevelopmental and behavioral
effects (Section 7.4.2).
In summary evidence is accumulating from epidemiologic studies for effects on LBW
and infant mortality, especially due to respiratory causes during the post-neonatal period.
The mean PM2.5 concentrations during the study periods ranged from 5.3-27.4 ug/ma.
Exposure to PM2.5 was usually associated with greater reductions in birth weight than
exposure to PM10. Several U.S. studies of PM10 investigating fetal growth reported 11-g
decrements in birth weight associated with PM10 exposure. The consistency of these results
strengthens the interpretation that particle exposure may be causally related to reductions
in birth weight. Similarly, animal evidence supported an association between PM2.5 and
PM10 exposure and adverse reproductive and developmental outcomes, but provided little
mechanistic information or biological plausibility for an association between long-term PM
exposure and adverse birth outcomes, including LBW, or infant mortality. Epidemiologic
studies do not consistently report associations between PM exposure and preterm birth,
growth restriction, birth defects or decreased sperm quality. New evidence from animal
toxicological studies on heritable mutations is of great interest, and warrants further
investigation. Overall, the epidemiologic and toxicological evidence is suggestive of a causal
relationship between long-term exposures to PM2.5 and reproductive and developmental
outcomes.
7.4.3.2. PM10-2.5
Evidence is inadequate to determine if a causal relationship exists between long-term
exposure to PM 10-2.5 or other PM components and developmental and reproductive outcomes
because studies have not been conducted in sufficient quantity or quality to draw any
conclusion. A single study found an association between PM10-2.5 and birthweight (-13 g
[95% CP -18.3 to -7.6] per 10 pg/m3 increase), but no such association for PM2.5 (Parker et
al„ 2008, 156013).
July 2009	7-98	DRAFT - DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
7.5. Cancer, Mutagenicity, and Genotoxicity
Evidence from epidemiologic and animal toxicological studies has been accumulating
for more than three decades regarding the mutagenicity and carcinogenicity of PM in the
ambient air. Diesel exhaust (DE) has been identified as one source of PM in ambient air,
and has been extensively studied for its carcinogenic potential. In 1989, the International
Agency for Research on Cancer (IARC) found that there was sufficient evidence that
extracts of DEP were carcinogenic in experimental animals and that there was limited
evidence for the carcinogenic effect of DE in humans (IARC, 1989, 002958). This conclusion
was based on studies in which organic extracts of DEP were used to evaluate the effects of
concentrates of the organic compounds associated with carbonaceous soot particles. These
extracts were applied to the skin or administered by IT or intrapulmonary implantation to
mice, rats, or Syrian hamsters and an excess of tumors on the skin, lung or at the site of
injection were observed.
In 2002, the U.S. EPA reviewed over 30 epidemiologic studies that investigated the
potential carcinogenicity of DE. These studies, on average, found that long-term
occupational exposures to DE were associated with a 40% increase in the relative risk of
lung cancer (U.S. EPA, 2002, 042866). In the same report the U.S. EPA concluded that
extensive studies with salmonella had unequivocally demonstrated mutagenic activity in
both particulate and gaseous fractions of DE. They further concluded that DE may present
a lung cancer hazard to humans (U.S. EPA, 2002, 042866). The particulate phase appeared
to have the greatest contribution to the carcinogenic effect. Both the particle core and the
associated organic compounds demonstrated carcinogenic properties, although a role for the
gas-phase components of DE could not be ruled out. Almost the entire diesel particle mass
is < 10 pm in diameter (PMio), with approximately 94% of the mass of these particles
<2.5 pm in diameter (PM2.5), including a subgroup with a large number of ultrafine
particles (U.S. EPA, 2002, 042866). U.S. EPA considered the weight of evidence for potential
human carcinogenicity for DE to be strong, even though inferences were involved in the
overall assessment, and concluded that DE is "likely to be carcinogenic to humans by
inhalation" and that this hazard applies to environmental exposures (U.S. EPA, 2002,
042866).
Two recent reviews of the mutagenicity (Claxton et al., 2004, 089008) and
carcinogenicity (Claxton and Woodall GM, 2007, 180391) of ambient air have characterized
the animal toxicological literature on ambient air pollution and cancer. The majority of
these toxicological studies have been conducted using IT or dermal routes of exposure.
Generally, the toxicological evidence reviewed in this ISA has been limited to inhalation
studies conducted with lower concentrations of PM (<2 mg/m3), relevant to current ambient
concentrations and the current regulatory standard (See Section 1.3). This ISA focuses on
toxicological studies which use the inhalation route of exposure, therefore it is possible that
July 2009
7-99
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
important evidence for the role of PM in mutagenicity, tumorigenicity, and/or
carcinogenicity may be missed. In order to accurately characterize the relationship between
PM and cancer and be consistent with the EPA Guidelines for Carcinogen Risk Assessment
(U.S. EPA, 2005, 086237), these reviews (that include studies that employ IT and dermal
routes of exposure) are summarized briefly.
Claxton et al. (2004, 089008) reviewed the mutagenicity of air in the Salmonella
(Ames) assay, and showed that hundreds of compounds identified in ambient air from
varying chemical classes are mutagenic and that the commonly monitored PAHs could not
account for the majority of mutagenicity associated with most airborne particles. They
concluded that the smallest particles have the highest toxicity per particulate mass, with
the PM2.5 size fraction having greater mutagenic and cytotoxic potential than the PM10 size
fraction, which had a higher mutagenic potential than the TSP size fraction. One study
reviewed by Claxton et al. (2004, 089008) found that the cytotoxic potential of PM2.5 was
higher in wintertime samples than in summertime samples. A series of studies on source
apportionment for ambient particle mutagenic activity reviewed by Claxton et al. (2004,
089008) indicate that mobile sources (cars and diesel trucks) account for most of the
mutagenic activity.
Claxton and Woodall (2007, 180391) reviewed many studies that examined the rodent
carcinogenicity of extracts of ambient PM samples! the PM was of various size classes, often
from TSP samples. Among a variety of mouse and rat strains, application methods, and
samples employed, the authors found no pattern that would suggest the routine use of a
particular strain or protocol would be more informative than another. The primary
conclusion that comes from the analysis of rodent carcinogenicity studies is that the most
polluted urban air samples tested to date are carcinogenic! the contribution of PM and
different size classes of PM to the carcinogenic effects of ambient air has not been
delineated. The differences in response by the various strains of inbred mice indicate that
the genetic background of an individual influences how likely a tumorigenic response is to
occur. Studies examining different components of ambient PM (e.g., PAHs) confirm that
ambient air contains multiple carcinogens, and that the carcinogenic potential of particles
from different airsheds can be quite different. Therefore, one would expect the incidence of
cancers related to ambient air exposure in different metropolitan areas to differ.
Numerous epidemiologic and animal toxicological studies of ambient PM and their
contributing sources have been conducted to assess the relative mutagenic or genotoxic
potential. Studies previously reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905)
provide evidence that ambient PM as well as PM from specific combustion sources (e.g.,
fossil fuels) is mutagenic in vivo and in vitro. Building on these results, data from recent
epidemiologic and animal toxicological studies that evaluated the carcinogenic, mutagenic
and/or genotoxic effects of PM, PM-constituents, and combustion emission source particles
are reviewed in this section.
July 2009
7-100
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
7.5.1. Epidemiologic Studies
The 2004 PM AQCD reported on original and follow-up analyses for three prospective
cohort studies that examined the relationship between PM and lung cancer incidence and
mortality. Based on these findings, as well as on the results from case-control and ecologic
studies, the 2004 PM AQCD concluded that long-term PM exposure may increase the risk of
lung cancer incidence and mortality. The largest of the three prospective cohort studies
included in the 2004 PM AQCD was the ACS study (Pope CAet al., 2002, 024689). This
study was the follow-up to the original ACS study (Pope CAet al., 1995, 045159), and
included a longer follow-up period and reported a statistically significant association
between PM2.5 exposure and lung cancer mortality.
A 14- to 16-yr prospective study conducted using the Six Cities Study cohort reported
a slightly elevated risk of lung cancer mortality for individuals living in the most polluted
city (mean PMio: 46.5 |ug/m3; mean PM2.5 29.6 ug/m3) as compared to the least polluted city
(mean PMio: 18.2 ug/m3; mean PM2.5 11.0 ug/m3) but the association was not statistically
significant (Dockery et al., 1993, 044457).
Re-analysis of the AHSMOG cohort, a study of non-smoking whites living in
California, concluded that elevated long-term exposure to PM10 was associated with lung
cancer incidence among both men and women (Beeson et al., 1998, 048890). The original
study had reported an excess of incident lung cancers only among women (Abbey et al.,
1991, 042668). Further reanalysis of this cohort revealed an association between PM10 and
lung cancer mortality among men but no association among women (Abbey et al., 1999,
047559). In addition, McDonnell et al. (2000, 010319) reported increases in lung cancer
mortality with long-term exposure to PM2.5 in the AHSMOG cohort! no association was seen
for PM10-2.5.
7.5.1.1. Lung Cancer Mortality and Incidence
The following sections will examine extensions of the above mentioned cohort studies
and new studies published since the 2004 PM AQCD. The section includes discussion of
both lung cancer incidence and mortality, as well as markers of exposure/susceptibility. A
summary of the mean PM concentrations reported for the new studies is presented in
Table 7.6. In addition, a summary of the associations for lung cancer mortality and
incidence are presented in Table 7.7 and Figure 7.8 (see Section 7.6) Further discussion of
all-cause and cause-specific mortality is presented in Section 7.6.
July 2009
7-101
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Table 7-6. Characterization of ambient PM concentrations from recent studies of cancer and long-
term exposures to PM.
Reference
Location
Pollutant
Mean Annual
Concentration (/^g/m31
Upper Percentile
Concentrations (/^g/m31
Brunekreef et al. (2009,191947)
The Netherlands
PM2.6
28.3
Max: 36.8
Bonner et al. (2005, 088993)
Western NY State
TSP
44

Jerret et al. (2005,189405)
Los Angeles, California
PM2.6

Max:27.1
Laden et al. (2006, 087605)
6 U.S. cities
PM2.6
Range of means across sites: 10.2-29.0
Avg of means across sites: 16.4

Naess et al. (2007, 090736)
Oslo, Norway
PM2.6
15
Max: 22


PM10
19
Max: 30
Palli et al. (2008,156837)
Florence, Italy
PM10
NR

Pedersen et al. (2006,156848)
Czech Republic
PM2.6

Max: 46-120


PM10

Max: 120-238.6
Sorensen et al. (2005,188901)
Copenhagen, Denmark
PM2.6
Range of means across sites: 12.6-20.7
Avg of means across sites: 16.7
75th: 24.3-27.7
Sram et al. (2007,188457)
Czech Republic
PM10

Max: 55


PM2.6

Max: 38
Sram et al. (2007,192084)
Czech Republic
PM10
Range of means across sites: 36.4-55.6
Avg of means across sites: 46.0



PM2.6
Range of means across sites: 24.8-44.4
Avg of means across sites: 34.6

Vineis et al. (2006,192089)
Multi-city, Europe
PM10
Range of means across sites: 19.9-73.4
Avg of means across sites: 35.4

Vinzents et al. (2005, 087482)
Copenhagen, Denmark
PM10
Range of means across sites: 16.9-23.5
Avg of means across sites: 20.2

A subset of the ACS cohort study from 1982-2000 that included only residents of Los
Angeles, California was used to examine the association between PM2.5 and lung cancer
mortality while adjusting for both individual and neighborhood covariates (Jerrett et al.,
2005, 189405). There was a positive association between PM2.5 and lung cancer mortality
when adjusting for 44 individual covariates [RR 1.44 (95% CI: 0.98-2.11) per 10 jug/m3
increase in PM2.5]. However, including all potential individual and neighborhood covariates
associated with mortality reduced the association [RR 1.20 (95% CI: 0.79-1.82) per 10 ug/m3
increase in PM2.5]. A recent re-analysis of the full ACS cohort by the Health Effects Institute
also demonstrated a positive association between PM2.5 and lung cancer mortality [RR 1.11
(95% CP 1.04-1.18)] (Krewski et al., 2009, 191193). The authors observed modification of
this risk by educational attainment, with those completing a high school degree or less
having greater risk. In addition to utilizing the ACS cohort for a nationwide analysis, this
same study conducted two regional assessments, one in the New York City area and the
other in the Los Angeles area. No association was detected between PM2.5 and lung cancer
July 2009
7-102
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
mortality in the analysis of the region included in the New York City analysis. A positive
association was observed in the Los Angeles-area analysis using an unadjusted model, but
this association did not persist after control for individual, ecologic, and co-pollutant
covariates.
The Six Cities Study was extended to include data from 1990-1998, a period including
1,368 deaths and 54,735 person-years (Laden et al., 2006, 087605). An elevated risk ratio
for lung cancer mortality was reported when the entire followup period (1974-1998) was
included in the analysis [RR 1.27 (95% CI 0.96, 1.69) per 10 |ug/m3 increase in average
annual PM2.5]. However, estimated decreases in PM2.5 were not associated with reduced
lung cancer mortality [RR 1.06 (95% CP 0.43-2.62) for every 10 ug/m3 reduction in PM2.5].
Naess et al. (2007, 090736) studied individuals aged 51-90 living in Oslo, Norway in
1992. Death certificate data were obtained for 1992-1998 and information on PM was
collected from 1992-1995. Women had a larger association of lung cancer mortality with
PM2.5 compared to men. Similar results were reported for PM10.
Most recently, Brunekreef et al. (2009, 191947) used the Netherlands cohort study
(NLCS) on diet and cancer to conduct a re-analysis of the research performed by Beelen et
al. (2008, 156263) examining the association between PM and both lung cancer mortality
and incidence. After ten years of followup, there was no association between PM2.5 and lung
cancer mortality for either the analysis of the full cohort (n = 105,296) [RR 1.06 (95% CI
0.82-1.38) per 10 |ug/m3 increase in PM2.5] or the case-cohort (n = 4075) [RR 0.87 (95% CP
0.52-1.47)]. There was also no association with black smoke or traffic density variables,
although living near a major roadway was associated with an elevated relative risk for lung
cancer in the full cohort analysis [RR 1.20 (95% CP 0.98-1.47)]. The association was not
present in the case-cohort analysis [RR 1.07 (95% CP 0.70-1.64)].
In addition to lung cancer mortality, Brunekreef et al. (2009, 191947) also examined
the association with lung cancer incidence using 11.3 years of followup data. In both the
full cohort and the case-cohort analyses no association was reported between PM2.5 and lung
cancer incidence [full cohort: RR 0.81 (95% CP 0.63-1.04); case-cohort: RR 0.67 (95% CP
0.41-1.10) per 10 jug/m3 increase in PM2.5]. The same was true for analyses of black smoke
and traffic density variables.
Table 7-7. Associations* between ambient PM concentrations from select studies of lung cancer
mortality and incidence.



Study
Cohort
Location
Yrs
Analysis subgroup Effect Estimate (95% CI)
MORTALITY PMzs
Dockerv et al. (1993, 044457)*
Six-Cities
Six cities across the U.S.
1974-1991
1.18(0.89-1.57)
Krewski et al. (2000,012281)*
Six-C it ies- Re-analy sis
Six cities across the U.S.
1974-1991
1.16(0.86-1.23)
Laden et al. (2006, 087605)
Six-Cities
Six cities across the U.S.
1974-1998
1.27 (0.96-1.69)
Beelen et al. (2008,156263)
NLCS
Netherlands
1987-1996
Full Cohort 1.06(0.82-1.38)
July 2009
7-103
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Cohort
Location
Yrs
Analysis subgroup
Effect Estimate (95% CI)
Beelen et al. (2008,156263)
NLCS
Netherlands
1987-1996
Case Cohort
0.87 (0.52-1.47)
Brunekreef et al. (2009,191947)
NLCS-Re-analysis
Netherlands
1987-1996
Full Cohort
1.06(0.82-1.38)
Brunekreef et al. (2009,191947)
NLCS-Re-analysis
Netherlands
1987-1996
Case Cohort
0.87 (0.52-1.47)
Pope et al. (1995, 045159)*
ACS
U.S.
1982-1989

1.01 (0.91-1.12)
Pope et al. (2002, 024689)*
ACS
U.S.
1982-2000

1.13(1.04-1.22)
Jerret et al. (2005,189405)
ACS-LA
Los Angeles
1982-2000
Intra-metro Los Angeles
1.44(0.98-2.11)
Krewski et al. (2009,191193)
ACS-Re-analysis
U.S.
1982-2000

1.11 (1.04-1.18)
Krewski et al. (2009,191193)
ACS-Re-analysis
New York City
1982-2000
Intra-metro New York City
0.90 (0.29-2.78)
Krewski et al. (2009,191193)
ACS-Re-analysis
Los Angeles
1982-2000
Intra-metro Los Angeles
1.31 (0.90-1.92)
McDonnell et al. (2000,010319)*
AHSM0G
California
1973-1977
Men
1.39 (0.79-2.46)
Naess et al. (2007, 090736)*

Oslo, Norway
1992-1998
Men, 51-70 yrs
1.18(0.93-1.52)
Naess et al. (2007, 090736)*

Oslo, Norway
1992-1998
Men, 71-90 yrs
1.18(0.93-1.52)
Naess et al. (2007, 090736)*

Oslo, Norway
1992-1998
Women, 51-70 yrs
1.83 (1.36-2.47)
Naess et al. (2007, 090736)*

Oslo, Norway
1992-1998
Women, 71-90 yrs
1.45 (1.05-2.02)
MORTALITY PMw
McDonnell et al. (2000,010319)*
AHSM0G
California
1973-1977
Men
1.23 (0.84-1.80)
Naess et al. (2007, 090736)*

Oslo, Norway
1992-1998
Men, 51-70 yrs
1.12(0.95-1.33)
Naess et al. (2007, 090736)*

Oslo, Norway
1992-1998
Men, 71-90 yrs
1.14(0.97-1.36)
Naess et al. (2007, 090736)*

Oslo, Norway
1992-1998
Women, 51-70 yrs
1.50 (1.23-1.84)
Naess et al. (2007, 090736)*

Oslo, Norway
1992-1998
Women, 71-90 yrs
1.29 (1.03-1.60)
INCIDENCE -PMzs
Beelen et al. (2008,155681)
NLCS
Netherlands
1987-1996
Full Cohort
0.81 (0.63-1.04)
Beelen et al. (2008,155681)
NLCS
Netherlands
1987-1996
Case Cohort
0.65 (0.41-1.04)
Brunekreef et al. (2009,191947)
NLCS-Re-analysis
Netherlands
1987-1996
Full Cohort
0.81 (0.63-1.04)
Brunekreef et al. (2009,191947)
NLCS-Re-analysis
Netherlands
1987-1996
Case Cohort
0.67 (0.41-1.10)
INCIDENCE - PMw
Beesn et al. (1998, 048890)
AHSM0G
California
1977-1992
Men
1.99 (1.32-3.00)
Vineis et al. (2006.192089)
GenAir
Europe
1993-1999
Case-Control
0.91 (0.70-1.18)
*per 10 /yg/m3 increase
tResults from the paper were standardized to10 /yg/m3 [For McDonnell et al. the non-standardized results were reported based on IQR increments (24.3 /yg/m3 for PM2.5 and 29.5 /yg/m3 for PM10).
For Naess et al. the original hazard ratios were calculated based on quartiles of PM exposure. The results were converted to10 /yg/m3 using the mean range of the four quartiles (3.95 /yg/m3 for
PM2.5 and 5.88 /yg/m3 for PM10)].
iStudy was included in the 2004 PM AQCD
The association between PM and incident lung cancers was examined in the European
Prospective Investigation into Cancer and Nutrition study (EPIC) (Vineis et al., 2006,
192089). Within this cohort, a nested case-control study, the GenAir study, included cases of
incident cancer and controls matched on age, gender, smoking status, country of
recruitment, and time between recruitment and diagnosis. Only non-smokers and former
smokers who had quit smoking at least ten years prior were included. The study included
113 cases and 312 controls. No association was seen between PMio and lung cancer [OR
0.91 (95% CP 0.70-1.18) per 10 |ug/m3]. The OR was elevated when cotinine, a marker for
cigarette exposure, was included in the model but the authors state that this is probably
due to small study size [OR 2.85 (95% CP 0.97-8.33) comparing > 11 jug/m3 to <11 |ug/m3].
July 2009
7-104
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Control for other potential confounders, such as BMI, education level, and intake of fruit
and vegetables, did not have a large impact on the estimate.
7.5.1.2.	Other Cancers
Bonner et al. (2005, 088993) conducted a population-based, case-control study of the
association between ambient exposure to polycyclic aromatic hydrocarbons (PAHs) in early
life and breast cancer incidence among women living in Erie and Niagara counties in the
state of New York. Cases (n = 1166 of which 841 were post-menopausal) were women with
primary breast cancer, and controls (n = 2105 of which 1495 were post-menopausal) were
frequency matched to the cases by age, race, and county of residence. TSP was used as a
proxy for PAH exposure. Annual average TSP concentrations (1959-1997) were obtained
from the New York State Department of Environmental Conservation for Erie and Niagara
Counties. Among postmenopausal women, exposure to high concentrations of TSP
(>140 |ug/m3) at birth was associated with an OR of 2.42 for breast cancer (95% CP 0.97-
6.09) relative to low concentrations of TSP (<84 jug/m3). ORs were elevated for pollution
exposures at age ofmenarche (OR: 1.45 [95% CP 0.74-2.87]) and age at first birth (OR: 1.33
[95% CP 0.87-2.06]) among postmenopausal women. Among premenopausal women,
exposure to high concentrations of TSP at birth was associated with an OR for breast
cancer incidence of 1.79 (95% CP 0.62-5.10) relative to low exposure levels, exposure at age
of menarche was associated with an OR of 0.66 (95% CP 0.38-1.16), and exposure at age of
first birth was associated with an OR of 0.52 (95% CP 0.22-1.20).
7.5.1.3.	Markers of Exposure or Susceptibility
Several studies looked at markers of exposure or susceptibility as the outcome
associated with short-term exposure. These studies are included here because they may be
relevant to the mechanism that leads to cancer associated with long-term exposures.
A study performed in the Czech Republic compared 53 male policemen working at
least eight hours per day outdoors in urban air with age- and sex-matched controls who
spent at least 90% of their day indoors (n = 52) (Sram et al., 2007, 188457). During the
sampling period, two monitors from downtown and suburban areas detected levels of air
pollutants in the following ranges: PMio 32-55 |ug/m3, P2.5 27"38 |ug/m3, c-PAHs 18-22 ng/m3,
and B[a]P 2.5-3.1 ng/m3 using a VAPS monitor (measurements taken with a HiVol monitor,
which has a lower flow rate, had a mean for PM10 of 62.6 jug/m3). c-PAHs detected on
personal monitors during sampling days had a mean of 12.04 ng/m3 among the policemen
and 6.17 ng/m3 among the controls. No difference in percent of chromosomal aberrations
was observed between the policemen and control group using conventional cytogenetic
analysis. However, using fluorescent in situ hybridization (FISH), a difference in
chromosomal aberrations between the policemen and control group was reported. For
example, the percentage of aberrant cells, as well as the genomic frequency of
July 2009
7-105
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
translocations per 100 cells, was about 1.4-fold greater in the policemen. This was largely
driven by a difference in chromosomal aberrations between nonsmoking policemen and
nonsmoking controls. A similar study that included only the policemen (n = 60), reported
that the mean exposure to c-PAHs and B[a]P for 40-50 days before sampling were
associated with chromosomal aberrations when analyzed with FISH (Sram et al., 2007,
192084). However, when included in a model with other covariates, the association with
these variables was null. No association was present with use of conventional cytogenetic
analysis.
Palli et al. (2008, 156837) investigated the correlation between ambient PMio
concentrations and individual levels of DNAbulky adducts. Study participants were 214
healthy adults aged 35-64 years at enrollment who resided in the city of Florence, Italy.
This study was conducted between 1993 and 1998. PMio exposure levels were based on
daily environmental measures provided by two types of urban monitoring stations (high-
traffic and lowtraffic). The researchers assessed correlation between DNAbulky adducts
measured in blood samples and PMio concentrations prior to blood sample collection. Time
windows of PMio exposure evaluated in this study were 0-5 days, 0-10 days, 0-15 days, 0-30
days, 0-60 days, and 0-90 days prior to blood sample collection. Overall, average PMio
concentrations decreased during the study period, with some fluctuations. Quantitative
values were not reported, but PMio appeared to range between approximately 30 and
100 ug/m3 lor high-traffic stations, and between approximately 20 and 50 ug/ma lor low-
traffic stations. This study found that levels of DNAbulky adducts among non-smoking
workers with occupational traffic exposure were positively correlated with cumulative PMio
levels from high-traffic stations during approximately 2 wk preceding blood sample
collection (0-5 days: r = 0.55, p = 0.03; 0-10 days: r = 0.58, p = 0.02; 0-15 days: r = 0.56,
p = 0.02). DNAbulky adducts were not associated with PMio levels among Florence
residents with no occupational exposure to vehicle emissions or among smokers. DNAbulky
adducts were not associated with PMio levels assessed by low-traffic urban monitoring
stations.
The association between personal exposure to water-soluble transition metals in PM2.5
and oxidative stress-induced DNA damage was investigated among 49 students from
Central Copenhagen, Denmark (Sorensen et al., 2005, 188901). Researchers assessed PM2.5
exposure by personal sampling over two weekday periods twice in 1-yr (November, 1999
and August, 2000), and determined the concentration of water-soluble transition metals (V.
Cr, Fe, Ni, Cu and Pt) in these samples. In addition, lymphocyte and 24-h urine samples
were analyzed for DNA damage by measuring 7-hydro-8-oxo-2'-deoxyguanosine (8-oxodG).
Mean concentrations and corresponding IQR of these metals differed between months of
sample collection. This study found that 8-oxodG concentration in lymphocytes was
significantly associated with V and Cr concentrations, with a 1.9% increase in 8-oxodG per
1 jug/L increase in V concentration and a 2.2% increase in 8-oxodG per 1 jag/L increase in Cr
July 2009
7-106
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
concentration; these associations were independent of the PM2.5 mass concentration. The
other transition metals were not significantly associated with the 8-oxodG concentration in
lymphocytes, and none of the six measured transition metals was associated with the 8*
oxodG concentration in urine.
Vinzents et al. (2005, 087482) investigated the association between ultrafine particles
(UFP) and PM10 concentrations with levels of purine oxidation and strand breaks in DNA
using a crossover design. Study participants were 15 healthy nonsmoking individuals with
a mean age of 25. UFP exposure was evaluated using number concentration obtained in the
breathing zone by portable instruments in six 18"h weekday periods from March to June
2003. Ambient concentrations for PM10 and UFP were also measured on all exposure days
at curbside street stations and at one urban background station. Oxidative DNA damage
was assessed by evaluating strand breaks and oxidized purines in mononuclear cells
isolated from venous blood the morning after exposure measurement. Mean number
concentration of UFPs (street station) was 30.4 x 103UFPs/mL (standard deviation [SD]:
1.38), mean mass concentration of PM10 at a background monitoring station was 16.9 |ug/m3
(SD: 1.53), and mean mass concentration of PM10 at a street station was 23.5 |ug/m3 (SD:
1.48). Mean personal exposure to UFPs was 32.4 x 103 UFPs/mL (SD: 1.49) while bicycling
(5 occasions), 19.6 x 103 UFPs/mL (SD: 1.78) during other outdoor activities (6 occasions),
and 13.4 x 103 UFPs/mL (SD: 1.96) while indoors (6 occasions). The regression coefficients of
the mixed-effects models looking at level of purine oxidation were estimated as 1.50 x 10~3
(95% CI: 0.59 x 10~3 to 2.42 x 10~3; p = 0.002) for cumulative outdoor exposure and 1.07 x
10"3 (95% CI: 0.37 x 10-3 to 1.77 x 10~3! p = 0.003) for cumulative indoor exposure. Neither
cumulative outdoor nor cumulative indoor exposures to UFPs were associated with strand
breaks. Neither ambient air concentrations of PM10 nor number concentrations of UFPs at
monitoring stations were significant predictors of DNA damage.
Additionally, a number of studies employed ecologic study designs, comparing the
prevalence of biomarkers in populations from more polluted locations to those in less
polluted locations. In a pilot study conducted in the Czech Republic (Pedersen et al., 2006,
156848), children age 5-11 provided 5 mL blood samples and the frequency of micronuclei in
peripheral blood lymphocytes was analyzed for cytogenetic effects. Significantly higher
frequencies of micronuclei were found in younger children living in Teplice (PM2.5
concentration = 120 ug/m") than in Prachatice (PM2.5 concentration = 46 ug/m3). The levels
of c-PAHs were also much higher in Teplice (nearly 30 ng/m3 in Teplice and about 15 ng/m3
in Prachatice). The difference in micronuclei frequencies observed in the children from the
two locations may be attributable to differences in exposure to air pollution, but could also
be due to differences in diet or other environmental exposures. This finding is noteworthy
considering micronuclei formation in peripheral blood lymphocytes is thought to be
biologically relevant for carcinogenesis.
July 2009
7-107
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Avogbe et al. (2005, 087811) showed that there appeared to be a correlation between
the level of oxidative DNA damage in individuals and exposure to ambient ultrafine
particulate matter. Formamidopyrimidine DNA glycosylase sensitive sites and the presence
of DNA strand breaks were assessed from blood and urine samples obtained from healthy,
non-smoking male volunteers that lived and worked in different areas of Cotonou, Benin.
Exposure to benzene was assessed by urinary excretion of S-phenylmercapturic acid. There
was a high degree of correlation between exposure to benzene and ultrafine particles and
the presence of DNA strand breaks and formamidopyrimidine DNA glycosylase sensitive
sites (rural subjects < suburban subjects < residents living near high traffic roads < taxi
drivers). Genotyping studies showed that the magnitude of the effects of benzene and
ultrafine particulates may be modified by polymorphisms in GSTP1 and NQOl genes.
Tovalin et al. (2006, 091322) evaluated the association between exposure to air
pollutants and the level of DNA damage using the single cell gel electrophoreis (comet)
assay. Mononuclear lymphocytes from outdoor and indoor workers from two areas in
Mexico, Mexico City (large city) and Puebla (medium size city), were evaluated. The
outcomes showed that the outdoor workers in Mexico City exhibited greater DNA damage
than indoor workers in the same region. Similar levels of DNA damage were observed
between indoor and outdoor workers in Puebla. The level of observed DNA damage was
correlated with exposure to ozone and PM2.5.
In summary, several recent studies have reported an association between lung cancer
mortality and long-term PM2.5 exposure. Although many of the estimates include the null in
the confidence interval, overall the results have shown a positive relationship. The two
recent studies that looked at lung cancer incidence did not report an association with PM2.5
(Brunekreef et al., 2009, 191947) or PM10 (Vineis et al., 2006, 192089). Studies of
exposure/susceptibility markers have reported inconsistent outcomes, with some markers
being associated with PM and others not.
7.5.2. Toxicological Studies
Over the past 30 years numerous mutagenicity and ge no toxicity studies of ambient
PM and their contributing sources have been conducted to assess the relative mutagenic or
genotoxic potential. Studies previously reviewed in the 2004 PM AQCD (U.S. EPA, 2004,
056905) provide compelling evidence that ambient PM and PM from specific combustion
sources (e.g., fossil fuels) are mutagenic in vivo and in vitro. Research cited in the 2004
AQCD demonstrated mutagenic activity of ambient PM from urban centers in California,
Germany and the Netherlands. These studies suggested that ubiquitous emission sources,
particularly motor vehicle emissions, rather than isolated point sources were largely
responsible for the mutagenic effects. In addition, the mutagenicity was dependent upon
the chemical composition of the PM with unsubstituted polyaromatic compounds and semi-
July 2009
7-108
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
polar compounds being highly mutagenic. Mutagenicity was also dependent on size, with
the fine fraction of urban PM having greater effects than the coarse fraction. Genotoxic
activity was demonstrated for ambient PM from two high traffic areas (one upwind and one
downwind) and a rural site. In addition, the 2004 AQCD reported that exhausts from
gasoline and diesel engines were mutagenic and that DE was more potent. More
mutagenicity was observed for exhaust from cold starts than starts at room temperature.
Both gaseous and particulate fractions of DE were found to be mutagenic. Sequential
fractionation of extracts from gasoline and diesel exhaust implicated the polar fractions,
especially nitrated polynuclear aromatic compounds, as contributing greatly to
mutagenicity. Among some of the other mutagenically active compounds found in the gas
phase of DE are ethylene, benzene, 1.3-butadiene, acrolein and several PAHs, all of which
are also present in gasoline exhaust. Also cited in the 2004 AQCD were studies
demonstrating mutagenic effects of emissions from wood/biomass burning, which were
primarily attributable to the organic fraction and not the condensate. It was noted that
wood smoke induced both frameshift mutations and base pair substitution but not DNA
adducts. Further, emissions from coal combustion in China were found to be mutagenic,
with both polar and aromatic fractions contributing to effects. Little data were available on
the mutagenicity of coal fly ash emissions from U.S. conventional combustion plants. In
conclusion, these studies provide evidence that ambient PM and combustion-derived PM
are mutagenic/genotoxic. The 2004 AQCD noted that there is not a simple relationship
between mutagenic potential and carcinogenic potential in animals or humans. No studies
evaluating carcinogenic effects of PM were reported in the 2004 AQCD.
Building on results of earlier studies in the 2004 PM AQCD, data from newly
published studies that evaluated the mutagenic, genotoxic and carcinogenic effects of PM,
1'M-constituents, and combustion emission source particles are reviewed. Pertinent studies
are described briefly in the following paragraphs. A summary table is provided in Annex D
(Tables D7-D9).
7.5.2.1. Mutagenesis and Genotoxicity
In Vitro Studies
In general, studies have focused on PM and PM extracts for mutagenicity testing
using bacteria and mammalian cell lines. PM and/or PM extracts from ambient air samples,
wood smoke, and coal, diesel, or gasoline combustion have all been reported to induce
mutation in S. typhimurium and in cultured human cells (Abou Chakra et al., 2007,
098819; Gabelova et al., 2007, 156457; Gabelova et al., 2007, 156458; Hannigan et al., 1997,
083598; Hornberg et al., 1998, 155849). In addition, effects associated with PM and PM-
associated constituents include induction of micronucleated reticulocytes (MN) formation,
DNA adduct formation, SCE, DNA strand breaks, frameshifts and inhibition of gap-
July 2009
7-109
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
junctional intercellular communication (Alink et al., 1998, 087159; Arlt et al., 2007, 097257;
Avogbe et al., 2005, 087811; Gabelova et al., 2007, 156457; Gabelova et al., 2007, 156458;
Healey et al., 2006, 156532; Hornberg et al., 1996, 087164; Hornberg et al., 1998, 155849;
Sevastyanova et al., 2007, 156969)
Constituents adsorbed onto individual particles play a large role in the genotoxic
potential of PM. Poma et al. (2006, 096903) showed that fine carbon black particles were
consistently less genotoxic than similar concentrations of PM2.5 extracts, suggesting that
the adsorbed components play a role in the genotoxic potential of PM. Total PAH and
carcinogenic PAH content is correlated with the genotoxic effects of PM (De Kok et al.,
2005, 088656; Sevastyanova et al., 2007, 156969). Comparison of different extracts (water
vs. organic) by Gutierrez-Castillo et al. (2006, 089030) indicated that water soluble extracts
were more genotoxic than the corresponding organic extracts. Sharma et al. (2007, 156975)
reported that mutagenic activity of extracted PM samples collected in and around a waste
incineration plant was attributed to the moderately polar and polar fractions. The polar and
crude fractions were mutagenic without metabolic activation, suggesting a direct mutagenic
effect. No mutagenic activity was observed from any of the nonpolar samples evaluated.
Arlt and colleagues (2007, 097257) have shown that the PM constituents 2-
nitrobenzanthrone and 3-nitrobenzanthrone were genotoxic in a variety of bacterial and
mammalian cell systems.
Conflicting data have been reported for the role of metabolic enzymes on the
genotoxicity of PM and their adsorbed constituents. Arlt et al. (2007, 097257) reported that
the PM constituent 2-nitrobenzanthrone (2-NB) was genotoxic in bacterial and mammalian
cells. However, metabolic activation with the human N-acetyl transferase 2 or
sulfotransferase (SULT1A1) enzyme was needed for the effect to be observed in human
cells. Erdinger et al. (2005, 156423) demonstrated that mutagenic activity was not affected
when metabolism was induced, de Kok et al. (2005, 088656) evaluated the relationship
between the physical, chemical, and genotoxic effects of ambient PM. TSP, PM10, and PM2.5
were sampled at different locations and the extracts were assessed for mutagenicity and
induction of DNA adducts in cells. Overall, induction of rat liver S9 metabolism generally
reduced the mutagenic potential via the Ames assay of the particle fractions and DNA
reactivity (induction of DNA adducts) was generally higher after metabolic activation.
Binkova et al. (2003, 156274) found that the addition of S9 increased PMicrdependent DNA
adduct formation.
Ambient Air
A limited number of studies evaluated the impact of the season on the genotoxic
effects of ambient PM. A few studies however have indicated that greater genotoxic effects
were associated with samples collected during the winter months compared to those
collected in the summer (Abou Chakra et al., 2007, 098819; Gabelova et al., 2007, 156457;
Gabelova et al., 2007, 156458). In contrast, Hannigan et al. (1997, 083598) indicated that
July 2009
7-110
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
no seasonal variation was observed. Studies have also shown that greater genotoxic effects
were associated with smaller particle size extracts (e.g., PM2.5>PMio) and from samples
collected in urban areas or closer to higher trafficked areas (Abou Chakra et al., 2007,
098819; Hornberg et al., 1998, 155849).
de Kok et al. (2005, 088656) found the direct mutagenicity (Ames assay) and the
direct DNA reactivity (DNA adduct formation) of the PM2.5 size fraction was significantly
higher than that of the larger size fractions (TSP, PM10) at most locations.
DNA damage was assessed by the Comet assay in A549 cells exposed to PM collected
from a high traffic area in Copenhagen, Denmark (TSP approximately 30 ug/m3) and
compared to the results from exposure of A549 cells to standard reference materials
(SRM1650 or SRM2975) at the same concentrations (2.5-250 |ug/ml) (Danielsen et al., 2008,
192092). All three particles induced strand breaks and oxidized purines in a dose-dependent
manner and there were no obvious differences in potency. In contrast, only the ambient PM
formed 8-oxo-7,8"dihydro-2'-deoxyguanosine (8-oxodG) when incubated with calf thymus
DNA, which may be due to the levels of transition metals.
Diesel and Gasoline Exhaust
Automobile DEP (A-DEP) was tested in S. typhimurium strains TA98, TA100, and its
derivatives (e.g., TA98NR and YG1021) and found to be more mutagenic than forklift DEP
(i.e., SRM2975) particles, based on PM mass. A-DEP had 227 times more PAH-type
mutagenic activity and 8-45 times more nitroarene-type mutagenic activity due to the
different conditions for generating and collecting the two DEP samples (DeMarini et al.,
2004, 066329). Using a diesel engine without an oxidation catalytic converter (OCC), the
diesel engine exhaust particle (DEP) extract produced the highest number of revertant
colonies in strains TA98 and TA100 with and without S9 at several tested loads when
compared to extracts from lowsulfur diesel fuel (LSDF), rapeseed oil methyl ester (RME),
and soybean oil methyl ester (SME). When an OCC was installed in the exhaust pipe of the
engine, all extracts reduced the number of revertant colonies in both strains with and
without S9 at partial loads but increased the number of revertant colonies without S9 at
rated power. At idling, DEP extracts increased the number of revertant colonies with and
without S9 (Bunger et al., 2006, 156303). In a separate study, engine emissions (particle
extracts and condensates) from rapeseed (canola) oil were found to produce greater
mutagenic effects in S. typhimurium strains TA98 and TA100 than DEP (Bunger et al.,
2007, 156304). Additionally, DE extract (DEE) from diesel fuel containing various
percentages of ethanol was also observed to induce mutational response in two Salmonella
strains. Base diesel fuel DEE and DEE from fuel with 20% ethanol caused more significant
DNA damage in rat fibrocytes L-929 cells than extracts containing 5, 10, or 15% ethanol
(Song et al., 2007, 188903).
DE and gasoline engine exhaust particles, as well as their semi-volatile organic
compound (SVOC) extracts, induced mutations in the two S. typhimurium strains YG1024
July 2009
7-111
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
and YG1029 in the absence and presence of S9; the PM extracts were more mutagenic than
the SVOC extracts. Additionally, all extracts except the DE extract induced DNA damage
and MN in Chinese hamster lung V79 cells (Liu et al., 2005, 192098). Another study
demonstrated that gasoline engine exhaust significantly increased colony formation in
TA98 with and without S9 (Zhang et al., 2007, 157186).
Jacobsen et al. (2008, 156597) used the FEl-Muta™ Mouse lung epithelial cell line to
investigate putative mechanisms of DEP-induced mutagenicity. Mutation ion frequencies
and ROS were determined after cells were incubated with 37.5 or 75 ug/ml DEP (SRM1650)
for 72-h (n = 8). The mutation frequency at the 75 ug/ml dose was significantly increased
(l.55-fold; p<0.00l) in contrast to cells treated with 37.5 |ug/ml DEP. DEP induced ROS
generation 1.6-1.9-fold in the epithelial cell cultures after 3 hours of exposure compared
with the 3-10-fold increase in ROS production previously reported for carbon black. The
authors concluded that the mutagenic activity of DEP is likely attributable to activity from
the organic fraction that both contains reactive species and can generate ROS.
In human A549 and CHO-K1 cells, the organic fraction of DEP significantly increased
the amount of Comet and MN formation, respectively, in the presence and absence of SKF-
525A (a CYP450 inhibitor) and S9, respectively (Oh and Chung, 2006, 088296). The organic
base and neutral fractions of DEP also significantly induced DNA damage but only without
SKF-525A, and all fractions but the moderately polar fraction (phthalates and PAH
oxyderivatives) induced MN formation with and without S9 (Bao et al., 2007, 097258).
Gasoline engine exhaust significantly induced DNA damage as measured in the Comet
assay and increased the frequency of MN in human A549 cells (Zhang et al., 2007, 157186).
In human-hamster hybrid (Al) cells, DEP (SRM 2975) dose-dependently increased the
mutation yield at the CD59 locus! this was significantly reduced by simultaneous treatment
with phagocytosis inhibitors (Bao et al., 2007, 097258).
Wood Smoke
The mutagenicity of wood smoke (WS) and cigarette smoke (CS) extracts was assayed
in Salmonella typhimuriuni strains TA98 and TA100 (Ames assay) using the pre-incubation
assay with exogenous metabolic activation (rat liver S-9). Extracts of both samples (62.5 or
125 ug TPM equivalent/ml) were equally mutagenic to strain TA98 but the WS extract was
less mutagenic than the CS extracts in strain TA100 (Iba et al., 2006, 156582).
In vivo studies
Ambient Air
The contribution of ambient urban roadside air exposure (4, 12, 24, 48 or 60 wk) to
DNA damage was examined in the lungs, nasal mucosa, and livers of adult male Wistar
rats in Kawasaki, Japan (Sato et al., 2003, 096615). Messenger RNA levels of CYP450
enzymes that catalyze the transformation of PAHs to reactive metabolites were also
July 2009
7-112
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
evaluated. Concentrations of gases were reported to be 12-182 ppb NO and 0-9 ppb NO2 in
the filtered air chamber and 33-280 ppb NO and 42-81 ppb NO2 in the experimental group
chamber. Suspended PM concentrations were 11-19 ug/m3 in the filtered air chamber and
42-100 |ug/m3 (average 63 ug/m3) in the experimental group chamber. Body weight
significantly decreased in exposed animals at 24, 48 and 60 wk. Exposure of 4 wk to urban
roadside air resulted in significant increases in multiple DNA adducts (lung, nasal, and
liver DNA adducts). With longer exposures, there were significant increases in lung (48 wk),
nasal (60 wk), and liver DNA adducts (60 wk). Changes were seen in CYP1A2 mRNA at
4 wk with a 2.3 fold increase in exposed animals compared to the control group with no
change observed at 60 wk; CYP1A1 mRNA was unchanged. These results indicate that
exposure to ambient air in this roadside area could induce DNA adduct formation, which
may be important for carcinogenicity. Earlier studies (Ichinose et al., 1997, 053264) have
shown that 8-oxo-dG, a DNA adduct, is elevated along with tumor formation in a dose-
dependent manner in mice administered DEP. The finding of adducts in the liver indicated
that deposition of PM and its associated PAHs in the lung can have indirect effects on
extrapulmonary organs. It should be noted that PM deposition on the fur and ingestion
during grooming cannot be ruled out as a possible exposure route.
Another animal toxicological study employed "non-carcinogenic" particles obtained
from pooled non-cancerous lung tissue collected during surgical lung resection from three
non-smoking male patients diagnosed with lung adenocarcinomas (Tokiwa et al., 2005,
191952). Particles were partially purified to remove organic compounds. Morphologically
the particles were similar to DE or ambient air PM and the organic extracts from the
particles were directly mutagenic in S. typhimurium tester strains TA98, YG1021 and
YG1024. BALB/c and ICR mice were intratrachaelly instilled with particles at doses of 0.25,
0.5, 1.0, or 2.0 mg/mouse. After 24 hours, 8-oxo-dG was measured in lung DNA and found to
be increased in ICR mice in a dose-dependent manner, reaching a maximum of-2.75 8-oxo-
dG/105 dG at the 2.0 mg dose. The response was statistically significant at doses of 0.5, 1.0,
and 2.0 mg. The increased 8-oxo-dG levels observed in vivo was reported to be likely due to
hydroxyl radicals presumed to be involved in phagocytosis of non-mutagenic particles by
inflammatory cells that could induce hydroxylation of guanine residue on DNA.
Diesel Exhaust
An in vivo study employed gtp delta transgenic mice carrying the lambda EG10 on
each Chromosome 17 from a C57BL/6J background to investigate the effects of DEP on
mutation frequency (Hashimoto et al., 2007, 097261). Mice were exposed via inhalation to
DEP or via IT instillation to DEP or DEP extract and lambda EG10 phages were rescued!
E. coli YG6020was infected with the phage and screened for 6-thioguanine resistance. The
mutagenic potency (mutation frequency per mg) caused by DEP extract was twice that of
DEP, suggesting that the mutagenicity of DEP is attributed primarily to compounds in the
extract, since «50% of the weight of DEP was provided by the extract. Interestingly, there
July 2009
7-113
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
was no difference in mutation frequency between the 1 and 3 mg/m3 DEP groups after
12 wk of exposure.
Wood Smoke
One recent study measured the effect of freshly generated hard wood smoke (HWS) on
CYP1A1 activity based on ethoxyresorufin Odeethylase in pulmonary microsomes
recovered from male Sprague-Dawley rats exposed to HWS by nose-only inhalation
exposure (Iba et al., 2006, 156582). CYP1A1 activity in rat lung explants treated with
extracts of the total PM (TPM) from HWS samples and from freshly generated cigarette
smoke (CS) was also evaluated. Unlike CS, HWS did not induce pulmonary CYP1A1
activity or mRNA (assessed by northern blot analysis) nor did extracts of HWS TPM induce
CYP1A1 protein (assessed by western blot analysis) in cultured rat lung explants. The
results suggest that unique constituents that are activated by CYP1A1 may be present in
CS but not HWS.
7.5.2.2. Carcinogenesis
Studies published prior to the 2004 AQCD that evaluated the carcinogenicity of
ambient air were reviewed by Claxton and Woodall (2007, 180391). Five studies involved
chronic inhalation exposures in rodents. No statistically significant increase in
tumorigenesis was observed following chronic exposure to urban air pollution in Los
Angeles (Gardner, 1966, 015129; Gardner et al., 1969, 015130; Wayne and Chambers, 1968,
038537). However in a study conducted in Brazil, urban air pollution was found to enhance
the formation of urethane-induced lung tumors in mice (Cury et al., 2000, 192100; Reymao
et al., 1997, 084653).
Two recent studies evaluated the carcinogenic potential of chronic inhalation
exposures to DE (Reed et al., 2004, 055625) and hardwood smoke (HWS) (Reed et al., 2006,
156043). Two indicators of carcinogenic potential, formation of MN and tumorogenesis were
measured in strain A/J mice, which is a mouse model that spontaneously develops lung
tumors. Exposure to DE or HWS at concentrations of 1000 pg/m3 and below did not cause
increased formation of MN or an increased rate of lung tumors in this cancer-prone rodent
model. These studies are described below.
Diesel Exhaust
A/J mice were exposed to 30, 100, 300 and 1000 pg/m3 DE for 6 h/day and 7 days/wk
for 6 months (Reed et al., 2004, 055625). The concentration of gases in this including NOx,
NO2, CO, SO2, NH3, methane, non-methane VOC, and FID total hydrocarbon ranged from
control to high dose group values of 0 to 50.4 ± 0.6 ppm, 0.2 ± 0.2 to 6.9 ± 3.3 ppm, 0.3 ±0.1
to 30.9 ± 4.5 ppm, not detectable to 955.2 ± 58.4 ppb, 176.5 ± 8.8 to 9.1 ± 0.2 |ug/m3, 1406.5 ±
253.2 to 2642.1 ± 455.9 |ug/m3, 134.0 ± 52.1 to 1578.6 ± 256.2 |ug/m3, 0.1 ± 0.1 to 2.2 ± 0.2
ppm, respectively. Particle sizes in the 4 exposure groups ranged from 0.10-0.15 pm mass
July 2009
7-114
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
median aerodynamic diameter with geometric standard deviations of 1.4-1.8. Following the
6 month exposure and a 6 month recovery period, mice were collected and MN formation in
blood and tumor multiplicity and tumor incidence were measured in lungs. No increases in
formation of MN or numbers of lung adenomas were observed in DE-exposed mice
compared with controls.
Wood Smoke
A/J mice were exposed to 30, 100, 300 and 1000 pg/m3 HWS for 30, 100, 300 and 1000
pg/m3 DE for 6 h/day and 7 days/wk for 6 month (Reed et al., 2006, 156043). Gaseous
components of the HWS included CO, NH3, and non-methane VOC with concentrations
from control levels to high dose HWS exposure ranging from 229 ± 31 to 14887.6 ± 832.3
ppm, 139.3 ± 2.3 to 54.9 ± 1.2 jug/m3 and 177.6 ± 10.4 to 3455.0 ± 557.2 |ug/m3, respectively.
Concentrations of NOx, NO2 and SO2 were reported to be null. Particle sizes in the 4 exposure groups
ranged from 0.25-0.36 pm mass median aerodynamic diameter with geometric standard
deviations of 2.0-3.3. Following the 6 month exposure and a 6-month recovery period, mice
were collected and MN formation in blood and tumor multiplicity and tumor incidence were
measured in lungs. No increases in formation of MN or numbers of lung adenomas were
observed in HWS-exposed mice compared with controls. However, HWS from this study was
mutagenic in the Ames reverse mutation assay.
In summary, numerous new in vitro studies confirm and extend findings reported in
the 2004 AQCD that ambient PM from urban sites and combustion-derived PM are
mutagenic and genotoxic. A small number of new studies were conducted in vivo. One of
these studies demonstrated increased mutagenic potency in mice exposed to DEP and DEP
extract. Another study found increased formation of 8-oxo-dG, a DNA adduct, following
intratracheal instillation of PM in mice. A chronic inhalation study of rats exposed to urban
roadside air reported increased formation of DNA adducts in nose, lung, and liver and
induction of CYP1A2. Inhalation exposure of rats to HWS failed to induce CYP1A1 in
another study. Finally, two chronic inhalation studies found no evidence of carcinogenic
potential for DE and HWS in a cancer-prone mouse model. Collectively, these results
provide some evidence, mainly from in vitro studies, to support the biological plausibility of
ambient PM-lung cancer relationships observed in epidemiology studies.
7.5.3. Epigenetic Studies and Other Heritable DNA mutations
Two epidemiologic epigenetic studies examined the effect of PM on DNA methylation. Both
studies examined methylation of Alu and long interspersed nuclear element-1 (LINE-1) sequences, which
are located in repetitive elements. In previous studies, methylation of these sequences has been linked to
global genomic DNA methylation content (Weisenberger et al., 2005, 192101; Yang et al., 2004,
192102).
July 2009
7-115
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
The first study included men age 55 and older who were part of the Normative Aging Study in the
Boston area. A stationary monitoring site located 1 km from the examination site was used to estimate
ambient PM2 5 exposure for the duration of the study (1999-2007). During the study period, the median
level of PM25j averaged over 7 day periods, was 9.8 |ig/m3 (interquartile range 8.0-12.0 |ig/m3). There
was no association between PM25 and Alu methylation. LINE-1 methylation was associated with PM2 5
measured over the 7 days before the examinations.
The second study included 63 healthy men aged 27 to 55 working at an electric furnace steel
plant (Tarantini et al., 2009, 192010). Blood samples were taken twice, once in the morning after two
days of not working and once in the morning after three full days of work. PMi0 was measured in 11 work
areas and individuals completed daily logs about the amount of time spent in each area. On average,
individuals had an estimated exposure of 233.4 |ig/m3 PM10 (range 73.4-1220.2 |ig/m3). Short-term
exposure did not alter the methylation of Alu and LINE-1. To examine effects of long-term exposure,
both blood samples were considered independent of time, and Alu and LINE-1 were examined with
respect to overall estimated PMi0 exposure using mixed effects models. There was a negative association
between increasing levels of PMi0 exposure and Alu and LINE-1 methylation, indicating that PM10 causes
epigenetic changes to occur with long-term exposure. This study also looked at levels of inducible nitric
oxide synthase gene (iNOS), which is a gene suppressed by DNA methylation. iNOS expression was not
associated with long term exposure to PMi0 but was affected by methylation in the short term.
Animal toxicology studies evaluating the effect of PM exposure on changes in the
epigenome and other non-epigenetic heritable DNA changes have only recently been
conducted. After earlier work showed increased germline mutation rates in herring gulls
nesting near steel mills on Lake Ontario (Yauk and Quinn, 1996, 157163) further work was
conducted to address air-dependent contribution to germline mutations by housing male
and female Swiss Webster mice in the same area and comparing mutation rates in those
animals with mutation rates of animals housed in a rural setting with less air pollution
(Somers et al., 2002, 078100). To determine if PM or the gaseous phase of the urban air was
responsible for heritable mutations, Yauk et al. (2008, 157164) exposed mature male
C57BlxCBAFl hybrid mice to either HEPA-filtered air or to ambient air in Hamilton,
Ontario, Canada for 3 or 10 wk, or 10 wk plus 6 wk of clean air exposure (16 wk). Sperm
DNA was monitored for expanded simple tandem repeat (ESTR) mutations, testicular
sample bulky DNA adducts, and DNA single or double strand breaks. In addition, male-
germ line (spermatogonial stem cell) DNA methylation was monitored post-exposure. This
area in Hamilton is near 2 steel mills and a major highway. Air composition showed mean
concentrations for TSP of 93.8 ± 17 |ug/m3, PAH of 8.3 ± 1.7 ng/m3, and metal of 3.6 ± 0.7
pg/m3. Mutation frequency at ESTR Ms6"hm locus in sperm DNA from mice exposed 3 or
10 wk did not show elevated ESTR mutation frequencies, but there was a significant
increase in ESTR mutation frequency at 16 wk compared to HEPA-filter control animals,
pointing to a PM-dependent mechanism of action. No detectable DNA adducts were
observed in testes samples at any of the time points monitored. To verify inhalation
July 2009
7-116
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
exposure to particles, DNA adducts were reported in the lungs of mice exposed for 3 wk to
ambient-air! no other time points showed detectable DNA adduct formation.
Hypermethylation of germ-line DNA was also observed in mice exposed to ambient air for
10 and 16 wk. These PM-dependent epigenetic modifications (hypermethylation) were not
seen in the halploid stage (3 wk) of spermatogenesis, but were nonetheless seen in early
stages of spermatogenesis (10 wk) and remained significantly elevated in mature sperm
even after removal of the mouse from the environmental exposure (16 wk). Thus, these
studies indicate that the ambient PM phase and not the gaseous phase is responsible for
the increased frequency of heritable DNA mutations and epigenetic modifications.
Based on the limited evidence from these epigenetics studies, long-term exposure to
PMio may result in epigenetic changes. PM2.5 also potentially affects some DNA methylation
content. As epigenetic research progresses, future studies examining the relationship
between PM and DNA methylation will be important in more thoroughly characterizing
these associations.
The effect of ambient PM on heritable DNA mutations and the epigenome has been
well characterized in a Canadian steel mill area. Mice exposed to ambient PM plus gases
developed paternally-derived heritable DNA mutations and epigenetic changes in sperm
DNA that were not observed in mice exposed to ambient air that was HEPA-filtered. This is
the first animal toxicology study showing heritable effects of PM exposure on DNA
mutation and the epigenome. Because the epigenetics field is so new, further work in this
emerging area will expand on these PM-dependent methylation changes to determine if the
results can be recapitulated at other urban sites.
7.5.4. Summary and Causal Determinations
The 2004 PM AQCD reported on original and follow-up analyses for three prospective
cohort studies that reported positive relationships between PM2.5 and lung cancer mortality.
Several recent, well-conducted epidemiologic studies have extended the evidence for a
positive association between PM2.5 and lung cancer mortality (Section 7.5.1.1). Generally,
studies have not reported associations between long-term exposure to PM2.5 or PM10 and
lung cancer incidence (Section 7.5.1.1). Animal toxicological studies did not focus on specific
size fractions of PM, but rather examined ambient PM, woodsmoke, and DEP (Section
7.5.2). A number of recent studies indicate that ambient urban PM, emissions from
wood/biomass burning, emissions from coal combustion, and gasoline and DE are
mutagenic and that PAHs are ge no toxic (Section 7.5.2). These findings are consistent with
earlier studies that concluded that ambient PM and PM from specific combustion sources
are mutagenic and genotoxic and provide biological plausibility for the results observed in
the epidemiologic studies. A limited number of epidemiologic and toxicological studies on
the epigenome demonstrate that PM induces changes in methylation (Section 7.5.3), a new
July 2009
7-117
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
area of research that will likely be expanded in the future. However, it has yet to be
determined how these alterations in the genome could influence the initiation and
promotion of cancer. Overall, the evidence is suggestive of a causal relationship between
relevant PM2.5 exposures and cancer, with the strongest evidence from the epidemiologic studies
of lung cancer mortality. This evidence is limited by the non-specific measure of PM size
fraction in some of the epidemiologic studies and most of the animal toxicological studies,
and the inconsistency in evidence with recent epidemiologic studies for an effect on cancer
incidence. There is no epidemiologic evidence for cancer related to long-term exposure to
PM in organs or systems other than the lung. The evidence is inadequate to assess the
association between PM10-2.5 and UFP exposures and cancer.
7.6. Mortality Associated with Long-term Exposure
In the 1996 PM AQCD, results were presented for three prospective cohort studies of
adult populations: the Six Cities Study (Dockery et al., 1993, 044457); the American Cancer
Society (ACS) Study (Pope CAet al., 1995, 045159); and the California Seventh Day
Adventist (AHSMOG) Study (Abbey et al., 1995, 000669). The 1996 AQCD concluded that
the chronic exposure studies, taken together, suggested associations between increases in
mortality and long-term exposure to fine PM, though there was no evidence to support an
association with PM10-2.5 (U.S. EPA, 1996, 079380).
Discussions of mortality and long-term exposure to PM in the 2004 PM AQCD
emphasized the results of four U.S. prospective cohort studies, but the greatest weight was
placed on the findings of the ACS and the Harvard Six Cities studies, which had each
undergone extensive independent reanalysis, and which were based on cohorts that were
broadly representative of the U.S. population. The 2004 PM AQCD concluded that the
results from the Seventh-Day Adventist (AHSMOG) cohort provided some suggestive (but
less conclusive) evidence for associations, while results from the Veterans Cohort provided
inconsistent evidence for associations between long-term exposures to PM2.5 and mortality.
Collectively, the 2004 PM AQCD found that these studies provided strong evidence that
long-term exposure to PM2.5 was associated with increased risk of human mortality. Effect
estimates for all-cause mortality ranged from 6 to 13% increased risk per 10 ug/m3 PM2.5,
while effect estimates for cardiopulmonary mortality ranged from 6 to 19% per 10 ug/m3
PM2.5. For lung cancer mortality, the effect estimate was a 13% increase per 10 ug/m3 PM2.5,
based upon the results of the extended analysis from the ACS cohort (Pope CAet al., 2002,
024689). With regard to thoracic coarse particles, the 2004 PM AQCD reported that no
association was observed between mortality and long-term exposure to PM10-2.5 in the ACS
study (Pope CAet al., 2002, 024689), while a positive but statistically non-significant
association was reported in males in the AHSMOG cohort (McDonnell et al., 2000, 010319).
Thus, the 2004 PM AQCD concluded that there was insufficient evidence for associations
July 2009
7-118
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
between long-term exposure to thoracic coarse particles and mortality. Overall, the 2004
PM AQCD concluded that there was strong epidemiologic evidence for associations between
long-term exposures to PM2.5 and excess all-cause and cardiopulmonary mortality.
At the time of the 2004 PM AQCD, only a limited number of the chronic-exposure
cohort studies had considered direct measurements of constituents of PM, other than
sulfates. With regard to source-oriented evaluations of mortality associations with long-
term exposure, the 2004 PM AQCD noted only the study by Hoek et al. (2002, 042364), in
which the authors concluded that long-term exposure to traffic-related air pollution may
shorten life expectancy. However, Hoek et al. (2002, 042364) also noted that living near a
major road might include other factors that contribute to mortality associations. There was
not sufficient evidence at the time of the 2004 PM AQCD to draw conclusions on effects
associated with specific components or sources of PM.
New epidemiologic evidence reports a consistent association between long-term
exposure to PM2.5 and increased risk of mortality. There is little evidence for the long-term
effects of PM10-2.5 on mortality. Although this section focuses on mortality outcomes in
response to long-term exposure to PM, it does not evaluate studies that examine the
association between PM and infant mortality. These studies are evaluated in Section 7.5-
"Reproductive, developmental, prenatal and neonatal outcomes associated with long-term
exposure to PM" because it is possible that in utero exposures contribute to infant
mortality. A summary of the mean PM concentrations reported for the studies characterized
in this section is presented in Table 7-8.
Table 7-8. Characterization of ambient PM concentrations from studies of mortality and long-term
exposures to PM.
Reference
Location
Mean Concentration
(*/fl/m31
Upper Percentile Concentrations (/^g/m31
PM2.5
Brunekreef et al. (2009,191947)
The Netherlands
28
95th: 32
99th: 33
Max: 37
Chen et al. (2005, 087942)
Multicity, CA
29.0

Eftim et al. (2008, 099104)
U.S.
13.6-14.1
Max: 19.1-25.1
Enstrom (2005, 087356)
CA
23.4
Max: 36.1
Goss et al. (2004, 055624)
U.S.
13.7
75th: 15.9
Janes et al. (2007, 090927)
U.S.
14.0

Jerrett et al. (2005,189405)
Los Angeles, CA

Max: 27.1
Krewski et al. (2009,191193)
U.S.
14.02
75th: 16.00
90th: 26.75
95th: 27.89
Max: 30.01
Laden et al. (2006, 087605)
Multicity, U.S.
10.2-29.0

July 2009

7-119
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Reference
Location
Mean Concentration
(j/g/m31
Upper Percentile Concentrations (/^g/m31
Lipfert et al. (2006, 088218)
U.S.
14.3

Miller et al. (2007, 090130)
U.S.
13.5
75th: 18.3
Max: 28.3
Pope et al. (2004, 055880)
U.S.
17.1

Schwartz et al. (2008,156963)
Multicity, U.S.
17.5
Max: 40
Zeper et al. (2007,157176)
U.S.

17.0
Zegeret al. (2008, 19 1 951)
U.S.
13.2
75th: 14.9
PM10-2.5
Chen et al. (2005, 087942)
Multicity, CA
25.4

Lipfert etal. (2006, 088218)
U.S.
16.0

PMio
Chen et al. (2005, 087942)
Multicity, CA
52.6

Gehring et al. (2006, 089797)
North Rhine,
Germany
43.7-48.0
Max: 52.5-56.1
Goss et al. (2004, 055624)
U.S.
24.8
75th: 28.9
Puett et al. (2008,156891)
NE U.S.
21.6

Zanobetti et al. (2008,156177)
U.S.
29.4

7.6.1. Recent Studies of Long-Term Exposure to PM and
Mortality
Studies since the last PM AQCD include results of new analyses and insights for the
ACS and Harvard Six Cities studies, further analyses from the AHSMOG and Veterans
study cohorts, as well as analyses of a Cystic Fibrosis cohort and subsets of the ACS from
Los Angeles and New York City In the original analyses of the Six Cities and ACS cohort
studies, no associations were found between long-term exposure to PM10-2.5 and mortality,
and the extended and followup analyses did not evaluate associations with PM10-2.5. The
historical and more recent results for PM2.5 of both the ACS and the Harvard Six Cities
studies are compiled in Figure 7-6. Moreover, since the last PM AQCD, there is a major new
cohort that investigates the effects of PM2.5on cardiovascular mortality in the literature^
the Women's Health Initiative (WHI) study (Miller et al., 2007, 090130). Most recently, an
ecological cohort study of the nation's Medicare population has been completed (Eftim et al.,
2008, 099104). These new findings further strengthen the evidence linking long-term
exposure to PM2.5 and mortality, while providing indications that the magnitude of the
PM2.5 -mortality association is larger than previously estimated (Figure 7-7). Two recent
reports from the AHSMOG and Veterans study cohorts have provided some limited
evidence for associations between long-term exposure to PM10-2.5 and mortality. The original
analyses of the AHSMOG cohort study found positive associations between long-term
July 2009
7-120
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
concentrations of PMio and 15-yr mortality due to natural causes and lung cancer (Abbey et
al., 1999, 047559). McDonnell et al. (2000, 010319) reanalyzed these data and concluded
that previously observed association of long-term ambient PMio concentrations with
mortality for males were best explained by a relationship of mortality with the fine fraction
of PMio rather than the coarse fraction of PMio. Recent reports from the AHSMOG study
cohort, as well as the Nurses' Health Study and a cohort of women in Germany have
provided some evidence for associations between long-term exposure to PMio and mortality
among women.
Harvard Six Cities: A follow-up study has used updated air pollution and mortality
data! an additional 1,368 deaths occurred during the follow-up period (1990-1998) vs. 1,364
deaths in the original study period (1974-1989) (Laden et al., 2006, 087605). Statistically
significant associations are reported between long-term exposure to PM2.5 and mortality for
data for the two periods (RR = 1.16 [95% CP 1.07-1.26] per 10 |u,g/m3 PM2.5). Of special note
is a statistically significant reduction in mortality risk reported with reduced long-term fine
particle concentrations (RR = 0.73 [95% CP 0.57-0.95] per 10 |u,g/m3 PM2.5). This is
equivalent to an RR of 1.27 for reduced mortality risks with reduced long-term PM2.5
concentrations. This reduced mortality risk was observed for deaths due to cardiovascular
and respiratory causes, but not for lung cancer deaths. The PM2.5 concentrations for recent
years were estimated from visibility data, which introduces some uncertainty in the
interpretation of the results from this study. Coupled with the results of the original
analysis (Dockery et al., 1993, 044457), this study strongly suggests that a reduction in fine
PM pollution yields positive health benefits.
ACS Extended Analyses/Reanalysis II: Two new analyses further evaluated the
associations of long-term PM2.5 exposures with risk of mortality in 50 U.S. cities reported by
Pope and colleagues (2002, 024689), adding new details about deaths from specific
cardiovascular and respiratory causes (Krewski, 2009, 190075; Pope CAet al., 2004,
055880). Pope et al. (2004, 055880) reported positive associations with deaths from specific
cardiovascular diseases, particularly ischemic heart disease, and a group of cardiac
conditions including dysrhythmia, heart failure and cardiac arrest (RR for cardiovascular
mortality = 1.12, 95% CI 1.08-1.15 per 10 jug/m3 PM2.5), but no PM associations were found
with respiratory mortality.
In an additional reanalysis that extended the follow-up period for the ACS cohort to
18 years (1982-2000) (Krewski et al., 2009, 191193), investigators found effect estimates
that were similar, though generally higher, than those reported in previous ACS analyses.
This reanalysis also included data for seven ecologic (neighborhood-level) contextual (i.e.,
not individual-level) covariates, each of which represents local factors known or suspected
to influence mortality, such as poverty level, educational attainment, and unemployment.
The effect estimate for all cause mortality, based on PM2.5 concentrations measured in 1999
and 2000 was 1.03 (95% CP 1.01-1.05). The corresponding effect estimates for deaths due to
July 2009
7-121
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
IHD and lung cancer were 1.15 (95% CI: 1.04-1.18) and 1.11 (95% CI: 1.04-1.18),
respectively. In earlier analyses of this cohort, investigators found that increasing education
levels appeared to reduce the effect of PM2.5 exposure on mortality. Results from this
reanalysis show a similar pattern, although with somewhat less certainty, for all causes of
death except IHD, for which the pattern was reversed. Overall, although the addition of
random effects modeling and contextual covariates to the ACS model made most effect
estimates higher (but some lower), they were not statistically different from the earlier ACS
effect CPD=Cardio-Pulmonary Disease! CVD=Cardiovascular Disease! IHD=Ischemic Heart
estimates. Thus, these new analyses, with their more extensive consideration of potentially
confounding factors, confirm the published ACS PJVk.s'mortality results to be robust.
California Cancer Prevention Study: In a cohort of elderly people in 11 California
counties (mean age 73 years in 1983), an association was reported for long-term PM2.5
exposure with all-cause deaths from 1973-1982 (RR = 1.04 [95% CP 1.01-1.07] per 10 |ug/m3
PM2.5) (Enstrom, 2005, 087356). However, no significant associations were reported with
deaths in later time periods when PM2.5 levels had decreased in the most polluted counties
(1983-2002) (RR = 1.00 [95% CP 0.98-1.02] per 10 |ug/m3 PM2.5). The PM2.5 data were
obtained from the EPA's Inhalation Particle Network (collected 1979-1983), and the
locations represented a subset of data used in the 50-city ACS study (Pope CAet al., 1995,
045159). However, the use of average values for California counties as exposure surrogates
likely leads to significant exposure error, as many California counties are large and quite
topographically variable.
July 2009
7-122
DRAFT-DO NOT CITE OR QUOTE

-------

Cohort
Study
Years
Mean



Original
Dockery et al. (1993, 044457)
1974-1991
18.6
i—>	
All Cause

Reanalysis
Krewski etal. (2000, 012281)
1974-1991
18.6



Temporal Changes
Villeneuve et al. (2002,042576)
1974-1991
18.6
1 "


Extended
Laden et al. (2006, 087605)
1974-1998
17.6
1 	*	


6-Cities Medicare
Eftim et al. (2008, 099104)
2000-2002
14.1
1


Original
Pope etal. (1995, 045159)
1982-1989
18.2
]


Reanalysis
Krewski et al.(2000, 012281)
1982-1989
18.2
j "Hh


Extended
Pope et al. (2002, 024689)
1979-1983
21.1
r«-


Extended
Pope et al. (2002, 024689)
1999-2000
140.
!^-

ACS
Intra-metro LA
Jerrett et al. (2005,189405)
1982-2000
19.0
)


ACS Medicare
Eftim et al. (2008, 099104)
2000-2002
13.6
~


Reanalysis II
Krewski et al. (2009,191193)
1982-2000
14.0



Reanalysis II - LA
Krewski et al. (2009,191193)
1982-2000
20.5
i i


Reanalysis II - NYC
Krewski etal. (2009,191193)
1982-2000
128	
i

SCS
Original
Dockery etal. (1993, 044457)
1974-1991
18.6
i
i 1
CPD

Reanalysis
Krewski etal. (2000, 012281)
1974-1991
18.6
i —•	
t


Original
Pope etal. (1995, 045159)
1982-1989
18.2
i m^mm


Reanalysis
Krewski etal. (2000, 012281)
1982-1989
18.2



Extended
Pope et al. (2002, 024689)
1979-1983
21.1
j™*"*

ACS
Extended
Pope et al. (2002, 024689)
1999-2000
14.0
J—«—


Intra-metro LA
Jerrett etal. (2005,189405)
1982-2000
19.0
* t


Reanalysis II
Krewski et al. (2009,191193)
1982-2000
14.0



Reanalysis II - LA
Krewski et al. (2009,191193)
1982-2000
20.5



Reanalysis II - NYC
Krewski et al. (2009,191193)
1982-2000
KM—•	
	1—
i

SCS
Extended
Laden et al. (2006, 087605)
1974-1998
17.6
i
i
CVD
arq
Reanalysis
Krewski et al. (2000,012281)
1982-1989
18.2
i


Extended
Pope et al. (2004, 055880)
1982-2000
17.1
i


Extended
Pope et al. (2004, 055880)
1982-2000
17.1

IHD

Intra-metro LA
Jerrett etal. (2005,189405)
1982-2000
19.0
i 9

ACS
Reanalysis II
Krewski et al. (2009,191193)
1982-2000
14.0



Reanalysis II - LA
Krewski et al. (2009,191193)
1982-2000
20.5



Reanalysis II - NYC
Krewski et al. (2009,191193)
1982-2000
12.8
i


Original
Dockery etal. (1993, 044457)
1974-1991
18.6
—1 '
Lung Cancer
SCS
Reanalysis
Krewski etal. (2000, 012281)
1974-1991
18.6
	> *	


Extended
Laden et al. (2006, 087605)
1974-1998
17.6
1 «	
]


Original
Pope et al. (1995, 045159)
1982-1989
18.2
1


Extended
Pope etal. (2002, 024689)
1979-1983
21.1
i i


Extended
Popeet al. (2002, 024689)
1999-2000
14.0
i —i—

AL/O
Intra-metro LA
Jerrett etal. (2005,189405)
1982-2000
19.0
i


Reanalysis II
Krewski et al. (2009,191193)
1982-2000
14.0
J -MR—


Reanalysis II - LA
Krewski et al. (2009,191193)
1982-2000
20.5
	J	*


Reanalysis II - NYC
Krewski et al. (2009,191193)
1982-2000
m	
—*	1

SCS
Extended
Laden et al. (2006, 087605)
1974-1998
17.6
¦—h—
i
Other

Extended
Pope et al. (2002, 024689)
1979-1983
21.1
i


Extended
Pope et al. (2002, 024689)
1999-2000
14.0



Intra-metro LA
Jerrett etal. (2005,189405)
1982-2000
19.0
i


Reanalysis II
Krewski et al. (2009,191193)
1982-2000
14.0


I	1	1	
OJ	U	f,5
Figure 7-6. Mortality risk estimates associated with long-term exposure to PM2.5 from the
Harvard Six Cities Study (SCS) and the American Cancer Society Study (ACS).
July 2009
7-123
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Cohort
Subset
Mean
McDonnell et al. (2000,010319)
AHSMOG
Males
32.0
Brunekreef et al. (2009,191947)
NLCS-AIR
Full Cohort
28.3
Brunekreef et al. (2009,191947)
NLCS-AIR
Case Cohort
28.3
Enstrom (2005, 087356)
CA Cancer Prevention
1973-1982
23.4
Enstrom (2005, 087356)
CA Cancer Prevention
1983-2002

Enstrom (2005, 087356)
CA Cancer Prevention
1973-2002

Jerrett et al. (2005,189405)
ACS-LA

19.0
Krewski et al. (2009,191193)
ACS Reanalvsis II - LA

20.5
Laden et al. (2006, 087605)
Harvard 6-Cities

16.0
Lipfert et al. (2006, 088218)
Veterans Cohort

14.3
Lipfert et al. (2006, 088756)
Veterans Cohort

14.3
Eftim et al. (2008,099104)
Medicare Cohort
ACS Sites
13.6
Eftim et al. (2008,099104)
Medicare Cohort
6-Cities sites
14.1
Krewski et al. (2009,191193)
ACS Reanalvsis II

14.0
Goss et al. (2004, 055624)
U.S. Cvstic Fibrosis

13.7
Zegeret al. (2008,191951)
MCAPS
65+, Eastern
14.0
Zegeret al. (2008,191951)
MCAPS
65+, Central
10.7
Zegeret al. (2008,191951)
MCAPS
65+, Western
13.1
Zegeret al. (2008,191951)
MCAPS
65-74, Eastern
14.0
Zegeret al. (2008,191951)
MCAPS
65-74, Central
10.7
Zegeret al. (2008,191951)
MCAPS
65-74, Western
13.1
Zegeret al. (2008,191951)
MCAPS
65+, Eastern
14.0
Zegeret al. (2008,191951)
MCAPS
75-84, Central
10.7
Zegeret al. (2008,191951)
MCAPS
75-84, Western
13.1
Zegeret al. (2008,191951)
MCAPS
85+, Eastern
14.0
Zegeret al. (2008,191951)
MCAPS
85+, Central
10.7
Zegeret al. (2008,191951)
MCAPS
85+, Western
13.1
Krewski et al. (2009,191193)
ACS Reanalvsis II - NYC

12.8
Brunekreef et al. (2009,191947)
NLCS-AIR
Full Cohort
28.3
Brunekreef et al. (2009, 191947)
NLCS-AIR
Case Cohort
28.3
Pope et al. (2004, 055880)
ACS

17.1
Laden et al. (2006, 087605)
Harvard 6-Cities

16.0
Naess et al. (2007, 090736)
Oslo, Norwav
Men, 51-70 vrs
14.3
Naess et al. (2007,090736)
Oslo, Norwav
Men, 71-90 vrs
14.3
Naess et al. (2007,090736)
Oslo, Norwav
Women, 51-70
14.3
Naess et al. (2007,090736)
Oslo, Norwav
Women 71-90 vrs
14.3
Miller et al. (2007,090130)
WHI
Females
13.5
Chen et al. (2005, 087942)
AHSMOG
Females
29.0
Chen et al. (2005, 087942)
AHSMOG
Males
29.0
Jerrett et al. (2005, 189405)
ACS-LA

19.0
Krewski etal. (2009,190075)
ACS Reanalvsis II - LA

20.5
Pope et al. (2004, 055880)
ACS

17.1
Krewski et al. (2009,191193)
ACS Reanalvsis II

14.0
Krewski et al. (2009,191193)
ACS Reanalvsis II -NYC

12.8
McDonnell etal. (2000,010319)
AHSMOG
Males
32.0
Jerrett et al. (2005,189405)
ACS-LA

19.0
Krewski etal. (2009,191193)
ACS Reanalvsis II - LA

20.5
Krewski etal. (2009,191193)
ACS Reanalvsis II

14.0
Krewski etal. (2009,191193)
ACS Reanalvsis II -NYC

12.8
Brunekreef et al. (2009,191947)
NLCS-AIR
Full Cohort
28.3
Brunekreef et al. (2009, 191947)
NLCS-AIR
Case Cohort
28.3
Laden et al. (2006, 087605)
Harvard 6-Cities

16.0
McDonnell etal. (2000, 010319)
AHSMOG
Males
32.0
Brunekreef et al. (2009,191947)
NLCS-AIR
Full Cohort
28.3
Brunekreef et al. (2009,191947)
NLCS-AIR
Case Cohort
28.3
Jerrett et al. (2005,189405)
ACS-LA

19.0
Krewski etal. (2009, 191193)
ACS Reanalvsis II - LA

20.5
Laden et al. (2006, 087605)
Harvard 6-Cities

16.0
Naess et al. (2007, 090736)
Oslo, Norwav
Men, 51 -70 vrs
14.3
Naess et al. (2007,090736)
Oslo, Norwav
Men, 71 -90 vrs
14.3
Naess et al. (2007,090736)
Oslo, Norwav
Women, 51 -70
14.3
Naess etal. (2007,090736)
Oslo, Norwav
Women 71-90
14.3
Krewski etal. (2009,191193)
ACS Reanalvsis II

14.0
Krewski et al. (2009,191193)
ACS Reanalvsis II -

12.8
Brunekreef et al. (2009,191947)
NLCS-AIR
Full Cohort
28.3
Brunekreef et al. (2009,191947)
NLCS-AIR
Case Cohort
28.3
Jerrett et al. (2005,189405)
ACS-LA

19.0
Laden et al. (2006, 087605)
Harvard 6-Cities

16.0
Krewski etal. (2009,191193)
ACS Reanalvsis II

14.0
CV = Cardiovascular; CHD= Coronary Heart [
lisease; IHD = Ischemic Heart Dise
iase; CPD = Cardio-Pulmonary Disease
All Cause
!~
t
f
+
»
cv
CHD
IHD
CPD
Respiratorv
Lung Cancer
t
[.
[.
» i
i
Other
01 1.0 1.1 21
Figure 7-7. Mortality risk estimates associated with long-term exposure to PM2.5 in recent cohort
studies.
July 2009
7-124
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
AHSMOG: In this analysis for the Seventh-Day Adventist cohort in California, a
positive, statistically significant, association with coronary heart disease mortality was
reported among females (92 deaths! RR = 1.42 [95% CI: 1.06-1.90] per 10 |u,g/m3 PM2.5), but
not among males (53 deaths! RR = 0.90 [95% CI: 0.76-1.05] per 10 |ug/m3 PM2.5) (Chen et al.,
2005, 087942). Associations were strongest in the subset of postmenopausal women (80
deaths! RR = 1.49 [95% CI: 1.17-1.89] per 10 |ug/m3 PM2.5). The authors speculated that
females may be more sensitive to air pollution-related effects, based on differences between
males and females in dosimetry and exposure. As was found with fine particles, a positive
association with coronary heart disease mortality was reported for PM10-25 and PM10 among
females (RR = 1.38 [95% CI: 0.97-1.95] per 10 |ug/m3 PMio-2.5;RR = 1.22 [95% CI: 1.01-1.47]
per 10 |ug/m3 PM10), but not for males (RR = 0.92 [95% CI: 0.66-1.29] per 10 |ug/m3 PM10-2.5;
RR = 0.94 [95% CI: 0.82-1.08] per 10 |ug/m3 PM10); associations were strongest in the subset
of postmenopausal women (80 deaths) (Chen et al., 2005, 087942).
U.S. Cystic Fibrosis cohort: A positive, but not statistically significant, association was
reported for PM2.5 in this cohort (RR = 1.32 [95% CI: 0.91-1.93] per 10 jug/m3 PM2.5) in a
study that primarily focused on evidence of exacerbation of respiratory symptoms (Goss et
al., 2004, 055624). No clear association was reported for PM10. However, only 200 deaths
had occurred in the cohort of over 11,000 people (average age in cohort was 18.4 years), so
the power of this study to detect associations was relatively low.
Women's Health Initiative (WHI) Study: This nationwide cohort study considered 65,893
postmenopausal women with no history of cardiovascular disease who lived in 36 U.S.
metropolitan areas from 1994 to 1998 (Miller et al., 2007, 090130). The study had a median
subject followup time of six years. Miller and colleagues assessed each woman's exposure to
air pollutants using the monitor located nearest to their residence. Hazard ratios were
estimated for the first cardiovascular event, adjusting for age, race or ethnic group,
smoking status, educational level, household income, body-mass index, and presence or
absence of diabetes, hypertension, or hypercholesterolemia. Overall, this study concludes
that "long-term exposure to fine particulate air pollution is associated with the incidence of
cardiovascular disease and death among postmenopausal women." In terms of effect size,
the study found that each increase of 10 |ug/m3 of PM2.5 was associated with a 24% increase
in the risk of a cardiovascular event (hazard ratio, 1.24 [95% CI: 1.09-1.41]) and a 76%
increase in the risk of death from cardiovascular disease (hazard ratio, 1.76 [95% CI: 1.25-
2.47]). While this study found results confirmatory to the ACS and Six Cities Study, it
reports much larger relative risk estimates per ug/ma PM2.5. In addition, since the study
included only women without pre-existing cardiovascular disease, it could potentially be a
healthier cohort population than that considered by the ACS and Six Cities Study. Indeed,
the WHI Study reported only 216 cardiovascular deaths in 349,643 women-years of
followup, or a rate of 0.075% deaths per year (Miller et al., 2007, 090130), while the ACS
Study reported that 10% of subjects died of cardiovascular disease over a 16 yr followup
July 2009
7-125
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
period, yielding a rate of 0.625% per year, or approximately 8 times the cardiovascular
mortality rate of the WHI population (Pope CAet al., 2004, 055880). Thus, PM2.5 impacts
may yield higher relative risk estimates in the WHI population because the PM2.5 risk is
being compared to a much lower prevailing risk of cardiovascular death in this select study
population.
The WHI study not only confirms the ACS and Six City Study associations with
mortality in yet another well characterized cohort with detailed individual-level
information, it also has been able to consider the individual medical records of the
thousands of WHI subjects over the period of the study. This has allowed the researchers to
examine not only mortality, but also related morbidity in the form of heart problems
(cardiovascular events) experienced by the subjects during the study. As reported in this
paper, this examination confirmed that there is an increased risk of cardiovascular
morbidity, as well (see Section 7.2.9). These morbidity co-associations with PM2.5 in the
same population lend even greater support to the biological plausibility of the air pollution-
mortality associations found in this study.
Medicare Cohort Studies: Using Medicare data, Eftim and co-authors (2008, 099104)
have assessed the association of PM2.5 with mortality for the same locations included in the
ACS and Six City Study. For these locations, they estimated the chronic effects of PM2.5 on
mortality for the period 2000-2002 using mortality data for cohorts of Medicare participants
and average PM2.5 levels from monitors in the same counties included in the two studies.
Using aggregate counts of mortality by county for three age groups, they estimated
mortality risk associated with air pollution adjusting for age and sex and area-level
covariates (education, income level, poverty, and employment), and controlled for potential
confounding by cigarette smoking by including standardized mortality ratios for lung
cancer and COPD. This study is, therefore, an ecological analysis, similar to past published
cross-sectional analyses, in that area-level covariates (education, income level, poverty, and
employment) are employed as controlling variables, since individual level information is not
available from the Medicare database (other than age and sex), which includes virtually all
Americans aged 65 or greater. Exposures are also ecological in nature, as central site data
are used as indices of exposure. These results indicated that a 10 ug/ma increase in the
yearly average PM2.5 concentration is associated with 10.9% (95% CP 9.0-12.8) and with
20.8% (95% CP 14.8-27.1) increases in all-cause mortality for the ACS and Six Cities Study
counties, respectively. The estimates are somewhat higher than those reported by the
original investigators, and there may be several possible explanations for this apparent
increase, especially that this is an older population than the ACS cohort. Perhaps the most
likely explanation is that the lack of personal confounder information (e.g., past personal
smoking information) led to an insufficient control for the effects of these other variables'
effects on mortality, inflating the pollution effect estimates somewhat, similar to what has
been found in the ACS analyses when only ecological-level control variables were included.
July 2009
7-126
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
The ability of the Eftim et al. (2008, 099104) study results to qualitatively replicate the
original individual-level cohort study (e.g., ACS and Six Cities Study) results suggests that
past ecological cross-sectional mortality study results may also provide useful insights into
the nature of the association, especially when used for consideration of time trends, or for
comparisons of the relative (rather than absolute) sizes of risks between different pollutants
or PM components in health effects associations.
Janes et al. (2007, 090927) use the same nationwide Medicare mortality data to
examine the association between monthly averages of fine particles (PM2.5) over the
preceding 12 months and monthly mortality rates in 113 U.S. counties from 2000 to 2002.
They decompose the association between PM2.5 and mortality into 2 components^ (l) the
association between "national trends" in PM2.5 and mortality! and (2) the association
between "local trends," defined as county-specific deviations from national trends. This
second component is posited to provide evidence as to whether counties having steeper
declines in PM2.5 also have steeper declines in mortality relative to the national trend. They
report that the exposure effect estimates are different at these 2 spatiotemporal scales,
raising concerns about confounding bias in these analyses. The authors assert that the
association between trends in PM2.5 and mortality at the national scale is more likely to be
confounded than is the association between trends in PM2.5 and mortality at the local scale
and, if the association at the national scale is set aside, that there is little evidence of an
association between 12-month exposure to PM2.5 and mortality in this analysis. However, in
response, Pope and Burnett (2007, 090928) point out that such use of long-term time trends
as the primary source of exposure variability has been avoided in most other air pollution
epidemiology studies because of such concerns about potential confounding of such time-
trend associations.
By linking monitoring data to the U.S. Medicare system by county of residence, Zeger
et al. (2007, 157176) analyzed Medicare mortality records, comprising over 20 million
enrollees in the 250 largest counties during 2000-2002. The authors estimated log-linear
regression models having age-specific county level mortality rates as the outcome and, as
the main predictor, the average PM2.5 pollution level in each county during 2000. Area-level
covariates were used to adjust for socio-economic status and smoking. The authors reported
results under several degrees of adjustment for spatial confounding and with stratification
into eastern, central and western U.S. counties. A 10 pg/m3 increase in PM2.5 was associated
with a 7.6% increase in mortality (95% CP 4.4-10.8). When adjusted for spatial
confounding, the estimated log-relative risks dropped by 50%. Zeger et al. (2007, 157176)
found a stronger association in the eastern counties than nationally, with no evidence of an
association in western counties.
In a subsequent report, Zeger et al. (2008, 191951) created a new retrospective cohort,
the Medicare Cohort Air Pollution Study (MCAPS), consisting of 13.2 million persons
residing in 4,568 ZIP codes in urban areas having geographic centroids within 6 miles of a
July 2009
7-127
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
PM2.5 monitor. Using this cohort, they investigated the relationship between 6-yr avg
exposure to PM2.5 and mortaility risk over the period 2000-2005. When divided by region,
the associations between long-term exposure to PM2.5 and mortality for the eastern and
central ZIP codes were qualitatively similar to those reported in the ACS and Six Cities
Study, with 11.4% (95% CP 8.8-14.1) and 20.4% (95% CP 15.0-25.8) increases per 10 |ug/m3
increase in PIVL.r, in the eastern and central regions, respectively. The MCAPS results
included evidence of differing PM2.5 relative risks by age and geographic location, where
risk declines with increasing age category until there is no evidence of an association
among persons > 85 years of age, and there is no evidence of a positive association for the
640 urban ZIP codes in the western region of the U.S.
Using hospital discharge data, Zanobetti et al. (2008, 156177) constructed a cohort of
persons discharged with chronic obstructive pulmonary disease using Medicare data
between 1985 and 1999. Positive associations in the survival analyses were reported for
single year and multiple-year lag exposures, with a hazard ratio for total mortality of 1.22
(95% CP 1.17-1.27) per 10 jug/m3 increase in PM10 over the previous 4 years.
Veterans cohort: A recent reanalysis of the Veterans cohort data focused on exposure to
traffic-related air pollution (traffic density based on traffic flow rate data and road segment
length) reported a stronger relationship between mortality with long-term exposure to
traffic than with PM2.5 mass (Lipfert et al., 2006, 088218). A significant association was
reported between total mortality and PM2.5 in single-pollutant models (RR = 1.12 [95% CP
1.04-1.20] per 10 jug/m3 PM2.5). This risk estimate is larger than results reported in a
previous study of this cohort. In multipollutant models including traffic density, the
association with PM2.5 was reduced and lost statistical significance. Traffic emissions
contribute to PM2.5 so it would be expected that the two would be highly correlated, and,
thus, these multipollutant model results should be interpreted with caution. In a
companion study, Lipfert et al. (2006, 088218) used data from EPA's fine particle speciation
network, and reported findings for PM2.5 which were similar to those reported by Lipfert et
al. (2006, 088218). In this study (Lipfert et al., 2006, 088218), a significant association was
reported between long-term exposure to PM10-2.5 and total mortality in a single-pollutant
model (RR = 1.07, 95% CI 1.01-1.12 per 10 |u,g/m3 PM10-2.5). However, the association became
negative and not statistically significant in a model that included traffic density. As it would
be expected that traffic would contribute to the thoracic coarse particle concentrations, it is
difficult to interpret the results of these multipollutant analyses.
Nurses' Health Study Cohort: The Nurses' Health Study (Puett et al., 2008, 156891) is
an ongoing prospective cohort study examining the relation of chronic PM10 exposures with
all-cause mortality and incident and fatal coronary heart disease consisting of 66,250
female nurses in MSAs in the northeastern region of the U.S. All cause mortality was
statistically significantly associated with average PM10 exposures in the time period 3-48
months preceding death. The association was strongest with average PM10 exposure in the
July 2009
7-128
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
24 months prior to death (hazard ratio 1.16 [95% CP 1.05-1.28]) and weakest with exposure
in the month prior to death (hazard ratio 1.04 [95% CP 0.98-1.11]). The association with
fatal CHD occurred with the greatest magnitude with mean exposure in the 24 months
prior to death (hazard ratio 1.42 [95% CP 1.11-1.81]).
Netherlands Cohort Study (NLCS): The Netherlands Cohort Study (Brunekreef et al.,
2009, 191947) estimates the effects of traffic-related air pollution on cause specific mortality
in a cohort of approximately 120,000 subjects aged 55 to 69 year at enrollment. For a 10
iig/ma increase in PM2.5 concentration, the relative risk for natural-cause mortality in the
full cohort was 1.06 (95% CP 0.97-1.16), similar in magnitude to the results reported by the
ACS. In a case-cohort analysis adjusted for additional potential confounders, there were no
associations between air pollution and mortality.
German Cohort: The North Rhine-Westphalia State Environment Agency (LUANRW)
initiated a cohort of approximately 4,800 women, and assessed whether long-term exposure
to air pollution originating from motorized traffic and industrial sources was associated
with total and cause-specific mortality (Gehring et al., 2006, 089797). They found that
cardiopulmonary mortality was associated with PM10 (RR = 1.52 [95% CP 1.09-2.15] per
10 |ug/m3 PM10).
7.6.2. Composition and Source-Oriented Analyses of PM
As discussed in the 2004 PM AQCD, only a very limited number of the chronic
exposure cohort studies have included direct measurements of chemical-specific PM
constituents other than sulfates, or assessments of source-oriented effects, their analyses.
One exception is the Veterans Cohort Study, which looked at associations with some
constituents, and traffic.
Veterans Cohort: Using data from EPA's fine particle speciation network, Lipfert et al.
(2006, 088756) reported a positive association for mortality with sulfates. Using 2002 data
from the fine particle speciation network, positive associations were found between
mortality and long-term exposures to nitrates, EC, Ni and V, as well as traffic density and
peak ozone concentrations. In multipollutant models, associations with traffic density
remained significant, as did nitrates, Ni and V in some models.
Netherlands Cohort Study: Beelen et al. (2008, 156263) studied the association between
long-term exposure to traffic-related air pollution and mortality in a Dutch cohort. They
used data from an ongoing cohort study on diet and cancer with 120,852 subjects who were
followed from 1987 to 1996. Exposure to BS, NO2, SO2, and PM2.5, as well as various
exposure variables related to traffic, were estimated at the home address. Cox analyses
were conducted in the full cohort, adjusting for age, sex, smoking, and area-level
socioeconomic status. Traffic intensity on the nearest road was independently associated
with mortality. Relative risks (95% confidence intervals) for a 10 ug/ma increase in BS
July 2009
7-129
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
concentrations (difference between 5th and 95th percentile) were 1.05 (95% CP 1.00-1.11)
for natural cause, 1.04 (95% CI: 0.95-1.13) for cardiovascular, 1.22 (95% CP 0.99-1.50) for
respiratory, 1.03 (95% CP 0.88-1.20) for lung cancer, and 1.04 (95% CI: 0.97-1.12) for
mortality other than cardiovascular, respiratory, or lung cancer. Results were similar for
NO2 and PM2.5, but no associations were found for SO2. Traffic-related air pollution and
several traffic exposure variables were associated with mortality in the full cohort, although
the relative risks were generally small. Associations between natural-cause and respiratory
mortality were statistically significant for NO2 and BS. These results add to the evidence
that long-term exposure to traffic-related particulate air pollution is associated with
increased mortality.
Given the general dearth of published source-oriented studies of the mortality impacts
of long-term PM exposure components, and given that the recent Medicare Cohort study
now indicates that such ecological cross-sectional studies can be useful for evaluating time
trends and/or comparisons across pollution components, it may well be that examining past
cross-sectional studies comparing source-oriented components of PM may be informative. In
particular, Ozkaynak and Thurston (1987, 072960), utilized the chemical speciation
conducted in the Inhalable Particle (IP) Network to conduct a chemical constituent and
source-oriented evaluation on long-term PM exposure and mortality in the U.S. They
analyzed the 1980 U.S. vital statistics and available ambient air pollution data bases for
sulfates and fine, inhalable, and TSP mass. Using multiple regression analyses, they
conducted a cross-sectional analysis of the association between various particle measures
and total mortality. Results from the various analyses indicated the importance of
considering particle size, composition, and source information in modeling of particle
pollution health effects. Of the independent mortality predictors considered, particle
exposure measures most related to the respirable fraction of the aerosols, such as fine
particles and sulfates, were most consistently and significantly associated with the reported
SMSA-specific total annual mortality rates. On the other hand, particle mass measures
that included thoracic coarse particles (e.g., total suspended particles and inhalable
particles) were often found to be non-significant predictors of total mortality. Furthermore,
based on the application of fine particle source apportionment, particles from industrial
sources and from coal combustion were indicated to be more significant contributors to
human mortality than fine soil-derived particles.
7.6.3. Within-City Effects of PM Exposure
Much of the exposure gradient in the national-scale cohort studies was due to city-to-
city differences in regional air pollution, raising the possibility that some or all of the
original PM-survival associations may have been driven instead by city-to-city differences
July 2009
7-130
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
in some unknown (non-pollution) confounder variable. This has been evaluated by three
recent studies.
ACS, Los Angeles: To investigate this issue, two new analyses using ACS data focused
on neighborhood-to-neighborhood differences in urban air pollutants, using data from 23
PM2.5 monitoring stations in the Los Angeles area, and applying interpolation methods
Jerrett et al. (2005, 189405) or land use regression methods (Krewski et al., 2009, 191193)
to assign exposure levels to study individuals. This resulted in both improved exposure
assessment and an increased focus on local sources of fine particle pollution. Significant
associations between PM2.5 and mortality from all causes and cardiopulmonary diseases
were reported with the magnitude of the relative risks being greater than those reported in
previous assessments. In general, the associations for PM2.5 and mortality using these two
methods for exposure assessment were similar, though the use of land use regression
resulted in somewhat smaller hazard ratios and tighter confidence intervals (see Table 7-9)
This indicates that city-to-city confounding was not the cause of the associations found in
the earlier ACS Cohort studies. This provides evidence that reducing exposure error can
result in stronger associations between PM2.5 and mortality than generally observed in
broader studies having less exposure detail.
Table 7-9:
Comparison of results from ACS intra-urban analysis of Los Angeles and new york city
using kriging or land use regression to estimate exposure
Cause of
Death
Los Angeles:
Hazard Ratio1 and 95% Confidence
Interval Using Kriging2
(Jerrett et al., 2005,189405)
Los Angeles:
Hazard Ratio1 and 95% Confidence
Interval Using Land Use Regression3
(Krewski et al., 2009,191193)
New York City:
Hazard Ratio1 and 95% Confidence
Interval Using Land Use Regression4
(Krewski et al., 2009,191193)
All Cause
1.11 (0.99-1.25)
1.13(1.01-1.25)
0.86 (0.63-1.18)
IHD
1.25(0.99-1.59)
1.26 (1.02-1.56)
1.56 (0.87-2.88)
CPD
1.07 (0.91-1.26)
1.09(0.94-1.26)
0.66 (0.41-1.08)
Lung Cancer
1.20 (0.79-1.82)
1.31 (0.90-1.92)
0.90 (0.29-2.78)
'Hazard ratios presented per 10/yg/m3 increase in PM2.5
2Model included parsimonious contextual covariates
3Model included parsimonious individual level (23) and ecologic (4) covariates
4Model included all 44 individual level and 7 ecologic covariates.
ACS, New York: Krewski et al. (2009, 191193) applied the same techniques used in the
land use regression analysis of Los Angeles to an investigation conducted in New York City.
Annual average concentrations were calculated for each of 62 monitors from 3 years of daily
monitoring data for 1999 through 2001. Those data were combined with land-use data
collected from traffic counting systems, roadway network maps, satellite photos of the study
area, and local government planning and tax-assessment maps to assign estimated
exposures to the ACS participants. The investigators did not observe elevated effect
estimates for all cause, CPD or lung cancer deaths, but IHD did show a positive association
with PM2.5 concentration. The difference between the 90th and 10th percentiles of the
3-yr avg PM2.5 concentration was 1.5 ug/m3 and the difference between the minimum and
July 2009
7-131
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
maximum values of the 3"yr avg PMa.5 concentration was 7.8 ug/nr1. This narrow range in
I'M2.5 exposure contrasts across the New York City metropolitan area may well account for
the inconclusive results in this city specific analysis. Relatively uniform exposures would
reduce the power of the statistical models to detect patterns of mortality relative to
exposure and estimate the association with precision.
WHI Study: This study also investigated the within- vs. between-city effects in its
cities. As shown in Figure 7*8, similar effects for both the within and between-city analyses
demonstrate that this association is not due to some other (non-pollution) confounder
differing between the various cities, strengthening confidence in the overall pollution-effect
estimates.
A Overall Effect
12-i
11-
10-
C Within-City Effect
12-i
11-
10-
B Between-City Effect
12-i
11-
10-
£	S
3	D
0	L.
w	re
¦B g
o 0
Q w
w £
3 o
js
O 3
30
linii|ii mini ii|uiii|iuii| niii|uni|iiiii|Ti ni|)ii ii|
Source: Miller et al. (200/, 090130)
Figure 7-8. Plots of the relative risk of death from cardiovascular disease from the Women's Health
Initiative study displaying the between-city and within-city contributions to the overall
association between PM2.5 and cardiovascular mortality windows of exposure-effects.
7.6.4. Effects of Different Long-term Exposure Windows
The delay between changes in exposure and changes in health has important policy
implications. Schwartz et al. (2008, 156963) investigated this issue using an extended
July 2009
7-132
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
followup of the Harvard Six Cities Study. Cox proportional hazards models were fit to
control for smoking, body mass index, and other covariates. Penalized splines were fit in a
flexible functional form to the concentration response to examine its shape, and the degrees
of freedom for the curve were selected based on Akaike's information criterion (AIC). The
researchers also used model averaging as an alternative approach, where multiple models
are fit explicitly and averaged, weighted by their probability of being correct given the data.
The lag relationship by model was averaged across a range of unconstrained distributed lag
models (i.e., same year, year prior, two years prior, etc.). Results of the lag comparison are
shown in Figure 7"9 indicating that the effects of changes in exposure on mortality are seen
within two years. The authors also noted that the concentration-response curve was linear,
clearly continuing below the level of the current U.S. air quality standard of 15 ug/m3.
1.20
ec
.10
> 1.05
cr 1.00
Year before death
Source: Schwartz et al. (2008,156963)
Figure 7-9. The model-averaged estimated effect of a 10-/vg/m3 increase in PM2.5 on all-cause
mortality at different lags (in years) between exposure and death. Each lag is estimated
independently of the others. Also shown are the pointwise 95% CIs for each lag, based
on jacknife estimates.
Similarly, the effect of long-term exposure to PMioon the risk of death in a large
multicity study of elderly subjects discharged alive following an admission for COPD found
the effect was not limited to the exposure in each year of follow-up, and had larger
cumulative effects spread over the follow-up year and 3 preceding years (Zanobetti et al.,
2008, 156177).
Roosli et al. (2005, 156923) took an alternative approach to determining the window
over which the mortality effects of long-term pollution exposures occurred. They fit the
model shown in Figure 7-10 using k = 0.5 based on the Utah Steel Strike (Pope CA, 1989,
044461) and the Ireland coal ban study (Clancy et al., 2002, 035270). They found that
July 2009
7-133
DRAFT-DO NOT CITE OR QUOTE

-------
1	roughly 75% of health benefits are observed in the first 5 years, as shown in Table 7-10.
2	These results are consistent with the findings of Schwartz et al. (2008, 156963). Puett et al.
3	(2008, 156891) also compared different long-term lags, with exposure periods ranging from
4	1 month to 48 months prior to death. They found statistically significant associations with
5	average PMio exposures in the time period 3-48 months prior to death, with the strongest
6	associations in the 24 months prior to death and the weakest with exposure in the 1 month
7	prior to death.
io
O —
2000 2002 2004 2006 2008
Time (years)
Source: Roosli et al. (2005,156923)
Figure 7-10. Time course of relative risk of death after a sudden decrease in air pollution exposure
during the year 2000, assuming a steady state model (solid line) and a dynamic model
(bold dashed line). The thin dashed line refers to the reference scenario. Table 7-10.
Distribution
of the effect of a hypothetical reduction of 10 yug/m3 PM10 in 2000 on all-cause mortality
2000-2009 in Switzerland.
Yr
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Proportion of total effect (%)

39.3
23.9
14.5
8.8
5.3
3.2
2.0
1.2
0.7
0.4
Relative risk (per 10/yg/m3 reduction in PMio)
1.0
0.9775
0.9863
0.9917
0.9950
0.9969
0.9981
0.9989
0.9993
0.9996
0.9997
Relative risk and proportion of total effect in each year are shown, assuming a time constant k of 0.5
Source: Roosli et al. (2005,156923)
July 2009
7-134
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
In the reanalysis of the ACS cohort, the investigators calculated time windows of
exposure as average concentrations during successive five-yr periods preceding the date of
death (Krewski et al., 2009, 191193). The investigators considered the time window with
the best-fitting model (judged by the AIC statistic) to be the period during which pollution
had the strongest influence on mortality. Overall, the differences between the time periods
were small and demonstrated no definitive patterns. High correlations between exposure
levels in the three periods may have reduced the ability of this analysis to detect any
differences in the relative importance of the time windows. The investigators did not
analyze any time periods smaller than five years, so the results are not directly comparable
to those reported by Schwartz et al. (2008, 156963), Roosli et al. (2005, 156923), and Puett
et al. (2008, 156891).
Generally, these results indicate a developing coherence of the air pollution mortality
literature, suggesting that the health benefits from reducing air pollution do not require a
long latency period and would be expected within a few years of intervention.
7.6.5. Summary and Causal Determinations
7.6.5.1. PM2.5
In the 1996 PM AQCD (U.S. EPA, 1996, 079380), results were presented for three
prospective cohort studies of adult populations: the Six Cities Study (Dockery et al., 1993,
044457); the American Cancer Society (ACS) Study (Pope CAet al., 1995, 045159); and the
California Seventh Day Adventist (AHSMOG) Study (Abbey et al., 1995, 000669). The 1996
AQCD concluded that the chronic exposure studies, taken together, suggested associations
between increases in mortality and long-term exposure to fine PM, though there was no
evidence to support an association with PM10-2.5 (U.S. EPA, 1996, 079380). Discussions of
mortality and long-term exposure to PM in the 2004 PM AQCD emphasized the results of
four U.S. prospective cohort studies, but the greatest weight was placed on the findings of
the ACS and the Harvard Six Cities studies, which had undergone extensive independent
reanalysis, and which were based on cohorts that were broadly representative of the U.S.
population. Collectively, the 2004 PM AQCD found that these studies provided strong
evidence that long-term exposure to PM2.5 was associated with increased risk of human
mortality.
The recent evidence is largely consistent with past studies, further supporting the
evidence of associations between long-term PM2.5 exposure and increased risk of human
mortality (Section 7.6) in areas with mean concentrations from 13.2 to 29 ug/ma
(Figure 7-7). New evidence from the Six Cities cohort study shows a relatively large risk
estimate for reduced mortality risk with decreases in PM2.5 (Laden et al., 2006, 087605).
The results of new analyses from the Six Cities cohort and the ACS study in Los Angeles
July 2009
7-135
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
suggest that previous and current studies may have underestimated the magnitude of the
association (Jerrett et al., 2005, 189405). With regard to mortality by cause-of-death, recent
ACS analyses indicate that cardiovascular mortality primarily accounts for the total
mortality association with PM2.5 among adults, and not respiratory mortality The recent
WHI cohort study shows even higher cardiovascular risks per pg/m.'i than found in the ACS
study, but this is likely due to the fact that the study included only women without pre-
existing cardiovascular disease (Miller et al., 2007, 090130). There is additional evidence for
an association between PM2.5 exposure and lung cancer mortality (Section7.5.1.l). The WHI
study also considered within vs. between city mortality, as well as morbidity co-associations
with PM2.5 in the same population. The first showed that the results are not due to between
city confounding, and the morbidity analyses show the coherence of the mortality
association across health endpoints, supporting the biological plausibility of the air
pollution-mortality associations found in these studies.
Results from a new study examining the relationship between life expectancy and
PM2.5 and the findings from a multiyear expert judgment study that comprehensively
characterizes the size and uncertainty in estimates of mortality reductions associated with
decreases in PM2.5 in the U.S draw conclusions that are consistent with an association
between long-term exposure to PM2.5 and mortality (Pope et al., 2009, 190107; Roman et
al., 2008, 156921). Pope et al. (2009, 190107) report that a decrease of 10 ug/m in the
concentration of PIVk.sis associated with an estimated increase in mean (± SE) life
expectancy of 0.61 ± 0.20 year. For the approximate period of 1980-2000, the average
increase in life expectancy was 2.72 years among the 211 counties in the analysis. The
authors note that reduced air pollution was only one factor contributing to increased life
expectancies, with its effects overlapping with those of other factors.
Roman et al. (2008, 156921) applied state-of-the-art expert judgment elicitation
techniques to develop probabilistic uncertainty distributions that reflect the broader array
of uncertainties in the concentration-response relationship. This study followed best
standard practices for expert elicitations based on the body of literature accumulated over
the past two decades. The resulting PM2.5 effect estimate distributions, elicited from 12 of
the world's leading experts on this issue, are shown in Figure 7-11. They indicate both
larger central estimates of mortality reductions for decreases in long-term PM2.5 exposure
in the U.S. (averaging almost 1% per pg/m3 PM2.5) than reported (for example) by the ACS
Study (i.e., 0.6% per jug/m3 PM2.5 in Pope et al. (2002, 024689), and a wider distribution of
uncertainty by each expert than provided by any one of the PM2.5 epidemiologic studies.
However, a composite uncertainty range of the overall mean effect estimate (i.e., based
upon all 12 experts' estimates, but not provided in Figure 7-11) would be much narrower,
and closer to that derived from the ACS study than indicated for any one expert shown in
Figure 7-11.
July 2009
7-136
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Group 1
Group 2
s
I
Range of PM2 5 Q -J	e	1	1	f	f	1	i	1	i	\	I	1	i	1	1	¦	f	i	
Concentration	4-30 4-10 >10-30 4*10 >10-30 4-30 4-30 4-30 4-16 >16-30 4-7 >7-30 4*30 4-30 4-30 4-30 pope OocKery
Causality Likelihood 99% 75% 99% 98% 98% 95% 95% 70% 35% 35% 100% 100% 99% 99% 95% 90% et al et al,
Expert	E	L	B	DIG	K	F	CJAH 2002 1993
Key: Closed circle = median; Open circle = mean; Box = interquartile range; Solid line = 90% credible interval
Source: Roman et al. (2008,156921)
Figure 7-11. Experts' mean effect estimates and uncertainty distributions for the PM2.5 mortality
concentration-response coefficient for a 1 yug/m3 change in annual average PM2.5
Overall, recent evidence supports the strong evidence reported in the 2004 PM AQCD
(EPA, 2004, 056905) that long-term exposure to PM2.5 is associated with an increased risk
of human mortality. When looking at the cause of death, the strongest evidence comes from
mortality due to cardiovascular disease, with additional evidence supporting an association
between PM2.5 and lung cancer mortality (Figure 7-7). Fewer studies evaluate the
respiratory component of cardiopulmonary mortality, and the evidence to support an
association with long-term exposure to PM2.5 and respiratory mortality is weak (Figure 7-7).
Together these findings are consistent and coherent with the evidence from epidemiologic,
controlled human exposure, and animal toxicological studies for the effects of short- and
long-term exposure to PM on cardiovascular effects presented in Sections 6.2 and 7.2,
respectively. Evidence of short- and long-term exposure to PM2.5 and respiratory effects
(Sections 6.3 and 7.3, respectively) and infant mortality (Section 7.4) are coherent with the
weak respiratory mortality effects. Additionally, the evidence for short- and long-term
cardiovascular and respiratory morbidity provides biological plausibility for mortality due
to cardiovascular or respiratory disease. The most recent evidence for the association
between long-term exposure to PM2.5 and mortality is particularly strong for women.
Collectively, the evidence is sufficient to conclude that the relationship between long-term
PM2.5 exposures and mortality is likely to be causal.
July 2009
7-137
DRAFT-DO NOT CITE OR QUOTE

-------
7.6.5.2. PMio-2.5
1	In the 2004 PM AQCD, results from the ACS and Six Cities study analyses indicated
2	that thoracic coarse particles were not associated with mortality. Evidence is still limited to
3	adequately characterize the association between PM10-2.5 and PM sources and/or
4	components. The new findings from AHSMOG and Veterans cohort studies provide limited
5	evidence of associations between long-term exposure to PM10-2.5 and mortality in areas with
6	mean concentrations from 16 to 25 jug/m3. The evidence for PM 10-2.5 is inadequate to determine
7	if a causal relationship exists between long-term exposures and mortality.
July 2009
7-138
DRAFT-DO NOT CITE OR QUOTE

-------
Chapter 7 References
Abbey DE; Lebowitz MD; Mills PE Petersen FF; Beeson WL; Burchette RJ. (1995). Long-term ambient
concentrations of particulates and oxidants and development of chronic disease in a cohort of nonsmoking
California residents. Presented at In- Phalen, R. F.i Bates, D. V., eds. Proceedings of the colloquium on
particulate air pollution and human mortality and morbidity; January 1994; Irvine, CA. Inhalation
Toxicol. 7: 19-34., Irvine, CA. 000669
Abbey DE; Mills PE Petersen FF; Beeson WL. (1991). Long-term ambient concentrations of total suspended
particulates and oxidants as related to incidence of chronic disease in California Seventh-Day Adventists.
Environ Health Perspect, 94- 43-50. 042668
Abbey DE; Nishino N; McDonnell WF; Burchette RJ; Knutsen SF; Beeson WL; Yang JX. (1999). Long-term
inhalable particles and other air pollutants related to mortality in nonsmokers. Am J Respir Crit Care
Med, 159: 373-382. 047559
Abou Chakra OR; Joyeux M; Nerriere E; Strub MP; Zmirou-Navier D. (2007). Genotoxicity of organic extracts of
urban airborne particulate matter- an assessment within a personal exposure study. Chemosphere, 66-
1375-81. 098819
Achenbach S; Daniel WG. (2001). Noninvasive coronary angiography--an acceptable alternative?. , 345- 1909"
1910. 156189
Ackermann-Liebrich U; Kuna-Dibbert B; Probst-Hensch NM; Schindler C; Dietrich DF; Stutz EZ; Bayer-
Oglesby L; Baum F; Brandli O; Brutsche M; Downs SH; Keidel D; Gerbase MWi Imboden M; Keller R;
Knopfli B; Kunzli Nellweger J-P; Leuenberger P; SALPADIA Team. (2005). Follow-up of the Swiss
Cohort study on air pollution and lung diseases in adults (SAPALDIA 2) 1991-2003- methods and
characterization of participants. Int J Public Health, 50- 245-263. 087826
Ackermann-Liebrich U; Leuenberger P; Schwartz J; Schindler C; Monn C; Bolognini B; Bongard JP; Brandli O;
Domenighetti G; Elsasser S; Grize L; Karrer W; Keller R; Keller-Wossidlo H; Kunzli N; Martin BW;
Medici TC; Pliger B; Wuthrich B; Zellweger JP; Zemp E. (1997). Lung function and long term exposure to
air pollutants in Switzerland. Am J Respir Crit Care Med, 155: 122-129. 077537
Agatston AS; Janowitz WR; Hildner FJ; Zusmer NR; Viamonte M Jr; Detrano R. (1990). Quantification of
coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol, 15: 827-832. 156197
Alink GM; Sjogren M; Bos RP; Doekes G; Kromhout H; Scheepers PTJ. (1998). Effect of airborne particles from
selected indoor and outdoor environments on gap-junctional intercellular communication. Toxicol Lett,
96/97: 209-213. 087159
Allen RW; Criqui MH; Diez Roux AV; Allison M; Shea S; Detrano R; Sheppard L; Wong N; Hinckley Stukovsky
K; Kaufman JD. (2009). Fine particulate air pollution, proximity to traffic, and aortic atherosclerosis:
The Multi-Ethnic Study of Atherosclerosis. Epidemiology, 20: 254-264. 189644
Allison MA; Cheung P; Criqui MH; Langer RD; Wright CM. (2006). Mitral and aortic annular calcification are
highly associated with systemic calcified atherosclerosis. Circulation, 113: 861-866. 155653
Allison MA; Criqui MH; Wright CM. (2004). Patterns and risk factors for systemic calcified atherosclerosis.
Arterioscler Thromb Vase Biol, 24: 331-336. 156210
Amato M; Montorsi P; Ravani A; Oldani E; Galli S; Ravagnani PM; Tremoli E; Baldassarre D. (2007). Carotid
intima-media thickness by B-mode ultrasound as surrogate of coronary atherosclerosis: correlation with
quantitative coronary angiography and coronary intravascular ultrasound findings. Eur Heart J, 28:
2094-2101. 155656
Annesi-Maesano I; Moreau D; Caillaud D; Lavaud F; Le Moullec Y; Taytard A; Pauli G; Charpin D. (2007).
Residential proximity fine particles related to allergic sensitisation and asthma in primary school
children. Respir Med, 10L 1721-1729. 093180
Arad Y; Spadaro LA; Goodman K; Lledo-Perez A; Sherman S; Lerner G; Guerci AD. (1996). Predictive value of
electron beam computed tomography of the coronary arteries. 19-month follow-up of 1173 asymptomatic
subjects. Circulation, 93: 1951-1953. 155661
Araujo JA; Barajas B; Kleinman M; Wang X; Bennett BJ; Gong KW; Navab M; Harkema J; Sioutas C; Lusis AJ;
Nel AE. (2008). Ambient particulate pollutants in the ultrafine range promote early atherosclerosis and
systemic oxidative stress. Circ Res, 102: 589-596. 156222
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009	7-139	DRAFT - DO NOT CITE OR QUOTE

-------
Ardehali R; Nasir Ki Kolandaivelu A; Budoff MJ; Blumenthal RS. (2007). Screening patients for subclinical
atherosclerosis with non-contrast cardiac CT. Atherosclerosis, 192- 235-242. 155662
Arlt VM; Glatt H; Gamboa da Costa G; Reynisson J; Takamura-Enya T; Phillips DH. (2007). Mutagenicity and
DNA adduct formation by the urban air pollutant 2-nitrobenzanthrone. , 98- 445-57. 097257
Auchincloss AH; Roux AV; Dvonch JT; Brown PL; Barr RG; Daviglus ML; Goff DC; Kaufman JD; O'Neill MS.
(2008). Associations between Recent Exposure to Ambient Fine Particulate Matter and Blood Pressure in
the Multi-Ethnic Study of Atherosclerosis (MESA). Environ Health Perspect, 116- 486-491. 156234
Avogbe PH; Ayi-Fanou L; Autrup H; Loft S; Fayomi B; Sanni A; Vinzents P; Moller P. (2005). Ultrafine
particulate matter and high-level benzene urban air pollution in relation to oxidative DNA damage.
Carcinogenesis, 26- 613-620. 087811
Avol EL; Gauderman WJ; Tan SM; London SJ; Peters JM. (2001). Respiratory effects of relocating to areas of
differing air pollution levels. Am J Respir Crit Care Med, 164: 2067-2072. 020552
Baccarelli A; Martinelli I; Pegoraro V; Melly S; Grillo P; Zanobetti A; Hou L; Bertazzi PA; Mannucci PM;
Schwartz J. (2009). Living near Major Traffic Roads and Risk of Deep Vein Thrombosis. Circulation, x: x.
188183
Baccarelli A; Martinelli I; Zanobetti A; Grillo P; Hou LF; Bertazzi PA; Mannucci PM; Schwartz J. (2008).
Exposure to Particulate Air Pollution and Risk of Deep Vein Thrombosis. Arch Intern Med, 168; 920-927.
157984
Baccarelli A; Zanobetti A; Martinelli I; Grillo P; Hou L; Giacomini S; Bonzini M; Lanzani G; Mannucci PM;
Bertazzi PA; Schwartz J. (2007). Effects of exposure to air pollution on blood coagulation. J Thromb
Haemost, 5: 252-260. 090733
Baker DG. (1998). Natural pathogens of laboratory mice, rats, and rabbits and their effects on research. Clin
Microbiol Rev, 11: 231-266. 156245
Bao L; Chen S; Wu L; Hei Tom K; Wu Y; Yu Z; Xu A. (2007). Mutagenicity of diesel exhaust particles mediated
by cell-particle interaction in mammalian cells. Toxicol Sci, 229: 91-100. 097258
Basu R; Woodruff TJ; Parker JD; Saulnier L; Schoendorf KC. (2004). Comparing exposure metrics in the
relationship between PM25 and birth weight in California. J Expo Sci Environ Epidemiol, 14: 391-396.
087896
Bayer-Oglesby L; Grize L; Gassner M; Takken-Sahli K; Sennhauser FH; Neu U; Schindler C; Braun-Fahrlander
C. (2005). Decline of ambient air pollution levels and improved respiratory health in Swiss children.
Environ Health Perspect, 113: 1632-1637. 086245
Beelen R; Hoek G; van den Brandt PA; Goldbohm RA; Fischer P; Schouten LJ; Armstrong B; Brunekreef B.
(2008). Long-term exposure to traffic-related air pollution and lung cancer risk. , 19: 702-710. 155681
Beelen R; Hoek G; van den Brandt PA; Goldbohm RA; Fischer P; Schouten LJ; Jerrett M; Hughes E; Armstrong
B; Brunekreef B. (2008). Long-term effects of traffic-related air pollution on mortality in a Dutch cohort
(NLCS-AIR study). Environ Health Perspect, 116: 196-202. 156263
Beeson WL; Abbey DE; Knutsen SF. (1998). Long-term concentrations of ambient air pollutants and incident
lung cancer in California adults: results from the AHSMOG study. Environ Health Perspect, 106: 813-
823.048890
Bell ML; Ebisu K; Belanger K. (2007). Ambient air pollution and low birth weight in Connecticut and
Massachusetts. Environ Health Perspect, 115: 1118-24. 091059
Bell ML; Kim JY; Dominici F. (2007). Potential confounding of particulate matter on the short-term association
between ozone and mortality in multisite time-series studies. Environ Health Perspect, 115: 1591-1595.
093256
Bhattacharyya N. (2009). Air quality influences the prevalence of hay fever and sinusitis. Laryngoscope, 119:
429-433. 180154
Bigazzi R; Bianchi S; Baldari D; Campese VM. (1998). Microalbuminuria predicts cardiovascular events and
renal insufficiency in patients with essential hypertension. J Hypertens, 16: 1325-1333. 156272
Binkova B; Cerna M; Pastorkova A; Jeli'nek R; Benes I; Novak J; Sram RJ. (2003). Biological activities of
organic compounds adsorbed onto ambient air particles: comparison between the cities of Teplice and
Prague during the summer and winter seasons 2000—2001. , 525: 43"59. 156274
Bobak M. (2000). Outdoor air pollution, low birth weight, and prematurity. Environ Health Perspect, 108: 173-
176. 011448
Bobak M; Leon DA. (1992). Air pollution and infant mortality in the Czech Republic, 1986-1988. Lancet, 8826:
1010-1014. 044415
Bobak M; Leon DA. (1999). The effect of air pollution on infant mortality appears specific for respiratory causes
in the postneonatal period. , 10: 666-670. 007678
July 2009
7-140
DRAFT-DO NOT CITE OR QUOTE

-------
Bonner MR; Han D; Nie J; Rogerson P; Vena JE; Muti P; Trevisan M; Edge SB; Freudenheim JL. (2005). Breast
cancer risk and exposure in early life to polycyclic aromatic hydrocarbons using total suspended
particulates as a proxy measure. Cancer Epidemiol Biomarkers Prev, 14- 53-60. 088993
Bracken MB; Triche EW; Belanger K; Saftlas A; Beckett WS; Leaderer BP. (2003). Asthma symptoms, severity,
and drug therapy- a prospective study of effects on 2205 pregnancies. Obstet Gynecol, 102- 739-753.
156288
Brauer M; Gehring U; Brunekreef B; De Jongste J; Gerritsen J; Rovers M; Wichmann H"E; Wijga A; Heinrich J.
(2006). Traffic-related air pollution and otitis media. Environ Health Perspect, 114: 1414-1418. 090757
Brauer M; Hoek G; Smit HA; De Jongste JC; Gerritsen J; Postma DS; Kerkhof M; Brunekreef B. (2007). Air
pollution and development of asthma, allergy and infections in a birth cohort. Eur Respir J, 29: 879-888.
090691
Brauer M; Hoek G; Van Vliet P; Meliefste K; Fischer PH; Wijga A; Koopman LP; Neijens HJ; Gerritsen J;
Kerkhof M; Heinrich J; Bellander T; Brunekreef B. (2002). Air pollution from traffic and the development
of respiratory infections and asthmatic and allergic symptoms in children. Am J Respir Crit Care Med,
166: 1092-1098. 035192
Brauer M; Lencar C; Tamburic L; Koehoorn M; Demers P; Karr C. (2008). A cohort study of traffic-related air
pollution impacts on birth outcomes. Environ Health Perspect, 116: 680-686. 156292
Brown JS; Zeman KL; Bennett WD. (2001). Regional deposition of coarse particles and ventilation distribution
in healthy subjects and patients with cystic fibrosis. , 14: 443-454. 012307
Brunekreef B; Beelen R; Hoek G; Schouten L; Bausch-Goldbohm S; Fischer P; Armstrong B; Hughes E; Jerrett
M; van den Brandt P. (2009). Effects of long-term exposure to traffic-related air pollution on respiratory
and cardiovascular mortality in the Netherlands: The NLCS-AIR Study. Health Effects Institute. Boston,
MA. 139. 191947
Budoff MJ; Takasu J; Katz R; Mao S; Shavelle DM; O'Brien KD; Blumenthal RS; Carr JJ; Kronmal R. (2005).
Reproducibility of CT measurements of aortic valve calcification, mitral annulus calcification, and aortic
wall calcification in the multi-ethnic study of atherosclerosis. Acad Radiol, 13: 166-172. 192105
Bunger J; Krahl J; Weigel A; Schroder O; Bruning T; Muller M; Hallier E; Westphal G. (2006). Influence of fuel
properties, nitrogen oxides, and exhaust treatment by an oxidation catalytic converter on the
mutagenicity of diesel engine emissions. Arch Toxicol, 80: 540-546. 156303
Bunger J; Schappler-Scheele B; Hilgers R; Hallier E. (2007). A 5-year follow-up study on respiratory disorders
and lung function in workers exposed to organic dust from composting plants. Int Arch Occup Environ
Health, 80: 306-312. 156304
Burchiel SW; Lauer FT; Dunaway SL; Zawadzki J; McDonald JD; Reed MD. (2005). Hardwood smoke alters
murine splenic T cell responses to mitogens following a 6-month whole body inhalation exposure. Toxicol
Appl Pharmacol, 202: 229-236. 088090
Burchiel SW; Lauer FT; McDonald JD; Reed MD. (2004). Systemic immunotoxicity in AJ mice following 6"
month whole body inhalation exposure to diesel exhaust. Toxicol Appl Pharmacol, 196: 337-345. 055557
Calderon-Garciduenas L; Vincent R; Mora-Tiscareno A; Franco-Lira M; Henriquez-Roldan C; Barragan-Mejia G;
Garrido-Garcia L; Camacho-Reyes L; Valencia-Salazar G; Paredes R; Romero L; Osnaya H; Villarreal"
Calderon R; Torres-Jardon. (2007). Elevated plasma endothelin-1 and pulmonary arterial pressure in
children exposed to air pollution. Environ Health Perspect, 115: 1248-1253. 091252
Chambless LE; Heiss G; Folsom AR; Rosamond W; Szklo M; Sharrett AR; Clegg LX. (1997). Association of
coronary heart disease incidence with carotid arterial wall thickness and major risk factors: the
Atherosclerosis Risk in Communities (ARIC) Study, 1987-1993. Am J Epidemiol, 146: 483-494. 156329
Chang J; Delfino RJ; Gillen D; Tjoa T; Nickerson B; Cooper D. (2008). Repeated respiratory hospital encounters
among children with asthma and residential proximity to traffic. J Occup Environ Med, 66: 90-98.
180393
Chen JC; Schwartz J. (2008). Metabolic syndrome and inflammatory responses to long-term particulate air
pollutants. Environ Health Perspect, 116: 612-617. 190106
Chen L; Yang W', Jennison BL; Goodrich A; Omaye ST. (2002). Air pollution and birth weight in northern
Nevada, 1991-1999. Inhal Toxicol, 14: 141-157. 024945
Chen LC; Hwang JS. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice
IV	Characterization of acute and chronic effects of ambient air fine particulate matter exposures on
heart-rate variability. Inhal Toxicol, 17: 209-216. 087218
Chen LC; Nadziejko C. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice
V	CAPs exacerbate aortic plaque development in hyperlipidemic mice. Inhal Toxicol, 17: 217-224. 087219
Chen LH; Knutsen SF; Shavlik D; Beeson WL; Petersen F; Ghamsary M; Abbey D. (2005). The association
between fatal coronary heart disease and ambient particulate air pollution: Are females at greater risk?.
Environ Health Perspect, 113: 1723-1729. 087942
July 2009
7-141
DRAFT-DO NOT CITE OR QUOTE

-------
Churg A; Brauer Mi del Carmen Avila-Casado Mi Fortoul Tli Wright JL. (2003). Chronic exposure to high levels
of particulate air pollution and small airway remodeling. Environ Health Perspect, 111- 714-718. 087899
Clancy Li Goodman Pi Sinclair Hi Dockery DW. (2002). Effect of air pollution control on death rates in Dublin,
Ireland- an intervention study. Lancet, 360- 1210-1214. 035270
Claxton LDi Matthews PPi Warren SH. (2004). The genotoxicity of ambient outdoor air, a review- Salmonella
mutagenicity. DNA Repair (Amst), 567- 347-399. 089008
Claxton LDi Woodall GM Jr. (2007). A review of the mutagenicity and rodent carcinogenicity of ambient air. ,
636: 36-94. 180391
Clifton VL; Giles WB; Smith R; Bisits AT; Hempenstall PA; Kessell CG; Gibson PG. (2001). Alterations of
placental vascular function in asthmatic pregnancies. Am J Respir Crit Care Med, 164: 546-554. 156360
Craven TE; Ryu JE; Espeland MA; Kahl FR; McKinney WM; Toole JE; McMahan MR; Thompson CJ; Heiss G;
Crouse JR, 3rd. (1990). Evaluation of the associations between carotid artery atherosclerosis and
coronary artery stenosis. A case-control study. , 82: 1230-1242. 155740
Cury PM; Lichtenfels AJ; Reymao MS; Conceicjao GM; Capelozzi VL; Saldiva PH. (2000). Urban levels of air
pollution modifies the progression of urethane-induced lung tumours in mice. Pathol Res Pract, 196: 627"
633. 192100
Dales R; Burnett RT; Smith-Doiron M; Stieb DM; Brook JR. (2004). Air pollution and sudden infant death
syndrome. , 113: 628-631. 087342
Dales R; Wheeler A; Mahmud M; Frescura AM; Smith-Doiron M; Nethery E; Liu L. (2008). The Influence of
Living Near Roadways on Spirometry and Exhaled Nitric Oxide in Elementary Schoolchildren. Environ
Health Perspect, 116: 1423. 156378
Danielsen PH; Loft S; Moller P. (2008). DNA damage and cytotoxicity in type II lung epithelial (A549) cell
cultures after exposure to diesel exhaust and urban street particles. Part Fibre Toxicol, 5- 6. 192092
De Kok TM; Hogervorst JG; Briede JJ; Van Herwijnen MH! Maas LM; Moonen EJ; Driece HA; Kleinjans JC.
(2005). Genotoxicity and physicochemical characteristics of traffic-related ambient particulate matter.
Environ Mol Mutagen, 46: 71-80. 088656
Deckert T; Yokoyama H; Mathiesen E; Ronn B; Jensen T; Feldt-Rasmussen B; Borch-Johnsen K; Jensen JS.
(1996). Cohort study of predictive value of urinary albumin excretion for atherosclerotic vascular disease
in patients with insulin dependent diabetes. Br Med J, 312: 871-874. 156389
DeMarini DM; Brooks LR; Warren SH; Kobayashi T; Gilmour MI; Singh P. (2004). Bioassay-directed
fractionation and Salmonella mutagenicity of automobile and forklift diesel exhaust particles. Environ
Health Perspect, 112: 814-819. 066329
Diez Roux AV; Auchincloss AH; Franklin TG; Raghunathan T; Barr RG; Kaufman J; Astor B; Keeler J. (2008).
Long-term exposure to ambient particulate matter and prevalence of subclinical atherosclerosis in the
Multi-Ethnic Study of Atherosclerosis. Am J Epidemiol, 167: 667-675. 156401
Dinneen SF; Gerstein HC. (1997). The association of microalbuminuria and mortality in non-insulin-dependent
diabetes mellitus. A systematic overview of the literature. , 157: 1413-1418. 156403
Dockery DW; Cunningham J; Damokosh AI; Neas LM; Spengler JD; Koutrakis P; Ware JH; Raizenne M;
Speizer FE. (1996). Health effects of acid aerosols on North American children: respiratory symptoms.
Environ Health Perspect, 104: 500-505. 077269
Dockery DW; Pope CA III; Xu X; Spengler JD; Ware JH; Fay ME; Ferris BG Jr; Speizer FE. (1993). An
association between air pollution and mortality in six US cities. , 329: 1753-1759. 044457
Downs SH; Schindler C; Liu L-JS; Keidel D; Bayer-Oglesby L; Brutsche MH; Gerbase MWi Keller R; Kunzli N;
Leuenberger P; Probst-Hensch NM; Tschopp J-M; Zellweger J-P; Rochat T; Schwartz J; Ackermann-
Liebrich U;. (2007). Reduced exposure to PM10 and attenuated age-related decline in lung function. , 15:
185-204. 092853
Dugandzic RDodds L; Stieb D; Smith-Doiron M. (2006). The association between low level exposures to ambient
air pollution and term low birth weight: a retrospective cohort study. , 5- 3. 088681
Eftim SE; Samet JM; Janes H; McDermott A; Dominici F. (2008). Fine Particulate Matter and Mortality: A
Comparison of the Six Cities and American Cancer Society Cohorts With a Medicare Cohort. , 19: 209"
216. 099104
Enstrom JE. (2005). Fine particulate air pollution and total mortality among elderly Californians, 1973-2002.
Inhal Toxicol, 17: 803-816. 087356
Erbel R; Mohlenkamp S; Kerkhoff G; Budde T; Schmermund A. (2007). Non-invasive screening for coronary
artery disease: calcium scoring. Heart, 93: 1620-1629. 155768
Erdinger L; Durr M; Hopker KA. (2005). Correlations between mutagenic activity of organic extracts of airborne
particulate matter, NOx and sulphur dioxide in southern Germany: results of a two-year study. Environ
Sci Pollut Res Int, 12: 10-20. 156423
July 2009
7-142
DRAFT-DO NOT CITE OR QUOTE

-------
Fedulov AV; Leme A; Yang Z; Dahl M> Lim R; Mariani TJ; Kobzik L. (2008). Pulmonary exposure to particles
during pregnancy causes increased neonatal asthma susceptibility. , 38- 57"67. 097482
Finch GL; Hobbs CH; Blair LF; Barr EB; Hahn FF; Jaramillo RJ; Kubatko JE; March TH; White RE Krone JR;
Menache MG; Nikula KJ; Mauderly JL; Van Gerpen J; Merceica MD! Zielinska B; Stankowski L; Burling
K; Howell S> Mauderly JL. (2002). Effects of subchronic inhalation exposure of rats to emissions from a
diesel engine burning soybean oil-derived biodiesel fuel. Inhal Toxicol, 14- 1017-1048. 054603
Floyd HS; Chen LC; Vallanat B; Dreher K. (2009). Fine ambient air particulate matter exposure induces
molecular alterations associated with vascular disease progression within plaques of atherosclerotic
susceptible mice. Inhal Toxicol, 21: 394-403. 190350
Forbes LJ; Patel MDi Rudnicka AR; Cook DG; Bush Ti Stedman JR; Whincup PH; Strachan DP; Anderson RH.
(2009). Chronic exposure to outdoor air pollution and markers of systemic inflammation. Epidemiology,
20: 245-253. 190351
Forman JP; Brenner BM. (2006). 'Hypertension' and 'microalbuminuria': the bell tolls for thee. Kidney Int, 69:
22-28. 156439
Foster PMD; Mylchreest E; Gaido KW; Sar M. (2001). Effects of phthalate esters on the developing reproductive
tract of male rats. Hum Reprod Update, 7: 231-235. 156442
Foster PMD; Thomas LV; Cook MWi Gangolli SD. (1980). Study of the testicular effects and changes in zinc
excretion produced by some n-alkyl phthalates in the rat. Toxicol Appl Pharmacol, 54: 392-398. 094701
Franklin SS; Sutton-Tyrrell K; Belle SH; Weber MA; Kuller LH. (1997). The importance of pulsatile components
of hypertension in predicting carotid stenosis in older adults. J Hypertens, 15: 1143-1150. 156446
Fryer ME; Collins CD. (2003). Model intercomparison for the uptake of organic chemicals by plants. Environ Sci
Technol, 37: 1617-1624. 156454
Fujimoto A; Tsukue N; Watanabe M; Sugawara I; Yanagisawa R; Takano H; Yoshida S; Takeda K. (2005).
Diesel exhaust affects immunological action in the placentas of mice. , 20: 431-440. 096556
Gardner MB. (1966). Biological effects of urban air pollution III Lung tumors in mice. Arch Environ Occup
Health, 12: 305-313. 015129
Gardner MB; Loosli CG; Hanes B; Blackmore W', Teebken D. (1969). Histopathologic findings in rats exposed to
ambient and filtered air. Arch Environ Occup Health, 19: 637-647. 015130
Gauderman WJ; Avol E; Gilliland F; Vora H; Thomas D; Berhane K; McConnell R; Kuenzli N; Lurmann F;
Rappaport E; Margolis H; Bates D; Peters J. (2004). The effect of air pollution on lung development from
10 to 18 years of age. , 351: 1057-1067. 056569
Gauderman WJ; Gilliland GF; Vora H; Avol E; Stram D; McConnell R; Thomas D; Lurmann F; Margolis HG;
Rappaport EB; Berhane K; Peters JM. (2002). Association between air pollution and lung function
growth in southern California children: results from a second cohort. Am J Respir Crit Care Med, 166:
76-84. 026013
Gauderman WJ; McConnell R; Gilliland F; London S; Thomas D; Avol E; Vora H; Berhane K; Rappaport EB;
Lurmann F; Margolis HG; Peters J. (2000). Association between air pollution and lung function growth in
southern California children. Am J Respir Crit Care Med, 162: 1383-1390. 012531
Gehrig R; Buchmann B. (2003). Characterising seasonal variations and spatial distribution of ambient PM10
and PM25 concentrations based on long-term Swiss monitoring data. Atmos Environ, 37: 2571-2580.
139678
Gehring U; Heinrich J; Kramer U; Grote V; Hochadel M; Sugiri D; Kraft M; Rauchfuss K; Eberwein HG;
Wichmann H"E. (2006). Long-term exposure to ambient air pollution and cardiopulmonary mortality in
women. , 17: 545-551. 089797
Geroulakos G; O'Gorman DJ; Kalodiki E; Sheridan DJ; Nicolaides AN. (1994). The carotid intima-media
thickness as a marker of the presence of severe symptomatic coronary artery disease. , 15: 781-785.
155788
Gerstein HC; Mann JF; Yi Q; Zinman B; Dinneen SF; Hoogwerf B; Halle JP; Young J; Rashkow A; Joyce C;
Nawaz S; Yusuf S. (2001). Albuminuria and risk of cardiovascular events, death, and heart failure in
diabetic and nondiabetic individuals. JAMA, 286: 421-426. 156466
Ghosh R; Rankin J; Pless-Mulloli T; Glinianaia S. (2007). Does the effect of air pollution on pregnancy outcomes
differ by gender? A systematic review. Environ Res, 105: 400-408. 091233
Gilboa SM; Mendola P; Olshan AF; Langlois PH; Savitz DA; Loomis D; Herring AH; Fixler DE. (2005). Relation
between ambient air quality and selected birth defects, seven county study, Texas, 1997-2000. Am J
Epidemiol, 162: 238-252. 087892
Girerd X; Mourad JJ; Acar C; Heudes D; Chiche S; Bruneval P; Mignot JP; Billaud E; Safar M; Laurent S.
(1994). Noninvasive measurement of medium-sized artery intima-media thickness in humans: in vitro
validation. J Vase Res, 31: 114-120. 156474
July 2009
7-143
DRAFT-DO NOT CITE OR QUOTE

-------
Glinianaia FV; Rankin J; Bell R; Pless-Mulloli Ti Howel D. (2004). Does particulate air pollution contribute to
infant death? A systematic review. Environ Health Perspect, 112- 1365-1371. 087898
Goss CH; Newsom SA; Schildcrout JS; Sheppard L; Kaufman JD. (2004). Effect of ambient air pollution on
pulmonary exacerbations and lung function in cystic fibrosis. Am J Respir Crit Care Med, 169- 816-821.
055624
Gotschi T> Heinrich J; Sunyer J; Kunzli N. (2008). Long-term effects of ambient air pollution on lung function -
A review. , 19: 690-701. 156485
Gotschi Ti Sunyer J; Chinn Si de Marco Ri Forsberg Bi Gauderman JWi Garcia-Esteban Ri Heinrich Ji
Jacquemin Bi Jarvis Di Ponzio Mi Villani Si Kunzli N. (2008). Air pollution and lung function in the
European Community Respiratory Health Survey. Int J Epidemiol, 37: 1349-1358. 180364
Gottipolu RRi Wallenborn JGi Karoly ED; Schladweiler MC; Ledbetter AD; Krantz T; Linak WP; Nyska A;
Johnson JA; Thomas R; Richards JE; Jaskot RH; Kodavanti UP. (2009). One-month diesel exhaust
inhalation produces hypertensive gene expression pattern in healthy rats. Environ Health Perspect, 117:
38-46. 190360
Gouveia N> Bremner SA; Novaes HMD. (2004). Association between ambient air pollution and birth weight in
Sao Paulo, Brazil. J Epidemiol Community Health, 58: 11-17. 055613
Greenland P; Kizilbash MA. (2005). Coronary computed tomography in coronary risk assessment. , 25: 3-10.
156496
Grigg J; Kulkarni N; Pierse N; Rushton L; O'Callaghan C; Rutman A. (2008). Black-pigmented material in
airway macrophages from healthy children: association with lung function and modeled PM10. 156499
Gunnison A; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice
VI Gene expression in heart and lung tissue. Inhal Toxicol, 17: 225-233. 087956
Gutierrez-Castillo ME; Roubicek DA; Cebrian-Garcia ME; De Vizcaya-Ruiz A; Sordo-Cedeno M; Ostrosky-
Wegman P. (2006). Effect of chemical composition on the induction of DNA damage by urban airborne
particulate matter. Environ Mol Mutagen, 47: 199-211. 089030
Gabelova A; Valovicova Z; Bacova G; Labaj J; Binkova B; Topinka J; Sevastyanova O; Sram RJ; Kalina I;
Habalova V; Popov TA; Panev T; Farmer PB. (2007). Sensitivity of different endpoints for in vitro
measurement of genotoxicity of extractable organic matter associated with ambient airborne particles
(PM10). , 620: 103-113. 156458
Gabelova A; Valovicova Z; Labaj J; Bacova G; Binkova B; Farmer Peter B. (2007). Assessment of oxidative DNA
damage formation by organic complex mixtures from airborne particles PM(10). , 620: 135-144. 156457
Ha E-H; Lee J-T; Kim H; Hong Y-C; Lee. (2003). Infant susceptibility of mortality to air pollution in Seoul,
South Korea. , 111: 284-290. 042552
Haland G; Carlsen KC; Sandvik L; Devulapalli CS; Munthe-Kaas MC; Pettersen M; Carlsen KH. (2006).
Reduced lung function at birth and the risk of asthma at 10 years of age. , 355: 1682-1689. 156511
Hamada K; Suzaki Y; Leme A; Ito T; Miyamoto K; Kobzik L; Kimura H. (2007). Exposure of pregnant mice to an
air pollutant aerosol increases asthma susceptibility in offspring. J Toxicol Environ Health A, 70: 688"
695. 091235
Hannigan MP; Cass GR; Penman BW; Crespi CL; Lafleur AL; Busby WF Jr; Thilly WG. (1997). Human cell
mutagens in Los Angeles air. Environ Sci Technol, 31: 438-447. 083598
Hansen C; Neller A; Williams G; Simpson R. (2006). Maternal exposure to low levels of ambient air pollution
and preterm birth in Brisbane, Australia. BJOG, 113: 935-941. 089818
Hansen C; Neller A; Williams G; Simpson R. (2007). Low levels of ambient air pollution during pregnancy and
fetal growth among term neonates in Brisbane, Australia. Environ Res, 103: 383-389. 090703
Harrod KS; Jaramillo RJ; Berger JA; Gigliotti AP; Seilkop SK; Reed MD. (2005). Inhaled diesel engine
emissions reduce bacterial clearance and exacerbate lung disease to Pseudomonas aeruginosa infection in
vivo. Toxicol Sci, 83: 155-165. 088144
Hashimoto AH; Amanuma K; Hiyoshi K; Sugawara Y; Goto S; Yanagisawa R; Takano H; Masumura K"L Nohmi
T; Aoki Y. (2007). Mutations in the lungs of gpt delta transgenic mice following inhalation of diesel
exhaust. , 48: 682-693. 097261
Healey K; Smith EC; Wild CP; Routledge MN. (2006). The mutagenicity of urban particulate matter in an
enzyme free system is associated with the generation of reactive oxygen species. , 602: 1-6. 156532
Heinrich J; Hoelscher B; Frye C; Meyer I; Pitz M; Cyrys J; Wjst M; Neas L; Wichmann H"E. (2002). Improved
air quality in reunified Germany and decreases in respiratory symptoms. , 13: 394-401. 034825
Heinrich J; Slama R. (2007). Fine particles, a major threat to children. Int J Hyg Environ Health, 210: 617-622.
156534
Heiss G; Sharrett AR; Barnes R; Chambless LE; Szklo M; Alzola C. (1991). Carotid atherosclerosis measured by
B-mode ultrasound in populations: associations with cardiovascular risk factors in the ARIC study. Am J
Epidemiol, 134: 250-256. 156535
July 2009
7-144
DRAFT-DO NOT CITE OR QUOTE

-------
Hiramatsu Ki Azuma A; Kudoh Si Desaki Mi Takizawa Hi Sugawara I. (2003). Inhalation of diesel exhaust for
three months affects major cytokine expression and induces bronchus-associated lymphoid tissue
formation in murine lungs. , 29- 607-622. 155846
Hirose Ti Morito Ki Kizu Ri Toriba Ai Hayakawa Ki Ogawa Si Inoue Si Muramatsu Mi Masamune Y. (2001).
Estrogenic/Antiestrogenic Activities of Benzo [a] pyrene Monohydroxy Derivatives. Eisei Kagaku, 47-
552-558. 156548
Hoek Gi Brunekreef Bi Goldbohm Si Fischer Pi Van den Brandt PA. (2002). Association between mortality and
indicators of traffic-related air pollution in the Netherlands' a cohort study. Lancet, 360- 1203-1209.
042364
Hoffmann Bi Moebus Si Kroger Ki Stang Ai Mohlenkamp Si Dragano Ni Schmermund Ai Memmesheimer Mi
Erbel Ri Jockel K"H. (2009). Residential exposure to urban pollution, ankle "brachial index, and
peripheral arterial disease. Epidemiology, 20- 280-288. 190376
Hoffmann Bi Moebus Si Mohlenkamp Si Stang Ai Lehmann Ni Dragano Ni Schmermund Ai Memmesheimer Mi
Mann Ki Erbel Ri Jockel K"Hi Heinz Nixdorf Recall Study Investigative Group. (2007). Residential
exposure to traffic is associated with coronary atherosclerosis. Circulation, 116- 489-496. 091163
Hoffmann Bi Moebus Si Stang Ai Beck E"Mi Dragano Ni Mohlenkamp Si Schmermund Ai Memmesheimer Mi
Mann Ki Erbel Ri Jockel K"Hi Heinz Nixdorf RECALL Study Investigative Group. (2006). Residence
close to high traffic and prevalence of coronary heart disease. Eur Heart J, 27- 2696-2702. 091162
Hoffmann MHi Shi Hi Schmitz BLi Schmid FTi Lieberknecht Mi Schulze Ri Ludwig Bi Kroschel Ui Jahnke Ni
Haerer Wi Brambs HJi Aschoff AJ. (2005). Noninvasive coronary angiography with multislice computed
tomography. JAMA, 293: 2471-2478. 156556
Hollander Mi Hak AEi Koudstaal PJi Bots MLi Grobbee DEi Hofman Ai Witteman JCi Breteler MM. (2003).
Comparison between measures of atherosclerosis and risk of stroke: the Rotterdam Study. , 34: 2367"
2372.156562
Hornberg Ci Maciuleviciute Li Seemayer NH. (1996). Sister chromatid exchanges in rodent tracheal epithelium
exposed in vitro to environmental pollutants. Toxicol Lett, 88: 45"53. 087164
Hornberg Ci Maciuleviciute Li Seemayer NHi Kainka E. (1998). Induction of sister chromatid exchanges (SCE)
in human tracheal epithelial cells by the fractions PM-10 and PM-2.5 of airborne particulates. Toxicol
Lett, 96: 215-220. 155849
Hougaard KSi Jensen KAi Nordly Pi Taxvig Ci Vogel Ui Saber ATi Wallin H. (2008). Effects of prenatal
exposure to diesel exhaust particles on postnatal development, behavior, genotoxicity and inflammation
in mice. Part Fibre Toxicol, 5: 3. 156570
Huang JYi Liao JWi Liu YCi Lu SYi Chou CPi Chan WHi Chen SUi Ueng TH. (2008). Motorcycle exhaust
induces reproductive toxicity and testicular interleukin-6 in male rats. Toxicol Sci, 103: 137-148. 156574
Huynh Mi Woodruff TJi Parker JDi Schoendorf KC. (2006). Relationships between air pollution and preterm
birth in California. Paediatr Perinat Epidemiol, 20: 454-461. 091240
Hwang B-Fi Jaakkola JJKi Lee Y"Li Lin Y"Ci Guo Y-LL. (2006). Relation between air pollution and allergic
rhinitis in Taiwanese schoolchildren. Respir Res, T- 23. 088971
Hwang J-Si Nadziejko Ci Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles
(CAPs) in mice: HI Acute and chronic effects of CAPs on heart rate, heart-rate fluctuation, and body
temperature. Inhal Toxicol, 17: 199-207. 087957
IARC. (1989). Diesel and gasoline engine exhausts. Lyon, France: International Agency for Research on Cancer.
002958
Iba MMi Fung Ji Chung Li Zhao Ji Winnik Bi Buckley BTi Chen LCi Zelikoff JTi Kou YR. (2006). Differential
inducibility of rat pulmonary CYP1A1 by cigarette smoke and wood smoke. , 606: 1-11. 156582
Ichinose Ti Yajima Yi Nagashima Mi Takenoshita Si Nagamachi Yi Sagai M. (1997). Lung carcinogenesis and
formation of 8-hydroxy-deoxyguanosine in mice by diesel exhaust particles. Carcinogenesis, 18: 185-192.
053264
Ishihara Yi Kagawa J. (2003). Chronic diesel exhaust exposures of rats demonstrate concentration and time-
dependent effects on pulmonary inflammation. , 15: 473-492. 096404
Islam Ti Gauderman WJ; Berhane K; McConnell R; Avol Ei Peters JM; Gilliland FD. (2007). The relationship
between air pollution, lung function and asthma in adolescents. Thorax, 62: 957-963. 090697
Izawa Hi Watanabe Gi Taya Ki Sagai M. (2007). Inhibitory effects of foods and polyphenols on activation of aryl
hydrocarbon receptor induced by diesel exhaust particles. , 14: 149-156. 190387
Jacobsen NRi Mrller Pi Cohn CAi Loft Si Vogel Ui Wallin H. (2008). Diesel exhaust particles are mutagenic in
FEl-Muta™ Mouse lung epithelial cells. , 641: 54-57. 156597
Jalaludin Bi Mannes Ti Morgan Gi Lincoln Di Sheppeard Vi Corbett S. (2007). Impact of ambient air pollution
on gestational age is modified by season in Sydney, Australia. , 6: 16. 156601
July 2009
7-145
DRAFT-DO NOT CITE OR QUOTE

-------
Janes H; Dominici Fi Zeger SL. (2007). Trends in air pollution and mortality- an approach to the assessment of
unmeasured confounding. , 18: 416-23. 090927
Janssen NAH; Brunekreef B; van Vliet P; Aarts F; Maliefste E Harssema H; Fischer P. (2003). The relationship
between air pollution from heavy traffic and allergic sensitization, bronchial byperresponsiveness, and
respiratory symptoms in Dutch schoolchildren. , 111: 1512-1518. 133555
Jerrett M; Burnett RT; Ma R; Pope 3rd CA; Krewski D; Newbold KB; Thurston G; Shi Y; Finkelstein N; Calle
EE; Thun MJ. (2005). Spatial analysis of air pollution and mortality in Los Angeles. Epidemiology, 16:
727-36. 189405
Jerrett M; Buzzelli M; Burnett RT; DeLuca PF. (2005). Particulate air pollution, social confounders, and
mortality in small areas of an industrial city. Soc Sci Med, 60: 2845-2863. 087381
Karr C; Lumley T; Schreuder A; Davis R; Larson T; Ritz B; Kaufman J. (2007). Effects of subchronic and chronic
exposure to ambient air pollutants on infant bronchiolitis. Am J Epidemiol, 165: 553-560. 090719
Karr C; Rudra C; Miller K; Gould T; Larson T; Sathyanarayana S; Koenig J. (2009). Infant exposure to fine
particulate matter and traffic and risk of hospitalization for RSV bronchiolitis in a region with lower
ambient air pollution. Environ Res, 109: 321-327. 191946
Karthikeyan VJ; Lip GYH. (2007). Peripheral artery disease and hypertension: the relation between ante-
brachial index and mortality. J Hum Hypertens, 21: 762-765. 156626
Kato A; Kagawa J. (2003). Morphological effects in rat lungs exposed to urban roadside air. Inhal Toxicol, 15:
799-818. 089563
Khattar RS; Swales JD; Dore C; Senior R; Lahiri A. (2001). Effect of aging on the prognostic significance of
ambulatory systolic, diastolic, and pulse pressure in essential hypertension. , 104: 783-789. 155896
Khoury Z; Schwartz R; Gottlieb S; Chenzbraun A; Stern S; Keren A. (1997). Relation of coronary artery disease
to atherosclerotic disease in the aorta, carotid, and femoral arteries evaluated by ultrasound. Am J
Cardiol, 80: 1429-1433. 156636
Kim JJ; Smorodinsky S; Lipsett M; Singer BC; Hodgson AT; Ostro B. (2004). Traffic-related air pollution near
busy roads: the East Bay children's Respiratory Health Study. Am J Respir Crit Care Med, 170: 520-526.
087383
Kim OJ; Ha EH; Kim BM; Park HS; Jung WJ; Lee BE; Suh YJ; Kim YJ; Lee JT; Kim H; Hong YC. (2007). PM10
and pregnancy outcomes: a hospital-based cohort study of pregnant women in Seoul. J Occup Environ
Med, 49: 1394-1402. 156642
Kizu R; Okamura K; Toriba A; Mizokami A; Burnstein KL; KLinge CM; Hayakawa K. (2003). Antiandrogenic
activities of diesel exhaust particle extracts in PC3/AR human prostate carcinoma cells. , 76: 299-309.
096196
Knight EL; Curhan GC. (2003). Albuminuria: moving beyond traditional microalbuminuria cut-points. Current
Opinion Nephrol Hypertens, 12: 283-284. 179900
Knobel HH; Chen CJ; Liang KY. (1995). Sudden infant death syndrome in relation to weather and
optimetrically measured air pollution in Taiwan. , 96: 1106-1110. 155905
Krewski D. (2009). Evaluating the effects of ambient air pollution on life expectancy. N Engl J Med, 360: 413-
415. 190075
Krewski D; Burnett RT; Goldberg MS; Hoover K; Siemiatycki J; Jerrett M; Abrahamowicz M; White WH.
(2000). Reanalysis of the Harvard Six Cities study and the American Cancer Society study of particulate
air pollution and mortality: a special report of the Institute's Particle Epidemiology Reanalysis Project.
012281
Krewski D; Jerrett M; Burnett RT; Ma R; Hughes E; Shi Y; Turner MC; Pope AC III; Thurston G; Calle EE;
Thun MJ. (2009). Extended follow-up and spatial analysis of the American Cancer Society study linking
particulate air pollution and mortality. Health Effects Institute. Cambridge, MA. 140. 191193
Kulkarni N; Pierse N; Rushton L; Grigg J. (2006). Carbon in airway macrophages and lung function in children.
, 355: 21-30. 089257
Kunzli N; Bridevaux P-O; Liu S; Garcia-Esteban R; Schindler C; Gerbase M; Sunyer J; keidel D; Rochat T.
(2009). Traffic-related air pollution correlates with adult-onset asthma among never-smokers. Thorax, x:
x. 191949
Kunzli N; Jerrett MJ Mack WJ; Beckerman B; LaBree L; Gilliland F; Thomas D; Peters J; Hodis HN. (2005).
Ambient air pollution and atherosclerosis in Los Angeles. Environ Health Perspect, 113: 201-206. 087387
Kunzli N; Perez L; Lurmann F; Hricko A; Penfold B; McConnell R. (2008). An Attributable Risk Model for
Exposures Assumed to Cause Both Chronic Disease and its Exacerbations. , 19(2): 179-185. 129258
Lacasana M; Esplugues A; Ballester F. (2005). Exposure to ambient air pollution and prenatal and early
childhood health effects. , 20: 183-199. 155914
Laden F; Schwartz J; Speizer FE; Dockery DW. (2006). Reduction in fine particulate air pollution and mortality:
extended follow-up of the Harvard Six Cities study. Am J Respir Crit Care Med, 173: 667-672. 087605
July 2009
7-146
DRAFT-DO NOT CITE OR QUOTE

-------
Lee BE; Ha EH; Park HS; Kim YJ; Hong YC; Kim H; Lee JT. (2003). Exposure to air pollution during different
gestational phases contributes to risks of low birth weight. Hum Reprod, 18- 638-643. 043202
Leem J-H; Kaplan BM; Shim YK; Pohl HR; Gotway CA; Bullard SM; Rogers JF; Smith MM; Tylenda CA. (2006).
Exposures to air pollutants during pregnancy and preterm delivery. Environ Health Perspect, 114- 905"
910. 089828
Lemos M; Mohallen SJ Macchione MJ Dolhnikoff M; Assunaao J; Godleski J; Saldiva P. (2006). Chronic exposure
to urban air Pollution induces structural alterations in murine pulmonary and coronary arteries. Inhal
Toxicol, 18: 247-253. 088594
Li Y-J; Kawada T; Matsumoto A; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2007). Airway inflammatory
responses to oxidative stress induced by low-dose diesel exhaust particle exposure differ between mouse
strains. , 33: 227-244. 155929
Lichtenfels AJFC; Gomes JB; Pieri PC; Miraglia SGEK; Hallak J; Saldiva PHN. (2007). Increased levels of air
pollution and a decrease in the human and mouse male-to-female ration in Sao Paulo, Brazil. Fertil
SterH, 87: 230-232. 097041
Lin C-M; Li C-Y; Mao I-F. (2004). Increased risks of term low "birth-weight infants in a petrochemical industrial
city with high air pollution levels. Arch Environ Occup Health, 59: 663-668. 089827
Lin CA; Pereira LAA; Nishioka DC; Conceicao GMS; Graga ALF; Saldiva PHN. (2004). Air pollution and
neonatal deaths in Sao Paulo, Brazil. Braz J Med Biol Res, 37: 765-770. 095787
Lipfert FW; Baty JD; Miller JP; Wyzga RE. (2006). PM2.5 constituents and related air quality variables as
predictors of survival in a cohort of U.S. military veterans. Inhal Toxicol, 18: 645-657. 088756
Lipfert FW; Wyzga RE; Baty JD; Miller JP. (2006). Traffic density as a surrogate measure of environmental
exposures in studies of air pollution health effects: long-term mortality in a cohort of US veterans. Atmos
Environ, 40: 154-169. 088218
Lipfert FW; Zhang J; Wyzga RE. (2000). Infant mortality and air pollution: a comprehensive analysis of US data
for 1990. J Air Waste Manag Assoc, 50: 1350-1366. 004103
Lippmann M; Gordon T; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles in
mice. IX. Integral assessment and human health implications of subchronic exposures of mice to CAPs.
Inhal Toxicol, 17: 255-261. 087452
Liu L-JS; Curjuric I; Keidel D; Heldstab J; Kunzli N; Bayer-Oglesby L; Ackermann-Liebrich U; Schindler C;
SAPALDIA team. (2007). Characterization of source-specific air pollution exposure for a large
population-based Swiss cohort (SAPALDIA). Environ Health Perspect, 115: 1638-1645. 093093
Liu S; Krewski D; Shi Y; Chen Y; Burnett R. (2007). Association between maternal exposure to ambient air
pollutants during pregnancy and fetal growth restriction. J Expo Sci Environ Epidemiol, 17: 426-432.
090429
Liu YQ; Keane M; Ensell M; Miller W', Kashon M; Ong TM; Mauderly J; Lawson D; Gautam M; Zielinska B;
Whitney K; Eberhardt J; Wallace W. (2005). In vitro genotoxicity of exhaust emissions of diesel and
gasoline engine vehicles operated on a unified driving cycle. J Environ Monit, T- 60-66. 192098
Loomis D; Castillejos M; Gold DR; McDonnell W; Borja-Aburto VH. (1999). Air pollution and infant mortality in
Mexico City. , 10: 118-123. 087288
Lopes FD; Pinto TS; Arantes-Costa FM; Moriya HT; Biselli PJ; Ferraz LF; Lichtenfels AJ; Saldiva PH; Mauad
T; Martins MA. (2009). Exposure to ambient levels of particles emitted by traffic worsens emphysema in
mice. Environ Res, 109: 544-551. 190430
Lund AK; Knuckles TL; Obot Akata C; Shohet R; McDonald JD; Gigliotti A; Seagrave JC; Campen MJ. (2007).
Gasoline exhaust emissions induce vascular remodeling pathways involved in atherosclerosis. , 95: 485"
94. 125741
Maciejczyk P; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in
mice: VIII source-related daily variations in in vitro responses to CAPs. Inhal Toxicol, 17: 243-253.
087456
Mackman N> Tilley RE; Key NS. (2007). Role of the extrinsic pathway of blood coagulation in hemostasis and
thrombosis. Arterioscler Thromb Vase Biol, 27: 1687-1693. 156723
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005). Outdoor
air pollution and stroke in Sheffield, United Kingdom: a small-area level geographical study. , 36: 239"
243. 088683
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005). Outdoor
air pollution, mortality, and hospital admissions from coronary heart disease in Sheffield, UK: a small-
area level ecological study. Eur Heart J, 26: 2543-2549. 090769
Maisonet M; Bush TJ; Correa A; Jaakkola JJK. (2001). Relation between ambient air pollution and low birth
weight in the northeastern United States. Environ Health Perspect, 1093: 351-356. 016624
July 2009
7-147
DRAFT-DO NOT CITE OR QUOTE

-------
Maisonet M> Correa A; Misra D; Jaakkola JJ. (2004). A review of the literature on the effects of ambient air
pollution on fetal growth. Environ Res, 95- 106-115. 156725
Mannes T> Jalaludin B; Morgan Gi Lincoln D; Sheppeard V; Corbett S. (2005). Impact of ambient air pollution
on birth weight in Sydney, Australia. Occup Environ Med, 62: 524-530. 087895
Matsumoto A; Hiramatsu Ki Li Yi Azuma A; Kudoh Si Takizawa Hi Sugawara I. (2006). Repeated exposure to
low-dose diesel exhaust after allergen challenge exaggerates asthmatic responses in mice. , 121: 227-235.
098017
Mauad Ti Rivero DHi de Oliveira RCi Lichtenfels AJi Guimaraes ETi de Andre PAi Kasahara Dli Bueno HMi
Saldiva PH. (2008). Chronic exposure to ambient levels of urban particles affects mouse lung
development. Am J Respir Crit Care Med, 178: 721-728. 156743
McConnell Ri Berhane Ki Gilliland Fi London SJi Vora Hi Avol Ei Gauderman WJi Margolis HGi Lurmann Fi
Thomas DCi Peters JM. (1999). Air pollution and bronchitic symptoms in southern California children
with asthma. Environ Health Perspect, 107: 757-760. 007028
McConnell Ri Berhane Ki Gilliland Fi Molitor Ji Thomas Di Lurmann Fi Avol Ei Gauderman WJi Peters JM.
(2003). Prospective study of air pollution and bronchitic symptoms in children with asthma. Am J Respir
Crit Care Med, 168: 790-797. 049490
McDonald JDi Reed MDi Campen MJi Barrett EGi Seagrave Ji Mauderly JL. (2007). Health effects of inhaled
gasoline engine emissions. Inhal Toxicol, 19 Suppl l: 107-116. 156746
McDonnell WFi Nishino-Ishikawa Ni Petersen FFi Chen LHi Abbey DE. (2000). Relationships of mortality with
the fine and coarse fractions of long-term ambient PM10 concentrations in nonsmokers. J Expo Sci
Environ Epidemiol, 10: 427-436. 010319
Medeiros Ai Gouveia N. (2005). Relacao entre baixo peso ao nascer e a poluicao do ar no Municipio de Sao Paulo
[Relationship between low birthweight and air pollution in the city of Sao Paulo, Brazil]. Rev Saude
Publica, 39: 65-72. 089824
Mehta RHi Anad Kumar TC. (1997). Declining semen quality in Bangaloreans: A preliminary report. Curr Sci,
72: 621-622. 157197
Menache MGi Miller FJi Raabe OG. (1995). Particle inhalability curves for humans and small laboratory
animals. Ann Occup Hyg, 39: 317-328. 006533
Meng YYi Wilhelm Mi Rull RPi English Pi Ritz B. (2007). Traffic and outdoor air pollution levels near
residences and poorly controlled asthma in adults. Ann Allergy Asthma Immunol, 98: 455-63. 093275
Miller KAi Siscovick DSi Sheppard Li Shepherd Ki Sullivan JHi Anderson GLi Kaufman JD. (2007). Long-term
exposure to air pollution and incidence of cardiovascular events in women. , 356: 447-458. 090130
Mogensen CE. (1984). Microalbuminuria predicts clinical proteinuria and early mortality in maturity-onset
diabetes. , 310: 356-360. 156769
Mohallem SVi de Araujo Lobo DJi Pesquero CRi Assuncao JVi de Andre PAi Saldiva PHi Dolhnikoff M. (2005).
Decreased fertility in mice exposed to environmental air pollution in the city of Sao Paulo. Environ Res,
98: 196-202. 088657
Mollet NRi Cademartiri F; de Feyter PJ. (2005). Non -invasive multislice CT coronary imaging. Heart, 91: 401-
407. 155988
Montauban van Swijndregt AD; De Lange EE; De Groot E; Ackerstaff RG. (1999). An in vivo evaluation of the
reproducibility of intima-media thickness measurements of the carotid artery segments using B-mode
ultrasound. Ultrasound Med Biol, 25: 323-330. 156777
Morgenstern V; Zutavern A; Cyrys J; Brockow I; Koletzko S; Kramer U; Behrendt H; Herbarth O; von Berg A;
Bauer CP; Wichmann HE; Heinrich J. (2008). Atopic diseases, allergic sensitization, and exposure to
traffic-related air pollution in children. Am J Respir Crit Care Med, 177: 1331-1337. 156782
Mortimer K; Neugebauer R; Lurmann F; Alcorn S; Balmes J; Tager I. (2008). Early-Lifetime exposure to air
pollution and allergic sensitizationin children with asthma. J Asthma, 45: 874-881. 187280
Naess O; Nafstad P; Aamodt G; Claussen B; Rosland P. (2007). Relation between concentration of air pollution
and cause-specific mortality: four-year exposures to nitrogen dioxide and particulate matter pollutants in
470 neighborhoods in Oslo, Norway. Am J Epidemiol, 165: 435-443. 090736
National Kidney Foundation. (2008). Chronic Kidney Disease: Evaluation, Classification, and Stratification. The
National Kidney Foundation . Kidney Disease Outcomes Quality Initiative. 156796
Newman AB; Sutton-Tyrrell K; Vogt MT; Kuller LH. (1993). Morbidity and mortality in hypertensive adults
with a low ankle/arm blood pressure index. JAMA, 270: 487-489. 156805
Nordling E; Berglind N; Melen E; Emenius G; Hallberg J; Nyberg F; Pershagen G; Svartengren M; Wickman M;
Bellander T. (2008). Traffic-related air pollution and childhood respiratory symptoms, function and
allergies. , 19: 401-8. 097998
July 2009
7-148
DRAFT-DO NOT CITE OR QUOTE

-------
O'Leary DH; Polak JF; Kronmal RA; Kittner SJ; Bond MG; Wolfson SK, Jr.; Bommer W> Price TR; Gardin JM;
Savage PJ. (1992). Distribution and correlates of sonographically detected carotid artery disease in the
Cardiovascular Health Study. The CHS Collaborative Research Group. , 23- 1752-1760. 156825
O'Leary DH; Polak JF; Kronmal RA; Manolio TA; Burke GL; Wolfson SK, Jr. (1999). Carotid-artery intima and
media thickness as a risk factor for myocardial infarction and stroke in older adults. Cardiovascular
Health Study Collaborative Research Group. , 340- 14-22. 156826
O'Neill MS; Diez-Roux AV; Auchincloss AH; Franklin TG; Jacobs Jnr DR; Astor BC; Dvonch JT; Kaufman J.
(2007). Airborne particulate matter exposure and urinary albumin excretion- The Multi-Ethnic Study of
Atherosclerosis. Occup Environ Med. 156006
Oei HH; Vhegenthart R; Hak AE; Iglesias del Sol A; Hofman A; Oudkerk M; Witteman JC. (2002). The
association between coronary calcification assessed by electron beam computed tomography and
measures of extracoronary atherosclerosis' the Rotterdam Coronary Calcification Study. J Am Coll
Cardiol, 39: 1745-1751. 156820
Oftedal B; Brunekreef B; Nystad W; Madsen C; Walker S"E; Nafstad P. (2008). Residential outdoor air pollution
and lung function in schoolchildren. , 19: 129-137. 093202
Oftedal B; Brunekreef B; Nystad W', Nafstad P. (2007). Residential outdoor air pollution and allergen
sensitization in schoolchildren in Oslo, Norway. , 37: 1632. 191948
Oh S-M; Chung K"H. (2006). Identification of mammalian cell genotoxins in respirable diesel exhaust particles
by bioassay-directed chemical analysis. Toxicol Lett, 161: 226-235. 088296
Ono N; Oshio S; Niwata Y; Yoshida S; Tsukue N; Sugawara I; Takano H; Takeda K. (2007). Prenatal exposure
to diesel exhaust impairs mouse spermatogenesis. Inhal Toxicol, 19: 275-281. 156007
Ozkaynak H; Thurston GD. (1987). Associations between 1980 US mortality rates and alternative measures of
airborne particle concentration. Risk Anal, T- 449-461. 072960
Palli D; Saieva C; Munnia A; Peluso M; Grechi D; Zanna I; Caini S; Decarli A; Sera F; Masala G. (2008). DNA
adducts and PM10 exposure in traffic-exposed workers and urban residents from the EPIC"Florence City
study. Sci Total Environ, 403: 105-112. 156837
Parker JD; Mendola P; Woodruff TJ. (2008). Preterm birth after the utah valley steel mill closure: a natural
experiment. , 19: 820-823. 156013
Parker JD; Woodruff TJ. (2008). Influences of study design and location on the relationship between particulate
matter air pollution and birthweight. Paediatr Perinat Epidemiol, 22: 214-227. 156846
Parker JD; Woodruff TJ; Basu R; Schoendorf KC. (2005). Air pollution and birth weight among term infants in
California. , 115: 121-128. 087462
Pedersen M; Vinzents P; Petersen JH; KLeinjans JC; Plas G; Kirsch-Volders M; Dostal M; Rossner P; Beskid O;
Sram RJ; Merlo DF; Knudsen LE. (2006). Cytogenetic effects in children and mothers exposed to air
pollution assessed by the frequency of micronuclei and fluorescence in situ hybridization (FISH): a family
pilot study in the Czech Republic. , 608: 112-120. 156848
Penard-Morand C; Charpin D; Raherison C; Kopferschmitt C; Caillaud D; Lavaud F; Annesi-Maesano I. (2005).
Long-term exposure to background air pollution related to respiratory and allergic health in
schoolchildren. Clin Exp Allergy, 35: 1279-1287. 087951
Penna MLF; Duchiade MP. (1991). Air pollution and infant mortality from pneumonia in the Rio de Janeiro
metropolitan area. Bull Pan Am Health Organ, 25: 47-54. 073325
Pereira LAA; Loomis D; Conceicao GMS; Braga ALF; Areas RM; Kishi HS; Singer JM; Bohm GM; Saldiva PHN.
(1998). Association between air pollution and intrauterine mortality in Sao Paulo, Brazil. Environ Health
Perspect, 106: 325-329. 007264
Peters A; Skorkovsky J; Kotesovec F; Brynda J; Spix C; Wichmann HE; Heinrich J. (2000). Associations
between mortality and air pollution in central Europe. Environ Health Perspect, 108: 283-287. 001756
Pierse N; Rushton L; Harris RS; Kuehni CE; Silverman M; Grigg J. (2006). Locally-generated particulate
pollution and respiratory symptoms in young children. Thorax, 61: 216-220. 088757
Pignoli P; Tremoli E; Poli A; Oreste P; Paoletti R. (1986). Intimal plus medial thickness of the arterial wall: a
direct measurement with ultrasound imaging. , 74: 1399-1406. 156026
Pinkerton KE; Zhou Y; Teague SV; Peake JL; Walther RC; Kennedy IM; Leppert VJ; Aust AE. (2004). Reduced
lung cell proliferation following short-term exposure to ultrafine soot and iron particles in neonatal rats:
key to impaired lung growth?. Inhal Toxicol, l: 73-81. 087465
Pires-Neto RC; Lichtenfels AJ; Soares SR; Macchione M; Saldiva PHN; Dolhnikoff M. (2006). Effects of Sao
Paulo air pollution on the upper airways of mice. , 101: 356-361. 096734
Poma A; Limongi T; Pisani C; Granato V; Picozzi P. (2006). Genotoxicity induced by fine urban air particulate
matter in the macrophages cell line RAW 264.7. Toxicol In Vitro, 20: 1023-1029. 096903
Pope CA 3rd; Ezzati M; Dockery DW. (2009). Fine-particulate air pollution and life expectancy in the United
States. N Engl J Med, 360: 376-386. 190107
July 2009
7-149
DRAFT-DO NOT CITE OR QUOTE

-------
Pope CA III. (1989). Respiratory disease associated with community air pollution and a steel mill, Utah Valley.
Am J Public Health, 79: 623-628. 044461
Pope CA IIL Burnett RT. (2007). Confounding in air pollution epidemiology: the broader context. , 18: 424-426.
090928
Pope CA IIL Burnett RT; Thun MJ; Calle EE; Krewski D; Ito K; Thurston GD. (2002). Lung cancer,
cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA, 287: 1132"
1141. 024689
Pope CA IIL Burnett RT; Thurston GD; Thun MJ; Calle EE; Krewski D; Godleski JJ. (2004). Cardiovascular
mortality and long-term exposure to particulate air pollution: epidemiological evidence of general
pathophysiological pathways of disease. , 109: 71-77. 055880
Pope CA IIL Thun MJ; Namboodiri MM; Dockery DW; Evans JS; Speizer FE; Heath CW Jr. (1995). Particulate
air pollution as a predictor of mortality in a prospective study of US adults. Am J Respir Crit Care Med,
151: 669-674. 045159
Puett RC; Schwartz J; Hart JE; Yanosky JD; Speizer FE; Suh H; Paciorek CJ; Neas LM; Laden F. (2008).
Chronic particulate exposure, mortality, and coronary heart disease in the nurses' health study. Am J
Epidemiol, 168: 1161-1168. 156891
Raizenne M; Neas LM; Damokosh AL Dockery DW; Spengler JD; Koutrakis P; Ware JH; Speizer FE. (1996).
Health effects of acid aerosols on North American children: pulmonary function. Environ Health
Perspect, 104: 506-514. 077268
Ramos C; Cisneros J; Gonzalez-Avila G; Becerril C; Ruiz V; Montano M. (2009). Increase of matrix
metalloproteinases in woodsmoke-induced lung emphysema in guinea pigs. Inhal Toxicol, 21: 119-132.
190116
Reed MD; Barrett EG; Campen MJ; Divine KK; Gigliotti AP; McDonald JD; Seagrave JC; Mauderly JL; Seilkop
SK; Swenberg JA. (2008). Health effects of subchronic inhalation exposure to gasoline engine exhaust.
Inhal Toxicol, 20: 1125-1143. 156903
Reed MD; Campen MJ; Gigliotti AP; Harrod KS; McDonald JD; Seagrave JC; Mauderly JL; Seilkop SK. (2006).
Health effects of subchronic exposure to environmental levels of hardwood smoke. Inhal Toxicol, 18: 523"
539. 156043
Reed MD; Gigliotti AP; McDonald JD; Seagrave JC; Seilkop SK; Mauderly JL. (2004). Health effects of
subchronic exposure to environmental levels of diesel exhaust. Inhal Toxicol, 16: 177-193. 055625
Resnick HE; Lindsay RS; McDermott MM; Devereux RB; Jones KL; Fabsitz RR; Howard BV. (2004).
Relationship of high and low ankle brachial index to all-cause and cardiovascular disease mortality: the
Strong Heart Study. , 109: 733-739. 156048
Reymao MSF; Cury PM; Lichtenfels AJFC; Lemos M; Battlehner CN; Conceicao GMS; Capelozzi VL; Montes
GS; Junior MF; Martins MA; Bohm GM; Saldiva PHN. (1997). Urban air pollution enhances the
formation of urethane-induced lung tumors in mice. Environ Res, 74: 150-158. 084653
Rich DQ; Demissie K; Lu SE; Kamat L; Wartenberg D; Rhoads GG. (2009). Ambient air pollutant
concentrations during pregnancy and the risk of fetal growth restriction. J Epidemiol Community
Health, In Press: 1-9. 180122
Ritz B; Wilhelm M. (2008). Ambient air pollution and adverse birth outcomes: methodologic issues in an
emerging field. ,102: 182-190. 156914
Ritz B; Wilhelm M; Hoggatt KJ; Ghosh JK. (2007). Ambient air pollution and preterm birth in the environment
and pregnancy outcomes study at the University of California, Los Angeles. Am J Epidemiol, 166: 1045"
52.096146
Ritz B; Wilhelm M; Zhao Y. (2006). Air pollution and infant death in southern California, 1989-2000. , 118: 493-
502. 089819
Ritz B; Yu F; Chapa G; Fruin S. (2000). Effect of air pollution on preterm birth among children born in Southern
California between 1989 and 1993. , 11: 502-511. 012068
Ritz B; Yu F; Fruin S; Chapa G; Shaw GM; Harris JA. (2002). Ambient air pollution and risk of birth defects in
Southern California. Am J Epidemiol, 155: 17-25. 023227
Rogers JF; Dunlop AL. (2006). Air pollution and very low birth weight infants: a target population?. , 118: 156"
164. 091232
Rojas-Martinez R; Perez-Padilla R; Olaiz-Fernandez G; Mendoza-Alvarado L; Moreno-Macias H; Fortoul T;
McDonnell W; Loomis D; Romieu I. (2007). Lung function growth in children with long-term exposure to
air pollutants in Mexico City. , 176: 377-84. 188508
Roman HA; Walker KD; Walsh TL; Conner L; Richmond HM; Hubbell BJ; Kinney PL. (2008). Expert judgment
assessment of the mortality impact of changes in ambient fine particulate matter in the U.S. Environ Sci
Technol, 42: 2268-2274. 156921
July 2009
7-150
DRAFT-DO NOT CITE OR QUOTE

-------
Romieu L Garcia-Esteban R; Sunyer J; Rios C> Alcaraz-Zubeldia Mi Velasco SRi Holguin F. (2008). The effect of
supplementation with omega-3 polyunsaturated fatty acids on markers of oxidative stress in elderly
exposed to PM(2.5). Environ Health Perspect, 116: 1237-1242. 156922
Romieu Ii Ramirez "Aguilar Mi Moreno-Macias Hi Barraza-Villarreal Ai Miller Pi Hernandez-Cadena Li
Carbajal"Arroyo LAi Hernandez-Avila M. (2004). Infant mortality and air pollution: modifying effect by
social class. J Occup Environ Hyg, 46: 1210-1216. 093074
Rosenlund Mi Bellander Ti Nordquist Ti Alfredsson L. (2009). Traffic-generated air pollution and myocardial
infarction. Epidemiology, 20: 265-71. 190309
Rosenlund Mi Berglind Ni Pershagen Gi Hallqvist Ji Jonson Ti Bellander T. (2006). Long-term exposure to
urban air pollution and myocardial infarction. , 17: 383-390. 089796
Ross R. (1999). Atherosclerosis--an inflammatory disease. , 340: 115-126. 156926
Rubes Ji Selevan SGi Evenson DP; Zudova D; Vozdova M; Zudova Z; Robbins WA; Perreault SD. (2005).
Episodic air pollution is associated with increased DNA fragmentation in human sperm without other
changes in semen quality. Hum Reprod, 20: 2776-2783. 078091
Ruggenenti P; Remuzzi G. (2006). Time to abandon microalbuminuria?. Kidney Int, 70' 1214-1222. 156933
Roosli M; Braun-Fahrlander C; Kunzli N; Oglesby L; Theis G; Camenzind M; Mathys P; Staehelin J. (2000).
Spatial variability of different fractions of particulate matter within an urban environment and between
urban and rural sites. J Air Waste Manag Assoc, 50: 1115-1124. 010296
Roosli M; Kunzli N> Braun-Fahrlander C> Egger M. (2005). Years of life lost attributable to air pollution in
Switzerland: dynamic exposure-response model. Int J Epidemiol, 34: 1029-1035. 156923
Roosli M; Theis G; Kunzli N; Staehelin J; Mathys P; Oglesby L; Camenzind M; Braun-Fahrlander C. (2001).
Temporal and spatial variation of the chemical composition of PM10 at urban and rural sites in the Basel
area, Switzerland. Atmos Environ, 35: 3701-3713. 108738
Sagiv SE Mendola P; Loomis D; Herring AH; Neas LM; Savitz DA; Poole C. (2005). A time-series analysis of air
pollution and preterm birth in Pennsylvania, 1997-2001. Environ Health Perspect, 113: 602-606. 087468
Salam MT Millstein J; Li Y-F; Lurmann FW; Margolis HG; Gilliland FD. (2005). Birth outcomes and prenatal
exposure to ozone, carbon monoxide, and particulate matter: results from the Children's Health Study.
Environ Health Perspect, 113: 1638-1644. 087885
Salonen JT Salonen R. (1991). Ultrasonographically assessed carotid morphology and the risk of coronary heart
disease. Arterioscler Thromb Vase Biol, 11: 1245-1249. 156938
Sato H; Suzuki Kazuo T; Sone H; Yamano Y; Kagawa J; Aoki Y. (2003). DNA-adduct formation in lungs, nasal
mucosa, and livers of rats exposed to urban roadside air in Kawasaki City, Japan. , 93: 36"44. 096615
Schatz M; Zeiger RS; Hoffman CP. (1990). Intrauterine growth is related to gestational pulmonary function in
pregnant asthmatic women. Kaiser-Permanente Asthma and Pregnancy Study Group. , 98: 389-392.
156073
Schikowski T; Sugiri D; Ranft U; Gehring U; Heinrich J; Wichmann HE; Kramer U. (2005). Long-term air
pollution exposure and living close to busy roads are associated with COPD in women. Respir Res, 22:
152-161. 088637
Schindler C; Keidel D; Gerbase MW; Zemp E; Bettschart R; Brandli O; Brutsche MH; Burdet L; Karrer W;
Knopfli B; Pons M; Rapp R; Bayer-Oglesby L; Kunzli N> Schwartz J; Liu L-JS; Ackermann-Liebrich U;
Rochat T; the SAPALDIA Team. (2009). Improvements in PM10 exposure and reduced rates of
respiratory symptoms in a cohort of swiss adults (SAPALDIA). , 179: 579-587. 191950
Schwartz J; Coull B; Laden F; Ryan L. (2008). The effect of dose and timing of dose on the association between
airborne particles and survival. Environ Health Perspect, 116: 64"69. 156963
Seagrave J; McDonald JD; Reed MD; Seilkop SK; Mauderly JL. (2005). Responses to subchronic inhalation of
low concentrations of diesel exhaust and hardwood smoke measured in rat bronchoalveolar lavage fluid.
Inhal Toxicol, 17: 657-670. 088000
Selevan SG; Borkovec L; Slott VL; Zudova Z; Rubes J; Evenson DP; Perreault SD. (2000). Semen quality and
reproductive health of young Czech men exposed to seasonal air pollution. Environ Health Perspect, 108:
887-894. 012578
Sevastyanova 0; Binkova B; Topinka J; Sram RJ; Kalina I; Popov T Novakova Z; Farmer PB. (2007). In vitro
genotoxicity of PAH mixtures and organic extract from urban air particles part II: human cell lines. , 620:
123-134. 156969
Sharma AK; Jensen KA; Rank J; White PA; Lundstedt S; Gagne R; Jacobsen NR; Kristiansen J; Vogel U; Wallin
H. (2007). Genotoxicity, inflammation and physico-chemical properties of fine particle samples from an
incineration energy plant and urban air. , 633: 95-111. 156975
Shaw LJ; Raggi P; Schisterman E; Berman DS; Callister TQ. (2003). Prognostic value of cardiac risk factors and
coronary artery calcium screening for all-cause mortality. Radiology, 228: 826-833. 156083
July 2009
7-151
DRAFT-DO NOT CITE OR QUOTE

-------
Shemesh J; Tenenbaum A; Fisman EZ; Apter Si Rath Si Rozenman Ji Itzchak Yi Motro M. (1996). Absence of
coronary calcification on double "helical CT scans' predictor of angiographic ally normal coronary arteries
in elderly women?. Radiology, 199- 665-668. 156085
Shinkura Ri Fijuyama Ci Akiba S. (1999). Relationship between ambient sulfur dioxide levels and neonatal
mortality near the Mt. Sakurajima volcano in Japan. , 9- 344-349. 156978
Silva PJi Erupe ME; Price D; Elias J; Malloy QG; Li Q; Warren B; Cocker DR, 3rd. (2008). Trimethylamine as
precursor to secondary organic aerosol formation via nitrate radical reaction in the atmosphere. Environ
Sci Technol, 42: 4689-4696. 156981
Slama R; Darrow L; Parker J; Woodruff TJ; Strickland M> Nieuwenhuijsen Mi Glinianaia Si Hoggatt KJi
Kannan Si Hurley Fi Kalinka Ji Sram Ri Brauer Mi Wilhelm Mi Heinrich Ji Ritz B. (2008). Meeting
report: atmospheric pollution and human reproduction. Environ Health Perspect, 116: 791-798. 156985
Slama Ri Morgenstern Vi Cyrys Ji Zutavern Ai Herbarth Oi Wichmann HEi Heinrich Ji LISA Study Group.
(2007).	Traffic-related atmospheric pollutants levels during pregnancy and offspring's term birth weight:
a study relying on a land-use regression exposure model. Environ Health Perspect, 115: 1283-1292.
093216
Smilde TJi Wollersheim Hi Van Langen Hi Stalenhoef AF. (1997). Reproducibility of ultrasonographic
measurements of different carotid and femoral artery segments in healthy subjects and in patients with
increased intima-media thickness. Clin Sci (Lond), 93: 317-324. 156988
Sokol RZi Kraft Pi Fowler IMi Mamet Ri Kim E> Berhane KT. (2006). Exposure to environmental ozone alters
semen quality. Environ Health Perspect, 114: 360-5. 098539
Solomon Pi Baumann Ki Edgerton E> Tanner Ri Eatough Di Modey W> Marin Hi Savoie D> Natarajan Si Meyer
MB. (2003). Comparison of integrated samplers for mass and composition during the 1999 Atlanta
supersites project. J Geophys Res, 108: 8423. 156994
Somers CM; McCarry BE; Malek F; Quinn JS. (2004). Reduction of particulate air pollution lowers the risk of
heritable mutations in mice. , 304: 1008-1010. 078098
Somers CM; Yauk CL; White PA; Parfett CLJ; Quinn JS. (2002). Air pollution induces heritable DNA
mutations. Proc Natl Acad Sci USA, 99: 15904-15907. 078100
Song CL; Zhou YC; Huang RJ; Wang YQ; Huang QF; Lu G; Liu KM. (2007). Influence of ethanol-diesel blended
fuels on diesel exhaust emissions and mutagenic and genotoxic activities of particulate extracts. , 149:
355-63. 188903
Sorensen M; Schins RP; Hertel O; Loft S. (2005). Transition metals in personal samples of PM2.5 and oxidative
stress in human volunteers. , 14: 1340-3. 188901
Sram R; Beskid O; Binkova B; Chvatalova I; Lnenickova Z; Milcova A; Solansky I; Tulupova E; Bavorova H;
Ocadlikova D. (2007). Chromosomal aberrations in environmentally exposed population in relation to
metabolic and DNA repair genes polymorphisms. , 620: 22"33. 188457
Sram RJ; Beskid O; Rossnerova A; Rossner P; Lnenickova Z; Milcova A; Solansky I; Binkova B. (2007).
Environmental exposure to carcinogenic polycyclic aromatic hydrocarbons-"the interpretation of
cytogenetic analysis by FISH. Toxicol Lett, 172: 12-20. 192084
Sram RJ; Binkova B; Dejmek J; Bobak M. (2005). Ambient air pollution and pregnancy outcomes: a review of
the literature. Environ Health Perspect, 113: 375-382. 087442
Stanojevic S; Wade A; Stocks J; Hankinson J; Coates AL; Pan H; Rosenthal M; Corey M; Lebecque P; Cole TJ.
(2008).	Reference ranges for spirometry across all ages: a new approach. Am J Respir Crit Care Med, 177:
253-260. 157007
Stacker R; Keaney JF Jr. (2004). Role of oxidative modifications in atherosclerosis. Physiol Rev, 84: 1381-1478.
157013
Strom KA; Garg BD; Johnson JT; D'Arcy JB; Smiler KL. (1990). Inhaled particle retention in rats receiving low
exposures of diesel exhaust. J Toxicol Environ Health, 29: 377-398. 157020
Sugamata M; Ihara T; Takano H; Oshio S; Takeda K. (2006). Maternal diesel exhaust exposure damages
newborn murine brains. Eisei Kagaku, 52: 82"84. 097166
Sugiri D; Ranft U; Schikowski T; Kramer U. (2006). The influence of large-scale airborne particle decline and
traffic-related exposure on children's lung function. Environ Health Perspect, 114: 282-288. 088760
Suh YJ; Ha EH; Park H; Kim YJ; Kim H; Hong YC. (2008). GSTM1 polymorphism along with PM10 exposure
contributes to the risk of preterm delivery. , 656: 62"67. 192077
Sun Q; Wang A; Jin X; Natanzon A; Duquaine D; Brook RD; Aguinaldo J-GS; Fayad ZA; Fuster V; Lippmann M;
Chen Lung C; Rajagopalan S. (2005). Long-term air pollution exposure and acceleration of
atherosclerosis and vascular inflammation in an animal model. , 294: 3003-3010. 186814
July 2009
7-152
DRAFT-DO NOT CITE OR QUOTE

-------
Sun Q; Yue P; Deiuliis JA; Lumeng CN; Kampfrath T; Mikolaj MB; Cai Y; Ostrowski MC; Lu B; Parthasarathy
S> Brook RD; Moffatt-Bruce SD; Chen LC; Rajagopalan S. (2009). Ambient air pollution exaggerates
adipose inflammation and insulin resistance in a mouse model of diet-induced obesity. Circulation, 119:
538-546. 190487
Sun Q; Yue P; Kirk RI; Wang A; Moatti D; Jin X; Lu B; Schecter AD; Lippmann M; Gordon T; Chen LC;
Rajagopalan S. (2008). Ambient air particulate matter exposure and tissue factor expression in
atherosclerosis. Inhal Toxicol, 20- 127-137. 157033
Sun Q; Yue P; Ying Z; Cardounel AJ; Brook RD; Devlin R; Hwang JS; Zweier JL; Chen LC; Rajagopalan S.
(2008). Air Pollution Exposure Potentiates Hypertension Through Reactive Oxygen Species-Mediated
Activation of Rho/ROCK. Arterioscler Thromb Vase Biol, 28- 1760-1766. 157032
Suwa T; Hogg JC; Quinlan KB; Ohgami A; Vincent R; Van Eeden SF. (2002). Particulate air pollution induces
progression of atherosclerosis. J Am Coll Cardiol, 39: 935-942. 028588
Tarantini L; Bonzini M; Apostoli P; Pegoraro V; Bollati V; Marinelli B; Cantone L; Rizzo G; Hou L; Schwartz J;
Bertazzi PA; Baccarelh A. (2009). Effects of particulate matter on genomic DNA methylation content and
iNOS promoter methylation. Environ Health Perspect, 117: 217-222. 192010
Tesfaigzi Y; McDonald JD; Reed MD; Singh SP; De Sanctis GT; Eynott PR; Hahn FF; Campen MJ; Mauderly
JL. (2005). Low-level subchronic exposure to wood smoke exacerbates inflammatory responses in allergic
rats. , 88: 505-513. 156116
Tesfaigzi Y; Singh SP; Foster JE; Kubatko J; Barr EB; Fine PM; McDonald JD; Hahn FF; Mauderly JL. (2002).
Health effects of subchronic exposure to low levels of wood smoke in rats. Toxicol Sci, 65: 115-125. 025575
Thurlbeck WM. (1982). Postnatal human lung growth. Thorax, 37: 564-571. 093260
Tokiwa H; Sera N; Nakanishi Y. (2005). Involvement of alveolar macrophages in the formation of 8-
oxodeoxyguanosine associated with exogenous particles in human lungs. Inhal Toxicol, 17: 577. 191952
Tong S; Colditz P. (2004). Air pollution and sudden infant death syndrome: a literature review. Paediatr Perinat
Epidemiol, 18: 327-335. 087883
Tovalin H; Valverde M; Morandi MT; Blanco S; Whitehead L; Rojas E. (2006). DNA damage in outdoor workers
occupationally exposed to environmental air pollutants. Occup Environ Med, 63: 230-236. 091322
Tozuka Y; Watanabe N; Ohsawa M; Toriba A; Kizu R; Hayakawa K. (2004). Transfer of polycyclic aromatic
hydrocarbons to fetuses and breast milk of rats exposed to diesel exhaust. Eisei Kagaku, 50: 497-502.
090864
Tsai S-S; Chen C"C; Hsieh H-J; Chang C-C; Yang C-Y. (2006). Air pollution and postneonatal mortality in a
tropical city: Kaohsiung, Taiwan. Inhal Toxicol, 18: 185-189. 090709
Tsukue N; Tsubone H; Suzuki AK. (2002). Diesel exhaust affects the abnormal delivery in pregnant mice and
the growth of their young. Inhal Toxicol, 14: 635-651. 030593
Tsukue N; Yoshida S; Sugawara I; Taked K. (2004). Effect of diesel exhaust on development of fetal
reproductive function in ICR female mice. Eisei Kagaku, 50: 174-80. 096643
U.S. EPA. (1996). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA/600/P-95/001aF-cF. 079380
U.S. EPA. (2002). Health assessment document for diesel engine exhaust. 042866
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA/600/P"99/002aF"bF. 056905
U.S. EPA. (2005). Guidelines for carcinogen risk assessment. U.S. Environmental Protection Agency. Just.
086237
U.S. EPA. (2006). Provisional Assessment of Recent Studies on Health Effects of Particulate Matter Exposure.
U.S. Environmental Protection AGency. Research Triangle Park, NC. 157071
Valentine R; Himmelstein MW. (2001). Overview of the acute, subchronic, reproductive, developmental and
genetic toxicology of "beta"-chloroprene. Chem Biol Interact, 135: 81-100. 019011
van der Meer IM; Bots ML; Hofman A; del Sol AI; van der Kuip DA; Witteman JC. (2004). Predictive value of
noninvasive measures of atherosclerosis for incident myocardial infarction: the Rotterdam Study.
Circulation, 109: 1089-1094. 156129
Veras MM; Damaceno-Rodrigues NR; Caldini EG; Maciel Ribeiro AA; Mayhew TM; Saldiva PH; Dolhnikoff M.
(2008).	Particulate urban air pollution affects the functional morphology of mouse placenta. Biol Reprod,
79: 578-584. 190493
Veras MM; Damaceno-Rodrigues NR; Guimaraes Silva RM; Scoriza JN; Saldiva PH; Caldini EG; Dolhnikoff M.
(2009).	Chronic exposure to fine particulate matter emitted by traffic affects reproductive and fetal
outcomes in mice. Environ Res, 109: 536-543. 190496
Villeneuve PJ; Goldberg MS; Krewski D; Burnett RT; Chen Y. (2002). Fine particulate air pollution and all-
cause mortality within the Harvard six-cities study: variations in risk by period of exposure. Ann
Epidemiol, 12: 568-576. 042576
July 2009
7-153
DRAFT-DO NOT CITE OR QUOTE

-------
Vineis P; Hoek G> Krzyzanowski M> Vigna-Taglianti Fi Veglia Fi Airoldi Li Autrup Hi Dunning Ai Garte Si
Hainaut Pi Malaveille Ci Matullo Gi Overvad Ki Raaschou-Nielsen Oi Clavel-Chapelon Fi Linseisen Ji
Boeing Hi Trichopoulou Ai Palli Di Peluso Mi Krogh Vi Tumino Ri Panico Si Bueno-De-Mesquita HBi
Peeters PHi Lund EEi Gonzalez CAi Martinez Ci Dorronsoro Mi Barricarte Ai Cirera Li Quiros JR.;
Berglund Gi Forsberg Bi Day NEi Key TJi Saracci Ri Kaaks Ri Riboli E. (2006). Air pollution and risk of
lung cancer in a prospective study in Europe. Int J Cancer, 119- 169-174. 192089
Vinzents PSi Moller Pi Sorensen Mi Knudsen LEi Herte LQi Jensen FPi Schibye Bi Loft S. (2005). Personal
exposure to ultrafine particles and oxidative DNA damage. Environ Health Perspect, 113- 1485-1490.
087482
Vogt MTi Cauley JAi Newman ABi Kuller LHi Hulley SB. (1993). Decreased ankle/arm blood pressure index
and mortality in elderly women. JAMA, 270- 465-469. 157100
Wallenborn JGi Evansky Pi Shannahan JHi Vallanat Bi Ledbetter AD; Schladweiler MC; Richards JHi
Gottipolu RRi Nyska A; Kodavanti UP. (2008). Subchronic inhalation of zinc sulfate induces cardiac
changes in healthy rats. Toxicol Appl Pharmacol, 232: 69"77. 191171
Walsh CR; Cupples LA; Levy D; Kiel DP; Hannan MJ Wilson PW; O'Donnell CJ. (2002). Abdominal aortic
calcific deposits are associated with increased risk for congestive heart failure: the Framingham Heart
Study. Am Heart J, 144: 733-739. 157103
Watanabe N. (2005). Decreased number of sperms and Sertoli cells in mature rats exposed to diesel exhaust as
fetuses. Toxicol Lett, 155: 51-58. 087985
Wayne LG; Chambers LA. (1968). Biological effects of urban air pollution: V a study of effects of Los Angeles
atmosphere on laboratory rodents. Arch Environ Occup Health, 16: 871-885. 038537
Weisenberger DJ; Campan M; Long TL Kim M; Woods C; Fiala E; Ehrlich M; Laird PW. (2005). Analysis of
repetitive element DNA methylation by MethyLight. Nucleic Acids Res, 33: 6823-36. 192101
Weitz JL Byrne J; Clagett GP; Farkouh ME; Porter JM; Sackett DL; Strandness DE Jr; Taylor LM. (1996).
Diagnosis and treatment of chronic arterial insufficiency of the lower extremities: a critical review.
Circulation, 94: 3026-3049. 156150
Wendelhag I; Gustavsson T; Suurkula M; Berglund G; Wikstrand J. (1991). Ultrasound measurement of wall
thickness in the carotid artery: fundamental principles and description of a computerized analysing
system. Clinical Physiol, 11: 565-577. 157135
Wendelhag I; Wiklund O; Wikstrand J. (1993). Atherosclerotic changes in the femoral and carotid arteries in
familial hypercholesterolemia. Ultrasonographic assessment of intima-media thickness and plaque
occurrence. Arterioscler Thromb Vase Biol, 13: 1404-1411. 157136
Wheeler AJ; Villeneuve P; Smith-Doiron M; Mahmud M; Dales R; Brook JR. (2006). Children's exposure to
ambient air pollution: the application of different exposure assessment methodologies. Epidemiology,
17(6): S33-S34. 103905
Wilhelm M; Ritz B. (2005). Local variations in CO and particulate air pollution and adverse birth outcomes in
Los Angeles County, California, USA. Environ Health Perspect, 113: 1212-1221. 088668
Willekes C; Brands PJ; Willigers JM; Hoeks AP; Reneman RS. (1999). Assessment of local differences in intima-
media thickness in the human common carotid artery. J Vase Res, 36: 222-228. 157147
Williams L; Ulrich C; Larson T; Wener M; Wood B; Campbell P; Potter J; McTiernan A; De Roos A. (2009).
Proximity to Traffic, Inflammation, and Immune Function among Women in the Seattle, Washington,
Area. Environ Health Perspect, 117: 373. 191945
Wilson PW; Kauppila LI; O'Donnell CJ; Kiel DP; Hannan M; Polak JM; Cupples LA. (2001). Abdominal aortic
calcific deposits are an important predictor of vascular morbidity and mortality. Circulation, 103: 1529"
1534. 156159
Wilson WE; Mage DT; Grant LD. (2000). Estimating separately personal exposure to ambient and nonambient
particulate matter for epidemiology and risk assessment: why and how. J Air Waste Manag Assoc, 50:
1167-1183. 010288
Witteman JC; Kok FJ; van Saase JL; Valkenburg HA. (1986). Aortic calcification as a predictor of
cardiovascular mortality. Lancet, 2' 1120-1122. 156161
Woodruff TJ; Darrow LA; Parker JD. (2008). Air pollution and postneonatal infant mortality in the United
States, 1999-2002. Environ Health Perspect, 116: 110-5. 098386
Woodruff TJ; Grillo J; Schoendorf KC. (1997). The relationship between selected causes of postneonatal infant
mortality and particulate air pollution in the United States. Environ Health Perspect, 105: 608-612.
084271
Woodruff TJ; Parker JD; Schoendorf KC. (2006). Fine particulate matter (PM25) air pollution and selected
causes of postneonatal infant mortality in California. Environ Health Perspect, 114: 785-790. 088758
July 2009
7-154
DRAFT-DO NOT CITE OR QUOTE

-------
Xu J; Lee ET; Devereux RB; Umans JG; Bella JN; Shara NM! Yeh J; Fabsitz RR; Howard BY. (2008). A
longitudinal study of risk factors for incident albuminuria in diabetic American Indians' the Strong
Heart Study. Am J Kidney Dis, 51: 415-424. 157157
Yang AS; Estecio MR; Doshi K; Kondo Y; Tajara EH; Issa JP. (2004). A simple method for estimating global
DNA methylation using bisulfite PCR of repetitive DNA elements. Nucleic Acids Res, 32: e38. 192102
Yang C-Y; Hsieh H-J; Tsai S_S; Wu T_N; Chiu H_F. (2006). Correlation between air pollution and postneonatal
mortality in a subtropical city: Taipei, Taiwan. J Toxicol Environ Health A, 69: 2033-2040. 090760
Yang C-Y; Tseng Y"T; Chang C"C. (2003). Effects of air pollution on birthweight among children born between
1995 and 1997 in Kaohsiung, Taiwan. J Toxicol Environ Health A, 66: 807-816. 087886
Yatera K; Hsieh J; Hogg James C; Tranfield E; Suzuki H; Shih C"H; Behzad Ali R; Vincent R; van Eeden
Stephan F. (2008). Particulate matter air pollution exposure promotes recruitment of monocytes into
atherosclerotic plaques. Am J Physiol Heart Circ Physiol, 294: H944-H953. 157162
Yauk C; Polyzos A; Rowan-Carroll A; Somers CM; Godschalk RW; Van Schooten FJ; Berndt ML; Pogribny IP;
Koturbash I; Williams A; Douglas GR; Kovalchuk O. (2008). Germ-line mutations, DNA damage, and
global hypermethylation in mice exposed to pa. Presented at . 157164
Yauk CL; Quinn JS. (1996). Multilocus DNA fingerprinting reveals high rate of heritable genetic mutation in
herring gulls nesting in an industrialized urban site. Presented at . 157163
Ying Z; Kampfrath T; Thurston G; Farrar B; Lippmann M; Wang A; Sun Q; Chen LC; Rajagopalan S. (2009).
Ambient particulates alter vascular function through induction of reactive oxygen and nitrogen species.
Toxicol Sci, l: 1-36. 190111
Yokota S; Mizuo K; Moriya N; Oshio S; Sugawara I; Takeda K. (2009). Effect of prenatal exposure to diesel
exhaust on dopaminergic system in mice. Neurosci Lett, 449: 38-41. 190518
Yoshida S; Ono N; Tsukue N; Oshio S; Umeda T; Takano H; Takeda K. (2006). In utero exposure to diesel
exhaust increased accessory reproductive gland weight and serum testosterone concentration in male
mice. , 13: 139-147. 097015
Yoshida S; Takedab K. (2004). The effects of diesel exhaust on murine male reproductive function. Eisei
Kagaku, 50: 210-4. 097760
Yoshida S; Yoshida M; Sugawara I; Takeda K. (2006). Mice strain differences in effects of fetal exposure to
diesel exhaust gas on male gonadal differentiation. , 13: 117-123. 156170
Zanobetti A; Bind MAC; Schwartz J. (2008). Particulate air pollution and survival in a COPD cohort. Environ
Health Perspect, 7: 48. 156177
Zanobetti A; Schwartz J. (2007). Particulate air pollution, progression, and survival after myocardial infarction.
Environ Health Perspect, 115: 769-775. 091247
Zeger S; Dominici F; McDermott A; Samet J. (2008). Mortality in the Medicare population and chronic exposure
to fine particulate air pollution in urban centers (2000-2005). Environ Health Perspect, 116: 1614.
191951
Zeger S; McDermott A; Dominici F; Samet J. (2007). Mortality in the medicare population and chronic exposure
to fine particulate air pollution. Environ Health Perspect, 116: 1614-1619. 157176
Zeman KL; Bennett WD. (2006). Growth of the small airways and alveoli from childhood to the adult lung
measured by aerosol-derived airway morphometry. J Appl Physiol, 100: 965-971. 157178
Zhang Z; Che W; Liang Y; Wu M; Li N; Shu Y; Liu F; Wu D. (2007). Comparison of cytotoxicity and genotoxicity
induced by the extracts of methanol and gasoline engine exhausts. Toxicol In Vitro, 21: 1058-1065.
157186
July 2009
7-155
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Chapter 8. Susceptible Subpopulations
8.1. Potentially Susceptible Subpopulations
Interindividual variation in human responses to air pollutants indicates that some
subpopulations are at increased risk for the detrimental effects of ambient exposure to PM
(Kleeberger and Ohtsuka, 2005, 130489). The NAAQS are intended to provide an adequate
margin of safety for both general populations and sensitive subpopulations, or those
subgroups potentially at increased risk for health effects in response to ambient air
pollution (see Section l.l). To facilitate the identification of subpopulations at the greatest
risk for PM-related health effects, studies have evaluated factors that contribute to the
susceptibility and/or vulnerability of an individual to PM. These terms have sometimes
been used interchangeably in the literature, and in other cases have been defined to
represent two different categories that could contribute to a subpopulation experiencing
increased risk to PM-related health effects, resulting in the lack of a clear and consistent
definition (see Table 8-1). Additionally, in some cases, "at-risk" has been used as a term
encompassing these concepts more generally.
In this ISA, the term 'susceptible' will be used to represent populations that have a
greater likelihood of experiencing health effects related to PM exposure. This increased
likelihood of response to PM can result from a multitude of factors, including genetic or
developmental factors, race, gender, age, lifestyle (e.g., smoking status and nutrition) or
preexisting disease states. Population-level susceptibility factors (e.g., socioeconomic status
[SES], which includes reduced access to health care and low educational attainment) are
also discussed within this section due to the fact their non-random distribution appears to
play a dominant role in influencing the pattern of health effects observed in response to PM
exposure (American Lung Association, 2001, 016626).
Table 8-1. Definitions of susceptible and vulnerable in the PM literature.

Definition
Reference
Susceptible: pred
Vulnerable: capal
isposed to develop a noninfectious disease
lie of being hurt: susceptible to injury or disease
Merriam-Webster (2009,192146)
Susceptible: greater likelihood of an adverse outcome given a specific exposure, in comparison with the general	American Lung Association (2001, 016626)
population. Includes both host and environmental factors (e.g., genetics, diet, physiologic state, age, gender,
social, economic, and geographic attributes).
Vulnerable: periods during an individual's life when they are more susceptible to environmental exposures.
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health
and Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in
the process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009	8-1	DRAFT - DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Definition
Reference
Susceptible: innate (e.g., genetic or developmental) or acquired (e.g., age, disease or smoking or smoking) factors
U.S. EPA. (2008,157072)
that make individuals more likely to experience effects with exposure to PM.

Vulnerable: PM-related effects due to factors including socioeconomic status (e.g., reduced access to health

care) or particularly elevated exposure levels.

Susceptible: greater or lesser biological response to exposure.
U.S. EPA (2009,192149)
Vulnerable: more or less exposed.

Vulnerable: to be susceptible to harm or neglect, that is, acts of commission or omission on the part of others
Adav, LA. (2001,192150)
that can wound.

Susceptible:may be those who are significantly more liable than the general population to be affected by a
U.S. EPA (2003,192145)
stressor due to life stage (e.g., children, the elderly, or pregnant women), genetic polymorphisms (e.g., the small

but significant percentage of the population who have genetic susceptibilities), prior immune reactions (e.g.,

individuals who have been "sensitized" to a particular chemical), disease state (e.g., asthmatics), or prior damage

to cells or systems (e.g., individuals with damaged ear structures due to prior exposure to toluene, making them

more sensitive to damage by high noise levels).

Vulnerable: differential exposure and differential preparedness (e.g., immunization).

Susceptible: intrinsic (e.g., age, gender, pre existing disease (e.g., asthma) and genetics) and extrinsic (previous
Kleeberger and Ohtsuka (2005,130489)
exposure and nutritional status) factors.

Susceptible:characteristics that contribute to increased risk of PM-related health effects (e.g., genetics, pre-
Pope and Dockerv (2006,156881)
existing disease, age, gender, race, socioeconomic status, healthcare availability, educational attainment, and

housing characteristics).

To examine whether PM differentially affects certain subpopulations, epidemiologic
studies conduct stratified analyses to identify the presence or absence of effect modification.
A thorough evaluation of potential effect modifiers may help identify subpopulations that
are more susceptible to PM. These analyses require the proper identification of confounders
and their subsequent adjustment in statistical models, which helps separate a spurious
association from a true causal association. Although the design of toxicological and
controlled human exposure studies do not allow for an extensive examination of effect
modifiers, the use of animal models of disease and the study of individuals with underlying
disease or genetic polymorphisms do allow for comparisons between subgroups. Therefore,
the results from these studies, combined with those results obtained through stratified
analyses in epidemiologic studies, contribute to the overall weight of evidence for the
increased susceptibility of specific subpopulations to PM.
This chapter discusses the epidemiologic, controlled human exposure, and
toxicological studies evaluated in Chapters 6 and 7 that provide information on potentially
susceptible subpopulations. The studies highlighted include only those studies that
presented stratified results (e.g., males vs. females or <65 vs. > 65). This approach allowed
for a comparison between subpopulations exposed to similar PM concentrations and within
the same study design. Although this chapter does not provide a comprehensive evaluation
of all the studies that examined the effect of PM on potentially susceptible subpopulations
it does summarize the trends in the associations or health effects observed within the
identified subpopulations. Table 8-2 provides an overview of the factors identified in the
current toxicological, controlled human exposure, and epidemiologic literature that have
been shown to contribute to the susceptibility of a subpopulation to PM-related health
effects.
July 2009
8-2
DRAFT-DO NOT CITE OR QUOTE

-------
Table 8-2. Susceptibility Factors.
Factor
Exposure
PM Size Fraction Evaluated
Children (<18)16
Short-term
PM2.6, PM10-2.E
i, PM10
Older Adults (a 65)
Short-term
PM2.6, PM10-2.E
i, PM10

Long-term
PM2.6

Pregnancy and Developmental Effects
Long-term
PM2.6

Gender
Short-term
PM2.6, PM10


Long-term
PM2.6, PM10-2.E
i, PM10
Race/Ethnicity
Short-term
PM2.6, PM10

Genetic polymorphisms
Short-term
PM2.6


Long-term
PM10

Cardiovascular Diseases
Short-term
PM2.6, PM10


Long-term
PM2.6

Respiratory Illnesses
Short-term
PM2.6, PM10


Long-term
PM10

Respiratory Contributions to Cardiovascular Effects
Shot-term
PM2.6, PM10

Diabetes
Short-term
PM10

Obesity
Short-term
PM2.6

Health Status (e.g., Nutrition)
Short-term
PM2.6

Socioeconomic Status (SES)
Short-term
PM2.6, PM10-2.E
i, PM10
Educational Attainment
Short-term
PM2.6, PM10


Long-term
PM2.6

Residential Location
Short-term
PM10

8.1.1. Age
8.1.1.1. Older Adults
1	Evidence for PM-related health effects in older adults spans epidemiologic, controlled
2	human exposure, and toxicological studies. The 2004 PM AQCD found evidence for
3	increased risk of cardiovascular effects in older adults exposed to PM (U.S. EPA, 2004,
4	056905). Older adults represent a potentially susceptible subpopulation due to the higher
5	prevalence of pre-existing cardiovascular and respiratory diseases found in this age range
6	compared to younger age groups. The increased susceptibility in this subpopulation can
7	primarily be attributed to the gradual decline in physiological processes as part of the aging
8	process (U.S. EPA, 2006, 192082). Therefore, some overlap exists between potentially
9	susceptible older adults and the subpopulation that encompases individuals with pre-
15 The age range that defines a child varies from study to study. In some cases it is <21 years old while in others it is <18
years old (Firestone et al., 2007, 192071). For the purposes of this exercise children are defined as those individuals <18 years old
because the majority of epidemiologic studies consider individuals under the age of 18 children.
July 2009
8-3
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
existing diseases (Kan et al., 2008, 156621). Epidemiologic studies that conduct age
stratified analyses primarily focus on the association between short-term exposure to PM
and cardiovascular morbidity, but additional studies have examined the association
between PM and respiratory morbidity and mortality.
In recent publications, the epidemiologic evidence for cardiovascular effects in older
adults in response to short-term exposure to PM10-2.5 and PlVk.sis limited, but taken together
with evidence from studies of PM10 (e.g., Larrieu et al., 2007, 093031; Le Tertre et al., 2002,
023746), supports the increased risk of cardiovascular morbidity in older adults. Host et al.
(2007, 155851) found an increase in cardiovascular disease (CVD) hospital admissions
(HAs) in individuals >65 compared to all ages for short-term exposure to both PM10-2.5 and
PM2.5. Barnett et al. (2006, 089770) analyzed data from several cities across Australia and
New Zealand and found that the excess risk of hospitalizations for cardiac diseases,
congestive heart failure (CHF), ischemic heart disease (IHD), myocardial infarction (MI),
and all CVD was greater among patients aged > 65 as compared to those individuals <65
years in response to short-term exposure to PM2.5. U.S.-based studies that examined the
association between short-term exposure to PM and cardiovascular morbidity primarily
found no evidence for increased risk among older adults. Metzger et al. (2004, 011222)
found no evidence of effect modification by age for cardiovascular outcomes and short-term
exposure to PM2.5 in Atlanta, Georgia, which is supported by the results from other studies
that focused on short-term exposure to PM10 (Fung et al., 2005, 074322; Zanobetti and
Schwartz, 2005, 088069). However, Pope et al. (2008, 191969) observed an increased risk of
HF hospital admissions in older adults (i.e., > 65) in Utah, but the study used a 14-day
lagged cumulative moving average of PM2.5, which is much longer than the lags examined
by the other U.S.-based studies. Although studies have not consistently found an
association between short-term exposure to PM and respiratory-related health effects in
older adults, some studies have reported an increase in respiratory hospital admissions in
individuals 65 years of age and older (e.g., Fung et al., 2005, 093262).
Additional evidence for an increase in cardiovascular and respiratory effects among
older adults has been observed in controlled human exposure and dosimetric studies.
Devlin et al. (2003, 087348) found that older subjects exposed to PM2.5 concentrated
ambient particles (CAPs) experienced significant decreases in heart rate variability (HRV)
(both in time and frequency) immediately following exposure, when compared to healthy
young subjects. In addition, Gong et al (2004, 055628) reported that older subjects
demonstrated significant decreases in HRV when exposed to PM2.5 CAPs, but this study did
not compare the response in older subjects to those elicited by young, healthy individuals.
However, the study did find that healthy older adults were more susceptible to decreases in
HRV compared to those with an underlying health condition (i.e., chronic obstructive
pulmonary disease [COPD]) in response to PM exposure (Gong et al., 2004, 055628).
Dosimetric studies have shown a depression of PM2.5 and PM10-2.5 clearance in all regions of
July 2009
8-4
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
the respiratory tract with increasing age beyond young adulthood in humans and
laboratory animals. These results suggest that older adults are also susceptible to PM-
related respiratory health effects (Section 4.3.4.1).
Animal toxicological studies have attempted to characterize the relationship between
age and PM-related health effects through the development of models that mimic the
physiological conditions associated with older individuals. For example, Nadziejko et al.
(2004, 055632) observed arrhythmias in older, but not younger, rats exposed to PM2.5 CAPs.
In addition, another study (Tankersley et al., 2004, 094378) that used a mouse model of
terminal senescence observed various cardiovascular-related responses including altered
baseline autonomic tone in response to carbon black exposure that may subsequently affect
the quality and severity of cardiovascular responses (Tankersley, 2007, 188859).
Reductions in cardiac fractional shortening and significant pulmonary vascular congestion
upon exposure to carbon black were also reported in older mice (Tankersley et al., 2008,
157043). Overall, these studies provide biological plausibility for the increase in
cardiovascular effects in older adults observed in the controlled human exposure and
epidemiologic studies.
Recent epidemiologic studies have also found that individuals >65 years old are more
susceptible to all-cause (non-accidental) mortality upon short-term exposure to both PM2.5
(Franklin et al., 2007, 188502; Ostro et al., 2006, 087991) and PM10 (Samoli et al., 2008,
188455; Zeka et al., 2006, 088749), which is consistent with the findings of the 2004 PM
AQCD. Of note are the results from Ostro et al. (2006, 087991) that reported a slight
increase in mortality for older adults compared to all ages in single-pollutant models, but a
robust effect estimate in co-pollutant models with gaseous pollutants (i.e., PM2.5+CO and
PM2.5+NO2). These results differ from those in the all ages model (i.e., attenuation of the
effect estimate in co-pollutant models with CO and NO2), which suggests that older adults
are more susceptible to PM exposures, even though the age-stratified effect estimates in
single-pollutant models did not significantly differ. Epidemiologic studies that examined the
association between mortality and long-term exposure to PM (i.e., PM2.5) have found results
contradictory to those obtained in the short-term exposure studies. Villeneuve et al. (2002,
042576), Naess et al. (2007, 090736) and Zeger et al. (2008, 191951) report evidence of
differing PM2.5 relative risks by age, where risk declines with increasing age starting at age
60 until there is no evidence of an association among persons > 85 years of age.
The evidence from epidemiologic, controlled human exposure, and toxicological
studies that focused on exposures to PM2.5, PM10-2.5, and PM10, provide coherence and
biological plausibility for the association between PM and cardiovascular morbidity in older
adults. The clear pattern of positive associations only being observed in epidemiologic
studies conducted in non-U.S. locations brings into question the influence of PM
composition on health effects. However, the difference in effects observed between U.S. and
non-U.S. studies could also be due to possible differences in the identification of CVD-
July 2009
8-5
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
related morbidity and mortality between the studies evaluated. The additional evidence
from epidemiologic studies that focus on mortality and respiratory morbidity in response to
short-term exposure to PM also indicate that older adults represent a susceptible
subpopulation. As the demographics of the U.S. population shift over the next 20 years with
a larger percentage of the population (i.e., 13% of the population in 2011 and a projected
20% in 2030) encompassing individuals > 65 years (U.S. Census Bureau, 2000, 157064). an
increase in the number of PM-related health effects (e.g., cardiovascular and respiratory
morbidity, and mortality) in individuals > 65 years old could occur.
8.1.1.2. Children
Children have generally been considered more susceptible to PM exposure due to
multiple factors including more time spent outdoors, greater activity levels, exposures
resulting in higher doses per body weight and lung surface area, and the potential for
irreversible effects on the developing lung (U.S. EPA, 2004, 056905). The 2004 PM AQCD
found that studies which stratify results by age typically report associations between PM
and respiratory-related health effects in children, specifically asthma (U.S. EPA, 2004,
056905). Of the recent epidemiologic studies evaluated, only a few have examined the
association between PM10-2.5 and PM2.5 and respiratory effects in children. Mar et al. (2004,
057309) found increased respiratory effects (e.g., wheeze, cough, lower respiratory
symptoms) in children 7-12 years of age compared to individuals 20-51 years of age in
response to exposure to both PM10-2.5 and PMa.sin Spokane, Washingon. In addition, Host et
al. (2007, 155851) found an increase in respiratory-related hospital admissions with short-
term exposure to PM10-2.5 among children ages 0-14 years in 6 French cities. Further
support for these effects is provided by the results from studies that focused on PM10 (Mar
et al., 2004, 057309; Peel et al., 2005, 056305). A recent toxicological study provides
biological plausibility for the increase in PM-related respiratory effects in children observed
in the epidemiologic studies. Mauad et al. (2008, 156743) using both prenatal and
postneonatal mice exposed to ambient PM2.5 in a "polluted chamber" found evidence for
changes in lung function and pulmonary injury (e.g., incomplete alveolarization).
Additionally, Pinkerton et al. (2004, 087465; 2008, 190471) found evidence suggesting that
the developing lung is more susceptible to PM by demonstrating that neonatal rats exposed
to iron-soot PM had a reduction in cell proliferation in the lung. Overall, the evidence from
epidemiologic studies that have examined the health effects associated with all size
fractions of PM and toxicological studies that have examined individual PM components
provide additional support to the hypothesis that children are more susceptible to
respiratory effects from short-term exposure to PM.
July 2009
8-6
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
8.1.2. Pregnancy and Developmental Effects
While the majority of the literature focuses on epidemiologic studies that examine the
potential health effects (e.g., low birth weight, growth restriction) attributed to in utero
exposure to PM (see Section 7.4), it is unclear if the health effects observed are due to
soluble fractions of PM that cross the placenta or physiological alterations in the pregnant
woman. In the case of exposure to PM, adverse health effects in the offspring could be
mediated by potentially greater susceptibility in the pregnant woman. For example, an
inflammatory response leads to differential activation of multiple genes involved in immune
response and regulation, cell metabolism, and proliferation all of which can lead to health
effects in the developing fetus (Fedulov et al., 2008, 097482). Toxicological studies have
recently examined whether exposure to air pollutants during pregnancy leads to increased
allergic susceptiblity in the offspring. Fedulov et al. (2008, 097482) used an animal model
to examine the effect of diesel exhaust particles (DEPs) along with an immunologically
"inert" particle (TKb) on pregnant mice. The authors found that pregnant mice exhibited a
local and systemic inflammatory response when exposed to either DEP or Ti02, which was
not observed in control, non-pregnant mice. In addition, the offspring of exposed pregnant
mice developed AHR and allergic inflammation. This study suggests that exposure to PM2.5,
and even relatively inert particles, during pregnancy can potentially lead to increased
allergic susceptibility in offspring and susbequently the development of asthma.
8.1.3. Gender
The 2004 PM AQCD did not find consistent evidence for a difference in health effects
by gender. However, there appeared to be gender differences in the localization of particles
when deposited in the respiratory tract and the deposition rate due to differences in body
size, conductive airway size, and ventilatory parameters (U.S. EPA, 2004, 056905). For
example, females have proportionally smaller airways and slightly greater airway
reactivity than males (Yunginger et al., 1992, 192071).
Few recent epidemiologic studies have conducted gender-stratified analyses when
examining the association between either short- or long-term exposure to PMio-2.5or PM2.5.
Similar to the studies evaluated in the 2004 PM AQCD, the current literature has not found
a consistent pattern of associations by gender for any health outcome. Pope et al. (2006,
189048) observed a slightly larger, non-significant, association between short-term
exposure to PM2.5 and daily HA for acute IHD events in males. An examination of gender-
specific effects by both Ostro et al. (2006, 087991) and Franklin et al. (2007, 188502) found
conflicting associations by gender for multiple cause-specific mortality outcomes. The
inconsistency in associations between males and females is further highlighted in studies
that examined the health effects associated with long-term exposure to PM10-2.5 and PM2.5.
July 2009
8-7
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Chen et al. (2005, 087942) found larger effects in females for congestive heart disease
(CHD) mortality upon long-term exposure to PMio a sin 3 California cities. Naess et al.
(2006, 189048). also observed slightly larger effect estimates in females for CVD and lung
cancer mortality upon long-term exposure to PM2.5, but for COPD mortality the greatest
association was found in males.
The majority of the epidemiologic studies that examined the association between
exposure to PM and gender focused on exposure to PM10. Although most of these studies do
not attribute the association to specific size fractions (i.e., PMio-2.5or PM2.5) or provide
insight as to whether one size fraction may be driving the observed effect, the studies of
PM10 provide further support that gender does not appear to differentially affect PM-related
health outcomes. Neither Zanobetti and Schwartz (2005, 088069) nor Wellenius et
al. (2006, 088748) found gender to be a significant effect modifier of the risk estimates
associated with short-term exposure to PM10 and cardiovascular hospital admissions. These
results are consistent with those found in other studies that examined the association
between short-term exposure to PM10 and both cardiovascular and respiratory hospital
admissions (Luginaah et al., 2005, 057327; Middleton et al., 2008, 156760). Additional
studies that examined the effects of short-term and long-term exposure to PM10 on
respiratory morbidity and mortality (Boezen et al., 2005, 087396; Chen et al., 2005, 087942;
Zanobetti and Schwartz, 2005, 088069; Zeka et al., 2006, 088749) found results that are
consistent with those reported in studies of PM10-2.5 and PM2.5 (i.e., gender is not likely to be
an effect modifier).
Although human clinical studies are not typically powered to detect differences in
response between males and females, one study did report significantly greater decreases in
blood monocytes, basophils, and eosinophils in females compared to males following
controlled exposures to UF EC (Frampton et al., 2006, 088665). Overall, the evidence from
primarily epidemiologic studies that examined the association between short- and long-
term exposure to PM10-2.5 and PM2.5, along with the supporting evidence from PM10 studies,
further confirms that although differences in dosimetry exist between males and females,
neither gender consistently exhibits a higher disposition for PM-related health effects.
8.1.4. RacefEthnicity
The 2004 PM AQCD did not evaluate the potential susceptibility of individuals of
different races and ethnicities to PM exposure. The results from epidemiologic studies
evaluated in this review that examined the potential effect modification of the PM-
morbidity and -mortality relationships by race and ethnicity varied depending on the study
location. In an analysis of the PM2.5-mortality relationship, Ostro et al. (2006, 087991)
stratified the association by race and ethnicity, and observed a positive and marginally
significant effect for whites and Hispanics, but not for blacks, in response to short-term
July 2009
8-8
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
exposure to PM2 .5 in 9 California counties. An additional analysis performed by Ostro et al. (2008,
097971) in 6 California counties using PM2.5 and various PM2.5 components, also found a
significant association between mortality, specifically cardiovascular mortality, and
Hispanic ethnicity (Ostro et al., 2008, 097971). It should be noted that neither study, Ostro
et al. (2006, 087991) nor Ostro et al. (2008, 097971), controlled for potential confounders
(e.g., SES factors and location of residence) of the association observed between PM2.5
exposure and Hispanic ethnicity. As a result, Ostro et al (2008, 097971) speculated that the
increased PM2.5-mortality risks observed for Hispanics could be due to a variety of factors
including, higher rates of non-high school graduates, obesity, no leisure-time activity, and
alcohol consumption within the Hispanic population in California. Additional evidence for
the potential susceptibility of individuals by race and ethnicity were derived from studies
on the health effects associated with short-term exposure to PM10. Wellenius et al. (2006,
088748) observed that race (i.e., white vs. other) did not significantly modify the association
between short-term exposure to PM10 and CHF hospital admissions. Additionally, Zeka et
al. (2006, 088749) did not observe any difference in mortality effect estimates when
stratifying by race (i.e., black and white) upon short-term exposure to PM10. To date,
dosimetric studies have not extensively examined differences in particle deposition between
races or ethnicities to confirm the epidemiologic findings. Although not extensively
analyzed, toxicological studies, have examined PM responses in different mouse and rat
strains have reported greater CV effects (Kodavanti et al., 2003, 051325; Tankersley et al.,
2007, 097910) and compromised host defense (Ohtsuka et al., 2000, 004409) for some
strains. These studies provide some support, in terms of biological plausibility, for
differences in PM-induced health effects by race or ethnicity. However, it is unclear how the
difference in the response to PM in different mouse or rat strains extrapolates to PM-
induced differences between races or ethnicities. Overall, the results from the studies that
examined the potential effect modification of PM associations by race and ethnicity provide
some evidence for increased risk of mortality in Hispanics upon short-term exposure to
PM2.5. However, the evidence for this association is from two studies conducted in
California, it is unclear if the studies adequately controlled for potential confounders, and
additional studies in other locations that stratified results by race and ethnicity have not
yet been conducted.
8.1.5. Gene-Environment Interaction
A consensus now exists that gene-environment interactions merit serious
consideration when examining the relationship between ambient exposures to air
pollutants and the development of health effects (Gilliland et al., 1999, 155792; Kauffmann
and Post Genome Respiratory Epidemiology, 2004, 090968). These potential interactions
were not evaluated in the 2004 PM AQCD. Inter-individual variation in human responses
July 2009
8-9
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
to air pollutants suggests that some subpopulations are at increased risk of detrimental
effects due to pollutant exposure, and it has become clear that the genetic makeup of an
individual can increase their susceptibility (Kleeberger and Ohtsuka, 2005, 130489).
Gene-environment interactions can result in health effects due to: genetic polymorphisms,
which result in the lack of a protein or a change that makes a functionally important
protein dysfunctional! or genetic damage in response to an exposure which potentially leads
to a health response (e.g., formation ofbenzo [a] pyrene DNAadducts in response to PM
exposure). In this review, the majority of studies examine gene-environment interactions
due to genetic polymorphisms. In order to establish useful links between polymorphisms in
candidate genes and adverse health effects, several criteria must be satisfied: the product of
the candidate gene must be significantly involved in the pathogenesis of the adverse effect
of interest! and polymorphisms in the gene must produce a functional change in either the
protein product or in the level of expression of the protein (U.S. EPA, 2008, 157075).
Further, the issue of confounding by other environmental exposures must be carefully
considered.
It has been hypothesized that the cardiovascular and respiratory health effects that
occur in response to short-term PM exposure are mediated by oxidative stress (see
Section 5.1.1). Research has examined this hypothesis by primarily focusing on the
glutathione-S transferase (GST) genes because they have common, functionally important
polymorphic alleles that significantly affect antioxidant defense function in the lung (e.g.,
homozygosity for the null allele at the GSTM1 and GSTT1 loci, homozygosity for the A105G
allele at the GSTP1 locus), and approximately half of the white population has a
polymorphic null allele, resulting a large potential study population (Schwartz et al., 2005,
086296). Exposure to free radicals and oxidants in air pollution leads to a cascade of events,
which can result in a reduction in glutathione (GSH), and an increase in the transcription
of GSTs. Individuals with genotypes that result in reduced or absent enzymatic activity are
likely to have reduced antioxidant defenses and potentially increased susceptibility to
inhaled oxidants and free radicals.
Numerous studies have examined the role of genetic polymorphisms on PM-related
cardiovascular health effects using the Normative Aging Study cohort. Schwartz et al.
(2005, 086296) and Chahine et al. (2007, 156327) found that individuals with null GSTM1
alleles had a larger decrease in HRV upon short-term exposure to PM2.5 compared to
individuals with at least one allele. Polymorphisms in the HO-1 promoter resulted in
lowered HRV upon short-term exposure to PM2.5 in individuals with the long repeat
polymorphism compared to those individuals with the short repeat polymorphism (Chahine
et al., 2007, 156327). In addition, Schneider et al. (2008, 191985) found that diabetic
individuals with null GSTM1 alleles had larger decrements in FMD (i.e., flow-mediated
dialation of the brachial artery), suggesting alterations in endothelial function. A controlled
human exposure study examined whether genetic polymorphisms increase the
July 2009
8-10
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
susceptibility of individuals to respiratory morbidity in response to PM exposure. Gilliland
et al. (2004, 156471) examined the effect of allergens and DEPs on individuals with either
null genotypes for GSTM1 and GSTT1 or GSTP1 codon 105 variants. The authors found
that individuals with the GSTM1 null or the GSTP11105 wildtype genotypes were more
susceptible to allergic inflammation upon exposure to allergen and DEPs. Additional genes
within the GST pathway have also been examined (e.g., NQOl), but the sample sizes are
relatively small, which prohibits the analysis of the potential effect modification of PM-
related health effects by these genes (e.g., Schneider et al., 2008, 191985).
The interaction between GST genes and PM exposure has recently been extended to
studies that examined the effect of PM exposure on birth outcomes. A recent study (Suh et
al., 2008, 192077) that examined the effect of high PM10 exposures during the third
trimester of pregnancy on the risk of preterm delivery, found that women with the GSTM1
null genotype were at an increased risk of preterm birth. When examining the interaction
between high PM10 concentrations during the third timester of pregnancy and the presence
of the GSTM1 null genotype on the risk of preterm delivery, there was evidence for a
synergistic gene-environment interaction in pregnant women. This effect could occur due to
oxidative stress induced by metals contained in PM10, which could be modified by
polymorphisms of the GSTM1 gene. This oxidative stress causes oxidative DNA damage in
fetal tissues, which may lead to preterm delivery by causing a reduction in placental blood
flow.
An examination of other genes outside the GST pathway have also been conducted to
determine if specific polymorphisms increase the susceptibility of individuals to PM.
Baccarelli et al. (2008, 157984) in the Normative Aging Study observed that individuals
with polymorphisms in MTHFR (C677T methylenetetrahydrofolate reductase), an
alteration associated with reduced enzyme activity, and cSHMT (cytoplasmic serine
hydroxymethyltransferase) (i.e., [CT/TT] MTHFR and [CC] cSHMT genotypes), alterations
associated with higher homocysteine levels, have a reduction in SDNN, upon exposure to
PM2.5. Peters et al. (2009, 191992) examined single nucleotide polymorphisms (SNPs) in the
fibrinogen gene in myocardial infarction survivors to assess whether exposure to PM10
altered physologic levels of fibrinogen, which has been implicated in promoting
atherothrombosis. The authors found that individuals with single nucleotide
polymorphisms (SNPs) in the fibrinogen gene have higher steady state fibrinogen levels
which when combined with the inflammatory effects (i.e., increased fibrinogen levels)
associated with exposure to PM could increase their risk of PM-related cardiovascular
health effects. These results taken together suggest that individuals with null alleles or
specific polymorphisms in genes that mediate the antioxidant response to oxidative stress,
regulate enzyme activity, or regulate physiological levels of inflammatory markers are more
susceptible to PM. However, in some cases genetic polymorphisms may actually reduce an
individual's susceptibility to PM-related health effects. For example, Park et al. (2006,
July 2009
8-11
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
189099) found that individuals with two hemochromatosis (HFE) polymorphisms (C282Y
and H63D), which result in an increase in iron uptake, had smaller reductions in HRV upon
exposure to PM2.5. This effect could possibly be due to the reduction in free iron that enters
oxidation-reduction (redox) reactions and the subsequent reduction in reactive oxygen
species (ROS).
More recently studies have begun to focus on epigenetic effects associated with PM
exposure (i.e., the effect of PM on DNA methylation) due to the fact that DNA methylation
can result in gene alterations. The limited number of epidemiologic studies that examined
epigenetic effects have found some evidence that long-term exposure to PM2.5 and PM10 can
influence DNA methylation (Baccarelli et al., 2009, 188183: Tarantini et al., 2009, 192010).
Additionally, a toxicological study found some evidence of hypermethylation of
spermatogonial stem cells in response to the PM component of ambient urban air (Yauk et
al., 2008, 157164). Although epigenetic effects have been observed in response to PM
exposure in some studies additional research is needed to more accurately characterize
these associations.
Overall, the evidence suggests that specific genetic polymorphisms can potentially
increase the susceptibility of an individual to PM exposure, but protective polymorphisms
also exist, which may diminish the health effects attributed to PM exposure in some
individuals. In addition, the studies that examine genetic polymorphisms or epigenetics can
potentially provide additional information that can aid in identifying the specific pathways
and mechanisms by which PM initiates health effects.
8.1.6. Pre-Existing Disease
In 2004, the National Research Council (NRC) published a report that emphasized
the need to evaluate the effect of air pollution on susceptible and subpopulations, including
those with respiratory illnesses and cardiovascular diseases (NRC, 2004, 156814). The 2004
PM AQCD included epidemiologic evidence suggesting that individuals with pre-existing
heart and lung diseases, as well as diabetes may be more susceptible to PM exposure. In
addition, toxicological studies that used animal models of cardiopulmonary diseases and
heightened allergic sensitivity found evidence of enhanced susceptibility. More recent
epidemiologic and human clinical studies have directly examined the effect of PM on
individuals with pre-existing diseases and toxicological studies have employed disease
models to identify whether exposure to PM disproportionately effects certain
subpopulations.
8.1.6.1. Cardiovascular Diseases
The potential effect of underlying cardiovascular diseases on PM-related health
responses has been examined using epidemiologic studies that stratify effect estimates by
July 2009
8-12
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
underlying conditions or secondary diagnoses, and toxicological studies that use animal
models to mimic the physiological conditions associated with various cardiovascular
diseases (e.g., MI, ischemia, and atherosclerosis). A limited number of controlled human
exposure studies have also examined the potential relationship between cardiovascular
diseases and exposure to PM in individuals with underlying cardiovascular conditions, but
these studies have provided somewhat inconsistent evidence for these associations.
The majority of the epidemiologic literature that examined the association between
short-term exposure to PM and cardiovascular outcomes focuses on cardiovascular-related
hospital admissions and emergency department (ED) visits. Hypertension is the pre-
existing condition that has been considered to the greatest extent when examining the
association between short-term exposure to PM and cardiovascular-related HAs and ED
visits. Pope et al. (2006, 091246) found no evidence of effect modification of the IHD ED
visit association with PIVL.r, in individuals with secondary hypertension in Utah. This is
consistent with the results of both Wellenius et al. (2006, 088748) in 7 U.S. cities and Lee et
al. (2008, 192076) in Taipei, which found that hypertension did not modify the association
between PMio and cardiovascular-related health outcomes. These results differ from those
presented by Peel et al. (2007, 090442), in Atlanta, which observed that exposure to PMio
resulted in an increase in ED visits for arrhythmias and CHF in individuals with
underlying hypertension. An additional study conducted by Park et al. (2005, 057331) in
Boston found that underlying hypertension increased associations between HRV,
specifically a reduction in the HF parameter, and short-term exposure to PM2.5.
Park et al. (2005, 057331). in the analysis mentioned above, examined other
underlying cardiovascular conditions and found associations between PM2.5 and HRV in
individuals with pre-existing IHD. In a toxicological study, Wellenius et al. (2003, 055691)
examined the effects of PM2.5 CAPs exposure on induced myocardial ischemia in dogs,
which mimics the effects associated with IHD. The authors found that exposure to PM2.5
prior to the induced ischemia increased ST-segment elevation, indicating greater ischemia
than air-exposed animals. A follow-up study implicated impaired myocardial blood flow in
the response (Bartoli et al., 2009, 179904).
Additional studies examined the effects of PM on cardiac function in individuals with
dysrhythmia. Peel et al. (2007, 090442) observed some evidence for an increase in ED visits
for IHD for individuals with secondary dysrhythmia and PMio exposure. However, when
examining CHF hospital admissions in 7 U.S. cities, Wellenius et al. (2006, 088748) found
no evidence for effect modification of PMio exposure in individuals with secondary
dysrhythmia.
Limited evidence is available from epidemiologic studies that examined other pre-
existing cardiovascular conditions, such as CHF and MI. Pope et al. (2006, 091246)
observed an increase in hospital admissions for acute IHD in individuals with underlying
CHF upon short-term exposure to PM2.5. However, Peel et al. (2007, 090442) did not find that
July 2009
8-13
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
underlying CHF contributed to an increase in the association between IHD ED visits and
short-term exposure to PMio. Zanobetti and Schwartz (2005, 088069) also examined the
potential effect modification of the association between PMio and cardiovascular-related
health effects in individuals with CHF, but used MI hospital admissions as the outcome of
interest. Underlying CHF was not found to increase MI hospital admissions for exposure to
PMio in the cohort of more than 300,000 hospital admissions.
Wellenius et al. (2006, 088748) examined the effect of previous diagnoses of acute MI
on the association between CHF hospital admissions and short-term exposure to PMio in 7
U.S. cities. In this study, Wellenius et al. (2006, 088748) found no evidence of effect
modification of the relationship between PMio and CHF hospital admissions by previous
acute MI. Toxiciological studies have provided additional evidence for the cardiovascular
health effects associated with exposure to PM in individuals with underlying MI. Anselme
et al. (2007, 097084) and Wellenius et al. (2006, 156152) examined the arrhythmic effects of
PM on rats that experienced an MI using two different models. Wellenius et al. (2006,
156152) used a post-myocardium sensitivity model (acute MI) and observed that exposure
to PM2.5 CAPs decreased ventricular premature beats and spontaneous supraventricular
ectopic beats. In contrast, the MI model of chronic heart failure (i.e., rats that experienced
an MI 3 months prior to exposure), demonstrated a prominent increase in the incidence of
premature ventricular contraction when exposed to DE (Anselme et al., 2007, 097084). The
discrepancy in effects observed between studies could be due to differences in the MI model
or the PM exposure (i.e., CAPs vs. DE).
Additional toxicological studies examined the association between PM and
pre-existing cardiovascular diseases using a murine model of atherosclerosis (ApoE"'"
mouse). For example, Campen et al. (2005, 083977; 2006, 096879) examined the heart rate
and ECG effects of acute exposure to PM on ApoE"'" mice. With DE, dramatic bradycardia
and T-wave depression were observed that were attributable to the gases (Campen et al.,
2005, 083977), while whole gasoline emissions induced T-wave alterations that required
particles (Campen et al., 2006, 096879). However, these studies along with others that used
this mouse model (see Section 6.2 and 7.2) did not compare the effects observed with the
ApoE '" mouse to other non-diseased mouse models, so it is unclear if the responses would
differ if other strains underwent the same experimental protocol.
Controlled human exposure studies that examined the effect of pre-existing diseases
on cardiovascular outcomes with exposure to PM are less consistent and difficult to
interpret in the context of the results from the epidemiologic and toxicological studies. Mills
et al. (2007, 091206; 2008, 156766) investigated the effects of dilute DE, or fine and
ultrafine CAPs, respectively, on subjects with coronary artery disease and prior MI.
Exposure to dilute DE was found to promote exercise-induced ST-segment changes
indicating myocardial ischemia, as well as inhibit endogenous fibrinolytic capacity Mills et
al. (2007, 091206). The physiological responses observed in Mills et al. provides a measure
July 2009
8-14
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
of coherency with the cardiovascular effects observed in epidemiologic studies, including
increases in hospital admissions and ED visits for IHD and stroke associated with exposure
to PM. An examination of fine and ultrafine CAPs that were low in combustion derived
particles, were not found to exhibit any significant effects on vascular function (Mills et al.,
2008, 156766). Routledge et al. (2006, 088674) reported no change in HRV in a group of
adults with coronary artery disease following exposure to ultrafine carbon particles, which
may be explained in part by the use of medication (beta blockers) among the majority of the
subjects.
Although the epidemiologic studies did not examine potential effect modification of
pre-existing cardiovascular conditions on effects of long-term exposure to PM, a few
toxicological studies exposed animals with underlying cardiovascular conditions to PM for
months. In studies that focused on the cardiovascular effects following subchronic exposure
to PM in ApoE"'" mice, relatively consistent physiological effects were observed across
studies. Araujo et al. (Araujo et al., 2008, 156222) exposed mice to ultrafine CAPs and
observed enhanced size of early atherosclerotic lesions. Similarly, Chen and Nadziejko
(2005, 087219) and Sun et al. (2005, 186814; Sun et al., 2008, 157033) exposed mice to
PM2.5 CAPs with the same results. An additional long-term exposure study observed a
decreasing trend in heart rate, physical activity, and temperature along with biphasic
responses in HRV (SDNN and rMSSD) upon exposure to CAPs (Chen and Hwang, 2005,
087218).
While the majority of the literature examines the potential modification of the
association between PM and non-fatal cardiovascular health effects, a few new studies have
also examined effect modification in mortality associations. Zeka et al. (2006, 088749) found
an increase in risk estimates for associations between PM10 and mortality in individuals
with underlying stroke, while Bateson et al. (2004, 086244) found evidence for effect
modification of the PM-mortality association in individuals with CHF.
Collectively, the evidence from epidemiologic and toxicological, and to a lesser extent,
controlled human exposure studies indicates increased susceptibility of individuals with
underlying cardiovascular diseases to PM exposure. Although the evidence for some
outcomes was inconsistent across epidemiologic and toxicological studies, this could be due
to a variety of issues including the PM size fraction used in the study along with the study
location. Even with these caveats, a large proportion of the U.S. population has been
diagnosed with cardiovascular diseases (i.e., approximately 51.6 million people with
hypertension, 24.1 million with heart disease, and 14.1 million with coronary heart disease
[see Table 8-3]), and therefore represents a large subpopulation that is potentially more
susceptible to PM exposure than the general population.
July 2009
8-15
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Table 8-3. Percent of the U.S. population with respiratory diseases, cardiovascular diseases, and
diabetes.



Age




Regional


Adults (18+)*

18-44
45-64
65-74
75+
NE
MW
s
w
Chronic Condition/ Disease
Number (x 106)
%
%
%
%
%
%
%
%
%
RESPIRA TORY DISEASES
Asthma*
24.2
11.0
11.5
10.5
11.7
9.3
11.7
11.5
10.5
10.8
Asthma (< 18 yrs)
6.8*
CD
CO
...
...
...
...
...
...
...
...
COPD
Chronic bronchitis
9.5
4.3
2.9
5.5
5.6
6.7
3.8
4.4
4.9
3.5
Emphysema
4.1
1.8
0.3
2.4
5.0
6.4
1.4
2.3
1.9
1.6
CARDIOVASCULAR DISEASES
All heart disease
24.1
10.9
3.6
12.3
26.1
36.3
10.8
12.7
10.9
9.2
Coronary heart disease
14.1
6.4
0.9
7.2
18.4
25.5
6.4
7.6
6.6
4.7
Hypertension
51.6
23.4
77
32.4
52.7
53.5
22.2
23.7
25.3
20.6
Stroke
5.6
2.6
0.5
2.4
7.6
11.2
2.1
2.8
2.9
2.2
Diabetes
17.1
7.8
2.6
10.4
18.2
17.9
7.2
8.1
8.0
7.4
* All data for adults except asthma prevalence data for children under 18 years of age, from CDC (2008,156324; 2008,156325). For adults prevalence data based off
adults responding to "ever told had asthma."
Source: Pleis and Lethbridge Cejku (2007,156875); CDC (2008,156324; 2008,156325).
8.1.6.2. Respiratory Illnesses
Investigators have examined the effect of pre-existing respiratory illnesses on
multiple health outcomes (e.g., mortality, asthma symptoms, CHF) in response to exposure
to ambient levels of PM. Animal models have been developed and/or human clinical studies
conducted to examine the possible PM effects on pre-existing respiratory conditions in a
controlled setting.
Epidemiologic studies have examined the effect of short-term exposure to PM on the respiratory health of
asthmatic individuals measuring a variety of respiratory outcomes. Asthmatic individuals were found to
have an increase in medication use (Rabinovitch et al., 2006, 088031). respiratory symptoms (i.e., asthma
symptoms, cough, shortness of breath, and chest tightness (Gent et al., 2003, 052885). and asthma
symptoms (Delfino et al., 2002, 093740; 2003, 050460) with short-term exposure to PM2.5; and morning
symptoms (Mortimer et al., 2002, 030281) and asthma attacks (Desqueyroux et al., 2002, 026052) with
short-term exposure to PMi0.
Toxicological studies that have used ovalbumin-induced allergic airway disease models provide
evidence which supports the findings of the epidemiologic literature. Morishita et al. (2004, 087979) used
this model to assess the health effects of PM2.5 components. In response to a short-term exposure to CAPs
from Detroit, an area with pediatric asthma rates three times the national average, rats with allergic airway
disease were found to preferentially retain PM derived from identified local combustion sources in
association with eosinophil influx and BALF protein content after an acute exposure (Morishita et al.,
2004, 087979). These findings suggest that individuals with allergic airways conditions are more
July 2009
8-16
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
susceptible to allergic airways responses upon exposure to PM2.5, which may be partially attributed to
increased pulmonary deposition and localization of particles in the respiratory tract (Morishita et al.,
2004, 087979). An additional study (Heidenfelder et al., 2009, 190026) examined whether genes are
differentially expressed upon exposure to PM. They found that exposure to CAPs increased the
expression of genes associated with inflammation and airway remodeling in rats with allergic airway
disease. Although the evidence is much more limited, not all of the toxicological studies evaluated that
examined the effect of underlying respiratory conditions on PM-related respiratory morbidity focused on
allergic airways disease. Using an animal model of emphysema (i.e., papain-treated mice) Lopes et al.
(2009, 190430) found that papain-treated mice exposed to urban ambient PM demonstrated a statistically
significant increase in mean linear intercept, a measure of airspace enlargement, compared to saline-
treated controls exposed to filtered air. These results provide preliminary evidence, which suggests that
non-allergic respiratory morbidities may also increase the susceptibility of an individual to PM-related
respiratory effects.
The results from the epidemiologic and toxicological studies that focused on underlying allergic
airways disease is supported by a series of controlled human exposure studies which have shown that
exposure to DEPs increases the allergic inflammatory response in atopic individuals (Bastain et al., 2003,
098690; Diaz-Sanchez et al., 1997, 051247; Nordenhall et al., 2001, 025185). However, not all controlled
human exposure studies have found evidence for differences between the respiratory effects exhibited by
healthy and asthmatic individuals. Studies by Gong et al. (2003, 042106; 2004, 055628; Gong et al.,
2008, 156483) reported that healthy and asthmatic subjects exposed to coarse, fine and ultrafine CAPs,
exhibited similar respiratory responses. However, it should be noted that these studies excluded moderate
and severe asthmatics that would be expected to show increased susceptibility to PM exposure.
In addition to examining the association between exposure to PM and respiratory
effects in asthmatics, some studies examined whether individuals with COPD represent a
potentially susceptible subpopulation. Desqueyroux et al. (2002, 026052) did not observe an
increase in the exacerbation16 of COPD in response to short-term exposure to PM2.5.
However, studies that examined the effect of PM on lung function in individuals with COPD
(Lagorio et al., 2006, 089800; Trenga et al., 2006, 155209) observed declines in FEVi, and
FEVi and FVC, respectively in response to PM10 and/or PM2.5. Silkoff et al. (2005, 087471)
observed associations between PM10 and a reduction in FEVI and PM2.5 and a reduction in
PEF, in those with COPD, but only during one winter of the analysis. Only one controlled
human exposure study examined the effects of PM on COPD subjects and found no
significant difference in respiratory effects between healthy and individuals with COPD
upon exposure to PM2.5 CAPs (Gong et al., 2004, 055628). On the other hand the results
from dosimetric studies have shown that COPD patients have increased dose rates and
impaired mucociliary clearance relative to age matched healthy subjects, suggesting that
16 Desqueyroux et al. (2002, 026052) defined a COPD exacerbation as (a) decrease in "vesicular" breath sound, (b) bronchial
obstruction, (c) tachycardia or arrhythmia, or (d) cyanosis.
July 2009
8-17
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
individuals with COPD are potentially at a greater risk of PM-related health effects
(Sections 4.2.4.5 and 4.3.4.3).
A few of the epidemiologic studies examined the effect of underlying respiratory
illnesses on the association between short- and long-term exposure to PM and mortality.
Using different pre-existing respiratory illnesses, Zeka et al. (2006, 088749) and De Leon et
al. (2003, 055688) found that short-term exposure to PMio increased the risk of
non-accidental mortality for pneumonia and circulatory mortality for all respiratory
illnesses, respectively. Additionally, Zanobetti et al. (2008, 156177) observed an association
between long-term exposure to PMio and mortality in individuals that had previously been
hospitalized for COPD. Although these studies do not examine additional size fractions of
PM, together they highlight the potential effect of underlying respiratory illnesses on the
PM-mortality relationship.
Overall, the epidemiologic, controlled human exposure, and toxiciological studies
evaluated provide biological plausibility for the increased health effects observed in
epidemiologic studies among asthmatic individuals in response to PM exposure. Although,
the evidence from studies that examined associations between PM and health effects in
individuals with COPD is inconsistent, taken together individuals with COPD and asthma
represent a large percent of the U.S. population, which may be more susceptible to PM-
related health effects (Table 8-3).
8.1.6.3. Respiratory Contributions to Cardiovascular Effects
Although the majority of health effects observed in individuals with pre-existing
respiratory illnesses were associated with respiratory illness exacerbations, studies also
examined whether underlying respiratory illnesses can lead to cardiovascular effects with
PM exposure. Controlled human exposure and toxicological studies have also observed some
cardiovascular effects in individuals with pre-existing respiratory illnesses. Gong et al.
(2003, 042106) observed acute responses in the cardiovascular system and systemic
circulation among asthmatic individuals after exposure to PM2.5 CAPs. However,
respiratory disease has not consistently been observed to affect cardiovascular response in
controlled human exposure studies. In a toxicological study, Batalha et al. (2002, 088109),
using a chronic bronchitis animal model, found that the pulmonary artery lumen-to-wall
ratio was decreased in rats exposed to PM2.5 CAPs, although the induced bronchitis didn't
seem to affect the response. The majority of epidemiologic studies that examined whether
underlying respiratory illnesses contributed to the mainfestation of PM-related
cardiovascular hospital admission or ED visits, did not report increases in effects for a
variety of cardiovascular outcomes (e.g., IHD, arrhythmias, CHF, MI) for individuals with
underlying respiratory infection (Wellenius et al., 2006, 088748), pneumonia (Zanobetti and
Schwartz, 2005, 088069), or COPD (Peel et al., 2007, 090442; Wellenius et al., 2005,
087483). However, Yeatts et al. (2007, 091266), in a panel study, found evidence for
July 2009
8-18
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
cardiovascular effects, specifically reductions in HRV parameters, in asthmatic adults upon
short-term exposure to PM10-2.5. It must be noted that most of the aforementioned
epidemiologic studies focused on exposure to PM10, and, therefore, it is unclear how these
results compare to those found in the controlled human exposure and toxicological studies
that focused on exposure to PM2.5 (e.g., CAPs). Thus, it is unclear if individuals with
underlying respiratory illnesses represent a subpopulation that is potentially susceptible to
PM-related cardiovascular effects.
8.1.6.4. Diabetes and Obesity
It has been hypothesized that the systemic inflammatory cascade leads to an increase
in cardiovascular risk (Dubowsky et al., 2006, 088750). As a result, individuals with
conditions linked to chronic inflammation (i.e., diabetes and obesity), have been examined
to determine whether diabetes or obesity facilitate the manifestation of PM-mediated
health effects, and, therefore, represent a potentially susceptible subpopulation.
Epidemiologic studies have examined whether diabetes modifies the association
between cardiovascular health effects and PM exposure, but these studies have primarily
focused on short-term exposure to PM10. Time-series studies have provided evidence
through an examination of hospital admission and ED visits and mortality, which suggests
an increase in health effects in diabetic individuals in response to PM exposure. Multicity
studies have found upwards of 75% greater risk of hospitalization for cardiac diseases in
individuals with diabetes upon to exposure to PM10 (Zanobetti and Schwartz, 2002,
034821). Studies conducted in Atlanta, Georgia have also found increased risk for
cardiovascular-related ED visits in diabetics, specifically for IHD, arrhythmias, and CHF
(Peel et al., 2007, 090442). Additional studies found some evidence that individuals with
diabetes are at increased risk of mortality upon exposure to PM10 (Zeka et al., 2006,
088749) and PM2.5 (Goldberg et al., 2006, 088641). However, some studies (both multicity
and single-city) have not observed a modification of the risk of cardiovascular ED visits and
hospital admissions in response to exposure to PM10 in diabetics (Pope et al., 2006, 091246;
Wellenius et al., 2006, 088748; Zanobetti and Schwartz, 2005, 088069).
Panel and cohort studies have been conducted to determine the physiological changes
that occur in individuals with diabetes in response to PM exposure. These studies examined
both changes in inflammatory markers along with specific physiological alterations in the
cardiovascular system. Schneider et al. (2008, 191985) in a panel study of 22 individuals
with type 2 diabetes mellitus in Chapel Hill, NC found evidence that ambient exposure to
PM2.5 enhanced the reduction in various markers of endothelial function. Liu et al. (2007,
156705) observed an increase in end-diastolic FMD and end-systolic FMD, and decreases in
end-diastolic basal diameter and end-systolic basal diameter in diabetics upon exposure to
PM10. The authors also observed positive associations with FMD and blood pressure in
diabetic individuals. A controlled human exposure study conducted by Carlsten et al. (2008,
July 2009
8-19
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
156323) found that DE did not elicit any prothrombotic effects in subjects with metabolic
syndrome, which consists of physiological alterations similar to those observed in both
diabetic and obese individuals. An examination of biomarkers found mixed results, with
Liao et al. (2005, 088677) observing an increase in vWF; Liu et al. (2007, 156705) observing
an increase in TBARS, but not CRP or TNF-a! and Dubowsky et al. (2006, 088750)
observing an increase in CRP and WBCs. Overall, it is unclear how differences in each of
the aforementioned biomarkers contribute to the potential overall cardiovascular effect
observed in diabetic individuals! however, an increase in inflammation, oxidative stress,
and acute phase response may contribute to cardiovascular effects. A recent toxicological
study, also demonstrated the potential for PM-related health effects in diabetics. Sun et al.
(2009, 190487) found that PM2.5 CAPs exposure for 4 months can exaggerate insulin
resistance, visceral adiposity, and inflammation in a diet-induced obesity mouse model.
Overall, epidemiologic studies have reported evidence for increased effects in
diabetics in response to PM exposure, with preliminary evidence for physiological
alterations from toxicological studies. However, the limited evidence from toxicological and
controlled human exposure studies along with the lack of studies that examined additional
PM size fractions warrants additional research to confirm these associations and to identify
the biological pathway(s) that may result in a greater response to PM in diabetics. This
potentially susceptible subpopulation is large, with an estimated 17.1 million diabetic
individuals in the U.S. (Table 8-3).
In addition to diabetes, obesity has been examined as a health condition with the
potential to lead to an increase in PM-related health effects. Only a few recent studies have
examined the potential effect modification of PM risk estimates by obesity. Schwartz et al.
(2005, 086296) reported a change in HRV in obese (i.e., BMI > 30 kg/m2) compared to non-
obese subjects, while Dubowsky et al. (2006, 088750) observed an increase in inflammatory
markers (i.e., CRP, IL-6, and WBC) in response to short-term exposure to PM2 .5 among obese
individuals. Additionally, Schneider et al. (2008, 191985) found some evidence for a larger
reduction in FMD in individuals with a BMI >30 kg/m3 in response to PM2.5 exposure.
These effects could be due, in part, to a higher PM dose rate in obese individuals, which has
been demonstrated in children by Bennett and Zeman (2004, 155686). These investigators
also reported that tidal volume and resting minute ventilation increased with body mass
index. Although a limited amount of research has been conducted to examine PM-related
health effects in obese individuals there is an increasing trend of individuals within the
U.S. that have been defined as overweight or obese (BMI > 25.0) (56-65% between NHANES
III and NHANES [1999-2002]).
July 2009
8-20
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
8.1.7. Socioeconomic Status
Socioeconomic status (SES) is a composite measure that usually consists of economic
status, measured by income! social status measured by education! and work status
measured by occupation (Dutton and Levine, 1989, 192052). Based on data from the U.S.
Census Bureau in 2006, from among commonly-used indicators of SES, about 12% of
individuals and 11% of families are below the poverty line (U.S., 2009, 192117). Although
the measure of SES is composed of a multitude of determinants, each of these linked factors
can influence an individual's susceptibility to PM-related health effects. Additionally, low
SES individuals have been found to have a higher prevalence of pre-existing diseases!
inadequate medical treatment! and limited access to fresh foods leading to a reduced intake
of antioxidant polyunsaturated fatty acids and vitamins, which can increase this
subpopulations susceptibility to PM (Kan et al., 2008, 156621).
SES and individual determinants of SES, such as educational attainment, are not
mutually exclusive and together can influence the susceptibility of a population. Within the
U.S. approximately 16% of the population does not have a high school degree and only 27%
have a bachelor's degree or higher level of education (U.S., 2009, 192148). Educational
attainment generally coincides with an individual's income level, which is correlated to
other surrogates of SES, such as residential environment (Jerrett et al., 2004, 087379). Low
SES, and surrogates of SES such as educational attainment, have been shown in some
studies to modify health outcomes of PM exposure for a population. Franklin et al. (2008,
155779) noted an increased risk in mortality associated with short-term exposure to PM2.5
and its components for individuals with low SES while additional analyses stratified by
education level have also observed consistent trends of increased mortality for PM2.5 and
PM2.5 species for individuals with low educational attainment (Ostro et al., 2006, 087991!
Ostro et al., 2008, 097971! Zeka et al., 2006, 088749) This is further supported by a
reanalysis of the ACS cohort (Krewski et al., 2009, 191193), which found moderate evidence
for increased lung cancer mortality in individuals with a high school education or less
compared to individuals with more than a high school education in response to long-term
exposure to PM2.5.
Epidemiologic studies have also examined additional surrogates of SES, such as
residential location and nutritional status to identify their influence on the susceptibility of
a subpopulation. Jerrett et al. (2004, 087379) examined the modification of acute mortality
effects due to particulate air pollution exposure by residential location in Hamilton, Canada
using educational attainment as a surrogate for SES. The authors found that the area of
the city with the highest SES characteristics displayed no evidence of effect modification
while the area with the lowest SES characteristics had the largest health effects. Likewise,
Wilson et al. (2007, 157149) examined the effect of SES on the association between
mortality and short-term exposure to PM in Phoenix, but used educational attainment and
July 2009
8-21
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
income to represent SES. When stratifying Phoenix into central, middle, and outer rings of
varying urban density central Phoenix, the area with the lowest SES, was found to exhibit
the greatest association with PM2.5. However, the association with urban density differed
when examining PM10-2.5, with the greatest effect being observed for the middle ring.
Yanosky et al. (2008, 192081) examined whether long-term exposure to traffic-related
pollutants, using NO2 as a surrogate, varied by SES at the block group level. The authors
found higher levels of NO2 associated with lower SES areas, which suggests that lower SES
individuals are disproportionately exposed to traffic-related pollutants, which includes PM.
Nutritional deficiencies have been associated with increased susceptibility to a variety
of infectious diseases and chronic health effects. Low SES may decrease access to fresh
foods, and thus be related to nutritional deficiencies that could increase susceptibility to
PM-related health effects. Baccarelli et al. (2008, 157984) examined the association
between exposure to PM2.5 and HRV in individuals with polymorphisms in MTHFR and
cSHMT genes, which are associated with reduced enzyme activity and increased risk of
CVD. The authors found that when individuals with these genetic polymorphisms increased
their intake (above median levels) of B6, B12, or methionine no PM2.5 effect on HRV was
observed.
8.1.8. Summary
Upon evaluating the association between short- and long-term exposure to PM and various health
outcomes, studies also attempted to identify subpopulations that are more susceptible to PM. These
studies did so by: conducting stratified analyses; examining individuals with an underlying health
condition; or developing animal models that mimic the physiological conditions associated with an
adverse health effect. These studies identified a multitude of factors that could potentially contribute to
whether an individual is susceptible to PM (Table 8-2). Although studies have primarily used exposures
to PM10 or PM2 5, the available evidence suggests that the identified factors may also enhance
susceptibility to coarse fraction particles.
The majority of observations made during the evaluation of the literature reviewed in this ISA are
consistent with those reported in the 2004 PM AQCD. An evaluation of age-related health effects
suggests that older adults have heightened responses for cardiovascular morbidity with PM exposure. In
addition, epidemiologic and toxicological studies provide evidence, which indicates that children are at an
increased risk of PM-related respiratory effects. It should be noted that the health effects observed in
children could be initiated by exposures to PM that occurred during key windows of development, such as
in utero. Studies that focus on exposures during development have reported inconsistent findings (see
Section 7.4.), but a recent toxicological study suggests that inflammatory responses during pregnancy due
to exposure to PM could result in health effects in the developing fetus.
Epidemiologic studies have also examined whether additional factors, such as gender, race, or
ethnicity modify the association between PM and morbidity and mortality outcomes. Consistent with the
July 2009
8-22
DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
findings of the 2004 PM AQCD, gender and race do not seem to modify the association between PM and
morbidity and mortality outcomes. However, some evidence, albeit from two studies conducted in
California, suggest that Hispanic ethnicity may modify the association between PM and mortality.
Recent epidemiologic and toxicological studies provided evidence that individuals with null alleles
or polymorphisms in genes that mediate the antioxidant response to oxidative stress (i.e., GSTM1),
regulate enzyme activity (i.e., MTHFR and cSHMT), or regulate physiological levels of inflammatory
markers (i.e., fibrinogen) are more susceptible to PM exposure. However, some studies have shown that
polymorphisms in genes (e.g., HFE) can have a protective effect upon PM exposure. Additionally,
preliminary evidence suggests that PM exposure can impart epigenetic effects (i.e., DNA methylation),
however, these require further investigation.
Collectively, the evidence from epidemiologic and toxicological, and to a lesser extent, controlled
human exposure studies indicate increased susceptibility of individuals with underlying cardiovascular
diseases and respiratory illnesses, specifically asthma, to PM exposure. Additional controlled human
exposure and toxicological studies provide some evidence for increased PM-related cardiovascular effects
in individuals with underlying respiratory health conditions. However, the results are not consistent with
epidemiologic studies, resulting in the need for further investigation.
Recently studies have begun to examine the influence of preexisting chronic inflammatory
conditions, such as diabetes and obesity, on PM-related health effects. These studies have found some
evidence for increased associations for cardiovascular outcomes along with physiological alterations in
markers of inflammation, oxidative stress, and acute phase response. However more research is needed to
thoroughly examine the effect of PM exposure on obese individuals and to identify the biological
pathway(s) that could lead to increased susceptibility of diabetic and obese individuals to PM.
There is also evidence that socioeconomic status (SES), measured using surrogates such as
educational attainment or residential location, modifies the association between PM and morbidity and
mortality outcomes. In addition, nutritional status, another surrogate of SES, has been shown to have
protective effects against PM exposure in individuals that have a higher intake in some vitamins and
nutrients.
Overall, the epidemiologic, controlled human exposure, and toxicological studies evaluated in this
review provide evidence for increased susceptibility for various subpopulations. Although the level of
evidence varies depending on the factor being evaluated collectively, it can be concluded that some
subpopulations are more susceptible than the general population.
July 2009
8-23
DRAFT-DO NOT CITE OR QUOTE

-------
Chapter 8 References
Aday LA. (2001). At Risk in America- The Health and Health Care Needs of Vulnerable Populations in the
United States . San Francisco, CA- Jossey-Bass, Inc. 192150
American Lung Association. (2001). Urban air pollution and health inequities' A workshop report. Environ
Health Perspect, 3: 357-374. 016626
Anselme Fi Loriot S; Henry J-P; Dionnet Fi Napoleoni J-Gi Thnillez Ci Morin J-P. (2007). Inhalation of diluted
diesel engine emission impacts heart rate variability and arrhythmia occurrence in a rat model of chronic
ischemic heart failure. Arch Toxicol, 81: 299-307. 097084
Araujo JAi Barajas Bi KLeinman Mi Wang X; Bennett BJi Gong KWi Navab Mi Harkema Ji Sioutas Ci Lusis AJi
Nel AE. (2008). Ambient particulate pollutants in the ultrafine range promote early atherosclerosis and
systemic oxidative stress. Circ Res, 102: 589-596. 156222
Baccarelli Ai Martinelli Ii Pegoraro Vi Melly Si Grillo Pi Zanobetti Ai Hou Li Bertazzi PAi Mannucci PMi
Schwartz J. (2009). Living near Major Traffic Roads and Risk of Deep Vein Thrombosis. Circulation, x: x.
188183
Baccarelli Ai Martinelli Ii Zanobetti Ai Grillo Pi Hou LFi Bertazzi PAi Mannucci PMi Schwartz J. (2008).
Exposure to Particulate Air Pollution and Risk of Deep Vein Thrombosis. Arch Intern Med, 168: 920-927.
157984
Barnett AGi Williams GMi Schwartz Ji Best TLi Neller AHi Petroeschevsky ALi Simpson RW. (2006). The
effects of air pollution on hospitalizations for cardiovascular disease in elderly people in Australian and
New Zealand cities. Environ Health Perspect, 114: 1018-1023. 089770
Bartoli CRi Wellenius GAi Coull BAi Akiyama Ii Diaz EAi Lawrence Ji Okabe Ki Verrier RLi Godleski JJ.
(2009). Concentrated ambient particles alter myocardial blood flow during acute ischemia in conscious
canines. Environ Health Perspect, 117: 333-337. 179904
Bastain TMi Gilliland FDi Li Y"Fi Saxon Ai Diaz-Sanchez D. (2003). Intraindividual reproducibility of nasal
allergic responses to diesel exhaust particles indicates a susceptible phenotype. Clin Immunol, 109: 130-
136. 098690
Batalha JRi Saldiva P Hi Clarke RWi Coull BAi Stearns RCi Lawrence Ji Murthy GGi Koutrakis Pi Godleski JJ.
(2002). Concentrated ambient air particles induce vasoconstriction of small pulmonary arteries in rats.
Environ Health Perspect, 110: 1191-1197. 088109
Bateson TFi Schwartz J. (2004). Who is sensitive to the effects of particulate air pollution on mortality? A case-
crossover analysis of effect modifiers. Epidemiology, 15: 143-149. 086244
Bennett WDi Zeman KL. (2004). Effect of body size on breathing pattern and fine-particle deposition in children.
J Appl Physiol, 97: 821-826. 155686
Boezen HMi Vonk JMi Van Der Zee SCi Gerritsen Ji Hoek Gi Brunekreef Bi Schouten JPi Postma DS. (2005).
Susceptibility to air pollution in elderly males and females. Eur Respir J, 25: 1018-1024. 087396
Campen MJi Babu NSi Helms GAi Pett Si Wernly Ji Mehran Ri McDonald JD. (2005). Nonparticulate
components of diesel exhaust promote constriction in coronary arteries from ApoE -/- mice. Toxicol Sci,
88: 95-102. 083977
Campen MJi McDonald JDi Reed MDi Seagrave J. (2006). Fresh gasoline emissions, not paved road dust, alter
cardiac repolarization in ApoE-/- mice. Cardiovasc Toxicol, 6: 199-210. 096879
Carlsten Ci Kaufman JDi Trenga CAi Allen Ji Peretz Ai Sullivan JH. (2008). Thrombotic markers in metabolic
syndrome subjects exposed to diesel exhaust. Inhal Toxicol, 20: 917-921. 156323
CDC. (2008). Table 3-1: Current Asthma Population Estimates, in thousands by Age, United States: National
Health Interview Survey, 2006. Retrieved , from . 156324
CDC. (2008). Table 4-1: Current Asthma Prevalence Percents by Age, United States: National Health Interview
Survey, 2006. Retrieved , from . 156325
Chahine Ti Baccarelli Ai Litonjua Ai Wright ROi Suh Hi Gold DRi Sparrow Di Vokonas Pi Schwartz J. (2007).
Particulate air pollution, oxidative stress genes, and heart rate variability in an elderly cohort. Environ
Health Perspect, 115: 1617-1622. 156327
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health
and Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in
the process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009
8-24
DRAFT-DO NOT CITE OR QUOTE

-------
Chen LC; Hwang JS. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice
IV	Characterization of acute and chronic effects of ambient air fine particulate matter exposures on
heart-rate variability. Inhal Toxicol, 17: 209-216. 087218
Chen LC; Nadziejko C. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice
V	CAPs exacerbate aortic plaque development in hyperlipidemic mice. Inhal Toxicol, 17- 217-224. 087219
Chen LH; Knutsen SF; Shavlik D; Beeson WL; Petersen F; Ghamsary M; Abbey D. (2005). The association
between fatal coronary heart disease and ambient particulate air pollution- Are females at greater risk?.
Environ Health Perspect, 113: 1723-1729. 087942
De Leon SF; Thurston GD; Ito K. (2003). Contribution of respiratory disease to nonrespiratory mortality
associations with air pollution. Am J Respir Crit Care Med, 167: 1117-1123. 055688
Delfino RJ; Gone H; Linn WS; Pellizzari ED; Hu Y. (2003). Asthma symptoms in Hispanic children and daily
ambient exposures to toxic and criteria air pollutants. Environ Health Perspect, 111: 647-656. 050460
Delfino RJ; Zeiger RS; Seltzer JM; Street DH; McLaren CE. (2002). Association of asthma symptoms with peak
particulate air pollution and effect modification by anti"inflammatory medication use. Environ Health
Perspect, 110: A607-A617. 093740
Desqueyroux H; Pujet J-C; Prosper M; Squinazi F; Momas I. (2002). Short-term effects of low-level air pollution
on respiratory health of adults suffering from moderate to severe asthma. Environ Res, 89: 29"37. 026052
Devlin RB; Ghio AJ; Kehrl H; Sanders G; Cascio W. (2003). Elderly humans exposed to concentrated air
pollution particles have decreased heart rate variability. Eur Respir J, 40: 76S-80S.. 087348
Diaz-Sanchez D; Tsien A; Fleming J; Saxon A. (1997). Combined diesel exhaust particulate and ragweed
allergen challenge markedly enhances human in vivo nasal ragweed-specific IgE and skews cytokine
production to a T helper cell 2"type pattern. J Immunol, 158: 2406-2413. 051247
Dubowsky SD; Suh H; Schwartz J; Coull BA; Gold DR. (2006). Diabetes, obesity, and hypertension may enhance
associations between air pollution and markers of systemic inflammation. Environ Health Perspect, 114:
992-998. 088750
Dutton DB; Levine S. (1989). Overview, methodological critique, and reformulation. In JP Bunker; DS Gomby;
BH Kehrer (Ed.),Pathways to health (pp. 29"69). Menlo Park, CA: The Henry J. Kaiser Family
Foundation. 192052
Fedulov AV; Leme A; Yang Z; Dahl M; Lim R; Mariani TJ; Kobzik L. (2008). Pulmonary exposure to particles
during pregnancy causes increased neonatal asthma susceptibility. , 38: 57"67. 097482
Firestone M; Moya J; Cohen-Hubal E; Zartarian V; Xue J. (2007). Identifying childhood age groups for exposure
assessments and monitoring. Risk Anal, 27: 701-714. 192071
Frampton MW; Stewart JC; Oberdorster G; Morrow PE; Chalupa D; Pietropaoli AP; Frasier LM; Speers DM;
Cox C; Huang LS; Utell MJ. (2006). Inhalation of ultrafine particles alters blood leukocyte expression of
adhesion molecules in humans. Environ Health Perspect, 114: 51-58. 088665
Franklin M; Koutrakis P; Schwartz J. (2008). The role of particle composition on the association between PM2.5
and mortality. Epidemiology, 19: 680-689. 155779
Franklin M; Zeka A; Schwartz J. (2007). Association between PM2.5 and all-cause and specific-cause mortality
in 27 US communities. , 17: 279-87. 188502
Fung KY; Luginaah I; Gorey KM; Webster G. (2005). Air pollution and daily hospital admissions for
cardiovascular diseases in Windsor, Ontario. Can J Public Health, 96: 29"33. 074322
Fung KY; Luginaah IKMG; Webster G. (2005). Air pollution and daily hospitalization rates for cardiovascular
and respiratory diseases in London, Ontario. Int J Environ Stud, 62: 677-685. 093262
Gent JF; Triche EW; Holford TR; Belanger K; Bracken MB; Beckett WS; Leaderer BP. (2003). Association of
low-level ozone and fine particles with respiratory symptoms in children with asthma. JAMA, 290: 1859"
1867. 052885
Gilliland FD; Li YF; Saxon A; Diaz-Sanchez D. (2004). Effect of glutathione-S-transferase Ml and PI genotypes
on xenobiotic enhancement of allergic responses: randomised, placebo-controlled crossover study. Lancet,
363: 119-125. 156471
Gilliland FD; McConnell R; Peters J; Gong Jr H. (1999). A Theoretical Basis for Investigating Ambient Air
Pollution and Children's Respiratory Health. Environ Health Perspect, 107: 403-407. 155792
Goldberg MS; Burnett RT; Yale JF; Valois MF; Brook JR. (2006). Associations between ambient air pollution
and daily mortality among persons with diabetes and cardiovascular disease. Environ Res, 100: 255-267.
088641
Gong H Jr; Linn WS; Terrell SL; Clark KW; Geller MD; Anderson KR; Cascio WE; Sioutas C. (2004). Altered
heart-rate variability in asthmatic and healthy volunteers exposed to concentrated ambient coarse
particles. Inhal Toxicol, 16: 335-343. 055628
July 2009
8-25
DRAFT-DO NOT CITE OR QUOTE

-------
Gong H Jr; Linn WS; Clark KW; Anderson KR; Sioutas C; Alexis NE; Cascio WE; Devlin RB. (2008). Exposures
of healthy and asthmatic volunteers to concentrated ambient ultrafine particles in Los Angeles. Inhal
Toxicol, 20: 533-545. 156483
Gong H Jri Linn WS; Sioutas C', Terrell SL; Clark KW; Anderson KR; Terrell LL. (2003). Controlled exposures of
healthy and asthmatic volunteers to concentrated ambient fine particles in Los Angeles. Inhal Toxicol,
15: 305-325. 042106
Heidenfelder BL; Reif DM; Harkema JR; Cohen Hubal EA; Hudgens EE; Bramble LA; Wagner JG; Morishita M;
Keeler GJ; Edwards SW; Gallagher JE. (2009). Comparative microarray analysis and pulmonary changes
in brown norway rats exposed to ovalbumin and concentrated air particulates. Toxicol Sci, 108- 207-221.
190026
Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V;
Prouvost H; Lefranc A. (2007). Short-term Associations between Fine and Coarse Particles and
Cardiorespiratory Hospitalizations in Six French Cities. Occup Environ Med. 155851
Jerrett M; Burnett RT; Brook J; Kanaroglou P; Giovis C; Finkelstein N; Hutchison B. (2004). Do socioeconomic
characteristics modify the short term association between air pollution and mortality? Evidence from a
zonal time series in Hamilton, Canada. J Epidemiol Community Health, 58- 31-40. 087379
Kan H; London SJ; Chen G; Zhang Y; Song G; Zhao N; Jiang L; Chen B. (2008). Season, sex, age, and education
as modifiers of the effects of outdoor air pollution on daily mortality in Shanghai, China- The Public
Health and Air Pollution in Asia (PAPA) Study. Environ Health Perspect, 116: 1183-1188. 156621
Kauffmann F; Post Genome Respiratory Epidemiology Group. (2004). Post-genome respiratory epidemiology: a
multidisciplinary challenge. Eur Respir J, 24: 471-480. 090968
Kleeberger SR; Ohtsuka Y. (2005). Gene-particulate matter-health interactions. Toxicol Appl Pharmacol, 207:
S276-S281. 130489
Kodavanti UP; Moyer CF; Ledbetter AD; Schladweiler MC; Costa DL; Hauser R; Christiani DC; Nyska A.
(2003). Inhaled environmental combustion particles cause myocardial injury in the Wistar Kyoto rat.
Toxicol Sci, 71: 237-245. 051325
Krewski D; Jerrett M; Burnett RT; Ma R; Hughes E; Shi Y; Turner MC; Pope AC IIL Thurston G; Calle EE;
Thun MJ. (2009). Extended follow-up and spatial analysis of the American Cancer Society study linking
particulate air pollution and mortality. Health Effects Institute. Cambridge, MA. 140. 191193
Lagorio S; Forastiere F; Pistelli R; Iavarone I; Michelozzi P; Fano V; Marconi A; Ziemacki G; Ostro BD. (2006).
Air pollution and lung function among susceptible adult subjects: a panel study. ,5: 11. 089800
Larrieu S; Jusot J-F; Blanchard M; Prouvost H; Declercq C; Fabre P; Pascal L; Le Tertre A; Wagner V; Riviere
S; Chardon B; Borelli D; Cassadou S; Eilstein D; Lefranc A. (2007). Short term effects of air pollution on
hospitalizations for cardiovascular diseases in eight French cities: The PSAS program. Sci Total Environ,
387: 105-112. 093031
Le Tertre A; Medina S; Samoli E; Forsberg B; Michelozzi P; Boumghar A; Vonk JM; Bellini A; Atkinson R; Ayres
JG; Sunyer J; Schwartz J; Katsouyanni K. (2002). Short term effects of particulate air pollution on
cardiovascular diseases in eight European cities. J Epidemiol Community Health, 56: 773-779. 023746
Lee IM; Tsai SS; Ho CK; Chiu HF; Wu TN; Yang CY. (2008). Air pollution and hospital admissions for
congestive heart failure: are there potentially sensitive groups?. Environ Res, 108: 348-353. 192076
Liao D; Heiss G; Chinchilli VM; Duan Y; Folsom AR; Lin HM; Salomaa V. (2005). Association of criteria
pollutants with plasma hemostatic/inflammatory markers: a population-based study. J Expo Sci Environ
Epidemiol, 15: 319-328. 088677
Liu L; Ruddy TD; Dalipaj M; Szyszkowicz M; You H; Poon R; Wheeler A; Dales R. (2007). Influence of personal
exposure to particulate air pollution on cardiovascular physiology and biomarkers of inflammation and
oxidative stress in subjects with diabetes. J Occup Environ Med, 49: 258-265. 156705
Lopes FD; Pinto TS; Arantes-Costa FM; Moriya HT; Biselli PJ; Ferraz LF; Lichtenfels AJ; Saldiva PH; Mauad
T; Martins MA. (2009). Exposure to ambient levels of particles emitted by traffic worsens emphysema in
mice. Environ Res, 109: 544-551. 190430
Luginaah IN; Fung KY; Gorey KM; Webster G; Wills C. (2005). Association of ambient air pollution with
respiratory hospitalization in a government designated "area of concern": the case of Windsor, Ontario.
Environ Health Perspect, 113: 290-296. 057327
Mar TF; Larson TV; Stier RA; Claiborn C; Koenig JQ. (2004). An analysis of the association between respiratory
symptoms in subjects with asthma and daily air pollution in Spokane, Washington. Inhal Toxicol, 16:
809-815. 057309
Mauad T; Rivero DH; de Oliveira RC; Lichtenfels AJ; Guimaraes ET; de Andre PA; Kasahara DI; Bueno HM;
Saldiva PH. (2008). Chronic exposure to ambient levels of urban particles affects mouse lung
development. Am J Respir Crit Care Med, 178: 721-728. 156743
July 2009
8-26
DRAFT-DO NOT CITE OR QUOTE

-------
Merriam-Webster. (2009). Merriam-Webster on-line- httpV/www.merriam-webster.com/medical/sussceptible,
httpV/merriam-webster.com/medical/vulerable. Retrieved 15-JUN-09, from httpV/www.merrian-
webster.com/medical/. 192146
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd KH; Mulholland JA; Ryan PB; Frumkin H.
(2004). Ambient air pollution and cardiovascular emergency department visits. Epidemiology, 15- 46"56.
044222
Middleton N; Yiallouros P; Kleanthous S; Kolokotroni O; Schwartz J; Dockery DW! Demokritou P; Koutrakis P.
(2008).	A 10-year time-series analysis of respiratory and cardiovascular morbidity in Nicosia, Cyprus' the
effect of short-term changes in air pollution and dust storms. , 7- 39. 156760
Mills NL; Robinson SD; Fokkens PH; Leseman DL; Miller MR; Anderson D; Freney EJ; Heal MR; Donovan RJ;
Blomberg A; Sandstrom T; MacNee W; Boon NA; Donaldson K; Newby DE; Cassee FR. (2008). Exposure
to concentrated ambient particles does not affect vascular function in patients with coronary heart
disease. Environ Health Perspect, 116- 709-715. 156766
Mills NL; Tornqvist H; Gonzalez MC; Vink E; Robinson SD; Soderberg S; Boon NA; Donaldson K; Sandstrom T;
Blomberg A; Newby DE. (2007). Ischemic and thrombotic effects of dilute diesel-exhaust inhalation in
men with coronary heart disease. , 357- 1075-1082. 091206
Morishita M; Keeler G; Wagner J; Marsik F; Timm E; Dvonch J; Harkema J. (2004). Pulmonary retention of
particulate matter is associated with airway inflammation in allergic rats exposed to air pollution in
urban Detroit. Inhal Toxicol, 16- 663-674. 087979
Mortimer KM; Neas LM; Dockery DW; Redline S; Tager IB. (2002). The effect of air pollution on inner-city
children with asthma. Eur Respir J, 19- 699-705. 030281
Nadziejko C; Fang K; Narciso S; Zhong M; Su WC; Gordon T; Nadas A; Chen LC. (2004). Effect of particulate
and gaseous pollutants on spontaneous arrhythmias in aged rats. Inhal Toxicol, 16- 373-380. 055632
Naess O; Nafstad P; Aamodt G; Claussen B; Rosland P. (2007). Relation between concentration of air pollution
and cause-specific mortality- four-year exposures to nitrogen dioxide and particulate matter pollutants in
470 neighborhoods in Oslo, Norway. Am J Epidemiol, 165- 435-443. 090736
Nordenhall C; Pourazar J; Ledin M"C; Levin J-O; Sandstrom T; Adelroth E. (2001). Diesel exhaust enhances
airway responsiveness in asthmatic subjects. Eur Respir J, 17- 909-915. 025185
NRC. (2004). Research priorities for airborne particulate matter- IV Continuing research progress. National
Academies Press . Washington, DC. 156814
Ohtsuka Y; Brunson KJ; Jedlicka AE; Mitzner W; Clarke RW; Zhang L"Y; Eleff SM; KLeeberger SR. (2000).
Genetic linkage analysis of susceptibility to particle exposure in mice. Am J Respir Cell Mol Biol, 22- 574"
581. 004409
Ostro B; Broadwin R; Green S; Feng W"Y; Lipsett M. (2006). Fine particulate air pollution and mortality in nine
California counties: results from CALFINE. Environ Health Perspect, 114: 29"33. 087991
Ostro BD; Feng WY; Broadwin R; Malig BJ; Green RS; Lipsett MJ. (2008). The impact of components of fine
particulate matter on cardiovascular mortality in susceptible subpopulations. Occup Environ Med, 65:
750-756. 097971
Park SK,O'Neill MS,Wright RO,Hu H,Vokonas PS,Sparrow D,Suh H,Schwartz J. (2006). HFE Genotype,
Particulate Air Pollution, and Heart Rate Variability: A Gene-Environment Interaction. Circulation, 114:
2798-2805. 189099
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Schwartz J. (2005). Effects of air pollution on heart rate
variability: The VA normative aging study. Environ Health Perspect, 113: 304-309. 057331
Peel JL; Metzger KB; Klein M; Flanders WD; Mulholland JA; Tolbert PE. (2007). Ambient air pollution and
cardiovascular emergency department visits in potentially sensitive groups. Am J Epidemiol, 165: 625"
633. 090442
Peel JL; Tolbert PE; Klein M; Metzger KB; Flanders WD; Knox T; Mulholland JA; Ryan PB; Frumkin H. (2005).
Ambient air pollution and respiratory emergency department visits. Epidemiology, 16: 164-174. 056305
Peters A; Greven S; Heid I; Baldari F; Breitner S; Bellander T; Chrysohoou C; Illig T; Jacquemin B; Koenig W.
(2009).	Fibrinogen Genes Modify the Fibrinogen Response to Ambient Particulate Matter. Am J Respir
Crit Care Med, 179: 484-491. 191992
Pinkerton KE; Zhou Y; Teague SV; Peake JL; Walther RC; Kennedy IM; Leppert VJ; Aust AE. (2004). Reduced
lung cell proliferation following short-term exposure to ultrafine soot and iron particles in neonatal rats:
key to impaired lung growth?. Inhal Toxicol, l: 73-81. 087465
Pinkerton KE; Zhou Y; Zhong C; Smith KR; Teague SV; Kennedy IM; Menache MG. (2008). Mechanisms of
particulate matter toxicity in neonatal and young adult rat lungs. Health Effects Institute. Boston, MA.
135. 190471
Pleis JR; Lethbridge-Cejku M. (2007). Summary health statistics for US adults: National Health Interview
Survey, 2006. CDC. Maryland. 156875
July 2009
8-27
DRAFT-DO NOT CITE OR QUOTE

-------
Pope C> Renlund D; Kfoury A; May H; Home B. (2008). Relation of heart failure hospitalization to exposure to
fine particulate air pollution. , 102- 1230-1234. 191969
Pope CA, 3rd,Muhlestein JB,May HT,Renlund DG,Anderson JL,Horne BD. (2006). Ischemic heart disease
events triggered by short-term exposure to fine particulate air pollution. , 114- 2443"8. 189048
Pope CA; Dockery DW. (2006). Health effects of fine particulate air pollution- Lines that connect. J Air Waste
Manag Assoc, 56: 709-742. 156881
Pope CA IIL Muhlestein JB; May HT; Renlund DG; Anderson JL; Home BD. (2006). Ischemic heart disease
events triggered by short-term exposure to fine particulate air pollution. Circulation, 114: 2443-2448.
091246
Rabinovitch N; Strand M; Gelfand EW. (2006). Particulate levels are associated with early asthma worsening in
children with persistent disease. Am J Respir Crit Care Med, 173: 1098-1105. 088031
Routledge HC; Manney S; Harrison RM; Ayres JG; Townend JN. (2006). Effect of inhaled sulphur dioxide and
carbon particles on heart rate variability and markers of inflammation and coagulation in human
subjects. Heart, 92: 220-227. 088674
Samoli E; Peng R; Ramsay T; Pipikou M; Touloumi G; Dominici F; Burnett R; Cohen A; Krewski D; Samet J.
(2008). Acute effects of ambient particulate matter on mortality in Europe and North America: results
from the APHENA study. Environ Health Perspect, 116: 1480. 188455
Schneider A; Neas L; Herbst M; Case M; Williams R; Cascio W', Hinderliter A; Holguin F; Buse J; Dungan K.
(2008). Endothelial dysfunction: associations with exposure to ambient fine particles in diabetic
individuals. Environ Health Perspect, 116: 1666. 191985
Schwartz J; Park SK; O'Neill MS; Vokonas PS; Sparrow D; Weiss S; Kelsey K. (2005). Glutathione-S-transferase
Ml, obesity, statins, and autonomic effects of particles: gene-by-drug-by-environment interaction. Am J
Respir Crit Care Med, 172: 1529-1533. 086296
Silkoff PE; Zhang L; Dutton S; Langmack EL; Vedal S; Murphy J; Make B. (2005). Winter air pollution and
disease parameters in advanced chronic obstructive pulmonary disease panels residing in Denver,
Colorado. J Allergy Clin Immunol, 115: 337-344. 087471
Suh YJ; Ha EH; Park H; Kim YJ; Kim H; Hong YC. (2008). GSTM1 polymorphism along with PM10 exposure
contributes to the risk of preterm delivery. , 656: 62"67. 192077
Sun Q; Wang A; Jin X; Natanzon A; Duquaine D; Brook RD; Aguinaldo J-GS; Fayad ZA; Fuster V; Lippmann M;
Chen Lung C; Rajagopalan S. (2005). Long-term air pollution exposure and acceleration of
atherosclerosis and vascular inflammation in an animal model. , 294: 3003-3010. 186814
Sun Q; Yue P; Deiuliis JA; Lumeng CN; Kampfrath T; Mikolaj MB; Cai Y; Ostrowski MC; Lu B; Parthasarathy
S; Brook RD; Moffatt-Bruce SD; Chen LC; Rajagopalan S. (2009). Ambient air pollution exaggerates
adipose inflammation and insulin resistance in a mouse model of diet-induced obesity. Circulation, 119:
538-546. 190487
Sun Q; Yue P; Kirk RI; Wang A; Moatti D; Jin X; Lu B; Schecter AD; Lippmann M; Gordon T; Chen LC;
Rajagopalan S. (2008). Ambient air particulate matter exposure and tissue factor expression in
atherosclerosis. Inhal Toxicol, 20: 127-137. 157033
Tankersley CG,Bierman A,Rabold R. (2007). Variation in heart rate regulation and the effects of particle
exposure in inbred mice. Inhal Toxicol, 19: 621-9. 188859
Tankersley CG; Bierman A; Rabold R. (2007). Variation in heart rate regulation and the effects of particle
exposure in inbred mice. , 19: 621-9. 097910
Tankersley CG; Campen M; Bierman A; Flanders SE; Broman KW; Rabold R. (2004). Particle effects on heart-
rate regulation in senescent mice. Inhal Toxicol, 16: 381-390. 094378
Tankersley CG; Champion HC; Takimoto E; Gabrielson K; Bedja D; Misra V; El-Haddad H; Rabold R; Mitzner
W. (2008). Exposure to inhaled particulate matter impairs cardiac function in senescent mice. Am J
Physiol Regul Integr Comp Physiol, 295: R252-263. 157043
Tarantini L; Bonzini M; Apostoli P; Pegoraro V; Bollati V; Marinelli B; Cantone L; Rizzo G; Hou L; Schwartz J;
Bertazzi PA; Baccarelli A. (2009). Effects of particulate matter on genomic DNA methylation content and
iNOS promoter methylation. Environ Health Perspect, 117: 217-222. 192010
Trenga CA; Sullivan JH; Schildcrout JS; Shepherd KP; Shapiro GG; Liu LJ; Kaufman JD; Koenig JQ. (2006).
Effect of particulate air pollution on lung function in adult and pediatric subjects in a Seattle panel
study. , 129: 1614-22. 155209
U.S. Census. (2009). Educational Attainment in the United States: 2007
(http://www.census.gov/prod/2009pubs/p20-560.pdf). U.S. Cenus Bureau, Economics and Statistics
Administration, U.S. Department of Commerce. Washington, DC. P20-560.
http://www.census.gov/prod/2009pubs/p20-560.pdf. 192148
U.S. Census. (2009). U.S. Census (2000)Poverty. Retrieved 01-JUL-09, from
http://www.census.gov/hhes/www.porvety/hispov/hstpov2.html. 192147
July 2009
8-28
DRAFT-DO NOT CITE OR QUOTE

-------
U.S. EPA. (2003). Framework for cumulative risk assessment. National Center for Environmental Assessment,
Office of Research and Development, U.S. Environmental Protection Agency. Washington, DC.
EPA/630/P-02/001F. 192145
U.S. EPA. (2009). Susceptible subpopulations. Retrieved 15"MAY"09, from
http '//www. epa.gov/nerl/goals/health/populations. html. 192149
U.S. Census Bureau PIO. (2000). Census Bureau proiects doubling of nation's population by 2100. Retrieved ,
from . 157064
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA/600/P-99/002aF-bF. 056905
U.S. EPA. (2006). Aging and Toxic Response: I ssues Relevant to Risk Assessment (Final). U.S. Environmental
Protection Agency. Washington, DC.httpV/cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=156648. 192082
U.S. EPA. (2008). Integrated review plan for the national ambient air quality standards for particulate matter.
U.S. Environmental Protection Agency, Office of Research and Development, National Center for
Environmental Assessment. Research Triangle Park, NC. 157072
U.S. EPA. (2008). Integrated Science Assessment for Sulfur Oxides?Health Criteria. U.S. Environmental
Protection Agency. Research Triangle Park, NC. 157075
Villeneuve PJ; Goldberg MS; Krewski D; Burnett RT; Chen Y. (2002). Fine particulate air pollution and all-
cause mortality within the Harvard six-cities study- variations in risk by period of exposure. Ann
Epidemiol, 12: 568-576. 042576
Wellenius GA; Bateson TF; Mittleman MA; Schwartz J. (2005). Particulate air pollution and the rate of
hospitalization for congestive heart failure among medicare beneficiaries in Pittsburgh, Pennsylvania.
Am J Epidemiol, 161: 1030-1036. 087483
Wellenius GA; Coull BA; Batalha JRF; Diaz EA; Lawrence J; Godleski JJ. (2006). Effects of ambient particles
and carbon monoxide on supraventricular arrhythmias in a rat model of myocardial infarction. Inhal
Toxicol, 18: 1077-1082. 156152
Wellenius GA; Coull BA; Godleski JJ; Koutrakis P; Okabe K; Savage ST. (2003). Inhalation of concentrated
ambient air particles exacerbates myocardial ischemia in conscious dogs. Environ Health Perspect, 111:
402-408. 055691
Wellenius GA; Schwartz J; Mittleman MA. (2006). Particulate air pollution and hospital admissions for
congestive heart failure in seven United States cities. Am J Cardiol, 97: 404-408. 088748
Wilson WE; Mar TF; Koenig JQ. (2007). Influence of exposure error and effect modification by socioeconomic
status on the association of acute cardiovascular mortality with particulate matter in Phoenix. J Expo Sci
Environ Epidemiol, 17: S11-S19. 157149
Yanosky JD; Schwartz J; Suh HH. (2008). Associations between measures of socioeconomic position and chronic
nitrogen dioxide exposure in Worcester, Massachusetts. , 71: 1593-1602. 192081
Yauk C; Polyzos A; Rowan-Carroll A; Somers CM; Godschalk RW; Van Schooten FJ; Berndt ML; Pogribny IP;
Koturbash I; Williams A; Douglas GR; Kovalchuk O. (2008). Germ-line mutations, DNA damage, and
global hypermethylation in mice exposed to particulate air pollution in an urban/industrial location. ,
105: 605-610. 157164
Yeatts K; Svendsen E; Creason J; Alexis N; Herbst M; Scott J; Kupper L; Williams R; Neas L; Cascio W', Devlin
RB; Peden DB. (2007). Coarse particulate matter (PM25-10) affects heart rate variability, blood lipids,
and circulating eosinophils in adults with asthma. Environ Health Perspect, 115: 709-714. 091266
Yunginger JW; Reed CE; O'Connell EJ; Melton LJ 3rd; O'Fallon WM; Silverstein MD. (1992). A community-
based study of the epidemiology of asthma. Incidence rates, 1964-1983. , 146: 888-894. 192074
Zanobetti A; Bind MAC; Schwartz J. (2008). Particulate air pollution and survival in a COPD cohort. Environ
Health Perspect, 7: 48. 156177
Zanobetti A; Schwartz J. (2002). Cardiovascular damage by airborne particles: are diabetics more susceptible?.
Epidemiology, 13: 588-592. 034821
Zanobetti A; Schwartz J. (2005). The effect of particulate air pollution on emergency admissions for myocardial
infarction: Amulticity case-crossover analysis. Environ Health Perspect, 113: 978-982. 088069
Zeger S; Dominici F; McDermott A; Samet J. (2008). Mortality in the Medicare population and chronic exposure
to fine particulate air pollution in urban centers (2000-2005). Environ Health Perspect, 116: 1614-1619.
191951
Zeka A; Zanobetti A; Schwartz J. (2006). Individual-level modifiers of the effects of particulate matter on daily
mortality. Am J Epidemiol, 163: 849-859. 088749
July 2009
8-29
DRAFT-DO NOT CITE OR QUOTE

-------
Chapter 9. Welfare Effects
9.1. Introduction
This chapter is a synthesis and evaluation of the most policy-relevant science used to
help form the scientific foundation for review of the secondary (welfare-based) NAAQS
aimed at protecting against welfare effects of ambient airborne PM. Specifically Chapter 9
assesses the effects of atmospheric PM on the environment, including: (a) effects on
visibility! (b) effects on climate! (c) ecological effects! and (d) effects on materials. These
sections initially highlight the conclusions from the 2004 PM AQCD (U.S. EPA, 2004,
056905). followed by an evaluation of recent publications and assessment of the expanded
body of evidence. In some sections, few new publications are available, and the discussion is
primarily a brief overview of the key conclusions from the previous review.
As discussed in Chapter 1, the effects of particulate NOx and SOx have recently been
evaluated in the ISA for Oxides of Nitrogen and Sulfur - Ecological Criteria (U.S. EPA,
2008, 157074). That ISA focused on the effects from deposition of gas- and particle-phase
pollutants related to ambient NOx and SOx concentrations that can lead to acidification
and nutrient enrichment, as well as on the potential for increased mercury methylation
from SO r deposition. Thus, emphasis in this document is placed on the effects of airborne
PM on visibility and climate, and on the deposition effects of PM constituents other than
NOx and SOx, primarily metals and carbonaceous compounds.
Chapter 2 of this assessment provides an integrative overview of the major welfare
effects evaluated. EPA's framework for causality, described in Chapter 1, is applied
throughout the evaluation and the causal determinations are highlighted.
Note- Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health
and Environmental Research Online) at http7/epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in
the process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009
9-1
DRAFT-DO NOT CITE OR QUOTE

-------
9.2. Effects on Visibility
9.2.1. Introduction
In recent years, most visibility research involved characterizing visibility conditions
and trends over broad regional scales, improving the understanding of the atmospheric
processes and pollutants responsible for the regional impacts, and attribution of visibility-
impairing pollutants to emission sources, source types, and regions. The motivation for
much of this work has come from the visibility protection provisions of the 1977 Clean Air
Act Amendments (CAAA) that called for the development of regulations to address
reduction of regional haze in 156 national parks and wilderness areas to natural conditions,
and from the subsequent RHR (RHR) promulgated in 1999 by EPA in response to the CAA
mandate. Implementation of the RHR entails planned emissions reductions to reach
natural haze conditions in these protected areas by 2064 in six ten-yr planning steps.
Haze conditions caused solely by PM from natural sources are generally much lower
than contemporary conditions. The largest difference is between natural and current
conditions for the inorganic salts ammonium sulfate and ammonium nitrate, with natural
concentrations taken to be just a few tenths of a |u,g/m3 each (Trijonis et al., 1990, 157058),
while current conditions of both over large regions of the country are an order of magnitude
or more larger (DeBell, 2006, 156388). However, natural source PM can be substantial on
an episodic basis for crustal mineral PM components during high windblown dust
conditions and for carbonaceous PM from biogenic combustion during wildfire and
prescribed burning episodes. The need for information to generate RHR implementation
plans has resulted in extensive use of continental-scale air quality simulation modeling and
assessment of expanded ambient monitoring data sets.
Unlike the substantial remote-area visibility investigations that have been conducted
in response to the RHR, relatively little work on urban visibility effects has been done in
recent years. For example, there has been relatively little new research on the optical and
human perceptual aspects of atmospheric visibility over the last decade or more. These
topics have been the subjects of numerous earlier investigations that have been
summarized in detail elsewhere (U.S. EPA, 1979, 157065; Latimer and Ireson, 1980,
035723; Middleton, 1952, 016324; Tombach and McDonald, 2004, 157054; Trijonis et al.,
July 2009
9-2
DRAFT-DO NOT CITE OR QUOTE

-------
1990, 157058; Watson et al., 2002, 035623), including past criteria documents on PM, SO2
and NOx (U.S. EPA, 1982, 017610; U.S. EPA, 1993, 017649; U.S. EPA, 2004, 056905).
In spite of this fact, the understanding of urban visibility conditions has continued to
improve. By applying a well established algorithm that relates PM and haze conditions to
data currently collected from routine filter-based PM chemical speciation monitors located
in numerous urban areas (Jayanty, 2003, 156605). and to data collected from the more
recently deployed high time- and size-resolved PM speciation monitors located in several
cities such as those in the PM Supersites program (Solomon and Hopke, 2008, 156997)
urban visibility conditions can be better characterized. Comparisons between urban and
remote area data in the same region afford the opportunity to differentiate between
regional and local visibility impacts. The availability of better size and time resolution PM
composition data, compared to that available from the routine monitoring programs,
reduces the number of simplifying assumptions required to estimate visibility conditions in
these areas, thereby reducing the uncertainty of the estimates. Thus, the state of the
science supporting urban visibility assessments continues to improve.
The background section below contains an overview of long-available information to
help provide context to the more recently published literature summarized in subsequent
sections.
9.2.2. Background
Air pollution-induced visibility impairment is caused by the loss of image-forming
light (i.e., signal) and the addition of non-image forming light (i.e., noise) between an
observer and the object being viewed. These changes to the light reaching the observer are
a result of light being scattered and absorbed by particles and gases in the sight path (see
the schematic in Figure 9-1). Electromagnetic theory developed to characterize the
interaction of light with matter (Mie, 1908, 155983) permits the calculation of light
scattering and absorption by particles and gas molecules where the index of refraction and
shape of particles by size are known (Van de Hulst, 1981, 191972).
The ability of human observers to visually detect distant objects or identify changes in
their appearance depends on the apparent contrast of the object against its background.
The apparent contrast is affected by changes in the transparency of the atmosphere caused
by air pollution as well as factors not related to air quality such as length of the sight path,
scenic lighting and the physical characteristics of the viewed object and other elements of
July 2009
9-3
DRAFT-DO NOT CITE OR QUOTE

-------
the scene. To rigorously determine the perceived visual effects of changes in the optical
properties of the atmosphere requires the use of radiative transfer modeling to determine
changes in light from the field of view experienced by the observer, followed by the use of
psychophysical modeling to determine the response to the light by the eye-brain system.
The complexity of such an approach discourages its common use.
Atmospheric light extinction is a fundamental atmospheric optics metric used to
characterize air pollution impacts on visibility. It is the fractional loss of intensity in a light
beam per unit distance due to scattering and absorption by the gases and particles in the
air. Light extinction (bext) can be expressed as the sum of light scattering by particles (/)-,/,),
scattering by gases (bs,g), absorption by particles (hn,) and absorption by gases (baJ. Light
extinction and its components are expressed in units of inverse length, typically either
inverse kilometers (km1) or, as will be the convention in this document, inverse
megameters (Mm1). Traditionally, for visibility-protection applications, the most sensitive
portion of the spectrum for human vision (550 11m) has been used to characterize light
extinction and its components.
July 2009
9-4
DRAFT-DO NOT CITE OR QUOTE

-------
•	Sunlight (Sun Angla)
•	Cloud Cover (Overcast, Puffy, etc.)
•	Skv
Image-forming
•	Sunlight (Sun Angle)
•	Cloud Cover (Overcast,
•	Skv	-A
Image-forming
Image-forming
light absorbed
Optical Characteristics of Illumination
Characteristics of Observer
Optical Characteristics of
tlcal Characteristics of
Intervenl
Viewed Target
•	Color
•	Contrast Detail (Texture)
•	Form
•	Brightness
•	Detection Thresholds
•	Psychological Response to
Incoming Light
•	Value Judgements
Light from clouds
scattered Into
sight path y
•	Light Added to Sight Path by
Particles and Gases
•	Image-Forming Light Subtracted
from Sight Path by Scattering
and Absorption
light scattered /
out of sight path
si
Sunlight J0
scattered Ught ref)ected
from ground
scattered into
Source: Malm (1999, 026037).
Figure 9-1. Important factors involved in seeing a scenic vista are outlined. Image-forming
information from an object is reduced (scattered and absorbed) as it passes through the
atmosphere to the human observer. Air light is also added to the sight path by
scattering processes. Sunlight, light from clouds, and ground-reflected light all impinge
on and scatter from particulates located in the sight path. Some of this scattered light
remains in the sight path, and at times it can become so bright that the image
essentially disappears. A final important factor in seeing and appreciating a scenic vista
are the characteristics of the human observer.
A parametric analysis has shown that a constant fractional change in light extinction
results in a similar perceptual change regardless of baseline conditions (Pitchford et al.,
1990, 156871). From this assessment, the deciview haze index, which is a log
transformation of light extinction, similar in many ways to the decibel index for acoustic
measurements, was developed (Pitchford and Malm, 1994, 044922). A one deciview (ldv)
change is about a 10% change in light extinction, which is a small change that is detectable
for sensitive viewing situations. The haze index in deciview units is an appropriate metric
for expressing the extent of haze changes where the perceptibility of the change is an issue.
The RHR has adopted the deciview haze index as the metric for tracking long-term haze
July 2009	9-5	DRAFT - DO NOT CITE OR QUOTE

-------
trends of visibility-protected federal lands (EPA, 2001, 157068). Light extinction and its
components are more useful metrics for characterizing the apportionment of haze to its
pollutant components due to the approximately linear relationship between pollutant
species concentrations and their contributions to light extinction.
July 2009
9-6
DRAFT-DO NOT CITE OR QUOTE

-------
starlight
absorbed
moonlight
scattering
by particles
starlight
scattered
urban light pollution
particle
forward
scatter
wildlife
affected by
unnaturally
bright
nighttime
sky
observer in a non-urban setting
reflective cloud
reflective cloud
starlight
absorbed
particle
side
scatter
moonlight f
scattering
by panicles
starlight
scattered I
bright urban nighttime sky
/ particle
forward
scatter
Figure 9-2. Schematic of remote-area (top) and urban (bottom) nighttime sky visibility showing the
effects of PM and light pollution.
Daytime visibility has dominated the attention of those who have studied the
visibility effects of air pollution, though nighttime visibility is also known to be affected by
air pollution. Stargazing is a popular human activity in urban and remote settings. The
July 2009
9-7
DRAFT-DO NOT CITE OR QUOTE

-------
reduction in visibility of the night sky is primarily dependent on the addition of light into
the sight path, the brightness of the night sky and the reduction in contrast of stars against
the background (see the schematic in Figure 9-2). These are controlled by the addition of
PM, which enhances scattering, and the addition of anthropogenic sources of light.
Scattering of anthropogenic light contributes to the "skyglow" within and over populated
areas, adding to the total sky brightness. The visual result is a reduction of the number of
visible stars and the disappearance of diffuse or subtle phenomena such as the Milky Way.
The extinction of starlight is a secondary and minor effect also caused by increased
scattering and absorption. Anthropogenic light sources include artificial outdoor lighting,
which varies dramatically across space. Natural sources include the Moon, planets, and
stars that have a predictable rhythm across time.
The nighttime visual environment has some important differences to note. Light
sources and ambient conditions are typically five to seven orders of magnitude dimmer at
night than in sunlight. Moonlight, like sunlight, introduces light throughout an observer's
sight path at a constant angle. On the other hand, dim starlight emanates from all over the
celestial hemisphere while artificial lights are concentrated in cities and illuminate the
atmosphere from below. Sight paths are often inclined upward at night as targets may be
nearby terrain features or celestial phenomena. Extinction behaves the same at night as
during the day, lowering the contrast of scenes through scattering and absorption!
nevertheless the different light sources will yield variable changes in visibility as compared
to what has been established for the daytime scenario. Little research has been conducted
on nighttime visibility. Even if the air quality-visibility interactions are shown to be similar
between day and night settings, the human psychophysical response at night is expected to
differ. Though recent advances in the ability to instrument and quantify nighttime scenes
(Duriscoe et al., 2007, 156411) have been made and can be utilized to evaluate nocturnal
visibility, the state of the science is not yet comparable to that associated with daytime
visibility impairment. The remainder of this document focuses exclusively on daytime
visibility.
9.2.2.1. l\lon-PM Visibility Effects
Light extinction due to the gaseous components of the atmosphere is relatively well
understood and well estimated for any atmospheric conditions. Absorption of visible light by
gases in the atmosphere is primarily by NO2, and can be directly and accurately estimated
July 2009
9-8
DRAFT-DO NOT CITE OR QUOTE

-------
from NO2 concentrations by multiplying by the absorption efficiency. Scattering by gases is
described by the Rayleigh scattering theory. Rayleigh scattering occurs in a pollution-free
atmosphere as a result of light scattering by the gas molecules that compose the
atmosphere (i.e., N2, O2, CO2, etc.) and depends only on the density of the atmosphere, with
highest values at sea level (~12 Mm1) and diminishing with elevation (8 Mm1 at ~4 km),
and varies somewhat at any elevation due to atmospheric temperature and pressure
variations. Rayleigh scattering can be accurately determined for any elevation and
meteorological conditions.
NO2 absorbs more light in the short wavelength blue portion of the spectrum than at
longer wavelengths. For this reason a plume or layer of NO2 removes more of the blue light
from the scene viewed through the layer giving a yellow or brown appearance to the layer
or plume. This filtering of blue light by NO2 can deepen the brown appearance of hazes over
urban areas, although it is not the sole cause of such discoloration (EPA, 1993, 017649). The
photopic-weighted absorption efficiency at the 550 nm wavelength is incorporated into the
revised version of the algorithm for estimating light extinction from aerosol data that is
used for implementing the RHR (Pitchford et al., 2007, 098066). However, NO2 is not
routinely measured at any of the monitoring sites representing visibility protected areas
where its impacts are assumed to be inconsequential compared to those of PM. At
background concentrations NO2 absorption is generally less than five percent of the light
scattering by clean air (Rayleigh scattering), making it unperceivable. Plume visibility
models are available to assess both achromatic and discoloration associated with NO2 light
absorption, for point source emissions (Latimer and Ireson, 1980, 035723! Seigneur et al.,
1984, 156965).
9.2.2.2. PM Visibility Effects
Particle light extinction is more complex than that caused by gaseous components.
PM is responsible for most visibility impairment except under near-pristine conditions,
where Rayleigh scattering is the largest contributor to light extinction or in plumes of
combustion sources that are well-controlled for particulate emissions (e.g., coal-fired power
plants with bag houses), where light absorption by NO2 may dominate the light extinction.
Light-absorbing carbon (e.g., DE soot and smoke) and some crustal minerals are the
only commonly occurring airborne particle components that absorb light. All particles
scatter light, and generally particle light scattering is the largest of the four light extinction
July 2009
9-9
DRAFT-DO NOT CITE OR QUOTE

-------
components. While a larger particle scatters more light than a similar shaped smaller
particle of the same composition, the light scattered per unit of mass concentration (i.e.,
mass scattering efficiency in units of Mm Vljig/m3] which reduces to m2/g) is greatest for
particles with diameters from ~0.3-1.0 pm. If the index of refraction, particle shape and
concentration as a function of particle size are well characterized, Mie theory can be used to
accurately calculate the light scattering and absorption by those particles. However, it is
rare that these particle properties are known, so assumptions are used in place of missing
information to develop a simplified calculation scheme that provides an estimate of the
particle light extinction from available data sets.
Particles composed of water soluble inorganic salts (i.e., ammoniated sulfate,
ammonium nitrate, sodium chloride, etc.) are hygroscopic in that they absorb water as a
function of relative humidity to form a liquid solution droplet. Aside from the chemical
consequences of this water growth, the droplets become larger when relative humidity
increases, resulting in increased light scattering. Hence, the same PM dry concentration
produces more haze. Figure 9-3 shows the effect of water growth as a function of relative
humidity on light scattering for two size distributions of ammonium nitrate and ammonium
sulfate particles as well as for internal and external mixtures (i.e., mixed within the same
particle and in separate particles, respectively) of the two components. This
figure illustrates a number of important points. The water growth effect is substantial with
an increase in light scattering by about a factor of ten between 40% and 97% relative
humidity for the same dry particle concentrations. The amount of scattering is significantly
dependent on the dry particle size distribution. However the growth curves for ammonium
sulfate, ammonium nitrate and mixtures of the two particle components are similar at any
of the dry particle size distributions. Water growth curves are also available for sodium
chloride, the major component in sea salt, which is an important PM component at coastal
locations.
July 2009
9-10
DRAFT-DO NOT CITE OR QUOTE

-------
0.1
c
0)
'o
*-l—
Q)
O
o
O)
c
*1—
QJ
ro
o
CO
0.01
0.001
i—'—i—->—i—'—i—'—i—'—i—'—r
	NHjNOj
-	(NH4)2SO„
	External Mixture
(55.5% nitrate + 44.5% sulfate) d/;
o Internal mixture
(0.3 (im, 1.5}^ ar/..-
40
50
(0.6 Mm, 1.5)
-JL-
60
70
% RH
80 90 100
Source: Tang (1996,157042).
Figure 9-3. Effect of relative humidity on light scattering by mixtures of ammonium nitrate and
ammonium sulfate.
Using Mie theory, the scattering and absorption of any wavelength of light by a
particle of known size and index of refraction (a function of the wavelength of light) can be
calculated (Van de Hulst, 1981, 191972). Particle density is used to convert the particle
light extinction to its mass extinction efficiency (i.e., the ratio of particle light extinction to
its mass). To expand the calculations from one particle at a time to the multitude of
particles in ambient aerosol, information about the aerosol size and composition
distributions are needed. Aerosol mixture refers to how the major components that make up
the particles are mixed. Methods have been developed to treat simple mixture models
ranging from external mixtures where the various components are assumed to be in
separate particles, to multi-component mixtures where individual particles contain several
components (Ouimette and Flagan, 1982, 025047). The latter includes internally mixed
particles where two or more components are mixed within the particles, and layered aerosol
with a core of one component covered by a shell of another component.
The Mie theory solution for an external mixture can be simplified to a linear
relationship where the light extinction is the sum of the mass concentration of each species
July 2009
9-11
DRAFT-DO NOT CITE OR QUOTE

-------
multiplied by its specific mass extinction efficiency (Ouimette and Flagan, 1982, 025047).
This formulation promotes the concept of apportioning the light extinction among the
various PM species. For internally mixed aerosol, the light extinction response to adding or
removing mass of any component to the aerosol is dependent on how such changes would
affect the particle size, density and index of refraction distribution of the aerosol. However,
a number of investigators have shown that the differences among the calculated light
extinction values using external and various internal mixture assumptions are generally
less than about 10% (Lowenthal et al., 1995, 045134; Ramsey, 1966, 013946; Sloane, 1983,
025039; Sloane, 1984, 025040; Sloane, 1986, 045954; Sloane and Wolff, 1985, 045953; Wolff,
1985, 044680). This provides a basis to accept the apportionment of light extinction to the
PM components calculated using an external mixture assumption as a meaningful
surrogate for their contributions.
Ambient aerosols are usually a complex and unknown combination of both internal
and external mixtures of the particle components. Despite these complexities, PM light
scattering can be accurately calculated for any relative humidity if the chemical
composition as a function of dry particle size is known (Hand et al., 2002, 190367; Malm
and Pitchford, 1997, 002519). However, most routinely available ambient monitoring
programs do not include data with sufficient detail to make such calculations. The
IMPROVE network with its greater than 150 remote area monitoring sites (DeBell, 2006,
156388) and the CSN (Jayanty, 2003, 156605) with its greater than 150 urban area
monitoring sites collect 24-h duration fine particle samples (PM2.5) that are analyzed for the
major PM components including SO42 , nitrate, and carbonaceous particulate. CSN also
analyzes for ammonium ion, but does not monitor coarse mass (PM10-2.5), while IMPROVE
measures coarse mass but does not analyze for ammonium ion. Neither data set has
sufficient size resolution to make Mie theory calculations of light extinction, nor does either
program routinely monitor NO2 concentrations, which would be required to calculate its
contribution to light extinction by absorption.
A simple algorithm similar in form to the linear equation that results from Mie theory
applied with an external mixture assumption is frequently used to estimate light extinction
from the concentrations of the major components. The concentration of each of the major
aerosol components is multiplied by a dry extinction efficiency value and for the hygroscopic
components (e.g., ammoniated sulfate and ammonium nitrate) an additional multiplicative
term to account for the water growth to estimate that components contribution to light
July 2009
9-12
DRAFT-DO NOT CITE OR QUOTE

-------
extinction. Both the dry extinction efficiency and water growth terms are developed by
some combination of empirical assessment and theoretical calculation using typical particle
size distributions associated with each of the major aerosol components, and they are
evaluated by comparing the algorithm estimates of light extinction with coincident optical
measurements. Summing the contribution of each component gives the estimate of total
light extinction. The most commonly used of these is referred to as the IMPROVE algorithm
because it was developed specifically to use the IMPROVE aerosol monitoring data and was
evaluated using IMPROVE optical measurements at the subset of sites that make those
measurements (Malm et al., 1994, 044920). The formula for the traditional IMPROVE
algorithm is shown below.
hex, ~ 3 x f (RH) x [Sulfate]
+ 3 x /(RH) x [Nitrate]
+ 4 x [Organic Mass]
+ 10 x [Elemental Carbon]
+ 1 x [Fine Soil]
+ 0.6 x [("oarse Mass]
+ 10
Equation 9-1
Light extinction (bext) is in units of Mm"1, the mass concentrations of the components
indicated in brackets are in pg/m3, and f(RH) is the unitless water growth term that
depends on relative humidity. The dry extinction efficiency for particulate organic mass is
larger than those for particulate SOr and nitrate principally because the density of the dry
inorganic compounds are higher than that assumed for the PM organic mass components.
Since IMPROVE does not include ammonium ion monitoring, the assumption is made that
all S042- is fully neutralized ammonium sulfate and all nitrate is assumed to be ammonium
nitrate. Though often reasonable, neither assumption is always true (see Section 9.2.3.1). In
the eastern U.S. during the summer there is insufficient ammonia in the atmosphere to
neutralize the SOr fully. Fine particle nitrates can include sodium or calcium nitrate,
which are the fine particle fraction of generally much coarser particles due to nitric acid
interactions with sea salt at near-coastal areas (sodium nitrate) or nitric acid interactions
with calcium carbonate in crustal aerosol (calcium nitrate). Despite the simplicity of the
algorithm, it performs reasonably well and permits the contributions to light extinction
July 2009
9-13
DRAFT-DO NOT CITE OR QUOTE

-------
from each of the major components (including the water associated with the SOr and
nitrate compounds) to be separately approximated.
The f(RH) terms inflate the particulate SO r and nitrate light scattering for high
relative humidity conditions. For relative humidity below 40% the f(RH) value is 1, but it
increases to 2 at -66%, 3 at -83%, 4 at -90%, 5 at -93% and 6 at -95% relative humidity.
The result is that both particulate SO r and nitrate are more efficient per unit mass than
any other aerosol component for relative humidity above -85% where its total light
extinction efficiency exceeds the 10m2/g associated with EC. Based on this algorithm,
particulate SO r and nitrate are estimated to have comparable light extinction efficiencies
(i.e., the same dry extinction efficiency and f(RH) water growth terms), so on a per unit
mass concentration basis at any specific relative humidity they are treated as equally
effective contributors to visibility effects. The strong relationship demonstrated between
dry light scattering and fine PM mass concentration or ambient light extinction and fine
PM mass concentration under low relative humidity conditions noted by a number of
investigators (Charlson et al., 1968, 095355; Chow et al., 2002, 037784; Chow et al., 2002,
036166; McMurry, 2000, 081517; Samuels et al., 1973, 070601; Waggoner and Weiss, 1980,
070152; Waggoner et al., 1981, 095453) is reasonable based on this algorithm when the PM
fractional composition is either relatively constant or varies most among PM components
with similar dry extinction efficiency values (e.g., SO42~, nitrate and organic mass
efficiencies).
9.2.2.3. Direct Optical Measurements
Light extinction and its components (i.e., scattering and absorption by particles and
gases) can be determined directly by optical measurements using commercially available
instruments (Trijonis et al., 1990, 157058). Though these measurements are all wavelength
dependent, the convention for visibility monitoring purposes is to make measurement at or
near 550iim, which is the wavelength of maximum eye response. Direct PM light extinction,
scattering and absorption measurements offer a number of advantages compared to
estimates using an algorithm applied to PM speciation data. The direct optical
measurements are considered more accurate because they do not depend on the assumed
particle characteristics (e.g., size, shape, density, component mixture, etc.) thought to be
associated with the major PM species. Also the optical measurements are made with high
time resolution (e.g., minutes to hourly) compared with the filter composition based
July 2009
9-14
DRAFT-DO NOT CITE OR QUOTE

-------
estimates that are typically 24-h duration, allowing the former to better characterize sub-
daily temporal patterns which can help in identifying influential source categories and
characterize atmospheric phenomenon. The higher time resolution attainable with direct
light extinction measurements are also more commensurate than the 24-h light extinction
estimates from PM samples with the short exposure time associated with perceived
visibility effects.
Path-averaged light extinction can be determined by long-path transmissometers that
monitors the intensity of light that has traversed a know distance from a known initial
intensity light source. Transmission (i.e., the ratio of the final to the initial light intensity)
is the natural logarithm of the product of the path-averaged light extinction and the
distance the light has traversed. Transmissometer path-length establishes the useful range
of light extinction over which the measurements can be accurately measured, with path-
lengths of 10km or more required for pristine conditions and less than 1km more
appropriate for hazier situations or to measure the visibility impacts associated with fogs or
precipitation events. The NP Service operated long-path transmissometers at up to 25
locations from 1986 through 2004 (DeBell, 2006, 156388), but have more recently
discontinued their use at all but one remote area location due to the cost of maintenance
and the difficulties of performing calibration. Transmissometers are currently in routine
service at five urban areas.
A number of instruments measure the light scattered by particles and gases from a
source of known intensity. These include forward scattering, back scattering, polar, and
integrating nephelometers. Of these the integrating nephelometers with its high sensitivity
and sample control options has been more widely used for air quality-related visibility and
PM monitoring purposes, while the robust design of the open air forward scattering
instruments have seen extensive use by the National Weather Service (NWS) Automated
Surface Observing System (ASOS) for characterizing visibility principally for
transportation safety purposes (NOAA, 1998). The potential utility of the ASOS visibility
network at about 900 locations for air quality monitoring has been established, but the lack
of resolution in the reported data is a serious impediment to this use of the data (Richards
et al„ 1996, 190476).
Integrating nephelometers draw air into a sample chamber, making it possible to
modify the sample either by changing its humidity or controlling the particle size range
that is measured. This feature makes it possible to use sample-controlled nephelometers to
July 2009
9-15
DRAFT-DO NOT CITE OR QUOTE

-------
investigate the effects of ambient PM size and water growth characteristic on light
scattering (Covert et al., 1972, 072055; Rood et al., 1987, 046397; Malm and Day, 2001,
190431). For instance the coarse particle contribution to light scattering can be estimated
using a nephelometer that alternately samples through a 2.5pm size selective inlet and a
10pm size selective inlet. For routine monitoring, integrating nephelometers are typically
either used to measure the PM component of light scattering when operated at ambient
relative humidity or to measure dry PM light scattering as a high-time resolution surrogate
for PM mass concentration when operated with a heater or other sample air drier.
Integrating nephelometers operated at ambient conditions by the IMPROVE program have
replaced the long-path transmissometer as the principal optical measurement at about 30
locations (DeBell, 2006, 156388).
PM light absorption can also be inferred from measured changes in the light
transmitted through a filter used to sample the PM compared to an identical clean filter
(Bond et al., 1999, 156281). Such measurements can be made subsequent to sampling
(Campbell et al., 1995, 190171) or continuously during sampling by using specifically
designed sampler (Hansen et al., 1982, 190368; Hansen et al., 1984, 002396). All of the
filter-based methods require adjustments to the optical measurements to account for filter
and sampled particle light scattering effects associated with particles concentrated on and
within the matrix of the filters (Bond et al., 1999, 156281). Often PM light absorption
measurements are used to infer BC concentration by assuming it is the dominant PM
contributor to light absorption with a near constant absorption efficiency (Allen et al., 1999,
048923; Babich et al., 2000, 156239). In fact commercially available aethalometers
incorporate an absorption efficiency value so they can directly report BC concentrations.
Like nephelometers, commercially available aethalometers can be obtained with either
single or multiple-wavelength measurement capabilities, where the multi-wavelength data
can be used to better characterize the PM. More recently these have been used to
distinguish BC that absorbs light strongly over the full visible light spectrum (e.g., DE)
from brown carbon that absorbs appreciably more at shorter wavelengths than at long
wavelengths (e.g., WS) (Andreae and Gelencser, 2006, 156215).
Other approaches to measure light absorption include a photoacoustic instrument
that measures the heating associated with absorbed light by suspended PM (Arnott et al.,
1999, 020650; Moosmiiller et al., 1998, 192073), as well as by the difference between light
extinction and light scattering measurements (Bond et al., 1999, 156281).
July 2009
9-16
DRAFT-DO NOT CITE OR QUOTE

-------
9.2.2.4. Value of Good Visual Air Quality
The term visual air quality (VAQ) is used here to refer to the visibility effects caused
solely by air quality conditions. For example, it excludes the reduced visibility caused by
fog. Two broadly different approaches have traditionally been used to define and quantify
the value of good VAQ. One approach assesses the monetary value associated with visibility
changes! the other assesses the psychological value of visual air quality. With respect to the
latter, reduced VAQ is considered an environmental stressor (Campbell, 1983, 190172) that
is associated with heightened amounts of anxiety, tension, anger, fatigue, depression, and
feelings of helplessness (Evans et al., 1987, 190347; Zeidner and Shechter, 1988, 191973).
Though the relationship between impaired VAQ and mental health is poorly understood,
there are greater emergency calls associated with psychiatric disturbances during periods
with reduced VAQ (Rotton and Frey, 1982, 190477). Studies have shown that reduced VAQ
affects people's behavior, including reductions in outdoor activities, and increased hostility
and aggressive behavior (Evans et al., 1982, 190521; Cunningham, 1979, 191974; Jones and
Bogat, 1978, 190396; Rotton et al., 1979, 190478).
The value of VAQ (both monetary and non-monetary) has been investigated in two
broadly different settings, non-recreational or urban settings and recreational settings,
such as the national parks and wilderness areas where visibility is protected by the RHR
(Trijonis et al., 1990, 157058). In urban settings, public surveys have shown that greater
than 80% of the participants are aware of poor VAQ conditions (Cohen et al., 1986, 190182),
though attitudes towards poor VAQ have been shown to vary by socio-economic status,
health, and length of residence in the urban setting (Barker, 1976, 072137). The economic
importance of urban visibility has been examined by a number of studies designed to
quantify the benefits (or willingness to pay) associated with potential improvements in
urban visibility. Urban visibility valuation research prior to 1997 was summarized in
Chestnut and Dennis (1997, 014525), and was also described in the 2004 Air Quality for PM
(EPA, 2004, 056905) and the 2005 OAQPS PM NAAQS Staff Paper (U.S. EPA, 2005,
090209). These reviews summarize 34 estimates (based on different cities or model
specifications) from six different studies. Since the mid 1990s, however, only one new
valuation study of urban visibility has been published (Beron et al., 2001, 156270)which is
summarized below (Section 9.1.4.6).
In recreational settings, experience based demand models have been developed using
on-site and mail-in surveys to judge the relative importance to national park visitors of
July 2009
9-17
DRAFT-DO NOT CITE OR QUOTE

-------
various park attributes including good VAQ, to assess visitor awareness of VAQ conditions,
and to explore possible relationships between VAQ and visitor satisfaction (Ross et al.,
1987, 037420; Ross et al., 1985, 044287). At the three western and two eastern national
parks where this survey was conducted, visitors rated the attribute identified as "clean,
clear air" among the most important features of the parks. A random sample of 1800
visitors at one of the parks (Grand Canyon) showed that visitor awareness of VAQ impacts
increased as measured visibility conditions decreased, and that overall park enjoyment and
satisfaction decreased with reduced VAQ. Grand Canyon visitors when asked to indicate
how they would budget their time (e.g., between visiting an archaeological site or a view
lookout point) indicated that they would to be willing to significantly alter their behavior to
experience views under improved VAQ (Malm et al., 1984, 044292).
9.2.3. Monitoring and Assessment
Monitoring and the assessment of monitoring data serve a number of goals with
regard to the visibility effects of PM, including improving the understanding of the
physio/chemical/optical properties of the aerosol, characterizing spatial and temporal air
quality patterns, and assessing the causes (i.e., pollution sources and atmospheric
processes) that are responsible for visibility impairment. Information generated by special
studies employing sophisticated instrumentation are typically needed to advance the
understanding of aerosol properties, while characterizing trends is the product of analyzing
routine monitoring data, whereas assessing the causes of haze usually involves a
weight-of-evidence approach applied to special study and/or routine monitoring data sets
plus the use of air quality simulation modeling. This section summarizes recently available
information that is based on monitoring data.
9.2.3.1. Aerosol Properties
Particle size is the most influential physical property of aerosols with respect to their
dry light extinction efficiency. Chemical composition by size is used to ascertain density
(needed to convert aerodynamic to physical size and to determine particle mass as a
function of size) and to identify the water growth characteristics of the aerosol (needed to
calculate the particle size, density and index of refraction at ambient RH). To characterize
aerosol properties of interest for visibility effects, field monitoring programs typically
include particle size distribution monitoring, high size resolution particle sampling with
July 2009
9-18
DRAFT-DO NOT CITE OR QUOTE

-------
subsequent compositional analysis, and optical monitoring. These generate data that
permit optical closure assessments where the light scattering and/or light extinction
estimates from the aerosol data are compared to corresponding optical data. Since
component contributions to visibility are generally assessed by applying the IMPROVE or
some similar algorithm to measured or modeled aerosol concentration data, this section will
include recent investigations that evaluate or address various assumptions inherent in the
use of these simple algorithms.
One component of the Big Bend Regional Aerosol and Visibility Observational
(BRAVO) Study, conducted at Big Bend NP, TX in the summer and fall of 1999, entailed use
of detailed measurements of aerosol chemical composition, size distribution, water growth,
and optical properties to characterize the aerosol and assess the relationship between
aerosol physical, chemical and optical properties (Malm et al., 2003, 190434; Schichtel et
al., 2004, 179902). Fine ammoniated sulfate during the BRAVO Study was about half the
fine particle mass concentration and was shown to be responsible for about 35% of the light
extinction. Rayleigh scattering was the second largest contributor at about 25%, followed by
coarse particle (about 18%), and organic compounds (about 13%). There was little fine
particle nitrate (less than 5% of the mass concentration) and most of it is apparently in the
form of sodium nitrate and two thirds of it was found in the coarse mode where it comprises
about 8% of the coarse particle mass concentration. Both the composition of the nitrate and
the fact of much of it being in the coarse size mode (2.5 pm >D>10 pm) are inconsistent
with the implied assumptions of the IMPROVE algorithm.
A year-long special study of coarse particle speciation was conducted at nine
IMPROVE remote area monitoring sites during 2003 to 2004 to provide additional
information about the geographic and seasonal variations in coarse particle composition
(Malm et al., 2007, 156730). The same sampling and analytical methodologies procedures
were used for the PMio samples as are routinely used on the IMPROVE PM2.5 samples. The
IMPROVE coarse particle speciation study did not include ammonium analysis, so SO r
and nitrate ions were assumed to be ammonium sulfate and ammonium nitrate. As
expected crustal minerals were the largest contributors to coarse mass overall (about 60%),
though at Mt. Rainier the fraction of coarse PM that was organic exceeded the crustal
mineral by nearly two to one (i.e., 59.2% compared to 33.5%) On average across sites the
organic particulate contributed significantly at about one quarter of the coarse mass, while
ammonium nitrate was the third largest contributor to coarse mass (about 8%). Seas salt
July 2009
9-19
DRAFT-DO NOT CITE OR QUOTE

-------
was negligible overall, but high at the one coastal site (i.e., 12% at Brigantine, NJ). The two
California sites, San Gorgonio and Sequoia, had the highest coarse nitrate concentrations,
0.74 pg/m3 and 0.69 pg/m3, and high fine nitrates concentrations on average, 2.66 pg/m3 and
2.14 pg/m3, respectively. Brigantine, a coastal site in New Jersey, had the highest fraction of
total nitrate in the coarse size range (36%). The authors speculate that Brigantine's
particulate nitrate is likely sodium nitrate, the result of nitric acid reactions with sodium
chloride. The nine-site average fraction of total nitrate in the coarse size range is 26%. By
contrast, coarse SOr concentrations are small with only about -1% of the total SOr in the
coarse fraction.
Routine IMPROVE monitoring data include the mass concentration, but not the
composition of the coarse PM fraction, so the algorithm used to estimate light extinction
does not include any provision for varied coarse PM composition as shown in this study.
This study shows that about 10% of the coarse mass across the nine monitoring sites is
composed of hygroscopic materials (i.e., ammonium sulfate, ammonium nitrate and sea
salt), which during high humidity conditions will scatter more light than estimated by the
current algorithm (e.g., -20% bias at -90% relative humidity). However, at coastal sites
such as the Brigantine, NJ, IMPROVE site where the combined concentration of the
inorganic salts (i.e., sea salt, nitrate and SOr ) constitute a significant fraction (-24% on
average) of the coarse mass concentration, the IMPROVE algorithm underestimation of
light extinction by coarse PM can be significant for high relative humidity conditions (-60%
at -90% relative humidity). The resulting underestimation of total light extinction is
typically much smaller since fine particle light extinction generally exceeds that
contributed by coarse particles. Another issue with regard to estimating light extinction
from coarse PM concentration when the composition is not crustal minerals, as has been
assumed, has to do with the lower average density of the coarse mode particles that results
in greater particle numbers and/or larger particles and therefore a greater light extinction
efficiency (Malm and Hand, 2007, 155962).
Special studies with more complete, higher time resolution and size resolved
particulate inorganic ion species chemistry and precursor gases were conducted at seven of
the nine sites with IMPROVE coarse particle speciation monitoring (Lee et al., 2008,
156686). This work confirmed the presence of sodium and calcium nitrate (referred to as
mineral nitrate) primarily in the coarse particle size range in addition to fine particle
ammonium nitrate where low temperatures, high humidity and excess ammonium (beyond
July 2009
9-20
DRAFT-DO NOT CITE OR QUOTE

-------
that required to neutralize the particulate S() r ) favored particle phase equilibrium.
Figure 9-4 is a map showing the locations and sample times and estimated composition of
the total particulate nitrate for the seven locations for this special study. Sites with a high
fraction of ammonium nitrate (e.g., San Gorgonio, Bondville, and Brigantine) have the
highest nitrate contributions to total mass concentration and haze, whereas sites with high
mineral nitrates tend to have low total nitrate contributions. This work shows that the
common assumption that particulate nitrate is in the fine particle size range and consists
principally of ammonium nitrate is not necessarily true.
Grand Catwon (May. 2003)
Bondville (February, 2003)
San Gib ijimio (April, 2003)'
Yosemile (July/Aug, 700?)
Uriyariliiie (Novanber, 2003)
—GRSM (July/Aug, 7004)^1
San Gorgonio (July, 2003)
Kg (tend (July Oci, 1998)
Source: Lee et al. (2008,156686)
Figure 9-4. Estimated fractions of total particulate nitrate during each field campaign comprised of
ammonium nitrate, reacted sea salt nitrate (shown as NaNCh), and reacted soil dust
nitrate (shown as CafNQsh}.
Extinction efficiencies for individual particle species can be theoretically calculated
from sized'resolved aerosol measurements and can be inferred using multiple linear
regression applied to aerosol composition and light extinction measurement data. In a
recent publication, Hand and Malm (2007, 155825) reviewed the literature since 1990 in
which aerosol mass scattering efficiency values were calculated or inferred. From these
they have compiled normalized dry scattering efficiency values for the individual species.
July 2009
9-21
DRAFT-DO NOT CITE OR QUOTE

-------
Based on 93 separate determinations including marine, remote continental and urban
areas data sets, the average dry mass scattering efficiency for ammonium sulfate is
2.5 ± 0.6 m2/g. Average values tended to be somewhat lower for the marine aerosol (~2 m2/g)
than for remote continental (~2.7 m2/g) and urban (2.6 m2/g) areas, and values also tended
to be lower for fairly clean arid locations compared with more humid polluted areas.
Based on 48 separate determinations including remote area and urban area data sets,
the average dry mass scattering efficiency for ammonium nitrate is 2.7 ± 0.5 m2/g (Hand
and Malm, 2007, 155825). Average values were higher in remote locations (2.8 ± 0.5 m 2/g)
compared to urban locations (2.2 ± 0.5 m2/g) though this might be accounted for by the
predominate use of multiple linear regression for the remote areas, which can be biased
high, compared to the use of theoretical calculations for the urban data sets.
Organic fine PM extinction efficiency of 3.9 ± 1.5 m2/g is based on 58 separate
determinations, though much higher values (~6 m2/g) resulted for locations influenced by
industrial andbiomass combustion sources (Hand and Malm, 2007, 155825). These organic
fine PM extinction efficiency values were adjusted to use a consistent ratio of organic mass
to OC (OC) of 1.8 for each determination of the mass concentration. This value is generally
associated with aged organic PM, while for more freshly emitted PM, such as in an urban
environment, a smaller ratio (e.g., 1.4) would be more appropriate. This could explain the
discrepancy between two approaches used to estimate the organic PM light extinction
efficiency for Phoenix (Hand and Malm, 2006, 156517), which resulted in a significantly
lower value where a site specific regression method was used compared to the value
obtained from a method optimized for remote-area monitoring (2.47 m2/g compared to 3.71
m2/g). However in Fresno both the mass balance and light scattering balance was improved
by using a ratio of 1.8 instead of 1.4 to estimate the organic compound mass (Watson and
Chow, 2007, 157127). Another possible or partial factor with respect to urban light
extinction efficiency for organic PM may be that the size distribution of freshly emitted
organic PM in urban areas extends significantly into the ultra-fine particle size range
(Demerjian and Mohnen, 2008, 156392) that is less efficient per mass concentration at light
scattering than the generally larger-sized aged organic PM such as from a distant forest fire
as was measured at the Baltimore Supersite.
Hand and Malm (2007, 155825) also reviewed and made recommendation for
extinction efficiencies for the other PM components including for mixed coarse mode
(1.0 ± 0.9 m2/g based on 51 determinations) and fine mode dust or soil (3.3 ± 0.6 m2/g based
July 2009
9-22
DRAFT-DO NOT CITE OR QUOTE

-------
on 23 determinations, but recommending 1.0 m2/g for use with data from realistic collection
efficiency samplers) and fine sea salt (4.5 ± 0.9 m2/g based on 25 determinations, but
recommending 1.0 m2/g to 1.3 m2/g for use with data from realistic collection efficiency
samplers). This work did not address light absorption efficiency of EC, CB, or crustal PM.
The Hand and Malm (2007, 155825) average dry mass light scattering efficiency
values are generally consistent with the values for the IMPROVE algorithm (as shown in
equation 9" 1). However the adoption of the IMPROVE algorithm by EPA for calculating the
haze metric used to track trends and assess the nominal pace of progress for the RHR (EPA,
2001, 157068) resulted in much greater scrutiny of its performance in estimating extinction
(Lowenthal and Kumar, 2003, 156712; Malm, 1999, 025037; Malm and Hand, 2007, 155962;
Ryan et al., 2005, 156934). Among the issues raised is that the algorithm tended to
underestimate the light extinction for the haziest conditions that occur principally during
the summer in the southeastern U.S. and overestimate for near pristine conditions that
tend to occur most often in the arid western U.S. Furthermore they showed the lack of mass
or light scattering closure at coastal sites due to sea salt that was not accounted for by the
IMPROVE algorithm. These assessments used mass concentration and light extinction
closure and regression analysis methods to infer that the dry extinction efficiency for the
major fine particle components would need to vary in order to avoid the biased estimates of
light extinction at the extremes. Theoretical calculations of S( )r dry extinction efficiencies
for 41 days of size-resolved chemical composition data for Big Bend, Texas, as part of the
BRAVO Study produced a range of results from ~2.4 m2/g to ~4.1 m2/g, with the larger dry
extinction efficiency values tending to be associated with higher ammonium sulfate mass
concentration and narrower size distributions (Schichtel et al., 2004, 179902).
In response to the technical concerns raised about the performance of the IMPROVE
algorithm, a revised algorithm was developed (Pitchford et al., 2007, 098066). The revised
version of the algorithm differs from the original algorithm by including a fine sea salt term
related to the measured chloride ion concentration, increases by about 30% the mass
concentration of the organic aerosol component by changing the ratio of organic compound
mass to OC mass from 1.4 to 1.8, uses site elevation dependent Rayleigh scattering in place
of 10 Mm 1 that had been used at every site, added a NO2 light absorption term and
employs a split component model for the secondary particulate components (i.e., SO42",
nitrate and organic species) with new water growth terms to better estimate their
July 2009
9-23
DRAFT-DO NOT CITE OR QUOTE

-------
extinction at the high and low extremes of the range. The revised algorithm is displayed
below in Equation 9-2.
bext « 2.2 x /' (lilI) x [SmallSulfate + 4.8 x fL(RH) x [Large Sulfate]
+ 2.4 x /' (lill) x [Small Nitratd + 5.1 x fL(RLl) x [Large NitratA
+ 2.8 x [SmallOrganic Mas A + 6.1 x [Large OrganicMasA
+ 10 x [ElementalCarboA
+ 1 x [FineSoiA
+ 1.7 xfjEB x [SeaSalA
+ 0.6 x [Coarse Mas A
+ RayleighScattering (Site Sped fib
+ 0.33 x [N02 (ppH}\
Equation 9-2
Small and large SO42 , nitrate and organic mass are used to refer to the splitting of
the concentrations of each of those three species into two size distributions. This approach
accounts for increased light extinction efficiency with mass by using a simple mixing model
that assume that each of these three components are comprised of an external mixture of
small and large particle size modes. Conceptually, the large mode particles represent aged
or cloud-processed aerosol, while the small mode particles represent relatively newly
generated particles from the gas phase precursors. The former are more likely to be
associated with high concentrations while the latter are likely to be at relatively low
concentration.
The geometric mean diameter and standard deviations assumed for these two size
modes are 0.5 pm and 1.5 for the large mode particles and 0.2 pm and 2.2 for the small
mode particles. Mie theory applied to these size distributions for the three species results in
dry extinction efficiencies for the small and large mode ammonium sulfate (2.2 m2/g and 4.8
m2/g), ammonium nitrate (2.4 m2/g and 5.1 m2/g) and organic mass (2.8 m2/g and 6.1 m 2/g)-
Water growth terms specifically derived for the small and large size distribution using the
upper branch of the hygroscopic growth curves for ammonium sulfate are applied to both
the SO42- and nitrate PM. No water growth is assumed for organic PM.
A simple empirically developed apportionment approach that was evaluated by
testing the new algorithms estimated light scattering at the 21 IMPROVE sites that have
nephelometer-measured light scattering data. For each sample, the fraction of the fine
particle component (SO42 , nitrate, or organic mass) that is assigned to the large mode is
calculated by dividing the total concentration of the component by 20 pg/m3 (e.g., if the total
July 2009
9-24
DRAFT-DO NOT CITE OR QUOTE

-------
fine particle nitrate concentration is 4 pg/m3, the large mode concentration is 1/5 of 4 pg/m3
or 0.8 pg/m3, leaving 3.2 pg/m3 in the small mode). If the total concentration of a component
exceeds 20 pg/m3, all of it is assumed to be in the large mode.
The performance of the original and revised IMPROVE algorithms was evaluated
using the data for 21 IMPROVE remote-area monitoring sites that also have nephelometer
monitoring of particle light scattering. Figures 9"5 and 9-6 are scatterplots of the estimated
versus measured light scattering for the two algorithms. The revised algorithm has
noticeably reduced bias at the upper and lower extremes. However, the new algorithm
estimates have somewhat reduced precision (i.e., the points are more broadly scattered).
States have adopted the new algorithm for the technical assessments that support their
RHR State Implementation Plans, but the revised algorithm was too recently developed to
be incorporated into any of the peer-reviewed technical literature reported on below. In
general the differences resulting from use of the original versus the revised IMPROVE
algorithm in identifying best and worst haze conditions and the apportionment of the
various PM components are small with exception of coastal locations where sea salt may be
a significant contributor.
July 2009
9-25
DRAFT-DO NOT CITE OR QUOTE

-------
100	150	200	250
Measured Bsp
Source: Pitchford, et al. (2007, 098066)
Figure 9-5. A scatter plot of the original IMPROVE algorithm estimated particle light scattering
versus measured particle light scattering.
150	200
Measured Bsp
Source: Pitchford, et al. (2007, 098066).
Figure 9-6. Scatter plot of the revised algorithm estimates of light scattering versus measured light
scattering.
July 2009
9-26
DRAFT-DO NOT CITE OR QUOTE

-------
9.2.3.2. Spatial Patterns
The IMPROVE network is the basis for much of what is known about particulate
species spatial and temporal patterns for remote areas of the U.S. Though IMPROVE
includes some urban monitoring sites, most of what is known about urban particle
speciation trends is based on the EPASpeciation Trend Network (STN) and other similarly
operated state particle speciation sites jointly referred to as the Chemical Speciation
Network (CSN) (Jayanty, 2003, 156605). The number of IMPROVE network sites has
increased considerably beginning in 2000, first to increase its ability to generate data
representative of the 156 visibility-protected national parks and wilderness areas, then
later as the states in the central U.S. requested additional remote-area monitoring to better
understand their contributions to regional haze. The expansion of the network into the
central U.S. significantly improved the understanding of spatial trends in a region of the
country that had little speciation monitoring. Except as otherwise noted most of the
information in this section was from the IMPROVE Report IV (DeBell, 2006, 156388) and
displays of data that are readily generated using the Visibility Information Exchange Web
Site (VIEWS). VIEWS, the ambient monitoring data system, is one of several websites (as
described in Table 9-1) sponsored by the Regional Planning Organizations (RPO) that
documents substantial, though often otherwise unpublished, technical information
generated to support implementation of the RHR.
Figure 9*7 shows maps of remote area light extinction estimates from PM speciation
data for two yr selected to demonstrate the additional information available due to the
expansion of the IMPROVE network into the central U.S. The locations of monitoring sites
supplying the data shown as color contours are shown as dot on the maps. Users of such
contour maps are usually cautioned that the contours are only there to guide the eye to
sites with similar measurements and that nothing should be implied about spatial patterns
where there are no monitoring sites. Certainly these plots give proof to the wisdom of such
warnings. Prior to 2001 there were no IMPROVE or any other remote-area aerosol
speciation monitoring sites in the central states between northern Minnesota and Michigan
to the north and Arkansas and Kentucky to the south. The lack of monitoring over such a
large region in the center of the country hid the presence of high average regional haze over
the Midwestern U.S. Smaller scale differences are seen in the rest of the country and some
of those are due to interannual variations as well as to better spatial resolution made
possible by a more dense monitoring network.
July 2009
9-27
DRAFT-DO NOT CITE OR QUOTE

-------
Table 9-1. Regional Planning Organization websites with visibility characterization and source
attribution assessment information.
..	Name and Web Address	RPO	Information Content and Comments
Information
RPO Home Pages Western Regional Air Partnership	WRAP	Organizational structure, plans, projects, reports and links to other sites with
additional information.
http:llwww.wrapair.orgl
Central Regional Air Planning Association CENRAP
http:llwww.cenrap.orgl
Midwest Regional Planning Organization MRPO
MANE-VU works in close cooperation with Northeast States for Coordinated Air
Use Management (NESCAUM) and Mid-Atlantic Regional Air Management
Association (MARAMA) to develop the technical information for RHR in the
Northeast. All three web sites contain unique technical support information.
Mid-Atlantic/Northeast Visibility Union	MANE-VU
http:llwww.manevu.orgl	NESCAUM
http:||www.nescaum.org|topics|regional-haze MARAMA
http:||www.marama.org|visibility|
Visibility - Air Quality
Monitoring Data
Visibility Information Exchange Web Site
http:Hvista.cira.colostate.edu/viewsl
All RPOs
All IMPROVE and most other PM speciation data, RHR compatible derived
parameters, and user-friendly tools to summarize and display data.
Emission Inventory
Data
Emissions Data Management System
http:ffwww.wrapedms.orgfdefault login.asp
WRAP
WRAP emission inventory data warehouse and tools that provides a consistent
approach to regional emissions tracking
Monitoring Data
Assessment
Causes of Haze Assessment
http:llwww.coha.dri.edu)
WRAP
CENRAP
Monitoring site-specific descriptive characterizations and maps, seasonal and
trends analysis, air flow analysis, & receptor modeling.
Visibility Modeling
U. of California-Riverside Modeling Center
http://pah.cert.ucr.edu/aqm/308/ WRAP
CENRAP
http://pah.cert.ucr.edu/aqm/cenrap/index.shtml vistas
Descriptions of input data, performance, and results of regional scale modeling
(CMAQ & CAMx) & source attribution for base and future yr regional haze.

http://pah.cert.ucr.edu/vistas/


Integrated Information
to Support RHR SIP
Preparations
Technical Support System
http://matar.cira.colostate.edu/tss/
WRAP
Provides access and common formats to display and summarize emissions
inventory information, monitoring data/ assessment and regional haze modeling
result to aid state and tribal analyst prepare RHR implementation plans.
http:llG4.27.125.175lmrpo.html
Visibility Improvement State and Tribal	VISTAS
Association of the Southeast
http:llwww.vistas-sesarm.orgl
July 2009
9-28
DRAFT-DO NOT CITE OR QUOTE

-------
Source: VIEWS (http://vista.cira.colostate.edu/views/)
Figure 9-7. IMPROVE network PM species estimated light extinction for 2000 (left) and for 2004
(right).
Source: VIEWS (http://vista.cira.colostate.edu/views/)
Figure 9-8. Mean estimated light extinction from PM speciation measurements for the first (top
left), second (top right), third (bottom left), and fourth (bottom right) calendar quarters of
2004.
July 2009
9-29
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 9"8 shows the seasonal pattern of PM species estimated light extinction using
maps of mean values for each of the calendar quarter for 2004. The first quarter has the
highest region of haze centered in the Midwestern U.S.! the warmer second and third
quarters have the region of highest haze over the Ohio River Valley! and the fourth quarter
is a composite with high haze in both the Midwest and Ohio River Valley. Smaller regions of
haze show up in the Columbia River Valley (border between Washington and Oregon) in the
colder first and fourth quarters and in Southern California in the warmer second and third
quarters.
The IMPROVE algorithm permits each PM component contribution to light extinction
to be separately estimated. Figures 9-9, 9-10 and 9-11 display the seasonal variation of the
percent contribution to aerosol light extinction by the various component estimates.
Figure 9"9 shows the contributions by SOr and nitrate particulate including the haze
enhancement caused by the absorbed water in humid conditions. As shown in Figures 9"9, a
large regional pattern of high contribution to haze by nitrate PM is centered in the
Midwest, and during the cooler months the nitrate PM is the dominant cause of haze in the
region responsible for a third to a half of the particulate light extinction. Midwestern
particulate nitrate is responsible for the regional pattern of the highest haze conditions
shifting from the Ohio River Valley during summer to the Midwest in the winter as shown
in Figure 9"8. Particulate nitrate is also a significant contributor to particulate light
extinction year-around in parts of California, where it generally contributes 20%-40%. The
Pacific Northwest, parts of Idaho and Utah experience large contributions to particulate
light extinction by nitrates during the colder seasons, with contributions of 20%-30%.
Figure 9-9 also shows that particulate S( )r is the predominate contributor in the eastern
U.S., where it contributes 40% or more on average and during the summer months up to
three quarters of the particulate light extinction over much of the East. In the western U.S.
particulate SO r generally contribute 20-50% of the particle light extinction. Regions of the
lowest fractional contributions by particulate S( )r and nitrate for any calendar quarter are
generally in the western U.S., and as are shown in the subsequent two figures have
significant contributions by crustal PM components (i.e., coarse mass and fine soil) and by
carbonaceous PM (i.e., organic mass and EC).
Figure 9" 10 shows the contributions to haze by the carbonaceous PM components (i.e.,
organic mass and EC). They show broadly similar patterns with the greatest contributions
in the western U.S. especially during the warmer months of the year. For the most part this
July 2009
9-30
DRAFT-DO NOT CITE OR QUOTE

-------
spatial pattern results from the dominant contributions to haze by SO r and nitrate PM in
the eastern half of the U.S., leaving relatively little for other component contributions. The
fractional contribution to haze by organic PM is generally two to five times that of EC. In
absolute terms, both carbonaceous components tend to have two to three times higher
concentrations in the eastern U.S. than in the non-coastal western states.
July 2009
9-31
DRAFT-DO NOT CITE OR QUOTE

-------
Source: VIEWS (http://vista.cira.colostate.edu/views/)
Figure 9-9. Percent contributions of ammonium nitrate (left column) and ammonium sulfate (right
column) to particulate light extinction for each calendar quarter of 2004 (first through
fourth quarter arranged from top to bottom). Note that the contour intervals are not the
same for the two species contributions.
July 2009
9-32
DRAFT-DO NOT CITE OR QUOTE

-------
Source: VIEWS (http://vista.cira.colostate.edu/views/)
Figure 9-10. Percent contributions of organic mass (left column) and EC (right column) to particulate
light extinction for each calendar quarter of 2004 (first through fourth quarter arranged
from top to bottom). Note that the contour intervals are not the same for the two species
contributions.
July 2009	9-33	DRAFT - DO NOT CITE OR QUOTE

-------
Source: VIEWS (http://vista.cira.colostate.edu/views/)
July 2009
9-34
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 9-11. Percent contributions of coarse mass (left column) and fine soil (right column) to
particulate light extinction for each calendar quarter of 2004 (first through fourth
quarter arranged from top to bottom). Note that the contour intervals are not the same
for the two species contributions.
Figure 9" 11 shows the contributions to haze by coarse mass and fine soil components.
As with the carbonaceous components, these crustal dom inated components have a similar
spatial pattern with regions of highest contribution to haze in the western U.S., and just as
for the carbonaceous PM, the explanation for low contributions in the eastern U.S. is the
dominant contributions to haze by SO r and nitrate PM leaving relatively little for other
components. The crustal components contribute more to haze in the arid regions of the west
including the southwestern deserts. In absolute terms, coarse mass concentrations are as
high in the rural areas of the center of the country (including Oklahoma, Arkansas, Kansas,
Missouri, and Iowa) as they are in the Desert Southwest. Typically coarse mass
contributions to haze exceed those of fine mass by a factor of 2-4.
9.2.3.3. Urban and Regional Patterns
Using a combination of IMPROVE and CSN data, it is possible to compare urban
PM2.5 concentrations and composition to corresponding remote-area regional values. These
are shown as paired color contours maps for IMPROVE and IMPROVE plus CSN (see
Figures 9-12 to 9-23). The degree of comparability of the data from these two networks was
assessed by an analysis of two yr of co-located IMPROVE and CSN data from six urban
areas. The CSN organic mass data were adjusted for a positive sampling artifact prior to
inclusion in this assessment, in a fashion similar to that used for the IMPROVE data set
(pages 29-30, DeBell, 2006, 156388). Note that the contour scales are different between the
two maps for each component pair of maps so that each contains as much information as
possible using ten concentration contours. To assess the degree to which urban areas have
higher PM component concentrations compared to regional background note how many
contour intervals surround the urban monitoring sites. The U.S. EPA (2004, 190219) used
the pairing of IMPROVE and CSN monitoring sites at 13 selected urban areas to separate
local and regional contributions of three major PM2.5 components as shown in Figure 9-24.
In Figures 9-12 and 9-13, urban PM2.5 concentrations are systematically higher than
those in the surrounding non-urban regions. The urban excess is generally much higher in
the western U.S. than in the East (e.g., there are five contour intervals separating Salt
Lake City from its remote regional area compared to only two for Indianapolis). This
July 2009
9-35
DRAFT-DO NOT CITE OR QUOTE

-------
implies that eastern and western urban PM2.5 concentrations and resulting visibility are
less different than the eastern and western regional concentrations and visibility.
• IMPROVE Site
Hawaii
Alaska
¦ IMPROVE Urban Site
Puerto Rico /
Virgin Islands
_ 146
10 9
9.68
8 47
7.26
6.05
4.84
3.63
h 2.42
¦ 1 21
¦o.OO
|jg/m3
Source: Debell (2006,156388).
Figure 9-12. IMPROVE Mean PM2.5 mass concentration determined by summing the major
components for the 2000-2004.
July 2009
9-36
DRAFT-DO NOT CITE OR QUOTE

-------
• IMPROVE Site
¦ IMPROVE Urban Site
Puerto Rico/
Virgin Islands
30 7
¦ 14.7
13 0
11.4
9.78
8.15
6,52
4 89
13.26
1.63
0.00
Mg/m3
Alaska
Hawaii
Source: Debell (2006,156388).
Figure 9-13. IMPROVE and CSN (STN) mean PM2.5 mass concentration determined by summing the
major components for 2000-2004
• IMPROVE Site
¦ IMPROVE Urban Site
Puerto Rico /
Virgin Islands
Alaska
Hawaii
Source: Debell (2006,156388).
Figure 9-14. IMPROVE mean ammonium nitrate concentrations for 2000-2004.
July 2009
9-37
DRAFT-DO NOT CITE OR QUOTE

-------
• IMPROVE
¦ IMPROVE
Site
Urban
Puerto Rico /
Virgin Islands
Alaska
Hawaii
Site
Source: Debell (2006,156388).
Figure 9-15. IMPROVE and CSN (STN) mean ammonium nitrate concentrations for 2000-2004.
® Alaska
• IMPROVE Site
¦ IMPROVE Urban Site
Hawaii
6.63
5.46
4.86
4.25
3.64
13.04
2.43
1.82
11.21
0.61
0.00
Mg/m3
O
Puerto Rico /
Virgin Islands
Source: Debell (2006,156388).
Figure 9-16. IMPROVE mean ammonium sulfate concentrations for 2000-2004.
July 2009
9-38
DRAFT-DO NOT CITE OR QUOTE

-------
A STN Site
• IMPROVE Site
¦ IMPROVE Urban Site
e e
Puerto Rico /
Virgin Islands
® Alaska
Hawaii
Source: Debell (2006,156388).
Figure 9-17. IMPROVE and CSN (STN) mean ammonium sulfate concentrations for 2000-2004.
Figures 9-14, 9-15, and 9-24 show the PM2.6 nitrate in remote and urban areas. Here
the western states have urban particulate nitrate concentrations that far exceed twice the
remote area regional concentrations. For the Central Valley of California and Los Angeles
areas, the urban excess of ammonium nitrate exceeds regional concentrations by 2 pg/m3 to
12 pg/m3. In the region of the Midwest nitrate bulge, the urban concentrations were less
than twice the regional concentrations for an annual urban excess of about 1 pg/m3.
Northeast and southeast of the Midwest nitrate bulge, annual urban particulate nitrate
concentrations are several tenths to about one ug/rrr above the remote area regional
concentrations, with warmer southern locations tending to have the smaller concentrations
of both regional and urban excess particulate nitrate.
As shown in Figures 9-16, 9-17, and 9-24, annual-averaged urban particulate SOr
concentrations are generally not much higher than the regional values, with urban excess
generally of less than about a half ug/m3. The exceptions apparent by comparing Figures
9-16 and 9-17 are in Texas and Louisiana where urban excess particulate SO r are greater
than 1 pg/m3, perhaps caused by local emissions (e.g., from oil refineries). Urban
contributions are a larger fraction of the total particulate S( ) r concentrations in the
western U.S. because the regional concentrations are much lower than in the East. The
modest additional particulate S( )r concentrations associated with urban areas suggests
July 2009
9-39
DRAFT-DO NOT CITE OR QUOTE

-------
that most particulate SOr is regionally distributed, and that IMPROVE and CSN
monitoring sites can be used together to enhance the ability to delineate particulate SOr
spatial distributions. For example, note that the additional data from urban sites shown in
Figure 9" 17 extends north and south distribution of the high particulate SO r loading
shown in Figure 9" 16 over Tennessee and Kentucky, as well as the high loadings over
southern Pennsylvania, eastern West Virginia and northern Virginia. (The color-contour
suggested dip in concentrations between the two eastern particulate SOr high
concentrations regions may not exist in the atmosphere, but this cannot be verified without
speciation monitoring sites in southern Ohio, the boarder of Kentucky and West Virginia
and western Virginia.)
Urban and remote area carbonaceous PM2.5 are displayed in Figures 9-18 and 9-19
(organic mass), 9-20 and 9-21 (EC), and 9-24 (total carbon = organic + EC concentration).
Just as with particulate nitrate both organic mass and EC concentrations are more than
twice the remote-area background concentrations for western urban monitoring locations.
One of the more interesting pairing of sites is for the Virgin Islands compared to the urban
site at San Juan Puerto Rico (see the map cutout Figures 9-18 through 9"2l). The San Juan
urban excess OC is moderate, while the EC value is among the most extreme inferred in
this manner. For eastern urban areas, approximately half the total carbon is local while the
other half is regional. In eastern urban areas, carbonaceous and SO r particulate are the
two major components of PM2.5, with roughly equal contributions, and account for over 80%
of the mass concentration. Edgerton et al. (2004, 156413) showed that carbonaceous PM2.5
is responsible for most of the urban excess above regional concentrations at four
urban/rural paired Southeastern Aerosol Research and Characterization (SEARCH)
monitoring sites in the southeastern U.S. However, the higher overall light extinction
efficiency for SOr resulting from its hydrophilicity gives it ~ 2:1 dominance in
responsibility for eastern urban light extinction.
Urban and remote area fine soil PM2.5 concentrations are displayed in Figures 9-22
and 9"23. Urban fine soil concentrations are at most a few tenths of a ug/ma higher than the
regional background concentrations and in some regions they are much less. Just as with
carbonaceous PM2.5, the Virgin Island, San Juan, Puerto Rico pair are interesting for fine
soil. In this case the interesting feature is that both of these island monitoring sites have
high concentrations of fine soil, which is caused by their being in the trans-Atlantic
transport path of dust from Africa (Prospero, 1996, 156889).
July 2009
9-40
DRAFT-DO NOT CITE OR QUOTE

-------
No urban-remote pair of coarse mass concentration maps is available because CSN
does not monitor coarse mass. In Malm et al. (2004, 156728) a map of annual mean coarse
mass concentration is shown for 2003 which includes the values for IMPROVE urban sites,
including two in the western U.S. with much more coarse mass than the nearby remote
areas monitoring sites (i.e., ~24 pg/m3 compared to ~9 pg/m3 for Phoenix, AZ, and ~6 pg/m3
compared to ~2 pg/m3 for Ptxget Sound, WA) and one eastern IMPROVE site at Washington,
DC with less coarse mass than the surrounding remote area values (~2 pg/m3 compared to
~4 pg/m3).
• IMPROVE Site
• Alaska
Hawaii
Puerto Rico /
Virgin Islands
— 5.79
¦ 3.72
3.30
2.89
2.48
2.07
1.65
1.24
10.83
0.41
0.00
pg/m3
¦ IMPROVE Urban Site
Source: Debell 12006,156388).
Figure 9-18. IMPROVE monitored mean organic mass concentrations for 2000-2004.
July 2009
9-41
DRAFT-DO NOT CITE OR QUOTE

-------
Hawaii
Alaska
• IMPROVE Site
¦ IMPROVE Urban Site
O
Puerto Rico/
Virgin Islands
12.4
6 47
5.75
5.03
4.31
3.60
2.88
2.16
11.44
0.72
0.00
|jg/m3
Source: Debell (2006,156388).
Figure 9-19. IMPROVE and CSN (STN) mean organic mass concentrations for 2000-2004.
• IMPROVE Site
0 96
0.44
0.39
0.34
0.29
0.25
0.20
0.15
10.10
0.05
0.00
pg/m3
Alaska
Hawaii
Puerto Rico I
Virgin Islands
v}
¦ IMPROVE Urban Site
Source: Debell (2006.156388).
Figure 9-20. IMPROVE mean EC concentrations for 2000-2004.
July 2009
9-42
DRAFT-DO NOT CITE OR QUOTE

-------
o
Hawaii
Alaska
• IMPROVE Site
¦ Urban IMPROVE Site
• e
Puerto Rico /
Virgin Islands
Source: Debell (2006.156388).
Figure 9-21. IMPROVE and CSN (STN) mean EC concentrations for 2000-2004.
• IMPROVE Site
Hawaii
¦ IMPROVE Urban Site
Puerto Rico /
Virgin Islands
3.08
¦ 1.44
1.28
1.12
0.96
0.80
0.64
0.48
_ 0.32
I 0.16
¦o.oo
|jg/m3
Source: Debell (2006,156388).
Figure 9-22. IMPROVE mean fine soil concentrations for 2000-2004.
July 2009
9-43
DRAFT-DO NOT CITE OR QUOTE

-------
A STN Site
• IMPROVE Site
•	• IMPROVE Site
¦ IMPROVE Urban Site
e ©
Puerto Rico/
Virgin Islands
IU.ida
0.14
0.00
4 80
1.29
0.72
0.57
(jg/m3
0.43
0.29
1.14
1.00
0.86
® Alaska
Hawaii
Source: Debell (2006,156388).
Figure 9-23. IMPROVE and CSN (STN) fine soil concentrations, 2000-2004.
Figure 9-25 shows the remote area coarse mass concentrations as measured by the
IMPROVE network. The pattern of high coarse mass concentrations from Oklahoma to
Iowa is comparable to the high concentrations in the desert southwest, though as shown in
Figure 9-11 it contributes a smaller relative share of the light extinction because of the
higher contributions to haze by particulate nitrate and sulfate in this agricultural region of
the country Comparing Figures 9"22 and 9-25 shows that the coarse mass and fine soil
concentration patterns are similar for the desert southwest but there is a much lower fine
soil to coarse mass concentration ratio for the agricultural center of the country suggesting
a regional difference in the size distribution of the coarse mass, perhaps due to differences
in suspendable soil materials.
9.2.3.4. Temporal Trends
Visibility trend analysis requires relatively long data records to avoid having
meteorologically driven interannual variability obscure more meaningful emissions "driven
air quality trends. A requirement for long-term data limits the number of monitoring sites
useful for trend analysis. Maps that show haze trends for IMPROVE sites for the 10-yr
period from 1995 through 2004 for the mean of the 20% best and the 20% worst haze days
where sites are required to have a minimum of 6 complete yr of data during the 10-yr
July 2009
9-44
DRAFT-DO NOT CITE OR QUOTE

-------
period are shown in Figures 9-26 and 9-27, respectively. The best haze days have improving
haze at most sites (32 of 47), no trend at several sites (10 of 47) and degrading visibility at
just one site (Great Sand Dunes, CO). The worst haze days have improving haze conditions
at several sites (13 of 47), no trend at most sites (30 of 47) and degrading visibility at a few
western sites (4 of 47).
July 2009
9-45
DRAFT-DO NOT CITE OR QUOTE

-------
Nitrates
Sulfates
Fresno

Missoula
I |

5aIt Lake City

WEST
Tulsa
¦
EAST
St. Louis
¦

Bi rmingham
XD

Indianapolis


Atlanta
X

Clweland


Charlotte
Richmond
=~
~ Regional
Contribution
Baltimore

~ Local
Contribution
New York City
_J	|

~i	1	1	r~
10
0 2 4 6 8 10 12
Annual Average Concentration
of Nitrates, ug/m3
Fresno
Missoula
Salt Lake City
Tulsa
St. Louis
Birmingham
Indianapolis
Atlanta
Cleveland
Charlotte
Richmond
Baltimore
New York City
XI
WEST
EAST
~	Regional
Contribution
~	Local
Contribution
i—r
2
~i—r
4 6 8
Annual Average Concentration
of Sulfates, ug/m3
"1—r
10
12
Carbon
Fresno
Missoula
Salt Lake City
Tulsa
St. Louis
Birmingham
Indianapolis
Atlanta
Cleveland
Charlotte
Richmond
Baltimore
New York City
0 2 4 6 8 10 12
Annual Average Concentration
of Carbon, ng/m3





WEST
I I
EAST



I I
|

|

| |

I
1 ~ Regional
Contribution
|
~ Local
—1 Contribution
I

i	1	1	1	1	1	1	r
Source: U.S. EPA (2004,190219)
Figure 9-24. Regional and local contributions to annual average PM2.5 by particulate SO42", nitrate
and total carbon (i.e., organic plus EC) for select urban areas based on paired IMPROVE
and CSN monitoring sites.
July 2009
9-46	DRAFT-00 NOT CITE OR QUOTE

-------
• IMPROVE Site
Hawaii
¦ IMPROVE Urban Site
Puerto Rico /
Virgin Islands
Alaska
¦ 22 1
8.92
7.93
6.94
5.95
4.96
3.96
2.97
11.98
0.99
0.00
(jg/rrv
Source: Debell (2006,156388).
Figure 9-25. IMPROVE mean coarse mass concentrations for 2000-2004.
Eight- ten- and sixteen-yr trends analysis conducted for the Western Regional Air
Partnership (WRAP) as part of the Causes of Haze Assessment (htto 7/www.wrauair.or)
show that improving trends for the 20% best haze days for the sites in the western U.S.
generally correspond to improving trends for all of the major components with the exception
of particulate nitrate. Trends assessment for the worst haze days at western sites show
consistent reductions in particulate S( )r- , but otherwise have mixed increasing and
decreasing haze component trends, many of which are not statistically significant.
Substantial interannual and shorter term spatial and temporal wildfire activity variations
have been shown to have a significant impact on the variability of haze in the Western U.S.
(Park et al., 2006, 190469; Spracklen et al., 2007, 190485). Edgerton et al. (2004, 156413)
showed a decreasing trend in PMa.e of about 18% (corresponding to 1 pg/m3 to 2 pg/m3) for
four urban-rural paired SEARCH sites in the Southeastern U.S. corresponding to similar
reductions in SO#" and carbonaceous particulate.
July 2009
9-47
DRAFT-DO NOT CITE OR QUOTE

-------
iMoosehorn

Greai Sai d Dunes
Weminuche
Smoky Mtns
¦ Handel
Petrified Forest

Ikefenokee
Improving Trend, p<=0.05
Improving Trend, 0.05
-------
Bridger
Improving Trend, p<=0.05
Improving Trend, 0.05
-------
than 10%. Others have shown similar decreasing particulate S( )r concentration trends
and a correspondence in trends between SO2 emissions and particulate SO42- concentration
by region (EPA, 2004, 056905; Holland et al., 1999, 092051).
Holland et al. (1999, 092051) developed and compared NOx emissions trends from
1989 to 1995 to corresponding trends in total nitrogen concentration (defined as particulate
nitrate plus gaseous nitric acid) for the eastern U.S. (states between Louisiana to
Minnesota and further east) based on data from 34 rural CASTNet dry deposition
monitoring sites. They found a decrease in nitrogen median values of about 8% associated
with a decrease of 5.4% in non-biogenic NOx emissions. Trends in haze associated with
particulate S( )r and nitrate concentrations should correspond fairly well with trends in
their concentration due to the simple relationship between concentration and light
extinction at any relative humidity. However, nitrogen as defined in the Holland et al.
(1999, 092051) trend analysis includes the nitrogen from particulate nitrate and gaseous
nitric acid, but nitric acid does not contribute to light extinction. For situations with limited
atmospheric ammonia or elevated temperatures, trends in nitrogen may be principally in
nitric acid with no net change in nitrate light extinction. Alternately with abundant
ammonia and low temperatures the trend in nitrogen may be principally in particulate
nitrate and the nitrate component of haze.
July 2009
9-50
DRAFT-DO NOT CITE OR QUOTE

-------
Western US
North Eastern US
1.15 5
m 0.75
0.5 	1	1—1	1	1	1	1	1	1	1	 0.46
1988 1990 1992 1994 1996 1998
South Middle US
3 	1	1	1	1	1	1	1	1	1	1	 54
1988 1990 1992 1994 1996 1998
South Eastern US
» 4.5
6.75
5,825
.2
M
3,375 |
O
m
2.25
1S88 1990 1982 1994 199® 1938
Sulfate Ion
1988 1S80 1992 1184 1988 1S98
SO: Emissions
Source: Malm et al. (2002,156727).
Figure 9-28. Ten-yr trends in the 80th percentile particulate S042" concentration based on IMPROVE
and CASTNet monitoring and net SO2 emissions from the National Emissions Trends
(NET) data base by region of the U.S.
Ten-yr trends (1994-2003) of particulate nitrate contribution to light extinction during
the 20% worst haze conditions conducted as part of the Causes of Haze Assessment (see the
link in Table 9"l) are shown in Figure 9-29. This indicates that haze from particulate
nitrates is increasing across the western U.S. at a rate of several Mm"1 per year in parts of
California and at a rate of several tenths of an Mm 1 across the Four-Corners states. A
similar particulate nitrate trends map (not shown here) for the 20% best haze conditions
that is available at the same web site shows decreasing particulate nitrate contribution to
light extinction at nearly all of the western monitoring sites. While statistically significant,
these trends for both the 20% worst and 20% best haze periods are influenced by an
unexplained nationwide period of depressed nitrate concentrations measured by the
July 2009
9-51
DRAFT-DO NOT CITE OR QUOTE

-------
IMPROVE network during a 4-yr period from the winter of 1996-97 through the winter of
2000-01. Extensive examinations of plausible monitoring methodological explanations have
failed to offer any evidence that the data are invalid (McDade, 2006, 192075), but no
satisfactory atmospheric or emissions-related explanation has been offered to account for
this four-yr depression of nitrate. Similar analyses of particulate nitrate haze trends are
not available for the rest of the country.
Maps of remote-area 8-, 10-, and 16-yr trends for carbonaceous and crustal PM species
based on IMPROVE monitoring are available for the western U.S. conducted as part of the
Causes of Haze Assessment. Generally these show a broad range of results (i.e., a mixture
of statistically significant upward or downward trends and insignificant trends often with
neighboring sites having opposing trends) that vary considerably depending on the number
of years selected (i.e., 8, 10, or 16) and whether trends are for the best, worst, or middle of
the haze distribution data. The scatter in these results is undoubtedly due to the high
inter annual variability and varying locations of wildfire and wind-suspended dust
emissions that dominate the remote-area concentrations of the carbonaceous and crustal
PM species in the western U.S.
July 2009
9-52
DRAFT-DO NOT CITE OR QUOTE

-------
Source: http://www.coha.dri.edu/
Figure 9-29. Map of 10-yr trends (1994-2003) in haze by particulate nitrate contribution to haze for
the worst 20% annual haze periods. The orientation, size and color of the arrows
indicate the direction, magnitude and statistical level of significance of the trends.
These consistent upward trends may be a misleading result due to an unexplained
sampling issue (see text for additional information).
9.2.3.5. Causes of Haze
In order to attribute haze to emissions from individual sources, source types, or source
regions, scientists generally apply any of a number of receptor and air quality simulation
modeling approaches and when using multiple approaches they reconcile the results of each
using a weight-of-evidence methodology Commonly this methodology has been applied to
the extensive datasets generated by special studies designed to estimate source-receptor
relationships for a few receptor locations or for individual emission sources (Pitchford et al.,
1999, 156873; Pitchford et al., 2005, 156874; Schichtel et al., 2005, 156957). More recently
the Regional Planning Organizations (RPOs) have sponsored extensive regional haze source
attribution assessments using weight-of-evidence methodologies to reconcile attribution
results for virtually all of the remote-area IMPROVE sites to support the development of
State Implementation Plans for the RHR. Additionally, a number of recent urban special
Legend
WN Slope
ft 0.064 - 0.093
ft 0.094 - 0.170
Virgin Islands
Hawaii
July 2009
9-53
DRAFT-DO NOT CITE OR QUOTE

-------
studies, including those sponsored by EPAPM Supersites program (Solomon and Hopke,
2008, 156997), have addressed the causes of and sources contributing to urban excess PM
concentrations above region concentrations. Attribution results uncertainties are generally
larger than those of the measurements upon which they are based because they also
include model uncertainties and assumptions, as well as issues of representativeness (e.g.,
How well does the data used in the analysis represent the typical emissions, air quality and
meteorological conditions of interest?). As such it is advisable to treat the attribution
results reported below as semi-quantitative.
The relative importance of the PM species that contribute to haze varies by region of
the U.S. and time of year as seen in Figures 9-9, 9-10 and 9-11, above. Generally haze in the
western half of the U.S. is not dominated by any one or two PM species. In the eastern half
of the U.S., SO42_, especially during summer and nitrate during the winter in the Midwest
are the dominant haze species. As described above, urban haze can be viewed as a
composite of the regional and local contributions where local contributions seem to be
dominated by carbonaceous and to a lesser extent nitrate and crustal PM components.
There have been far fewer urban investigations that explicitly consider visibility impacts,
though there are numerous studies of urban PM source attribution. The order of discussion
below on the cause of haze is by region beginning in the western U.S. and proceeding to the
east, analogous to dominant air flow patterns across the lower 48 states and will include
information from urban studies along side those of remote-area haze investigations.
Based on modeling of an episode (September 23-25, 1996) in the California South
Coast Air Basin (SCAB) and another episode (January 4-6, 1996) in the San Joaquin Valley
(SJV) by Ying and Kleeman (2006, 098359), about 80% of the particulate SC>42~ for both
regions are from upwind sources (i.e., thought to be principally from offshore sources
including marine shipping, long-range transport and natural marine sources), with most of
the remaining associated with diesel and high-sulfur fuel combustion. Kleeman et al. (1999,
011286), using a combination of measurements and modeling, showed that the upwind
particulate SO r source region for the SCAB was over the Pacific Ocean (confirmed by
measurements on Santa Catalina Island) and that these particles subsequently grew with
accumulation of additional secondary aerosol material, principally ammonium nitrate as
they traversed the SCAB. The majority of the nitric acid that forms particulate nitrate in
the SCAB is from diesel and gasoline combustion (-63%), while much of the ammonia is
from agricultural sources (-40%), catalyst equipped gasoline combustion (-16%), and
July 2009
9-54
DRAFT-DO NOT CITE OR QUOTE

-------
upwind sources (-18%). The majority of the OC found in SCAB was attributed in this study
to primary emissions by transportation-related sources including diesel (-13%) and
gasoline (-44%) engines and paved road dust (-12%). At the Fullerton site in the middle of
the SCAB the concentration of locally generated organics is roughly double that of the
locally generated nitrates (-5.6 pg/m3 compared to -2.4 pg/m3), while at Riverside on the
east edge of the SCAB and near the large agricultural sources of ammonia emissions, the
particulate nitrate concentrations are nearly double that of organic PM (-17 pg/m3
compared to -10 pg/m3).
Ying and Kleeman (2006, 098359) showed that during the winter 1996 episode in the
SJV most of the nitric acid that forms particulate nitrate is from upwind sources (-57%)
with diesel and gasoline combustion contributing most of the rest (30%), while much of the
ammonia is from upwind sources (-39%) and a combination of area, soil and fertilizer
sources (-52%). In an assessment of PM particle size and composition in the SJV during the
winter of 2000-2001, Herner et al. (2006, 135981) showed that fresh emissions of
carbonaceous PM from combustion sources in urban locations (Sacramento, Modesto, and
Bakersfield, CA) move quickly from ultrafine particle size (i.e., diameter -0.1 pm) to
accumulation mode by condensation with accumulation mode (i.e., diameter -0.5 pm)
particles, and that secondary nitrate particle formation occurs preferentially on the surface
of hydrated ammonium sulfate particles during the afternoon when gas-phase nitric acid is
at peak photo-chemical production from NOx. Given the abundance of ammonia emissions
and low ambient temperatures, particulate nitrate production in this way is only limited by
the availability of nitric acid. Due to the cool winter conditions there was little secondary
organic aerosol production during this study. Sea salt was shown to dominate the larger
coarse particle mode during on-shore wind at the background coastal monitoring site at
Bodega Bay, north of San Francisco, CA.
Using a regression analysis to find the dependence of particulate SOr concentration
measured over a 3-yr period (2000-2002) at 84 western IMPROVE monitoring sites on the
modeled transport trajectories to the sites for each sample period, Xu et al. (2006, 102706)
were able to infer the source regions that supplied particulate SOr in the western U.S.
Among the source regions included in this analysis was the near coastal Pacific Ocean (i.e.,
a 300 km zone off the coast of California, Oregon, and Washington). Up to half of the
particulate S( )r measured at Southern California monitoring sites was associated with
this source region. As shown in Figure 9"30 the zone of impact from this source region
July 2009
9-55
DRAFT-DO NOT CITE OR QUOTE

-------
included large regions of California, Arizona, and Nevada. The authors made the case that
high sulfur content fuel used in marine shipping and port emissions may be largely
responsible. As a result, the WRAP RPO emissions inventory was modified to include
marine shipping and a Pacific Offshore source region was added to source attribution by air
quality simulation modeling.
The SOr attribution results of the WRAP air quality modeling (results available on
the Technical Support System (TSS) website, see Table 9-1 for the web-link) credit the
Pacific offshore source region with somewhat smaller contributions than those from the
trajectory regression work by Xu et al. (2006, 102706) with concentrations at the peak
impact site in California that are about 45% compared to 50% by regressions and even
greater differences for more distant monitoring sites. Based on the modeling attribution the
Pacific offshore source region was responsible for 10%-20% of the nitrate measured in
Southern California.
^ •
1
Contribution of the Pacific Coast to Sulfate Concentration
<0.1
O.t - 0.2
0.2 - 0.4
¦	0.4 - 0.6
¦	0.6-0.8
¦	0.8- 1.4
Source: Xuetal. (2006,102706).
Figure 9-30. Contributions of the Pacific Coast area to the ammonium sulfate (Llgjm3)
at 84 remote-area monitoring sites in western U.S. based on trajectory regression for all
sample periods from 2000-2002 (dots denote locations of the IMPROVE aerosol
monitoring sites).
A coordinated effort by federal, state, and county air quality organizations to
determine the causes of haze in the Columbia River Gorge (a deep and narrow gap in the
July 2009
9-56
DRAFT-DO NOT CITE OR QUOTE

-------
Cascade Mountains on the Washington/Oregon border) through extensive multiyear
measurements and high spatial resolution air quality modeling of typical episodes
demonstrated the multitude of emission sources that contribute to its impairment
(Pitchford et al., 2008, 180159). During the summer, gorge winds are generally from the
west and relatively dry More than half of the haze during a typical summer episode is from
a combination of international and other distant sources (-22% at the western end of the
gorge) plus regional natural sources including wildfire and secondary organic PM from
biogenic emissions (-39% at the eastern end of the gorge). The Portland/Vancouver
metropolitan area was responsible for a significant amount of the haze during the summer
(-20% on in the western end of the gorge), while sources within the gorge were responsible
for a moderate amount of haze (-6% and -9% at the western and eastern ends of the gorge).
The wind is much more often from the east during the winter. The highest haze conditions
in the gorge are during the winter and are associated with fog conditions that rapidly
convert precursor gaseous emissions of NOx and SO2 from local and regional combustion
sources and NH4 from local and regional agricultural activities to secondary nitrate and
SO42- PM that persist as a post-fog intense haze. Contributions by these sources east of the
gorge contribute -57% of the haze on the eastern end of the gorge, with half of the nitrate
and SO r particulate from electric utility emissions and most of the rest from
transportation sources. Other sources contributing during the winter haze at the eastern
end of the gorge are from sources outside the modeling domain (i.e., most of Washington
and Oregon) and within the gorge (-23% and -10%, respectively).
An assessment of concurrent measurements at the nearby Mt. Hood IMPROVE
monitoring site (45 km south of the Columbia River at 1531m ASL) shows that Columbia
River Gorge haze conditions and especially the wintertime high nilrale/SO r contributions
to haze are not typical of the generally higher elevation remote areas of the region
(Pitchford et al., 2008, 180159) . However Gorge-like high wintertime nitrate and SO42- are
found at the Hells Canyon IMPROVE site, which is similarly situated in a narrow canyon of
the Snake River almost 400 km east of the Gorge (from the VIEWS web site, see Table 9-1),
implying that there may be a substantial vertical concentration gradient during winter in
this complex terrain.
Several example monitoring locations distributed across the northern and southern
portions of the western U.S. have been selected to illustrate the attribution results from the
WRAP-sponsored attribution analysis tools that estimate the relative responsibility for
July 2009
9-57
DRAFT-DO NOT CITE OR QUOTE

-------
haze of the various PM species by source region and source type. The selected sites include
Olympic NP (NP), WA; Yellowstone NP, WY; and Badlands NP, SD across the north, and
San Gorgonio Wilderness (W), CA; Grand Canyon NP, AZ and Salt Creek W, NM across the
south as shown in Figure 9"31.
Olympic N.P
Grand Canyon N.P.
2000-2004 Baseline Average
20% Worst Days
IMPROVE Aerosol Extinction (Mm-1)
Ammonium Sulfate
^ Ammonium Nitrate
Organic Material
^ Elemental Carbon
Soil
(Aj Coarse Material
20 60	100
* <3 years
Yellowstone N.P.
San Gorgonio W.
Badlands N.P.
Salt Creek W.
Source: From the TSS website.
Figure 9-31. Shows the IMPROVE monitoring sites in the WRAP region with at least three yr of valid
data and identifies the six sites selected to demonstrate the apportionment tools. Pie
diagrams show the composition for the mean of the 20% worst haze conditions and the
mean light extinction (Mm1) by the size of the circle (see figure key).
WRAP-sponsored CMAQ modeling for 2002 used virtual tracers of SOs and NOx
emissions that tracked the source region and category through the transport and
transformation processes to particulate IKS#" and nitrate. This was used to produce pie
diagrams of particulate SO#-* and nitrate attribution results by source region for each of
July 2009
9-58
DRAFT-DO NOT CITE OR QUOTE

-------
these sites as shown in Figure 9-32 (produced using the TSS, see Table 9-1). Based on these
sites, over half of the particulate SO r in remote areas of the Pacific coastal states is from
outside of the U.S. (Pacific offshore and outside of the domain). The outside of the domain
values were derived by simulating the fate of the boundary condition concentrations, which
for the WRAP air quality modeling were obtained using output from the GEOS-CHEM
global air quality model (Fiore et al., 2003, 047805). The SO42- fraction from the region
labeled outside of domain was approximately uniform throughout the western U.S. with
site-to-site variation in the fraction caused mostly by variations in total SO r
concentration. The more northerly sites have impacts from Canadian emissions, while the
southern sites have impacts from Mexican emissions. Half of the Salt Creek, New Mexico
SO42- is from the domestic source emissions further to the east, which also contribute about
20% to Badlands particulate S( ) r concentrations. A breakout of the emission sources from
within the WRAP region by source type (not shown) has most of the emissions from point
sources, with the combination of motor vehicle, area and wildfire emissions contributing
from a few percent at the furthest eastern sites to about half at San Gorgonio.
By comparison, the particulate nitrate is much more from domestic regional emission
sources, with ~60%-80% being from emissions within the WRAP region. For the west coast
sites about 25% of the nitrate is from a combination of Pacific offshore emissions (i.e.,
marine shipping) and outside domain regions. Canadian emissions are responsible for about
10%-30% of the particulate nitrate for the three northern sites, but Mexican emissions do
not contribute appreciably to particulate nitrate for the three southern sites. Motor vehicles
are the largest contributing NOx source category responsible for particulate nitrate for
these six WRAP sites, with a combination of point, area and wildfire source categories also
contributing from about 10%"50% of the WRAP regional emissions.
WRAP only used the virtual tracer approach to investigate source locations and
categories for SO2 and NOx emissions. A different type of virtual tracer modeling tool was
used to track the various OC compounds and sort them into three groups for 2002. The first
group labeled primary organics includes all of the organics that are emitted directly as PM
from any source type and location. The second group labeled anthropogenic secondary
organics is PM produced in the atmosphere by aromatic VOCs. The third category labeled
biogenic secondary organics is PM produced in the atmosphere by biogenic VOCs. Organic
PM in the biogenic secondary category includes those that would functionally be considered
man-made emissions (e.g., those from agricultural crops and urban landscapes), though in
July 2009
9-59
DRAFT-DO NOT CITE OR QUOTE

-------
most remote areas of the west these man-made VOC emissions are small compared to those
of the natural biogenic sources. Figures 9-33, 9-34 and 9-35 show the monthly averaged
apportionment of organic PM for the six selected monitoring locations.
July 2009
9-60
DRAFT-DO NOT CITE OR QUOTE

-------





(a)
Olympic NP
Yellowstone NP
Badlands NP
WW





Pacific Offshore
2002
2002
2002
CENRAP
Sulfate 1.1 uq/m3
Sulfate 0.6 ua/m3
Sulfate 1.2 ua/m3
Eastern U.S.





Canada





Mexico
San Gorgonio W
Grand Canyon NP
Salt Creek W
Outside Domain
2352
2002
2002

Sulfate 0
.8 ua/m3
Sulfate 0.5 ua/m3
Sulfate 1A uq/m3






(b)
Olympic NP
Yellowstone NP
Badlands NP




/

WRAP
Pacific Offshore
2002
2002
2002
CENRAP
Nitrate 1.6 ua/m3
Nitrate 0.6 ua/m3
Nitrate t .2 ua/m3
Eastern U.S.





Canada





Mexico
San Gorgonio W
Grand Canyon NP
Salt Creek W
/ 1
Outside Domain
2002
2002
2002

Nitrate 1.5 ua/m3
Nitrate 0.4 ua/m3
Nitrate 0.5 ua/m3

Figure 9-32. Particulate SO42' (a} and nitrate (b) source attribution by region using CAMx
modeling for six western remote area monitoring sites: top left to right Olympic HIP, WA;
Yellowstone RIP, WY; Badlands NP, SD; bottom left to right San Gorgonio W, OA; Grand
Canyon NP, AZ; and Salt Creek W, NM. WRAP includes ND, SD, WY, CO, NM and all
states further west. CENRAP includes all states east of WRAP and west of the
Mississippi River including MN. Eastern U.S. includes all states east of CENRAP. The
Pacific Offshore extends 300km to the west of CA, OR, and WA. Outside Domain refers
to the modeling domain, which extend hundreds of kilometers into the Pacific and
Atlantic Oceans and from Hudson Bay Canada to just north of Mexico City. This
figure was assembled from site-specific diagrams produced on the TSS web site (see
Table 9 1) for 2002.
July 2009
9-61
DRAFT-DO NOT CITE OR QUOTE

-------
Organic Aerosol for All Days
Class I Area - Olympic NP, WA
3.20
^ 2.80
CO
f,2-40
2.00
0
1° 1.60
8 1.20
° 0.80
0.40
0.00
Organic Aerosol for All Days
Class I Areas - San Gorgonio W, CA: San Jacinto W, CA
2.80
f? 2-40
-i
g1 2.00
| 1.60
In
1	1.20
o
3 0.80
0.40
0.00
Source: From the TSS website, see Table 9-1.
Figure 9-33. Monthly averaged model predicted organic mass concentration apportioned into primary
and anthropogenic and biogenic secondary PM categories for the Olympic RIP (top) and
San Gorgonio Wilderness (bottom) monitoring sites.
Based on the modeling results for these six sites and confirmed by measurements
(see, e.g., Figure 9-10), a west to east decreasing gradient of organic mass exists with
annual concentrations from ~2 pg/m3 for the coastal state sites to ~1 pg/m3 for the
intermountain west sites to less than 1 pg/m3 for the sites just east of the Rocky Mountains,
discounting the large fire impacts for July at Yellowstone NP which raised its annual mean
to ~2 pg/m3. At all of these remote-area sites anthropogenic secondary PM is estimated to
be a small fraction of the organic mass, with the largest fractional contribution at the San
Gorgonio monitoring site immediately downwind of the major Southern California urban
areas, yet having less than 10% of the monthly mean organic mass from anthropogenic
Lllllllllllll
CN	CO
i/j	to
CO	CO
Anthro. Secondary
Biogenic Secondary
| Anthro. & Bio. Primary
July 2009
9-62
DRAFT-DO NOT CITE OR QUOTE

-------
secondary PM. Of the six selected monitoring sites, San Gorgonio has the highest fraction of
the organic PM from primary emissions (-57%), followed by Yellowstone (-55%), then the
two eastern-most sites (Badlands -42% and Salt Creek 41%), and with Grand Canyon and
Olympic national parks the lowest fraction by primary emissions (-37%). Yellowstone NP
would have had the lowest fraction of organic PM by primary emissions had it not been for
the month of July (the 11 month mean is 29%) when wild fire smoke contributed. Results
from recent chamber and field studies, and modeling would seem to call into question
apportionment of primary and secondary carbon done by traditional air quality model
simulations of OC, as described above, due to the combined effects of extensive evaporation
of semivolatile primary emissions when diluted and to photochemical reactions of low
volatility gas phase species that substantially increases the amount of secondary organic
PM (Robinson et al., 2007, 191975).
July 2009
9-63
DRAFT-DO NOT CITE OR QUOTE

-------
14.00
12.00
10.00
8.00
6.00
4.00
2.00
0.00
Organic Aerosol for All Days
Class I Areas - Grand Teton NP, WY: Red Rock Lakes NWRW, MT: Teton W, WY: Yellowstone NP.WY
l 11 l 1 1 l
*3-	Lj-J	OIi
1.80
1.50
1.20
0.90
0.60
0.30
0.00
Organic Aerosol for All Days
Hopi Point#!





1


















LJ-J	CD	h-	CO	CD	O
] Anthro. Secondary
Biogenic Secondary
] Anthro. & Bio. Primary
Source: From the TSS website, see Table 9-1.
Figure 9-34. Monthly averaged model predicted organic mass concentration apportioned into primary
and anthropogenic and biogenic secondary PM categories for the Yellowstone RIP (top)
and Grand Canyon (Hopi Point) (bottom) monitoring sites.
July 2009
9-64
DRAFT-DO NOT CITE OR QUOTE

-------
0.80
0.70
co 0.60
-I
3 0.50
| 0.40
| 0.30
C
8 0.20
0.10
0.00
Organic Aerosol for All Days
Class I Area - Badlands NP, SD
I MM
ilA ii	i i i" i
Organic Aerosol for All Days
Class I Area - Salt Creek NWRW, NM
£ 0.40
0.20
0.00
I I I I I I I I I I I I I
] Anthro. Secondary
Tfr	W	CD	hw
Biogenic Secondary
J Anthro. & Bio. Primary
Source: From the TSS website, see Table 9-1.
Figure 9-35. Monthly averaged model predicted organic mass concentration apportioned into primary
and anthropogenic and biogenic secondary PM categories for the Badland RIP (top) and
Salt Creek Wilderness (bottom) monitoring sites.
Radiocarbon (14C) dating techniques were used to group ambient PM carbon into fossil
and contemporary source categories at 12 IMPROVE monitoring sites across the U.S., 8 of
which are in the WRAP region (Schichtel et al., 2008, 156958). Results of this work showed
that contemporary carbon accounts for about half the carbon in urban areas, 70%"97% in
near-urban areas (i.e., San Gorgonio) and 82%-100% in remote areas. Comparing these
radiocarbon dating results with the WRAP virtual tracer modeling results for organic
aerosol (above), and presuming that the modeled anthropogenic secondary organic is fossil
carbon and the biogenic secondary is contemporary carbon, suggests that a large fraction of
July 2009
9-65
DRAFT-DO NOT CITE OR QUOTE

-------
the model-determined regional primary organic PM is from contemporary carbon sources
(e.g., smoke from wildfires).
Schichtel et al. (2008, 156958) compared radiocarbon measurements at two sets of
urban/rural paired sites in the west (Mount Rainer/Seattle, and Tonto/Phoenix). Figure 9".'Hi
shows that most of total carbon urban excess (i.e., urban site concentration minus the
regional site concentration) in the summer is from fossil carbon sources (87% and 79%,
respectively), while in the winter there is a surprisingly high fraction of the urban excess at
both sites that is from contemporary carbon sources (41% and 47%, respectively). This
implies that urban, and therefore anthropogenic, activities generate almost as much PM2.5
carbon from contemporary sources (e.g., residential wood combustion) as from the fossil
sources during the winter for these two western urban areas.
July 2009
9-66
DRAFT-DO NOT CITE OR QUOTE

-------
6
5
4
3
2
1
0
6
5
4
3
2
1
0
Summer
I
ro o
O |-
|— TO
O
1*-
CO
£

O)
a.
ID
in






II


C0


X

E
O)
a.
CO
C\
o
E
0
c
o
o
5 -
E 4
3 3 H
|2
° 1
0
Summer
I
sS
H ®
O
ss
3



1

CO
E

2





CO

Ol



II

V

xs
¦
xs




Fossil
i
E
0)
c
O
o
Winter
TO S
(J
c/)
o
E
a)
c
o
O
Winter
Source: Schichtel et al. (2008,156958).
Comparison of carbon concentrations between Seattle (Puget Sound site) and Mt. Rainer
(left) and between Phoenix and Tonto (right) showing the background site concentration
(gray) and the urban excess concentration (black) for total, fossil and contemporary
carbon during the summer and winter studies.
9-67
DRAFT-DO NOT CITE OR QUOTE

-------
Proctor Maple
0.94
Great Smoky
Mountains
0.90
Tonto
Fraction Contemporary
Puget Sound
0.51
Mount
Sula
Briganline
0.79
O
| > 0.90 + 0.24
0.75 * 0.90 + 0.22
0.60 - 0.75 * 0.21
0.45 - 0.60 + 0.20
¦ < 0.45 + 0.19
Phoenix
0 56 0.84
O O
Proctor Maple
0.87
Rocky
Mountains
0.93
Tonto
Fraction Contemporary
Puget Sound
0.53
Mount
Kairif:r
BrlganOne
0.78
O
¦	> 0.90 t 0,15
0.75 - 0.90 + 0.14
0.60 - 0.75 t 0.13
0.45 - 0.60 + 0.12
¦	< 0.45 + 0.11
Phoenix
0.49 0.71
Q O
Great Smoky
Mounts ins
0.82
O
Figure 9-37. Average contemporary fraction of PM2.5 carbon for the summer (top) and winter
(bottom), estimated from IMPROVE monitoring data (6/04 to 2/06) based on EC/TC ratios.
The contemporary values from radiocarbon dating for the 12 monitoring sites are
indicated in by colored circles with the site names. Color contours are shown to aid in
showing sites with similar values. Site locations are indicated by circles for remote area
sites and triangles for urban sites.
July 2009
9-68
DRAFT-DO NOT CITE OR QUOTE

-------
Contemporary carbon estimates for all of the IMPROVE network monitoring sites for
data from two summer seasons (June, July and August, 2004/2005) and two winter seasons
(December, 2004/2005, January and February, 2005/2006) were calculated from the
measured EC to total carbon (EC/TC) ratios using the 12-site empirical relationship
between radiocarbon determined contemporary carbon fraction and IMPROVE measured
EC/TC ratio (Schichtel et al., 2008, 156958). The results are displayed in color contour maps
in Figure 9"37, which also shows the radiocarbon determined contemporary carbon for the
12 sites. The lowest contemporary carbon estimates (<60%) in both seasons are for urban
areas. In the rural West, most of the sites have over 90% of their PM carbon from
contemporary carbon sources during the summer and from 60%-over 90% during the
winter. In the rural East, most of the sites have 45%-90% of their PM carbon from
contemporary carbon sources during the summer and from 60%-over 90% in the winter.
Schichtel et al. (2008, 156958) showed a strong relationship between the site-averaged
EC/TC ratios and the site-averaged fraction of fossil carbon separately for the summer and
winter data sets (i.e., R2 of 0.71 and 0.87, respectively). Using regression analysis they
estimated that the summer and winter EC/TC ratios associated with purely fossil carbon
were 0.35 ± 0.039 and 0.46 ± 0.028, respectively, and for purely contemporary carbon the
EC/TC ratios were 0.12 ± 0.011 and 0.19 ± 0.0095. These ratios are shown to be consistent
with corresponding ratios from the literature for source testing primary fossil and
contemporary combustion sources respectively. They are also shown to be consistent with
the 90 percentile value of the EC/TC ratio from the urban IMPROVE monitoring sites (0.41
and 0.44 for summer and winter) and the 10th percentile values of the EC/TC ratio for
remote areas IMPROVE monitoring sites (0.07 and 0.16 for summer and winter), which
they argue are dominated by fossil and contemporary carbon, respectively.
The largest sources of contemporary carbon are primary emissions from biomass
burning and secondary organic aerosol from biogenic precursor gases (e.g., terpenes from
conifer forests). Schichtel et al. (2008, 156958) estimated the 12-site overall contribution by
secondary organic PM to the summer contemporary carbon fraction as 36 ± 6.4% by
assuming the EC/TC ratio for contemporary carbon during the winter represented the ratio
of primary emissions only (i.e., no secondary organic PM formation in the winter) and that
the EC/TC ratio for primary emissions is independent of seasons. This approach should
provide a lower bound estimate of the secondary OC species. The same method applied to
the fossil carbon fraction yielded an estimate of 23 ± 10% of the fossil carbon PM from
July 2009
9-69
DRAFT-DO NOT CITE OR QUOTE

-------
secondary organic formation in the atmosphere during the summer. These estimates
correspond to over 40% of the contemporary and over 35% of the fossil OC being from
secondary PM formation.
July 2009
9-70
DRAFT-DO NOT CITE OR QUOTE

-------
100.00
90.00
80.00
70.00
60.00
"E
o 50.00
Q_
40.00
30.00
20.00
10.00
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Area - Olympic NP, WA
o.oo l-1
WRAPTSS-3Q1J2EB
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility1 Days
Class I Area - Olympic NP, WA
70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00















¦
Jl













Bb

WRAPTSS-301flE5 ^

-------
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Areas - San Gorgonio W, OA: San Jacinto W, OA
no. no
HU LIU
a u. i J u
7IJ. IJIJ
aij.uu
" 50.00
40.00
30.00
20.00
10.00
0.00 ¦
WRAP TSS - 301J2IJB
CA - 2000-04
2018 PRP
PO - 2000-04
2018 PRP
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days
Class I Areas - San Gorgonio W, CA: San Jacinto W, CA
100.00
90.00
a U. U U
7IJ. IJIJ
bU.UU
" 50.00
40.00
30.00
20.00
10.00
0.00 ¦
WRAP TSS - 301J2IJB
CA - 2000-04
2018 PRP
PO - 2000-04
2018 PRP
Bra Dust ¦ Road Dust | On-Road Mobile ~ WRAP Area O&G ~ Biogenic BAnthroFire
~ Fugitive Dust Off-Road Mobile ~ Off-Shore []Area	~ Natural Fire |] Point
Source: From the TSS website (see Table 9-1).
Figure 9-39. Results of the weighted emissions potential tool applied to primary OC emissions (top)
and EC emissions (bottom) for the baseline and projected 2018 emissions
inventories for San Gorgonio W. Only source regions (WRAP states and other regions)
with the largest estimated contributions are shown (i.e., California and Pacific Off-Shore
from left to right). The scale is normalized (i.e., unitless) one over distance weighted
emissions multiplied by trajectory residence time.
July 2009
9-72
DRAFT-DO NOT CITE OR QUOTE

-------
WRAP applied a weighted emissions potential analysis tool that combined gridded
emissions data with back-trajectory analysis that simulated the transport pathway to the
various monitoring sites to infer likely source region and emission categories for the 20%
best and 20% worst haze conditions for each of the IMPROVE PM speciation monitoring
locations in the West. Unlike the virtual tracer approach that uses a full regional air
quality simulation model, this method does not explicitly account for chemistry or removal
processes and it does not incorporate the sophisticated dispersion estimates (i.e., it uses one
over distance weighting for dispersion), so it should be considered a screening tool that has
been found to be helpful in identifying the likely sources contributing to haze. More
information on this approach is available elsewhere (see the link to the TSS in Table 9-1).
Primary organic and EC PM species results from the weighted emissions potential tool for
the worst 20% haze days using the 2000-2004 base years' emissions and trajectories, and
the same trajectories with 2018-projected emissions for each of the six selected western
monitoring locations are shown in Figures 9"38 through 9-43.
For Olympic NP (Figure 9"38), most of the primary organic as well as most of the EC
PM is likely to be from the state of Washington during the worst haze days. This is because
the multi-day trajectories that transport emissions on its worst days tend to be short
(within 200 km based on maps available on TSS, see Table 9-1). Area sources, which include
emissions from residential wood heating, watercraft, non-mobile urban and other sources
too small to be labeled as point sources, are the big contributors to primary organic, while
on- and off-road mobile emissions plus area sources are large contributors to the EC at
Olympic NP The 2018 projected growth in area sources and decrease in emissions of mobile
source emissions is anticipated to increase the haze by primary organic while reducing the
haze by EC at Olympic NP The same analysis applied to San Gorgonio (Figure 9-39), is
similar in that the majority of the emissions with the potential to contribute to primary
organic and EC PM is from the home state, California in this case. However, the likely
importance of natural fire emissions for carbonaceous PM species sites is substantially
greater at San Gorgonio W. than it was for Olympic NP
The weighted emissions potential results applied to Yellowstone NP and Grand
Canyon NP (Figures 9-40 and 9"4l) show the likely dominance of natural fire emissions in
the intermountain western U.S. to primary organic and EC PM during worst haze
conditions for these two locations. Numerous states have emissions that have the potential
to contribute noticeably to these carbonaceous species, due to relatively long multi-day
July 2009
9-73
DRAFT-DO NOT CITE OR QUOTE

-------
trajectories (500 km to 1000 km) on worst haze days, though for both sites the home state
has the greatest potential based on this inverse distance weighting approach. On- and
off-road mobile sources in Arizona and California have significant potential to contribute to
Grand Canyon carbonaceous particles, especially EC concentrations, probably due to some
of the trajectories being over the populated areas of these two states to the south and
southwest of Grand Canyon.
July 2009
9-74
DRAFT-DO NOT CITE OR QUOTE

-------
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Areas - Grand Teton NP, WY: Red Rock Lakes NWRW, MT: Teton W, WY: Yellowstone NP, WY
70.00
60.00
50.00
1 40.00
30.00
20.00
10.00
WRAP TSS - 3Q1JHDB
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days

70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00
WRAP TSS - 30U2DB
BWBDust ¦ Road Dust | On-Road Mobile ~ WRAP Area O&G ~ Biogenic BAnthroFire
~ Fugitive Dust Off-Road Mobile ~off-Shore []Area	~ Natural Fire |]Point
Source: From the TSS website (see Table 9-1).
Figure 9-40. Results of the weighted emissions potential tool applied to primary OC emissions (top)
and EC emissions (bottom) for the baseline and projected 2018 emissions inventories for
Yellowstone RIP. Only source regions (WRAP states and other regions) with the largest
estimated contributions are shown (i.e., California, Idaho, Montana, Oregon, Utah,
Washington, and Wyoming from left to right). The scale is normalized (i.e., unitless) one
over distance weighted emissions multiplied by trajectory residence time.
July 2009
9-75
DRAFT-DO NOT CITE OR QUOTE

-------
60.00
50.00
40.00
IE
^ 30.00 -H
20.00
10.00
0.00
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Area - Grand Canyon NP, AZ
WRAP TSS- 4/1ODS
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days
Class I Area - Grand Canyon NP, AZ
50.00
45.00 +
40.00
35.00
30.00 4-
25.00
20.00
15.00
10.00
5.00 --
0.00
WRAP TSS -
|WB Dust | Road Dust | On-Road Mobile QWRAP Area O&G D Biogenic Q Anthro Fire
~ Fugitive Dust ~ Off-Road Mobile ~ Off-Shore DArea	~ Natural Fire |] Point
Source: From the TSS website (see Table 9-1).
Figure 9-41. Results of the weighted emissions potential tool applied to primary OC emissions (top)
and EC emissions (bottom) for the baseline and projected 2018 emissions inventories for
Grand Canyon RIP. Only source regions (WRAP states and other regions) with the largest
estimated contributions are shown (i.e., Arizona, California, Mexico, New Mexico,
Nevada, Oregon, Pacific Off-shore and Utah from left to right). The scale is normalized
(i.e., unitless) one over distance weighted emissions multiplied by trajectory residence
time.
July 2009
9-76
DRAFT-DO NOT CITE OR QUOTE

-------
30.00
27.00
24.00
21.00
18.00
0)
2 15.00
a)
CL
12.00
9.00
6.00
3.00
0.00
WRAPTSS-aaiflDDS
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Area - Badlands IMP, SD


T—r

=
•¦=1-
Cl
"¦=}¦
Cl
¦•=1-
CL

Cl

Cl
-3-
Cl

Cl
¦*3"
o
Ctl
O
C£
O
CE
O
C£
O
CC
O

O

O
o
Cl
o
Cl
O
Cl
o
Cl
O
Cl
O
CL
o
LL
o
o
CO
o
CO
o
CO
o
CO
O
CO
O
CO
o
CO
o
o

o

o

o

O

o

o

o
CM
O
CN
CN
O
CN
CN
O
CN
CN
O
CN
CM
O
CN
CN
o
CN
CN
o
CN
CN
M

<



2

o

to

Q

1—
<



<

LU

o

3



2
a:
O
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days
Class I Area - Badlands NP, SD
30.00
27.00
24.00
21.00
18.00
0J
£ 15.00
CD
CL
12.00
9.00
6.00
3.00
0.00
mU

' 	
~n
Q

— =
aa
Ct O 0£ O
M
<
WRAP TSS-3GimE
<
'w'
CL
"¦=*-
Cl
"=1"
Cl

CL

Cl

Cl
¦¦=1-
Cl
-¦=1-
Cl
¦¦=1-
Cl
-¦=1-
Cl

Cl
•*=*"
Cl
Ltl
o
or
O
or
O
ctr
o
C£
O
Ctl
o
IT
o
CE
o
cc
O

O
C£
O
Ltl
Cl
o
Cl
o
CL
O
CL
o
CL
O
CL
o
CL
o
CL
o
Cl
O
CL
o
Cl
o
CL
CO
o
CO
o
CO
o
CO
o
CO
O
CO
o
CO
o
CO
o
CO
O
CO
o
CO
o
CO

o

o

o

o

O

o

o

o

o

o

o

o
CN
o
CN
O
CN
o
CN
O
CM
O
CN
O
CM
O
CN
O
CN
O
CN
O
CN
O
CN
i
CN
i
CN
i
CN
r
CN
i
CN
i
CN
i
CN
i
CN
i
CN
i
CN
i
CN

o

CO

Q

1—

Ci

>

CC

Cl

1—

<




C_5

3
LU





-Z.

IZ

o

to

Ci

£


HWBDust
Fugitive Dust
| Road Dust
Off-Road Mobile
| On-Road Mobile WRAP Area O&G Biogenic | Anthro Fire
Off-Shore [ Area	C Natural Fire | Point
Source: From the TSS website (see Table 9-1).
Figure 9-42. Results of the weighted emissions potential tool applied to primary 0C emissions (top)
and EC emissions (bottom) for the baseline and projected 2018 emissions inventories for
Badlands RIP. Only source regions (WRAP states and other regions) with the largest
estimated contributions are shown (i.e., Arizona, California, Canada, CenRAP, Colorado,
eastern U.S., Idaho, Montana, North Dakota, Nevada, Oregon, South Dakota, Utah,
Washington, and Wyoming from left to right). The scale is normalized (i.e., unitless) one
over distance weighted emissions multiplied by trajectory residence time.
July 2009
9-77
DRAFT-DO NOT CITE OR QUOTE

-------
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Area - Salt Creek NWRW, NM








n

































a
u





	











































-
J
L.




















=r

sJ
—	—
M-


~~
nn


40.00

36.00

32.00

28.00

24 00
r

a."

a5
20.00

16.00

12.00

8.00

4.00

0.00
WRAP TSS-3G1OT33
36.00
33.00
30.00
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days
Class I Area - Salt Creek NWRW, NM
27.00
24. UU
21 .00
S 18.00
is.uu
12.00


























































nn

— —
= =
	 			
an


1=1
<
- 3C31SDS
<
U
o
u
X
LU
IE
O
BWB Dust | Road Dust | On-Road Mobile _| WRAP Area O&G Biogenic BAnthroFire
Fugitive Dust Off-Road Mobile Off-Shore	Area	Natural Fire [J] Point
July 2009
9-78
DRAFT-DO NOT CITE OR QUOTE

-------
Source: From the TSS website (see Table 9-1).
Figure 9-43. Results of the weighted emissions potential tool applied to primary OC emissions (top)
and EC emissions (bottom) for the baseline and projected 2018 emissions
inventories for Salt Creek W. Only source regions (WRAP states and other regions) with
the largest estimated contributions are shown (i.e., Arizona, California, CenRAP,
Colorado, eastern U.S., Idaho, Mexico, Montana, New Mexico, Nevada, Oregon, Utah,
and Wyoming from left to right). The scale is normalized (i.e., unitless) one over distance
weighted emissions multiplied by trajectory residence time.
For the most easterly of the selected WRAP sites, Badlands NP and Salt Creek, the
weighted emissions potential results for primary organic and EC (Figures 9-42 and 9-43)
show potential contributions from a greater number of states and multi-state regions than
for selected sites further to the west. This may be due in part to trajectories associated with
worst haze conditions for these two sites being moderately long (-500 km) and in multiple
directions. Natural fire emissions have the greatest potential to contribute to organic
species PM at Badlands NP, but are less likely to be dominant at Salt Creek or at either site
in its contribution to EC PM concentrations. The contributions by emissions from area and
mobile sources from the home states and states to the east (Central States Regional Air
Partnership states are labeled "CEN" in the figures) are potentially greater than by natural
fire! this is especially true for contributions to EC PM.
WRAP applied the weighted emissions potential tool to assess likely source types and
regions contributing to coarse mass concentrations. The results for the six selected
monitoring sites (not shown) are as follows. Most dust at Olympic NP is likely to be from
fugitive dust sources in Washington state, while at San Gorgonio it is likely more from road
dust with smaller amounts from fugitive dust sources. The amount from wind-blown dust is
small for both of these far westerly sites. Wind-blown dust is likely the largest source
contributing to coarse mass at Grand Canyon NP, Badlands NP and Salt Creek Wilderness
with most of it originating in the home-state for those sites. The weighted emissions
potential results for coarse mass at Yellowstone are different from those of the other five
selected sites in that Idaho and Montana each have a higher potential to contribute to
coarse mass on the worst haze days than the home state (Wyoming), and that wind-blown
and road dust both contribute substantially as does fugitive dust and natural fire.
In another WRAP-sponsored effort to better understand the causes of remote area
haze in the western U.S., each of the worst haze days for all western IMPROVE monitoring
sites where dust (defined as the sum of coarse mass and fine soil PM) was the largest
July 2009
9-79
DRAFT-DO NOT CITE OR QUOTE

-------
contributor to light extinction was separately assessed to categorize the most likely dust
source (Kavouras et al., 2007, 156630; Kavouras et al., 2009, 191976) and the Causes of
Haze Website - see Table 9-1). Elemental composition was used to assess the likelihood that
the dust was associated with long-range transport from Asia. A regression analysis at each
site between dust concentrations and coincident local wind speed was used to generate
site-specific estimates of local windblown dust for each sample period. Finally, back
trajectory analysis combined with maps constructed of wind erosion potential
(i.e., developed by combining soil types and land cover classifications) are used in a manner
similar to the weighted emissions potential analysis to identify the likelihood of regionally
transported wind-blown dust as the source. These assessments were conducted on each of
the 610 so-called "worst dust haze days" at 70 monitoring sites for data from 2001-2003 to
classify each day by its likely contributions from Asian dust, local windblown dust, upwind
transport and undetermined. The undetermined category includes those sample periods
that failed to be classified into one of the other three source categories suggesting that
mechanically suspended dust activities (e.g., unpaved road dust, agricultural, construction
and mining activities) may be responsible.
Of the 610 "worst dust haze days" at the 70 WRAP monitoring sites, 55 sample
periods are classified as Asian dust influenced, almost exclusively in the spring! 201 sample
period are classified as local windblown dust, mostly in the spring but some in all seasons!
240 sample periods are classified as upwind transported dust, with a broader seasonal
distribution centered on summer and few instances during winter! and 114 are in the
undetermined category with most in summer and least in winter. Most dust days occurred
in the deserts of Arizona, New Mexico, Colorado, western Texas and southern California,
and these were dominated by local and regionally transported wind-blown dust (e.g., 84%
for Salt Creek W). Asian dust caused only a few of the worst dust days during the 3-yr
assessment period, though it is an important source (i.e., 10-40% of the worst dust days) for
sites in the more northern regions of the West with greater vegetative land-cover where
local and regionally transported wind-blown dust was infrequent. The frequency of worst
dust events classified as undetermined was greatest for sites in the vicinity of large urban
and agricultural areas such as those in California and southern Arizona.
July 2009
9-80
DRAFT-DO NOT CITE OR QUOTE

-------
Sulfate Haze Source Attribution
Organics + LAC + Nitrates +
/ Fine Soil + Coarse
~	Carbon ¦ Other Mexico
¦ Texas	¦ Eastern US
~	Western US ¦ Other
Clear Air
August 9
September 9
October 9
Source: Pitchford et al. (2005,156874)
Figure 9-44. BRAVO study haze contributions for Big Bend NP, TX during a 4-mo period in 1999.
Shown are impacts by various particulate SO*2" sources, as well as the total light
extinction (black line) and Rayleigh or clear air light scattering.
July 2009
9-81
DRAFT-DO NOT CITE OR QUOTE

-------
"ACAD 1
2.5
2.0
1.7
1.5
1.2
1.0
ug/rn3
2004 Annual
amrnN03f
Figure 9-45. Maps of spatial patterns for average annual particulate nitrate measurements (top), and
for ammonia emissions for April 2002 from the WRAP emissions inventory (bottom).
July 2009
9-82
DRAFT-DO NOT CITE OR QUOTE

-------
N02
3.125
1.875
2.500
1.500
WRAP 36k base02a All Sources Emissions	WRAP 36k base02a All Sources Emissions
2002 Yearly Total	2002 Yearly Total
PAbJE	December 31,2002 0:00:00	PAbJE	December 31,2002 0:00:00
mcnc	Min= 0.000 at (4,1), Max= 4.995 at (130,68)	MCnc	Min= 0.000 at (4,1), Max= 4.041 at (130,68)
1 5.000112
4.375
¦ 3.000112
I 2.625
3.750
2.250
1.875
1.125
1.250
0.625
0.000 1
LOG(tons/yeaf)
0.750
0.375
0.000 1
LOG(tons/yeat)
Figure 9-46. Maps of spatial patterns of annual NO (left) and NO2 (right) emissions for 2002 from the
WRAP emissions inventory.
Source attribution of the particulate S()¦>'-- contribution to haze at Big Bend NP, TX
was a primary motivation for the BRAVO study. Schichtel et al. (2005, 156957) showed that
during the four-month field monitoring study (July-October, 1999) SOa emissions sources in
the U.S. and Mexico were responsible for -55% and -38% of the particulate SO;:; .
respectively. Among U.S. source regions, Texas was responsible for -16%, eastern U.S. -
30%, and the western U.S. -9%. A large coal fired power plant, the Carbon facility in
Mexico, just south of Eagle Pass, TX, was responsible for about -19%, making it the largest
single contributor. Pitchford et al. (2005, 156874) put these results into the context of other
component contributions to regional haze, plus seasonal and longer-term variations in haze
by particulate components. Figure 9-44 shows the temporal variation of the contributions
by the various SO2 emissions source regions plus the Carbon facility during the BRAVO
study period. The largest particulate SOr peak haze periods are dominated by infrequent
large contribution by emission sources in TX and eastern U.S., while Mexican sources
including the Carbon facility are more frequent contributors to haze, but at generally lower
light extinction values. Particulate nitrate contributions to haze at Big Bend NP are among
the lowest measured in the U.S. (-3% of light extinction on average and for worst haze
episodes).
Nitrate concentrations are a significant contributor to light extinction further to the
north of Texas in the center of the country. While SO r can be in particulate form though
not fully neutralized by ammonia, nitric acid from NOx emissions requires neutralization
July 2009
9-83
DRAFT-DO NOT CITE OR QUOTE

-------
by ammonium to become particulate ammonium nitrate. One way to explore the causes of
the Midwest nitrate bulge is to compare its spatial distribution with the spatial
distributions of NOx and ammonia emissions. Figure 9-45 shows a map of the annual
average particulate nitrate concentrations (top) with a map of ammonia emissions directly
below it. Animal agriculture is responsible for most of the ammonia emissions in the
Midwest. The striking similarity between the ambient particulate nitrate concentration and
the ammonia emissions spatial patterns with regional maximum centered on Iowa is in
contrast to the NOx (i.e., NO + NO2) emissions spatial patterns, shown in Figure 9-46. NOx
emissions are high over a broad region of the country associated with the larger population
densities and greater numbers of fossil fuel electric generation plant generally to the east of
the Midwest nitrate bulge. While both ammonia and nitric acid are needed to form
particulate ammonium nitrate, the maps suggest the Midwest nitrate bulge is due
primarily to the abundance of free ammonia (i.e., the amount beyond what is required to
neutralize the acidic particulate SOr ). By contrast the region to the east of the Midwest
nitrate bulge should have plenty of nitric acid given the higher emissions of NOx, but
apparently has a deficiency of free ammonia. The few eastern monitoring sites with locally
high particulate nitrate (near southeastern PA) are located within a small region of high
density animal agricultural that shows up as a high ammonia emissions region in
Figure 9-45. Note that California's South Coast and Central Valley have both high
ammonia and high NOx emissions, explaining the high particulate nitrate contribution to
haze there.
To better understand the role of ammonia in the formation of the Midwest nitrate
bulge, the Midwest RPO and Central States Regional Air Partnership deployed a
measurement program from late 2003 through early 2005 at 10 locations (9 rural and 1
urban) in the region (see Figure 9*47) to monitor particulate SO r , nitrate, and ammonium
ions, plus the precursor gases sulfur dioxide, nitric acid, and ammonia (Kenski et al., 2004,
192078; Sweet et al., 2005, 180038). These data have been used as input for thermodynamic
equilibrium modeling to assess the changes in PM concentrations that would result from
changes to precursor concentrations (Blanchard and Tanenbaum, 2006, 192181; Blanchard
et al., 2007, 098659). Blanchard and Tanenbaum (2006, 192181) and Blanchard et al. (2007,
098659) conclude that the current conditions at nine of the ten sites are near the point of
transition between the precursor species (nitric acid and ammonia) that limits the
formation of particulate nitrate. If excess ammonia increases, either by greater ammonia
July 2009
9-84
DRAFT-DO NOT CITE OR QUOTE

-------
emissions or by anticipated decreases in SO2 emissions, then nitric acid concentration
would need to be reduced (via lower NOx emissions) in order to reduce the particulate
nitrate concentration.
Given the comparability of particulate S() r and nitrate with regard to their light
extinction efficiencies, their visibility impacts are proportional to the sum of their mass
concentrations. A reduction in SO r caused by SO2 emission reductions would reduce the
particulate S() r concentration, though according to the thermodynamic equilibrium
modeling for these sites the particulate nitrate concentration will be increased somewhat.
However the total particulate SO r plus nitrate concentration would be reduced so
visibility impacts would be decreased. At current ammonium concentrations the predicted
response of changes to SC>42~ and nitric acid concentrations (i.e., SO2 and NOx emissions
changes) are similar in respect to the resulting magnitude of changes to the total
particulate S() r plus nitrate concentrations. At all but two sites the total particulate S() r
plus nitrate concentrations would decrease if either ammonia or nitric acid where reduced.
Figure 9-47. Midwest ammonia monitoring network.
A further degree of complications in understanding the response of particulate nitrate
to changes in precursor concentrations results from the temperature and humidity
dependence of the partition between particulate ammonium nitrate and the disassociated
Source: Kenski et al. (2004,192078)
July 2009
9-85
DRAFT-DO NOT CITE OR QUOTE

-------
gaseous nitric acid and ammonia. This dependence causes seasonal and even diurnal
differences in the expected responses of particulate nitrate concentrations to changes in
precursor concentrations. As expected during the colder times of the year the total
particulate concentrations are more sensitive to changes in ammonia and nitric acid
concentrations than during the warmer seasons when SO-r ¦ concentrations are greater.
As shown in Figure 9-<18, results of an air transport assessment to identify emission
source areas associated with high particulate nitrate at five monitoring locations in the
East (four remote-area sites and Toronto, Canada) implicate the high ammonia emissions
region of the Midwest as a common source region (Canada-US Air Quality Committee,
Subcommittee 2: Scientific, 2004, 190519). This assessment does not preclude local sources
of the precursor gases responsible for particulate ammonium nitrate, but does suggest that
long-range transport of particulate nitrate or ammonia from the high emissions region of
the Midwest is also contributing to eastern nitrate episodes.
Upwind Probability Fields for Ammonium Nitrate
Lye Brook
Source: Canada-U.S. Air Committee (2004,190519).
Figure 9-48. Upwind transport probability fields associated with high particulate
nitrateconcentrations measured at Toronto, Canada; Boundary Water Canoe Area, MN;
Shenandoah NP, VA; Lye Brook, VT; and Great Smoky Mountains RIP, TN.
July 2009
9-86
DRAFT-DO NOT CITE OR QUOTE

-------
In a similar air transport assessment for measurements at Underbill, VT and at
Brigantine, N J, Hopke et al. (2005, 156567) identified separate regions associated with
particulate SO#~ accompanied by trace particulate components associated with coal
burning (e.g., selenium) and accompanied by trace particulate components associated with
oil burning (e.g., vanadium). As shown in Figure 9-49, the coal-burning related particulate
S( ) i- for these two monitoring sites is associated with long-range transport from the Ohio
River Valley, while oil-burning related particulate SOi2 is from more nearby emissions in
the high population region of coastal New York, New Jersey, Massachusetts, and
Connecticut.
COAL
Hopke, et al. (2005,156567).
Figure 949. Trajectory probability fields for periods with high particulate SO42' measured at
Underbill, VT and Brigantine, NJ (shown as white stars) associated with oil-burning
trace components (left) and with coal-burning trace components (right). Shown for
comparison are the interpolated SO2 emissions areal density contours for oil combustion
sources (emissions times 10) and coal combustion sources, displayed as yellow and red
contour lines, respectively.
The Regional Aerosol Intensive Network (RAIN) was established by MANE-VU to
generate enhanced continuous visibility, plus fine particle mass and composition monitoring
data at a string of three monitoring locations along the transport path from the Ohio River
Valley to coastal Maine (NESCAUM, , 156802). The dominant role of particulate SO*- in
the northeast is well demonstrated by a scatter plot of RAIN data that shows the
July 2009
9-87
DRAFT-DO NOT CITE OR QUOTE

-------
relationship between particulate S() r extinction, calculated using the IMPROVE
algorithm plotted against directly measured particle light scattering for hourly data over a
eight month period beginning in July 2004 at the Acadia NP, ME monitoring site (see
Figure 9-50). Particulate S() r explains 90% of the variance in particulate light scattering
even though it is responsible for only about 64% of the total light extinction (annual
averaged value from the VIEWS web site). Adding the contribution by the second-largest
regional contributor to light extinction, particulate OC with about 14%, does not improve
the variance explained, but does increase the slope to 0.78. The noticeable difference
between these two plots is that the particulate S() r alone underestimates light scattering
during low haze periods (points on the plot are below the regression line for light scattering
<70Mm"1), while the agreement is improved with the addition of particulate OC
contributions to haze (regression slope is nearer to one and reduced bias for low haze
periods). This implies that particulate S() r and any other co-varying PM species are
largely responsible for the highly impacted periods, while OC and other co-varying PM
species contribute more during the less extreme haze periods. Particulate nitrate
contribution to light extinction at Acadia is about 10% on average.
Nephelometer Bip versus SULFATE Bs
Nephelometer Bsp vs (SULFATE + OMC) Bsp
y = 0.64x - 1.87
R' = 0.90

2-Hour Nephelometer Bsp (1/Mmeters)
9
UJ
»-
<



y = 0.78x + 4.57

R2 = 0.89 OS ~
n = 1627 °

0
^ S-"b





££> 0

Mb
0
SO 100 150 200 250 300 3.
2-Hr Nephelometer (1/Mmeters)
Source: RAIN Preliminary Data Analysis Report (NESCAUM,, 156802)
Figure 9-50. Scatter plots of particulate SO42 (left) and particulate nitrate and organic mass (right)
versus nephelometer measured particle light scattering for Acadia RIP, ME.
Particulate nitrate concentrations are considerably lower in the SOr "dominated
warmer southeastern U.S. than in the Northeast and upper Midwest. Blanchard et al.
(2007, 098659) conducted thermodynamic equilibrium modeling on data from the eight
July 2009
9-88
DRAFT-DO NOT CITE OR QUOTE

-------
SEARCH monitoring sites and found that total particulate nitrate plus S( )r is much more
responsive to changes in SO2 concentrations than to changes in nitric acid concentrations,
which in turn is more responsive than changes in ammonia concentrations.
The VISTAS RPO commissioned an emissions sensitivity study using CMAQ
modeling on winter and summer 2009 emissions projected from the 2002 emissions
inventory (NCDENR, 2007, 156798). Figure 9"51 contains bar plots for two North Carolina
Class I areas that indicate projected changes in light extinction for the worst haze day due
to 30% emissions reductions by particulate species, source types and location across the
Southeastern states modeling domain (i.e., as far west as Texas, as far north as
Pennsylvania, as far south as the Florida Keys, and as far east as -300 km from the North
Carolina coast). Great Smoky Mountains in the southern Appalachian Mountains has the
greatest sensitivity to changes by SO2 emissions from electrical generation units (EGU) and
to a lesser extent other SO2 emission sources in the region. Reductions of NOx emissions
from ground or point sources are not nearly as effective as SO2 reductions in reducing the
light extinction at Great Smoky Mountains. This is due principally to the worst days at
Great Smoky Mountains occurring during the summer, when temperatures are too high to
support high particulate nitrate concentrations. For the same reason, ammonia emission
reductions are also ineffective. Swanquarter Wilderness, NC is a coastal location where
some of the worst haze days are during the winter and include contributions from
particulate ammonium nitrate. Both SO2 and ammonia emissions reductions would be
effective at reducing worst haze days at the Swanquarter Wilderness, though NOx
emissions are not as effective presumably because the atmosphere is ammonia-limited for
particulate nitrate production.
July 2009
9-89
DRAFT-DO NOT CITE OR QUOTE

-------
Great Smoky Mtns, TN (20% Worst Days)

-12.00 r

-10.00 --

-8.00 -

-6.00 -
1
-4.00 -
CD
<
-2.00 -
0.00 ^
//,•,

¦	Bio VOC.
~	Anthro VOC
¦	BCs
¦	MRPO
¦	M-VU
~	CEN
¦	VISTAS
¦	wv
~	VA
~	TN
~	SC
¦	NC
¦	MS
~	KY
~	GA
¦	FL
~	AL
Swanquarter, NC (20% Worst Days)
¦	Bio VOC
~	Anthro VOC
¦	BCs
¦	MRPO
¦	M-VU
~	CEN
~	VISTAS
¦	WV
~	VA
~	TN
~	SC
¦	NC
~	MS
~	KY
~	GA
~	FL
~	AL
Source: NCDENR (2007.156798).
Figure 9-51. CMAQ air quality modeling projections of visibility responses on the 20% worst haze
days at Great Smoky Mountains RIP, l\IC (top) and Swanquarter Wilderness, l\IC (bottom)
to 30% reductions from a projected 2009 emission inventory of visibility-reducing
pollutants by source category and geographic areas.
9.2.4. Urban Visibility Valuation and Preference
The Clean Air Act §302(h) defines public welfare to include the effects of air pollution
on "...visibility, ... and personal comfort and wellbeing." Though good visibility conditions in
Class I (e.g., NPs) and wilderness areas have long been recognized as important to the
public welfare (see discussions in EPA (2004, 056905; 2005, 090209) and Chestnut and
July 2009
9-90
DRAFT-DO NOT CITE OR QUOTE

-------
Dennis (1997, 014525), visibility conditions in urban areas also contribute to the public
welfare. Although visibility impairment may be caused by either natural or manmade
conditions (or both), it is only impairment that occurs as a result of air pollution (either
alone or in combination with water vapor or other atmospheric conditions) that can be
mitigated by regulations such as the RHR (40 CFR 51.300 through 309) or the Secondary
National Ambient Air Quality Standards (NAAQS). The term visual air quality (VAQ) is
used here to refer to the visibility effects caused solely by air quality conditions, so for
example it excludes the reduced visibility caused by fog. Visibly poor air quality causes
people to be concerned about substantive health risks, but degraded VAQ adversely affects
people in additional ways. These include the aesthetic and wellbeing benefits of better
visibility, improved road and air safety, and enhanced recreation in activities like hiking
and bicycling. Because the human health impacts of air pollution are assessed under the
Primary NAAQS, it is necessary to separate out these non-health components associated
with the visibility condition produced by a given amount of air pollution when assessing the
need for additional regulation to protect the public welfare effect of visibility under the
Secondary NAAQS. The degree to which previous human preference and valuation studies
for VAQ have adequately made this distinction and separation is an important issue in
applying results from available studies in a Secondary NAAQS (or benefits estimation for
any policy effecting VAQ) context. The remainder of this discussion is focused on those
aesthetic and wellbeing qualities associated with a given VAQ in urban areas.
The term "urban visibility" is used to refer to VAQ throughout a city or metropolitan
area. Urban visibility includes the VAQ conditions in all locations that people experience in
their daily lives, including scenes such as residential streets and neighborhood parks,
commercial and industrial areas, highway and commuting corridors, central downtown
areas, and views from elevated locations providing a broad overlook of the metropolitan
area. Thus urban visibility includes VAQ conditions in major cities and smaller towns and
encompasses all the VAQ an individual resident sees on a regular basis. Visibility
conditions in urban and suburban locations are therefore distinct from visibility in rural or
wilderness settings such as the Class 1 areas defined by the Clean Air Act, which include
NPs and similar natural settings.
Visibility has direct significance to people's enjoyment of daily activities and their
overall wellbeing. Visibility conditions can be described both as an aesthetic quality as well
as a scientifically measurable set of atmospheric conditions. Due to the subjective nature of
July 2009
9-91
DRAFT-DO NOT CITE OR QUOTE

-------
aesthetics, people's preferences with respect to visibility are difficult to express or quantify,
but people have expressed in many different ways that they enjoy and value a clear view. A
number of social science disciplines have undertaken to link perceived urban visibility to an
array of effects reflecting the overall desire for good VAQ, and the benefits of improving
currently degraded VAQ. This wide range of diverse studies have identified types of benefits
of good VAQ in addition to those directly connected with air-pollution related health effects
such as respiratory diseases and premature mortality.
For example, psychological research has demonstrated that people are emotionally
affected by low VAQ such that their overall sense of wellbeing is diminished (e.g.,
Bickerstaff and Walker, 2001, 156271). Researchers have also shown that perception of
pollution is correlated with stress, annoyance, and symptoms of depression (Evans and
Jacobs, 1982, 179899; Jacobs et al., 1984, 156596; Mace et al., 2004, 180255). Sociological
research has demonstrated that VAQ is deeply intertwined with a "sense of place," effecting
people's sense of the desirability of a neighborhood quite apart from the actual physical
conditions of the area (e.g., ABT, 2002, 156186; Day, 2007, 156386; Elliot et al., 1999,
010716; Gilliland et al., 2004, 156471). Public policy research finds that people think it is
important to protect visibility, and accept the concept of setting standards to protect
visibility (e.g., ABT, 2001, 156185; BBC Research & Consulting, 2002, 156258; Ely et al.,
1991, 156417; Pryor, 1996, 056598). Finally, economic valuation research has measured the
amount of money that people are willing to pay to protect or improve both urban visibility
(e.g., summary review in Beron et al., 2001, 156270; Chestnut and Dennis, 1997, 014525)
and natural locations such as NPs and other locations defined by the Clean Air Act as Class
I visibility area (e.g., summary review in Chestnut and Dennis, 1997, 014525).
Urban visibility has been examined in two types of studies directly relevant to the
NAAQS review process: urban visibility preference studies and urban visibility valuation
studies. The purpose of the remainder of this section is to review four urban preference
studies, as well as one new urban visibility valuation study not previously discussed in
previous EPA Criteria Documents or OAQPS Staff Papers.
Both types of studies are designed to evaluate individuals' desire (or demand) for good
VAQ where they live, using different metrics to evaluate demand. Urban visibility
preference studies examine individuals' demand by investigating what amount of visibility
degradation is unacceptable while economic studies examine demand by investigating how
much one would be willing to pay to improve visibility.
July 2009
9-92
DRAFT-DO NOT CITE OR QUOTE

-------
9.2.4.1. Urban Visibility Preference Studies
One group of urban visibility research projects focused on identifying preferences for
urban VAQ without necessarily estimating the economic value of improving visibility. This
group of preference studies used a common focus group method to estimate the visibility
impairment conditions that respondents described as "acceptable." The specific definition of
acceptable was largely left to each individual respondent, allowing each to identify their
own preferences.
There are three completed studies that used this method, and one additional pilot
study (designed as a survey instrument development project) that provided additional
information (Table 9-2). The completed studies were conducted in Denver, Colorado (Ely et
al., 1991, 156417), two cities in British Columbia, Canada (Pryor, 1996, 056598), and
Phoenix, Arizona (BBC Research & Consulting, 2002, 156258). The pilot study was
conducted in Washington, DC (ABT, 2001, 156185).
Each study collected information in a focus group setting, presenting slides depicting
various visibility conditions. All four studies used photographs of a single scene from the
study's city! each photo included images of the broad downtown area and spreading out to
the hills or mountains composing the scene's backdrop. The maximum sight distance under
good conditions varied by city, ranging from 8 kilometers in Washington, DC to mountains
hundreds of kilometers away in Denver. Multiple photos of the same scene were used to
present approximately 20 different visibility impairment conditions. The Denver and
British Columbia studies used actual photographs taken in the same location to depict
various visibility conditions. The Phoenix and Washington, DC pilot study used
photographs prepared using the WinHaze software from Air Resource Specialists (ARS).
WinHaze is a computer-imaging software program that simulates visual air quality
differences of various scenes, allowing the user to "degrade" an original near-pristine
visibility condition photograph to create a photograph of each desired VAQ condition.
A common characteristic of the three visibility preference studies was that each were
conducted in the West where distant mountains were shown in the photograph used to elicit
local participant responses about visibility. Among other issues, the Washington D.C. pilot
study was the first step in a process to expand the results to other regions where typical
scenes may have different sensitivity to perceived visibility changes in PM air quality and
where participants may have different acceptable visibility preference values.
July 2009
9-93
DRAFT-DO NOT CITE OR QUOTE

-------
One notable finding of the three visibility preference studies and the one pilot study is
the general degree of consistency in the median preferences for an acceptable amount of
visibility degradation. The range of median acceptable preference values from the four
studies is 19-25 dv, the preferred measure of visibility impairment. Measured in terms of
visual range (VR), these median acceptable values are between 59 and 32 km.
Table 9-2.
Summary of urban visibility preference studies.


Denver, CO
Phoenix, AZ
2 British Columbia cities
Washington, DC (pilot)
Report Date
1991
2003
1996
2001
Duration of session

45 min
50 mins
2 h
Compensation
None (civic groups)
$50
None (class room exercise)
$50
# focus group sessions
17
27 total at 6 locations,
Including 3 in Spanish
4
1
# participants
214
385
180
9
Age range
adults
18-65 +
University students
27-58
Annual or seasonal
Wintertime
Annual
Summertime
Annual
# total scenes
presented
Single scene of downtown
with mountains in
background
Single scene of downtown and
mountains, 42 km maximum
distance
Single scene from each city
Single scene of DC Mall and
downtown, 8 km maximum sight
# of total visibility
conditions presented
20 conditions (+ 5
duplicates)
21 conditions (+ 4 duplicates)
20 conditions (10 each from each city)
20 conditions (+ 5 duplicates)
Source of slides
Actual photos taken between
9am and 3pm
WinHaze
Actual photos taken at 1 pm or 4pm
WinHaze
Medium of presentatior
i Slide projection
Slide projection
Slide projection
Slide projection
Ranking scale used
7 point scale
7 point scale
7 point scale
7 point scale
Visibility range
presented
11 to 40 dv
15 to 35 dv
13 to 25 dv (Chilliwack) 13.5 to 31.5 dv
(Abbotsford)
9 to 38 dv
Health issue directions
Ignore potential health
impacts; visibility only
Judge solely on visibility, do not
consider health
Judge solely on visibility, do not consider
health
Health never mentioned, "Focus
only on visibility"
Key Questions asked
a) Rank VAQ (1-7 scale)
a) Rank VAQ (1-7 scale)
a) Rank VAQ (1-7 scale)
a) Rank VAQ (1-7 scale)

b) Is each slide "acceptable"
b) Is each slide "acceptable"
b) Is each slide "acceptable"
b) Is each slide "acceptable"

c)"How much haze is too
much?"
c) How many days a yr would
this picture be "acceptable"

c) if this hazy, how many hs
would it be acceptable (3 slides
only)




d) valuation question
Mean dvfound
"acceptable"
20.3 dv
23 to 25 dv
— 23 dv(Chilliwack),
— 20 dv (range 20-25)


~ 19 dv(Abbotsford)

July 2009
9-94
DRAFT-DO NOT CITE OR QUOTE

-------
9.2.4.2.	Denver, Colorado Urban Visibility Preference Study
The Denver urban visibility preference study (Ely et al., 1991, 156417) was conducted
on behalf of the Colorado Department of Public Health and Environment (CDPHE). The
study conducted a series of focus group sessions with 17 civic and community groups in
which a total of 214 individuals were asked to rate slides. The slides depicted varying
values of VAQ for a well-known Denver vista, including a broad view of downtown Denver
with the mountains to the west composing the scene's background. The participants were
instructed to base their judgments on three factors^
1.	The standard was for an urban area, not a pristine national park area where the standards might be
more strict;
2.	The value of an urban visibility standard violation should be set at a VAQ value considered to be
unreasonable, objectionable, and unacceptable visually; and
3.	Judgments of standards violations should be based on visibility only, not on health effects.
Participants were shown 25 randomly ordered slides of actual photographs. The
visibility conditions presented in the slides ranged from 11 to 40 dv, approximating the 10th
to 90th percentile of wintertime visibility conditions in Denver. The participants rated the
25 slides based on a scale of 1 (poor) to 7 (excellent), with 5 duplicates included. They were
then asked to judge whether the slide would violate what they would consider to be an
appropriate urban visibility standard (i.e., whether the amount of impairment was
"acceptable" or "unacceptable"). The individual's judgment of a slide's VAQ and whether the
slide violated a visibility standard were highly correlated (Pearson correlation coefficient
>80%), as were the VAQ ratings and the yes/no "acceptable" response. The participant's
median response was that a visibility condition of 20.3 dv (extinction coefficient
bext = 76Mm1, or VR ~51 km) was judged as "acceptable." The CDPHE subsequently
established a Denver visibility standard at this value (defined as bext = 76Mm"1), based on
the median 50% acceptability findings from the study.
9.2.4.3.	Phoenix, Arizona Urban Visibility Preference Study
The Phoenix urban visibility preference study (BBC Research & Consulting, 2002,
156258) was conducted on behalf of the Arizona Department of Environmental Quality. The
Phoenix study patterned its focus group survey process after the Denver study. The study
July 2009
9-95
DRAFT-DO NOT CITE OR QUOTE

-------
included 385 participants in 27 separate focus group sessions. Participants were recruited
using random digit dialing to obtain a sample group designed to be demographically
representative of the larger Phoenix population. Focus group sessions were held at six
neighborhood locations throughout the metropolitan area to improve the participation rate.
Three sessions were held in Spanish in one region of the city with a large Hispanic
population (25%), although the final overall participation of native Spanish speakers (18%)
in the study was modestly below the targeted value. Participants received $50 as an
inducement to participate.
Participants were shown a series of 25 images of the same vista of downtown Phoenix,
with South Mountain in the background at a distance of about 40 km. Photographic slides
of the images were developed using WinHaze. The visibility impairment conditions ranged
from 15-35 dv (the extinction coefficient, bext, range was approximately 45 Mm"1 to 330 Mm"
1, or a visual range of 87-12 km). Participants first individually rated the randomly shown
slides on a VAQ scale of 1 (unacceptable) to 7 (excellent). Participants were instructed to
rate the photographs solely on visibility, and to not base their decisions on either health
concerns or what it would cost to have better visibility. Next, the participants individually
rated the randomly ordered slides as "acceptable" or "unacceptable," defined as whether the
visibility in the slide is unreasonable or objectionable. Better visibility conditions (15 dv and
20 dv) were judged "acceptable" by 90% of all participants. At 24 dv nearly half of all
participants thought the VAQ was "unacceptable," with almost three-quarters judging 26 dv
as unacceptable.
The Phoenix urban visibility study formed the basis of the decision of the Phoenix
Visibility Index Oversight Committee for a visibility index for the Phoenix Metropolitan
Area (Arizona Department of Environmental, 2000, 019164). The Phoenix Visibility Index
establishes an indexed system with 5 categories of visibility conditions, ranging from
"Excellent" (14 dv or less) to "Very Poor" (29 dv or greater). The "Good" range is 15-20 dv.
The environmental goal of the Phoenix urban visibility program is to achieve continued
progress through 2018 by moving the number of days in lower quality categories into better
quality categories.
9.2.4.4. British Columbia, Canada Urban Visibility Preference Study
The British Columbia urban visibility preference study (Pryor, 1996, 056598) was
conducted on behalf of the Ministry of Environment. The study conducted focus group
July 2009
9-96
DRAFT-DO NOT CITE OR QUOTE

-------
sessions that were also developed following the methods used in the Denver study.
Participants were students at the University of British Columbia, who participated in one
of four focus group sessions with between 7 and 95 participants. A total of 180 respondents
completed surveys (29 did not complete the survey).
Participants in the study were shown slides of two suburban locations in British
Columbia: Chilliwack and Abbotsford. Using the same general protocol as the Denver study
Pryor found that responses from this study found the acceptable level of visibility was 23 dv
in Chilliwack and 19 dv in Abbotsford. Pryor (1996, 056598) discusses some possible
reasons for the variation in standard visibility judgments between the two locations.
Factors discussed include the relative complexity of the scenes, potential bias of the sample
population (only University students participated), and the different amounts of
development at each location. Abbotsford (population 130,000) is an ethnically diverse
suburb adjacent to the Vancouver Metro area, while Chilliwack (population 70,000) is an
agricultural community 100 km east Vancouver in the Frazier Valley.
The British Columbia urban visibility preference study is being considered by the B.C.
Ministry of the Environment as a part of establishing urban and wilderness visibility goals
in British Columbia.
9.2.4.5. Washington, DC Urban Visibility Pilot Preference Study
The Washington, DC urban visibility pilot study (ABT, 2001, 156185) was conducted
on behalf of the EPA, and was designed to be a pilot focus group study, an initial
developmental trial run of a larger study. The intent of the pilot study was to study both
focus group method design and potential survey questions. Due to funding limitations, only
a single focus group session was held, consisting of one extended session with 9
participants. No further urban visibility focus group sessions were held in Washington, DC.
Due to the small number of participants, it is not possible to make statistical
inferences about the opinions of the general population. The study does, however, provide
additional useful information about urban visibility studies, potentially helping to both
better understand previous studies as well as design future studies.
The study also adopted the general Denver study method, modifying it as appropriate
to be applicable in an eastern urban setting which has substantially different visibility
conditions than any of the three western locations of the other preference studies.
Washington's (and the entire East) visibility is typically substantially worse than western
July 2009
9-97
DRAFT-DO NOT CITE OR QUOTE

-------
cities, and has different characteristics. Washington's visibility impairment is primarily a
uniform whitish haze dominated by sulfates, relative humidity values are higher, the low
lying terrain provides substantially shorter maximum sight distances, and many residents
are not well informed that anthropogenic emissions impair visibility on hazy days.
The Washington focus group session included questions on valuation, as well as on
preferences. The focus group was asked to state its preferences measured in an increase in
the general cost of living for certain increments of improvement in visibility on a typical
summer day. A general cost of living approach is one payment vehicle approach that can be
used in willingness to pay studies, especially for environmental issues arising from multiple
diverse emission sources (e.g., transportation, electricity generation, industry, etc.) making
a specific price increase potentially misleading.
The first part of the focus group session was designed to be an hour long, and was
comparable to the focus group sessions in the Denver and Phoenix studies. A single scene
was used; a panoramic shot of the Potomac River, Washington mall and downtown
Washington, DC. In the first part of the session people were asked to rate the VAQ of 25
photographs (prepared using WinHaze, and projected on a large screen), judge the
acceptability of visibility condition in each slide, and answer the valuation questions. The
second half of the session, however, was a moderated discussion session about the format
and content of the first phase of the session. In this moderated discussion, participants
were asked about their understanding of each question asked in the first half of the session.
Particular issues in designing a focus group session were also explored. Important
participant comments included:
1.	Participants had been asked how they reacted to the initial direction to base their answers only on
visibility, but health was never explicitly mentioned by the focus group moderator. Participants
strongly agreed with the decision to not mention that health effects are associated with visibility
impairment. They understood the directions as meaning they should ignore health issues, and said
their answers would have been different if they included health as well as visibility in their
judgments.
2.	Differentiating between haze and weather conditions was difficult. Weather was not discussed in
the focus group session, and the photographs were WinHaze altered photos with identical weather
conditions. Participants mentioned they were still confused about the role of weather and
humidity in the different visibility conditions presented in the photos.
July 2009
9-98
DRAFT-DO NOT CITE OR QUOTE

-------
3.	Questions about how many hours an impairment level would be acceptable were confusing. Most
participants were normally indoors during most of the day, so questions about duration of outdoor
conditions were difficult to answer.
4.	Participants strongly agreed that not mentioning the purpose of the study, or the sponsor, until the
very end (after all the questions were answered) was viewed as very important. Most felt this
information would have influenced their answers.
9.2.4.6. Urban Visibility Valuation Studies
The one recent urban visibility benefit assessment not included in earlier reviews is
"The Benefits of Visibility Improvement: New Evidence from the Los Angeles Metropolitan
Area" (Beron et al., 2001, 156270). Rather than a contingent valuation method (CVM)
technique used in the majority of other urban visibility valuation studies, Beron et al.
(2001, 156270) used a housing market hedonic technique. The housing hedonic methods
were used in previous urban visibility studies by Murdoch and Thayer (1988, 156788) and
Trijonis et al. (1985, 078468). A housing market hedonic study views a housing unit as
composed of a bundle of attributes, and uses housing sale price data from a large number of
units in a metropolitan area to estimate the value of each component. Hedonic pricing has
been used to estimate economic values for environmental effects that have a direct effect on
housing market values. It relies on the measurement of differentials in property values
under various environmental quality conditions including air pollution, visibility and other
environmental amenities such as access to nearby beaches and parks, as well as by physical
attributes of the house and attributes of the neighborhood.
Beron et al. (2001, 156270) obtained data on approximately 840,000 owner-occupied,
single family housing sales between 1980 and 1995 from the California South Coast Air
Basin (composed of Los Angeles and Orange Counties, and the portions of Riverside and
San Bernardino Counties in the greater metropolitan area). The real estate data included
information on the sale price of the house, 13 housing attributes (square footage, number of
bathrooms, etc.), 9 neighborhood attributes (percent poverty, school quality, FBI crime
index, etc.), and three air pollution variables: ozone, particulates (measured by total
suspended particulates, or TSP), and visibility. Visibility was measured as the annual
average of visual range, measured in miles, and was obtained from seven airports within
the study region. The visibility range was from 12.4 miles (Los Angeles International
July 2009
9-99
DRAFT-DO NOT CITE OR QUOTE

-------
Airport, 1991) to 31.9 miles (Palm Springs Airport, 1995). Ozone data (39 monitors) and
TSP data (40 monitors) were obtained from the South Coast Air Quality Management
District. Annual mean values for each year were calculated for ozone and TSP.
Beron et al. (2001, 156270) presented results for a hypothetical basin-wide 20%
visibility improvement, or an increase from 15.3 to 18.4 miles, which is equivalent to
approximately 27.6 dv to 25.8 dv. The initial results reflect the change in the purchase price
of a house associated with this difference in VAQ, which can be interpreted as a present
value of a stream of annual values over the lifetime of the house. The authors therefore
selected a time horizon (30 yr) and an interest rate (8%) to calculate an annual per
household benefit per dv ranging from $484 to $1,756. The Beron results are higher than
the CVM-based values summarized in Chestnut and Dennis (1997, 014525), which ranged
from $12 to $132 per dv. It should be noted that the $132 CVM values cited by Chestnut
and Dennis (1997, 014525) is from a study in the Los Angeles area (Brookshire, 1979,
156298). The Beron et al. (2001, 156270) results are also higher than the Trijonis et al.
(1990, 157058) hedonic study in the Los Angeles area, which had a range of $134 to $360
per dv per year. All values reported here are in terms of 1994 prices.
A critical question for all urban visibility valuation studies is the extent to which the
estimated values strictly reflect preferences for visibility, and do not include a component of
preferences for reducing health risk from air pollution. The ability to isolate the value of
visibility from within the collection of intertwined benefits from visual air quality, which is
inherently multi-attributed, is a challenge for all visibility valuation studies. Each study
attempts to isolate visibility from other effect categories, but different studies take different
approaches.
Beron et al. (2001, 156270) include two measures of air pollution directly related to
health effects in their housing market hedonic study, ozone and particulates (using TSP as
the metric for particulates), as well as visibility. They argue that the presence of the two
health-related pollution conditions results in an estimated hedonic demand function for
visibility that successfully separates the health component of demand for overall air quality
from the visibility component. An alternative interpretation is that the estimated visibility
function still includes a component of health risk because the housing market data does not
support completely isolating the demand for visibility (due to correlated variables, omitted
variables, measurement error, model specification error, etc.) from demand for health risk
reductions measured by the two health related air quality metrics.
July 2009
9-100
DRAFT-DO NOT CITE OR QUOTE

-------
A key issue in interpreting the Beron et al. (2001, 156270) results is whether the
objective measures of air quality characteristics (e.g., visibility, PM concentrations, etc.)
capture people's perceptions of the different aspects of air quality in a given location. To the
extent the people simultaneously use what they see regarding VAQ as an indicator of the
overall air quality including potential health risks, then including all the measures in the
equation is not necessarily sufficient to isolate one effect from the other.
9.2.5. Summary of Effects on Visibility
Visibility impairment is caused by light scattering and absorption by suspended
particles and gases. NO2 is the only commonly occurring atmospheric pollutant gas that
absorbs visible spectrum radiation, though in most situations the amount of light
absorption by NO2 is overwhelmed by the higher amounts of particulate light extinction
(i.e., the combination of scattering and absorption) usually accompanying high NO2
concentrations. Light scattering by gases in a pollutant-free atmosphere provides a limit to
visibility in pristine conditions and is the largest contributor to the total light extinction
during the least visibility-impaired periods in remote regions of the western U.S. There is
strong and consistent evidence that PM is the overwhelming source of visibility impairment
in both urban and remote areas. Elemental carbon (EC) and some crustal minerals are the
only commonly occurring airborne particle components that absorb light. All particles
scatter light, and generally light scattering by particles is the largest of the four light
extinction components. Although a larger particle scatters more light than similarly shaped
smaller particle of the same composition, the light scattered per unit of mass is greatest for
particles with diameters from approximately 0.3-1.0 pm.
For studies where detailed data on particle composition by size data are available,
accurate calculations of light extinction can be made. However, routinely available PM
speciation data can be used to make reasonable estimates of light extinction using
relatively simple algorithms that multiply the concentrations of each of the major PM
species by its dry extinction efficiency and by a water growth term that accounts for particle
size change as a function of relative humidity for hygroscopic species (e.g., SO42-, nitrate,
and sea salt). This permits the visibility impairment associated with each of the major PM
components to be separately approximated from PM speciation monitoring data. There are
six major PM components^ PM2.5 SC>42~ usually assumed to be ammonium sulfate, PM2.5
July 2009
9-101
DRAFT-DO NOT CITE OR QUOTE

-------
nitrate usually assumed to be ammonium nitrate, PM2.5 OC compound, PM2.5 EC, PM2.5
crustal material (referred to as fine soil), and PM10-2.5 or coarse mass.
Direct optical measurement of light extinction measured by transmissometer, or by
combining the PM light scattering measured by integrating nephelometers with the PM
light absorption measured by an aethalometer offer a number of advantages compared to
algorithm estimates of light extinction based on PM composition and relative humidity
data. The direct measurements are not subject to the uncertainties associated with
assumed scattering and absorption efficiencies used in the PM algorithm approach. The
direct measurements have higher time resolution (i.e., minutes to hours), which is more
commensurate with the visibility effects compared with calculated light extinction using
routinely available PM speciation data (i.e., 24-h duration).
Particulate SO r and nitrate are produced in the atmosphere from gaseous
precursors, making them secondary PM species. They both have comparable light
extinction efficiencies (haze impacts per unit mass concentration) at any relative humidity
value, their light scattering per unit mass concentration increases with increasing relative
humidity, and at sufficiently high humidity values (RH>85%) they are the most efficient
particulate species contributing to haze. Particulate S( )r is the dominant source of
regional haze in the eastern U.S. (>50% of the particulate light extinction) and an
important contributor to haze elsewhere in the country (>20% of particulate light
extinction).
Particulate nitrate is a minor component of remote-area regional haze in the
non-California western and eastern U.S., but an important contributor in much of
California and in the upper Midwestern U.S. especially during winter when it is the
dominant contributor to particulate light extinction. While both nitric acid (a reaction
product of NOx emissions) and ammonia are needed to form ammonium nitrate, the
apparent reason for the Midwest nitrate bulge (i.e., region of high winter PM nitrate) is an
abundance of atmospheric ammonia in this region principally from agricultural emissions.
There is evidence that transport from the Midwest nitrate bulge region is responsible for
some of the ammonium nitrate episodes experienced in downwind regions far to the east.
Urban particulate nitrate concentrations are significantly elevated above surrounding
remote-area background concentrations with the largest urban contributions in the western
U.S. Particulate ammonium nitrate concentrations in California and the Midwestern
nitrate bulge region are an order of magnitude greater than estimated natural ammonium
July 2009
9-102
DRAFT-DO NOT CITE OR QUOTE

-------
nitrate concentrations. Thermodynamic and air quality simulation modeling show that
particulate nitrate concentrations are sensitive to changes in either NOx emissions (from a
combination of mobile and point sources) or ammonia emissions (principally from
agricultural sources), with the responsiveness of particulate nitrate to emissions changes
depending on the relative abundance of ammonia and nitric acid in the atmosphere.
EC and OC have the highest dry extinction efficiencies of the major PM species and
are responsible for a large fraction of the haze especially in the Northwestern U.S., though
absolute concentrations are as high in the eastern U.S. Both are a product of incomplete
combustion of fuels, including those used in internal combustion processes (gasoline and
diesel emissions) and open biomass burning (smoke from wild and prescribed fire). Organic
compound PM species are also produced by atmospheric transformation of precursor
gaseous emissions. Smoke plume impacts from large wildfires dominate many of the worst
haze periods in the western U.S. Carbonaceous PM is generally the largest component of
urban excess PM2.5 (i.e., the difference between urban and regional background
concentration). Western urban areas have more than twice the average concentrations of
carbonaceous PM than remote areas sites in the same region. In eastern urban areas PM2.5
is dominated by about equal concentrations of carbonaceous and SOr components, though
the usually high relative humidity in the East causes the hydrated SO r particles to be
responsible for about twice as much of the urban haze as that caused by the carbonaceous
PM.
Radiocarbon dating of carbonaceous PM from twelve sites (eight in the West, two of
which are urban) showed that about half of the urban area carbonaceous PM is from
contemporary as opposed to fossil sources, while in remote areas the fraction that is
contemporary ranges from 82%-100%. Summer urban excess carbonaceous PM is
dominated by fossil carbon for the two western urban areas (Phoenix, AZ and Seattle, WA),
but nearly half of the winter urban excess for these two urban areas are from contemporary
carbon sources (e.g., residential wood combustion). An empirical relationship between the
radiocarbon analysis results and the more widely measured elemental and OC data set was
used to estimate the fraction of contemporary carbon at about 150 monitoring locations
nationwide. The highest fraction of contemporary carbon is for the western remote areas
sites during the summer (>90% contemporary) and the least was for eastern urban areas
during the summer (<45% contemporary). Winter tended to have less extreme fractions of
contemporary carbon for both remote and urban areas. A lower bound estimate of 40% of
July 2009
9-103
DRAFT-DO NOT CITE OR QUOTE

-------
the contemporary and 35% of the fossil carbon is from secondary conversion of gaseous
precursor during the summer at the twelve radiocarbon monitoring sites, suggesting that
primary carbonaceous PM whether from fossil or contemporary sources represent less than
two thirds of the total carbonaceous PM.
PM2.5 crustal material (referred to as fine soil) and coarse mass (i.e., PM10 minus
PM2.5) are significant contributors to haze for remote areas sites in the arid Southwestern
U.S. where they contribute a quarter to a third of the haze, with coarse mass usually
contributing twice that of fine soil. Coarse mass concentrations are as high in the Central
Great Plains as in the Southwestern deserts though there are no corresponding high
concentrations of fine soil as in the Southwest. Also, the relative contribution to haze by the
high coarse mass in the Great Plains is much smaller because of the generally higher haze
values caused by the high concentrations of S( )r and nitrate PM in that region.
A comprehensive assessment of the 610 worst haze sample periods over a 3-yr period
in the western U.S. where dust is the major contributor categorized each site/sampler
period into four causal groups: Asian dust, local windblown dust, transported regional
windblown dust, and undetermined dust (i.e., not in one of the three other groups). Most
dust days occurred at sites in Arizona, New Mexico, Colorado, western Texas, and southern
California, and these were dominated by local and regionally transported wind-blown dust.
Asian dust caused only a few of the worst dust days during the 3-yr assessment period,
though it is an important source of dust for the more northerly regions of the West
(responsible for 10%-40% of their worst dust periods) were there is rarely any windblown
dust probably due to the greater ground cover. The frequency of worst dust events classified
as undetermined was greatest for sites in the vicinity of large urban and agricultural areas
such as those in California and Southern Arizona.
Visibility has direct significance to people's enjoyment of daily activities and their
overall sense of wellbeing. A number of social science disciplines have undertaken to link
perceived urban visibility to an array of effects reflecting the overall desire for good VAQ,
and the benefits of improving currently degraded VAQ. For example, psychological research
has demonstrated that people are emotionally affected by poor VAQ such that their overall
sense of wellbeing is diminished. Urban visibility has been examined in two types of studies
directly relevant to the NAAQS review process: urban visibility preference studies and
urban visibility valuation studies. Both types of studies are designed to evaluate
individuals' desire for good VAQ where they live, using different metrics. Urban visibility
July 2009
9-104
DRAFT-DO NOT CITE OR QUOTE

-------
preference studies examine individuals' preferences by investigating the amount of
visibility degradation considered unacceptable, while economic studies examine the value
an individual places on improving VAQ by eliciting how much the individual would be
willing to pay for different amounts of VAQ improvement.
There are three urban visibility preference studies and one additional pilot study
(designed as a survey instrument development project) that have been conducted to date
that provide useful information on individuals' preferences for good VAQ in the urban
setting. The completed studies were conducted in Denver, Colorado (Ely et al., 1991,
156417), two cities in British Columbia, Canada (Pryor, 1996, 056598) and Phoenix, Arizona
(BBC Research & Consulting, 2002, 156258). The pilot study was conducted in Washington,
DC (ABT, 2002, 156186). One notable finding of the three visibility preference studies and
the one pilot study is the general degree of consistency in the median preferences for an
acceptable amount of visibility degradation. The range of median acceptable visibility
preference values from the four studies is 19-25 dv. Measured in terms of visual range (VR),
these median acceptable values are between 59 km and 32 km.
The economic importance of urban visibility has been examined by a number of
studies designed to quantify the benefits (or willingness to pay) associated with potential
improvements in urban visibility. Urban visibility valuation research prior to 1997 was
summarized in Chestnut and Dennis (1997, 014525), and was also described in the 2004
PM AQCD (EPA, 2004, 056905) and the 2005 OAQPS PM NAAQS Staff Paper (U.S. EPA,
2005, 090209). Since the mid-1990s, little new information has become available regarding
urban visibility valuation.
Collectively, the evidence is sufficient to conclude that a causal relationship exists
between PM and visibility impairment.
9.3. Effects on Climate
While most of this ISA is restricted to consideration of the emissions, transport and
transformation, resulting concentrations, and effects from PM in the U.S., because the
effects endpoint here is climate, a larger spatial domain is needed. However, this
assessment is not intended to be comprehensive even as a survey of the enormous range
and volume of science related to climate effects from PM; rather, particular attention has
been paid to data relevant to the U.S.
July 2009
9-105
DRAFT-DO NOT CITE OR QUOTE

-------
The two principal sources for material in this section are Chapter 2, "Changes in
Atmospheric Constituents and in Radiative Forcing," (Forster et al., 2007, 092936) in the
comprehensive Working Group I report in the Fourth Assessment Report (AR4) from the
Intergovernmental Panel on Climate Change (IPCC), Climate Change 2007: The Physical
Science Basis (Intergovernmental Panel on Climate Change, 2007, 092765), hereafter IPCC
AR4; and the U.S. Climate Change Science Program Synthesis and Assessment Product 2.3,
"Atmospheric Aerosol Properties and Climate Impacts," by Chin et al. (2009, 192130).
hereafter CCSP SAP2.3. The EPA is a constituent agency member of the U.S. federated
CCSP along with NOAA and NASA, which led production of CCSP SAP2.3 incorporating
significant sections from EPA data and reports related particularly to U.S. emissions and
measurements. Sections from each of these recent comprehensive reports are included here
in their entirety or as emended as noted where they represent the most thorough summary
of the climate effects of aerosols. (In the sections included from IPCC AR4 and CCSP
SAP2.3, 'aerosols' is more frequently used than "PM" and that word is retained.)
9.3.1. The Climate Effects of Aerosols
Section 9.3.1 comes directly from CCSP SAP2.3 Chapter 1 Section 1.2, with section, table,
and figure numbers changed to be internally consistent with this ISA.
Aerosols exert a variety of impacts on the environment. Aerosols (sometimes
referred to particulate matter or "PM," especially in air quality applications), when
concentrated near the surface, have long been recognized as affecting pulmonary
function and other aspects of human health. Sulfate and nitrate aerosols play a role in
acidifying the surface downwind of gaseous sulfur and odd nitrogen sources. Particles
deposited far downwind might fertilize iron-poor waters in remote oceans, and
Saharan dust reaching the Amazon Basin is thought to contribute nutrients to the
rainforest soil.
Aerosols also interact strongly with solar and terrestrial radiation in several
ways. Figure 9"52 offers a schematic overview. First, they scatter and absorb sunlight
(McCormick and Ludwig, 1967, 190528; Mitchell, 1971, 190546! Charlson and Pilat,
1969, 190025); these are described as "direct effects" on shortwave (solar) radiation.
Second, aerosols act as sites at which water vapor can accumulate during cloud
droplet formation, serving as cloud condensation nuclei or CCN. Any change in
number concentration or hygroscopic properties of such particles has the potential to
modify the physical and radiative properties of clouds, altering cloud brightness
(Twomey, 1977, 190 5 33) and the likelihood and intensity with which a cloud will
precipitate (Liou and Ou, 1989, 190407; Albrecht, 1989, 045783; e.g., Gunn and
Phillips, 1957, 190595). Collectively changes in cloud processes due to anthropogenic
aerosols are referred to as aerosol indirect effects. Finally, absorption of solar
radiation by particles is thought to contribute to a reduction in cloudiness, a
phenomenon referred to as the semi-direct effect. This occurs because absorbing
aerosol warms the atmosphere, which changes the atmospheric stability, and reduces
surface flux.
July 2009
9-106
DRAFT-DO NOT CITE OR QUOTE

-------
The primary direct effect of aerosols is a brightening of the planet when viewed
from space, as much of Earth's surface is dark ocean, and most aerosols scatter more
than 90% of the visible light reaching them. The primary indirect effects of aerosols on
clouds include an increase in cloud brightness, change in precipitation and possibly an
increase in lifetime; thus the overall net impact of aerosols is an enhancement of
Earth's reflectance (shortwave albedo). This reduces the sunlight reaching Earth's
surface, producing a net climatic cooling, as well as a redistribution of the radiant and
latent heat energy deposited in the atmosphere. These effects can alter atmospheric
circulation and the water cycle, including precipitation patterns, on a variety of length
and time scales (e.g., Ramanathan et al., 2001, 042681; Zhang et al., 2006, 190933).
Scattering & Unperturbed Increased CDNC
Drizzle
Increased cloud height Increased cloud
absorption of
radiation
^ Direct effects J
cloud
(constant LWC)
(Twomey, 1974)
Cloud albedo effect/
1st indirect effect/ J
I Twomey effect I
suppression. (Pincus & Baker, 1994) lifetime
Increased LWC	(Albrecht, 1989)
Vca
Cloud lifetime effect/ 2nd indirect effect/ Albrecht effect
Heating causes
cloud burn-off
(Ackerman et al., 2000)

Semi-direct effect
Surface
Source: IPCC (2007,190988) modified from Haywood and Boucher (2000,156531).
Figure 9-52. Aerosol radiative forcing. Airborne particles can affect the heat balance of the
atmosphere, directly, by scattering and absorbing sunlight, and indirectly, by altering
cloud brightness and possibly lifetime. Here small black dots represent aerosols, circles
represent cloud droplets, and straight lines represent short-wave radiation, and wavy
lines, long-wave radiation. LWC is liquid water content, and CDNC is cloud droplet
number concentration. Confidence in the magnitudes of these effects varies
considerably (see Chapter 3). Although the overall effect of aerosols is a net cooling at
the surface, the heterogeneity of particle spatial distribution, emission history, and
properties, as well as differences in surface reflectance, mean that the magnitude and
even the sign of aerosol effects vary immensely with location, season and sometimes
inter-annually. The human-induced component of these effects is sometimes called
"climate forcing."
Several variables are used to quantify the impact aerosols have on Earth's energy
balance; these are helpful in describing current understanding, and in assessing
possible future steps.
For the purposes of this report, aerosol radiative forcing (RF) is defined as the net
energy flux (downwelling minus upwelling) difference between an initial and a
perturbed aerosol loading state, at a specified level in the atmosphere. (Other
quantities, such as solar radiation, are assumed to be the same for both states.) This
difference is defined such that a negative aerosol forcing implies that the change in
Top of the
atmosphere
Indirect effect
on ice clouds
and contrails
July 2009
9-107
DRAFT-DO NOT CITE OR QUOTE

-------
aerosols relative to the initial state exerts a cooling influence, whereas a positive
forcing would mean the change in aerosols exerts a warming influence.
There are a number of subtleties associated with this definition-
(1)	The initial state against which aerosol forcing is assessed must be specified.
For direct aerosol radiative forcing, it is sometimes taken as the complete absence of
aerosols. IPCC AR4 (2001, 156587) uses as the initial state their estimate of aerosol
loading in 1750. That year is taken as the approximate beginning of the era when
humans exerted accelerated influence on the environment.
(2)	A distinction must be made between aerosol RF and the anthropogenic
contribution to aerosol RF. Much effort has been made to distinguishing these
contributions by modeling and with the help of space-based, airborne, and surface-
based remote sensing, as well as in situ measurements. These efforts are described in
subsequent chapters (of the CCSP SAP2.3).
(3)	In general, aerosol RF and anthropogenic aerosol RF include energy associated
with both the shortwave (solar) and the long-wave (primarily planetary thermal
infrared) components of Earth's radiation budget. However, the solar component
typically dominates, so in this document, these terms are used to refer to the solar
component only, unless specified otherwise. The wavelength separation between the
short" and long-wave components is usually set at around three or four micrometers.
(4)	The IPCC AR4 (2007, 190988) defines radiative forcing as the net downward
minus upward irradiance at the tropopause due to an external driver of climate
change. This definition excludes stratospheric contributions to the overall forcing.
Under typical conditions, most aerosols are located within the troposphere, so aerosol
forcing at TOA and at the tropopause are expected to be very similar. Major volcanic
eruptions or conflagrations can alter this picture regionally, and even globally.
(5)	Aerosol radiative forcing can be evaluated at the surface, within the
atmosphere, or at top-of-atmosphere (TOA). In this document, unless specified
otherwise, aerosol radiative forcing is assessed at TOA.
(6)	As discussed subsequently, aerosol radiative forcing can be greater at the
surface than at TOA if the aerosols absorb solar radiation. TOA forcing affects the
radiation budget of the planet. Differences between TOA forcing and surface forcing
represent heating within the atmosphere that can affect vertical stability, circulation
on many scales, cloud formation, and precipitation, all of which are climate effects of
aerosols. In this document, unless specified otherwise, these additional climate effects
are not included in aerosol radiative forcing.)
(7)	Aerosol direct radiative forcing can be evaluated under cloud-free conditions or
under natural conditions, sometimes termed "all-sky" conditions, which include
clouds. Cloud-free direct aerosol forcing is more easily and more accurately calculated;
it is generally greater than all-sky forcing because clouds can mask the aerosol
contribution to the scattered light. Indirect forcing, of course, must be evaluated for
cloudy or all-sky conditions. In this document, unless specified otherwise, aerosol
radiative forcing is assessed for all-sky conditions.
(8)	Aerosol radiative forcing can be evaluated instantaneously, daily (24 h)
averaged, or assessed over some other time period. Many measurements, such as
those from polar-orbiting satellites, provide instantaneous values, whereas models
usually consider aerosol RF as a daily average quantity. In this document, unless
specified otherwise, daily averaged aerosol radiative forcing is reported.
(9)	Another subtlety is the distinction between a "forcing" and a "feedback." As
different parts of the climate system interact, it is often unclear which elements are
"causes" of climate change (forcings among them), which are responses to these
causes, and which might be some of each. So, for example, the concept of aerosol
effects on clouds is complicated by the impact clouds have on aerosols; the aggregate is
often called aerosol-cloud interactions. This distinction sometimes matters, as it is
July 2009
9-108
DRAFT-DO NOT CITE OR QUOTE

-------
more natural to attribute responsibility for causes than for responses. However,
practical environmental considerations usually depend on the net result of all
influences. In this report, "feedbacks" are taken as the consequences of changes in
surface or atmospheric temperature, with the understanding that for some
applications, the accounting may be done differently
RFTerrm
	1	
l'; -	• ¦- ¦ ¦ j
Lemuel vied
p eisihoiHe 1
fijO "
i luic^i Lei i>
s
DO
~
CL
£
-T-
Diane	StTiUMplwK
I nepOipTW riC
watw
vapour f-uiri CH,
S:..-fari- : ;h-r>do
R-nc L r.:r |if|ii
C3« SlKf#
IDirctii cifn.
A0-j|
Glctml
(VLa-fc-rtl
CfflWronjJ

-------
cd
-O
o
>
cu
&
warning
Irrrg Wrr. jTrsrrnnqitf
.^aes and azanft
lliLk.TPC Ul|B^
aeiltoiiiw.i
kfi»J nirrcsoi
~ - r.^-w H3*T«riL
I i Jr. .il Tj/G
|otA antfrcMMfpjfc rgefiHsrvB farong 1
t, g A u 1 * W fN
A	0
Radiative Forcing {W nr3)
Source: Adapted from IPCC (2007,190988).
Figure 9-54. Probability distribution functions (PDFs) for anthropogenic aerosol and GHG RFs.
Dashed red curve: RF of long-lived greenhouse gases plus ozone; dashed blue curve: RF
of aerosols (direct and cloud albedo RF); red filled curve: combined anthropogenic RF.
The RF range is at the 90% confidence interval.
In summary, aerosol radiative forcing, the fundamental quantity about which this
report is written, must be qualified by specifying the initial and perturbed, aerosol
states: for which the radiative flux difference is calculated, the altitude at which the
quantity is assessed, the wavelength regime considered, the temporal averaging, the
cloud conditions, and whether total or only human-induced contributions are-
considered- The definition given here, qualified as needed, is used throughout the
report.
Although the possibility that aerosols affect climate was recognized more than 40
yr ago, the measurements needed to establish the magnitude of such effects, or even
whether specific aerosol types warm or cool the surface , were lacking. Satellite
instruments capable of at least crudely monitoring aerosol amount globally were first
deployed in the late 1970s. But Scientific focus on this Subject grew substantially in
the 1990s (e.g.j.Charlson and Wigley, 1994, 192011. 1991, 045793. 1992, 045794:
Penner et al., 1992, 045825). in part because it was recognized that reproducing the
observed temperature trends over the industrial period with climate models requires
including net global cooling by aerosols in the calculation (IP<3C, 1995, 190991: 1996,
190990). along with the warming influence of enhanced atmospheric greenhouse gas
(i ill1 i) concentrations - mainly carbon dioxide, methane, nitrous oxide,
chlorofluoroearbons, and ozone.
Improved satellite instruments, ground- and ship-based surface monitoring, more
sophisticated chemical transport and climate models, and field campaigns that
brought all these elements together with aircraft remote sensing and in situ sampling
for focused, coordinated study, began to fill in some of the knowledge gaps. By the
Fourth IP<3C Assessment Report, the scientific community consensus held that in
global average, the sum of direct and indirect top-of-atmosphere (T0A) forcing by
anthropogenic aerosols is negative (cooling) of about "1.3 W/m2 ("2.2 to -0.5 W/m2).
This is significant compared to the positive forcing by anthropogenic GHSs (including
ozone), about 2.9 ± 0.3 W/m2 (IPCC (Intergovernmental Panel on Climate, 2007,
190988). However, the spatial distribution of the gases and aerosols are very different,
and they do not simply exert compensating influences on climate.
July 2009
9-110
DRAFT-DO NOT CITE OR QUOTE

-------
r
E
I
LLi
The ,IPCC aerosol forcing assessments are based largely on model calculations,
constrained as much as possible by observations. At present., aerosol influences are
not yet quantified adequately, according to Figure SHU, as scientific understanding is
designated as "Medium-Low" and "Low" for the direct and indirect climate forcing,
respectively. The 11 If V AIM (2007, 190988) concluded that uncertainties associated
with changes in Earth's radiation budget .due to anthropogenic aerosols make the
largest contribution to the overall uncertainty in radiative forcing of climate change
among the factors assessed over the industrial period (Figure U-~< I).
120
• 1ftQ
40
-60
-AO
*0
0
t
1
TOA
Surfac*
1
I
i
I
South
Antvri&n Africa
tftomass burning
&OUIH A.< 0,1 A ¦*- 0 .2 A -0,1. Nurlli Europe
0.35 America
M
x

Asi.i
Mineral a li 5t
PalLunon
Source: Adapted from Zhou et al. (2005,156183).
Figure 9-55. The clear-sky forcing efficiency En, defined as the diurnally averaged aerosol direct
radiative effect (W/m2) per unit A0D at 550 nm, calculated at both TOA and the surface,
for typical aerosol types over different geographical regions. The vertical black lines
represent ± one standard deviation of Ed for individual aerosol regimes and A is surface
broadband albedo.
Although AOD, aerosol properties, aerosol vertical distribution, and surface,
reflectivity all contribute to aerosol radiative forcing, AOD usually varies on regional
scales more than the other aerosol quantities involved. Forcing efficiency (Et) defined
as a ratio of direct aerosol radiative forcing to ADD at 550 nm, reports the sensitivity
of aerosol radiative forcing to AOD, and is useful for isolating the influences of particle
properties and other factors from that of AOD. /•-' is expected to exhibit a range of
values globally, because it is governed mainly by aerosol size distribution and
chemical composition (which determine aerosol single-scattering albedo and phase
function), surface reflectivity, and solar irradiance, each of which exhibits pronounced
spatial and temporal variations. To assess aerosol RF JKJgs multiplied by the ambient
AOD,
Figure 9-55 shows a range of A;. derived from AERONET surface sun photometer
network measurements of aerosol loading and particle properties;, representing
July 2009
9-111
DRAFT-DO NOT CITE OR QUOTE

-------
different aerosol and surface types, and geographic locations. It demonstrates how
aerosol direct solar radiative forcing (with initial state taken as the absence of aerosol)
is determined by a combination of aerosol and surface properties. For example, _£rdue
to southern African biomass burning smoke is greater at the surface and smaller at
TOA than South American smoke because the southern African smoke absorbs
sunlight more strongly, and the magnitude of Eiior mineral dust for several locations
varies depending on the underlying surface reflectance. Figure 9"55 illustrates one
further point, that the radiative forcing by aerosols on surface energy balance can be
much greater than that at TOA. This is especially true when the particles have SSA
substantially less than 1, which can create differences between surface and TOA
forcing as large as a factor of five (e.g., Zhou, 2005, 156183).
Table 9-3. Top-of-atmosphere, cloud-free, instantaneous direct aerosol radiative forcing dependence
on aerosol and surface properties. Here TWP, SGP, and NSA are the Tropical West Pacific
island. Southern Great Plains, and North Slope Alaska observation stations maintained by
the DOE ARM program, respectively. Instantaneous values are given at specific solar
zenith angle. Upper and middle parts are from McComiskey et al. (2008,190523).
Representative, parameter-specific measurement uncertainty upper bounds for producing
1 W/m2 instantaneous TOA forcing accuracy are given in the lower part, based on
sensitivities at three sites from the middle part of the table.
Parameters
TWP
SGP
NSA
AEROSOL PROPERTIES (AOD, SSA, GJ, SOLAR ZENITH ANGLE (SZA), SURFACE ALBEDO (A), AND AEROSOL DIRECT RF A T TOA (F)
AOD
0.05
0.1
0.05
SSA
0.97
0.95
0.95
g
0.8
0.6
0.7
A
0.05
0.1
0.9
SZA
30
45
70
F (W/m2)
¦2.2
¦6.3
2.6
SENSITIVITY OF CLOUD FREE, INSTANTANEOUS, TOA DIRECT AEROSOL RADIATIVE FORCING TO AEROSOL AND SURFACE PROPERTIES, W/M2 PER UNIT
CHANGE IN PROPERTY
rfFMAOD)
¦45
¦64
51
rfFMSSA)
¦11
¦50
¦60
rfF/rfg
13
23
2
rfFM
8
24
6
REPRESENTATIVE MEASUREMENT UNCERTAINTY UPPER BOUNDS FOR PRODUCING 1 W/M2 ACCURACY OF AEROSOL RF
AOD
0.022
0.016
0.020
SSA
0.091
0.020
0.017
g
0.077
0.043

A
0.125
0.042
0.167
Table 9"3 presents estimates of cloud-free, instantaneous, aerosol direct RF
dependence on AOD, and on aerosol and surface properties, calculated for three sites
maintained by the U.S. Department of Energy's Atmospheric Radiation Measurement
(ARM) program, where surface and atmospheric conditions span a significant range of
July 2009
9-112
DRAFT-DO NOT CITE OR QUOTE

-------
natural environments (McComiskey et al., 2008, 190523). Here aerosol RF is
evaluated relative to an initial state that is the complete absence of aerosols. Note
that aerosol direct RF dependence on individual parameters varies considerably,
depending on the values of the other parameters, and in particular, that aerosol RF
dependence on AOD actually changes sign, from net cooling to net warming, when
aerosols reside over an exceedingly bright surface. Sensitivity values are given for
snapshots at fixed solar zenith angles, relevant to measurements made, for example,
by polar-orbiting satellites.
The lower portion of Table 9"3 presents upper bounds on instantaneous
measurement uncertainty, assessed individually for each of AOD, SSA, g, and A, to
produce a 1 W/m2 top-of atmosphere, cloud-free aerosol RF accuracy. The values are
derived from the upper portion of the table, and reflect the diversity of conditions
captured by the three ARM sties. Aerosol RF sensitivity of 1 W/m2 is used as an
example! uncertainty upper bounds are obtained from the partial derivative for each
parameter by neglecting the uncertainties for all other parameters. These estimates
produce an instantaneous AOD measurement uncertainty upper bound between about
0.01 and 0.02, and SSA constrained to about 0.02 over surfaces as bright as or
brighter than the ARM Southern Great Plains site, typical of mid-latitude, vegetated
land. Other researchers, using independent data sets, have derived ranges of _£rand
aerosol RF sensitivity similar to those presented here, for a variety of conditions
(Zhou, 2005, 156183! Yu, 2006, 156173; Christopher and Jones, 2008, 189985).
These uncertainty bounds provide a baseline against which current and expected
near-future instantaneous measurement capabilities are assessed in Chapter 2 (of the
CCSP SAP2.3). Model sensitivity is usually evaluated for larger-scale (even global)
and longer-term averages. When instantaneous measured values from a randomly
sampled population are averaged, the uncertainty component associated with random
error diminishes as something like the inverse square root of the number of samples.
As a result, the accuracy limits used for assessing more broadly averaged model
results corresponding to those used for assessing instantaneous measurements would
have to be tighter, as discussed in Chapter 4 (of the CCSP SAP2.3).
In summary, much of the challenge in quantifying aerosol influences arises from
large spatial and temporal heterogeneity, caused by the wide variety of aerosol
sources, sizes and compositions, the spatial non-uniformity and intermittency of these
sources, the short atmospheric lifetime of most aerosols, and the spatially and
temporally non-uniform chemical and microphysical processing that occurs in the
atmosphere. In regions having high concentrations of anthropogenic aerosol, for
example, aerosol forcing is much stronger than the global average, and can exceed the
magnitude of GHG warming, locally reversing the sign of the net forcing. It is also
important to recognize that the global-scale aerosol TOA forcing alone is not an
adequate metric for climate change (National Research, 2005, 057409). Due to aerosol
absorption, mainly by soot, smoke, and some desert dust particles, the aerosol direct
radiative forcing at the surface can be much greater than the TOA forcing, and in
addition, the radiative heating of the atmosphere by absorbing particles can change
the atmospheric temperature structure, evolution, and possibly large-scale dynamical
systems such as the monsoons (Kim et al., 2006, 190917; Lau et al., 2008, 190229). By
realizing aerosol's climate significance and the challenge of charactering highly
variable aerosol amount and properties, the U.S. Climate Change Research Initiative
(CCRI) identified research on atmospheric concentrations and effects of aerosols
specifically as a top priority (National Research, 2001, 053303).
July 2009
9-113
DRAFT-DO NOT CITE OR QUOTE

-------
9.3.2. Overview of Aerosol Measurement Capabilities
9.3.2.1. Satellite Remote Sensing
Section 9.3.2 with the exception of the final paragraph, comes directly from CCSP
SAP2.3 Chapter 2, Section 2.2 with section, table, and figure numbers changed to be
internally consistent with this ISA.
A measurement-based characterization of aerosols on a global scale can be
realized only through satellite remote sensing, which is the only means of
characterizing the large spatial and temporal heterogeneities of aerosol distributions.
Monitoring aerosols from space has been performed for over two decades and is
planned for the coming decade with enhanced capabilities (King et al., 1999, 190635;
Forster et al., 2007, 092936; Lee et al., 2006, 190358; Mischenko et al., 2007, 190543).
Table 9"4 summarizes major satellite measurements currently available for the
tropospheric aerosol characterization and radiative forcing research.
Early aerosol monitoring from space relied on sensors that were designed for
other purposes. The Advanced Very High Resolution Radiometer (AVHRR), intended
as a cloud and surface monitoring instrument, provides radiance observations in the
visible and near infrared wavelengths that are sensitive to aerosol properties over the
ocean (Husar et al., 1997, 045900; Mishchenko et al., 1999, 190541). Originally
intended for ozone monitoring, the ultraviolet (UV) channels used for the Total Ozone
Mapping Spectrometer (TOMS) are sensitive to aerosol UV absorption with little
surface interferences, even over land (Torres et al., 1998, 190503). This UV-technique
makes TOMS suitable for monitoring biomass burning smoke and dust, though with
limited sensitivity near the surface (Herman et al., 1997, 048393) and for retrieving
aerosol single-scattering albedo from space (Torres et al., 2005, 190507). (Anew
sensor, the Ozone Monitoring Instrument (OMI) aboard Aura, has improved on such
UVtechnique advantages, providing higher spatial resolution and more spectral
channels; see (Veihelmann et al., 2007, 190627). Such historical sensors have provided
multi-decadal climatology of aerosol optical depth that has significantly advanced the
understanding of aerosol distributions and long-term variability (e.g.,Geogdzhayev et
al., 2002, 190574; Massie et al., 2004, 190492; Mishchenko and Geogdzhayev, 2007,
190545; Mishchenko et al., 2007, 190542; Torres et al., 2002, 190505; Zhao et al., 2008,
190935).
Over the past decade, satellite aerosol retrievals have become increasingly
sophisticated. Now, satellites measure the angular dependence of radiance and
polarization at multiple wavelengths from UV through the infrared (IR) at fine spatial
resolution. From these observations, retrieved aerosol products include not only
optical depth at one wavelength, but also spectral optical depth and some information
about particle size over both ocean and land, as well as more direct measurements of
polarization and phase function. In addition, cloud screening is much more robust
than before and onboard calibration is now widely available. Examples of such new
and enhanced sensors include the MODerate resolution Imaging Spectroradiometer
(MODIS, see Box 2.1 of CCSP SAP2.3), the Multi-angle Imaging SpectroRadiometer
(MISR, see Box 2.2 of CCSP SAP2.3), Polarization and Directionality of the Earth's
Reflectance (POLDER, see Box 2.3 of CCSP SAP2.3), and OMI, among others. The
accuracy for AOD measurement from these sensors is about 0.05 or 20% of AOD
(Kahn RA Gaitley et al., 2005, 190966; Remer et al., 2005, 190221) and somewhat
better over dark water, but that for aerosol microphysical properties, which is useful
for distinguishing aerosol air mass types, is generally low. The Clouds and the Earth's
Radiant Energy System (CERES, see Box 2.4 of CCSP SAP2.3) measures broadband
solar and terrestrial radiances. The CERES radiation measurements in combination
July 2009
9-114
DRAFT-DO NOT CITE OR QUOTE

-------
with satellite retrievals of aerosol optical depth can be used to determine aerosol
direct radiative forcing.
Table 9-4. Summary of major satellite measurements currently available for the tropospheric aerosol
characterization and radiative forcing research.
Category
Properties
Sensor/platform
Parameters
Spatial coverage Temporal coverage
Column-integrated
Loading
AVHRR/NOAA-series
MISR/Terra
OMI/Aura
Optical depth
TOMS/Nimbus, ADEOSI, EP
POLDER-1,2, PARASOL
MODIS/Terra. Aqua
-daily coverage of global
ocean
1981 present
-daily coverage of global
land and ocean
1979-2001
1997-present
2000-present (Terra)
2002-present (Aqua)
— weekly coverage of
global land and ocean,
including bright desert and
nadir sun-glint
-daily coverage of global
land and ocean
2000-present
2005-present
Size,
AVHRR/NOAA-series
POLDER-1,2, PARASOL
MODIS/T erra. Aqua
Angstrom exponent
Global ocean
1981 -present
Fine-mode fraction,
Angstrom exponent, non-
spherical fraction
Global land and ocean
1997-present
Fine-mode fraction
Angstrom exponent
Global land and ocean
(better quality over ocean)
Effective radium
Global ocean
Asymmetry factor
MISR/Terra	Angstrom exponent, small,
medium large fractions,
non-spherical fractions
Global land and ocean
2000-present (Terra)
2002-present (Aqua)
2000-present
Absorption
TOMS/Nimbus, ADEOSI, EP
OMI/Aura
Absorbing aerosol index,
single-scattering albedo,
absorbing optical depth
Global land and ocean
MISR/Terra	Single-scattering albedo (2-
4 bins)
1979-2001
2005-present
2000-present
Vertical-resolved
Loading, size, and
shape
GLAS/ICESat
CAUOP/CAUPSO
Extinction/backscatter
Extinction/backscatter,
color ratio, depolarization
ratio
Global land and ocean, 16-
day repeating cycle, single-
nadir measurement
2003-present
(~ 3 months/yr)
2006-present
Complementary to these passive sensors, active remote sensing from space is also
now possible and ongoing (see Box 2.5 of CCSP SAP2.3). Both the Geoscience Laser
Altimeter System (GLAS) and the Cloud and Aerosol Lidar with Orthogonal
Polarization (CALIOP) are collecting essential information about aerosol vertical
distributions. Furthermore, the constellation of six afternoon-overpass spacecrafts (as
illustrated in Figure 9-60), the so-called A-Train (Stephens and et, 2002, 190412)
makes it possible for the first time to conduct near simultaneous (within 15 minutes)
July 2009	9-115	DRAFT - DO NOT CITE OR QUOTE

-------
measurements of aerosols, clouds, and radiative fluxes in multiple dimensions with
sensors in complementary capabilities.
The improved accuracy of aerosol products (mainly AOD) from these new-
generation sensors, together with improvements in characterizing the earth's surface
and clouds, can help reduce the uncertainties associated with estimating the aerosol
direct radiative forcing (Yu H, et, 2006, 156173); and references therein). The
retrieved aerosol microphysical properties, such as size, absorption, and non-spherical
fraction can help distinguish anthropogenic aerosols from natural aerosols and hence
help assess the anthropogenic component of aerosol direct radiative forcing (Bellouin
et al., 2005, 155684; Christopher et al., 2006, 155729; Kaufman et al., 2005, 155891;
Yu H, et, 2006, 156173). However, to infer aerosol number concentrations and examine
indirect aerosol radiative effects from space, significant efforts are needed to measure
aerosol size distribution with much improved accuracy, characterize aerosol type,
account for impacts of water uptake on aerosol optical depth, and determine the
fraction of aerosols that is at the level of the clouds (Kapustin et al., 2006, 190961;
Rosenfeld, 2006, 190233). In addition, satellite remote sensing is not sensitive to
particles much smaller than 0.1 micrometer in diameter, which comprise of a
significant fraction of those that serve as cloud condensation nuclei.
Finally, algorithms are being developed to retrieve aerosol absorption or SSAfrom
satellite observations (e.g., Kaufman et al., 2002, 190955; Torres et al., 2005, 190507).
The NASA Glory mission, scheduled to launch in 2009 and to be added to the A"Train,
will deploy a multi-angle, multispectral polarimeter to determine the global
distribution of aerosol and clouds. It will also be able to infer microphysical property
information, from which aerosol type (e.g., marine, dust, pollution, etc.) can be
inferred for improving quantification of the aerosol direct and indirect forcing on
climate (Mischenko et al., 2007, 190543).
In summary, major advances have been made in both passive and active aerosol
remote sensing from space in the past decade, providing better coverage, spatial
resolution, retrieved AOD accuracy, and particle property information. However, AOD
accuracy is still much poorer than that from surface "based sun photometers
(0.01-0.02), even over vegetated land and dark water where retrievals are most
reliable. Although there is some hope of approaching this level of uncertainty with a
new generation of satellite instruments, the satellite retrievals entail additional
sensitivities to aerosol and surface scattering properties. It seems unlikely that
satellite remote sensing could exceed the sun photometer accuracy without
introducing some as-yet-unspecified new technology. Spacebased lidars are for the
first time providing global constraints on aerosol vertical distribution, and multi-angle
imaging is supplementing this with maps of plume injection height in aerosol source
regions. Major advances have also been made during the past decade in
distinguishing aerosol types from space, and the data are now useful for validating
aerosol transport model simulations of aerosol air mass type distributions and
transports, particularly over dark water. But particle size, shape, and especially SSA
information has large uncertainty; improvements will be needed to better distinguish
anthropogenic from natural aerosols using space-based retrievals. The particle
microphysical property detail required to assess aerosol radiative forcing will come
largely from targeted in situ and surface remote sensing measurements, at least for
the near-future, although estimates of measurement-based aerosol RF can be made
from judicious use of the satellite data with relaxed requirements for characterizing
aerosol microphysical properties.
MODerate resolution Imaging Spectroradiometer. MODIS performs near
global daily observations of atmospheric aerosols. Seven of 36 channels (between 0.47
and 2.13 ]im) are used to retrieve aerosol properties over cloud and surface-screened
areas (Martins et al., 2002, 190470; Li et al., 2004, 190386). Over vegetated land,
MODIS retrieves aerosol optical depth at three visible channels with high accuracy of
± 0.05 ± 0.2t (Kaufman and Fraser, 1997, 190958; Chu et al., 2002, 190001; Remer et
al., 2005, 190221; Levy et al., 2007, 190379). Most recently a deep blue algorithm (Hsu
July 2009
9-116
DRAFT-DO NOT CITE OR QUOTE

-------
et al., 2004, 190622) has been implemented to retrieve aerosols over bright deserts on
an operational basis, with an estimated accuracy of 20-30%. Because of the greater
simplicity of the ocean surface, MODIS has the unique capability of retrieving not
only aerosol optical depth with greater accuracy, i.e., ± 0.03 ± 0.05x (Tanre et al., 1997,
190452; Remer et al., 2002, 190218; 2005, 190221; Remer and et, 2008, 190224). but
also quantitative aerosol size parameters (e.g., effective radius, fine-mode fraction of
AOD) (Kaufman et al., 2002, 190956; Remer et al., 2005, 190221; Kleidman et al.,
2005, 190175). The fine-mode fraction has been used as a tool for separating
anthropogenic aerosol from natural ones and estimating the anthropogenic aerosol
direct climate forcing (Kaufman et al., 2005, 155891). Figure 9"56 shows composites of
MODIS AOD and fine-mode fraction that illustrate seasonal and geographical
variations of aerosol types. Clearly seen from the figure is heavy pollution over East
Asia in both months, biomass burning smoke over South Africa, South America, and
Southeast Asia in August, heavy dust storms over North Atlantic in both months and
over Arabian Sea in August, and a mixture of dust and pollution plume swept across
North Pacific in April.
Multi-Angle Imaging SpectroRadiometer. MISR, aboard the sun-
synchronous, polar orbiting satellite Terra, measures upwelling solar radiance in four
visible-near-IR spectral bands and at nine view angles spread out in the forward and
aft directions along the flight path (Diner et al., 2002, 189967). It acquires global
coverage about once per week. A wide range of along-track view angles makes it
feasible to more accurately evaluate the surface contribution to the TOA radiances
and hence retrieve aerosols over both ocean and land surfaces, including bright desert
and sunglint regions (Diner et al., 1998, 189962; Kahn RAGaitley et al., 2005,
190966; Martonchik et al., 1998, 190472; 2002, 190490). MISR AODs are within 20%
or ± 0.05 of coincident AERONET measurements (Abdou et al., 2005, 190028; Kahn et
al., 2005, 189961). The MISR multi -angle data also sample scattering angles ranging
from about 60° to 160° in midlatitudes, yielding information about particle size (Chen
et al., 2008, 189984; Kahn et al., 1998, 190970; 2001, 190969; 2005, 190966) and
shape (Kalashnikova and Kahn, 2006, 190962). The aggregate of aerosol
microphysical properties can be used to assess aerosol airmass type, a more robust
characterization of MISR-retrieved particle property information than individual
attributes. MISR also retrieves plume height in the vicinity of wildfire, volcano, and
mineral dust aerosol sources, where the plumes have discernable spatial contrast in
the multi-angle imagery (Kahn et al., 2007, 190964). Figure 9"57 is an example that
illustrates MISR's ability to characterize the load, optical properties, and stereo
height of near-source fire plumes.
POLarization and Directionality of the Earth's Reflectance, polder is a
unique aerosol sensor that consists of wide field-of-view imaging spectro-radiometer
capable of measuring multi-spectral, multi-directional, and polarized radiances
(Deuze et al., 2001, 192013). The observed radiances can be exploited to better
separate the atmospheric contribution from the surface contribution over both land
and ocean. POLDER "1 and "2 flew onboard the ADEOS (Advanced Earth Observing
Satellite) from November 1996 to June 1997 and April to October of 2003, respectively.
A similar POLDER instrument flies on the PARASOL satellite that was launched in
December 2004.
Figure 9"58 shows global horizontal patterns of AOD and Angstrom exponent over
the oceans derived from the POLDER instrument for June 1997. The oceanic AOD
map (Figure 9"58.a) reveals near-coastal plumes of high AOD, which decrease with
distance from the coast. This pattern arises from aerosol emissions from the
continents, followed by atmospheric dispersion, transformation, and removal in the
downwind direction. In large-scale flow fields, such as the trade winds, these
continental plumes persist over several thousand kilometers. The Angstrom exponent
shown in Figure 9"58 exhibits a very different pattern from that of the aerosol optical
depth; specifically, it exhibits high values downwind of industrialized regions and
regions of biomass burning, indicative of small particles arising from direct emissions
July 2009
9-117
DRAFT-DO NOT CITE OR QUOTE

-------
from combustion sources and/or gas-to-particle conversion, and low values associated
with large particles in plumes of soil dust from deserts and in sea salt aerosols.
July 2009
9-118
DRAFT-DO NOT CITE OR QUOTE

-------
Source: Adapted from (Chin et al., 2007,190062); original figure from Yoram Kaufman and Reto Stockli.
July 2009
9-119
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 9-56. A composite of MODIS/Terra observed aerosol optical depth (at 550 nm, green light near
the peak of human vision) and fine-mode fraction that shows spatial and seasonal
variations of aerosol types. Industrial pollution and biomass burning aerosols are
predominately small particles (shown as red), whereas mineral dust and sea salt consist
primarily of large particles (shown as green). Bright red and bright green indicate heavy
pollution and dust plumes, respectively.
a	b	c
Source: taken from Kahn et al. (2007,190964).
Figure 9-57. Oregon fire on September 4, 2003 as observed by MISR: (a) MISR nadir view of the fire
plume, with five patch locations numbered and wind-vectors superposed in yellow; (b)
MISR aerosol optical depth at 558 nm; and (c) MISR stereo height without wind
correction for the same region.
9.3.2.2. Focused Field Campaigns
Over the past two decades, numerous focused field campaigns have examined the
physical, chemical, and optical properties and radiative forcing of aerosols in a variety
of aerosol regimes around the world, as listed in Table 9-5. These campaigns, which
have been designed with aerosol characterization as the main goal or as one of the
major themes in more interdisciplinary studies, were conducted mainly over or
downwind of known continental aerosol source regions, but in some instances in low-
aerosol regimes, for contrast. During each of these comprehensive campaigns , aerosols
were studied in great detail, using combinations of in situ and remote sensing
observations of physical and chemical properties from various platforms (e .g. , aircraft,
ships, satellites, and ground-based stations) and numerical modeling. In spite of their
relatively short duration, these field studies have acquired comprehensive data sets of
July 2009
9-120
DRAFT-DO NOT CITE OR QUOTE

-------
Angstrom Exponent
(X = -d In rId In >.
regional aerosol properties that have been used to understand the properties and
evolution of aerosols within the atmosphere and to improve the climatology of aerosol
microphysical properties used in satellite retrieval algorithms and CTMs.
a
Optical Depth
>. 865 nrn
Reproduced with permission of Laboratoire d'Optique Atmospherique (LOA), Lille, FR; Laboratoire des Sciences du Climat et de I'Environnement
(LSCE), Gif sur Yvette, FR; Centre National d'etudes Spatiales (CNES), Toulouse, FR; and NAtional Space Development Agency (NASDA), Japan.
Figure 9-58. Global maps at 18 km resolution showing monthly average (a) AOD at 865 nm and (hj
o
Angstrom exponent of AOD over water surfaces only for June, 1997, derived from
radiance measurements by the POLDER.
9.3.2.3. Ground-Based In Situ Measurement Networks
Major U.S."operated surface in situ and remote Sensing networks for tropospheric
aerosol characterization and climate, forcing research are listed in Table !H>. These
surface in situ stations provide information about long-term changes and trends in
aerosol concentrations and properties, the influence of regional sources on aerosol
properties, climatologies of aerosol radiative properties, and data for testing models
and satellite aerosol retrievals. The. MQAAEarth System Research Laboratory (ESRL)
aerosol monitoring network consists of baseline, regional, and mobile Stations. These
near-surface measurements include submicrometer and sub-10 micrometer scattering
and absorption coefficients from which the extinction coefficient and single-scattering
albedo can be derived. Additional measurements include particle concentration and, at
selected sites, OCN concentration, the hygroscopic growth factor, and chemical
composition.
July 2009
9-121
DRAFT-DO NOT CITE OR QUOTE

-------
Several of the stations, which are located across North America and world-wide,
are in regions where recent focused field campaigns have been conducted. The
measurement protocols at the stations are similar to those used during the field
campaigns. Hence, the station data are directly comparable to the field campaign data
so that they provide a longer-term measure of mean aerosol properties and their
variability, as well as a context for the shorter-duration measurements of the field
campaigns.
The Interagency Monitoring of Protected Visual Environment (IMPROVE), which
is operated by the NP Service Air Resources Division, has stations across the U.S.
located within national parks (Malm et al., 1994, 044920). Although the primary focus
of the network is air pollution, the measurements are also relevant to climate forcing
research. Measurements include fine and coarse mode (PM2.5 and PM10) aerosol mass
concentration; concentrations of elements, sulfate, nitrate, organic carbon, and
elemental carbon! and scattering coefficients.
In addition, to these U.S.-operated networks, there are other national and
international surface networks that provide measurements of aerosol properties
including, but not limited to, the World Meteorological Organization (WMO) Global
Atmospheric Watch (GAW) network
(httpy/www.wmo.int/pages/prog/arep/gaw/monitoring.html). the European Monitoring
and Evaluation Programme (EMEP) (http 7/www.emep .int/). the Canadian Air and
Precipitation Monitoring Network (CAPMoN) (http V/www.msc-
smc.ec.gc.ca/capmon/index e.cfm). and the Acid Deposition Monitoring Network in
East Asia (EANET) (httpV/www.eanet.cc/eanet.html).
Clouds and the Earth's Radiant Energy System, ceres measures
broadband solar and terrestrial radiances at three channels with a large footprint
(e.g., 20 km for CERES/Terra) (Wielicki et al., 1996, 190637). It is co-located with
MODIS and MISR aboard Terra and with MODIS on Aqua. The observed radiances
are converted to TOAirradiances or fluxes using the Angular Distribution Models
(ADMs) that are functions of viewing angle, sun angle, and scene type (Loeb and Kato,
2002, 190432; Loeb et al., 2005, 190436; Zhang et al., 2005, 190929). Such estimates of
TOA solar flux in clear-sky conditions can be compared to the expected flux for an
aerosol-free atmosphere, in conjunction with measurements of aerosol optical depth
from other sensors (e.g., MODIS and MISR) to derive the aerosol direct radiative
forcing (Christopher et al., 2006, 155729; Loeb and Manalo-Smith, 2005, 190433;
Patadia et al., 2008, 190558; Zhang and Christopher, 2003, 190928; Zhang et al., 2005,
190930). The derived instantaneous value is then scaled to obtain a daily average. A
direct use of the coarse spatial resolution CERES measurements would exclude
aerosol distributions in partly cloudy CERES scenes. Several approaches that
incorporate coincident, high spatial and spectral resolution measurements (e.g.,
MODIS) have been employed to overcome this limitation (Loeb and Manalo-Smith,
2005, 190433; Zhang et al., 2005, 190930).
Active Remote Sensing of Aerosols. Following the success of a
demonstration of lidar system aboard the U.S. Space Shuttle mission in 1994, i.e.,
Lidar In-space Technology Experiment (LITE) (Winker et al., 1996, 190914). the
Geoscience Laser Altimeter System (GLAS) was launched in early 2003 to become the
first polar orbiting satellite lidar. It provides global aerosol and cloud profiling for a
one-month period out of every three-to-six months. It has been demonstrated that
GLAS is capable of detecting and discriminating multiple layer clouds, atmospheric
boundary layer aerosols, and elevated aerosol layers (e.g.,Spinhirne et al., 2005,
190410). The Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations
(CALIPSO), launched on April 28, 2006, is carrying a lidar instrument (Cloud and
Aerosol Lidar with Orthogonal Polarization - CALIOP) that has been collecting
profiles of the attenuated backscatter at visible and near-infrared wavelengths along
with polarized backscatter in the visible channel (Winker et al., 2003, 192017).
CALIOP measurements have been used to derive the above-cloud fraction of aerosol
extinction optical depth (Chand et al., 2008, 189974). one of the important factors
July 2009
9-122
DRAFT-DO NOT CITE OR QUOTE

-------
determining aerosol direct radiative forcing in cloudy conditions. Figure 9-59 shows
an event, of trans-Atlantic transport. of Saharan dust captured by CALIPSO. Flying in
formation with the Aqua, AURA, POLDER, and OloudSat satellites, the vertically
resolved information is expected to greatly improve passive aerosol and cloud
retrievals as well as allow the retrieval of vertical distributions of aerosol extinction,
fine- and coarse-mode separately (Huneeus and 0 Boucher, 2007, 190624' Kaufman et
al.. 2003, 190954; Leon et al., 2003, 190366).
Source: Liu et al. (2008,156709).
Figure 9-59. A dust event that originated in the Sahara desert on 17 August 2007 and was
transported to the Gulf of Mexico. Red lines represent back trajectories indicating the
transport track of the dust event. Vertical images are 532 nm attenuated backscatter
coefficients measured by 0ALI0P when passing over the dust transport track. The letter
"D" designates the dust layer, and "S" represents smoke layers from biomass burning in
Africa (17-19 August) and South America (22 August). The track of the high-spectral-
resolution-lidar (HSRL) measurement is indicated by the white line superimposed on the
28 August CALIPSO image. The HSRL track is coincident with the track of the 28 August
CALIPSO measurement off the coast of Texas between 28.75°N and 29.08°N.
July 2009
9-123
DRAFT-DO NOT CITE OR QUOTE

-------
9.3.2.4. In Situ Aerosol Profiling Programs
In addition to long-term ground based measurements, regular long-term aircraft
in situ measurements recently have been implemented at several locations. These
programs provide a statistically significant data set of the vertical distribution of
aerosol properties to determine spatial and temporal variability through the vertical
column and the influence of regional sources on that variability. In addition, the
measurements provide data for satellite and model validation. As part of its long-term
ground measurements, NOAAhas conducted regular flights over Bondville, Illinois
since 2006. Measurements include light scattering and absorption coefficients, the
relative humidity dependence of light scattering, aerosol number concentration and
size distribution, and chemical composition. The same measurements with the
exception of number concentration, size distribution, and chemical composition were
made by NOAA during regular overflights of DOE ARM's Southern Great Plains
(SGP) site from 2000-2007 (IOM et al., 2004, 190058)
(http'//www. esrl.noaa.gov/gmd/aero/net/index.html).
In summary of Sections 9.3.2.2, 9.3.2.3, and 9.3.2.4, in situ measurements of
aerosol properties have greatly expanded over the past two decades as evidenced by
the number of focused field campaigns in or downwind of aerosol source regions all
over the globe, the continuation of existing and implementation of new sampling
networks worldwide, and the implementation of regular aerosol profiling
measurements from fixed locations. In addition, in situ measurement capabilities
have undergone major advancements during this same time period. These
advancements include the ability to measure aerosol chemical composition as a
function of size at a time resolution of seconds to minutes (e.g.,Jayne et al., 2000,
190978). the development of instruments able to measure aerosol absorption and
extinction coefficients at high sensitivity and time resolution and as a function of
relative humidity (e.g.,Baynard et al., 2007, 151669; Lack et al., 2006, 190203). and
the deployment of these instruments across the globe on ships, at ground-based sites,
and on aircraft. However, further advances are needed to make this newly developed
instrumentation more affordable and turn "key so that it can be deployed more widely
to characterize aerosol properties at a variety of sites worldwide.
July 2009
9-124	DRAFT-DO NOT CITE OR QUOTE

-------
Figure 9-60. A constellation of five spacecraft that overfly the Equator at about 1:30 PM, the so-
called A-Train, carries sensors having complementary capabilities, offering
unprecedented opportunities to study aerosols from space in multiple dimensions.
Table 9-5. List of major intensive field experiments that are relevant to aerosol research in a variety
of aerosol regimes around the globe conducted in the past two decades.
Intensive Field Experiments
Aerosol Regimes	Major References
Name	Location	Time Period
Anthropogenic aerosol and TARFOX	North Atlantic	July 1996	Russell et al. (1999, 190363)
boreal forest from North
America and West Europe NEAQS	North Atlantic	July-August 2002	Ouinn and Bates (2003, 049189)
SCAR-A	North America	1993	Remeret al. (1997,190216)
CLAMS	East Coast of U.S.	July-August 2001	Smith et al. (2005,190401)
INlf X NA, ICARTT	North America	Summer 2004	Fehsenfeld et al. (2006, 190531)
DOE AI0P	Northern Oklahoma	May 2003	Ferrare et al. (2006,190561)
MILAGR0
Mexico City, Mexico
March 2006
Molina etal. (2008,192019)
TexAQS/GoMACCS
Texas and Gulf of Mexico
August-September 2006
Jiang et al. (2008, 156609);
Luet al, (2008, 190455)
ARCTS
North-central Alaska to
Greenland (Arctic haze)
March-April 2008
http:llwww.espo.nasa.gov/arctasl
ARCTAS
Northern Canada (smoke)
June-July 2008

MINOS
Mediterranean region
July-August 2001
Lelieveld et al. (2002,190361)
LACE98
Lindberg, Germany
July-August 1998
Ansmann et al, (2002)
July 2009
9-125
DRAFT-DO NOT CITE OR QUOTE

-------
Intensive Field Experiments

Aerosols99
Atlantic
January-February 1999
Bates et al. (2001, 043385)
Brown haze in South Asia
INDOEX
Indian subcontinent and
Indian Ocean
January-April 1998 & 1999
Ramanathan et al. (2001,190196)

ABC
South and East Asia
Ongoing
Ramanathan and Crutzen (2003,190198)
Anthropogenic aerosol and
EAST-AIRE
China
March-April 2005
Li et al. (2007,190392)
desert dust mixture from
East Asia
INTEX-B
Northeastern Pacific
April 2006
Singh et al. (2008,190394)

ACE-Asia
East Asian and Northwest
Pacific
April 2001
Huebert et al. (2003,190623);
Seinfeld et al. (2004,190388)

TRACE-P

March-April 2001
Jacob et al. (2003,190987)

PEM-West A & B
Western Pacific off East
Asia
September-October 1991
February-March 1994
Hoell et al. (1996,190607; 1997, 057373)
Biomass burning smoke in
BASE-A
Brazil
1989
Kaufman et al. (1992, 044557)
the tropics
SCAR-B
Brazil
August-September 1995
Kaufman et al. (1998, 089989)

LBA-SMOCC
Amazon basin
September-November 2002
Andreae et al. (2004,155658)

SAFARI2000
South Africa and South
August-September 2000
King et al. (2003, 094395)

SAFARI92
Atlantic
September-October 1992
Lindesav et al. (1996,190403)

TRACE-ADABEX
South Atlantic
September-October 1992
Fishman et al. (1996,190566)

DABEX
West Africa
January-February 2006
Haywood et al. (2008,190602)
Mineral dusts from North
SAMUM
Southern Morocco
May-June 2006
Heintzenberg et al. (2008,190605)
Africa and Arabian Peninsula
SHADE
West coast of North Africa
September 2000
Tanre et al. (2003,190454)

PRIDE
Puerto Rico
June-July 2000
Reid et al. (2003,190213)

UAE2
Arabian Peninsula
August-September 2004
Reid et al. (2008,190214)
Remote oceanic aerosol
ACE-1
Southern Oceans
December 1995
Bates et al. (1998,190063);
Quinn and Coffman (1998,190918)
Source: (Yu, 2006,156173) (Updated)
Table 9-6. Summary of major U.S. surface in situ and remote sensing networks for the tropospheric
aerosol characterization and radiative forcing research. All the reported quantities are
column-integrated or column-effective, except as indicated.
Surface Network
Measured/Derived Parameters
Loading Size, Shape Absorption Chemistry
Spatial Temporal
Coverage Coverage
In situ NOAA ESRL aerosol monitoring
Near-surface
Angstrom
Single-scattering
Chemical
5 baseline 1976 onward
(http://www.esrl.noaa.gov/gmd/aero/)
extinction
exponent,
albedo,
composition in
stations, several

coefficient,
hemispheric
absorption
selected sites and
regional

optical depth,
backscatter
coefficient
periods
stations,

CN/CNN number
fraction,


aircraft and

concentrations
asymmetry
factor,
hygroscopic
growth


mobile
platforms
July 2009
9-126
DRAFT-DO NOT CITE OR QUOTE

-------
Measured/Derived Parameters
NPSfEPA IMPROVE
(http:ff\rist8.eira.cnlo8tatB.8du)iroprov8j)
Near-surface
mass
concentrations
and derived
extinction
coefficients by
species
Fine and coarse
separately
Single-scattering
albedo,
absorption
coefficient
Ions, ammonium
SO42, ammonium
nitrate organics,
EC, fine soil
156 national
parks and
wilderness
areas in the
U.S.
1S88 onward
Remote NASA AERONET (http://aeronet-Hsfc.nasa.nov) Optical depth
Sensing
DOE ARM (http:ffwww.arm.gov)
NOAA SURFRAD
(http:ffwww.srrb. noaa.govfsurfradf)
AERONET-MAN
(http:ffaeronet.gsfc.nasa.govfmaritime aerosol
network.html)
Fine-mode
fraction,
Angstrom
exponents,
asymmetry
factor, phase
function, non-
spherical fraction
Single-scattering
albedo,
absorption optical
depth, refractive
indices
N/A
N/A
N/A
N/A
— 200 sites 1993 onward
over global land
and islands
6 sites and 1 1989 onward
mobile facility in
N. America,
Europe, and
Asia
7 sites in the
U.S.
1995 onward
Global Ocean 2004-present
periodically
NASA MP1.NET (http:ffmplnet.gsfc.n3sa.govf)
Vertical profiles N/A
of backscatterf
extinction
coefficient
N/A
N/A
— 30 sites in
major
continents,
usually
co-located with
AERONET and
ARM sites
2000 onward
Figure 9-61. Geographical coverage of active AERONET sites in 2006.
9.3.2.5. Ground-Based Remote Sensing Measurement Networks
The Aerosol Robotic Network (AERONET) program is a federated ground"based
remote sensing network of well-calibrated sun photometers and radiometers
(http 7/aeronet .gsfc .nasa.gov).
July 2009
9-127
DRAFT-DO NOT CITE OR QUOTE

-------
AERONET includes about 200 sites around the world, covering all major
tropospheric aerosol regimes (Holben et al., 1998, 155848; 2001, 190618). as
illustrated in Figure 9"61. Spectral measurements of sun and sky radiance are
calibrated and screene d for cloud-free conditions (Smirnov et al., 2000, 190397).
AERONET stations provide direct, calibrated measurements of spectral AOD
(normally at wavelengths of 440, 670, 870, and 1020 nm) with an accuracy of ± 0.015
(Eck et al., 1999, 190390). In addition, inversion-based retrievals of a variety of
effective, column-mean properties have been developed, including aerosol single-
scattering albedo, size distributions, fine-mode fraction, degree of non-sphericity,
phase function, and asymmetry factor (Dubovik and King, 2000, 190197! Dubovik et
al., 2000, 190177; Dubovik et al., 2002, 190202; O'Neill et al., 2003, 180187). The SSA
can be retrieved with an accuracy of ± 0.03, but only for AOD >0.4 (Dubovik et al.,
2002, 190202). which precludes much of the planet. These retrieved parameters have
been validated or are undergoing validation by comparison to in situ measurements
(e.g.,Haywood et al., 2003, 190599; Leahy et al., 2007, 190232; Magi et al., 2005,
190468).
Recent developments associated with AERONET algorithms and data products
include-
¦	simultaneous retrieval of aerosol and surface properties using combined AERONET
and satellite measurements (Sinyuk and et, 2007, 190395) with surface reflectance
taken into account (which significantly improves AERONET SSA retrieval accuracy)
(Eck et al., 2008, 190409);
¦	the addition of ocean color and high frequency solar flux measurements; and
¦	the establishment of the Maritime Aerosol Network (MAN) component to monitor
aerosols over the World oceans from ships of- opportunity (Smirnov et al., 2006,
190400).
Because of consistent calibration, cloud-screening, and retrieval methods,
uniformly acquired and processed data are available from all stations, some of which
have operated for over 10 years. These data constitute a high-quality, ground-based
aerosol climatology and, as such, have been widely used for aerosol process studies as
well as for evaluation and validation of model simulation and satellite remote sensing
applications (e.g., Chin et al., 2002, 189996; Kahn RAGaitley et al., 2005, 190966;
Remer et al., 2005, 190221; Yu et al., 2003, 156171; 2006, 156173). In addition,
AERONET retrievals of aerosol size distribution and refractive indices have been used
in algorithm development for satellite sensors (Levy et al., 2007, 190377; Remer et al.,
2005, 190221). A set of aerosol optical properties provided by AERONET has been
used to calculate the aerosol direct radiative forcing (Procopio et al., 2004, 190571;
Zhou M, et, 2005, 156183). which can be used to evaluate both satellite remote
sensing measurements and model simulations.
AERONET measurements are complemented by other ground-based aerosol
networks having less geographical or temporal coverage, such as the Atmospheric
Radiation Measurement (ARM) network (Ackerman and Stokes, 2003, 192080).
NOAA's national surface radiation budget network (SURFRAD) (Augustine et al.,
2008, 189913) and other networks with multifilter rotating shadowband radiometer
(MFRSR) (Harrison et al., 1994, 045805; Michalsky et al., 2001, 190537). and several
lidar networks including-
¦	NASA Micro Pulse Lidar Network (MPLNET) (Welton et al., 2001, 157133) (2002,
190631);
¦	Regional East Atmospheric Lidar Mesonet (REALM) in North America (Hoff and
McCann, 2002, 190612; 2004, 190617);
¦	European Aerosol Research Lidar Network (EARLINET) (Matthias, 2004, 155971); and
¦	Asian Dust Network (AD-Net) (e.g.,Murayama T, et, 2001, 155992).
Obtaining accurate aerosol extinction profile observations is pivotal to improving
aerosol radiative forcing and atmospheric response calculations. The values derived
from these lidar networks with state-of-the-art techniques (Schmid et al., 2006,
190372) are helping to fill this need.
July 2009
9-128
DRAFT-DO NOT CITE OR QUOTE

-------
9.3.2.6. Synergy of Measurements and Model Simulations
Individual approaches discussed above have their own strengths and limitations,
and are usually complementary. None of these approaches alone is adequate to
characterize large spatial and temporal variations of aerosol physical and chemical
properties and to address complex aerosol-climate interactions. The best strategy for
characterizing aerosols and estimating their radiative forcing is to integrate
measurements from different satellite sensors with complementary capabilities from
in situ and surface based measurements. Similarly, while models are essential tools
for estimating regional and global distributions and radiative forcing of aerosols at
present as well as in the past and the future, observations are required to provide
following, several synergistic approaches to studying aerosols and their radiative
forcing are discussed.
Closure Experiments. During intensive field studies, multiple platforms and
instruments are deployed to sample regional aerosol properties through a well-
coordinated experimental design. Often, several independent methods are used to
measure or derive a single aerosol property or radiative forcing. This combination of
methods can be used to identify inconsistencies in the methods and to quantify
uncertainties in measured, derived, and calculated aerosol properties and radiative
forcings. This approach, often referred to as a closure experiment, has been widely
employed on both individual measurement platforms (local closure) and in studies
involving vertical measurements through the atmospheric column by one or more
platforms (column closure) (Quinn et al., 1996, 192021 j Russell et al., 1997, 190359).
Past closure studies have revealed that the best agreement between methods
occurs for submicrometer, spherical particles such that different measures of aerosol
optical properties and optical depth agree within 10-15% and often better (e.g.,Clarke
et al., 1996, 190003; Collins et al., 2000, 190059; Schmid and et, 2000, 190369; Quinn
et al., 2004, 190937). Larger particle sizes (e.g., sea salt and dust) present inlet
collection efficiency issues and non-spherical particles (e.g., dust) lead to differences in
instrumental responses. In these cases, differences between methods for determining
aerosol optical depth can be as great as 35% (Doherty et al., 2005, 190027; Wang et al.,
2003, 157106). Closure studies on aerosol clear-sky DRF reveal uncertainties of about
25% for sulfate/carbonaceous aerosol and 60% for dust-containing aerosol (Bates et al.,
2006, 189912). Future closure studies could integrate surface- and satellite-based
radiometric measurements of AOD with in situ optical, microphysical, and aircraft
radiometric measurements for a wide range of situations. There is also a need to
maintain consistency in comparing results and expressing uncertainties (Bates et al.,
2006, 189912).
Constraining Models with In Situ Measurements. In situ measurements of
aerosol chemical, microphysical, and optical properties with known accuracy, based in
part on closure studies, can be used to constrain regional CTM simulations of aerosol
direct forcing, as described by Bates et al. (2006, 189912). A key step in the approach
is assigning empirically derived optical properties to the individual chemical
components generated by the CTM for use in a Radiative Transfer Model (RTM).
Specifically, regional data from focused, short-duration field programs can be
segregated according to aerosol type (sea salt, dust, or sulfate/carbonaceous) based on
measured chemical composition and particle size. Corresponding measured optical
properties can be carried along in the sorting process so that they, too, are segregated
by aerosol type. The empirically derived aerosol properties for individual aerosol
types, including mass scattering efficiency, single-scattering albedo, and asymmetry
factor, and their dependences on relative humidity, can be used in place of assumed
values in CTMs. Short-term, focused measurements of aerosol properties (e.g., aerosol
concentration and AOD) also can be used to evaluate CTM parameterizations on a
regional basis, to suggest improvements to such uncertain model parameters, such as
emission factors and scavenging coefficients (e.g., Koch et al., 2007, 190185).
Improvements in these parameterizations using observations yield increasing
July 2009
9-129
DRAFT-DO NOT CITE OR QUOTE

-------
confidence in simulations covering regions and periods where and when
measurements are not available. To evaluate the aerosol properties generated by
CTMs on broader scales in space and time, satellite observations and long-term in situ
measurements are required.
Improved Model Simulations with Satellite Measurements. Global
measurements of aerosols from satellites (mainly AOD) with well-defined accuracies
offer an opportunity to evaluate model simulations at large spatial and temporal
scales. The satellite measurements can also be used to constrain aerosol model
simulations and hence the assessment of aerosol DRF through data assimilation or
objective analysis process (e.g., Collins et al., , 189987; Yu et al., 2003, 156171 j 2004,
190926; 2006, 156173; Liu et al., 2005, 190414; Zhang et al., 2008, 190932). Both
satellite retrievals and model simulations have uncertainties. The goal of data
integration is to minimize the discrepancies between them, and to form an optimal
estimate of aerosol distributions by combining them, typically with weights inversely
proportional to the square of the errors of individual descriptions. Such integration
can fill gaps in satellite retrievals and generate global distributions of aerosols that
are consistent with ground-based measurements (Collins et al., , 189987; Yu et al.,
2003, 156171; 2006, 156173; Liu et al., 2005, 190414). Recent efforts have also focused
on retrieving global sources of aerosol from satellite observations using inverse
modeling, which may be valuable for reducing large aerosol simulation uncertainties
(Dubovik et al., 2007, 190211). Model refinements guided by model evaluation and
integration practices with satellite retrievals can then be used to improve aerosol
simulations of the pre- and post-satellite eras. Current measurement-based
understanding of aerosol characterization and radiative forcing is assessed in
Section 9.3.3 through intercomparisons of a variety of measurement-based estimates
and model simulations published in literature. This is followed by a detailed
discussion of major outstanding issues in Section 9.3.4.
July 2009
9-130
DRAFT-DO NOT CITE OR QUOTE

-------
E
s
Source: Bates et al. !2008,189912).
Figure 9-62. Comparison of the mean concentration (Clg/m3) and standard deviation of the modeled
(STEM) aerosol chemical components with shipboard measurements during INDOEX,
ACE Asia, and ICARTT.
Further complexity is added when attempting to relate surface PM2.5 to aerosol optical
depths. The main approach to derive surface PM2.5 from satellite optical depths from MISR
is based on the use of model derived profiles to determine ratios of aerosol optical depth to
surface PM2.5 (Liu and Koutrakis, 2007, 187007; 2007, 192180; Van Donkelaar et al., 2006,
192108). Van Donkelaar et al. (2006, 192108) derived r = 0.69 (MODIS) and 0.58 (MISR)
with annual average ground based PM2.5 across the U.S. and Canada, For comparison, r
between AERONET total AOD and surface measurements was 0.71. On average, MODIS
tended to overestimate surface PM2.5 by about 5 ug/m:;, while MISR estimates were biased
high by about 3 ,ug/m3. Liu et al. (2007, 187007; 2007, 192180) used MISR derived fractional
AODs in their analysis and found improvement in the retrievals when fractional AODs
were used instead of total AOD, allowing for better fits to the radiance data. They found
that fractional AODs can explain 13-62% of the variability in PM2.5 and its components in
the eastern U.S. and 28-56% of the variability in the western U.S. The models tended to
m	SKM -BD€>.
~	Otrtervftfebf* - WOO
m	STEW ACE Abu
B viHwtt -
m	ST£M-IC4mr
C-3	CfcHfirvalmn - ItAHTT
Arnmon/wCTi
m	STEM WDEIEX
E=3	GfcMrwADOH MDGEX
m	STEM - ACE-Aam
S3	OfcBifvirten - ACE-A«*
mp	gTSM CAFrTT
EZJ	OhsBrvBtxni CARTT
¦II m,
Sulfate Ammonium	BC
S^al0
POM
Sea Salt
July 2009
9-131
DRAFT-DO NOT CITE OR QUOTE

-------
underpredict PM2.5 by ~7"8% in both the East and West. The relative errors in surface PM2.5
were estimated to be 30% in the East and 34% in the West. For AODs >0.15 (nominal
continental background values), dust particles could be distinguished from other particles
with an estimated error of 4%. Performance improves substantially over polluted urban
areas because they have much larger AOD. For example, Gupta et al. (2006, 137694)
derived a Pearson r between MODIS AOD and surface PM2.5 of 0.96 for several urban areas
around the world.
9.3.3. Assessments of Aerosol Characterization and Climate
Forcing
Sections 9.3.3 through 9.3.6 come directly from CCSP SAP2.3 Chapters 2, Section 2.3
through Chapter 3, Section 3.8, with section, table, and figure numbers changed to be
internally consistent with this ISA.
This section focuses on the assessment of measurement-based aerosol
characterization and its use in improving estimates of the direct radiative forcing on
regional and global scales. In situ measurements provide highly accurate aerosol
chemical, microphysical, and optical properties on a regional basis and for the
particular time period of a given field campaign. Remote sensing from satellites and
ground-based networks provide spatial and temporal coverage that intensive field
campaigns lack. Both in situ measurements and remote sensing have been used to
determine key parameters for estimating aerosol direct radiative forcing including
aerosol single scattering albedo, asymmetry factor, optical depth Remote sensing has
also been providing simultaneous measurements of aerosol optical depth and
radiative fluxes that can be combined to derive aerosol direct radiative forcing at the
TOA with relaxed requirement for characterizing aerosol properties. Progress in using
both satellite and surface-based measurements to study aerosol-cloud interactions and
aerosol indirect forcing is also discussed.
9.3.3.1. The Use of Measured Aerosol Properties to Improve Models
The wide variety of aerosol data sets from intensive field campaigns provides a
rigorous "testbed" for model simulations of aerosol properties and distributions and
estimates of DRF. As described in Section 9.3.2.6, in situ measurements can be used to
constrain regional CTM simulations of aerosol properties, DRF, anthropogenic
component of DRF, and to evaluate CTM parameterizations. In addition, in situ
measurements can be used to develop simplifying parameterizations for use by CTMs.
Several factors contribute to the uncertainty of CTM calculations of size-
distributed aerosol composition including emissions, aerosol removal by wet
deposition, processes involved in the formation of secondary aerosols and the chemical
and microphysical evolution of aerosols, vertical transport, and meteorological fields
including the timing and amount of precipitation, formation of clouds, and relative
humidity. In situ measurements made during focused field campaigns provide a point
of comparison for the CTM-generated aerosol distributions at the surface and at
discrete points above the surface. Such comparisons are essential for identifying areas
where the models need improvement.
July 2009
9-132
DRAFT-DO NOT CITE OR QUOTE

-------
(Winter).
5.2 iigrtn
ISurnmur).
14.0 ug.m
PSf,, NY
12
13 pgkVn
Ho-tiS-ioik
T*
Wrtraai* gli
^ 2.1 Mg>Pn'
«
1 & yfli'm'1
12
TORCH 1. UK TOftCH 2, UK
T.ifi jig'm'
Hyyiia-la
Fnitencl,
2.0 M9>'ni
QUEST Tawny*, Q««m*ny
Finland ,
2.S Mg/m	liM^Wt
31 pgftn
0
Mexico City
2.8
Duke Forest
NC
BS^yn"
ort Coast NE
US
Jungfimijoclh
Switzerland
• 0
Chejii Island Okinawa Island
Korea	Japan
Mace Head
Ireland
Fukuc Island
Japan
| |Qrgarfccs	Sutf-ato	Nitrate jj^H ^nirnci^irn |HB Chiorkfe	Urtwn Downwind Urban Rural - ftomofic
Source: Adapted from Zhang et al. (2007,189998).
Figure 9-63. Location of aerosol chemical composition measurements with aerosol mass
spectrometers. Colors for the labels indicate the type of sampling location: urban areas
(blue), < 100 miles downwind of major cites (black), and rural/remote areas > 100 miles
downwind (pink). Pie charts show the average mass concentration and chemical
composition: organics (green), Sth2' (red), nitrate (blue), ammonium (orange), and chloride
(purple), of non-refractory PMi.
Figure 9-62 shows a comparison of submicrometer and supermicromet'er aerosol
chemical components measured during INT.'1 >MX. AOEAsia, andlOARTT onboard a
ship and the same values; calculated with the STEM Model (e.g., Bates et al.. 2004,
189958. Garmichael et al., 2002, 148319: Carmichael et al., 2003, 190042: Streets et
al.. 2006, 157019: Tang ami el. 2003, 190441; Tang and et, 2004. 190445). To permit
direct comparison of the measured and modeled values;, the model was driven by
analyzed meteorological data and sampled at the times and locations of the shipboard
measurements every 80 rain along the cruise track. The best agreement was found for
submicrometer sulfate and BO. The agreement was best for sulfate; this is attributed
to greater accuracy in emissions, chemical conversion, and removal for this
component:. Underestimation of dust and sea salt is most likely due to errors in model-
calculated emissions. Large discrepancies between the modeled and measured values
occurred for submicrometer particulate organic matter (POM) (INDOEX), and for
particles in the supermicrometer size range such as dust (AGE-Asia), and sea salt (all
regions). The model underestimated the total mass of the supermicrometer aerosol by
about a factor of 3. POM makes up a large and variable fraction of aerosol mass
throughout the anthropogenically influenced northern hemisphere, ami yel models
have severe problems in properly representing this type of aerosol. Much of this
discrepancy follows from the models inability to represent the formation of secondary
organic aerosols (SOA) from the precursor volatile organic: compounds (VQC) . Figure
9-63 shows a summary of the results from aerosol mass spectrometer measurements;
July 2009
9-133
DRAFT-DO NOT CITE OR QUOTE

-------
at 30 sites over North America, Europe, and Asia. Based on aircraft measurements of
urban-influenced air over New England, de Gouw et al. (2005, 190020) found that
POM was highly correlated with secondary anthropogenic gas phase species
suggesting that the POM was derived from secondary anthropogenic sources and that
the formation took one day or more,
1 -]¦
0
10
2D
30
P^otDCtwmiGal age (h I
i i i r
0,0 0,5 1.0 1.5
Acetylene (ppbv)
1	1	1	[	
0 10 20 30
Iso-propyl nitrate (pptv)
Source: Adapted from de Guow et al. (2005,190020).
Figure 9-64. Scatterplots of the submicrometer POM measured during NEAQS versus A) acetylene
and Bj iso-propyl nitrate. The colors of the data points in A) denote the photochemical
age as determined by the ratios of compounds of known OH reactivity. The gray area in
A) shows the range of ratios between submicrometer POM and acetylene observed by
Kirchstetter et al. (1999, 010642) in tunnel studies.
Figure 9"64 shows seni lorplois of submicrometer POM versus acetylene (a gas
phase primary emitted \ < ><' species) and isopropyl nitrate (a secondary gas phase
organic species formed by atmospheric reactions). The increase in submicrometer
POM with increasing photochemical age could not be explained by the removal of
VOO alone, which are its traditionally recognized precursors. This result suggests
that other species must have contributed and/or that the mechanism for POM
formation is more efficient than assumed by models. Similar results were obtained
from the 2006 MILAGEQ field campaign conducted in Mexico City (IvLeinman et al.,
2008, 190074). and comparisons of GCM results with several long-term monitoring
stations also showed that the model underestimated organic aerosol concentrations
(Koch et; al,, 2007, 190185). Recent laboratory work suggests that isoprene may be a
major SOAsource missing from previous atmospheric models (Kroll et al., 2006,
190195; Henze and Seinfeld, 2006) 190606), but underestimating sources from certain
economic sectors may also play a role (Koch et al., 2007, 190185). Models also have
difficulty in representing the vertical distribution of organic aerosols, underpredicting
their occurrence in the free 'troposphere (FT) (Heakl et al., 2005, 190603). While
organic aerosol presents models with some of their greatest challenges, even the
distribution of well-characterized sulfate aerosol is not always estimated correctly in
models (Shindell el al.. 2008, 190391).
Comparisons of DRF and its anthropogenic component calculated with assumed
optical properties and values constrained by in situ measurements Can help identify
July 2009
9-134
DRAFT-DO NOT CITE OR QUOTE

-------
areas of uncertainty in model parameterizations. In a study described by Bates et al.
(2006, 189912). two different CTMs (MOZART and STEM) were used to calculate dry
mass concentrations of the dominant aerosol species (sulfate, organic carbon, BC, sea
salt, and dust).
In situ measurements were used to calculate the corresponding optical properties
for each aerosol type for use in a radiative transfer model. Aerosol DRF and its
anthropogenic component estimated using the empirically derived and a priori optical
properties were then compared. The DRF and its anthropogenic component were
calculated as the net downward solar flux difference between the model state with
aerosol and of the model state with no aerosol. It was found that the constrained
optical properties derived from measurements increased the calculated AOD (34 ±
8%), TOA DRF (32 ± 12%), and anthropogenic component of TOA DRF (37 ± 7%)
relative to runs using the a priori values. These increases were due to larger values of
the constrained mass extinction efficiencies relative to the a priori^ alues. In addition,
differences in AOD due to using the aerosol loadings from MOZART versus those from
STEM were much greater than differences resulting from the apriorivs. constrained
RTM runs. In situ observations also can be used to generate simplified
parameterizations for CTMs and RTMs thereby lending an empirical foundation to
uncertain parameters currently in use by models. CTMs generate concentration fields
of individual aerosol chemical components that are then used as input to radiative
transfer models (RTMs) for the calculation of DRF. Currently, these calculations are
performed with a variety of simplifying assumptions concerning the RH dependence of
light scattering by the aerosol. Chemical components often are treated as externally
mixed each with a unique RH dependence of light scattering. However, both model
and measurement studies reveal that POM, internally mixed with water-soluble salts,
can reduce the hygroscopic response of the aerosol, which decreases its water content
and ability to scatter light at elevated relative humidity (e.g., Saxena et al., 1995,
077273; Carrico et al., 2005, 190052). The complexity of the POM composition and its
impact on aerosol optical properties requires the development of simplifying
parameterizations that allow for the incorporation of information derived from field
measurements into calculations of DRF (Quinn et al., 2005, 156033). Measurements
made during INDOEX, ACE-Asia, and ICARTT revealed a substantial decrease in
iasp (RH) with increasing mass fraction of POM in the accumulation mode. Based on
these data, a parameterization was developed that quantitatively describes the
relationship between POM mass fraction and iosp (RH)for accumulation mode sulfate-
POM mixtures (Quinn et al., 2005, 156033). This simplified parameterization may be
used as input to RTMs to derive values of iosp (RH) based on CTM estimates of the
POM mass fraction. Alternatively, the relationship may be used to assess values of
£>sp (RH) currently being used in RTMs.
9.3.3.2. Intercomparisons of Satellite Measurements and Model Simulation of Aerosol
Optical Depth
As aerosol DRF is highly dependent on the amount of aerosol present, it is of first-
order importance to improve the spatial characterization of AOD on a global scale.
This requires an evaluation of the various remote sensing AOD data sets and
comparison with model-based AOD estimates. The latter comparison is particularly
important if models are to be used in projections of future climate states that would
result from assumed future emissions. Both remote sensing and model simulation
have uncertainties and satellite-model integration is needed to obtain an optimum
description of aerosol distribution.
Figure 9"65 shows an intercomparison of annual average AOD at 550 nm from
two recent satellite aerosol sensors (MODIS and MISR), five model simulations
(GOCART, GISS, SPRINTARS, LMDZ-LOA, LMDZ-INCA) and three satellite-model
integrations (MO_GO, MI_GO, MO_MI_GO). These model-satellite integrations are
conducted by using an optimum interpolation approach (Yu et al., 2003, 156171) to
July 2009
9-135
DRAFT-DO NOT CITE OR QUOTE

-------
constrain GOCART simulated AOD with that from MODIS, MISR, or MODIS over
ocean and MISR over land, denoted as MO_GO, MI_GO, and MO_MI_GO,
respectively. MODIS values of AOD are from Terra Collection 4 retrievals and MISR
AOD is based on early post launch retrievals. MODIS and MISR retrievals give a
comparable average AOD on the global scale, with MISR greater than MODIS by
0.01-0.02 depending on the season. However, differences between MODIS and MISR
are much larger when land and ocean are examined separately- AOD from MODIS is
0.02-0.07 higher over land but 0.03-0.04 lower over ocean than the AOD from MISR.
Several major causes for the systematic MODIS-MISR differences have been
identified, including instrument calibration and sampling differences, different
assumptions about ocean surface boundary conditions made in the individual retrieval
algorithms, missing particle property or mixture options in the look-up tables, and
cloud screening (Kahn et al., 2007, 190963). The MODIS-MISR AOD differences are
being reduced by continuous efforts on improving satellite retrieval algorithms and
radiance calibration. The new MODIS aerosol retrieval algorithms in Collection 5
have resulted in a reduction of 0.07 for global land mean AOD (Levy et al., 2007,
190379). and improved radiance calibration for MISR removed ~40% of AOD bias over
dark water scenes (Kahn et al., 2005, 190965).
The annual and global average AOD from the five models is 0.19 ± 0.02 (mean ±
standard deviation) over land and 0.13 ± 0.05 over ocean, respectively. Clearly, the
model-based mean AOD is smaller than both MODIS" and MISRderived values
(except the GISS model). A similar conclusion has been drawn from more extensive
comparisons involving more models and satellites (Kinne et al., 2006, 155903). On
regional scales, satellite-model differences are much larger. These differences could be
attributed in part to cloud contamination (Kaufman et al., 2005, 155891; Zhang et al.,
2005,	190931) and 3D cloud effects in satellite retrievals (Kaufman et al., 2005,
155891! Wen et al., 2006, 179964) or to models missing important aerosol
sources/sinks or physical processes (Koren et al., 2007, 190192). Integrated satellite-
model products are generally in-between the satellite retrievals and the model
simulations, and agree better with AERONET measurements (e.g.,Yu et al., 2003,
156171). As in comparisons between models and in situ measurements (Bates et al.,
2006,	189912). there appears to be a relationship between uncertainties in the
representation of dust in models and the uncertainty in AOD, and its global
distribution.
For example, the GISS model generates more dust than the other models
(Figure 9-66), resulting in a closer agreement with MODIS and MISR in the global
mean (Figure 9"65). However, the distribution of AOD between land and ocean is quite
different from MODIS" and MISRderived values.
Figure 9"66 shows larger model differences in the simulated percentage
contributions of individual components to the total aerosol optical depth on a global
scale, and hence in the simulated aerosol single-scattering properties (e.g., single-
scattering albedo, and phase function), as documented in Kinne et al. (2006, 155903).
This, combined with the differences in aerosol loading (as characterized by AOD)
determines the model diversity in simulated aerosol direct radiative forcing, as
discussed later. However, current satellite remote sensing capability is not sufficient
to constrain model simulations of aerosol components.
July 2009
9-136
DRAFT-DO NOT CITE OR QUOTE

-------
Lind ¦Ocean ¦ Global
Figure 9-65. Comparison of annual mean aerosol optical depth (AOD)
100%
80%
60%
40%
20%
0%
&
iSS
DU
POM
IBC
SU


&

-------
Global Distributions. Figure 9-67 shows global distributions of aerosol optical
depth at 550 nm (left panel) and diurnally averaged clear-sky TOA DRF (right panel)
for March-April-May (MAM) based on the different approaches. The DRF at the
surface follows the same pattern as that at the TOA but is significantly larger in
magnitude because of aerosol absorption. It appears that different approaches agree
on large-scale patterns of aerosol optical depth and the direct radiative forcing. In this
season, the aerosol impacts in the Northern Hemisphere are much larger than those
in the Southern Hemisphere. Dust outbreaks andbiomass burning elevate the optical
depth to more than 0.3 over large parts of North Africa and the tropical Atlantic. In
the tropical Atlantic, TOA cooling as large as "10 W/m2 extends westward to Central
America. In eastern China, the optical depth is as high as 0.6-0.8, resulting from the
combined effects of industrial activities and biomass burning in the south, and dust
outbreaks in the north. The Asian impacts also extend to the North Pacific, producing
a TOA cooling of more than "10 W/m2. Other areas having large aerosol impacts
include Western Europe, midlatitude North Atlantic, and much of South Asia and the
Indian Ocean. Over the "roaring forties" in the Southern Hemisphere, high winds
generate a large amount of sea salt. Elevated optical depth, along with high solar
zenith angle and hence large backscattering to space, results in a band of TOA cooling
of more than "4 W/m2. However, there is also some question as to whether thin cirrus
(e.g., Zhang et al., 2005, 190931) and unaccounted-for whitecaps contribute to the
apparent enhancement in AOD retrieved by satellite. Some differences exist between
different approaches. For example, the early post-launch MISR retrieved optical
depths over the southern hemisphere oceans are higher than MODIS retrievals and
GOCART simulations. Over the "roaring forties," the MODIS derived TOA solar flux
perturbations are larger than the estimates from other approaches.
Table 9-7. Summary of approaches to estimating the aerosol direct radiative forcing in three
categories: (A) satellite retrievals; (B) satellite-model integrations; and (C) model
simulations.
Category Product	Brief Description	Identified Sources of Uncertainty	Major References
A Satellite MODIS Using MODIS retrievals of a linked set of Radiance calibration, cloudaerosol discrimination,	Remer and Kaufman (2006,
retrievals	AOD, Wo, and phase function consistently instantaneous-to-diurnal scaling, RTM parameterizations 190222)
in conjunction with a radiative transfer
model (RTM) to calculate TOA fluxes that
best match the observed radiances.
M0DIS A Splitting MODIS AOD over ocean into Satellite AOD and FMF retrievals, overestimate due to Bellouin et al. (2005,155684)
mineral dust, sea salt, and biomass-burning summing up the compositional direct forcing, use of a single
and pollution; using AER0NET	AER0NET site to characterize a large region
measurements to derive the size
distribution and single-scattering albedo for
individual components.
CERESA Using CERES fluxes in combination with Calibration of CERES radiances, large CERES footprint,
standard MODIS aerosol.
CERES B Using CERES fluxes in combination with
N0AA NESDIS aerosol from MODIS
radiances.
CERES C Using CERES fluxes in combination with
MODIS (ocean) and MISR (non-desert land)
aerosol with new angular models for
aerosols.
satellite AOD retrieval, radiance-to-flux conversion (ADM),
' instantaneous-to-diurnal scaling, narrow-to-broadband
conversion
Loeb and Manalo-Smith (2005,
190433); Loeb and Kato (2002,
190432)
Zhang et al. (2005, 086743;
2005,157185); Zhang and
Christopher (2003,190928);
Christopher et al. (2006,155729);
Patadia et al. (2008,190558)
July 2009
9-138
DRAFT-DO NOT CITE OR QUOTE

-------
Category
Product
Brief Description
identified Sources of Uncertainty
Major References

POLDER
Using POLDER AOD in combination with
prescribed aerosol models (similar to
MODIS).
Similar to MODIS
Boucher and Tanre (2000,
190041); Bellouin et al. (2003,
189911)
B. Satellite-
MODISG
Using GOCART simulations to fill AOD gaps
Propagation of uncertainties associated with both satellite
"Aerosol single-scattering albedo
model
integrations
MISRG
in satellite retrievals.
retrievals and model simulations (but the model-satellite
integration approach does result in improved AOD quality for
and asymmetry factor are taken
from GOCART simulations

MOGO
Integration of MODIS and GOCART AOD.
MO GO, and O MI GO)
*Yu et al. (2003,156171:2004,

MOMIGO
Integration of GOCART AOD with retrievals
from MODIS (Ocean) and MISR (Land).

190926; 2006,156173)

SeaWiFS
Using SeaWiFS AOD and assumed aerosol
models.
Similar to MODIS G and MISR G, too weak aerosol
absorption
Chouet al. (2002,190008)
C. Model
GOCART
Offline RT calculations using monthly avg
Emissions, parameterizations of a variety of sub-grid aerosol
Chin et al. (2002,189996); Yu et
simulations

aerosols with a time step of 30 min
(without the presence of clouds).
processes (e.g., wet and dry deposition, cloud convection,
aqueous-phase oxidation), assumptions on aerosol size,
al. (2004,190926)

SPRINTARS Online RT calculations every 3 hrs (cloud
fraction - 0).
absorption, mixture, and humidification of particles,
Meteorology fields, not fully evaluated surface albedo
schemes, RT parameterizations
Takemura et al. (2002,190438;
2005,190439)

GISS
Online model simulations and weighted by
clear-sky fraction.

Koch and Hansen (2005,190183);
Koch et al. (2006,190184)

LMDZ-INCA
Online RT calculations every 2 hrs (cloud
fraction - 0).

Balkanski et al. (2007,189979);
Schulz et al. (2006,190381);
Kinne et al. (2006,155903)

LMDZ-LOA
Online RT calculations every 2 hrs (cloud
fraction - 0).

Reddv et al. (2005,190207;
2005,190208)
Source: Adapted
I from Yu et al. (2006,156173).


July 2009
9-139
DRAFT-DO NOT CITE OR QUOTE

-------
MOD S
MODIS
MISR
j%v*.
¦^k *mAfy «¦ r*.r v.*-"
¦ • 'V'r"
MO_MI_GO
- - •
'***'' * " * -lllr
55" V	•
GOCART
CERES A
: . 	
: vi>,
• - A-'~^-r-|S. -*-'..
, ,
MO Ml GO

-^Tv^
GOCART
^:',2


.0 .05 .1 .15 .2 .3 .4 .5 .6 .8 I. -30 -20 -IS -10 -8 -6 -4 -2 0 5Wm
Figure 9-67. Geographical patterns of seasonally (MAM) averaged aerosol optical depth at 550 nm
(left panel) and the diurnally averaged clear-sky aerosol direct radiative (solar spectrum)
forcing (W|m2) at the T0A (right panel) derived from satellite (Terra) retrievals. MODIS
(Remer et al., 2005,190221; Remer and Kaufman, 2006,190222); MISR (Kahn RA
Gaitley et al., 2005,190966); and CERESA (Loeb and Manalo-Smith, 2005,110433);
GOCART simulations (Chin et al., 2002,189996; Yu et al., 2004,190926); and GOCART
M0DIS-MISR integrations (M0 Ml GO) (Yu H, et, 2006,156173).
July 2009
9-140
DRAFT - DO NOT CITE OR QUOTE

-------
Global Mean. Figure 9-68 summarizes the measurement- and model-based
.estimates of clear-sky annual-averaged DRF at both the TOA and surface from 60°S to
60°N. Seasonal DRF values for individual estimates are summarized in Table. 9-8 and
Table 9"9 for ocean and land, respectively. Mean, median and standard error e
(s=a/(n-1)1/2), where ois standard deviation and n is the number of methods) are
calculated for measurement- and model-based.estimates separately. Note that
although the standard deviation or standard error reported here is not a fully rigorous
measure of a true experimental uncertainty, it is indicative of the uncertainty because
independent approaches with independent sources of errors are used (see Table 9-7; in
the modeling community, this is called the "diversity ;" see Section 9.3.6).
Ocean: For the TOA DRF, a majority of measurement-based and satellite-model
integration "based estimates agree with each other within about 10%. On annual
average, the measurement-based estimates give, the DRF of -.5.5 ± 0.2 \Y. m:: (mean ± e)
at the TOA and -8.7 ±0.7 W in- at the surface. This suggests that the ocean surface,
cooling is. about 60% larger than the coohng at the TOA- Model simulations give wide
ranges of DRF estimates at both the TOA and surface. The ensemble of five models
gives the annual average DRF (mean ± s) of -3.2 ± 0.6 W/m2 and '4,9 ±0.8 W/m2 at the
TOA and surface, respectively. On: average, the surface cooling is about 37%.larger
than the TOA cooling, smaller than the measurement-based estimate of surface and
TOA difference of 60%, However, the 'measurement-based' estimate of surface DRF is
actually a: calculated value, using poorly constrained particle properties.
Land: It remains challenging to use satellite measurements alone for
characterizing complex aerosol properties over land surfaces with high accuracy. As
such , DRF estimates over land have to rely largely on model simulations and satellite-
model integrations. On a global and annual average, the satellite-model integrated
approaches derive a mean DRF of -4.9 W/m2 at the TOA and -11.9 W/m2 at the surface
respectively. The surface, coohng is more than a factor of 2 larger than the. TOA cooling
because of aerosol absorption. Note that the TOA DRF of -4.9 W/m2 agrees quite well
with the most recent satellite-based estimate of -5.1 ±1.1 W/m2 over non-desert land
based on coincident measurements of MISR AOD and GERES solar flux (I '.Ml adia el
al., 2008, 190558). For comparisons, an ensemble of five model simulations derives a
DRF (mean ± s) over land of -3.0 ± 0.6 W/m2 at the TOA and -7.6 ±0.9 W/m2 at the
surface, respectively. Seasonal variations of DRF over land, as derived from both
measurements and models, are larger than those over ocean.
- Jb
e
5
iu -4
K
o
<
D
I" -2

" MOD
-IS
41
1
a
I
-I
oBS
MOO
OCEAN
LAND
OCEAH
LANE
Figure 9-68. Summary of observation- and model-based (denoted as 0BS and MOD, respectively)
estimates of clear-sky, annual average DRF at the TOA and at the surface. The box and
vertical bar represent median and standard error, respectively (taken from Yu H, et,
2006,156173).
July 2009
9-141
DRAFT-DO NOT CITE OR QUOTE

-------
Table 9-8. Summary of seasonal and annual average clear-sky DRF (W/m2) at the TOA and the surface
(SFC) over global OCEAN derived with different methods and data.
DJF	MAM	JJA	SON	ANN
Products 	
TOA SFC TOA SFC TOA SFC TOA SFC TOA	SFC
MODIS	-5.9	-5.8	-6.0	-5.8	-5.9
MODIXA*
¦6.0
¦8.2
¦6.4
¦8.9
¦6.5
¦9.3
¦6.4
¦8.9
¦6.4
¦8.9
CERESA
¦5.3

¦6.1

¦5.4

¦5.1

¦5.5

CERESB
¦3.8

¦4.3

¦3.5

¦3.6

¦3.8

CERESC
¦5.3

¦5.4

¦5.2



¦5.3

MODISG
¦5.5
¦9.1
¦5.7
¦10.4
¦6.0
¦10.6
¦5.5
¦9.8
¦5.7
¦10.0
MISRG**
¦6.4
¦10.3
¦6.5
¦11.4
¦7.0
¦11.9
¦6.3
¦10.9
¦6.5
¦11.1
M0G0
¦4.9
¦7.8
¦5.1
¦9.3
¦5.4
¦9.4
¦5.0
¦8.7
¦5.1
¦8.8
M0MIG0
¦4.9
¦7.9
¦5.1
¦9.2
¦5.5
¦9.5
¦5.0
¦8.6
¦5.1
¦8.7
POLDER
¦5.7

¦5.7

¦5.8

¦5.6

¦5.7
¦5.2***
¦7.7***
SeaWiFS
¦6.0
¦6.6
¦5.2
¦5.8
¦4.9
¦5.6
¦5.3
¦5.7
¦5.4
¦5.9
Obs. Mean
¦5.4
¦8.3
¦5.6
¦9.2
¦5.6
¦9.4
¦5.4
¦8.8
¦5.5
¦8.7
Obs. Median
¦5.5
¦8.1
¦5.7
¦9.3
¦5.5
¦9.5
¦5.4
¦8.8
¦5.5
¦8.8
Obs. a
0.72
1.26
0.64
1.89
0.91
2.10
0.79
1.74
0.70
1.65
Obs. ~
0.23
0.56
0.20
0.85
0.29
0.94
0.26
0.78
0.21
0.67
GOCART
¦3.6
¦5.7
¦4.0
¦7.2
¦4.7
¦8.0
¦4.0
¦6.8
¦4.1
¦6.9
SPRINTARS
¦1.5
¦2.5
¦1.5
¦2.5
¦1.9
¦3.3
¦1.5
¦2.5
¦1.6
¦2.7
GISS
¦3.3
¦4.1
¦3.5
¦4.6
¦3.5
¦4.9
¦3.8
¦5.4
¦3.5
¦4.8
LMDZ-INCA
¦4.6
¦5.6
¦4.7
¦5.9
¦5.0
¦6.3
¦4.8
¦5.5
¦4.7
¦5.8
LMDZLOA
¦2.2
¦4.1
¦2.2
¦3.7
¦2.5
¦4.4
¦2.2
¦4.1
¦2.3
¦4.1
Mod. Mean
¦3.0
¦4.4
¦3.2
¦4.8
¦3.5
¦5.4
¦3.3
¦4.9
¦3.2
¦4.9
Mod Median
¦3.3
¦41.
¦3.5
¦4.6
¦3.5
¦4.9
¦3.8
¦5.4
¦3.5
¦4.8
Mod. cr
1.21
1.32
1.31
1.84
1.35
1.82
1.36
1.63
1.28
1.60
Mod. £
0.61
0.66
0.66
0.92
0.67
0.91
0.68
0.81
0.64
0.80
Mod.; Obs.
.60
.51
.61
.50
.64
.52
.70
.61
.64
.55
July 2009
9-142
DRAFT-DO NOT CITE OR QUOTE

-------
DJF	MAM	JJA	SON	ANN
* High bias may result from adding the DRF of individual components to derive the total DRF (Bellouin et al., 2005,155684).
** High bias most likely results from an overall overestimate of 20% in early post-launch MISR optical depth retrievals (Kahn RA Gaitley et al., 2005,190966).
*** Bellouin et al. (2003,189911) use AERONET retrieval of aerosol absorption as a constraint to the method in Boucher and Tanre (2000,190041), deriving aerosol direct radiative forcing both
at the T0A and the surface.
Sources of data: M0DIS (Remer & Kaufman, 2006), M0DIS_A (Bellouin et al., 2005,155684), POLDER (Boucher and Tanre, 2000,190041; Bellouin et al., 2003,189911), CERES_A and
CERES B (Loeb and Manalo-Smith, 2005,190433), CERES_C (Zhang et al., 2005,190930), M0DIS_G, MISR_G, M0_G0, M0_MI_G0 (Yu et al., 2004,190926; 2006,156173), SeaWiFS (Chou
et al., 2002,190008), GOCART (Chin et al., 2002,189996; Yu et al., 2004,190926), SPRINTARS (Takemura et al., 2002,190438), GISS (Koch and Hansen, 2005,190183; Koch et al., 2006,
190184), LMDZ-INCA (Kinne et al., 2006,155903; Schulz et al., 2006,190381), LMDZ-LOA (Reddy et al., 2005,190207; Reddy et al., 2005,190208). Mean, median, standard deviation (ct), and
standard error (£) are calculated for observations (Obs) and model simulations (Mod) separately. The last row is the ratio of model median to observational median (taken from Yu H, et, 2006,
156173).
Table 9-9. Summary of seasonal and annual average clear-sky DRF (W/m2) at the TOA and the surface
(SFC) over global LAND derived with different methods and data.
DJF	MAM	JJA	SON	ANN

TOA
SFC
TOA
SFC
TOA
SFC
TOA
SFC
TOA
SFC
MODISG
¦4.1
¦9.1
¦5.8
¦14.9
¦6.6
¦17.4
¦5.4
¦12.8
¦5.5
¦13.5
MISRG
¦3.9
¦8.7
¦5.1
¦13.0
¦5.8
¦14.6
¦4.6
¦101.7
¦4.9
¦11.8
MOGO
¦3.5
¦7.5
¦5.1
¦12.9
¦5.8
¦14.9
¦4.8
¦10.9
¦4.8
¦11.6
MOMIGO
¦3.4
¦7.4
¦4.7
¦11.8
¦5.3
¦13.5
¦4.3
¦9.7
¦4.4
¦10.6
Obs. Mean
¦3.7
¦8.2
¦5.2
¦13.2
¦5.9
¦15.1
¦4.8
¦11.0
¦4.9
¦11.9
Obs. Median
¦3.7
¦8.1
¦5.1
¦13.0
¦5.8
¦14.8
¦4.7
¦10.8
¦4.9
¦11.7
Obs. a
0.33
0.85
0.46
1.29
0.54
1.65
0.46
1.29
0.45
1.20
Obs. ~
0.17
0.49
0.26
0.74
0.31
0.85
0.27
0.75
0.26
0.70
GOCART
02.9
¦6.1
¦4.4
¦10.9
¦4.8
¦12.3
¦4.3
¦9.3
¦4.1
¦9.7
SPRINTARS
¦1.4
¦4.0
¦1.5
¦4.6
¦2.0
¦6.7
¦1.7
¦5.2
¦1.7
¦5.1
GISS
¦1.6
¦3.9
¦3.2
¦7.9
¦3.6
¦9.3
¦2.5
¦6.6
¦2.8
¦7.2
LMDZ-INCA
¦3.0
¦5.8
¦4.0
¦9.2
¦6.0
¦13.5
¦4.3
¦8.2
¦4.3
¦9.2
LMDZ-LOA
¦1.3
¦5.4
¦1.8
¦6.4
¦2.7
¦8.9
¦2.1
¦6.7
¦2.0
¦6.9
Mod. Mean
¦2.0
¦5.0
¦3.0
¦7.8
¦3.8
¦10.1
¦3.0
¦7.2
¦3.0
¦7.6
Mod Median
¦1.6
¦5.4
¦3.2
¦7.9
¦3.6
¦9.3
¦2.5
¦6.7
¦2.8
¦7.2
Mod. CT
0.84
1.03
1.29
2.44
1.61
2.74
1.24
1.58
1.19
1.86
Mod. £
0.42
0.51
0.65
1.22
0.80
1.37
0.62
0.79
0.59
0.93
Mod.; Obs.
0.43
0.67
0.63
0.61
0.62
0.62
0.53
0.62
0.58
0.62
Sources of data: M0DISG, MISRG, M0G0, MOMIGO (Yu et al., 2004,190926; 2006,156173), GOCART (Chin et al., 2002,189996; Yu et al., 2004,190926),
SPRINTARS (Takemura et al., 2002,190438), GISS (Koch and Hansen, 2005,190183; Koch et al., 2006,190184), LMDZ-INCA (Balkanski et al., 2007, 189979; Kinne et
al., 2006,155903; Schulz et al., 2006,190381), LMDZ-LOA (Reddy et al., 2005,190207; Reddy et al., 2005,190208). Mean, median, standard deviation (cr), and
standard error (£) are calculated for observations (Obs) and model simulations (Mod) separately. The last row is the ratio of model median to observational median. (Taken
from Yu H, et, 2006,156173).
July 2009
9-143
DRAFT-DO NOT CITE OR QUOTE

-------
The above analyses show that, on a global average, the measurement-based
estimates of DRF are 55-80% greater than the model-based estimates. The differences
are even larger on regional scales. Such measurement-model differences are a
combination of differences in aerosol amount (optical depth), single-scattering
properties, surface albedo, and radiative transfer schemes (Yu H, et, 2006, 156173). As
discussed earlier, MODIS retrieved optical depths tend to be overestimated by about
10-15% due to the contamination of thin cirrus and clouds in general (Kaufman et al.,
2005,	155891). Such overestimation of optical depth would result in a comparable
overestimate of the aerosol direct radiative forcing. Other satellite AOD data may
have similar contamination, which however has not yet been quantified. On the other
hand, the observations may be measuring enhanced AOD and DRF due to processes
not well represented in the models including humidification and enhancement of
aerosols in the vicinity of clouds (Koren et al., 2007, 190192).
From the perspective of model simulations, uncertainties associated with
parameterizations of various aerosol processes and meteorological fields, as
documented under the AEROCOM and Global Modeling Initiative (GMI) frameworks
(Kinne et al., 2006, 155903; Textor et al., 2006, 190456; Liu et al., 2007, 190427).
contribute to the large measurement-model and model-model discrepancies. Factors
determining the AOD should be major reasons for the DRF discrepancy and the
constraint of model AOD with well evaluated and bias reduced satellite AOD through
a data assimilation approach can reduce the DRF discrepancy significantly. Other
factors (such as model parameterization of surface reflectance, and model-satellite
differences in single-scattering albedo and asymmetry factor due to satellite sampling
bias toward cloud-free conditions) should also contribute, as evidenced by the
existence of a large discrepancy in the radiative efficiency (Yu H, et, 2006, 156173).
Significant effort will be needed in the future to conduct comprehensive assessments.
9.3.3.4. Satellite-Based Estimates of Anthropogenic Component of Aerosol Direct
Radiative Forcing
Satellite instruments do not measure the aerosol chemical composition needed to
discriminate anthropogenic from natural aerosol components. Because anthropogenic
aerosols are predominantly sub-micron, the fine-mode fraction derived from POLDER,
MODIS, or MISR might be used as a tool for deriving anthropogenic aerosol optical
depth. This could provide a feasible way to conduct measurement-based estimates of
anthropogenic component of aerosol direct radiative forcing (Kaufman et al., 2002,
190956). Such method derives anthropogenic AOD from satellite measurements by
empirically correcting contributions of natural sources (dust and maritime aerosol) to
the sub-micron AOD (Kaufman et al., 2005, 155891). The MODISbased estimate of
anthropogenic AOD is about 0.033 over oceans, consistent with model assessments of
0.030-0.036 even though the total AOD from MODIS is 25-40% higher than the
models (Kaufman et al., 2005, 155891). This accounts for 21 ± 7% of the MODIS"
observed total aerosol optical depth, compared with about 33% of anthropogenic
contributions estimated by the models. The anthropogenic fraction of AOD should be
much larger over land (i.e., 47 ± 9% from a composite of several models) (Bellouin et
al., 2005, 155684). comparable to the 40% estimated by Yu et al. (2006, 156173).
Similarly, the non-spherical fraction from MISR or POLDER can be used to separate
dust from spherical aerosol (Kahn et al., 2001, 190969; Kalashnikova and Kahn, 2006,
190962). providing another constraint for distinguishing anthropogenic from natural
aerosols.
There have been several estimates of anthropogenic component of DRF in recent
years. Table 9"10 lists such estimates of anthropogenic component of TOA DRF that
are from model simulations (Schulz et al., 2006, 190381) and constrained to some
degree by satellite observations (Kaufman et al., 2005, 155891; Bellouin et al., 2005,
155684; Bellouin et al., 2008, 189999; Chung et al., 2005, 155733; Christopher et al.,
2006,	155729; Matsui et al., 2006, 190495; Yu H, et, 2006, 156173; Quaas et al., 2008,
July 2009
9-144
DRAFT-DO NOT CITE OR QUOTE

-------
190916^ Zhao et al., 2008, 190936). The satellite-based clear-sky DRF by
anthropogenic aerosols is estimated to be -1.1 ± 0.37 W/m2 over ocean, about a factor
of 2 stronger than model simulated "0.6 W/m2. Similar DRF estimates are rare over
land, but a few studies do suggest that the anthropogenic DRF over land is much
more negative than that over ocean (Yu H, et, 2006, 156173; Bellouin et al., 2005,
155684; Bellouin et al., 2008, 189999). On global average, the measurement-based
estimate of anthropogenic DRF ranges from "0.9 to "1.9 W/m2, again stronger than the
model-based estimate of "0.8 W/m2. Similar to DRF estimates for total aerosols,
satellite-based estimates of anthropogenic component of DRF are rare over land.
On global average, anthropogenic aerosols are generally more absorptive than
natural aerosols. As such the anthropogenic component of DRF is much more negative
at the surface than at TOA. Several observation-constrained studies estimate that the
global average, clear-sky, anthropogenic component of DRF at the surface ranges from
-4.2 to "5.1 W/m2 (Yu et al., 2004, 190926; Bellouin et al., 2005, 155684; Chung et al.,
2005,	155733; Matsui et al., 2006, 190495), which is about a factor of 2 larger in
magnitude than the model estimates (e.g., Reddy et al., 2005, 190208).
Uncertainties in estimates of the anthropogenic component of aerosol DRF are
greater than for the total aerosol, particularly over land. An uncertainty analysis (Yu
H, et, 2006, 156173) partitions the uncertainty for the global average anthropogenic
DRF between land and ocean more or less evenly. Five parameters, namely fine-mode
fraction (if) and anthropogenic fraction of fine-mode fraction C/af) over both land and
ocean, and rover ocean, contribute nearly 80% of the overall uncertainty in the
anthropogenic DRF estimate, with individual shares ranging from 13-20% (Yu H, et,
2006,	156173). These uncertainties presumably represent a lower bound because the
sources of error are assumed to be independent. Uncertainties associated with several
parameters are also not well defined. Nevertheless, such uncertainty analysis is
useful for guiding future research and documenting advances in understanding.
Table 9-10. Estimates of anthropogenic components of aerosol optical depth (Tant) and clear-sky DRF at
the TOA from model simulations
Data Sources

Ocean

Land

Global
Estimated uncertainty or
Tant
DRF
(W/m2)
Tant
DRF
(W/m2)
Tant
DRF
(W/m2)
model diversity for DRF
Kaufman et al. (2005,155891)
0.033
¦1.4




30%
Bellouin et al. (2005,155684)
0.028
¦0.8
0.13

0.062
¦1.9
15%
Chunq et al. (2005,155733)





¦1.1

Yu et al. (2006,156173)
0.031
¦1.1
-.088
¦1.8
0.048
¦1.3
47% (ocean), 845 (land), and 62%
(global)
Christopher et al. (2006,155729)

¦1.4




65%
Matsui and Pielke (2006,190495)

¦1.6




30°S-30°N oceans
Quaas et al. (2008, 190916)

¦0.7

¦1.8

¦0.9
45%
Bellouin et al. (2008,189999)
0.021
¦0.6
0.107
¦3.3
0.043
¦1.3
Update to Bellouin et al. (2005)
with M0DIS Collection 5 data
Zhao et al. (2008,190936)

¦1.25




35%
Schulz et al. (2006,190381)
0.022
¦0.59
0.065
¦1.14
0.036
¦0.77
30-40%; same emissions
prescribed for all models
July 2009
9-145
DRAFT-DO NOT CITE OR QUOTE

-------
Ocean
Land
Global
Sources: Schulz et al., (2006,190381) approaches constrained by satellite observations, Kaufman et al. (2005,155891); Bellouin et al. (2005,155684) 2008; Chung et al. (2005,155733); Yu et
al. (2006,156173); Christopher et al. (2006,155729); Matsui and Pielke (2006,190495); Quaas et al. (2008,190916); Zhao et al. (2008,190936).
9.3.3.5. Aerosol-Cloud Interactions and Indirect Forcing
Satellite views of the Earth show a planet whose albedo is dominated by dark
oceans and vegetated surfaces, white clouds, and bright deserts. The bright white
clouds overlying darker oceans or vegetated surface demonstrate the significant effect
that clouds have on the Earth's radiative balance. Low clouds reflect incoming
sunlight back to space, acting to cool the planet, whereas high clouds can trap
outgoing terrestrial radiation and act to warm the planet. In the Arctic, low clouds
have also been shown to warm the surface (Garrett and Zhao, 2006, 190570). Changes
in cloud cover, in cloud vertical development, and cloud optical properties will have
strong radiative and therefore, climatic impacts. Furthermore, factors that change
cloud development will also change precipitation processes. These changes may alter
amounts, locations and intensities of local and regional rain and snowfall, creating
droughts, floods and severe weather.
Cloud droplets form on a subset of aerosol particles called cloud condensation
nuclei (CCN). In general, an increase in aerosol leads to an increase in CCN and an
increase in drop concentration. Thus, for the same amount of liquid water in a cloud,
more available CCN will result in a greater number but smaller size of droplets
(Twomey, 1977, 190533). A cloud with smaller but more numerous droplets will be
brighter and reflect more sunlight to space, thus exerting a cooling effect. This is the
first aerosol indirect radiative effect, or "albedo effect". The effectiveness of a particle
as a CCN depends on its size and composition so that the degree to which clouds
become brighter for a given aerosol perturbation, and therefore the extent of cooling,
depends on the aerosol size distribution and its size "dependent composition. In
addition, aerosol perturbations to cloud microphysics may involve feedbacks; for
example, smaller drops are less likely to collide and coalesce; this will inhibit growth,
suppressing precipitation, and possibly increasing cloud lifetime (Albrecht, 1989,
045783). In this case clouds may exert an even stronger cooling effect.
A distinctly different aerosol effect on clouds exists in thin Arctic clouds (LWP <25
g m-2) having low emissivity. Aerosol has been shown to increase the longwave
emissivity in these clouds, thereby warming the surface (Lubin and Vogelmann, 2006,
190466; Garrett and Zhao, 2006, 190570).
Some aerosol particles, particularly black carbon and dust, also act as ice nuclei
(IN) and in so doing, modify the microphysical properties of mixed-phase and ice-
clouds. An increase in IN will generate more ice crystals, which grow at the expense of
water droplets due to the difference in vapor pressure over ice and water surfaces. The
efficient growth of ice particles may increase the precipitation efficiency. In deep
convective, polluted clouds there is a delay in the onset of freezing because droplets
are smaller. These clouds may eventually precipitate, but only after higher altitudes
are reached that result in taller cloud tops, more lightning and greater chance of
severe weather (Rosenfeld and Lansky, 1998, 190230; Andreae et al., 2004, 155658).
The present state of knowledge of the nature and abundance of IN, and ice formation
in clouds is extremely poor. There is some observational evidence of aerosol influences
on ice processes, but a clear link between aerosol, IN concentrations, ice crystal
concentrations and growth to precipitation has not been established. This report
therefore only peripherally addresses ice processes. More information can be found in
a review by the WMO/IUGG International Aerosol-Precipitation Scientific Assessment
(Levin and Cotton, 2008, 190375).
July 2009
9-146
DRAFT-DO NOT CITE OR QUOTE

-------
In addition to their roles as CCN and IN, aerosols also absorb and scatter light,
and therefore they can change atmospheric conditions (temperature, stability, and
surface fluxes) that influence cloud development and properties (Ackerman et al.,
2000,	002987; Hansen et al., 1997, 043104). Thus, aerosols affect clouds through
changing cloud droplet size distributions, cloud particle phase, and by changing the
atmospheric environment of the cloud.
9.3.3.6. Remote Sensing of Aerosol-Cloud Interactions and Indirect Forcing
The AVHRR satellite instruments have observed relationships between columnar
aerosol loading, retrieved cloud microphysics, and cloud brightness over the Amazon
Basin that are consistent with the theories explained above (Feingold et al., 2001,
190544; Kaufman and Fraser, 1997, 190958; Kaufman and Nakajima, 1993, 190959).
but do not necessarily prove a causal relationship. Other studies have linked cloud
and aerosol microphysical parameters or cloud albedo and droplet size using satellite
data applied over the entire global oceans (Han et al., 1998, 190594; Nakajima et al.,
2001,	190552; Wetzel and Stowe, 1999, 190636). Using these correlations with
estimates of aerosol increase from the pre-industrial era, estimates of anthropogenic
aerosol indirect radiative forcing fall into the range of "0.7 to "1.7 W/m2 (Nakajima et
al., 2001, 190552).
Introduction of the more modern instruments (POLDER and MODIS) has allowed
more detailed observations of relationships between aerosol and cloud parameters.
Cloud cover can both decrease and increase with increasing aerosol loading (Kaufman
et al., 2005, 155891; Koren et al., 2004, 190187; Koren et al., 2005, 190188; Matheson
et al., 2005, 190494; Sekiguchi et al., 2003, 190385; Yu and et, 2007, 093173). The
same is true of LWP (Han et al., 2002, 049181; Matsui et al., 2006, 190498). Aerosol
absorption appears to be an important factor in determining how cloud cover will
respond to increased aerosol loading (Jiang and Feingold, 2006, 190976; Kaufman and
Koren, 2006, 190951; Koren et al., 2008, 190193). Different responses of cloud cover to
increased aerosol could also be correlated with atmospheric thermodynamic and
moisture structure (Yu and et, 2007, 093173). Observations in the MODIS data show
that aerosol loading correlates with enhanced convection and greater production of ice
anvils in the summer Atlantic Ocean (Koren et al., 2005, 190188). which conflicts with
previous results that used AVHRR and could not isolate convective systems from
shallow clouds (Sekiguchi et al., 2003, 190385).
In recent years, surface-based remote sensing has also been applied to address
aerosol effects on cloud microphysics. This method offers some interesting insights,
and is complementary to the global satellite view. Surface remote sensing can only be
applied at a limited number of locations, and therefore lacks the global satellite view.
However, these surface stations yield high temporal resolution data and because they
sample aerosol below, rather than adjacent to clouds they do not suffer from "cloud
contamination". With the appropriate instrumentation (lidar) they can measure the
local aerosol entering the clouds, rather than a column-integrated aerosol optical
depth. Under well-mixed conditions, surface in situ aerosol measurements can be
used. Surface re mote-sen sing studies are discussed in more detail below, although the
main science issues are common to satellite remote sensing.
Feingold et al. (2001, 190544) used data collected at the ARM Southern Great
Plains (SGP) site to allow simultaneous retrieval of aerosol and cloud properties. A
combination of a Doppler cloud radar and a microwave radiometer was used to
retrieve cloud drop effective radius profiles in non-precipitating (radar reflectivity Z
<-17 dBZ), ice-free clouds. Simultaneously, sub-cloud aerosol extinction profiles were
measured with a lidar to quantify the response of drop sizes to changes in aerosol
properties. Cloud data were binned according to liquid water path (LWP) as measured
with a microwave radiometer, consistent with Twomey's (1977, 190533) conceptual
view of the aerosol impact on cloud microphysics. With high temporal/spatial
resolution data (on the order of 20s or 100s of meters), realizations of aerosol-cloud
July 2009
9-147
DRAFT-DO NOT CITE OR QUOTE

-------
interactions at the large eddy scale were obtained, and quantified in terms of the
relative decrease in is in response to a relative increase in aerosol extinction (dn
iWcin extinction), as shown in Figure 9-69. Examining the dependence in this way
reduces reliance on absolute measures of cloud and aerosol parameters and minimizes
sensitivity to measurement error, provided errors are unbiased. This formulation
permitted these responses to be related to cloud microphysical theory Restricting the
examination to updrafts only (as determined from the radar Doppler signal) permitted
examination of the role of updraft in determining the response of |i to changes in
aerosol (via changes in drop number concent rat ion Ml). Analysis of data from 7 days
showed that turbulence intensifies the aerosol impact on cloud microphysics.
In addition to radar/microwave radiometer retrievals of aerosol and cloud
properties, measurements of cloud optical depth by surface based radiometers such as
the MFRSR (Michalsky et al., 2001, 190537) have been used in combination with
measurements of cloud I AVI' by microwave radiometer to measure an average value of
during dayhght. when the solar elevation angle is sufficiently high (Min and
Harrison, 1996,190538)., Using this retrieval, Kim et al. (2003, 155899) performed
analyses of the response to changes in aerosol at the same continental site, using a
surface measurement of the aerosol light scattering coefficient instead of using
extinction near cloud base as a proxy for QQN. Variance in LWP was shown to explain
most of the variance in cloud optical depth , exacerbating detection of an aerosol effect.
Although a decrease in w as observed with increasing scattering coefficient, the
relation was not strong, indicative of other influences on re and/or decoupling between
the surface and cloud layer. A similar study was conducted by Garrett et al. (2004,
190568) at a location in the Arctic.
1
i I i i mi i j i i i i ia
April 3 1998

_
: -
IE
¦
0,07
# LWP: 100—110 g
" 0.09
• IWP: 1 10-121 g m'1
0.09
- LWP 121-133 g m"2
	J—
.1—1-JLJ.llil 1 J. _I_L_LJ_I_L
0.01	0.10	1.00
Aerosol extinction (km1)
Source: Adapted from Feingold et al. (2003,190551).
Figure 9-69. Scatter plots showing mean cloud drop effective radius (re) vs. aerosol extinction
coefficient (unit: km i) for various liquid water path (LWP) bands on April 3,1998 at ARM
SGP site.
July 2009
9-148
DRAFT-DO NOT CITE OR QUOTE

-------
They suggested that summertime Arctic clouds are more sensitive to aerosol
perturbations than clouds at lower latitudes. The advantage of the MFRSR/microwave
radiometer combination is that it derives re from cloud optical depth and LWP and it
is not as sensitive to large drops as the radar is. A limitation is that it can be applied
only to clouds with extensive horizontal cover during daylight hours.
More recent data analyses by Feingold et al. (2003, 190551). Kim et al. (2008,
130785) and McComiskey et al. (2008, 190525) at a variety of locations, and modeling
work (Feingold, 2003, 190547) have investigated (i) the use of different proxies for
cloud condensation nuclei, such as the light scattering coefficient and aerosol index!
(ii) sensitivity of cloud microphysical/optical properties to controlling factors such as
aerosol size distribution, entrainment, LWP, and updraft velocity! (iii) the effect of
optical- as opposed to radar-retrievals of drop size! and (iv) spatial heterogeneity.
These studies have reinforced the importance of LWP and vertical velocity as
controlling parameters. They have also begun to reconcile the reasons for the large
discrepancies between various approaches, and platforms (satellite, aircraft in situ,
and surface-based remote sensing). These investigations are important because
satellite measurements that use a similar approach are being employed in GCMs to
represent the albedo indirect effect (Quaas and Boucher, 2005, 190573). In fact, the
weakest albedo indirect effect in IPCC (2007, 189430) derives from satellite
measurements that have very weak responses of n> to changes in aerosol. The
relationship between these aerosol-cloud microphysical responses and cloud radiative
forcing has been examined by McComiskey and Feingold (2008, 190 5 1 7). They
showed that for plane-parallel clouds, a typical uncertainty in the logarithmic
gradient of a /6-aerosol relationship of 0.05 results in a local forcing error of "3 to "10
W/m2, depending on the aerosol perturbation. This sensitivity reinforces the
importance of adequate quantification of aerosol effects on cloud microphysics to
assessment of the radiative forcing, i.e., the indirect effect. Quantification of these
effects from remote sensors is exacerbated by measurement errors. For example, LWP
is measured to an accuracy of 25 g m-2 at best, and since it is the thinnest clouds (i.e.,
low LWP) that are most susceptible (from a radiative forcing perspective) to changes
in aerosol, this measurement uncertainty represents a significant uncertainty in
whether the observed response is related to aerosol, or to differences in LWP. The
accuracy and spatial resolution of satellite "based LWP measurements is much poorer
and this represents a significant challenge. In some cases important measurements
are simply absent, e.g., updraft is not measured from satellite-based remote sensors.
Finally, cloud radar data from CloudSat, along with the A-train aerosol data, is
providing great opportunity for inferring aerosol effects on precipitation (e.g.,
Stephens and Haynes, 2007, 190413). The aerosol effect on precipitation is far more
complex than the albedo effect because the instantaneous view provided by satellites
makes it difficult to establish causal relationships.
9.3.3.7. In Situ Studies of Aerosol-Cloud Interactions
In situ observations of aerosol effects on cloud microphysics date back to the
1950s and 1960s (Gunn and Phillips, 1957, 190595! Squires, 1958, 045608! Warner,
1968, 157114! Warner and Twomey, 1967, 045616! Radke et al., 1989, 156034! Leaitch
et al., 1992, 045270! Brenguier et al., 2000, 189966! to name a few). These studies
showed that high concentrations of CCN from anthropogenic sources, such as
industrial pollution or the burning of sugarcane can increase cloud droplet number
concentration A&, thus increasing cloud microphysical stability and potentially
reducing precipitation efficiency. As in the case of remote sensing studies, the causal
link between aerosol perturbations and cloud microphysical responses (e.g., /& or AS)
is much better established than the relationship between aerosol and changes in cloud
fraction, LWC, and precipitation (see also Levin and Cotton, 2008, 190375).
In situ cloud measurements are usually regarded as "ground truth" for satellite
retrievals but in fact there is considerable uncertainty in measured parameters such
July 2009
9-149
DRAFT-DO NOT CITE OR QUOTE

-------
liquid water content (LWC), and size distribution, which forms the basis of other
calculations such as drop concentration, re and extinction. It is not uncommon to see
discrepancies in LWC on the order of 50% between different instruments, and cloud
drop size distributions are difficult to measure, particularly for droplets <10 jim where
Mie scattering oscillations generate ambiguities in drop size. Measurement
uncertainty in ib from in situ probes is assessed, for horizontally homogeneous clouds,
to be on the order of 15" 20%, compared to 10% for MODIS and 15-20% for other
spectral measurements (Feingold et al., 2003, 190551). As with remote measurements
it is prudent to consider relative (as opposed to absolute) changes in cloud
microphysics related to relative changes in aerosol. An added consideration is that in
situ measurements typically represent a very small sample of the atmosphere akin to
a thin pencil line through a large volume. For an aircraft flying at 100 m/s and
sampling at 1 Hz, the sample volume is on the order of 10 cm3. The larger spatial
sampling of remote sensing has the advantage of being more representative but it
removes small-scale (i.e., sub sampling volume) variability, and therefore, may
obscure important cloud processes.
Measurements at a wide variety of locations around the world have shown that
increases in aerosol concentration lead to increases in A&. However the rate of this
increase is highly variable and always sub-linear, as exemplified by the compilation of
data in Ramanathan et al. (2001, 042681). This is because, as discussed previously, A&
is a function of numerous parameters in addition to aerosol number concentration,
including size distribution, updraft velocity (Leaitch et al., 1996, 190354). and
composition. In stratocumulus clouds, characterized by relatively low vertical velocity
(and low supersaturation) only a small fraction of particles can be activated whereas
in vigorous cumulus clouds that have high updraft velocities, a much larger fraction of
aerosol particles is activated. Thus the ratio of AS to aerosol particle number
concentration is highly variable.
In recent years there has been a concerted effort to reconcile measured A&
concentrations with those calculated based on observed aerosol size and composition,
as well as updraft velocity. These so-called "closure experiments" have demonstrated
that on average, agreement in A& between these approaches is on the order of 20%
(Conant et al., 2004, 190010). This provides confidence in theoretical understanding of
droplet activation, however, measurement accuracy is not high enough to constrain
the aerosol composition effects that have magnitudes <20%.
One exception to the rule that more aerosol particles result in larger A&is the
case of giant CCN (sizes on the order of a few microns), which, in concentrations on
the order of 1 cm3 (i.e., ~ 1% of the total concentration) can lead to significant
suppression in cloud supersaturation and reductions in A& (O'Dowd et al., 1999,
090414). The measurement of these large particles is difficult and hence the
importance of this effect is hard to assess. These same giant CCN, at concentrations
as low as 1/liter, can significantly affect the initiation of precipitation in moderately
polluted clouds (Johnson, 1982, 190973) and in so doing alter cloud albedo (Feingold et
al., 1999, 190540).
The most direct link between the remote sensing of aerosol-cloud interactions
discussed in Section 9.3.3.6 and in situ observations is via observations of
relationships between drop concentration A& and CCN concentration. Theory shows
that if /fe-CCN relationships are calculated at constant LWP or LWC, their logarithmic
slope is "1/3 that of the A&-CCN logarithmic slope (i.e., dn/fe/dnCCN = "1/3 dnA&/
dnCCN). In general, A&-CCN slopes measured in situ tend to be stronger than
equivalent slopes obtained from remote sensing — particularly in the case of satellite
remote sensing (McComiskey and Feingold, 2008, 190517). There are a number of
reasons for this- (i) in situ measurements focus on smaller spatial scales and are more
likely to observe the droplet activation process as opposed to remote sensing that
incorporates larger spatial scales and includes other processes such as drop
coalescence that reduce A&, and therefore the slope of the A&- CCN relationship
(McComiskey et al., 2008, 190525). (ii) Satellite remote sensing studies typically do
July 2009
9-150
DRAFT-DO NOT CITE OR QUOTE

-------
not sort their data by LWP, and this has been shown to reduce the magnitude of the
/fe-CCN response (Feingold, 2003, 190547).
In conclusion, observational estimates of aerosol indirect radiative forcings are
still in their infancy. Effects on cloud microphysics that result in cloud brightening
have to be considered along with effects on cloud lifetime, cover, vertical development
and ice production. For in situ measurements, aerosol effects on cloud microphysics
are reasonably consistent (within ~ 20%) with theory but measurement uncertainties
in remote sensing of aerosol effects on clouds, as well as complexity associated with
three-dimensional radiative transfer, result in considerable uncertainty in radiative
forcing. The higher order indirect effects are poorly understood and even the sign of
the microphysical response and forcing may not always be the same. Aerosol type and
specifically the absorption properties of the aerosol may cause different cloud
responses. Early estimates of observationally based aerosol indirect forcing range
from "0.7 to -1.7 W/m2 (Nakajima et al., 2001, 190552) and "0.6 to -1.2 W/m2
(Sekiguchi et al., 2003, 190385). depending on the estimate for aerosol increase from
pre-industrial times and whether aerosol effects on cloud fraction are also included in
the estimate.
9.3.4. Outstanding Issues
Despite substantial progress, as summarized in Sections 9.3.2 and 9.3.3, most
measurement-based studies so far have concentrated on influences produced by the
sum of natural and anthropogenic aerosols on solar radiation under clear sky
conditions. Important issues remain-
¦	Because accurate measurements of aerosol absorption are lacking and land surface
reflection values are uncertain, DRF estimates over land and at the ocean surface are
less well constrained than the estimate of TOA DRF over ocean.
¦	Current estimates of the anthropogenic component of aerosol direct radiative forcing
have large uncertainties, especially over land.
¦	Because there are very few measurements of aerosol absorption vertical distribution,
mainly from aircraft during field campaigns, estimates of direct radiative forcing of
above-cloud aerosols and profiles of atmospheric radiative heating induced by aerosol
absorption are poorly constrained.
¦	There is a need to quantify aerosol impacts on thermal infrared radiation, especially for
dust.
¦	The diurnal cycle of aerosol direct radiative forcing cannot be adequately characterized
with currently available, sun-synchronous, polar orbiting satellite measurements.
¦	Measuring aerosol, cloud, and ambient meteorology contributions to indirect radiative
forcing remains a major challenge.
¦	Long-term aerosol trends and their relationship to observed surface solar radiation
changes are not well understood.
The current status and prospects for these areas are briefly discussed below.
Measuring Aerosol Absorption and Single-Scattering Albedo: Currently,
the accuracy of both in situ and remote sensing aerosol SSA measurements is
generally ± 0.03 at best, which implies that the inferred accuracy of clear sky aerosol
DRF would be larger than 1 W/m2 (see Chapter 1 of the CSSP SAP2.3). Recently
developed photoacoustic (Arnott et al., 1999, 020650) and cavity ring down extinction
cell (Strawa et al., 2002, 190421) techniques for measuring aerosol absorption produce
SSA with improved accuracy over previous methods. However, these methods are still
experimental, and must be deployed on aircraft. Aerosol absorption retrievals from
satellites using the UV"technique have large uncertainties associated with its
sensitivity to the height of the aerosol layer(s) (Torres et al., 2005, 190507). and it is
unclear how the UV results can be extended to visible wavelengths. Views in and out
of sunglint can be used to retrieve total aerosol extinction and scattering, respectively,
thus constraining aerosol absorption over oceans (Kaufman et al., 2002, 190955).
July 2009
9-151
DRAFT-DO NOT CITE OR QUOTE

-------
However, this technique requires retrievals of aerosol scattering properties, including
the real part of the refractive index, well beyond what has so far been demonstrated
from space. In summary, there is a need to pursue a better understanding of the
uncertainty in SSAfrom both in situ measurements and remote sensing retrievals
and, with this knowledge, to synthesize different data sets to yield a characterization
of aerosol absorption with well-defined uncertainty (Leahy et al., 2007, 190232).
Laboratory studies of aerosol absorption of specific known composition are also needed
to interpret in situ measurements and remote sensing retrievals and to provide
updated database of particle absorbing properties for models.
Estimating the Aerosol Direct Radiative Forcing over Land: Land surface
reflection is large, heterogeneous, and anisotropic, which complicates aerosol
retrievals and DRF determination from satellites. Currently, the aerosol retrievals
over land have relatively lower accuracy than those over ocean (Section 9.3.2.5) and
satellite data are rarely used alone for estimating DRF over land (Section 9.3.3).
Several issues need to be addressed, such as developing appropriate angular models
for aerosols over land (Patadia et al., 2008, 190558) and improving land surface
reflectance characterization. MODIS andMISR measure land surface reflection
wavelength dependence and angular distribution at high resolution (Moody et al.,
2005, 190548; Martonchik et al., 1998, 190484; Martonchik et al., 2002, 190490). This
offers a promising opportunity for inferring the aerosol direct radiative forcing over
land from satellite measurements of radiative fluxes (e.g., CERES) and from critical
reflectance techniques (Fraser and Kaufman, 1985, 190567; Kaufman, 1987, 190960).
The aerosol direct radiative forcing over land depends strongly on aerosol absorption
and improved measurements of aerosol absorption are required.
Distinguishing Anthropogenic from Natural Aerosols: Current estimates
of anthropogenic components of AOD and direct radiative forcing have larger
uncertainties than total aerosol optical depth and direct radiative forcing, particularly
over land (see Section 9.3.3.4), because of relatively large uncertainties in the
retrieved aerosol microphysical properties (see Section 9.3.2). Future measurements
should focus on improved retrievals of such aerosol properties as size distribution,
particle shape, and absorption, along with algorithm refinement for better aerosol
optical depth retrievals. Coordinated in situ measurements offer a promising avenue
for validating and refining satellite identification of anthropogenic aerosols (Anderson
et al., 2005, 189993; 2005, 189991). For satellite-based aerosol type characterization,
it is sometimes assumed that all biomass-burning aerosol is anthropogenic and all
dust aerosol is natural (Kaufman et al., 2005, 155891). The better determination of
anthropogenic aerosols requires a quantification of biomass burning ignited by
lightning (natural origin) and mineral dust due to human induced changes of land
cover/land use and climate (anthropogenic origin). Improved emissions inventories
and better integration of satellite observations with models seem likely to reduce the
uncertainties in aerosol source attribution.
Profiling the Vertical Distributions of Aerosols: Current aerosol profile
data are far from adequate for quantifying the aerosol radiative forcing and
atmospheric response to the forcing. The data have limited spatial and temporal
coverage, even for current spaceborne lidar measurements. Retrieving aerosol
extinction profile from lidar measured attenuated backscatter is subject to large
uncertainties resulting from aerosol type characterization. Current space-borne Lidar
measurements are also not sensitive to aerosol absorption. Because of lack of aerosol
vertical distribution observations, the estimates of DRF in cloudy conditions and dust
DRF in the thermal infrared remain highly uncertain (Schulz et al., 2006, 190381;
Sokolik et al., 2001, 190404; Lubin et al., 2002, 190463). It also remains challenging to
constrain the aerosol-induced atmospheric heating rate increment that is essential for
assessing atmospheric responses to the aerosol radiative forcing (e.g., Yu et al., 2002,
190923; Feingold et al., 2005, 190550; Lau et al., 2006, 190223). Progress in the
foreseeable future is likely to come from (l) better use of existing, global, space-based
backscatter lidar data to constrain model simulations, and (2) deployment of new
instruments, such as high-spectral-resolution lidar (HSRL), capable of retrieving both
July 2009
9-152
DRAFT-DO NOT CITE OR QUOTE

-------
extinction and backscatter from space. The HSRL lidar system will be deployed on the
EarthCARE satellite mission tentatively scheduled for 2013
(http7/asimov/esrin.esi.it/esaLP/ASESMYNW9SC Lpearthcare_l.html).
Characterizing the Diurnal Cycle of Aerosol Direct Radiative Forcing:
The diurnal variability of aerosol can be large, depending on location and aerosol type
(Smirnov et al., 2002, 190398). especially in wildfire situations, and in places where
boundary layer aerosols hydrate or otherwise change significantly during the day.
This cannot be captured by currently available, sun-synchronous, polar orbiting
satellites. Geostationary satellites provide adequate time resolution (TSI et al., 2002,
190031; Wang et al., 2003, 157106). but lack the information required to characterize
aerosol types. Aerosol type information from low earth orbit satellites can help
improve accuracy of geostationary satellite aerosol retrievals (Costa et al., 2004,
190006; 2004, 192022). For estimating the diurnal cycle of aerosol DRF, additional
efforts are needed to adequately characterize the anisotropy of surface reflection (Yu
et al., 2004, 190926) and daytime variation of clouds.
Studying Aerosol-Cloud Interactions and Indirect Radiative Forcing:
Remote sensing estimates of aerosol indirect forcing are still rare and uncertain.
Improvements are needed for both aerosol characterization and measurements of
cloud properties, precipitation, water vapor, and temperature profiles. Basic processes
still need to be understood on regional and global scales. Remote sensing observations
of aerosol-cloud interactions and aerosol indirect forcing are for the most part based
on simple correlations among variables, from which cause-and-effects cannot be
deduced. One difficulty in inferring aerosol effects on clouds from the observed
relationships is separating aerosol from meteorological effects, as aerosol loading itself
is often correlated with the meteorology. In addition, there are systematic errors and
biases in satellite aerosol retrievals for partly cloud-filled scenes. Stratifying aerosol
and cloud data by liquid water content, a key step in quantifying the albedo (or first)
indirect effect, is usually missing. Future work will need to combine satellite
observations with in situ validation and modeling interpretation. A methodology for
integrating observations (in situ and remote) and models at the range of relevant
temporal/spatial scales is crucial to improve understanding of aerosol indirect effects
and aerosol-cloud interactions.
Quantifying Long-Term Trends of Aerosols at Regional Scales: Because
secular changes are subtle and are superposed on seasonal and other natural
variability, this requires the construction of consistent, multi-decadal records of
climate-quality data. To be meaningful, aerosol trend analysis must be performed on a
regional basis. Long-term trends of aerosol optical depth have been studied using
measurements from surface remote sensing stations (e.g., Hoyt and Frohlich, 1983,
190621; Augustine et al., 2008, 189913; Luo et al., 2001, 190467) and historic satellite
sensors (Massie et al., 2004, 190492; Mishchenko et al., 2007, 190542; Mishchenko
and Geogdzhayev, 2007, 190545; Zhao et al., 2008, 190935). An emerging multiyear
climatology of high quality AOD data from modern satellite sensors (e.g., Remer and
et, 2008, 190224; Kahn RAGaitley et al., 2005, 190966) has been used to examine the
interannual variations of aerosol (e.g., Koren et al., 2007, 190189; Mishchenko and
Geogdzhayev, 2007, 190545) and contribute significantly to the study of aerosol
trends. Current observational capability needs to be continued to avoid any data gaps.
A synergy of aerosol products from historical, modern and future sensors is needed to
construct as long a record as possible. Such a data synergy can build upon
understanding and reconciliation of AOD differences among different sensors or
platforms (Jeong et al., 2005, 190977). This requires overlapping data records for
multiple sensors. A close examination of relevant issues associated with individual
sensors is urgently needed, including sensor calibration, algorithm assumptions, cloud
screening, data sampling and aggregation, among others.
Linking Aerosol Long-Term Trends with Changes of Surface Solar
Radiation: Anaiy sis of the long-term surface solar radiation record suggests
significant trends during past decades (e.g., Stanhill and Cohen, 2001, 042121; Wild et
July 2009
9-153
DRAFT-DO NOT CITE OR QUOTE

-------
al., 2005, 156156; Pinker et al., 2005, 190569; Alpert et al., 2005, 190047). Although a
significant and widespread decline in surface total solar radiation (the sum of direct
and diffuse irradiance) occurred up to 1990 (so-called solar dimming), a sustained
increase has been observed during the subsequent decade. Speculation suggests that
such trends result from decadal changes of aerosols and the interplay of aerosol direct
and indirect radiative forcing (Stanhill and Cohen, 2001, 042121; Wild et al., 2005,
156156; Streets et al., 2006, 190425; Norris and Wild, 2007, 190555; Ruckstuhl and et,
2008, 190356). However, reliable observations of aerosol trends are required test these
ideas. In addition to aerosol optical depth, changes in aerosol composition must also
be quantified, to account for changing industrial practices, environmental regulations,
and biomass burning emissions (Novakov et al., 2003, 048398; Streets et al., 2004,
190423; Streets and Aunan, 2005, 156106). Such compositional changes will affect the
aerosol SSAand size distribution, which in turn will affect the surface solar radiation
(e.g., Qian et al., 2007, 190572). However such data are currently rare and subject to
large uncertainties. Finally, a better understanding of aerosol-radiation-cloud
interactions and trends in cloudiness, cloud albedo, and surface albedo is badly needed
to attribute the observed radiation changes to aerosol changes with less ambiguity.
9.3.5. Concluding Remarks
Since the concept of aerosol-radiation-climate interactions was first proposed
around 1970, substantial progress has been made in determining the mechanisms and
magnitudes of these interactions, particularly in the last 10 yr. Such progress has
greatly benefited from significant improvements in aerosol measurements and
increasing sophistication of model simulations. As a result, knowledge of aerosol
properties and their interaction with solar radiation on regional and global scales is
much improved. Such progress plays a unique role in the definitive assessment of the
global anthropogenic radiative forcing, as "virtually certainly positive" in IPCC AR4
(Haywood and Schulz, 2007, 190600).
In Situ Measurements of Aerosols: New in situ instruments such as aerosol
mass spectrometers, photoacoustic techniques, and cavity ring down cells provide high
accuracy and fast time resolution measurements of aerosol chemical and optical
properties. Numerous focused field campaigns and the emerging ground-based aerosol
networks are improving regional aerosol chemical, microphysical, and radiative
property characterization. Aerosol closure studies of different measurements indicate
that measurements of submicrometer, spherical sulfate and carbonaceous particles
have a much better accuracy than that for dust-dominated aerosol. The accumulated
comprehensive data sets of regional aerosol properties provide a rigorous "test bed"
and strong constraint for satellite retrievals and model simulations of aerosols and
their direct radiative forcing.
Remote Sensing Measurements of Aerosols: Surface networks, covering
various aerosol regimes around the globe, have been measuring aerosol optical depth
with an accuracy of 0.01-0.02, which is adequate for achieving the accuracy of 1 W/m2
for cloud-free TOADRF. On the other hand, aerosol microphysical properties retrieved
from these networks, especially SSA, have relatively large uncertainties and are only
available in very limited conditions. Current satellite sensors can measure AOD with
an accuracy of about 0.05 or 15-20% in most cases. The implementation of multi-
wavelength, multi-angle, and polarization measuring capabilities has also made it
possible to measure particle properties (size, shape, and absorption) that are essential
for characterizing aerosol type and estimating anthropogenic component of aerosols.
However, these microphysical measurements are more uncertain than AOD
measurements.
Observational Estimates of Clear-Sky Aerosol Direct Radiative
Forcing: Closure studies based on focused field experiments reveal DRF
uncertainties of about 25% for sulfate/carbonaceous aerosol and 60% for dust at
July 2009
9-154
DRAFT-DO NOT CITE OR QUOTE

-------
regional scales. The high-accuracy of MODIS, MISR and POLDER aerosol products
and broadband flux measurements from CERES make it feasible to obtain
observational constraints for aerosol TOA DRF at a global scale, with relaxed
requirements for measuring particle microphysical properties. Major conclusions from
the assessment are-
¦	A number of satellite-based approaches consistently estimate the clear-sky diurnally
averaged TOA DRF (on solar radiation) to be about "5.5 ± 0.2 W/m2 (mean ± standard
error from various methods) over global ocean. At the ocean surface, the diurnally
averaged DRF is estimated to be "8.7 ± 0.7 W/m2. These values are calculated for the
difference between today's measured total aerosol (natural plus anthropogenic) and the
absence of all aerosol.
¦	Overall, in comparison to that over ocean, the DRF estimates over land are more poorly
constrained by observations and have larger uncertainties. A few satellite retrieval and
satellite-model integration yield the overland clear-sky diurnally averaged DRF of "4.9
± 0.7 W/m2 and -11.8 ± 1.9 W/m2 at the TOA and surface, respectively. These values
over land are calculated for the difference between total aerosol and the complete
absence of all aerosol.
¦	Use of satellite measurements of aerosol microphysical properties yields that on a
global ocean average, about 20% of AOD is contributed by human activities and the
clear-sky TOA DRF by anthropogenic aerosols is -1.1 ± 0.4 W/m2. Similar DRF
estimates are rare over land, but a few measurement-model integrated studies do
suggest much more negative DRF over land than over ocean.
¦	These satellite "based DRF estimates are much greater than the model-based estimates,
with differences much larger at regional scales than at a global scale.
Measurements of Aerosol-Cloud Interactions and Indirect Radiative
Forcing: In situ measurement of cloud properties and aerosol effects on cloud
microphysics suggest that theoretical understanding of the activation process for
water cloud is reasonably well-understood. Remote sensing of aerosol effects on
droplet size associated with the albedo effect tends to underestimate the magnitude of
the response compared to in situ measurements. Recent efforts trace this to a
combination of lack of stratification of data by cloud water, the relatively large spatial
scale over which measurements are averaged (which includes variability in cloud
fields, and processes that obscure the aerosol-cloud processes), as well as
measurement uncertainties (particularly in broken cloud fields). It remains a major
challenge to infer aerosol number concentrations from satellite measurements. The
present state of knowledge of the nature and abundance of IN, and ice formation in
clouds is extremely poor.
Despite the substantial progress in recent decades, several important issues
remain, such as measurements of aerosol size distribution, particle shape, absorption,
and vertical profiles, and the detection of aerosol long-term trend and establishment
of its connection with the observed trends of solar radiation reaching the surface, as
discussed in Section 9.3.4. Furthering the understanding of aerosol impacts on climate
requires a coordinated research strategy to improve the measurement accuracy and
use the measurements to validate and effectively constrain model simulations.
Concepts of future research in measurements are discussed in Chapter 4 "Way
Forward" (of the CCSP SAP2.3).
9.3.6. Modeling the Effect of Aerosols on Climate
9.3.6.1. Introduction
The IPCC Fourth Assessment Report (AR4) (Intergovernmental Panel on
Climate, 2007, 190988) concludes that man's influence on the warming climate is in
the category of "very likely". This conclusion is based on, among other things, the
July 2009	9-155	DRAFT - DO NOT CITE OR QUOTE

-------
ability of models to simulate the global and, to some extent, regional variations of
temperature over the past 50-100 yr. When anthropogenic effects are included, the
simulations can reproduce the observed warming (primarily for the past 50 yr); when
they are not, the models do not get very much warming at all. In fact, all of the models
runs for the IPCC AIM assessment (more than 20) produce this distinctive result,
driven by the greenhouse gas increases that have been observed to occur.
These results were produced in models whose average global warming associated
with a doubled C02 forcing of 4 W/m2 was about 3°C. This translates into a climate
sensitivity (surface temperature change per forcing) of about 0.75°C/(W/m2). The
determination of climate sensitivity is crucial to projecting the future impact of
increased greenhouse gases, and the credibility of this projected value relies on the
ability of these models to simulate the observed temperature changes over the past
century. However, in producing the observed temperature trend in the past, the
models made use of very uncertain aerosol forcing. The greenhouse gas change by
itself produces warming in models that exceeds that observed by some 40% on average
(IPCC et al., 2007, 189430! IPCC (Intergovernmental Panel on Climate, 2007,
190988). Cooling associated with aerosols reduces this warming to the observed level.
Different climate models use differing aerosol forcings, both direct (aerosol scattering
and absorption of short and longwave radiation) and indirect (aerosol effect on cloud
cover reflectivity and lifetime), whose magnitudes vary markedly from one model to
the next. Kiehl (2007, 190949) using nine of the IPCC (2007, 189430; 2007, 190988)
AE.4 climate models found that they had a factor of three forcing differences in the
aerosol contribution for the 20th century. The differing aerosol forcing is the prime
reason why models whose climate sensitivity varies by almost a factor of three can
produce the observed trend. It was thus concluded that the uncertainty in IPCC
(2007, 190988) anthropogenic climate simulations for the past century should really
be much greater than stated (Schwartz et al., 2007, 188938; Kerr, 2007, 190950).
since, in general, models with low/high sensitivity to greenhouse warming used
weaker/stronger aerosol cooling to obtain the same temperature response (Kiehl,
2007, 190949). Had the situation been reversed and the low/high sensitivity models
used strong/ weak aerosol forcing, there would have been a greater divergence in
model simulations of the past century.
July 2009
9-156
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 9-70. Sampling the Arcic Haze. Pollution and smoke aerosols can travel long distances, from
mid-latitudes to the Arctic, causing "Arctic Haze." Photo taken from the NASA DC-8
aircraft during the ARCTAS field experiment over Alaska in April 2008. Credit: Mian
Chin, NASA.
Therefore, the fact that a model has accurately reproduced the global
temperature change in the past does not imply that its future forecast is accurate.
This state of affairs will remain until a firmer estimate of radiative forcing (RF) by
aerosols, in addition to that by greenhouse gases, is available.
Two different approaches are used to assess the aerosol effect on climate,
"Forward modeling" studies incorporate different aerosol types and attempt to
explicitly Calculate the. aerosol RE, From this approach, IPCC (2007, 190988)
concluded that the best estimate of the global aerosol direct RF (compared with
preindustrial times) is -0,5 (-0.9 to -0.1) W/m2. The RF due to the cloud albedo or
brightness effect (also referred to as first indirect or Twomey effect) is estimated to be
-0.7 (-1.8 to -0,3) W/m2. Mo estimate was specified for the effect associated with cloud
lifetime. The total negative RF due to aerosols according to IPOG (2007, 190988)
estimates is then -1.3 (-2.2 to -0,5) W/m2. In comparison, the positive radiative: forcing
(RF) from greenhouse gases (including tropospheric ozone) is estimated to be +2.9 ±
0.3 W/m2; hence tropospheric aerosols reduce the influence: from greenhouse gases by
about 45% (15-85%). This approach however inherits large uncertainties in aerosol
amount, composition, and physical and optical properties in modeling of atmospheric
aerosols. The consequences of these uncertainties are discussed in the next section.
July 2009
9-157
DRAFT-DO NOT CITE OR QUOTE

-------
The other method of calculating aerosol forcing is called the "inverse approach" —
it is assumed that the observed climate change is primarily the result of the known
climate forcing contributions. If one further assumes a particular climate sensitivity
(or a range of sensitivities), one can determine what the total forcing had to be to
produce the observed temperature change. The aerosol forcing is then deduced as a
residual after subtraction of the greenhouse gas forcing along with other known
forcings from the total value. Studies of this nature come up with aerosol forcing
ranges of "0.6 to -1.7 W/m2 (Knutti et al., 2002, 190178; Knutti et al., 2003, 190180);
IPCC AR.4 Chap.9); -0.4 to -1.6 W/m2 (Gregory et al., 2002, 190593); and -0.4 to -1.4
W/m2 (Stott, 2006, 190419). This approach however provides a bracket of the possible
range of aerosol forcing without the assessment of current knowledge of the
complexity of atmospheric aerosols.
This chapter of the CCSP SAP2.3 reviews the current state of aerosol RF in the
global models and assesses the uncertainties in these calculations. First
representation of aerosols in the forward global chemistry and transport models and
the diversity of the model simulated aerosol fields are discussed; then calculation of
the aerosol direct and indirect effects in the climate models is reviewed; finally the
impacts of aerosols on climate model simulations and their implications are assessed.
9.3.6.2. Modeling of Atmospheric Aerosols
The global aerosol modeling capability has developed rapidly in the past decade.
In the late 1990s, there were only a few global models that were able to simulate one
or two aerosol components, but now there are a few dozen global models that simulate
a comprehensive suite of aerosols in the atmosphere. As introduced in Chapter 1 (of
the CCSP SAP2.3), aerosols consist of a variety of species including dust, sea salt,
sulfate, nitrate, and carbonaceous aerosols (black and organic carbon) produced from
natural and man-made sources with a wide range of physical and optical properties.
Because of the complexity of the processes and composition, and highly
inhomogeneous distribution of aerosols, accurately modeling atmospheric aerosols and
their effects remains a challenge. Models have to take into account not only the
aerosol and precursor emissions, but also the chemical transformation, transport, and
removal processes (e.g., dry and wet depositions) to simulate the aerosol mass
concentrations. Furthermore, aerosol particle size can grow in the atmosphere
because the ambient water vapor can condense on the aerosol particles. This
"swelling" process, called hygroscopic growth, is most commonly parameterized in the
models as a function of relative humidity.
Estimates of Emissions. Aerosols have various sources from both natural and
anthropogenic processes. Natural emissions include wind-blown mineral dust, aerosol
and precursor gases from volcanic eruptions, natural wild fires, vegetation, and
oceans. Anthropogenic sources include emissions from fossil fuel and biofuel
combustion, industrial processes, agriculture practices, and human "induced biomass
burning.
Following earlier attempts to quantify manmade primary emissions of aerosols
(Turco et al., 1983, 190529; Penner et al., 1993, 045457) systematic work was
undertaken in the late 1990s to calculate emissions of black carbon (BC) and organic
carbon (OC), using fuel-use data and measured emission factors (Liousse et al., 1996,
078158; Cooke and Wilson, , 190046; Cooke et al., 1999, 156365). The work was
extended in greater detail and with improved attention to source-specific emission
factors in Bond et al. (2004, 056389). which provides global inventories of BC and OC
for the year 1996, with regional and source-category discrimination that includes
contributions from industrial, transportation, residential solid-fuel combustion,
vegetation and open biomass burning (forest fires, agricultural waste burning, etc.),
and diesel vehicles.
July 2009
9-158
DRAFT-DO NOT CITE OR QUOTE

-------
Emissions from natural sources—which include wind-blown mineral dust,
wildfires, sea salt, and volcanic eruptions—are less well quantified, mainly because of
the difficulties of measuring emission rates in the field and the unpredictable nature
of the events. Often, emissions must be inferred from ambient observations at some
distance from the actual source. As an example, it was concluded (Lewis and
Schwartz, 2004, 192023) that available information on size-dependent sea salt
production rates could only provide order-of-magnitude estimates. The natural
emissions in general can vary dramatically over space and time.
Aerosols can be produced from trace gases in the atmospheric via chemical
reactions, and those aerosols are called secondary aerosols, as distinct from primary
aerosols that are directly emitted to the atmosphere as aerosol particles. For example,
most sulfate and nitrate aerosols are secondary aerosols that are formed from their
precursor gases, sulfur dioxide (SO2) and nitrogen oxides (NO and N02, collectively
called NOx), respectively. Those sources have been studied for many years and are
relatively well known. By contrast, the sources of secondary organic aerosols (SOA)
are poorly understood, including emissions of their precursor gases (called volatile
organic compounds, VOC) from both natural and anthropogenic sources and the
atmospheric production processes.
Globally, sea salt and mineral dust dominate the total aerosol mass emissions
because of the large source areas and/or large particle sizes. However, sea salt and
dust also have shorter atmospheric lifetimes because of their large particle size, and
are radiatively less active than aerosols with small particle size, such as sulfate,
nitrate, BC, and particulate organic matter (POM, which includes both carbon and
non-carbon mass in the organic aerosol), most of which are anthropogenic in origin.
Because the anthropogenic aerosol RF is usually evaluated (e.g., by the IPCC) as
the anthropogenic perturbation since the pre-industrial period, it is necessary to
estimate the historical emission trends, especially the emissions in the pre-industrial
era. Compared to estimates of present-day emissions, estimates of historical emission
have much larger uncertainties. Information for past years on the source types and
strengths and even locations are difficult to obtain, so historical inventories from
preindustrial times to the present have to be based on limited knowledge and data.
Several studies on historical emission inventories of BC and OC (e.g., Novakov et al.,
2003, 048398; Ito and Penne, 2005, 190626; Bond et al., 2007, 190050; Fernandes et
al., 2007, 190554; Junker and Liousse, 2008, 190971). SO2 (Stern, 2005, 190416). and
various species (Van Aardenne et al., 2001, 055564; Dentener et al., 2006, 088434) are
available in the literature; there are some similarities and some differences among
them, but the emission estimates for early times do not have the rigor of the studies
for present-day emissions. One major conclusion from all these studies is that the
growth of primary aerosol emissions in the 20th century was not nearly as rapid as
the growth in CO2 emissions. This is because in the late 19th and early 20th
centuries, particle emissions such as BC and POM were relatively high due to the
heavy use of biofuels and the lack of particulate controls on coal-burning facilities;
however, as economic development continued, traditional biofuel use remained fairly
constant and particulate emissions from coal burning were reduced by the application
of technological controls (Bond et al., 2007, 190050). Thus, particle emissions in the
20th century did not grow as fast as CO2 emissions, as the latter are roughly
proportional to total fuel use—oil and gas included. Another challenge is estimating
historical biomass burning emissions. A recent study suggested about a 40% increase
in carbon emissions from biomass burning from the beginning to the end of last
century (Mouillot et al., 2006, 190549). but it is difficult to verify.
July 2009
9-159
DRAFT-DO NOT CITE OR QUOTE

-------
Table 9-11. Anthropogenic emissions of aerosols and precursors for 2000 and 1750.
Source
Species*
Emission*
Emission 1750
2000 (Tg/yr)
(Tg/yr)
Biomass burning
BC
3.1
1.03

POM
34.7
12.8

S
4.1
1.46
Biofuel
BC
9.1
0.39

POM
9.6
1.56

S

0.12
Fossil fuel
BC
3.0


POM
3.2


S
98.9

# Data source for 2000 emission: biomass burning - Global Fire Emission Dataset (GFED); biofuel BC and POM - Speciated Pollutant Emission Wizard (SPEW); biofuel sulfur - International Institute
for Applied System Analysis (NASA); fossil fuel BC and POM - SPEW; fossil fuel sulfur - Emission Database for Global Atmospheric Research (EDGAR) and NASA. Fossil fuel emission of sulfur (S) is
the sum of emission from industry, power plants, and transportation listed in Dentener et al. (2006, 088434).
" S = sulfur, including SO2 and particulate S042". Most emitted as SO2, and 2.5% emitted as S042".
Source: Adapted from Dentener et al. (2006, 088434).
As an example, Table 9"11 shows estimated anthropogenic emissions of sulfur, BC
and POM in the present day (year 2000) and pre "industrial time (1750) compiled by
Dentener et al. (2006, 088434) These estimates have been used in the Aerosol
Comparisons between Observations and Models (AeroCom) project (Experiment B,
which uses the year 2000 emission! and Experiment PRE, which uses pre "industrial
emissions), for simulating atmospheric aerosols and anthropogenic aerosol RF. The
AeroCom results are discussed below and in Section 9.3.6.3.
Aerosol Mass Loading and Optical Depth. In the global models, aerosols are
usually simulated in the successive steps of sources (emission and chemical
formation), transport (from source location to other area), and removal processes (dry
deposition, in which particles fall onto the surface, and wet deposition by rain) that
control the aerosol lifetime. Collectively, emission, transport, and removal determine
the amount (mass) of aerosols in the atmosphere. Aerosol optical depth (AOD), which
is a measure of solar or thermal radiation being attenuated by aerosol particles via
scattering or absorption, can be related to the atmospheric aerosol mass loading as
follows-
AOD = MEE • M
Equation 9-3
where M is the aerosol mass loading per unit area (g m-2), MEE is the mass
extinction efficiency or specific extinction in unit of m2/g, which is
MEE = 3Qest «J
47rpreff
Equation 9-4
where Qex£ is the extinction coefficient (a function of particle size distribution and
refractive index), reff is the aerosol particle effective radius, p is the aerosol particle
density, and f is the ratio of ambient aerosol mass (wet) to dry aerosol mass M. Here,
M is the result from model-simulated atmospheric processes and MEE embodies the
aerosol physical (including microphysical) and optical properties. Since Qex^varies
July 2009
9-160
DRAFT-DO NOT CITE OR QUOTE

-------
with radiation wavelength, so do MEE and AOD. AOD is the quantity that is most
commonly obtained from remote sensing measurements and is frequently used for
model evaluation (see Chapter 2 of the CCSP SAP2.3). AOD is also a key parameter
determining aerosol radiative effects.
Here the results from the recent multiple global" model studies by the AeroCom
project are summarized, as they represent the current assessment of model-simulated
atmospheric aerosol loading, optical properties, and RF for the present-day. AeroCom
aims to document differences in global aerosol models and com pare the model output
to observations. Sixteen global models participated in the AeroCom Experiment A
(AeroCom-A), for which every model used their own configuration, including their own
choice of estimating emissions (Kinne et al., 2006, 155903; Textor et al., 2006,
190456). Five major aerosol types' sulfate, BC, POM, dust, and sea salt, were included
in the experiments, although some models had additional aerosol species. Of those
major aerosol types, dust and sea-salt are predominantly natural in origin, whereas
sulfate, BC, and POM have major anthropogenic sources.
Table 9"12 summarizes the model results from the AeroCom-A for several key
parameters' Sources (emission and chemical transformation), mass loading, lifetime,
removal rates, and MEE and AOD at a commonly used, mid-visible, wavelength of 550
nanometer (nm). These are the globally averaged values for the year 2000. Major
features and conclusions are'
¦	Globally, aerosol source (in mass) is dominated by sea salt, followed by dust, sulfate,
POM, and BC. Over the non-desert land area, human activity is the major source of
sulfate, black carbon, and organic aerosols.
¦	Aerosols are removed from the atmosphere by wet and dry deposition. Although sea
salt dominates the emissions, it is quickly removed from the atmosphere because of its
large particle size and near-surface distributions, thus having the shortest lifetime.
The median lifetime of sea salt from the AeroCom-A models is less than half a day,
whereas dust and sulfate have similar lifetimes of 4 days and BC and POM 6"7 days.
¦	Globally, small-particle-sized sulfate, BC, and POM make up a little over 10% of total
aerosol mass in the atmosphere. However, they are mainly from anthropogenic
activity, so the highest concentrations are in the most populated regions, where their
effects on climate and air quality are major concerns.
¦	Sulfate and BC have their highest MEE at mid-visible wavelengths, whereas dust is
lowest among the aerosol types modeled. That means for the same amount of aerosol
mass, sulfate and BC are more effective at attenuating (scattering or absorbing) solar
radiation than dust. This is why the sulfate AOD is about the same as dust AOD even
though the atmospheric amount of sulfate mass is 10 times less than that of the dust.
¦	There are large differences, or diversities, among the models for all the parameters
listed in Table 9"12. The largest model diversity, shown as the % standard deviation
from the all-model-mean and the range (minimum and maximum values) in Table 9"12,
is in sea salt emission and removal; this is mainly associated with the differences in
particle size range and source parameterizations in each model. The diversity of sea
salt atmospheric loading however is much smaller than that of sources or sinks,
because the largest particles have the shortest lifetimes even though they comprise the
largest fraction of emitted and deposited mass.
Table 9-12. Summary of statistics of AeroCom Experiment A results from 16 global models.
Quantity
Mean
Median
Range
Stddev/mean*
SOURCES (TG/YR!
SO42-
179
186
98-232
22%
BC
11.9
11.3
7.8-19.4
23%
Organic matter
96.6
96.0
53-138
26%
Dust
1840
1640
672-4040
49%
July 2009

9-161

DRAFT-DO NOT CITE OR QUOTE

-------
Quantity
Mean
Median
Range
Stddev/mean*
Sea salt
16600
6280
2180-121000
199%
REMOVAL RATE (DAYJ
S042-
0.25
0.24
0.19-0.39
18%
BC
0.15
0.15
0.066-0.19
21%
Organic matter
0.16
0.16
0.09-0.23
24%
Dust
0.31
0.25
0.14-0.79
62%
Sea salt
5.07
2.50
0.95-35.0
188%
LIFETIME (DAY)
S042-
4.12
4.13
2.6-5.4
18%
BC
7.12
6.54
5.3-15
33%
Organic matter
6.54
6.16
4.3-11
27%
Dust
4.14
4.04
1.3-7.0
43%
Sea salt
0.48
0.41
0.03-1.1
58%
MASS LOADING (TG)
S042-
1.99
1.98
0.92-2.70
25%
BC
0.24
0.21
0.046-0.51
42%
Organic matter
1.70
1.76
0.46-2.56
27%
Dust
19.2
20.5
4.5-29.5
40%
Sea salt
7.52
6.37
2.5-13.2
54%
MEE AT550 NM (HfG'l
S042-
11.3
9.5
4.2-28.3
56%
BC
9.4
9.2
5.3-18.9
36%
Organic matter
5.7
5.7
3.7-9.1
26%
Dust
0.99
0.95
0.46-2.05
45%
Sea salt
3.0
3.1
0.97-7.5
55%
AOD AT550 NM
S042-
0.035
0.034
0.015-0.051
33%
BC
0.004
0.004
0.002-0.009
46%
Organic matter
0.018
0.019
0.006-0.030
36%
Dust
0.032
0.033
0.012-0.054
44%
Sea salt
0.033
0.030
0.02-0.067
42%
TOTAL AOT AT550 NM
0.124
0.127
0.0560.151
18%
* Stddev/mean was used as the term "diversity" in Textor et al. (2006,190456).
Source: Textor et al. (2006,190456) and Kinne et al. (2006,155903), and AeroCom website http:llnansen.ipsl.iussieu.fr/AEROCOMIdata.html
¦ Among the key parameters compared in Table 9"12, the models agree best for
simulated total AOD — the % of standard deviation from the model mean is 18%, with
the extreme values just a factor of 2 apart. The median value of the multi-model
simulated global annual mean total AOD, 0.127, is also in agreement with the global
mean values from recent satellite measurements. However, despite the general
agreement in total AOD, there are significant diversities at the individual component
level for aerosol optical thickness, mass loading, and mass extinction efficiency. This
indicates that uncertainties in assessing aerosol climate forcing are still large, and they
depend not only on total AOD but also on aerosol absorption and scattering direction
(called asymmetry factor), both of which are determined by aerosol physical and optical
July 2009
9-162
DRAFT-DO NOT CITE OR QUOTE

-------
properties. In addition, even with large differences in mass loading and MEE among
different models, these terms could compensate for each other (Equation 9-1) to
produce similar AOD. This is illustrated in Figure 9-71. For example, model LO and LS
have quite different mass loading (44 and 74 mg m-2, respectively), especially for dust
and sea salt amount, but they produce nearly identical total AOD (0.127 and 0.128,
respectively).
¦ Because of the large spatial and temporal variations of aerosol distributions, regional
and seasonal diversities are even larger than the diversity for global annual means.
To further isolate the impact of the differences in emissions on the diversity of
simulated aerosol mass loading, identical emissions for aerosols and their precursor
were used in the AeroCom Experiment B exercise in which 12 of the 16 AeroConrA
models participated (Textor and et, 2007, 190458). The comparison of the results and
diversity between AeroConrA and —B for the same models showed that using
harmonized emissions does not significantly reduce model diversity for the simulated
global mass and AOD fields, indicating that the differences in atmospheric processes,
such as transport, removal, chemistry, and aerosol microphysics, play more important
roles than emission in creating diversity among the models. This outcome is
somewhat different from another recent study, in which the differences in calculated
clear-sky aerosol RF between two models (a regional model STEM and a global model
MOZART) were attributed mostly to the differences in emissions (Bates et al., 2006,
189912). although the conclusion was based on only two model simulations for a few
focused regions. It is highly recommended from the outcome of AeroCom "A and "B
that, although more detailed evaluation for each individual process is needed, multi-
model ensemble results, e.g., median values of multi-model output variables, should
be used to estimate aerosol RF, due to their greater robustness, relative to individual
models, when compared to observations (Textor et al., 2006, 190456) Textor and et,
2007, 190458; Schulz et al., 2006, 190381).
9.3.6.3. Calculating Aerosol Direct Radiative Forcing
The three parameters that define the aerosol direct RF are the AOD, the single
scattering albedo (SSA), and the asymmetry factor (g), all of which are wavelength
dependent. AOD is indicative of how much aerosol exists in the column, SSA is the
fraction of radiation being scattered versus the total attenuation (scattered and
absorbed), and the g relates to the direction of scattering that is related to the size of
the particles (see Chapter 1 of the CCSP SAP2.3). An indication of the particle size is
provided by another parameter, the Angstrom exponent (A), which is a measure of
differences of AOD at different wavelengths. For typical tropospheric aerosols, A tends
to be inversely dependent on particle size! larger values of A are generally associated
with smaller aerosols particles. These parameters are further related; for example, for
a given composition, the ability of a particle to scatter radiation decreases more
rapidly with decreasing size than does its ability to absorb, so at a given wavelength
varying A can change SSA. Note that AOD, SSA, g, A, and all the other parameters in
Equations 9.1 and 9.2 vary with space and time due to variations of both aerosol
composition and relative humidity, which influence these characteristics.
July 2009
9-163
DRAFT-DO NOT CITE OR QUOTE

-------
s
w
c
a

iJ
- pi ku
LO UL 3P«T Ml EH NF QtQG IM OMGO Gl TM
MODEL
A19 £'
a
-a
a
a
i
m
£
*5
VJ
c
lu
I
1 QC»
fl(»
80
TO
GO
50
40
30
20
10
ft
m
I. 1, ,
¦ ¦ i ¦ ¦ ¦ ¦ ¦ «i
U3 LS UL 3P C" Ml EH NF OT OG EM CMCG G| TM
MODEL
5


35
DO
•PQJV
•3a
¦su
Source: Kinne et al. (2006,155903)
Figure 9-71. Global annual averaged AOD (upper panel) and aerosol mass loading (lower panel) with
their components simulated by 15 models in AeroCom- A (excluding one model which
only reported mass). SU=S0*'~, BC=black carbon, POM=particulate OC, DU=dust,
SS=sea salt. Model abbreviations: L0 = L0A (Lille, Fra), LS=LSCE (Paris, Fra), UL=ULAQ
(L'Aquila, Ita), SP=SPRINTARS (Kyushu, Jap), CT=ARQM (Toronto, Can), MI=MIRAGE
(Richland, USA), EH=ECHAM5 (MPI Hamburg, Ger), NF=CCM Match (NCAR Boulder,
USA), 0T=0slo-CTM (Oslo, Nor), 0G=0LS0-GCM (Oslo, Nor) (prescribed background for
DU and SS], IM=IMPACT (Michigan, USA), GM=GFDLMozart (Princeton, NJ, USA),
G0=G0CART (NASA-GSFC, Washington DC, USA), GI=GISS (NASA-GISS, New York,
USA), TM=TM5 (Utrecht, Net). Also shown in the upper panel are the averaged
observation data from AERONET (Ae) and the satellite composite (S*).
In the recent AeroCom project. aerosol direct RF for the solar spectral
wavelengths (or shortwave) was assessed based on the 9 models that participated in
both Experiment B and PEE in which identical, prescribed emissions for present (year
2000) and predndustrial time (year 1750) listed in Table 9-11 were used across the
models (Schulz et al., 2006:, 190381). The anthropogenic direct RF was obtained by
subtracting Aero- Com-PRE from AeroCom-B simulated results. Because dust and sea
salt are predominantly from natural sources, they were not included in the
anthropogenic RF assessment although the land use practice can contribute to dust
emissions as "anthropogenic". Other aerosols that were not considered in the AeroCom
forcing assessment were natural sulfate (e.g., from volcanoes or ocean) and POM (e.g.,
from biogenic hydrocarbon oxidation), as well as nitrate. The aerosol direct forcing in
July 2009
9-164
DRAFT-DO NOT CITE OR QUOTE

-------
the AeroCom assessment thus comprises three major anthropogenic aerosol
components sulfate, BC, and POM.
Table 9-13. S042~ mass loading, MEE and AOD at 550 nm, shortwave radiative forcing at the top of the
atmosphere, and normalized forcing with respect to AOD and mass. All values refer to
anthropogenic perturbation.
Model
Mass load MEE
(mg m2) (m2g')
AOD att
550 nm
TOA Forcing
(W/m2)
Forcing/AOD
(W/m2)
Forcing/Mass
(W g1)
PUBLISHED SINCE IPCC 2001
ACCM3
2.23

¦0.56

¦251
B GEOSCHEM
1.53 11.8
0.018
¦0.33
¦18
¦216
C GISS
3.30 6.7
0.022
¦0.65
¦30
¦197
DGISS
3.27

¦0.96

¦294
E GISS*
2.12

¦0.57

¦269
F SPRINTARS
1.55 9.7
0.015
¦0.21

¦135
GLMD
2.76

¦0.42

¦152
HLOA
3.03 9.9
0.03
¦0.41
¦14
¦135
IGATORG
3.06

¦0.32

¦105
JPNNL
5.50 7.6
0.042
¦0.44
¦10
¦80
K UIO-CTM
1.79 10.6
0.019
¦0.37
¦19
¦207
L UIO-GCM
2.28

¦0.29

¦127
AEROCOM: IDENTICAL EMISSIONS USED FOR YR 2000 AND 1750
MUMI 2.64 7.6
0.02
¦0.58
¦29
¦220
N UIO-CTM 1.70 11.2
0.019
¦0.36
¦19
¦212
0 LOA 3.64 9.6
0.035
¦0.49
¦14
¦135
PLSCE 3.01 7.6
0.023
¦0.42
¦18
¦140
Q ECHAMS-HAM 2.47 6.5
0.016
¦0.46
¦29
¦186
R GISS** 1.34 4.5
0.006
¦0.19
¦32
¦142
S UIO-GCM 1.72 7.0
0.012
¦0.25
¦21
¦145
T SPRINTARS 1.19 10.9
0.013
¦0.16
¦12
¦134
UULAQ 1.62 12.3
0.02
¦0.22
¦11
¦136
Average A-L 2.70 9.4
0.024
¦00.46
¦18
¦181
Average M-U 2.15 8.6
0.018
¦0.35
¦21
¦161
Minimum A-U 1.19 4.5
0.006
¦0.96
¦32
¦294
Maximum A-U 5.50 12.3
0.042
¦0.16
¦10
¦80
Std devA-L 1.09 1.9
0.010
0.202
7
68
StddevM-U 0.83 2.6
0.008
0.149
8
35
%Stddev/avg A-L 40% 20%
41%
44%
38%
385
%Stddev/avg M-U 39% 30%
45%
43%
37%
22%
July 2009
9-165
DRAFT-DO NOT CITE OR QUOTE

-------
Mass load	MEE	AODatt TOA Forcing Forcing/AOD Forcing/Mass
(mg m2)	(m2g ')	550 nm	(W/m2)	(W/m2)	(W g')
Model abbreviations: CCM3 = Community Climate Model; GEOSCHEM-Goddard Earth Observing System-Chemistry; GISS = Goddard nstitute for Space Studies; SPRINTARS = Spectral Radiation-
Transport Model for Aerosol Species; LMD = Laboratoire de Meteorologie Dynamique; L0A=Laboratoire d'Optique Atmospherique; GAT0RG = Gas, Aerosol Transport and General circulation model;
PNNL=Pacific Northwest National Laboratory; UI0-CTM = Univeristy of Oslo CTM; UIO-GCM = University of Oslo GCM; UMI = University of Michigan; LSCE = Laboratoire des Sciences du Climat et de
('Environment; ECHAMS5-HAM = European Centre Hamburg with Hamburg Aerosol Module; ULAQ=University of IL'Aquila.
Source: Adapted from IPCC AR4 (Intergovernmental Panel on Climate, 2007,190988) and Schulz et al. (2006,190381)
Table 9-14. Particulate organic matter (POM) and BC mass loading, AOD at 550 nm, shortwave
radiative forcing at the top of the atmosphere, and normalized forcing with respect to AOD
and mass.
Model
Mass
load
(mg m 2)
MEE
(m2g')
AOD at
550 nm
TOA
Forcing
(W/m2)
Forcing/
AOD
(W/m2)
Forcing/
Mass
(W g')
Mass
load (mg
m2)
MEE
(m2g')
AOD at
550 nm
TOA
Forcing
(W/m2)
Forcing/
AOD
(W/m2)
Forcing/
Mass
(W g')
PUBLISHED SINCE IPCC 2001
A SPRINTARS



¦0.24

¦107



0.36


BL0A
2.33
6.9
0.016
¦0.25
¦16
¦140
0.37


0.55


C GISS
1.86
9.1
0.017
¦0.26
¦15
¦161
0.29


0.61


DGISS
1.86
8.1
0.015
¦0.30
¦20
¦75
0.29


0.35


E GISS*
2.39


¦0.18

¦92
0.39


0.50


F GISS
2.49


¦0.23

¦101
0.43


0.53


G SPRINTARS
2.67
10.9
0.029
¦0.27
¦9
¦23
0.53


0.42


HGAT0RG
2.56


¦0.06

¦112
0.39


0.55


1 M0ZGN
3.03
5.9
0.018
¦0.34
¦19







JCCM






0.33


0.34


K UI0-CTM






0.30


0.19


AEROCOM: IDENTICAL EMISSIONS FOR YR 2000 & 1750
L UMI
1.16
5.2
0.0060
¦0.23
¦38
¦198
0.19
6.8
1.29
0.25
194
1316
M UI0-CTM
1.12
5.2
0.0058
¦0.16
¦28
¦143
0.19
7.1
1.34
0.22
164
1158
NL0A
1.41
6.0
0.0085
¦0.16
¦19
¦113
0.25
7.9
1.98
0.32
162
1280
0LSCE
1.50
5.3
0.0079
¦0.17
¦22
¦113
0.25
4.4
1.11
0.30
270
1200
PECHAMS-HAM
1.00
7.7
0.0077
¦0.10
¦13
¦100
0.16
7.7
1.23
0.20
163
1250
Q GISS**
1.22
4.9
0.0060
¦0.14
¦23
¦115
0.24
7.6
1.83
0.22
120
917
R UIO-GCM
0.88
5.2
0.0046
¦0.06
¦13
¦68
0.19
10.3
1.95
0.36
185
1895
S SPRINTARS
1.84
10.9
0.0200
¦0.10
¦5
¦54
0.37
9.5
3.50
0.32
91
865
TULAQ
1.71
4.4
0.0075
¦0.09
¦12
¦53
0.38
7.6
2.90
0.08
28
211
Average A-K
2.40
8.2
0.019
-0.24
-16
-102
0.37


0.44

1242
Average L-T
1.32
6.1
0.008
-0.13
-19
-106
0.25
7.7
1.90
0.25
153
1121
Minimum A-T
0.88
4.4
0.005
-0.34
-38
-198
0.16
4.4
1.11
0.08
28
211
Maximum A-T
3.03
10.9
0.029
-0.06
-5
-23
0.53
10.3
3.50
0.61
270
2103
Std dev A-K
0.39
1.7
0.006
0.09
4
41
0.08


0.06

384
Std dev L-T
0.32
2.0
0.005
0.05
10
46
0.08
1.6
0.82
0.09
68
450
%Stddev/avg A-K
16%
21%
30%
36%
26%
41%
22%


23%

31%
July 2009
9-166
DRAFT-DO NOT CITE OR QUOTE

-------
Model
Mass
load
(mg m2)
MEE
(m2g')
AOD at
550 nm
TOA
Forcing
(W/m2)
Forcing/
AOD
(W/m2)
Forcing/
Mass
(W g')
Mass
load (mg
m2)
MEE
(m2g')
AOD at
550 nm
TOA
Forcing
(W/m2)
Forcing/
AOD
(W/m2)
Forcing/
Mass
(W g')
%Stddev/avg L-T
25%
33%
56%
39%
52%
43%
32%
21%
43%
34%
45%
40%
Source: Based on IPCC AR4 (2007,190988) and Schulz et al. (2006,190381).
The IPCC AR.4 (Intergovernmental Panel on Climate, 2007, 190988) assessed
anthropogenic aerosol RF based on the model results published after the IPCC TAR. in
2001, including those from the AeroCom study discussed above. These results
(adopted from IPCC AR4) are shown in Table 9"13 for sulfate and Table 9"14 for
carbonaceous aerosols (BC and POM), respectively. All values listed in Table 9"13 and
9.14 refer to anthropogenic perturbation, i.e., excluding the natural fraction of these
aerosols. In addition to the mass burden, MEE, and AOD, Table 9"13 and 9.14 also list
the "normalized forcing", also known as "forcing efficiency", one for the forcing per
unit AOD, and the other the forcing per gram of aerosol mass (dry). For some models,
aerosols are externally mixed, that is, each aerosol particle contains only one aerosol
type such as sulfate, whereas other models allow aerosols to mix internally to
different degrees, that is, each aerosol particle can have more than one component,
such as black carbon coated with sulfate. For models with internal mixing of aerosols,
the component values for AOD, MEE, and forcing were extracted (Schulz et al., 2006,
190381).
Considerable variation exists among these models for all quantities in Table 9"13
and 9.14. The RF for all the components varies by a factor of 6 or more- Sulfate from
0.16 to 0.96 W/m2, POM from -0.06 to -0.34 W/m2, and BC from +0.08 to +0.61 W/m2,
with the standard deviation in the range of 30 to 40% of the ensemble mean. It should
be noted that although BC has the lowest mass loading and AOD, it is the only aerosol
species that absorbs strongly, causing positive forcing to warm the atmosphere, in
contrast to other aerosols that impose negative forcing to cool the atmosphere. As a
result, the net anthropogenic aerosol forcing as a whole becomes less negative when
BC is included. The global average anthropogenic aerosol direct RF at the top of the
atmosphere (TOA) from the models, together with observation-based estimates (see
Chapter 2 of the CCSP SAP2.3), is presented in Figure 9"72. Note the wide range for
forcing in Figure 9"72. The comparison with observation-based estimates shows that
the model estimated forcing is in general lower, partially because the forcing value
from the model is the difference between present-day and pre "industrial time,
whereas the observation-derived quantity is the difference between an atmosphere
with and without anthropogenic aerosols, so the "background" value that is subtracted
from the total forcing is higher in the models. The discussion so far has dealt with
global average values. The geographic distributions of multi-model aerosol direct RF
has been evaluated among the AeroCom models, which are shown in Figure 9"73 for
total and anthropogenic AOD at 550 nm and anthropogenic aerosol RF at TOA, within
the atmospheric column, and at the surface. Globally, anthropogenic AOD is about
25% of total AOD (Figure 9"73a and b) but is more concentrated over polluted regions
in Asia, Europe, and North America and biomass burning regions in tropical southern
Africa and South America. At TOA, anthropogenic aerosol causes negative forcing over
mid-latitude continents and oceans with the most negative values ("1 to "2 W/m2) over
polluted regions (Figure 9-73C). Although anthropogenic aerosol has a cooling effect at
the surface with surface forcing values down to "10 W/m2 over China, India, and
tropical Africa (Figure 9-73E), it warms the atmospheric column with the largest
effects again over the polluted and biomass burning regions. This heating effect will
change the atmospheric circulation and can affect the weather and precipitation (e.g.,
Kim et al., 2006, 190917).
July 2009
9-167
DRAFT-DO NOT CITE OR QUOTE

-------
i
H
O
T
E-XI
8
B
t.
a
*j
ti
E
c
u
<
Asro&ul Direct Radiative Forcing
nr
Bfileuiftci* tiCi)
CNng u-f .
iu fJT # \iWt\
rasE 1 !• imp i—
-------
biomass burning) is "0.22 ± 0.18 W/m2 averaged globally, exerting a net cooling effect.
This value would represent the low-end of the forcing magnitude, since some
potentially significant anthropogenic aerosols, such as nitrate and dust from human
activities are not included because of their highly uncertain sources and processes.
IPCC AR4 had adjusted the total anthropogenic aerosol direct RF to "0.5 ± 0.4 W/m2 by
adding estimated anthropogenic nitrate and dust forcing values based on limited
modeling studies and by considering the observation-based estimates (see Chapter 2 of
the CCSP SAP2.3).
¦	Both sulfate and POM negative forcing whereas BC causes positive forcing because of
its highly absorbing nature. Although BC comprises only a small fraction of
anthropogenic aerosol mass load and AOD, its forcing efficiency (with respect to either
AOD or mass) is an order of magnitude stronger than sulfate and POM, so its positive
shortwave forcing largely offsets the negative forcing from sulfate and POM. This
points out the importance of improving the model ability to simulate each individual
aerosol components more accurately, especially black carbon. Separately, it is
estimated from recent model studies that anthropogenic sulfate, POM, and BC forcings
at TOA are "0.4, "0.18, +0.35 W/m2, respectively. The anthropogenic nitrate and dust
forcings are estimated at "0.1 W/m2 for each, with uncertainties exceeds 100%
(Intergovernmental Panel on Climate, 2007, 190988).
¦	In contrast to long-lived greenhouse gases, anthropogenic aerosol RF exhibits
significant regional and seasonal variations. The forcing magnitude is the largest over
the industrial and biomass burning source regions, where the magnitude of the
negative aerosol forcing can be of the same magnitude or even stronger than that of
positive greenhouse gas forcing.
¦	There is a large spread of model-calculated aerosol RF even in the global annual
averaged values. The AeroCom study shows that the model diversity at some locations
(mostly East Asia and African biomass burning regions) can reach ± 3 W/m2, which is
an order of magnitude above the global averaged forcing value of "0.22 W/m2. The large
diversity reflects the low level of current understanding of aerosol radiative forcing,
which is compounded by uncertainties in emissions, transport, transformation,
removal, particle size, and optical and microphysical (including hygroscopic) properties.
¦	In spite of the relatively small value of forcing at TOA, the magnitudes of
anthropogenic forcing at the surface and within the atmospheric column are
considerably larger- "1 to "2 W/m2 at the surface and +0.8 to +2 W/m2 in the
atmosphere. Anthropogenic aerosols thus cool the surface but heat the atmosphere, on
average. Regionally, the atmospheric heating can reach annually averaged values
exceeding 5 W/m2 (Figure 9-73D). These regional effects and the negative surface
forcing are expected to exert an important effect on climate through alteration of the
hydrological cycle.
July 2009
9-169
DRAFT-DO NOT CITE OR QUOTE

-------
Figure 9-73. Aerosol optical thickness and anthropogenic shortwave all-sky radiative forcing from
the AeroOom study. Shown in the figure: total AOD (A) and anthropogenic AOD (B) at
550 nm, and radiative forcing at TOA (C), atmospheric column (D)f and surface (E).
O.&O
~.SO
0.70
0.60
0.50
0.40
0.30
0.20
~.SO
0.00
0.0ft
0.04
0,02
0,00
- 1-SU .135 K> -45 0
Lonffftata
1.0
0.5
0,3
0.0
-0.2
-Q.S
-i.a
-2.a
Vi/m2
5.0
*.0
3.a
2.0
1.0
0.5
0.5
0.0
-0.2
-a.o
-i.a
-2.a
-j.a
-+.~
-5.0
Source: Schulz et al. (2006,190381) and AeroCom image catalog (http:// nansen.ipsl.jussieu.fr/AEROCOMIdata.html).
LnnplutH
D
-130 135 40 -45 0 45 M 1S6 lflO
Inn^tuin
WI'UKWM
0.3000
Q.2KJD
0.10Q0
~.0750
0.0500
o.oaso
0.0100
0.0050
0.DD25
0.0010
Q.OOQO
July 2009
9-170
DRAFT-DO NOT CITE OR QUOTE

-------
Raciialivs Forcing tram CJolkI Albodn Eftecl
lti
mri
t
Ci
a
¦jH
lv-l
¦>1
a
2
¦s

			[— ' f 1 t ' 1
Mnomy fJanai «td 3QQ11

HBfJfe? iWMimw m . KWibl
1
CS»SO 'Tlntirav''' Mid Rammr. 2Q011
'n
(AW u< - XSOni
\
CSIRQ ami Uu. 2Q0®|
1
1
t MOJ (fliwul 4 WJM i

vMCZj^OLDES {Oftwi m m
i
1 1
GFDL I«hUtqM Hi. smabj
1
FDHAM ^L«^frn Hi m , Kffil
•H
PMNL 
.Tl»,
IWDZ'MODia i.Oiim m Bimtfw ztwii
i #
L I
LA* fiitt i Cho ¦ ami £0001
1
UU V Chop qjljJ PfT*g|. 30061

UM 2 Qrrt2 and Furwaf. L>jl'ji
r J
H
IJMJJ i!Swr nnd Ptennw. XBel

UMJ) >;ci4«i and FsniMLr, JKi

UM_» |
t— -1
1 a 1 a 1 I 1
-SLO -1_S -t,Q -&5
RadiativB Fcfrci'ip (Wm"2J
Source: IPCC (2007. 190988).
Figure 9-74. Radiative forcing from the cloud albedo effect {1st aerosol indirect effect) in the global
climate models used from IPCC (2007,190988), Chapter 2, Figure 2.14, of the IPCC AR4.
Species included in the lower panel are S(h2", sea salt, organic and BC, dust and
nitrates; in the top panel, only Sth2", sea salt and OC are included.
July 2009
9-171
DRAFT-DO NOT CITE OR QUOTE

-------
9.3.6.4. Calculating Aerosol Indirect Forcing
Aerosol Effects on Clouds. A subset of the aerosol particles can act as cloud
condensation nuclei (CCN) and/or ice nuclei (IN). Increases in aerosol particle
concentrations, therefore, may increase the ambient concentrations of CCN and IN,
affecting cloud properties. For a fixed cloud liquid water content, a CCN increase will
lead to more cloud droplets so that the cloud droplet size will decrease. That effect
leads to brighter clouds, the enhanced albedo then being referred to as the "cloud
albedo effect" (Twomey, 1977, 190533). also known as the first indirect effect. If the
droplet size is smaller, it may take longer to rainout, leading to an increase in cloud
lifetime, hence the "cloud lifetime" effect (Albrecht, 1989, 045783). also called the
second indirect effect. Approximately one-third of the models used for the IPCC 20th
century climate change simulations incorporated an aerosol indirect effect, generally
(though not exclusively) considered only with sulfates.
Shown in Figure 9"74 are results from published model studies indicating the
different RF values from the cloud albedo effect. The cloud albedo effect ranges from -
0.22 to "1.85 W/m2; the lowest estimates are from simulations that constrained
representation of aerosol effects on clouds with satellite measurements of drop size vs.
aerosol index. In view of the difficulty of quantifying this effect remotely (discussed
later), it is not clear whether this constraint provides an improved estimate. The
estimate in the IPCC AR4 ranges from +0.4 to -1.1 W/m2, with a "best-guess" estimate
of 0.7 W/m2.
The representation of cloud effects in GCMs is considered below. However, it is
becoming increasingly clear from studies based on high resolution simulations of
aerosol-cloud interactions that there is a great deal of complexity that is unresolved in
climate models. This point is examined below (High Resolution Modeling).
Most models did not incorporate the "cloud lifetime effect." Hansen et al. (2005,
059087) compared this latter influence (in the form of time-averaged cloud area or
cloud cover increase) with the cloud albedo effect. In contrast to the discussion in
IPCC (2007, 189430). they argue that the cloud cover effect is more likely to be the
dominant one, as suggested both by cloud-resolving model studies (Ackerman et al.,
2004, 190056) and satellite observations (Kaufman et al., 2005, 155891). The cloud
albedo effect may be partly offset by reduced cloud thickness accompanying aerosol
pollutants, producing a meteorological (cloud) rather than aerosol effect (see the
discussion in Lohman and Feichter, 2005, 155942). The distinction between
meteorological feedback and aerosol forcing can become quite opaque! as noted earlier,
the term feedback is restricted here to those processes that are responding to a change
in temperature. Nevertheless, both aerosol indirect effects were utilized in Hansen et
al. (2005, 059087). with the second indirect effect calculated by relating cloud cover to
the aerosol number concentration, which in turn is a function of sulfate, nitrate, black
carbon and organic carbon concentration. Only the low altitude cloud influence was
modeled, principally because there are greater aerosol concentrations at low levels,
and because low clouds currently exert greater cloud RF. The aerosol influence on high
altitude clouds, associated with IN changes, is a relatively unexplored area for models
and as well for process-level understanding.
Hansen et al. (2005, 059087) used coefficients to normalize the cooling from
aerosol indirect effects to between "0.75 and "1 W/m2, based on comparisons of
modeled and observed changes in the diurnal temperature range as well as some
satellite observations. The response of the GISS model to the direct and two indirect
effects is shown in Figure 9-75. As parameterized, the cloud lifetime effect produced
somewhat greater negative RF (cooling), but this was the result of the coefficients
chosen. Geographically, it appears that the "cloud cover" effect produced slightly more
cooling in the Southern Hemisphere than did the "cloud albedo" response, with the
reverse being true in the Northern Hemisphere (differences on the order of a few
tenths °C).
July 2009
9-172
DRAFT-DO NOT CITE OR QUOTE

-------
Model Experiments. There are many different factors that can explain the
large divergence of aerosol indirect effects in models (Figure 9-744). To explore this in
more depth, I Vnner et al. (2006, 190564) used three general circulation models to
analyze the differences between models for the first indirect effect, as well as a
combined first plus second indirect effect. The models all had different cloud and/or
convection parameterizations. In the first experiment, the monthly average aerosol
mass and size distribution of, effectively, sulfate aerosol were prescribed, and all
models followed the same prescription for parameterizing the cloud droplet number
concentration (OIDNC) as a function of aerosol concentration. In that sense, the only
difference among the models was their separate cloud formation and radiation
schemes. The different models all produced similar droplet effective radii, and
therefore shortwave cloud forcing, and change in net outgoing whole sky radiation
between pre-indust rial times and the present. Hence the first indirect effect was not a
strong function of the cloud or radiation scheme. The results for this and the following
experiments are presented in Figure 9-76, where the experimental results are shown
sequentially from left to right for the whole sky effect and in Table 9-15 for the clear-
sky and cloud forcing response as well,
AIE: CIdAlb {I Run
AIE: CSdAlb
AIE: CldCov
Cloud Cover (%)
Direct	.16
Aerosol Direct arid Indirect Effects
Planetary Albedo (%)
.14
ATs (°C)
Direct {I E3 Run) -.32
All- r. V > K. ; : 4,
V -* j
-23'5 a -I -5-.2 2 .5 ! 2 5 19 -5 -2 -I -.S-.2 -.1,1 .2 .5 I 2 5
Source: Hansen et al. (2005,059087).
Figure 9-75. Anthropogenic impact on cloud cover, planetary albedo, radiative flux at the surface
(while holding sea surface temperatures and sea ice fixed} and surface air temperature
change from the direct aerosol forcing (top row), the first indirect effect (second row)
and the second indirect effect (third row). The temperature change is calculated from
years 81-120 of a coupled atmosphere simulation with the GISS model.
The change in cloud forcing is the difference between whole sky and clear sky
outgoing radiation in the present day minus pre-industrial simulation. The large
differences seen between experiments 5 and 6 are due to the inclusion of the clear sky
component of aerosol scattering and absorption (the direct effect) in experiment 6.
July 2009
9-173
DRAFT-DO NOT CITE OR QUOTE

-------
In the second experiment, the aerosol mass and size distribution were again
prescribed, but now each model used its own formulation for relating aerosols to
droplets. In this case one of the models produced larger effective radii and therefore a
much smaller first indirect aerosol effect (Figure 9.76, Table !> I However, even in
the two models where the effective radius change and net global forcing were similar,
the spatial patterns of cloud forcing differ, especially over the biomass burning regions
of Africa and South America.
The third experiment allowed the models to relate the change in droplet size to
change in precipitation efficiency (i.e. < they were now also allowing the second indirect
.effect. — smaller droplets being less efficient rain producers — as well asthe first). The;
models utilized the same relationship for autoconversion of cloud droplets to
precipitation. Changing the precipitation efficiency results in all models producing an
increase in cloud liquid water path, although the effect on cloud fraction was smaller
than in the previous experiments. The net result was to increase the negative
radiative forcing in all three models, albeit with different magnitudes^ for two of the
models the net impact on outgoing shortwave radiative increased by about 20%,
whereas in the third model (which hadthe much smaller first indirect effect), it was
magnified by a factor of three..
Exp I Exp 2 Exp 3 Exp 4 Exp S Exp 6
(_AM-CHwl
LMD
i, cess
Source: Adapted from Penner et al. (2006,190564).
Figure 9-76. Global average present-day short wave cloud forcing at TOA (top) and change in whole
sky net outgoing shortwave radiation (bottom) between the present-day and pre-
industrial simulations for each model in each experiment.
July 2009
9-174
DRAFT-DO NOT CITE OR QUOTE

-------
Table 9-15. Differences in present day and pre-industrial outgoing solar radiation (W/m2) in the
different experiments.
Model
EXP 1
EXP 2
EXP 3
EXP 4
EXP 5
EXP 6
WHOLESKY
CAM-Oslo
¦0.648
¦0.726
¦0.833
¦0.580
¦0.365
¦0.518
LMD-Z
¦0.682
¦0.597
¦0.722
¦1.194
¦1.479
¦1.553
CCSR
¦0.739
¦0.218
¦0.733
¦0.350
¦1.386
¦1.386
CLEARSKY
CAM-Oslo
¦0.063
¦0.066
¦0.026
0.014
¦0.054
¦0.575
LMD-Z
¦0.054
0.019
0.030
¦0.066
¦0.126
¦1.034
CCSR
0.018
¦0.007
¦0.045
¦0.008
0.018
¦1.160
CLOUDFORCING
CAM-Oslo
¦0.548
¦0.660
¦0.807
¦0.595
¦0.311
0.056
LMD-Z
¦0.628
¦0.616
¦0.752
¦1.128
¦1.353
¦0.518
CCSR
¦0.757
¦0.212
¦0.728
¦0.345
¦1.404
¦0.200
EXP1: tests cloud formation and radiation schemes
EXP2: tests formulation for relating aerosols to droplets
EXP3: tests inclusion of droplet size influence on precipitation efficiency
EXP4: tests formulation of droplet size influence on precipitation efficiency
EXP5: tests model aerosol formulation from common sources
EXP6: added the direct aerosol effect
Source: Adapted from Penner et al. (2006,190564).
In the fourth experiment, the models were now each allowed to use their own
formulation to relate aerosols to precipitation efficiency. This introduced some
additional changes in the whole sky shortwave forcing (Figure 9-76).
In the fifth experiment, models were allowed to produce their own aerosol
concentrations, but were given common sources. This produced the largest changes in
the RF in several of the models. Within any one model, therefore, the change in
aerosol concentration has the largest effect on droplet concentrations and effective
radii. This experiment too resulted in large changes in RF.
In the last experiment, the aerosol direct effect was included, based on the full
range of aerosols used in each model. While the impact on the whole-sky forcing was
not large, the addition of aerosol scattering and absorption primarily affected the
change in clear sky radiation (Table 9" 15).
The results of this study emphasize that in addition to questions concerning cloud
physics, the differences in aerosol concentrations among the models play a strong role
in inducing differences in the indirect effect(s), as well as the direct one.
Observational constraints on climate model simulations of the indirect effect with
satellite data (e.g., MODIS) have been performed previously in a number of studies
(e.g., Storlevmo et al., 2006, 190418; Lohmann et al., 2006, 190451; Quaas et al., 2006,
190915! Menon et al., 2008, 190534). These have been somewhat limited since the
satellite retrieved data used do not have the vertical profiles needed to resolve aerosol
and cloud fields (e.g., cloud droplet number and liquid water content); the temporal
resolution of simultaneous aerosol and cloud product retrievals are usually not
available at a frequency of more than one a day; and higher level clouds often obscure
low clouds and aerosols. Thus, the indirect effect, especially the second indirect effect,
remains, to a large extent, unconstrained by satellite observations. However,
improved measurements of aerosol vertical distribution from the newer generation of
July 2009
9-175
DRAFT-DO NOT CITE OR QUOTE

-------
sensors on the A-train platform may provide a better understanding of changes to
cloud properties from aerosols. Simulating the top-of-atmosphere reflectance for
comparison to satellite measured values could be another way to compare model with
observations, which would eliminate the inconsistent assumptions of aerosol optical
properties and surface reflectance encountered when compared the model calculated
and satellite retrieved AOD values.
Additional Aerosol Influences. Various observations have empirically related
aerosols injected from biomass burning or industrial processes to reductions in
rainfall (e.g., Warner, 1968, 157114! Eagan et al., 1974, 190231; Andreae et al., 2004,
155658; Rosenfeld, 2000, 002234). There are several potential mechanisms associated
with this response.
In addition to the two indirect aerosol effects noted above, a process denoted as
the "semidirect" effect involves the absorption of solar radiation by aerosols such as
black carbon and dust. The absorption increases the temperature, thus lowering the
relative humidity and producing evaporation, hence a reduction in cloud liquid water.
The impact of this process depends strongly on what the effective aerosol absorption
actually is! the more absorbing the aerosol, the larger the potential positive forcing on
climate (by reducing low level clouds and allowing more solar radiation to reach the
surface). This effect is responsible for shifting the critical value of SSA (separating
aerosol cooling from aerosol warming) from 0.86 with fixed clouds to 0.91 with varying
clouds (Hansen et al., 1997, 043104). Reduction in cloud cover and liquid water is one
way aerosols could reduce rainfall.
More generally, aerosols can alter the location of solar radiation absorption within
the system, and this aspect alone can alter climate and precipitation even without
producing any change in net radiation at the top of the atmosphere (the usual metric
for climate impact). By decreasing solar absorption at the surface, aerosols (from both
the direct and indirect effects) reduce the energy available for evapotranspiration,
potentially resulting in a decrease in precipitation. This effect has been suggested as
the reason for the decrease in pan evaporation over the last 50 yr (Roderick and
Farquhar, 2002, 042788). The decline in solar radiation at the surface appears to have
ended in the 1990s (Wild et al., 2005, 156156). perhaps because of reduced aerosol
emissions in industrial areas (Kruger and Grasl, 2002, 190200). although this issue is
still not settled.
Energy absorption by aerosols above the boundary layer can also inhibit
precipitation by warming the air at altitude relative to the surface, i.e., increasing
atmospheric stability. The increased stability can then inhibit convection, affecting
both rainfall and atmospheric circulation (Ramanathan et al., 2001, 042681! Chung
and Zhang, 2004, 190054). To the extent that aerosols decrease droplet size and
reduce precipitation efficiency, this effect by itself could result in lowered rainfall
values locally.
In their latest simulations, Hansen et al. (2007, 190597) did find that the indirect
aerosol effect reduced tropical precipitation! however, the effect is similar regardless of
which of the two indirect effects is used, and also similar to the direct effect. So it is
likely that the reduction of tropical precipitation is because of aerosol induced cooling
at the surface and the consequent reduced evapotranspiration. Similar conclusions
were reached by Yu et al. (2002, 190923) and Feingold et al. (2005, 190550). In this
case, the effect is a feedback and not a forcing.
The local precipitation change, through its impacts on dynamics and soil
moisture, can have large positive feedbacks. Harvey (2004, 190598) concluded from
assessing the response to aerosols in eight coupled models that the aerosol impact on
precipitation was larger than on temperature. He also found that the precipitation
impact differed substantially among the models, with little correlation among them.
Recent GCM simulations have further examined the aerosol effects on
hydrological cycle. Ramanathan et al. (2005, 190199) showed from fully coupled
ocean-atmosphere GCM experiments that the "solar dimming" effect at the surface,
July 2009
9-176
DRAFT-DO NOT CITE OR QUOTE

-------
i.e., the reduction of solar radiation reaching the surface, due to the inclusion of
absorbing aerosol forcing causes a reduction in surface evaporation, a decrease in
meridional sea surface temperature (SST) gradient and an increase in atmospheric
stability, and a reduction in rainfall over South Asia. Lau and Kim (2006, 190226)
examined the direct effects of aerosol on the monsoon water cycle variability from
GCM simulations with prescribed realistic global aerosol forcing and proposed the
"elevated heat pump" effect, suggesting that atmospheric heating by absorbing
aerosols (dust and black carbon), through water cycle feedback, may lead to a
strengthening of the South Asia monsoon. These model results are not necessarily at
odds with each other, but rather illustrate the complexity of the aerosol-monsoon
interactions that are associated with different mechanisms, whose relative importance
in affecting the monsoon may be strongly dependent on spatial and temporal scales
and the timing of the monsoon. These results may be model dependent and should be
further examined.
High Resolution Modeling. Largely by its nature, the representation of the
interaction between aerosol and clouds in GCMs is poorly resolved. This stems in
large part from the fact that GCMs do not resolve convection on their large grids
(order of several hundred km), that their treatment of cloud microphysics is rather
crude, and that as discussed previously, their representation of aerosol needs
improvement. Superparametrization efforts (where standard cloud parameterizations
in the GCM are replaced by resolving clouds in each grid column of the GCM via a
cloud resolving model) (e.g.,Grabowski, 2004, 190590) could lead the way for the
development of more realistic cloud fields and thus improved treatments of aerosol
cloud interactions in large-scale models. However, these are just being incorporated in
models that resolve both cloud and aerosols. Detailed cloud parcel models have been
developed to focus on the droplet activation problem (that asks under what conditions
droplets actually start forming) and questions associated with the first indirect effect.
The coupling of aerosol and cloud modules to dynamical models that resolve the large
turbulent eddies associated with vertical motion and clouds [large eddy simulations
(LES) models, with grid sizes of ~100 m and domains ~10 km] has proven to be a
powerful tool for representing the details of aerosol-cloud interactions together with
feedbacks (e.g., Feingold et al., 1994, 190535! Kogan et al., 1994, 190186; Stevens et
al., 1996, 190417! Feingold et al., 1999, 190540; Ackerman et al., 2004, 190056). This
section explores some of the complexity in the aerosol indirect effects revealed by such
studies to illustrate how difficult parameterizing these effects properly in GCMs could
really be.
The First Indirect Effect. The relationship between aerosol and drop
concentrations (or drop sizes) is a key piece of the first indirect effect puzzle. (It should
not, however, be equated to the first indirect effect which concerns itself with the
resultant RF). A huge body of measurement and modeling work points to the fact that
drop concentrations increase with increasing aerosol. The main unresolved questions
relate to the degree of this effect, and the relative importance of aerosol size
distribution, composition and updraft velocity in determining drop concentrations (for
a review, see Mcfiggans et al., 2006, 190532). Studies indicate that the aerosol number
concentration and size distribution are the most important aerosol factors. Updraft
velocity (unresolved by GCMs) is particularly important under conditions of high
aerosol particle number concentration.
Although it is likely that composition has some effect on drop number
concentrations, composition is generally regarded as relatively unimportant compared
to the other parameters (Fitzgerald, 1975, 095417; Feingold et al., 2003, 190551;
Ervens et al., 2005, 190527; Dusek et al., 2006, 155756). Therefore, it has been stated
that the significant complexity in aerosol composition can be modeled, for the most
part, using fairly simple parameterizations that reflect the soluble and insoluble
fractions (e.g., Rissler et al., 2004, 190225). However, composition cannot be simply
dismissed. Furthermore, chemical interactions also cannot be overlooked. Alarge
uncertainty remains concerning the impact of organic species on cloud droplet growth
kinetics, thus cloud droplet formation. Cloud drop size is affected by wet scavenging,
July 2009
9-177
DRAFT-DO NOT CITE OR QUOTE

-------
which depends on aerosol composition especially for freshly emitted aerosol. And
future changes in composition will presumably arise due to biofuels/biomass burning
and a reduction in sulfate emissions, which emphasizes the need to include
composition changes in models when assessing the first indirect effect. The simple
soluble/insoluble fraction model may become less applicable than is currently the case.
The updraft velocity, and its change as climate warms, may be the most difficult
aspect to simulate in GCMs because of the small scales involved. In GCMs it is
calculated in the dynamics as a grid box average, and parameterized on the small
scale indirectly because it is a key part of convection and the spatial distribution of
condensate, as well as droplet activation. Numerous solutions to this problem have
been sought, including estimation of vertical velocity based on predicted turbulent
kinetic energy from boundary layer models (Lohmann et al., 1999, 190443; Larson et
al., 2001, 190212) and PDF representations of subgrid quantities, such as vertical
velocity and the vertically-integrated cloud liquid water ('liquid water path,' or LWP)
(Pincus and Klein, 2000, 190565! Golaz et al., 2002, 190587; 2002, 190589! Larson et
al., 2005, 190220). Embedding cloud-resolving models within GCMs is also being
actively pursued (Grabowski et al., 1999, 190592; Randall et al., 2003, 190201).
Numerous other details come into play! for example, the treatment of cloud droplet
activation in GCM frameworks is often based on the assumption of adiabatic
conditions, which may overestimate the sensitivity of cloud to changes in CCN
(Sotiropoulou et al., 2007, 190405! Sotiropoulou et al., 2006, 190406). This points to
the need for improved theoretical understanding followed by new parameterizations.
Other Indirect Effects. The second indirect effect is often referred to as the
"cloud lifetime effect", based on the premise that non-precipitating clouds will live
longer. In GCMs the "lifetime effect" is equivalent to changing the representation of
precipitation production and can be parameterized as an increase in cloud area or
cloud cover (e.g., Hansen et al., 2005, 059087). The second indirect effect hypothesis
states that the more numerous and smaller drops associated with aerosol
perturbations, suppress collision-induced rain, and result in a longer cloud lifetime.
Observational evidence for the suppression of rain in warm clouds exists in the form
of isolated studies (e.g., Warner, 1968, 157114) but to date there is no statistically
robust proof of surface rain suppression (Levin and Cotton, 2008, 190375). Results
from ship-track studies show that cloud water may increase or decrease in the tracks
(Coakley and Walsh, 2002, 192025) and satellite studies suggest similar results for
warm boundary layer clouds (Han et al., 2002, 049181). Ackerman et al. (2004,
190056) used LES to show that in stratocumulus, cloud water may increase or
decrease in response to increasing aerosol depending on the relative humidity of the
air overlaying the cloud. Wang et al. (2003, 157106) showed that all else being equal,
polluted stratocumulus clouds tend to have lower water contents than clean clouds
because the small droplets associated with polluted clouds evaporate more readily and
induce an evaporation-entrainment feedback that dilutes the cloud. This result was
confirmed by Xue and Feingold (2006, 190920) and Jiang and Feingold (2006, 190976)
for shallow cumulus, where pollution particles were shown to decrease cloud fraction.
Furthermore, Xue et al. (2008, 190921) suggested that there may exist two regimes'
the first, a precipitating regime at low aerosol concentrations where an increase in
aerosol will suppress precipitation and increase cloud cover (Albrecht, 1989, 045783);
and a second, non-precipitating regime where the enhanced evaporation associated
with smaller drops will decrease cloud water and cloud fraction.
The possibility of bistable aerosol states was proposed earlier by Baker and
Charlson (1990, 190016) based on consideration of aerosol sources and sinks. They
used a simple numerical model to suggest that the marine boundary layer prefers two
aerosol states' a clean, oceanic regime characterized by a weak aerosol source and less
reflective clouds! and a polluted, continental regime characterized by more reflective
clouds. On the other hand, study by Ackerman et al. (1994, 189975) did not support
such a bistable system using a somewhat more sophisticated model. Further
observations are needed to clarify the nature of cloud/aerosol interactions under a
variety of conditions.
July 2009
9-178
DRAFT-DO NOT CITE OR QUOTE

-------
Finally, the question of possible effects of aerosol on cloud lifetime was examined
by Jiang et al. (2006, 133165). who tracked hundreds of cumulus clouds generated by
LES from their formative stages until they dissipated. They showed that in the model
there was no effect of aerosol on cloud lifetime, and that cloud lifetime was dominated
by dynamical variability.
It could be argued that the representation of these complex feedbacks in GCMs is
not warranted until a better understanding of the processes is at hand. Moreover,
until GCMs are able to represent cloud scales, it is questionable what can be obtained
by adding microphysical complexity to poorly resolved clouds. Abetter representation
of aerosol-cloud interactions in GCMs therefore depends on the ability to improve
representation of aerosols and clouds, as well as their interaction, in the hydrologic
cycle. This issue is discussed further in the next chapter.
9.3.6.5. Aerosol in the Climate Models
Aerosol in the IPCC AR4 Climate Model Simulations. To assess the
atmospheric and climate response to aerosol forcing, e.g., changes in surface
temperate, precipitation, or atmospheric circulation, aerosols, together with
greenhouse gases should be an integrated part of climate model simulation under the
past, present, and future conditions. Table 9" 16 lists the forcing species that were
included in 25 climate modeling groups used in the IPCC AR4 (2007, 190988)
assessment. All the models included long-lived greenhouse gases, most models
included sulfate direct forcing, but only a fraction of those climate models considered
other aerosol types. In other words, aerosol RF was not adequately accounted for in
the climate simulations for the IPCC AR4. Put still differently, the current aerosol
modeling capability has not been fully incorporated into the climate model
simulations. As pointed out in Section 9.3.6.4, fewer than one-third of the models
incorporated an aerosol indirect effect, and most considered only sulfates.
The following discussion compares two of the IPCC AR4 climate models that
include all major forcing agencies in their climate simulation- the model from the
NASA Goddard Institute for Space Studies (GISS) and from the NOAA Geophysical
Fluid Dynamics Laboratory (GFDL). The purpose in presenting these comparisons is
to help elucidate how modelers go about assessing their aerosol components, and the
difficulties that entail. A particular concern is how aerosol forcings were obtained in
the climate model experiments for IPCC AR4. Comparisons with observations have
already led to some improvements that can be implemented in climate models for
subsequent climate change experiments (e.g., Koch et al., 2006, 190184. for GISS
model). This aspect is discussed further in Chapter 4 of the CCSP SAP2.3.
July 2009
9-179
DRAFT-DO NOT CITE OR QUOTE

-------
Table 9-16. Forcings used in IPCC AR4 simulations of 20th century climate change. This table is
adapted from SAP 1.1 Table 5.2 (compiled using information provided by the participating
modeling centers, see http://www.pcmdi.llnl.gov/ipcc/model-documentation/ipcc
model documentation.php) plus additional information from that website. Eleven different
forcings are listed: well-mixed greenhouse gases (G), tropospheric and stratospheric ozone
(0), SO42" aerosol direct (SD) and indirect effects (S), black carbon (BC) and organic carbon
aerosols (OC), mineral dust (MD), sea salt (SS), land use/land cover (LU), solar irradiance
(SO), and volcanic aerosols (V). Check mark denotes inclusion of a specific forcing. As used
here, "inclusion" means specification of a time-varying forcing, with changes on
interannual and longer timescales.

Model
Country
G
0
SD
SI
BC
OC
MD
SS
LU
so
V
1
BCC-CMI
China
¦V
¦V
V








2
BCCR-BCM2.0
Norway
-V

¦V



-V
V



3
CCSM3
USA
¦V

¦V

¦V
V



V
V
4
CGCM3.1 (T47)
Canada
¦V

¦V








5
CGCM3.1 (T63)
Canada
¦V

¦V








6
CNRM-CM3
France
¦V
¦V









7
CSIR0-Mk3.0
Australia
-V

¦V








8
CSIR0-Mk3.5
Australia
-V

¦V








9
ECHAMS/MPI-OM
Germany
¦V

¦V
V







10
ECHO-G
Germany/
Korea
~
~
~
~





~
~
11
FG0ALS-g1.0
China
~

~








12
GFDL-CM2.0
USA
~
~
~

~



~
~
~
13
GFDL-CM2.1
USA
~
~
~

~



~
~
~
14
GISS-AOM
USA
~

~




~



15
GISS-EH
USA
~
~
~
~
~
~
~
~
~
~
~
16
GISS-ER
USA
~
~
~
~
~
~
~
~
~
~
~
17
INGV-SXG
Italy
~
~
~








18
INM-CM3.0
Russia
~

~






~

19
IPSL-CM4
France
~

~
~







20
MICR0C3.2(hires)
Japan
~
~
~

~
~
~
~
~
~
~
21
MICR0C3.2(medres)
Japan
~
~
~

~
~
~
~
~
~
~
22
MRI-CGCM2.3.2
Japan
~

~






~
~
23
PCM
USA
¦V
¦V
¦V






V
V
24
UKM0-HasCM3
U.K.
¦V
-V
¦V
¦V







25
UKM0-HadGEM1
U.K.
¦V
¦V

¦V





V
V
The GISS Model . There have been many different configurations of aerosol
simulations in the GISS model over the years, with different emissions, physics
July 2009
9-180
DRAFT-DO NOT CITE OR QUOTE

-------
packages, etc., as is apparent from the multiple GISS entries in the preceding figures
and tables. There were also three different GISS GCM submissions to IPCC AR4,
which varied in their model physics and ocean formulation.(Note that the aerosols in
these three GISS versions are different from those in the AeroCom simulations
described in Sections 9.3.6.2 and 9.3.6.3.) The GCM results discussed below all relate
to the simulations known as GISS model ER (Schmidt and et, 2006, 190373) (see
Table 9" 16). Although the detailed description and model evaluation have been
presented in Liu et al. (2006, 190422). below are the general characteristics of aerosols
in the GISS ER:
Aerosol lields-"Yhe aerosol fields used in the GISS ER is a prescribed
"climatology" which is obtained from chemistry transport model simulations with
monthly averaged mass concentrations representing conditions up to 1990. Aerosol
species included are sulfate, nitrate, BC, POM, dust, and sea salt. Dry size effective
radii are specified for each of the aerosol types, and laboratory-measured phase
functions are employed for all solar and thermal wavelengths. For hygroscopic
aerosols (sulfate, nitrate, POM, and sea salt), formulas are used for the particle
growth of each aerosol as a function of relative humidity, including the change in
density and optical parameters. With these specifications, the AOD, single scattering
albedo, and phase function of the various aerosols are calculated. While the aerosol
distribution is prescribed as monthly mean values, the relative humidity component of
the extinction is updated each hour. The global averaged AOD at 550 nm is about
0.15.
Global distribution-'When comparing with AOD from observations by multiple
satellite sensors of MODIS, MISR, POLDER, and AVHRR and surface based
sunphotometer network AERONET (see Chapter 2 of the CCSP SAP2.3 for detailed
information about data), qualitative agreement is apparent, with generally higher
burdens in Northern Hemisphere summer, and seasonal variations of smoke over
southern Africa and South America, as well as wind blown dust over northern African
and the Persian Gulf. Aerosol optical depth in both model and observations is smaller
away from land. There are, however, considerable discrepancies between the model
and observations. Overall, the GISS GCM has reduced aerosol optical depths
compared with the satellite data (a global, clear-sky average of about 80% compared
with MODIS and MISR data), although it is in better agreement with AERONET
ground-based measurements in some locations (note that the input aerosol values
were calibrated with AERONET data). The model values over the Sahel in Northern
Hemisphere winter and the Amazon in Southern Hemisphere winter are excessive,
indicative of errors in the biomass burning distributions, at least partially associated
with an older biomass burning source used (the source used here was from (Liousse et
al., 1996, 078158)).
Seasonal variation-' A comparison of the seasonal distribution of the globalAOD
between the GISS model and satellite data indicates that the model seasonal
variation is in qualitative agreement with observations for many of the locations that
represent major aerosol regimes, although there are noticeable differences. For
example, in some locations the seasonal variations are different from or even opposite
to the observations.
Particle size parameter' The Angstrom exponent Of), which is determined by the
contrast between the AOD at two or more different wavelengths and is related to
aerosol particle size (discussed in Section 9.3.6.3). This parameter is important
because the particle size distribution affects the efficiency of scattering of both short
and long wave radiation, as discussed earlier. A from the GISS model is biased low
compared with AERONET, MODIS, and POLDER data, although there are technical
differences in determining the A. This low bias suggests that the aerosol particle size
in the GISS model is probably too large. The average effective radius in the GISS
model appears to be 0.3-0.4 ]im, whereas the observational data indicates a value
more in the range of 0.2-0.3 jim (Liu et al., 2006, 190422).
July 2009
9-181
DRAFT-DO NOT CITE OR QUOTE

-------
Single scattering albedo-The model-calculated SSA (at 550 nm) appears to be
generally higher than the AERONET data at worldwide locations (not enough
absorption), but lower than AERONET data in Northern Africa, the Persian Gulf, and
the Amazon (too much absorption). This discrepancy reflects the difficulties in
modeling BC, which is the dominant absorbing aerosol, and aerosol sizes. Global
averaged SSA at 550 nm from the GISS model is at about 0.95.
Aerosol direct RF- The GISS model calculated anthropogenic aerosol direct
shortwave RF is "0.56 W/m2 at TOAand "2.87 W/m2 at the surface. The TOAforcing
(upper left, Figure 9-77) indicates that, as expected, the model has larger negative
values in polluted regions and positive forcing at the highest latitudes. At the surface
(lower left, Figure 9-77) GISS model values exceed "4 W/m2 over large regions. Note
there is also a longwave RF of aerosols (right column), although they are much weaker
than the shortwave RF.
There are several concerns for climate change simulations related to the aerosol
trend in the GISS model. One is that the aerosol fields in the GISS AR4 climate
simulation (version ER) are kept fixed after 1990. In fact, the observed trend shows a
reduction in tropospheric aerosol optical thickness from 1990 through the present, at
least over the oceans (Mishchenko and Geogdzhayev, 2007, 190545). Hansen et al.
(2007, 190597) suggested that the deficient warming in the GISS model over Eurasia
post" 1990 was due to the lack of this trend. Indeed, a possible conclusion from the
Penner et al. (2002, 190562) study was that the GISS model overestimated the AOD
(presumably associated with anthropogenic aerosols) poleward of 30°N. However,
when an alternate experiment reduced the aerosol optical depths, the polar warming
became excessive (Hansen et al., 2007, 190597). The other concern is that the GISS
model may underestimate the organic and sea salt AOD, and overestimate the
influence of black carbon aerosols in the biomass burning regions (deduced from
Penner et al., 2002, 190562; Liu et al., 2006, 190422). To the extent that is true, it
would indicate the GISS model underestimates the aerosol direct cooling effect in a
substantial portion of the tropics, outside of biomass burning areas. Clarifying those
issues requires numerous modeling experiments and various types of observations.
The GFDL Model . A comprehensive description and evaluation of the GFDL
aerosol simulation are given in Ginoux et al. (2006, 190582). Below are the general
characteristics'
Aerosol Gelds-"Yhe aerosols used in the GFDL climate experiments are obtained
from simulations performed with the MOZART 2 model (Model for Ozone and Related
chemical Tracers) (Horowitz et al., 2003, 057770! Horowitz, 2006, 190620). The
exceptions were dust, which was generated with a separate simulation of MOZART 2,
using sources from Ginoux et al. (2001, 190579) and wind fields from NCEP/NCAR
reanalysis data! and sea salt, whose monthly mean concentrations were obtained from
a previous study by Haywood et al. (1999, 040453). It includes most of the same
aerosol species as in the GISS model (although it does not include nitrates), and, as in
the GISS model, relates the dry aerosol to wet aerosol optical depth via the model's
relative humidity for sulfate (but not for organic carbon); for sea salt, a constant
relative humidity of 80% was used. Although the parameterizations come from
different sources, both models maintain a very large growth in sulfate particle size
when the relative humidity exceeds 90%.
Global distributions-' Overall, the GFDL global mean aerosol mass loading is
within 30% of that of other studies (Chin et al., 2002, 189996; Tie and et, 2005,
190459; Reddy et al., 2005, 190207). except for sea salt, which is 2 to 5 times smaller.
However, the sulfate AOD (0.1) is 2.5 times that of other studies, whereas the organic
carbon value is considerably smaller (on the order of 1/2). Both of these differences are
influenced by the relationship with relative humidity. In the GFDL model, sulfate is
allowed to grow up to 100% relative humidity, but organic carbon does not increase in
size as relative humidity increases. Comparison of AOD with AVHRR and MODIS
data for the time period 1996-2000 shows that the global mean value over the ocean
(0.15) agrees with AVHRR data (0.14) but there are significant differences regionally,
July 2009
9-182
DRAFT-DO NOT CITE OR QUOTE

-------
with the model overestimating the value in the northern mid latitude oceans and
underestimating it in the southern ocean. Comparison with MODIS also shows good
agreement globally (0.1-5.), but in this case indicates large disagreements over land,
with the model producing excessive A1 >1) over industrialized countries and
underestimating the effect over Worn ass burning regions. Overall, the global averaged
AOD at 550 nm is 0.17, which is higher than the maximum values in the AeroConvA
experiments (Table 9-12/ and exceeds the observed value too (Ae and S* in Figure
9-71).
net thermal BOA
ml_.ji_.l_j. i i —	.. i i i —
A -J -2 -I 0 \ 1 3 4 5 -5 -A -J -.2 -J 0 .1 3 5.7 I.I
Source: Figure provided by A. Lacis, GISS.
Figure 9-77. Direct radiative farcing by anthropogenic aerosols in the GISS model (including sulfates,
BC, 0C and nitrates). Short wave forcing at TOA and surface are shown in the top left
and bottom left panels. The corresponding thermal forcing is indicated in the right hand
panels.
Composition¦ Comparison of GPDL modeled species with in situ data over North
America, Europe, and over oceans has revealed that the sulfate is overestimated in
spring and summer and underestimated in winter in many regions, including Europe
and North America, Organic and black carbon aerosols are also overestimated in
polluted regions by a factor of two, whereas organic carbon aerosols are elsewhere
underestimated by factors of 2 to •">. Dust concentrations at the surface agree with
observations to within a factor of 2 in most places where significant dust exists,
although over the southwest U.S. it is a factor of 10 too large. Surface concentrations
of sea salt are underestimated by more than a factor of 2, Over the oceans, the
excessive sulfate AOD compensates for the low se.a salt values except in the southern
oceans. Size and single -scattering albedo* No specific comparison was given for
July 2009
9-183
DRAFT-DO NOT CITE OR QUOTE

-------
particle size or single-scattering albedo, but the excessive sulfate would likely produce
too high a value of reflectivity relative to absorption except in some polluted regions
where black carbon (an absorbing aerosol) is also overestimated.
As in the case of the GISS model, there are several concerns with the GFDL
model. The good global-average agreement masks an excessive aerosol loading over
the Northern Hemisphere (in particular, over the northeast U.S. and Europe) and an
underestimate over biomass burning regions and the southern oceans. Several model
improvements are needed, including better parameterization of hygroscopic growth at
high relative humidity for sulfate and organic carbon! better sea salt simulations!
correcting an error in extinction coefficients; and improved biomass burning emissions
inventory (Ginoux et al., 2006, 190582).
Comparisons between GISS and GFDL Model. Both giss and gfdl
models were used in the IPCC AR.4 climate simulations for climate sensitivity that
included aerosol forcing. It would be constructive, therefore, to compare the
similarities and differences of aerosols in these two models and to understand what
their impacts are in climate change simulations. Figure 9"78 shows the percentage
AOD from different aerosol components in the two models.
Sulfate-"The sulfate AOD from the GISS model is within the range of that from all
other models (Table 9"13), but that from the GFDL model exceeds the maximum value
by a factor of 2.5. An assessment in SAP 3.2 (CCSP, 2008, 192028; Shin dell et al.,
2008, 190393) also concludes that GFDL had excessive sulfate AOD compared with
other models. The sulfate AOD from GFDL is nearly a factor of 4 large than that from
GISS, although the sulfate burden differs only by about 50% between the two models.
Clearly, this implies a large difference in sulfate MEE between the two models.
BC and POM-' Compared to observations, the GISS model appears to overestimate
the influence of BC and POM in the biomass burning regions and underestimate it
elsewhere, whereas the GFDL model is somewhat the reverse- it overestimates it in
polluted regions, and underestimates it in biomass burning areas. The global
comparison shown in Table 9"14 indicates the GISS model has values similar to those
from other models, which might be the result of such compensating errors. The GISS
and GFDL models have relatively similar global-average black carbon contributions,
and the same appears true for POM.
Sea salt-"Yhe GISS model has a much larger sea salt contribution than does
GFDL (or indeed other models).
Global and regional distributions-' Overall, the global averaged AOD is 0.15 from
the GISS model and 0.17 from GFDL. However, as shown in Figure 9"78, the
contribution to this AOD from different aerosol components shows greater disparity.
For example, over the Southern Ocean where the primary influence is due to sea salt
in the GISS model, but in the GFDL it is sulfate. The lack of satellite observations of
the component contributions and the limited available in situ measurements make
the model improvements at aerosol composition level difficult.
Climate simulations-' With such large differences in aerosol composition and
distribution between the GISS and GFDL models, one might expect that the model
simulated surface temperature might be quite different. Indeed, the GFDL model was
able to reproduce the observed temperature change during the 20th century without
the use of an indirect aerosol effect, whereas the GISS model required a substantial
indirect aerosol contribution (more than half of the total aerosol forcing; (Hansen et
al., 2007, 190597). It is likely that the reason for this difference was the excessive
direct effect in the GFDL model caused by its overestimation of the sulfate optical
depth. The GISS model direct aerosol effect (see Section 9.3.6.6) is close to that
derived from observations (Chapter 2 of the CCSP SAP2.3); this suggests that for
models with climate sensitivity close to 0.75°C/(W/m2) (as in the GISS and GFDL
models), an indirect effect is needed.
July 2009
9-184
DRAFT-DO NOT CITE OR QUOTE

-------
Blank Cii'Wui
OrawulL Caption
Sea Salt
33.30
Figure 9-78. Percentage of aerosol optical depth in the GISS, left, based on Liu et al.
(2006,190422), provided by A. Lacis, GISS, and GFDL, right, from Ginoux et al. (2006,
190582). Models associated with the different components: S(h2~ (1st row), BC (2nd
row), 00 (3rd row), sea-salt (4th row), dust (5th row), and nitrate (last row). Nitrate not
available in GFDL model). Numbers on the GISS panels are global average, but on the
GFDL panels are maximum and minimum.
min-13 I Sulphate ADD Frartkm
ma* =95 9
rnln=OQO Sea Say AOD Fraction max^B.?
mirigt-S< OCjADD Fraction , rrtftx= 37.0
nun=0.9fc BC AOD Fraction max—19.6
rnwv»D 06 Dim AOD franrtpn. max BO,7
180 I2GW 6DW 0
July 2009
9-185
DRAFT-DO NOT CITE OR QUOTE

-------
Additional Considerations. Long wave aerosol forcing- So far only the aerosol
RF in the shortwave (solar) spectrum has been discussed. Figure 9"77 (right column)
shows that compared to the shortwave forcing, the values of anthropogenic aerosol
long wave (thermal) forcing in the GISS model are on the order of 10%. Like the
shortwave forcing, these values will also be affected by the particular aerosol
characteristics used in the simulation.
Aerosol vertical distribution-' Vertical distribution is particularly important for
absorbing aerosols, such as BC and dust in calculating the RF, particularly when
longwave forcing is considered (e.g., Figure 9-77) because the energy they reradiate
depends on the temperature (and hence altitude), which affects the calculated forcing
values. Several model inter-comparison studies have shown that the largest difference
among model simulated aerosol distributions is the vertical profile (e.g., Lohmann et
al., 2001, 190448; Penner et al., 2002, 190562; Textor et al., 2006, 190456)). due to the
significant diversities in atmospheric processes in the models (e.g., Table 9"12). In
addition, the vertical distribution also varies with space and time, as illustrated in
Figure 9"79 from the GISS ER simulations for January and July showing the most
probable altitude of aerosol vertical locations. In general, aerosols in the northern
hemisphere are located at lower altitudes in January than in July, and vice versa for
the southern hemisphere.
Mixing state-'Most climate model simulations incorporating different aerosol
types have been made using external mixtures, i.e., the evaluation of the aerosols and
their radiative properties are calculated separately for each aerosol type (assuming no
mixing between different components within individual particles). Observations
indicate that aerosols commonly consist of internally mixed particles, and these
"internal mixtures" can have very different radiative impacts. For example, the GISS"
1 (internal mixture) and GISS-2 (external mixture) model results shows very different
magnitude and sign of aerosol forcing from slightly positive (implying slight warming)
to strong negative (implying significant cooling) TOA forcing (Figure 9-72), due to
changes in both radiative properties of the mixtures, and in aerosol amount. The more
sophisticated aerosol mixtures from detailed microphysics calculations now being
used/developed by different modeling groups may well end up producing very different
direct (and indirect) forcing values.
Cloudy sky vs. clear sty-'The satellite or AERONET observations are all for clear
sky only because aerosol cannot be measured in the remote sensing technique when
clouds are present. However, almost all the model results are for all-sky because of
difficulty in extracting cloud-free scenes from the GCMs. So the AOD comparisons
discussed earlier are not completely consistent. Because AOD can be significantly
amplified when relative humidity is high, such as near or inside clouds, all-sky AOD
values are expected to be higher than clear sky AOD values. On the other hand, the
aerosol RF at TOA is significantly lower for all-sky than for clear sky conditions; the
IPCC AR4 and AeroCom RF study (Schulz et al., 2006, 190381) have shown that on
average the aerosol RF value for all-sky is about 1/3 of that for clear sky although
with large diversity (63%). These aspects illustrate the complexity of the system and
the difficulty of representing aerosol radiative influences in climate models whose
cloud and aerosol distributions are somewhat problematic. And of course aerosols in
cloudy regions can affect the clouds themselves, as discussed in Section 9.3.6.4.
July 2009
9-186
DRAFT-DO NOT CITE OR QUOTE

-------
m i 	;	 i i
100 360 420 480 540 400 660 720 780 540 900
Source: A. Lacis, GISS.
Figure 9-79. Most probable aerosol altitude (in pressure, hPa) from the GISS model in January (top)
and July (bottom).
9.3.6.6. Impacts of Aerosols on Climate Model Simulations
Surface Temperature Change. It was noted in the introduction that aerosol
cooling is essential in order for models to produce the observed global"temperature
rise over the last century, at least models with climate sensitivities in the range of 3°C
for doubled C02 (or ~0.759C;/(W/m2)) . The implications of this are discussed here in
somewhat more detail. Hansen et Si, (2007, 190597) show that in the CSlSS model,
well-mixed greenhouse gases produce a warming of close to 1°C between 1880 and the
present (Table !)-1 7). The direct effect of tropospheric aerosols as calculated in that
model produces cooling of close to "0.3°C between those same years, while the indirect
effect (represented in that study as cloud cover change) produces an additional cooling
of similar magnitude (note that the general model result quoted in IPOG AR4 is that
the indirect RF is twice that of the direct effect).
July 2009
9-187	DRAFT-DO NOT CITE OR QUOTE

-------
Table 9-17. Climate forcings (1880-2003) used to drive GISS climate simulations, along with the
surface air temperature changes obtained for several periods.
Forcing Agent

Forcing w M-2 (1880-2003)


AT Surface °C (yr to 2003)


Fi
Fa
Fs
Fe
1880
1900
1950
1979
Well-mixed GHGs
2.62
2.50
2.65
2.72
0.96
0.93
0.74
0.43
Stratospheric H2O


0.06
0.05
0.03
0.01
0.05
0.00
0s
0.44
0.28
0.26
0.23
0.p08
0.05
0.00
¦0.01
Land use


¦0.09
¦0.09
¦0.05
¦0.p07
¦0.04
¦0.02
Snow albedo
0.05
0.05
0.14
0.14
0.03
0.00
0.02
¦0.01
Solar irradiance
0.23
0.24
0.23
0.22
0.07
0.07
0.01
0.02
Stratospheric aerosols
0.00
0.00
0.00
0.00
¦0.08
¦0.03
¦0.06
0.04
Trop. aerosol direct forcing
¦0.41
¦0.38
¦0.52
¦0.60
¦0.28
¦0.23
¦0.18
¦0.10
Trop. aerosol indirect forcing


¦0.87
¦0.77
¦0.27
¦0.29
¦0.14
¦0.05
Sum of above


1.86
1.90
0.49
0.44
0.40
0.30
All forcings at once


1.77
1.75
0.53
0.61
0.44
0.29
Source: Hansen et al. (2007,190597). Instantaneous (/I), adjusted (fa), fixed SST (fs) and effective (fe) forcings are defined in Hansen et al. (2005,059087).
The time dependence of the total aerosol forcing used as well as the individual
species components is shown in Figure 9"80. The resultant warming, 0.53 (± 0.04) °C
including these and other forcings (Table 9-17), is less than the observed value of
0.6-0.7°C from 1880-2003. Hansen et al. (2007, 190597) further show that a reduction
in sulfate optical thickness and the direct aerosol effect by 50%, which also reduced
the aerosol indirect effect by 18%, produces a negative aerosol forcing from 1880-2003
of "0.91 W/m2 (down from "1.37 W/m2 with this revised forcing). The model now warms
0.75°C over that time. Hansen et al. (2007, 190597) defend this change by noting that
sulfate aerosol removal over North America and western Europe during the 1990s led
to a cleaner atmosphere. Note that the comparisons shown in the previous
section suggest that the GISS model already underestimates aerosol optical depths! it
is thus trends that are the issue here.
The magnitude of the indirect effect used by Hansen et al. (2005, 190596) is
roughly calibrated to reproduce the observed change in diurnal temperature cycle and
is consistent with some satellite observations. However, as Anderson et al. (2003,
054820) note, the forward calculation of aerosol negative forcing covers a much larger
range than is normally used in GCMsi the values chosen, as in this case, are
consistent with the inverse reasoning estimates of what is needed to produce the
observed warming, and hence generally consistent with current model climate
sensitivities. The authors justify this approach by claiming that paleoclimate data
indicate a climate sensitivity of close to 0.75 (± 0.25)°C/(W/m2), and therefore
something close to this magnitude of negative forcing is reasonable. Even this stated
range leaves significant uncertainty in climate sensitivity and the magnitude of the
aerosol negative forcing. Furthermore, IPCC (2007, 189430) concluded that
paleoclimate data are not capable of narrowing the range of climate sensitivity,
nominally 0.375-1.13 °C/(W/m2), because of uncertainties in paleoclimate forcing and
response! so from this perspective the total aerosol forcing is even less constrained
than the GISS estimate. Hansen et al. (2007, 190597) acknowledge that "an equally
good match to observations probably could be obtained from a model with larger
July 2009
9-188
DRAFT-DO NOT CITE OR QUOTE

-------
sensitivity and smaller net forcing, or a model with smaller sensitivity and larger
forcing".
NiiTrnifS
Sulfa: pi
- RC Industrial
UC tJiomiis
OC InduarnaJ
OC fluoriasi
NiMBS
S«j Ifirc-s.
i«?4 1590 IKS (910
Sum ejf AJjiivn Anraxrjk
- Tbtt I
Sua Salt
Soili'Dua
—	- BC laduacml
-	- - SC. Btamaw
- OC Industrial
OC &OT1IS

Slack Carbon forcings
B.eil«set»vie Aerosol FortKvgs
Ali Aerosat?
Source: Hansen et al. (2007,190597).
Figure 9-80. Time dependence of aerosol optical thickness (left) and climate forcing (right). Note that
as specified, the aerosol trends are all "flat" from 1990-2000.
The GFDL model results for global mean ocean temperature change (down to
3 km depth) for the time period 186Q-20D0 is shown in Figure 9"81, along with the
different contributing factors (Delworth et al., 2005, 190055). This is the same GFDL
model whose aerosol distribution was discussed previously. The aerosol forcing
produces a cooling on the order of 50% that of greenhouse warming (generally similar
to that calculated by the GISS model, Table 9-17). Note that this was achieved
without any aerosol indirect effect .
The general model response noted by IPCO, as discussed in the introduction, was
that the total aerosol forcing of -1,3 W/m2 reduced the greenhouse forcing of near 3
W/m2 by about 45%, in the neighborhood of the GFDL and GISS forcings. Since the
average model sensitivity was close to 0.75 °C/(W/m2), similar to the sensitivities of
these models, the necessary negative forcing is therefore similar. The agreement
cannot therefore be used to validate the actual aerosol effect until climate sensitivity
itself is better known.
Is I here some way to distinguish between greenhouse gas and aerosol forcing that
would allow the observational record to indicate how much of each was really
occurring? This question of attribution has been the subject of numerous papers, and
the full scope of the discussion is beyond the range of this (OCSP SAP2.3) report. It
might be briefly noted that Zhang et al. (2006, 157722) using results from Several
climate models and incluchng both spatial and temporal patterns, found that the
climate responses to greenhouse gases and sulfate aerosols are correlated, and
separation is possible only occasionally, especially at global scales. This conclusion
appears to be both model and method"dependent^ using time-space distinctions as
opposed to trend detection may work differently in different models (Gillett et al.,
July 2009
9-189
DRAFT - DO NOT CITE OR QUOTE

-------
2002, 190576). Using multiple models helps primarily by providing larger-ensemble
sizes for statistics (Gillett at al., 2002. 190578). However, even separating between the
effects of different aerosol types is difficult, Jones et al. (2005, 155885) concluded that
currently the pattern of temperature change due to black carbon is indistinguishable
from the sulfate aerosol pattern. In contrast, Hansen et al, (2005, 059087) found that
absorbing aerosols produce a different global response than other forcings, and so may
be distinguishable , Overall, the similarity in response to all these very different
forcings is undoubtedly due to the importance of climate feedbacks in amplifying the
forcing, whatever its nature.
¦
fi£L lilHd	IBSiti 1UL IrMJ «.J
Source: Delworth et al. (2005,190055).
Figure 9-81. Change in global mean ocean temperature (left axis) and ocean heat content (right axis)
for the top 3000 m due to different forcings in the GFDL model. WMGG includes all
greenhouse gases and ozone; NATURAL includes solar and volcanic aerosols (events
shown as green triangles on the bottom axis). Observed ocean heat content changes are
shown as well.
Distinctions in the climate response do appear to arise in the vertical, where
absorbing aerosols produce warming that is exhibited throughout the troposphere and
into the stratosphere, whereas reflective aerosols cool the troposphere but warm the
stratosphere (Hansen et al., 2005, 0590871 IPOfJ (intergovernmental Panel on
Climate, 2007, 190988). Delworth et al. (2005, 190055) noted I hat in the ocean, the
cooling effect of aerosols extended to greater depths, due to the thermal instability
associated with cooling the ocean surface, Hen.ce. the temperature response at.le.yels
both above and below the surface may provide an additional constraint on the
magnitudes of each of these forcings, as may the difference between Northern and
Southern Hemisphere changes UJ'< V et al., 2007, 189430. Chapter 9). The profile of
atmospheric temperature response will be useful to the extent that the vertical profile
of aerosol absorption, an important parameter to measure, is known.
Implications for Climate Model Simulations. The comparisons in Sections
9:3.6.2 and 9.3.6.3 suggest, that there are large differences in model calculated aerosol
distributions, mainly because of the large uncertainties in modeling the aerosol
atmospheric processes in addition to the uncertainties in emissions. The, fact that the
total optical depth is in better agreement between models than the individual
components means that even with similar optical depths, the aerosol direct forcing
effect can be quite different, as shown in the Aeroi \>m studies. Because the diversity
July 2009
9-190
DRAFT-DO NOT CITE OR QUOTE

-------
among models and discrepancy between models and observations are much larger at
the regional level than in global average, the assessment of climate response (e.g.,
surface temperature change) to aerosol forcing would be more accurate for global
average than for regional or hemispheric differentiation. However, since aerosol
forcing is much more pronounced on regional than on global scales because of the
highly variable aerosol distributions, it is insufficient or even misleading to just get
the global average right. The indirect effect is strongly influenced by the aerosol
concentrations, size, type, mixing state, microphysical processes, and vertical profile.
As shown in previous sections, very large differences exist in those quantities even
among the models having similar AOD. Moreover, modeling aerosol indirect forcing
presents more challenges than direct forcing because there is so far no rigorous
observational data, especially on a global scale, that one can use to test the model
simulations. As seen in the comparisons of the GISS and GFDL model climate
simulations for IPCC AR4, aerosol indirect forcing was so poorly constrained that it
was completely ignored by one model (GFDL) but used by another (GISS) at a
magnitude that is more than half of the direct forcing, in order to reproduce the
observed surface temperature trends. A majority of the climate models used in IPCC
AR4 do not consider indirect effects; the ones that did were mostly limited to highly
simplified sulfate indirect effects (Table 9"16). Improvements must be made to at least
the degree that the aerosol indirect forcing can no longer be used to mask the
deficiencies in estimating the climate response to greenhouse gas and aerosol direct
RF.
9.3.6.7. Outstanding Issues
Clearly there are still large gaps in assessing the aerosol impacts on climate
through modeling. Major outstanding issues and prospects of improving model
simulations are discussed below.
Aerosol composition. Many global models are now able to simulate major
aerosol types such as sulfate, black carbon, and POM, dust, and sea salt, but only a
small fraction of these models simulate nitrate aerosols or consider anthropogenic
secondary organic aerosols. And it is difficult to quantify the dust emission from
human activities. As a result, the IPCC AR4 estimation of the nitrate and
anthropogenic dust TOA forcing was left with very large uncertainty. The next
generation of global models should therefore have a more comprehensive suite of
aerosol compositions with better-constrained anthropogenic sources.
Aerosol absorption. One of the most critical parameters in aerosol direct RF
and aerosol impact on hydrological cycles is the aerosol absorption. Most of the
absorption is from BC despite its small contribution to total aerosol short and long-
wave spectral ranges, whereas POM absorbs in the near UV. The aerosol absorption or
SSA, will have to be much better represented in the models through improving the
estimates of carbonaceous and dust aerosol sources, their atmospheric distributions,
and optical properties.
Aerosol indirect effects. The activation of aerosol particles into CCN depends
not only on particle size but chemical composition, with the relative importance of size
and composition unclear. In current aerosol-climate modeling, aerosol size distribution
is generally prescribed and simulations of aerosol composition have large
uncertainties. Therefore the model estimated "albedo effect" has large uncertainties.
How aerosol would influence cloud lifetime/ cover is still in debate. The influence of
aerosols on other aspects of the climate system, such as precipitation, is even more
uncertain, as are the physical processes involved. Processes that determine aerosol
size distributions, hygroscopic growth, mixing state, as well as CCN concentrations,
however, are inadequately represented in most of the global models. It will also be
difficult to improve the estimate of indirect effects until the models can produce more
realistic cloud characteristics.
July 2009
9-191
DRAFT-DO NOT CITE OR QUOTE

-------
Aerosol impacts on surface radiation and atmospheric heating.
Although these effects are well acknowledged to play roles in modulating atmospheric
circulation and water cycle, few coherent or comprehensive modeling studies have
focused on them, as compared to the efforts that have gone to assessing aerosol RF at
TOA. They have not yet been addressed in the previous IPCC reports. Here, of
particular importance is to improve the accuracy of aerosol absorption.
Long-term trends of aerosol. To assess the aerosol effects on climate change
the long-term variations of aerosol amount and composition and how they are related
to the emission trends in different regions have to be specified. Simulations of
historical aerosol trends can be problematic since historical emissions of aerosols have
shown large uncertainties—as information is difficult to obtain on past source types,
strengths, and even locations. The IPCC AR4 simulations used several alternative
aerosol emission histories, especially for BC and POM aerosols.
Climate modeling. Current aerosol simulation capabilities from CTMs have not
been fully implemented in most models used in IPCC AR4 climate simulations.
Instead, a majority employed simplified approaches to account for aerosol effects, to
the extent that aerosol representations in the GCMs, and the resulting forcing
estimates, are inadequate. The oversimplification occurs in part because the modeling
complexity and computing resource would be significantly increased if the full suite of
aerosols were fully coupled in the climate models.
Observational constraints. Model improvement has been hindered by a lack
of comprehensive datasets that could provide multiple constraints for the key
parameters simulated in the model. The extensive AOD coverage from satellite
observations and AERONET measurements has helped a great deal in validating
model-simulated AOD over the past decade, but further progress has been slow. Large
model diversities in aerosol composition, size, vertical distribution, and mixing state
are difficult to constrain, because of lack of reliable measurements with adequate
spatial and temporal coverage (see Chapter 2 of the CCSP SAP2.3).
Aerosol radiative forcing. Because of the large spatial and temporal
differences in aerosol sources, types, emission trends, compositions, and atmospheric
concentrations, anthropogenic aerosol RF has profound regional and seasonal
variations. So it is an insufficient measure of aerosol RF scientific understanding,
however useful, for models (or observation-derived products) to converge only on
globally and annually averaged TOARF values and accuracy. More emphasis should
be placed on regional and seasonal comparisons, and on climate effects in addition to
direct RF at TOA.
9.3.6.8. Conclusions
From forward modeling studies, as discussed in the IPCC (2007, 190988) , the
direct effect of aerosols since pre "industrial times has resulted in a negative RF of
about "0.5 ± 0.4 W/m2. The RF due to cloud albedo or brightness effect is estimated to
be "0.7 (-1.8 to -0.3) W/m2. Forcing of similar magnitude has been used in some
modeling studies for the effect associated with cloud lifetime, in lieu of the cloud
brightness influence. The total negative RF due to aerosols according to IPCC (2007,
190988) estimates is therefore -1.3 ("2.2 to -0.5) W/m2. With the inverse approach, in
which aerosols provide forcing necessary to produce the observed temperature change,
values range from -1.7 to "0.4 W/m2 (Intergovernmental Panel on Climate, 2007,
190988) . These results represent a substantial advance over previous assessments
(e.g., IPCC TAR), as the forward model estimated and inverse approach required
aerosol TOA forcing values are converging. However, large uncertainty ranges
preclude using the forcing and temperature records to more accurately determine
climate sensitivity.
There are now a few dozen models that simulate a comprehensive suite of
aerosols. This is done primarily in the CTMs. Model inter-comparison studies have
July 2009
9-192
DRAFT-DO NOT CITE OR QUOTE

-------
shown that models have merged at matching the global annual averaged AOD
observed by satellite instruments, but they differ greatly in the relative amount of
individual components, in vertical distributions, and in optical properties. Because of
the great spatial and temporal variations of aerosol distributions, regional and
seasonal diversities are much larger than that of the global annual mean. Different
emissions and differences in atmospheric processes, such as transport, removal,
chemistry, and aerosol microphysics, are chiefly responsible for the spread among the
models. The varying component contributions then lead to differences in aerosol direct
RF, as aerosol scattering and absorption properties depend on aerosol size and type.
They also impact the calculated indirect RF, whose variations are further amplified by
the wide range of cloud and convective parameterizations in models. Currently, the
largest aerosol RF uncertainties are associated with the aerosol indirect effect. Most
climate models used for the IPCC AR4 simulations employed simplified approaches,
with aerosols specified from stand-alone CTM simulations. Despite the uncertainties
in aerosol RF and widely varying model climate sensitivity, the IPCC AR4 models
were generally able to reproduce the observed temperature record for the past
century. This is because models with lower/higher climate sensitivity generally used
less/more negative aerosol forcing to offset the greenhouse gas warming. An equally
good match to observed surface temperature change in the past could be obtained
from a model with larger climate sensitivity and smaller net forcing, or a model with
smaller sensitivity and larger forcing (Hansen et al., 2007, 190597). Obviously, both
greenhouse gases and aerosol effects have to be much better quantified in future
assessments.
Progress in better quantifying aerosol impacts on climate will only be made when
the capabilities of both aerosol observations and representation of aerosol processes in
models are improved. The primary concerns and issues discussed in this chapter of
the CCSP SAP2.3 include-
¦	Better representation of aerosol composition and absorption in the global models
¦	Improved theoretical understanding of subgrid-scale processes crucial to aerosol-cloud
interactions and lifetime
¦	Improved aerosol microphysics and cloud parameterizations
¦	Better understanding of aerosol effects on surface radiation and hydrological cycles
¦	More focused analysis on regional and seasonal variations of aerosols
¦	More reliable simulations of aerosol historic long-term trends
¦	More sophisticated climate model simulations with coupled aerosol and cloud processes
¦	Enhanced satellite observations of aerosol type, SSA, vertical distributions, and aerosol
radiative effect at TO A; more coordinated field experiments to provide constraints on
aerosol chemical, physical, and optical properties. Progress in better quantifying
aerosol impacts on climate will only be made when the capabilities of both aerosol
observations and representation of aerosol processes in models are improved.
9.3.7. Fire as a Special Source of PM Welfare Effects
Much interest has developed in defining more precisely the role of pyrogenic C in the
boreal C cycle. This is due to: (l) the resistance of pyrogenic C to decomposition! (2) its
influence on soil processes! and (3) the absorption of solar radiation by soot aerosols
(Preston and Schmidt, 2006, 156030).
Preston and Schmidt (2006, 156030) reviewed the current state of knowledge
regarding atmospheric emissions of pyrogenic C in the boreal zone. They considered
chemical structures, analytical methods, formation, characteristics in soil, loss
mechanisms, and longevity. Biomass is largely converted to gaseous forms during burning,
July 2009
9-193
DRAFT-DO NOT CITE OR QUOTE

-------
but up to several percent is converted to pyrogenic C, and this includes charcoal and BC.
Charcoal is defined visually! BC is defined chemically by its resistance to oxidation in the
laboratory.
Andreae and Gelencser (2006, 156215) reviewed a different category of
light-absorbing carbon, referred to as brown carbon. Operational methods used to measure
all light-absorbing carbon are estimated to have a factor of 2 at present.
Within the boreal zone, fire is a critical driver of ecosystem process and nutrient
cycling (Hicke et al., 2003, 156545). For example, Bachelet et al. (2005, 156241) estimated
that 61% of the C gained in Alaska by primary production of boreal forests between 1922
and 1996 was lost to fire.
An updated modeling effort to evaluate the radiative effects of aerosols was presented
by Stier et al. (2007, 157012). Inclusion of refractive indices recommended by (2005,
155696) significantly increased aerosol RF and resulted in better agreement with
sun-photometer estimates. Although this stage of climate modeling improved the
representation of aerosols, large uncertainties remain regarding the effects of aerosol
mixing and aerosol-cloud interactions. Furthermore, Stier et al. (2007, 157012) emphasized
that these types of modeling efforts are dependent upon emission estimates that are likely
to vary by a factor of 2 or more.
One important reason for the acknowledged uncertainty in estimating global
emissions of carbonaceous aerosols is the influence of intermittent fires that can occur at
scales large enough to affect hemispheric aerosol concentrations. To better quantify the
effects of large-scale fire, Generoso et al. (2007, 155786) used satellite observations of boreal
fires in Russia in 2003 to evaluate the performance of a global chemistry and transport
model in simulating aerosol optical thickness, transport, and deposition. Emissions
estimates of BC and OC were adjusted in the model to better match satellite observations of
pollutant transport over the North Pacific. This resulted in an increase in optical thickness
and BC deposition by a factor of 2. The adjusted model estimated that the fires contributed
16-33% of the optical thickness and 40-56% of BC deposition north of 75° N in the spring
and summer of 2003.
Large fires also occurred over the Iberian Peninsula and Mediterranean coast during
2003. Ameso-scale atmospheric transport model was used with ground-based
measurements and satellite optical measurements to characterize the dispersion of emitted
smoke particles and quantify radiative effects across Europe (Hodzic et al., 2007, 156553).
July 2009
9-194
DRAFT-DO NOT CITE OR QUOTE

-------
The modeled wildfire emissions resulted in increases in PM2.5 concentrations from 20 to
200%. The increased aerosol concentration was estimated to increase radiative forcing by
10-35 W/m2 during the period of strong fire influence. Absorption of radiation by BC was
also estimated to decrease rates of photolysis by 30%. In this simulation, all particles were
assumed to be internally mixed, and secondary aerosol formation was not considered.
Meteorological conditions in Europe during the exceptionally hot summer of 2003 were
linked to enhanced photochemically derived pollutants, increased wild fires, and elevated
aerosol concentrations in an analysis by (Vautard et al., 2007, 106012).
In addition to incidental fires, routine biomass burning, usually associated with
agriculture in eastern Europe, also has been shown to contribute to hemispheric
concentrations of carbonaceous aerosols. In the spring of 2006, the most severe air pollution
levels in the Arctic to date were recorded (Stohl et al., 2006, 156100). Atmospheric transport
modeling coupled with satellite fire detection data identified biomass burning for
agriculture as the primary cause of the high pollution levels. Concentrations of PM2.5
peaked during the pollution episode at values of an order of magnitude greater than those
recorded prior to the episode. The increased transport of pollution into the Arctic during
2006 was attributed to weather conditions that delayed preparations for crop planting into
May. Weather patterns favorable for pollutant transport into the Arctic were related to
unusually warm weather in late April and May, when the majority of agricultural biomass
burning took place that year.
In the summer of 2004, 2.7 million ha were burned by wildfire in Alaska and 3.1
million ha were burned in Canada. Effects on atmospheric air quality were measured
throughout the Arctic, although the concentrations of particulates varied considerably.
Aerosol optical depths were also increased at all measurement stations, which indicated
that the fires were likely to have had a significant effect on the atmospheric radiation
budget for the Arctic (Stohl et al., 2006, 156100). At one site, a pronounced drop in albedo
was observed due presumably to high deposition of light absorbing particulates on the snow
surface by the North American fires in 2004.
Investigations of the effects of large fires on climate forcing have typically focused on
the absorptive effects of BC. However, these fires also release large amounts of CO2 and
CH4, as well as light scattering compounds such as OC, and can enhance cloud formation.
These fires also increase radiative surface absorption through BC deposition on snow and
ice, and alter surface albedo and ecosystem energy budgets within the burn perimeter.
July 2009
9-195
DRAFT-DO NOT CITE OR QUOTE

-------
Randerson et al. (2006, 156038) estimated the net climate forcing of greenhouse gases,
aerosols, BC deposition on snow and ice and changes in albedo for the year subsequent to a
fire and for 80 yr in the future in interior Alaska. The net effect of the fire in the first year
was an increase in radiative forcing, but over the 80-yr recovery period, average net annual
radiative forcing was decreased by the fire.
9.3.8. Radiative Effects of Volcanic Aerosols
Section 9.3.8.1. comes directly from IPCC AR4 Chapter 2, Section 2.7.2, with section,
table, and figure numbers changed to be internally consistent with this ISA.
9.3.8.1. Explosive Volcanic Activity
Radiative Effects of Volcanic Aerosols. Volcanic sulfate aerosols are formed
as a result of oxidation of the sulfur gases emitted by explosive volcanic eruptions into
the stratosphere. The process of gas-to-particle conversion has an e-folding time of
roughly 35 days (Bluth et al., 1992, 192029; Read et al., 1993, 192031). The e-folding
time (by mass) for sedimentation of sulfate aerosols is typically about 12 to 14 months
(Baran and Foot, 1994, 192032; Barnes and Hofmann, 1997, 192044; Bluth et al.,
1997, 192045; Lambert et al., 1993, 192231). Also emitted directly during an eruption
are volcanic ash particulates (siliceous material). These are particles usually larger
than 2 jim that sediment out of the stratosphere fairly rapidly due to gravity (within
three months or so), but could also play a role in the radiative perturbations in the
immediate aftermath of an eruption. Stratospheric aerosol data incorporated for
climate change simulations tends to be mostly that of the sulfates (Ammann et al.,
2003, 192057; Hansen et al., 2002, 049177; Ramachandran et al., 2000, 192050; Sato
et al., 1993, 192046; Stenchikov et al., 1998, 192049; Tett et al., 2002, 192053). As
noted in the Second Assessment Report (SAR) and the TAR, explosive volcanic events
are episodic, but the stratospheric aerosols resulting from them yield substantial
transitory perturbations to the radiative energy balance of the planet, with both
shortwave and longwave effects sensitive to the microphysical characteristics of the
aerosols (e.g., size distribution).
Long-term ground-based and balloon-borne instrumental observations have
resulted in an understanding of the optical effects and microphysical evolution of
volcanic aerosols (Deshler et al., 2003, 192058; Hofmann et al., 2003, 192062).
Important groundbased observations of aerosol characteristics from pre-satellite era
spectral extinction measurements have been analysed by Stothers (2001, 192233;
2001, 192232). but they do not provide global coverage. Global observations of
stratospheric aerosol over the last 25 years have been possible owing to a number of
satellite platforms, for example, TOMS and TOVS have been used to estimate SO2
loadings from volcanic eruptions (Krueger et al., 2000, 192234; Prata et al., 2003,
192235). The Stratospheric Aerosol and Gas Experiment (SAGE) and Stratospheric
Aerosol Measurement (SAM) projects (e.g., McCormick and Trepte, 1987, 192328)
have provided vertically resolved stratospheric aerosol spectral extinction data for
over 20 years, the longest such record. This data set has significant gaps in coverage
at the time of the El Chichon eruption in 1982 (the second most important in the 20th
century after Mt. Pinatubo in 1991) and when the aerosol cloud is dense; these gaps
have been partially filled by lidar measurements and field campaigns (e.g., Antuna et
al., 2003, 192251; Thomason and Peter, 2006, 192248).
July 2009
9-196
DRAFT-DO NOT CITE OR QUOTE

-------
Volcanic aerosols transported in the atmosphere to polar regions are preserved in
the ice sheets, thus recording the history of the Earth's volcanism for thousands of
years (Kruysse, 1971, 192236; Mosley-Thompson et al., 2003, 192255; Palmer et al.,
2002, 192319). However, the atmospheric loadings obtained from ice records suffer
from uncertainties due to imprecise knowledge of the latitudinal distribution of the
aerosols, depositional noise that can affect the signal for an individual eruption in a
single ice core, and poor constraints on aerosol microphysical properties. The best-
documented explosive volcanic event to date, by way of reliable and accurate
observations, is the 1991 eruption of Mt. Pinatubo. The growth and decay of aerosols
resulting from this eruption have provided a basis for modeling the RF due to
explosive volcanoes. There have been no explosive and climatically significant volcanic
events since Mt. Pinatubo. As pointed out in Ramaswamy et al. (2001, 156899).
stratospheric aerosol concentrations are now at the lowest concentrations since the
satellite era and global coverage began in about 1980. Altitude dependent
stratospheric optical observations at a few wavelengths, together with columnar
optical and physical measurements, have been used to construct the time-dependent
global field of stratospheric aerosol size distribution formed in the aftermath of
volcanic events. The wavelength "dependent stratospheric aerosol single-scattering
characteristics calculated for the solar and longwave spectrum are deployed in climate
models to account for the resulting radiative (shortwave plus longwave) perturbations.
Using available satellite- and ground-based observations, Hansen et al. (2002,
049177) constructed a volcanic aerosols data set for the 1850-1999 period (Sato et al.,
1993, 192046). This has yielded zonal mean vertically resolved aerosol optical depths
for visible wavelengths and column average effective radii. Stenchikov et al. (2006,
192260) introduced a slight variation to this data set, employing UARS observations
to modify the effective radii relative to Hansen et al. (2002, 049177). thus accounting
for variations with altitude. Ammann et al. (2003, 192057) developed a data set of
total aerosol optical depth for the period since 1890 that does not include the
Krakatau eruption. The data set is based on empirical estimates of atmospheric
loadings, which are then globally distributed using a simplified parameterization of
atmospheric transport, and employs a fixed aerosol effective radius (0.42 ]im) for
calculating optical properties. The above data sets have essentially provided the bases
for the volcanic aerosols implemented in virtually all of the models that have
performed the 20th-century climate integrations (Stenchikov et al., 2006, 192260).
Relative to Sato et al. (1993, 192046). the Ammann et al. (2003, 192057) estimate
yields a larger value of the optical depth, by 20 to 30% in the second part of the 20th
century, and by 50% for eruptions at the end of 19th and beginning of 20th century, for
example, the 1902 Santa Maria eruption (Figure 9"82).
The global mean RF calculated using the Sato et al. (1993, 192046) data yields a
peak in radiative perturbation of about "3 W/m2 for the strong (rated in terms of
emitted SO2) 1860 and 1991 eruptions of Krakatau andMt. Pinatubo, respectively.
The value is reduced to about "2 W/m2 for the relatively less intense El Chichon and
Agung eruptions (Hansen et al., 2002, 049177). As expected from the arguments
above, Ammann's RF is roughly 20 to 30% larger than Sato's RF.
Not all features of the aerosols are well quantified, and extending and improving
the data sets remains an important area of research. This includes improved
estimates of the aerosol size parameters (Bingen et al., 2004, 192262). a new approach
for calculating aerosol optical characteristics using SAGE and UARS data (Bauman et
al., 2003, 192265). and intercomparison of data from different satellites and
combining them to fill gaps (Randall et al., 2001, 192268). While the aerosol
characteristics are better constrained for the Mt. Pinatubo eruption, and to some extent
for the El Chichon and Agung eruptions, the reliability degrades for aerosols from
explosive volcanic events further back in time as there are few, if any, observational
constraints on their optical depth and size evolution.
The radiative effects due to volcanic aerosols from major eruptions are manifest
in the global mean anomaly of reflected solar radiation; this variable affords a good
July 2009
9-197
DRAFT-DO NOT CITE OR QUOTE

-------
estimate of radiative effects that can actually be tested against observations.
However, unlike RF, this variable contains effects due to feedbacks (e.g., changes in
cloud distributions) so that it is actually more a signature of the climate response. In
the case of the Mt, Pinal ubo eruption, with apeak global visible optical depth of about
0,18, simulations yield a large negative perturbation as noted above of about -3 W/m2
(Hansen et al., 2002, 049177; Ramachandran et al.. 2:000, 192050) (see also Section 9.2
of the IPCO AR4), This modeled estimate of reflected solar radiation compares
reasonably with Kli'lJS observations (Minnis et al., 1993, 190539). However, the ERBS
observations were for a relatively short duration, and the model-observation
comparisons are likely affected by differing cloud effects in simulations and
measurements, It is interesting to note (Stenchikov et al. , 2006.. 192260) that, in the
Mt, Pinatubo case, the Goddard Institute for Space Studies (GISS) models that use
the Sato et al. (1993, 192046) data: yield an even greater solar reflection than the
National Qenter for Atmospheric Research (NCAR) model that uses the larger
(Ammann jgfe al., 2003, 192057) opt ical depth estimate.
0.2 <
0.18-
o.ia-
0.1
£ 0.12-
O.
<3
o M-
§
o"0.oe-
0.06-
0.04-
0.02-
0^
Figure 9-82. Visible (wavelength 0.55 Dm) optical depth estimates of stratospheric SO*2" aerosols
formed in the aftermath of explosive volcanic eruptions that occurred between 1860 and
2000. Results are shown from two different data sets that have been used in recent
climate model integrations. Note that the Ammann et al. (2003,192057) data begins in
1890.
Volcanic Aerosol Total Visible Optical Depth
Sato et al. (1993)
Ammann et al. (2003)
1870 1880 1800 1000 1010 1020 1030 1040 1090 1060 1970 1900 1W0
Time (years)
July 2009
9-198
DRAFT-DO NOT CITE OR QUOTE

-------
Thermal, Dynamical and Chemistry Perturbations Forced by Volcanic
Aerosols. Four distinct mechanisms have been invoked with regards to the climate
response to volcanic aerosol RF. First, these forcings can directly affect the Earth's
radiative balance and thus alter surface temperature. Second, they introduce
horizontal and vertical heating gradients; these can alter the stratospheric
circulation, in turn affecting the troposphere. Third, the forcings can interact with
internal climate system variability (e.g., El Nino-Southern Oscillation, North Atlantic
Oscillation, Quasi" Biennial Oscillation) and dynamical noise, thereby triggering,
amplifying or shifting these modes (see Section 9.2 of the IPCC AR4) (Yang and
Schlesinger, 2001, 192270; Stenchikov et al., 2004, 192274). Fourth, volcanic aerosols
provide surfaces for heterogeneous chemistry affecting global stratospheric ozone
distributions (Chipperfield et al., 2003, 192275) and perturbing other trace gases for a
considerable period following an eruption. Each of the above mechanisms has its own
spatial and temporal response pattern. In addition, the mechanisms could depend on
the background state of the climate system, and thus on other forcings (e.g., due to
well-mixe d gases) (Meehl et al., 2004, 192279). or interact with each other.
The complexity of radiative "dynamical response forced by volcanic impacts
suggests that it is important to calculate aerosol radiative effects interactively within
the model rather than prescribe them (Andronova et al., 1999, 192286; Broccoli et al.,
2003, 192283). Despite differences in volcanic aerosol parameters employed, models
computing the aerosol radiative effects interactively yield tropical and global mean
lower-stratospheric warmings that are fairly consistent with each other and with
observations (Ramachandran et al., 2000, 192050; Hansen et al., 2002, 049177; Yang
and Schlesinger, 2001, 192270; Stenchikov et al., 2004, 192274; Ramaswamy et al.,
2006, 192284); however, there is a considerable range in the responses in the polar
stratosphere and troposphere. The global mean warming of the lower stratosphere is
due mainly to aerosol effects in the longwave spectrum, in contrast to the flux changes
at the TOAthat are essentially due to aerosol effects in the solar spectrum. The net
radiative effects of volcanic aerosols on the thermal and hydrologic balance (e.g.,
surface temperature and moisture) have been highlighted by recent studies (Free and
Angell, 2002, 192281; Jones et al., 2003, 192278) (see Chapter 6 of the IPCC AR4). See
Chapter 9 (of the IPCC AR4) for significance of the simulated responses and model-
observation comparisons for 20th-century eruptions. A mechanism closely linked to
the optical depth perturbation and ensuing warming of the tropical lower stratosphere
is the potential change in the cross-tropopause water vapour flux (Joshi and Shine,
2003, 192327) (see Section 2.3.7 of the IPCC AR4).
Anomalies in the volcanic-aerosol induced global radiative heating distribution
can force significant changes in atmospheric circulation, for example, perturbing the
equator-to-pole heating gradient (Ramaswamy et al., 2006, 192273; Stenchikov et al.,
2002, 192277) (see Section 9.2 of the IPCC AR4) and forcing a positive phase of the
Arctic Oscillation that in turn causes a counterintuitive boreal winter warming at
middle and high latitudes over Eurasia and North America (Miller et al., 2005,
192258; Perlwitz and Graf, 2001, 192271; Perlwitz and Harnik, 2003, 192264; Rind et
al., 2005, 192261; Shin dell et al., 2003, 192069; Shindell et al., 2004, 192267;
Stenchikov et al., 2002, 192277; Stenchikov et al., 2004, 192274; Stenchikov et al.,
2006, 192260).
Stratospheric aerosols affect the chemistry and transport processes in the
stratosphere, resulting in the depletion of ozone (Brasseur and Granier, 1992, 192256;
Chipperfield et al., 2003, 192275; Solomon et al., 1996, 192252; Tie et al., 1994,
192253). Stenchikov et al. (2002, 192277) demonstrated a link between ozone
depletion and Arctic Oscillation response; this is essentially a secondary radiative
mechanism induced by volcanic aerosols through stratospheric chemistry.
Stratospheric cooling in the polar region associated with a stronger polar vortex
initiated by volcanic effects can increase the probability of formation of polar
stratospheric clouds and therefore enhance the rate of heterogeneous chemical
destruction of stratospheric ozone, especially in the NH (Tabazadeh et al., 2002,
July 2009
9-199
DRAFT-DO NOT CITE OR QUOTE

-------
192250). The above studies indicate effects on the stratospheric ozone layer in the
wake of a volcanic eruption and under conditions of enhanced anthropogenic halogen
loading. Interactive microphysics-chemistry-climate models (Dameris et al., 2005,
192238; Rozanov et al., 2002, 192247; Rozanov et al., 2004, 192245; Shindell et al.,
2003, 192069; Timmreck et al., 2003, 192254) indicate that aerosol-induced
stratospheric heating affects the dispersion of the volcanic aerosol cloud, thus
affecting the spatial RF. However the models' simplified treatment of aerosol
microphysics introduces biases; further, they usually overestimate the mixing at the
tropopause level and intensity of meridional transport in the stratosphere (Douglass
et al., 2003, 057260; Schoeberl et al., 2003, 057262). For present climate studies, it is
practical to utilize simpler approaches that are reliably constrained by aerosol
observations.
Because of its episodic and transitory nature, it is difficult to give a best estimate
for the volcanic RF, unlike the other agents. Neither a best estimate nor a level of
scientific understanding was given in the TAR. For the well-documented case of the
explosive 1991 Mt. Pinatubo eruption, there is a good scientific understanding.
However, the limited knowledge of the RF associated with prior episodic, explosive
events indicates a low level of scientific understanding.
9.3.9. Other Special Sources and Effects
International shipping has been identified as an additional source of carbonaceous
aerosols. Simulations with a climate model that included aerosol effects and 3 different
emissions inventories showed that shipping contributed 2.3-3.6% of the total S( ) r
atmospheric aerosol content and 0.4-1.4% of the total BC atmospheric aerosol content,
based on global means in 2000. This modeling also showed that aerosol optical thickness
over the Indian Ocean, the Gulf of Mexico, and the northeastern Pacific Ocean varied by 8
to 10%. The corresponding all-sky (that includes both cloudy and clear skies) direct
radiative forcings ranged from -0.011 W/m2 to -0.013 W/m2. The greatest effect of aerosols
emitted from global shipping is likely to be an increase in cloud formation and the resulting
change in reflectivity of shortwave radiation. Aerosols from shipping were estimated to
contribute 17-39% of the total anthropogenic aerosol radiation forcing effect.
When BC is deposited to the surface of ice or snow, solar absorption and heating occur
at the surface. This can melt additional snow or ice at the surface and the reflectivity of the
surface can change. Both factors affect aspects of climate. Jacobson (2003, 155866; 2004,
155870) and Jacobson et al. (2004, 180362) estimated the warming due to fossil fuel BC and
organic matter using the Gas, Aerosol, Transport, Radiation, General Circulation,
Mesoscale and Ocean Model (GATOR-GCMOM). The modeling effort included consideration
of the BC cycle, accounting for emissions, transport, aerosol coagulation, aerosol growth,
cloud activation, aerosol-cloud coagulation, cloud-cloud coagulation, rainout, washout, dry
deposition, and processes of precipitated and dry-deposited BC in snow and sea ice. Results
July 2009
9-200
DRAFT-DO NOT CITE OR QUOTE

-------
suggested that BC absorption in snow and sea ice increased near-surface temperatures over
a 10-yr simulation by about 0.06°K (Jacobson, 2003, 189421).
BC soot is a potentially important agent of climate warming in the Arctic, and
northern boreal wildfires may contribute substantially to this effect. Soot is approximately
twice as effective as CO2 in altering surface air temperature, and can reduce sea ice
formation and snow albedo (Hansen and Nazarenko, 2004, 156521).
Kim et al. (2005, 155900) investigated the relationships between northern boreal
wildfires and reductions in Arctic sea ice and glacial coverage. They modeled the
FROSTFIRE boreal forest control burn (Hinzman et al., 2003, 155845) with respect to BC
aerosol transport, dispersion, and deposition. Model results suggested that boreal wildfires
could be a major source of BC soot to sea ice and glaciers in Alaska. This may exacerbate
summer melting of sea ice and reduce recruitment of first-year ice into multiyear ice,
thereby leading to an overall reduction in sea ice. Similarly, increased BC soot on glaciers
would be expected to increase summer melting and lead to an overall reduction in glacial
coverage (Kim et al., 2005, 155900).
Jacobson (2002, 155865) proposed, based on model simulations with 12 identifiable
effects of aerosol particles on climate, emission reductions of fossil fuel particulate BC and
associated organic matter could potentially slow warming for a specific period more than
reduction of CO2 or CH4 for a specific period. Jacobson's (2006, 156599) calculations
suggested that fossil fuel BC plus organic matter emissions reductions could eliminate 8 to
18% of total anthropogenic warming, and 20 to 45% of net warming after accounting for
aerosol cooling, within a period of 3-5 yr (Chock et al., 2003, 155727). See also conflicting
discussions (Feichter et al., 2003, 155772; Penner, 2003, 156851); and further responses
(Jacobson, 2003, 155867; Jacobson, 2003, 155868; Jacobson, 2003, 155869; Penner, 2003,
156851).
Bond and Sun (2005, 156282) reviewed published data regarding the warming
potential of BC, compared with CO2 and other GHG. Climatic effects of GHG are generally
compared on the basis of top-of-the-atmosphere, globally averaged changes in radiative
balance. On that basis, BC is one of the largest individual warming agents, after CO2 and
perhaps CH4 (Bond and Sun, 2005, 156282; Jacobson, 2000, 056378; Sato et al., 2003,
156947).
Reddy and Boucher (2007, 156042) conducted an analysis that provided regional
estimates of BC emissions from fossil fuels andbiofuels. These estimates indicated that
July 2009
9-201
DRAFT-DO NOT CITE OR QUOTE

-------
East and Southeast Asia contributed over 50% of the global BC burden and its associated
direct radiative forcing. Europe was found to be the largest BC contributor in the northern
latitudes. The indirect effect of BC deposition on snow was also estimated to be highest for
Europe.
To improve understanding of the role of aerosols in climate forcing, Chung and
Seinfeld (2002, 155732) estimated the global distribution of BC, primary organic particles
(those directly emitted from combustion), secondary organic particles (primary organic
compounds partially oxidized in the atmosphere), and SOr aerosols to model the overall
radiative forcing of these groups of compounds. The model was run with the assumption
that the BC particles do not combine with OC or SO r particles (termed an external
mixture), and with the assumption that the particles are represented by a core of BC
surrounded by a shell of light scattering aerosols. Modeling results suggested an overall
radiative cooling effect from aerosols ranging from -0.39 to "0.78 W/m2.
Roberts and Jones (2004, 156052) used a climate modeling approach to compare
possible effects of BC on climate warming to those attributable to emissions from
greenhouse gases. Results suggested that the warming effect from atmospheric BC aerosols
may not be large relative to that from greenhouse gases. A different modeling approach by
Roeckner et al. (2006, 156920) evaluated the effects of BC and primary OC on climate
under two scenarios of carbonaceous aerosol emissions. In the first scenario, BC and
primary OC emissions decreased over Europe and China, but increased at lower latitudes.
In the second scenario, emissions were frozen at 2000 levels. The effects of both scenarios
on mean global temperature were found to be small, but higher aerosol emissions at low
latitudes did result in atmospheric heating and corresponding land surface cooling that led
to increased precipitation and runoff in this simulation.
Study of BC effects in tropical climates was undertaken by Wang (2007, 156147).
Substantial effects of direct radiative forcing by BC on the tropical Pacific were shown in
model results that were similar to the El Nino Southern Oscillation activities both in the
nature and scale of effects with enhancement of the Indian monsoon circulation. The model
suggested that atmospheric heating by radiation absorption by BC can form temperature
and pressure anomalies that favor propagation of convection from western to central and
eastern Pacific. More work will be needed to distinguish between the aerosol signal and
natural factors in controlling tropical precipitation in this region.
July 2009
9-202
DRAFT-DO NOT CITE OR QUOTE

-------
Table 9-18.
Overview of the different aerosol indirect effects and their sign of the net radiative flux
change at the top of the atmosphere (TOA).
Effect
Cloud Types
Affected
Process
Sign of Change in
TOA Radiation
Potential
Magnitude
Scientific
Understanding
Cloud albedo
effect
All clouds
For the same cloud water or ice content more but smaller cloud
particles reflect more solar radiation
Negative
Medium
Low
Cloud lifetime
effect
All clouds
Smaller cloud particles decrease the precipitation efficiency
thereby presumably prolonging cloud lifetime
Negative
Medium
Very low
Semi-direct effect
All clouds
Absorption of solar radiation by absorbing aerosols affects static
stability and the surface energy budget, and may lead to an
evaporation of cloud particles
Positive or Negative
Small
Very low
Glaciation indirect
effect
Mixed-phase
clouds
An increase in IN increases the precipitation efficiency
Positive
Medium
Very low
Thermodynamic
effect
Mixed-phase
clouds
Smaller cloud droplets delay freezing causing super-cooled clouds
to extend to colder temperatures
Positive or Negative
Medium
Very low
Source: Denman (2007,156394)
Table 9-19. Overview of the different aerosol indirect effects and their implications for the global
mean net shortwave radiation of the surface Fsfc (columns 2-4) and for precipitation
(columns 5-7).
Effect
Sign of Change
in Fsfc
Potential
Magnitude
Scientific
Understanding
Sign of Change in
Precipitation
Potential
Magnitude
Scientific
Understanding
Cloud albedo effect
Negative
Medium
Low
n.a.
n.a.
n.a.
Cloud lifetime effect
Negative
Medium
Very low
Negative
Small
Very low
Semi-direct effect
Negative
Large
Very low
Negative
Large
Very low
Glaciation indirect
effect
Positive
Medium
Very low
Positive
Medium
Very low
Thermodynamic
effect
Positive or Negative
Medium
Very low
Positive or Negative
Medium
Very low
Source: Denman (2007,156394)
There are several other kinds of climate effects from aerosol PM. None is well
understood or well quantified. The semi-direct effect, which involves absorption of solar
radiation by soot particles followed by re-emission as thermal radiation, is expected to heat
the air mass and increase its static stability relative to the surface. The semi-direct effect
can also cause evaporation of cloud droplets, thereby partially offsetting the cloud albedo
indirect effect. The glaciation effect involves an increase in IN, which is expected to cause
rapid glaciation of a super-cooled liquid water cloud due to the differences in vapor pressure
over ice and water. Unlike cloud droplets, these ice crystals can quickly reach precipitation
size, with the potential to turn a non-precipitating cloud into a precipitating cloud. The
thermodynamic effect involves a delay in freezing by the smaller cloud droplets, which can
July 2009
9-203
DRAFT-DO NOT CITE OR QUOTE

-------
cause super cooled clouds to occur under colder temperatures. The possible consequences to
radiative flux of all of the processes are outlined in Table 9-18 (top of the atmosphere
effects) and Table 9-19 (surface radiative and precipitation effects) (Denman et al., 2007,
156394), though significant uncertainties remain. Nevertheless, the individual processes
cannot be considered in isolation because of the numerous feedbacks, and because
atmospheric aerosol concentrations and climate are intimately coupled (Denman et al.,
2007, 156394; Dentener et al., 2006, 088434).
9.3.9.1. Glaciers and Snowpack
Organic compounds are incorporated into snow by wet and dry deposition processes
(Lei and Wania, 2004, 127880; Roth et al., 2004, 056431). Atmospherically deposited
organics appear to be ubiquitous in snowpacks at appreciable concentrations (Grannas et
al., 2007, 156492). Examples include PAHs, phthalates, alkanes, phenols, low molecular
weight carbonyls, POPs, and low molecular weight organic acids (Halsall, 2004, 155822;
Nakamura et al., 2000, 156792; Villa et al., 2003, 156139). Humic-like substances found in
the snowpack may release VOCs into the atmosphere via photo-oxidation (Grannas et al.,
2004, 155803; Grannas et al., 2007, 156492). Several thousand organic species were
identified by Grannas et al. (2006, 155804), based on molecular weight, from a single ice
core collected in Russia. Little information is available, however, regarding the chemical
properties of these chemical constituents. In addition to the diversity of chemicals that are
deposited into the snowpack, there are also biological organisms, including bacteria and
algae. Their role in influencing snow chemistry and volatilization processes is not
understood (Grannas et al., 2007, 156492).
Recent research has explored connections between the atmosphere and the cryosphere
(land or sea covered by snow or ice). A seasonal maximum of 40% of the Earth's land surface
is covered by snow or ice, as well as several percent of the oceans. Particulate deposition to
snow and ice surfaces can affect melting rates. Deposition of PM to glacial ice surfaces can
affect the subsequent rate of melting. A thin cover of debris contributes to accelerated
melting. A thicker cover of debris, such as may result from a volcanic eruption, retards
melting. The difference is due to the changing balance between enhanced absorption of
shortwave radiation by PM and conductive heat flow (insulation) through a buildup of
material having low heat conduction (Kirkbride and Dugmore, 2003, 156645).This issue is
particularly important for deposition of large quantities of volcanic material. To a lesser
July 2009
9-204
DRAFT-DO NOT CITE OR QUOTE

-------
extent, however, the same principles apply to PM deposition derived from air pollution.
Under a thin layer of debris, ablation rates are higher than for clean ice. However, as the
thickness of the debris layer increases, ablation rates systematically decline (Nicholson and
Benn, 2006, 156806). The threshold debris thickness that separates ablation increase from
decrease is site specific and depends on local climate and the nature of the debris particles.
Nicholson and Benn (2006, 156806) presented a surface energy balance model to calculate
ice melt beneath a surface debris layer, based on meteorological data and basic debris
characteristics. Modeled melting rates matched observed rates, suggesting that the model
produced useful results.
Long-range atmospheric transport of PM delivers a large fraction of the total input of
POPs to the Arctic region (Halsall, 2004, 155822). These contaminants can accumulate in
Arctic food webs and have become the focus of international research and concern.
Nevertheless, fate and transport of POPs within terrestrial and marine Arctic ecosystems
are not well understood and are strongly affected by the presence of snow and ice. Sea ice
provides a barrier to air-water exchange, and this hinders volatilization and re-emission of
previously deposited contaminants (Halsall, 2004, 155822). Thus, the effects of greenhouse
gasses and PM on climate in the Arctic region have feedbacks to POP fate, transport, and
toxicity. The transfer of POPs among the major abiotic environmental compartments in the
Arctic are summarized in Figure 9-83 from Halsall (2004, 155822). Recent studies detailing
rate and transport of POPs are summarized in Table 9-20.
Table 9-20. Recent studies highlighting POP occurrence and fate in the major arctic
compartments.
ATMOSPHERE
1
Annual time-series of 0C and PCB concentrations in the Norwegian Arctic
Oehme et al. (1996,156001)
2
Long-term analysis of the chlordane-group and their input to the Arctic with changing sources
Bidleman et al. (2002,155691)
3
PAH occurrence at monitoring sites across the Arctic, seasonality and gas/particle partitioning
Halsall et al. (1997,155821)
4
PCB occurrence at monitoring sites across the Arctic, spatial differences and seasonality
Stern et al. (1997,156096)
5
Long-term analysis of PCB and 0C trends in the Canadian Arctic and seasonal patterns
Hung et al. (2001,155856; 2002,155857)
6
Trans-Pacific LRAT and impact of Asian sources on the western Canadian Arctic
Bailey et al. (2000,155670)
FRESHWATER
1
Annual avg water concentrations in major Russian rivers for selected 0C pesticides
Alexeeva et al. (2001,155651)
8
Long-term (decades) PCB deposition profile in Arctic lake sediments
Muir et al. (1996,155991)
9
Mass balance of selected OCs in Canadian Arctic lake conducted with data collected over 3 yrs
Helm et al. (2002,155835)
10
Examining the biodegradation of HCHs in Canadian Arctic watersheds
Helm et al. (2000,155834)
July 2009
9-205
DRAFT-DO NOT CITE OR QUOTE

-------
ATMOSPHERE
MARINE
11
Transport and entry of p-HCH into western Arctic Ocean via Pacific surface waters
Li etal. (2002,156691)
12
Occurrence of current use pesticides in air, fog and surface seawater in the western Arctic Ocean
Chermyak et al. (1996,155726)
13
Resolving petrogenic and anthropogenic PAH input to marine sediments in coastal Arctic seas
Yunkeretal. (1996,156175)
14
Quantifying abiotic and biotic degradation of HCHs in the Arctic Ocean water column
Harneret al. (2000,155829)
15
PCBs and OCs in surface ocean water-Bering and Chukchi seas
Strachan et al. (2001,156103)
16
Spatial patterns of HCHs and toxaphene in Arctic Ocean surface water
Jantunen and Bidleman (1998,155877)
SNOW/AIR FRESHWA TER
17
PAHs (and inorganics) in surface snow layers (snowpit) at Summit, Greenland
Masclet et al. (2000,155966)
18
PAHs measured in snow and ice layers on Agassiz ice cap, Ellesmere Island, Canada
Peters et al. (1995,156856)
19
Modeling OC behaviour and fate in the surface seasonal snow pack at Amituk Lake, Canada
Wania et al. (1998,156148)
20
OCs, PCBs and PAHs in snow and ice of the Ob-Yenisey watershed of the Russian Arctic
Melnikov et al. (2003,156753)
OCEAN/AIR
21
Transfer of a-HCH across the air/water interface in the western Arctic ocean
Jantunen and Bidleman (1996,155876)
22
Calculated seasonality of OC air/water fluxes in the Canadian high Arctic
Hargrave et al. (1997,155827)
OCEAN/ICE
23
Transport potential of contaminants across the Arctic ocean via sea-ice drift
Pfirman et al. (1997,156864)
24
The importance of eastern Arctic sea-ice drift as a source of contaminants to the Norwegian sea
Korsneset al. (2002,156657)
Source: Halsall (2004,155822)
July 2009
9-206
DRAFT-DO NOT CITE OR QUOTE

-------
Atmosphere
LRAT

(21-22)
(19)
LROT
Snow/sea ice
Snow/ice caps
Terrestrial (20)
Catchment <7-1 O)
(11-16) JL?
Surface
Ocean nA
rivers
Deep Ocean
Figure 9-83. The transfer of POPs between the major abiotic compartments of the Arctic. Shaded
arrows represent inputs/outputs of POPs to the Arctic. The numbers refer to selected
studies detailing the occurrence and behavior of POPs, and are listed in Table 9-20.
Question marks represent those areas that are least well understood. LRAT-long range
atmospheric transport; LROT - long range oceanic transport.
9.3.9.2. Radiative Forcing by Anthropogenic Surface Albedo Change: BC in Snow and
Ice
Section 9.3.9.2. comes directly from IPCC AR4 Chapter 2, Section 2.5.4, with section,
table, and figure numbers changed to be internally consistent with this ISA.
The presence of soot particles in snow could cause a decrease in the albedo of
snow and affect snowmelt. Initial estimates by Hansen et al. (2000, 042683) suggested
that BC could thereby exert a positive RF of +0.2 W/m2. This estimate was refined by
Hansen and Nazarenko (2004, 156521). who used measured BC concentrations within
snow and ice at a wide range of geographic locations to deduce the perturbation to the
surface and planetary albedo, deriving an RF of +0.15 W/m2. The uncertainty in this
estimate is substantial due to uncertainties in whether BC and snow particles are
internally or externally mixed, in BC and snow particle shapes and sizes, in voids
within BC particles, and in the BC imaginary refractive index. Jacobson (2004,
155870) developed a global model that allows the BC aerosol to enter snow via
precipitation and dry deposition, thereby modifying the snow albedo and emissivity.
They found modeled concentrations of BC within snow that were in reasonable
agreement with those from many observations. The model study found that BC on
snow and sea ice caused a decrease in the surface albedo of 0.4% globally and 1% in
July 2009	9-207	DRAFT - DO NOT CITE OR QUOTE

-------
the NH, although RFs were not reported. Hansen et al. (2005, 059087) allowed the
albedo change to be proportional to local BC deposition according to Koch (2001,
192054) and presented a further revised estimate of 0.08 W/m2. They also suggested
that this RF mechanism produces a greater temperature response by a factor of 1.7
than an equivalent C02 RF, that is, the 'efficacy' may be higher for this RF mechanism
(see Section 2.8.5.7 of the IPCC AR4). This report adopt s a best estimate for the BC on
snow RF of +0.10 ± 0.10 W/m2, with a low level of scientific understanding
(Section 2.9, Table 2.11, of the IPCC AR4).
9.3.9.3. Effects on Local and Regional Climate
Most effects of PM on climate, as assessed by IPCC (Stohl et al., 2007, 157015) and
summarized in this assessment, focus on global-scale processes and responses. In addition,
it is also possible that PM emissions contribute to local and regional climate changes. These
might include short-term cycles in rainfall or temperature and rainfall suppression,
especially near cities and for orographic precipitation. Rainfall suppression, in particular, is
believed to exacerbate water supply problems which are substantial in many regions,
especially in the western U.S.
Aerosol particles, directly and through cloud enhancement, may reduce near-surface
wind speeds locally. Slower winds, in turn, reduce evaporation. The overall impact can be a
reduction in local precipitation. Jacobson and Kaufman (2006, 090942) investigated the
effects of PM on spatially-distributed wind speeds and resulting feedbacks to precipitation
using the GATOR-GCMOM (Jacobson, 2001, 155864) and supporting evidence from satellite
data. The study focused on the South Coast Air Basin (SCAB) in California during
February and August, 2002-2004. The modeled precipitation decrease over land in
California was 2% of the baseline 1.5 mm/day due to emissions of anthropogenic aerosol
particle and precursor gasses in the SCAB domain. However, the reduction over much of
the Sierra Nevada, where most precipitation falls, was up to 0.5 mm/day, or 4 to 5% of the
baseline 10 to 13 mm/day in that mountainous region (Jacobson, 2006, 156599). The
probable mechanism was described as follows. Aerosol particles and aerosol-enhanced
clouds reduce wind speeds below them by stabilizing the air, reducing the vertical transport
of horizontal momentum. In turn, the reduced wind speeds, and associated reduced
evaporation and increased cloud lifetime, contributes to reduced local and regional
precipitation (Jacobson, 2006, 156599).
Effects of air pollution on regional precipitation were quantified by Givati and
Rosenfeld (2004, 156475). They found a 15-25% reduction in the orographic component of
precipitation downwind of major coastal urban areas during the 20th century. Their study
focused on orographically-forced clouds because these short-lived, shallow clouds are
July 2009
9-208
DRAFT-DO NOT CITE OR QUOTE

-------
expected to exhibit the largest effect of air pollution on precipitation. Substantially larger
precipitation suppression due to aerosol particulate pollution was found between Fresno
and Sacramento in California by Givati and Rosenfeld (2004, 156475). Precipitation losses
over topographical barriers in the Sierra Nevada amounted to 15%-25% of the annual
precipitation at elevations less than 2,000 m. This precipitation suppression occurred
mainly in the relatively shallow orographic clouds within the cold air mass of cyclones. The
suppression that occurred on the upslope side of the mountains was coupled with similar
percentage (but lower absolute volume) enhancement on the drier downslope eastern side
(Givati and Rosenfeld, 2004, 156475). Similar results were found in studies by Griffith et al.
(2005, 156497), Jirak and Cotton (2006, 156612), Rosenfeld and Givati (2006, 156924), and
Rosenfeld et al. (2007, 156057). At all of these study locations (California, Israel, Utah,
Colorado, China), orographic precipitation decreased by 15-30% downwind of pollution
sources, likely due to creation of more and smaller cloud droplets and resulting suppression
of precipitation.
The study of Givati and Rosenfeld (2004, 156475) was the first to quantify the
microphysical effect of mesoscale precipitation. Following the findings of Givati and
Rosenfeld (2004, 156475), the effects of aerosol air pollution on precipitation at high
elevation sites in the Front Range of Colorado adjacent to urban areas were investigated by
Jirak and Cotton (2006, 156612). Examination of precipitation trends showed that the ratio
of upslope precipitation during easterly flows at high elevation west of Denver and Colorado
Springs to the upwind urban sites decreased by about 30% over the past half century. These
results provide further support for the hypothesis that aerosol pollution suppresses
orographic precipitation downwind of pollution source areas.
Griffith et al. (2005, 156497) found similar reductions in mountainous precipitation in
Utah, downwind of Salt Lake City and Provo. The ratio of precipitation at mountain
stations located in rural settings in Utah and Nevada remained stable, supporting the
hypothesis that air pollution decreases Ro (the ratio of precipitation at the downwind site to
precipitation at the upwind pollution source) over the mountains to the east of Salt Lake
City.
Rosenfeld and Givati (2006, 156924) extended the investigation of the suppression of
precipitation by aerosol pollutants to a larger scale by examining the ratio between
precipitation amounts over the hills to precipitation over upwind lowland areas throughout
the western U.S. from the Pacific Coast to the Rocky Mountains. They found in these paired
July 2009
9-209
DRAFT-DO NOT CITE OR QUOTE

-------
analyses a pattern of decreasing precipitation by as much as 24% from the Mexican border
to central California, with no decrease in northern California and Oregon and smaller
decrease of 14% in Washington east of Seattle and Puget Sound. Similar decreases were
found over Arizona and New Mexico (Rosenfeld, 2006, 190233), Utah (Griffith et al., 2005,
156497), and the east slope of the Colorado Rockies (Jirak and Cotton, 2006, 156612).
Suppression of winter orographic precipitation appears to occur up to hundreds of
kilometers inland of coastal urban areas (Rosenfeld, 2006, 190233). Decreases in this
precipitation ratio occurred during winter orographic precipitation, but not during
convective summer precipitation over the same mountain ranges. This finding agrees with
the expectation that aerosol-induced changes in the rate of precipitation formation would
cause a decrease in precipitation from shallow and short-lived orographic clouds, but not
necessarily from deeper and longer-lived thermally-driven convective clouds.
Results of these studies of aerosol effects on orographic precipitation suggest that
human-caused air pollution, and fine particles in particular, have had a large effect on
precipitation well beyond the local scales of the pollution sources (Rosenfeld, 2006, 190233).
9.3.10. Summary of Effects on Climate
Aerosols affect climate through direct and indirect effects. The direct effect is
primarily realized as planet brightening when seen from space because most aerosols
scatter most of the visible spectrum light that reaches them. The IPCC AR4 reported that
the radiative forcing from this direct effect was -0.5 (±0.4) W/m2 and identified the level of
scientific understanding of this effect as 'Medium-low'. The global mean direct radiative
forcing from individual aerosol components varies from strongly negative for S( ) r to
positive for BC with weaker positive or negative effects for other components, all of which
can vary strongly over space and time and with aerosol size. The indirect effects are
primarily realized as an increase in cloud brightness (termed the "first indirect" or
"Twomey" effect), changes in precipitation, and possible changes in cloud lifetime. The IPCC
AR4 reported that the radiative forcing from the Twomey effect was -0.7 (range: -1.1 to +4)
and identified the level of scientific understanding of this effect as "Low'". The other
indirect effects from aerosols were not considered to be radiative-forcing.
Taken together, direct and indirect effects from aerosols increase Earth's shortwave
albedo or reflectance, thereby reducing the radiative flux reaching Earth's surface from the
Sun. This produces net climate cooling from aerosols. The current scientific consensus
July 2009
9-210
DRAFT-DO NOT CITE OR QUOTE

-------
reported by IPCC AIM is that the direct and indirect radiative forcing from anthropogenic
aerosols computed at the top of the atmosphere, on a global average, is about -1.3 (range: -
2.2 to -0.5) W/m2. Although the magnitude of this negative radiative forcing appears large
in comparison to the analogous IPCC AIM estimate of positive radiative forcing from
anthropogenic GHG of about 2.9 (±0.3) W/m2, the spatial and temporal distributions of
these two very different radiative forcing agents are dissimilar! therefore, they do not
simply cancel and regional differences can be large. These differences result from the much
shorter atmospheric lifetime of aerosols than for the radiatively important trace gases,
implying that the radiative effects of aerosols respond much more quickly to changes in
emissions than do the effects from the gas-phase forcing agents Moreover, the effect of
present-day aerosols is to cool Earth's surface but, on average, to heat the atmosphere
itself within the atmospheric column, the radiative forcing effect from aerosols is estimated
to range from +0.8 to +2 W/m2.
Considerable progress has been made with in situ and remotely-sensed aerosols
concentrations including the MODIS, MISR, POLDER, and OMI satellite instruments! see
the discussion in Section 3.4.1.6. The accuracy for aerosol optical depth (AOD) measured
with these global-coverage remote sensing instruments is on the order of 0.05 or 20% of the
AOD, but is still much lower than for the limited-area surface-based sun photometers
which have accuracy in the range of 0.01-0.02. The differences remaining between surface
and remotely sensed AOD and between estimates computed from measurements and from
numerical model predictions are important because AOD is a significant element in
determining radiative forcing. Hence, uncertainty and error in AOD measurements and
modeling propagate into the range of estimates for total radiative forcing reported here.
Numerical modeling of aerosol effects on climate has also sustained remarkable
progress since the 2004 PM AQCD (EPA, 2004, 056905), though model solutions still
display large heterogeneity in their estimation of the direct radiative forcing effect from
anthropogenic aerosols. Differences among models are due in large measure to differences
in: emissions of PM and precursors to secondary PM formation! in the representation of
aerosol microphysical and optical processes! in regional- and global-scale transport and
transformation; as well as in the effects of aerosol radiative forcing. The clear-sky direct
radiative forcing over ocean due to anthropogenic aerosols is estimated from satellite
instruments to be on the order of -1.1 (±0.37) W/m2 while model estimates are -0.6 W/m2.
The models' low bias over ocean is carried through for the global average: global average
July 2009
9-211
DRAFT-DO NOT CITE OR QUOTE

-------
direct radiative forcing from anthropogenic aerosols is estimated from measurements to
range from -0.9 to -1.9 W/m2, larger than the estimate of -0.8 W/m2 from the models. Spatial
heterogeneity in radiative forcing is expected to exert significant effects on regional climate,
but because the effects of climate warming and cooling are not strictly co-located spatially
or temporally with radiative forcing or with emissions in particular for precursors for
secondary PM formation, assessments of effects for sub-global domains are even more
uncertain than the global averages reported here.
Overall, the evidence is sufficient to conclude that a causal relationship exists between
PM and effects on climate, including both direct effects on radiative forcing and indirect effects
that involve cloud feedbacks that influence precipitation formation and cloud lifetimes.
9.4. Ecological Effects of PM
9.4.1. Introduction
PM is heterogeneous with respect to chemical composition and size, therefore it can
cause a variety of ecological effects, which have been previously described by the U.S. EPA
(2004, 056905) and by Grantz et al. (2003, 155805). Atmospheric PM has been defined, for
regulatory purposes, mainly by size fractions and less clearly so in terms of chemical
nature, structure, or source. Both fine and coarse-mode particles may affect plants and
other organisms, however PM size classes do not necessarily relate to ecological effects
(U.S. EPA, 1996, 079380). More often the chemical constituents drive the ecosystem
response to PM (Grantz et al., 2003, 155805).
The previous PM assessment (U.S. EPA, 2004, 056905) included the acidifying effects
of particulate N and S. The2008 NOxSOx ISA (U.S. EPA, 2008, 157074) assessed the effects
of particle- and gas-phase N and S pollution on acidification, N enrichment, and Hg
methylation. Acidification of ecosystems is driven primarily by deposition resulting from
SOx, NOx, and NHx pollution. Acidification from the deposition resulting from current
emission levels causes a cascade of effects that harm susceptible aquatic and terrestrial
ecosystems, including slower growth and injury to forests and localized extinction of fishes
and other aquatic species. In addition to acidification, atmospheric deposition of reactive N
resulting from current NOx and NHx emissions along with other non-atmospheric sources
(e.g., fertilizers and wastewater), causes a suite of ecological changes within sensitive
July 2009
9-212
DRAFT-DO NOT CITE OR QUOTE

-------
ecosystems. These include increased primary productivity in most N-limited ecosystems,
biodiversity losses, changes in C cycling, and eutrophication and harmful algal blooms in
freshwater, estuarine, and ocean ecosystems. In some watersheds, additional S042~from
atmospheric deposition increases Hg methylation rates by increasing both the number and
activity of S-reducing bacteria. Methylmercury is a powerful toxin that can bioaccumulated
to toxic amounts in food webs at higher trophic levels. Effects of particulate metals and
organics on ecosystems and ecosystem receptors are the focus in this PM ISA.
This assessment of PM effects on ecosystems considers both direct and indirect
exposure pathways. Atmospheric PM may affect ecological receptors directly following
deposition on surfaces or indirectly by changing the soil chemistry or by changing the
amount of radiation reaching the Earth's surface. Indirect effects acting through the soil
are often thought to be most significant because they can alter nutrient cycling and inhibit
nutrient uptake (EPA, 2004, 056905). The U.S. EPA (2004, 056905) reported that the effects
of PM can be both chemical and physical. Physical effects of particle deposition on
vegetation may include abrasion and radiative heating. Chemical effects may be more
significant (U.S. EPA, 2008, 157074).
In general, anthropogenic stressors can result in damaged ecosystems that do not
recover readily (Odum, 1993, 076742; Rapport and Whitford, 1999, 004595). Ecosystems
sometimes lack the capacity to adapt to an anthropogenic stress and are unable to maintain
their normal structure and functions unless the stressor is removed (Rapport and Whitford,
1999, 004595). These stresses result in a process of ecosystem degradation marked by a
decrease in biodiversity, reduced primary and secondary production, and a lower capacity to
recover and return to the original ecosystem state. In addition, there can be an increased
prevalence of disease, reduced nutrient cycling, increased dominance of exotic species, and
increased dominance by smaller, short-lived opportunistic species (Odum, 1985, 039482;
Rapport and Whitford, 1999, 004595).
Ecosystems are often subjected to multiple stressors, of which atmospheric PM
deposition is only one. Additional stressors are also important, including O3 exposure,
climatic variation, natural and human disturbance, the occurrence of invasive non-native
plants, native and non-native insect pests, disease, acidification, and eutrophication among
others. PM deposition interacts with these other stressors to affect ecosystem patterns and
processes.
July 2009
9-213
DRAFT-DO NOT CITE OR QUOTE

-------
The possible effects of particulate (and other) air pollutants on ecosystems have been
categorized by Guderian (1977, 004150) as follows:
¦	accumulation of pollutants in plants and other ecosystem components (such as soil
and surface- and groundwater),
¦	damage to consumers as a result of pollutant accumulation,
¦	changes in species diversity because of shifts in competition,
¦	disruption of biogeochemical cycles,
¦	disruption of stability and reduction in the ability to self-regulate,
¦	breakdown of stands and associations, and
¦	expansion of denuded zones.
The general conclusion of the last PM assessment (U.S. EPA, 2004, 056905) was that
ecosystem response to PM can be difficult to determine because the changes are often
subtle. For example, changes in the soil may not be observed until pollutant deposition has
occurred for many decades, except in the most severely polluted areas around heavily
industrialized point sources. The presence of co-occurring pollutants generally makes it
difficult to attribute ecological effects to PM alone or to one constituent in the deposited
PM. In other words, the potential for alteration of ecosystem function and structure exists
but can be difficult to quantify except in cases of extreme amounts of deposition, especially
when there are other pollutants present in the ambient air that may produce additive or
synergistic responses.
New information on the ecological effects of coarse and fine particle PM is presented
in the following discussion in the context of effects that were known from the last PM
AQCD (EPA, 2004, 056905). The general effects of the chemical constituents of PM are
discussed; however, a rigorous assessment of each chemical constituent (e.g., Hg, Cd, Pb,
etc.) is not given. Both direct and indirect effects will be discussed and the strength of the
scientific evidence will be evaluated using the causality framework.
July 2009
9-214
DRAFT-DO NOT CITE OR QUOTE

-------
9.4.1.1. Ecosystem Scale, Function, and Structure
Information presented in this section was collected at multiple scales, ranging from
the physiology of a given species to population, community, and ecosystem-level
investigations. For this assessment, "ecosystem" is defined as a functional entity consisting
of interacting groups of living organisms and their abiotic (chemical and physical)
environment. Ecosystems cover a hierarchy of spatial scales and can comprise the entire
globe, biomes at the continental scale, or small, well-circumscribed systems such as a small
pond.
Ecosystems have both structure and function. Structure may refer to a variety of
measurements including the species richness, abundance, community composition and
biodiversity as well as landscape attributes. Competition among and within species and
their tolerance to environmental stresses are key elements of survivorship. When
environmental conditions are shifted, for example, by the presence of anthropogenic air
pollution, these competitive relationships may change and tolerance to stress may be
exceeded. "Function" refers to the suite of processes and interactions among the ecosystem
components and their environment that involve nutrient and energy flow as well as other
attributes including water dynamics and the flux of trace gases. Plant processes including
photosynthesis, nutrient uptake, respiration, and C allocation, are directly related to
functions of energy flow and nutrient cycling. The energy accumulated and stored by
vegetation (via photosynthetic C capture) is available to other organisms. Energy moves
from one organism to another through food webs, until it is ultimately released as heat.
Nutrients and water can be recycled. Air pollution alters the function of ecosystems when
elemental cycles or the energy flow are altered. This alteration can also be manifested in
changes in the biotic composition of ecosystems.
There are at least three levels of ecosystem response to pollutant deposition: (l) the
individual organism and its environment! (2) the population and its environment! and (3)
the biological community composed of many species and their environment (Billings, 1978,
034165). Individual organisms within a population vary in their ability to withstand the
stress of environmental change. The response of individual organisms within a population
is based on their genetic constitution, stage of growth at time of exposure to stress, and the
microhabitat in which they are growing (Levine and Pinto, 1998, 029599). The range within
which organisms can exist and function determines the ability of the population to survive.
Those able to cope with the stresses survive and reproduce. Competition among different
July 2009
9-215
DRAFT-DO NOT CITE OR QUOTE

-------
species results in succession (community change over time) and, ultimately, produces
ecosystems composed of populations of species that have the capability to tolerate the
stresses (Guderian R, 1985, 019325; Rapport and Whitford, 1999, 004595). Available
information on individual, population and community response to PM will be discussed.
9.4.1.2. Ecosystem Services
Ecosystem structure and function may be translated into ecosystem services.
Ecosystem services identify the varied and numerous ways that ecosystems are important
to human welfare. Ecosystems provide many goods and services that are of vital importance
for the functioning of the biosphere and provide the basis for the delivery of tangible
benefits to human society. Hassan et al. (2005, 092759) define these to include supporting,
provisioning, regulating, and cultural services^
¦	Supporting services are necessary for the production of all other ecosystem services.
Some examples include biomass production, production of atmospheric O2, soil
formation and retention, nutrient cycling, water cycling, and provisioning of habitat.
Biodiversity is a supporting service that is increasingly recognized to sustain many
of the goods and services that humans enjoy from ecosystems. These provide a basis
for three higher-level categories of services.
¦	Provisioning services, such as products (Gitay et al., 2001, 092761), i.e., food
(including game, roots, seeds, nuts and other fruit, spices, fodder), fiber (including
wood, textiles), and medicinal and cosmetic products (including aromatic plants,
pigments).
¦	Regulating services that are of paramount importance for human society such as (a)
C sequestration, (b) climate and water regulation, (c) protection from natural
hazards such as floods, avalanches, or rock-fall, (d) water and air purification, and
(e) disease and pest regulation.
¦	Cultural services that satisfy human spiritual and aesthetic appreciation of
ecosystems and their components.
July 2009
9-216
DRAFT-DO NOT CITE OR QUOTE

-------
9.4.2. Deposition of PM
Deposition of PM is discussed in Chapter 3.3.3. Additional material specifically
related to ecosystems is discussed in this section.
9.4.2.1. Forms of Deposition
Research summarized by the previous NAAQS PM assessment illustrated the
complexity of deposition processes. Airborne particles, their gas-phase precursors, and their
transformation products are removed from the atmosphere by wet and dry deposition
processes. These deposition processes transfer PM pollutants to other environmental media
where they can alter the structure, function, diversity, and sustainability of complex
ecosystems. Dry deposition of PM is most effective for coarse particles. These include
primary geologic materials and elements such as iron and manganese. By contrast, wet
deposition is more effective for fine particles of secondary atmospheric origin and elements
such as cadmium, chromium, lead, nickel, and vanadium (Reisinger, 1990, 046737; Smith,
1990, 084015; U.S. EPA, 2004, 056905). The relative magnitudes of the different deposition
modes vary with ecosystem type, location, elevation, and chemical burden of the
atmosphere (U.S. EPA, 2004, 056905). There are differences in the deposition behavior of
fine and coarse particles. Coarse particles generally settle nearer their site of formation
than do fine particles. In addition, the chemical constitution of individual particles is
correlated with size. For example, much of the base cation and heavy metal burden is
present on coarse particles.
Fine PM is often a secondary pollutant that forms within the atmosphere, rather than
being directly emitted from a pollution source. It derives from atmospheric gas-to-particle
conversion reactions involving nucleation, condensation, and coagulation, and from
evaporation of water from contaminated fog and cloud droplets. Fine PM may also contain
condensates of VOCs, volatilized metals, and products of incomplete combustion, including
polycyclic aromatic hydrocarbons (PAH) and BC (soot) (EPA, 2004, 056905).
Fine PM may act as a carrier for materials such as herbicides that are phytotoxic.
Fine PM provides much of the surface area of particles suspended in the atmosphere,
whereas coarse PM provides much of the mass of airborne particles. Surface area can
influence ecological effects associated with the oxidizing capacity of fine particles, their
interactions with other pollutants, and their adsorption of organic compounds. Fine and
July 2009
9-217
DRAFT-DO NOT CITE OR QUOTE

-------
coarse particles also respond to changes in atmospheric humidity, precipitation, and wind,
and these can alter their deposition characteristics.
Coarse PM is mainly a primary pollutant, having been emitted from pollution sources
as fully formed particles derived from abrasion and crushing processes, soil disturbances,
desiccation of marine aerosol emitted from bursting bubbles, hygroscopic fine PM
expanding with humidity to coarse mode, and/or gas condensation directly onto preexisting
coarse particles. Suspended primary coarse PM may contain iron, silica, aluminum, and
base cations from soil, plant and insect fragments, pollen, fungal spores, bacteria, and
viruses, as well as fly ash, brake lining particles, and automobile tire fragments. Coarse
mode particles can be altered by chemical reactions and/or physical interactions with
gaseous or liquid contaminants.
Exposure to a given mass concentration of PM may lead to widely differing phytotoxic
and other environmental outcomes depending upon the particular mix of PM constituents
involved. Especially important in this regard are S and N components of PM, which are
addressed in the 2008 NOxSOx ISA, and effects of particulate heavy metals and organic
contaminants. This variability has not been characterized adequately. Though effects of
specific chemical fractions of PM have been described to some extent, there has been
relatively little research aimed at defining the effects of unspeciated PM on plants or
ecosystems.
9.4.2.2. Components of PM Deposition
Trace Metals
Atmospheric deposition can be the primary source of some metals to some watersheds.
Metal inputs can include the primary crustal elements (Al, Ca, K, Fe, Mg, Si, Ti) and the
primary anthropogenic elements (Cu, Zn, Cd, Cr, Mn, Pb, V, Hg). The crustal elements are
derived largely from weathering and erosion, whereas the anthropogenic elements are
derived from combustion, industrial sources, and other man-made sources (Goforth and
Christoforou, 2006, 088353).
Heavy metal deposition to ecosystems depends on their location as well as upwind
emissions source strength. The deposition velocity tends to be dependent on particle size
and chemical species. Larger particles deposit more efficiently than smaller particles.
Heavy metals preferentially associate with fine particles. Fine particles also have the
July 2009
9-218
DRAFT-DO NOT CITE OR QUOTE

-------
longest atmospheric residence times. Depending on climate and topography, fine particles
may remain airborne for days to months and may be transported thousands of kilometers
from their source.
Ecosystems immediately downwind of major heavy metal emissions sources may
receive locally heavy dry deposition. Trace element investigations conducted in roadside,
industrial, and urban environments have also shown that substantial amounts of
particulate heavy metals can accumulate on surfaces.
A significant trace metal component of PM is mercury (Hg). Mercury is toxic and can
move readily through environmental compartments. Atmospheric and depositional inputs of
Hg include both natural and anthropogenic sources. Natural geologic contributions to Hg in
the environment include geothermal and volcanic activity, geologic metal deposits, and
organic-rich sedimentary rocks. These natural emissions combine with anthropogenic
emissions from such sources as power plants, landfills, sewage sludge, mine waste, and
incineration (Gustin, 2003, 155816; Schroeder and Munthe, 1998, 014559). Emissions from
natural sources are controlled by geologic features, including substrate Hg content, rock
type, the degree of hydrothermal activity, and the presence of heat sources (Gustin, 2003,
155816). The significance of natural Hg sources relative to anthropogenic sources varies
geographically. For example, Nevada occurs within a global mercuriferous belt, with area
emissions about three times higher than the value assumed for global modeling (Gustin,
2003, 155816). In Nevada, natural and anthropogenic Hg emissions are approximately
equal (Gustin, 2003, 155816).
The U.S. EPA (1997, 157066) compiled an assessment of the sources and
environmental effects of Hg in the U.S. A variety of factors were found to influence Hg
deposition, fate and transport (Table l). Such factors relate in particular to speciation of the
Hg that is emitted, the forms in which it is deposited from the atmosphere, and
transformations that occur within the atmosphere and within the aquatic, transitional, and
terrestrial compartments of the receiving watershed. There have been studies that have
reconstructed, from lake sediment records, the atmospheric depositional history of trace
metals and PAHs in lakes adjacent to coal-fired power plants. For example, Donahue et al.
(2006, 155751) analyzed sediment from Wababun Lake, which is located in Alberta, Canada
in proximity (within 35 km) to 4 power plants built since 1950. Trace metal concentrations
of Hg, Cu, Pb, As, and Se in lake sediment increased by 1.2- to 4-fold. The total PAH flux to
surface sediments was 730 to 1100 pg/m2/yr, which was two to five times higher than in 2
July 2009
9-219
DRAFT-DO NOT CITE OR QUOTE

-------
lakes situated 20 km to the north and 70 km to the south. Further discussion of Hg effects
on ecosystems can be found in Section 9.4.5.
Table 9-21. Factors potentially important in estimating mercury exposure.
Factor	Importance and Possible Effect on Mercury Exposure
Type of anthropogenic source of mercury
Different combustion and industrial process sources are anticipated to have different local scale impacts due to

physical source characteristics (e.g., stack height), the method of waste generation (e.g., incineration or mass burn) or

mercury control devices and their effectiveness.
Mercury emission rates from stack
Increased emissions will result in a greater chance of adverse impacts on environment.
Mercury species emitted from stack
More soluble species will tend to deposit closer to the source.
Form of mercury emitted from stack	Transport properties can be highly dependent on form.
Deposition differences between vapor and	Vapor-phase forms may deposit significantly faster than particulate-bound forms,
particulate-bound mercury
Transformations of mercury after emission from Relatively nontoxic forms emitted from source may be transformed into more toxic compounds,
source
Transformation of mercury in watershed soil	Reduction and revolatilization of mercury in soil limits the buildup of concentration.
Transport of mercury from watershed soils to water	Mercury in watershed soils can be a significant source to water bodies and subsequently to fish,
body
Transformation of mercury in water body	Reduction, methylation, and demethylation of mercury in water bodies affect the overall concentration and the MHg
fraction, which is bioaccumulated in fish.
Facility locations	Effects of meteorology and terrain may be significant.
Location relative to local mercury source	Receptors located downwind are more likely to have higher exposures. Influence of distance depends on source type.
Contribution from non local sources of mercury	Important to keep predicted impacts of local sources in perspective.
Uncertainty	Reduces confidence in ability to estimate exposure accurately.
Source: Modified from U.S. EPA (1997,157066)
Organics
Organic compounds that may be associated with deposited PM include persistent
organic pollutants (POPs), pesticides, SOCs, polyaromatic hydrocarbons (PAHs) and flame
retardants among others. Organic compounds partition between gas and particle phases,
and organic particulate deposition depends largely on the particle sizes available for
adsorption (EPA, 2004, 056905). Dry deposition of organic materials is often dominated by
the coarse fraction. Gas-particle phase interconversions are important in determining the
amount of dry deposition.
Most persistent organic pollutants (POPs) enter the biosphere via human activities,
including synthetic pesticide application, output of polychlorinated dibenzo dioxins (PCDD)
from incinerators, and accidental release of PCBs from transformers (Lee, 2006, 088968).
Once they are introduced into the environment, their accumulation and magnification in
biological systems are determined by physiochemical properties and environmental
July 2009
9-220
DRAFT-DO NOT CITE OR QUOTE

-------
conditions (Section 9.4.5.7). Uptake by plants can occur at the soil/plant interface and at the
air/plant interface. For lipophilic POPs, such as PCDDs and PCBs, the air/plant response
route generally dominates (Lee, 2006, 088968; Thomas et al., 1998, 156118), but uptake
through above-ground plant tissue also occurs. In a study of zucchini (Cucurbitapepo), Lee
et al. (2006, 088968) found chlordane pesticide components in all vegetation tissues
examined: root, stem, leaves, fruits.
Many pesticides and SOCs are carcinogenic or estrogenic and pose potential threats to
aquatic and terrestrial biota. Although deposition of SOCs was previously reported for the
Sierra Nevada Mountains in California and the Rocky Mountains in Colorado, little was
previously known about the occurrence, distribution, or sources of SOCs in alpine,
sub-Arctic, and Arctic ecosystems in the western U.S. The snowpack is efficient at
scavenging of both particulate and gas phase pesticides from the atmosphere (Halsall, 2004,
155822; Lei and Wania, 2004, 127880). Analysis of pesticides in snowpack samples from
seven national parks in the western U.S. by Hageman et al. (2006, 156509) illustrated the
deposition and fate of 47 pesticides and their degradation products. Correlation analysis
with latitude, temperature, elevation, PM, and two indicators of regional pesticide use
suggested that regional patterns in historic and current agricultural practices are largely
responsible for the distribution of pesticides in the national parks. Pesticide deposition to
parks in Alaska was attributed to long-range atmospheric transport.
PAHs include hundreds of different compounds that are characterized by possessing
two or more fused benzene rings. They are widespread contaminants in the environment,
and are formed by incomplete combustion of fossil fuels and other organic materials. Eight
PAHs are considered carcinogenic and 16 are classified by EPA as priority pollutants. They
are common air pollutants in metropolitan areas, derived from vehicular traffic and other
urban sources. Especially high concentrations have been found near Soderberg aluminum
production industries and areas where wood heating during winter is common. Other
sources, in addition to gasoline and diesel engines, include forest fires and various forms of
fossil fuel combustion (Sanderson and Farant, 2004, 156942).
The behavior of PAHs is strongly determined by their chemical characteristics,
especially their nonpolarity and hydrophobicity. They readily adsorb to particulates in the
air and to sediments in water. Srogi (2007, 180049) provided a thorough review of PAH
concentrations in various environmental compartments and their use for assessing
environmental risks and possible effects on ecosystems and human health.
July 2009
9-221
DRAFT-DO NOT CITE OR QUOTE

-------
Deposition and fate of PAH has been an important area of research. Because they are
carcinogenic, PAHs are important environmental contaminants. Root-soil behavior of PAHs
is an area of active study. Soil-bound PAHs are associated with soil organic matter and are
therefore generally not easily available for root uptake. PAHs are readily adsorbed to root
surfaces but there seems to be little movement to the interior of the root or movement up to
the shoots (Gao and Zhu, 2004, 155782). Paddy rice is the main food crop planted in China.
As an aquatic plant having aerial roots, the movement of PAHs into rice roots may be
different than their movement into more widely studied land-grown food crops. PAH
concentrations in the rice roots were more correlated with the water and air compartments
than with the soil (Jiao et al., 2007, 155879).
The total PAH concentration in grasses adjacent to a highway have been measured to
be about eight times higher than in grasses from reference sites not close to a highway
(Crepineau et al., 2003, 155741). Howe et al. (2004, 155854) found that concentrations of
PAHs and hexachlorobenzene (HCB) in spruce (Picea spp.) needles at 36 sites in eastern
Alaska varied by an order of magnitude. Samples collected near the city of Fairbanks
generally had higher concentrations than samples collected from rural areas. The relative
importance of combustion sources versus petrogenic sources was highest in the near-coastal
areas, as reflected in variation in the concentration of ratios of isomeric PAHs.
Use of flame retardants has increased in recent years in response to fire product
safety regulations. However, some flame retardant chemicals are toxic and are readily
transported atmospherically. Use of some has been banned in Europe and some of the
United States because of their persistence and tendency to bioaccumulate (Hoh et al., 2006,
190378).
Base Cations
With respect to ecosystem effects from PM deposition, the inclusion of base cations
(especially Ca, Mg, and K) in atmospheric deposition is generally considered to be a positive
effect. Base cations are important plant nutrients that are in some locations present in
short supply and that are further depleted by the acidic components of deposition.
Increased base cation deposition can help to ameliorate adverse effects of acidification of
soils and surface waters and reduce the toxicity of inorganic Al to plant roots and aquatic
biota. These topics are covered in detail in the recent 2000 NOxSOx ISA (U.S. EPA, 2008,
157074). Calcium supply is also well known to be important for breeding success in
July 2009
9-222
DRAFT-DO NOT CITE OR QUOTE

-------
passerine bird species. Eggshell thickness, egg size, clutch size, and hatchability of pied
flycatcher (Ficedula hypoleuca) were found to be depressed near the Cu smelter at
Harjavalta, SW Finland (Eeva and Lehikoinen, 2004, 155762). Availability of Ca-rich food
to the birds was estimated by counting snail shells in the nests postfledging. The number of
snail shells correlated positively with the Ca concentration of nestling feces and adult
breeding success. In addition, the negative impact of Cu on the number of fledglings was
stronger at locations where Ca concentration was low (Eeva and Lehikoinen, 2004, 155762).
Although the effects of base cation deposition inputs to terrestrial ecosystems are
most commonly considered to be positive, under very high base cation deposition, plant
health can be adversely affected. Dust that is high in base cations can settle on leaves and
other plant structures and remain for extended periods of time. This is especially likely in
arid environments because rainfall can serve to wash dry deposited materials off the
foliage. Extended dust coverage can result in a variety of adverse impacts on plant
physiology (Grantz et al., 2003, 155805). For example, van Heerden et al. (2007, 156131)
documented decreased chlorophyll content, inhibition of CO2 assimilation, and uncoupling
of the oxygen-evolving complex in desert shrubs exposed to high limestone dust deposition
near a limestone quarry in Namibia.
Based on the Integrated Forest Study (IFS) data, the U.S. EPA (2004, 056905)
concluded that particulate deposition has a greater effect on base cation inputs to soils than
on base cation losses associated with the inputs of sulfur, nitrogen, and H+. These
atmospheric inputs of base cations have considerable significance, not only to the base
cation status of these ecosystems, but also to the potential of incoming precipitation to
acidify or alkalize the soils in these ecosystems. This topic is discussed in detail in the
recent NOxSOx ISA (U.S. EPA, 2008, 157074).
9.4.2.3. Magnitude of Dry Deposition
Using Vegetation for Estimating Atmospheric Deposition
Whereas direct real-time measurement of deposition or air concentrations of
atmospheric contaminants is desirable, it is not always practical (Howe et al., 2004,
155854). Instead, passive time-integrative methods are frequently used. These can involve
analysis of vegetative tissues as a record of pollutant exposure, or analysis of lake sediment
cores or ice cores to determine changes in pollutant input over time. There is a general
July 2009
9-223
DRAFT-DO NOT CITE OR QUOTE

-------
assumption that the concentration of an analyte in vegetation reflects the time-integrated
concentration of that analyte in the air. The development of deposition layers in sediment
or ice cores allows the possibility of determining the effects of changes in the atmospheric
concentration over periods of years, decades, or longer.
Biomonitoring methods are important in air pollution assessment and provide a
complement for more typical instrumental analyses. It is well known that mosses can
accumulate large amounts of heavy metals in response to atmospheric deposition. Mosses
accumulate dissolved materials and PM deposited from the atmosphere and have been used
extensively in Europe as surrogate collectors for estimating bulk (wet plus dry) deposition
of metals. The ease and low cost of this method has enabled regional assessments to be
conducted throughout Europe.
Despite its wide use, however, several papers have pointed out complications in the
use of mosses to quantify metal deposition rates. Zechmeister (1998, 156178) found that the
uptake efficiency for 12 heavy metals in 3 species of moss was similar, but that uptake
efficiency in a fourth species was uncorrelated with the other species for about half the
metals considered. Zechmeister (1998, 156178) also showed that productivity of an
individual species can vary greatly among sites. To calculate atmospheric deposition of
metals from accumulation in mosses, both the metal concentration and the rate of biomass
production is needed. Further complication was shown in the study of Shakya et al. (2008,
156081), which revealed that accumulation of Cu, Zn and Pb decreased chlorophyll content.
Sites with greater deposition amounts may therefore have lower rates of productivity than
cleaner sites.
Differences in uptake efficiencies among species and productivity among sites has led
to the use of a single moss species placed in mesh bags that can be distributed to areas
where that species of moss does not grow naturally. Studies to standardize this passive
deposition monitoring approach have been limited. Adamo et al. (2007, 155644) evaluated
the effects of washing with water, oven drying, and acid washing as preteatments and
found little difference in uptake efficiencies, although the ratio of the collecting surface area
to mass was found to be a key factor in uptake efficiency.
Couto et al. (2004, 155739) investigated dry versus bulk deposition of metals using
transplanted moss bags. This study showed that at some sites dry deposition exceeded bulk
deposition, a likely outcome of wash-off of dry deposited particles. This study also
documented intercationic displacement and leaching as a result of acidic precipitation. The
July 2009
9-224
DRAFT-DO NOT CITE OR QUOTE

-------
authors concluded that the accumulated metal concentration represented an unstable
equilibrium between inputs and outputs of elements that were a function of the local
environment and weather during the exposure period. They also concluded that it was not
possible to extrapolate calibrations between metal accumulation in moss and atmospheric
deposition of metals to areas with different weather conditions, precipitation pH, and air
contaminant concentrations. Zechmeister et al. (2003, 157175) also presented results
demonstrating the problems with dry deposited particles that can be washed off by rain.
These studies indicate that moss is not a completely effective collector of total particle
deposition. Deposition estimates from moss accumulation probably represent values that
fall between wet deposition and total deposition.
A European moss biomonitoring network has been in place since 1990 (Harmens et
al., 2007, 155828). Sampling surveys are repeated every 5 yr. The survey conducted in
2005/2006 occurred in 32 countries at over 7000 sites. The network reports metal
concentrations associated with live moss tissue. Trends analysis of these data showed
statistically significant decreases over time in moss concentrations for As, Cu, V, and Zn.
Trends were not observed for Cr, Fe, or Ni. Results for individual countries participating in
the survey have also been published. In Hungary, major pollution sources were readily
detected by moss sampling (Otvos, 2003). Somewhat higher metal concentrations in mosses
in 1997 than in other European countries were attributed to the use of a different moss
species in the Hungarian survey (Otvos, 2003). Similar sampling in Romania showed
regions with contamination that were among the highest in Europe. These results were
consistent with known air quality problems in Romania (Lucaciu et al., 2004, 155947).
Because particulate deposition is not well characterized using this method, spatial patterns
and temporal trends for particulate metal deposition in Europe only provide crude
estimates of relative deposition patterns.
The use of moss to assess heavy metal deposition has received much less attention in
the U.S. than in Europe. A study conducted in the Blue Ridge Mountains, VA, found that
metal concentrations in moss were related to elevation and canopy species at some sites
(Schilling and Lehman, 2002, 113075). However, metal concentrations in moss were not
related to concentrations in the O horizon of the soil. Other measurement methods for trace
metal deposition were not available to compare with moss concentrations.
Epiphytic lichens have also been used to evaluate heavy metal accumulation. Helena
et al. (2004, 155833) found substantially increased concentrations of metals in lichens
July 2009
9-225
DRAFT-DO NOT CITE OR QUOTE

-------
transplanted from a relatively clean region to an area in proximity to a metal smelter. The
presence of specific species of bryophyte or lichen can serve as an effective bioindicator of
metal contamination (Cuny et al., 2004, 155742). In some studies, tree bark has been used
as a biomonitor for atmospheric deposition of heavy metals (Baptista et al., 2008, 155673;
Pacheco and Freitas, 2004, 156011; Rusu et al., 2006, 156062).
Biomonitoring using mosses, lichens, or other types of vegetation has been well
established as a means of identifying spatial patterns in atmospheric deposition of heavy
metals in relation to power plants, industry, and other point and regional emissions
sources. More recently, a number of studies (Lopez Alonso et al., 2002, 155943; 2003,
155944; 2003, 155945) have used cattle that have been reared predominantly on local
forage as a means of monitoring atmospheric inputs of Cu, Ar, Zn, and Hg. For example, Hg
emissions from coal fired power plants in Spain had a substantial effect on Hg
accumulation by calves (Lopez Alonso et al., 2003, 155944). Accumulation of Hg by cattle
extended to -140-200 km downwind from the source.
Yang and Zhu (2007, 156168) investigated the effectiveness of pine needles as passive
air samplers for SOCs, such as PAHs, that are partially or completely particle-associated in
the atmosphere. PAH distribution patterns are complicated by their properties, which span
a broad range of octanol-air partition coefficients. This allows them to be present in both
vapor and particle phases. In addition, the air-plant partitioning of PAHs is affected by air
temperature and atmospheric stability (Yang and Chen, 2007, 092847). DeNicola et al.
(2005, 155747) documented the suitability of a Mediterranean evergreen oak (Quercus ilex)
to serve as a passive biomonitor for atmospheric contamination with PAH in Italy.
Deposition to Canopies
Tree canopies have been shown to increase dry deposition from the atmosphere,
including deposition of PM. Dry deposition rates in the canopy are commonly estimated by
the difference between throughfall deposition and deposition measured by an open collector,
although the use of this approach to specifically quantify particulate deposition is
complicated by gaseous deposition to leaf surfaces and, for some elements, leaching and
uptake. Avila and Rodrigo (2004, 155664) found that trace metal deposition in throughfall
in a Spanish oak forest were higher than bulk deposition for Cu, Pb, Mn, V, and Ni, but not
for Cd and Zn. This study also found that dry deposition of Cu, Pb, Zn, Cd and V occurred,
but that canopy uptake of Zn and Cd also occurred. Leaching of Mn and Ni from the foliage
July 2009
9-226
DRAFT-DO NOT CITE OR QUOTE

-------
was observed as well. Leaching of Ni, Cu, Mn, Rb, and Sr from a red spruce-balsam fir
canopy by acidic cloud water was also measured in a study by Lawson et al. (2003, 089371).
These studies suggest that leaching of trace metals from forest canopies varies with tree
species and the acidity of precipitation. Throughfall therefore cannot be assumed to
represent total deposition of heavy metals without evaluating uptake and leaching at the
specific study site.
Physical models have provided an alternative to estimating dry deposition to canopies
with throughfall measurements. Recently, Pryor and Binkowski (2004, 116805) identified
an additional complication in that models typically hold particle size constant.
Nevertheless, there maybe significant modification of particle size distributions during the
deposition process. Condensation processes in the vicinity of the canopy can increase
particle size and may explain discrepancies between observations and modeled dry
deposition that is based on air sampling of particulates above the canopy.
The use of pine and oak canopies as bioindicators of atmospheric trace metal pollution
was investigated by Aboal et al. (2004, 155642). Metal concentrations in leaves were found
to be one to three orders of magnitude lower than in mosses collected in this study. As an
ecosystem pool, metals in leaves were likely to be much more important than those in
mosses. The authors concluded, however, that these tree species were not effective
bioindicators of atmospheric deposition of heavy metals.
The effectiveness of tree canopies in capturing particulates was investigated as a
method for improving air quality by Freer-Smith et al. (2004, 156451). This study showed
that with consideration of planting design, location of pollution source, and tree species,
planting of trees can be effective at reducing particulate air pollution. However, this
approach does not address the possible effects of the captured pollution on trees, soils and
surface waters.
High-elevation forests generally receive larger particulate deposition loadings than
equivalent low elevation sites. Higher wind speeds at high elevation enhance the rate of
aerosol impaction. Orographic effects enhance rainfall intensity and composition and
increase the duration of occult deposition. High-elevation forests are often dominated by
coniferous species with needle-shaped leaves that enhance impaction and retention of PM
delivered by all three deposition modes.
July 2009
9-227
DRAFT-DO NOT CITE OR QUOTE

-------
Deposition to Soil
As with mosses, accumulation of heavy metals in surface soils provides a general
reflection of the spatial distribution of industrial pollution. The distribution of toxic
elements in urban soils has been an important area of study (Madrid et al., 2002, 155956;
Markiewicz Patkowska et al., 2005, 155963). Generally, Cu, Pb, Zn, and Ni have
accumulated in urban soils compared with their rural counterparts (Yuangen et al., 2006,
156174). In the study of Romic and Romic (2003, 156055), relationships were found between
urban activities and concentrations of metals in soils in developed areas surrounding
Zagreb, Croatia. Goodarzi et al. (2002, 155801) compared deposition estimated by moss
bags to concentrations of metals in A-horizon soils in the vicinity of a large smelter.
Statistically significant correlations were observed between the moss bag deposition
estimates and the soil metal concentrations for Cd, Pb, Zn, and in some cases also Cu.
These correlations suggested that atmospheric deposition of metals caused elevated metal
concentrations in the upper mineral horizon of these soils. No correlations were found for
Hg or As in this study.
Studies have also looked at metal accumulation in peat because of the tendency of
most metals to be immobilized through binding with organic matter. Steinnes et al. (2005,
156095) presented geographical patterns of metal concentrations in surface peat
throughout Norway that corresponded to pollution sources, although the peat samples were
collected in 1979. Zaccone et al. (2007, 179930) found that variations of metal
concentrations with depth in a single Swiss peat core corresponded with the depositional
history that would be expected from the industrial revolution, although Cs137 activity
exhibited a distribution in the profile that was not fully consistent with the Chernobyl
nuclear reactor accident. A detailed study of Finish peat showed that relationships between
depth profiles of metal concentrations and deposition history can match well for some
metals at some sites, but not well for the same metals at other sites (Roberts et al., 2003,
156051). They also found that Zn and Cd accumulation rates were independent of
deposition history at each of three study sites.
Metal deposition to soil is also a significant concern adjacent to roadways. Urban
stormwater can be rich in heavy metals and other contaminants derived from atmospheric
deposition, and can be a major source of pollutant inputs to water bodies in urban settings.
Urban stormwater runoff can also be toxic to aquatic biota, partly due to trace metal
concentrations (Greenstein et al., 2004, 155808! Sabin et al., 2005, 088300; Schiff et al.,
July 2009
9-228
DRAFT-DO NOT CITE OR QUOTE

-------
2002, 156959). These processes are largely a function of the impervious nature of much of
the ground surface in urban areas (i.e., buildings, roads, sidewalks, parking lots,
construction sites). Dry-deposited pollutants can build up, especially in arid and semi-arid
environments, and then be washed into surface waters with the first precipitation event.
The concentrations of Cd, Ca, Cu, Pb, and Zn in road runoff were found to be significantly
higher during winter in Sweden. This seasonal pattern was attributed to the intense
wearing of the pavement that occurred during winter due to the use of studded tires in
combination with chemical effects of deicing salts (Backstrom et al., 2003, 156242).
9.4.3. Direct Effects of PM on Vegetation
Exposure to airborne PM can lead to a range of phytotoxic responses, depending on
the particular mix of deposited particles. This was well known at the time of the previous
PM criteria assessment, as summarized below. Most direct phytotoxic effects occur in
severely polluted areas surrounding industrial point sources, such as limestone quarries
and other mining activities, cement kilns, and metal smelting facilities (U.S. EPA, 2006,
090110). Experimental application of PM constituents to foliage typically elicits little
response at the more common ambient concentrations. The diverse chemistry and size
characteristics of ambient PM and the lack of clear distinction between effects attributed to
phytotoxic particles and to other air pollutants further confound understanding of the
direct effects on foliar surfaces.
Deposition of PM can cause the accumulation of heavy metals on vegetative surfaces.
Low solubility limits foliar uptake and direct heavy metal toxicity because trace metals
must be brought into solution before they can enter into the leaves or bark of vascular
plants. In those instances when trace metals are absorbed, they are frequently bound in
leaf tissue and are lost when the leaf drops off (Hughes, 1981, 156578).
Depending on the size of the particles, the PM deposited on the leaf surface can affect
the plant's metabolism and photosynthesis by blocking light, obstructing stomatal
apertures, increasing leaf temperature and altering pigment and mineral content (Naidoo
and Chirkoot, 2004, 190449) (see Section 9.4.3.1.). Fine PM has been shown to enter the
leaf through the stomata and penetrate into the mesophyll layers where it alters leaf
chemistry (Da Silva et al., 2006, 190190). Kuki et al. (2008, 155346) also showed increased
leaf permeability and increased activity of enzymes in response to fine PM lead (see
discussion in Section 9.4.5.).
July 2009
9-229
DRAFT-DO NOT CITE OR QUOTE

-------
Studies of the direct toxic effects of particles on vegetation have not yet advanced to
the stage of reproducible exposure experiments. In general, phytotoxic gases are deposited
more readily assimilated more rapidly and lead to greater direct injury of vegetation as
compared with most common particulate materials. The dose-response functions obtained
in early experiments following the exposure of plants to phytotoxic gases generally have not
been observed following the application of particles (EPA, 2004, 056905).
9.4.3.1. Effects of Coarse-mode Particles
The current state of scientific knowledge regarding the direct effects of coarse PM on
plants has not changed since publication of the previous PM criteria assessment (EPA,
2004, 056905). The summary provided here is taken from that report. In many rural areas
and some urban areas, the majority of the mass in the coarse particle mode derives from
the elements silicon, aluminum, calcium, and iron, suggesting a crustal origin as fugitive
dust from disturbed land, roadways, agriculture tillage, or construction activities. Large
particles tend to deposit near their source (Grantz et al., 2003, 155805) and rapid
sedimentation of coarse particles tends to restrict their direct effects on vegetation largely
to roadsides and forest edges, which often receive the greatest deposition (EPA, 2004,
056905)
Dust
Dust can cause both physical and chemical effects. Consequences are often mediated
via impacts on leaf cuticles and waxes. Deposition of inert PM on above-ground plant
organs sufficient to coat them with a layer of dust may result in changes in radiation
received, a rise in leaf temperature, and the blockage of stomata. Crust formation can
reduce photosynthesis and the formation of carbohydrates needed for normal growth,
induce premature leaf-fall, damage leaf tissues, inhibit growth of new tissue, and reduce
starch storage. Dust may decrease photosynthesis, respiration, and transpiration! and it
may result in the condensation and reactivity of gaseous pollutants with PM, thereby
causing visible injury symptoms and decreased productivity (EPA, 2004, 056905). Leaves
with trichomes may be more prone to the accumulation of dust on leaf surfaces (Kuki et al.,
2008, 155346).
The chemical composition of PM is usually the key phytotoxic factor leading to plant
injury. For example, cement-kiln dust liberates calcium hydroxide on hydration. It can then
July 2009
9-230
DRAFT-DO NOT CITE OR QUOTE

-------
penetrate the epidermis and enter the mesophyll, causing an increase in leaf surface pH. In
turn, surface pH can be important for surface microbial colonization and wax formation and
degradation.
Salt
Sea-salt particles can serve as nuclei for the absorption and subsequent reaction of
other gaseous and particulate air pollutants. Direct effects on vegetation reflect these
inputs and salt injury caused by the sodium and chloride that constitute the bulk of these
particles. The source of most salt spray is aerosolized ocean water. Sea salt can cause
damage to plants! however, it is not covered in this assessment because it is not of
anthropogenic origin. However particulate salt may be input to an ecosystem from deicing
salt.
Injury to vegetation from the application of deicing salt is caused by salt spray blown
or drifting from the highways (Viskari and Karenlampi, 2000, 019101). The most severe
injury is often observed nearest the highway. Conifers planted near roadway margins in the
eastern U.S. often exhibit foliar injury due to toxic amounts of saline aerosols deposited
from deicing solutions (EPA, 2004, 056905).
Exposure of vegetation to atmospheric PM deposition can lead to varying levels of
effects, depending on PM deposition amounts and the chemical make-up of the deposited
materials. Nevertheless, most of the well-documented examples of direct PM effects have
been caused largely by coarse particles deposited in close proximity to industrial point
sources, such as limestone quarries, cement kilns, and metal smelting facilities (Grantz et
al., 2003, 155805). Fine particles tend to have wider atmospheric distribution, and their
direct effects have not been as clearly demonstrated.
9.4.4. PM and Diffuse Light Effects
Atmospheric PM can affect ambient radiation, which can be considered in both its
direct and diffuse components. Foliar interception by canopy elements occurs for both up-
and down-welling radiation. Therefore, the effect of atmospheric PM on atmospheric
turbidity influences canopy processes both by radiation attenuation and by changing the
efficiency of radiation interception in the canopy through conversion of direct to diffuse
radiation (Hoyt, 1978, 046638). Diffuse radiation is more uniformly distributed throughout
the canopy and increases canopy photosynthetic productivity by distributing radiation to
July 2009
9-231
DRAFT-DO NOT CITE OR QUOTE

-------
lower leaves. The enrichment in photosynthetically active radiation (PAR) present in
diffuse radiation appears to offset a portion of the effect of an increased atmospheric albedo
due to atmospheric particles.
The effects of regional haze on the yield of crops because of reduction in solar
radiation were examined by Chameides et al. (1999, 011184) in China, where regional haze
is especially severe. They estimated that approximately 70% of crops were being depressed
by at least 3 to 5% by regional scale air pollution and its associated haze (Chameides et al.,
1999, 011184; EPA, 2004, 056905).
The net effect of PM on photosynthesis depends on the balance between the reduction
in total PAR (which decreases photosynthesis) and the increase in the diffuse fraction of
PAR (which tends to increase photosynthesis). The 'global dimming' period occurred
between 1950 and 1980 and was characterized by reduced PAR and increased diffuse light
caused by anthropogenic aerosols. Mercado et al. (2009, 190444) estimate the effects of
variations in diffuse light on the terrestrial carbon sink during the last century using a
global model. The results indicate that the terrestrial carbon sink increased by
approximately 25% during the "global dimming'" period, likely driven by increased diffuse
light despite decreased PAR. However under a future scenario in which S( ) r and BC
aerosols decline, the diffuse-radiation and the associated terrestrial C sink also decline.
9.4.5. Effects of Trace Metals on Ecosystems
Trace metals may enter the ecosystems as both fine and coarse particles. All but 10 of
the 90 elements that comprise the inorganic fraction of the soil occur at concentrations of
<0.1% (1000 pg/g) and are termed "trace" elements or trace metals. Trace metals with a
density greater than 6 g/cm3, referred to as "heavy metals" (e.g., Cd, Cu, Pb, Cr, Hg, Ni, Zn),
are of particular interest because of their potential toxicity to plants and animals. Although
some trace metals are essential for vegetative growth or animal health, they are all toxic in
large quantities. Most trace elements exist in the atmosphere in particulate form as metal
oxides (Ormrod, 1984, 046892). Aerosols containing trace elements derive predominantly
from industrial activities. Generally, only the heavy metals Cd, Cr, Ni, and Hg are released
from stacks in the vapor phase (McGowan et al., 1993, 046731). Atmospherically deposited
PM can interact with a variety of biogeochemical processes. The potential pathways of
accumulation of trace metals in terrestrial ecosystems, as well as the possible consequences
of trace metal deposition on ecosystem functions, are summarized in Figure 9"84 (U.S. EPA,
July 2009
9-232
DRAFT-DO NOT CITE OR QUOTE

-------
2004, 056905). A number of mass balance approaches (Macleod et al., 2005, 155954; Toose
and Mackay, 2004, 156123), and metal speciation and transport models (Bhavsar et al.,
2004, 155689; Bhavsar et al., 2004, 155690; Gandhi et al., 2007, 155781) have been
developed in recent years.
Atmospheric Pb is a component of PM in some regions. The effects of Pb on
ecosystems were discussed in the 2006 Pb AQCD (U.S. EPA, 2006, 090110). which
concluded that, due to the deposition of Pb from past human practices (e.g., leaded gasoline,
ore smelting) and the long residence time of Pb in many aquatic and terrestrial ecosystems,
a legacy of environmental Pb burden exists, over which is superimposed much lower
contemporary atmospheric Pb loadings. The potential for ecological effects of the combined
legacy and contemporary Pb burden to occur is a function of the bioavailability or
bioaccessibility of the Pb. This, in turn, is highly dependent upon numerous site factors
(e.g., soil OC content, pH, water hardness). Although the more localized ecosystem impacts
observed around smelters are often striking, effects generally cannot be attributed solely to
Pb, because of the presence of many other stressors (e.g., other heavy metals, oxides of
sulfur and nitrogen) that can also act singly or in concert with Pb to cause readily
observable environmental impacts (U.S. EPA, 2008, 157074; EPA, 2004, 056905).
Effects of fine particle trace elements were described by the U.S. EPA (2004, 056905).
and some additional more recent research has also been conducted, especially on the topic
of vegetative uptake of trace elements from the soil. The state of scientific understanding as
presented by the U.S. EPA (2004, 056905) as well as a discussion of more recent research
findings are presented below.
July 2009
9-233
DRAFT-DO NOT CITE OR QUOTE

-------
9. Retranslocation
Atmosphere
1. Wet/Dry Deposition
Above-
Ground
Storage,
Metabolism
2. Foliar Uptake
Plant Surface
Phyllosphere
3, Lrtterfall, Res us pension,
Deposition, Leaching,
Stem Flow
8, Volatile Loss
4. Translocation
Biologically
Available >
Biologically
Unavailable
10, Root
\ Turnover
5. Mass Flow,
_ Diffusion
11. Mineralization
IX.
Soil Organic
Upper Soil
VI.
Root Storage,
Metabolism
7. Leaching
Rhizosphere/
Rhizoplane
6. Root
Uptake
12. Weathering
VII.
Lower Soi
Primary
Minerals
5. Mass Flow,
Diffusion
7. Leaching
VIII.
Groundwater
Source: U.S. EPA (2004, 056905)
Figure 9-84. Relationship of plant nutrients and trace metals with vegetation. Compartments (roman
numerals) represent potential storage sites; whereas arrows (Arabic numerals)
represent potential transfer routes.
9.4.5.1. Direct Effects of Metals
Direct effects of trace elements on vegetation can result from their deposition and
residence on foliar surfaces. Direct metal phytotoxicity can occur only if the metal can move
from the surface into the leaf or directly from the soil into the root. Low solubility limits
entry into plant tissue. Trace metals absorbed into leaf tissue are eventually transferred to
the soil litter layer where they can affect litter decomposition, an important source of soil
nutrients. Fungi and microorganisms living on leaves aid in leaf decomposition after leaves
are dropped to the forest floor. Changes in litter decomposition processes in response to
July 2009
9-234
DRAFT-DO NOT CITE OR QUOTE

-------
metal toxicity can influence nutrient cycling in the soil and limit the supply of essential
nutrients.
Foliar Effects
Da Silva et al. (2006, 190190) have shown that PM had anatomical and physiological
effects on plants growing near an iron pelletization factory in Brazil. The effects of PM
occurred due to foliar uptake. Structural characteristics such as peltate trichomes, probably
formed a barrier lessening the penetration of metallic iron into the mesophyll in some
species. Iron was shown to penetrate the trichomes, epidermic cells (adaxial and abaxial
surfaces), stomata, xylem cells, collenchyma, endodermis and mesophyll tissues. Once
entering the stomata, the PM penetrates within the mesophyll, it may modify the chemical
balance of the mesophyll (Da Silva et al., 2006, 190190).
A greenhouse study evaluated the combined effects of iron dust on restinga vegetation
(coastal vegetation of Brazil) that commonly grows near iron ore industries simulated field
conditions. Kuki et al. (2008, 155346) found that iron dust had differing effects on gas
exchange, chlorophyll content, iron content and antioxidant enzyme activity on two plant
species common to the restinga. Schinus terebinthifolius (an invasive exotic in the U.S.)
was not affected by the iron dust. However, Sophora tomentosa showed increased iron
content and membrane permeability to the leaves, increased activity of antioxidant
enzymes. These results showed that the plants used different strategies to cope with PM
pollution, S. terebinthifolius avoided stress, while S. tomentosa used antioxidant enzyme
systems to partially neutralize oxidative stress.
Toxicity to Mosses and Lichens
At the time of the most recent air quality criteria report for PM (EPA, 2004, 056905),
trace metal toxicity to lichens had been demonstrated in relatively few cases. Nash (1975,
016763) documented Zn toxicity in the vicinity of a Zn smelter near Palmerton, PA.
Experimental data had suggested that lichen tolerance to Zn and Cd generally ranges
between 200 and 600 ppm (Nash TH, 1975, 016763).
The effects of deposited metals on the mosses have not been well studied. Tremper et
al. (2004, 156126) exposed mosses of two species to roadside conditions and sampled them
over a period of three months. Under field conditions, chlorophyll concentrations in moss
tissue were not affected by metal contamination and accumulation
July 2009
9-235
DRAFT-DO NOT CITE OR QUOTE

-------
Mosses and lichens readily take up metals from atmospheric deposition. Otnyukova
(2007, 156009) demonstrated vertical gradients within a coniferous forest canopy in the
fruticose lichen genus Usnea with respect to lichen thallus morphology and heavy metal
concentration. Abnormal thalli at the tree-top level contained higher concentrations of Al,
Fe, Zn, F, Sr, and Pb. This vertical pattern within the tree canopy is in general accordance
with known deposition of PM to plants (Otnyukova, 2007, 156009).
There is an extensive literature on the use of mosses and lichens for estimating
deposition (biomonitors) and indicating metal exposure in ecosystems (bioindicators) (see
Section 9.4.2.3.).
9.4.5.2. Effects on Soil Chemistry
Trace metals are naturally found in small amounts in soils, ground water, and
vegetation. Many are essential micronutrients required for growth by plants and animals.
Naturally occurring mineralization can produce metal concentrations in soils and
vegetation that are high compared to atmospheric sources. Many metals are bound by
chemical processes in the soil, reducing their availability to biota. However, epiphytic or
parasitic root colonizing microorganisms can solubilize and transport metals for root uptake
(Lingua et al., 2008, 155935). It can be difficult to assess the extent to which observed
heavy metal concentrations in soil are of anthropogenic origin. This is because soil parent
material, pedogenesis, and anthropogenic inputs all influence the amounts and distribution
of trace elements in soil. Trace element concentrations in some natural soils that are remote
from air pollution can be higher than soils derived from other parent materials that receive
anthropogenic inputs (Burt et al., 2003, 155709). The general effects of metals from
atmospheric deposition are presented in the following discussion.
There is not a standard method available for quantifying the bioavailability of heavy
metals in soil. A variety of models, isotopic studies, and sequential extraction methods have
been used (Shan et al., 2003, 156972; Collins et al., 2003, 155737; Feng et al., 2005,
155774). Total metal concentration in soil does not give a good indication of potential
biological effects because soils vary in their ability to bind metals in forms that are not
bioavailable. There are various methods available for assessing bioavailability of metals,
but soils are heterogeneous and there is no ideal method for evaluating what conditions the
soil biota experience. Almas et al. (2004, 155654) argued that the actual measurement of
biological effects is the best criterion for determining bioavailability. In particular, the
July 2009
9-236
DRAFT-DO NOT CITE OR QUOTE

-------
replacement of metal-sensitive microorganisms by metal-tolerant organisms within each
functional group may be one of the most sensitive indicators of metal exposure. An increase
in microbial trace metal tolerance per se would not be problematic if it was not for the fact
that this increase in tolerance is generally accompanied by a decrease in microbial diversity
(Lakzian et al., 2002, 156671; Almas AR et al., 2004, 155654).
Heavy metals deposited from the atmosphere to forests accumulate either in the
organic forest floor or in the upper mineral soil layers and metal concentration tends to
decrease with soil depth. The accumulation of heavy metals in soil is influenced by a variety
of soil characteristics, including pH, Fe and Al oxide content, amount of clay and organic
material, and cation exchange capacity (CEC) (Hernandez et al., 2003, 155841). Thus, the
pattern of distribution of heavy metals in soils depends on both the soil characteristics and
the metal characteristics.
Burt et al. (2003, 155709) investigated the concentrations and chemical forms of trace
metals in smelter-contaminated soils collected in the Anaconda and Deer Lodge Valley area
of Montana, one of the major mining districts of the world for over a century (1864-1983).
The relative distributions of trace metals within the more soluble soil extraction forms were
similar to their respective total concentrations. This suggested a relationship between the
concentrations of total trace elements and concentrations of soluble mobile fractions.
Sequential extractions do not provide direct characterization of trace metal speciation, but
rather an indication of chemical reactivity (Burt et al., 2003, 155709; Ramos et al., 1994,
046736). Soluble and exchangeable forms are considered readily mobile and bioavailable.
Those bound to clay minerals or organic matter are considered generally unavailable.
There is concern that Pb contamination of forest soil could move into groundwater.
This would be important in view of the large quantity of Pb deposited from the atmosphere
in the 1960s and 1970s in response to combustion of leaded gasoline. This issue was
investigated by Watmough et al. (2004, 077809) who applied a stable isotope (207Pb) to the
forest floors of white pine (Pinus strobus) and sugar maple (Acer saccharuni) stands. Added
Pb was rapidly lost from the forest floor, likely due to high litter turnover in these forest
types. However, Pb concentrations in the upper 30 cm of mineral soil were strongly
correlated with soil OM, suggesting that Pb does not readily move down the soil profile to
the ground water, but rather is associated with the organic content of the upper soil layers
(Watmough et al., 2004, 077809).
July 2009
9-237
DRAFT-DO NOT CITE OR QUOTE

-------
The upper soil layers are typically an active site of litter decomposition and plant root
uptake, both processes may be affected by metal components of PM. Surface litter
decomposition is reduced in soils having high metal concentrations. This is likely due to the
sensitivity to metals of microbial decomposers and reduced palatability of plant litter
having high metal concentration (Johnson and Hale, 2008, 155881). Root decomposition is a
key component of nutrient cycling. Johnson and Hale (2008, 155881) measured in situ fine
root decomposition at Sudbury, Ontario and Rouyn-Noranda, Quebec. Elevated soil metal
concentrations (Cu, Ni, Pb, Zn) did not necessarily reduce fine root decomposition. Only at
sites having high concentrations of metals did decomposing roots show increased metal
concentrations over time.
9.4.5.3. Effects on Soil Microbes and Plant Uptake via Soil
Upon entering the soil environment, PM pollutants can alter ecological processes of
energy flow and nutrient cycling, inhibit nutrient uptake, change ecosystem structure, and
affect ecosystem biodiversity. Many of the most important effects occur in the soil. The soil
environment is one of the most dynamic sites of biological interaction in nature. It is
inhabited by microbial communities of bacteria, fungi, and actinomycetes. These organisms
are essential participants in the nutrient cycles that make elements available for plant
uptake. Changes in the soil environment that influence the role of the bacteria and fungi in
nutrient cycling determine plant and ultimately ecosystem response.
Many of the major indirect plant responses to PM deposition are chiefly soil-mediated
and depend on the chemical composition of the individual components of deposited PM.
Effects may result in changes in biota and in soil conditions that affect ecological processes,
such as nutrient cycling and uptake by plants.
The soil environment is rich in biota. Bacteria and fungi are usually most abundant in
the rhizosphere, the soil around plant roots that all mineral nutrients must pass through.
Bacteria and fungi benefit from the nutrients that are present in root exudates and make
mineral nutrients available for plant uptake. The soil-mediated ecosystem impacts of PM
are largely determined by effects on the growth of bacteria and mycorrhizal fungi that are
involved in nutrient cycling and plant nutrient uptake.
July 2009
9-238
DRAFT-DO NOT CITE OR QUOTE

-------
Soil Nutrient Cycling
Accumulation of heavy metals in litter can interfere with nutrient cycling.
Microorganisms are responsible for decomposition of organic matter, which contributes to
soil fertility. Toxic effects on the microflora can be caused by Zn, Cd, and Cu. The U.S. EPA
(2004, 056905) judged that addition of only a few mg of Zn per kg of soil can inhibit
sensitive microbial processes. Enzymes involved in the cycling of N, P, and S (especially
arylsulfatase and phosphatase) seem to be most affected (Kandeler et al., 1996, 094392).
Soil organic matter cycling is known to be sensitive to disturbance due to heavy metal
pollution. This can cause increased litter accumulation at sites close to metal emissions
point sources. The relative importance of the various processes that might be responsible
for this observation is poorly known. Boucher et al. (2005, 155699) conducted CO2 evolution
studies in microcosms having metal-rich and metal-poor plant materials. Their results
suggested that there was a pool of less readily decomposable C that appeared to be
preferentially preserved in the presence of high metal (Zn, Pb, Cd) concentrations in the
leaves of the metallophyte Arabidopsis halleri. An additional possibility is that increased
lignification of the cell walls increased the amount of insoluble C (Mayo et al., 1992,
155974).
Yuangen et al. (2006, 156174) found that urban soil basal respiration rates were
positively correlated with soil acetic acid-extractable Cd, Cu, Ni, and Zn. The soil microbial
biomass was negatively correlated with the concentrations of Pb fractions, but not with
other metals. Overall microbial biomass was lower for urban soils as compared with rural
soils (Yuangen et al., 2006, 156174).
Metal Toxicity to Microbial Communities
It is believed that increased accumulation of litter in metal-contaminated areas is due
to the effects of metal toxicity on microorganisms. Smith (1991, 042566) reported the effects
of Cd, Cu, Ni, and Zn on the symbiotic activity of fungi, bacteria, and actinomycetes. In
particular, the formation of mycorrhizae has been shown to be reduced when Zn, Cu, Ni,
and Cd were added to the soil.
Most studies of the effects of heavy metals on soils have been conducted under
laboratory conditions. However, Oliveira and Pampulha (2006, 156827) performed a field
study to evaluate long-term changes in soil microbiological characteristics in response to
heavy metal contamination. Dehydrogenase activity, soil ATP content, and enumeration of
July 2009
9-239
DRAFT-DO NOT CITE OR QUOTE

-------
major soil microbial groups illustrated the effects of contamination. There was a marked
decrease in total numbers of the different microbial groups. In particular, asymbiotic
nitrogen-fixers and heterotrophic bacteria were found to be sensitive. Dehydrogenase
activity was confirmed to be a good assay for determining the effect of heavy metals on
physiologically active soil microbial biomass.
The toxic effects of heavy metals on soil microorganisms are well known. However,
less is known about the relative sensitivity of different types of soil microorganisms
(Rajapaksha et al., 2004, 156035). Vaisvalavicius et al. (2006, 157080) assessed the toxicity
of high concentrations of Pb (839 mg/kg), Zn (844 mg/kg), and Cu (773 mg/kg) in the upper 0
to 0.1 m soil layer. Microbial abundance of all groups was reduced and enzymatic activity
was lower than for uncontaminated soil. In particular, actinomycetes, oligonitrophobic and
mineral N assimilating bacteria were most affected.
Effects of heavy metals in soil on microbes depends on soil pH, organic content, and
the type of heavy metal exposure (Kucharski and Wyszkowska, 2004, 156662). Some studies
have shown that heavy metals inhibit microbial activity in soil (Smejkalova et al., 2003,
156987; Vasundhara et al., 2004, 156133). However, Wyszkowska et al. (2007, 179948)
showed that heavy metals can either inhibit or stimulate the growth of soil microbes.
Populations of Azotobacter spp. decreased, but populations of oligotrophic and copiotrophic
bacteria, actinomyces, and fungi increased in response to heavy metal exposure. Acute
metal stress causes a decrease in microbial biomass as metal-sensitive microbes are
inhibited (Joynt et al., 2006, 155887).
Studies of the impacts of metal stress on the microbial community composition in soil
have generally been based on microbial culturing techniques that can select only a subset of
the natural soil population of microbes. More recent culture-independent studies have been
conducted using phospholipids or nucleic acid biomarkers to reveal information regarding
changes in microbial community structure (Joynt et al., 2006, 155887). Using this approach,
Joynt et al.(2006, 155887) demonstrated that soils contaminated with both metals (Pb, Cr)
and organic solvent compounds over a period of several decades had undergone changes in
community composition, but still contained a phytogenetically diverse group of bacteria.
This may reflect adaptation to the potentially toxic conditions through such processes as
natural selection, gene exchange, and immigration. Comparison between a severely
contaminated soil with a similar soil that had much lower amounts of contamination
showed considerably lower microbial diversity in the contaminated soil, particularly for
July 2009
9-240
DRAFT-DO NOT CITE OR QUOTE

-------
asymbiotic nitrogen fixers and heterotrophic bacteria (Oliveira and Pampulha, 2006,
156827).
As pollution increases, it is expected that the more sensitive species will be lost and
the more tolerant species remain. This gives rise to the concept of pollution-induced
community tolerance (PICT), which has been demonstrated for populations of bacteria and
fungi (Davis et al., 2004, 155744). These researchers assessed the effects of long-term Zn
exposure on the metabolic diversity and tolerance to Zn of a soil microbial community
across a gradient of Zn pollution. PICT was found to correlate better with total soil Zn than
with the concentration of Zn in soil pore water.
Soil Microbe Interactions with Plant Uptake of Metals
Atmospherically-deposited metals accumulate in upper soil horizons where fine roots
are most developed. The availability for plant uptake of metals in soil depends on metal
speciation and soil pH. In addition, metal binding to dissolved organic matter (DOM)
reduces bioavailability (Sauve, 2001, 156948). Because organic matter typically decreases
with soil depth, the affinity of metals for organic matter can influence metal bioavailability
at different soil depths. Fine roots (<2 mm diameter) provide the major site of uptake and
transport to the above-ground plant and generally contain a large proportion of the total
metals found in plants (Gordon and Jackson, 2000, 155802).
Fine roots are often colonized by mycorrhiza and interact with other soil microbes.
Recent published evidence supports that mycorrizha and bacteria influence plant uptake
and tolerance of metals. Mycorrhiza are fungi that colonize plant roots to form a symbiosis.
Mycorrhiza uptake nutrients from the soil and transfer them to the plant in exchange for
carbon from the plant. Like plants, some species and strains of mycorrhiza are more
tolerant of metals in the soil (Ray et al., 2005, 190473), so that unpolluted and polluted
sites may host different species and strains of mycorrhiza (Vogel-Mikus et al., 2005,
190501).
Mycorrhiza have been observed to cause a range of effects on plants. In some cases,
plants colonized with mycorrhiza showed improved nutrient uptake and decreased metal
uptake (Berthelsen et al., 1995, 078058; Vogel-Mikus et al., 2006, 190502; Nogueira et al.,
2004, 190460). Mycorrhiza have been shown to accumulate metals and act as a sink
(Berthelsen et al., 1995, 078058; Carvalho et al., 2006, 155715) often preventing the metals
in the roots from allocation to shoots (Kaldorf et al., 1999, 190399; Zhang et al., 2005,
July 2009
9-241
DRAFT-DO NOT CITE OR QUOTE

-------
192083; Soares and Siqueira, 2008, 190482). For example, estuarine salt marshes are often
located close to urban and industrial areas and receive elevated amounts of trace metal
contaminants from point and nonpoint (including atmospheric deposition) sources.
Vegetation is important in the retention and accumulation of heavy metals in salt marshes.
Carvalho et al. (2006, 155715) conducted experiments on the effects of arbuscular
mycorrhizal fungi (AMF) on the uptake of Cd and Cu by Aster tripolium, a common plant
species in polluted salt marshes and a host of AMF. Carvalho et al. (2006, 155715) found
that AMF colonization increased metal accumulation in the root system of Aster tripolium
without enhancing translocation to the shoot. By trapping toxic metals in the roots, this
plant species may reduce the extent of vegetative stress caused by metal exposure and act
as an effective sink for these metals. In a review paper Christie et al. (2004, 190171)
concluded that mycorrhiza may directly improve plant tolerance to metals by binding and
immobilizing metals and indirectly improve plant tolerance by improving uptake of
nutrients that increase plant growth.
There is recent evidence that bacteria and mycorrhiza act together to improve plant
tolerance to metals. Like mycorrhiza some bacteria are more tolerant to metals than others
(Vivas et al., 2003, 190499). Combined inoculation of Trifnlium sp. by the arbuscular
mycorrhiza Glomus mosseae and the bacterium Brevivacillus sp. conferred tolerance to Cd
by increasing nutrient status and rooting development and by creasing Cd uptake by the
plant (Vivas et al., 2003, 190499). A similar results was observed for Zn uptake (Vivas et al.,
2006, 190500).
In some cases mycorhizzae will not prevent metal uptake (Weissenhorn et al., 1995,
073826). In fact mycorrhiza may facilitate the accumulation of metals in plants and
mycorrhiza will enhance the translocation of metals from the root to the shoot (Vogel-Mikus
et al., 2005, 190501; Citterio et al., 2005, 190176; Zimmer et al., 2009, 192085). There is
evidence of variable responses depending on the combination of mycorrhiza and bacteria
species. Zimmer et al. (2009, 192085) recently showed that the willow tree Salix sp.
colonized with the ectomycorrhizal fungus Hebeloma crustuliniforme and the bacteria
Microcuccus luters increased total Cd and Zn accumulation due to enhanced mycorrhizal
formation. In these cases where soil microbes cause increased metal accumulation, there is
a potential to use the system for phytoremediation.
Plants also vary in the extent to which they take up heavy metals from the soil.
Variability has been shown to occur in response to different plant species and different
July 2009
9-242
DRAFT-DO NOT CITE OR QUOTE

-------
metals. For example, Szabo and Fodor (2006, 156109) exposed winter wheat (Triticum
aestivuni), maize (Zen mays) and sunflower (Helianthus annuus) to a variety of
micro-pollutants. Cadmium accumulation was significant in both vegetative and
reproductive plant parts. Vegetative winter wheat accumulated substantial amounts of Hg,
but the other species did not. Lead, Cu, and Zn showed only moderate enrichment in crops
(Szabo and Fodor, 2006, 156109).
There is some evidence to support that shallowrooted plant species are most likely to
take up metals from the soil (Martin and Coughtrey, 1981, 047727). However, there is little
evidence confirming this observation. It may be more likely that shallow roots of species are
likely to take up metals because the metal often accumulates in shallow soil layers. Even
though atmospheric PM will usually deposit on soils before being taken up by plants, it
could also be deposited to aquatic systems with subsequent transfer to terrestrial plants.
Contamination of stream sediments by heavy metals can impact adjacent terrestrial
ecosystems when high flows cause resuspension and subsequent streamside deposition of
sediment particles. For example, Ozdilek et al. (2007, 156010) showed that metal
concentrations in vegetation along the Blackstone River in Massachusetts and Rhode
Island were generally inversely related to the distance from the riverbank, with higher
metal concentrations in plant tissues located near the river. The ability of plants to take up
metals from soil is an important part of metal cycling in the environment. This uptake
process allows the metals to enter the food web, where they might exert mutagenic,
carcinogenic, and teratogenic effects (Hunaiti et al., 2007, 156579).
9.4.5.4. Plant Response to Metals
Some metals, including Cu, Co, Ni, and Zn, are essential micronutrients needed for
plant growth. Others, including Hg, Cd, and Pb are not essential for plants. Though all
heavy metals can be directly toxic at sufficiently high concentrations, only Cu, Ni, and Zn
have been documented as being frequently toxic to plants (EPA, 2004, 056905), while
toxicity due to Cd, Co, and Pb has been observed less frequently (Smith, 1990,
046896).Toxic doses depend on the type of ion, ion concentration, plant species and the
stage of plant growth (Memon and Schroder, 2009, 190442). Toxicity response is also
dependent on the nutritional status of the plant and the development of mycorrhizae
(Strandberg et al., 2006, 156105). Plants respond to high concentrations of metals in soil
through a variety of mechanisms and there are substantial differences among plant species
July 2009
9-243
DRAFT-DO NOT CITE OR QUOTE

-------
in their response to heavy metal exposure. Mechanisms of metal tolerance included
exclusion or excretion rates, genetics (Patra et al., 2004, 081976; Yang et al., 2005, 192104),
mycorrhizal interactions (Gohre and Paszkowski, 2006, 190355), storage capability and
accumulation (Clemens, 2006, 190179), and various cellular detoxification mechanisms
(Hall, 2002, 190365; Gratao et al., 2005, 190364).
One of the most important mechanisms that increases plant tolerance to metals is
chelation with phytochelatins, such as metallothioneins and peptide ligands that are
synthesized within the plant from glutathione (Memon and Schroder, 2009, 190442).
Phytochelatins are intracellular metal-binding peptides that act as specific indicators of
metal stress. Because they are produced by plants as a response to sublethal concentrations
of heavy metals, they can indicate that heavy metals play a role in forest decline (Gawel et
al., 1996, 012278). Phytochelatin concentrations have previously been measured in
coniferous trees in the northeastern U.S. The U.S. EPA (2004, 056905) and Grantz et al.
(2003, 155805) summarized studies indicating that both the number of dead red spruce
trees and phytochelatin concentrations increased sharply with elevation in the
northeastern U.S. Red spruce stands showing varying degrees of decline indicated a
systematic and significant increase in phytochelatin concentrations associated with the
extent of tree injury. These data suggest that metal stress might contribute to tree injury
and forest decline in the northeastern U.S. The extent to which low to moderate amounts of
heavy metal deposition, which might occur at locations that are not in close proximity to a
large point source, contribute to adverse impacts on forest vegetation is not known.
Although the phytochelatin data suggest a linkage, more direct experimental data would be
needed to confirm such a finding.
In general plant growth is negatively correlated with trace metal and heavy metal
concentration in soils and plant tissue (Audet and Charest, 2007, 190169). Trace metals,
particularly heavy metals can influence forest growth. Growth suppression of foliar
microflora has been shown to result from Fe, Al, and Zn. These three metals can also inhibit
fungal spore formation, as can Cd, Cr, Mg, and Ni (see Smith, 1990, 046896). Metals cause
stress and decreased photosynthesis (Kucera et al., 2008, 190408) and disrupt numerous
enzymes and metabolic pathways (Strydom et al., 2006, 190486). Excessive concentrations
of metals result in phytotoxicity through: (i) changes in the permeability of the cell
membrane! (ii) reactions of sulfydryl (-SH) groups with cations! (iii) affinity for reacting
July 2009
9-244
DRAFT-DO NOT CITE OR QUOTE

-------
with phosphate groups and active groups of ADP or ATP; and (iv) replacement of essential
ions (Patra et al., 2004, 081976).
In addition to disrupting photosynthesis and other metabolic pathways, metals have
been shown to alter frost hardiness and impair nutrition. A recent review by Taulavuori
(2005, 190489) suggests that metal- induced stress reduces frost hardiness of plants, a
particular concern at high elevation sites. Kim et al. (2003, 155899) found decreased
concentration of K in needles and Ca in stems of Pinus sylvestris seedlings exposed to Cd
addition. This response suggests a disturbance of nutrition in response to Cd.
Pollutant-caused needle loss can reduce the interception of pollutants from the atmosphere,
and therefore reduce their concentrations in stemflow. This may be responsible for the
observation that species diversity of lichens is sometimes higher on trees affected by
die-back (Hauck, 2003, 155830).
Plant foliage can accumulate elemental Hg over time in response to air exposure and
concentrations in soil (Frescholtz et al., 2003, 190352; Ericksen et al., 2003, 155769). A
mesocosm experiment was conducted by Ericksen et al. (2003, 155769) where aspen trees
were grown in gas-exchange chambers in Hg-enriched soil (12.3 ± 1.3 pg/g) and the Hg
content in the foliage was analyzed. Foliar Hg increased with leaf age for two to three
months and then stabilized at leaf concentrations near 150 ng/g. About 80% of the Hg found
in above-ground biomass was present in the leaves. The concentration of Hg in trees grown
in the same mesocosms in containers of low Hg soil (0.03 ± 0.01 pg/g) exhibited foliar Hg
concentrations that were similar to those of trees grown in Hg-enriched soil. Almost all of
the foliar Hg originated from the atmosphere. Clearly, plant foliage can be a major sink for
airborne Hg, which can subsequently enter the soil after litterfall (Ericksen et al., 2003,
155769). However, this study did not determine the extent to which atmospheric Hg was
dry-deposited on the foliage, as opposed to gaseous uptake through the stomata. Foliar/air
Hg exchange has been shown to be dynamic and bi-directional (Millhollen et al., 2006,
190447). These investigators compared foliar Hg accumulation over time in three tree
species with fluxes measured using a plant gas-exchange system subsequent to soil
amendment with HgCb. Root tissue Hg concentrations were strongly correlated with soil
Hg concentrations, suggesting that below-ground accumulation of Hg by roots may be an
important process in the biogeochemical cycling of Hg in soil systems. Nevertheless,
measured foliar Hg fluxes indicated that deposition of atmospheric Hg constituted the
dominant flux of Hg to the leaf surface (Millhollen et al., 2006, 190447). Grigal (2003,
July 2009
9-245
DRAFT-DO NOT CITE OR QUOTE

-------
155811) also found that Hg in vegetation is derived almost exclusively from the atmosphere.
Mercury uptake from soil is limited, partly because roots adsorb Hg but transport it to
foliage very poorly (Grigal, 2002, 156498). Grigal (2003, 155811) provided a thorough review
of the sequestration of Hg in forest and peatland ecosystems. A fundamental aspect of Hg
cycling is its strong relationship to organic matter. For that reason, peatlands sequester
much larger quantities of Hg than would be expected on the basis of their land area. Thus,
if global climate change affects C storage, it may indirectly affect Hg storage because of the
strong relationship between Hg and organic matter (Grigal, 2003, 155811).
Arbuscular mycorrhizal (AM) fungi can play important roles in mitigating toxicity of
heavy metals in plants. For example, AM symbiosis is known to be involved in plant
adaptation to As-contaminated soils. Higher plants that are adapted to As contaminated
soils are generally associated with mycorrhizal fungi (Gonzalez-Chavez et al., 2002,
155800). It has also been shown that AM symbioses can influence plant coexistence and
community diversity (O'Connor, 2002). Some plants associated with AM fungi can
successfully colonize sites that are heavily contaminated by heavy metals (Pennisi, 2004,
156018).
Dong et al. (2008, 192106) cultivated white clover (Trifolium repens) and ryegrass
(Lolium perenne) in As-contaminated soil (water extractable As 82.7 mg/kg). The growth
and P nutrition of both species largely depended on AM symbiosis. The AM-inoculated
plants showed selective uptake and transfer of P over As.
PM pollution has the potential to alter species composition over long time scales. Kuki
et al. (2009, 190111) showed that early establishment stages of Sophora tomentosa species
were negatively affected by the combination of iron ore and acidifying particles. The
deleterious effects of the PM included deficient germination and toxic concentrations in
roots. In contrast, S. terebinthifolius was not affected by the PM revealing species
resistance to the pollution. The difference among species response suggests that over a long
time period the imbalance will likely change the species composition (Kuki et al., 2009,
190111).
The process of removing toxins from soil or water using photoautotrophs is referred to
as phytoremediation. Some plant species have good ability to extract heavy metals from
soil, thereby offering potential for phytoremediation (Clemens, 2006, 190179; Hooda, 2007,
190382; Padmavathiamma and Li, 2007, 190465) . For example, several species of willow
CSk/ixspp.) accumulate high concentrations of Zn and Cd in aboveground biomass
July 2009
9-246
DRAFT-DO NOT CITE OR QUOTE

-------
(Lunackova et al., 2003, 155948; Meers et al., 2007, 155977; Rosselli et al., 2003, 156058). A
first estimation of the order of magnitude of potential metal removal by willow was 2 to 27
kg/ha/yr of Zn and 0.25 to 0.65 kg/ha/yr for Cd (Meers et al., 2007, 155977). Build-up of high
concentrations of trace metals in soil is difficult to remediate because of the long residence
times of metals in the environment. Plants that survive on heavy metal-contaminated soils
have been studied to elucidate the mechanisms that allow them to tolerate such conditions
and interactions between soil contamination and vegetation composition (Becker and
Brandel, 2007, 156260; Hall, 2002, 190365). There are numerous other plants that have
been investigated for application to phytoremediation. Plants that hyperaccumulate metals
have special potential for remediation of metal-contaminated sites. About 400 species have
been reported. Brassicaceae has the largest numbers of taxa, with 11 genera and 87 species
known to hyperaccumulate one or more metal contaminants (Prasad and De Oliveira
Freitas, 2003, 156885).
Plant uptake is often the first step for a metal to enter higher levels of the food web.
Consumers of vegetation may often receive heavy loading of metals from their diets. Metals
may also bioaccumulate in some species and tissue concentrations are magnified at the
higher trophic levels, so-called biomagnifications (see Section 9.4.5.7.on Biomagnification).
9.4.5.5. Effects on Aquatic Ecosystems
The atmospheric deposition of PM into the ocean has important implications for
primary productivity and carbon sequestration. This is because metals in PM deposition
limit phytoplankton growth in parts of the ocean (Crawford et al., 2003, 156370). In
particular, Fe and Zn can influence the productivity of algae that are involved in CaC03
production. The production of both particulate organic C and CaC03 drive the ocean's
biological carbon pump (Shulz et al., 2004, 156087). Thus, in oceanic areas of trace metal
limitation, changes in trace metal atmospheric deposition can affect biogenic calcification,
with potential consequences for CO2 partitioning between the ocean and atmosphere.
A study by Sheesley et al. (2004, 156084) illustrated the value ofbioassay procedures
to provide an initial screening of ambient PM toxicity. They used two species of green algae
and two extraction methods to compare the toxicities of atmospheric PM collected at two
urban/industrial sites and one rural site near the southern shore of Lake Michigan.
Toxicities varied by site, by extraction solvent, and by bioassay. Results suggested that
toxicity was not related to the total mass of PM in the extract, but to the chemical
July 2009
9-247
DRAFT-DO NOT CITE OR QUOTE

-------
components of the PM. It is noteworthy that the concentrations of contaminants in PM in
this type of short-term and acute toxicity testing are much higher than would be found in
the natural environment. Thus, the purpose of this type of testing is to provide an initial
screening-level comparison of relative toxicities of atmospheric PM from different source
areas. It does not provide the data that would be needed to assess risk (Sheesley et al.,
2004,156084)-
9.4.5.6. Effects on Animals
There has been little work focusing on animal indicators of PM effects in the field.
However, there have been several recent studies on snails, amphibians, earthworms, and
bivalves that are discussed below.
Bioindicator organisms can be especially useful for monitoring PM effects over
geographical and temporal scales. Terrestrial invertebrates have been used to monitor
contaminants in both air and soil. Snails {Helix spp.) accumulate trace metals and
agrochemicals, and can be used as effective biomonitors for urban air pollution (Beeby and
Richmond, 2002, 155680)(Viard et al., 2004, 055675)(Regoli et al., 2006, 156046).
Demonstrated biological effects include growth inhibition, impairment of reproduction, and
induction of metallothioneins that are involved in metal detoxification (Gomot'de Vaufleury
and Kerhoas, 2000,155798)(Regoli et al., 2006, 156046). The use of sentinel species to
detect the effects of complex mixtures of air pollutants is of particular value because the
chemical constituents are difficult to characterize, exhibit varying bioavailability, and are
subject to various synergistic effects.
Regoli et al. (2006, 156046) caged land snails (Helix aspersa) at five locations in the
urban areas of Ancona, Italy. After four weeks of exposure to ambient air pollution, the
snails were analyzed for trace metals and PAHs. Biomarkers were measured that
correlated with contaminant accumulation, including concentrations of metallothioneins,
activity of biotransformation enzymes, and peroxisomal proliferation. In addition,
indicators of oxidative stress were measured, such as oxyradical scavenging capacity, onset
of cellular damage, and loss of DNA integrity. Results documented substantial
accumulation of metals and PAHs in snail digestive tissues in urban areas having high
traffic congestion. Cellular reactivity was also found, suggesting that this species is an
effective bioindicator for multipollutant air quality and PM monitoring.
July 2009
9-248
DRAFT-DO NOT CITE OR QUOTE

-------
Some amphibian ecotoxicological research has focused on heavy metal exposure.
Contaminant uptake can occur by oral, pulmonary, and dermal exposure (Lambert, 1997,
155916)(Johnson et al., 1999, 155880)(James et al., 2004, 155874). This is potentially
important because of documented declines in amphibian populations in the U.S. and
elsewhere in recent decades (Houlahan et al., 2000, 155853). Toads were shown to be fairly
tolerant of Cd exposure (James et al., 2004, 155874). It is not clear whether current
amounts of terrestrial metal contamination pose an increased risk to amphibians in
general.
Estuarine and marine bivalves provide potential bioindicators for Hg
bioaccumulation. For example, Coelho et al. (2006, 190181) investigated Hg concentrations
in Scrobiculariaplana, a long-lived, deposit-feeding bivalve in southern Europe. Annual
bioaccumulation rates were shown to be strongly correlated with Hg concentrations in
suspended particulate matter (SPM), a response to their deposit-feeding tactics (Verdelhos
et al., 2005, 190497). The ability to predict annual accumulation rates for indicator species,
such as this bivalve, may facilitate management actions to avoid deleterious effects on
humans through consumption of bivalves above a certain age/size class.
Earthworms often constitute a large percentage of soil animal biomass and they are
considered to be relatively sensitive indicators of soil metal contamination. They are
continuously exposed to the soil via dermal contact in the soil solution or ingestion of large
quantities of soil pore water, polluted food and/or soil particles (Lanno et al., 2004, 190415).
Hobbelen et al. (2006, 190371) determined the important metal pools for bioaccumulation
by earthworms Lumbricus rubellus, which live in the upper 5cm of soil and Aporrectodea
caliginosa, which live in the upper 25 cm of soil. Soil concentration explained much of
earthworm concentrations, however Cd concentration in A. calinginosa was best explained
by pore water concentrations and no variable tested explained Zn tissue concentrations.
Massicotte et al. (2003, 155968) compared the cell viability and phagocytic potential of
three earthworm species (Lumbricus terrestris, Eisenia andrei, and Aporrectodea
tuberculata) in response to atmospheric emissions of metals from a cement factory in
Quebec, Canada. Cell viability actually increased in proximity (0.5 km) to the cement
factory for A. tuberculata, and this might have been due to beneficial effects of increased Ca
deposition. There were no significant differences observed for the other two species
(Massicotte et al., 2003, 155968).
July 2009
9-249
DRAFT-DO NOT CITE OR QUOTE

-------
Biogeochemical cycling of Hg in the Arctic has been investigated, in part because
observed Hg concentrations in marine animals may pose health risks for local human
populations. The lifetime of gaseous elemental Hg (GEM) in the atmosphere, which
constitutes about 95% of atmospheric Hg, is generally about 1-yr (Lin and Pehkonen, 1999,
190426). However, during spring (typically March through June), the lifetime of GEM in
the Arctic is much shorter, and atmospheric GEM can be depleted in less than one day
during atmospheric Hg depletion episodes (MADE) (Lindberg et al., 2002, 190429) (Skov et
al., 2004, 190481). During the AMDE, GEM is rapidly oxidized to reactive gaseous Hg that
can be deposited to the ground surface (Skov et al., 2004, 190481). Because of the increased
solar flux to the Arctic during spring and seasonal melting of sea ice, there may be an
increased efficiency of Hg bio accumulation in Arctic food webs than would be expected
based on data collected at mid-latitudes. Skov et al. (2004, 190481) developed a simple
parameterization for AMDE and included it in the Danish Eulerian Hemispheric Model
(DEHM). The model was shown to reproduce the general structure of AMDE, suggesting
that the limiting factor for AMDE may be the surface temperature of sea ice.
9.4.5.7. Biomagnification across trophic levels
Biomagnification is the progressive accumulation of chemicals with increasing trophic
level (LeBlanc, 1995, 155921). Organic Hg is the most likely metal to biomagnify, in part
because organisms can efficient assimilate methylmercury and it is slowly eliminated
(Croteau et al., 2005, 156373; Reinfelder et al., 1998, 156047). Of the trace metals, there is
also evidence that Cd, Pb, Zn, Cu and Se biomagnify.
The study of trophic transfer and biomagnification is limited by the difficulty in
discriminating food webs and the uncertainty associated with assignment of trophic
position to individual species (Croteau et al., 2005, 156373). Use of stable isotopes can help
to establish linkages. However, it is difficult to determine the extent to which
biomagnification occurs in a given ecosystem without thoroughly investigating physiological
biodynamics, habitat, food web structure, and trophic position of relevant species. Thus,
development of an understanding of ecosystem complexity is necessary to determine what
species might be at greatest risk from toxic metal exposure (Croteau et al., 2005, 156373).
July 2009
9-250
DRAFT-DO NOT CITE OR QUOTE

-------
Terrestrial
Bioaccumulation of heavy metals can occur through the plant-herbivore and
litter-detrivore food webs. The U.S. EPA (2004, 056905) concluded that Cd and Zn can
bioaccumulate in earthworms. Other invertebrates inhabiting soil litter may also
accumulate metals. Although food web accumulation of a metal may not result in mortality,
it might reduce breeding potential or result in other non-lethal effects that adversely affect
organism responses to environmental cues.
Metal accumulation in litter can be found mainly around brass works and Pb and Zn
smelters. Organisms that feed on earthworms living in soils with elevated metal
concentrations may also accumulate Pb and Zn. Increased concentrations of heavy metals
have been found in a variety of mammals living in areas with elevated heavy metal
concentrations in the soils.
The transfer of metals from plants to terrestrial snails is an interesting system for
biomagnifications because snails accumulate metals in their soft tissue and can contribute
significantly to the transfer of pollutants to primary consumers and terrestrial predators
(Dallinger et al., 2001, 192109). Notten et al. (2005, 190461) studied the transfer of Cu, Zn,
Cd and Pb in terrestrial soil-plant-snail food chains in metal-polluted soils of the
Netherlands. The food chain included perennial plant species Urtica dioica and the
herbivorous snail Cepaea nemoralis. The transfer of metal from the soil to the plant
compartment was low (coefficient of determination R2 = 0.20). Total concentration of metals
in soils was a poor predictor of leaf concentration. Low metal concentration in the leaves
was thought to be due to low pore water metal concentrations and was also thought to be
partly caused by low translocation from roots within the plant. The Cu, Zn and Cd
concentrations in the snails were always higher than concentrations in the leaves
indicating bioaccumulation. The metal transfer from the leaf to snail was highest among all
routes tested, suggesting that transfer from diet is important. Similar results were found by
Beeby and Richmond (2002, 155680) with the snail Helix aspersa and the plant Taraxacum
sp. for Zn, Pb, Cd, but not for Cu.
Many types of predators including shrews, thrushes and beetle larvae include snails
as part of their diet (Gomot-De Vaufleury and Pihan, 2002, 190357; Seifert et al., 1999,
190480). Seifert et al. (1999, 190480) found the shrews eating snails with elevated Cd had
critical levels of Cd in their kidneys. Scheifler et al. (2007, 190379) found that Cd in snails
lead to toxic levels in beetle larvae that caused increased amounts of mortality.
July 2009
9-251
DRAFT-DO NOT CITE OR QUOTE

-------
Aquatic
In general, it has been assumed that metal biomagnification in aquatic ecosystems is
an exception rather than the rule (Gray, 2002, 155806). More recent research has
demonstrated aquatic biomagnification of certain metals. For example, Stewart et al. (2004,
156097) used stable isotopes of C and N to show biomagnification of Se in San Francisco
Bay food webs. Croteau et al. (2005, 156373) identified trophic position of estuarine
organisms and food web structure in the delta of San Francisco Bay to document Cd
biomagnification in invertebrates that live on macrophytes and also in fish. Concentrations
of Cd were biomagnified 15 times within two trophic links in each food web. In contrast, no
tendency towards biomagnification was observed for Cu.
In aquatic ecosystems, biomagnification of trace metals does not necessarily occur.
Nguyen et al. (2005, 155997) found biodiminution for most metals in Lake Balaton,
Hungary, with the exception of slight enrichment of Zn from PM to zooplankton and of Cd
from sediment to mussel.
Once transported to aquatic ecosystems, trace metals often preferentially bind to
sediment particles. Some of these sediment-bound metals may be unavailable to biota! in
contrast, metals bound to sediment organic matter may exhibit varying degrees of
bioavailability (Di Toro et al., 2005, 155750). Piol et al. (2006, 156028) studied the
bioavailability of sediment-bound Cd to the freshwater oligochaete Lumbriculus variegatus.
They found that Cd uptake depended on the amount of free dissolved Cd(II), and the Cd
contribution from sedimentary particles to biological uptake was negligible.
Marine bivalve mollusks bioaccumulate trace metals and other contaminants
(LaBrecque et al., 2004, 155913) and therefore may be used as bioindicators of
contamination. In addition, they constitute an important link to human health by virtue of
their importance as a food source (Cheggour et al., 2005, 155723! Li et al., 2002, 156691).
9.4.5.8. Effects near Smelters and Roadsides
The high PM concentrations in proximity to mining, smelting, roadsides and other
industrial sources result in heavy metal loadings that may be particularly damaging to
nearby ecosystems.
July 2009
9-252
DRAFT-DO NOT CITE OR QUOTE

-------
Smelters
The Harjavalta region is one of the most intensively studied heavy metal polluted
areas in the world. Kiikkila et al. (2003, 156637) reviewed available data on heavy metal
deposition and environmental effects in this area. Emissions from the smelter were as high
as 1100 t/yr of dust, 140 t/yr Cu, 96 t/yr Ni, 162 t/yr Zn, and 94 t/yr Pb in 1987. Deposition
amounts decreased substantially after 1990, to only a few percent of the amounts that
occurred during the 1980s.
Kiikkila (2003, 156637) investigated the effects of heavy metal pollution in proximity
to a Cu-Ni smelter at Harjavalta, Finland. The deposition of heavy metals increased within
30 km of the smelter. Only slight changes in the understory vegetation were observed at
distances greater than 8 km from the smelter. At 4 km distance, species composition of
vegetation, insects, birds, and soil microbiota changed and tree growth was reduced. Within
about 1 km, only the most resistant organisms were surviving.
The number of soil organisms clearly decreased and their community structure was
altered close to the Harjavalta smelter (Kiikkila, 2003, 156637). However, this effect was
only pronounced within about 2 km of the smelter. This suggests that the soil microfauna
are relatively resistant to metal pollution effects.
Soil microbial activity decreased close to the Harjavalta smelter (Kiikkila, 2003,
156637), as reflected by microbial respiration, distribution of species within physiological
groups, and microbial and fungal biomass. The fungi appeared to be more sensitive to metal
contamination than the bacteria (Pennanen et al., 1996, 156016). The rate of litter
decomposition decreased, causing an accumulation of needle litter on top of the forest floor
near the smelter (Fritze et al., 1989, 079635).
Inhibition of nutrient cycling and displacement by Cu and Ni of base cations from
cation exchange sites on the soil resulted in a decrease in base cation concentrations in the
organic soil layer (Derome and Lindroos, 1998, 155749; Kiikkila, 2003, 156637) close to the
Harjavalta smelter. In addition, Mg, Ca, and Mn concentrations in Scots pine needles were
low, and this was attributed by Kiikkila (2003, 156637) to the toxic effects of Cu and Ni to
plant fine roots and also to ectomycorrhizal root tips (Helmisaari et al., 1999, 155836).
Nutrient translocation during fall was also affected close to the smelter! as a consequence
needle concentrations of K were relatively high (Nieminen et al., 1999, 155998).
Tree growth (Scots pine) has been poor (Malkonen et al., 1999, 155961) and most
vegetation was absent within 0.5 km of the smelter. Effects on plant species occurrence
July 2009
9-253
DRAFT-DO NOT CITE OR QUOTE

-------
close to the smelter were almost entirely negative. In contrast, some animal species
responded positively, including a leaf miner, three species of aphid, and some ants, beetles,
and spiders.
Salemaa et al. (2004, 156069) investigated heavy metal concentrations in understory
plant species growing at varying distances from the Harjavalta Cu-Ni smelter. Heavy metal
concentrations (except Mn) were highest in bryophytes, followed by lichens, and were
lowest in vascular plants. Vascular plants are generally able to restrict the uptake of toxic
elements, and therefore were able to grow closer to the smelter than lichens. A pioneer moss
(Pohlia nutans) was unusual in that it survived close to the smelter despite its
accumulation of high amounts of Cu and Ni.
Changes in breeding success of cavity-nesting passerine birds close to the Harjavalta
smelter were attributed by Kiikkila (2003, 156637) to habitat changes in response to metal
toxicity. In particular, there was an apparent decrease in the proportion of green insect
larvae in the diet of nestlings. In addition, pollution stress was inferred from increased
heavy metals and decreased Ca in the diet of the pied fly catcher (Ficedula hypoleuca)
(Eeva et al., 2000, 155761).
Documentation of effects on individual species, such as was reported above, does not
reveal what the impacts might be on ecosystem function. Nevertheless, the mere fact that
multiple species, operating at different trophic levels, have been shown to be affected by the
ambient deposition in proximity to the smelter suggests that effects on ecosystem function
may indeed have occurred. More research is needed, however, to fully evaluate effects on
function as opposed to abundance of individual species.
Roadsides
Heavy metal particles are important constituents of road dust. These particles
accumulate on the road surface from brake linings, road paint, tire debris, DE, road
construction materials, and catalyst materials. Road dust can be suspended in the
atmosphere and contribute metals to soil, air, and urban runoff (Adachi and Tainosho, 2004,
081380; Davis et al., 2001, 024933; Smolders and Degryse, 2002, 156091). In particular, Zn
oxide comprises 0.4 to 4.3% of tire tread (Smolders and Degryse, 2002, 156091) and tire
wear is a substantial source of environmental Zn pollution. Adachi and Tainosho (2004,
081380) used a field emission screening electron microscope equipped with an energy
dispersive x-ray spectrometer to characterize heavy metal particles embedded in tire dust.
July 2009
9-254
DRAFT-DO NOT CITE OR QUOTE

-------
Samples were classified into four likely source categories, based on cluster analysis. Based
on morphology and chemical composition, the samples were identified as having derived
from yellow paint (CrPbCh particles), brake dust (particulate Ti, Fe, Cu, Sb, Zr, Ba and
heavy minerals [Y, Zr, La, Ce]), and tire tread (Zn oxide).
Since publication of EPA's 2004 PM criteria assessment, some additional research has
been conducted on the effects of windblown PM. Effects on physical, chemical, and
biological attributes of both plants and animals have been documented (Englert, 2004,
087939; Gleason et al., 2007, 155794; Kappos et al., 2004, 087922). Experiments by Gleason
et al. (Gleason et al., 2007, 155794) suggest that most direct effects on plants of windblown
PM originating from on-road surfaces occur within 40 m of the source. Windblown PM from
roads or agriculture can cover plant photosynthetic structures (Sharifi et al., 1999, 156082),
cause impact damage (Armbrust and Retta, 2002, 156225), or interfere with physiological
mechanisms (Burkhardt et al., 2002, 155708). As previously discussed in Section 9.4.4.6
land snails in urban areas have been shown to be a good indicator of traffic pollution.
9.4.6. Organic Compounds
VOCs in the atmosphere are partitioned between the gas and particle phases. As
described by the U.S. EPA (2004, 056905), the partitioning depends on vapor pressure,
temperature, surface area of the particles, and the nature of the particles and of the
chemical being adsorbed. A wide variety of organic contaminants are deposited from the
atmosphere. These include chemicals such as DDT, PCBs, and PAHs.
Important organic atmospheric contaminants are generally those that are transported
long distances in the atmosphere, subsequently deposited into remote locations, and
bioaccumulated to sufficient concentrations that they can affect humans, wildlife, or other
biota (Swackhamer et al., 2004, 190488). Certain physical and chemical properties facilitate
the movement of these contaminants from land and water surfaces into the atmosphere,
provide stability, and enhance accumulation in lipids. Some, including the relatively small
(up to 4 rings) PAHs degrade relatively rapidly in the atmosphere or at the surface
subsequent to atmospheric deposition. Below is a summary of the findings of the U.S. EPA
(2004, 056905), followed by discussion of more recent research findings.
Plants may be used as passive monitors to compare the deposition of organic
compounds between sites. Vegetation can be used semi-quantitatively to indicate organic
pollutant amounts if the mechanism of accumulation is considered. Organic compounds can
July 2009
9-255
DRAFT-DO NOT CITE OR QUOTE

-------
enter the plant via the roots or be deposited as particles on the leaves and be taken up
through the cuticle or stomata. The pathways depend on the chemical and its physical
properties. These include, for example, lipophilicity, water solubility, vapor pressure, and
Henry's law constant. Environmental conditions can also be important, including
temperature and organic content of soil, plant species, and the foliar surface area and lipid
content.
Organic particulates in the atmosphere are diverse in their makeup and sources.
Vegetation itself is an important source of hydrocarbon aerosols. Terpenes, particularly
a-pinene, 6-pinene, and limonene, released from tree foliage may react in the atmosphere to
form submicron particles. These naturally generated organic particles contribute
significantly to the blue haze aerosols formed naturally over forested areas (Geron et al.,
2000, 019095; U.S. EPA, 2004, 056905).
The low water solubility and high lipo-affinity of many organic xenobiotics control
their interaction with the vegetative components of natural ecosystems. Foliar surfaces are
covered with a waxy cuticle layer that helps reduce moisture loss and short-wave radiation
stress. This epicuticular wax consists largely of long-chain esters, polyesters, and paraffins,
which accumulate lipophilic compounds. Organic air contaminants in the particulate or
vapor phase can be adsorbed to, and accumulate in, the epicuticular wax of leaf surfaces.
Direct uptake of organic contaminants through the cuticle and vapor-phase uptake through
the stomata are not well characterized for most trace organics.
Topographic and vegetative characteristics exert different influence on deposition
modes. In general, dry deposition is most affected by plant morphology (Grantz et al., 2003,
155805). The potential effects of PM on vegetation include the full range of biological
organization, with exposures occurring through the soil and through vegetative surfaces. In
general, soil-mediated exposure is thought to be more significant (Grantz et al., 2003,
155805). Soil acts as an important storage compartment for POPs, including PCBs and
PAHs. There is a continuous process of partitioning between the soil pool and the
atmosphere, and this controls the regional and global transport of these compounds (Backe
et al., 2004, 155668; Wania and Mackay, 1993, 157110). Over time, POPs move towards
equilibrium between the environmental compartments, and this process can be described
using the fugacity concept (Backe et al., 2004, 155668; Mackay, 1991, 042941). Fugacity
reflects the tendency of a chemical constituent to escape one environmental compartment
and move to another. When an equilibrium distribution is achieved, the fugacity quotient
July 2009
9-256
DRAFT-DO NOT CITE OR QUOTE

-------
values in each compartment will be equal. Soil/air partitioning is controlled by a variety of
factors. These include soil properties, such as organic matter content, moisture, porosity,
texture, and structure, as well as the physiochemical properties of the pollutant, including
vapor pressure and water solubility.
The accumulation of PAHs in vegetation, due to their lipophilic nature, could
contribute to human and other animal exposure via food consumption. As a result, plant
uptake of PAHs has been an important area of research (Gao and Zhu, 2004, 155782). Most
bioaccumulation of PAHs by plants occurs by leaf uptake (Tao et al., 2006, 156112). Root
uptake also occurs. It appears that roots preferentially accumulate the lower molecular
weight PAHs due to their greater water solubility (Wild and Jones, 1992, 156155).
Various models have been developed to simulate plant uptake of organic
contaminants. The simple partition-limited model of Chiou et al. (2001, 156342) has been
further expanded to increase complexity and to include root uptake pathways (e.g., Fryer
and Collins, 2003, 156454; Yang et al., 2005, 192104; Zhu et al., 2004, 156184).
In evaluating receptor choice for studies of contaminant exposure to plants, and also
remediation potential, it is important to consider differences among species. For example,
Parrish et al. (2006, 156014) assessed the bioavailability of PAHs in soil. During the first
growing season, zucchini (Cucurbitapepo ssp. pepo) accumulated significantly more PAHs
than did other related plant species, including up to three orders of magnitude greater
concentrations of the six-ring PAHs. Parrish et al. (2006, 156014) also noted differences in
PAH uptake by two different species of earthworm.
The leaves of Quercus ilex have been shown to readily accumulate PAHs in situ.
Young leaves accumulated PAHs within three weeks of bud break. Mature leaves showed
seasonality, with higher PAH concentrations during winter (Alfani et al., 2005, 154319).
It is difficult to discriminate between PAHs that are adsorbed to plant root surfaces as
opposed to those that are actually taken up by the roots. In general, soil bound PAHs are
associated with soil organic matter and are therefore not readily available for root uptake
(Fismes et al., 2002, 141156; Jiao et al., 2007, 155879). Wild et al. (2005, 156156) used
two-photon excitation microscopy to visualize the uptake and transport of two PAHs
(anthracene and phenanthrene) from a contaminated soil into living wheat and maize roots.
Jiao et al. (2007, 155879) developed a sequential extraction method to discriminate between
PAH adsorption in rice roots.
July 2009
9-257
DRAFT-DO NOT CITE OR QUOTE

-------
Maize roots and tops of plants have been shown to directly accumulate PAHs from
aqueous solution and from air in proportion to exposure amounts. Root concentration
factors are log-linear functions of log-based octanol-water partition coefficients (log Kow);
similarly, leaf concentration factors are log-linear functions of log-based octanol-air
partition coefficients (log Koa) (Lin et al., 2007, 155933). Although the bulk concentrations of
PAHs in various plant tissues can differ greatly, the observed differences disappear after
they are normalized to lipid content (Lin et al., 2007, 155933). This suggests that the lipid
content of different plant tissues may influence PAH distribution within the plant.
Previously, there was relatively little information available regarding incorporation of
atmospherically deposited PAHs into aquatic food webs. It is known that PAHs can be
transferred to higher trophic levels, including fish, and that this transfer can be mediated
by aquatic invertebrates, which generally comprise an important part of fish diets. High
mountain lakes offer an effective receptor for quantification of biomagnification in aquatic
ecosystems from atmospheric PM deposition. There are typically no sources of organic
contaminants in their watersheds, and atmospheric inputs dominate as sources of
contamination. In addition, such lakes tend to have relatively simple food webs. Vives et al.
(2005, 157099) investigated PAH content of brown trout (Salmo trutta) and their food items.
Total PAH concentrations tended to be highest in organisms that occupy littoral habitats,
and lowest in pelagic organisms. This may reflect more efficient transfer of PAHs to
underlying sediments in shallower water and associated degradation within the water
column.
Some atmospheric organic contaminants have been shown to accumulate in biota at
remote locations. For example, polybrominated diphenyl ethers (PBDEs), which are man-
made chemicals used as flame retardants in materials manufacturing, have been found to
accumulate in lichens and mosses collected at King George Island, maritime Antarctica
(Yogui and Sericano, 2008, 189971). Because contaminant concentrations were not
statistically different at sites close to and distant from human facilities in Antarctica, the
authors concluded that long-range atmospheric transport was the likely primary source of
PBDEs to King George Island. Law et al. (2003, 190420) reviewed available data for
accumulation of PBDEs and other brominated flame retardants in wildlife. These
compounds have become widely distributed in the environment, including in the deep-
water, oceanic food webs.
July 2009
9-258
DRAFT-DO NOT CITE OR QUOTE

-------
Ohyama et al. (2004, 190462) chose salmonid fish, mainly rainbow trout
(Oncorhyncus mykiss), as an indicator species to evaluate the transport and
bioaccumulation of organochloride compounds in the northern and central Sierra Nevada.
They found that elevation was an important factor affecting residual concentrations of
polychlorinated biphenyls (PCBs) in fish muscle tissue. On this basis, Ohyama et al. (2004,
190462) concluded that PCB residue in rainbow trout, a widely distributed salmonid
species, provided a good monitoring tool for studying the effects of mountainous topography
on the long-range transport and distribution of persistent organic pollutants.
Semivolatile compounds can undergo repeated volatilization on surfaces, such as
plant foliage, in response to diel changes in temperature. As a consequence, such
compounds can be deposited, re-emitted, and re-deposited multiple times. This behavior can
cause these compounds to move large distances in a leap-frog fashion. It is believed that
POPs can be atmospherically transported throughout the world because of their volatility
and response to changes in temperature. This "global distillation theory" (Holmqvist et al.,
2006, 190380; Wania and Mackay, 1993, 157110) predicts that POPs in the northern
hemisphere are generally transported towards the Arctic, and in the southern hemisphere
they are transported toward the Antarctic. In general, POP concentrations measured in the
Arctic are higher than in the Antarctic. They have been detected in all levels of the Arctic
food web (Oehme et al., 1995, 011267). Bioconcentration of organochlorines has been shown
in the Arctic food web, including fish, seals, and polar bears (Oehme et al., 1995, 011267).
Concentrations measured in Arctic polar bears are especially high (Arctic Monitoring and
Assessment Programme, 2004, 190168).
Holmqvist et al. (2006, 190380) measured levels of PCBs in longfin eels (Anguilla
dieffenbachil) in 17 streams on the west coast of South Island, New Zealand. The PCBs
were at low levels, and were believed to originate from atmospheric transport from
industrial areas in Asia. Characteristics of the longfin eel that make it susceptible to
bioaccumulation of lipophilic persistent pollutants include high lipid content (up to 40%),
long lifespan (up to 90 yr), and position near the top of the food chain (Holmqvist et al.,
2006, 190380).
Long-range transport of atmospherically deposited contaminants can be augmented
by biotransport. A good example of this phenomenon was documented by Ewald et al. (1998,
190348), who showed that biotransport by migrating sockeye salmon (Oncorhynchus nerka)
in the Copper River watershed, Alaska, had a greater influence than atmospheric transport
July 2009
9-259
DRAFT-DO NOT CITE OR QUOTE

-------
on bioaccumulation of PCBs and DDT in lake food webs. Organic pollutants accumulated by
salmon during their ocean residence were effectively transferred 410 km inland to their
spawning lake. Arctic grayling (Thymallus arcticus) in the salmon spawning lake were
found to contain organic pollutants more than twice as high as arctic grayling in a near-by
salmon-free lake. The pollutant composition of the grayling in the salmon spawning lake
was similar to that of the migrating salmon (Ewald et al., 1998, 190348). suggesting that
salmon migration contributed to bioaccumulation of organic contaminants in the lake used
for spawning by the salmon.
An assessment of the ecological effects of airborne metals and SOCs was conducted for
eight NPs by the Western Airborne Contaminants Assessment Project (WACAP) (Landers et
al., 2008, 191181). From 2002-2007, WACAP researchers conducted analysis of the
biological effects of airborne contaminants in seven ecosystem compartments^ air, snow,
water, sediments, lichens, conifer needles and fish. The goals were to identify where the
pollutants were accumulating, identify ecological indicators for those pollutants causing
ecological harm, and to determine the source of the air masses most likely to have
transported the contaminants to the parks.
The results from WACAP were summarized by Landers et al. (2008, 191181). which
concluded that bioaccumulation of SOCs were observed throughout park ecosystems.
Vegetation tended to accumulate PAHs, CUPs, and HCHs. Conifer needles were a good
indicator of pesticides, however the ecological consequences of this accumulation are un-
examined. SOCs in vegetation and air showed different patterns, possibly because each
medium absorbs different types of SOCs with varying efficiencies. Mean ammonium nitrate
concentration in ambient fine particulates <2.5 urn diameter was a good predictor of
dacthal, endosulfan, chloradane, trifluralin, DDT and PAH concentrations in vegetation.
Concentrations of SOCs were five to seven orders of magnitude higher in fish tissue
than in sediments. Fish accumulated more PCBs, chlordanes, DDT and dieldrin than
vegetation. Fish lipid and age were the most reliable predictors of SOC concentrations.
Most fish appeared normal during field necropsies! however, individuals with both male
and female reproductive organs were collected at two sites. The incidence of this condition
has increased since the pre-organic pollutant era. Additionally, elevated concentrations of
vitellogenin, a female protein involved in egg production, were found in male fish from
three sites, and directly related to the concentration of several organochlorines at one site.
July 2009
9-260
DRAFT-DO NOT CITE OR QUOTE

-------
The lake sediment records showed steadily increasing mercury deposition over time at
lakes in two parks, Mt. Ranier NP and Rocky Mountain NR Apportionment of the mercury
to its atmospheric sources is not quantified at this time! however, the pattern in the
sediment suggests a local source rather than a global source. Mercury concentrations in fish
exceeded contaminant health thresholds for some piscivorous fish, mammals and birds in
most parks. The average mercury concentration in fish from one site and individual fish
from three additional sites exceeded the U.S. EPA contaminant health thresholds for
humans.
Although this assessment focuses on chemical species that are components of PM it
does not specifically assess the effects of particulate vs. gas-phase forms, therefore in most
cases it is difficult to apply the results to this assessment based on particulate
concentration and size fraction.
9.4.7. Summary of Ecological Effects of PM
Ecological effects of PM include direct effects to metabolic processes of plant foliage!
contribution to total metal loading resulting in alteration of soil biogeochemistry, plant and
animal growth and reproduction! and contribution to total organics loading resulting in
bioaccumulation and biomagnification across trophic levels. These effects were well-
characterized in the 2004 PM AQCD (U.S. EPA, 2004, 056905). Thus, the summary below
builds upon the conclusions provided in that review.
PM deposition comprises a heterogeneous mixture of particles differing in origin, size,
and chemical composition. Exposure to a given concentration of PM may, depending on the
mix of deposited particles, lead to a variety of phytotoxic responses and ecosystem effects.
Moreover, many of the ecological effects of PM are due to the chemical constituents (e.g.,
metals and organics) and their contribution to total loading within an ecosystem.
Investigations of the direct effects of PM deposition on foliage have suggested little or
no effects on foliar processes, unless deposition levels were higher than is typically found in
the ambient environment. However, consistent and coherent evidence of direct effects of PM
has been found in heavily polluted areas adjacent to industrial point sources such as
limestone quarries, cement kilns, and metal smelters (Sections 9.4.3. and 9.4.5.8.). Where
toxic responses have been documented, they generally have been associated with the
acidity, trace metal content, surfactant properties, or salinity of the deposited materials.
July 2009
9-261
DRAFT-DO NOT CITE OR QUOTE

-------
An important characteristic of fine particles is their ability to affect the flux of solar
radiation passing through the atmosphere directly, by scattering and absorbing solar
radiation, and, indirectly, by acting as cloud condensation nuclei (CCN) that, in turn,
influence the optical properties of clouds. Regional haze has been estimated to diminish
surface solar visible radiation. Crop yields can be sensitive to the amount of sunlight
received, and crop losses have been attributed to increased airborne particle concentrations
in some areas of the world. PM has been observed to cause a decrease in photosynthetically
active radiation (PAR) via thick haze occurring in China that decreases plant growth in the
diffuse light portion of PAR. However, a global model showed that PM can increase the
diffuse light fraction of PAR. On a global scale, the diffuse light fraction of PAR has been
shown to increase growth. Consequently, it was shown that when PM is decreased, plant
growth and C storage are also decreased. Further research is needed to determine net
effects of PM alteration of light conditions on the growth of vegetation in the U.S.
The deposition of PM onto vegetation and soil, depending on its chemical composition,
can produce responses within an ecosystem. The ecosystem response to pollutant deposition
is a direct function of the level of sensitivity of the ecosystem and its ability to ameliorate
resulting change. Many of the most important ecosystem effects of PM deposition occur in
the soil. Upon entering the soil environment, PM pollutants can alter ecological processes of
energy flow and nutrient cycling, inhibit nutrient uptake, change ecosystem structure, and
affect ecosystem biodiversity. The soil environment is one of the most dynamic sites of
biological interaction in nature. It is inhabited by microbial communities of bacteria, fungi,
and actinomycetes, in addition to plant roots and soil macro-fauna. These organisms are
essential participants in the nutrient cycles that make elements available for plant uptake.
Changes in the soil environment can be important in determining plant and ultimately
ecosystem response to PM inputs.
There is strong and consistent evidence from field and laboratory experiments that
metal components of PM alter numerous aspects of ecosystem structure and function.
Changes in the soil chemistry, microbial communities and nutrient cycling, can result from
the deposition of trace metals. Exposures to trace metals are highly variable, depending on
whether deposition is by wet or dry processes. Although metals can cause phytotoxicity at
high concentrations, few heavy metals (e.g., Cu, Ni, Zn) have been documented to cause
direct phytotoxicity under field conditions. Exposure to coarse particles and elements such
as Fe and Mg are more likely to occur via dry deposition, while fine particles are more likely
July 2009
9-262
DRAFT-DO NOT CITE OR QUOTE

-------
to contain elements such as Ca, Cr, Pb, Ni, and V. Ecosystems immediately downwind of
major emissions sources can receive locally heavy deposition inputs. Phytochelatins
produced by plants as a response to sublethal concentrations of heavy metals are indicators
of metal stress to plants. Increased concentrations of phytochelatins across regions and at
greater elevation have been associated with increased amounts of forest injury in the
northeastern U.S.
Overall, the ecological evidence is sufficient to conclude that a causal relationship is
likely to exist between deposition of PM and a variety of effects on individual organisms and
ecosystems, based on information from the previous review and limited new findings in this
review. However, in many cases, it is difficult to characterize the nature and magnitude of
effects and to quantify relationships between ambient concentrations of PM and ecosystem
response due to significant data gaps and uncertainties as well as considerable variability
that exists in the components of PM and their various ecological effects.
9.5. Effects on Materials
Effects of air pollution on materials are related to both aesthetic appeal and physical
damage. Deposited particles, primarily carbonaceous compounds, cause soiling of building
materials and culturally important items, such as statues and works of art. Physical
damage from dry deposition of PM also can accelerate natural weathering processes. The
major deterioration phenomenon affecting building materials in response to atmospheric
deposition is probably sulfation, leading to secondary salt crystallization which forms
gypsum (Marinoni et al., 2003, 092520).
This section (a) summarizes information on exposure-related effects on materials
associated with particulate pollutants as addressed in the 2004 PM AQCD (U.S. EPA, 2004,
056905) and (b) presents relevant information derived from very limited research conducted
and published since completion of that document. Most recent work on this topic has been
conducted outside the U.S.
There is a variety of factors that contribute to the deterioration of monuments and
buildings of cultural significance. They include^ (l) biodeterioration processes! (2)
weathering of materials exposed to the air! and (3) air pollution from both anthropogenic
and natural sources (Herrera and Videla, 2004, 155843). Because of the diversity in climate,
July 2009
9-263
DRAFT-DO NOT CITE OR QUOTE

-------
proximity to marine aerosol sources, and pollution of various types, the magnitude and
relative importance of these causal agents vary by location.
Much existing literature on damage to structural materials of cultural heritage has
not seriously considered the importance of biodeterioration processes and the relationship
that often exists between environmental characteristics and the microbial communities
that colonize monuments and buildings. In general, high humidity, high temperature, and
air pollution often enhance the biodeterioration hazard. Herrera and Videla (2004, 155843)
concluded that heterotrophic bacteria, fungi, and cyanobacteria were the main microbial
colonizers of buildings that they investigated in Latin America. Their analyses suggested
that the major deterioration mechanism of limestone at the Mayan site of Uxmal in a
non-polluted rural environment was biosolubilization induced by metabolic acids produced
by bacteria and fungi. The rock decay at Tulum, near the seashore, was mainly attributed
to the marine influence. At Medellin, it appeared that biodeterioration effects from microbes
synergistically enhanced the effects of atmospheric factors on material decay. Deterioration
of structural material in the Cathedral of La Plata, located in a mixed urban/industrial
environment, was attributed mainly to atmospheric pollutants (Herrera and Videla, 2004,
155843).
Ambient particles can cause soiling of man-made surfaces. Soiling generally is
considered an optical effect. Soiling changes the reflectance from opaque materials and
reduces the transmission of light through transparent materials. Soiling can represent a
significant detrimental effect, requiring increased frequency of cleaning of glass windows
and concrete structures, washing and repainting of structures, and, in some cases, reduces
the useful life of the object. Particles, especially carbon, may also help catalyze chemical
reactions that result in the deterioration of materials (U.S. EPA, 2004, 056905).
Soiling is dependent on atmospheric particle concentration, particle size distribution,
deposition rate, and the horizontal or vertical orientation and texture of the exposed surface
(Haynie, 1986, 157198). The chemical composition and morphology of the particles and the
optical properties of the surface being soiled will determine the time at which soiling is
perceived by human observers (Nazaroff and Cass, 1991, 044577).
Ferm et al. (2006, 155135) reported development of a simple passive particle collector
for estimating dry deposition to objects of cultural heritage. The observed mass of deposited
particles mainly belonged to the coarse particulate mode. The sampler collects particles
July 2009
9-264
DRAFT-DO NOT CITE OR QUOTE

-------
from all directions. It replicates at least some of the complexity of particle deposition to
actual objects, and is easier to analyze than a precious object (Ferm et al., 2006, 155135).
Soiling of urban buildings constitutes a visual nuisance that leads to the loss of
architectural value. Soiling can include reversible darkening of the building surfaces and
also irreversible damage. Water runoff patterns on the building surfaces are influenced by
the type of surface material, architectural elements, and climate. Therefore, soiling does not
occur uniformly across the building. Public perception of soiling entails complex
interactions between the extent of soiling, architecture, and aesthetics (Grossi and
Brimblecombe, 2004, 155813).
One of the most significant air pollution damage features affecting urban buildings
and monuments is the formation of black crusts. Quantification of different forms of carbon
in black crusts is difficult. There is often a carbonate component which is derived from the
building material, plus OC and EC, derived from air pollution. EC is considered to be a
tracer for combustion sources, whereas OC may derive from multiple sources, including
atmospheric deposition of primary and secondary pollutants, and the decay of protective
organic treatments (Bonazza et al., 2005, 155695). Bonazza et al. (2005, 155695) quantified
OC and EC in damage layers on European cultural heritage structures. OC predominated
over EC at almost all locations investigated. Traffic appeared to be the major source of fine
carbonaceous particles, with organic matter as the main component (Putaud et al., 2004,
055545). Viles and Gorbushina (2003, 156138) found that soiling in Oxford, U.K. showed a
relationship with traffic and NO2 concentrations.
In addition to the soiling effects of EC, much soiling appears to be largely of
microbiological origin (Viles and Gorbushina, 2003, 156138). Microbial biofilms, composed
mainly of fungi, can stain exposed rock surfaces with yellow, orange, brown, gray, or black
colors. Microorganisms may be able to trap PM more efficiently than the stone surface
itself. In addition, microbial growth may be stimulated by organic or nutrient constituents
in PM deposition.
Viles et al. (2002, 156137) investigated the nature of soiling on limestone tablets in
relation to ambient air pollution and climate at 3 contrasting sites in Great Britain over
periods of 1-8 yr. Spectrophotometer and microscope observations suggested that there were
not consistent trends in soiling over time at the study sites. Each site behaved differently in
terms of the temporal development of soiling and the differences between sheltered and
July 2009
9-265
DRAFT-DO NOT CITE OR QUOTE

-------
exposed limestone tablets. In addition, organisms played important roles in the soiling
response, even at the highly polluted site.
Some work has been conducted on public perception regarding the lightness of historic
buildings and the aesthetic need for cleaning subsequent to soiling by air pollution.
Brimblecombe and Grossi (2005, 155703) found a strong relationship between the perceived
lightness of a building and the opinion that it was dirty. This relationship was used to
establish levels of blackening that might be publicly acceptable.
Recently, the importance of organic contaminant deposition to the overall air pollution
damage to building materials has been recognized. Low molecular weight organic anions
such as formate, acetate, and oxalate are ubiquitous in black crusts in damage layers on
stones and mortars sampled from monuments and buildings throughout Europe (Sabbioni
et al., 2003, 049282). This has been observed at urban, suburban, and rural sites.
9.5.1.	Effects on Paint
Studies have evaluated the soiling effects of particles on painted surfaces (U.S. EPA,
2004, 056905). Particles composed of EC, acids, and various other constituents are
responsible for the soiling of structural painted surfaces. Coarse-mode particles (>2.5 pm)
initially contribute more soiling of horizontal and vertical painted surfaces than do
fine-mode particles (<2.5 pm), but are more easily removed by rain (Haynie and Lemmons,
1990, 044579). Rain interacts with coarse particles, dissolving the particle and leaving
stains on the painted surface (Creighton et al., 1990, 044578; Haynie and Lemmons, 1990,
044579). Particle deposition contributes to increased frequency of cleaning of painted
surfaces and physical damage to the painted surface. Air pollution affects the durability of
paint finishes by promoting discoloration, chalking, loss of gloss, erosion, blistering, and
peeling (U.S. EPA, 2004, 056905). There have been no new developments in this field
subsequent to the review of the U.S. EPA (2004, 056905).
9.5.2.	Effects on Metal Surfaces
Metals undergo natural weathering processes. The effects of air pollutants on natural
weathering processes depend on the nature of the pollutant(s), the deposition rate, and the
presence of moisture (U.S. EPA, 2004, 056905). Pollutant effects on metal surfaces are
governed by such factors as the presence of protective corrosion films and surface
July 2009
9-266
DRAFT-DO NOT CITE OR QUOTE

-------
electrolytes, the orientation of the metal surface, and surface moisture. Surface moisture
facilitates particulate deposition and promotes corrosive reactions. Formation of
hygroscopic salts increases the duration of surface wetness and enhances corrosion.
A corrosion film, such as for example the rust layer on a metal surface, provides some
protection against further corrosion. Its effectiveness in retarding the corrosion process is
affected by the solubility of the corrosion layer and the pollutant exposure. Other than the
effects of acidifying compounds, there has not been additional research conducted in recent
years on the effects of PM deposition on metal corrosion.
9.5.3. Effects on Stone
Air pollutants can enhance the natural weathering processes on building stone. The
development of crusts on stone monuments has been attributed to the interaction of the
stone's surface with pollutants, wet or dry deposition of atmospheric particles, and dry
deposition of gypsum particles. Because of a greater porosity and specific surface, mortars
have a high potential for reacting with environmental pollutants (Zappia et al., 1998,
012037).
Most research evaluating the effects of air pollutants on stone structures has
concentrated on gaseous pollutants (U.S. EPA, 2004, 056905). The dark color of gypsum is
attributed to soiling by carbonaceous particles. A lighter gray colored crust is attributed to
soil dust and metal deposits (Ausset et al., 1998, 040480; Camuffo, 1995, 076278; Lorusso et
al., 1997, 084534; Moropoulou et al., 1998, 040485). Lorusso et al. (1997, 084534) attributed
the need for frequent cleaning and restoration of historic monuments in Rome to exposure
to total suspended particulates.
Grossi et al. (2003, 155812) investigated the black soiling rates of building granite,
marble, and limestone in two urban environments with different climates. Horizontal
specimens were exposed, both sheltered and unsheltered from rainfall. Limestone showed
soiling proportional to the square root of the time of exposure, but granite and marble did
not.
Black soiling is caused mainly by particulate EC (PEC). For that reason, it is most
prevalent in urban environments due to the formation of carbonaceous fine particles from
the incomplete combustion of fossil fuels. Traffic emissions, especially from diesel engines,
and wood burning are important sources of PEC (Grossi et al., 2003, 155812).
July 2009
9-267
DRAFT-DO NOT CITE OR QUOTE

-------
Kamh (2005, 155888) studied the effects of weathering on Conway Castle, an
historical structure in Great Britain built about 1289 AC. The weathering was identified as
honeycomb, blackcrust, exfoliation, and discoloration, with white salt efflorescence at some
parts. These features are diagnostic for salt weathering (Goudie et al., 2002, 156486), and
this was confirmed by laboratory analyses, including scanning electron microscopy and
x-ray diffraction. The authors concluded that the salt was derived from three sources^ sea
spray, chemical alteration of the carbonate in mortar into SOr salts by acidic deposition,
and wet deposition of air pollutants on the stone surface. The salt content on the rock
surface fills the rock pores and then exerts high pressure on the rock texture due to
hydration of the salt in the cold humid environment. In particular, CaSCk and Na2S04 exert
enough pressure on hydration as to deteriorate construction rock at both the micro- and
macroscale (Moses and Smith, 1994, 156785).
9.5.4. Summary of Effects on Materials
Building materials (metals, stones, cements, and paints) undergo natural weathering
processes from exposure to environmental elements (wind, moisture, temperature
fluctuations, sunlight, etc.). Metals form a protective film of oxidized metal (e.g., rust) that
slows environmentally induced corrosion. However, the natural process of metal corrosion is
enhanced by exposure to anthropogenic pollutants. For example, formation of hygroscopic
salts increases the duration of surface wetness and enhances corrosion.
A significant detrimental effect of particle pollution is the soiling of painted surfaces
and other building materials. Soiling changes the reflectance of opaque materials and
reduces the transmission of light through transparent materials. Soiling is a degradation
process that requires remediation by cleaning or washing, and, depending on the soiled
surface, repainting. Particulate deposition can result in increased cleaning frequency of the
exposed surface and may reduce the usefulness of the soiled material. Attempts have been
made to quantify the pollutant exposure at which materials damage and soiling have been
perceived. However, to date, insufficient data are available to advance the knowledge
regarding perception thresholds with respect to pollutant concentration, particle size, and
chemical composition. Nevertheless, the evidence is sufficient to conclude that a causal
relationship exists between PM and effects on materials.
July 2009
9-268
DRAFT-DO NOT CITE OR QUOTE

-------
Chapter 9 References
Abdou W; Diner D; Martonchik J; Bruegge C; Kahn R; Gaitley B; Crean K; Remer L; Holben B. (2005).
Comparison of coincident Multiangle Imaging Spectroradiometer and Moderate Resolution Imaging
Spectroradiometer aerosol optical depths over land and ocean scenes containing Aerosol Robotic Network
sites. , 110. 190028
Aboal JR; Fernandez JA; Carballeira A. (2004). Oak leaves and pine needles as biomonitors of airborne trace
elements pollution. Environ Exp Bot, 51: 215-225. 155642
ABT. (2001). Assessing Public Opinions on Visibility Impairment Due to Air Pollution: Summary Report.
156185
ABT. (2002). Sense of Place and Stewardship: Final Focus Group Report. 156186
Ackerman A; Kirkpatrick M; Stevens D; Toon O. (2004). The impact of humidity above stratiform clouds on
indirect aerosol climate forcing. Nature, 432: 1014-1017. 190056
Ackerman AS; Toon OB; Stevens DE; Heymsfield AJ; Ramanathan V; Welton EJ. (2000). Reduction of tropical
cloudiness by soot. , 288: 1042-1047. 002987
Ackerman T; Stokes G. (2003). The atmospheric radiation measurement program. , 56: 38"44. 192080
Adachi K; Tainosho Y. (2004). Characterization of heavy metal particles embedded in tire dust. Environ Int, 30:
1009-1017. 081380
Adamo P; Crisafulli P; Giordano S; Minganti V; Modenesi P; Monaci F; Pittao E; Tretiach M; Bargagli R. (2007).
Lichen and moss bags as monitoring devices in urban areas. Part II: Trace element content in living and
dead biomonitors and comparison with synthetic materials. , 146: 392-399. 155644
Albrecht B. (1989). Aerosols, cloud microphysics, and fractional cloudiness. , 245: 1227-1230. 045783
Alexeeva LB; Strachan WMJ; Shlychkova W; Nazarova AA; Nikanorov AM; Korotova LG. (2001).
Organochlorine pesticides and trace metal monitoring of Russian rivers flowing to the Arctic Ocean. Mar
Pollut Bull, 43: 71-85. 155651
Alfani A; De Nicola F; Maisto G; Prati MV. (2005). Long-term PAH accumulation after bud break in Quercus
ilex L leaves in a polluted environment. Atmos Environ, 39: 307-314. 154319
Allen GA; Lawrence J; Koutrakis P. (1999). Field validation of a semi-continuous method for aerosol black
carbon (aethalometer) and temporal patterns of summertime hourly black carbon measurements in
southwestern PA. Atmos Environ, 33: 817-823. 048923
Almas AR; Bakken LR; Mulder J. (2004). Changes in tolerance of soil microbial communities in Zn and Cd
contaminated soils. Soil Biol Biochem, 36: 805-813. 155654
Alpert P; Kishcha P; Kaufman Y; Schwarzbard R. (2005). Global dimming or local dimming?: Effect of
urbanization on sunlight availability. , 32: 17. 190047
Ammann CM; Meehl GA; Washington WM; Zender CS. (2003). A monthly and latitudinally varying volcanic
forcing dataset in simulations of 20th century climate. Geophys Res Lett, 30: 1657. 192057
Anderson HR; Atkinson RW! Bremner SA; Marston L. (2003). Particulate air pollution and hospital admissions
for cardiorespiratory diseases: are the elderly at greater risk?. Eur Respir J, 40: 39s-46s. 054820
Anderson T; Charlson R; Bellouin N; Boucher O; Chin M; Christopher S; Haywood J; Kaufman Y; Kinne S;
Ogren J. (2005). An "A-Train" strategy for quantifying direct climate forcing by anthropogenic aerosols. ,
86: 1795-1809. 189993
Anderson T; Wu Y; Chu D; Schmid B; Redemann J; Dubovik O. (2005). Testing the MODIS satellite retrieval of
aerosol fine-mode fraction. , 110: D18204. 189991
Andreae MO; Gelencser A. (2006). Black carbon or brown carbon? The nature of light-absorbing carbonaceous
aerosols. , 6: 3131-3148. 156215
Andreae MO; Rosenfeld D; Artaxo P; Costa AA; Frank GP; Longo KM; Silva-Dias MAF. (2004). Atmospheric
science: smoking rain clouds over the Amazon. , 303: 1337-1341. 155658
Andronova NG; Rozanov EV; Yang F; Schlesinger ME; Stenchikov GL. (1999). Radiative forcing by volcanic
aerosols from 1850 to 1994. J Geophys Res, 104: 16807—16826. 192286
Antuna JC; Robock A; Stenchikov G; Zhou J; David C; Barnes J; Thomason L. (2003). Spatial and temporal
variability of the stratospheric aerosol cloud produced by the 1991 Mount Pinatubo eruption. J Geophys
Res, 108: 4624. 192251
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009
9-269
DRAFT-DO NOT CITE OR QUOTE

-------
Arctic Monitoring and Assessment Programme (AMAP). (2004). AMAP Assessment 2002- Persistent Organic
Pollutants in the Arctic. AMAP. Oslo, Norway. 190168
Arizona Department of Environmental Quality. (2000). Phoenix metropolitan area visibility briefing paper.
019164
Armbrust DV; Retta A. (2002). Wind and sandblast damage to growing vegetation. , 39: 273-284. 156225
Arnott WP; Moosmuller H; Rogers CF; Jin T> Bruch R. (1999). Photoacoustic spectrometer for measuring light
absorption by aerosol; instrument description. Atmos Environ, 33: 2845-2852. 020650
Audet P; Charest C. (2007). Heavy metal phytoremediation from a meta-analytical perspective. , 147: 231-237.
190169
Augustine J; Hodges G> Dutton Ei Michalsky J; Cornwall C. (2008). An aerosol optical depth climatology for
NOAA's national surface radiation budget network (SURFRAD). , 113: D11204. 189913
Ausset P; Bannery Fi Del Monte M> Lefevre RA. (1998). Recording of pre "industrial atmospheric environment by
ancient crusts on stone monuments. Atmos Environ, 32: 2859-2863. 040480
Avila A; Rodrigo A. (2004). Trace metal fluxes in bulk deposition, throughfall and stemflow at two evergreen
oakstands in NE Spain subject to different exposure to the industrial environment. Atmos Environ, 38:
171-180. 155664
Babich P; Wang PY; Allen G> Sioutas C> Koutrakis P. (2000). Development and Evaluation of a Continuous
Ambient PM2.5 Mass Monitor. Aerosol Sci Technol, 32: 309-324. 156239
Bachelet D; Lenihan J; Neilson R; Drapek R; Kittel T. (2005). Simulating the response of natural ecosystems
and their fire regimes to climatic variability in Alaska. Can J For Res, 35: 2244-2257. 156241
Backe C; Cousins IT; Larsson P. (2004). PCB in soils and estimated soil-air exchange fluxes of selected PCB
congeners in the south of Sweden. Environ Pollut, 128: 59"72. 155668
Bailey R; Barrie LA; Halsall CJ; Fellin PJ Muir DCG. (2000). Atmospheric organochlorine pesticides in the
western Canadian Arctic: evidence of trans-Pacific transport. J Geophys Res, 105: 11805-11811. 155670
Baker M; Charlson R. (1990). Bistability of CCN concentrations and thermodynamics in the cloud-topped
boundary layer. 190016
Balkanski Y; Schulz M; Claquin TJ Guibert S. (2007). Reevaluation of mineral aerosol radiative forcings
suggests a better agreement with satellite and AERONET data. , !'¦ 95. 189979
Baptista MS; Vasconcelos MTSD; Cabral JP; Freitas MC; Pacheco AMG. (2008). Copper, nickel and lead in
lichen and tree bark transplants over different periods of time. Environ Pollut, 151: 408-413. 155673
Baran AJ; Foot JS. (1994). New application of the operational sounder HIRS in determining a climatology of
sulphuric acid aerosol from the Pinatubo eruption. J Geophys Res, 99: 25673-25679. 192032
Barker ML. (1976). Planning for environmental indices: observer appraisals of air quality. In K Craig; E Zube
(Ed.),Perceiving Environmental Quality (pp. 175-203). New York: Plenum Press. 072137
Barnes JE; Hofmann DJ. (1997). Lidar measurements of stratospheric aerosol over Mauna Loa Observatory.
Geophys Res Lett, 24: 1923?1926. 192044
Bates T; Anderson T; Baynard T; Bond T; Boucher O; Carmichael G; Clarke A; Erlick C', Guo H; Horowitz L.
(2006). Aerosol direct radiative effects over the northwest Atlantic, northwest Pacific, and North Indian
Oceans: estimates based on in-situ chemical and optical measurements and chemical transport modeling.
, 6: 175-362. 189912
Bates T; Huebert B; Gras J; Griffiths B; Durkee P. (1998). The International Global Atmospheric Chemistry
(IGAC) Project's First Aerosol Characterization Experiment (ACE l)"Overview. , 103: 16297-16318.
190063
Bates T; Quinn P; Coffman D; Covert D; Miller T; Johnson J; Carmichael G; Uno I; Guazzotti S; Sodeman D.
(2004). Marine boundary layer dust and pollutant transport associated with the passage of a frontal
system over eastern Asia. , 109. 189958
Bates TS; Quinn PK; Coffman DJ; Johnson JE; Miller TL; Covert DS; Wiedensohler A; Leinert S; Nowak A;
Neususs C. (2001). Regional physical and chemical properties of the marine boundary layer aerosol
across the Atlantic during Aerosols99: an overview. J Geophys Res, 106: 20,767-20,782. 043385
Bauman JJ; Russell PB; Geller MA; Hamill P. (2003). A stratospheric aerosol climatology from SAGE II and
CLAES measurements: 2. Results and comparisons. J Geophys Res, 108: 4383. 192265
BaumgarcLner RE Jr; Isil SS; Bowser JJ; Fitzgerald KM. (1999). Measurements of rural sulfur dioxide and
particle sulfate: analysis of CASTNet data, 1987 through 1996. J Air Waste Manag Assoc, 49: 1266-1279.
011308
Baynard T; Lovejoy ER; Pettersson A; Brown SS; Lack D; Osthoff H; Massoli P; Ciciora S; Dube WP;
Ravishankara AR. (2007). Design and application of a pulsed cavity ring-down aerosol extinction
spectrometer for field measurements. Aerosol Sci Technol, 41: 447-462. 151669
July 2009
9-270
DRAFT-DO NOT CITE OR QUOTE

-------
BBC Research & Consulting. (2002). Phoenix Area Visibility Survey. Draft Report. Available at
http '//www. azdeq.gov/environ/air/download/vis_021903f.pdf. 156258
Becker Ti Brandel M. (2007). Vegetation-environment relationships in a heavy metal-dry grassland complex. ,
42: H-28. 156260
Beeby A; Richmond L. (2002). Evaluating Helix aspersa as a sentinel for mapping metal pollution. , l: 261-270.
155680
Bellouin N> Boucher 0> Haywood J; Reddy MS. (2005). Global emissions of aerosol direct radiative forcing from
satellite measurements. , 438: 1138-1141. 155684
Bellouin N> Boucher 0> Tanre D> Dubovik O. (2003). Aerosol absorption over the clear-sky oceans deduced from
POLDER-1 and AERONET observations. , 30: 1748. 189911
Bellouin N> Jones A; Haywood J; Christopher S. (2008). Updated estimate of aerosol direct radiative forcing
from satellite observations and comparison against the Hadley Centre climate model. , 113. 189999
Beron K> Murdoch J; Thayer M. (2001). The Benefits of Visibility Improvement: New Evidence from the Los
Angeles Metropolitan Area. , 22: 319-337. 156270
Berthelsen BO; Olsen RA; Steinnes E. (1995). Ectomycorrhizal heavy metal accumulation as a contributing
factor to heavy metal levels in organic surface soils. Sci Total Environ, 170: 141-149. 078058
Bhavsar SP; Diamond ML; Evans LJ; Gandhi N; Nilsen J; Antunes P. (2004). Development of a coupled metal
speciation—fate model for surface aquatic systems. Environ Toxicol Chem, 23: 1376— 1385. 155689
Bhavsar SP; Diamond ML; Gandhi N> Nilsen J. (2004). Dynamic coupled metal transport—speciation model:
application to assess a zinc-contaminated lake. Environ Toxicol Chem, 23: 2410—2420. 155690
Bickerstaff K; Walker G. (2001). Public understandings of air pollution: the 'localisation'of environmental risk.
Global Environ Change, 11: 133-145. 156271
Bidleman TF; Jantunen LMM; Helm PA; Bronstrom-Lunden E; Juntto S. (2002). Chlordane isomers and
enantiomers suggest changing sources in Arctic air. Environ Sci Technol, 36: 539-544. 155691
Billings WD. (1978). Plants and the ecosystem. Belmont, CA: Wadsworth Publishing Company, Inc. 034165
Bingen C; Fussen D; Vanhellemont F. (2004). A global climatology of stratospheric aerosol size distribution
parameters derived from SAGE II data over the period 1984?2000: 2. Reference data . J Geophys Res,
102: DQ6202. 192262
Blanchard CL; Tanenbaum S. (2006). Weekday/Weekend differences in ambient air pollutant concentrations in
atlanta and the southeastern United States. , 56: 271-284. 192181
Blanchard CL; Tanenbaum S; Hidy GM. (2007). Effects of sulfur dioxide and oxides of nitrogen emission
reductions on fine particulate matter mass concentrations: regional comparisons. J Air Waste Manag
Assoc, 57: 1337-50. 098659
Bluth GJS; Doiron SD; Schnetzler CC; Krueger AJ; Walter LS. (1992). Global tracking of the S02 clouds from
the June, 1991 Mount Pinatubo eruptions . Geophys Res Lett, 19: 151-154. 192029
Bluth GJS; Rose WL Sprod IE. (1997). Stratospheric loading of sulfur from explosive volcanic eruptions. , 105:
671-683. 192045
Bonazza A; Sabbioni C; Ghedini N. (2005). Quantitative data on carbon fractions in interpretation of black
crusts and soiling on European built heritage. Atmos Environ, 39: 2607-2618. 155695
Bond T; Bhardwaj E; Dong R; Jogani R; Jung S; Roden C; Streets D; Trautmann N. (2007). Historical emissions
of black and organic carbon aerosol from energy-related combustion, 1850-2000. Global Biogeochem
Cycles, 21. 190050
Bond TC; Anderson TL; Campbell D. (1999). Calibration and Intercomparison of Filter-Based Measurements of
Visible Light Absorption by Aerosols. Aerosol Sci Technol, 30: 582-600. 156281
Bond TC; Bergstrom RW. (2005). Light absorption by carbonaceous particles: an investigative review. Aerosol
Sci Technol, 39: 1-41. 155696
Bond TC; Streets DG; Yarber KF; Nelson SM; Woo J-H; KLimont Z. (2004). A technology-based global inventory
of black and organic carbon emissions from combustion. J Geophys Res, 109: D14203. 056389
Bond TC; Sun H. (2005). Can reducing black carbon emissions counteract global warming?. Environ Sci Technol,
39: 5921-5926. 156282
Boucher 0; Tanre D. (2000). Estimation of the aerosol perturbation to the Earth's radiative budget over oceans
using POLDER satellite aerosol retrievals. Geophys Res Lett, 27: 1103-1106. 190041
Boucher U; Balabane M; Lamy I; Cambier P. (2005). Decomposition in soil microcosms of leaves of the
metallophyte Arabidopsis halleri: Effect of leaf-associated heavy metals on biodegradation. , 135: 187-
194. 155699
Brasseur G; Granier C. (1992). Mount Pinatubo aerosols, chlorofluorocarbons, and ozone depletion. , 257: 1239"
1242. 192256
July 2009
9-271
DRAFT-DO NOT CITE OR QUOTE

-------
Brenguier J; Chuang Pi Fouquart Yi Johnson Di Parol Fi Pawlowska Hi Pelon J; Schuller Li Schroder Fi Snider
J. (2000). An overview of the ACE-2 CLOUDYCOLUMN closure experiment. , 52: 815-827. 189966
Brimblecombe Pi Grossi CM. (2005). Aesthetic thresholds and blackening of stone buildings. Sci Total Environ,
349: 175-189. 155703
Broccoli AJi Dixon KWi Delworth TLi Knutson TRi Stouffer RJi Zeng F. (2003). Twentieth-century temperature
and precipitation trends in ensemble climate simulations including natural and anthropogenic forcing. J
Geophys Res, 108: 4798. 192283
Brookshire D. (1979). Methods Development for Assessing Air Pollution Control Benefits. Volume 2'
Experiments in Valuing Nonmarket Goods. A Case Study of Alternative Benefit Measures of Air
Pollution in the South Coast Air Basin of Southern California	 156298
Burkhardt J; Kaiser Hi Kappen Li Goldbach HE. (2002). The possible role of aerosols on stomatal conductivity
for water vapour. , 2' 351-364. 155708
Burt R; Wilson MA; Mays MD; Lee CW. (2003). Major and trace elements of selected pedons in the USA. J
Environ Qual, 32: 2109-2121. 155709
Backstrom Mi Nilsson Ui Hakansson Ki Allard Bi Karlsson S. (2003). Speciation of heavy metals in road runoff
and roadside total deposition. Water Air Soil Pollut, 147: 343-366. 156242
Campbell Di Copeland Si Cahill T. (1995). Measurement of Aerosol Absorption Coefficient from Teflon Filters
using Integrating Plate and Integrating Sphere Techniques. Aerosol Sci Technol, 22: 287-292. 190171
Campbell FW. (1983). Ambient Stressors. , 15: 355-380. 190172
Camuffo D. (1995). Physical weathering of stones. Presented at In: Saiz-Jimenez, C., ed. The deterioration of
monuments: proceedings of the 2nd international symposium on biodeterioration and biodegradationi
January 1994i Sevilla, Spain. Sci. Total Environ. 167: 1-14. 076278
Canada-US Air Quality Committee, Subcommittee 2' Scientific Cooperation. (2004). Canada-United States
transboundary particulate matter science assessment. Meteorological Service of Canada. Toronto,
Ontario, Canada. En56"203/2004E. http://www.msc-smc.ec.gc.ca/saib/smog/transboundary/toc_e.html.
190519
Carmichael Gi Tang Yi Kurata Gi Uno Ii Streets Di Thongboonchoo Ni Woo Ji Guttikunda Si White Ai Wang T.
(2003). Evaluating regional emission estimates using the TRACE-P observations. , 108: 10.1029. 190042
Carmichael GRi Calori Gi Hay ami Hi Uno Ii Cho SYi Engardt Mi Kim SBi Ichikawa Yi Ikeda Yi Woo JHi Ueda
Hi Amann M. (2002). The MICS-Asia study: model intercomparison of long-range transport and sulfur
deposition in East Asia. Atmos Environ, 36: 175-199. 148319
Carrico Ci Kreidenweis Si Malm Wi Day Di Lee Ti Carrillo Ji McMeeking Gi Collett J. (2005). Hygroscopic
growth behavior of a carbon-dominated aerosol in Yosemite National Park. Atmos Environ, 39: 1393"
1404. 190052
Carvalho LMi Cacjador Ii Martins-Lou
-------
Chen W; Kahn R; Nelson D; Yau Ki Seinfeld J. (2008). Sensitivity of multiangle imaging to the optical and
microphysical properties of biomass burning aerosols. , 113. 189984
Chernyak SM; Rice CP; McConnell LL. (1996). Evidence of currently used pesticides in air, ice, fog, seawater
and surface microlayer in the Bering and Chukchi Seas. Mar Pollut Bull, 32- 410-419. 155726
Chestnut LG; Dennis RL. (1997). Economic benefits of improvements in visibility- acid rain Provisions of the
1990 Clean Air Act Amendments. J Air Waste Manag Assoc, 47: 395-402. 014525
Chin M> Diehl T; Ginoux P; Malm W. (2007). Intercontinental transport of pollution and dust aerosols:
implications for regional air quality. , T- 5501-5517. 190062
Chin M> Ginoux P; Kinne Si Torres Oi Holben Bi Duncan Bi Martin Ri Logan Ji Higurashi Ai Nakajima T.
(2002). Tropospheric aerosol optical thickness from the GOCART model and comparisons with satellite
and Sun photometer measurements. J Trace Elem Med Biol, 59: 461-483. 189996
Chin Mi Kahn RAi Remer LAi Yu Hi Rind Di Feingold Gi Quinn PKi Schwartz SEi Streets DGi DeCola Pi
Halthorne R. (2009). Atmospheric aerosol properties and climate impacts. Synthesis and assesment
product 2.3. A report by the U.S. Climate Change ScienceProgram and the Subcommittee on Global
Change Research. National Aeronautics and Space Administration. Washington, DC. 192130
Chiou CTi Sheng Gi Manes M. (2001). A partition-limited model for the plant uptake of organic contaminants
from soil and water. Environ Sci Technol, 35: 1437—1444. 156342
Chipperfield MP; Randel WJi Bodeker GEi Johnston P. (2003). Global ozone: past and future. In Scientific
Assessment of Ozone Depletion: 2002 (pp. 4.1?4.90). Geneva: World Meteorological Organization. 192275
Chock DPi Song Q; Hass H; Schell B; Ackermann I. (2003). Comment on "Control of fossil-fuel particulate black
carbon and organic matter, possibly the most effective method of slowing global warming" by Jacobson,
M. Z. J Geophys Res, 108: ACH12.1-ACH13.4. 155727
Chou M; Chan P; Wang M. (2002). Aerosol radiative forcing derived from SeaWiFS-retrieved aerosol optical
properties. J Trace Elem Med Biol, 59: 748-757. 190008
Chow JC; Watson JG; Edgerton SA; Vega E. (2002). Chemical composition of PM10 and PM25 in Mexico City
during winter 1997. Sci Total Environ, 287: 177-201. 036166
Chow JC; Watson JG; Lowenthal DH; Richards LW. (2002). Comparability between PM25 and particle light
scattering measurements. Environ Monit Assess, 79: 29"45. 037784
Christie P; Li XL; Chen BD. (2004). Arbuscular mycorrhiza can depress translocation of zinc to shoots of host
plants in soils moderately polluted with zinc. Plant Soil, 261: 209-217. 190174
Christopher S; Jones T. (2008). Short -wave aerosol radiative efficiency over the global oceans derived from
satellite data. , 60: 636-640. 189985
Christopher SA; Zhang J; Kaufman YJ; Remer L. (2006). Satellite "based assessment of the top of the
atmosphere anthropogenic aerosol radiative forcing over cloud-free oceans. Geophys Res Lett, 111:
L15816.155729
Chu D; Kaufman Y; Ichoku C; Remer L; Tanre D; Holben B. (2002). Validation of MODIS aerosol optical depth
retrieval over land. , 29: 8007. 190001
Chung C; Zhang G. (2004). Impact of absorbing aerosol on precipitation: Dynamic aspects in association with
convective available potential energy and convective parameterization closure and dependence on aerosol
heating profile. , 109. 190054
Chung CE; Ramanathan V; Kim D; Podgorny IA. (2005). Global anthropogenic aerosol direct forcing derived
from satellite and ground-based observations. J Geophys Res, 110: D24207. 155733
Chung SH; Seinfeld JH. (2002). Global distribution and climate forcing of carbonaceous aerosols. J Geophys Res,
107: 4407. 155732
Citterio S; Prato N; Fumagalli P; Aina R; Massa N; Santagostino A; Sgorbati S; Berta G. (2005). The arbuscular
mycorrhizal fungus Glomus mosseae induces growth and metal accumulation changes in Cannabis sativa
L. Chemosphere, 59: 21-29. 190176
Clarke A; Porter J; Valero F; Pilewskie P. (1996). Vertical profiles, aerosol microphysics, and optical closure
during the Atlantic Stratocumulus Transition Experiment: Measured and modeled column optical
properties. , 101: 4443-4453. 190003
Clemens S. (2006). Toxic metal accumulation, responses to exposure and mechanisms of tolerance in plants.
Biochimie, 88: 1707-1719. 190179
Coakley JA; Walsh CD. (2002). Limits to the aerosol indirect radiative effect derived from observations of ship
tracks. J Trace Elem Med Biol, 59: 668-680. 192025
Coelho JP; Rosa M; Pereira E; Duarte A; Pardal MA. (2006). Pattern and annual rates of Scrobicularia plana
mercury bioaccumulation in a human induced mercury gradient (Ria de Aveiro, Portugal). , 69: 629-635.
190181
July 2009
9-273
DRAFT-DO NOT CITE OR QUOTE

-------
Cohen Si Evans GWi Stokols Di Krantz DS. (1986). Behavior, Health, and Environmental Stress. New York-
Plenum Press. 190182
Collins Di Johnsson Hi Seinfeld J; Flagan R; Gasso Si Hegg Di Russell Pi Schmid Bi Livingston J; Ostrom E.
(2000). In situ aerosol -size distributions and clear-column radiative closure during ACE-2. , 52- 498-525.
190059
Collins RNi Merrington G> McLaughlin MJ; Morel JL. (2003). Organic ligand and pH effects on isotopically
exchangeable cadmium in polluted soils. Soil Sci Soc Am J, 67- 112-121. 155737
Collins W> Rasch P; Eaton B> Khattatov Bi Lamarque J; Zender C. 0- Simulating aerosols using a chemical
transport model with assimilation of satellite aerosol retrievals' Methodology for INDOEX. ,106. 189987
Conant W> VanReken Ti Rissman T> Varutbangkul V; Jonsson H> Nenes A; Jimenez J; Delia A; Bahreini R;
Roberts G. (2004). Aerosol-cloud drop concentration closure in warm cumulus. , 109. 190010
Cooke W; Wilson J. (). A global black carbon aerosol model. , 101. 190046
Cooke WF; Liousse C> Cachier Hi Feichter J. (1999). Construction of a 1 0 x 1 0 fossil fuel emission data set for
carbonaceous aerosol and implementation and radiative impact in the ECHAM 4 model. J Geophys Res,
104: 22137-22162. 156365
Costa Mi Silva Ai Levizzani V. (2004). Aerosol characterization and direct radiative forcing assessment over the
ocean. Part I: Methodology and sensitivity analysis. , 43: 1799-1817. 190006
Costa MJi Levizzani V; Silva AM. (2004). Aerosol characterization and direct radiative forcing assessment over
the ocean. Part IP Application to test cases and validation. , 43: 1818-1833. 192022
Couto JA; Fernandez JA; Aboal JR.; Carballeira A. (2004). Active biomonitoring of element uptake with
terrestrial mosses: a comparison of bulk and dry deposition. Sci Total Environ, 324: 211-222. 155739
Covert DS; Charlson RJ; Ahlquist NC. (1972). A study of the relationship of chemical composition and humidity
to light scattering by aerosols. ,11: 968-976. 072055
Crawford DW; Lipsen MS; Purdie DA; Lohan MC; Statham PJ; Whitney FA; Putland JN; Johnson WK
Sutherland N; Peterson TD; Harrison PJ; Wong CS. (2003). Influence of zinc andiron enrichments on
phytoplankton growth in the northeastern subarctic Pacific. Limnol Oceanogr, 48: 1583-1600. 156370
Creighton PJ; Lioy PJ; Haynie FH; Lemmons TJ; Miller JL; Gerhart J. (1990). Soiling by atmospheric aerosols
in an urban industrial area. J Air Waste Manag Assoc, 40: 1285-1289. 044578
Croteau MN! Luoma SN; Stewart AR. (2005). Trophic transfer of metals along freshwater food webs: Evidence
of cadmium biomagnification in nature. Limnol Oceanogr, 50: 1511-1519. 156373
Crepineau C; Rychen G; Feidt C; Le Roux Y; Lichtfouse E; Laurent F. (2003). Contamination of pastures by
polycyclic aromatic hydrocarbons (PAHs) in the vicinity of a highway. J Agric Food Chem, 51: 4841-4845.
155741
Cunningham M. (1979). Weather, mood, and helping behavior: Quasi experiments with the sunshine
Samaritan. , 37: 1947-1956. 191974
Cuny D; Denayer FO; De Foucault B; Schumacker R; Colein P; Van Haluwyn C. (2004). Patterns of metal soil
contamination and changes in terrestrial cryptogamic communities. Environ Pollut, 129: 289-297. 155742
Da Silva LC; Oliva MA; Azevedo AA; De Araujo JM. (2006). Responses of restinga plant species to pollution
from an iron pelletization factory. , 175: 241-256. 190190
Dallinger R; Wang Y; Berger B; Mackay EA; Kagi JH. (2001). Spectroscopic characterization of metallothionein
from the terrestrial snail, Helix pomatia. , 268: 4126-4133. 192109
Dameris M; Grewe V; Ponater M; Deckert R; Eyring V; Mager F; Matthes S; Schnadt C; Stenke A; Steil B; Bruhl
C; Giorgetta MA. (2005). Long-term changes and variability in a transient simulation with a chemistry-
climate model employing realistic forcing. , 5- 2121-2145. 192238
Davis AP; Shokouhian M; Ni S. (2001). Loading estimates of lead, copper, cadmium, and zinc in urban runoff
from specific sources. Chemosphere, 44: 997-1009. 024933
Davis MRH; Zhao FJ; McGrath SP. (2004). Pollution-induced community tolerance of soil microbes in response
to a zinc gradient. Environ Toxicol Chem, 23: 2665-2672. 155744
Day R. (2007). Place and the experience of air quality. Health Place, 13: 249-260. 156386
De Gouw J; Middlebrook A; Warneke C; Goldan P; Kuster W', Roberts J; Fehsenfeld F; Worsnop D; Canagaratna
M; Pszenny A. (2005). Budget of organic carbon in a polluted atmosphere: Results from the New England
Air Quality Study in 2002. ,110. 190020
De Nicola F; Maisto G; Prati MV; Alfani A. (2005). Temporal variations in PAH concentrations in Quercus ilex
L. (holm oak) leaves in an urban area. Chemosphere, 61: 432-440. 155747
DeBell L. (2006). Spatial and seasonal patterns and temporal variaibility of haze and its constituents in the
United States: Report IV. ICooperative Institute for Research in the Atmosphere. Colorado State
University Fort Collins, CO. 156388
July 2009
9-274
DRAFT-DO NOT CITE OR QUOTE

-------
Delworth T> Ramaswamy V; Stenchikov G. (2005). The impact of aerosols on simulated ocean temperature and
heat content in the 20th century. , 32. 190055
Demerjian KL; Mohnen VA. (2008). Synopsis of the temporal variation of particulate matter composition and
size. J Air Waste Manag Assoc, 58- 216-233. 156392
Denman KL; Brasseur G> Chidthaisong A; Ciais P; Cox PM; Dickinson RE; Hauglustaine D; Heinze C; Holland
E; Jacob D; Lohmann U; Ramachandran S; da Silva Dias PL; Wofsy SC; Zhang X. (2007). Couplings
Between Changes in the Climate System and Biogeochemistry. • . 156394
Dentener F; Stevenson D; Ellingsen K; Van Noije T; Schultz MJ Amann M; Atherton CJ Bell NJ Bergmann D;
Bey L Bouwman L; Butler T> Cofala J; Collins B; Drevet J; Doherty R; Eickhout B; Eskes H; Fiore A;
Gauss Mi Hauglustaine Di Horowitz Li Isaksen ISA; Josse B; Lawrence M; Krol M; Lamarque JE;
Montanaro V; Muller JF; Peuch VH; Pitari G; Pyle J; Rast S; Rodriguez J; Sanderson M; Savage NH;
Shindell D; Strahan S; et al. (2006). The global atmospheric environment for the next generation.
Environ Sci Technol, 40: 3586-3594. 088434
Derome J; Lindroos A"J. (1998). Effects of heavy metal contamination on macronutrient availability and
acidification parameters in forest soil in the vicinity of the Harjavalta Cu-Ni smelter, SW Finland.
Environ Pollut, 99: 225-232. 155749
Deshler T; Hervig ME; Hofmann DJ; Rosen JM; Liley JB. (2003). Thirty years of in situ stratospheric aerosol
size distribution measurements from Laramie, Wyoming (41°N), using balloon-borne instruments . J
Geophys Res, 108: 4167. 192058
Deuze J; Breon F; Devaux C; Goloub P; Herman M; Lafrance B; Maignan F; Marchand A; Nadal F; Perry G.
(2001). Remote sensing of aerosols over land surfaces from POLDER-ADEOS-1 polarized measurements.
, 106: x. 192013
Di Toro DM; McGrath JH; Berry WJ; Paquin PR; Mathew R; Wu KB; Santore RC. (2005). Predicting sediment
metal toxicity using a sediment biotic ligand model: Methodology and initial application. Environ Toxicol
Chem, 24: 2410-2427. 155750
Diner D; Beckert J; Bothwell G; Rodriguez J. (2002). Performance of the MISR instrument during its first 20
months in Earth orbit. IEEE Trans Geosci Remote Sens, 40: 1449-1466. 189967
Diner D; Beckert J; Reilly T; Bruegge C; Conel J; Kahn R; Martonchik J; Ackerman T; Davies R; Gerstl S.
(1998). Multi-angle Imaging SpectroRadiometer(MISR) instrument description and experiment overview.
IEEE Trans Geosci Remote Sens, 36: 1072-1087. 189962
Doherty S; Quinn P; Jefferson A; Carrico C; Anderson T; Hegg D. (2005). A comparison and summary of aerosol
optical properties as observed in situ from aircraft, ship, and land during ACE-Asia. , 110. 190027
Donahue WF; Allen EW! Schindler DW. (2006). Impacts of coal-fired power plants on trace metals and polycyclic
aromatic hydrocarbons (PAHs) in lake sediments in central Alberta, Canada. J Paleolimnol, 35: 111-128.
155751
Dong Y; Zhu YG; Smith FA; Wang Y; Chen B. (2008). Arbuscular mycorrhiza enhanced arsenic resistance of
both white clover (Trifolium repens Linn.) and ryegrass (Lolium perenne L.) plants in an arsenic-
contaminated soil. Environ Pollut, 155: 174-181. 192106
Douglass AR; Schoeberl MR; Rood RB; Pawson S. (2003). Evaluation of transport in the lower tropical
stratosphere in a global chemistry and transport model. J Geophys Res, 108D9. 057260
Dubovik O; Holben B; Eck TF; Smirnov A; Kaufman YJ; King MD; Tanre D;Slutsker I. (2002). Variability of
absorption and optical properties of key aerosol types observed in worldwide locations . J Trace Elem
Med Biol, 59: 590-608. 190202
Dubovik O; King MD. (2000). A flexible inversion algorithm for retrieval of aerosol optical properties from sun
and sky radiance measurements. J Geophys Res, 105: 20673-20696. 190197
Dubovik O; Lapyonok T; Kaufman Y; Chin M; Ginoux P; Sinyuk A. (2007). Retrieving global sources of aerosols
from MODIS observations by inverting GOCART model. , 7: 3629-3718. 190211
Dubovik O; Smirnov A; Holben BN; King MDiKaufman YJ; Eck TF; Slutsker I. (2000). Accuracy assessments of
aerosol optical properties retrieved from AERONET sun and sky radiance measurements. J Geophys Res,
105: 9791-9806. 190177
Duriscoe DM; Luginbuhl CB; Moore CA. (2007). Measuring Night-Sky Brightness with a Wide-Field CCD
Camera. , 119: 192-213. 156411
Dusek R; Frank GP; Hildebrandt L; Curtius J; Schneider J; Walter S; Chand D; Drewnick F; Hings S; Jung D;
Borrmann S; Andreae MO. (2006). Size matters more than chemistry for cloud-nucleating ability of
aerosol particles. 1375-1378. 155756
Eagan RC; Hobbs PV; Radke LF. (1974). Measurements of cloud condensation nuclei and cloud droplet size
distributions in the vicinity of forest fires. , 13: 553-557. 190231
July 2009
9-275
DRAFT-DO NOT CITE OR QUOTE

-------
Eck TF; Holben BN; Reid JS; Dubovik O; Smirnov A; O'Neill NT; Slutsker I; Kinne S. (1999). Wavelength
dependence of the optical depth of biomass burning, urban, and desert dust aerosols . J Geophys Res,
104: 31333-31350. 190390
Eck TF; Holben BN; Reid JS; Sinyuk A; Dubovik O; Smirnov A; Giles D; O'Neill NT; Tsay SC; Ji Q; Mandoos
AAI; Khan MR; Reid EA; Schafer JS; Sorokine M; Newcomb W', Slutsker I. (2008). Spatial and temporal
variability of column- integrated aerosol optical properties in the southern Arabian Gulf and United Arab
Emirates in summer. J Geophys Res, 113: in press. 190409
Edgerton ES; Hartsell BE; Jansen JJ; Hansen DA; Waid CJ; Kandasamy K. (2004). 5-Year Trend Analysis of
PM2.5 Data from the SEARCH Network. Presented at , . 156413
Eeva T; Lehikoinen E. (2004). Rich calcium availability diminishes heavy metal toxicity in Pied Flycatcher.
Funct Ecol, 18: 548-553. 155762
Eeva T; Tanhuanpaa S; Rabergh C; Airaksinen S; Nikinmaa M; Lehikoinen E. (2000). Biomarkers and
fluctuating asymmetry as indicators of pollution- induced stress in two hole-nesting passerines. Funct
Ecol, 14: 235-243. 155761
Elliot SJ; Cole DC; Krueger P; Voorberg N; Wakefield S. (1999). The power of perception: health risk attributed
to air pollution in an urban industrial neighborhood. Risk Anal, 19: 621-634. 010716
Ely DW; Leary JT; Stewart TR; Ross DM. (1991). The Establishment of the Denver Visibility Standard.
Presented at . 156417
Englert N. (2004). Fine particles and human health - a review of epidemiological studies. Toxicol Lett, 149: 235"
242. 087939
EPA. (2000). National air pollutant emission trends, 1900-1998. U.S. Environmental Protection Agency.
Washington, D.C.. . . 012211
EPA. (2001). Draft Guidance for tracking progress under the regional haze rule. U.S. Environmental Protection
Agency. Washington, D.C.. . . 157068
Ericksen JA; Gustin MS; Schorran DE; Johnson DW; Lindberg SE; Coleman JS. (2003). Accumulation of
atmospheric mercury in forest foliage. Atmos Environ, 37: 1613-1622. 155769
Ervens B; Feingold G; and SMKreidenweis SM. (2005). The influence of water-soluble organic carbon on cloud
drop number concentration. J Geophys Res, 110: D18211.l-D18211.14. 190527
Evans GW; Jacobs SV. (1982). Air Pollution and Human Behavior. In Evans, GW (Ed.),Environmental Stress
(pp. 105-132). New York, NY: Cambridge University Press. 179899
Evans GW; Jacobs SV; Dooley D; Catalano R. (1987). The Interaction of Stressful Life Events and Chronic
Strains on Community Mental Health. , 15: 23"24. 190347
Evans GW; Jacobs SV; Frager NB. (1982). Behavioral responses to air pollution. In A Baum! J Singer
(Ed.),Advances in Environmental Psychology (pp. 237-269). Hillsdale, NJ: Erlbaum. 190521
Ewald G; Larsson P; Linge H; Okla L; Szarzi M. (1998). Biotransport of organic pollutants to an inland Alaska
lake by migrating sockeye salmon (Oncorhynchus nerka). , 51: 40-47. 190348
FehsenfeldFC | Ancellet G | Bates TS | Goldstein AH | Hardesty RM | Honrath R | Law KS; Lewis AC;
Leaitch R; McKeen S; Meagher J; Parrish DD; Pszenny AAP; Russell PB; Schlager H; Seinfeld J; Talbot
R; Zbinden R. (2006). International consortium for atmospheric research on transport and transformation
(ICARTT): North America to Europe -overview of the 2004 summer field study . J Geophys Res, 111:
D23S01.1-D23S01.36. 190531
Feichter J; Sausen R; GraBl H; Fiebig M. (2003). Comment on "Control of fossil-fuel particulate black carbon
and organic matter, possibly the most effective method of slowing global warming" by M. Z. Jacobson. J
Geophys Res, 108 (D24), 4767, doi:i0.1029/2002JD003223. 155772
Feingold G. (2003). Modeling of the first indirect effect: Analysis of measurement requirements. Geophys Res
Lett, 30: ASC7.1-ASC7.4. 190547
Feingold G; Furrer R; Pilewskie P; Remer LA; Min Q; Jonsson; Haflidi. (2003). Aerosol indirect effect studies at
southern great plains during the May 2003 intensive operations period . J Geophys Res, 111: D5. 190551
Feingold G; Jiang H; Harrington J. (2005). On smoke suppression of clouds in Amazonia. Geophys Res Lett, 32:
L02804.1-L02804.4. 190550
Feingold G; Remer LA; Ramaprasad J; Kaufman YJ. (2001). Analysis of smoke impact on clouds in Brazilian
biomass burning regions: An extension of Twomey's approach. J Geophys Res, 106: 22907-22922. 190544
Feingold G; Stevens B; Cotton WR; Walko RL. (1994). An explicit microphysics/LES model designed to simulate
the Twomey Effect. , 33: 207-233. 190535
Feingold GW; Cotton WR; Kreidenweis SM; Davis JT. (1999). The impact of giant cloud condensation nuclei on
drizzle formation in stratocumulus: Implications for cloud radiative properties. J Trace Elem Med Biol,
56: 4100-4117. 190540
July 2009
9-276
DRAFT-DO NOT CITE OR QUOTE

-------
Feng MH; Shan XQ; Zhang SZ; Wen B. (2005). Comparison of a rhizosphere-based method with other one-step
extraction methods for assessing the bioavailability of soil metals to wheat. Chemosphere, 59- 939-949.
155774
Ferm Mi Watt Ji O'Hanlon Si De Santis F- Varotsos C. (2006). Deposition measurement of particulate matter in
connection with corrosion studies. Anal Bioanal Chem, 384- 1320-30. 155135
Fernandes SDi Trautmann NMi Streets DGi Roden CAi Bond TC. (2007). Global biofuel use, 1850-2000. Global
Biogeochem Cycles, 21: GB2019.1-GB2019.15. 190554
Ferrare Ri Feingold Gi Ghan Si Ogren J Schmid Bi Schwartz SEi Sheridan P. (2006). Preface to special section:
Atmospheric radiation measurement program May 2003 intensive operations period examining aerosol
properties and radiative influences . J Geophys Res, 111: D05S01. 190561
Fiore AMi Jacob DJi Mathur Ri Martin RV. (2003). Application of empirical orthogonal functions to evaluate
ozone simulations with regional and global models. J Geophys Res, 108D14: . 047805
Fishman Ji Hoell JMi Bendura RDi McNeal RJi Kirchhoff V. (1996). NASA GTE TRACE A experiment
(September- October 2002): Overview. J Geophys Res, 101: 23865-23880. 190566
Fismes Ji Perrin-Ganier Ci Empereur-Bissonnet Pi Morel JL. (2002). Soil-to-root transfer and translocation of
polycyclic aromatic hydrocarbons by vegetables grown on industrial contaminated soils. , 31: 1649-1656.
141156
Fitzgerald JW. (1975). Approximation formulas for the equilibrium size of an aerosol particle as a function of its
dry size and composition and the ambient relative humidity. , 14: 1044-1049. 095417
Forster Pi Ramaswamy Vi Artaxo Pi Betss Ri Fahey DWi Haywood Ji Lean Ji Lowe DCi Myhre Gi Nganga Ji
Prinn Ri Raga Gi Schultz Mi Van Dorland R. (2007). Changes in atmospheric constituents and in
radiative forcing. In Solomon, S.i Qin, D.i Manning, M.i Chen, Z.i Marquis, M.i Averyt, K. B.i Tignor, M.i
Miller, H. L. (Ed.),Climate Change 2007: the physical science basis. Contribution of Working Group I to
the fourth assessment report of the intergovernmental panel on: . 092936
Fraser Ri Kaufman Y. (1985). The relative importance of aerosol scattering and absorption in remote sensing.
IEEE Trans Geosci Remote Sens, GE-23: 625-633. 190567
Free Mi Angell JK. (2002). Effect of volcanoes on the vertical temperature profile in radiosonde data. J Geophys
Res, 107: 4101. 192281
Freer-Smith PHi El-khatib Ai Taylor G. (2004). Capture of particulate pollution by trees: a comparison of
species typical of semi-arid areas (Ficus nitida and Eucalyptus globulus) with European and North
American species. Water Air Soil Pollut, 155: 173-187. 156451
Frescholtz TFi Gustin MSi Schorran DEi Fernandez GCJ. (2003). Assessing the source of mercury in the foliar
tissue of quaking aspen. Environ Toxicol Chem, 22: 2114-2119. 190352
Fritze Hi Niini Si Mikkola Ki Makinen A. (1989). Soil microbial effects of a Cu-Ni smelter in southwestern
Finland. , 8: 87-94. 079635
Fryer MEi Collins CD. (2003). Model intercomparison for the uptake of organic chemicals by plants. Environ Sci
Technol, 37: 1617-1624. 156454
Gandhi Ni Bhavsar SPi Diamond MLi Kuwabara JSi Marvin-DiPasquale Mi Krabbenhoft DP. (2007).
Development of a mercury speciation, fate, andbiotic uptake (BIOTRANSPEC) model: Application to
Lahontan Reservoir (Nevada, USA). Environ Toxicol Chem, 26: 2260-2273. 155781
Gao YZi Zhu LZ. (2004). Plant uptake, accumulation and translocation of phenanthrene andpyrene in soils.
Chemosphere, 55: 1169-1178. 155782
Garrett Ti Zhao Ci Dong X; Mace Gi Hobbs P. (2004). Effects of varying aerosol regimes on low-level Arctic
stratus. Geophys Res Lett, 31: L17105.1-L17105.4. 190568
Garrett TJi Zhao C. (2006). Increased Arctic cloud longwave emissivity associated with pollution from mid-
latitudes. Nature, 440: 787-789. 190570
Gawel JEi Ahner BA; Friedland AJ; Morel FMM. (1996). Role for heavy metals in forest decline indicated by
phytochelatin measurements. , 381: 64"65. 012278
Generoso S; Bey I; Attie J-L; Breon F"M. (2007). A satellite- and model-based assessment of the 2003 Russian
fires: impact on the arctic region. J Geophys Res, 112: 5302. 155786
Geogdzhayev L Mishchenko M; Rossow W; Cairns B; Lacis AA. (2002). Global two-channel AVHRR retrievals of
aerosol properties over the ocean for the period of NOAA-9 observations and preliminary retrievals using
NOAA-7 and NOAA-11 data. J Trace Elem Med Biol, 59: 262-278. 190574
Geron C; Rasmussen R; Arnts RR; Guenther A. (2000). A review and synthesis of monoterpene speciation from
forests in the United States. Atmos Environ, 34: 1761-1781. 019095
Gillett NP; Hegerl GC; Allen MR; Stott PA; Schnur R. (2002). Reconciling two approaches to the detection of
anthropogenic influence on climate. J Clim, 15: 326-329. 190576
July 2009
9-277
DRAFT-DO NOT CITE OR QUOTE

-------
Gillett NP; Zwiers FW; Weaver AJ; Hegerl GC; Allen MR; Stott PA. (2002). Detecting anthropogenic influence
with a multi-model ensemble. J Geophys Res, 29- 31.1-31.4. 190578
Gilliland FD; Li YF; Saxon A; Diaz-Sanchez D. (2004). Effect of glutathione-S-transferase Ml and PI genotypes
on xenobiotic enhancement of allergic responses' randomised, placebo-controlled crossover study. Lancet,
363: 119-125. 156471
Ginoux P; Chin M> Holben B; Lin SJ; Tegen L Prospero JM; Dubovik O. (2001). Sources and distributions of
dust aerosols simulated with the GOCART model. J Geophys Res, 20: 20255-20273. 190579
Ginoux P; Horowitz LW; Ramaswamy V; Geogdzhayev IV; Holben BN; Stenchikov GJ Tie X. (2006). Evaluation
of aerosol distribution and optical depth in the geophysical fluid dynamics laboratory coupled model
CM2.1 for present climate. J Geophys Res, 111: D22210. 190582
Gitay Hi Brown Si Easterling Wi Jallow B. (2001). Ecosystems and their goods and services. In Climate change
2001: impacts, adaptation and vulnerability. Contribution of working group II to the third assessment
report of the Intergovernmental Panel on Climate Change (pp. 237-342). Cambridge, United Kingdom:
Cambridge University Press. 092761
Givati Ai Rosenfeld D. (2004). Quantifying precipitation suppression due to air pollution. , 43: 1038-1056.
156475
Gleason SMi Faucette DTi Toyofuku MM; Torres CA; Bagley CF. (2007). Assessing and mitigating the effects of
windblown soil on rare and common vegetation. Environ Manage, 40: 1016-1024. 155794
Goforth MR; Christoforou CS. (2006). Particle size distribution and atmospheric metals measurements in a
rural area in the South Eastern USA. Sci Total Environ, 356: 217-227. 088353
Gohre V; Paszkowski U. (2006). Contribution of the arbuscular mycorrhizal symbiosis to heavy metal
phytoremediation. Planta, 223: 1115-1122. 190355
Golaz JC; Larson VE; Cotton WR. (2002). A PDF based model for boundary layer clouds. Part I: Method and
model description. J Trace Elem Med Biol, 59: 3540-3551. 190587
Golaz JC; Larson VE; Cotton WR. (2002). A PDF-Based Model for Boundary Layer Clouds. Part II: Model
Results. J Trace Elem Med Biol, 59: 3552-3571. 190589
Gomot-de Vaufleury A; Kerhoas I. (2000). Effects of cadmium on the reproductive system of the land snail Helix
aspersa. , 64: 434-442. 155798
Gomot-De Vaufleury A; Pihan F. (2002). Methods for toxicity assessment of contaminated soil by oral or dermal
uptake in land snails: Metal bioavailability and bioaccumulation. Environ Toxicol Chem, 21: 820-827.
190357
Gonzalez-Chavez C; Harris PJ; Dodd J; Meharg AA. (2002). Arbuscular mycorrhizal fungi confer enhanced
arsenate resistance on Holcus lanatus. New Phytol, 155: 163-171. 155800
Goodarzi F; Sanei H; Garrett RG; Duncan WF. (2002). Accumulation of trace elements on the surface soil
around the Trail smelter, British Columbia, Canada. , 43: 29"38. 155801
Gordon W; Jackson R. (2000). Nutrient concentrations in fine roots. Ecology, 81: 275-280. 155802
Goudie AS; Elaine W; Viles HA. (2002). The roles of salt and fog in weathering: a laboratory simulation of
conditions the northern Atacama Desert, Chile. Catena, 48: 255—266. 156486
Grabowski WW. (2004). An improved framework for superparameterization. [embedding of cloud-resolving
model in each column of large-scale model for small-scale and mesoscale processes representation] . J
Trace Elem Med Biol, 61: 1940-1952. 190590
Grabowski WW; Wu X; Moncrieff MW. (1999). Cloud resolving modeling of tropical cloud systems during phase
III of GATE. Part III: Effects of cloud microphysics . J Trace Elem Med Biol, 56: 2384-2402. 190592
Grannas AM; Hockaday WC; Hatcher PG; Thompson LG; Mosley-Thompson E. (2006). New revelations on the
nature of organic matter in ice cores. J Geophys Res, 111, D04304, doi:i0.1029/2005JD006251. 155804
Grannas AM; Jones AE; Dibb J; Ammann M; Anastasio C; Beine HJ; Bergin M; Bottenheim J; Boxe CS; Carver
G; Chen G; Crawford JH; Domine F; Frey MM; Guzman MI; Heard DE; Helmig D; Hoffmann MR;
Honrath RE; Huey LG; Hutterli M; Jacobi HW; Klan P; Lefer B; McCo. (2007). An overview of snow
photochemistry: evidence, mechanisms and impacts. , T- 4329-4373. 156492
Grannas AM; Shepson PB; Filley TR. (2004). Photochemistry and nature of organic matter in Arctic and
Antarctic snow. Global Biogeochem Cycles, 18(l), GB1006, doi:i0.1029/2003GB002133. 155803
Grantz DA; Garner JHB; Johnson DW. (2003). Ecological effects of particulate matter. Environ Int, 29: 213-239.
155805
Gratao PL; Polle A; Lea PJ; Azevedo RA. (2005). Making the life of heavy metal-stressed plants a little easier.
Funct Plant Biol, 32: 481-494. 190364
Gray JS. (2002). Biomagnification in marine systems: The perspective of an ecologist. Mar Pollut Bull, 45: 46"52.
155806
July 2009
9-278
DRAFT-DO NOT CITE OR QUOTE

-------
Greenstein D; Tiefenthaler L; Bay S. (2004). Toxicity of parking lot runoff after application of simulated rainfall.
Arch Environ Contam Toxicol, 47: 199-206. 155808
Gregory JM; Stouffer RJ; Raper SCB; Stott PA; Rayner NA. (2002). An observationally based estimate of the
climate sensitivity. J Clim, 15: 3117-3121. 190593
Griffith DA; Solak ME; Yorty DP. (2005). Is air pollution impacting winter orographic precipitation in Utah?
Weather modification association. J Weather Modif, 37: 14-20. 156497
Grigal DF. (2002). Inputs and outputs of mercury from terrestrial watersheds: a review. , 10: 1—39. 156498
Grigal DF. (2003). Mercury sequestration in forests and peatlands: A review. J Environ Qual, 32: 393-405.
155811
Grossi CM; Brimblecombe P. (2004). Aesthetics of simulated soiling patterns on architecture. Environ Sci
Technol, 38: 3971-3976. 155813
Grossi CM; Esbert RM; Diaz-Pache F; Alonso FJ. (2003). Soiling of building stones in urban environments. , 38:
147-159. 155812
Guderian R ed. (1985). Air pollution by photochemical oxidants: formation, transport, control, and effects on
plants. New York: Springer-Verlag. 019325
Guderian R. (1977). Accumulation of pollutants in plant organs. In Air pollution: phytotoxicity of acidic gases
and its significance in air pollution controlBerlin, Germany: Springer-Verlag. 004150
Gunn R; Phillips BB. (1957). An experimental investigation of the effect of air pollution on the initiation of rain.
J Trace Elem Med Biol, 14: 272-280. 190595
Gupta P; Christopher SA; Wang J; Gehrig R; Lee Y; Kumar N. (2006). Satellite remote sensing of particulate
matter and air quality assessment over global cities. Atmos Environ, 40: 5880-5892. 137694
Gustin MS. (2003). Are mercury emissions from geologic sources significant? A status report. Sci Total Environ,
304: 153-167. 155816
Hageman KJ; Simonich SL; Campbell DH; Wilson GR; Landers DH. (2006). Atmospheric deposition of current-
use and historic-use pesticides in snow at national parks in the western United States. Environ Sci
Technol, 40: 3174-3180. 156509
Hall JL. (2002). Cellular mechanisms for heavy metal detoxification and tolerance. J Exp Bot, 53: 1-11. 190365
Halsall CJ. (2004). Investigating the occurrence of persistent organic pollutants (POPs) in the Arctic: their
atmospheric behaviour and interaction with the seasonal snow pack. Environ Pollut, 128: 163-175.
155822
Halsall CJ; Barrie LA; Fellin P; Muir DCG; Rovinski FY; Kononov EY. (1997). Spatial and temporal variation of
polycyclic aromatic hydrocarbons in the arctic atmosphere. Environ Sci Technol, 31: 3593-3599. 155821
Han Q; Rossow WB; Chou J; Welch RM. (1998). Global survey of the relationship of cloud albedo and liquid
water path with droplet size using ISCCP. J Clim, 11: 1516-1528. 190594
Han Q; Rossow WB; Zeng J. (2002). Three different behaviors on liquid water path of clouds in aerosol-cloud
interactions. J Trace Elem Med Biol, 59: 726-735. 049181
Hand JL; Kreidenweis SM; Sherman DE; Collett JL Jr; Hering SV; Day DE; Malm WC. (2002). Aerosol size
distributions and visibility estimates during the big bend regional aerosol visibility and observational
study (BRAVO). Atmos Environ, 36: 5043-5055. 190367
Hand JL; Malm WC. (2006). Review of the IMPROVE Equation for Estimating Ambient Light Extinction
Coefficients-Final Report	 156517
Hand JL; Malm WC. (2007). Review of aerosol mass scattering efficiencies from ground-based measurements
since 1990. J Geophys Res, 112: . 155825
Hansen AC; Zhang Q; Lyne PWL. (2005). Ethanol-diesel fuel blends--a review. Bioresour Technol, 96: 277-285.
059087
Hansen AD; Rosen AH; Novakov T. (1982). Real-Time Measurement of the Absorption Coefficient of Aerosol
Particles. Appl Opt, 21: 3060-3062. 190368
Hansen ADA; Rosen H; Novakov T. (1984). The aethalometer - an instrument for the real-time measurement of
optical absorption by aerosol particles. Sci Total Environ, 36: 191-196. 002396
Hansen J; Nazarenko L. (2004). Soot climate forcing via snow and ice albedos. Presented at , . 156521
Hansen J; Sato M; Mazarenko L; Ruedy R; Lacis A; Koch D; Tegen I; Hall T; Shindell D; Santer B; Stone P;
Novakov T; Thomason L; Wang R; Wang Y; Jacob D; Hollandsworth S; Bishop L; Logan J; Thompson A;
Stolarski R; Lean J; Willson R; Levitus S; Antonov J; Rayner N; Parker D; Christy J. (2002). Climate
forcings in Goddard Institute for Space Studies S12000 simulations. J Geophys Res, 107. 049177
July 2009
9-279
DRAFT-DO NOT CITE OR QUOTE

-------
Hansen J; Sato M> Ruedy R; Kharecha P; Lacis A; Miller R; Nazarenko L; Lo K> Schmidt GA; Russell G> Aleinov
L Bauer S> Baum Ei Cairns Bi Canuto V; Chandler M> Cheng Y> Cohen A; Del Genio A; Faluvegi G>
Fleming E; Friend A; Hall T; Jackman C; Jonas J; Kelley M; Kiang NY; Koch D; Labow G; Lerner J;
Menon S> Novakov T> Oinas V; Perlwitz JA; Perlwitz Jui Rind Di Romanou A; Schmunk R; Shindell D>
Stone P; Sun Si Streets Di Tausnev Ni Thresher Di Unger Ni Yao Mi Zhang S. (2007). Climate
simulations for 1880-2003 with GISS modelE. , 29- 661-696. 190597
Hansen Ji Sato Mi Ruedy Ri Lacis Ai Oinas V. (2000). Global warming in the twenty-first century- an
alternative scenario. Proc Natl Acad Sci USA, 97: 9875-9880. 042683
Hansen Ji Sato Mi Ruedy Ri Nazarenko Li Lacis Ai Schmidt GAi Russell Gi Aleinov Ii Bauer Mi Bauer Si Bell Ni
Cairns Bi Canuto Vi Chandler Mi Cheng Yi Del Genio Ai Faluvegi Gi Fleming Ei Friend Ai Hall Ti
Jackman Ci Kelley Mi Kiang Ni Koch Di Lean Ji Lerner Ji Lo Ki Menon Si Miller Ri Minnis Pi Novakov
Ti Oinas Vi Perlwitz JA; Perlwitz Jui Rind D; Romanou A; Shindell D; Stone P; Sun Si Tausnev Ni
Thresher Di Wielicki Bi Wong Ti Yao Mi Zhang S. (2005). Efficacy of climate forcings. J Geophys Res,
110: D18104.190596
Hansen JEi Sato Mi Ruedy R. (1997). Radiative forcing and climate response. J Geophys Res, 102: 6831-6864.
043104
Hargrave BTi Barrie LAi Bidleman TFi Welch HE. (1997). Seasonality in exchange of organochlorines between
arctic air and seawater. Environ Sci Technol, 31: 3258-3266. 155827
Harmens Hi Norris DAi Koerber GRi Buse Ai Steinnes Ei Riihling A. (2007). Temporal trends in the
concentration of arsenic, chromium, copper, iron, nickel, vanadium and zinc in mosses across Europe
between 1990 and 2000. Atmos Environ, 41: 6673-6687. 155828
Harner Ti Jantunen LMMi Bidleman TFi Barrie LAi Kylin Hi Strachan WMJi Macdonald RW. (2000). Microbial
degradation is a key elimination pathway of hexachlorocyclohexanes from the Arctic Ocean. Geophys Res
Lett, 27: 1155-1158. 155829
Harrison Li Michalsky Ji Berndt J. (1994). Automated multifilter rotating shadow-band radiometer: an
instrument for optical depth and radiation measurements. Appl Opt, 33: 5118-5125. 045805
Harvey LDD. (2004). Characterizing the annual-mean climatic effect of anthropogenic C02 and aerosol
emissions in eight coupled atmosphere-ocean GCMs. , 23: 569-599. 190598
Hassan Ri Scholes Ri Ash N. (2005). Ecosystems and human well-being: current state and trends, volume 1.
092759
Hauck M. (2003). Epiphytic lichen diversity and forest dieback: The role of chemical site factors. Bryologist, 106:
257-269. 155830
Haynie FH. (1986). Environmental factors affecting corrosion of weathering steel. Presented at Materials
degradation caused by acid rain: developed from the 20th state-of-the-art symposium of the American
Chemical Society; June 1985; Arlington, VA., Washington, DC. 157198
Haynie FH; Lemmons TJ. (1990). Particulate matter soiling of exterior paints at a rural site. Aerosol Sci
Technol, 13: 356-367. 044579
Haywood J; Boucher O. (2000). Estimates of the direct and indirect radiative forcing due to tropospheric
aerosols: A review. Rev Geophys, 38: 513-543. 156531
Haywood J; Francis P; Osborne S; Glew M; Loeb N; Highwood E; Tanre D; Myhre G; Formenti P; Hirst E.
(2003). Radiative properties and direct radiative effect of Saharan dust measured by the C-130 aircraft
during SHADE: 1. Solar spectrum : The Saharan Dust Experiment (SHADE). , 108: SAH 4-1. 190599
Haywood J; Schulz M. (2007). Causes of the reduction in uncertainty in the anthropogenic radiative forcing of
climate between IPCC (2001) and IPCC (2007). Geophys Res Lett, 34: L20701. 190600
Haywood JM; Pelon J; Formenti P; Bharmal N; Brooks M; Capes G; Chazette P; Chou C; Christopher S; Coe H;
Cuesta J; Derimian Y; Desboeufs K; Greed G; Harrison M; Heese B; Highwood EJ; Johnson B; Mallet M;
Marticorena B; Marsham J; Milton S; Myhre G; Osborne SR; Parker DJ; Rajot JL; Schulz M; Slingo A;
Tanre D; Tulet P. (2008). Overview of the dust and biomass burning experiment and African monsoon
multidisciplinary analysis special observing Period"0. J Geophys Res, 113: D00C17. 190602
Haywood JM; Ramaswamy V; Soden BJ. (1999). Tropospheric aerosol climate forcing in clear-sky satellite
observations over the oceans. , 283: 1299-1303. 040453
Heald CL; Jacob DJ; Park RJ; Russell LM; Huebert BJ; Seinfeld JH; Liao H; Weber RJ. (2005). A large organic
aerosol source in the free troposphere missing from current models. Geophys Res Lett, 32: L18809.1"
L18809.4. 190603
HEL Ackerman A; Toon O; Hobbs P. (1994). Reassessing the dependence of cloud condensation nucleus
concentration on formation rate. Nature. 189975
Heintzenberg J. (2008). The SAMUM-1 experiment over Southern Morocco: overview and introduction. , 61: 2-
11. 190605
July 2009
9-280
DRAFT-DO NOT CITE OR QUOTE

-------
Helena P; Franc B; Cvetka RL. (2004). Monitoring of short-term heavy metal deposition by accumulation in
epiphytic lichens (Hypogymnia Physodes (L.) Nyl.). , 49' 223-230. 155833
Helm PA; Diamond ML; Semkin R; Bidleman TF. (2000). Degradation as a loss mechanism in the fate of a-
hexachlorocyclohexane in Arctic watersheds. Environ Sci Technol, 34- 812-818. 155834
Helm PA; Diamond ML; Semkin R; Strachan WMJ; Teixeira C', Gregor D. (2002). A mass balance model
describing multi-year fate of organochlorine compounds in a high arctic lake. Environ Sci Technol, 36-
996-1003. 155835
Helmisaari H"S; Makkonen K; Olsson M; Viksna A; Malkonen E. (1999). Fine-root growth, mortality and heavy
metal concentrations in limed and fertilized Pinus silvestris (L.) stands in the vicinity of a Cu_Ni smelter
in SW Finland. Plant Soil, 209- 193-200. 155836
Henze DK; Seinfeld JH. (2006). Global secondary organic aerosol from isoprene oxidation. Geophys Res Lett, 33-
L09812.1-L09812.4. 190606
Herman JR; Bhartia PK; Torres O; Hsu C; Seftor C; Celarier E. (1997). Global distribution of UV-absorbing
aerosols from Nimbus 7/TOMS data. J Geophys Res, 102- 16911-16922. 048393
Hernandez L; Probst A; Probst JL; Ulrich E. (2003). Heavy metal distribution in some French forest soils-
Evidence for atmospheric contamination. Sci Total Environ, 312- 195-219. 155841
Herner JD; Ying Q; Aw J; Gao O; Chang DPY; Kleeman MJ. (2006). Dominant Mechanisms that Shape the
Airborne Particle Size and Composition Distribution in Central California. Aerosol Sci Technol, 40- 827"
844. 135981
Herrera LK; Videla HA. (2004). The importance of atmospheric effects on biodeterioration of cultural heritage
constructional materials. Int Biodeterior Biodegradation, 54- 125-134. 155843
Hicke JA; Asner GP; Kasischike ES; French NHF; Randerson JT; Collatz GJ; Stocks BJ; Tucker CJ; Los SO;
Field CB. (2003). Postfire response of North American boreal forest net primary productivity analysed
with satellite observations. , 9: 1145-1157. 156545
Hinzman LD; Fukuda M; Sandberg DV; Chapin III FS; Dash D. (2003). FROSTFIRE: an experimental approach
to predicting the climate feedbacks from the changing boreal fire regime. J Geophys Res, 108 (Dl), 8153,
doi:i0.1029/2001JD000415: . 155845
Hobbelen PHF; Koolhaas JE; van Gestel CAM. (2006). Bioaccumulation of heavy metals in the earthworms
Lumbricus rubellus and Aporrectodea caliginosa in relation to total and available metal concentrations
in field soils. , 144: 639-646. 190371
Hodzic A; Madronich S; Bohn B; Massie S; Menut L; Wiedinmyer C. (2007). Wildfire particulate matter in
Europe during summer 2003: meso-scale modeling of smoke emissions, transport and radiative effects. ,
7: 4043-4064. 156553
Hoell JM; Davis DD; Liu SC; Newell R; Shipham M; Akimoto H; McNeal RJ; Bemdura RJ; Drewry JW. (1996).
Pacific exploratory mission-west A (PEM-WEST A): September- October, 1991. J Geophys Res, 101: 1641"
1653. 190607
Hoell JM; Davis DD; Liu SC; Newell RE; Akimoto H; McNeal RJ; Bendura RJ. (1997). The Pacific exploratory
mission-west phase B: February-March, 1994. J Geophys Res, 102: 28223-28239. 057373
Hoff R; EngeLCox J; Krotkov N; Palm S; Rogers R; McCann K; Sparling L; Jordan N; Torres O; Spinhirne J.
(2004). Long-range transport observations of two large forest fire plumes to the northeastern U.S. In
Proceedings of the 22nd International Laser Radar Conference (pp. 683-686.). Paris: European Space
Agency. 190617
Hoff RM; McCann KJ. (2002). Regional East Atmospheric Lidar Mesonet: REALM. In L. Bissonette; G Roy; G
Vallee (Ed.),Lidar Remote Sensing in Atmospheric and Earth Sciences (pp. 281-284). Val-Belair, Quebec:
Def. R&D Can. Valcartier. 190612
Hofmann D; Barnes J; Dutton E; Deshler T; Jager H; Keen R; Osborn M. (2003). Surface-based observations of
volcanic emissions to stratosphere. In Volcanism and the Earth's Atmosphere (pp. 57—73). Washington,
DC: American Geophysical Union. 192062
Hoh E; Zhu L; Hites RA. (2006). Dechlorane plus, a chlorinated flame retardant, in the Great Lakes. Environ
Sci Technol, 40: 1184-1189. 190378
Holben BN; Eck TF; Slutsker I; Tanre D; Buis JP; Setzer A; Vermote E; Reagan JA; Kaufman YJ; Nakajima T;
Lavenu F; Jankowiak I; Smirnov A. (1998). AERONET: A federated instrument network and data
archive for aerosol characterization. , 66: 1-16. 155848
Holben BN; Smirnov A; Eck TF; Slutsker I; Abuhassan N; Newcomb WW; Schafer JS; Tanre D; Chatenet B;
Lavenu F. (2001). An emerging groundbased aerosol climatology: aerosol optical depth from AERONET.
J Geophys Res, 106: 12067-12098. 190618
Holland EA; Dentener FJ; Braswell BH; Sulzman JM. (1999). Contemporary and pre-industrial global reactive
nitrogen budgets. , 46: 7-43. 092051
July 2009
9-281
DRAFT-DO NOT CITE OR QUOTE

-------
Holmqvist Mi Stenroth Pi Berglund Oi Nystrom Pi Olsson Ki Jellyman Di Mcintosh AR; Larsson P. (2006). Low
levels of persistent organic pollutants (POPs) in New Zealand eels reflect isolation from atmospheric
sources. Environ Pollut, 141: 532-538. 190380
Hooda V. (2007). Phytoremediation of toxic metals from soil and waste water. J Environ Biol, 28: 367-376.
190382
Hopke PKi Zhou Li Poirot RL. (2005). Reconciling Trajectory Ensemble Receptor Model Results with Emissions.
Environ Sci Technol, 39: 7980-7983. 156567
Horowitz L. (2006). Past, present, and future concentrations of tropospheric ozone and aerosols: Methodology,
ozone evaluation, and sensitivity to aerosol wet removal. J Geophys Res, 111: D22211.l-D22211.16.
190620
Horowitz LW; Walters Si Mauzerall DLi Emmons LKi Rasch PJi Granier Ci Tie Xi Lamarque J-Fi Schultz MGi
Tyndall GSi Orlando JJi Brasseur GP. (2003). A global simulation of tropospheric ozone and related
tracers: description and evaluation of MOZART, version 2. J Geophys Res, 108: in press. 057770
Houlahan JEi Findlay CSi Schmidt BRi Meyer AHi Kuzmin SL. (2000). Quantitative evidence for global
amphibian population declines. , 404: 752—755. 155853
Howe TSi Billings Si Stolzberg RJ. (2004). Sources of polycyclic aromatic hydrocarbons and hexachlorobenzene
in spruce needles of eastern Alaska. Environ Sci Technol, 38: 3294-3298. 155854
Hoyt Di Frohlich C. (1983). Atmospheric transmission at Davos, Switzerland 1909-1979. Clim Change, 5: 61-71.
190621
Hoyt DV. (1978). A model for the calculation of solar global insolation. , 21: 27"35. 046638
Hsu NCi Tsay SCi King MDi Herman JR. (2004). Aerosol properties over bright-reflecting source regions. IEEE
Trans Geosci Remote Sens, 42: 557-569. 190622
Huebert BJi Bates Ti Russell PBi Shi Gi Kim YJi Kawamura Ki Carmichael Gi Nakajima T. (2003). An overview
of ACE-Asia: Strategies for quantifying the relationships between Asian aerosols and their climatic
impacts. J Geophys Res, 108: 8633. 190623
Hughes MK. (1981). Cycling of trace metals in ecosystems. : . 156578
Hunaiti AAi Al-Oqlah Ai Shannag NMi Abukhalaf IKi Silvestrov NAi Von Deutsch DAi Bayorh MA. (2007).
Toward understanding the influence of soil metals and sulfate content on plant thiols. J Toxicol Environ
Health A, 70: 559-567. 156579
Huneeus Ni O Boucher O. (2007). One-dimensional variational retrieval of aerosol extinction coefficient from
synthetic LIDAR and radiometric measurements. J Geophys Res, 112: D14303.l-D14303.14. 190624
Hung HHi Halsall CJi Blanchard P. (2001). Are PCBs in the Canadian Arctic atmosphere declining? Evidence
from 5 years of monitoring. Environ Sci Technol, 35: 1303-1311. 155856
Hung HHi Halsall CJi Blanchard Pi Li HHi Fellin Pi Stern Gi Rosenberg B. (2002). Temporal trends of
organochlorine pesticides in the Canadian Arctic atmosphere. Environ Sci Technol, 36: 862-868. 155857
Husar RBi Prospero JMi Stowe LL. (1997). Characterization of tropospheric aerosols over the oceans with the
NOAA advanced very high resolution radiometer optical thickness operational product. J Geophys Res,
102: 16,889-16,909. 045900
Intergovernmental Panel on Climate Change (IPCC). (2007). Climate change 2007: the physical science basis
Contribution of Working Group I to the fourth assessment report of the Intergovernmental Panel on
Climate Change. 092765
IOMi Andrews Ei Sheridan Pi Ogren Ji Ferrare R. (2004). In situ aerosol profiles over the Southern Great Plains
cloud and radiation test bed site: 1. Aerosol optical properties. , 109: D06208. 190058
IPCC (Intergovernmental Panel on Climate Change). (1996). Radiative forcing of climate change. In Climate
Change 1995 (pp. np). New York: Cambridge University Press. 190990
IPCC (Intergovernmental Panel on Climate Change). (2007). Changes in Atmospheric Constituents and in
Radiative forcing. In Climate Change 2007 (pp. np). New York: Cambridge University Press. 190988
IPCC. (1995). Radiative forcing of climate change and an evaluation of the IPCC IS92 emission scenarios. In
Climate Change 1994 (pp. np). New York: Cambridge University Press. 190991
IPCC. (2001). Climate Change 2001: The Scientific Basis. Cambridge University Press. Cambridge. 156587
IPCCi IOMi Inoue K"ii Takano Hi Yanagisawa Ri Sakurai Mi Ueki Ni Yoshikawa T. (2007). Effects of diesel
exhaust particles on cytokine production by splenocytes stimulated with lipopolysaccharide. J Appl
Toxicol, 27: 95-100. 189430
Ito Ai Penne JE. (2005). Historical emissions of carbonaceous aerosols from biomass and fossil fuel burning for
the period 1870-2000. Global Biogeochem Cycles, 19: GB2028.1-GB2028.14. 190626
Jacob DJi Crawford JH; Kleb MM; Connors VS; Bendura RJ; Raper JL; Sachse GW; Gille JC; Emmons L; Heald
CL. (2003). The Transport and Chemical Evolution over the Pacific (TRACE-P) aircraft mission: design,
execution, and first results. J Geophys Res, 108: 9000. 190987
July 2009
9-282
DRAFT-DO NOT CITE OR QUOTE

-------
Jacobs SV; Evans GW; Catalano R; Dooley D. (1984). Air pollution and depressive symptomatology- Exploratory
analyses of intervening psychosocial factors. , 7- 260-272. 156596
Jacobson MZ. (2000). A physically-based treatment of elemental carbon optics' implications for global direct
forcing of aerosols. Geophys Res Lett, 27- 217-220. 056378
Jacobson MZ. (2001). GATOR-GCMM: a global through urban scale air pollution and weather forecast model- 1.
Model design and treatment of subgrid soil, vegetation, roads, rooftops, water, sea ice, and snow. J
Geophys Res, 106: 5385-5402. 155864
Jacobson MZ. (2002). Control of fossil-fuel particulate black carbon and organic matter, possibly the most
effective method of slowing global warming. J Geophys Res, 107: 4410. 155865
Jacobson MZ. (2003). Development of mixed-phase clouds from multiple aerosol size distributions and the effect
of the clouds on aerosol removal. J Geophys Res, 108 D8, 4245, doi:i0.1029/2002JD002691. 155866
Jacobson MZ. (2003). Reply to comment by D. P. Chock et al. on "Control of fossil-fuel particulate black carbon
and organic matter, possibly the most effective method of slowing global warming". J Geophys Res, 108
D24, 4770, doi:i0.1029/2003JD003707. 155867
Jacobson MZ. (2003). Reply to comment by J. E. Penner on "Control of fossil-fuel particulate black carbon and
organic matter, possibly the most effective method of slowing global warming". J Geophys Res, 108 (D24),
4772, doi:i0.1029/2003JD003403. 155868
Jacobson MZ. (2003). Reply to comment by J. E. Penner on "Control of fossil-fuel particulate black carbon and
organic matter, possibly the most effective method of slowing global warming. J Geophys Res, 108 (D24),
4772, doi:i0.1029/2003JD003403. 189421
Jacobson MZ. (2003). Reply to comment by J. Feichter et al. on "Control of fossil-fuel particulate black carbon
and organic matter, possibly the most effective method of slowing global warming". J Geophys Res, 108
D24, 4768, doi:i0.1029/2002JD003299,. 155869
Jacobson MZ. (2004). Climate response of fossil fuel and biofuel soot, accounting for soot's feedback to snow and
sea ice albedo and emissivity. J Geophys Res, 109, D21201, doi:i0.1029/2004JD004945. 155870
Jacobson MZ. (2006). Effects of externally-through-internally-mixed soot inclusions within clouds and
precipitation on global climate. J Phys Chem B, 110: 6860-6873. 156599
Jacobson MZ; Kaufman YJ. (2006). Wind reduction by aerosol particles. Geophys Res Lett, 33. 090942
Jacobson MZ; Seinfeld JH; Carmichael GR; Streets DG. (2004). The effect on photochemical smog of converting
the U.S. fleet ofgasoline vehicles to modern diesel vehicles. Geophys Res Lett, 31: L02116. 180362
James SM; Little EE; Semlitsch RD. (2004). Effects of multiple routes of cadmium exposure on the hibernation
success of the American toad (Bufo americanus). Arch Environ Contam Toxicol, 46: 518-527. 155874
Jantunen LMM; Bidleman TF. (1996). Air-water gas exchange of hexachlorocyclohexanes (HCHs) and the
enantiomers of crHCH in Arctic regions. J Geophys Res, 101: 28837-28846. 155876
Jantunen LMM; Bidleman TF. (1998). Organochlorine pesticides and enantiomers of chiral pesticides in Arctic
Ocean water. Arch Environ Contam Toxicol, 35. 155877
Jayanty R. (2003). Overview of PM2.5 chemical speciation network program. Presented at . 156605
Jayne JT; Leard DC; Zhang X; Davidovits P; Smith KA; Kolb CE; Worsnop DR. (2000). Development of an
aerosol mass spectrometer for size and composition analysis of submicron particles. Aerosol Sci Technol,
33: 49-70. 190978
Jeong MJ; Li Z; Chu DA; Tsay SC. (2005). Quality and Compatibility Analyses of Global Aerosol Products
Derived from the Advanced Very High Resolution Radiometer and Moderate Resolution Imaging
Spectroradiometer. J Geophys Res, 110: np. 190977
Jiang H; Feingold G. (2006). Effect of aerosol on warm convective clouds: Aerosol-cloud-surface flux feedbacks in
a new coupled large eddy model. J Geophys Res, 111: D01202. 190976
Jiang J; Oberdrster G; Elder A; Gelein R; Mercer P; Biswas P. (2008). Does nanoparticle activity depend upon
size and crystal phase?. , 2: 33"42. 156609
Jiang W; Smyth S; Giroux E; Roth H; Yin D. (2006). Differences between CMAQ fine mode particle and PM 25
concentrations and their impact on model performance evaluation in the lower Fraser valley. Atmos
Environ, 40: 4973-4985. 133165
Jiao XC; Xu FL; Dawson R; Chen SH; Tao S. (2007). Adsorption and absorption of polycyclic aromatic
hydrocarbons to rice roots. , 148: 230-235. 155879
Jirak IL; Cotton WR. (2006). Effect of air pollution on precipitation along the Front Range of the Rocky
Mountains. , 45: 236-245. 156612
Johnson D; Hale B. (2008). Fine root decomposition and cycling of Cu, Ni, Pb, and Zn at forest sites near
smelters in Sudbury, ON, and Rouyn-Noranda, QU, Canada. , 14: 41-53. 155881
Johnson DB. (1982). The role of giant and ultragiant aerosol particles in warm rain initiation. , 39: 448-460.
190973
July 2009
9-283
DRAFT-DO NOT CITE OR QUOTE

-------
Johnson MS; Franke LS; Lee RB; Holladay SD. (1999). Bioaccumulation of 2,4,6-trinitrotoluene and
polychlorinated biphenyls through two routes of exposure in a terrestrial amphibian- Is the dermal route
significant?. Environ Toxicol Chem, 18: 873-878. 155880
Jones GS; Jones A; Roberts DL; Stott PA; Williams KD. (2005). Sensitivity of global-scale climate change
attribution results to inclusion of fossil fuel black carbon aerosol. Geophys Res Lett, 32, L14701,
doi:i0.1029/2005GL023370. 155885
Jones JW; Bogat GA. (1978). Air pollution and human aggression. Psychol Rep, 43: 721-722. 190396
Jones PD; Moberg A; Osborn TJ; Briffa KR. (2003). Surface climate responses to explosive volcanic eruptions
seen in long European temperature records and mid-to-high latitude tree-ring density around the
Northern Hemisphere. In Robock A; Oppenheimer C (Ed.),Volcanism and the earth's atmosphere (pp.
239-254). Washington, DC: American Geophysical Union. 192278
Joshi M; Shine K. (2003). A GCM study of volcanic eruption s as a cause of increased stratospheric water vapor.
J Clim, 16: 3525-3534. 192327
Joynt J; Bischoff M; Turco R; Konopka A; Nakatsu CH. (2006). Microbial community analysis of soils
contaminated with lead, chromium and petroleum hydrocarbons. Microb Ecol, 51: 209-219. 155887
Junker C; Liousse C. (2008). A global emission inventory of carbonaceous aerosol from historic records of fossil
fuel and biofuel consumption for the period 1860-1997. ,8- 1195-1207. 190971
Kahn R; Banerjee P; McDonald D. (2001). The sensitivity of multiangle imaging to natural mixtures of aerosols
over ocean. J Geophys Res, 106: 18219-18238. 190969
Kahn R; Banerjee P; McDonald D; Diner D. (1998). Sensitivity of multiangle imaging to aerosol optical depth,
and to pure-particle size distribution and composition over ocean. J Geophys Res, 103: 32195-32213.
190970
Kahn R; Gaitley B; Martonchik J; Diner D; Crean K; Holben B. (2005). Multiangle Imaging Spectroradiometer
(MISR) global aerosol optical depth validation based on 2 years of coincident Aerosol Robotic Network
(AERONET) observations. J Geophys Res, 110: D10S04. 189961
Kahn R; Li W', Martonchik J; Bruegge C; Diner D; Gaitley B; Abdou W; Dubovik 0; Holben B; Smirnov A; Jin Z;
Clark D. (2005). MISR low-light-level calibration, and implications for aerosol retrieval over dark water. ,
62: 1032-1052. 190965
Kahn R; Li W', Moroney C; Diner D; Martonchik J; Fishbein E. (2007). Aerosol source plume physical
characteristics from space-based multiangle imaging. J Geophys Res, 112: D11205. 190964
Kahn RA Gaitley BJ; Martonchik JV; Diner DJ Crean KA; Holben B. (2005). MISR global aerosol optical depth
validation based on two years of coincident AERONET observations. J Geophys Res, 110: D10S04.
190966
Kahn RA Garay MJ; Nelson DL; Yau KK Bull M; Gaitley BJ; Martonchik JV; Levy R. (2007). Satellite-derived
aerosol optical depth over dark water from MISR and MODIS: Comparisons with AERONET and
implications for climatological studies. J Geophys Res, 112: D18205. 190963
Kalashnikova O; Kahn R. (2006). Ability of multiangle remote sensing observations to identify and distinguish
mineral dust types: Part 2 Sensitivity over dark water. J Geophys Res, 111: D11207. 190962
Kaldorf M; Kuhn AJ; Schroder WH; Hildebrandt U; Bothe H. (1999). Selective element deposits in maize
colonized by a heavy metal tolerance conferring arbuscular mycorrhizal fungus. J Plant Physiol, 154:
718-728. 190399
Kamh GME. (2005). The impact of landslides and salt weathering on Roman structures at high latitudes -
Conway Castle, Great Britain: A case study. , 48: 238-254. 155888
Kandeler E; Kampichler C; Horak O. (1996). Influence of heavy metals on the functional diversity of soil
microbial communities. , 23: 299-306. 094392
Kappos AD; Bruckmann P; Eikmann T; Englert N; Heinrich U; Hoppe P; Koch E; Krause GH; Kreyling WG;
Rauchfuss K; Rombout P; Schulz-KLemp V; Thiel WR; Wichmann HE. (2004). Health effects of particles
in ambient air. Int J Hyg Environ Health, 207: 399-407. 087922
Kapustin VN; Clarke AD; Shinozuka Y; Howell S; V Brekhovskikh V; Nakajima T; Higurashi A. (2006). On the
determination of a cloud condensation nuclei from satellite: Challenges and possibilities. J Geophys Res,
111: DQ4202.190961
Kaufman Y. (1987). Satellite sensing of aerosol absorption. J Geophys Res, 92: 4307-4317. 190960
Kaufman Y; Boucher O; Tanre D; Chin M; Remer L; Takemura T. (2005). Aerosol anthropogenic component
estimated from satellite data. Geophys Res Lett, 32: L17804. 155891
Kaufman Y; Fraser R. (1997). The effect of smoke particles on clouds and climate forcing. , 277: 1636-1639.
190958
Kaufman Y; Haywood J; Hobbs P; Hart W', KLeidman R; Schmid B. (2003). Remote sensing of vertical
distributions of smoke aerosol off the coast of Africa. Geophys Res Lett, 30: 1831. 190954
July 2009
9-284
DRAFT-DO NOT CITE OR QUOTE

-------
Kaufman YD; Tanre D; Boucher O. (2002). A satellite view of aerosols in the climate system. Nature, 419- 215"
223. 190956
Kaufman YJ; Hobbs PV; Kirchoff VWJH; Artaxo P; Remer LA; Holben BN; King MD; Ward DE; Prins EM;
Longo KM; Mattos LF; Nobre CA; Spinhirne JD; Ji Q; Thompson AM; Gleason JF; Christopher SA; Tsay
S"C. (1998). Smoke, clouds, and radiation-'Brazil (SCAR-B) experiment. J Geophys Res, 103- 31783"
31808. 089989
Kaufman YJ; Koren I. (2006). Smoke and pollution aerosol effect on cloud cover. , 313- 655-658. 190951
Kaufman YJ; Martins JV; Remer LA; Schoeberl MR; Yamasoe MA. (2002). Satellite retrieval of aerosol
absorption over the oceans using sunglint. Geophys Res Lett, 29- 34.1-34.4. 190955
Kaufman YJ; Nakajima T. (1993). Effect of Amazon smoke on cloud microphysics and albedo—Analysis from
satellite imagery. , 32: 729-744. 190959
Kaufman YJ; Setzer A; Ward D; Tanre D; Holben BN; Menzel P; Pereira MC; Rasmussen R. (1992). Biomass
Burning Airborne and Spaceborne Experiment in the Amazonas (BASE-A). J Geophys Res, 97: 14581-
14599. 044557
Kavouras IG; Etyemezian V; DuBois DW! Xu J; Pitchford M. (2009). Source reconciliation of atmospheric dust
causing visibility impairment in Class I areas of the western United States. , 114: x. 191976
Kavouras IG; Etyemezian V; Xu J; DuBois DW; Green M; Pitchford M. (2007). Assessment of the local
windblown component of dust in the western United States. J Geophys Res, 112: . 156630
Kenski DM; Gay D; Fitzsimmons S. (2004). Ammonia and its role in midwestern haze. Presented at Regional
and Global Perspectives on Haze: Causes, Consequences and Controversies Visibility Specialty
Conference, Asheville, NC. 192078
Kerr R. (2007). Another global warming icon comes under attack. , 317: 28"29. 190950
Kiehl JT. (2007). Twentieth century climate model response and climate sensitivity. Geophys Res Lett, 34: np.
190949
Kiikkila O. (2003). Heavy-metal pollution and remediation of forest soil around the Harjavalta Cu_Ni smelter, in
SW Finland. , 37: 399-415. 156637
Kim CG; Bell JNB; Power SA. (2003). Effects of soil cadmium on Pinus sylvestris L. seedlings. Plant Soil, 257:
443-449. 155899
Kim MK; Lau KM; Chin M; Kim KM; Sud Y; Walker GK. (2006). Atmospheric teleconnection over Eurasia
induced by aerosol radiative forcing during boreal spring. Proc Natl Acad Sci USA, 19: 4700-4718.
190917
Kim SW; Yoon SC; Kim J. (2008). Columnar Asian dust particle properties observed by sun/sky radiometers
from 2000 to 2006 in Korea. Atmos Environ, 42: 492-504. 130785
Kim Y; Hatsushika H; Muskett RR; Yamazaki K. (2005). Possible effect of boreal wildfire soot on Arctic sea ice
and Alaska glaciers. Atmos Environ, 39: 3513-3520. 155900
King M; Kaufman Y; Tanre DiNakajima T. (1999). Remote sensing of tropospheric aerosols from space: Past,
present, and future. , 80: 2229-2259. 190635
King MD; Platnick S; Moeller CC; Revercomb HE; Chu DA. (2003). Remote sensing of smoke, land, and clouds
from the NASA ER-2 during SAFARI 2000. J Geophys Res, 108: SAF38.1-SAF38.12. 094395
Kinne S; Schulz M; Textor C; Guibert S; Balkanski Y; Bauer SE; Berntsen T; Berglen TF; Boucher O; Chin M;
Collins W; Dentener F; Diehl T; Easter R; Feichter J; Fillmore D; Ghan S; Ginoux P; Gong S; Grini A;
Hendricks J; Herzog M; Horowitz L; Isaksen I; Iversen T; Kirkevag A; Kloster S; Koch D; Kristjansson
JE; Krol M; Lauer A; Lamarque JF; Lesins G; Liu X; Lohmann U; Montanaro V; Myhre G; Penner J;
Pitari G; Reddy S; Seland O; Stier P; Takemura T; Tie X. (2006). An AeroCom initial assessment—optical
properties in aerosol component modules of global models. , 6: 1815-1834. 155903
Kirchstetter TW; Harley RA; Kreisberg NM; Stolzenberg MR; Hering SV. (1999). On-road measurement of fine
particle and nitrogen oxide emissions from light" and heavy-duty motor vehicles. Atmos Environ, 33:
2955-2968. 010642
Kirkbride MP; Dugmore JA. (2003). Glaciological response to distal tephra fallout from the 1947 eruption of
Hekla, south Iceland. , 49: 420-428. 156645
Kleeman MJ; Hughes LS; Allen JO; Cass GR. (1999). Source contributions to the size and composition
distribution of atmospheric particles: southern California in September 1996. Environ Sci Technol, 33:
4331-4341. 011286
Kleidman R; O'Neill N; Remer L; Kaufman Y; Eck T; Tanre D; Dubovik O; Holben B. (2005). Comparison of
Moderate Resolution Imaging Spectroradiometer (MODIS) and Aerosol Robotic Network (AERONET)
remote-sensing retrievals of aerosol fine mode fraction over ocean. J Geophys Res, 110: 1"16. 190175
July 2009
9-285
DRAFT-DO NOT CITE OR QUOTE

-------
KLeinman MT; Araujo JA; Nel A; Sioutas Ci Campbell A; Cong PQ; Li H; Bondy SC. (2008). Inhaled ultrafine
particulate matter affects CNS inflammatory processes and may act via MAP kinase signaling pathways.
Toxicol Lett, 178: 127-130. 190074
Knutti R; Stocker TF; Joos F; Plattner GK. (2002). Constraints on radiative forcing and future climate change
from observations and climate model ensembles. Nature, 416: 719-723. 190178
Knutti R; Stocker TF; Joos F> Plattner GK. (2003). Probabilistic climate change projections using neural
networks. , 21: 257-272. 190180
Koch D. (2001). Transport and direct radiative forcing of carbonaceous and sulfate aerosols in the GISS GCM. J
Geophys Res, 106: 20311-20332. 192054
Koch D; Bond TC; Streets D; Unger Ni van der Werf GR. (2007). Global impact of aerosols from particular
source regions and sectors. J Geophys Res, 112: D02205. 190185
Koch D; Hansen J. (2005). Distant origins of Arctic black carbon: A Goddard Institute for Space Studies ModelE
experiment. J Geophys Res, 110: D04204. 190183
Koch D; Schmidt G> Field C. (2006). Sulfur, sea salt, and radionuclide aerosols in GISS ModelE. J Geophys Res,
111: DQ6206.190184
Kogan YL; Lilly DK; Kogan ZN; Filyushkin V. (1994). The effect of CCN regeneration on the evolution of
stratocumulus cloud layers. , 33: 137-150. 190186
Koren L Kaufman YJ; Remer LA; Martins JV. (2004). Measurement of the effect of Amazon smoke on inhibition
of cloud formation. Science, 303: 1342-1345. 190187
Koren I; Kaufman YJ; Rosenfeld D; Remer LA; Rudich Y. (2005). Aerosol invigoration and restructuring of
Atlantic convective clouds. Geophys Res Lett, 32: L14828. 190188
Koren I; Martins JV; Remer LA; Afargan H. (2008). Smoke invigoration versus inhibition of clouds over the
Amazon. , 321: 946-949. 190193
Koren I; Remer KA; Kaufman YJ; Rudich Y; Martins JV. (2007). On the twilight zone between clouds and
aerosols. Geophys Res Lett, 34: L08805. 190192
Koren I; Remer LA; Longo K. (2007). Reversal of trend of biomass burning in the Amazon. Geophys Res Lett, 34:
L20404. 190189
Korsnes R; Pavlova O; Godtliebsen F. (2002). Assessment of potential transport of pollutants into the Barents
Sea via sea ice--an observational approach. Mar Pollut Bull, 44: 861-869. 156657
Kroll JH; Ng NL; Murphy SM; Flagan RC; Seinfeld JH. (2006). Secondary organic aerosol formation from
isoprene photooxidation. Environ Sci Technol, 40: 1869-1877. 190195
Krueger AJ; Schaefer SJ; Krotkov N; Bluth G; Barker S. (2000). Ultraviolet remote sensing ofvolcanic
emissions. In Remote Sensing of Active Volcanism (pp. 25-43). Washington, DC: American Geophysical
Union. 192234
Kruger O; Grasl H. (2002). The indirect aerosol effect over Europe. Geophys Res Lett, 29: 1925. 190200
Kruysse A. (1971). Acute inhalation toxicity of acrolein in hamsters. Central Institute for Nutrition and Food
Research. The Netherlands. R 3516. 192236
Kucera T; Horakova H; Sonska A. (2008). Toxic metal ions in photoautotrophic organisms. , 46: 481-489. 190408
Kucharski J; Wyszkowska J. (2004). Inter-relationship between number of microorganisms and spring barley
yield and degree of soil contamination with copper. , 50: 243—249. 156662
Kuki KN; Oliva MA; Costa AC. (2009). The simulated effects of iron dust and acidity during the early stages of
establishment of two coastal plant species. , 196: 287-295. 190411
Kuki KN; Oliva MA; Pereira EG; Costa AC; Cambraia J. (2008). Effects of simulated deposition of acid mist and
iron ore particulate matter on photosynthesis and the generation of oxidative stress in Schinus
terebinthifolius Radii and Sophora tomentosa L. , 403(l"3): 207-214. 155346
LaBrecque JJ; Benzo Z; Alfonso JA; Cordoves Manuelita Quintal PR; Gomez CV Marcano E. (2004). The
concentrations of selected trace elements in clams, Trivela mactroidea along the Venezuelan coast in the
state of Miranda. Mar Pollut Bull, 49: 659-667. 155913
Lack D; Lovejoy E; Baynard T; Pettersson A; Ravishankara A. (2006). Aerosol absorption measurements using
photoacoustic spectroscopy: sensitivity, calibration, and uncertainty developments. Aerosol Sci Technol,
40: 697-708. 190203
Lakzian A; Murphy P; Turner A; Beynon JL; Giller KE. (2002). Rhizobium leguminosarum bv. viciae
populations in soils with increasing heavy metal contamination: abundance, plasmid profiles, diversity
and metal tolerance. Soil Biol Biochem, 34: 519-529. 156671
Lambert A; Grainger RG; Remedios JJ; Rodgers CD; Corney M; Taylor FW. (1993). Measurements of the
evolution of the Mt. Pinatubo aerosol cloud by ISAMS. Geophys Res Lett, 20: 1287-1290. 192231
July 2009
9-286
DRAFT-DO NOT CITE OR QUOTE

-------
Lambert MRK. (1997). Environmental effects of heavy spillage from a destroyed pesticide store near Hargeisa
(Somaliland) assessed during the dry season, using reptiles and amphibians as bioindicators. Arch
Environ Contam Toxicol, 32- 80—93. 155916
Landers DH; Simonich SL; Jaffe DA; Geiser LH; Campbell DH; Schwindt AR; Schreck CB; Kent ML; Hafner
WD; Taylor HE; Hageman KJ; Usenko S; Ackerman LK; Schrlau JE; Rose NL; Blett TF; Erway MM.
(2008). The Fate, Transport and Ecological Impacts of Airborne Contaminants in Western National
Parks (USA). U.S. Environmental Protection Agency, Office of Research and Development, NHEERL,
Western Ecology Division. Corvallis, Oregon. 191181
Lanno R; Wells J; Conder J; Bradham K; Basta N. (2004). The bioavailability of chemicals in soil for
earthworms. Ecotoxicol Environ Saf, 57- 39"47. 190415
Larson VE; Golaz J-C; Jiang H; Cotton WR. (2005). Supplying local microphysics parameterizations with
information about subgrid variability- Latin hypercube sampling. J Trace Elem Med Biol, 62- 4010-4026.
190220
Larson VE; Wood R; Field PR; Golaz J-C. (2001). Small-scale and mesoscale variability of scalars in cloudy
boundary layers' One-dimensional probability density functions. J Trace Elem Med Biol, 58- 1978-1996.
190212
Latimer DA; Ireson RG. (1980). Workbook for estimating visibility impairment. U.S. Enviornmental Protection
Agency. Research Triangle Park, NC. EPA-450/4-80-031. 035723
Lau K-M; Kim K"M. (2006). Observational relationships between aerosol and Asian monsoon rainfall, and
circulation. Geophys Res Lett, 33: L21810. 190226
Lau K-M; Kim K"M; Walker G; Sud G. (2008). A GCM study of the possible impacts of Saharan dust heating on
the water cycle and climate of the tropical Atlantic and Caribbean regions. Proc Natl Acad Sci U S A, in
press^ in press. 190229
Lau K; Kim M; Kim K. (2006). Asian summer monsoon anomalies induced by aerosol direct forcing—the role of
the Tibetan Plateau. , 36: 855-864. 190223
Law RJ; Alaee M; Allchin CR; Boon JP; Lebeuf M; Lepom P; Ster GA. (2003). Levels and trends of
polybrominated diphenylethers and other brominated flame retardants in wildlife. Environ Int, 29: 757"
770. 190420
Lawson ST; Scherbatskoy TD; Malcolm EG; Keeler GJ. (2003). Cloud water and throughfall deposition of
mercury and trace elements in a high elevation spruce-fir forest at Mt Mansfield, Vermont. J Environ
Monit, 5: 578-583. 089371
Leahy L; Anderson T; Eck T; Bergstrom R. (2007). A synthesis of single scattering albedo of biomass burning
aerosol over southern Africa during SAFARI 2000. Geophys Res Lett, 34: L12814. 190232
Leaitch WR; Banic CM; Isaac GA; Couture MD; Liu PSK; Gultepe I; et al. (1996). Physical and chemical
observations in marine stratus during the 1993 North Atlantic Regional Experiment: Factors controlling
cloud droplet number concentrations. J Geophys Res, 101: 29123-29135. 190354
Leaitch WR; Isaac GA; Strapp JW; Banic CM; Wiebe HA. (1992). The relationship between cloud droplet
number concentrations and anthropogenic pollution: observations and climatic implications. J Geophys
Res, 97: 2463-2474. 045270
LeBlanc GA. (1995). Trophic-level differences in the bioconcentration of chemicals: Implications in assessing
environmental biomagnification. Environ Sci Technol, 29: 154-160. 155921
Lee JK. (2006). Toxicity and tissue distribution of magnetic nanoparticles in mice [erratum]. Toxicol Sci, 90:
267. 088968
Lee T; Yu XY; Ayres B; Kreidenweis SM; Malm WC; Collett JL. (2008). Observations of fine and coarse particle
nitrate at several rural locations in the United States. Atmos Environ, 42: 2720-2732. 156686
Lee TE; Miller SD; Turk FJ; Schueler C; Julian R; Deyo S; et al. (2006). The NPOESS VIIRS day/night visible
sensor. , 87: 191-199. 190358
Lei YD; Wania F. (2004). Is rain or snow a more efficient scavenger of organic chemicals?. Atmos Environ, 38:
3557-3571. 127880
Lelieveld J; Berresheim H; Borrmann S; Crutzen PJ; Dentener FJ; Fischer H; et al. (2002). Global air pollution
crossroads over the Mediterranean . , 298: 794-799. 190361
Levin Z; Cotton WR. (2008). Report from the WMO/ IUGG International Aerosol Precipitation Science,
Assessment Group (IAPSAG). In Aerosol pollution impact on precipitation: A scientific review (pp. 482).
Geneva, Switzerland: World Meteorological Organization. 190375
Levine JS; Pinto JP. (1998). The production of CO by biomass burning. Presented at . 029599
Levy R; Remer L; Dubovik O. (2007). Global aerosol optical properties and application to MODIS aerosol
retrieval over land. J Geophys Res, 112: D13210. 190377
July 2009
9-287
DRAFT-DO NOT CITE OR QUOTE

-------
Levy Ri Remer Li Mattoo Si Vermote Ei Kaufman Y. (2007). Second-generation algorithm for retrieving aerosol
properties over land from MODIS spectral reflectance. J Geophys Res, 112- D13211. 190379
Lewis ERi Schwartz SE. (2004). Sea salt aerosol production- Mechanisms, methods, measurements and models'
A critical review. Washington, DC- American Geophysical Union. 192023
Li R-Ri Kaufman YJi Hao WMi Salmon JMi Gao B"C . (2004). A technique for detecting burn scars using
MODIS data. IEEE Trans Geosci Remote Sens, 42- 1300-1308. 190386
Li Y-F; MacDonald RW; Jantunen LMMi Harner T; Bidleman TFi Stachen WMJ. (2002). The transport of 8"
hexachlorocyclohexane to the western Arctic Ocean- a contrast to aHCH. Sci Total Environ, 291- 229-246.
156691
Li Zi Chen Hi Cribb Mi Dickerson Ri Holben Bi Li Ci et al. (2007). Preface to special section on East Asian
studies of tropospheric aerosols' An international regional experiment (EAST-AIRE). J Geophys Res, 112-
D22S00.190392
Lin CJi Pehkonen SO. (1999). The chemistry of atmospheric mercury- a review. Atmos Environ, 33- 2067-2070.
190426
Lin Hi Tao Si Zuo Qi Coveney RM. (2007). Uptake of polycyclic aromatic hydrocarbons by maize plants. Environ
Pollut, 148: 614-619. 155933
Lindberg SEi Brooks Si Linn CJi Scott KJi Landis MSi Stevens RKi Goodsite Mi Richter A. (2002). Dynamic
oxidation of gaseous mercury in the Arctic troposphere at polar sunrise. Environ Sci Technol, 36: 1245"
1256. 190429
Lindesay JAi Andreae MOi Goldammer JGi Harris Gi Annegarn HJi Garstang Mi et al. (1996). International
Geosphere Biosphere Programme/International Global Atmospheric Chemistry SAFARI "92 field
experiment: Background and overview. J Geophys Res, 23521-23530: 101. 190403
Lingua Gi Franchin Ci Todeschini Vi Castiglione Si SBiondii Burlando Bi Parravicini Vi PTorrigianii GBerta.
(2008). Arbuscular mycorrhizal fungi differentially affect the response to high zinc concentrations of two
registered poplar clones. , 153: 137-147. 155935
Liou KNi Ou S"C. (1989). The role of cloud microphysical processes in climate: an assessment from a one-
dimensional perspective. J Geophys Res, 94: 8599-8607. 190407
Liousse C> Penner JEi Chuang Ci Walton JJi Eddleman Hi Cachier H. (1996). A global three-dimensional model
study of carbonaceous aerosols. J Geophys Res, 101: 19,411-19,432. 078158
Liu Hi Pinker Ri Holben B. (2005). A global view of aerosols from merged transport models, satellite, and
ground observations. J Geophys Res, 110: D10S15. 190414
Liu J; Ballaney Mi Al-Alem Ui Quan Ci Jin Xi Perera Fi Chen LCi Miller RL. (2008). Combined Inhaled Diesel
Exhaust Particles and Allergen Exposure Alter Methylation of T Helper Genes and IgE Production In
Vivo. Toxicol Sci, 102: 76-81. 156709
Liu Li Lacis AAi Carlson BEi Mishchenko ML Cairns B. (2006). Assessing Goddard Institute for Space Studies
ModelE aerosol climatology using satellite and ground-based measurements. J Geophys Res, 111:
D20212.190422
Liu Xi Penner Ji Das Bi Bergmann Di Rodriguez Ji Strahan Si et al. (2007). Uncertainties in global aerosol
simulations: Assessment using three meteorological data sets. J Geophys Res, 112: D11212. 190427
Liu Yi Koutrakis P. (2007). Estimating Fine Particulate Matter Component Concentrations and Size
Distributions Using Satellite- Retrieved Fractional Aerosol Optical Depth: Part 1—Method Development.
J Air Waste Manag Assoc, 57: 1351-1359. 187007
Liu, Yi Franklin, Mi Kahn, Ri Koutrakis, P. (2007). Using aerosol optical thickness to predict ground-level PM
sub(2) sub(.) sub(5) concentrations in the St. Louis area: A comparison between MISR and MODIS . , 107:
33-44. 192180
Loeb Ni Kato S. (2002). Top-of-atmosphere direct radiative effect of aerosols over the tropical oceans from the
Clouds and the Earth's Radiant Energy System (CERES) satellite instrument. Proc Natl Acad Sci USA,
15: 1474-1484. 190432
Loeb Ni Manalo-Smith N. (2005). Top-of-Atmosphere direct radiative effect of aerosols over global oceans from
merged CERES and MODIS observations. J Clim, 18: 3506-3526. 190433
Loeb NGi Kato Si Loukachine Ki Smith NM. (2005). Angular distribution models for top-of-atmosphere radiative
flux estimation from the Clouds and the Earth's Radiant Energy System instrument on the Terra
Satellite, part I: Methodology. , 22: 338—351. 190436
Lohman Ui Feichter J. (2005). Global indirect aerosol effects: a review. , 5- 715-737. 155942
Lohmann Ui Feichter Ji Chuang CCi Penner JE. (1999). Prediction of the number of cloud droplets in the
ECHAM GCM. J Geophys Res, 104: 9169-9198. 190443
Lohmann Ui Koren Ii Kaufman YJ. (2006). Disentangling the role of microphysical and dynamical effects in
determining cloud properties over the Atlantic. Geophys Res Lett, 33: L09802. 190451
July 2009
9-288
DRAFT-DO NOT CITE OR QUOTE

-------
Lohmann Ui Leaitch WR; Barrie L; Law K; Yi Y> Bergmann D; et al. (2001). Vertical distributions of sulfur
species simulated by large scale atmospheric models in COSAM- Comparison with observations. , 53- 646
- 672. 190448
Lorusso Si Marabelli Mi Troili M. (1997). Air pollution and the deterioration of historic monuments. J Environ
Pathol Toxicol Oncol, 16: 171-173. 084534
Lowenthal DHi Kumar N. (2003). PM2.5 mass and light extinction reconstruction in IMPROVE. J Air Waste
Manag Assoc, 53: 1109-1120. 156712
Lowenthal DHi Rogers CFi Saxena Pi Watson JGi Chow JC. (1995). Sensitivity of estimated light extinction
coefficients to model assumptions and measurement errors. Atmos Environ, 29: 751-766. 045134
Lu M-Li Feingold Gi Jonsson Hi Chuang Pi Gates Hi Flagan RCi Seinfeld JH. (2008). Aerosol-cloud
relationships in continental shallow cumulus. J Geophys Res, 113: D15201. 190455
Lubin Di Vogelmann AM. (2006). A climatologically significant aerosol longwave indirect effect in the Arctic.
Nature, 439: 453-456. 190466
Lubin DSi Satheesh Si McFarquar Gi Heymsfield A. (2002). Longwave radiative forcing of Indian Ocean
tropospheric aerosol. J Geophys Res, 107: 8004. 190463
Lucaciu Ai Timofte Li Culicov Oi Frontasyeva MVi Oprea Ci Cucu-Man Si Mocanu Ri Steinnes E. (2004).
Atmospheric deposition of trace elements in Romania studied by the moss biomonitoring technique. , 49:
533-548. 155947
Lunackova Li Masarovicova Ei Kraova Ki Stresko V. (2003). Response of fast growing woody plants from family
Salicaceae to cadmium treatment. , 70: 576-585. 155948
Luo Yi Lu Di Zhou X> Li W> He Q. (2001). Characteristics of the spatial distribution and yearly variation of
aerosol optical depth over China in last 30 years. J Geophys Res, 106: 14501. 190467
Leon Ji Tanre D; Pelon J; Kaufman Yi Haywood Ji Chatenet B. (2003). Profiling of a Saharan dust outbreak
based on a synergy between active and passive remote sensing. J Geophys Res, 108: 8575. 190366
Lopez Alonso Mi Benedito JLi Miranda Mi Castillo Ci Hernandez Ji Shore RF. (2002). Cattle as biomonitors of
environmental semi-metal and trace metal concentrations in Galicia (NW Spain). Arch Environ Contam
Toxicol, 43: 103-108. 155943
Lopez Alonso Mi Benedito JLi Miranda Mi Castillo Ci Hernandez Ji Shore RF. (2003). Mercury concentrations in
cattle from NW Spain. Sci Total Environ, 302: 93-100 . 155945
Lopez Alonso Mi Benedito JLi Miranda Mi Fernandez JAi Castillo Ci Hernandez Ji Shore RF. (2003). Large-
scale spatial variation in mercury concentrations in cattle in NW Spain. Environ Pollut, 125: 173-181.
155944
Mace BLi Bell PA; Loomis RJ. (2004). Visibility and Natural Quiet in National Parks and Wilderness Areas:
Psychological Considerations. , 36: 5-31. 180255
Mackay D. (1991). Multimedia environmental models: the fugacity approach. Chelsea, MI: Lewis Publishers.
042941
Macleod M; McKone TE; Mackay D. (2005). Mass balance for mercury in the San Francisco Bay area. Environ
Sci Technol, 39: 6721-6729. 155954
Madrid L; Diaz-Barrientos E> Madrid F. (2002). Distribution of heavy metal contents of urban soils in parks of
Seville. Chemosphere, 49: 1301—1309. 155956
Magi B; Hobbs P; Kirchstetter T> Novakov T> Hegg D; Gao Si et al. (2005). Aerosol properties and chemical
apportionment of aerosol optical depth at locations off the United States East Coast in July and August
2001. J Trace Elem Med Biol, 62: 919-933. 190468
Malm W; Bell P; McGlothin GE. (1984). Field testing a methodology for assessing the importance of good visual
air quality. Presented at . 044292
Malm WC. (1999). Introduction to visibility. 025037
Malm WC; Day DE. (2001). Estimates of aerosol species scattering characteristics as a function of relative
humidity. Atmos Environ, 35: 2845-2860. 190431
Malm WC; Day DE; Kreidenweis SM; Collett JL; Lee T. (2003). Humidity dependent optical properties of fine
particle during the Big Bend Regional Aerosol and Visibility Observational Study (BRAVO). , 108: 4279.
190434
Malm WC; Hand JL. (2007). An examination of the physical and optical properties of aerosols collected in the
IMPROVE program. Atmos Environ, 41: 3407-3427. 155962
Malm WC; Pitchford ML. (1997). Comparison of calculated sulfate scattering efficiencies as estimated from size-
resolved particle measurements at three national locations. Atmos Environ, 31: 1315-1325. 002519
Malm WC; Pitchford ML; McDade C; Ashbaugh LL. (2007). Coarse particle speciation at selected locations in
the rural continental United States. Atmos Environ, 41: 2225-2239. 156730
July 2009
9-289
DRAFT-DO NOT CITE OR QUOTE

-------
Malm WC; Schichtel BA; Ames RB; Gebhart KA. (2002). A ten-Year spatial and temporal trend of sulfate across
the United States. J Geophys Res, : . 156727
Malm WC; Schichtel BA; Pitchford ML; Ashbaugh LL; Eldred RA. (2004). Spatial and monthly trends in
speciated fine particle concentration in the United States. J Geophys Res, 109- 3306. 156728
Malm WC; Trijonis J; Sisler J; Pitchford M; Dennis RL. (1994). Assessing the effect of S02 emission changes on
visibility. Atmos Environ, 28: 1023-1034. 044920
Marinoni N> Birelli MP; Rostagno C; Pavese A. (2003). The effects of atmospheric multipollutants on modern
concrete. Atmos Environ, 37: 4701-4712. 092520
Markiewicz Patkowska J; Hursthouse A; Przybyla-Kij H. (2005). The interaction of heavy metals with urban
soils: sorption behaviour of Cd, Cu, Cr, Pb and Zn with a typical mixed brownfield deposit. Environ Int,
31: 513-521. 155963
Martin MH; Coughtrey PJ. (1981). Impact of metals on ecosystem function and productivity. In Lepp, N. W.
(Ed.),Effect of heavy metal pollution on plants: volume 2, metals in the environment (pp. 119-158).
Barking, United Kingdom: Applied Science Publishers. 047727
Martins J; Tanre D; Remer L; Kaufman Y; Mattoo S; Levy R; et al. (2002). MODIS cloud screening for remote
sensing of aerosol over oceans using spatial variability. Geophys Res Lett, 29: MOD4. 190470
Martonchik J; Diner D; Crean K; Bull M. (2002). Regional aerosol retrieval results from MISR. IEEE Trans
Geosci Remote Sens, 40: 1520-1531. 190490
Martonchik J; Diner D; Pinty B; Verstraete M; Myneni R; Knjazikhin Y; Gordon H. (1998). Determination of
land and ocean reflective, radiative, and biophysical properties using multiangle imaging. IEEE Trans
Geosci Remote Sens, 36: 1266-1281. 190484
Martonchik JD; Diner D; Kahn R; Verstraete M; Pinty B; Gordon H; Ackerman T. (1998). Techniques for the
Retrieval of aerosol properties over land and ocean using multiangle data. IEEE Trans Geosci Remote
Sens, 36: 1212-1227. 190472
Masclet PP; Hoyau V; Jaffrezo JL; Cachier H. (2000). Polycyclic aromatic hydrocarbons deposition on the ice
sheet of Greenland. Part I. Superficial snow. Atmos Environ, 34: 3195-3207. 155966
Massicotte R; Robidoux PY; Sauve S; Flipo D; Fournier M; Trottier B. (2003). Immune response of earthworms
(Lumbricus terrestris, Eisenia andrei and Aporrectodea tuberculata) following in situ soil exposure to
atmospheric deposition from a cement factory. J Environ Monit, 5- 774-779. 155968
Massie S; Torres O; Smith S. (2004). Total ozone mapping spectrometer (TOMS) observations of increases in
Asian aerosol in winter from 1979 to 2000. J Geophys Res, 109: D18211. 190492
Matheson MA Coakley JA; Tahnk WR. (2005). Aerosol and cloud property relationships for summertime
stratiform clouds in the northeastern Atlantic from Advanced Very High Resolution Radiometer
observations. J Geophys Res, 110: D24204. 190494
Matsui T; Masunaga H; Kreidenweis SM; Pielke RA; Tao W"K; Chin M; et al. (2006). Satellite-based assessment
of marine low cloud variability associated with aerosol, atmospheric stability, and the diurnal cycle. J
Geophys Res, 111: D17204. 190498
Matsui T; Pielke R;. (2006). Measurement-based estimation of the spatial gradient of aerosol radiative forcing.
Geophys Res Lett, 33: LI 1813. 190495
Matthias I, et al.. (2004). Multiyear aerosol observations with dual wavelength Raman lidar in the framework of
EARLINET. J Geophys Res, 109, D13203, doi:i0.1029/2004JD004600: . 155971
Mayo JM; Legge AH; Yeung EC; Krupa SV; Bogner JC. (1992). The effects of sulphur gas and elemental sulphur
dust deposition on Pinus contorta x Pinus banksiana: Cell walls and water relations. , 76: 43-50. 155974
McComiskey A; Feingold G. (2008). Quantifying error in the radiative forcing of the first aerosol indirect effect.
Geophys Res Lett, 35: L20810. 190517
McComiskey A; Feingold G; Frisch AS; Turner D; Miller M; Chiu JC; et al. (2008). An assessment of aerosol-
cloud interactions in marine stratus clouds based on surface remote sensing. J Geophys Res, 114:
D09203.190525
McComiskey A; Schwartz SE; Schmid B; Guan H; Lewis ER; Ricchiazzi P; et al. (2008). Direct aerosol forcing:
Calculation from observables and sensitivity to inputs. J Geophys Res, 113: D09202. 190523
McCormick M; Trepte C. (1987). Polar stratospheric optical depth observed between 1978 and 1985. J Geophys
Res, 92: 4297-4306. 192328
McCormick R; Ludwig J. (1967). Climate modification by atmospheric aerosols. , 9: 1358 - 1359. 190528
McDade C. (2006). Summary of IMPROVE Nitrate Measurements. Retrieved 29-JUL-09, from
http://vista.cira.colostate.edu/improve/publications/GrayLit/gray_literature.htm. 192075
Mcfiggans G; Artaxo P; Baltensperger U; Coe H; Facchini MC;Feingold G; et al. (2006). The effect of physical
and chemical aerosol properties on warm cloud droplet activation. , 6: 2593-2649. 190532
McGowan TF; Lipinski GE; Santoleri JJ. (1993). New rules affect the handling of waste fuels. 122-128. 046731
July 2009
9-290
DRAFT-DO NOT CITE OR QUOTE

-------
McMurry PH. (2000). A review of atmospheric aerosol measurements. Atmos Environ, 34- 1959-1999. 081517
Meehl GA; Washington WM; Ammann CM; Arblaster JM; Wigley TML; Tebaldi C. (2004). Combinations of
natural and anthropogenic forcings in twentieth-century climate. J Clim, 17- 3721—3727. 192279
Meers E> Vangronsveld J; Tack FMG; Vandecasteele B; Ruttens A. (2007). Potential of five willow species (Salix
spp.) for phytoextraction of heavy metals. Environ Exp Bot, 60- 57"68. 155977
Melnikov Si Carroll Ji Gorshkov Ai Vlasov Si Dahle S. (2003). Snow and ice concentrations of selected persistent
organic pollutants in the Ob-Yenisey River watershed. Sci Total Environ, 306- 27"37. 156753
Memon ARi Schroder P. (2009). Implications of metal accumulation mechanisms to phytoremediation. Environ
Sci Pollut Res Int, 16: 162-175. 190442
Menon Si Del Genio ADi Kaufman Yi Bennartz Ri Koch Di Loeb Ni Orlikowski D. (2008). Analyzing signatures
of aerosol cloud interactions from satellite retrievals and the GISS GCM to constrain the aerosol indirect
effect. J Geophys Res, 113: D14S22. 190534
Mercado LMi Bellouin Ni Sitch Si Boucher Oi Huntingford Ci Wild Mi Cox PM. (2009). Impact of changes in
diffuse radiation on the global land carbon sink. Nature, 458: 1014-U1087. 190444
Michalsky Ji Schlemmer Ji Berkheiser Wi. (2001). Multiyear measurements of aerosol optical depth in the
Atmospheric Radiation Measurement and Quantitative Links program. J Geophys Res, 106: 12099"
12108. 190537
Middleton WEK. (1952). Vision through the atmosphere. Toronto, ON, Canada: University of Toronto Press.
016324
Mie G. (1908). Beitrage zur Optik triiber Medien, speziell kolloidaler Metallosungen. , 330: 377-445. 155983
Miller RLi Schmidt GAi Shindell DT. (2005). Forced annular variations in the 20th century Intergovernmental
Panel on Climate Change Fourth Assessment Report models. J Geophys Res, 111: D18101. 192258
Millhollen AGi Gustin MSi Obrist D. (2006). Foliar mercury accumulation and exchange for three tree species.
Environ Sci Technol, 40: 6001-6006. 190447
Min Qi Harrison LC. (1996). Cloud properties derived from surface MFRSR measurements and comparison with
GEOS results at the ARM SGP site. Geophys Res Lett, 23: 1641- 1644. 190538
Minnis Pi Harrison EFi Stowe LLi Gibson GGi Denn FMi Doelling DRi Smith WL Jr. (1993). Radiative Climate
Forcing by the Mount Pinatubo Eruption. , 259: 1411-1415. 190539
Mischenko Mi Cairns Bi Kopp G> Schueler CFi Fafaul BAi Hansen JEi et al. (2007). Accurate monitoring of
terrestrial aerosols and total solar irradiance. , 88: 677-691. 190543
Mishchenko ML Geogdzhayev IV. (2007). Satellite remote sensing reveals regional tropospheric aerosol trends. ,
15: 7423-7438. 190545
Mishchenko ML Geogdzhayev IV; Cairns Bi Rossow WBi Lacis AA. (1999). Aerosol retrievals over the ocean by
use of channels 1 and 2 AVHRR data: sensitivity analysis and preliminary results. Appl Opt, 38: 7325"
7341. 190541
Mishchenko ML Geogdzhayev IV; Rossow WBi Cairns Bi Carlson BE; Lacis AA; Liu L; Travis LD. (2007). Long-
term satellite record reveals likely recent aerosol trend. , 315: 1543. 190542
Mitchell JM Jr. (1971). The effect of atmospheric aerosols on climate with special reference to temperature near
the Earth's surface. , 10: 703-714. 190546
Molina LT; Madronich S; Gaffney JS; Singh HB. (2008). Overview of MILAGRO/INTEX-B Campaign.
International Global Atmospheric Chemistry (Newsletter). Seattle,
WA.http ://www .igac.noaa.gov/newsletter/igac38/Apr_2008_IGAC_38 .pdf. 192019
Moody E; King M; Platnick S; Schaaf C; Gao F. (2005). Spatially complete global spectral surface albedos: value-
added datasets derived from Terra MODIS land products. IEEE Trans Geosci Remote Sens, 43: 144-158.
190548
Moosmiiller H; Arnott WP; Rogers CF; Chow JC; Frazier CA; Sherman LE; Dietrich DL. (1998). Photoacoustic
and filter measurements related to aerosol light-absorption during the nothern front range air quality
study (Colorado 1996/1997). J Geophys Res, 103: 28149-28157. 192073
Moropoulou A; Bisbikou K; Torfs K; Van Grieken R; Zezza FJ Macri F. (1998). Origin and growth of weathering
crusts on ancient marbles in industrial atmosphere. Atmos Environ, 32: 967-982. 040485
Moses CA; Smith BJ. (1994). Limestone weathering in the supratidal zone: an example from Mallorca. : . 156785
Mosley-Thompson E; Mashiotta TA; Thompson LG. (2003). High resolution ice core records of Late Holocene
volcanism: Current and future contributions from the Greenland PARCA cores. In Volcanism and the
Earth's Atmosphere (pp. 153—164). Washington, DC: American Geophysical Union. 192255
Mouillot F; Narasimha A; Balkanski Y; Lamarque J-F. (2006). Global carbon emissions from biomass burning in
the 20th century. Geophys Res Lett, 33: L01801. 190549
July 2009
9-291
DRAFT-DO NOT CITE OR QUOTE

-------
Muir DCG; Omelchenko A; Grift NP; Savoie DA; Lockhart WL; Wilkinson P; Brunskill GJ. (1996). Spatial
trends and historical deposition of polychlorinated biphenyls in Canadian midlatitude and arctic lake
sediments. Environ Sci Technol, 30- 3609-3617. 155991
Murayama T, et al. (2001). Ground-based network observation of Asian dust events of April 1998 in East Asia. J
Geophys Res, 106: 18345-18360. 155992
Murdoch JC; Thayer MA. (1988). Hedonic price estimation of variable urban air quality. J Environ Econ
Manage, 15: 143-146. 156788
Malkonen Ei Derome J; Fritze H; Helmisaari H"S; Kukkola Mi Kyto Mi Saarsalmi Ai Salemaa M. (1999).
Compensatory fertilization of Scots pine stands polluted by heavy metals. , 55: 239—268. 155961
Naidoo Gi Chirkoot D. (2004). The effects of coal dust on photosynthetic performance of the mangrove,
Avicennia marina in Richards Bay, South Africa. , 127: 359-366. 190449
Nakajima Ti Higurashi Ai Kawamoto Ki Penner JE. (2001). A possible correlation between satellite-derived
cloud and aerosol microphysical parameters. Geophys Res Lett, 28: 1171-1174. 190552
Nakamura Ki Nakawo Mi Ageta Yi Goto-Azuma Ki Kamiyam K. (2000). Post-depositional loss of nitrate in
surface snow layers of the Antarctic ice sheet. , 17: 11-16. 156792
Nash TH III. (1975). Influence of effluents from a zinc factory on lichens. Ecol Monogr, 45: 183-198. 016763
National Research Council. (2001). Climate change science: an analysis of some key questions. 053303
National Research Council. (2005). Radiative forcing of climate change: expanding the concept and addressing
uncertainties. 057409
Nazaroff WWi Cass GR. (1991). Protecting museum collections from soiling due to the deposition of airborne
particles. Atmos Environ, 25: 841-852. 044577
NCDENR. (2007). Technical Analyses Supporting Regional Haze State Implementation Plan	 156798
NESCAUM. (0). Regional Aerosol Intensive Network (RAIN) Preliminary Data Analysis	 156802
Nguyen HL; Leermakers MJ Elskens Mi De Ridder Fi Doan THi Baeyens W. (2005). Correlations, partitioning
and bioaccumulation of heavy metals between different compartments of Lake Balaton. Sci Total
Environ, 341: 211-226. 155997
Nicholson Li Benn DI. (2006). Calculating ice melt beneath a debris layer using meteorological data. , 52: 463"
470. 156806
Nieminen Ti Derome Ji Helmisaari H"S. (1999). Interactions between precipitation and Scots pine canopies
along a heavy-metal pollution gradient. Environ Pollut, 106: 129—137. 155998
Nogueira MAi Magalhaes GCi Cardoso EJBN. (2004). Manganese toxicity in mycorrhizal and phosphorus-
fertilized soybean plants. , 27: 141-156. 190460
Norris Ji Wild M. (2007). Trends in aerosol radiative effects over Europe inferred from observed cloud cover,
solar "dimming", and solar "brightening". J Geophys Res, 112: D08214. 190555
Notten MJMi Oosthoek AJPi Rozema Ji Aerts R. (2005). Heavy metal concentrations in a soil-plant-snail food
chain along a terrestrial soil pollution gradient. Environ Pollut, 138: 178-190. 190461
Novakov Ti Ramanathan Vi Hansen JEi Kirchstetter TWi Sato Mi Sinton JEi Sathaye JA. (2003). Large
historical changes of fossil-fuel black carbon aerosols. Geophys Res Lett, 30. 048398
O'Dowd Ci McFiggans Gi Creasey DJi Pirjola Li Hoell Ci Smith MHi Allan BJi Plane JMCi Heard DEi Lee JDi
Pilling MJi Kulmala M. (1999). On the photochemical production of new particles in the coastal boundary
layer. Geophys Res Lett, 26: 1707-1710. 090414
O'Neill NT; Eck TF; Smirnov A; Holben BN; Thulasiraman S. (2003). Spectral discrimination of coarse and fine
mode optical depth. J Geophys Res, 108: 4559. 180187
Odum EP. (1985). Trends expected in stressed ecosystems. , 35: 419-422. 039482
Odum EP. (1993). Major ecosystem types of the world. In Odum, E. P. (Ed.),Ecology and our endangered life-
support systemsSunderland, MA: Sinauer Associates, Inc. 076742
Oehme M; Biseth A; Schlabach M; Wiig O. (1995). Concentrations of polychlorinated dibenzo-p-dioxins,
dibenzofurans and non-ortho substituted biphenyls in polar bear milk from Svalbard (Norway). Environ
Pollut, 90: 401-407. 011267
Oehme M; Haugen J-E; Schlabach M. (1996). Seasonal changes and relation between levels of organochlorines
in Arctic ambient air: First results from an all-round-year monitoring program at Ny-Alesund, Norway.
Environ Sci Technol, 30: 2294-2304. 156001
Ohyama K; Angermann J; Dunlap DY; Matsumura F. (2004). Distribution of polychlorinated biphenyls and
chlorinated pesticide residues in trout in the Sierra Nevada. J Environ Qual, 33: 1752-1764. 190462
Oliveira A; Pampulha ME. (2006). Effects of long-term heavy metal contamination on soil microbial
characteristics. , 102: 157-161. 156827
Ormrod DP. (1984). Impact of trace element pollution on plants. In Treshow, M. (Ed.),Air pollution and plant
lifeChichester, United Kingdom: John Wiley & Sons Ltd. 046892
July 2009
9-292
DRAFT-DO NOT CITE OR QUOTE

-------
Otnyukova T. (2007). Epiphytic lichen growth abnormalities and element concentrations as early indicators of
forest decline. , 146: 359-365. 156009
Ouimette JR.; Flagan RC. (1982). The extinction coefficient of multicomponent aerosols. Atmos Environ, 16:
2405-2420. 025047
Ozdilek HG; Mathisen PP; Pellegrino D. (2007). Distribution of heavy metals in vegetation surrounding the
Blackstone River, USA: Considerations regarding sediment contamination and long term metals
transport in freshwater riverine ecosystems. J Environ Biol, 28: 493-502. 156010
Pacheco AMG; Freitas MC. (2004). Are lower epiphytes really that better than higher plants for indicating
airborne contaminants? An insight into the elemental contents of lichen thalli and tree bark by INAA.
Journal of Radioanal Chem, 259: 27"33. 156011
Padmavathiamma PK; Li LY. (2007). Phytoremediation technology: Hyper-accumulation metals in plants. , 184:
105-126. 190465
Palmer AS; Morgan VI; Curran MAJ; van Ommen TD; Mayewski PA. (2002). Antarctic volcanic flux ratios from
Law Dome ice cores. Ann Glaciol, 35: 329-332. 192319
Park RJ; Jacob DJ; Kumar N> Yantosca RM. (2006). Regional visibility statistics in the United States: Natural
and transboundary pollution influences, and implications for the regional haze rule. Atmos Environ, 40:
5405-5423. 190469
Parrish ZD; White JC; Isleyen M; Gent MPN; Iannucci-Berger W; Eitzer BD; Kelsey JW; Mattina MI. (2006).
Accumulation of weathered polycyclic aromatic hydrocarbons (PAHs) by plant and earthworm species.
Chemosphere, 64: 609-618. 156014
Patadia F; Gupta P; Christopher SA. (2008). First observational estimates of global clear-sky shortwave aerosol
direct radiative effect over land. J Geophys Res, 35: L04810. 190558
Patra M; Bhowmik N> Bandopadhyay B; Sharma A. (2004). Comparison of mercury, lead and arsenic with
respect to genotoxic effects on plant systems and the development of genetic tolerance. Environ Exp Bot,
52: 199-223. 081976
Pennanen T; Frostegard A; Fritze H; Baath E. (1996). Phospholipid fatty acid composition and heavy metal
tolerance of soil microbial communities along two heavy metal-polluted gradients in coniferous forests.
Appl Environ Microbiol, 62: 420-428. 156016
Penner J; Quaas J; Storelvmo T; Takemura T; Boucher O; Guo H; et al. (2006). Model intercomparison of
indirect aerosol effects. , 6: 3391-3405. 190564
Penner JE. (2003). Comment on "Control of fossil-fuel particulate black carbon and organic matter, possibly the
most effective method of slowing global warming. J Geophys Res, 108: 4771. 156851
Penner JE; Dickinson RE; O'Neill CA. (1992). Effects of aerosol from biomass burning on the global radiation
budget. , 256: 1432-1433. 045825
Penner JE; Eddleman H; Novakov T. (1993). Towards the development of a global inventory for black carbon
emissions. Atmos Environ, 27: 1277-1295. 045457
Penner JE; Zhang SY; Chin M; Chuang CC; Feichter J; Geogdzhayev IV; et al. (2002). A comparison of model"
and satellite derived aerosol optical depth and reflectivity. J Trace Elem Med Biol, 59: 441-460. 190562
Pennisi E. (2004). The secret life of fungi. , 304: 1620-1622. 156018
Perlwitz J; Graf HF. (2001). Trosposphere-stratosphere dynamic coupling under strong and weak polar vortex
conditions . Geophys Res Lett, 28: 271-274. 192271
Perlwitz J; Harnik N. (2003). Observational evidence of a stratospheric influence on the troposphere by
planetary wave reflection. J Clim, 16: 3011-3026. 192264
Peters AJ; Gregor DJ; Teixeira CF; Jones NP; Spencer C. (1995). The recent depositional trend of polycyclic
aromatic hydrocarbons and elemental carbon to the Agassiz Ice cap, Ellesmere Island, Canada. Sci Total
Environ, 160/161: 167-179. 156856
Pfirman SL; Kogeler JW; Rigor I. (1997). Potential for rapid transport of contaminants from the Kara Sea. Sci
Total Environ, 202: 111-122. 156864
Pincus R; Klein SA. (2000). Unresolved spatial variability and microphysical process rates in large-scale models.
J Geophys Res, 105: 27059-27065. 190565
Pinker R; Zhang B; Dutton E. (2005). Do satellites detect trends in surface solar radiation?. , 308: 850-854.
190569
Piol MN; Lopez AG; Mino LA; Dos Santos Afonso M; Verrengia Guerrero NR. (2006). The impact of particle-
bound cadmium on bioavailability and bioaccumulation: A pragmatic approach. Environ Sci Technol, 40:
6341-6347. 156028
Pitchford M; Green M; Kuhns H; Tombach I; Malm W', Scruggs M; Farber R; Mirabella V; White WH; McDade
C. (1999). Project MOHAVE Final Report. 156873
July 2009
9-293
DRAFT-DO NOT CITE OR QUOTE

-------
Pitchford M> Maim W; Schichtel B; Kumar Ni Lowenthal D; Hand J. (2007). Revised algorithm for estimating
light extinction from IMPROVE particle speciation data. J Air Waste Manag Assoc, 57' 1326-36. 098066
Pitchford ML; Green MC; Morris R; Emery C; Sakata R; Swab C', Mairose PT. (2008). Columbia River Gorge Air
Quality Study- Science Summary Report. Southwest Clean Air Agency. Vancouver, WA. 180159
Pitchford ML; Malm WC. (1994). Development and applications of a standard visual index. Atmos Environ, 28-
1049-1054. 044922
Pitchford ML; Polkowsky BV; McGown MR; Malm WC; Molenar JV; Mauch L. (1990). Percent change in
extinction coefficient- A proposed approach for federal visibility protection strategy. • . 156871
Pitchford ML; Schichtel BA; Gebhart KA; Barna MG; Malm WC; Tombach IH; Knipping EM. (2005).
Reconciliation and Interpretation of the Big Bend National Park Light Extinction Source Apportionment-
Results from the Big Bend Regional Aerosol and Visibility Observational Study- Part II. J Air Waste
Manag Assoc, 55: 1726-1732. 156874
Prasad MNV; De Oliveira Freitas HM. (2003). Metal hyperaccumulation in plants - Biodiversity prospecting for.
Electron J Biotechnol, 6: 110-146. 156885
Prata A; Rose W', Self S; O'Brien D. (2003). Global, long-term sulphur dioxide measurements from TOVS data: a
new tool for studying explosive volcanism and climate. In Volcanism and the Earth's Atmosphere (pp.
75-92). Washington, DC: American Geophysical Union. 192235
Preston CM; Schmidt MWI. (2006). Black (pyrogenic) carbon: a synthesis of current knowledge and
uncertainties with special consideration of boreal regions. , 3: 397-420. 156030
Procopio AS; Artaxo P; Kaufman YJ; Remer LA; Schafer JA; Holben BN. (2004). Multiyear analysis of
Amazonian biomass burning smoke radiative forcing of climate. J Geophys Res, 31: L03108. 190571
Prospero JM. (1996). Saharan dust transport over the North Atlantic Ocean and Mediterranean: an overview. In
Guerzoni S; Chester R (Ed.),The Impact of Desert Dust Across the Mediterranean (pp. 133—152). The
Netherlands: Kluwer Academic Publishers. 156889
Pryor SC. (1996). Assessing public perception of visibility for standard setting exercises. Atmos Environ, 30:
2705-2716. 056598
Pryor SC; Binkowski FS. (2004). An analysis of the time scales associated with aerosol processes during dry
deposition. Aerosol Sci Technol, 38: 1091-1098. 116805
Putaud J-P; Raes F; Van Dengenen R; Bruggemann E; Facchini M"C; Decesari S; Fuzzi S; Gehrig R; Huglin C;
Laj P; Lorbeer G; Maenhaut W', Mihalopoulos N; Muller K; Querol X; Rodriguez S; Schneider J; Spindler
G; ten Brink H; Torseth K; Wiedensohler A. (2004). A European aerosol phenomenology--2' chemical
characteristics of particulate matter at kerbside, urban, rural and background sites in Europe. Atmos
Environ, 38: 2579-2595. 055545
Qian Y; Wang W; Leung L; Kaiser D. (2007). Variability of solar radiation under cloud-free skies in China: The
role of aerosols. Geophys Res Lett, 34: L12804. 190572
Quaas J; Boucher O. (2005). Constraining the first aerosol indirect radiative forcing in the LMDZ GCM using
POLDER and MODIS satellite data. Geophys Res Lett, 32: L17814. 190573
Quaas J; Boucher O; Bellouin N; Kinne S. (2008). Satellite- based estimate of the direct and indirect aerosol
climate forcing. J Geophys Res, 113: D05204. 190916
Quaas J; Boucher O; Lohmann U. (2006). Constraining the total aerosol indirect effect in the LMDZ GCM and
ECHAM4 GCMs using MODIS satellite data. , 5: 9669-9690. 190915
Quinn PK; Anderson T; Bates T; Dlugi R; Heintzenberg J; Von Hoyningen-Huene W; Kumula M; Russel P;
Swietlicki E. (1996). Closure in tropospheric aerosol-climate research : A review and future needs for
addressing aerosol direct shortwave radiative forcing. , 69: 547-577. 192021
Quinn PK; Bates TS. (2003). North American, Asian, and Indian haze: similar regional impacts on climate?.
Geophys Res Lett, 30. 049189
Quinn PK; Bates TS; Baynard T; Clarke AD; Onasch TB; Wang W', Rood MJ; Andrews E; Allan J; Carrico CM;
Coffman D; Worsnop D. (2005). Impact of particulate organic matter on the relative humidity dependence
of light scattering: A simplified parameterization. Geophys Res Lett, 32: 1-4. 156033
Quinn PK; Coffman D; Kapustin V; Bates TS; Covert DS. (1998). Aerosol optical properties in the marine
boundary layer during ACE 1 and the underlying chemical and physical aerosol properties. J Geophys
Res, 103: 16547-16563. 190918
Quinn PK; Coffman DJ; Bates TS; Welton EJ; Covert DS; Miller TL; et al. (2004). Aerosol optical properties
measured aboard the Ronald H. Brown during ACE "Asia as a function of aerosol chemical composition
and source region. J Geophys Res, 109: 109. 190937
Radke LF; Coakley JA; King MD. (1989). Direct and remote sensing observations of the effects of ships on
clouds. , 246: 1146-1149. 156034
July 2009
9-294
DRAFT-DO NOT CITE OR QUOTE

-------
Rajapaksha R; MT"K; Baath E. (2004). Metal toxicity affects fungal and bacterial activities in soil differently.
Appl Environ Microbiol, 5- 2966 — 2973. 156035
Ramachandran S> Ramaswamy V; Stenchikov GL; Robock A. (2000). Radiative impact of the Mount Pinatubo
volcanic eruption- Lower stratospheric response. J Geophys Res, 105- 24409-24429. 192050
Ramanathan V; Crutzen P. (2003). Atmospheric Brown "Clouds". Atmos Environ, 37- 4033-4035. 190198
Ramanathan V; Crutzen P; Lelieveld J; et al. (2001). Indian Ocean Experiment- An integrated analysis of the
climate forcing and effects of the great Indo-Asian haze. J Geophys Res, 106- 28371-38398. 190196
Ramanathan V; Crutzen PJ; Kiehl JT; Rosenfeld D. (2001). Aerosols, climate and the hydrological cycle. , 294-
2119-2124. 042681
Ramanathan V; et al. (2005). Atmospheric brown clouds' Impact on South Asian climate and hydrologic cycle. ,
102: 5326-5333. 190199
Ramaswamy V; Boucher Oi Haigh J; Hauglustaine D; Haywood J; Myhre Gi Nakajima Ti Shi G> Solomon Si
Betts RE. (2001). Radiative Forcing of Climate Change: PNNL"SA"39648. New York: Cambridge
University Press. 156899
Ramaswamy Vi Ramachandran Si Stenchikov Gi Robock A. (2006). A model study of the effect of Pinatubo
volcanic aerosols on the stratospheric temperatures+. In Kiehl JTi Ramanathan V (Ed.),Frontiers of
climate modeling (pp. 152—178). Cambridge, UK: Cambridge University Press. 192273
Ramaswamy Vi Schwarzkopf MDi Randel WJi Santer BDi Soden BJi Stenchikov GL. (2006). Anthropogenic and
natural influences in the evolution of lower stratospheric cooling. , 311: 1138-1141. 192284
Ramos Li Hernandez LMi Gonzalez MJ. (1994). Sequential fractionation of copper, lead, cadmium and zinc in
soils from or near Donana National Park. J Environ Qual, 23: 50-57. 046736
Ramsey JM. (1966). Concentrations of carbon monoxide at traffic intersections in Dayton, Ohio. Arch Environ
Occup Health, 13: 44-46. 013946
Randall CEi Bevilacqua RMi Lumpe JDi Hoppel KW. (2001). Validation of POAM III aerosols: Comparison to
SAGE II and HALOE. J Geophys Res, 106: 27525-27536. 192268
Randall Di Khairoutdinov Mi Arakawa Ai Grabowski W. (2003). Breaking the cloud parameterization deadlock.
, 84: 1547-1564. 190201
Randerson JTi Liu Hi Flanner MGi Chambers SDi Jin Yi Hess PGi Pfister Gi Mack MCi Treseder KKi Welp LRi
Chapin FSi Harden JWi Goulden MLi Lyons Ei Neff JCi Schuur EAGi Zender CS. (2006). The impact of
boreal forest fire on climate warming. , 314: 1130-1132. 156038
Rapport DJi Whitford WG. (1999). How ecosystems respond to stress: common properties of arid and aquatic
systems. , 49: 193-203. 004595
Ray Pi Reddy UG; Lapeyrie Fi Adholeya A. (2005). Effect of coal ash on growth and metal uptake by some
selected ectomycorrhizal fungi in vitro. , T- 199-216. 190473
Read WG; Froidevaux L; Waters JW. (1993). Microwave limb sounder measurement of stratospheric S02 from
the Mt. Pinatubo Volcano. Geophys Res Lett, 20: 1299-1302. 192031
Reddy Mi Boucher Oi Balkanski Yi Schulz M. (2005). Aerosol optical depths and direct radiative perturbations
by species and source type. Geophys Res Lett, 32: L12803. 190208
Reddy Mi Boucher Oi Bellouin Ni Schulz Mi Balkanski Yi Dufresne Ji Pham M. (2005). Estimates of multi-
component aerosol optical depth and direct radiative perturbation in the LMDZT general circulation
model. J Geophys Res, 110: D10S16. 190207
Reddy MSi Boucher O. (2007). Climate impact of black carbon emitted from energy consumption in the world's
regions. Geophys Res Lett, 34: 1802. 156042
Regoli Fi Gorbi Si Fattorini Di Tedesco Si Notti Ai Machella Ni Bocchetti R; Benedetti Mi Piva F. (2006). Use of
the land snail Helix aspersa sentinel organism for monitoring ecotoxicologic effects of urban pollution: An
integrated approach. , 114: 63"69. 156046
Reid J. (2008). An overview of UAE2 flight operations: Observations of summertime atmospheric
thermodynamic and aerosol profiles of the southern Arabian Gulf. J Geophys Res, 113: D14213. 190214
Reid Ji Kinney J; Wesphal D; et al. (2003). Analysis of measurements of Saharan dust by airborne and ground-
based remote sensing methods during the Puerto Rico Dust Experiment (PRIDE). J Geophys Res, 108:
8586. 190213
Reinfelder JR; Fisher NS; Luoma SN; Nichols JW; Wang W"X. (1998). Trace element trophic transfer in aquatic
organisms: A critique of the kinetic model approach. Sci Total Environ, 219: 117-135. 156047
Reisinger LM. (1990). Analysis of airborne particles sampled in the southern Appalachian Mountains. Water Air
Soil Pollut, 50: 149-162. 046737
Remer L; et al. (2008). An emerging aerosol climatology from the MODIS satellite sensors. J Geophys Res, 113:
D14S01.190224
July 2009
9-295
DRAFT-DO NOT CITE OR QUOTE

-------
Remer L; Gassoi Hegg D; Kaufman Y> Holben B. (1997). Urban/industrial aerosol- ground based sun/sky
radiometer and airborne in situ measurements. J Geophys Res, 102- 16849-16859. 190216
Remer L; Kaufman Y. (2006). Aerosol direct radiative effect at the top of the atmosphere over cloud free ocean
derived from four years of MODIS data. , 6: 237-253. 190222
Remer L; Kaufman Y> Tanre D; Mattoo Si Chu Di Martins Ji Li Ri Ichoku Ci Levy Ri Kleidman Ri Eck Ti
Vermote Ei Holben B. (2005). The MODIS aerosol algorithm, products and validation. J Trace Elem Med
Biol, 62: 947-973. 190221
Remer Li Tanre Di Kaufman Yi Ichoku Ci Mattoo Si Levy Ri Chu Di Holben Bi Dubovik Oi Smirnov Ai Martins
Ji Li Ri Ahman Z. (2002). Validation of MODIS aerosol retrieval over ocean. Geophys Res Lett, 29: 8008.
190218
Richards LW; Dye TS; Arthur M; Byars MS. (1996). Analysis of ASOS Data for Visibility Purposes, Final Report
. Systems Applications International, Inc.. San Rafael, CA. STI-996231-1610-FR. 190476
Rind D; Perlwitz J; Lonergan P. (2005). AO/NAO response to climate change: 1. Respective influences of
stratospheric and tropospheric climate changes. J Geophys Res, 110: D12107. 192261
Rissler J; Swietlicki E; Zhou J; Roberts G; Andreae MO; Gatti LV; Artaxo P. (2004). Physical properties of the
sub-micrometer aerosol over the Amazon rain forest during the wet-todry season transition—comparison
of modeled and measured CCN concentrations. , 4- 2119-2143. 190225
Roberts DL; Jones A. (2004). Climate sensitivity to black carbon aerosol from fossil fuel combustion. J Geophys
Res, 109: 6202. 156052
Roberts ES; Richards JH; Jaskot R; Dreher KL. (2003). Oxidative stress mediates air pollution particle-induced
acute lung injury and molecular pathology. Inhal Toxicol, 15: 1327-1346. 156051
Robinson A; Donahue N; Shrivastava M; Weitkamp E; Sage A; Grieshop A; Lane T; Pierce J; Pandis S. (2007).
Rethinking organic aerosols: Semivolatile emissions and photochemical aging. , 315: 1259. 191975
Roderick ML; Farquhar GD. (2002). The cause of decreased pan evaporation over the past 50 years. , 298: 1410"
1411. 042788
Roeckner E; Stier P; Feichter J; KLoster S; Esch M; Fischer-Bruns L. (2006). Impact of carbonaceous aerosol
emissions on regional climate change. , 27: 553-571. 156920
Romic M; Romic D. (2003). Heavy metals distribution in agricultural topsoils in urban area. , 43: 795-805.
156055
Rood MJ; Covert DS; Larson TV. (1987). Hygroscopic properties of atmospheric aerosol in Riverside, California.
Tellus B Chem Phys Meteorol, 39B: 383-397. 046397
Rosenfeld D. (2000). Suppression of rain and snow by urban and industrial air pollution. , 287: 1793-1796.
002234
Rosenfeld D. (2006). Aerosols, clouds, and climate. , 312: 1. 190233
Rosenfeld D; Dai J; Yu X; Yao Z; Xu X; Yang X; Du C. (2007). Inverse relations between amounts of air pollution
and orographic precipitation. , 315: 1396-1398. 156057
Rosenfeld D; Givati A. (2006). Evidence of orographic precipitation suppression by air pollution-induced aerosols
in the western United States. , 45: 893-911. 156924
Rosenfeld D; Lansky I. (1998). Satellite-based insights into precipitation formation processes in continental and
maritime convective clouds. , 79: 2457-2476. 190230
Ross DM; Malm WC; Loomis RJ. (1985). The psychological valuation of good visual air quality by national park
visitors. Presented at Presented at: 78th annual meeting of the Air Pollution Control Association; June;
Detroit, MI. Pittsburgh, PA: Air Pollution Control Association; paper no. 85-10.4. 044287
Ross DM; Malm WC; Loomis RJ. (1987). An examination of the relative importance of park attributes at several
national parks. In Bhardwaja, P. S. (Ed.),Visibility protection: research and policy aspects, an APCA
international specialty conferenceGrand Teton National Park, WY; Pittsburgh, PA: Air Pollution Control
Association. 037420
Rosselli W; Keller C; Boschi K. (2003). Phytoextraction capacity of trees growing on a metal contaminated soil.
Plant Soil, 256: 265-272. 156058
Roth CM; Goss K-U; Schwarzenbach RP. (2004). Sorption of diverse organic vapors to snow. Environ Sci
Technol, 38: 4078-4084. 056431
Rotton J; Barry T; Milligan M; Fitzpatrick M. (1979). The air pollution experience and interpersonal aggression.
, 9: 397-412. 190478
Rotton J; Frey J. (1982). Atmospheric conditions, seasonal trends, and psychiatric emergencies. In Replications
and Extensions (pp. Unknown). Washington, DC: American Psychological Association. 190477
Rozanov EV; Schlesinger ME; Andronova NG; Yang F; Malyshev SL; Zubov VA; Egorova TA; Li B. (2002).
Climate/chemistry effects of the Pinatubo volcanic eruption simulated by the UIUC
stratosphere/troposphere GCM with interactive photochemistry. J Geophys Res, 107: 4594. 192247
July 2009
9-296
DRAFT-DO NOT CITE OR QUOTE

-------
Rozanov EV; Schlesinger ME; Egorova TA; Li B; Andronova N> Zubov VA. (2004). Atmospheric response to the
observed increase of solar UV radiation from solar minimum to solar maximum simulated by the
University of Illinois at Urbana-Champaign climate-chemistry model . J Geophys Res, 109- D01110.
192245
Ruckstuhl C> et al. (2008). Aerosol and cloud effects on solar brightening and recent rapid warming. Geophys
Res Lett, 35: L12708. 190356
Russell P; Kinne S> Bergstrom R. (1997). Aerosol climate effects: local radiative forcing and column closure
experiments. J Geophys Res, 102: 9397-9407. 190359
Russell P; Livingston J; Hignett P; Kinne S> Wong J; Chien A; Bergstrom R; Durkee P; Hobbs P. (1999). Aerosol-
induced radiative flux changes off the United States mid"Atlantic coast: comparison of values calculated
from sun photometer and in situ data with those measured by airborne pyranometer. J Geophys Res,
104: 2289-2307. 190363
Rusu A-M; Jones GC; Chimonides PDJ; Purvis OW. (2006). Biomonitoring using the lichen Hypogymnia
physodes and bark samples near Zlatna, Romania immediately following closure of a copper ore-
processing plant. Environ Pollut, 143: 81-88. 156062
Ryan PA; Lowenthal D; Kumar N. (2005). Improved light extinction reconstruction in interagency monitoring of
protected visual environments. J Air Waste Manag Assoc, 55: 1751-1759. 156934
Sabbioni C; Ghedini N; Bonazza A. (2003). Organic anions in damage layers on monuments and buildings.
Atmos Environ, 37: 1261-1269. 049282
Sabin LD; Lim JH; Stolzenbach KD; Schiff KC. (2005). Contribution of trace metals from atmospheric deposition
to stormwater runoff in a small impervious urban catchment. Water Res, 39: 3929-3937. 088300
Salemaa M; Derome J; Helmisaari HS; Nieminen T; Vanha-Majamaa I. (2004). Element accumulation in boreal
bryophytes, lichens and vascular plants exposed to heavy metal and sulfur deposition in Finland. Sci
Total Environ, 324: 141-160. 156069
Samuels HJ; Twiss SJ Wong EW. (1973). Visibility, light scattering and mass concentration of particulate
matter: report of the California Tri-City Aerosol Sampling Project. 070601
Sanderson EG; Farant JP. (2004). Indoor and outdoor polycyclic aromatic hydrocarbons in residences
surrounding a Soderberg aluminum smelter in Canada. Environ Sci Technol, 38: 5350-5356. 156942
Sato M; Hansen J; Koch D; Lucis A; Ruedy R; Dubovik O; Holben B; Chin M; Novakov T. (2003). Global
atmospheric black carbon inferred from AAEONET. Presented at Proceedings of the National Academy of
Science. 156947
Sato M; Hansen JE; McCormick MP; Pollack JB. (1993). Stratospheric aerosol optical depths, 1850-1990. J
Geophys Res, 98: 22987-22994. 192046
Sauve S. (2001). Speciation of metals in soils. Presented at , Pensacola, FL. 156948
Saxena P; Hildemann LM; McMurry PH; Seinfeld JH. (1995). Organics alter hygroscopic behavior of
atmospheric particles. J Geophys Res, 100: 18,755-18,770. 077273
Schichtel BA; Gebhart K; Barna MG; Malm WC; Green MC. (2004). Big Bend Regional Aerosol and Visibility
Observational (BRAVO) Study Results: Air Quality Data and Source Attribution Analyses Results from
the National Park Service	 179902
Schichtel BA; Gebhart KA; Malm WC; Barna MG; Pitchford ML; Knipping EM; Tombach IH. (2005).
Reconciliation and interpretation of big bend national park particulate sulfur source apportionment:
Results from the big bend regional aerosol and visibility observational study-part I. J Air Waste Manag
Assoc, 55: 1709-1725. 156957
Schichtel BA; Malm WC; Bench G; Fallon S; McDade CE; Chow JC; Watson JG. (2008). Fossil and
contemporary fine particulate carbon fractions at 12 rural and urban sites in the United States. J
Geophys Res, 113: . 156958
Schiff K; Bay S; Stransky C. (2002). Characterization of stormwater toxicants from an urban watershed to
freshwater and marine organisms. , 4- 215-227. 156959
Schilling JS; Lehman ME. (2002). Bioindication of atmospheric heavy metal deposition in the Southeastern US
using the moss Thuidium delicatulum. Atmos Environ, 36: 1611-1618. 113075
Schmid B; et al. (2000). Clearsky closure studies of lower tropospheric aerosol and water vapor during ACE"2
using airborne sunphotometer, airborne in situ, space-borne, and ground-based measurements. , 52: 568"
593. 190369
Schmid B; Ferrare R; Flynn C; et al. (2006). How well do stateof- the-art techniques measuring the vertical
profile of tropospheric aerosol extinction compare?. J Geophys Res, 111: 0. 190372
Schmidt GA; et al. (2006). Present-day atmospheric simulations using GISS Model E: Comparison to in situ,
satellite and reanalysis data. J Clim, 19: 153-192. 190373
July 2009
9-297
DRAFT-DO NOT CITE OR QUOTE

-------
Schoeberl MR; Douglass AR; Zhu Z; Pawson S. (2003). A comparison of the lower stratospheric age spectra
derived from a general circulation model and two data assimilation systems. J Geophys Res, 108D3.
057262
Schroeder WHJ Munthe J. (1998). Atmospheric mercury -- an overview. Atmos Environ, 32- 809-822. 014559
Schulz Mi Textor Ci Kinne Si et al. (2006). Radiative forcing by aerosols as derived from the AeroCom present-
day and preindustrial simulations. , 6- 5225-5246. 190381
Schwartz Ji Zanobetti Ai Bateson TFi Schwartz Ji Sarnat JAi Coull BAi Wilson WE. (2007). Effects of exposure
measurement error on particle matter epidemiology- a simulation using data from a panel study in
Baltimore, MD. , 17: S2. 188938
Seifert Mi Anke Si Holzinger Si Jaritz Mi Arnhold Wi Anke M. (1999). Cadmium and strontium content of mice,
shrews, and some invertebrates. , 17: 357-365. 190480
Seigneur Ci Johnson CDi Latimer DAi Bergstrom RWi Hogo H. (1984). Users manual for the Plume Visibility
Model (PLUVUE II). Final report 23 Feb-29 Aug 83	 156965
Seinfeld JHi Carmichael GRi Arimoto Ri et al. (2004). ACEAsia: Regional climatic and atmospheric chemical
effects of Asian dust and pollution. , 85: 367-380. 190388
Sekiguchi Mi Nakajima Ti Suzuki Ki et al. (2003). A study of the direct and indirect effects of aerosols using
global satellite data sets of aerosol and cloud parameters. J Geophys Res, 108 D22: 4699. 190385
Shakya Ki Chettri MKi Sawidis T. (2008). Impact of heavy metals (copper, zinc, and lead) on the chlorophyll
content of some mosses. Arch Environ Contam Toxicol, 54: 412-421. 156081
Shan XQi Wang ZW; Wang WS; Zhang SZ; Wen B. (2003). Labile rhizosphere soil solution fraction for prediction
of bioavailability of heavy metals and rare earth elements to plants. Anal Bioanal Chem, 375: 400-407.
156972
Sharifi MR; Gibson AC; Rundel PW. (1999). Phenological and physiological responses of heavily dusted creosote
bush (Larrea tridentata) to summer irrigation in the Mojave Desert. , 14: 369—378. 156082
Sheesley RJ; Schauer JJ; Hemming JD; Barman MA; Geis SW; Tortorelli JJ. (2004). Toxicity of ambient
atmospheric particulate matter from the Lake Michigan (USA) airshed to aquatic organisms. Environ
Toxicol Chem, 23: 133-140. 156084
Shindell DT; Chin M; Dentener F; et al. (2008). A multi-model assessment of pollution transport to the Arctic. ,
8: 5353-5372. 190391
Shindell DT; Levy H; Schwarzkopf II MD; Horowitz LW; Lamarque JF; Faluvegi G. (2008). Multimodel
projections of climate change from short-lived emissions due to human activities. J Geophys Res, 113:
D11109.190393
Shindell DT; Schmidt GA; Mann ME; Faluvegi G. (2004). Dynamic winter climate response to large tropical
volcanic eruptions since 1600. J Geophys Res, 109: D05104. 192267
Shindell DT; Schmidt GA; Miller RL; Mann M. (2003). Volcanic and solar forcing of climate change during the
preindustrial era. J Clim, 16: 4094—4107. 192069
Shulz KG; Zondervan I; Gerringa LJA; Timmermans KR; Veldhuls MJW; Riebesell U. (2004). Effect of trace
metal availability on coccolithophorid calcification. , 430: 673-676. 156087
Singh HB; Brune WH; Crawford JH; Flocke F; Jacob DJ. (2008). Chemistry and Transport of Pollution over the
Gulf of Mexico and the Pacific: Spring 2006 INTEX-B Campaign Overview and First Results. , 0: 0.
190394
Sinyuk A; et al. (2007). Simultaneous retrieval of aerosol and surface properties from a combination of
AERONET and satellite data. , 107: 90-108. 190395
Skov H; Christensen JH; Goodsite ME; Heidam NZ; Jensen B; Wahlin P; Geernaert G. (2004). Fate of elemental
mercury in the Arctic during atmospheric mercury depletion episodes and the load of atmospheric
mercury to the Arctic. Environ Sci Technol, 38: 2373-2382. 190481
Sloane CS. (1983). Optical properties of aerosols--comparison of measurements with model calculations. Atmos
Environ, 17: 409-416. 025039
Sloane CS. (1984). Optical properties of aerosols of mixed composition. Atmos Environ, 18: 871-878. 025040
Sloane CS. (1986). Effect of composition on aerosol light scattering efficiencies. Atmos Environ, 20: 1025-1037.
045954
Sloane CS; Wolff GT. (1985). Prediction of ambient light scattering using a physical model responsive to relative
humidity: validation with measurements from Detroit. Atmos Environ, 19: 669-680. 045953
Smejkalova M; Mikanova O; Boruvka L. (2003). Effects of heavy metal concentrations on biological activity of
soil microorganisms. , 49: 321?326. 156987
Smirnov A; Holben B; Eck T; Dubovik O; Slutsker I. (2000). Cloud screening and quality control algorithms for
the AERONET database. , 73: 337-349. 190397
July 2009
9-298
DRAFT-DO NOT CITE OR QUOTE

-------
Smirnov A; Holben B; Eck T> Slutsker L Chatenet B; Pinker R. (2002). Diurnal variability of aerosol optical
depth observed at AERONET (Aerosol Robotic Network) sites. Geophys Res Lett, 29: 2115. 190398
Smirnov A; Holben B; Sakerin Si et al. (2006). Shipbased aerosol optical depth measurements in the Atlantic
Ocean, comparison with satellite retrievals and GOCART model. Geophys Res Lett, 33: L14817. 190400
Smith Jr WLi et al. (2005). EOS Terra aerosol and radiative flux validation: An overview of the Chesapeake
Lighthouse and aircraft measurements from satellites (CLAMS) experiment. J Trace Elem Med Biol, 62:
903-918. 190401
Smith WH. (1990). Forest nutrient cycling: toxic ions. In Air pollution and forests: interactions between air
contaminants and forest ecosystemsNew York, NY: Springer-Verlag. 046896
Smith WH. (1990). Forests as sinks for air contaminants: vegetative compartment. In Air pollution and forests:
interactions between air contaminants and forest ecosystems (pp. 147-180). New York, NY: Springer-
Verlag. 084015
Smith WH. (1991). Air pollution and forest damage. Chem Eng News, 6945: 30-43. 042566
Smolders Ei Degryse F. (2002). Fate and effect of zinc from tire debris in soil. Environ Sci Technol, 36: 3706—
3710. 156091
Soares CRFSi Siqueira JO. (2008). Mycorrhiza and phosphate protection of tropical grass species against heavy
metal toxicity in multi-contaminated soil. , 44: 833-841. 190482
Sokolik Ii Winker Di Bergametti Gi et al. (2001). Introduction to special section: outstanding problems in
quantifying the radiative impacts of mineral dust. J Geophys Res, 106: 18015-18027. 190404
Solomon PA; Hopke PK. (2008). A Special I ssue of JA&WMA Supporting Key Scientific and Policy-and Health-
Relevant Findings from EPA's Particulate Matter Supersites Program and Related Studies: An
Integration and Synthesis of Results. J Air Waste Manag Assoc, 58: 137. 156997
Solomon S; Portmann RW! Garcia RR; Thomason LW; Poole LR; McCormick MP. (1996). The role of aerosol
variations in anthropogenic ozone depletion at northern midlatitudes. J Geophys Res, 101: 6713-6727.
192252
Sotiropoulou REP; Medina J; Nenes A. (2006). CCN predictions: is theory sufficient for assessments of the
indirect effect?. Geophys Res Lett, 33: L05816. 190406
Sotiropoulou REP; Nenes A; Adams PJ; Seinfeld JH. (2007). Cloud condensation nuclei prediction error from
application of Kohler theory: Importance for the aerosol indirect effect. J Geophys Res, 112: D12202.
190405
Spinhirne J; Palm S; Hart W; Hlavka D; Welton E. (2005). Cloud and Aerosol Measurements from the GLAS
Space Borne Lidar: initial results. Geophys Res Lett, 32: L22S03. 190410
Spracklen DV; Logan JA; Mickley LJ; Park RJ; Yevich R; Westerling AL; Jaffe DA. (2007). Wildfires drive
interannual variability of organic carbon aerosol in the western U.S. in summer. Geophys Res Lett, 34:
L16816. 190485
Squires P. (1958). The microstructure and colloidal stability of warm clouds: Part I - the relation between
structure and stability. , 10: 256-261. 045608
Srogi. (2007). Monitoring of environmental exposure to polycyclic aromatic hydrocarbons: a review. , 5' 169-195.
180049
Stanhill G; Cohen S. (2001). Global dimming: a review of the evidence for a widespread and significant
reduction in global radiation with discussion of its probable causes and possible agricultural
consequences. , 107: 255-278. 042121
Steinnes E; Hvatum 00; Bolviken B; Varskog P. (2005). Atmospheric supply of trace elements studied by peat
samples from ombrotrophic bogs. J Environ Qual, 34: 192-197. 156095
Stenchikov G; Hamilton K; Robock A; Ramaswamy V; Schwarzkopf MD. (2004). Arctic Oscillation response to
the 1991 Pinatubo eruption in the SKYHI GCM with a realistic quasi-biennial oscillation. J Geophys Res,
109: D03112. 192274
Stenchikov G; Hamilton K; Stouffer RJ; Robock A; Ramaswamy V; Santer B; Graf H"F. (2006). Arctic Oscillation
response to volcanic eruptions in the IPCC AR4 climate models. J Geophys Res, 111: D07107. 192260
Stenchikov G; Robock A; Ramaswamy V; Schwarzkopf MD; Hamilton K; Ramachandran S. (2002). Arctic
Oscillation response to the 1991 Mount Pinatubo eruption: Effects of volcanic aerosols and ozone
depletion. J Geophys Res, 107: D24. 192277
Stenchikov GL; Kirchner I; Robock A; Graf H"F; Antuna JC; Grainger RG; Lambert A; Thomason L. (1998).
Radiative forcing from the 1991 Mount Pinatubo volcanic eruption. J Geophys Res, 103: 13837-13857.
192049
Stephens G; et al. (2002). The CloudSat mission and the A-Train. , 83: 1771-1790. 190412
Stephens GL; Haynes JM. (2007). Near global observations of the warm rain coalescence process. Geophys Res
Lett, 34: L20805. 190413
July 2009
9-299
DRAFT-DO NOT CITE OR QUOTE

-------
Stern DI. (2005). Global sulfur emissions from 1850 to 2000. Chemosphere, 58- 163-175. 190416
Stern GA; Halsall CJ; Barrie LA; Muir DCG; Fellin P; Rosenberg B. (1997). Polychlorinated biphenyls in arctic
air. 1. Temporal and spatial trends' 1992-1994. Environ Sci Technol, 31- 3619-3628. 156096
Stevens B; Feingold G> Walko RL; Cotton WR. (1996). On elements of the microphysical structure of numerically
simulated non-precipitating stratocumulus. J Trace Elem Med Biol, 53- 980-1006. 190417
Stewart AR; Luoma SN; Schlekat CE; Doblin MA; Hieb KA. (2004). Food web pathway determines how
selenium affects aquatic ecosystems. Environ Sci Technol, 38- 4519-4526. 156097
Stier P; Seinfeld JH; Kinne S; Boucher O. (2007). Aerosol absorption and radiative forcing. , 7- 5237-5261.
157012
Stohl A; Andrews E; Burkhart JF; Forster C; Herber A; Hoch SW; Kowal D; Lunder C; Mefford T; Ogren JA;
Sharma S; Spichtinger N; Stebel K; Stone R; Strom J; Torseth K; Wehrli C; Yttri KE. (2006). Pan-Arctic
enhancements of light absorbing aerosol concentrations due to North American boreal forest fires during
summer 2004. J Geophys Res, 111, D22214, doi:i0.1029/2006JD007216: . 156100
Stohl A; Berg T; Burkhart JF; Fjeeraa AM; Forster C; Herber A; Hov 0; Lunder C; McMillan WW; Oltmans S;
Shiobara M; Dimpson D; Solberg S; Stebel K; Strom J; Torseth K; Treffeisen R; Virkkunen K; Yttri KE.
(2007). Arctic smoke - record high air pollution levels in the European Arctic due to agricultural fires in
Eastern Europe in spring 2006. , 7: 511-534. 157015
Storlevmo T; Kristjansson JE; Myhre G; Johnsud M; Stordal F. (2006). Combined observational and modeling
based study of the aerosol indirect effect. , 6- 3583-3601. 190418
Stothers RB. (2001). A chronology of annual mean effective radii of stratospheric aerosols from volcanic
eruptions during the twentieth century as derived from ground-based spectral extinction measurements.
J Geophys Res, 106: 32043-32049. 192233
Stothers RB. (2001). Major optical depth perturbations to the stratosphere from volcanic eruptions: Stellar
extinction period, 1961—1978. J Geophys Res, 106: 2993—3003. 192232
Stott PA. (2006). Observational constraints on past attributable warming and predictions of future global
warming. J dim, 19: 3055-3069. 190419
Strachan WMJ; Burniston DA; Williamson M; Bohdanowicz H. (2001). Spatial difference in persistent
organochlorine pollutant concentrations between the Bering and Chukchi Seas. Mar Pollut Bull, 43: 132-
142. 156103
Strandberg B; Axelsen JA; Pedersen MB; Jensen J; Attrill MJ. (2006). Effect of a copper gradient on plant
community structure. Environ Toxicol Chem, 25: 743-753. 156105
Strawa A; Castaneda R; Owano T; Baer P; Paldus B. (2002). The measurement of aerosol optical properties
using continuous wave cavity ring-down techniques. , 20: 454-465. 190421
Streets D; Bond T; Lee T; Jang C. (2004). On the future of carbonaceous aerosol emissions. J Geophys Res, 109:
D24212.190423
Streets D; Wu Y; Chin M. (2006). Two-decadal aerosol trends as a likely explanation of the global
dimming/brightening transition. Geophys Res Lett, 33: L15806. 190425
Streets DG; Aunan K. (2005). The importance of China's household sector for black carbon emissions. Geophys
Res Lett, 32. 156106
Streets DG; Zhang Q; Wang L; He K; Hao J; Wu Y; Tang Y; Carmichael GR. (2006). Revisiting China?s CO
emissions after the Transport and Chemical Evolution over the Pacific (TRACE-P) mission: Synthesis of
inventories, atmospheric modeling, and observations. J Geophys Res, 111: D14306. 157019
Strydom C; Robinson C; Pretorius E; Whitcutt JM; Marx J; Bornman MS. (2006). The effect of selected metals
on the central metabolic pathways in biology: A review. , 32: 543-554. 190486
Swackhamer DL; Paerl HW! Eisenreich SJ; Hurley J; Hornbuckle KC; McLachlan M; Mount D; Muir D;
Schindler D. (2004). Impacts of atmospheric pollutants on aquatic ecosystems. , 12: 1-24. 190488
Sweet C; Caughey M; Gay D. (2005). Midwest Ammonia Monitoring Project: Summary for October 2003 through
November 2004. Illinois State Water Survey. Illinois. 180038
Szabo L; Fodor L. (2006). Uptake of microelements by crops grown on heavy metal-amended soil. Commun Soil
Sci Plant Anal, 37: 2679-2689. 156109
Tabazadeh A; Drdla K; Schoeberl MR; Hamill P; Toon OB. (2002). Arctic " ozone hole" in a cold volcanic
stratosphere. , 99: 2609-2612. 192250
Takemura T; Nakajima T; Dubovik O; Holben B; Kinne S. (2002). Single -scattering albedo and radiative forcing
of various aerosol species with a global three-dimensional model. Proc Natl Acad Sci USA, 15: 333-352.
190438
Takemura T; Nozawa T; Emori S; Nakajima T; Nakajima T. (2005). Simulation of climate response to aerosol
direct and indirect effects with aerosol transport-radiation model. J Geophys Res, 110: D02202. 190439
July 2009
9-300
DRAFT-DO NOT CITE OR QUOTE

-------
Tang IN. (1996). Chemical and size effects of hygroscopic aerosols on light scattering coefficients. J Geophys
Res, 101: 245-219. 157042
Tang Y> et al. (2003). Influences of biomass burning during the Transport and Chemical Evolution Over the
Pacific (TRACE-P) experiment identified by the regional chemical transport model. J Geophys Res, 108:
8824. 190441
Tang Y> et al. (2004). Three-dimensional simulations of inorganic aerosol distributions in East Asia during
spring 2001. J Geophys Res, 109: D19S23. 190445
Tanre D; et al. (2003). Measurement and modeling of the Saharan dust radiative impact: Overview of the
Saharan Dust Experiment (SHADE). J Geophys Res, 108: 8574. 190454
Tanre D; Kaufman Yi Herman Mi Mattoo S. (1997). Remote sensing of aerosol properties over oceans using the
MODIS/ EOS spectral radiances. J Geophys Res, 102: 16971-16988. 190452
Tao Si Jiao Xi Chen Si Xu Fi Li Yi Liu F. (2006). Uptake ofvapor and particulate polycyclic aromatic
hydrocarbons by cabbage. Environ Pollut, 140: 13—15. 156112
Taulavuori Ki Prasad MNVi Taulavuori Ei Laine K. (2005). Metal stress consequences on frost hardiness of
plants at northern high latitudes: a review and hypothesis. , 135: 209-220. 190489
Tett SFBi Jones GSi Stott PA; Hill DC; Mitchell JFB; Allen MR; Ingram WJ; Johns TC; Johnson CE; Jones A;
Roberts DL; Sexton DMH; Woodage MJ. (2002). Estimation of natural and anthropogenic contributions
to twentieth century temperature change. J Geophys Res, 107: 4306. 192053
Textor C; et al. (2007). The effect of harmonized emissions on aerosol properties in global models—an AeroCom
experiment. , T- 4489-4501. 190458
Textor C; Schulz M; Guibert S; et al. (2006). Analysis and quantification of the diversities of aerosol life cycles
within AEROCOM. , 6: 1777-1813. 190456
Thomas GO; Smith KEC; Sweetman AJ; Jones KC. (1998). Further studies of the air-pasture transfer of
polychlorinatedbiphenyls. , 102: 119-128. 156118
Thomason LW; Peter T. (2006). Assessment of Stratospheric Aerosol Properties (ASAP): Report on the
Assessment Kick"Off Workshop, Paris, France, 4"6 November 2001. Presented at Assessment of
Stratospheric Aerosol Properties , Paris, France. 192248
Tie X; et al. (2005). Assessment of the global impact of aerosols on tropospheric oxidants. J Geophys Res, 110: 0.
190459
Tie XX; Brasseur GP; Briegleb B; Granier C. (1994). Two-dimensional simulation of Pinatubo aerosol and its
effect on stratospheric ozone. J Geophys Res, 99: 20545-20562. 192253
Timmreck C; Graf HF; Steil B. (2003). Aerosol chemistry interactions after the Mt. Pinatubo eruption. In
Robock A; Oppenheimer C (Ed.),Volcanism and the earth's atmosphere (pp. 213-227). Washington, DC:
American Geophysical Union. 192254
Tombach I; McDonald K. (2004). Visibility and Radiative Balance Effects. In P.H. McMurry MFS, and J.S.
Vickery (Ed.),Particulate Matter Science for Policy Makers: A NARSTO Assessment: Cambridge
University Press. 157054
Toose LK; Mackay D. (2004). Adaptation of fugacity models to treat speciating chemicals with constant species
concentration ratios. Environ Sci Technol, 38: 4619—4626. 156123
Torres O; Bhartia P; Herman J; Ahmad Z; Gleason J. (1998). Derivation of aerosol properties from satellite
measurements of backscattered ultraviolet radiation: Theoretical bases. J Geophys Res, 103: 17009"
17110. 190503
Torres O; Bhartia P; Herman J; Sinyuk A; Ginoux P; Holben B. (2002). A long-term record of aerosol optical
depth from TOMS observations and comparison to AERONET measurements. J Trace Elem Med Biol,
59: 398-413. 190505
Torres O; Bhartia P; Sinyuk A; Welton E; Holben B. (2005). Total Ozone Mapping Spectrometer measurements
of aerosol absorption from space: Comparison to SAFARI 2000 groundbased observations. J Geophys Res,
110: D10S18. 190507
Tremper AH; Agneta M; Burton S; Higgs DEB. (2004). Field and laboratory exposures of two moss species to low
level metal pollution. , 49: 111-120. 156126
Trijonis J; Thayer M; Murdoch J; Hagemen R. (1985). Air quality benefit analysis for Los Angeles and San
Francisco based on housing values and visibility	 078468
Trijonis JC; Malm WC; Pitchford M; White WH; Charlson R. (1990). Acidic deposition: State of science and
technology. Report 24. Visibility: Existing and historical conditions-causes and effects. Final report.
157058
TSL Christopher S; Zhang J. (2002). Daytime variation of shortwave direct radiative forcing of biomass burning
aerosols from GOES-8 imager. J Trace Elem Med Biol, 59: 681-691. 190031
July 2009
9-301
DRAFT-DO NOT CITE OR QUOTE

-------
Turco RP; Toon OB; Whitten RC; Pollack JB; Hamill P. (1983). The global cycle of particulate elemental carbon:
a theoretical assessment. In Pruppacher HR (Ed.),Precipitation Scavenging, Dry Deposition, and
Resuspension (pp. 1337" 1351). New York: Elsevier Science. 190529
Twomey S. (1977). The influence of pollution on the shortwave albedo of clouds. J Trace Elem Med Biol, 34:
1149-1152. 190533
U.S. EPA,Zhang Q,Jimenez J,Canagaratna M,Allan J,Coe H,Ulbrich I,Alfarra M,Takami A,Middlebrook A,Sun
Y. (2007). Ubiquity and dominance of oxygenated species in organic aerosols in anthropogenically-
influenced Northern Hemisphere midlatitudes. , 34: L13801. 189998
U.S. EPA. (1979). Protecting Visibility, an EPA Report to Congress. U.S. Environmental Protection Agency.
Washington, D.C.. 157065
U.S. EPA. (1982). Air quality criteria for particulate matter and sulfur oxides. U.S. Environmental Protection
Agency. Washington, D.C.. 017610
U.S. EPA. (1993). Air quality criteria for oxides of nitrogen. U.S. Environmental Protection Agency.
Washington, D.C.. 017649
U.S. EPA. (1996). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA/600/P-95/001aF-cF. 079380
U.S. EPA. (1997). Mercury Study Report to Congress. 157066
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA/600/P-99/002aF-bF. 056905
U.S. EPA. (2005). Review of the national ambient air quality standards for particulate matter: Policy
assessment of scientific and technical information OAQPS staff paper. U.S. EPA. Research Triangle
Park, North Carolina. 090209
U.S. EPA. (2006). Air quality criteria for lead. 090110
U.S. EPA. (2008). Integrated science assessment for oxides of nitrogen and sulfur: Ecological criteria. EPA.
Research Triangle Park, NC. EPA/600/R-08/082F. 157074
US Environmental Protection Agency. (2004). The Particle Pollution Report: Current Understanding of Air
Quality and Emissions through 2003. US EPA. Research Triangle Park, NC. EPA 454-R-04-002. 190219
Vaisvalavicius R; Motuzas A; Prosycevas L Levinskaite L; Zakarauskaite D; Grigaliuniene K; Butkus V. (2006).
Effect of heavy metals on microbial communities and enzymatic activity in soil column experiment. Arch
Agron SoH Sci, 52: 161-169. 157080
Van Aardenne JA; Dentener FJ; Olivier JGJ; Klein Goldewijk CGM; Lelieveld J. (2001). A l"degree" x l"degree"
resolution data set of historical anthropogenic trace gas emissions for the period 1980-1990. Global
Biogeochem Cycles, 15: 909-928. 055564
Van de Hulst H. (1981). Light scattering by small particles. New York: Dover. 191972
van Heerden PDR; Kriiger GHJ; Kilbourn Louw M. (2007). Dynamic responses of photosystem II in the Namib
Desert shrub, Zygophyllum prismatocarpum, during and after foliar deposition of limestone dust. , 146:
34-45. 156131
Van Donkelaar A; Martin RV; Park RJ. (2006). Estimating ground-level PM2.5 using aerosol optical depth
determined from satellite remote sensing . J Geophys Res, 111: D21. 192108
Vasundhara G> Jayashree G> Muraleedhara-Kurup G. (2004). Sequestration of nickel and copper by Azotobacter
chroococcum SB1. , 72: 1122-1127. 156133
Vautard R; Maidi Mi Menut Li Beekmann Mi Colette A. (2007). Boundary layer photochemistry simulated with
a two-stream convection scheme. Atmos Environ, 41: 8275-8287. 106012
Veihelmann Bi Levelt PFi Stammes Pi Veefkind JP. (2007). Simulation study of the aerosol information content
in OMI spectral reflectance measurements. , 7' 3115-3127. 190627
Verdelhos Ti Neto JMi Marques JCi Pardal MA. (2005). The effect of eutrophication abatement on the bivalve
Scrobicularia plana. , 63: 261-268. 190497
Viard Bi Pihan Fi Promeyrat Si Pihan J-C. (2004). Integrated assessment of heavy metal (Pb, Zn, Cd) highway
pollution: bioaccumulation in soil, graminaceae and land snails. Chemosphere, 55: 1349-1314. 055675
Viles HAi Gorbushina AA. (2003). Soiling and microbial colonisation on urban roadside limestone: A three year
study in Oxford, England. , 38: 1217-1224. 156138
Viles HA; Taylor MP; Yates TJS; Massey SW. (2002). Soiling and decay of N.M.E.P. limestone tablets. Sci Total
Environ, 292: 215-229. 156137
Villa S; Vighi M; Maggi V; Finizio A; Bolzacchini E. (2003). Historical trends of organochlorine pesticides in an
Alpine glacier. , 46: 295-311. 156139
Viskari E"L; Karenlampi L. (2000). Roadside Scots pine as an indicator of deicing salt use - a comparative study
from two consecutive winters. Water Air Soil Pollut, 122: 405-419. 019101
July 2009
9-302
DRAFT-DO NOT CITE OR QUOTE

-------
Vivas A; Barea JM; Biro B; Azcon R. (2006). Effectiveness of autochthonous bacterium and mycorrhizal fungus
on Trifolium growth, symbiotic development and soil enzymatic activities in Zn contaminated soil. J Appl
Microbiol, 100: 587-598. 190500
Vivas A; Voros A; Biro B; Barea JM; Ruiz-Lozano JM; Azcon R. (2003). Beneficial effects of indigenous Cd-
tolerant and Cd-sensitive Glomus mosseae associated with a Cd-adapted strain of Brevibacillus sp in
improving plant tolerance to Cd contamination. , 24- 177-186. 190499
Vives I; Grimalt JO; Ventura M; Catalan J. (2005). Distribution of polycyclic aromatic hydrocarbons in the food
web of a high mountain lake, Pyrenees, Catalonia, Spain. Environ Toxicol Chem, 24- 1344-1352. 157099
Vogel-Mikus K; Drobne D; Regvar M. (2005). Zn, Cd and Pb accumulation and arbuscular mycorrhizal
colonisation of pennycress Thlaspi praecox Wulf. (Brassicaceae) from the vicinity of a lead mine and
smelter in Slovenia. , 133- 233-242. 190501
Vogel-Mikus K; Pongrac P; Kump P; Necemer M; Regvar M. (2006). Colonisation of a Zn, Cd and Pb
hyperaccumulator Thlaspi praecox Wulfen with indigenous arbuscular mycorrhizal fungal mixture
induces changes in heavy metal and nutrient uptake. , 139- 362-371. 190502
Waggoner AP; Weiss RE. (1980). Comparison of fine particle mass concentration and light scattering extinction
in ambient aerosol. Atmos Environ, 14- 623-626. 070152
Waggoner AP; Weiss RE; Ahlquist NC; Covert DS; Will S; Charlson RJ. (1981). Optical characteristics of
atmospheric aerosols. Atmos Environ, 15- 1891-1909. 095453
Wang C. (2007). Impact of direct radiative forcing of black carbon aerosols on tropical convective precipitation.
Geophys Res Lett, 34: 5709. 156147
Wang Y-Z; Ingram JL; Walters DM; Rice AB; Santos JH; Van Houten B; Bonner JC. (2003). Vanadium-induced
STAT-1 activation in lung myofibroblasts requires H202 and P38 MAP kinase. Free Radic Biol Med, 35:
845-855. 157106
Wania F; Hoff JT; Jia CQ; Mackay D. (1998). The effects of snow and ice on the environmental behaviour of
hydrophobic organic chemicals. , 102: 43-51. 156148
Wania F; Mackay D. (1993). Global fractionation and cold condensation of low volatility organochlorine
compounds in polar regions. Ambio, 22: 10-18. 157110
Warner J. (1968). A reduction in rainfall associated with smoke from sugar-cane fires—an inadvertent weather
modification?. , T- 247-251. 157114
Warner J; Twomey S. (1967). The production of cloud nuclei by cane fires and the effect on cloud droplet
concentration. J Trace Elem Med Biol, 24: 704-706. 045616
Watmough SA; Hutchinson TC; Dillon PJ. (2004). Lead dynamics in the forest floor and mineral soil in south-
central Ontario. , 71: 43-68. 077809
Watson JG; Chow JC. (2007). Receptor models for source apportionment of suspended particles. In Introduction
to Environmental Forensics (pp. x). x: x. 157127
Watson JG; Zhu T; Chow JC; Engelbrecht J; Fujita EM; Wilson WE. (2002). Receptor modeling application
framework for particle source apportionment. Chemosphere, 49: 1093-1136. 035623
Weissenhorn I; Leyval C; Berthelin J. (1995). Bioavailability of heavy metals and abundance of arbuscular
mycorrhiza in a soil polluted by atmospheric deposition from a smelter. , 19: 22"28. 073826
Welton E; Voss K; Quinn P; Flatau P; Markowicz K; Campbell J; Spinhirne J; Gordon H; Johnson J. (2002).
Measurements of aerosol vertical profiles and optical properties during INDOEX 1999 using micro-pulse
lidars. J Geophys Res, 107: 8019. 190631
Welton EJ; Campbell JR; Spinhirne JD; Scott VS. (2001). Global monitoring of clouds and aerosols using a
network of micro-pulse lidar systems. : . 157133
Wen T; Wang Y; Chang SY; Liu G. (2006). On-line Measurement of Water-Soluble Ions in Ambient Particles. ,
23: 586-592. 179964
Wetzel MA; Stowe LL. (1999). Satellite-observed patterns in stratus microphysics, aerosol optical thickness, and
shortwave radiative forcing. J Geophys Res, 104: 31287-31299. 190636
Wielicki B; Barkstrom B; Harrison E; Lee R; Smith G; Cooper J. (1996). Clouds and the Earth's radiant energy
system (CERES): An Earth observing system experiment. , 77: 853-868. 190637
Wild E; Dent J; Thomas GO; Jones KC. (2005). Direct observation of organic contaminant uptake, storage, and
metabolism within plant roots. Environ Sci Technol, 39: 3695-3702. 156156
Wild SR; Jones KC. (1992). Polynuclear aromatic hydrocarbons uptake by carrots grown in sludge amended soil.
J Environ Qual, 21: 217- 225. 156155
Winker D; Couch R; McCormick M. (1996). An overview of LITE: NASA's Lidar In-Space Technology
Experiment. , 84: 164-180. 190914
Winker DM; Pelon J; McCormick MP. (2003). The CALIPSO mission: spaceborne lidar for observation of
aerosols and clouds. , 4893: 1-11. 192017
July 2009
9-303
DRAFT-DO NOT CITE OR QUOTE

-------
Wolff GT. (1985). Characteristics and consequences of soot in the atmosphere. Environ Int, 11- 259-269. 044680
Wyszkowska J; Boros Ei Kucharski J. (2007). Effect of interactions between nickel and other heavy metals on
the soil microbiological properties. , 53- 544-552. 179948
Xu J; DuBois D; Pitchford M; Green M; Etyemezian V. (2006). Attribution of sulfate aerosols in Federal Class I
areas of the western United States based on traiectory regression analysis. Atmos Environ, 40- 3433"
3447. 102706
Xue H; Feingold G. (2006). Large eddy simulations of tradewind cumuli- Investigation of aerosol indirect effects.
J Trace Elem Med Biol, 63: 1605-1622. 190920
Xue H; Feingold G; Stevens B. (2008). Aerosol effects on clouds, precipitation, and the organization of shallow
cumulus convection. J Trace Elem Med Biol, 65: 392" 406. 190921
Yang CY; Chen CJ. (2007). Air pollution and hospital admissions for chronic obstructive pulmonary disease in a
subtropical city: Taipei, Taiwan. J Toxicol Environ Health A, 70' 1214-9. 092847
Yang F; Schlesinger M. (2001). Identification and separation of Mount Pinatubo and El Nino-Southern
Oscillation land surface temperature anomalies. J Geophys Res, 106: 14757—14770. 192270
Yang XE; Jin XF; Feng Yi Islam E. (2005). Molecular mechanisms and genetic basis of heavy metal
tolerance/hyperaccumulation in plants. , 47: 1025-1035. 192104
Yang Z; Zhu L. (2007). Performance of the partition-limited model on predicting ryegrass uptake of polycyclic
aromatic hydrocarbons. Chemosphere, 67: 402-409. 156168
Ying Q; Kleeman MJ. (2006). Source contributions to the regional distribution of secondary particulate matter
in California. Atmos Environ, 40: 736-752. 098359
Yogui G> Sericano J. (2008). Polybrominated diphenyl ether flame retardants in lichens and mosses from King
George Island, maritime Antarctica. Chemosphere, 73: 1589-1593. 189971
Yu H, et al. (2006). A review of measurement-based assessments of the aerosol direct radiative effect and
forcing. , 6: 613-666. 156173
Yu H; Dickinson R; Chin M; Kaufman Y; Zhou M; Zhou L; Tian Y; Dubovik O; Holben B. (2004). The direct
radiative effect of aerosols as determined from a combination of MODIS retrievals and GOCART
simulations. J Geophys Res, 109: D03206. 190926
Yu H; Liu S; Dickinson R. (2002). Radiative effects of aerosols on the evolution of the atmospheric boundary
layer. J Geophys Res, 107: 4142. 190923
Yu IJ; Park JD; Park ES; Song KS; Han KT; Han JH; Chung YH; Choi BS; Chung KH; Cho MH. (2003).
Manganese distribution in brains of Sprague-Dawley rats after 60 days of stainless steel welding-fume
exposure. Neurotoxicology, 24: 777-785. 156171
Yui et al. (2007). Translocation and effects of gold nanoparticles after inhalation exposure in rats. , 2' 713-717.
093173
Yuangen Y> Campbell CD; Clark L; Camerson CM; Paterson E. (2006). Microbial indicators of heavy metal
contamination in urban and rural soils. Chemosphere, 63: 1942-1952. 156174
Yunker MB; Snow don LR; MacDonald RW; Smith JN; Fowler MG; Skibo DN. (1996). Polycyclic aromatic
hydrocarbon composition and potential sources for sediment samples from the Beufort and Barents Seas.
Environ Sci Technol, 30: 1310-1320. 156175
Zaccone C; Cocozza C; Cheburkin AK; Shotyk W; Miano TM. (2007). Highly Organic Soils as "Witnesses" of
Anthropogenic Pb, Cu, Zn, and 137Cs Inputs During Centuries. Water Air Soil Pollut, 186: 263-271.
179930
Zappia G; Sabbioni C; Riontino C; Gobbi G; Favoni O. (1998). Exposure tests of building materials in urban
atmosphere. Sci Total Environ, 224: 235-244. 012037
Zechmeister HG. (1998). Annual growth of four pleurocarpous moss species and their applicability for
biomonitoring heavy metals. Environ Monit Assess, 52: 441-451. 156178
Zechmeister HG; Hohenwallner D; Riss A; Hanus-Illnar A. (2003). Variation in heavy metal concentrations in
the moss species Abietinella abietina (Hedw.) Fleisch. according to sampling time, within site variability
and increase in biomass. Sci Total Environ, 301: 55"65. 157175
Zeidner M; Shechter M. (1988). Psycological responses to air polution: Some personality and demographic
correlates. , 8: 191-208. 191973
Zhang J; Christopher S. (2003). Longwave radiative forcing of Saharan dust aerosols estimated from MODIS,
MISR, and CERES observations on Terra. Geophys Res Lett, 30: 2188. 190928
Zhang J; Christopher S; Remer L; Kaufman Y. (2005). Shortwave aerosol radiative forcing over cloud-free
oceans from Terra. I: Angular models for aerosols. J Geophys Res, 110: D10S23. 190929
Zhang J; Christopher S; Remer L; Kaufman Y. (2005). Shortwave aerosol radiative forcing over cloud-free
oceans from Terra. II: Seasonal and global distributions. J Geophys Res, 110: D10S24. 190930
July 2009
9-304
DRAFT-DO NOT CITE OR QUOTE

-------
Zhang J; Reid JS; Holben BN. (2005). An analysis of potential cloud artifacts in MODIS over ocean aerosol
optical thickness products. Geophys Res Lett, 32- L15803. 190931
Zhang J; Reid JS; Westphal DL; Baker NL; Hyer EJ. (2008). A system for operational aerosol optical depth data
assimilation over global oceans. J Geophys Res, 113- D10208. 190932
Zhang KM; Wexler AS; Niemeier DA; Zhu YF; Sioutas W; Sioutas C. (2005). Evolution of particle number
distribution near roadways Part III- traffic on-road size resolved particulate emission factors. Atmos
Environ, 39: 4155-4166. 086743
Zhang Q; Jimenez JL; Canagaratna MR; Jayne JT; Worsnop DR. (2005). Time- and size-resolved chemical
composition of submicron particles in Pittsburgh: Implications for aerosol sources and processes. J
Geophys Res, 110: 1-19. 157185
Zhang X; Zwiers FW; Stott PA. (2006). Multi-model multisignal climate change detection at regional scale. J
Clim, 19: 4294-4307. 190933
Zhang XH; Zhu YG; Chen BD; Lin AJ; Smith SE; Smith FA. (2005). Arbuscular mycorrhizal fungi contribute to
resistance of upland rice to combined metal contamination of soil. , 28: 2065-2077. 192083
Zhang XZ; Sun HW; Zhang ZY. (2006). [Bioaccumulation of titanium dioxide nanoparticles in carp]. , 27: 1631-5.
157722
Zhao TXP; Laszlo I; Guo W', Heidinger A; Cao C; Jelenak A; Tarpley D; Sullivan J. (2008). Study of long-term
trend in aerosol optical thickness observed from operational AVHRR satellite instrument. J Geophys Res,
113: DQ7201.190935
Zhao TXP; Yu H; Laszlo I; Chin M; Conant WC. (2008). Derivation of component aerosol direct radiative forcing
at the top of atmosphere for clear-sky oceans. , 109: 1162-1186. 190936
Zhou M, et al.. (2005). A normalized description of the direct effect of key aerosol types on solar radiation as
estimated from aerosol robotic network aerosols and moderate resolution imaging spectroradiometer
albedos. J Geophys Res, 110, D19202, doi:i0.1029/2005JD005909: . 156183
Zhu Y; Hinds WC; Shen S; Sioutas C. (2004). Seasonal trends of concentration and size distribution of ultrafine
particles near major highways in Los Angeles. Aerosol Sci Technol, 38: 5-13. 156184Zimmer D; Baum C;
Leinweber P; Hrynkiewicz K; Meissner R. (2009). Associated bacteria increase the phytoextraction of
cadmium and zinc from a metal-contaminated soil by mycorrhizal willows. ,11: 200-213. 192085
July 2009
9-305
DRAFT-DO NOT CITE OR QUOTE

-------
Annex A. Atmospheric Science
A.1. Ambient Air Particle Monitoring
A.1.1. Measurements and Analytical Specifications
Table A-1.
Summary of integrated and continuous samplers included in the field comparison.
Abbreviation
i Instrument
Manufacturer / Research Institute
INTEGRATED PARTICLE OR GAS/PARTICLE INSTRUMENTS
Dichot
Dichotomous Sampler with Virtual Impactor
Andersen Instruments (Smyrna, GA
AND-241 Dichot
Thermo Andersen Series 241 Dichotomous Sampler
Andersen Instruments
AND-246 Dichot
Thermo Andersen SA-246B Dichotomous Sampler
Andersen Instruments
AI\ID-hlV0L10FRM
Thermo Andersen GMW-1200 HiVol PMio FRM Sampler
Andersen Instruments
ARA-PCM
ARA Particle Composition Monitor
Atmospheric Research and Analysis Inc. (Piano, TX)
CMU
CMU Speciation Sampler
Carnegie Mellon University (CMU), (Pittsburgh, PA)
DRI-SFS
DRI Sequential Filter Sampler
Desert Research Institute (Reno, NV)
HEADS (or HI)
Harvard EPA Annular Denuder System (or Harvard Impactor)
Harvard School of Public Health (Boston, MA)
IMPROVESS"
IMPROVE Speciation Sampler
URG Corp. (Chapel Hill, NC)
URG-3000Nb
Modined IMPROVE Module C Sampler for Carbon
URG Corp.
MASS-400b
URG Mass Aerosol Speciation Sampler Model 400
URG Corp.
MASS-450b
URG MASS Model 450
URG Corp.
MiniVol
Battery-Powered Portable Low-Volume Sampler
Air Metrics Inc. (Eugene, OR)
PC-BOSS
Particle Concentrator-Brigham Young University Organic
Brigham Young University (Provo, UT)
SAMPLING SYSTEM
PQ-200 FRM
BGI PQ-200 FRM Sampler
BGI Inc. (Waltham, MA)
PQ-200 FRMA
BGI PQ-200A FRM Audit Sampler
BGI Inc.
R&P-ACCU
R&P-Automated Cartridge Collector Unit Sampler
Rupprecht & Patashnick, Co. (Albany, NY
R&P-2000 FRM
R&P Partisol-2000 FRM Sampler
Rupprecht & Patashnick, Co.
R&P-2000 FRMA
R&P Partisol-2000 FRM Audit Sampler
Rupprecht & Patashnick, Co.
R&P-2025 Dichot"
R&P Partisol 2025 Dichotomous Sequential Air Sampler
Rupprecht & Patashnick, Co.
R&P-2025 FRM
R&P Partisol-Plus Model 2025 PM2.6 Sequential Samplers
Rupprecht & Patashnick, Co.
R&P-2300b
R&P Partisol 2300 Chemical Speciation Sampler
Rupprecht & Patashnick, Co.
Note- Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health
and Environmental Research Online) at http • I lev a. go v/her o. HERO is a database of scientific literature used by U.S. EPA in
the process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009
A-l
DRAFT-DO NOT CITE OR QUOTE

-------
Abbreviation
Instrument
Manufacturer / Research Institute
RAAS-100 FRM	Thermo Andersen Reference Ambient Air Sampler Model 100	Andersen Instruments
FRM SAMPLER
RAAS-200 FRM
Thermo Andersen RAAS Model 200 FRM Audit Sampler
Andersen Instruments
RAAS-300 FRM
Thermo Andersen RAAS Model 300 FRM Sampler
Andersen Instruments
RAAS-400b
Thermo Andersen RAAS Model 400 Speciation Sampler
Andersen Instruments
SASSb
MetOne Spiral Ambient Speciation Sampler
Met One Instruments (Grants Pass, OR)
SCS
PM2.E Sequential Cyclone Sampler
New York University (New York, NY)
URG-PCMb
URG Particle Composition Monitor
URG Corp. (Chapel Hill, NC)
VAPS
URG Versatile Air Pollution Sampler
URG Corp.
CONTINUOUS MASS INSTRUMENTS
BAM
B-Attenuation Monitor Model 1020
Met One Instruments
nano-BAM
Met One BAM Model 1020 with 150 nm impactor
Met One Instruments
CAMM
Continuous Ambient Mass Monitor
Developed by Harvard School of Public Health, comercialized by
Thermo Andersen Instruments; now withdrawn from market
RAMS
Real-Time Ambient Mass Sampler (modined Tapered Element Oscillation
Microbalance with diffusion denuder and Nanon dryer)
Brigham Young University
TEOM
Tapered Element Oscillating Microbalance
Rupprecht 81 Patashnick, Co.
30 °C-TE0M
TEOM operated at 30 °C
Rupprecht 81 Patashnick, Co.
50 °C-TE0M
TEOM operated at 50 °C
Rupprecht 81 Patashnick, Co.
SESTEOM
TEOM 1400a Series with Sample Equilibration System
Rupprecht 81 Patashnick, Co.
D-TEOM
Differential TEOM
Rupprecht 81 Patashnick, Co.
FDMS-TEOM
Filter Dynamics Measurement System TEOM
Rupprecht 81 Patashnick, Co.
ACCU-TEOM
TEOM 1400 Series with an automated cartridge collection unit
Rupprecht 81 Patashnick, Co.
CONTINUOUS PARTICLE LIGHT SCA TTERING INSTRUMENTS
Dust Trak
Dust Trak nephelometer
TSI Inc. (Shoreview, MN)
EcoTech
EcoTech Model M9003 nephelometer
EcoTech Pty Ltd., Australia (American EcoTech, Warren, Rl)
NGN
NGN-2 nephelometer
Optec Inc. (Lowell, Ml)
RR-M903
Radiance Research Nephelometer Model M903
Radiance Research Inc. (Seattle, WA)
CONTINUOUS ELEMENT INSTRUMENTS
GFAAS
Graphite Furnace Atomic Absorption Spectrometry-aerosol collection
as preconcentrate slurry
University of Maryland (College Park, MD)
SEAS
Semicontinuous Elements in Aerosol Sampler
University of Maryland
CONTINUOUS NITRA TE INSTRUMENTS
ADI-N
Aerosol Dynamics Inc. dash volatilization analyzer
Aerosol Dynamics Inc. (Berkeley, CA)
ARAN
Atmospheric Research and Analysis NO3 analyzer
Atmospheric Research and Analysis Inc.
R&P-8400N
R&P-8400N Flash Volatilization Continuous NO3' Analyzer
Rupprecht 81 Patashnick, Co.
CONTINUOUS SULFA TE INSTRUMENTS
ADI-S
Aerosol Dynamics Inc. Flash Volatilization Analyzer
Aerosol Dynamics Inc.
CASM
Continuous Ambient Sulfate Monitor (prototype of the TE-5020 by
Thermo Electron [Franklin, MA])
Harvard School of Public Health
July 2009
A-2
DRAFT-DO NOT CITE OR QUOTE

-------
Abbreviation
Instrument
Manufacturer / Research Institute
R&P-8400S	R&P-8400S Flash Volatilization Continuous SO42' Analyzer	Rupprecht & Patashnick, Co.
TE-5020
Thermo Electron Model 5020 SO42' Particulate Analyzer
Thermo Electron Corp. (Franklin, MA)
CONTINUOUS MUL Tl ION INSTRUMENTS
AIM
Ambient Ion Monitor Model 9000 (CI',N02',N03',P043',
S042", NH4+,Na+,Mg2+,K+,Ca2+)
URG Corp.
~ionex-IC
Dionex Ion Chromatograph (F, CI", NO2', Br',N03', PO43', SO42', Li+,
NH4+ ,Na+,Mg2+,K+,Ca2+)
Dionex Corp.
ECN
Energy Research Center of the Netherlands IC-based sampler (CI , NO3',
S042+,NH4+,Na+,Mg2+,K+,Ca2+)
Energy Research Center of the Netherlands (Petten, the
Netherlands
PILS-IC
Particle into Liquid Sampler, coupled with IC (CI, NO2', NO3', PO43',
S042+,NH4+,Na+,Mg2+,K+,Ca2+)
Georgia Institute of Technology (Atlanta, GA)
TT
Texas Tech IC-based sampler (NO3', S042+)
Texas Tech University (Lubbock, TX)
CONTINUOUS CARBON INSTRUMENTS
OC and EC
ADI-C
ADI Flash Volatilization Carbon Analyzer
Aerosol Dynamics Inc.
RU-OGI
Rutgers University/Oregon Graduate Institute in-situ carbon analyzer
(OC, EC)
Rutgers University (Camden, NJHOregon Graduate Institute
(Beaverton, OR)
R&P-5400
R&P-5400 continuous ambient carbon analyzer
Rupprecht 81 Patashnick, Co.
Sunset OCEC
Sunset Semi-Continuous Real-Time Carbon Aerosol Analysis Instrument
Sunset Laboratory, Inc. (Tigard, OR)
BC
Aethalometer

Magee Scientinc Co. (Berkeley, CA)
AE-16
Magee AE-16 aethalometer (BC)
Magee Scientinc Co.
AE-20
Magee AE-20 dual wavelength aethalometer (BC)
Magee Scientinc Co.
AE-21
Magee AE-21 dual-wavelength aethalometer (BC)
Magee Scientinc Co.
AE-31
Magee AE-31 seven color aethalometer (BC)
Magee Scientinc Co.
DRI-PA
DRI Photoacoustic Analyzer (BC)
Droplet Measurement Technologies, Inc. (Boulder, CO)
MAAP
Multi-Angle Absorption Photometer, Model 5012 (BC)
Thermo Scientinc Corp. (Franklin, MA)
PSAP
Particle Soot Absorption Photometer (BC)
Radiance Research Inc. (Seattle, WA)
OTHER CARBON
PASPAH
Photo-lonization Monitor for PAHs (Model PAS 2000)
EcoChem Analytics (League City, TX)
PILS-WSOC
PILS-WSOC Analyzer, combination of PILS and total organic analyzer
(TOA)
Georgia Institute of Technology
PARTICLE SIZING INSTRUMENTS FOR MASS AND CHEMICAL SPECIA TION
DRUM-3
Davis Rotating-Drum Uniform Size-Cut Monitor (0.1-2.5/jm in three
stages)
University of California-Davis (Davis, CA)
DRUM-8
Davis Rotating-Drum Uniform Size-Cut Monitor (0.09- > 5.0/jm in
eight stages)
University of California-Davis
ELPI
Electrical Low Pressure Impactor (0.007-10 //m in 12 stages)
Dekati (Tampere, Finland)
LPI
Low Pressure Impactor (0.03-10 /jm in 13 stages)
Aerosol Dynamics, Inc.
MOUDI
Micro Orince Uniform Deposit Impactor
MSP Corp. (Minneapolis, MN)
MOUDI-100
MOUDI Model 100 (0.18-18/jm in eight stages)
MSP Corp.
M0UDI-110
MOUDI Model 110 (0.056-18/jm in 10 stages)
MSP Corp.
July 2009
A-3
DRAFT-DO NOT CITE OR QUOTE

-------
Abbreviation
Instrument
Manufacturer / Research Institute
Nano-MOUDI
Nano M0UDI (0.010-0.056//m in three stages coupled to M0UDI
Model 110)
MSP Corp.
PARTICLE NUMBER/ VOLUME INSTRUMENTS
APS
Aerodynamic Particle Sizer
TSI Inc.
APS-3320
TSI Model 3320 (0.5-20//m)
TSI Inc.
APS-3321
TSI Model 3321 (0.5-20 fjm; replaced TSI Model 3320)
TSI Inc.
DMA
Differential Mobility Analyzer
TSI Inc.
DMA-3081
TSI Model 3081 (0.01-1.0 m)
TSI Inc.
DMA-3085
TSI Model 3085 (0.002-0.15//m)
TSI Inc.
EEPS
Engine Exhaust Particle Sizer (EEPS 0.056-0.56//m)
TSI Inc.
FMPS
Fast Mobility Particle Sizer (FMPS 0.056-0.56//m)
TSI Inc.
GRIMM-1108
Optical Particle Counter (OPC; 0.3-20 //m)
GRIMM Technologies, Inc. (Douglasville, GA)
SMPS
Scanning Mobility Particle Sizer
TSI Inc.
SMPS-3936
TSI Model 3936L (0.01-1.0//m)
TSI Inc.
Nano-SMPS-3936 TSI Model 3936N (0.002-0.15 //m)
TSI Inc.
SMPS + C
SMPS and Condensation Nucleus Counter (0.005-0.35 or 0.01-0.875
/jm)
GRIMM Technologies, Inc.
SMPS-custom
DMA Model 3071 and CPC Model 3010
TSI Inc.
WPS
Wide-Range Particle Spectrometer (0.01-10.0//m)
MSP Corp.
SINGLE PARTICLE INSTRUMENTS
AMS
Aerosol Mass Spectrometer (0.04-2//m)
Aerodyne Research Inc. (Billerica, MA)
AT0FMS
Aerosol Time of Flight Mass Spectrometer (0.3-2.5 //m)
TSI Inc.
CNC, CPC
Condensation Nucleus Counters, Condensation Particle Counter
Various vendors
DAASS
Dry-Ambient Aerosol Size Spectrometer consisting of two SMPS and
One APS (0.003—10 //m)
Carnegie Mellon University
LIBS
Laser-Induced Breakdown Spectroscopy
National Research Council, Industrial Materials Institute
(Boucherville, Quebec, Canada)
PALMS
Particle Analysis by Laser Mass Spectrometer (0.22-2.5//m)
NOAA (Boulder, CO)
RSMS-II
Rapid Single Particle Mass Spectrometer -II (0.035-1.1 //m)
University of Delaware (Newark, DE)
RSMS-III
Rapid Single Particle Mass Spectrometer RSMS-III (0.01 -2.0 //m)
University of Delaware
LABORA TORY INSTRUMENTS
DRI Model 2001
DRI Model 2001 Thermal/Optical Carbon Analyzer (OC, EC, Eight Carbon
Fractions with renectance and transmittance laser correction)
Atmoslytic, Inc. (Calabasas, CA)
SEM
Scanning Electron Microscopy
Various vendors
8Now with Thermo Scientific, Franklin, MA.
bEPA-approved speciation sampler used in the Speciation Trends Network (STN).
eNow commercialized by Applikon Analytical, the Netherlands, and marketed under the name "MARGA" (Monitor for Aerosols and Gases in Ambient Air).
dNot available.
Source: Chow et al. (2008,156355)
July 2009
A-4
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-2. Summary of PM2.5 and PM10 FRM and FEM samplers.
Manufacturer3
Sampler
Name
Size
Cut"
Description
FRM or
FEMC
Designation #
FRN
BGI Inc.
PQ-100
PM10
Louvered PMio inlet; operates at flow rate of 16.7 L/min; 24-h integrated
sampler; uses a mass flow meter to adjust to equivalent volumetric flow at
FRM
RFPS-1298-124
Vol. 63, p. 69625,
12/17/98
BGI Inc.
PQ-200
PM10
ambient temperature and pressure.
FRM
RFPS-1298-125
Vol. 63, p. 69625,
12/17/98
BGI Inc.
PQ-200
PM2.B
Identical to PMio sampler, but uses a WINS9 impactor downstream of the
PMio inlet for PM2.6 fractionation at 16.7 L/min; 24-h integrated sampler.
FRM
RFPS-0498-116
Vol. 63, p. 18911,
04/16/98
Vol. 63, p. 31993,
06/11/98
BGI Inc.
PQ-200VSCC
PM2.B
Same as BGI PQ200 PM2.6 sampler, but with BGI VSCC instead of WINS
FEM (II)
EQPM-0202-142 Vol. 67, p. 15567,

or PQ-200A-

impactor; PQ200A is a portable audit sampler, similar in design to PQ-200,


04/02/02

VSCC

but more compact in nature.



R&P
R&P-2000
PM10
R&P Partisol FRM Model 2000 PMio sampler with louvered PMio inlet;
operates at flow rate of 16.7 L/min; 24-h integrated sampler; uses a mass
flow meter to adjust to equivalent volumetric flow at ambient temperature
and pressure; single-channel sampler.
FRM
RFPS-1298-126
Vol. 63, p. 69625,
12/17/98
R&P
R&P-2000
PM2.B
R&P Partisol FRM Model 2000 PM2.6 sampler, identical to PMio sampler, but
uses a WINS impactor downstream of the PMio inlet for PM2.6 fractionation
FRM
RFPS-0498-117
Vol. 63, p. 18911,
04/16/98
R&P
R&P2000A
PM2.6
at 16.7 L/min; 24-h integrated sampler; R&P2000A is a portable audit
sampler.
FRM
RFPS-0499-129
Vol. 64, p. 19153,
04/19/99
R&P
R&P-2025
PM10
R&P Partisol-Plus Model 2025 PMio sequential sampler with louvered PMio
inlet; operates at 16.7 L/min; 24-h integrated sampler; uses a mass flow
meter to adjust to equivalent volumetric flow at ambient temperature and
pressure; sequential sampler with a capacity of 16 filter cassettes, allowing
for two weeks of unattended daily sampling; filter exchange is performed
pneumatically.
FRM
RFPS-1298-127
Vol. 63, p. 69625,
12/17/98
R&P
R&P-2025
PM2.B
R&P Partisol-Plus Model 2025 PM2.6 sequential sampler, identical to R&P-
2025 PMio sampler, but uses a WINS impactor downstream of the PMio inlet
for PM2.6 fractionation at 16.7 L/min.
FRM
RFPS-0498-118
Vol. 63, p. 18911,
04/16/98
R&P
R&P2000-
PM2.B
Same as R&P-2000 PM2.6 sampler, but with BGI VSCC, instead of WINS
FEM (II)
EQPM-0202-143 Vol. 67, p. 15567,

VSCC

impactor for PM2.6 separation.


04/02/02
R&P
R&P2000A-
PM2.B
Same as R&P-2000A PM2.6 sampler, but with BGI VSCC instead of WINS
FEM (II)
EQPM-0202-144 Vol. 67, p. 5567,

VSCC

impactor for PM2.6 separation.


04/02/02
R&P
R&P-2025-
PM2.B
Same as R&P-2025 PM2.6 sampler, but with BGI VSCC instead of WINS
FEM (II)
EQPM-0202-145 Vol. 67, p. 15567,

VSCC

impactor, for PM2.6 separation.


04/02/02
Andersen
RAAS-100
PM10
Andersen Instruments, Inc. Model RAAS10-100 PMio sampler with louvered
PMio inlet; operates at flow rate of 16.7 L/min; 24-h integrated sampler;
volumetric flow measured by dry test meter at pump outlet modulates pump
speed to maintain flow rate; single-channel.
FRM
RFPS-0699-130
Vol. 64, p. 33481,
06/23/99
Andersen
RAAS-100
PM2.B
Graseby Andersen Model RAAS2.5-100 PM2.6 sampler, similar to RAAS-100
PMio with a WINS impactor for PM2.6 separation.
FRM
RFPS-0598-119
Vol 63, p. 31991,
06/11/98
Andersen
RAAS200A
PM10
Andersen Instruments, Inc. Model RAAS10-200 and RAAS2.5-100 Audit
Samplers, portable compact version; similar to RAAS-100.
FRM
RFPS-0699-131
Vol. 64, p. 33481,
06/23/99
Andersen
RAAS-200A
PM2.B

FRM
RFPS-0299-128
Vol. 64, p. 12167,
03/11/99
Andersen
RAAS-300
PM10
Andersen Instruments, Inc. Model RAAS10-300, sequential sampler with
louvered PMio inlet, operates at 16.7 L/min; capacity to hold eight
filter-holders for multiple day operation.
FRM
RFPS-0699-132
Vol. 64, p. 33481,
06/23/99
July 2009
A-5
DRAFT-DO NOT CITE OR QUOTE

-------
Manufacturer3
Sampler
Name
Size
Cut"
Description
FRM or
FEMC
Designation# FRN
Andersen
RAAS-300
PM2.6
Graseby Andersen Model RAAS2.5-300 PM2.6 sampler, similar to RAAS-300
PM10 sampler with a WINS impactor for PM2.6 separation.
FRM
RFPS-0598-120 Vol. 63, p. 31991,
06111198
Thermo Scientific,
Inc.
CAPS
PM2.6
Model 605 Computer Assisted Particle Sampler (CAPS), 24-h integrated. Not
available commercially.
FRM
RFPS-1098-123 Vol. 63, p. 8036,
10129198
Thermo Scientific,
Inc.
RAAS 100-
VSCC
PM2.6
Same as RAAS-100 PM2.6 sampler, but with BGI VSCC, instead of WINS
impactor.
FEM (II)
EQPM-0804-153 Vol. 69, p. 47924,
08106104
Thermo Scientific,
Inc.
RAAS 200-
VSCC
PM2.6
Same as RAAS-200 PM2.6 sampler, but with BGI VSCC instead of WINS
impactor.
FEM (II)
EQPM-0804-154 Vol. 69, p. 47924,
08106104
Thermo Scientific,
Inc.
RAAS 300-
VSCC
PM2.6
Same as RAAS-300 PM2.6 sampler, but with BGI VSCC instead of WINS
impactor.
FEM (II)
EQPM-0804-155 Vol. 69, p. 47925,
08106104
URG Corp.
MASS-100
PM2.6
Model MASS100 PM2.6 sampler with louvered PM10 inlet followed by WINS
impactor, operates at 16.7 L/min; 24-h integrated, volumetric flow measured
by dry test meter at pump outlet modulates pump speed to maintain flow
rate; single channel.
FRM
RFPS-0400-135 Vol. 65, p. 26603,
05108100
URG Corp.
MASS-300
PM2.6
Model MASS300 PM2.6 sampler with louvered PM10 inlet followed by WINS
impactor, operates at 16.7 L/min; 24-h integrated, sequential sampler with
circular tray holding six filters.
FRM
RFPS-0400-136 Vol. 65, p. 26603,
05108100
Tisch Environ-
mental, Inc.
TE-6070
HiVol
PM10
Model TE-6070 PM10 High-Volume Sampler, with TE-6001 PM10 size
selective inlet; 8" x 10" filter holder.
FRM
RFPS-0202-141 Vol. 67, p. 15566,
04102102
Met One
BAM
PM10
Models BAM 1020, GBAM 1020, BAM 1020-1, and GBAM 1020-1, with
BX-802 inlet; glass-fiber filter tape with 1-h filter change frequency.
FEM
EQPM-0798-122 Vol. 63, p. 41253,
08103198
8 BGI Inc.: BGI Incorporated, Waltham, MA. R&P: Rupprecht & Patashnick Company, Inc., Albany, NY, now Thermo Scientific, Inc., Franklin, MA. Andersen: Graseby Andersen, later Andersen
Instruments, Inc., Smyrna, GA, now Thermo Scientific, Inc., Franklin, MA. Thermo Environmental Instruments, Inc., now Thermo Scientific, Inc., Franklin, MA. URG Corp.: URG Corporation, Chapel
Hill, NC. Tisch Environmental, Inc., Cleves, OH. Met One Instruments, Inc., Grants Pass, OR
bThe efficiency of an inlet3 is determined by its 50% cut-point (d50, the diameter at which half of the particles penetrate through the inlet, while the other half is retained by the inlet) and
eFRM: Federal Reference Method; FEM: Federal Equivalent Method. Roman numeral within parenthesis indicates FEM class.
8Particle separation in WINS is achieved by means of a single-jet round nozzle with flow directed into an impaction reservoir. The impaction surface consists of a Gelman Type A/E glass-fiber filter
immersed in 1 mL of Dow Corning (Midland, Ml) 704 diffusion pump oil housed in a reservoir.
Note: The geometric standard deviation (GSD, which is an indicator of the sharpness of the separation, and is derived by the square root of the ratio of particle diameters at penetrations of 16%
and 84%, [d16/d84]0.5).
Source: Chow (1995, 077012); Watson and Chow (2001,157123).
Table A-3. Measurement and analytical specifications for filter analysis of mass, elements, ions, and
carbon.
Observable
Analytical
Accuracy'
Precision11
Minimum Detectable
Limit (MDL)
Interferences
Comparability
Data
Completeness
PM2.6 mass
± 5%'
± 10%4
0.04/yg/m3 cto
~1//gfm3d6'e
Electrostatic charges need to be Within 20% "
neutralized before measurement;
positive (e.g., 0C adsorption)
and negative artifacts (e.g.,
nitrate volatilization)
90 to 100% 0-7
July 2009
A-6
DRAFT-DO NOT CITE OR QUOTE

-------
Observable
Analytical
Accuracy'
Precision
Minimum Detectable
Limit (MDL)
Interferences
Comparability
Data
Completeness
Elements
± 2-5%4
± 10% 4
XRF: 0.4-30 ng/m3 0 8 PIXE: Volatile compounds may evap- 10 to 30% depending 90 to 100%h E
6-360 ng/m3dB ICP/MS:
0.004-25 nglm310 0.05-
11.7 ng/m39'" AAS: 0.02-
7.15 nglm312
orate from filters due to vacuum on species
in XRF and PIXE Potential
contamination during extraction
and incomplete extraction effi-
ciency for ICP-MS and AAS
Matrix interference and peak
overlap may occur on heavily
loaded samples.
Nitrate
± 6% with spiked
± 5 to 10% on repli-
0.06//g/m3 8 to 0.2//g/m3d
Subject to volatilization from
Within 35% and
85 to 100% 0-7

concentrations on
cate analysis4'13'14 co-
1,6,17
Teflon or quartz-fiber filters
probably greater4


Teflon4 and ± 1-14%
located precision ± 5-





on nylon filters13
7%,4',e




Sulfate
±5% 4
±6 to 10% V4'16
0.06 //g/m3" to 0.2 //g/m3 d
1,6,13
n/a
Typically within 10%;
MOUDIs13 to 20%
lower than speciation
samplers4,17,0
85 to 100%
6,7,26,21
Ammonium
±5% 4
± 10% 4
0.06 //g/m3 0 to
0.07//g/m3 d,'e
Subject to volatilization from
Teflon or quartz-fiber filters
Within 30% 4
86 to 100% e'7
OC, EC,TC
± 5% for TC and
OC. No standard
OC: ± 20%
OC: 0.1 //g/m3fto
0.8//g/m3 d
Subject to adsorption (positive
artifact) and volatilization
OC: Within 20 to 50%
86 to 100%8-7

exists to determine
EC accuracy
EC: ± 20%
EC: 0.03//g/m3"to
0.1 //g/m3f
(negative artifact) of organic
gases to and from quartz-fiber
filters
EC: Within 20 to
200%



TC: ± 10% 4
TC: 0.8//g/m3d,'e

TC: Within 20% 4'17'22

Total mass of
DRI Model 2001
DRI Model 2001
DRI Model 2001 Carbon
Extraction efficiency and volume
Within 17% 20
n/a
WSOC
Carbon
Analyzer: ± 5%23
TOA: ± 3-7%24,26
Carbon
Analyzer: ± 10%23
Sunset Carbon
Analyzer: ± 3%2e
TOA: ±5-10% 27
Analyzer: 0.1 ¦ 0.23//g
C/m323
Sunset Carbon Analyzer:
0.05-0.22//g C/m3 20'28
Elemental High TOC II:
0.05//g C/m320
TOA: 0.12//g C/m3 20
reduction steps


Elements in
carbon: 1.5%;
± 2%30
carbon: 0.3//g/m3
Contamination during sample
n/a
n/a
water soluble
hydrogen: 3%;

hydrogen:0.09 //g/m3
drying step


matter: carbon,
nitrogen: 3%; sulfur:

nitrogen: 0.03 //g/m3



hydrogen,
5% 30

sulfur: 0.10 //g/m3 30



nitrogen, and






sulfur






Dissolved organic
n/a
± 5-30%31
0.001 jjg N/m3 while
Concentration of inorganic
Good correlation
n/a
nitrogen


inorganic nitrogen is low; a
0.071 jjg N/m3 while
inorganic nitrogen is high31
nitrogen
between UV and
persulfate oxidation
methods (R2 - 0.87)
31

Neutral polyols
GC/MS: ± 4-8% 32
GC/MS: ± 23% 33'34
GC/MS: Levoglucosan: 10
GCMS: Extraction recovery
IC/PAD: Good
n/a
and polyether

Typically ± 20%,
ng/m340
interfered by sample matrix
correlation



ranged from ± 10
to ± 30%i32'36'30'37'38
HPLC/MS: ± 5-26%39
2.08 ng/m3'31
0.01-0.03 ng/m3 33'41
HPLC/MS: 9-648 pg/m330
Derivatization efficiency
IC/PAD: Overlapping peaks in
chromatogram
(R2 - 0.97) with
HPLC/MS: and
(R2 - 0.89) with
GC/MS Method42

Mono- and
Di-carboxylic
acids
n/a	GC/MS: ±5-11% on GC/MS: 0.04-1.12 ng/m3
,in 10:0.01-0.12 nglm347
3 replicates, ± 8 [
avg43,44
IC: ± 10-15%
45
GC/MS: Extraction recovery
interfered by sample matrix
Derivatization efficiency
IC: Overlapping peaks in
chromatogram
GC/MS: Within 50%
for less volatile
compounds40
n/a
July 2009
A-7
DRAFT-DO NOT CITE OR QUOTE

-------
Analytical n . . h Minimum Detectable . „ .Data
Observable . a Precision .. Interferences Comparability
	Accuracy	Limit (MDL)			 Completeness
Amino acids	n/a	±9%48	1.65-23.6	Derivatization efficiency	n/a	n/a
P9'm	Stability of derivatives
Overlapping peaks in
chromatogram
Mass of humic-	n/a	n/a	0.083 ng/m3'49	Separation efficiency	n/a	n/a
like substances
(HULIS)
8	Accuracy is the ability of analytical methods to quantify the observable of a standard reference material correctly; it does not refer to measurement accuracy if no standards available.50
b Refers to precision of co-located measurements, unless specified otherwise
e Based on 1 /vg/filter limit of detection for 24-h samples, assuming a flow rate of 16.7 L/min
d Based on field blanks collected with FRM samplers; /yg/filter converted to /yg/m3 basis assuming a flow rate of 16.7 L/min for 24-h
8Based on h of a 47-mm filter extracted in 15 mL deionized-distilled water (DDW) for 24-h samples, assuming a flow rate of 16.7 L/min
1 Based on 0.2 /yg/cm2 detection limit and 13.8 cm2 deposit area for a 47-mm filter, assuming a flow rate of 16.7 L/min for 24-h
9	Based on 24-h samples at a flow rate of 16.7 L/min and analyzed by XRF
h Except for samples from one FRM sampler at Atlanta Supersite, for which data recovery was 50%7; reason not reported.
'Reported as uncertainty in literature
1 Based on 24-h samples at a flow rate of 16.7 L/min
k Based on 13.8 cm2 deposit area for a 47-mm filter and extracted into a final volume of 200/yL, assuming a flow rate of 16.7 L/min for 24-h and molecular weight of amino acid = 150
' Based on 13.8 cm2 deposit area for a 47-mm filter and extracted into a final volume of 200/yL, assuming a flow rate of 16.7 L/min for 24-h
n/a: Not available
'Chow (1995,077012); 2Watson and Chow (2001,157123);3 Watson et al. (1983, 045084); 4Fehsenfeld et al. (2004,156432); 'Solomon et al. (2001,156993); GMikel (2001,156762); 7Mikel
(2001,156762); 8Watson et al. (1999, 020949); 9Solomon and Sioutas (2006,156995); ,0Graney et al. (2004,053756); "Tanaka et al. (1998,157041); ,2Pancras et al. (2005, 098120); ,3John
et al. (1988,045903); "Hering and Cass (1999, 084958); ,5Fritz et al. (1989,077387); 1GHering et al. (1988,036012); "Solomon et al. (2003,156994); ,8Cabada et al. (2004,148859); 19Fine et
al. (2003,155775); 20Hogrefe et al. (2004, 099003); 2,Drewnick et al. (2003, 099160); 22Watson et al. (2005,157125); 23Ho et al. (2006,156552); 24Decesari et al. (2005,144536); 25Mayol-
Bracero et al. (2002, 045010); 2GYang et al. (2003,156167);21 Tursic et al. (2006,157063); 28Mader et al.(2004,156724); 29Xiao and Liu (2004, 056801); 30Kiss et al. (2002,156646); 3,Cornell
and Jickells (1999,156367);32 Zheng et al. (2002, 026100); 33Fraser et al. (2002,140741);34 Fraser et al. (2003,042231) 35Schauer er al. (1996, 051162); 30Fine et al. (2004); 3JYue et al.
(2004); 38Rinehart et al. (2006,115184); 39Wan and Yu (2006,157104); 40Poore (2000, 012839); 4,Fraser et al. (2003a); 42Engling et al. (2006,156422); 43Yu et al. (2005); 44Tran et al. (2000);
45Yao et al. (2004,102213); 4GLi and Yu (2005,156692); Penning et al. (2003,156539); 48Zhang and Anastasio (2003,157182); 49Emmenegger et al. (2007,156418); 50Watson et al. (1989,
157119)
Source: Chow et al. (2008,156355
Table A-4. Measurement and analytical specifications for filter analysis of organic species.
Organic
Species
Analytical
Accuracy
Precision
MDL
Interferences
TD Solvent
Extraction
TD
Solvent
Extraction
TD
Solvent
Extraction
TD
Solvent
Extraction
Comparability
PAHs
± 2.8-
24.1 %B1
±4.4-
29.4%62
13.8-
26.5%63
± 0.5-
12.9%64
0.05-
4.83%66
Z-score
values 0 to -
1.9 60
± 4-8%32
± 6.5-
22%"
Avg ± 3.2%,	Avg ± 8%,
ranged	ranged from ± 3.8
from ± 0.05 to	to ± 15%6e
±11.5%55	+ 23%60 Avg
± 2.6%, ranged
from ± 0.6 to
± 9.5%67
typically
± 20%, ranged
from ± 10 to
± 30%c32*37
0.016-0.48
ng/m3 8 68
0.030-0.45
ng/m3 8 66
0.83-1.66
ng/m3b38
0.033-3.85
ng/m3 b 60
0.01-0.03
ng m ' '
0.76-276
pgfm3b67
Fragmentation of
labile com-
pounds
Possible contami-
nants from solvents
and complicated
extraction
procedures Loss of
volatile compounds
during the ex-
traction and pre-
treatment steps
Possible carryover
from injection port
R2s for solvent
extraction were
0.9 5 68, 0.9 7 66,
and 0.98 60
n-Alkanes
n/a
± 4-8
Avg ± 3.2%,
ranged
from ± 0.05
to ± 11.5%66
± 23%60
Typically ± 20%,
from ± 10
to ± 30%c 32,36 37
0.081-0.86
ng/m3 8 68
0.061-0.97
ng/m3 8 66
0.01-0.03
.3 33,34,37
Same as PAHs Same as PAHs
ng/m3
R2s for solvent
extraction are 0.94
68, and 0.98 66'60
July 2009
A-8
DRAFT-DO NOT CITE OR QUOTE

-------
Analytical
Accuracy
Precision
MDL
Interferences
Hopanes
n/a
n/a
Avg ± 3.2%,
ranged
from ± 0.05
to ± 11.5%66
± 23%6e
Typically ± 20%,
from ± 10
to ± 30%c 32,36 37
0.030-0.14
ng/m3a66
0.83-1.66
ng/m3b38
0.01-0.03
ng/m3 33,41
0.01 ng/m337
Same as PAHs
Same as PAHs
R2s for solvent
extraction are 0.99
66 and 0.998 60
Steranes
n/a
n/a
Avg ± 3.2%,
n/a
0.018-
0.83-1.66
Same as PAHs
Same as PAHs
R2s for solvent



ranged

0.063
ng/m3 b 00


extraction are 0.97



from ± 0.05

ng/m3a66



66 and 0.998 60
to ± 11.5%61
Organic acids
(including n-
alkanoic acids, n-
alkenoic acids,
alkane dicarboxylic
acids, aromatic
carboxylic acids,
resin acids)
n/a
±4-8
± 10
to ± 29%66
± 24%41
± 23%6e
Typically ± 20%,
from ± 10
to ± 30%c 32,36 37
Mono-
carboxylic
acids (C8,
C12, and
C16):
0.79, 2.0,
and 3.2
ng/m3 8 64
0.01-0.03
nq/m3 33,41
Fragmentation of
labile compounds
Loss of polar
species due to
absorption onto
the surface of the
injector
Improper station-
ary phase column
used during TD
analysis
Incomplete ther-
mal desorption of
analytes because
of strong affinity
with filter matrix
Possible conta-
minants from sol-
vents and com-
plicated extraction
procedures
Loss of volatile
compounds during
the extraction and
pretreatment steps
Possible carryover
from injection port
Low derivatization
efficiency
Correlation with
solvent extraction
method
R2 - 0.731 60
Polyols and sugars.
n/a
± 4-8%32
n/a
± 23%56
n/a
Levoglucosa: Same as organic
Same as organic
n/a
including guaiacol
and substituted



Typically ± 20%,

10 ng/m3 01 acids
acids

guaiacols, syringol



from ± 10

2.08 ng/m3b


and substituted



O
l+
GO
O




syringols.





0.01-0.03


anhydrosugars





ng/m3 33,41


8 Assumes 2.9 cm2 filter used in analysis from a deposit area of 13.8 cm2, and sample collection at a flow rate of 16.7 Llmin for 24-h
b Assumes sample collection at a flow rate of 16.7 Llmin for 24-h.
e Reported as uncertainty in literature.
d Assumes a final extract volume of 1 mL and sample collection at a flow rate of 16.7 Llmin for 24-h. nla: Not available
Source: 'Chow (1995,077012); 2Watson and Chow (2001,157123);3 Watson et al. (1983, 045084); "Fehsenfeld et al. (2004,156432); 'Solomon et al. (2001,156993); sMikel (2001,156762);
'Mikel (2001,156762); "Watson et al. (1999,020949); 'Solomon and Sioutas (2006,156995); '"Graney et al. (2004, 053756); "Tanaka et al. (1998,157041); "Pancras et al. (2005,098120);
"John et al. (1988,045903); "Hering and Cass (1999, 084958); "Fritz et al. (1989, 077387); '"Hering et al. (1988, 036012); "Solomon et al. (2003,156994); "Cabada et al. (2004,148859);
"Fine et al. (2003,155775); 20Hogrefe et al. (2004, 099003); 2,Drewnick et al. (2003, 099160); 22Watson et al. (2005,157125); 23Ho et al. (2006,156552); 24Decesari et al. (2005,144536);2S
Mayol-Bracero et al. (2002, 045010); 2sYang et al. 2003;21 Tursic et al. (2006,157063); 2"Mader et al.(2004,156724); 29Xiao and Liu (2004, 056801); 3°Kiss et al. (2002,156646); ''Cornell and
Jickells (1999,156367);32 Zheng et al. (2002, 026100); 33Fraser et al. (2002,140741);34 Fraser et al. (2003, 042231) 3sSchauer er al. (1996, 051162); 3sFine et al. (2004); 3,Yue et al. (2004);
3BRinehart et al. (2006,115184); 39Wan and Yu (2006,157104); 40Poore (2000, 012839); "Fraser et al. (2003a); 42Engling et al. (2006,156422); 43Yu et al. (2005); 44Tran et al. (2000); 4sYao et al.
(2004,102213); 4sLi and Yu (2005,156692); "Henning et al. (2003,156539); 4"Zhang and Anastasio (2003,157182); "Enmenegger et al. (2007,156418); s0(Watson et al., 1989,157119);
"Greaves et al. (1985,156494); "Waterman et al. (2000,157116); s3Waterman et al. (2001,157117); s4Falkovich and Rudich (2001,156427); ssChow et al. (2007,156354); "Miguel et al.
(2004,123260); "Crinmins and Baker (2006, 097008); s"Ho and Yu (2004,156551); s9Jeon et al. (2001,016636); ""Mazzoleni et al. (2007, 098038); s,Poore (2000, 012839); ^Butler et al.
(2003);	G3Chow et al. (2006c); "Russell et al. (2004, 082453); ssGrover et al. (2006); "Graver et al. (2005, 090044); "Schwab et al. (2006b); ""Hauck et al. (2004); s9Jaques et al. (2004,
155878); '"Rupprecht and Patashrick (2003,157207); "Pang et al (2002, 030353); "Eatough et al. (2001,010303); ,3Lee et al. (2005,128139); ,4Lee et al. (2005,156680); ,sBabich et al.
(2000,156239);; ,sLee et al. (2005c); "Lee et al. (2005,128139); "Anderson and Ogren (1998,156213); ,9Chung et al. (2001,156357); ""Kidwell and Ondov (2004,155898); "'Litt^ow et al.
(2004,126616); ""Weber et al. (2003,157129); "'Harrison et al. (2004); "Rattigan et al. (2006,115897); "sWittig et al. (2004,103413); ""Vaughn et al. (2005,157089); "'Chow et al. (2005,
099030); ""Weber et al (2001,024640); ""Schwab et al. (2006a); "°Lim et al. (2003,156697); "'Watson and Chow (2002, 037873); ""Venkatachari et al. (2006,105918); '"Bae et al. (2004a);
"Arhami et al. (2006,156224); "'Park et al. (2005,156843); ""Bae et al. (2004b); "Chow et al. (2006a); ""Arnott et al. (2005); ""Bond et al. (1999); '""Virkkula et al. (2005,157097); '"'Petzold et
al. (2002,156863); ™Park et al. (2006); "3Arnott et al. (1999, 020650); '"Peters et al. (2001); '"'Pitchford et al. (1997,156872); '""Rees et al. (2004, 097164); '"'Watson et al. (2000); '""Lee et
al. (2005,156680); '""Hering et al. (2004,155837); ""Watson et al. (1998); "'Chakrabarti et al. (2004); "2Mathai et al. (1990,156741); '"Kidwell and Ondov (2001, 017092); '"Stanier et al.
(2004);	'"Khlystov et al. (2005,156635); ""Takahama et al. (2004,157038); "'Chow et al. (2005,156348); ""Zhang et al. (2002); ""Subramanian et al. (2004, 081203); ,2°Chow et al. (2006b);
"'Birch and Cary (1996); ,22Birch (1998, 024953); ,23Birch and Cary (1996); ""NIOSH (1996,156810); "sNI0SH (1999,156811); ""Chow et al. (1993); "'Chow et al. (2007); ""Ellis and Novakov
(1982,156416); ""Peterson and Richards (2002,156861); ""Schauer et al. (2003); "'Middlebrook et al. (2003, 042932); ,32Wenzel et al. (2003,157139); "3Jimenez et al. (2003); ,34Phares et al.
(2003,156866); "sQin and Prather (2006,156895); ""Zhang et al. (2005); "'Bein et al. (2005,156265); ""Drewnick et al. (2004a); ""Drewnick et al. (2004b); ""Lake et al. (2003,156669);
"'Lake et al. (2004,088411)
July 2009
A-9
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-5. Measurement and analytical specifications for continuous mass and mass surrogate
instruments.
Instrument and
Measurement Principle
Averaging
Time
Analytical „ . . h
Precision
Accuracy
MDL
Data
Interferences Comparability „ , .
Completeness
INERTIA INSTRUMENTS
TEOM Air is drawn through a size- 10 min-24
selective inlet onto the filter mounted h
on an oscillating hollow tube. The
oscillation frequency changes with
mass loading on the filter, which is
used to calculate mass concentration
by calibrating measured frequency
with standards.
loss of semi-
volatile species
± 0.75%c ± 5//g/m for 0.01//g, which is Loses semi-volatile Underestimated 99% to 92%
10-minavgcd 0.06//g/m3for 1-h species at both FRM mass by 20 6
±1.5//g/m3for av9c	30°Cand50°C. to 35% 02 «
1-h avgc-'
SESTE0M, while
less sensitive to
relative humidity,
does not com-
Dletelv eliminate
FDMSTE0M. A self-referencing 1-h - 24 h ± 0.75%c < 1O%06
TEOM with a filter at 4 °C that
accounts for volatile species. It is
equipped with a diffusion Nafion
dryer to remove particle-bound water.
The Teflon (PTFE)-coated borosilicate
glass-fiber filter that is maintained at
4 °C removes particles during the
reference flow cycle. The flow
alternates between a base and refer-
ence flow every 6 min. If a negative
mass is measured during the
reference flow, due to loss of
volatiles from the filter, it is added to
the mass made during the prior parti-
cle-laden samples to obtain total
PM2.6 concentration.
0.01 //g, which is	n/a	9 to 30% higher 95 to 99%06'08
0.06//g/m3	than FRM mass 57 to 65%07
for 1-h avgc	Within 10% of
mass by D- TEOM,
PC- BOSS, RAMS
and BAM ee'07
Differential Tapered Element	1-h - 24 h ± 0.75%c < 10%Be6'e9'70 0.01 //g, or	n/a	Within 10% of 86%®
Oscillating Microbalance (D-TEOM)	0.06//g/m3 for 1-h	FDMS-TEOM 06 00
Similar to FDMS, but an electrostatic	av9
precipitator is used in place of the
glass-fiber filter to remove particles
during the 6 min reference flow cycle.
RAMS	10 min-24
A TEOM with a cyclone inlet, dif- ^
fusion denuders, and Nafion dryer.
Particles are collected on a "sand-
wich" filter (Teflon followed by car-
bon-impregnated glass-fiber filter) on
the tapered oscillating element. The
various denuders remove gas phase
organic compounds, nitric acid, sulfur
dioxide, nitrogen dioxide, ammonia,
and ozone, which could otherwise be
adsorbed by the TEOM filter.
n/a < 10%f71 ± 1 to 2 //g/m3 for
30-min avg 72
n/a	10 to 20% higher	n/a
than avg®72FRM
mass73'74
July 2009
A-10
DRAFT-DO NOT CITE OR QUOTE

-------
Instrument and
Measurement Principle
Averaging
Time
Analytical „ . . h
Precision
Accuracy
MDL
Data
Interferences Comparability „ , .
Completeness
PRESSURE DROP INSTRUMENT
Continuous Ambient Mass Monitor 1 -h — 24 h
(CAMM)
Air is drawn through a Teflon-
membrane filter tape and the pressure
drop across the filter is monitored
continuously. The proportion of
pressure drop to aerosol loading is
related to the PM concentration. The
filter tape advances every 30-60 min
to minimize volatilization and ad-
sorption artifacts during sampling.
n/a
28.1% for 1-h
avg
15.9% for 24-h
avg
(~ 3.5 //g/m3)
< 5 //g/m3 for 1 h
avg 76
Needs effective
sealing for good
performance; even
slight leaks may
result in highly
variable baseline.
Probably less
sensitive than D-
TEOMor RAMS.
Varied perfor-
mance: within 2%
of SES-TEOMand
FRM at Houston,
TX, while not
correlated with D-
TEOMor FRM at
Rubidoux, OA.76"
n/a
BA TTENUA TION INSTRUMENT
B Attenuation Monitor (BAM)
B rays electrons) are passed through
a quartz-fiber filter tape on which
particles are collected. The loss of
electrons (B attenuation) caused by
the particle loading on the filter is
converted to mass concentration,
after subtraction of blank filter
attenuation.
1-h-24 h
± 3 //g for ±2//g/m3c'h
5/yg/m3 for 1-h avg1 Water absorption
Up to 30% higher
24-h avg
by particles may
than FRM mass
concentrations
result in higher
and within 2% of
< 100/yg/m3
mass measure-
FDMS-TEOM 03'07
and 2% for 100
ments; maybe

to 1,000/yg/m3
important at RH

± 8/yg
>85%

< 100/yg/m3


and 8% for 100


to 1000/yg/m3


(1-h


93 to 99%e'e6'07
LIGHT SCA TTERING INSTRUMENT
Nephelometers (including DustTrak)
A light source illuminates the sample
air and the scattered light is detected
at an angle (usually 90°) relative to
the source. The signal is related to
the concentration of the particles
giving an estimate of the particle light
scattering coefficient. Zero air
calibrations can be performed using
particle-free air.
5 min - 24
h
n/a
Nephelometers:
< 5% for TSI
and NGNi
nephelometers
78,79
DustTrak:
Greater of 0.1%
or 1 /yg/m3c'h
Nephelometer:
< 1.5 Mm-1
DustTrak: ± 1 //g|m3
for 24-h avg'
Conversion factor
to calculate mass
concentration from
bscat may vary
depending on
particle size, shape
and composition.
Light scattering by
DustTrak propor-
tional to dp 6 for
dp < 0.25/ym 70
Typically good
correlation with
SES-TEOM and D-
TEOM (R2
>0.80).
Comparability
depends on
conversion factor
used.
>80 to 98% for
NGN2, RR-M903
and GreenTek
Nephelometers6
>80% for
DustTrak 006 to
98% for GRIMM
optical particle
counter06
July 2009
A-11
DRAFT-DO NOT CITE OR QUOTE

-------
Instrument and Averaging Analytical n . . h „ .... Data
. „ . . , . a Precision MDL Interferences Comparability -
Measurement Principle	Time Accuracy	_	 Completeness
a Accuracy is the ability of analytical methods to quantify the observable of a standard reference material correctly; does not refer to measurement accuracy, since no standards available,
b Refers to precision of co-located measurements, unless specified otherwise,
c Manufacturer-specified measurement parameter.
d Details not available on how the precision was obtained and whether it refers to co-located precision,
e Includes a combination of estimates: based on co-located precision and based on regression slopes,
f Co-located precision with respect to PC-BOSS reconstructed PM2.5 mass,
g Using glass-fiber "sandwich" filter,
h Specified as "resolution" by the manufacturer.
i Co-located precision estimate based on regression slope for NGN nephelometer (slope = 1.01, intercept = -1.04/yglm3, R2 = 0.99).
j Specified as "Zero stability" by the manufacturer,
nla: Not available.
Source: 'Chow (1995,077012): 2Watson and Chow (2001,157123);3 Watson et al. (1983, 045084); "Fehsenfeld et al. (2004,150432); 'Solomon et al. (2001,150993); sMikel (2001,150702);
'Mikel (2001,150702); "Watson et al. (1999,020949); 'Solomon and Sioutas (2000,150995); "Graney et al. (2004, 053750); "Tanaka et al. (1998,157041); "Pancras et al. (2005,098120);
"John et al. (1988,045903); "Hering and Cass (1999, 084958); "Fritz et al. (1989, 077387); '"Hering et al. (1988, 030012); "Solomon et al. (2003,150994); "Cabada et al. (2004,148859);
"Fine et al. (2003,155775); 2°Hogrefe et al. (2004, 099003); 2,Drewnick et al. (2003, 099100); 22Watson et al. (2005,157125); 23Ho et al. (2000,150552); 24Decesari et al. (2005,144530);2S
Mayol-Bracero et al. (2002, 045010); 2sYang et al. 2003;21 Tursic et al. (2000,157003); 2"Mader et al.(2004,150724); 29Xiao and Liu (2004, 050801); 30Kiss et al. (2002,150040); 3,Cornell and
Jickells (1999,150307);32 Zheng et al. (2002, 020100); 33Fraser et al. (2002,140741);34 Fraser et al. (2003, 042231) 3sSchauer er al. (1990, 051102); 3sFine et al. (2004); 3,Yue et al. (2004);
3BRinehart et al. (2000,115184); 39Wan and Yu (2000,157104); 40Poore (2000, 012839); "Fraser et al. (2003a); 42Engling et al. (2000,150422); 43Yu et al. (2005); 44Tran et al. (2000); 4sYao et al.
(2004,102213); 4sLi and Yu (2005,150092); "Henning et al. (2003,150539); 4"Zhang and Anastasio (2003,157182); "Enmenegger et al. (2007,150418); s0(Watson et al., 1989,157119);
"Greaves et al. (1985,150494); "Waterman et al. (2000,157110); s3Waterman et al. (2001,157117); s4Falkovich and Rudich (2001,150427); ssChow et al. (2007,150354); "Miguel et al.
(2004,123200); "Crinmins and Baker (2000, 097008); s"Ho and Yu (2004,150551); s9Jeon et al. (2001,010030); 60Mazzoleni et al. (2007, 098038); 6,Poore (2000, 012839); ^Butler et al.
(2003); s3Chow et al. (2000c); "Russell et al. (2004, 082453); ssGrover et al. (2000); ssGrover et al. (2005, 090044); "Schwab et al. (2000b); ""Hauck et al. (2004); s9Jaques et al. (2004,
155878); '"Rupprecht and Patashrick (2003,157207); "Pang et al (2002, 030353); "Eatough et al. (2001,010303); ,3Lee et al. (2005,128139); ,4Lee et al. (2005,150080); ,sBabich et al.
(2000,150239); ,sLee et al. (2005c); "Lee et al. (2005,128139); '"Anderson and Ogren (1998,150213); ,9Chung et al. (2001,150357); ""Kidwell and Ondov (2004,155898); "'Lithgow et al.
(2004,120010); ""Weber et al. (2003,157129); "'Harrison et al. (2004); "Rattigan et al. (2000,115897); "sWittig et al. (2004,103413); ""Vaughn et al. (2005,157089); "'Chow et al. (2005,
099030); ""Weber et al (2001,024040); ""Schwab et al. (2000a); "°Lim et al. (2003,150097); "'Watson and Chow (2002,037873); ""Venkatachari et al. (2000,105918); "3Bae et al. (2004a);
"Arhami et al. (2000,150224); "'Park et al. (2005,150843).
Table A-6. Measurement and analytical specifications for continuous elements.
Instrument and
Measurement Averaging Time
Principle	
Analytical
Accuracy3
Precision
MDL
Interferences Comparability
Data
Completeness
Semi-continuous
Elements in Aerosol
System (SEAS)
Particles are
collected at 30-min
interval for
subsequent
laboratory atomic
absorption analysis
for elements.
Aerosol collection is
through
condensational
growth by direct
steam injection.
The grown particles
are separated from
the airstream using
virtual impactor.
The droplets
accumulate in a
slurry that is
pumped to a
separate sample
vial for each time
period.
15-30 min
± 10%b for Mn, 20 to 43%c8°
Al: 440 pg
Spectral
Fe, Ni, Cu, Zn, Se,
Cr: 6.7 pg
interferences limit
Cd, and Sb
Mn: 9.9 pg
the number of
± 20%bfor Cr, As,
Fe: 85 pg
elements detected
and Pb 80
Ni: 42 pg
simultaneously
Cu: 26 pg


Zn: 43 pg


As: 27 pg


Se: 33 pg


Cd: 3.2 pg


Sb 160 pg


Pb: 31 pg80

n; a
n/a
July 2009
A-12
DRAFT-DO NOT CITE OR QUOTE

-------
Instrument and
Measurement Averaging Time
Principle	
Analytical
Accuracy3
Precision
MDL
Interferences Comparability
Data
Completeness
Laser-Induced
Breakdown
Spectroscopy
(LIBS)
Used for in-situ
single particle
analysis. A high-
power pulsed laser
is projected into
particles producing
high-temperature
plasma. Photons
emission from
relaxing atoms in
the excited states
provides
characteristics of
individual elements.
A few seconds
n/a
n/a
Na: 143 fg
Mg: 53 fg
Al: 184 fg
Ca: 50 fg
Cr: 166 fg
Mn: 176 fg
Cu: 15 fg81
n/a
n/a
n/a
8 Accuracy is the ability of analytical methods to quantify the observable of a standaid reference material correctly; does not refer to measurement accuracy, since no standards are available.
b Based on analysis of standard reference material (SRM) 1643d from National Institute of Standards and Technology (NIST).
c Based on error propagation,
n/a: Not available
Source: ""(Kidwell and Ondov, 2004,155898); "'(Lithgow et al., 2004,126616).
Table A-7. Measurement and analytical specifications for continuous NO3.
Instrument and Measurement Averaging Analytical
	Principle	Time Accuracy
Precision
MDL
Interferences Comparability
Data
Completion
FLASH VOLA TIZA TION INSTRUMENTS
Aerosol Dynamics Inc. continuous nitrate
analyzer (ADIN)
Particle collection by humidification and
impaction followed by flash volatilization and
detection of the evolved gases in a
chemiluminescent NOx analyzer.
10 min
n/a
n/a
0.1 /yg/m3 for
10-min avg 82
n/a
Within 30% of filter
and continuous NO3 -
. See Weber et al.82
for details.
93%7
Rupprecht and Patashnick continuous nitrate
analyzer (R&P-8400N)
Particle collection by impaction followed by
flash volatilization and detection of the
evolved gases in a chemiluminescent NOx
analyzer. A carbon honeycomb denuder,
installed at the inlet to the Nafion humidifier
removes nitric acid and ammonia vapor.
10 min
n/a
6.3%-23%b
0.17 to
0.3 /yg/m3for
24-h avg 83'84
0.24/yg/m3 to
0.45/yg/m3 for
10-min avg
Conversion and
volatilization efficiency
appears to depend on
ambient composition;
extent of underestimation
increases with higher
concentrations.84,80
20 to 45% lower
than filter NOs"20'82"
85,87
>80 to
>94%w
July 2009
A-13
DRAFT-DO NOT CITE OR QUOTE

-------
Instrument and Measurement Averaging	Analytical „ Data
„ . . .	. Precision MDL Interferences Comparability	„ , ..
	Principle	Time	Accuracy	_		Completion
DENUDERDIFFERENCE INSTRUMENT
Atmospheric Research and Analysis nitrate
analyzer (ARAN)
Sampled air passes through a 350°C
molybdenum (Mo) mesh that converts
particulate nitrate into NO. A pre-split stream
with a Teflon filter installed upstream of an
identical converter (i.e., particle-free air) is
used as a reference. NO in both streams is
quantified by chemiluminescence and their
difference determines the particulate nitrate
concentration. The instrument inlet contains a
potassium iodide- coated denuder to remove
HNOs and NO?.
30 sec
n/a
n/a
0.5/yg/m3 for
30-sec avg 82
n/a
Within 30% of filter
and continuous NO3 -
. See Weber et al.82
for details.
76%7
SAMPLE DISSOLUTION FOLLOWED BY IC ANAL YSIS INSTRUMENTS
Energy Research Center of the Netherlands
(EON) 10-based ion analyzer Oollects particles
into water drops using a steam jet aerosol
collector, via cyclone. The combined flow from
collected droplets containing dissolved aerosol
components and wall steam condensate is
directed to an anion 10 for analysis of nitrate.
Interfering gases are pre-removed by a rotating
wet annular denuder system.
1-h
n/a
n/a
0.1//g/m3
n/a
Within 30% of filter
and continuous NO3 -
. See Weber et al.82
for details.
100%7
Texas Tech University (TT) ion analyzer
Particles in the sample stream are processed
through a cyclone and a parallel plate wet
denuder, then collected alternatively on one of
two 2.5 cm pre-washed glass fiber filters for a
period of 15 min. The particles on the freshly
sampled filter are automatically extracted for
6.5 min with water and analyzed for nitrate by
IC.
15-30 min n/a
n/a
0.010//g/m3 n/a
Within 30% of filter
and continuous NO3 -
. See Weber et al.82
for details.
97%7
Particle into Liquid Sampler-Ion	1h	n/a	10%-15%c 0.05 to	Consistent water quality Within 10% of nylon- 65 to 70%2°
Chromatography (PILS-IC)	7'82'88 0.1 //g/m3 is essential for good filter NO3 - and 37%
Ambient particles are mixed with saturated	2°'82'88	Precision-	higher than R&P-
o4nni\i 20
water vapor to produce droplets collected by
impaction. The resulting liquid stream is
analyzed with an IC to quantify aerosol ionic
components.
Dionex-IC The gas-denuded air stream enters 1-h	n/a	14%de6 n/a	Consistent water quality Bias of < 10% n/a
the annular channel of a concentric nozzle,	is essential for good relative to filter NO3
where deionized water generates a spray that	precision.	-0E
entrains the particles. The flow is then drawn
through a 0.5//m pore size PTFE filter. The
remaining solution is aspirated by a peristaltic
pump and sent to IC for ion analysis.
Ambient Ion Monitor (AIM; Model 9000) Air is 1-h	n/a	n/a	0.1 //g/m3 for n/a	n/a	n/a
drawn through a size-selective inlet into a li-	1-h avg9
quid diffusion denuder where interfering gases
are removed. The stream enters a super-
saturation chamber where the resulting
droplets are collected through impaction. The
collected particles and a fraction of the
condensed water are accumulated until the
particles can be injected into IC for hourly
analysis.
July 2009
A-14
DRAFT-DO NOT CITE OR QUOTE

-------
Instrument and Measurement Averaging Analytical „ Data
„ . . . . Precision MDL Interferences Comparability „ , ..
	Principle	Time Accuracy	_	 Completion
PARTICLE MASS SPECTROMETER INSTRUMENT
Aerosol Mass Spectrometer (AMS)	A few
Air stream is drawn through an aerodynamic secon[
lens and focused into a beam in a vacuum
chamber. This aerosol beam is chopped by a
mechanical chopper and the flight time of the
particles through a particle-sizing chamber is
determined by the time-resolved mass
spectrometer measurement. The particle
impacts onto a 600 °C heated plate where it
decomposes and is analyzed by a quadruple
mass spectrometer. The nitrate ion, along with
other ions, is detected by the mass
spectrometer.
8 Accuracy is the ability of analytical methods to quantify the observable of a standaid reference material correctly; does not refer to measurement accuracy, since no standards are available.
b Overall uncertainty estimated by error propagation.
c Uncertainty estimated from uncertainties in flow rates and calibrations; does not refer to co-located precision.
d Co-located precision with respect to PC-BOSS PM2.5 total particulate NO3 (the sum of the denuded front filter [non-volatilized NO3-] and HN03-absorbing backup filter [volatilized NO3]).
e Manufacturer specified measurement parameter
nla: Not available.
Source: 'Chow (1995,077012); 2Watson and Chow (2001,157123);3 Watson et al. (1983, 045084); "Fehsenfeld et al. (2004,150432); 'Solomon et al. (2001,150993); sMikel (2001,150702);
'Mikel (2001,150702); "Watson et al. (1999,020949); 'Solomon and Sioutas (2000,150995); '"Graney et al. (2004, 053750); "Tanaka et al. (1998,157041); "Pancras et al. (2005,098120);
"John et al. (1988,045903); "Hering and Cass (1999, 084958); "Fritz et al. (1989, 077387); '"Hering et al. (1988, 030012); "Solomon et al. (2003,150994); '"Cabada et al. (2004,148859);
"Fine et al. (2003,155775); 2"Hogrefe et al. (2004, 099003); 2'Drewnick et al. (2003, 099100); 22Watson et al. (2005,157125); 23Ho et al. (2000,150552); 24Decesari et al. (2005,144530);2S
Mayol-Bracero et al. (2002, 045010); 2syang et al. 2003;21 Tursic et al. (2000,157003); 2"Mader et al.(2004,150724); 29Xiao and Liu (2004, 050801); 3"Kiss et al. (2002,150040); ''Cornell and
Jickells (1999,150307);32 Zheng et al. (2002, 020100); 33Fraser et al. (2002,140741);34 Fraser et al. (2003, 042231) 3sSchauer er al. (1990, 051102); 3sFine et al. (2004); 3,yue et al. (2004);
3BRinehart et al. (2000,115184); 39Wan and yu (2000,157104); 40Poore (2000, 012839); "Fraser et al. (2003a); 42Engling et al. (2000,150422); 43yu et al. (2005); 44Tran et al. (2000); 4iTao et al.
(2004,102213); 4sLi and yu (2005,150092); "Henning et al. (2003,150539); 4"Zhang and Anastasio (2003,157182); "Enmenegger et al. (2007,150418); s0(Watson et al., 1989,157119);
"Greaves et al. (1985,150494); "Waterman et al. (2000,157110); s3Waterman et al. (2001,157117); s4Falkovich and Rudich (2001,150427); ssChow et al. (2007,150354); "Miguel et al.
(2004,123200); "Crinmins and Baker (2000, 097008); s"Ho and yu (2004,150551); s9Jeon et al. (2001,010030); 60Mazzoleni et al. (2007, 098038); 6,Poore (2000, 012839); ^Butler et al.
(2003);	G3Chow et al. (2000c); "Russell et al. (2004, 082453); ssGrover et al. (2000); "Graver et al. (2005, 090044); "Schwab et al. (2000b); ""Hauck et al. (2004); s9Jaques et al. (2004,
155878); '"Rupprecht and Patashrick (2003,157207); "Pang et al (2002, 030353); ,2Eatough et al. (2001,010303); ,3Lee et al. (2005,128139); ,4Lee et al. (2005,150080); ,sBabich et al.
(2000,150239); ,sLee et al. (2005c); "Lee et al. (2005,128139); '"Anderson and Ogren (1998,150213); ,9Chung et al. (2001,150357); ""Kidwell and Ondov (2004,155898); "'Lithgow et al.
(2004,120010); ""Weber et al. (2003,157129); "'Harrison et al. (2004); "Rattigan et al. (2000,115897); "sWittig et al. (2004,103413); ""Vaughn et al. (2005,157089); "'Chow et al. (2005,
099030); ""Weber et al (2001,024040); ""Schwab et al. (2000a); "°Lim et al. (2003,150097); "'Watson and Chow (2002,037873); ""Venkatachari et al. (2000,105918); 93Bae et al. (2004a);
"Arhami et al. (2000,150224); "'Park et al. (2005,150843); ""Bae et al. (2004b); "Chow et al. (2000a); ""Arnott et al. (2005); ""Bond et al. (1999); '""Virkkula et al. (2005,157097); '"'Petzold et
al. (2002,150803); ,02Park et al. (2000); "3Arnott et al. (1999, 020050); '"Peters et al. (2001); '"'Pitchford et al. (1997,150872); '""Rees et al. (2004, 097104); '"'Watson et al. (2000); '""Lee et
al. (2005,150080); '""Hering et al. (2004,155837); ""Watson et al. (1998); "'Chakrabarti et al. (2004); "2Mathai et al. (1990,150741); '"Kidwell and Ondov (2001, 017092); '"Stanier et al.
(2004);	"sKhlystov et al. (2005,150035); ""Takahama et al. (2004,157038); "'Chow et al. (2005,150348); ""Zhang et al. (2002); ""Subramanian et al. (2004, 081203); ,2°Chow et al. (2000b);
"'Birch and Cary (1990); ,22Birch (1998, 024953); ,23Birch and Cary (1990); ""NIOSH (1990,150810); "sNI0SH (1999,150811); ""Chow et al. (1993); "'Chow et al. (2007); ""Ellis and Novakov
(1982,150410); ""Peterson and Richards (2002,150801); ""Schauer et al. (2003); "'Middlebrook et al. (2003, 042932); ,32Wenzel et al. (2003,157139); "3Jimenez et al. (2003); ,34Phares et al.
(2003,150800); "sQin and Prather (2000,150895); ""Zhang et al. (2005); "'Bein et al. (2005,150205); ""Drewnick et al. (2004a); ""Drewnick et al. (2004b); ""Lake et al. (2003,150009);
"'Lake et al. (2004,088411).
n/a	n/a	0.03/yg/m Subject to interferences Within 10% of nylon- 94 to 98%
from fragments of other filter NO3 -, and
species with mass to within 15% of PILS-
charge ratio in the same IC and 30% of R&P-
range as fragments of 8400N20
nitrate. Highly refractory
materials are not
detected.
Table A-8. Measurement and analytical specifications for continuous SO42 .
Instrument and Measurement Principle Averaging Analytical precjsjon |\/|Q|_ Interferences Comparability „ ",ata
	_	Time Accuracy	_	 Completeness
FLASH VOLA TILIZA TION INSTRUMENTS
Aerosol Dynamics, Inc. continuous sulfate analyzer (ADI- 10 min
S)
Particle collection by impaction followed by flash
volatilization and detection of the evolved gases by a UV-
fluorescence SO2 analyzer.
n/a	n/a	0.4/yg/m n/a	Within 15% of 100%
82	filter and
continuous
S042"
See Weber et al.
82 for details
July 2009
A-15
DRAFT-DO NOT CITE OR QUOTE

-------
Instrument and Measurement Principle
Averaging Analytical
Time Accuracy'
Precision MDL Interferences Comparability
Data
Completeness
Rupprecht and Patashnick continuous sulfate analyzer
(R&P-8400S)
Particle collection by impaction followed by flash
volatilization and detection of the evolved gases by a UV-
fluorescence SO2 analyzer. An activated carbon denuder
at the inlet to the Nafion humidifier removes SO2.
10 min
n/a
25% on avg
<	15% at
conc.
>9/yg/m3
and >30%
at conc.
<	2/yg/m3
0.48 /yg/m3
S042 to SO2
conversion and
volatilization
efficiency appears
to depend on
ambient
composition 84
10 to 30% lower
than filter SO42'
84 to
95%0'20'21'8
THERMAL REDUCTION INSTRUMENTS
Continuous Ambient Sulfate Monitor (CASM) Sampled air 15 min
passes through a NayCO:; coated annular denuder to
remove ambient SO2 and is subsequently split into
independent sample and filter flows. The sample flow
passes through a quartz tube containing a stainless steel
rod maintained at 1000 °C that reduces sulfate to SO2.
The flow then passes through a PTFE filter and into a
trace-level SO2 fluorescence analyzer.
n/a
n/a
n/a
n/a
Up to 25% lower
than filter SO42'
and within 6% of
R&P8400S, PILS-
IC andAMS20'21
80 to 98%20-21
Thermo Electron Model 5020 sulfate particulate analyzer 15 min
(TE-5020)
The commercial version of CASM, with slight changes in
the sample flow path.
n/a
< 10% c8
0.3/yg/m3
for 24-h
avg 80
0.5/yg/m3
for 15-min
avgd
S042 to SO
conversion
efficiency
on ambient
composition
-20% lower 88 to 9O%80
than filter SO42'80
SAMPLE DISSOLUTION FOLLOWED BY IC ANAL YSIS INSTRUMENTS
Energy Research Center of the Netherlands (ECN) IC-
based ion analyzer
Entrains particles into water drops using the steam jet
aerosol collector. The drops are collected using a cyclone
and the combined flow from collected droplets containing
dissolved aerosol components and wall steam
condensate is directed to an anion IC for analysis of
sulfate. Interfering gases are pre-removed by a rotating
wet annular denuder system.
1-h
n/a
n/a
n/a
n/a
Within 15% of
filter and
continuous SO42'
See Weber et al.
82 for details.
100%7
Texas Tech University (TT) ion analyzer
Particles in the sample stream, after being processed
through a cyclone and a parallel plate wet denuder, are
collected alternatively on one of two 2.5 cm pre-washed
glass fiber filters for a period of 15 min. The particles on
the freshly sampled filter are automatically extracted for
6.5 min with water and analyzed for sulfate by IC.
30 min
n/a
n/a
n/a
n/a
Within 15% of
filter and
continuous SO42'
See Weber et al.
82 for details.
100%7
Particle into Liquid Sampler-Ion Chromatography (PILS-
IC)
Ambient particles are mixed with saturated water vapor
to produce droplets collected by impaction. The resulting
liquid stream is analyzed with an IC to quantify aerosol
ionic components.
1-h
n/a
10%-15%B
0.1 to
O.'Vg/m3
Consistent water
quality is essential
for good precision.
Within 30% of
filter and other
continuous SO42'
65 to 7O%20-21
~ionex-IC
The gas-denuded air stream enters the annular channel of
a concentric nozzle, where deionized water generates a
spray that entrains the particles. The flow is then drawn
through a 0.5/ym pore size PTFE filter. The remaining
solution is aspirated by a peristaltic pump and sent to IC
for ion analysis.
1-h	n/a	11 %f06 n/a	Consistent water Within 10% of
quality is essential filter SO42'86
for good precision.
n/a
July 2009
A-16
DRAFT-DO NOT CITE OR QUOTE

-------
Instrument and Measurement Principle
Averaging Analytical
Time Accuracy'
Precision MDL Interferences Comparability
Data
Completeness
Ambient Ion Monitor (AIM; Model 9000)
Air is drawn through a size-selective inlet into a liquid
diffusion denuder where interfering gases are removed.
The stream enters a super saturation chamber where the
resulting droplets are collected through impaction. The
collected particles and a fraction of the condensed water
are accumulated until the particles can be injected into IC
for hourly analysis.
1-h
n/a
n/a
0.1 /yg/m3
for 1-h
avgd
n/a
n/a
n/a
PARTICLE MASS SPECTROMETER
Aerosol Mass Spectrometer (AMS)
Airstream is drawn through an aerodynamic lens and
focused into a beam in a vacuum chamber. This aerosol
beam is chopped by a mechanical chopper and the flight
time of the particles through a particle-sizing chamber is
determined by the time-resolved mass spectrometer
measurement. The particle impacts onto a 600 °C
heated plate where it decomposes and is analyzed by a
quadruple mass spectrometer. The sulfate ion, along with
other ions, is detected by the mass spectror
A few
seconds
n/a
n/a
n/a
Subject to
Up to 30% lower 93 to 98%20-21
interferences from
than filter SO42'
fragments of other
and within 5% of
species with mass
R&P8400S, PILS-
to charge ratio in
IC and CASM 20'21
the same range as

fragments of

sulfate. Highly

refractory

materials are not

detected.

8 Accuracy is the ability of analytical methods to quantify the observable of a standaid reference material correctly; does not refer to measurement accuracy, since no standards available.
b Overall uncertainty estimated by error propagation.
c Co-located precision estimate based on regression slope (slope = 0.95, intercept = 0.01 to 0.2, R2 > 0.98).
d Manufacturer specified measurement parameter.
e Uncertainty estimated from uncertainties in flow rates and calibrations; does not refer to co-located precision,
f Co-located precision with respect to PC-BOSS PM2.5 SO42
n/a: Not available
Source: 'Chow (1995,077012); 2Watson and Chow (2001,157123);3 Watson et al. (1983, 045084); "Fehsenfeld et al. (2004,156432); 'Solomon et al. (2001,156993); sMikel (2001,156762);
'Mikel (2001,156762); "Watson et al. (1999,020949); 'Solomon and Sioutas (2006,156995); ,0Graney et al. (2004, 053756); "Tanaka et al. (1998,157041); "Pancras et al. (2005,098120);
"John et al. (1988,045903); "Hering and Cass (1999, 084958); "Fritz et al. (1989, 077387); '"Hering et al. (1988, 036012); "Solomon et al. (2003,156994); '"Cabada et al. (2004,148859);
"Fine et al. (2003,155775); 2°Hogrefe et al. (2004, 099003); ''Drewnick et al. (2003, 099160); 22Watson et al. (2005,157125); 23Ho et al. (2006,156552); 24Decesari et al. (2005,144536);2S
Mayol-Bracero et al. (2002, 045010); 2"yang et al. 2003;21 Tursic et al. (2006,157063); 2"Mader et al.(2004,156724); 29Xiao and Liu (2004, 056801); "Kiss et al. (2002,156646); ''Cornell and
Jickells (1999,156367);32 Zheng et al. (2002, 026100); 33Fraser et al. (2002,140741);34 Fraser et al. (2003, 042231) 3sSchauer er al. (1996, 051162); 3sFine et al. (2004); 3,yue et al. (2004);
"Rinehart et al. (2006,115184); 39Wan and yu (2006,157104); 40Poore (2000, 012839); "Fraser et al. (2003a); 42Engling et al. (2006,156422); 43yu et al. (2005); 44Tran et al. (2000); 4syao et al.
(2004,102213); 4sLi and yu (2005,156692); "Henning et al. (2003,156539); 4"Zhang and Anastasio (2003,157182); "Enmenegger et al. (2007,156418); s0(Watson et al., 1989,157119);
"Greaves et al. (1985,156494); "Waterman et al. (2000,157116); s3Waterman et al. (2001,157117); s4Falkovich and Rudich (2001,156427); ssChow et al. (2007,156354); "Miguel et al.
(2004,123260); "Crinmins and Baker (2006, 097008); s"Ho and yu (2004,156551); s9Jeon et al. (2001,016636); "Mazzoleni et al. (2007, 098038); s,Poore (2000, 012839); ^Butler et al.
(2003);	"3Chow et al. (2006c); "Russell et al. (2004, 082453); ssGrover et al. (2006); "Graver et al. (2005, 090044); "Schwab et al. (2006b); "Hauck et al. (2004); s9Jaques et al. (2004,
155878); '"Rupprecht and Patashrick (2003,157207); "Pang et al (2002, 030353); "Eatough et al. (2001,010303); ,3Lee et al. (2005,128139); ,4Lee et al. (2005,156680); ,sBabich et al.
(2000,156239); ,sLee et al. (2005c); "Lee et al. (2005,128139); '"Anderson and Ogrend998,156213); ,9Chung et al. (2001,156357); "Kidwell and Ondov (2004,155898); "'Litt^ow et al.
(2004,126616); ""Weber et al. (2003,157129); "'Harrison et al. (2004); "Rattigan et al. (2006,115897); "sWittig et al. (2004,103413); ""Vaughn et al. (2005,157089); "'Chow et al. (2005,
099030); ""Weber et al (2001,024640); ""Schwab et al. (2006a); "Lim et al. (2003,156697); "'Watson and Chow (2002,037873); ""Venkatachari et al. (2006,105918); 93Bae et al. (2004a);
"Arhami et al. (2006,156224); ^Park et al. (2005,156843); 9"Bae et al. (2004b); "Chow et al. (2006a); "Arnott et al. (2005); "Bond et al. (1999); '""Virkkula et al. (2005,157097); '"'Petzold et
al. (2002,156863); ,02Park et al. (2006); "3Arnott et al. (1999, 020650); '"Peters et al. (2001); '"Pitchford et al. (1997,156872); '"Rees et al. (2004, 097164); '"'Watson et al. (2000); '""Lee et
al. (2005,156680); '"Hering et al. (2004,155837); ""Watson et al. (1998); "'Chakrabarti et al. (2004); "2Mathai et al. (1990,156741); '"Kidwell and Ondov (2001, 017092); '"Stanier et al.
(2004);	"sKhlystov et al. (2005,156635); ""Takahama et al. (2004,157038); "'Chow et al. (2005,156348); ""Zhang et al. (2002); '"Subramanian et al. (2004, 081203); ,2"Chow et al. (2006b);
"'Birch and Cary (1996); ,22Birch (1998, 024953); ,23Birch and Cary (1996); ""NIOSH (1996,156810); "sNI0SH (1999,156811); ""Chow et al. (1993); "'Chow et al. (2007); ""Ellis and Novakov
(1982,156416); ""Peterson and Richards (2002,156861); ""Schauer et al. (2003); "'Middlebrook et al. (2003, 042932); ,32Wenzel et al. (2003,157139); "3Jimenez et al. (2003); ,34Phares et al.
(2003,156866); "s0in and Prather (2006,156895); ""Zhang et al. (2005); "'Bein et al. (2005,156265); ""Drewnick et al. (2004a); ""Drewnick et al. (2004b); ""Lake et al. (2003,156669);
"'Lake et al. (2004,088411)
July 2009
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-9. Measurement and analytical specifications for ions other than NO3 and SO42-.
Instrument & Measurement Principle
Averaging Analytical
Minimum
Data
Time
, Precision Detectable Interferences Comparability „
AcCUraCV	Limit (Mdl)	Completeness
SAMPLE DISSOLUTION FOLLOWED BY IC ANAL YSIS INSTRUMENTS
N02'by Particle into Liquid Sampler-Ion
Chromatography (PILS-IC)
Ambient particles are mixed with saturated
water vapor to produce droplets collected by
impaction. The resulting liquid stream is analyzed
with an IC to quantify aerosol ionic components.
1-h
n/a
10%b
0.14/yg/m3
Consistent water
quality is essential
for good precision
n/a
n/a
NH4+ by Particle into Liquid Sampler-Ion
Chromatography (PILS-IC)
Ambient particles are mixed with saturated
water vapor to produce droplets collected by
impaction. The resulting liquid stream is analyzed
with an IC to quantify aerosol ionic components.
1-h
n/a
10%b
0.05 /yg/m3
Consistent water
quality is essential
for good precision
— 5% lower than
all-sampler avgc at
Atlanta 7
n/a
CI", Na+, K+, Ca++ by Particle into Liquid
Sampler-Ion Chromatography (PILS-IC)
Ambient particles are mixed with saturated
water vapor to produce droplets collected by
impaction. The resulting liquid stream is analyzed
with an IC to quantify aerosol ionic components.
1-h
n/a
10%b
0.1 /yg/m3
Consistent water
quality is essential
for good precision
n/a
n/a
CI , NO?", NOs-, PO43, SO42 , NH4+, Na+, Mg++,
K+, Ca++ by Ambient Ion Monitor (AIM; Model
9000)
Air is drawn through a size-selective inlet into a
liquid diffusion denuder where interfering gases
are removed. The stream enters a super
saturation chamber where the resulting droplets
are collected through impaction. The collected
particles and a fraction of the condensed water
are accumulated until the particles can be
injected into IC for hourly analysis.
1-h
n/a
n/a
0.1 /yg/m3 for
1-h avgd
n/a
n/a
n/a
July 2009
A-18
DRAFT-DO NOT CITE OR QUOTE

-------
Avera in Anal tical	Minimum
Instrument & Measurement Principle	„ Precision Detectable Interferences Comparability „
Time Accuracy	Completeness
8 Accuracy is the ability of analytical methods to quantify the observable of a standard reference material correctly; does not refer to measurement accuracy, since no standards are available.
b Uncertainty estimated from uncertainties in flow rates and calibrations; does not refer to co-located precision.
c All-sampler avg appears to include a combination of 10 integrated and 3 continuous samplers, although specific details are missing 7. Performance evaluations at sites dominated by semi-volatile
ammonium nitrate are needed.
d Manufacturer specified measurement parameter
Source: 'Chow (1995,077012); 2Watson and Chow (2001,157123);3 Watson et al. (1983, 045084); "Fehsenfeld et al. (2004,156432); 'Solomon et al. (2001,156993); sMikel (2001,156762);
'Mikel (2001,156762); "Watson et al. (1999,020949); 'Solomon and Sioutas (2006,156995); '"Graney et al. (2004, 053756); "Tanaka et al. (1998,157041); "Pancras et al. (2005,098120);
"John et al. (1988,045903); "Hering and Cass (1999, 084958); "Fritz et al. (1989, 077387); '"Hering et al. (1988, 036012); "Solomon et al. (2003,156994); "Cabada et al. (2004,148859);
"Fine et al. (2003,155775); 2°Hogrefe et al. (2004, 099003); ''Drewnick et al. (2003, 099160); 22Watson et al. (2005,157125); 23Ho et al. (2006,156552); 24Decesari et al. (2005,144536);2S
Mayol-Bracero et al. (2002, 045010); 2sYang et al. 2003;21 Tursic et al. (2006,157063); 2"Mader et al.(2004,156724); 29Xiao and Liu (2004, 056801); 3°Kiss et al. (2002,156646); ''Cornell and
Jickells (1999,156367);32 Zheng et al. (2002, 026100); 33Fraser et al. (2002,140741);34 Fraser et al. (2003, 042231) 3sSchauer er al. (1996, 051162); 3sFine et al. (2004); 3,Yue et al. (2004);
3BRinehart et al. (2006,115184); 39Wan and Yu (2006,157104); 40Poore (2000, 012839); "Fraser et al. (2003a); 42Engling et al. (2006,156422); 43Yu et al. (2005); 44Tran et al. (2000); 4sYao et al.
(2004,102213); 4sLi and Yu (2005,156692); "Henning et al. (2003,156539); 4"Zhang and Anastasio (2003,157182); "Enmenegger et al. (2007,156418); s0(Watson et al., 1989,157119);
"Greaves et al. (1985,156494); "Waterman et al. (2000,157116); s3Waterman et al. (2001,157117); s4Falkovich and Rudich (2001,156427); ssChow et al. (2007,156354); "Miguel et al.
(2004,123260); "Crinmins and Baker (2006, 097008); s"Ho and Yu (2004,156551); s9Jeon et al. (2001,016636); ""Mazzoleni et al. (2007, 098038); s,Poore (2000, 012839); ^Butler et al.
(2003);	s3Chow et al. (2006c); "Russell et al. (2004, 082453); ssGrover et al. (2006); ssGrover et al. (2005, 090044); "Schwab et al. (2006b); ""Hauck et al. (2004); s9Jaques et al. (2004,
155878); '"Rupprecht and Patashrick (2003,157207); "Pang et al (2002, 030353); "Eatough et al. (2001,010303); ,3Lee et al. (2005,128139); ,4Lee et al. (2005,156680); ,sBabich et al.
(2000,156239); ,sLee et al. (2005c); "Lee et al. (2005,128139); '"Anderson and Ogren (1998,156213); ,9Chung et al. (2001,156357); ""Kidwell and Ondov (2004,155898); "'Lithgow et al.
(2004,126616); ""Weber et al. (2003,157129); "'Harrison et al. (2004); "Rattigan et al. (2006,115897); "sWittig et al. (2004,103413); ""Vaughn et al. (2005,157089); "'Chow et al. (2005,
099030); ""Weber et al (2001,024640); ""Schwab et al. (2006a); "°Lim et al. (2003,156697); "'Watson and Chow (2002,037873); ""Venkatachari et al. (2006,105918); "3Bae et al. (2004a);
"Arhami et al. (2006,156224); "'Park et al. (2005,156843); ""Bae et al. (2004b); "Chow et al. (2006a); ""Arnott et al. (2005); ""Bond et al. (1999); '""Virkkula et al. (2005,157097); '"'Petzold et
al. (2002,156863); ™Park et al. (2006); "3Arnott et al. (1999, 020650); '"Peters et al. (2001); '"'Pitchford et al. (1997,156872); '""Rees et al. (2004, 097164); '"'Watson et al. (2000); '""Lee et
al. (2005,156680); '""Hering et al. (2004,155837); ""Watson et al. (1998); "'Chakrabarti et al. (2004); "2Mathai et al. (1990,156741); '"Kidwell and Ondov (2001, 017092); '"Stanier et al.
(2004);	"sKhlystov et al. (2005,156635); ""Takahama et al. (2004,157038); "'Chow et al. (2005,156348); ""Zhang et al. (2002); ""Subramanian et al. (2004, 081203); ,2°Chow et al. (2006b);
"'Birch and Cary (1996); ,22Birch (1998, 024953); ,23Birch and Cary (1996); ""NIOSH (1996,156810); "sNI0SH (1999,156811); ""Chow et al. (1993); "'Chow et al. (2007); ""Ellis and Novakov
(1982,156416); ""Peterson and Richards (2002,156861); ""Schauer et al. (2003); "'Middlebrook et al. (2003, 042932); ,32Wenzel et al. (2003,157139); "3Jimenez et al. (2003); ,34Phares et al.
(2003,156866); "sQin and Prather (2006,156895); ""Zhang et al. (2005); "'Bein et al. (2005,156265); ""Drewnick et al. (2004a); ""Drewnick et al. (2004b); ""Lake et al. (2003,156669);
"'Lake et al. (2004,088411)
Table A-10. Measurement and analytical specifications for continuous carbon.
Instrument and Measurement Principle
Averaging Analytical
Time Accuracy3
Minimum
Precision Detectable Interferences Comparability
Limit
Data
Completeness
PARTICLE COLLECTION ONIMPACTOR FOLLOWED BY FLASH VOLATILIZA TION INSTRUMENT
Aerosol Dynamic Inc. continuous carbon analyzer
(ADI-C)
Particle collection by impaction followed by flash
oxidation and detection of the evolved gases by a
non-dispersive infrared CO2 analyzer. OC is
estimated as twice the oxidizable carbon. EC is not
quantified.
10 min
n/a
n; a
0C:2//g/m3 n/a
EC, TC: not
applicable,
since it
measures only
OC 90
15 to 22% lower
OC than that by
R&P-5400 and Rll-
OGI
83%7
PARTICLE COLLECTION ONFIL TER/IMPACTOR FOLLOWED BYHEA TING/ANAL YSIS INSTRUMENTS
Rupprecht and Patashnick 5400 continuous ambient
carbon analyzer (R&P-5400)
Particles collected on an impactor, which is heated
to 275 °C to 350 °C, then to 700 °C after sample
collection is complete. Evolved C02 is measured by
an infrared detector. OC is defined as the carbon
measured at the lower temperature, and EC is the
remaining carbon measured at the higher
temperature.
1-h
n/a
n/a
OC: 0.5/yg/m3
EC: 0.5/yg/m3
TC: 0.5/yg/m3
n/a
20 to 60% lower
TC than filter TC
by TOR or
TOT.9',92
56 to 6O%0'01
July 2009
A-19
DRAFT-DO NOT CITE OR QUOTE

-------
Instrument and Measurement Principle
Averaging Analytical
Time Accuracy3
Precision
Minimum
Detectable
Limit
Interferences Comparability
Data
Completeness
Rutgers University-Oregon Graduate Institute (RU-
OGI) in-situ thermal/optical transmittance carbon
analyzer.
Air is sampled through a quartz-fiber filter for 1 -h
and then analyzed by heating through different
temperature steps to determine OC and EC. Sample
flow is pre-split into two identical systems that
alternate every hour between sampling and analysis
mode to achieve continuous measurements.
30 min
n/a
3%b'
0.3/yg/m3 n/a
0.5/yg/m3
0.4/yg/m3
8% higher 00 and
20% lower EC than
R&P5400 00
86%7
Sunset semi-continuous realtime carbon aerosol
analysis instrument (Sunset OCEC)
Particles collected on a quartz-fiber filter are subject
to heating temperature ramps following the NIOSH
5040 TOT protocol and the resulting CO2 is analyzed
by nondispersive infrared (NDIR) detector to quantify
OC and EC. Instrument is alternated between
sampling and analytical mode.
1-h
n/a
OC: 10%c
EC: 20%c
TC: 10%c
OC: n/a
EC: n/a
TC: 0.4/yg/m3
(1-havg)96
n/a
Within 7 to 25% of
filter OC and EC
and within 15% for
TC. Wide variation
due to differentces
in temperature and
analysis protocols.
80 to 89%6,95
LIGHT ABSORPTION INSTRUMENTS
Aethalometer (AE-16, AE-21, AE-31) 5 min
n/a
5 to
BC e:
Subject to multiple
Within ± 25% of 75 to 90%6
Attenuation of light transmitted through a quartz-

10%d'7'97
0.1 /yg/m300
scattering effects
RU-OGI, Sunset
fiber filter tape that continuously samples aerosol is



by particle and
and filter EC by
TOR/TOT.00"02
measured and converted to a BC mass concentration



filter matrix
using cjabs of 14625/A (m2/g).



resulting in




absorption en-
hancement. Em-
pirical corrections
have been proposed
98 that can correct
for such effects.

Particle Soot Absorption Photometer (PSAP)
Attenuation of light transmitted through a glass-
fiber filter that continuously samples aerosol is
measured to quantify light absorption (bate).
1 min
n/a
6 to
BC f:
0.1 /yg/m300
Instrument includes
an empirical
correction for
scattering and
loading effects 99
and adjustments
have been proposed
for the three wave-
length model100
— 50% lower than
AE-16, RU-OGI and
R&P-5400 EC.00
n/a
Multi-Angle Absorption Photometer (MAAP) Light
transmittance at 0° and reflectance from a glass-
fiber filter at 130° and 165° from the illumination
direction are used in a radiative transfer model to
estimate bate and is converted to BC using crate of
6.6 m2/g.
1 min
n/a
12%°'101
BC h:
0.05 /yg/m3
(or bate - 0.33
Mm-1 for 10-
min avg)
0.02 /yg/m3
(or bate - 0.13
Mm-1 for 30-
min avg)101
The instrument is
designed to mini-
mize multiple scat-
tering and loading
effects by mea-
suring both trans-
mittance and
reflectance and
using a two-stream
approximation
radiative transfer
model to calculate
Within 18% of
filter EC by
IMPROVETOR
(R2 - 0.96) and up
to 40% higher than
Sunset EC. 102
n/a
July 2009	A-20	DRAFT - DO NOT CITE OR QUOTE

-------
Instrument and Measurement Principle
Averaging Analytical
Time Accuracy3
Precision
Minimum
Detectable
Limit
Interferences Comparability
Data
Completeness
DRI Photoacoustic Analyzer (DRI-PA)
Light absorption by particles in air results in a
heating of the surrounding air. The expansion of the
heated air produces an acoustic (sound wave) signal
which is detected by a microphone to determine bate,
which is converted to BC using (Tabs - 5 m2/g for
the 1047 nm instrument and (Tabs - 10 m2/g for the
532 nm instrument.
5 sec
n/a
n/a
BC i:
At 532 nm, absor-
Good correlation
0.04 /yg/m3(or
bance by NO2
(R2 >0.80), but
babs = 0.4
interferes with that
more than 40%
Mm-1 for 10-
by particles. Ac-
lower than
min avg) at
counted by either
aethalometer,
532 nm103
removing NO2 from
MAAP and filter

sample line using
IMPROVE TOR

denuders or by
EC. Suggests need

doing a periodic
for a different (Tabs.

background
102

(particle-free air)


subtraction.

n/a
PHOTOIONIZA TION INSTRUMENTS
Photoionization monitor for 91 %6T polycyclic
aromatic hydrocarbons (PAS PAH) The air stream is
exposed to UV radiation, which ionizes the particle-
bound PAH molecules. The charged particles are
collected on a filter element and the piezoelectric
current is proportional to the particle-bound PAH.
5 min
n/a
n/a
-3ng/m3i,k n/a
n/a
> 91 %6t
8 Accuracy is the ability of analytical methods to quantify the observable of a standairi reference material correctly; does not refer to measurement accuracy, since no standards are available.
b No specific details on how the precision was estimated; appears to be based on replicate analysis, may not represent overall co-located measurement precision
c Co-located precision estimates based on variation in avg ratios of replicate analysis using laboratory instrument and regression slopes (Slopes for OC = 1.01, EC = 0.82, TC = 0.94; R2 = 0.97 -
0.99) of co-located field measurements.
d Estimated using co-located AE-21 and AE-31 BC measurements at Fresno, CA.97
e While the default manufacturer recommended conversion factor (or mass absorption efficiency, aabs) is 16.6 m2/g at 880 nm, Lim et al. (2003) assumed a value of 12.6 m2/g.
f Assuming a aabs of 10 m2/g.
B Co-located precision estimate based on the variability of the avg ratio (0.99 ± 0.12).
h Assuming a aabs of 6.5 rrv7g.
'Assuming a aabs of 10 m2/g at 532 nm and 5 m2/g at 1047 nm.
1 Specified by manufacturer as "lower threshold"; needs to be calibrated with site-specific PAH. Typically used as a relative measure in terms of electrical output in femtoamps.
k Manufacturer specified measurement parameter
n/a: Not available.
Source: 'Chow (1995,077012); 2Watson and Chow (2001,157123);3 Watson et al. (1983, 045084); "Fehsenfeld et al. (2004,156432); 'Solomon et al. (2001,156993); sMikel (2001,156762);
'Mikel (2001,156762); "Watson et al. (1999,020949); 'Solomon and Sioutas (2006,156995); ,0Graney et al. (2004, 053756); "Tanaka et al. (1998,157041); "Pancras et al. (2005,098120);
"John et al. (1988,045903); "Hering and Cass (1999, 084958); "Fritz et al. (1989, 077387); '"Hering et al. (1988, 036012); "Solomon et al. (2003,156994); '"Cabada et al. (2004,148859);
"Fine et al. (2003,155775); 20Hogrefe et al. (2004, 099003); 2,Drewnick et al. (2003, 099160); 22Watson et al. (2005,157125); 23Ho et al. (2006,156552); 24Decesari et al. (2005,144536);2S
Mayol-Bracero et al. (2002, 045010); 2sYang et al. 2003;21 Tursic et al. (2006,157063); 2"Mader et al.(2004,156724); 29Xiao and Liu (2004, 056801); "Kiss et al. (2002,156646); ''Cornell and
Jickells (1999,156367);32 Zheng et al. (2002, 026100); 33Fraser et al. (2002,140741);34 Fraser et al. (2003, 042231) 3sSchauer er al. (1996, 051162); 3sFine et al. (2004); 3,Yue et al. (2004);
3BRinehart et al. (2006,115184); 39Wan and Yu (2006,157104); 40Poore (2000, 012839); "Fraser et al. (2003a); 42Engling et al. (2006,156422); 43Yu et al. (2005); 44Tran et al. (2000); 4sYao et al.
(2004,102213); 4sLi and Yu (2005,156692); "Henning et al. (2003,156539); 4"Zhang and Anastasio (2003,157182); "Enmenegger et al. (2007,156418); s0(Watson et al., 1989,157119);
"Greaves et al. (1985,156494); "Waterman et al. (2000,157116); s3Waterman et al. (2001,157117); s4Falkovich and Rudich (2001,156427); ssChow et al. (2007,156354); "Miguel et al.
(2004,123260); "Crinmins and Baker (2006, 097008); s"Ho and Yu (2004,156551); s9Jeon et al. (2001,016636); "Mazzoleni et al. (2007, 098038); s,Poore (2000, 012839); ^Butler et al.
(2003);	G3Chow et al. (2006c); "Russell et al. (2004, 082453); ssGrover et al. (2006); "Graver et al. (2005, 090044); "Schwab et al. (2006b); "Hauck et al. (2004); s9Jaques et al. (2004,
155878); '"Rupprecht and Patashrick (2003,157207); "Pang et al (2002, 030353); ,2Eatough et al. (2001,010303); ,3Lee et al. (2005,128139); ,4Lee et al. (2005,156680); ,sBabich et al.
(2000,156239); ,sLee et al. (2005c); "Lee et al. (2005,128139); '"Anderson and Ogren (1998,156213); ,9Chung et al. (2001,156357); "Kidwell and Ondov (2004,155898); "'Lithgow et al.
(2004,126616); ""Weber et al. (2003,157129); "'Harrison et al. (2004); "Rattigan et al. (2006,115897); "sWittig et al. (2004,103413); ""Vaughn et al. (2005,157089); "'Chow et al. (2005,
099030); ""Weber et al (2001,024640); ""Schwab et al. (2006a); "Lim et al. (2003,156697); "'Watson and Chow (2002,037873); ""Venkatachari et al. (2006,105918); "3Bae et al. (2004a);
"Arhami et al. (2006,156224); "'Park et al. (2005,156843); "Bae et al. (2004b); "Chow et al. (2006a); "Arnott et al. (2005); ""Bond et al. (1999); '"Virkkula et al. (2005,157097); '"'Petzold et
al. (2002,156863); ,02Park et al. (2006); "3Arnott et al. (1999, 020650); '"Peters et al. (2001); '"'Pitchford et al. (1997,156872); '"Rees et al. (2004, 097164); '"'Watson et al. (2000); '""Lee et
al.(2005,156680); '""Hering et al. (2004,155837); ""Watson et al. (1998); "'Chakrabarti et al. (2004); "2Mathai et al. (1990,156741); '"Kidwell and Ondov (2001, 017092); '"Stanier et al.
(2004);	"sKhlystov et al. (2005,156635); ""Takahama et al. (2004,157038); "'Chow et al. (2005,156348); ""Zhang et al. (2002); ""Subramanian et al. (2004, 081203); ,2°Chow et al. (2006b);
"'Birch and Cary (1996); ,22Birch (1998, 024953); ,23Birch and Cary (1996); ""NIOSH (1996,156810); ,2sNI0SH (1999,156811); ""Chow et al. (1993); "'Chow et al. (2007); ""Ellis and Novakov
(1982,156416); ""Peterson and Richards (2002,156861); ""Schauer et al. (2003); "'Middlebrook et al. (2003, 042932); ,32Wenzel et al. (2003,157139); "'Jimenez et al. (2003); ,34Phares et al.
(2003,156866); "sQin and Prather (2006,156895); ""Zhang et al. (2005); "'Bein et al. (2005,156265); ""Drewnick et al. (2004a); ""Drewnick et al. (2004b); ""Lake et al. (2003,156669);
"'Lake et al. (2004,088411)
July 2009
A-21
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-11. Summary of mass measurement comparisons.
Site I Period I Sampler I Configuration
Summary of Findings
1.	Birmingham, AL (11/04/96 To 11/23/96)
2.	Denver-Adams City, CO (12/11/96 To 1/7/97)
3.	Bakersfield, CA (1/21/97 To 3/19/97)
4.	Denver-Welby, Co (12/12/96 To 12/21/96)
5.	Phoenix, AZ (12/06/96 To 12/21/96)
6.	Azusa, CA (3/25/97 To 5/19/97)
7.	Research Triangle Park (RTP), NC (1/17/97 To 8/14/97)
8.	Rubidoux, CA (1/6/99 To 2/26/99)
9.	Atlanta, GA (8/3/99 To 8/31/99)
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE'
DENUDER"
RAAS2.5-100 PM2.6
FRM
16.7

Teflon (n/a)
None
RAAS2.5-300 PM2.1,
FRM
16.7

Teflon (n/a)
None
RAAS2.5-200 PM2.1,
FRM
16.7

Teflon (n/a)
None
R&P Partisol 2000
PM2.6 FRM
16.7

Teflon (n/a)
None
R&P Partisol-plus
2025 PM2.6 FRM
16.7

Teflon (n/a)
None
BGI PQ200 PM2.6
FRM
16.7

Teflon (n/a)
None
Sierra Instruments
SA-244 Dichot
16.7

Teflon (n/a)
None
IMPROVE PM2.6
22.8

Teflon (n/a)
None
Harvard PM2.6
Impactor
10

Teflon (n/a)
None
Airmetrics battery
powered PM2.6
MiniVol
5

Teflon (n/a)
None
ATLANTA SUPERSITE, GA: 8/3/99 TO 9/1/99
4 km NW of downtown, within 200 m of a bus maintenance yard and several warehouse
facilities, representative of a mixed commercial-residential neighborhood.
Sampler
Flow
Rate
(L/Min)
Filter Type"

Denuder1
R&P-2000 FRM
16.7
Teflon (P)

None
RAAS-100 FRM
16.7
Teflon (P)

None
RAAS-400
24
Teflon (P)

None
SASS
6.7
Teflon (P)

None
MASS-400
16.7
Teflon (P)

Na2C03
R&P-2300
10
Teflon (P)

None
R&P-2025 Dichot:
PM2.6
15
Teflon (P)

None
PM 10-2.6
1.67
Polycarbonate
None Na2C03/Citric
Peters et al., (2001,017108)'": Pitchford et al., (1997,156872)1"5
dataset
Co-located precision (CV) for the RAAS2.5-100 samplers ranged from 1.5%
at Bakersfield to 6.2% at Birmingham.
In Birmingham, CV for two co-located Harvard Impactor was 1% and for
three Dichots was 6.2%. The IMPROVE samplers had greater variability, with
a CV of 11.3% (Denver-Adam City) and 10.8% (Bakersfield).
Partisol and RAAS showed the strongest pairwise comparison
' (slope — 1.0 ± 0.06, intercept — 0.26 ± 1.81, and correlation - 1.0),
within the EPA equivalency criteria. Strong relationships (correlation > 0.96;
slope - 0.9 - 1.12, intercept < 3ct) were observed for other samplers in
reference to the RAAS.
" At Denver-Welby, 6 RAAS samplers were deployed (3 with and 3 without
temperature compensation for flow control). The units with temperature
compensation had a positive bias relative to the non-temperature
compensated units.
Non-FRM samplers did not meet the EPA equivalency criteria, despite strong
-	linear relationships with the FRM sampler.
Peters et al.10*: RTP 97 dataset
¦	CV was 1.7%, 2.3%, 3.4%, 6.4% for the PQ200, Partisol 2000, RAAS2.5-
100, and Dichot, respectively. Dichot flows were valve controlled and set
visually by the operator using rotameters.
Good one-to-one correspondence was observed for FRM comparisons. The
FRM averages were within -1.2% to 3.2%, within the acceptable ± 10%
-	range
Peters et al. 104: Rubidoux 99 and Atlanta 99 dataset
¦	In Rubidoux, the precision for PQ200 was 6.1 %, higher than at RTP 97. In
Atlanta, the grouped data from PQ200, RAAS2.5-300, and Partisol yielded a
' precision of 1.7%.
Linear regression results met the EPA equivalency criteria for all FRMs.
Solomon et al.17
PM2.5 mass from individual samplers was compared to all-sampler avgs,
called the filter relative reference (filter RR) value. Overall agreements were
within ± 20% of filter RR.
FRM samplers were within 3.5% of filter RR.
Avg mass measured by RAAS-400, SASS and URG-PCM were within ± 10%
of filter RR. Avg mass measured by MASS-400, R&P-2300 and R&P-2025
dichot were greater than filter RR but within ± 20%. Avg mass measured by
PC-BOSS (BYU) and ARA-PCM were lower than filter RR within ±10%.
All samplers except PC-BOSS (TVA) had R2 > 0.80, relative to filter RR.
While avg mass for each sampler was within 20%, daily variability was
>50% of filter RR.
Glycerol in the Na2C03 denuder may have contaminated the filter in the
MASS-400 sampler resulting in higher PM2.6 values.
PC-BOSS samplers removed particles < 0.1 pm aerodynamic diameter from
PM2.6 measurements. Corrections were made using sulfate (SO42 )
concentrations in the major flow or immediately after the PM2.6 inlet, but
before the flow split-up. This was insufficient to bring PC-BOSS mass close
to filter RR. PC-BOSS was also equipped with upstream denuders ahead of
July 2009
A-22
DRAFT-DO NOT CITE OR QUOTE

-------
Site I Period I Sampler I Configuration
URG-PCM
16.7
Teflon (P)

Acid

ARA-PCM
16.7
Teflon (n/a)

Na2C03/Citric acid
PC-BOSS (operated
by TVA)
105
Teflon (W)

CIF

PC-BOSS (operated
by BYU)
150
Teflon (W)

CIF

PM2.5 CONTINUOUS
SAMPLER
FLOW
BATE
(L/MIN)
INLET TEMPEBATUBE DBYEB
OTHEB
TEOM
16.7
30 °C
Nafion
PM2.6
ATLANTA SUPEBSITE, GA: 11/21/01 TO 12/23/01
PM2.5 SAMPLEB
FLOW
BATE
(L/MIN)
FILTEB TYPE1

DENUDES1
R&P-2025 FRM
16.7
Teflon (n/a)

None


PM2.5 CONTINUOUS
SAMPLEB
FLOW
BATE
(L/MIN)
INLET
TEMPEBATUBE
DBYEB

OTHEB
TEOM
16.7
30 °C
Nafion

PM2.6
SESTEOM
16.7
30 °C
Nafion

PM2.6
CAMM
0.3
n/a
Nafion

PM2.6
RAMS
16.7
30 °C
Nafion

PM2.6
TEA & CIF
denuders With
particle
concentrator
Summary of Findings
the filters, which may have enhanced loss of semi-volatile components,
resulting in a lower mass on the filter.
Butler et al.62
The sum of individual species accounted for —78% of the RAAS-100 FRM
PM2.6 mass concentration.
TEOM explained ~ 82 to 92% of the species sum of RAAS with R2 - 0.86.
Lee et al.73
RAMS PM2.E adjusted using particle concentrator efficiency of 0.5.
Good correlation between SESTEOM and Radiance Research M903s
(R2 - 0.80), while medium correlation was found between CAMM and
Radiance Research M903 (R2 - 0.64) or RAMS and Radiance Research
M903 (R2 - 0.63).
CAMM - (0.75 ± 0.03) SESTEOM + (2.51 ± 0.51); R2 - 0.78; N - 196
RAMS - (0.85 ± 0.06) SESTEOM + (5.34 ± 1.04); R2 - 0.52; N - 96
RAMS - (0.91 ± 0.07) CAMM + (5.71 ± 1.20); R2 - 0.43; N - 196
Semi-volatile material explains the difference between RAMS and SES
TEOM.
CAMM - (0.75 ± 0.08) R&P-2025 FRM + (2.47 ± 1.02); R2 - 0.76;
N - 31
RAMS - (0.97 ± 0.22) R&P-2025 FRM + (2.39 ± 3.42); R2 - 0.64;
N - 13
SES TEOM - (1.07 ± 0.05) R&P-2025 FRM + (-1.34 ± 0.71); R2 - 0.95;
N - 26
CAMM vs. FRM yielded lower slopes (0.75) with high intercepts.
Radiance Research
M903
n/a
n/a
Nafion
bscat
Radiance Research
M903
n/a
n/a
None
bscat
PITTSBURGH SUPEBSITE, PA: 7/1/01 to 6/1/02 6 km east of downtown in a park on the Cabada et al;1S: Bees et al.1
top of a \
SAMPLEB
FLOW
BATE
(L/MIN)
FILTEB TYPEa
DENUDEB
MOUDI-110
30
Teflon (P,d)
None
And-241 Dichot
16.7
Teflon (P
None
R&P-2000 PM2.6
FRM
16.7
Teflon (W)
None

PM2.5 CONTINUOUS
SAMPLEB
FLOW
BATE
(L/MIN)
INLET TEMPEBATUBE
DBYEB OTHEB
SESTEOM
16.7
30 °C
Nafion PM2.6
DAASS
n/a
30 °C
Nafion or PM2.6
None
M0UDI PM10 - 0.80 Dichot PM10, R2 - 0.85
M0UDI PM2.6 - 1.03 Dichot PM2.6, R2 - 0.78
M0UDI PM2.6 - 1.01 FRM PM2.6, R2 - 0.78
Dichot PM2.6 - 0.97 FRM PM2.1, + 0.02; R2 - 0.94
Good agreement for PM2.6 FRM, Dichot, and M0UDI. Lower slope for PM10
suggests loss of coarse particles in the MOUDI sampler.
Ultrafine (< 100 nm) mass (PM0.10) measurements had high uncertainties
(-30%)
Ultrafine mass by MOUDI showed no correlation with ultrafine volume
(V0.10) by DAASS. Ratio of PM0.10/PM2.6 mass ratio showed reasonable
agreement with volume ratio (V0.10/V2.5, R2 - 0.55, slope - 0.76). Bounce
of large particles to smaller stages in MOUDI was small, since mass ratio
(PM0.10/PM2.6) did not exceed volume ratio (V0.10/V2.5). Low correlation
between ultrafine mass and volume could be due to the ultrafine mass
measurement uncertainty or due to fundamental differences in the
measurement methods employed by MOUDI and DAASS. Ambient conditions
and characteristics of the aerosols (such as non-spherical shapes of fresh
particles) could also influence these estimates.
July 2009
A-23
DRAFT-DO NOT CITE OR QUOTE

-------
Site I Period / Sampler / Configuration	Summary of Findings	
Rees etal.106
SES-TEOM PM2.6 - 1.02 FRM PM2.6 + 0.65; R2 - 0.95
Volatilization did not affect SES-TEOM performance when PM2.6 mass >20-
30/yglm3. When ambient temperature was < -6 °C, and when mass was
low, SES-TEOM was lower (up to 50%) than FRM or Dichot.
FRESNO SUPERSITE, CA and other CRPAQS sites; 12/2/99 to 2/3/01. Some
comparisons included data till 12/29/03 . Fresno Supersite was located 5.5 km northeast of
downtown in a mixed residential-commercial neighborhood. 107
SAMPLER
FLOW
RATE
(L/MIN)
FILTER TYPEa
DENUDER
RAAS-100 PM2.6
FRM
16.7
Teflon (P)
None
RAAS-300 PM2.6
FRM
16.7
Teflon (P)
None
R&P-2000 PM2.6
FRM
16.7
Teflon (P)
None
R&P-2025 PM2.6
FRM
16.7
Teflon (P)
None
RAAS-400 PM2.6
24
Teflon (P)
None
SASS PM2.6
6.7
Teflon (P)
None
And-246 Dichot
PM2.6
15
Teflon (P)
None
PM 10-2.6
1.67
Teflon (P)
None
DRI-SFS PM2.1,
113
Teflon (P)
None
MiniVol PM2.6
5
Teflon (P)
None
MOUDI-100
30
FEPb Teflon (P)
None
And-hlVOL PM10 FRM
1130
Teflon (P)
None
Chow etal.63
PM2.6 measurements from the 11 filter samplers were within ~ 20% of each
other, except for MiniVols, which were 20 to 30% lower than RAAS-300
FRM.
All the FRM samplers were within ± 10% of each other.
All the filter samplers were well correlated with each other (R2 > 0.901.e
DRI-SFS (with HNO3 denuder) and And-246 Dichot PM2.6 were lower (~ 5%
and 7%, respectively, on avg) than FRM, possibly due to nitrate (NO3-)
volatilization.
Poor correlation (R2) found between TE0M PM2.6 concentrations and RAAS-
100 FRM. TE0M PM2.6 was lower than RAAS-100 FRM by 22%. Heating of
TE0M inlet to 50 °C resulted in loss of semi-volatile components such as
ammonium nitrate (NH4N03) and possibly some semi-volatile organic
compounds.
TE0M PM10 concentrations were 28% lower than the And-hlV0L10 FRM on
avg, ranging from 13% in summer to 43% in winter.
TE0M was neither equivalente nor comparablee to the FRM sampler for
PM2.6 or PM10.
BAM PM2.6 concentrations showed high correlation (R2 > 0.90) with the
RAAS-100 and RAAS-300 FRM samplers, with slopes ranging from 0.92 to
0.97. BAM PM2.6 was typically higher than FRM (17 to 30%) except at
Bakersfield, CA, where it was 21 % lower, suggesting a BAM calibration
difference between Bakersfield and other sites.
BAM PM10 concentrations were 26% higher than And-hlVOL PM10 FRM
concentration on avg (R2 >0.92).
Higher BAM measurements were attributed to water absorption by
hygroscopic particles. BAM PM2.6 and PM10 deviations were larger for
concentrations < 25/yg/m3.
Grover et al.65
(0.88 ± 0.04) FDMS-TE0M + (6.7 ± 4.3); R2 - 95;
(1.11 ± 0.07) D-TE0M + (7.5 ± 6.1); R2 - 0.90; n - 29
¦	(0.80 ± 0.01) TE0m30C + (1.1 ± 3.1); R2 - 0.91;
¦	(0.50 ± 0.01) FDMS-TE0M -(1.7 ± 6.9); R2 - 0.68;
PC-BOSS PM2.6 ¦
n - 29
PC-BOSS PM2.6 ¦
TE0M50C PM2.6 ¦
n - 507
TE0m30C PM2.1, ¦
n - 516
Heated GRIMM PM concentrations were lower than FDMS-TEOM and
ambient temperature GRIMM, suggesting loss of semi-volatile matter.
Data recovery was greater than 95% for all continuous instruments, except
for D-TEOM, which had 86% recovery.
Reasonable agreement was seen between FDMS-TEOM, D-TEOM, BAM, and
GRIMM PM2.E when semi-volatile matter was dominated by NH4NO3.
However, the FDMS-TEOM was higher than the other instruments during
high concentration periods, associated with days with a high fraction of
semi-volatile organic compounds (SVOCs). Possible differences in SVOCs may
have contributed to the differences between FDMS and other instruments.
CONTINUOUS SAMPLER
FLOW RATE (L/MIN)
INLET
TEMPERATURE
DRYER
OTHER
TEOM
16.7
50 °C
None
PM2.6 and PM10
BAM
16.7
Ambient
None
PM2.6 and PM10
July 2009
A-24
DRAFT-DO NOT CITE OR QUOTE

-------
Site I Period I Sampler I Configuration
Summary of Findings
SAMPLER
FLOW RATE (LfMIN) FILTER TYPEa
PC-BOSS PM2.6
150
Teflon (W)
DENUDERb
CIF
CONTINUOUS SAMPLER
FLOW RATE (L/MIN)
INLET
TEMPERATURE
DRYER
OTHER

TEOM
16.7
50 °C
None
PM2.6

TEOM
16.7
30 °C
None
PM2.6

FDMSTEOM
16.7
30 °C
Nafion
PM2.6

D-TEOM
16.7
30 °C
Nafion
PM2.6

GRIMM1100
1.2
Ambient
None
bscat

GRIMM1100
1.2
80 °C heater,
resulting in aerosol
temperature
Heater
bscat

BAM
16.7
Ambient
None
PM2.6

HOUSTON SUPERSITE, TX; 1/1/00 to 2/28/02
The Houston Supersite included three sites located in southeast Texas including one on the grounds of a
municipal airport at the edge of a small community, one adjacent to the highly industrial ship channel and one on
the grounds of a middle school in a suburban community.
Russell et al.Lee et al.103
Good correlations between 24-h SES-TEOM PM2.6 and
R&P-2025 FRM mass.
CAMM - (0.93 + 0.03) RAMS + (3.14 + 0.74);
R2 - 0.81
PM2.5 SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE1

DENUDER
R&P-2025 FRM
16.7
Teflon (n/a)

None
SES-TEOM - (0.92 ± 0.03) RAMS + (1.52 ± 0.77);
. R2 - 0 80





SES-TEOM - (1.01 ± 0.03) CAMM + (-1.91 ± 0.79);
CONTINUOUS SAMPLER
FLOW RATE (L/MIN)
INLET
TEMPERATURE
DRYER
OTHER"
R2 - 0.83
Correlation of Radiance Research M903 and SES-TEOM
- was good (R2 - 0.95), while that of Radiance Research
M903 with CAMM or RAMS was poor (R2 ~ 0.4).
TEOM
16.7
50 °C
None
PM2.6
SES-TEOM
16.7
30 °C
Nafion
PM2.6
Aug-Sep '00
RAMS > SES-TEOM at high temperature and low RH
(< 60%), suggesting loss of water and particulate N03-
-from SES-TEOM.
CAMM - (1.02 ± 0.08) R&P-2025 + (1.62 ± 1.35);
R2 - 0.89
CAMM
0.3
Ambient
Nafion
PM2.6
Aug-Sep '00
RAMS
16.7
30 °C
Nafion
PM2.6
TEA & CIF
denuders; Aug-
Sep '00
RAMS - (1.10 ± 0.08) R&P-2025 + (0.68 ± 1.28);
R2 - 0.89
SES-TEOM - (1.09 ± 0.07) R&P-2025 + (0.21 ± 1.27);
R2 - 0.94
Radiance Research M903
n/a
n/a
Nafion
Bscat Aug-Sep
'00
Integrated mass < Continuous PM2.6 mass. Difference
possibly related to loss of SVOCs and N03- from
integrated sampler
LOS ANGELES SUPERSITE, CA; 9/01 to 8/02
The Los Angeles Supersite consisted of multiple sampling locations in the South Coast Air Basin to provide wide
geographical and seasonal coverage, including urban "source" sites and downwind "receptor" sites.
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE1
DENUDER"
R&P-2025 Dichot
PM2.6
15
Teflon (P)
None
PM 10-2.6
16.7
n/a
None
MOUDI-110
30
Teflon (P
None
HEADS PM2.1,
10
Teflon (n/a)
NaHCOs
Jaques et al.E9; Hering et al.109
Dichot PM2.6 - 0.83 MOUDI + 1.23; R2 - 0.83 (n - 37)
Dichot PM2.E showed higher N03- loss than MOUDI,
consistent with anodized aluminum surfaces serving as
efficient denuders that remove volatilized N03-.2.110.
D-TEOM PM2.6 - 1.18 MOUDI - 1.28; R2 - 0.86
(n - 20)
Over-estimation of D-TEOM may be due to particle losses
in the MOUDI.
PM2.6 by D-TEOM during ESP-off phase (net artifact
effect) tracked well with the N03- concentrations.
NOs- vaporization from the TEOM was caused by the
temperature of the TEOM filter (— 30 - 50 0C) rather
July 2009
A-25
DRAFT-DO NOT CITE OR QUOTE

-------
Site I Period I Sampler I Configuration
Summary of Findings
CONTINUOUS SAMPLER FLOW RATE (L/MIN) INLET	DRYER
TEMPERATURE
OTHER
D-TEOM
16.7
30 °C
Nafion
PM2.6
Nano-BAM
16.7
Ambient
None
— 150 nm cut-
(BAM-1020 with d50



point at 16.7
148 ± 10 nm inlet)



L/min
SMPS-3936
0.3
Ambient
None
Number to mass
assuming
spherical
particles of 1.6
g/cc density
than the pressure drop across the filter.
Vaporization from the TEOM had a time constant between
10 to 100 min, depending on ambient and TEOM filter
temperatures; the vapor pressure, and the extent of vapor
saturation upstream and downstream of the TEOM filter.
The mass measured during 5 min periods (ESP-on and off
cycle in D-TEOM) provides an estimate of the dynamic
vaporization losses.
Chakrabarti et al.111
Good agreement between MOUDI PM0.15 and Nano-BAM
PM0.15 (MOUDI PM0.15 - 0.97 Nano-BAM PM0.15 +
0.60; R2 - 0.92; n - 24)
Nano-BAM captured peak PM0.15 concentrations not
quantified by SMPS. Potential particle agglomeration
(with resulting high surface areas) caused SMPS to
include particles in the accumulation- rather than
ultrafine-mode, since mobility diameter is a function of
surface area.
RUBIDOUX, CA; 08/15/01 to 09/07/01, 07/01/03 to 07/31/03. Rubidoux is located in the eastern section of
the South Coast Air Basin (SoCAB) in the north-west corner of Riverside County, 78 km downwind of the
central Los Angeles metropolitan area and in the middle of the remaining agricultural production area in SoCAB.
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE1

DENUDER"
PC-BOSS PM2.6
150
Teflon (W)

CIF
R&P-2025 PM2.6 FRM
16.7
Teflon (n/a)

None

CONTINUOUS SAMPLER
FLOW RATE (L/MIN)
INLET
TEMPERATURE
DRYER
OTHER
TEOM
16.7
50 °C
None
PM2.6
FDMS-TEOM
16.7
30 °C
Nafion
PM2.6
D-TEOM
16.7
30 °C
Nafion
PM2.6
RAMS
16.7
30 °C
Nafion
PM2.6
Denuders used
CAMM
0.3
n/a
None
PM2.6
Radiance Research M903
n/a
n/a
Nafion
bscat
Radiance Research M903 n/a
n/a
None
bscat
Grover et al.66 (2003 measurements):
D-TEOM - (0.98 ± 0.02) FDMS-TEOM + (-0.6 ± 5.3);
R2-0.85; n - 426; excludes 38 data points when FDMS-
TEOM PM2.6 was higher than DTEOM PM2.6 by
— 21 /yg/m3.
RAMS - (0.93 ± 0.02) FDMS-TEOM + (2.4 ± 8.2);
R2 - 0.81; n - 337
FDMS-TEOM - (0.96 ± 0.06) PC-BOSSconstructed mass
+ (-0.3 ± 3.9); R2 - 0.90; n - 33
R&P-2025 FRM - (0.96 ± 0.06) FDMS-TEOM +
(-9.3 ± 3.9); R2 - 0.90; n - 29
The R&P-2025 FRM PM2.6 was, on avg, ~ 32% lower
than FDMSTEOM. Losses of NH4NO3 and organics can
account for the difference.
TEOM @ 50 °C PM2.6 was consistently lower than FDMS-
TEOM, DTEOM or RAMS and was, on avg, ~ 50% lower
than FDMS TEOM. This difference is due to loss of semi-
volatile NOs- and organics from the heated TEOM.
FDMS-TEOM and D-TEOM needed little attention from
site operators.
Lee et al.76 ( 2 0 01 measurements)
D-TEOM PM2.6 and Radiance Research M903s light
July 2009	A-26	DRAFT - DO NOT CITE OR QUOTE

-------
Site I Period I Sampler I Configuration
Summary of Findings
scattering (with and without dryers) showed good
correlation.
D-TEOM - (3.69 ± 0.09) Radiance Research M903no-
dryer + (2.74 ± 0.89); R2 - 0.84; n - 299
D-TEOM - (3.79 ± 0.10) Radiance Research M903dryed
+ (4.08 ± 0.84); R2 - 0.83; n - 312
Radiance Research M903no-dryer - (1.03 ± 0.01)
Radiance Research M903dryed + (0.34 ± 0.05);
R2 - 0.98; n - 513; absorbed water did not affect
relationship to PM2.6.
CAMM and RAMS compared poorly (R2 - 0 to 0.25) with
D-TEOM, Radiance Research M903s and among
themselves.
RAMS correlated well with D-TEOM for PM2.6
> 30/yglm3 due to RAMS's efficient particle collection of
larger particle sizes (historically associated with high
mass loadings at this site) in the PM2.6 size range.
D-TEOM PM2.6 correlated well with ADI-N sized NO3
(R2 - 0.62) and 0C by Sunset OCEC (R2 - 0.61)
suggesting that D-TEOM measured PM2.6 mass with
minimum loss of SVOCs. RAMS showed R2 of 0.20 (IMO3-)
to 0.30 (OC), while CAMM showed no correlation.
LINDON, UT; 01/29/03 to 02/12/03



Grover etal.66
SAMPLER FLOW RATE (L/MIN)
FILTER TYPE1

DENUDER1
RAMS required regular maintenance.
" RAMS - (0.92 ± 0.03) FDMS-TEOM + (1.3 ± 3.9);
R2 - 0.69; n - 332
PC-BOSS PM2.1, 150
Teflon (W)

CIF




PC-BOSS constructed mass - (0.89 ± 0.21) FDMS-
-TEOM + (1.8 ± 2.8); R2 - 0.66; n - 11
TEOM @ 30 °C PM2.6 was consistently lower than FDMS-
. TEOM and the difference was consistent with
CONTINUOUS SAMPLER FLOW RATE (L/MIN)
INLET
TEMPERATURE
DRYER
OTHER
TEOM 16.7
30 °C
None
PM2.6
concentrations SVOCs and NH4NO3 measured by PC-
- BOSS.
FDMS-TEOM 16.7
30 °C
Nafion
PM2.6
RAMS 16.7
30 °C
Nafion
PM2.6 Denuder
used

PHILADELPHIA, PA; 07/02/01 to 08/01/01 At water treatment center in a grassy field surrounded by mixed
deciduous and pine trees on three sides and a river on the other. Within 0.5 km of Interstate I-95 and within 30
km from downtown Philadelphia.
Lee et al.73
Radiance Research M903dryer - (0.78 ± 0.01) Radiance
Research M903no dryer + (0.30 ± 0.03); R2 - 0.95
SAMPLER FLOW RATE (L/MIN)
FILTER TYPE1

DENUDER"
Radiance Research M903s vs. CAMM, R2 - 0.78
Harvard Impactor PM2.B 10
Teflon (n/a)

n/a
Radiance Research M903s vs. RAMS, R2 - 0.63
Radiance Research M903svs. SES-TEOM, R2 - 0.72








CAMM - (0.60 + 0.03) SES-TEOM + (2.0 + 0.42);
R2 - 0.71; N - 185
RAMS - (0.71 ± 0.04) SES-TEOM + (2.51 ± 0.59);
CONTINUOUS SAMPLER FLOW RATE (L/MIN)
INLET
TEMPERATURE
DRYER
OTHER
SES-TEOM 16.7
35 °C
Nafion
PM2.6
R2 - 0.63; N - 185
- RAMS - (0.93 ± 0.06) CAMM + (2.44 ± 0.68);
R2 - 0.55; N - 185
CAMM 0.3
n/a
Nafion
PM2.6
RAMS 16.7
30 °C
Nafion
PM2.6 TEA & CIF
denuders With
particle
concentrator
Both RAMS and CAMM under-measured ambient PM2.6.
CAMM - (0.70 ± 0.06) HI + (0.16 ± 0.96); R2 - 0.87;
N - 22
SES-TEOM - (1.0 ± 0.10) HI + (-0.68 ± 1.74);
Radiance Research M903 n/a
n/a
Nafion
bscat
R2 - 0.89; N - 15
Radiance Research M903 n/a
n/a
None
bscat

BALTIMORE SUPERSITE, MD; 05/17/01 to 06/11/01. Located near;
1 freeway and bus yard.
Lee et al.73
SAMPLER FLOW RATE (L/MIN)
FILTER TYPE

DENUDER
Radiance Research M903dryed - (0.65 ± 0.02) Radiance
July 2009
A-27
DRAFT-DO NOT CITE OR QUOTE

-------
Site 1 Period / Sampler / Configuration


Summary of Findings
RAAS-100 PM2.1, FRM 16.7
Teflon

None
Research M903no dryer + (1.80 ± 0.20); R2 - 0.75,
suggesting influence from particle-bound water.
High correlation (R2 - 0.75) between Radiance Research




CONTINUOUS SAMPLER FLOW RATE (L/MIN)
INLET
TEMPERATURE
DRYER
OTHER
M903s.
Poor correlation among the continuous instruments.
SESTEOM 16.7
35 °C
Nafion
PM2.6
Radiance Research M903s did not follow PM2.6
concentrations measured by other continuous
instruments.
CAMM 0.3
n/a
Nafion
PM2.6
RAMS 16.7
30 °C
Nafion
PM2.6 TEA & CIF
denuders; No
particle
" CAMM - (0.32 ± 0.07) SES TEOM + (9.45 ± 1.61);
R2 - 0.14; N - 120
RAMS - (0.82 ± 0.10) SES-TEOM + (6.41 ± 2.09);
R2 - 0.38; N - 120
Radiance Research M903 n/a
n/a
Nafion
bscat
RAMS - (0.71 ± 0.12) CAMM + (11.3 ± 2.23);
Radiance Research M903 n/a
n/a
None
bscat
' R2 - 0.21; N - 120
CAMM - (0.80 ± 0.29) RAAS-100 FRM +
(¦0.83 ± 5.85); R2 - 0.60; N - 7
RAMS - (1.05 ± 0.12) RAAS-100 FRM + (4.80 ± 2.60);
R2 - 0.90; N - 11
SESTEOM - (0.86 ± 0.10) RAAS-100 FRM +
(2.96 ± 1.99); R2 - 0.90; N - 10
SEATTLE, WA; 01/28/01 to 02/21/01
Urban area near major highway and interstate, 8 km southeast of downtown.
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE'
1
DENUDER"
MASS PM2.6
16.7
Teflon (n/a)

NayCO:; denuder

CONTINUOUS SAMPLER
FLOW RATE (L/MIN)
INLET
TEMPERATURE
DRYER
OTHER
SESTEOM
16.7
30 °C
Nafion
PM2.6
CAMM
0.3
Ambient
Nafion
PM2.6
RAMS
16.7
30 °C
Nafion
PM2.6 TEA & CIF
denuders
Radiance Research M903
n/a
n/a
Nafion
bscat
Radiance Research M903
n/a
n/a
None
bscat
NEW YORK SUPERSITE, NY; 01/01/03 to 12/31/04
Urban site located at Queens College, NY, about 14 km west of Manhattan, within 2 km of freeways, and within
12 km of international airports. A rural site was located at Pinnacle State Park surrounded by golf course, picnic
areas, undeveloped forest lands, and no major cities within 15 km.
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE'
1
DENUDER"
R&P-2025 PM2.6 FRM
16.7
Teflon (n/a)

None
R&P-2300 PM2.6
16.7
Teflon (n/a)

None

CONTINUOUS SAMPLER
FLOW RATE (L/MIN)
INLET
TEMPERATURE
DRYER
OTHER
TEOM
16.7
50 °C
None
PM2.6
FDMS-TEOM
16.7
30 °C
Nafion
PM2.6
BAM
16.7
"smart" heater on @ RH >44%
PM2.6
Lee et al.108
Radiance Research M903dryed - 0.94 ± 0.00 Radiance
Research M903no dryer; R2 - 1.0.
Correlation of Radiance Research M903 vs. SES TEOM,
R2 - 0.80, while that of Radiance Research M903 with
CAMM was R2 - 0.84 and with RAMS was R2 - 0.72.
CAMM - (1.07 ± 0.05) RAMS + (1.03 ± 0.55);
R2 - 0.61
SES TEOM - (0.95 ± 0.03) RAMS + (1.24 ± 0.38);
R2 - 0.72
SES TEOM - (0.87 ± 0.03) CAMM + (0.55 ± 0.37);
R2 - 0.74
SES TEOM likely lost semi-volatile organic matter.
Continuous PM2.6 samplers were similar to filter PM2.6
sampler. Number of samples was small (~ 7).
Some SES TEOM mass values were less than MASS filter
values suggesting that loss of mass is likely for a SES-
TEOM at 30 °C, particularly during the cold season.
Schwab et al.67
FDMS-TE0M had operational difficulties resulting in low
data capture (65% at urban site and 57% at rural site).
BAM had data captures greater than 95% at both sites.
Urban site:
BAM - (1.02 ± 0.02) FDMS-TEOM + 1.72; R2 - 0.93;
n - 244
FDMS-TEOM - (1.25 ± 0.02) FRM - (0.63 ± 0.26);
R2 - 0.95; n - 238
BAM - (1.28 ± 0.03) FRM + (1.27 ± 0.38); R2 - 0.88;
n - 320
Rural site:
FDMS-TEOM - (1.09 ± 0.02) FRM - (0.004 ± 0.18);
R2 - 0.95; n - 349
PM2.6 FDMS-TEOM >FRM >TE0M50°C, suggesting
that FRM captured a fraction, but not all, of the volatile
components. TE0M50°C volatilizes PM2.6, particularly
during winter.
July 2009	A-28	DRAFT - DO NOT CITE OR QUOTE

-------
Site I Period / Sampler / Configuration
Summary of Findings
aFilter Manufacturer in parentheses - W: Whatman, Clifton, NJ; P: Pall-Gelman, Ann Arbor, Ml; S: Schleicher & Schnell. Keene, NH; nla: not available or not reported.
bNa2C(k Sodium carbonate; NaHC03: Sodium bicarbonate CIF: Charcoal Impregnated Filter; FEP: Fluorinated Ethylene Propylene copolymer; TEA: Triethanolamine; TSP: Total Suspended PM.
e37 mm filter.
d37-mm after-filter for stages smaller than 0.16 fjm and 47-mm for higher stages.
"Equivalence requires correlation coefficient (r) > 0.97, linear regression slope 1.0 ± 0.05 and an intercept 0 ± 1 /yglm3; Comparability requires r > 0.9 and linear regression slope equal 1 within 3
standard errors and intercept equal zero within 3 standard errors; Predictability requires r > 0.9. 91,112
Source: 'Chow (1995,077012); 2Watson and Chow (2001,157123);3 Watson et al. (1983, 045084); "Fehsenfeld et al. (2004,156432); 'Solomon et al. (2001,156993); 6Mikel (2001,156762);
'Mikel (2001,156762); "Watson et al. (1999,020949); 'Solomon and Sioutas (2006,156995); '"Graney et al. (2004, 053756); "Tanaka et al. (1998,157041); "Pancras et al. (2005,098120);
"John et al. (1988,045903); "Hering and Cass (1999, 084958); "Fritz et al. (1989, 077387); '"Hering et al. (1988, 036012); "Solomon et al. (2003,156994); '"Cabada et al. (2004,148859);
"Fine et al. (2003,155775); 2°Hogrefe et al. (2004, 099003); 2,Drewnick et al. (2003, 099160); 22Watson et al. (2005,157125); 23Ho et al. (2006,156552); 24Decesari et al. (2005,144536);2S
Mayol-Bracero et al. (2002, 045010); 26Yang et al. 2003;21 Tursic et al. (2006,157063); 2"Mader et al.(2004,156724); 29Xiao and Liu (2004, 056801); 30Kiss et al. (2002,156646); ''Cornell and
Jickells (1999,156367);32 Zheng et al. (2002, 026100); 33Fraser et al. (2002,140741);34 Fraser et al. (2003, 042231) 3sSchauer er al. (1996, 051162); 3sFine et al. (2004); 3,Yue et al. (2004);
3"Rinehart et al. (2006,115184); 39Wan and Yu (2006,157104); 40Poore (2000, 012839); "Fraser et al. (2003a); 42Engling et al. (2006,156422); 43Yu et al. (2005); 44Tran et al. (2000); 4sYao et al.
(2004,102213); 4sLi and Yu (2005,156692); "Henning et al. (2003,156539); 4"Zhang and Anastasio (2003,157182); "Enmenegger et al. (2007,156418); s0(Watson et al., 1989,157119);
"Greaves et al. (1985,156494); "Waterman et al. (2000,157116); s3Waterman et al. (2001,157117); s4Falkovich and Rudich (2001,156427); ssChow et al. (2007,156354); "Miguel et al.
(2004,123260); "Crinmins and Baker (2006, 097008); s"Ho and Yu (2004,156551); s9Jeon et al. (2001,016636); ""Mazzoleni et al. (2007, 098038); s,Poore (2000, 012839); ^Butler et al.
(2003); G3Chow et al. (2006c); "Russell et al. (2004, 082453); ssGrover et al. (2006); ssGrover et al. (2005, 090044); "Schwab et al. (2006b); ""Hauck et al. (2004); s9Jaques et al. (2004,
155878); '"Rupprecht and Patashrick (2003,157207); "Pang et al (2002, 030353); "Eatough et al. (2001,010303); ,3Lee et al. (2005b); ,4Lee et al. (2005,156680); ,sBabich et al. (2000,
156239); ,6Lee et al. (2005c); "Lee et al. (2005b); '"Anderson and Ogren (1998,156213);,9Chung et al. (2001,156357); ""Kidwell and Ondov (2004,155898); "'Lithgow et al. (2004,126616);
•"Weber et al. (2003,157129); B3Harrison et al. (2004); "Rattigan et al. (2006,115897); BSWittig et al. (2004,103413); ""Vaughn et al. (2005,157089); "'Chow et al. (2005b); "Weber et al
(2001, 024640); "Schwab et al. (2006a); "Lim et al. (2003,156697); 9,Watson and Chow (2002,037873); ^Venkatachari et al. (2006,105918); 93Bae et al. (2004a); "Arhami et al. (2006,
156224); 9sPark et al. (2005,156843); ""Bae et al. (2004b); 9,Chow et al. (2006a); ""Arnott et al. (2005); "Bond et al. (1999); '""Virkkula et al. (2005,157097); ""Petzold et al. (2002,156863);
,02Park et al. (2006); ""Arnott et al. (1999, 020650); ""Peters et al. (2001,017108); "Pitchford et al. (1997,156872); '""Rees et al. (2004,097164); ""Watson et al. (2000); '""Lee et al.(2005,
156680); ,09Hering et al. (2004,155837); ""Watson et al. (1998); '"Chakrabarti et al. (2004); "2Mathai et al. (1990,156741); '"Kidwell and Ondov (2001, 017092); '"Stanier et al. (2004);
'"Khlystov et al. (2005,156635); ""Takahama et al. (2004,157038); '"Chow et al. (2005,156348); ""Zhang et al. (2002); ""Subramanian et al. (2004, 081203); ,2°Chow et al. (2006b); "'Birch
and Cary (1996); ,22Birch (1998, 024953); ,23Birch and Cary (1996); ,24NI0SH (1996,156810); ,2sNI0SH (1999,156811); ,2"Chow et al. (1993); ,2,Chow et al. (2007); ""Ellis and Novakov (1982,
156416); ""Peterson and Richards (2002,156861); ""Schauer et al. (2003); "'Middlebrook et al. (2003, 042932); ,32Wenzel et al. (2003,157139);133Jimenez et al. (2003); ,34Phares et al. (2003,
156866); "sQin and Prather (2006,156895); ""Zhang et al. (2005); "'Bein et al. (2005,156265); ""Drewnick et al. (2004a); ""Drewnick et al. (2004b); "Lake et al. (2003,156669); ,4,Lake et
al. (2004,088411)
Table A-12. Summary of element and liquid water content measurement comparisons.
SITE/PERIOD/SAMPLER
SUMMARY OF FINDINGS
College Park, MD; 11/18/1999 to 11/19/1999, 11/22/1999
Adjacent to a parking lot in the University of Maryland campus,
influenced by motor vehicles, coal-fired power plants and
incinerators —21 km southwest of site and regionally transported
material.
Concentrated Slurry/Graphite Furnace Atomic Absorption
Spectrometry (GFAAS) (collectively known as Semi-
Continuous Elements in Aerosol Sampler, SEAS)
Ambient air is pulled in at a flow rate of 170 L/min. Particles are
grown using steam injection to about 3 to 4 pm in diameter, which
are then concentrated and separated from the air stream in the form
of a slurry using impactors. The slurry is collected in glass sample
vials, which are subsequently analyzed by GFAAS in the laboratory.
Kidwell and Ondov (2004,155898; 2001, 017092)
Overall collection efficiency (of the entire system) measured using latex particles was 40% for
particles initially 0.1 to 0.5 pm in diameter, increasing with size to 68% for particles 3 //m in
diameter. Major losses were in the virtual impactor major flow channel and in the condensers.
Six elements were detected simultaneously, limited by spectral interference and the minimum
detectable limit (MDL). Twelve elements (Al, Cr, Mn, Fe, Ni, Cu, In, As, Se, Cd, Sb, and Pb) were
measured.
MDLs ranged from 3.2 picogram (pg - 10'12 gram) to 440 pg.
Comparison with NIST standards showed good agreement, except for Al, Cr and Fe, due to poor
atomization. The method was valid for dissolved solutions, but not for large particles (>
Overall avg relative standard deviation (RSD) was 20 to 43% by error propagation, mainly due to the
collection and analytical efficiencies.
There were possible memory effects due to particle adhesion to impactor collection surfaces.
Lower MDLs may be possible through redesign and introduction of a wash cycle between samples. A
2.5 fjm inlet might improve analytical efficiency by removing coarse particles.
Pittsburgh Supersite, PA; 08/26/2002 to 09/02/2002
6 km east of downtown in a park on the top of a hill.
Laser Induced Breakdown Spectroscopy (LIBS)
Ambient air was concentrated using a PM2.6 inlet and a virtual
impactor. The concentrated stream was transported through a
Teflon tube to the sample cell of the LIBS system. The sample cell
was excited using a Nd: YAG laser. The resulting plasma was
collected and focused into a spectrometer, generating spectra
characteristic of different elements.
Lithgow et al. (2004,126616)
Calibration was done by sampling particle-laden streams with known metal concentrations. Good
linear fits with correlation coefficients 0.97 to 0.99
Seven metals (Na, Mg, Al, Ca, Cr, Mn, and Cu) were analyzed.
The MDLs were in the order of femtograms (fg - 10'16 gram) per sample.
This system has the capability of identifying the components, quantifying them and also giving a
particle size distribution. Mass was underestimated because of missing small particles.
July 2009
A-29
DRAFT-DO NOT CITE OR QUOTE

-------
SITE I PERIOD I SAMPLER
SUMMARY OF FINDINGS
Pittsburgh Supersite, PA; 07/01/2001 to 08/31/2001,
01/01/2002 to 07/01/2002.
6 km east of downtown in a park on the top of a hill.
Dry Ambient Aerosol Size Spectrometer (DAASS)
Measures the aerosol size distribution (using nano-SMPS, SMPS and
APS) alternatively, at ambient relative humidity (RH) (ambient
channel) and at low RH (18 ± 6%) (dry channel). A comparison of
the two size distributions provides information on the water
absorption and change in size due to RH.
Stanier et al. (2004, 095955); Khlystov et al. (2005,156635)
Measured water content ranging from less than 1 //g|m3 to 30 //g|m3, constituting < 5% to 100% of
the dry aerosol mass.
Small differences between dry and ambient channels of the DAASS. Number concentrations were
within 5% of each other.
Additional sources of error are associated with temperature differences between measured outdoor
ambient temperature and the temperature at which the ambient measurement channel was
maintained. Although the measurement system was placed in a ventilated enclosure, it was ~4 °C
higher than ambient temperature during July 2001. During winter, the system was maintained at a
minimum temperature of 9 °C, while the outdoor temperature dropped to -5 °C. This caused
differences in RH sensed by the system in the ambient channel versus the actual outdoor RH.
RH differences cause underestimation of the particle number at sizes < 200 nm and an
overestimation at sizes > 200 nm. This causes the volume growth factor to be higher by 2 to 14%,
with the highest bias occurring at high RH and low temperature (92% outside RH and -5 °C|.
The difference in temperature might also lead to evaporation of semi-volatile components such as
NH4N03. For the winter period, it was estimated that, for the worst case, the volume growth factor
would be underestimated by about 10% for 60 ¦ 90% RH.
Insufficient purging of dry air between the dry and ambient cycles (implying the need for supplemental
vacuum power during the vent stages) causes uncertainties in estimated growth factors. Correction
factors were between 0.97 and 1.03.
Water content estimated by DAASS can be used to evaluate the thermodynamic models. For the
Pittsburgh study, the models underestimated the water content by 37%.
Data from DAASS showed that the aerosol was wet even at ambient RH less than 30%.
Table A-13. Summary of PM2.5 l\IO3 measurement comparisons.
SITE I PERIOD I SAMPLER I CONFIGURATION
SUMMARY OF FINDINGS
ATLANTA SUPERSITE, GA: 8/3/99 to 9/1/99 4 km NW of downtown, within 200 m of a bus maintenance yard and
several warehouse facilities, representative of a mixed commercial-residential neighborhood.
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE1
DENUDER"
R&P-2000 FRM
16.7
Quartz (P)
None
RAAS-400
24
Nylon (P)
MgO
SASS
6.7
Nylon (P)
MgO
MASS-400
16.7
Teflon (P)-Nylon (P
NayCO:;
MASS-450
16.7
Quartz (P)
None
R&P-2300
10
Nylon (P)
Na2C03
VAPS
15
Polycarbonatec (front 81 back-up)
NayCO:;
URG-PCM
16.7
Teflon (P)-Cellulose-fiber (W
NayCO:!
ARA-PCM
16.7
Teflon (n/a)-Nylon (n/a)
NayCOii/Citric acid
PC-BOSS (TVA)
105
Teflon (W)-
Nylon (P)
CIF
PC-BOSS (BYU)
150
Teflon (W)-
Nylon (P)
CIF
PC-BOSS (BYU)
150
Quartz (P)-
CIF (S)
CIF
MOUDI-100
30
Teflon (n/a-
Quartz (n/a
None
Solomon et al
PM2.E N0:i from each sampler was compared to
the all-sampler avgs, called the filter relative
reference (filter RR) value. Overall agreements
were within 30-35% of filter RR.
Wide scatter from paired comparisons, possibly
due to volatilized NO3 , differences in denuder
design and filter types, and low concentrations
(close to analytical uncertainty).
A small positive artifact (few tenths of /yg/m3)
might be present when using Na2
CO:; impregnated filters, due to possible
collection (and subsequent oxidation) of HONO
and NO2 on carbonate-impregnated filters. In
addition, glycerol in Na2C03 coated denuders
may contaminate the filters downstream.
PM2.6 NOs" R&P-2000 FRM and M0UDI-100
samplers are consistently lower than other
samplers.
Weber et al.82
Hourly PM2.6 N0:i were compared to all-sampler
averages (continuous RR), similar to the
approach used for integrated filter samplers.
Overall agreements were within ± 20-30%
(or ± 0.2 /yg/m3) except for ARA-N.
Except for ARA N, good correlations
(R2 - 0.70 to 0.90) were found during the
July 2009
A-30
DRAFT-DO NOT CITE OR QUOTE

-------
SITE I PERIOD I SAMPLER / CONFIGURATION
SUMMARY OF FINDINGS
CONTINUOUS SAMPLER FLOW RATE (L/MIN)
DENUDER
ANALYSIS METHOD"
ADI-N 1
Activated Carbon
NOx Chemiluminescence
ARAN 3
Potassium iodide (Kl)
and dual sodium
chlorite (NaCIOz)
NOx Chemiluminescence
PILS-IC 5
Two URG annular glass
denuders in series
containing citric acid
and CaC03
IC
ECN 16.7
Rotating annular wet
denuder system
IC
TT 5
Wet parallel plate
denuder
IC
PITTSBURGH SUPERSITE, PA; 7/1/01 to 8/1/02 6km east of downtown in a park on the top of a hill
SAMPLER FLOW RATE (L/MIN)
FILTER TYPE1
DENUDER"
MOUDI-110 30
Teflon (W
Teflon (W)-
None
CMU 16.7
Nylon (W)
MgO/Citric acid
R&P-2000 FRM 16.7
Teflon (W)
None
second half of the study. The poor performance
of ARA N was probably due to an inefficient
denuder (25-60% efficient) resulting in high
background.
Large discrepancies between continuous and
filter RR, probably due to low ambient
concentrations (study avg - 0.5 /yg/m3) near
the detection limit ( — 0.1 //g|m3, except for
ARA-N, which had 0.5//g|m3).
The ARA-N was within 13%, ADI-N, ECN and
PILS-IC within 18% and TT within 26% of filter
RR (all < 0.2/yg/m3 difference).
Filter samples showed more variability (Relative
Standard Deviation, RSD - 22%) than
continuous measurements (RSD - 13%). This is
probably due to sampling artifacts in filter
samples; NOs'volatilization in continuous
monitors is expected to be minimal due to
shorter averaging times and rapid stabilization
in solutions.
Cabada et al.Is; Takahama et al.116
More than 70% (— 0.5 /yglm3) of NO3 mass
was lost from MOUDI samplers during summer.
MOUDI NOs - 0.27 CMU; R2 - 0.40; Summer
MOUDI NOs - 0.99 CMU; R2 - 0.49; winter
85
Wittig et al.
Avg conversion efficiency to NOx (tested using
NH4NO3 solution) was 0.85 ± 0.08. Gas
analyzer efficiency was stable at 0.99 ± 0.04.
Corrections were made for instrument offset,
software calculation error, conversion
efficiency, gas analyzer efficiency, vacuum
drift, and sample flow drift. The overall avg
correction was 8%, ranging from -62% to 93%.
Data Recovery > 80%. Data loss was
associated with vacuum pump failures and
excessive flash strip breakage.
R&P-8400N - 0.83 CMU + 0.20//gfm3;
R2 - 0.84
Under estimation in the R&P-8400N could be
due to incomplete particle collection or
incomplete conversion of various forms of NO3'.
Used co-located filter measurements for final
calibration.
FRESNO SUPERSITE, CA and other CRPAQS sites; 12/2/99 to 2/3/01
Located 5.5 km northeast of downtown in a mixed residential-commercial neighborhood.
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE1
DENUDER
DRISFS
113
Quartz (Pellulose
AI2O3
RAAS-400
24
Quartz (P)-Nylon (P)
NayCO:!
RAAS-400
24
Quartz (P)-Quartz (P)
None
RAAS-100 FRM
16.7
Quartz (P)
None

CONTINUOUS SAMPLER
FLOW RATE (L/MIN)
DENUDER
ANALYSIS METHOD"
R&P-8400N
Activated Carbon
NOx Chemiluminescence
Chow et al.87
Maximum NO3' volatilization was observed
during summer (Jun - Aug), while the lowest
volatilization was observed during winter
(Dec-Feb).
Seasonal avg volatilized NO3' in particulate NO3'
(PN0:i , the sum of non-volatilized and volatilized
NO3) ranged from less than 10% during winter
to more than 80% during summer.
Volatilized NH4NO3 accounted for 44% of actual
PM2.E mass (i.e., measured mass plus volatilized
NH4NO3) in Fresno during summer.
Front-quartz non-volatilized NO3' concentrations
were similar for DRISFS (0.52 ± 0.26 /yglm3)
July 2009	A-31	DRAFT - DO NOT CITE OR QUOTE

-------
SITE I PERIOD I SAMPLER / CONFIGURATION
SUMMARY OF FINDINGS
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE1
DENUDER
PC-BOSS
150
Teflon (W)- Nylon (P)
CIF

CONTINUOUS SAMPLER
FLOW RATE (L/MIN)
DENUDER
ANALYSIS METHOD"
R&P-8400N
5
Activated Carbon
NOx Chemiluminescence
Dionex-IC
5
Parallel plate wet denuder
IC
and RAAS-100 FRM (0.81 ± 0.33 //g/m3) for
warm months (May-Sep). With preceding
denuders, the DRI-SFS
PNO3 concentration (3 ± 1.9 //g/m3) was much
higher than the RAAS100 FRM NO3', suggesting
that the FRM sampler removed gaseous nitric
acid (HNO3) resulting in NO3' volatilization. FRM
Teflon-membrane filters are subject to similar
N03'losses.
Chow etal.117
High correlation (R2 >0.90) between 24-h avg
R&P-8400N NOs and SFS filter NOs"
concentrations, but R&P-8400N NOs'was 7 to
25% lower than filter NO3'.
Limited comparison (n < 15) with filter
samples at Bakersfield showed that the slopes
were close to unity during early morning hours,
while they decreased during the afternoon
hours, indicating possible loss of NO3 by the
R&P-8400N instrument.
The R&P-8400N required substantial
maintenance and careful operation.
Grover et al.65
Dionex-IC NOs - (0.71 ± 0.04) PC-BOSS NOs
+ (3.2 ± 1.1); R2 - 0.91; n - 29
R&P-8400N - (1.10 ± 0.06) PC-BOSS NOs ¦
(0.8 ± 1.8); R2 - 0.93; n - 29
R&P-8400N - (0.55 ± 0.01) Dionex-IC +
(1.4 ± 1.8); R2 - 0.75; n - 493
R&P-8400N measured less than DIONEX IC,
particularly at high RH. R&P-8400N may suffer
incomplete flash vaporization under conditions
of high RH.
BALTIMORE SUPERSITE, MD; 2/14/02 to 11/30/02
Adjacent to a parking lot in the University of Maryland campus, influenced by motor vehicles, coal-fired power plants and
incinerators —21 km southwest of site and regionally transported material.
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE1
DENUDER
SASS
6.7
Nylon (n/a)
MgO

CONTINUOUS SAMPLER
FLOW RATE (L/MIN)
DENUDER
ANALYSIS METHOD"
R&P-8400N
Activated Carbon
NOx Chemiluminescence
Harrison etal.33
Corrections were made to R&P-8400N data for
software calculation error, conversion
efficiency, gas analyzer efficiency, vacuum drift
and sample flow drift.
The relative uncertainty of R&P-8400N
measurements averaged 8.7%, ranging from
6.3% to 23%.
Data capture >95%.
R&P-8400N underestimated SASS filter NO3' by
July 2009	A-32	DRAFT - DO NOT CITE OR QUOTE

-------
SITE I PERIOD I SAMPLER / CONFIGURATION
SUMMARY OF FINDINGS
~ 33%, attributed to variations in conversion
efficiency, matrix effects, and impaction
efficiency. This suggested a true conversion
efficiency of 68% as compared to an avg
conversion efficiency of R&P-8400N to NOx
(tested using potassium nitrate solution) of
0.90 ± 0.04.
Large errors occurred when the concentrations
were near the detection limit, when the
temperature difference (between instrument
and ambient) was large, and when the ambient
relative humidity (RH)was < 40%. Ridged
flash strips produced lower dissociation losses
than flat strips.
Reliable measurements were obtained when the
instrument-outdoor temperature differences
were minimal and when grooved/ridged flash
strips were used. A co-located filter
measurement was used for final corrections.
NEW YORK SUPERSITE, NY; 06/29/01 to 08/05/01 and 07/09/02 to 08/07/02
Urban site located at Queens College, NY, about 14 km west of Manhattan, within 2 km of freeways, and within 12 km of
international airports. Rural site located at Whiteface mountain, 600 m above sea level, in a clearing surrounded by
deciduous and evergreen trees and no major cities within 20 km of the site.
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE1
DENUDER
R&P-2300
10
Nylon (n/a)
Na2C03

CONTINUOUS SAMPLER
FLOW RATE (L/MIN)
DENUDER
ANALYSIS METHOD"
R&P-8400N
5
Activated Carbon
NOx Chemiluminescence
PILS-IC
5
Na2C03 and citric
acid
IC
AMS
0.1
None
Mass Spectrometry
NEW YORK SUPERSITE, NY; 10/01 to 07/05 (urban), 07/02 to 07/05 (rural) Urban site located at a school in South
Bronx, NY in a residential area, within a few kilometers away from major highways and a freight yard (experiencing
significant truck traffic). Rural site located at Whiteface mountain, 600 m above sea level, in a clearing surrounded by
deciduous and evergreen trees and no major cities within 20 km of the site.
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPEA
DENUDER"
R&P-2300
10
Nylon (n/a)
NayCO:!
TEOM-ACCU
16.7
Zefluor
None

CONTINUOUS SAMPLER
FLOW RATE (L/MIN)
DENUDER
ANALYSIS METHOD"
R&P-8400N
5
Activated Carbon
NOx Chemiluminescence
Hogrefe et al.20
Data completeness: 86 ¦ 88% for R&P-8400N,
94 -98% for AMS, and 65 ¦ 70% for PILS-IC.
Some PILS measurements were invalidated
owing to larger aqueous flow caused by bigger
tubing. Larger aqueous flow and inconsistent
water quality affected NO3' concentrations.
R&P-8400N NO3' was lower than R&P-2300
filter NOs". PILS-IC was within 5% of R&P-2300
filter NO3' concentrations.
At the urban site, AMS was within 10% of the
filter N03concentration. At the rural site, AMS
had a slope of 0.51 and R2 of 0.46, compared
with filter NO3'.
Rattigan et al.84
Data capture was more than 94%.
Data were adjusted for span and zero drifts,
conversion efficiency, flow drift, and blanks.
R&P-8400N NO3' was systematically lower
than R&P-2300 filter N03over all concentration
ranges, except at < 1 /yglm3.
Urban: R&P-8400N - 0.59 R&P-2300 NOs +
0.28; R2 - 0.88; n - 305
Rural: R&P-8400N - 0.73 R&P-2300 NOs +
0.01; R2 - 0.90; n — 161; however
concentrations were low with 95% of data
< 1 Z^g/m3.
Required weekly or biweekly maintenance by
trained personnel.
LOS ANGELES SUPERSITE, CA; 7/13/01 to 9/15/01 (Rubidoux) and 9/15/01 to 2/10/02 (Claremont)
Multiple sampling locations in the South Coast Air Basin (SoCAB), including urban "source" sites and downwind "receptor"
sites.
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPEa
DENUDER'
MOUDI
30
Teflon (P
None
HEADS
10
Teflon (n/a) ¦
Carbonate


GF-GF

Fine et al.19
MOUDI - 0.68 HEADS: R2 - 0.88
ADI-N Sized - 0.80 HEADS: R2 - 0.79
ADI-N Sized - 1.12 MOUDI: R2 - 0.53
ADI-N NO3' showed better agreement with
HEADS at lower concentrations, the ADI-N
deviated (biased low) from the HEADS
concentrations at higher NO3 concentrations.
July 2009
A-33
DRAFT-DO NOT CITE OR QUOTE

-------
SITE I PERIOD I SAMPLER I CONFIGURATION
SUMMARY OF FINDINGS
This deviation was attributed to NO3'
vaporization, loss of NO3' associated with
particles less than 0.1 fjm not collected by the
ADI-N sampler, or loss of particles in the ADI-N
inlet tubing.
The underestimation of NO3' by MOUDI
compared to HEADS may be due to NO3'
volatilization from MOUDI stages, since SO42'
comparisons showed MOUDI to explain 85% of
HEADS S042"
ADI-N and MOUDI showed better correlation
(R2 - 0.67) for the 1 to 2 /jm size range NO3
relative to other size ranges (R2< - 0.56).
This is possibly due to NO3' in the form of non-
volatilized sodium nitrate (NaN03) than
volatilized NH4NO3 in the 1-2 //m size range.
Single particle analysis also indicated this
possibility of NaN03 in the 1 to 2/ym range.
Grover et al.66
R&P-8400N - (0.65 ± 0.07) PC-BOSS +
(3.3 ± 2.4); R2 - 0.73; n - 31
At higher concentrations (No numerical value
reported), R&P-8400N N03'was lower than PC-
BOSS NO3 , possibly due to incomplete
volatilization of NH4NO3 in R&P-8400N at
higher concentrations (and higher relative
humidity).
At the urban site, the continuous instruments
correlated well with filter NO3 - measurements
and among themselves (R2 a 0.89). At the rural
site, R2 ranged from 0.61 to 0.83, except for
the AMS versus R&P2300 comparison, with an
R2 of 0.46.
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE"
DENUDER"
PC-BOSS
150
Teflon (W)-Nylon (P)
CIF

CONTINUOUS SAMPLER
FLOW RATE (L/MIN)
DENUDER
ANALYSIS METHOD"
R&P-8400N
5
Activated Carbon
NOx Chemiluminescence
R&P-8400N
5
Activated Carbon
NOx Chemiluminescence
PILS-IC
5
Na2C03 and Citric acid
IC
AMS
0.1
None
Mass Spectrometry
July 2009	A-34	DRAFT - DO NOT CITE OR QUOTE
CONTINUOUS SAMPLER
FLOW RATE (L/MIN)
DENUDER
ANALYSIS METHOD"
ADI-N Sized
0.9
Activated Carbon
NOx Chemiluminescence
RUBIDOUX, CA; 07/01/03 to 07/31/03
Located in the eastern section of SoCAB in the north-west corner of Riverside County, 78 km downwind of the central Los
Angeles metropolitan area and in the middle of the remaining agricultural production area in SoCAB.

-------
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE"
DENUDED
aFilter Manufacturer in parenthesis - W: Whatman, Clifton, NJ; P: Pall-Gelman, Ann Arbor, Ml; S: Schleicher & Schnell. Keene, NH; nla: not available or not reported.
bAb03: Aluminum oxide; GF: Na2C03 impregnated Glass Fiber Filters; IC: Ion chromatography; MgO: Magnesium oxide; Na2C03: Sodium carbonate; NaHC03: Sodium bicarbonate NOx: Oxides of
nitrogen; CIF: Charcoal Impregnated Filter; FEP: Fluorinated Ethylene Propylene copolymer; TEA: Triethanolamine; TSP: Total Suspended PM.
cNa2C03 impregnated.
d37-mm filter.
Source: 'Chow (1995,077012); 2Watson and Chow (2001,157123);3 Watson et al. (1983, 045084); "Fehsenfeld et al. (2004,150432); 'Solomon et al. (2001,150993); sMikel (2001,150702);
'Mikel (2001,150702); "Watson et al. (1999,020949); 'Solomon and Sioutas (2000,150995); '"Graney et al. (2004, 053750); "Tanaka et al. (1998,157041); "Pancras et al. (2005,098120);
"John et al. (1988,045903); "Hering and Cass (1999, 084958); "Fritz et al. (1989, 077387); '"Hering et al. (1988, 030012); "Solomon et al. (2003,150994); "Cabada et al. (2004,148859);
"Fine et al. (2003,155775); 2°Hogrefe et al. (2004, 099003); 2,Drewnick et al. (2003, 099100); 22Watson et al. (2005,157125); 23Ho et al. (2000,150552); 24Decesari et al. (2005,144530);2S
Mayol-Bracero et al. (2002, 045010); 2sYang et al. 2003;21 Tursic et al. (2000,157003); 2"Mader et al.(2004,150724); 29Xiao and Liu (2004, 050801); 3°Kiss et al. (2002,150040); 3,Cornell and
Jickells (1999,150307);32 Zheng et al. (2002, 020100); 33Fraser et al. (2002,140741);34 Fraser et al. (2003, 042231) 3sSchauer er al. (1990, 051102); 3sFine et al. (2004); 3,Yue et al. (2004);
3BRinehart et al. (2000,115184); 39Wan and Yu (2000,157104); 40Poore (2000, 012839); "Fraser et al. (2003a); 42Engling et al. (2000,150422); 43Yu et al. (2005); 44Tran et al. (2000); 4sYao et al.
(2004,102213); 4sLi and Yu (2005,150092); "Henning et al. (2003,150539); 4"Zhang and Anastasio (2003,157182); "Enmenegger et al. (2007,150418); s0(Watson et al., 1989,157119);
"Greaves et al. (1985,150494); "Waterman et al. (2000,157110); s3Waterman et al. (2001,157117); s4Falkovich and Rudich (2001,150427); ssChow et al. (2007,150354); "Miguel et al.
(2004,123200); "Crinmins and Baker (2000, 097008); s"Ho and Yu (2004,150551); s9Jeon et al. (2001,010030); 60Mazzoleni et al. (2007, 098038); 6,Poore (2000, 012839); ^Butler et al.
(2003); s3Chow et al. (2000c); "Russell et al. (2004, 082453); ssGro»er et al. (2000); ssGro»er et al. (2005, 090044); "Schwab et al. (2000b); ""Hauck et al. (2004); s9Jaques et al. (2004,
155878); '"Rupprecht and Patashrick (2003,157207); "Pang et al (2002, 030353); "Eatough et al. (2001,010303); ,3Lee et al. (2005b); ,4Lee et al.(2005,150080); ,sBabich et al. (2000,
150239); ,sLee et al. (2005c); "Lee et al. (2005b); '"Anderson and Ogren (1998,150213);,9Chung et al. (2001,150357); ""Kidwell and Ondo» (2004,155898); "'Lithgow et al. (2004,120010);
""Weber et al. (2003,157129); "'Harrison et al. (2004); "Rattigan et al. (2000,115897); "sWittig et al. (2004,103413); ""Vaughn et al. (2005,157089); "'Chow et al. (2005b); ""Weber et al
(2001, 024040); "Schwab et al. (2000a); 90Lim et al. (2003,150097); 9,Watson and Chow (2002,037873); ^Venkatachari et al. (2000,105918); 93Bae et al. (2004a); "Arhami et al. (2000,
150224);95Park et al. (2005,150843); ""Bae et al. (2004b); 9,Chow et al. (2000a); ""Arnott et al. (2005); "Bond et al. (1999); '""Virkkula et al. (2005,157097); ""Petzold et al. (2002,150803);
,02Park et al. (2000); ,raArnott et al. (1999, 020050); ""Peters et al. (2001); ,osPitchford et al. (1997,150872); ,osRees et al. (2004, 097104); ""Watson et al. (2000); ,0BLee et al.(2005,150080);
'""Hering et al. (2004,155837); ""Watson et al. (1998); '"Chakrabarti et al. (2004); "2Mathai et al. (1990,150741); '"Kidwell and Ondo» (2001,017092); '"Stanier et al. (2004); '"Khlystov et
al. (2005,150035); ""Takahama et al. (2004,157038); '"Chow et al. (2005,150348); ""Zhang et al. (2002); ""Subramanian et al. (2004, 081203); ,20Chow et al. (2000b); "'Birch and Cary
(1990);122Birch (1998,024953); ,23Birch and Cary (1990); ""NIOSH (1990,150810); "sNI0SH (1999,150811); ,2"Chow et al. (1993); "'Chow et al. (2007); ""Ellis and No»ako» (1982,150410);
""Peterson and Richards (2002,150801); ""Schauer et al. (2003); ,3,Middlebrook et al. (2003,042932); ,32Wenzel et al. (2003,157139);133Jimenez et al. (2003); "4Phares et al. (2003,150800);
"sQin and Prather (2000,150895); "Zhang et al. (2005); "'Bein et al. (2005,150205); ""Drewnick et al. (2004a); ""Drewnick et al. (2004b); ""Lake et al. (2003,150009); "'Lake et al. (2004,
088411)
Table A-14. Summary of PM2.5 SO42 measurement comparisons
SITE/PERIOD/SAMPLER/ CONFIGURATION
SUMMARY OF FINDINGS
ATLANTA SUPERSITE, GA: 08/03/99 to 09/01/99
4 km NW of downtown, within 200 m of a bus maintenance yard and several warehouse facilities, representative of;
mixed commercial-residential neighborhood.
SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE1
DENUDER"
R&P-2000 FRM
16.7
Quartz (P)
None
RAAS-400
24
Teflon (P)
None
SASS
6.7
Teflon (P)
None
MASS-450
16.7
Quartz (P)
None
R&P-2300
10
Quartz (P)
NOne
VAPS
15
Quartz (P)
XAD-4
URG-PCM
16.7
Teflon (P)-Cellulose-fiber (W

ARA-PCM
16.7
Teflon (n/a)
NayCOii/Citric acid
ARA-PCM
16.7
Nylon (n/a)
Na2C03/Citric acid
PC-BOSS (TVA)
105
Teflon (W)
CIF
PC-BOSS (TVA)
105
Quartz (P)
CIF
PC-BOSS (BYU)
150
Teflon (W)
CIF
PC-BOSS (BYU)
150
Quartz (P)
CIF
MOUDI-100
30
Teflon (n/a-
Quartz (n/a
None
Solomon et al.17
PM2.6 S042'from each sampler was compared to all-
sampler averages, called the filter relative reference
(filter RR) value. The samplers agreed to within
10% of filter RR, except for the PC-BOSS (TVA)
and MOUDI-100.
While avg mass was within 10%, daily variability
was >50% of filter RR.
All samplers, except for the PC-BOSS (TVA),
correlated well (R2 > 0.90) with daily filter RR.
PC-BOSS (TVA) had instrument leaks.
The R&P-2000 FRM, on avg, agreed within 1% of
filter RR.
MOUDI-100 was ~ 13% low compared to filter RR.
Weber et al.32; Zhang et al.113
Hourly PM2.6 S042'were compared to all-sampler
averages (continuous RR), similar to the approach
used for filter samplers. Overall agreement was
within 16% or 2 /yglm3.
Good correlations (R2 - 0.76 to 0.94) were found
during the second half of the study, except for TT
versus ADI.
Good correlation (R2 - 0.84) was found between
continuous and filter-based S042': Continuous
RR - (1.15 ±0.15), Filter RR +(0.41 ± 1.73)
Variability among continuous SO42 instruments
July 2009
A-35
DRAFT-DO NOT CITE OR QUOTE

-------
SITE/PERIOD/SAMPLER/ CONFIGURATION
SUMMARY OF FINDINGS
CONTINUOUS SAMPLER
FLOW RATE
(L/min)
DENUDER
ANALYSIS METHOD"
ADI-S
2.7
Activated Carbon
SO2, UV Fluorescence
PILS-IC
5
Two URG annular glass
denuders in series
containing citric acid 81
CaC03
IC
ECN
16.7
Rotating annular wet
denuder system
IC
TT
5
Wet parallel plate denuder
IC
(RSD - 13%) was similar to that for NO3
instruments. Filter sample variability was low
(RSD - 8%) indicating more uniformity among
samplers.
The ECN and TT instruments were within 15%,
PILS-IC was within 20% and ADI-S was within
26% of filter RR.
PITTSBURGH SUPERSITE, PA; 070/1/01 to 08/01/02
6km east of downtown in a park on the top of a hill
SAMPLER
FLOW RATE
FILTER TYPE1
DENUDER'

(L/MIN)


MOUDI-110
30
Teflon (W
None
CMU
16.7
Teflon (W)
MgO/Citric acid
R&P-2000 FRM	16.7	Teflon (W) None
CONTINUOUS SAMPLER	FLOW RATE DENUDER	ANALYSIS METHOD"
(L/min)
R&P-8400S	5	Activated Carbon	SO2 UV Fluorescence
Cabada et al„ ls; Takahama et al„ 116
MOUDI S042" 0.80 CMU; R2 - 0.95; Summer
MOUDI S042" 0.97 CMU; R2 - 0.48; winter
Wittig et al.85
Avg conversion efficiency to SO2 (tested using
ammonium sulfate [(NHihSOi] solution) was
0.65 ± 0.07. Gas analyzer efficiency was stable at
0.99 ± 0.06.
Corrections were made for instrument offset,
software calculation error, conversion efficiency,
gas analyzer efficiency, vacuum drift, and sample
flow drift. The overall correction was, on avg, -1 %
and ranged from -90% to 100% for individual
samples.
Data Recovery > 90%. Data loss was associated
with vacuum pump failures or excessive flash strip
breakage.
R&P-8400S (S042 ) - 0.71 CMU + 0.42/yg/m3;
R2 - 0.83
Underestimation is attributed to incomplete particle
collection or incomplete conversion of various forms
of S042"
Used co-located filter measurements for final
calibration.
LOS ANGELES SUPERSITE, CA; 07/13/01 to 09/15/01 (Rubidoux) and 09/15/01 to 02/10/02 (Claremont)
Multiple sampling locations in the South Coast Air Basin (SoCAB), including urban "source" sites and downwind
"receptor" sites.
Fine et al.19
MOUDI explained 85% of HEADS SO42 (R2 - 0.89;
n - 40)
SAMPLER
FLOW RATE
FILTER TYPE1
DENUDER

(L/MIN)


MOUDI
30
Teflon (P
None
HEADS
10
Teflon (n/a) -GFC-GFC
Carbonate
NEW YORK SUPERSITE, NY; 06/29/01 to 08/05/01 and 07/09/02 to 08/07/02	Drewnick et al. Hogrefe et al.20
Urban site located at Queens College, NY, about 14 km west of Manhattan, within 2 km of freeways, and within 12	pata comp|eteness- 89 ¦ 93% for R&P-8400S 94 ¦
km of international airports. Rural site located at Whiteface mountain, 600m above sea level, in a clearing surrounded	gg% jor 81-98% for CASM and 65-70% for
by deciduous and evergreen trees and no major cities within 20 km of the site.	PILS-IC
SAMPLER
FLOW RATE
(L/MIN)
FILTER TYPE1
DENUDER"
The urban site data showed good correlations
(R2 - 0.87 to 0.94) with slopes ranging from 0.97
R&P-2300
10
Nylon (n/a)
NayCO:!
to 1.01. Al the rural site, the variability was large
(R2 - 0.73 to 0.91) with slopes ranging from 0.76
SCS
42
Zefluor (n/a)
None
to 1.32. SO4 from PILS-IC was overestimated by
July 2009
A-36
DRAFT-DO NOT CITE OR QUOTE

-------
SITE/PERIOD/SAMPLER/ CONFIGURATION
SUMMARY OF FINDINGS
TEOM-ACCU
16.7
Zefluor (n/a)
None

CONTINUOUS SAMPLER
FLOW
RATE
(L/min)
DENUDER
ANALYSIS METHOD"
R&P-8400S
5
Activated Carbon
SO2 UV Fluorescence
PILS-IC
5
Na2C03 and Citric acid
IC
AMS
0.1
None
Mass Spectrometry
CASM
5
Na2C03 and Carbon and a
Nafion dryer
SO2 UV Fluorescence
~ 25% when compared to the AMS at the rural
site.
Filter samples were within 5% of each other,
except for comparison of ACCU with R&P-2300 at
the rural site, with high correlations (R2 - 0.97 to
1.0). ACCU underestimated SO42'by —15%.
Continuous versus six-h SCS filter comparisons
showed high R2 (0.91 to 0.95) at the urban site.
Continuous instruments consistently measured
lower SO42'concentrations compared to the SCS
filter measurements (slopes 0.68 to 0.73)
On avg, 85% of the filter-based S042'was measured
by the continuous instruments with consistent
relationships. At the rural site, PILS-IC
overestimated SO42' concentrations (slopes 1.11 to
1.15), AMS and R&P-8400S showed slopes of
0.71-0.74 against SCS and ACCU, while it ranged
from 0.53- 0.68 against R&P-2300.
Error estimates:
Sampling losses: 2-3% for AMS and PILS-IC, 5-10%
for R&P-8400S and none for CASM.
Continuous instruments probably experienced more
inlet transport losses (~ 25%) than filter samplers
due to longer inlet lines.
Small (< 2%) positive artifact was found in filters.
NEWYORK SUPERSITE, NY; 10/01 to 07/05 (urban), 07/02 to 07/05 (rural)
Urban site located at a school in South Bronx, NY in a residential area, within a few kilometers from major highways
and a freight yard (experiencing significant truck traffic). Rural site located at Whiteface mountain, 600m above sea
level, in a clearing surrounded by deciduous and evergreen trees and no major cities within 20 km of the site. The
study by Schwab et al.89 was based at a rural site located at Pinnacle State Park surrounded by golf course, picnic
areas and undeveloped forest lands and no major cities within 15 km.
INTEGRATED SAMPLER
FLOW RATE (L/MIN)
FILTER TYPE1
DENUDER"
R&P-2300
10
Nylon (n/a)
NayCO:!
TEOM-ACCU
16.7
Zefluor
None

CONTINUOUS SAMPLER
FLOW RATE (L/min)
DENUDER
ANALYSIS METHOD"
R&P-8400S
5
Activated Carbon
SO2 pulsed fluorescence
TE-5020
(07114104 to 11101104)
5
Na2C03
SO2 pulsed fluorescence
Rattigan etal.84
Data capture was above 85%. Data loss was
primarily due to frequent flash strip failures, every
2 weeks and without warning.
Data were adjusted for span and zero drifts,
measured conversion efficiency, flow drift, and
blanks.
Calibrations used aqueous standards of (NH4)2S04
and oxalic acid solution in 1: 4 ratio. Lower
fractions of oxalic acid showed lower conversion
efficiencies.
Urban South Bronx site:
R&P-8400S - 0.82 TEOM-ACCU + 1.15;
R2 - 0.84; n - 513
R&P-8400S - 0.74 R&P-2300 + 1.14; R2- 0.81;
n - 322
Rural Whiteface mountain:
R&P-8400S - 0.75 TEOM-ACCU + 0.22;
R2 - 0.95; n - 207
R&P-8400S - 0.78 R&P-2300 + 0.17;
R2 - 0.85; n - 198
Required weekly or biweekly maintenance by
trained personnel
Schwab et al.39
TE-5020 - 0.78 ACCU - 0.2; R2 - 0.94
Similar studies at St. Louis, MO, show slopes near
unity. This suggests that the instrument is sensitive
to aerosol composition.
Low maintenance and calibration requirements for
TE-5020 compared to PILS-IC and R&P-8400S.
July 2009	A-37	DRAFT - DO NOT CITE OR QUOTE

-------
SITE/PERIOD/SAMPLER/CONFIGURATION
SUMMARY OF FINDINGS
FRESNO SUPERSITE.CA; 12/01/03 to 12/23/03
Located 5.5 km northeast of downtown in a mixed residential-commercial neighborhood. Flow Sampler (L/min) Filter
a	b
Type Denuder
SAMPLER	FLOW RATE (L/MIN) FILTER TYPE1	DENUDER"
PC-BOSS	150	Teflon (W)-Nylon (P) CIF
Grover et al.65
Dionex-IC S042" (1.03 ± 0.03) PC-BOSS S04 +
(0.2 ± 0.3); R2 - 0.98; n - 27
R&P-8400S S042" (0.95 ± 0.05) Dionex-IC SO4 +
(0.3 ± 0.6); R2 - 0.68; n - 195
CONTINUOUS SAMPLER FLOW RATE (L/min)	DENUDER	ANALYSIS METHOD"
R&P-8400S	5	Activated Carbon	SO2 pulsed fluorescence
Dionex-IC	5	Parallel plate wet	IC
denuder
aFilter Manufacturer in parentheses - W: Whatman, Clifton, NJ; P: Pall-Gelman, Ann Arbor, Ml; S: Schleicher & Schnell. Keene, NH; nla: not available.
bAb03: Aluminum oxide; IC: Ion chromatography; CIF: Charcoal Impregnated Filter; FEP: Fluorinated Ethylene Propylene copolymer; MgO: Magnesium oxide; Na2C(k Sodium carbonate; NaHC03:
Sodium bicarbonate NOx: Oxides of nitrogen; SO2: Sulfur dioxide; TEA: Triethanolamine; TSP: Total Suspended PM; UV: Ultraviolet; XAD-4: Hydrophobic, non-polar polyaromatic resin.
eNa2C03 impregnated.
d37-mm filter.
Source: 'Chow (1995,077012); 2Watson and Chow (2001,157123);3 Watson et al. (1983, 045084); "Fehsenfeld et al. (2004,150432); 'Solomon et al. (2001,150993); sMikel (2001,150702);
'Mikel (2001,150702); "Watson et al. (1999,020949); 'Solomon and Sioutas (2000,150995); '"Graney et al. (2004, 053750); "Tanaka et al. (1998,157041); "Pancras et al. (2005,098120);
"John et al. (1988,045903); "Hering and Cass (1999, 084958); "Fritz et al. (1989, 077387); '"Hering et al. (1988, 030012); "Solomon et al. (2003,150994); '"Cabada et al. (2004,148859);
"Fine et al. (2003,155775); 2"Hogrefe et al. (2004, 099003); 2'Drewnick et al. (2003, 099100); 22Watson et al. (2005,157125); 23Ho et al. (2000,150552); 24Decesari et al. (2005,144530);2S
Mayol-Bracero et al. (2002, 045010); 2sYang et al. 2003;21 Tursic et al. (2000,157003); 2"Mader et al.(2004,150724); 29Xiao and Liu (2004, 050801); 3"Kiss et al. (2002,150040); ''Cornell and
Jickells (1999,150307);32 Zheng et al. (2002, 020100); 33Fraser et al. (2002,140741);34 Fraser et al. (2003, 042231) 3sSchauer er al. (1990, 051102); 3sFine et al. (2004); 3,Yue et al. (2004);
3BRinehart et al. (2000,115184); 39Wan and Yu (2000,157104); 40Poore (2000, 012839); "Fraser et al. (2003a); 42Engling et al. (2000,150422); 43Yu et al. (2005); 44Tran et al. (2000); 4sYao et al.
(2004,102213); 4sLi and Yu (2005,150092); "Henning et al. (2003,150539); 4"Zhang and Anastasio (2003,157182); "Enmenegger et al. (2007,150418); s0(Watson et al., 1989,157119);
"Greaves et al. (1985,150494); "Waterman et al. (2000,157110); s3Waterman et al. (2001,157117); s4Falkovich and Rudich (2001,150427); ssChow et al. (2007,150354); "Miguel et al.
(2004,123200); "Crinmins and Baker (2000, 097008); s"Ho and Yu (2004,150551); s9Jeon et al. (2001,010030); ""Mazzoleni et al. (2007, 098038); s,Poore (2000, 012839); ^Butler et al.
(2003); G3Chow et al. (2000c); "Russell et al. (2004, 082453); ssGrover et al. (2000); "Graver et al. (2005, 090044); "Schwab et al. (2000b); ""Hauck et al. (2004); s9Jaques et al. (2004,
155878); '"Rupprecht and Patashrick (2003,157207); "Pang et al (2002, 030353); ,2Eatough et al. (2001,010303); ,3Lee et al. (2005b); ,4Lee et al. (2005,150080); ,sBabich et al. (2000,
150239); ,GLee et al. (2005c); "Lee et al. (2005b); '"Anderson and Ogren (1998,150213);,9Chung et al. (2001,150357); ""Kidwell and Ondov (2004,155898); "'Lithgow et al. (2004,120010);
'"Weber et al. (2003,157129); "'Harrison et al. (2004); "Rattigan et al. (2000,115897); "sWittig et al. (2004,103413); ""Vaughn et al. (2005,157089); "'Chow et al. (2005b); ""Weber et al
(2001, 024040); ""Schwab et al. (2000a); ""Lim et al. (2003,150097); "'Watson and Chow (2002,037873); ""Venkatachari et al. (2000,105918); 93Bae et al. (2004a); "Arhami et al. (2000,
150224);95Park et al. (2005,150843); ""Bae et al. (2004b); 9,Chow et al. (2000a); 9"Arnott et al. (2005); "Bond et al. (1999); '""Virkkula et al. (2005,157097); '"'Petzold et al. (2002,150803);
""Park et al. (2000); ""Arnott et al. (1999, 020050); ""Peters et al. (2001); ,osPitchford et al. (1997,150872); '""Rees et al. (2004, 097104); ""Watson et al. (2000); '""Lee et al. (2005,150080);
'""Hering et al. (2004,155837); ""Watson et al. (1998); "'Chakrabarti et al. (2004); "2Mathai et al. (1990,150741); '"Kidwell and Ondov (2001,017092); '"Stanier et al. (2004); '"Khlystov et
al. (2005,150035); ""Takahama et al. (2004,157038); '"Chow et al. (2005,150348); ""Zhang et al. (2002); ""Subramanian et al. (2004, 081203); ,2°Chow et al. (2000b); "'Birch and Cary
(1990);122Birch (1998,024953); ,23Birch and Cary (1990); ""NIOSH (1990,150810); ,2sNI0SH (1999,150811); ,2"Chow et al. (1993); "'Chow et al. (2007); ""Ellis and Novakov (1982,150410);
""Peterson and Richards (2002,150801); ""Schauer et al. (2003); "'Middlebrook et al. (2003,042932); ,32Wenzel et al. (2003,157139); "'Jimenez et al. (2003); "4Phares et al. (2003,150800);
"sQin and Prather (2000,150895); ""Zhang et al. (2005); "'Bein et al. (2005,150205); ""Drewnick et al. (2004a); ""Drewnick et al. (2004b); ""Lake et al. (2003,150009); "'Lake et al. (2004,
088411)
Table A-15. Summary of PM2.5 carbon measurement comparisons.
SITE/PERIOD/SAMPLER/ CONFIGURATION
SUMMARY OF FINDINGS
ATLANTA SUPERSITE, GA: 08/03/99 to 09/01/99
4 km NW of downtown, within 200 m of a bus maintenance yard and several warehouse facilities, representative
of a mixed commercial-residential neighborhood.
SAMPLER
FLOW RATE
(L/MIN)
FILTER
TYPE1
DENUDER"
ANALYSIS METHOD'
R&P-2000 FRM
16.7
Quartz (P)
None
NIOSH 5040-T0T
RAAS-400
24
Quartz (P)
None
NIOSH 5040-T0T
SASS
6.7
Quartz (P)-
Quartz (P)
None
NIOSH 5040-T0T
MASS-450
16.7
Quartz (P)
None
NIOSH 5040-T0T
Solomon et al.17
Organic Carbon (OC);
PM2.E OC from each sampler was compared to the all-
sampler avg, called the relative reference (RR) value.
The samplers agreed to within 20 to 50% of RR. Only
front filter OC is reported without artifact correction.
Denuded samplers showed lower OC (20 to 35%) than
RR, while non-denuded sampler OC was higher (5 to
35%).
Among non-denuded samplers, as filter face velocity
decreased, OC increased, with the exception of R&P-
2300.
July 2009
A-38
DRAFT-DO NOT CITE OR QUOTE

-------
SITE/PERIOD/SAMPLER/ CONFIGURATION
SUMMARY OF FINDINGS
R&P-2300
10
Quartz (P)-
Quartz (P)
None

NIOSH 5040-T0T
VAPS
15
Quartz (P)
XAD-4

NIOSH 5040-T0T
URG-PCM
16.7
Quartz (P)-
Quartz (P)
XAD-4

Front: NIOSH 5040-T0T;
Backup: custom-TOTd
ARA-PCM
16.7
Quartz (n/a)-
Quartz (n/a)
CIF

IMPROVETOR
PC BOSS (TVA)
150
Quartz (P)-
CIF (n/a)
CIF

Front: IMPROVE TOR;
Backup: TPV
PC BOSS (BYU)
150
Quartz (P)-CIF
(S)
CIF

TPB
MOUDI-100
30
Al Foilf-Quartz
(n/a
None

Custom-TOR to suit Alc

CONTINUOUS
SAMPLER
FLOW RATE
(L/min)
DENUDER
OC
EC
COMMENTS
ADI-C
2.7
Activated
Carbon
Not known
n/a
Part of SO42' instrument
W/CO2 non-dispersive
infrared (NDIR) analyzer;
data corrected for avg field
blank; OC - 2 oxidized OC
RU-OGI
16.1
None
700 in He
850 in 2% O2
TOT; Dynamic blank for
adsorption correction
R&P-5400
16.7
None
275 in air
750 in air
No pyrolysis correction
PSAP
1.26
None

babs@
565
nm
10m2/g factor
AE-16
4
None

babs@
880 nm
12.6 m2/g factor
OC positive artifacts ranged from 2 to 4/yg/m
EC:
PM2.E EC from each sampler was compared to the all-
sampler avg, called the relative reference (RR) value.
The samplers agreed to within 20 to 200% of RR.
TOT samples showed less EC than RR by 15 to 30%,
while TOR samples showed more EC than RR by 40 to
90%. PCBOSS (BYU) > RR value by 140%. EC by TOR
is —twice EC by TOT.
Major difference in EC is due to the carbon analysis
protocol and optical monitoring correction
(i.e., transmittance, reflectance).
Lim et al. (2003)
TC concentrations measured by the RU-OGI and R&P-
5400 correlated reasonably well (R2 - 0.83), with a
slope of 0.96. The ratio of the mean RU-OGI to mean
R&P-5400 TC was 1.02.
R&P-5400 OC was 8% lower than the RU-OGI
(R2 - 0.73), while the R&P-5400 EC was 20% higher
than RU-OGI (R2 - 0.74).
OC measured by ADI-C was lower than R&P-5400 and
RUOGI by 15% and 22%, respectively.
EC from PSAP and AE-16 correlated well (R2 - 0.97).
PSAP was lower by ~ 50%, compared with AE-16,
R&P-5400 and RU-OGI.
EC measured byAE-16was —12% higher than RU-
OGI. Calibration factors for the light absorption
instruments need to be adjusted for better correlation.
Calibration factor might be non-linear over the range of
absorbance measured.
The mean OC from R&P-5400 and RU-OGI were within
10% of filter RR values. Mean ADI-C OC was 14%
lower than filter RR OC.
EC from continuous instruments was 2-2.5 times filter
RR EC; continuous TC was also greater than filter RR
TC by 17% (R&P-400) to 27% (RU-OGI).
PITTSBURGH SUPERSITE, PA; 06/01/01 to 07/31/02
Six km east of downtown in a park on the top of a hill.
SAMPLER FLOW FILTER TYPE / PACK1	DENUDER	ANALYSIS METHOD'
CMU Custom-1 16.7 Non-denuded	Teflon (P/W)- None	NIOSH 5040-T0T
sample	Quartz (P) (QBT)
16.7 Non-denuded	Quartz (P)-Quartz None	NIOSH 5040-T0T
sample	(P) (QBQ)
CMU Custom-2 16.7 Denuded sample Denuder-Quartz Activated Carbon NIOSH 5040-T0T
(P)-CIG (S)
16.7 Dynamicblank Teflon (P/W)- Activated Carbon NIOSH 5040-T0T
(DYN!	Denuder-Quartz
(P)-CIG (S)
Subramanian et al.119
Particulate OC (POC) was estimated from denuded
sample (Quartz OC + CIG OC) after subtracting DYN
POC.
Denuder efficiency (1 -DYN POC/UDB POC) was
94 ± 3%. No seasonal variability or deterioration in
denuder performance was observed.
Positive artifact due to denuder breakthrough was
18.3 ± 12.5% of the denuded sample POC.
Negative artifact (CIGsample-CIGDYN) was, on avg,
6.3 ± 6.2% of POC.
Positive artifact was 34 ± 10% from QBT, and was
13 ±5% from QBQ. QBT > >QBQ.
July 2009
A-39
DRAFT-DO NOT CITE OR QUOTE

-------
SITE/PERIOD/SAMPLER/ CONFIGURATION
SUMMARY OF FINDINGS
16.7
Non-denuded
blank fUDBJ
Teflon (P/W)-
Quartz (P)-CIG
(S)
None
NIOSH 5040-T0T
QBT over-corrected the positive artifact by 20%. OC
volatilization from the front Teflon filter that
subsequently-adsorbed on the back-up quartz filter,
resulted in an overestimation of the positive artifact.
Non-denuded QBQ provided a more representative
estimate of the positive artifact on the non-denuded
front quartz filter for 24-h samples. However, it was
not suitable for 4-6 h samples, because the filters were
not in equilibrium with the air stream.
Positive artifact dominated when sampling with a non-
denuded quartz filter.
Comparison of 24-h avg non-denuded front quartz OC
versus denuded POC over the year showed an intercept
of 0.53/yg/m3, indicative of a positive artifact on
quartz filter samples.
The artifacts were higher in summer on an absolute
basis; however, they showed no seasonal variation
when expressed as a fraction of POC.
ST. LOUIS SUPERSITE, IL, MO; 01/01/02 to 12/31/02
Three km east of St. Louis, MO City center, also impacted by industrial sources, and located in a mixed residential
light commercial neighborhood.
SAMPLER
FLOW
RATE
(Lmin)
FILTER
TYPE/PACK1
DENUDER1
I
ANALYSIS METHOD'
University of
24
Quartz (P)
None

ACE Asia TOT
Wisconsin Custom-1

Denuder-Quartz
(P)
CIF

ACE Asia TOT
University of
Wisconsin Custom-2
24
Denuder-Quartz
(P)
CIF

ACE Asia TOT


Teflon (n/a)-
Denuder-Quartz
(P)
CIF

ACE Asia TOT

CONTINUOUS
SAMPLER
FLOW
RATE
(L/min)
DENUDER
OC
EC
COMMENTS
Sunset OCEC
8
CIF
340,
500,615,
870 °C in
100% He
550, 625, 700,
775, 850, 900
°C in 2% O2,
98% He
ACE Asia TOT; CH4 FID
detector
Bae et al. 93 96
Denuder breakthrough was 0.17 ± 0.15/yg/m3, and
constituted less than 5% of annual avg OC
concentration.
Non-denuded OC - (1.06 ± 0.02) x denuded OC +
(0.34 ± 0.10)
Equivalence of OC intercept and denuder breakthrough
implies that the low-level artifact is caused by denuder
breakthrough.
Non-denuded EC - (1.04 ± 0.03) x denuded EC +
(0.07 ± 0.03), indicating negligible EC artifact.
Results suggested higher summer-time OC artifact, on
an absolute basis.
Comparison of continuous Sunset TC and OC with 24-h
filter samples showed good correlations (R2) of 0.89
and 0.90, respectively.
Continuous Sunset TC in//g|m3 - (0.97 ± 0.02) x
filter TC + (0.83 ± 0.11), indicating comparability with
the filter measurements.
Continuous Sunset OC - (0.93 ± 0.02) x filter OC +
(0.94 ± 0.09)
Positive intercept was interpreted to be a blank
correction for the continuous measurements.
EC comparison was poor with large scatter in data
(R2 - 0.60), probably due to low EC concentrations
(avg - 0.70/yg/m3), close to the detection limit
(0.5/yg/m3).
FRESNO SUPERSITE, CA and other CRPAQS sites; 12/02/99 to 02/03/01, 12/1/03 to 11/30/04
Fresno Supersite was located 5.5 km northeast of downtown in a mixed residential-commercial neighborhood.
SAMPLER
FLOW
RATE
(Lmin)
FILTER TYPE/PACK1
DENUDER"
ANALYSIS METHOD'
DRI-SFS
113
Quartz (P)
None
IMPROVETOR


Teflon (P)-Quartz
None
IMPROVETOR
RAAS-400
24
(P) (QBT) Quartz (P)-
Quartz (P) (QBQ)
None
IMPROVETOR
RAAS-400
24
Quartz (P)-Quartz (P)
(QBQ)
XAD-4 / CIF
IMPROVETOR
Watson and Chow91; Chow et al.Chow et al.
12°; Watson et al.'; Parket al.102
Non-denuded RAAS-400 and RAAS-100 FRM measured
equivalent TC. DRI-SFS, RAAS-400 and RAAS-100
FRM samplers showed comparability for front filter TC,
OC and EC measurements.
Positive OC artifact was 1.62 ± 0.58/yglm3 (~ 24%
of non-denuded front quartz OC) from QBT, and
1.12 ± 0.91 //g/m3 (— 17% of non-denuded front
quartz OC) from QBQ. QBT >>QBQ
Results from CRPAQS showed, on avg, a positive OC
artifact of 34% (of the non-denuded front quartz OC)
from QBT and 17.5% (of the non-denuded front quartz
July 2009
A-40
DRAFT-DO NOT CITE OR QUOTE

-------
SITE/PERIOD/SAMPLER/ CONFIGURATION
SUMMARY OF FINDINGS
RAAS-100 FRM 16.7 Quartz (P)
None
IMPROVE TOR
CONTINUOUS
SAMPLER
FLOW RATE
(L/min)
DENUDER
OC EC
COMMENTS
R&P-5400
16.7
None
275 °C in air 750 °C in
air
No pyrolysis correction
Sunset OCEC
8.5
CIG
250,500,650, 650,750,
850 °C in He 850, 940
°C in 2%
O2 in He
Transmittance
MAAP
16.7
None
babs 0
670 nm
Transmittance 6.5 m2/g
factor
AE-16
6.8
None
babs 0
880 nm
Transmittance
14625/A m2/g factor, where
AE-21
6.8
None
babs 0
370, 880
nm
A is in nm
AE-31
6.8
None
babs @
370, 470,
520, 590,
660, 880
and 950
nm

DRI-PA
3
None
babs @
1047 nm
Absorption, 5 m2/g factor

SAMPLER
FLOW RATE
(L/min)
FILTER
TYPE/PACK1
DENUDER"
ANALYSIS METHOD'
PC-BOSS
150
Quartz (P)-CIG (S)t CIF
TPV
OC) from QBQ.
Positive artifact was higher during summer than winter.
Negative artifact was, on avg, 0.61 ± 0.58 /yglm3
(— 10% of POC) at Fresno. Over all the CRPAQS sites,
it ranged from 2.3% in winter to 11 % in summer, with
an avg of 4.9%.
Positive artifact is estimated to be 0.5 /yglm3.
No difference in denuded quartz backup OC was found
between using XAD and CIF denuders.
Comparison of R&P-5400 TC, OC, and EC against filter
samples showed poor correlation (R2 < 0.55).
TC from R&P-5400 was 40-60% higher than filter TC
by TOR. None of the R&P-5400 versus TOR filter
comparisons were comparable or predictable, due to
several frequent instrument malfunctions during the
experiment and the small data set (~ 35 data points).
IMPROVETOR EC was consistently 20-25% higher
than aethalometer BC.
IMPROVE TOR EC was comparable to MAAP BC.
Comparison of light absorption (babs) from DRI-PA (1047
nm), MAAP (670 nm), and AE (880 nm) analyzers with
the filter IMPROVE TOR EC, gave a crabs of 2.3, 5.5
and 10 m2fg, differing from the default conversion
factors of 5, 6.5, and 16.6 m2/g used for each
instrument at the specified wavelength.
Grover et al.65
R&P-5400 TC - (0.50 ± 0.01) Sunset TC +
(3.6 ± 1.5); R2 - 0.73; n - 480
Sunset TC - (0.63 ± 0.05) PC-BOSS TC +
(4.1 ± 3.2); R2 - 0.86; n - 29
R&P-5400 TC - (0.41 ± 0.02) PC-BOSS TC +
(6.7 ± 1.6); R2- 0.91; n - 29
CONTINUOUS
SAMPLER
R&P-5400
Sunset OCEC
FLOW RATE
(L/min)
DENUDER1 OC
EC
COMMENTS
16.7
None
375 °C in air 750 0C in No pyrolysis
air
8.0
CIG
250, 500, 650, 750, NIOSH 5040 T0T
650, 850 °C 850 °C in ND|R C02 detector
in He
2% 0? &
98% He
BALTIMORE SUPERSITE, MD; 02/15/2002 to 11/30/2002
East of downtown in an urban residential area. Within 91 m of bus maintenance facility.
SAMPLER
FLOW RATE
(L/min)
FILTER
TYPE/PACK1
DENUDER'
ANALYSIS METHOD'
SASS
6.7
Quartz (P)-
Quartz (P)
None
STNTOT

CONTINUOUS
SAMPLER
FLOW RATE
(L/min)
DENUDER"
OC EC
COMMENTS
Park et al.95
Data capture 93.8%
Compared to SASS, Sunset underestimated OC and EC
by 22% and ~ 11.5%, respectively.
Higher OC in SASS was attributed to the absence of a
denuder (i.e., positive artifact by gaseous adsorption)
and to temperature differences between the STNTOT
and Sunset TOT carbon analysis temperature
protocols.
EC discrepancy was probably related to the differences
July 2009
A-41
DRAFT-DO NOT CITE OR QUOTE

-------
SITE/PERIOD/SAMPLER/ CONFIGURATION
SUMMARY OF FINDINGS
Sunset OCEC
Carbon 600 °C, then 870 °C in TOT; CH4 FID detector;
870 °C in He 2% O2 in Denuder breakthrough ~
He 0.5 - 1 //g C/m3; Used 0.5
to correct OC
concentrations
in temperature protocol.
G rover et al.66
RUBIDOUX, CA; 07/13/03 to 07/26/03
Rubidoux is located in the eastern section of the South Coast Air Basin (SoCAB) in the north-west corner of gunset gcEC TC - (0 90 + 0 06) PC-BOSS
Riverside County, 78 km downwind of the central Los Angeles metropolitan area and in the middle of the
remaining agricultural production area in SoCAB.
(2.0 ±2.1); R2 = 0.93; n = 21
SAMPLER
FLOW
RATE
(L/min)
FILTER
TYPE/PACK
a
DENUDER"
ANALYSIS METHOD0
PC-BOSS
150
Quartz (P)-CIG (S)
CIF
TPB (CIG heated to 450 °C in N2)

CONTINUOUS
SAMPLER
FLOW
RATE
(L/min)
DENUDER1'
OC
EC
COMMENTS
Sunset OCEC
8
CIF
n/a
n/a
TOT; NDIR detector; NIOSH 5040
protocol
Sunset OCEC
8
CIF
n/a
Not meas-
ured
TOT; has blank quartz filter before
entering analyzer. Used as "blank"
stream for quantifying OC artifacts;
3-step analysis only in He.
NEW YORK SUPERSITE, NY; 01/12/04 to 02/05/04
Urban site located at Queens College, NY, about 14 km west of Manhattan, within 2 km of freeways, and
within 12 km of international airports.
INTEGRATED
SAMPLER
FLOW
RATE
(L/min)
FILTER
TYPE/PACK
a
DENUDER"
ANALYSIS METHOD0
R&P-2300
10
Quartz

None
STN_TOT

CONTINUOUS
SAMPLER
FLOW
RATE
(L/min)
DENUDER"
OC
EC
COMMENTS
R&P-5400
16.7
None
340 °C
in air
750 °C in air
No pyrolysis correction
Sunset OCEC
n/a
CIF
600,
870 °C
in He
870 "Cat 10%
O2 in He
Transmittance
AE-20
n/a
None

babs @ 370,
880 nm
Transmittance, 14625/0 nf/g factor,
where D is in nm
AMS
n/a
None
n/a
n/a
~ 1 pm cut-point
Sunset TC was adjusted for carbon artifacts
measured by second (blank) instrument.
Venkatachari et al.
Regression of OC from Sunset OCEC against PM2.5
mass concentration yielded an intercept of
1.14 pg/m3, which was used as a measure of the
positive artifact on the Sunset data. The Sunset OC
data was corrected for this artifact.
AE-20 BC concentrations were -86% of Sunset EC
and R&P2300 filter EC concentrations.
AE-20 versus R&P-5400 showed high scatter.
Sunset Optical EC = 0.58 ± 0.05 Sunset Thermal
EC; R2 = 0.86; n = 506
Sunset Optical EC = 0.62 ± 0.05 AE-20 BC;
R2 = 0.96; n = 539
R&P-5400 TC tracked filter TC closely, but differed
widely for OC and EC.
Sunset OC = (0.75 ± 0.76) R&P-2300 OC +
(0.08 ±0.36); R2 = 0.67;n = 16
Sunset OC = (0.98 ±0.11) R&P-5400 OC -
(0.47 ±0.17); R2 = 0.44; n = 327
R&P-5400 OC = (0.60 ± 0.47) R&P-2300 OC +
(0.58 ±0.82); R2 = 0.58;n = 17
July 2009
A-42
DRAFT-DO NOT CITE OR QUOTE

-------
RUBIDOUX, CA; 07/13/03 to 07/26/03
Rubidoux is located in the eastern section of the South Coast Air Basin (SoCAB) in the north-west corner of
Riverside County, 78 km downwind of the central Los Angeles metropolitan area and in the middle of the
remaining agricultural production area in SoCAB.
Organic matter measurements by AMS showed
reasonable correlation (R2 = 0.76) with filter (R&P-
2300) OC, while being poorly correlated with
continuous OC by Sunset (R2 = 0.32) and R&P-
5400 (R2 = 0.36)
Sunset EC = (1.21 ± 0.44) R&P-2300 EC -
(0.03 ±0.13); R2 = 0.94; n = 16
Sunset EC = (1.35 ± 0.12) R&P-5400 EC +
(0.06 ±0.04); R2 = 0.61; n = 327
R&P-5400 EC = (0.49 ± 0.46) R&P-2300 EC +
(0.09 ±0.26); R2 = 0.77;n = 15
Sunset TC = (0.86 ± 0.39) R&P-2300 TC -
(0.06 ±0.69); R2 = 0.77;n = 16
Sunset TC = (1.31 ± 0.10) R&P-5400 TC -
(1.15 ±0.15); R2 = 0.59; n = 327
R&P-5400 TC = (0.77 ± 0.58) R&P-2300 TC +
(0.35 ±1.37); R2 = 0.83;n = 16
aFilter Manufacturer in parentheses - W: Whatman, Clifton, NJ; P: Pall-Gelman, Ann Arbor, Ml; S: Schleicher & Schnell. Keene, NH; nla: not available. QBT: quartz backup filter behind Teflon front
filter. QBQ: quartz backup filter behind Quartz front filter.
bAb03: Aluminum oxide; IC: Ion chromatography; CIF: Charcoal Impregnated Filter; CIG: Charcoal Impregnated Glass-Fiber Filter; FEP: Fluorinated Ethylene Propylene copolymer; MgO: Magnesium
oxide; Na2C03: Sodium carbonate; NaHC03: Sodium bicarbonate NOx: Oxides of nitrogen; SO2: Sulfur dioxide; TEA: Triethanolamine; TSP: Total Suspended PM; UV: Ultraviolet; XAD-4: (hydrophobic,
non-polar polyaromatic resin.
eNI0SH 5040_T0T: National Institute of Occupational Safety and Health Method 5040 Thermal Optical Transmittance Protocol.12,1,22, ]n m'0C: 250, 500, 050, 850 °C for 0C1, 0C2, 0C3, and
0C4 fractions, respectively, for 00, 00, 00, 90 sec respectively, in 100% He atmosphere. EC: 050, 750,850, 940 °C for EC1, EC2, EC3, and EC4 fractions, respectively, 30, 30, 30, > 120 sec
respectively, in 98% He and 2% O2 atmosphere. OPT: Pyrolysis correction by transmittance. IMPR0VE_T0R: Interagency Monitoring of Protected Visual Environments Thermal Optical Reflectance
Protocol. ""0C fractions: 120, 250,450, 550 °C for 001, 002, 003, and 0C4 fractions, respectively, until a well defined peak has evolved at each step, with a time limit of min 80 sec and max of
580 sec, in 100% He atmosphere. EC fractions: 550, 700, 800 °C for EC1, EC2, and EC3 fractions, respectively, until a well defined peak has evolved at each step, with a time limit of min 80 sec
and max of 580 sec, in 2% O2 and 98% He atmosphere. 0PR: Pyrolysis correction for pyrolyzed organic carbon (OP) by reflectance. 0C = 0C1 +0C2 + 0C3 + 0C4 + 0P EC = EC1+ EC2 + EC3-0P
TC = 0C + EC. IMPR0VE_A TOR: "' Note that as of May, 2007, the U.S. EPA is switching samples from the Speciation Trends Network thermal optical transmittance protocol to the IMPR0VE_A
protocol. 0C: 140, 280, 480, 580 °C for 0C1, 002, 003, and 0C4, fractions, respectively, until a well defined peak has evolved at each step, with a time limit of 80 sec and max of 580 sec, in
100% He atmosphere EC: 580, 740, 840 °C for EC1, EC2, and EC3 fractions, respectively, until a well defined peak has evolved at each step, with a time limit of min 80 sec and max of 580 sec, in
2% O2 and 98% He atmosphere. 0PR: Pyrolysis correction for pyrolyzed organic carbon (OP) by reflectance. OPT: Pyrolysis correction by transmittance. TPV: Temperature Programmed
Volatilization. ""For CIF Filters: Heated from 50 °C to 300 °C at a ramp rate of 10 °C/min in N2. For Quartz filters: Heated from 50 °C to 800 °C at a ramp rate of 28 °C/min in 70% N2 and
30% O2; EC estimated from high temperature peak (> 450 °C) on thermogram obtained from quartz-fiber filter analysis; No pyrolysis correction. STN_T0T: Speciation Trends Network Thermal
Optical Transmittance Protocol.129 0C: 310, 480, 015, 920 °C for 00, 00, 00, 90 sec respectively, in 100% He atmosphere. EC: 000, 075, 750, 825, 920 °C for 45, 45, 45, 45,120 sec
respectively, in 98% He and 2% O2 atmosphere. ACE Asia TOT: Aerosol Characterization Experiments in Asia Thermal Optical Transmittance Protocol.138 0C: 340, 500, 015, 870 °C for 00, 00, 00,
90 sec respectively, in 100% He atmosphere. EC: 550, 025, 700, 775, 850,900 °C for45, 45, 45, 45, 45,120 sec respectively, in 98% He, 2% O2. Pyrolysis correction by transmittance.
dCustom TOT: XA0-4 impregnated quartz, analyzed in He-only atmosphere with a maximum temperature 176 °C; EC is not measured.
eCustom TOR to suit Al substrate; details not reported.
f37-mm filter
Source: 'Chow (1995,077012); 2Watson and Chow (2001,157123);3 Watson et al. (1983, 045084); "Fehsenfeld et al. (2004,150432); 'Solomon et al. (2001,150993); 8Mikel (2001,150702);
'Mikel (2001,150702); "Watson et al. (1999,020949); 'Solomon and Sioutas (2000,150995); '"Graney et al. (2004, 053750); "Tanaka et al. (1998,157041); "Pancras et al. (2005,098120);
"John et al. (1988,045903); "Hering and Cass (1999, 084958); "Fritz et al. (1989, 077387); '"Hering et al. (1988, 030012); "Solomon et al. (2003,150994); '"Cabada et al. (2004,148859);
"Fine et al. (2003,155775); 20Hogrefe et al. (2004, 099003); 2,0rewnick et al. (2003, 099100); 22Watson et al. (2005,157125); 23Ho et al. (2000,150552); 240ecesari et al. (2005,144530);2S
Mayol-Bracero et al. (2002, 045010); 2sYang et al. 2003;21 Tursic et al. (2000,157003); 2"Mader et al.(2004,150724); 29Xiao and Liu (2004, 050801); 30Kiss et al. (2002,150040); ''Cornell and
Jickells (1999,150307);32 Zheng et al. (2002, 020100); 33Fraser et al. (2002,140741);34 Fraser et al. (2003, 042231) 3sSchauer er al. (1990, 051102); 3sFine et al. (2004); 3,Yue et al. (2004);
3BRinehart et al. (2000,115184); 39Wan and Yu (2000,157104); 40Poore (2000, 012839); "Fraser et al. (2003a); 42Engling et al. (2000,150422); 43Yu et al. (2005); 44Tran et al. (2000); 4sYao et al.
(2004,102213); 4sLi and Yu (2005,150092); "Henning et al. (2003,150539); 4"Zhang and Anastasio (2003,157182); "Enmenegger et al. (2007,150418); s0(Watson et al., 1989,157119);
"Greaves et al. (1985,150494); "Waterman et al. (2000,157110); s3Waterman et al. (2001,157117); s4Falkovich and Rudich (2001,150427); ssChow et al. (2007,150354); "Miguel et al.
(2004,123200); "Crinmins and Baker (2000, 097008); s"Ho and Yu (2004,150551); s9Jeon et al. (2001,010030); 60Mazzoleni et al. (2007, 098038); 6,Poore (2000, 012839); ^Butler et al.
(2003); s3Chow et al. (2000c); "Russell et al. (2004, 082453); ssGrover et al. (2000); "Graver et al. (2005, 090044); "Schwab et al. (2000b); ""Hauck et al. (2004); s9Jaques et al. (2004,
155878); '"Rupprecht and Patashrick (2003,157207); "Pang et al (2002, 030353); ,2Eatough et al. (2001,010303); ,3Lee et al. (2005b); ,4Lee et al. (2005,150080); ,sBabich et al. (2000,
150239); ,sLee et al. (2005c); "Lee et al. (2005b); '"Anderson and Ogren (1998,150213); ,9Chung et al. (2001,150357); ""Kidwell and Ondov (2004,155898); "'Lithgow et al. (2004,120010);
'"Weber et al. (2003,157129); "'Harrison et al. (2004); "Rattigan et al. (2000,115897); "sWittig et al. (2004,103413); ""Vaughn et al. (2005,157089); "'Chow et al. (2005b); ""Weber et al
(2001, 024040); ""Schwab et al. (2000a); ""Lim et al. (2003,150097); "'Watson and Chow (2002,037873); ""Venkatachari et al. (2000,105918); "3Bae et al. (2004a); "Arhami et al. (2000,
150224);95Park et al. (2005,150843); ""Bae et al. (2004b); "'Chow et al. (2000a); ""Arnott et al. (2005); "Bond et al. (1999); '""Virkkula et al. (2005,157097); '"'Petzold et al. (2002,150803);
""Park et al. (2000); ""Arnott et al. (1999, 020050); ""Peters et al. (2001); ,osPitchford et al. (1997,150872); '""Rees et al. (2004, 097104); ""Watson et al. (2000); '""Lee et al. (2005,150080);
'""Hering et al. (2004,155837); ""Watson et al. (1998); "'Chakrabarti et al. (2004); "2Mathai et al. (1990,150741); '"Kidwell and Ondov (2001,017092); '"Stanier et al. (2004); '"Khlystov et
al. (2005,150035); ""Takahama et al. (2004,157038); "'Chow et al. (2005,150348); ""Zhang et al. (2002); ""Subramanian et al. (2004, 081203); ,2°Chow et al. (2000b); "'Birch and Cary
(1990);122Birch (1998,024953); ,23Birch and Cary (1990); ""NIOSH (1990,150810); ,2sNI0SH (1999,150811); ,2"Chow et al. (1993); "'Chow et al. (2007); ""Ellis and Novakov (1982,150410);
""Peterson and Richards (2002,150801); ""Schauer et al. (2003); "'Middlebrook et al. (2003,042932); ,32Wenzel et al. (2003,157139); "'Jimenez et al. (2003); "4Phares et al. (2003,150800);
"sQin and Prather (2000,150895); ""Zhang et al. (2005); "'Bein et al. (2005,150205); ""Orewnick et al. (2004a); ""Orewnick et al. (2004b); ""Lake et al. (2003,150009); "'Lake et al. (2004,
088411)
July 2009
A-43
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-16. Summary of particle mass spectrometer measurement comparisons.
Spectrometer
Inlet Characteristics3 (Flow
Rate [L/Min] Size Inlet Dryer
Aerodynamic Diameter, /jm
Particle Sizing Method)
Volatilization/
Ionization
Method3
Hit Rates'1 Mass Spectrometer'
Particle Analysis/
Classification
Other
PALMS	nfa
PM2.6 cyclone
Nation (17 days) I None (4 days)
0.35-2.5
Light scattering
LDI,
ArF 193 nm
2x109 to 5x109
W/cm2
14 to 100%,
overall 87%
Single T0F reflectron; Ion
polarity needs to be pre-
selected
Peak ID/regression tree
analysis
AT0FMS
1
None
None
02-2.5
Aerosol T0F
LDI, Nd: YAG 266
nm laser
~ 1x108 W|cm2
25-30%,	Dual T0F reflectron;
occasionally as Detects both positive and
low as 5% negative ions
Aerosol T0F
RSMS-II	nfa
None
Nation
0.015-1.3
Aerodynamic focusing Need to pre-
select sizes to be analyzed
LDI, Arf laser, 193
nm
1x108 to 2x108
W/cm2
nfa
Single linear T0F; Ion
polarity needs to be pre-
selected
Peak ID/artificial neural
network
Pure sulfuric acid
(H2S04),
(NH4>2S04, and
water (H2O)
have relatively
' high ionization
thresholds (i.e.
difficult to
ionize). Fraction
of molecules
ionized in the
particles is on
the order of 10'6
to 10e.
AMS
nfa
PM2.6 cyclone
None
0.05-2.5
Aerosol TOF
T ~ 550 °C/ El
nfa
~uadrupole;
Mass weighted size
distributions on pre-
selected positive ions
only.
ID using standard
El ionization databases
Does not detect/
analyze highly
refractory
materials such
as metals, sea
salt, soot etc.
Fraction of
molecules ionized
in the particles is
on the order of
106 to 10"7
Middlebrook et al.131; Wenzel et al.I32; Jimenez et al.133
Particle sizing is approximate in PALMS, while ATOFMS, RSMS-II and AMS provide relatively accurate particle sizing.
Particle transmission in AMS is ~ 100% (i.e., it uses all particles in the sampled air) between 60 and 600 nm, while that for PALMS, ATOFMS and RSMS-II range from 10'e
for submicron particles to 2% for supermicron (> 0.8 //m) particles.
AMS has fewer matrix effects (due to separate volatilization and ionization steps) compared to single-step LDI instruments.
While four major particle classifications (organic/ SO42', sodium/potassium sulfate, soot/hydrocarbon and mineral) were observed by all three laser instruments, they differed
in the classification frequencies. Differences in frequencies that are detected and grouped are related to the differences in the laser ionization conditions (e.g., wavelength),
particle transmission, sizing method and the way the spectra were classified.
Shorter ionization wavelengths are able to produce ions more easily than longer ones.
Low hit rates in ATOFMS corresponded to periods of high SO42' concentrations. Low hit rates in PALMS were related to a variety of factors including high SO42'
concentrations, differing laser fluence and laser position relative to particle beam. Use of a dryer in PALMS enhanced ionization of particles that were difficult to ionize at
high ambient RH.
The RSMS-II and ATOFMS were less sensitive to S042'and hence may have fewer organic/S042'particles (i.e., underestimate SO42', pure sulfuric acid etc.).
The PALMS, ATOFMS and RSMS (laser based instruments) are qualitative, while the AMS can be quantitative. The relative ratio of ion intensities from the laser
instruments, however, may be indicative of relative concentrations, thus giving semi-quantitative information.
Comparison of the ratio of NO3 to SO4 peaks with the results from the semi continuous instruments showed better correlation with the AMS (R2 - 0.93) than PALMS
(R2 - 0.65 for non-dry particles to 0.70 for dry particles). While reasonable correlations between the PALMS and the composite semi-continuous data indicate the
possibility for calibration of laser-based data for certain ions, the calibration factors may vary depending on the particle matrix, water content and laser ionization
parameters, and averaging the spectra according to these factors may minimize these effects.
Comparison of AMS SO4 with PILS SOi showed good correlation (R2 - 0.79), and the data uniformly scattered around a 1:1 line. NO3 comparison was poor (R2 - 0.49)
because of the low signal to noise ratio at low concentrations
The continuum between particle classifications indicates that the particles were not adequately represented by non-overlapping classifications.
July 2009
A-44
DRAFT-DO NOT CITE OR QUOTE

-------
Spectrometer
Inlet Characteristics3 (Flow
Rate [L/Min] Size Inlet Dryer
Aerodynamic Diameter, /jm
Particle Sizing Method)
Volatilization/
Ionization
Method3
Hit Rates'1 Mass Spectrometer'
Particle Analysis/
Classification
Other
HOUSTON SUPERSITE, TX; 08/23/00 to 09/18/00
Houston Regional Monitoring Site was located < 1.0 km north of the Houston ship channel, where chemical and other industries are present. The site was located between
a railway to the south and a chemical plant to the north. Major freeways were located just to the north and east of the sampling site.
SPECTROMETER INLET CHARACTERISTICS
(FLOW RATE [L/MIN]
SIZE INLET
DRYER AERODYNAMIC
DIAMETER,/jM PARTICLE SIZING
METHOD)
VOLATILIZATION!
IONIZATION
METHOD"
HIT RATES'
MASS
SPECTROMETER'
PARTICLE ANALYSIS! OTHER
CLASSIFICATION
RSMS-II	n!a
None
Nafion
0.035 - 1.14
Aerodynamic focusing; Need to pre-
select sizes to be analyzed
LDI, ArF laser,
193 nm
nfa
Single linear TOF; Ion
polarity needs to be pre-
selected
Peak ID/artificial neural
network
At each size
point, aerosol
was sampled in
each cycle for
either 10 min or
until mass
spectra for 30
particles per
major class were
collected,
whichever came
first.
Phares et al.134
27,000 spectra were classified using a neural network into 15 particle types
Fifteen particle type mass spectra were presented along with their size distribution, avg time of day occurrence, and wind direction dependence
Major classes were a K dominant, Si/Silicon Oxide, Carbon, Sea Salt, Fe, Zn, Amines, Lime, Vanadium, Organic Mineral, Pb and K, Al, and a Pb salt particle type.
FRESNO SUPERSITE, CA: 11/30/00 to 2/4/01
Urban location in a residential neighborhood.
SPECTROMETER INLET CHARACTERISTICS	VOLATILIZATION! HIT RATES" MASS	PARTICLE ANALYSIS! OTHER
(FLOW RATE [L/MIN]	IONIZATION	SPECTROMETER'	CLASSIFICATION
SIZE INLET	METHOD"
DRYER AERODYNAMIC
DIAMETER,/jM PARTICLE SIZING
METHOD)
ATOFMS
1
None
None
0.3-2.5
Aerodynamic
LDI, ND: YAG
266 nm
n!a
Dual reflectron TOF
Peak ID/artificial
neural network
ATOFMS
unsealed
detected
particles tracked
B attenuation
monitor PM2.6
mass
concentration
Qin and Prather135
Biomass burning particles reached a maximum at night and a minimum during the day. These particles were less than 1 //m in diameter and accounted for more than 60% of
the particles detected at night.
Another particle class characterized by high mass carbon fragments had a similar diurnal pattern. These particles were larger than 1 //m and were interpreted as biomass
particles that have undergone gas to particle conversion of semi-volatile species followed by dissolution in a water droplet.
PITTSBURGH SUPERSITE, PA; 09/07/02 TO 09/22/02 FOR AMS; 09/20/01 to 09/26/02 for RSMS-III
6 km east of downtown in a park on the top of a hill
July 2009
A-45
DRAFT-DO NOT CITE OR QUOTE

-------
Spectrometer
Inlet Characteristics3 (Flow
Rate [L/Min] Size Inlet Dryer
Aerodynamic Diameter, /jm
Particle Sizing Method)
Volatilization/
Ionization
Method3
Hit Rates'1 Mass Spectrometer'
Particle Analysis/
Classification
Other
SPECTROMETER INLET CHARACTERISTICS
(FLOW RATE [L/MIN]
SIZE INLET
DRYER AERODYNAMIC DIAMETER,/jM
PARTICLE SIZING METHOD)
VOLATILIZATION!
IONIZATION
METHOD"
MASS
SPECTROMETER"
OTHER
AMS	1.4 eels	T - 600 °C/ El
PM2.6 cyclone
None
0.05-1.0
Aerosol TOF
Quadrupole;	Particle size-cut of — 1 fjm
Mass weighted size distributions on pre-
selected positive ions only.
RSMS-IM	n/a
None
Nation
0.03-1.1
Aerodynamic focusing; Need to pre select
sizes to be analyzed.
LDI, ArF laser, 193 Dual TOF feflectron; Detects both positive
nm	and negative ions
At each size point, aerosol was
sampled in each cycle for either 10
min or until mass spectra for 30
particles per major class were
collected, whichever came first
136	137
Zhang etal. ; Bein et al.
The AMS observed 75% of the SO42' measured by R&P-8400S (R2 - 0.69).
Collection efficiency (CE) of 0.5 used for SO42', NO3 and NH4+ and 0.7 for organics to correct mass concentrations for incomplete detection. Use of a constant CE
irrespective of size and shape may overestimate accumulation mode (mostly, oxygenated) organics (true CE ~ 0.5) and underestimate smaller mode (primary) organics (true
CE ~ 1.0).
Comparison of AMS organics (organic matter, OM) with OC measured by a continuous Sunset OCEC instrument showed good correlation (R2 - 0.88) with a slope of 1.69.
A 24-h avg comparison, showed a slope of 1.45. These values are in the typical range of 1.2 to 2.0 for OMfOC ratios.
AMS could be used along with the SMPS to estimate particle density. The AMS did not always agree with SMPS, probably due to non-spherical particles (irregular) such as
soot from fresh traffic emissions, whose mass may be overestimated by the SMPS.
Comparison of AMS mass with the MOUDI, showed differences for aerodynamic diameters > 600 nm, probably due to the AMS transmission being less than unity for
particles larger than 600 nm.
For RSMS-III, 54% of the detected particles were assigned to one class (carbonaceous ammonium nitrate). This class was preferentially detected during the colder months
and was detected from many different wind directions.
The next largest RSMS-III class was EC/OC/K class at 11%, and is believed to be from biomass burning.
An unidentified organic carbon RSMS-III class (3.3% of all detected particles) was seen to be highly dependent on wind direction dependence and was primarily detected
during August and September of 2002. These particles likely originated from a landfill.
NEW YORK SUPERSITE; 06/30/01 to 08/05/01 (urban); 07/09/02 to 08/07/02: (rural)
Urban Site: Queens College, Queens, New York, located at the edge of a parking lot and within 1 km from expressways and highways in New York City Metropolitan area.
Rural Site: Whiteface Mountain, New York, located in a cleared area surrounded by mix of deciduous and evergreen trees, ~ 2 km away from the closest highway with no
major cities within 20 km.
SPECTROMETER
INLET CHARACTERISTICS
(FLOW RATE [L/MIN]
SIZE INLET
DRYER AERODYNAMIC
DIAMETER,/jM PARTICLE
SIZING METHOD)
VOLATILIZATION!
IONIZATION
METHOD"
MASS
SPECTROMETER'
OTHER
AMS
0.1
PM2.6 cyclone
None
0.02-2.5
Aerosol TOF
T - 700 °C| El
Quadrupole;
Mass weighted size
distributions on pre selected
positive ions only.
Data are 10-min averages
July 2009
A-46
DRAFT-DO NOT CITE OR QUOTE

-------
Inlet Characteristics3 (Flow
Rate [L/Min] Size Inlet Dryer v®latlllzatlon'	Particle Analysis/
Spectrometer .	.	Ionization Hit Rates Mass Spectrometer „. ... ..	Other
Aerodynamic Diameter, um	Classification
n J- I n- ¦ n fl -i j\	Method3
Particle Sizing Method)
Drewnick et al.138'139; Hogrefe et al.20
Transport losses were 1.3% on avg.
Inlet losses (at the inlet of AMS) were 1.9%, on avg, ranging from 11 % for a 20 nm particle to 9% for a 2.5 fjm particle, with a minimum of 0.7% for a 350 nm particle
Overall measurement uncertainty of particle diameter was —11%.
The AMS was reliable with proper calibration, care, and maintenance. Valid 10 min averages were obtained for all components more than 93% of the time.
The mass to charge ratios (m/z) of fragments from different components may overlap (e.g., NH+, a fragment of NH4+ and CH3+, a fragment of organic species, have
mfz - 15) resulting in an interference (called as isobaric interference) Interfering signals were not used to calculate concentrations. This loss in concentration was adjusted
by applying a correction factor determined from laboratory studies.
Typical interferences were from fragments of organic species, water and oxygen.
With adjustments, the SO42', NO:; , and ammonium concentrations measured by the AMS were consistently lower than that measured by other co-located instruments,
probably due to incomplete focusing of the (NH4)2S04 and NH4NO3 particles by the aerodynamic lens.
At the urban site, AMS NO3 was within 10% of the filter NO3 concentration. At the rural site, it had a slope of 0.51 and R2 of 0.46.
AMS SO4showed good agreement with R&P-8400S at both the rural and urban locations (R2 - 0.89 to 0.92, slope - 0.99, n - 407 to 695) and was within 70 to 85% of
filter SO42' concentration.
Comparison of the total non-refractory mass measured by the AMS with the PM2.6 TE0M mass (operated at 50 °C or with dryer) at the urban location, showed good
correlation (R2 - 0.91) with near zero intercept (0.22 /yglm3). On avg, the AMS observed 64% of the mass measured by the TE0M.
The unexplained mass (36%) was attributed to transport losses, transmission and optical losses, and refractory components in the aerosol sample (e.g., metals, EC). The
mass closure was within the estimated uncertainty of the AMS mass measurements (5 to 10%).
BALTIMORE SUPERSITE, MD; 04/01/02 to 11/30/02
East of downtown in an urban residential area. Within 91 m of a bus maintenance facility.
SPECTROMETER	INLET CHARACTERISTICS VOLATILIZATION/	MASS	OTHER
(FLOW RATE [L/MII\I]	IONIZATION	SPECTROMETER"
SIZE INLET	METHOD"
DRYER AERODYNAMIC
DIAMETER,/jM PARTICLE
SIZING METHOD)
RSMS-III	0.2 - 18, based on particle size
chosen
None
Nafion
0.045- 1.3
Aerodynamic focusing; Need to
pre select sizes to be analyzed
Lakeet al.140"1
Utilizing both positive and negative ion detection enables detection of more species. However, detection efficiencies of negative ions decreased for smaller particles.
S04+ concentration (number or mass) was not accurately quantified.
RSMS-III was most efficient in 0.050 to 0.77 fjm range.
Particle compositions could be related to specific source categaories.
LDI, ArF laser, 193 nm	TOF with dual ion polarity At each size set point,
aerosol was sampled in
each cycle for either 10
min or until mass spectra
from 30 particles were
collected, whichever
came first.
July 2009
A-47
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-17. Summary of particle mass spectrometer measurement comparisons.
Inlet Characteristics3 (Flow
Rate [L/Min] Size Inlet Volatilization/
Spectrometer Dryer Aerodynamic Ionization
Diameter, /jm Particle Method3
	Sizing Method)	
Hit Rates'1
Mass
Spectrometer'
Particle
Analysis/
Classification
Other
PALMS
n/a
PM2.6 cyclone
Nation (17 days) I
0.35-2.5
Light scattering
None (4 days)
LDI,
ArF 193 nm
2x109 to 5x109
W/cm2
14 to 100%,
overall 87%
Single T0F reflectron;
Ion polarity needs to be
preselected
Peak ID/regression
tree analysis
AT0FMS
1
None
None
02-2.5
Aerosol T0F
LDI, Nd: YAG 266
nm laser
~ 1x108 W|cm2
25-30%, Dual T0F reflectron;
occasionally as Detects both positive
low as 5% and negative ions
Aerosol T0F
Pure sulfuric acid (H2SO4),
(NH4>2S04, and water (H2O)
have relatively
high ionization thresholds (i.e.
difficult to ionize). Fraction of
molecules ionized in the
particles is on the order of 10 !l
to 100.
RSMS-II	n/a
None
Nafion
0.015-1.3
Aerodynamic focusing Need to
pre select sizes to be analyzed
LDI, Arf laser, 193 n/a
nm
1x108 to 2x108
W/cm2
Single linear T0F; Ion
polarity needs to be pre-
selected
Peak ID/artificial
neural network
AMS
n/a
PM2.6 cyclone
None
0.05-2.5
Aerosol TOF
T — 550 °C/EI
n/a
~uadrupole;
ID using standard Does not detect! analyze highly
Mass weighted size El ionization
distributions on pre- databases
selected positive ions
only.
refractory materials such as
metals, sea salt, soot etc.
Fraction of molecules ionized in
the particles is on the order of
106 to 10"7
Middlebrook et al.131; Wenzel et al.I32; Jimenez et al.133
Particle sizing is approximate in PALMS, while ATOFMS, RSMS-II and AMS provide relatively accurate particle sizing.
Particle transmission in AMS is ~ 100% (i.e., it uses all particles in the sampled air) between 60 and 600 nm, while that for PALMS, ATOFMS and RSMS-II range from 10'e
for submicron particles to 2% for supermicron (> 0.8 //m) particles.
AMS has fewer matrix effects (due to separate volatilization and ionization steps) compared to single-step LDI instruments.
While four major particle classifications (organic/ SO42', sodium/potassium sulfate, soot/hydrocarbon and mineral) were observed by all three laser instruments, they differed
in the classification frequencies. Differences in frequencies that are detected and grouped are related to the differences in the laser ionization conditions (e.g., wavelength),
particle transmission, sizing method and the way the spectra were classified.
Shorter ionization wavelengths are able to produce ions more easily than longer ones.
Low hit rates in ATOFMS corresponded to periods of high SO42' concentrations. Low hit rates in PALMS were related to a variety of factors including high SO42'
concentrations, differing laser fluence and laser position relative to particle beam. Use of a dryer in PALMS enhanced ionization of particles that were difficult to ionize at
high ambient RH.
The RSMS-II and ATOFMS were less sensitive to S042'and hence may have fewer organic/S042'particles (i.e., underestimate SO42', pure sulfuric acid etc.).
The PALMS, ATOFMS and RSMS (laser based instruments) are qualitative, while the AMS can be quantitative. The relative ratio of ion intensities from the laser
instruments, however, may be indicative of relative concentrations, thus giving semi-quantitative information.
Comparison of the ratio of NO3 to SO4 peaks with the results from the semi continuous instruments showed better correlation with the AMS (R2 - 0.93) than PALMS
(R2 - 0.65 for non-dry particles to 0.70 for dry particles). While reasonable correlations between the PALMS and the composite semi-continuous data indicate the
possibility for calibration of laser-based data for certain ions, the calibration factors may vary depending on the particle matrix, water content and laser ionization
parameters, and averaging the spectra according to these factors may minimize these effects.
Comparison of AMS SO4 with PILS SOi showed good correlation (R2 - 0.79), and the data uniformly scattered around a 1: 1 line. NO3 comparison was poor (R2 - 0.49)
because of the low signal to noise ratio at low concentrations
The continuum between particle classifications indicates that the particles were not adequately represented by non-overlapping classifications.
HOUSTON SUPERSITE, TX; 08/23/00 to 09/18/00
Houston Regional Monitoring Site was located < 1.0 km north of the Houston ship channel, where chemical and other industries are present. The site was located between
a railway to the south and a chemical plant to the north. Major freeways were located just to the north and east of the sampling site.
July 2009
A-48
DRAFT-DO NOT CITE OR QUOTE

-------
Spectrometer
Inlet Characteristics3 (Flow
Rate [L/Min] Size Inlet
Dryer Aerodynamic
Diameter, /jm Particle
Sizing Method)
Volatilization/
Ionization
Method3
Hit Rates'1
Mass
Spectrometer'
Particle
Analysis/
Classification
Other
SPECTROMETER
INLET CHARACTERISTICS
(FLOW RATE [L/MIN]
SIZE INLET
DRYER AERODYNAMIC
DIAMETER,/jM PARTICLE
SIZING METHOD)
VOLATILIZATION!
IONIZATION
METHOD"
HIT RATES'
MASS
SPECTROMETER'
PARTICLE
ANALYSIS!
CLASSIFICATION
OTHER
RSMS-II
n/a
None
Nafion
0.035 - 1.14
Aerodynamic focusing; Need to
pre select sizes to be analyzed
LDI, ArF laser,
193 nm
nfa
Single linear TOF; Ion
polarity needs to be pre-
selected
Peak ID/artificial
neural network
At each size point, aerosol was
sampled in each cycle for either
10 min or until mass spectra
for 30 particles per major class
were collected, whichever
came first.
Phares et al.134
27,000 spectra were classified using a neural network into 15 particle types
Fifteen particle type mass spectra were presented along with their size distribution, avg time of day occurrence, and wind direction dependence
Major classes were a K dominant, Si/Silicon Oxide, Carbon, Sea Salt, Fe, Zn, Amines, Lime, Vanadium, Organic Mineral, Pb and K, Al, and a Pb salt particle type.
FRESNO SUPERSITE, CA: 11/30/00 to 2/4/01
Urban location in a residential neighborhood.
SPECTROMETER INLET CHARACTERISTICS VOLATILIZATION! HIT RATES" MASS	PARTICLE	OTHER
(FLOW RATE [L/MIN]	IONIZATION	SPECTROMETER' ANALYSIS!
SIZEINLET	METHOD"	CLASSIFICATION
DRYER AERODYNAMIC
DIAMETER,/jM PARTICLE
SIZING METHOD)
ATOFMS
1
None
None
0.3-2.5
Aerodynamic
LDI, ND: YAG
266 nm
n)a
Dual reflectron TOF
Peak ID/artificial
neural network
ATOFMS unsealed detected
particles tracked B attenuation
monitor PM2.6 mass
concentration
Qin and Prather135
Biomass burning particles reached a maximum at night and a minimum during the day. These particles were less than 1 //m in diameter and accounted for more than 60% of
the particles detected at night.
Another particle class characterized by high mass carbon fragments had a similar diurnal pattern. These particles were larger than 1 //m and were interpreted as biomass
particles that have undergone gas to particle conversion of semi-volatile species followed by dissolution in a water droplet.
PITTSBURGH SUPERSITE, PA; 09/07/02 TO 09/22/02 FOR AMS; 09/20/01 to 09/26/02 for RSMS-III
6 km east of downtown in a park on the top of a hill
SPECTROMETER INLET CHARACTERISTICS	VOLATILIZATION! MASS SPECTROMETER"	OTHER
(FLOW RATE [L(MIN]	IONIZATION
SIZE INLET	METHOD"
DRYER AERODYNAMIC DIAMETER,/jM
PARTICLE SIZING METHOD)
AMS	1.4 eels	T ¦ 600 °C/ El
PM2.6 cyclone
None
0.05-1.0
Aerosol TOF
Quadrupole;	Particle size-cut of — 1 fjm
Mass weighted size distributions on
pre selected positive ions only.
July 2009
A-49
DRAFT-DO NOT CITE OR QUOTE

-------
Inlet Characteristics3 (Flow
Rate [L/Min] Size Inlet Volatilization/
Spectrometer Dryer Aerodynamic Ionization
Diameter, /jm Particle Method3
	Sizing Method)	
Hit Rates'1
Mass
Spectrometer'
Particle
Analysis/
Classification
Other
RSMS-III	n/a
None
Nation
0.03-1.1
Aerodynamic focusing; Need to pre-
select sizes to be analyzed.
LDI.ArF laser, 193
nm
Dual TOF feflectron; Detects both
positive and negative ions
At each size point, aerosol was sampled in
each cycle for either 10 min or until mass
spectra for 30 particles per major class were
collected, whichever came first
136	137
Zhang etal. ; Bein et al.
TheAMS observed 75% of the SO42' measured by R&P-8400S (R2 - 0.69).
Collection efficiency (CE) of 0.5 used for SO42', NO3 and NH4+ and 0.7 for organics to correct mass concentrations for incomplete detection. Use of a constant CE
irrespective of size and shape may overestimate accumulation mode (mostly, oxygenated) organics (true CE ~ 0.5) and underestimate smaller mode (primary) organics (true
CE ~ 1.0).
Comparison of AMS organics (organic matter, 0M) with 0C measured by a continuous Sunset 0CEC instrument showed good correlation (R2 - 0.88) with a slope of 1.69.
A 24-h avg comparison, showed a slope of 1.45. These values are in the typical range of 1.2 to 2.0 for 0IW0C ratios.
AMS could be used along with the SMPS to estimate particle density. The AMS did not always agree with SMPS, probably due to non-spherical particles (irregular) such as
soot from fresh traffic emissions, whose mass may be overestimated by the SMPS.
Comparison of AMS mass with the M0UDI, showed differences for aerodynamic diameters > 600 nm, probably due to the AMS transmission being less than unity for
particles larger than 600 nm.
For RSMS-III, 54% of the detected particles were assigned to one class (carbonaceous ammonium nitrate). This class was preferentially detected during the colder months
and was detected from many different wind directions.
The next largest RSMS-III class was EC/0C/K class at 11%, and is believed to be from biomass burning.
An unidentified organic carbon RSMS-III class (3.3% of all detected particles) was seen to be highly dependent on wind direction dependence and was primarily detected
during August and September of 2002. These particles likely originated from a landfill.
NEW YORK SUPERSITE; 06/30/01 to 08/05/01 (urban); 07/09/02 to 08/07/02: (rural)
Urban Site: Queens College, Queens, New York, located at the edge of a parking lot and within 1 km from expressways and highways in New York City Metropolitan area.
Rural Site: Whiteface Mountain, New York, located in a cleared area surrounded by mix of deciduous and evergreen trees, ~ 2 km away from the closest highway with no
major cities within 20 km.
SPECTROMETER	INLET CHARACTERISTICS VOLATILIZATION!	MASS	OTHER
(FLOW RATE [L/MIN]	l0NIZATI0N	SPECTROMETER"
SIZE INLET	METHOD"
DRYER AERODYNAMIC
DIAMETER,/jM PARTICLE
SIZING METHOD)
AMS
0.1
PM2.6 cyclone
None
0.02-2.5
Aerosol TOF
T - 700 °C/ El
Quadrupole;
Mass weighted size
distributions on pre-
selected positive ions
only.
Data are 10-min averages
July 2009
A-50
DRAFT-DO NOT CITE OR QUOTE

-------
Inlet Characteristics3 (Flow
Rate [L/Min] Size Inlet
Spectrometer Dryer Aerodynamic
Diameter, /jm Particle
Sizing Method)
Drewnick et al.138'139; Hogrefe et al.20
Transport losses were 1.3% on avg.
Inlet losses (at the inlet of AMS) were 1.9%, on avg, ranging from 11 % for a 20 nm particle to 9% for a 2.5 fjm particle, with a minimum of 0.7% for a 350 nm particle
Overall measurement uncertainty of particle diameter was —11%.
The AMS was reliable with proper calibration, care, and maintenance. Valid 10 min averages were obtained for all components more than 93% of the time.
The mass to charge ratios (m/z) of fragments from different components may overlap (e.g., NH+, a fragment of NH4+ and CH3+, a fragment of organic species, have
mfz - 15) resulting in an interference (called as isobaric interference) Interfering signals were not used to calculate concentrations. This loss in concentration was adjusted
by applying a correction factor determined from laboratory studies.
Typical interferences were from fragments of organic species, water and oxygen.
With adjustments, the SO42', NO:; , and ammonium concentrations measured by the AMS were consistently lower than that measured by other co-located instruments,
probably due to incomplete focusing of the (NH4)2S04 and NH4NO3 particles by the aerodynamic lens.
At the urban site, AMS NO3 was within 10% of the filter NO3 concentration. At the rural site, it had a slope of 0.51 and R2 of 0.46.
AMS SO4showed good agreement with R&P-8400S at both the rural and urban locations (R2 - 0.89 to 0.92, slope - 0.99, n - 407 to 695) and was within 70 to 85% of
filter SO42' concentration.
Comparison of the total non-refractory mass measured by the AMS with the PM2.6 TE0M mass (operated at 50 °C or with dryer) at the urban location, showed good
correlation (R2 - 0.91) with near zero intercept (0.22 /yglm3). On avg, the AMS observed 64% of the mass measured by the TE0M.
The unexplained mass (36%) was attributed to transport losses, transmission and optical losses, and refractory components in the aerosol sample (e.g., metals, EC). The
mass closure was within the estimated uncertainty of the AMS mass measurements (5 to 10%).
BALTIMORE SUPERSITE, MD; 04/01/02 to 11/30/02
East of downtown in an urban residential area. Within 91 m of a bus maintenance facility.
SPECTROMETER	INLET CHARACTERISTICS VOLATILIZATION!	MASS	OTHER
(FLOW RATE [L/MIN]	l0NIZATI0N	SPECTROMETER"
SIZE INLET	METHOD"
DRYER AERODYNAMIC
DIAMETER, fjM PARTICLE
SIZING METHOD)
RSMS-III	0.2 - 18, based on particle
size chosen
None
Nafion
0.045- 1.3
Aerodynamic focusing; Need
to pre select sizes to be
analyzed
Lakeet al.140"1
Utilizing both positive and negative ion detection enables detection of more species. However, detection efficiencies of negative ions decreased for smaller particles.
S04+ concentration (number or mass) was not accurately quantified.
RSMS-III was most efficient in 0.050 to 0.77 fjm range.
Particle compositions could be related to specific source categaories.
Volatilization/ ..	Particle
Mass
Ionization Hit Rates'1 „ r	Analysis/ Other
Method3 Spectrometer	c|assjfication
LDI, ArF laser, 193 nm	TOF with dual ion polarity At each size set point, aerosol was
sampled in each cycle for either 10 min
or until mass spectra from 30 particles
were collected, whichever came first.
July 2009
A-51
DRAFT-DO NOT CITE OR QUOTE

-------
Inlet Characteristics3 (Flow
Rate [L/Min] Size Inlet Volatilization/	„„	Particle
Macs
Spectrometer Dryer Aerodynamic	Ionization Hit Rates'1	Analysis/	Other
Diameter, //m Particle Method3 Classification
	Sizing Method)	
aEI: Electon Impact; LDI: Laser Desorption / Ionization
bHit rate refers to the number of particles with a mass spectrum as a fraction of the number of particles detected. It does not apply to RSMS and AMS because there is no separate detection
eT0F: Time fo Flight
ATLANTA SUPERSITE, GA; 08/03/99 to 09/01/99
4 km NW of downtown, within 200 m of a bus maintenance yard and several warehouse facilities, representative of a mixed commercial-residential neighborhood.
Source: 'Chow (1995,077012); 2Watson and Chow (2001,157123);3 Watson et al. (1983, 045084); "Fehsenfeld et al. (2004,156432); 'Solomon et al. (2001,156993); sMikel (2001,156762);
'Mikel (2001,156762); "Watson et al. (1999,020949); 'Solomon and Sioutas (2006,156995); '"Graney et al. (2004, 053756); "Tanaka et al. (1998,157041); "Pancras et al. (2005,098120);
"John et al. (1988,045903); "Hering and Cass (1999, 084958); "Fritz et al. (1989, 077387); '"Hering et al. (1988, 036012); "Solomon et al. (2003,156994); "Cabada et al. (2004,148859);
"Fine et al. (2003,155775); 2°Hogrefe et al. (2004, 099003); 2,Drewnick et al. (2003, 099160); 22Watson et al. (2005,157125); 23Ho et al. (2006,156552); 24Decesari et al. (2005,144536);2S
Mayol-Bracero et al. (2002, 045010); 2sYang et al. 2003;21 Tursic et al. (2006,157063); 2"Mader et al.(2004,156724); 29Xiao and Liu (2004, 056801); 3°Kiss et al. (2002,156646); 3,Cornell and
Jickells (1999,156367);32 Zheng et al. (2002, 026100); 33Fraser et al. (2002,140741);34 Fraser et al. (2003, 042231) 3sSchauer er al. (1996, 051162); 3sFine et al. (2004); 3,Yue et al. (2004);
3BRinehart et al. (2006,115184); 39Wan and Yu (2006,157104); 40Poore (2000, 012839); "Fraser et al. (2003a); 42Engling et al. (2006,156422); 43Yu et al. (2005); 44Tran et al. (2000); 4sYao et al.
(2004,102213); 4sLi and Yu (2005,156692); "Henning et al. (2003,156539); 4"Zhang and Anastasio (2003,157182); "Enmenegger et al. (2007,156418); s0(Watson et al., 1989,157119);
"Greaves et al. (1985,156494); "Waterman et al. (2000,157116); s3Waterman et al. (2001,157117); s4Falkovich and Rudich (2001,156427); ssChow et al. (2007,156354); "Miguel et al.
(2004,123260); "Crinmins and Baker (2006, 097008); s"Ho and Yu (2004,156551); s9Jeon et al. (2001,016636); 60Mazzoleni et al. (2007, 098038); 6,Poore (2000, 012839); ^Butler et al.
(2003); G3Chow et al. (2006c); "Russell et al. (2004, 082453); ssGro»er et al. (2006); ssGro»er et al. (2005, 090044); "Schwab et al. (2006b); ""Hauck et al. (2004); s9Jaques et al. (2004,
155878); '"Rupprecht and Patashrick (2003,157207); "Pang et al (2002, 030353); "Eatough et al. (2001,010303); ,3Lee et al. (2005b); ,4Lee et al. (2005,156680); ,sBabich et al. (2000,
156239); ,sLee et al. (2005c); "Lee et al. (2005b); '"Anderson and Ogren (1998,156213); ,9Chung et al. (2001,156357); ""Kidwell and Ondo» (2004,155898); "Lithgow et al. (2004,126616);
""Weber et al. (2003,157129); "'Harrison et al. (2004); "Rattigan et al. (2006,115897); "sWittig et al. (2004,103413); ""Vaughn et al. (2005,157089); "'Chow et al. (2005b); ""Weber et al
(2001, 024640); "Schwab et al. (2006a); 90Lim et al. (2003,156697); 9,Watson and Chow (2002,037873); ^Venkatachari et al. (2006,105918); 93Bae et al. (2004a); "Arhami et al. (2006,
156224); 9sPark et al. (2005,156843); ""Bae et al. (2004b); 9,Chow et al. (2006a); ""Arnott et al. (2005); "Bond et al. (1999); '""Virkkula et al. (2005,157097); ""Petzold et al. (2002,156863);
,02Park et al. (2006); ,raArnott et al. (1999, 020650); ""Peters et al. (2001); ,osPitchford et al. (1997,156872); ,osRees et al. (2004, 097164); ""Watson et al. (2000); ,0BLee et al. (2005,156680);
'""Hering et al. (2004,155837); ""Watson et al. (1998); '"Chakrabarti et al. (2004); "2Mathai et al. (1990,156741); '"Kidwell and Ondo» (2001,017092); '"Stanier et al. (2004); '"Khlystov et
al. (2005,156635); ""Takahama et al. (2004,157038); '"Chow et al. (2005,156348); ""Zhang et al. (2002); ""Subramanian et al. (2004, 081203); ,2°Chow et al. (2006b); "'Birch and Cary
(1996);122Birch (1998,024953); ,23Birch and Cary (1996); ""NIOSH (1996,156810); "sNI0SH (1999,156811); ,2"Chow et al. (1993); "'Chow et al. (2007); ""Ellis and No»ako» (1982,156416);
""Peterson and Richards (2002,156861); ""Schauer et al. (2003); ,3,Middlebrook et al. (2003,042932); ,32Wenzel et al. (2003,157139);133Jimenez et al. (2003); "4Phares et al. (2003,156866);
"sQin and Prather (2006,156895); "Zhang et al. (2005); "'Bein et al. (2005,156265); ""Drewnick et al. (2004a); ""Drewnick et al. (2004b); ""Lake et al. (2003,156669); "'Lake et al. (2004,
088411)
Source: Chow et al. (2008,156355
Table A-18. Summary of key parameters for TD-GC/MS and pyrolysis-GC/MS.
Reference
Sample Type
TD Unit
Analytical Instrument
Total Analysis
Time
TD-GC/MS WITHRESISTIVELYHEATED EXTERNAL OVEN
Greaves et al. (1985, 156494;
1987,156495); Veltkamp et al.
(1996,081594)
Aerosol sample and NIST SRM 1649
A cylindrical aluminum block containing
a heating cartridge connected to a
thermocouple
HP 5892A GC/MS in El mode
ambient sample:
55.5 min NIST
standard: 45.5 min
Waterman et al. (2000,157116)
NIST SRM 1640a
External oven mounted on the top of
the GC/MS system
HP 5890 GC/Fisons
MD 800 MS, scan
range: 40-520 amu
90 min
Waterman et al. (2001,157117)
NIST SRM 1649a
Same as above
HP 5890 GC/Fisons
MD 800 MS, scan
range: m/z 40 to 520
90 mins
Sidhu et al. (2001,155202)
Aerosol collected on glass nber niters
from combustion of alternative diesel
fuel.
A stainless steel tube (0.635 cm
O.D.laced in a GC oven
Two GCs and one MS. The nrst
GC is used as the TE unit. The
second GC separates the
desorbent.
Ua
Havs et al. (2003,156529;
2004,156530); Dong et al.
(2004,156409)
Aerosol collected from residential wood
combustion, residential oil furnace and
~replace appliance
A glass tube placed in an external oven
(TDS2 Gerstel Inc.)
Aglient 6890 GC/5793 MSD,
scan range: 50 to 500 amu
99 min
July 2009
A-52
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Sample Type
TD Unit
Analytical Instrument
Total Analysis
Time
CURIE POINT TDGC/MS
Jeon et al. (2001, 016636)
High-volume PMio ambient samples
collected along the U.S.fMexico border
Curie point pyrolyzer
HP 5890 GC/5792 MSD
Ua
Neususs et al. (2000,156804)
Ambient aerosol collected during the 2nd
Aerosol Characterization Experiment
Curie point pyrolyzer
Fisons Trio 1000
35 min
IN INJECTION PORT TED GC/MS
Helmiq et al. (1990,156536)
Aerosol samples collected on glass-fiber
filters at a forest site
GC injector port, with modified septum
cap
Carlo Erba Mega 5160 GC/VG
250/70 SE MS, scan range: 45-
400 amu
47 min
Hall et al. (1999,156512)
NIST SRM 1649
Micro-scale sealed vessel placed inside
the injector port
HP 5890 GC/Fisons MD 800
MS, scan range: 40-500 amu
82.5 min
Blanchard and Hopper (1997)
(1997,157195); Blanchard et al.
(2002,189737)
Aerosol samples collected on quartz-and-
glass filters in Ontario
A GC injection port was added with
three minor components, including a
small T-connector, 3-way valve, and
needle valve
HP 5892A GC/5972A MS in El
mode
71 min
Falkovich and Rudich (2001,
156427);	Falkovich et al. (2004,
156428);	Graham et al. (2004,
156490)
NIST SRM 1649a; urban aerosols
collected with an 8-stage impactor in
Tel-Aviv, Israel
Direct Sample Introduction (DSI) device
(ChromatoProbe, Varian Co.)
Varian Saturn 3400 GC/MS
64.2 min
Ho and Yu (2004,156551); Yanq
et al. (2005,102388)
Ambient aerosol samples collected on
Teflon-impregnated glass-fiber filters in
Hong Kong and on quartz filters at
Nanjing, China
Conventional GC injection port. No
modification of GC injector and liner
HP 5890 GC/5791 MSD, scan
range: 50-650 amu
41.5 min
TD-GCX GC-MS
Welthagen et al. (2003,
104056); Schnelle-Kreis et al.
(2005,112944)
Ambient samples in Augsburg, Germany
Injection port Optic III with autoloader
(ATAS-GL, Veldhoven, NL)
Agilent 6890 GC/LECO Pegasus
III TOF/MS with a LECO
Pegasus 4D GCxGC modulator
175 min
Hamilton et al. (2004,156516)
PM2.E aerosol collected in London
Conventional GC injection port
The same as above, scan range:
20-350 amu
93.7 min
Hamilton et al. (2005, 088173)
Secondary organic aerosol formed during
the photo-oxidation of toluene with OH
radicals
The same as above
The same above
102.5 min
IN SITU SEMI CONTINUOUS AND CONTINUOUS TD SYSTEMS
Williams et al. (2006,156157)
In situ aerosol samples collected in
Berkley, CA
Collection-TE cell with conventional GC
injection port
Agilent 6890 GC/5793 MSD,
scan range: 29-550 amu
59 min
PYROLYSIS TD GC/MS
Voorheeset al. (1991,157101)
PM0.6 and PM>o.4b collected on quartz
fiber in pristine regions of Colorado
A tube furnace directly interfaced to an
GC/MS
Extrel Simulscan GC/MS, scan
range: 35-450 amu
31.7 min
Subbalakshmi et al. (2000,
157023)
Ambient aerosol collected on glass-fiber
filters in Jakarta, Indonesia
A pyroinjector
Agilent 6890 GC/5973 MS,
scan range: 50-550 amu
63.5 min
Fabbri et al. (2002,156426)
PM10 collected on glass-fiber filters in an
industrial are of Italy
A pyrolyzer directly connected to the
GC injector port through an interface
heated at 250° C
Varian 3400 GC/Saturn II ion
trap MS, scan range: 45-400
amu
57 min
Blazso et al. (2003,156278)
PM2.6 collected on quartz-fiber filters
and size-segregated aerosol sampled
collected on A1 foils in Brazil
A pyrolyzer
Agilent 6890 GC/5973 MS
30.3 min
Labbanet al. (2006,156665)
PM10 of re-suspended soil collected on
quartz-fiber filters
Curie point pyrolyzer
HP 5890 GC/5972 MS
25.5. min
July 2009
A-53
DRAFT-DO NOT CITE OR QUOTE

-------
Reference	Sample Type	TD Unit	Analytical Instrument Total Analysis
Total analysis time could not be determined because of insufficient experimental details
Source: Chow et al. (2007,157209)
A. 1.2. Networks
Table A-19. Relevant Spatial Scales for PM10l PM2.5, and PM102.5 Measurement
Spatial
Scales
PM10
PM2.5
PMio 2.5
Microscale
(~ 5 - 100 m)
This scale would typify areas such as downtown
street canyons, traffic corridors, and fence line
stationary source monitoring locations where the
general public could be exposed to maximum
PM10 concentrations. Microscale particulate
matter sites should be located near inhabited
buildings or locations where the general public
can be expected to be exposed to the
concentration measured. Emissions from
stationary sources such as primary and
secondary smelters, power plants, and other
large industrial processes may, under certain
plume conditions, likewise result in high ground
level concentrations at the microscale. In the
latter case, the microscale would represent an
area impacted by the plume with dimensions
extending up to approximately 100 meters. Data
collected at microscale sites provide information
for evaluating and developing hot spot control
measures.
This scale would typify areas such as downtown
street canyons and traffic corridors where the
general public would be exposed to maximum
concentrations from mobile sources. In some
circumstances, the microscale is appropriate for
particulate sites; community-oriented SLAMS sites
measured at the microscale level should, however,
be limited to urban sites that are representative of
long-term human exposure and of many such
microenvironments in the area. In general,
microscale particulate matter sites should be located
near inhabited buildings or locations where the
general public can be expected to be exposed to the
concentration measured. Emissions from stationary
sources such as primary and secondary smelters,
power plants, and other large industrial processes
may, under certain plume conditions, likewise result
in high ground level concentrations at the
microscale. In the latter case, the microscale would
represent an area impacted by the plume with
dimensions extending up to approximately 100
meters. Data collected at microscale sites provide
information for evaluating and developing hot spot
control measures. Unless these sites are indicative
of population-oriented monitoring, they may be more
appropriately classified as SPM.
This scale would typify relatively small areas
immediately adjacent to: Industrial sources;
locations experiencing ongoing construction,
redevelopment, and soil disturbance; and heavily
traveled roadways. Data collected at microscale
stations would characterize exposure over areas
of limited spatial extent and population exposure,
and may provide information useful for evaluating
and developing source-oriented control measures.
Middle Scale Much of the short-term public exposure to coarse
(~100- fraction particles (PMio) is on this scale and on
500 ml	"1S neighborhood scale. People moving through
downtown areas or living near major roadways or
stationary sources, may encounter particulate
pollution that would be adequately characterized
by measurements of this spatial scale. Middle
scale PMio measurements can be appropriate for
the evaluation of possible short-term exposure
public health effects. In many situations,
monitoring sites that are representative of micro-
scale or middle-scale impacts are not unique and
are representative of many similar situations.
This can occur along traffic corridors or other
locations in a residential district. In this case, one
location is representative of a neighborhood of
small scale sites and is appropriate for evaluation
of long-term or chronic effects. This scale also
includes the characteristic concentrations for
other areas with dimensions of a few hundred
meters such as the parking lot and feeder streets
associated with shopping centers, stadia, and
office buildings. In the case of PMio, unpaved or
seldomly swept parking lots associated with
these sources could be an important source in
addition to the vehicular emissions themselves.
People moving through downtown areas, or living
near major roadways, encounter particle
concentrations that would be adequately
characterized by this spatial scale. Thus,
measurements of this type would be appropriate for
the evaluation of possible short-term exposure public
health effects of particulate matter pollution. In
many situations, monitoring sites that are
representative of microscale or middle-scale impacts
are not unique and are representative of many
similar situations. This can occur along traffic
corridors or other locations in a residential district.
In this case, one location is representative of a
number of small scale sites and is appropriate for
evaluation of long-term or chronic effects. This scale
also includes the characteristic concentrations for
other areas with dimensions of a few hundred
meters such as the parking lot and feeder streets
associated with shopping centers, stadia, and office
buildings.
People living or working near major roadways or
industrial districts encounter particle
concentrations that would be adequately
characterized by this spatial scale. Thus,
measurements of this type would be appropriate
for the evaluation of public health effects of
coarse particle exposure. Monitors located in
populated areas that are nearly adjacent to large
industrial point sources of coarse particles
provide suitable locations for assessing maximum
population exposure levels and identifying areas
of potentially poor air quality. Similarly, monitors
located in populated areas that border dense
networks of heavily-traveled traffic are
appropriate for assessing the impacts of
resuspended road dust. This scale also includes
the characteristic concentrations for other areas
with dimensions of a few hundred meters such as
school grounds and parks that are nearly
adjacentto major roadways and industrial point
sources, locations exhibiting mixed residential
and commercial development, and downtown
areas featuring office buildings, shopping
centers, and stadiums.
July 2009
A-54
DRAFT-DO NOT CITE OR QUOTE

-------
Spatial
Scales
PMio
PM2.5
PM102.5
Neighborhood
Scale
( — 500 m-
4 l
-------
Network
Lead
Agency
Number of
Sites
Initiated
Measurement
Parameters
Location of Information and/or Data
SLAMS1 - State and Local
EPA
-3000
1978
O3, NOxINO?, SO2,
http:ffwww.epa.govfairfoaqpsfqafmonprog.html
Ambient Monitoring



PM2.BIPM10, CO, Pb

Stations





STN-PM2.6 Speciation
EPA
300
1999
PM2.6, PM2.6 speciation,
http:))www.epa.qov)ttnamti1 fspecqen.html
Trends Network



major Ions, metals

PAMS-Photochemical
EPA
75
1994
O3, NOx/NOy, CO, speciated
http:ffwww.epa. govfairfoaqpsfpamsf
Assessment Monitoring



VOCs, carbonyls, surface

Network



meteorology 81 Upper Air

IMPROVE—Interagencv
NPS
110 plus
1988
PM2.6/PM10, major ions,
http://vista.cira.colostate.edu/IMPROVE/
Monitoring of Protected

67 protocol

metals, light extinction,

Visual Environments

sites

scattering coefficient

CASTNet - Clean Air
EPA
80 +
1987
O3, SO2, major ions,
http:/)www.epa.qov)castnet)
Status and Trends



calculated dry deposition,

Network



wet deposition, total
deposition for
sulfur/nitrogen, surface
meteorology

GPMN-Gaseous Pollutant
NPS
33
1987
O3, N0x/N0/N02, SO2, CO,
http://www2.nature.nps.goV/air/Monitoring/network.cfm#data
Monitoring Network



surface meteorology, (plus
enhanced monitoring of CO,
NO, NOx, NOy, and SO2 plus
canister samples for VOC at
3 sites)

POMS-Portable Ozone
NPS
14
2002
O3, surface meteorology,
http://www2.nature.nps.qov/air/studies/port03.cfm
Monitoring Stations



with CASTNet-protocol
filter pack (optional) sulfate,
nitrate, ammonium, nitric
acid, sulfur dioxide

Passive Ozone Sampler
NPS
43
1995
O3 dose (weekly)
http://www2.nature.nps.qov/air/Studies/Passives.cfm
Monitoring Program





NADP/NTN—National
USGS
200 +
1978
Major Ions from
http://nadp.sws.uiuc.edu/
Atmospheric Deposition



precipitation chemistry

Program / National Trends





Network





NADP/MDN—National
None
90 +
1996
Mercury from precipitation
http://nadp.sws.uiuc.edu/mdn/
Atmospheric Deposition



chemistry

Program / Mercury





Deposition Network





AIRMoN—National
NOAA
8
1992
Major Ions from
http://nadp.sws.uiuc.edu/AIRMoN/
Atmospheric Deposition



precipitation chemistry

Program / Atmospheric
Integrated Research
Monitoring Network



Note: some sites began in
1976 as part of the DOE
MAP3S program; early data
are archived on NADP and
ARL servers.

IADIM—Integrated
EPA
20
1990
PAHs, PCBs, and
http://www.epa.qov/qlnpo/monitorinq/air/
Atmospheric Deposition



organochlorine compounds

Network



are measured in air and
precipitation samples

NAPS—National Air
Canada
152 +
1969
SO2, CO, Os, NO, NO2, NOx,
http://www.etc-cte.ec.qc.ca/NAPS/index e.html
Pollution Surveillance



VOCs, SVOCs, PM10, PM2.6,

Network



TSP, metals

July 2009
A-56
DRAFT-DO NOT CITE OR QUOTE

-------
Network
Lead
Agency
Number of
Sites
Initiated
Measurement
Parameters
Location of Information and/or Data
CAPMoN-Canadian Air
Canada
29
2002
03, NO, NO?, NOy, PAN,
http://www.msc.ec.qc.ca/capmon/index e.cfm
and Precipitation



NHs, PM2.6, PM10 and coarse

Monitoring Network



fraction mass, PM2.6
speciation, major ions for
particles and trace gases,
precipitation chemistry for
major ions

Mexican Air Quality
Mexico
52-62
Late
O3, NOx, CO, SO2, PM10,
http://www.ine.gob.mx/dgicur/calaire/indicadores.html
Network


1960's
TSP.VOC

Mexican City Ambient Air
Mexico
49
Late
O3, NOx, CO, SO2, PM10,
http://www.ine.gob.mx/dgicur/calaire/indicadores.html
Quality Monitoring


1960's
TSP.VOC

Network





AIR TOXICS MONITORING NETWORKS
NATTS—National Air
EPA
23
2005
VOCs, Carbonyls, PM10
http://www.epa.gov/ttn/amtic/airtoxpg.html
Toxics Trends Stations



metalsc, Hg

State/Local Air Toxics
EPA
250 +
1987
VOCs, Carbonyls, PM10
http://www.epa.gov/ttn/amtic/airtoxpg.html
Monitoring



metalsc, Hg

NDAMN-National Dioxin
EPA
34
1998 -
CDDs, CDFs, dioxin-like
http://cfpub.epa.qov/ncea/CFM/recordisplav.cfm7deid-54811
Air Monitoring Network


2005
PCBs

TRIBAL MONITORING NETWORKS
Tribal Monitoring9
EPA
120 +
1995
O3, N0x|N02, SO2,
PM2.BIPM10, CO, Pb
http://www.epa. qov/air/tribal/airproqs.html#ambmon
Industry / Research Networks
New Source Permit
None
variable
variable
O3, N0x|N02, SO2,
Contact specific industrial facilities
Monitoring



PM2.BIPM10, CO, Pb

HRM Network-Houston
None
9
1980
O3, NOx, PM2.BIPM10, CO,
http://hrm.radian.com/houston/how/index.htm
Regional Monitoring



SO2, Pb, VOCs, surface

Network



meteorology

ARIES / SEARCH—Aerosol
None
8
1992
O3, NO/NO2/NOY, SO2, CO,
http://www.atmospheric-research.com/studies/SEARCH/index.html
Research Inhalation



PM2.6/PM10, PM2.6

Epidemiology Study /



speciation, major Ions, NH3,

SouthEastern Aerosol



HNO3, scattering

Research and



coefficient, surface

Characterization Study



meteorology

experiment





SOS - SERON-Southern
EPA
-40
1990
O3, NO, NOy, VOCs, CO,
http://www.ncsu.edu/sos/pubs/sos3/State of SOS 3.pdf
Oxidant Study -



surface meteorology

Southeastern Regional





Oxidant Networks





NA TIONAL/GLOBAL RADIA TION NETWORKS
RadlMet—formerly
EPA
200 +
1973
Radionuclides and radiation
http://www.epa.qov/enviro/html/erams/
Environmental Radiation





Ambient Monitoring





System (ERAMS)





SASP ¦¦ Surface Air
DHS
41
1963
89Sr, 90Sr, naturally
http://www.eml.st.dhs.qov/databases/sasp/
Sampling Program



occurring radionuclides,
7Be, 21 OPb

NEWNET—Neighborhood
DOE
26
1993
Ionizing gamma radiation,
http://newnet.lanl.gov/
Environmental Watch



surface meteorology

Network





Solar Radiation Networks
July 2009
A-57
DRAFT-DO NOT CITE OR QUOTE

-------
Network
Lead
Agency
Number of
Sites
Initiated
Measurement
Parameters
Location of Information and/or Data
UV Index - EPA Sunrise
EPA
-50 U.S.
2002
Calculated UV radiation
http:llwww.epa.govlsunwiseluvindex.html
Program'

cities

index

UV Net ¦¦ Ultraviolet
EPA
21
199512004
Ultraviolet solar radiation
http://www.epa.gov/uvnet/access.html
Monitoring Program



(UV-B and UV-A bands),
irradiance, ozone, NO2

NEUBrew (NOAA-EPA
NOAA
6
2005
Ultraviolet solar radiation
http://www.esrl.noaa.gov/gmd/grad/neubrew/
Brewer Spectrophotometer



(UV-B and UV-A bands),

UV and Ozone Network



irradiance, ozone, SO2

UV-B Monitoring and
USDA
35
1992
Ultraviolet-B radiation
http://uvb.nrel.colostate.edu/UVB/index.isf
Research Program





SURFRAD - Surface
NOAA
7
1993
Solar and infrared radiation,
http://www.srrb.noaa.gov/surfrad/index.html
Radiation Budget Network



direct and diffuse solar
radiation, photosynthetically
active radiation, UVB,
spectral solar, and
meteorological parameters

AERONET-Aerosol
NASA co-
22 + other
1998
Aerosol spectral optical
http://aeronet.gsfc.nasa.gov/index.html
RObotic NETwork
MPLNET - Micro-pulse
Lidar Network
located
networks
participants
8
2000
depths, aerosol size
distributions, and
precipitable water
Aerosols and cloud layer
heights
http://mplnet.gsfc.nasa.gov/
PRIMENet ¦¦ Park Research
NPS
14
1997
ozone, wet and dry
http://www.cfc.umt.edu/primenet/Assets/Announcements/99PReport.pdf
& Intensive Monitoring of
Ecosystems NETwork®



deposition, visibility, surface
meteorology, and ultraviolet
radiation

aNCore is a network proposed to replace NAMS, as a component of SLAMS; NAMS are currently designated as national trends sites.
bsurface meteorology includes wind direction and speed, temperature, precipitation, relative humidity, solar radiation (PAMS only).
ePMio metals may include arsenic, beryllium, cadmium, chromium, lead, manganese, nickel, and others.
dSome networks listed separately may also serve as subcomponents of other larger listed networks; as a result, some double counting of the number of individual monitors is likely.
"The number of sites indicated for tribal monitoring is actually the number of monitors, rather than sites. The number of sites with multiple monitors is probably < 80.
1 Sunrise program estimates UV exposure levels through modeling - does not include measurements.
9NEUBREW is a subset Original UV brewer network (UV Net); PRIMENET participated in UV Net program.
July 2009
A-58
DRAFT-DO NOT CITE OR QUOTE

-------
All Monitor Distribution with Respect to Population Density
2005 Population Density
Atlanta PM2.5 Monitors (15 km buffer)
Population perSq Km
0 - 89
90 - 177
178 - 886
887 - 1772
1773 - 4431
4432 - 17722
] Kilometers
0 5 10 20 30 40 50
] Kilometers
0 15 30 60 90 120 150
Figure A-1. I M ¦ monitor distribution in comparison with population density, Atlanta, GA.
July 2009	A-59	DRAFT - DO NOT CITE OR QUOTE

-------
Kilometers
1 Kilometers
Figure A-2. PMio monitor distribution in comparison with population density, Atlanta, GA.
July 2009
DRAFT-DO NOT CITE OR QUOTE
2005 Population Density
Atlanta PM10 Monitors (15 km buffer)
Population per Sq Km
0-89
90 -177
178-886
887- 1772
1773-4431
4432-17722

-------
] Kilometers
0 5 10 20 30 40 50
1 Kilometers
0 15 30 60 90 120 150
Figure A-3. PM2.5 monitor distribution in comparison with population density, Birmingham, AL.
2005 Population Density
Birmingham PM2.5 Monitors {15 km buffer)
Population per Sq Km
4-23
24-47
48 - 235
236 - 469
470- 1173
1174-4692
July 2009	A-61	DRAFT - DO NOT CITE OR QUOTE

-------
2005 Population Density
Birmingham PM10 Monitors (15 km buffer)
Population per Sq Km
4-23
24-47
48 - 235
236 - 469
470-1173
1174-4692
] Kilometers
0 5 10 20 30 40 50
1 Kilometers
0 15 30 60 90 120 150
Figure A-4. PMio monitor distribution in comparison with population density, Birmingham, AL.
July 2009
A-62
DRAFT-DO NOT CITE OR QUOTE

-------
1 Kilometers
0 15 30 60 90 120 150
Figure A-5. PM2,5 monitor distribution in comparison with population density, Chicago, IL.
2005 Population Density
Chicago PM2 5 Monitors (15 km buffer)
Population per Sq Km
0-220
221 -441
442 - 2204
2205 - 4407
4408- 11019
11020-44074
July 2009	A-63	DRAFT - DO NOT CITE OR QUOTE

-------
D Kilometers
0 5 10 20 30 40 50
2005 Population Density
Chicago PM10 Monitors (15 km buffer)
Population per Sq Km
0-220
221 - 441
442 - 2204
2205 - 4407
4408- 11019
11020-44074
J Kilometers
0 15 30 60 90 120 150
Figure A-6. PMio monitor distribution in comparison with population density, Chicago, IL.
July 2009
A-64
DRAFT-DO NOT CITE OR QUOTE

-------
] Kilometers
0 5 10 20 30 40 50
2005 Population Density
Denver PM2.5 Monitors {15 km buffer)
Population per Sq Km
0-67
68-135
136-673
674-1347
1348-3364
3365- 13456
] Kilometers
0 15 30 60 90 120 150
Figure A-7. PM2,5 monitor distribution in comparison with population density, Denver, CO.
July 2009	A-65	DRAFT - DO NOT CITE OR QUOTE

-------
e)
*
4
$
] Kilometers
0 5 10 20 30 40 50
1
2005 Population Density
Denver PM10 Monitors (15km buffer)
Population per Sq Km
| 0-67
68-135
136-673
674- 1347
1348 - 3364
I 3365-13456
] Kilometers
0 15 30 60 90 120 150
Figure A-8. PMio monitor distribution in comparison with population density, Denver, CO.
July 2009	A-66	DRAFT - DO NOT CITE OR QUOTE

-------
Kilometers
] Kilometers
0 15 30 60 90 120 150
Figure A-9. PM2.5 monitor distribution in comparison with population density, Detroit, Ml.
2005 Population Density
Detroit PM2.5 Monitors (15 km buffer)
Population per Sq Km
| 0-164
Q 165-327
328-1637
1638 - 3274
3275 -8185
8186-32741
July 2009
A-67
DRAFT-DO NOT CITE OR QUOTE

-------
Kilometers

2005 Population Density
Detroit PM10 Monitors (15 km buffer)
Population per Sq Km
| 0-164
| 165-327
328-1637
1638 - 3274
3275-8185
I 8186- 32741
0 15 30
60
90
120
] Kilometers
150
Figure A-10. PMio monitor distribution in comparison with population density, Detroit, Ml.
July 2009
A-68
DRAFT-DO NOT CITE OR QUOTE

-------
of
f /*

3 Kilometers
0 5 10 20 30 40 50
] Kilometers
2005 Population Density
Houston PM2.5 Monitors (15 km buffer)
Population per Sq Km
| 0- 182
| 183-364
365- 1822
1823-3644
3645-9109
I 9110-36435
0 15 30 60 90 120 150
Figure A-11. PMio monitor distribution in comparison with population density, Detroit, Ml.
July 2009
A-69
DRAFT-DO NOT CITE OR QUOTE

-------
Kilometers
0 5 10 20 30 40 50
2005 Population Density
Houston PM10 Monitors (15 km buffer)
Population per Sq Km
I 0 - 182
9110-36435
Kilometers
0 15 30 60 90 120 150
Figure A-12. PMio monitor distribution in comparison with population density, Houston, TX.
July 2009	A-70	DRAFT - DO NOT CITE OR QUOTE

-------
] Kilometers
0 15 30 60 90 120 150
Figure A-13. I M ¦ monitor distribution in comparison with population density, Los Angeles, CA.
3 Kilometers
5 10 20 30 40 50
2005 Population Density
Los Angeles PM2.5 Monitors (15 km buffer)
Population per Sq Km
0-271
272 - 542
543-2711
2712-5422
5423- 13556
13557-54222
July 2009	A-71	DRAFT - DO NOT CITE OR QUOTE

-------
~ Kilometers
5 10 20 30 40 50
I Kilometers
0 15 30 60 90 120 150
Figure A-14. PMio monitor distribution in comparison with population density, Los Angeles, CA.
2005 Population Density
Los Angeles PM10 Monitors (15km buffer)
Population per Sq Km
0-271
272 - 542
543 - 2711
2712 - 5422
5423- 13556
13557-54222
July 2009	A-72	DRAFT - DO NOT CITE OR QUOTE

-------
0 5 10 20 30 40 50
2005 Population Density
New York PM2.5 Monitors (15 km buffer)
Population per Sq Km
I 0 - 832
|' 833- 1664
1665-8319
8320 - 16637
16638 -41593
I 41594- 166371
3 Kilometers
0 1530 60 90 120 150
Figure A-15. PM2.5 monitor distribution in comparison with population density. New York City, NY.
July 2009
A-73
DRAFT-DO NOT CITE OR QUOTE

-------
•'*-> •	
i Kilometers
0 5 10 20 30 40 50
2005 Population Density
I	 New York PM10 Monitors (15 km buffer)
Population per Sq Km
| 0-832
| 833 - 1664
1665- 8319
8320 - 16637
16638 -41593
I 41594 - 166371
] Kilometers
0 15 30 60 90 120 150
Figure A-16. PMio monitor distribution in comparison with population density, New York City, NY.
July 2009
A-74
DRAFT-DO NOT CITE OR QUOTE

-------
Kilometers
2005 Population Density
Philadelphia PM2.5 Monitors (15 km buffer)
Population per Sq Km
I 0-183
9145 - 36577
] Kilometers
0 15 30 60 90 120 150
Figure A-17. I M ¦ monitor distribution in comparison with population density, Philadelphia, PA.
July 2009
A-75
DRAFT-DO NOT CITE OR QUOTE

-------
1 M	I Kilometers
0 5 10	20	30	40	50
2005 Population Density
Philadelphia PM10 Monitors (15 km buffer)
Population perSq Km
¦ 0 -183
Kilometers
0 15 30 60 90 120 150
Figure A-18. PMio monitor distribution in comparison with population density, Philadelphia, PA.
July 2009	A-76	DRAFT - DO NOT CITE OR QUOTE

-------
«r
] Kilometers
0 5 10 20 30 40 50
2005 Population Density
Phoenix PM2.5 Monitors (15 km buffer)
Population per Sq Km
| 0-80
| 81 - 159
160-795
796- 1591
1592-3977
¦ 3978- 15907
] Kilometers
0 15 30 60 90 120 150
Figure A-19. PMz.s monitor distribution in comparison with population density. Phoenix, AZ.
July 2009	A-77	DRAFT - DO NOT CITE OR QUOTE

-------
Kilometers
~ Kilometers
0 15 30 60 90 120 150
2005 Population Density
Phoenix PM10 Monitors (15 km buffer)
lation per Sq Km
| 0-80
81 -159
160-795
796-1591
1592-3977
3978- 15907
July 2009
A-78
DRAFT-DO NOT CITE OR QUOTE

-------
0 5 10
20	30
40
I Kilometers
50
n>
r
2005 Population Density
	 Pittsburgh PM2 5 Monitors (15 km buffer)
Population per Sq Km
| 6 - 204
| 205 - 409
410-2045
2046 - 4090
4091 - 10225
I 10226-40898
0 15 30
] Kilometers
60	90
120 150
Figure A-21. PM2,5 monitor distribution in comparison with population density, Pittsburgh, PA.
July 2009
A-79
DRAFT-DO NOT CITE OR QUOTE

-------
0 5 10 20

30
40
zd Kilometers
50
2005 Population Density
Pittsburgh PM10 Monitors (15 km buffer)
Population per Sq Km
| 6 - 204
| 205 - 409
410-2045
2046 - 4090
4091 - 10225
I 10226-40898
0 15 30
60
90
120
3 Kilometers
150
Figure A-22. PMio monitor distribution in comparison with population density, Pittsburgh, PA.
July 2009
A-80
DRAFT-DO NOT CITE OR QUOTE

-------
I Kilometers
0 5 10 20 30 40 50
TO
2005 Population Density
Riverside PM2.5 Monitors (15 km buffer)
Population perSq Km
| 0-89
| 90-178
179-892
893- 1784
1785-4460
¦ 4461 - 17838
] Kilometers
012.35 50 75 100 125
Figure A-23. PM.-.i monitor distribution in comparison with population density. Riverside, OA.
July 2009
A-81
DRAFT-DO NOT CITE OR QUOTE

-------
1 Kilometers
0 5 10 20 30 40 50
Of
c
3
2005 Population Density
Riverside PM10 Monitors (15 km buffer)
Population per Sq Km
| 0-89
| 90-178
179-892
893- 1784
1785 -4460
¦ 4461 -17838
1 Kilometers
0)2.95 50 75 100 125
Figure A-24. PMio monitor distribution in comparison with population density, Riverside, CA.
July 2009
A-82
DRAFT-DO NOT CITE OR QUOTE

-------
] Kilometers
0 5 10 20 30 40 50
] Kilometers
0 15 30 60 90 120 150
Figure A-25. PM2.5 monitor distribution in comparison with population density, Seattle, WA.
2005 Population Density
Seattle PM2.5 Monitors (15km buffer)
ation per Sq Km
0-120
121 -240
241 -1201
1202 - 2402
2403 - 6006
6007 - 24022
July 2009	A-83	DRAFT - DO NOT CITE OR QUOTE

-------
~ Kilometers
0 5 10 20 30 40 50
2005 Population Density
Seattle PM10 Monitors (15 Km buffer)
Population per Sq Km
0-120
121 - 240
241 -1201
1202 - 2402
2403 - 6006
6007 - 24022
I Kilometers
Figure A-26. PM2.5 monitor distribution in comparison with population density, Seattle, WA.
July 2009	A-84	DRAFT - DO NOT CITE OR QUOTE

-------
' v
I Kilometers
0 5 10 20 30 40 50
2005 Population Density
St Louis PM2.5 Monitors (15 km buffer)
Population per Sq Km
| 0-54
55-109
110-544
545-1088
1089 - 2720
I 2721 - 10878
3 Kilometers
0 15 30 60 90 120 150
Figure A-27. PM2.5 monitor distribution in comparison with population density, St. Louis, M0.
July 2009
A-85
DRAFT-DO NOT CITE OR QUOTE

-------
r
] Kilometers
0 5 10 20 30 40 50
2005 Population Density
St Louis PM10 Monitors (15 km buffer)
Population per Sq Km
| 0- 54
[ 55-109
110 - 544
545 - 1088
1089-2720
I 2721 - 10878
~ Kilometers
0 15 30 60 90 120 150
Figure A-28. PMio monitor distribution in comparison with population density, St. Louis, MO,
July 2009
A-86
DRAFT-DO NOT CITE OR QUOTE

-------
A.2.
A.2.1.
Ambient PM Concentration
Speciation Trends Network Site Data

Copper — Annual

0 0


•° • 0
•
2* e
\ .
X. •
0
0
° ' * 'V„-;: *
•••
0 • 0 «° f ••
° « 0 0 40
!• ^
3 °8 ° J
. • •« 7°"°
©
. T v •
-•
,0"
~
s
0 0 #
•
ccpper(ug/m3): <=0.002
>0.002 to 0.004 >0.004 to 0.006 >0.006- 0.008
>o.ooe

Copper - Spring


o* •
0 •
"" ' ' : * f
° * ® 5 us *
t 0
x. •
0
• % °
* °a
f *• •
»
•
CQBHNW"n3) <*c.a»
O m m
>0 00? too ocm >oaMtooooe >0 oos-o one >oooe

Copper - Fall


•° •
1" 8
1.
Xe •
•
•
S • •
.
° • © ®°"o °
% " °»9 ofq#' o
*	•0 IV?
.• bps
f " *
•
0 • #
cocp«ftua«>3) c»o oro
>Q CO? to 0 UW >0 OW to 0 006 >0006-0006 >DOtB
Copper - Summer
0
e
V


• o**c
•e	°<*>
o D ^
<* °
o _ u A©® O
i
cnppffiuym3l '-onii? >0 OB? to POM	¦QQiMtn OOCfe >0 0(^-0 008 -POOS
Copper - Whiter

£b e
I •
Xc
0 • a	"®
o e
o
0 V*
(Q«ar;tO»3J oOBK> >a nco to0flew >00MW0ta6 >0M-Q0CB >0008
Figure A-29. Three-vr avg of 24-h PM2.5 Ou concentrations measured at CSN sites across the U.S.,
2005-2007.
July 2009
A-87
DRAFT-DO NOT CITE OR QUOTE

-------
Iron - Annual
©
oo
CP
o*
0°
» •
*• *

o • «„
O	O o®1

>P ° O
r°o° <
o	o	O
lron(ugAn3): <=0.05	>0.05 to 0.075	>0.075 lo 0.1
Iran - Spring
M
i
. • O' o9- *
A	• =• '
°
• . P o*~ ©
05 to 0075	>0 075 Id 01
• •
>0.10125	>0.126
•* >
• •»
j0.05»CHJ75	•0.075(00-1	>01-0125	»0115
Iron - Summer
5 o v
s
°Q
I •
•
• t
o»
a*
0°. £*
O B oj
<8o° 0
O	o
® O cgoo?
V
• ••
cgOa*

ft?.


<^0 05	"-O05 to 00 75
O
>0 075toQ1
>0.1-0125	>0125
Iron - Winter
C»
V
"m c "®L
° °-0075 «»0.1	>01*0-125	>0135
Three-yr avg of 24-h PM2.5 iron concentrations measured at DSN sites across the U.S.,
2005-2007
A-8
DRAFT-DO NOT CITE OR QUOTE

-------
&
K
Ni - Annual
9
\
o
©
o° '
o
® cg©°
oo
O O
0 O
6
O 0
O°°o °°#
O O
8f*
G
o
Nickel[ug/m3): <=0.0005 >0.0005 to 0.001 >0.001 to 0.0015 >0.0015-0.002 >0.002
NI - Spring
• 0(9
® » #oZ°° '•
°° . r,o o >
¦ •• °
o
^ °
N.chC^U9frn3> <=00005 >0.0005 to OjOOI »0.001 to 0.0015 >00015-0002 >0002
NI - Fall
9
S
6	f-	¦ A ^
® 0 ePO O ® ©
oo	^ c
© -%•!
NiCk^ug^) « =00005 >0 0005 to 0 001 >0.001 to 0.0015 > 00015-0 002 >0.002
NI - Summer
K
*s
-o

»icka(ua»3); <-0.0005 >oooo5too.ooi >ooottoooots >0.0019-0002 .0.00:
NI - Winter
O
r v. ;•
5	- °4 .?4*>
°»9 °?
O ' •
X
§ )•
0 0©
o
o%*>
Nlckel(ogAn3). c^o.0005 > 0 0005 toOOOl >00011000015 >00015-0 002 >0 002
Figure A-31. Three-vr avg of 24-h PM2.5 nickel concentrations measured at CSN sites across the U.S.,
2005-2007
July 2009
A-89
DRAFT-DO NOT CITE OR QUOTE

-------

Lead -
Annual

& 0 0



0 0
0 0
&
e © «
a
kR
0 0
o°° 3
O O
0 0 a
„ 8
• 0 0
0 V-
0 °a & °
0 
S&
tml
r%>
O
O
0 \ik'
* & go
t 0 a 0
0 OO o
0 u
a
<*><>
6 0 a>
0
0
0 O
O *
Q0
0 0 °o
•
I }°8
O O
O 00/ 9 •JP
o 0 0
0 0
0

0 A * 3-
0 V'.


\ #*"

. ®

0



0
0
• •
Lea0.005 to 0.01
>0.01 10 0.015
>0.015-0.02 >0.02
DO
%
Lead - Spring
° °° \ < a* J*
o *3 O r
<9 u	OB f
® Q? V 0
a a	^ O
. * •••
L^aJ(n^rn3j <=0005	>0005 »Q01	>001100015	>0010-002	>0,02
Lead - Fall
•s
d
BO
%
a
%0
U»00150 0?	>00?
Lead - Summer
A
<% « ® " oJb°
•	D«VsSr
0 CPA
s 4..;°
• *» i
a -
(P°
LC3djUfl»i>3l <3)005	>0005 M 0.01	>001 to 0.015	>C01frOQ2	>002
Lead - Winter
e
%
.. • •• ^ °z$T
« <^3 **

 0OO5ioQOt	sfiOlMO-OlS	>001W»0?
Figure A-32. Three-vr avg of 24-h PM2.5 lead concentrations measured at CSN sites across the U.S.
2005-2007
July 2009	A-90	DRAFT - DO NOT CITE OR QUOTE

-------
Selenium - Annual
% <£
oO
GO ° o

«b„
9
O cPg
@ cg©o
O "(V op
°c oGf
o 0 b 0 0
® O	O
Oo ®
Selemum(ug/m3): <=0.001 >0.001 to 0.002 >0.002 to 0.003 >0.003-0.004 >0.004
Selenium - Spring
O
o
%
o •
j9 0 * °J»
««:w
8 <9«>=- "
"•5 of-l".
0 • D » 0 o
SriB>tum(u^rt3) *-Q 001 >0.001 to 0 002 ¦" 00." jo QQ03 .-0 OCO-O 00*1 >0 OM
t-o
» ,
Fall
* 9 # o°cO o
°"2 ¦>"?.£ *.
0 . • *«S °«°
so	o	9	9
SeieniumCugmiS) <=0001 >0 001 to 0 003 >0 003io0003 >0003-0 004 >0 001
Selenium - Summer
jgS
ore.
°» «>0° W
8
J	^ C5 O
ow o 0 oir o
S<*erHum(u9*ft3) <=0.001 >0.001 to 0002 >0 00210 0003 >0 003-0 004 >OOW
Selenium - Winter
°Q
a
0.
'• „ ^„
0 001 to 0 003 >0 Oft? In 0 003 >0 0030 004 >0 OOj
Figure A-33. Three-yr avg of 24-h PM2.5 selenium concentrations measured at CSN sites across the
U.S., 2005 2007
July 2009
A-91
DRAFT-DO NOT CITE OR QUOTE

-------
Vanadium - Annual
<9
o o
8
V» f ° c
4. o ® %o O,
0 *g°
1 £°° 00
o
o o	^3 rp
°e o o °£
a o„0 0 °
Vana(fum(ug/hn3): *=0.004 >0.004 to 0,006 >0.006 to 0.008 >0.008-0.01 >0.01
Vanadium - Spring
O
. *f* =
s ;•
o _» a> jP o
°o • 0 ojr ©
. .

so	O	•	#
vanadlum(ug/m3> <*0(KM >0 004to0 0fl6 >0006 to 0 008 >0008-0 01 >001
Vanadium — Fall

%
9 „ ffi
« » + t
* f £
& a o
© 0
I
9 •
vanadium (ug/in 3) <=0.0(M > 0 004 to 0 0
>0 00610 0 008 >0 008-0 01 >0 01
Vanadium - Summar

.y
* i	» -
o o	°
O	° o®	
00(Mto0 006 >OOO6toOOO0 >0 006-0 01 >0 01
•%
<>0
I.
a
Vanadium - Winter
o rf> I	» -
* « J."" •
O 0 ® a° a*® ff> «
° o	"® • D	'
. • »o!

Vana*um{ugfrn3) <=0 004 > 0004 to 0 006 >000610 0 003 >0 008-0 01 >0.01
Figure A-34. Three-yr avg of 24-h PM2.5 vanadium concentrations measured at DSN sites across the
U.S., 2005 2007
July 2009
A-92
DRAFT-DO NOT CITE OR QUOTE

-------
A.2.2. Intraurban Variability
The following figures and tables exemplify the intraurban variability among PM2.5,
PMio'2.5 and PM10 measurements for select CSAs/CBSAs (2005-2007) including Atlanta,
Birmingham, Chicago, Denver, Detroit, Houston, New York City, Philadelphia, Phoenix,
Riverside, Seattle and St. Louis. Maps are included to show monitor locations relative to
major roadways. Box plots show the median and interquartile range of concentrations with
whiskers extending to the 5th and 95th percentiles at each site during (l) winter
(December-February); (2) spring (March-May); (3) summer (June-August); and (4) fall
(September-November). Tables of inter-sampler comparison statistics and scatter plots of
inter-sampler correlation vs. distance illustrate variability present in each area.
July 2009
A-93
DRAFT-DO NOT CITE OR QUOTE

-------
July 2009
DRAFT-DO NOT CITE OR QUOTE
Figure A-35. Pib.s monitor distribution and major highways, Atlanta, GA.
• Atlanta PM2.5 Monitors
Atlanta Interstates
Atlanta Major Highways
Atlanta
0 10 20 40 60 80 100
3 Kilometers

-------
Site A	13-063-0091
Site B	13-067-0003
SiteC	13-067-0004
Site D	13-089-0002
SiteE	13-089-2001
Site F	13-121-0032
SiteG	13-135-0002
SiteH	13-139-0003
Site I	13-223-0003

0
A
B
C
D
E
F
G
H
I
Mean
1
16.2
16.2
15.4
15.3
15.2
15.7
15.2
13.9
14.4
Obs
2
351
352
339
1014
946
1036
221
336
344
SD
3
7.5
7.9
7.7
7.2
7.6
8.2
7.1
6.9
7.6
40 -
30 -
CT>
3.
c.
O 20 '
4-»
ro
+-'
c
en
u
c
o
° 10
1=wiriter
2=spring Q _
3=summer
4=fall	1234 1234 1234 1234 1234 1234 1234 1234 123 4
Figure A-36. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Atlanta, GA.
Table A-t. Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Atlanta, GA.

A
B
C
D
E
F
G
H
I
A
1.00
0.88
0.87
0.93
0.89
0.91
0.85
0.72
0.85

(0.0,0.00)
(5.2,0.11)
(6.2,0.12)
(3.9,0.11)
(5.3,0.12)
(4.6,0.11)
(6.9,0.15)
(8.7,0.19)
(7.2,0.15)

351
330
310
330
315
334
207
319
326
B

1.00
0.96
0.89
0.88
0.91
0.88
0.78
0.88


(0.0,0.00)
(4.1,0.08)
(5.7,0.12)
(4.6,0.10)
(3.6,0.08)
(5.6,0.13)
(9.0,0.17)
(6.5,0.13)


352
309
327
314
333
205
313
321
C


1.00
0.87
0.86
0.88
0.85
0.79
0.90



(0.0,0.00)
(5.2,0.12)
(5.6,0.11)
(4.4,0.10)
(5.8,0.13)
(7.9,0.17)
(4.5,0.11)



339
315
304
324
193
298
303
D


CT
1.00
0.89
0.80
0.87
0.74
0.82




(0.0,0.00)
(4.8,0.12)
(3.7,0.11)
(5.8,0.13)
(8.3,0.18)
(7.3,0.15)




1014
883
978
208
314
322
E




1.00
0.79
0.88
0.74
0.83





(0.0,0.00)
(3.8,0.11)
(5.3,0.12)
(7.8,0.17)
(6.4,0.14)





946
904
208
305
309
F

R



1.00
0.88
0.70
0.84


(P90, COD)



(0.0,0.00)
(5.3,0.12)
(8.5,0.19)
(6.3,0.14)


N



1036
213
321
327
G






1.00
0.73
0.79







(0.0,0.00)
(8.8,0.17)
(7.4,0.15)







221
195
198
H







1.00
0.76








(0.0,0.00)
(8.7,0.17)








336
309
I 1.00
(0.0,0.00)
344
July 2009
A-95
DRAFT-DO NOT CITE OR QUOTE

-------
1.00
0.80
~ ~ ~
~ «	ti
~ 4*
~
0.60
o
o
0.40
0.20
0.00 ^	1	1	1	1	1	1	1	1	1	
0	10	20	30	40	50	60	70	80	90	100
Distance Between Samplers (km)
Figure A-37. PM2.5 inter-sampler correlations as a function of distance between monitors for Atlanta,
GA.
July 2009
A-96
DRAFT-DO NOT CITE OR QUOTE

-------
a
r
Birmingham PM2.5 Monitors
Birmingham Interstates
Birmingham Major Highways
Birmingham
0 10 20 40 60 80
100
n Kilometers
Figure A-38. FM2.5 monitor distribution and major highways, Birmingham, AL
July 2009
A-97
DRAFT-DO NOT CITE OR QUOTE

-------
Site A
Site B
SiteC
Site D
Site E
Site F
SiteG
Site H
Site!
Site J
AQS Site ID
01-073-0023
01-073-1005
01-073-1009
01-073-1010
01-073-2003
01-073-2006
01-073-5002
01-073-5003
01-117-0006
01-127-0002
A
Mean 18.9
Obs 1087
SD 102
50
40
30
20
c
Ol
u
c
o
u
10
1=winter
2=spring
3=summer 0
4=fall
B
15.9
363
363
C
14 1
359
8.2
D
16.2
182
8.1
E
17.7
1079
9.4
8.2
15,4
364
7.8
G
14.8
363
7.9
H
14.9
360
8.3
I
14.6
361
7.6
J
14 4
351
7.5
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
Figure A-3S. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Birmingham, AL.
July 2009
A-98
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-2. Inter-sampler correlation statistics for each pair of pm2.5 monitors reporting to AQS for
Birmingham, AL.

A
B
C
D
E
F
G
H
I
J
A
1.00
0.91
0.86
0.91
0.88
0.91
0.87
0.88
0.88
0.84

(0.0,0.00)
(10.4,0.15)
(13.7,0.21)
(9.7,0.13)
(8.1,0.13)
(10.8,0.15)
(12.6,0.18)
(11.7,0.18)
(12.3,0.18)
(12.5,0.19)

1087
360
356
182
1072
361
360
357
358
348
B

1.00
0.93
0.93
0.85
0.96
0.91
0.93
0.93
0.89


(0.0,0.00)
(5.3,0.12)
(4.7,0.09)
(8.3,0.15)
(3.6,0.08)
(5.4,0.11)
(5.1,0.11)
(4.9,0.10)
(6.1,0.12)


363
356
181
359
358
360
355
358
348
C


1.00
0.93
0.81
0.93
0.91
0.94
0.90
0.90



(0.0,0.00)
(5.9,0.13)
(10.1,0.20)
(4.6,0.12)
(4.3,0.12)
(4.0,0.10)
(4.9,0.12)
(4.9,0.11)



359
180
355
354
355
350
353
343
D



1.00
0.88
0.96
0.95
0.95
0.93
0.89




(0.0,0.00)
(7.9,0.12)
(3.6,0.08)
(3.8,0.09)
(4.7,0.10)
(4.7,0.10)
(6.1,0.12)




182
179
179
181
179
180
174
E




1.00
0.87
0.85
0.85
0.86
0.81





(0.0,0.00)
(8.1,0.15)
(8.7,0.16)
(8.8,0.17)
(9.2,0.16)
(10.6,0.18)





1079
360
359
356
357
347
F

R



1.00
0.95
0.95
0.95
0.90


(P90, COD)



(0.0,0.00)
(3.9,0.09)
(4.1,0.10)
(3.4,0.09)
(5.6,0.11)


N



364
359
354
357
348
G






1.00
0.96
0.92
0.89







(0.0,0.00)
(3.3,0.08)
(4.5,0.10)
(4.9,0.11)







363
356
359
350
H







1.00
0.91
0.93








(0.0,0.00)
(5.0,0.11)
(4.3,0.09)








360
354
344
I








1.00
0.87









(0.0,0.00)
(5.8,0.12)









361
349
J









1.00
(0.0,0.00)
351
40	50	60
Distance Between Samplers (km)
100
Figure A-40. PM2.5 inter-sampler correlations as a function of distance between monitors for
Birmingham, AL.
July 2009
A-99
DRAFT-DO NOT CITE OR QUOTE

-------
m
r
* Chicago PM2.5 Monitors
	 Chicago Interstates
Chicago Major Highways
Chicago
0 10 20 40 60 80
100
3 Kilometers
Figure A-41. PM2.5 monitor distribution and major highways, Chicago, It
July 2009
A-100
DRAFT-00 NOT CITE OR QUOTE

-------
Site A
Site B
SiteC
SiieD
Site E
Site F
SiteG
SteH
Site I
Site J
SiteK
AGS Site 10
17-031-0022
17-031-0050
17-031-0052
17-031-0057
17-031-0076
17-031-1016
17-031-2001
17-031-3103
17-031-3301
17-031-4007
17-0314201
Mean	15.3
Obs	178
SO	88
50 •
40
fr-
E
30
e
0
S 20
s
c
u
1	10
1---winter
2=spring
3=summer
4=fall
B
C
D
E
F
G
H
i
J
K
14 7
15.6
15.1
14.8
16,4
146
15.9
152
12.6
13.2
343
984
333
351
345
350
335
361
356
361
84
89
90
6.2
91
82
85
8.3
7.8
8,2
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234 123 4
Site L
Site M
SteN
SfeO
SfeP
Site Q
SieR
SteS
SieT
SieU
AQS Site ID
17-031-6005
17-043-4002
17-089-0003
17-089-0007
17-097-1007
17-111-0001
17-197-1002
17-197-1011
18-089-0006
18-089-0022
sr
£
t
Mean 15.1
Obs 331
SD 87
50
40 ~
c
O
c
at
u
c
o
u
30 -
20 -
10 -
1=w inter
2=spring
3=summer
4=fall
M
14.0
179
85
N
13.5
176
8.5
0
143
174
8.9
P
12.1
181
8.2
Q
12.3
347
7.5
R
14.0
175
8.5
S
11.7
164
7.2
T
144
330
7.7
U
15.6
351
8.2
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
July 2009
A-101
DRAFT-DO NOT CITE OR QUOTE

-------
SiteV
SiteW
SiteX
SrteY
SiteZ
AGS Site ID
18-089-0026
18-089-0027
18-089-1003
18-089-2004
18-089-2010 50
SiteAA 18-091-0011
Site AB 18-091-0012
Site AC 18-127-0020
Site AD 18-127-0024 40
Site AE 55-059-0019
ST
E
cn
3 30
c
.2
TO
£ 20
u
c
Q
u
10 H

V
W
X
Y
Z
AA
AB
AC
AD
AE
Wean
16.5
13.7
14,1
140
140
12.4
12.6
12.6
13,4
12,9
Obs
333
328
334
340
347
532
348
336
346
355
SD
88
7,4
77
7,5
8.1
7.5
7.5
7,5
8.1
7,8
1-winter
2=spring
3=summer
4-fail
123 4 1234 1234 12 34 1234 1234 1234 1234 1234 1234
Figure A-42. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Chicago, IL.
Table A-3. Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Chicago, IL.
A	B	C	D	E	F	G	H	I	J	K	L	M	N	0
A
1.00 0.98
0.93
0.94
0.97
0.95
0.97
0.94
0.96
0.91
0.95
0.95
0.91
0.92
0.89

(0.0,0.00) (3.1,0.08)
(5.5,0.12)
(4.7,0.11)
(3.9,0.09)
(5.7,0.13)
(3.9,0.09)
(4.6,0.12)
(4.2,0.11)
(6.8,0.16)
(5.8,0.14)
(4.6,0.12)
(5.7,0.15)
(6.6,0.15)
(6.0,0.16)

178 156
176
149
154
154
151
156
164
163
166
141
165
152
156
B
1.00
0.94
0.95
0.97
0.95
0.97
0.95
0.96
0.93
0.93
0.95
0.92
0.93
0.90

(0.0,0.00)
(4.6,0.11)
(3.6,0.10)
(3.3,0.08)
(5.2,0.13)
(2.7,0.07)
(4.3,0.11)
(3.4,0.09)
(6.3,0.16)
(6.5,0.15)
(4.0,0.10)
(5.1,0.15)
(5.8,0.14)
(5.2,0.15)

343
320
276
300
296
296
289
312
315
306
288
157
152
150
C

1.00
0.96
0.92
0.91
0.90
0.94
0.92
0.90
0.91
0.92
0.88
0.92
0.86


(0.0,0.00)
(4.4,0.11)
(5.7,0.11)
(4.8,0.11)
(6.0,0.12)
(4.3,0.11)
(5.5,0.11)
(8.8,0.18)
(7.2,0.17)
(4.5,0.12)
(7.5,0.16)
(7.9,0.16)
(7.5,0.17)


984
313
325
318
324
312
336
332
337
311
178
175
173
D


1.00
0.94
0.93
0.94
0.95
0.94
0.93
0.93
0.92
0.89
0.96
0.88



(0.0,0.00)
(3.8,0.10)
(4.2,0.12)
(3.8,0.10)
(4.1,0.13)
(3.3,0.10)
(6.2,0.15)
(5.2,0.14)
(3.6,0.10)
(5.3,0.14)
(5.1,0.13)
(4.5,0.15)



333
286
280
283
270
299
296
289
273
151
146
145
E



1.00
0.95
0.98
0.95
0.98
0.92
0.92
0.95
0.95
0.94
0.92




(0.0,0.00)
(5.0,0.11)
(2.4,0.06)
(4.5,0.11)
(2.6,0.07)
(5.8,0.16)
(5.7,0.15)
(4.4,0.10)
(4.8,0.11)
(5.0,0.11)
(4.6,0.13)




351
306
304
292
320
321
313
286
159
154
152
F




1.00
0.95
0.95
0.96
0.89
0.91
0.94
0.94
0.94
0.94





(0.0,0.00)
(5.1,0.12)
(4.5,0.12)
(4.5,0.10)
(8.5,0.20)
(7.9,0.19)
(5.7,0.12)
(7.0,0.15)
(7.9,0.17)
(7.9,0.16)





345
301
294
322
323
311
285
161
157
154
G





1.00
0.95
0.97
0.90
0.91
0.94
0.95
0.95
0.95






(0.0,0.00)
(4.9,0.12)
(3.0,0.07)
(6.3,0.15)
(5.8,0.14)
(4.7,0.10)
(4.2,0.11)
(5.0,0.12)
(4.4,0.12)






350
284
315
318
309
287
154
149
148
H






1.00
0.95
0.91
0.92
0.94
0.93
0.94
0.91







(0.0,0.00)
(4.3,0.11)
(7.4,0.19)
(6.4,0.18)
(4.4,0.13)
(6.4,0.16)
(7.1,0.16)
(5.9,0.17)







335
311
309
302
275
164
157
156
I







1.00
0.90
0.92
0.96
0.96
0.95
0.93








(0.0,0.00)
(6.7,0.17)
(5.9,0.16)
(3.9,0.10)
(4.6,0.12)
(5.3,0.13)
(4.6,0.14)








361
341
328
304
173
169
166
J








1.00
0.92
0.90
0.91
0.94
0.89
July 2009	A-102	DRAFT - DO NOT CITE OR QUOTE

-------
A	B	C	D	E	F	G	H	I	J	K	L	M	N	0

(0.0,0.00)
(4.7,0.13)
(7.0,0.17)
(5.7,0.14)
(4.4,0.12)
(5.4,0.16)

356
330
304
171
165
164
K

1.00
0.93
0.94
0.96
0.92

R
(0.0,0.00)
(5.9,0.15)
(5.2,0.13)
(4.0,0.10)
(4.9,0.15)

(P90, COD)
361
292
173
166
167
L
N

1.00
0.94
0.95
0.92



(0.0,0.00)
(6.4,0.13)
(5.9,0.13)
(6.0,0.14)



331
147
142
142
M



1.00
0.97
0.95




(0.0,0.00)
(3.9,0.09)
(2.7,0.11)




179
160
165
N




1.00
0.95





(0.0,0.00)
(3.8,0.11)





176
152
0





1.00
(0.0,0.00)
174
1.

P
Q
R
S
T
U
V
W
X
Y
Z
AA
AB
AC
AD
AE
A
0.90
0.89
0.91
0.83
0.96
0.83
0.93
0.95
0.96
0.95
0.98
0.94
0.93
0.95
0.94
0.88

(8.0,0.19)
(7.5,0.19)
(5.6,0.16)
(8.0,0.24)
(4.4,0.11)
(7.2,0.16)
(5.9,0.13)
(4.7,0.12)
(4.4,0.10)
(4.5,0.12)
(3.4,0.10)
(5.8,0.17)
(6.8,0.17)
(6.0,0.16)
(5.4,0.15)
(7.1,0.17)

166
151
157
145
154
162
159
149
156
160
160
154
159
158
159
162
B
0.90
0.90
0.91
0.83
0.95
0.81
0.94
0.95
0.96
0.96
0.98
0.94
0.92
0.93
0.87
0.87

(8.0,0.20)
(6.7,0.17)
(5.5,0.16)
(8.0,0.24)
(3.9,0.10)
(6.7,0.15)
(5.2,0.11)
(5.0,0.11)
(4.0,0.10)
(4.2,0.11)
(2.9,0.09)
(5.9,0.17)
(6.8,0.17)
(6.5,0.18)
(5.3,0.16)
(7.2,0.17)

159
290
153
143
292
310
300
289
292
300
309
288
308
299
305
311
C
0.89
0.91
0.86
0.78
0.90
0.76
0.89
0.90
0.90
0.90
0.93
0.87
0.88
0.88
0.84
0.79

(10.2,0.22)
(8.3,0.19)
(7.1,0.17)
(10.4,0.25)
(6.9,0.13)
(8.5,0.18)
(6.4,0.13)
(7.7,0.15)
(7.1,0.14)
(7.9,0.15)
(6.7,0.15)
(8.6,0.20)
(8.9,0.20)
(9.6,0.21)
(8.7,0.18)
(8.5,0.20)

180
324
172
164
309
327
315
305
311
317
323
491
323
313
323
333
D
0.90
0.94
0.89
0.80
0.92
0.74
0.91
0.94
0.93
0.93
0.95
0.92
0.90
0.90
0.85
0.87

(7.6,0.19)
(6.5,0.16)
(5.5,0.14)
(8.6,0.22)
(5.0,0.12)
(7.8,0.19)
(5.9,0.13)
(5.4,0.13)
(5.4,0.13)
(5.8,0.13)
(4.8,0.13)
(6.4,0.17)
(6.9,0.18)
(7.0,0.19)
(6.2,0.17)
(6.9,0.17)

153
278
147
135
280
294
283
273
282
284
292
274
291
287
290
297
E
0.91
0.92
0.94
0.87
0.94
0.77
0.94
0.96
0.95
0.96
0.96
0.93
0.91
0.90
0.86
0.87

(8.3,0.18)
(5.6,0.16)
(4.9,0.12)
(7.1,0.20)
(4.1,0.10)
(7.5,0.17)
(5.6,0.12)
(4.3,0.10)
(4.4,0.11)
(3.9,0.09)
(3.9,0.11)
(5.8,0.17)
(6.9,0.17)
(6.8,0.18)
(6.3,0.16)
(7.3,0.17)

160
294
155
142
300
320
310
299
303
310
317
292
314
304
313
318
F
0.91
0.92
0.92
0.87
0.90
0.74
0.93
0.92
0.92
0.92
0.93
0.91
0.90
0.87
0.83
0.84

(10.5,0.23)
(8.6,0.20)
(8.5,0.17)
(10.0,0.25)
(6.9,0.14)
(9.2,0.19)
(5.4,0.11)
(8.2,0.16)
(7.2,0.15)
(7.6,0.15)
(6.3,0.16)
(8.5,0.22)
(9.4,0.21)
(9.1,0.23)
(8.3,0.20)
(9.3,0.21)

163
295
159
144
302
320
308
297
305
311
316
292
317
306
317
322
G
0.91
0.91
0.95
0.88
0.94
0.76
0.95
0.97
0.96
0.97
0.96
0.92
0.92
0.90
0.87
0.85

(7.9,0.19)
(5.9,0.16)
(4.1,0.12)
(7.1,0.21)
(3.9,0.10)
(7.5,0.17)
(4.7,0.10)
(3.7,0.09)
(3.6,0.09)
(3.5,0.08)
(3.4,0.11)
(5.7,0.17)
(6.8,0.16)
(6.8,0.18)
(5.9,0.15)
(7.5,0.17)

156
292
154
140
293
315
303
293
296
303
312
288
311
300
308
314
H
0.91
0.93
0.92
0.82
0.92
0.78
0.92
0.91
0.92
0.92
0.93
0.90
0.89
0.88
0.83
0.88

(9.3,0.23)
(7.1,0.20)
(6.6,0.17)
(9.6,0.26)
(5.7,0.13)
(7.5,0.17)
(6.1,0.13)
(6.8,0.15)
(5.9,0.15)
(6.7,0.14)
(5.9,0.15)
(7.7,0.22)
(8.1,0.22)
(8.3,0.22)
(8.1,0.20)
(7.6,0.20)

165
284
158
145
287
307
297
288
292
299
303
281
301
293
301
307
I
0.91
0.92
0.95
0.87
0.93
0.78
0.94
0.95
0.94
0.95
0.95
0.92
0.91
0.89
0.85
0.86

(8.2,0.21)
(6.1,0.17)
(4.7,0.12)
(7.7,0.22)
(4.2,0.10)
(7.1,0.17)
(5.0,0.10)
(4.6,0.11)
(4.8,0.11)
(4.7,0.10)
(4.2,0.13)
(6.5,0.18)
(6.8,0.18)
(7.4,0.19)
(6.6,0.17)
(7.0,0.18)

175
314
168
154
318
338
327
316
322
328
335
306
334
323
334
339
J
0.92
0.91
0.90
0.85
0.89
0.73
0.89
0.89
0.89
0.89
0.91
0.89
0.88
0.87
0.82
0.87

(5.6,0.14)
(5.1,0.14)
(6.2,0.16)
(6.7,0.21)
(6.1,0.17)
(8.6,0.22)
(8.7,0.20)
(6.0,0.16)
(6.3,0.16)
(6.2,0.16)
(5.9,0.16)
(5.6,0.16)
(5.8,0.16)
(6.3,0.17)
(6.3,0.17)
(6.4,0.15)

173
313
167
153
319
341
329
317
327
329
337
307
335
327
336
340
K
0.94
0.93
0.93
0.86
0.91
0.78
0.89
0.91
0.90
0.91
0.92
0.92
0.91
0.90
0.85
0.91

(5.2,0.12)
(4.2,0.12)
(6.1,0.16)
(7.2,0.20)
(5.1,0.16)
(8.0,0.21)
(8.4,0.19)
(5.2,0.15)
(5.5,0.15)
(5.2,0.15)
(5.4,0.14)
(5.2,0.15)
(4.9,0.15)
(5.3,0.16)
(5.3,0.16)
(5.1,0.13)

176
298
169
155
310
327
319
304
313
319
325
301
325
315
323
328
L
0.90
0.92
0.93
0.86
0.93
0.75
0.92
0.93
0.93
0.92
0.95
0.92
0.92
0.90
0.85
0.89

(9.3,0.20)
(6.7,0.17)
(6.7,0.14)
(8.9,0.21)
(5.5,0.12)
(8.2,0.19)
(6.2,0.13)
(5.7,0.13)
(5.8,0.13)
(6.2,0.13)
(5.0,0.13)
(7.2,0.17)
(7.6,0.17)
(7.4,0.18)
(7.3,0.16)
(7.1,0.17)

151
285
144
132
285
301
290
282
286
293
299
277
299
286
294
301
M
0.92
0.95
0.96
0.88
0.91
0.74
0.90
0.94
0.92
0.93
0.92
0.91
0.91
0.88
0.89
0.89

(6.2,0.16)
(4.5,0.14)
(3.4,0.09)
(6.3,0.19)
(5.9,0.14)
(9.0,0.22)
(8.0,0.17)
(4.7,0.12)
(5.2,0.14)
(5.0,0.13)
(5.8,0.16)
(6.4,0.17)
(5.5,0.15)
(6.9,0.19)
(6.2,0.17)
(6.9,0.17)

175
157
165
152
162
171
166
159
164
168
169
158
168
165
168
171
N
0.92
0.98
0.94
0.89
0.92
0.81
0.91
0.92
0.91
0.93
0.93
0.91
0.91
0.88
0.89
0.90

(5.4,0.13)
(2.8,0.08)
(4.1,0.12)
(5.8,0.17)
(6.2,0.13)
(7.9,0.20)
(8.3,0.17)
(4.9,0.12)
(5.6,0.14)
(4.6,0.13)
(4.9,0.15)
(5.4,0.15)
(4.9,0.14)
(6.5,0.17)
(5.4,0.16)
(6.0,0.14)

162
151
153
140
156
165
160
157
158
162
165
153
162
158
161
165
0
0.88
0.92
0.93
0.86
0.89
0.75
0.89
0.92
0.90
0.91
0.90
0.89
0.90
0.87
0.87
0.87

(7.5,0.18)
(4.9,0.15)
(3.8,0.13)
(6.5,0.20)
(5.9,0.15)
(8.8,0.22)
(7.6,0.17)
(5.1,0.13)
(5.8,0.15)
(5.5,0.14)
(5.6,0.16)
(6.1,0.17)
(5.7,0.16)
(7.0,0.19)
(6.2,0.17)
(7.0,0.18)

166
152
157
145
155
166
161
154
158
161
162
151
161
159
162
166
P
1.00
0.92
0.90
0.84
0.88
0.73
0.87
0.89
0.89
0.89
0.90
0.89
0.90
0.89
0.89
0.92

(0.0,0.00)
(5.2,0.13)
(7.1,0.17)
(7.2,0.20)
(8.5,0.20)
(12.0,0.26)
(10.9,0.24)
(6.7,0.18)
(7.4,0.18)
(6.9,0.18)
(7.6,0.19)
(6.1,0.16)
(5.7,0.15)
(6.3,0.17)
(6.4,0.17)
(5.7,0.13)

181
159
166
152
164
174
168
160
166
169
171
158
170
167
170
173
Q

1.00
0.92
0.85
0.88
0.71
0.89
0.90
0.88
0.90
0.90
0.89
0.88
0.86
0.82
0.91


(0.0,0.00)
(5.4,0.16)
(7.2,0.19)
(6.1,0.18)
(9.3,0.24)
(9.1,0.21)
(5.5,0.16)
(6.5,0.16)
(5.5,0.16)
(6.3,0.16)
(5.3,0.16)
(5.3,0.16)
(6.3,0.18)
(5.9,0.17)
(5.5,0.14)


347
154
139
290
309
296
289
294
302
306
292
303
293
303
310
R


1.00
0.91
0.93
0.76
0.90
0.94
0.93
0.94
0.91
0.92
0.92
0.88
0.89
0.89



(0.0,0.00)
(5.8,0.18)
(5.1,0.13)
(8.6,0.22)
(7.5,0.17)
(4.4,0.11)
(5.0,0.13)
(4.0,0.12)
(5.8,0.17)
(6.2,0.17)
(5.6,0.16)
(7.1,0.19)
(6.4,0.17)
(7.1,0.17)



175
143
157
167
161
153
160
161
164
153
164
160
163
166
S



1.00
0.83
0.66
0.81
0.86
0.83
0.87
0.83
0.84
0.84
0.81
0.82
0.80




(0.0,0.00)
(8.5,0.22)
(11.3,0.28)
(11.6,0.26)
(6.7,0.20)
(8.0,0.21)
(7.2,0.19)
(7.3,0.22)
(6.1,0.21)
(7.4,0.20)
(7.8,0.23)
(7.1,0.22)
(9.0,0.22)




164
144
153
148
143
146
148
151
141
151
149
148
153
T




1.00
0.81
0.93
0.95
0.97
0.96
0.97
0.92
0.91
0.92
0.87
0.85





(0.0,0.00)
(5.9,0.15)
(6.2,0.12)
(3.4,0.10)
(3.2,0.09)
(2.9,0.08)
(3.2,0.12)
(5.2,0.17)
(5.5,0.16)
(5.4,0.18)
(4.9,0.15)
(6.6,0.18)





330
318
307
297
302
305
315
284
312
311
313
319
U





1.00
0.77
0.79
0.81
0.81
0.81
0.78
0.76
0.79
0.74
0.69






(0.0,0.00)
(7.6,0.17)
(6.6,0.17)
(6.0,0.15)
(6.3,0.16)
(6.4,0.17)
(8.1,0.22)
(8.4,0.22)
(7.2,0.21)
(7.0,0.19)
(10.0,0.23)






351
327
319
322
326
336
305
334
324
333
338
V






1.00
0.96
0.97
0.95
0.95
0.93
0.91
0.90
0.88
0.83







(0.0,0.00)
(5.9,0.11)
(4.8,0.10)
(5.8,0.12)
(5.7,0.14)
(7.7,0.20)
(8.6,0.20)
(8.3,0.21)
(6.9,0.17)
(9.3,0.22)







339
306
314
316
325
292
323
314
321
325
W







1.00
0.98
0.98
0.96
0.95
0.93
0.92
0.89
0.85








(0.0,0.00)
(2.8,0.06)
(2.5,0.07)
(3.6,0.11)
(4.5,0.15)
(4.8,0.15)
(5.4,0.16)
(3.9,0.13)
(6.9,0.17)
July 2009
A-103
DRAFT-DO NOT CITE OR QUOTE

-------

P
Q
R
S
T
U
V
W
X
Y
Z
AA
AB
AC
AD
AE








328
299
306
312
281
310
306
311
316
X








1.00
0.98
0.97
0.95
0.93
0.94
0.91
0.85









(0.0,0.00)
(2.3,0.07)
(3.3,0.10)
(4.6,0.14)
(4.9,0.14)
(4.9,0.15)
(3.6,0.11)
(6.8,0.17)









334
311
318
286
319
305
316
321
Y









1.00
0.97
0.95
0.93
0.92
0.89
0.85










(0.0,0.00)
(3.6,0.11)
(4.7,0.16)
(5.0,0.15)
(5.3,0.17)
(4.4,0.14)
(6.7,0.18)










340
322
296
322
311
321
326
Z










1.00
0.95
0.93
0.94
0.89
0.86











(0.0,0.00)
(4.6,0.15)
(5.3,0.15)
(4.9,0.15)
(4.1,0.14)
(6.8,0.17)











347
305
331
321
328
335
AA











1.00
0.98
0.97
0.89
0.88












(0.0,0.00)
(2.4,0.07)
(2.9,0.08)
(3.2,0.11)
(5.9,0.17)












532
305
287
300
304
AB












1.00
0.96
0.89
0.86













(0.0,0.00)
(3.1,0.09)
(3.7,0.11)
(6.5,0.17)













346
317
328
333
AC













1.00
0.91
0.85














(0.0,0.00)
(2.8,0.10)
(6.7,0.17)














336
320
322
AD














1.00
0.79















(0.0,0.00)
(7.2,0.18)















346
332
AE















1.00
(0.0,0.00)
355

~ **• ~ ~ .*?+ «, * *
~ ~ ~ • ~ * ~ ~*~
~
~
10	20	30	40	50	60
Distance Between Samplers (km)
70
80
90
100
Figure A-43. PM2.5 inter-sampler correlations as a function of distance between monitors for Chicago,
IL.
July 2009
A-104
DRAFT-DO NOT CITE OR QUOTE

-------
q)\

• Denver PM2.5 Monitors
	 Denver Interstates
Denver Major Highways
Denver
0 10 20 40 60 80
100
I Kilometers
Figure A-44. FM2.5 monitor distribution and major highways, Denver, CO.
July 2009
A-105
DRAFT-DO NOT CITE OR QUOTE

-------
Site A
Site B
SiteC
Site D
Site E
Site F
SiteG
Site H
AGS Site ID
08-001-0008
08-005-0005
08-013-0003
08-013-0012
08-031-0002
08-031-0023
08-123-0008
08-123-0008
Mean
Obs
SD
40 -
30
cn
a.
c
o
M
£ 20
c

U
c
o
u
10
1=winter
2 -spring
3=summer ~
4=fail
A
B
C
D
E
F
6
11.6
12.9
13.8
13.9
13.5
13.7
14.0
356
308
342
308
303
350
1049
7.5
9.4
9.1
9.5
9.9
8.9
8.9
II §
1234 1234 1234 1234 1234 1234 1234 1234
Figure A-45. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Denver, CO.
Table A-4. Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Denver, CO.

A
B
C
D
E
F
G
H
A
1.00
0.74
0.84
0.68
0.86
0.91
0.76
0.83

(0.0,0.00)
(6.0,0.21)
(5.4,0.17)
(7.9,0.26)
(4.1,0.14)
(3.0,0.11)
(5.9,0.19)
(4.6,0.14)

369
353
347
332
362
339
341
325
B

1.00
0.58
0.76
0.92
0.84
0.50
0.49


(0.0,0.00)
(5.7,0.19)
(3.9,0.17)
(3.2,0.13)
(4.4,0.17)
(7.8,0.23)
(6.6,0.21)


363
344
328
356
336
337
323
C


1.00
0.74
0.71
0.75
0.83
0.88



(0.0,0.00)
(4.4,0.19)
(4.5,0.17)
(5.4,0.18)
(3.5,0.14)
(3.7,0.13)



361
326
354
336
333
320
D



1.00
0.82
0.77
0.54
0.57




(0.0,0.00)
(5.6,0.21)
(6.0,0.24)
(7.2,0.24)
(6.4,0.24)




354
347
332
318
305
E




1.00
0.94
0.64
0.60





(0.0,0.00)
(2.3,0.09)
(7.1,0.21)
(5.6,0.18)


R


1046
969
353
330
F

(P90, COD)



1.00
0.68
0.69


N



(0.0,0.00)
(6.6,0.21)
(5.9,0.17)






1006
333
317
G






1.00
0.88







(0.0,0.00)
(3.4,0.13)







359
313
H







1.00
July 2009
A-106
DRAFT-DO NOT CITE OR QUOTE

-------
A E
) C
D E
F G
H
(0.0,0.00)
334
10	20	30	40	50	60	70	80	90	100
Distance Between Samplers (km)
Figure A-46. PM2.5 inter-sampler correlations as a function of distance between monitors for Denver,
CO.
July 2009
A-107
DRAFT-DO NOT CITE OR QUOTE

-------
q!\
r
0 10 20
40
60
80
100
~ Kilometers
Figure A-47. PM2.5 monitor distribution and major highways, Detroit, Ml.
• Detroit PM2.5 Monitors
Detroit Interstates
Detroit Major Highways
Detroit
July 2009
A-108
DRAFT-DO NOT CITE OR QUOTE

-------
Obs
SO
SO •
AGS Site ID
Site A	26-049-0021 Mean
SfeB	26-099-0009
SiteG	26-115-0005
SifeD	26-125-0001
Site F	26-147-0005
Site F	26-161-0008
SiteG	26-163-0001
40 •
c
o
Qj 20 •
c
o
A
11.6
356
7.5
B
C
D
E
F
G
12.9
13.8
13.9
13.5
13.7
14.0
306
342
308
303
350
1049
9.4
9.1
9,5
9.9
8.9
8.9
1=winter
2-spring

4=fall
1 2 3 4 1
2 3 4 12 3 4
12 3 4
12 3 4
12 3 4
12 3 4

AOS Site ID

H
I
J
K
L
M
Site H
26-163-0015
Mean
155
150
14 4
134
17.2
143
Sitel
26-163-0016
Obs






342
572
308
301
344
342
Site J
26-163-0019
SD
Site K
26-163-0025
94
10.5
99
92
10.1
8.8
Site L 26-163-0033
Site M 26-163-0036
40
30
ar
E
1
c
o
S 20
2
o
10
1=winter
2=spring
3=surnroer
4=fail	1 2 3 4 1234 1234 1234 1234 1234
Figure A-48. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Detroit, Ml.
July 2009
A-109
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-5. Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Detroit, Ml.

A B
C
D
E
F
G
H
I
J
K
L
M
A
1.00 0.91
0.86
0.91
0.89
0.90
0.89
0.88
0.89
0.91
0.92
0.87
0.88

(0.0,0.00) (5.9,0.17)
(7.8,0.19)
(6.7,0.17)
(7.6,0.18)
(5.9,0.18)
(8.1,0.20)
(8.3,0.22)
(8.0,0.19)
(7.3,0.17)
(5.5,0.16)
(11.0,0.26)
(7.8,0.21)

356 299
333
301
296
341
349
334
284
301
293
336
333
B
1.00
0.90
0.94
0.92
0.92
0.93
0.90
0.92
0.91
0.92
0.89
0.91

(0.0,0.00)
(6.8,0.17)
(5.3,0.14)
(5.9,0.16)
(5.8,0.17)
(6.2,0.18)
(7.5,0.21)
(5.8,0.18)
(4.9,0.16)
(5.4,0.17)
(10.2,0.24)
(6.1,0.19)

306
286
296
290
294
300
288
277
297
286
292
288
C

1.00
0.90
0.87
0.91
0.93
0.90
0.91
0.90
0.89
0.87
0.93


(0.0,0.00)
(7.0,0.16)
(8.8,0.20)
(5.5,0.15)
(5.9,0.14)
(7.2,0.17)
(6.3,0.16)
(6.2,0.14)
(6.2,0.16)
(10.4,0.20)
(4.9,0.13)


342
289
284
326
335
320
273
286
279
321
319
D


1.00
0.93
0.94
0.96
0.92
0.94
0.94
0.94
0.91
0.92



(0.0,0.00)
(6.3,0.15)
(4.5,0.14)
(4.3,0.13)
(5.8,0.16)
(4.5,0.12)
(3.8,0.11)
(3.6,0.13)
(8.2,0.18)
(6.2,0.15)



308
292
296
303
291
281
297
291
290
290
E



1.00
0.90
0.90
0.89
0.90
0.90
0.90
0.87
0.87




(0.0,0.00)
(7.5,0.18)
(7.3,0.20)
(8.2,0.22)
(7.0,0.19)
(6.4,0.18)
(6.9,0.18)
(10.7,0.25)
(7.7,0.21)




303
291
297
286
276
292
284
288
288
F




1.00
0.95
0.90
0.92
0.92
0.95
0.89
0.93





(0.0,0.00)
(4.5,0.13)
(6.2,0.17)
(5.7,0.15)
(5.2,0.14)
(3.9,0.12)
(9.8,0.21)
(5.7,0.15)





350
343
326
280
297
288
329
326
G





1.00
0.94
0.95
0.92
0.93
0.90
0.95






(0.0,0.00)
(5.1,0.14)
(4.9,0.12)
(4.5,0.14)
(5.6,0.16)
(8.2,0.18)
(4.7,0.12)






1049
336
549
302
295
337
335
H






1.00
0.93
0.91
0.89
0.91
0.91







(0.0,0.00)
(4.8,0.15)
(5.4,0.15)
(6.9,0.18)
(7.6,0.16)
(6.1,0.15)







342
273
290
288
321
319
I







1.00
0.92
0.90
0.92
0.93


R





(0.0,0.00)
(4.4,0.13)
(6.1,0.14)
(7.9,0.18)
(5.8,0.14)


(P90, COD)





572
279
271
274
274
J

N






1.00
0.91
0.90
0.91









(0.0,0.00)
(5.3,0.15)
(8.1,0.17)
(5.6,0.13)









308
288
291
291
K









1.00
0.88
0.91










(0.0,0.00)
(9.5,0.21)
(6.3,0.16)










301
281
283
L










1.00
0.91











(0.0,0.00)
(8.5,0.17)











344
322
M











1.00
(0.0,0.00)
342
July 2009
A-110
DRAFT-DO NOT CITE OR QUOTE

-------
1
0.6
o
o
0.4
0.2
10	20	30	40	50	60	70	80	90	100
Distance Between Samplers (km)
Figure A-49. PM2.5 inter-sampler correlations as a function of distance between monitors for Detroit,
Ml.
July 2009
A-111
DRAFT-DO NOT CITE OR QUOTE

-------
d\
r
• Houston PM2.5 Monitors
	 Houston Interstates
Houston Major Highways
Houston
0 10 20 40 60 80 100
i Kilometers
Figure A-50. PM2.5 monitor distribution and major highways, Houston, TX.
July 2009
A-112
DRAFT-DO NOT CITE OR QUOTE

-------
Site A
Site B
AQS Site ID
48-201-0058
48-201-1035
Mean
Obs
SD
40
30
£
cn
c
o
4-»
ro
4->
c

-------
Table A-6. Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Houston, TX.

A
B
A
1.00
0.66

(0.0,0.00)
(10.0,0.24)

326
310
B

1.00
(0.0,0.00)
1016
R
(P90, COD)
N
1
0.8
0.6
o
o
0.4
0.2
0 	,	,	,	,	,	,	,	,	,	
0	10	20	30	40	50	60	70	80	90	100
Distance Between Samplers (km)
Figure A-52. PM2.5 inter-sampler correlations as a function of distance between monitors for Houston,
TX.
July 2009
A-114
DRAFT-DO NOT CITE OR QUOTE

-------
• New York PM2.5 Monitors
	 New York Interstates
New York Major Highways
New York
0 10 20 40 60 80 100
~ Kilometers
PM25 monitor distribution and major highways, New York City, NY.
A-115
DRAFT-DO NOT CITE OR QUOTE

-------
Site A
Site B
SiteC
Site D
Site E
Site F
Site G
SiteH
Site!
Site J
AQS Site ID
09-001-0010
09-001-1123
09-001-3005
09-001-9003
09-005-0005
09-009-0026
09-009-0027
09-009-1123
09-009-2008
09-009-2123
E
I
c
o
Mean
Obs
SD
50
40
A
13.2
349
8.7
B
12.5
339
8,3
C
12.3
332
8.1
D
11.2
585
7,6
E
8.0
341
6.8
F
121
341
8.0
G
12.4
992
8.1
H
12.8
338
8.5
i
11.1
344
7.5
J
12.8
352
8.2
30
t 20
Site K
Site L
SiteM
Site N
Site 0
SiteP
Site Q
SiteR
SiteS
Site T
c
at
u
c
o
u
1= winter
2=spring
3=summer
4—fall
AQS Site ID
34-003-0003
34-013-0015
34-017-1003
34-021-0008
34-021-8001
34-023-0008
34-027-0004
34-027-3001
34-029-2002
34-031-0005
10

12 3 4
12 3 4
12 3 4
12 3 4
12 3 4
12 3 4
12 3 4
12 3 4
12 3 4
1 2 3

K
L
M
N
O
P
Q
R
S
T
Mean
13.2
133
13.7
12.3
108
12.1
11.3
10.0
10.8
12.9
Obs
345
334
559
545
313
336
336
357
550
330
SD
8.9
8.8
8.5
7.7
7,0
7.7
7.7
7.4
6.8
8,7
50
40
30
3
c
o
+-~
fT5
£ 20
c
m
u
c
o
u
10
l=winter
2-spring
3=summer
4—fall
i
III
¦
ill
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
July 2009
A-116
DRAFT-DO NOT CITE OR QUOTE

-------

AQS Site ID

u
V
W
X
Y
Z
AA
AB
AC
AD
SiteV
34-039-0006
Mean
14.5
13.2
12.9
15.5
13.0
13.0
14.0
11.4
15.9
13.5
SiteW
34-039-2003
Obs










1017
352
332
349
359
1059
342
337
357
363
SiteX
36-005-0080








Site Y
36-005-0083
SO
8.7
8.4
8.4
S.1
8.2
8.2
8.4
7.5
8.9
8.4
SiteZ 38-005-0110 50
Site AA 36-047-0122
SifeAB 38-059-0008
Site AC 36-081-0056
Site AD 36-061-0079 40 -
ST
£
oi
3 30 -
1=winter
2=spring
3=summer 0 -j	|	|	|	|	|	|	|	|	
4=fall	1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
SiteAE 36-081-0128
SiteAF 36-071-0002
Site AG 36-081-0124
Site AH 36-085-0055
Site Al 38-085-0067
Site A J 36-119-1002
AE
Mean 15 3
Chs 34;
60
50
40
30
ST
E
en
3
c
o
Z3
res
£ 20
u
c
o
w
10
1 =winter
2=spring
3=summer 0
4—fall
AF
10 B
342
76
AG
116
m
15
AH
113
337
8 0
Al
114
33b
AJ
11 7
355
7 a
12 3 4 12 3 4
234 1234 1234 1234
Figure A-54. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for New
York City, NY.
July 2009
A-117
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-7. Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
New York City, NY.
Site
A
B
C
D
E
F
G
H
I
J
K
L
M
N
0
P
Q
R
A
1.00
0.89
0.97
0.97
0.82
0.96
0.96
0.96
0.96
0.93
0.91
0.91
0.92
0.88
0.84
0.87
0.89
0.84

(0.0,
0.00)
(5.3,
0.15)
(3.6,
0.09)
(4.8,
0.11)
111.8,
0.33)
(3.8,0.11)
(4.0,0.11)
(3.4,0.10)
(4.6,
0.12)
(5.1,0.12)
(5.8,0.12)
(5.7,0.12)
(5.5,0.13)
(6.6,0.16)
(9.1,
0.19)
(8.3,0.16)
(7.6,
0.16)
(9.3,
0.21)

349
322
316
322
325
328
321
324
326
335
329
316
331
301
296
321
318
316
B

1.00
0.93
0.91
0.78
0.91
0.92
0.91
0.91
0.92
0.83
0.84
0.85
0.82
0.79
0.82
0.82
0.78


(0.0,
0.00)
(4.5,
0.13)
(5.3,
0.14)
110.4,
0.32)
(4.7,0.13)
(4.6,0.13)
(4.6,0.14)
(5.0,
0.14)
(4.5,0.13)
(7.3,0.17)
(7.1,0.17)
(7.8,0.19)
(7.2,0.19)
(7.7,
0.20)
(7.6,0.18)
(6.6,
0.18)
(8.4,
0.22)


339
312
315
319
316
313
313
315
330
319
305
321
291
292
310
307
305
C


1.00
0.98
0.82
0.96
0.95
0.96
0.97
0.94
0.91
0.91
0.91
0.89
0.84
0.88
0.89
0.84



(0.0,
0.00)
(3.4,
0.08)
(10.8,
0.32)
(3.9,0.10)
(4.1,0.11)
(3.6,0.10)
(4.0,
0.11)
(4.8,0.11)
(5.7,0.13)
(5.8,0.14)
(6.5,0.15)
(5.4,0.15)
(6.9,
0.17)
(6.3,0.14)
(6.2,
0.15)
(8.2,
0.20)



332
314
309
310
308
307
310
319
314
299
316
287
289
307
305
297
D



1.00
0.85
0.96
0.96
0.94
0.96
0.92
0.90
0.89
0.91
0.88
0.87
0.89
0.90
0.86




(0.0,
0.00)
(8.4,0.29)
(3.4,0.11)
(3.8,0.11)
(5.0,0.13)
(3.0,
0.10)
(5.5,0.13)
(7.1,0.15)
(6.9,0.15)
(6.7,0.18)
(6.3,0.17)
(6.5,
0.16)
(6.0,0.15)
(5.5,
0.14)
(6.6,
0.18)




565
314
316
532
315
313
325
319
308
517
506
288
311
309
330
E




1.00
0.82
0.82
0.79
0.83
0.81
0.80
0.77
0.76
0.76
0.79
0.78
0.87
0.87





(0.0,0.00)
(10.0,
0.31)
(10.7,
0.33)
(11.4,
0.33)
(8.8,
0.28)
(10.3,
0.32)
(12.5,
0.34)
(13.0,
0.34)
(13.8,
0.39)
(11.6,
0.35)
(9.1,
0.30)
(10.4,
0.32)
(7.9,
0.28)
(7.3,
0.24)





341
321
313
317
319
330
322
305
323
294
291
316
311
305
F





1.00
0.99
0.98
0.98
0.94
0.88
0.89
0.89
0.86
0.85
0.88
0.87
0.83






(0.0,0.00)
(2.1,0.07)
(2.9,0.09)
(2.8,
0.09)
(4.7,0.11)
(6.7,0.14)
(6.8,0.15)
(6.8,0.16)
(6.4,0.17)
(6.8,
0.18)
(6.1,0.15)
(7.3,
0.16)
(7.5,
0.21)






341
314
319
321
328
321
308
323
293
295
312
310
308
G






1.00
0.96
0.98
0.93
0.88
0.89
0.89
0.84
0.84
0.86
0.87
0.82







(0.0,0.00)
(2.9,0.10)
(3.6,
0.11)
(5.2,0.12)
(7.1,0.15)
(6.7,0.15)
(6.9,0.16)
(6.9,0.18)
(8.0,
0.19)
(7.6,0.16)
(8.1,
0.17)
(8.4,
0.23)







992
315
319
326
319
309
526
513
286
310
306
327
H







1.00
0.98
0.94
0.88
0.89
0.89
0.84
0.82
0.85
0.85
0.79








(0.0,0.00)
(3.7,
0.10)
(3.7,0.10)
(7.1,0.14)
(7.1,0.14)
(6.6,0.16)
(6.7,0.18)
(8.1,
0.20)
(7.8,0.17)
(7.5,
0.17)
(9.2,
0.23)








338
320
324
318
303
321
292
285
310
307
304
I








1.00
0.95
0.89
0.90
0.89
0.87
0.85
0.87
0.88
0.83









(0.0,
0.00)
(4.1,0.11)
(7.0,0.16)
(7.0,0.16)
(7.7,0.20)
(6.4,0.18)
(6.6,
0.17)
(6.5,0.16)
(6.5,
0.15)
(7.6,
0.19)









344
327
324
307
323
296
291
313
313
310
J









1.00
0.87
0.87
0.87
0.84
0.79
0.82
0.84
0.79










(0.0,0.00)
(7.0,0.16)
(7.2,0.16)
(8.5,0.17)
(6.9,0.18)
(7.9,
0.20)
(8.1,0.18)
(7.5,
0.17)
(9.0,
0.22)










352
332
316
334
303
299
321
322
316
K










1.00
0.95
0.93
0.88
0.86
0.90
0.92
0.86



R
IP90,







(0.0,0.00)
(3.4,0.09)
(4.5,0.12)
(6.4,0.15)
(7.5,
0.17)
(5.7,0.13)
(5.8,
0.14)
(8.7,
0.20)



COD)







345
314
330
301
296
317
319
312
L


N








1.00
0.97
0.91
0.86
0.94
0.93
0.87












(0.0,0.00)
(4.1,0.10)
(6.4,0.14)
(8.0,
0.18)
(5.2,0.12)
(5.9,
0.13)
(8.3,
0.20)












334
321
289
288
309
303
301
M












1.00
0.91
0.86
0.93
0.92
0.85













(0.0,0.00)
(5.5,0.14)
(8.4,
0.21)
(6.7,0.15)
(7.5,
0.18)
(9.7,
0.25)













559
499
300
326
318
337
N













1.00
0.93
0.95
0.91
0.88














(0.0,0.00)
(4.7,
0.14)
(4.1,0.11)
(5.8,
0.15)
(7.2,
0.20)














545
270
293
292
316
0














1.00
0.93
0.91
0.94















(0.0,
0.00)
(4.3,0.12)
(4.9,
0.14)
(4.3,
0.14)















313
294
287
279
P















1.00
0.94
0.91
















(0.0,0.00)
(4.9,
0.12)
(5.5,
0.16)
















336
308
303
Q
















1.00
0.95

















(0.0,
0.00)
(3.8,
0.13)

















336
307
R	1.00
(0.0,
0.00)
357
S
T
U
V
W
X
Y
Z
AA
AB
AC
AD
AE
AF
AG
AH
Al
AJ
A 0.75
0.89
0.90
0.90
0.88
0.92
0.94
0.93
0.93
0.88
0.89
0.94
0.88
0.89
0.93
0.90
0.89
0.96
(10.4,
(6.1,
(7.1,
(6.0,
(7.2,
(7.2,
(4.0,
(4.7,
(5.5,
(7.6,
(7.3,
(4.4,
(7.2,
(6.7,
(5.1,
(6.2,
(7.5,
(4.4,
0.21)
0.13)
0.15)
0.13)
0.15)
0.16)
0.11)
0.12)
0.13)
0.18)
0.18)
0.11)
0.19)
0.16)
0.12)
0.15)
0.16)
0.12)
323
315
337
299
316
332
342
348
325
320
340
346
326
323
299
317
318
338
B 0.68
0.84
0.83
0.83
0.84
0.84
0.88
0.85
0.84
0.81
0.81
0.86
0.81
0.92
0.84
0.87
0.86
0.88
July 2009
A-118
DRAFT-DO NOT CITE OR QUOTE

-------
S	T	U	V	W	X	V	Z	AA AB AC AD AE	AF	AG AH Al AJ

(10.8,
(5.9,
(8.6,
(6.5,
(6.8,
(9.0,
(5.9,
(6.8,
(7.3,
(7.9,
(8.4,
(7.0,
(8.8,
(5.8,
(6.6,
(6.8,
17.1,
(5.5,

0.23)
0.16)
0.20)
0.18)
0.18)
0.21)
0.16)
0.17)
0.18)
0.20)
0.22)
0.17)
0.23)
0.16)
0.17)
0.18)
0.18)
0.15)

314
307
328
290
305
325
334
338
317
313
331
336
315
316
292
311
309
329
c
0.76
0.89
0.88
0.89
0.89
0.92
0.95
0.93
0.92
0.89
0.88
0.93
0.88
0.89
0.93
0.91
0.89
0.96

(8.5,
16.1,
(7.8,
(6.4,
(6.0,
(7.8,
(4.4,
(5.4,
(5.6,
16.1,
(7.5,
(5.3,
(7.4,
(6.7,
(4.7,
(5.7,
(6.0,
(3.5,

0.20)
0.14)
0.18)
0.15)
0.15)
0.18)
0.11)
0.13)
0.15)
0.16)
0.20)
0.12)
0.20)
0.15)
0.11)
0.14)
0.15)
0.10)

307
304
321
283
297
317
326
331
311
306
325
330
308
312
282
305
302
322
D
0.80
0.88
0.88
0.87
0.88
0.89
0.94
0.92
0.91
0.89
0.87
0.92
0.85
0.90
0.92
0.90
0.91
0.96

(7.7,
(7.3,
18.1,
17.1,
(6.9,
(9.7,
(5.6,
(6.2,
(7.0,
(5.9,
(9.6,
(6.6,
(9.2,
(5.4,
(4.8,
(6.5,
(5.3,
(3.7,

0.19)
0.16)
0.20)
0.17)
0.17)
0.21)
0.14)
0.15)
0.17)
0.16)
0.23)
0.15)
0.23)
0.14)
0.12)
0.16)
0.14)
0.10)

509
306
537
326
304
324
332
548
315
313
330
336
315
313
496
308
310
328
E
0.67
0.79
0.74
0.76
0.75
0.75
0.80
0.78
0.76
0.73
0.74
0.79
0.70
0.88
0.77
0.75
0.80
0.84

(9.8,
111.3,
114.9,
111.7,
112.1,
115.2,
111.5,
113.1,
(13.9,
110.1,
115.7,
113.1,
(15.0,
(7.6,
111.3,
112.5,
(9.4,
(9.8,

0.32)
0.34)
0.40)
0.36)
0.36)
0.41)
0.34)
0.36)
0.38)
0.33)
0.43)
0.35)
0.42)
0.26)
0.32)
0.36)
0.31)
0.29)

315
306
329
290
307
324
334
340
319
314
332
338
316
316
294
310
309
331
F
0.79
0.87
0.86
0.86
0.87
0.89
0.93
0.91
0.90
0.90
0.87
0.91
0.86
0.89
0.93
0.89
0.89
0.94

(7.9,
(6.7,
(8.5,
(6.8,
(6.6,
(8.2,
(5.0,
(6.4,
(6.7,
(5.6,
(8.4,
(6.3,
(8.0,
(6.4,
(4.6,
(5.7,
(5.3,
14.1,

0.19)
0.15)
0.19)
0.16)
0.16)
0.19)
0.12)
0.14)
0.15)
0.16)
0.20)
0.14)
0.21)
0.16)
0.12)
0.15)
0.15)
0.12)

316
306
329
293
309
325
335
340
320
317
334
339
319
317
290
312
310
332
G
0.77
0.87
0.87
0.86
0.88
0.89
0.92
0.90
0.90
0.88
0.86
0.91
0.85
0.88
0.90
0.89
0.88
0.93

(8.7,
(6.3,
(7.8,
(7.0,
(6.3,
(8.3,
(5.4,
(5.7,
17.1,
(7.5,
18.1,
(6.4,
(8.2,
(6.7,
(5.2,
16.1,
16.1,
(5.0,

0.21)
0.15)
0.18)
0.16)
0.15)
0.17)
0.13)
0.14)
0.15)
0.17)
0.19)
0.14)
0.20)
0.17)
0.13)
0.15)
0.16)
0.14)

513
304
928
327
303
319
329
958
315
308
327
333
314
311
856
312
309
325
H
0.74
0.86
0.86
0.87
0.87
0.90
0.92
0.89
0.90
0.88
0.87
0.91
0.86
0.88
0.90
0.88
0.86
0.93

(9.6,
(6.6,
(8.4,
17.1,
(6.9,
(7.5,
(5.2,
(5.6,
(6.4,
(7.3,
(7.9,
(5.7,
(7.3,
(6.9,
(5.6,
(6.8,
(6.4,
(5.0,

0.22)
0.15)
0.18)
0.16)
0.16)
0.17)
0.13)
0.14)
0.15)
0.19)
0.19)
0.13)
0.20)
0.17)
0.14)
0.16)
0.17)
0.13)

314
304
326
289
306
322
331
337
315
310
329
335
313
313
290
308
308
327
I
0.76
0.88
0.87
0.88
0.88
0.90
0.93
0.92
0.91
0.87
0.87
0.92
0.86
0.90
0.93
0.89
0.89
0.94

18.1,
17.1,
(8.7,
(7.4,
(6.9,
(9.4,
(5.7,
(6.5,
(7.2,
(6.2,
(9.6,
(6.3,
(9.2,
(5.5,
15.1,
(6.7,
(5.8,
14.1,

0.20)
0.17)
0.21)
0.17)
0.17)
0.22)
0.15)
0.16)
0.18)
0.17)
0.24)
0.16)
0.24)
0.13)
0.14)
0.18)
0.16)
0.12)

315
308
332
293
309
326
334
343
323
313
332
338
318
319
296
310
311
330
J
0.67
0.84
0.85
0.86
0.86
0.88
0.90
0.88
0.86
0.81
0.85
0.89
0.84
0.90
0.88
0.87
0.84
0.91

(11.1,
(6.6,
(9.0,
(6.7,
(6.8,
(8.8,
(6.1,
[7.1,
[7.3,
(8.2,
(9.0,
(6.9,
(8.9,
(6.4,
(6.4,
[7.5,
[7.7,
(5.6,

0.22)
0.16)
0.19)
0.16)
0.17)
0.19)
0.14)
0.16)
0.17)
0.19)
0.21)
0.15)
0.22)
0.16)
0.15)
0.16)
0.18)
0.14)

327
316
343
301
318
337
345
351
330
324
343
349
327
329
301
321
320
341
K
0.74
0.94
0.92
0.95
0.92
0.92
0.93
0.92
0.93
0.84
0.90
0.93
0.88
0.89
0.90
0.91
0.89
0.92

(10.9,
(3.9,
(5.7,
(3.4,
(4.3,
(6.0,
(3.8,
(4.2,
(4.3,
(8.5,
(6.2,
(3.8,
(6.2,
(7.4,
(5.9,
(5.0,
(6.5,
(4.8,

0.21)
0.11)
0.14)
0.10)
0.12)
0.15)
0.12)
0.12)
0.12)
0.19)
0.17)
0.11)
0.18)
0.17)
0.13)
0.13)
0.15)
0.13)

320
317
336
302
317
330
339
344
324
318
338
343
321
321
294
318
314
335
L
0.78
0.94
0.97
0.98
0.95
0.94
0.93
0.94
0.95
0.86
0.91
0.95
0.91
0.89
0.92
0.96
0.92
0.92

(9.8,
(3.9,
(4.5,
(2.9,
(4.0,
(6.3,
(4.5,
(4.2,
14.1,
18.1,
(6.3,
(4.2,
(5.9,
(6.8,
(5.4,
(4.0,
(6.4,
(5.4,

0.20)
0.11)
0.12)
0.08)
0.10)
0.15)
0.12)
0.11)
0.11)
0.18)
0.17)
0.10)
0.17)
0.17)
0.12)
0.11)
0.14)
0.13)

313
303
325
292
303
314
323
333
306
305
322
327
309
306
283
305
299
319
M
0.80
0.93
0.97
0.97
0.96
0.95
0.95
0.96
0.97
0.88
0.94
0.96
0.93
0.88
0.95
0.96
0.93
0.93

(9.9,
(5.4,
(3.8,
(3.5,
(4.7,
(4.9,
(4.7,
(3.5,
(3.4,
(8.3,
(4.5,
(3.5,
(4.5,
(8.5,
(5.0,
(4.3,
(6.8,
(5.7,

0.22)
0.13)
0.09)
0.09)
0.11)
0.11)
0.12)
0.10)
0.09)
0.20)
0.12)
0.10)
0.13)
0.21)
0.14)
0.10)
0.16)
0.17)

504
318
534
341
319
331
342
545
326
320
339
345
326
323
484
319
318
338
N
0.88
0.86
0.90
0.91
0.91
0.86
0.90
0.89
0.91
0.89
0.85
0.90
0.85
0.85
0.91
0.92
0.92
0.90

(6.4,
(6.5,
(6.5,
(5.7,
(4.5,
(8.2,
(5.9,
(5.4,
(5.3,
(5.6,
18.1,
(5.3,
(7.7,
(7.5,
(4.9,
(4.6,
(5.3,
(5.4,

0.17)
0.16)
0.15)
0.13)
0.13)
0.18)
0.14)
0.15)
0.14)
0.17)
0.18)
0.14)
0.18)
0.20)
0.14)
0.13)
0.14)
0.16)

492
287
519
313
290
301
309
529
297
289
308
313
294
292
477
292
293
306
0
0.87
0.82
0.86
0.87
0.88
0.84
0.87
0.86
0.88
0.89
0.82
0.88
0.81
0.84
0.88
0.89
0.92
0.88

(5.6,
(7.2,
(9.9,
(7.3,
(6.4,
111.1,
(6.7,
(8.6,
(8.2,
(5.2,
(10.3,
(8.4,
(10.6,
(7.0,
16.1,
16.1,
(4.7,
(5.4,

0.16)
0.18)
0.22)
0.18)
0.16)
0.24)
0.18)
0.19)
0.20)
0.15)
0.25)
0.18)
0.25)
0.18)
0.16)
0.18)
0.14)
0.16)

295
289
302
280
284
299
308
312
292
295
307
311
290
290
269
290
283
304
P
0.86
0.89
0.92
0.94
0.95
0.88
0.90
0.89
0.92
0.89
0.87
0.91
0.87
0.87
0.91
0.94
0.95
0.91

(6.2,
(6.2,
(7.4,
(5.0,
(4.0,
(8.9,
(5.9,
(6.8,
(6.4,
(5.3,
(8.9,
(6.4,
(8.3,
(6.7,
(5.5,
(4.7,
(3.5,
(5.0,

0.15)
0.14)
0.17)
0.12)
0.11)
0.19)
0.14)
0.15)
0.14)
0.14)
0.21)
0.14)
0.20)
0.16)
0.13)
0.12)
0.10)
0.14)

312
307
325
296
305
319
329
335
312
309
327
333
313
311
285
307
306
326
Q
0.79
0.92
0.91
0.93
0.92
0.89
0.90
0.90
0.90
0.83
0.90
0.91
0.87
0.91
0.89
0.90
0.91
0.92

18.1,
(5.0,
(8.2,
(6.2,
(5.4,
(9.7,
(6.3,
(7.3,
(7.3,
(6.9,
(9.9,
(6.9,
(9.5,
(4.8,
(6.0,
(6.3,
(5.3,
(5.5,

0.19)
0.14)
0.20)
0.14)
0.15)
0.22)
0.16)
0.17)
0.18)
0.19)
0.24)
0.16)
0.24)
0.14)
0.15)
0.17)
0.14)
0.13)

313
303
327
287
304
321
328
335
314
306
329
332
311
312
287
303
302
324
R
0.82
0.86
0.84
0.85
0.87
0.82
0.87
0.84
0.86
0.83
0.81
0.86
0.78
0.88
0.86
0.87
0.90
0.88

(6.5,
(7.6,
(10.9,
(8.2,
(7.0,
111.6,
(8.6,
(10.0,
19.1,
(7.0,
111.2,
(9.5,
111.0,
(6.2,
(7.6,
(7.8,
(5.7,
(6.3,

0.20)
0.21)
0.26)
0.21)
0.20)
0.28)
0.21)
0.23)
0.24)
0.21)
0.30)
0.22)
0.30)
0.17)
0.20)
0.22)
0.17)
0.17)

330
296
347
291
304
314
323
355
309
301
324
327
309
308
304
298
302
320
S
1.00
0.69
0.77
0.75
0.78
0.72
0.78
0.79
0.82
0.88
0.74
0.78
0.73
0.69
0.85
0.82
0.88
0.79

(0.0,
110.4,
(10.5,
(10.5,
(9.2,
112.5,
(8.6,
(9.3,
(9.4,
(5.0,
111.6,
(9.9,
111.5,
(10.3,
(7.2,
(8.5,
(5.5,
18.1,

0.00)
0.22)
0.24)
0.21)
0.19)
0.25)
0.19)
0.21)
0.20)
0.16)
0.26)
0.20)
0.25)
0.22)
0.17)
0.18)
0.14)
0.19)

550
306
525
324
306
325
336
536
319
314
333
339
322
316
478
310
312
331
T

1.00
0.92
0.93
0.93
0.92
0.91
0.91
0.90
0.80
0.87
0.93
0.85
0.89
0.88
0.90
0.86
0.91


(0.0,
(6.0,
(4.5,
(4.8,
(6.7,
(5.2,
(4.9,
(5.6,
(8.4,
(6.9,
(4.9,
(7.2,
(6.3,
(6.2,
(5.8,
(7.3,
(5.9,


0.00)
0.15)
0.12)
0.13)
0.16)
0.14)
0.13)
0.14)
0.20)
0.18)
0.12)
0.20)
0.17)
0.14)
0.14)
0.17)
0.14)


330
319
293
301
313
323
329
308
303
321
327
306
308
281
306
298
319
U


1.00
0.98
0.97
0.92
0.91
0.93
0.94
0.84
0.91
0.94
0.90
0.86
0.90
0.96
0.91
0.90



(0.0,
(3.9,
(5.0,
(5.4,
(6.9,
[5.2,
(4.9,
(9.9,
(5.0,
[5.2,
(5.9,
(9.6,
(6.9,
(5.8,
(8.4,
[7.5,



0.00)
0.10)
0.12)
0.12)
0.15)
0.13)
0.12)
0.22)
0.12)
0.12)
0.14)
0.23)
0.17)
0.12)
0.18)
0.19)



1017
341
325
337
347
987
332
326
346
351
330
330
878
325
323
343
V



1.00
0.97
0.93
0.92
0.92
0.94
0.83
0.90
0.94
0.91
0.88
0.89
0.96
0.91
0.91




(0.0,
(2.8,
16.1,
(5.0,
(4.4,
(4.2,
(8.2,
(6.6,
(4.3,
16.1,
(7.0,
(5.8,
(4.0,
(6.4,
(5.3,




0.00)
0.09)
0.14)
0.13)
0.12)
0.12)
0.19)
0.16)
0.11)
0.16)
0.18)
0.15)
0.10)
0.15)
0.15)




352
288
300
307
351
294
290
305
311
294
290
301
291
287
304
W




1.00
0.90
0.91
0.91
0.92
0.85
0.89
0.93
0.88
0.87
0.89
0.96
0.92
0.90
July 2009
A-119
DRAFT-DO NOT CITE OR QUOTE

-------
S T U V W
X
Y
Z
AA
AB
AC
AD
AE
AF
AG
AH
Al
AJ
(0.0,
0.00)
(7.0,
0.16)
(5.5,
0.13)
(5.3,
0.13)
(4.8,
0.13)
(7.0,
0.18)
(6.8,
0.17)
(5.0,
0.12)
(6.9,
0.18)
(6.8,
0.18)
(5.1,
0.14)
[3.7,
0.10)
(4.9,
0.13)
(5.0,
0.15)
332
316
325
331
310
309
323
328
308
310
281
304
303
320
X
1.00
0.96
0.97
0.95
0.86
0.94
0.97
0.93
0.88
0.93
0.92
0.88
0.93

(0.0,
0.00)
(5.8,
0.13)
(4.4,
0.11)
(5.0,
0.11)
(10.0,
0.23)
(3.3,
0.09)
(4.5,
0.11)
14.1,
0.11)
(9.8,
0.24)
(6.9,
0.17)
(6.5,
0.14)
(9.2,
0.20)
(8.2,
0.19)

349
344
349
328
324
342
348
326
326
301
319
319
340
Y

1.00
0.97
0.96
0.90
0.93
0.98
0.93
0.89
0.97
0.93
0.92
0.97


(0.0,
0.00)
(3.2,
0.08)
(3.9,
0.09)
(6.5,
0.16)
(5.4,
0.15)
(2.8,
0.08)
(5.4,
0.15)
(6.5,
0.18)
(3.3,
0.09)
(4.9,
0.12)
(5.3,
0.13)
(3.5,
0.11)


359
359
338
333
352
358
337
335
308
328
329
350
Z


1.00
0.97
0.90
0.94
0.98
0.92
0.88
0.95
0.93
0.91
0.95



(0.0,
0.00)
(2.9,
0.09)
(7.2,
0.17)
(4.4,
0.13)
11.8,
0.07)
(4.6,
0.14)
(7.8,
0.19)
(4.0,
0.10)
(4.7,
0.11)
16.1,
0.14)
(4.9,
0.14)



1059
342
337
357
363
341
342
919
337
335
355
AA



1.00
0.92
0.94
0.98
0.93
0.87
0.97
0.95
0.94
0.95




(0.0,
0.00)
17.1,
0.18)
(3.8,
0.11)
(2.9,
0.07)
14.1,
0.11)
18.1,
0.20)
(4.0,
0.11)
(4.3,
0.10)
(6.6,
0.15)
15.1,
0.15)




342
317
336
341
319
319
292
313
312
335
AB




1.00
0.85
0.89
0.86
0.79
0.95
0.90
0.93
0.90





(0.0,
0.00)
(9.2,
0.24)
(7.2,
0.17)
(9.0,
0.23)
18.1,
0.20)
14.1,
0.13)
16.1,
0.16)
(3.8,
0.12)
(5.5,
0.16)





337
330
337
316
313
291
310
310
329
AC





1.00
0.95
0.98
0.84
0.96
0.91
0.89
0.91






(0.0,
0.00)
(4.4,
0.13)
(3.0,
0.08)
110.4,
0.26)
(6.6,
0.18)
(6.6,
0.15)
(9.3,
0.21)
(7.4,
0.22)






357
356
334
336
304
326
326
348
AD






1.00
0.93
0.89
0.97
0.95
0.93
0.96







(0.0,
0.00)
(4.6,
0.14)
17.1,
0.18)
(4.0,
0.10)
(4.4,
0.10)
(6.0,
0.13)
(4.6,
0.13)







363
341
339
311
333
333
354
AE







1.00
0.82
0.94
0.92
0.89
0.89








(0.0,
0.00)
(10.0,
0.26)
(6.2,
0.18)
(5.6,
0.15)
(8.4,
0.20)
(8.0,
0.22)








341
319
290
313
314
332
AF








1.00
0.86
0.87
0.87
0.91









(0.0,
0.00)
(7.0,
0.16)
17.1,
0.18)
(6.4,
0.16)
(5.5,
0.14)









342
289
310
313
331
AG









1.00
0.93
0.94
0.96










(0.0,
0.00)
(4.8,
0.12)
(4.5,
0.11)
(3.7,
0.11)










951
289
283
304
AH










1.00
0.97
0.92











(0.0,
0.00)
14.1,
0.10)
(4.9,
0.15)











337
307
327
_A|	1.00 0.92
(0.0, (4.8,
	0.00) 0.14)
	335 324
AJ	L00_
(0.0,
	om_
	355
July 2009
A-120
DRAFT-DO NOT CITE OR QUOTE

-------
~ s^/i 7yGv"<:/V* ~~~~	Ct* ^
~ * .	~ ~~ *~ ~ * ~.~~ ~
~~* ~ .*
~
10	20	30	40	50	60	70	80	90	100
Distance Between Samplers (km)
Figure A-55 PM2.5 inter-sampler correlations as a function of distance between monitors for New
York City, NY.
July 2009
A-121
DRAFT-DO NOT CITE OR QUOTE

-------
0/
r
• Philadelphia PM2.5 Monitors
	 Philadelphia Interstates
Philadelphia Major Highways
Philadelphia
0 10 20 40 60 80
100
~ Kilometers
Figure A-56. PM2,5 monitor distribution and major highways, Philadelphia, PA.
July 2009
A-122
DRAFT-DO NOT CITE OR QUOTE

-------
AOS Site ID	A
Site A 10-003-1003 yearl 134
Site B 10-003-1007 0bs
SiteC 10-003-1012
SiteD 10-003-2004	11
Site E 24-015-0003 60 '
Site F 34-007-0003
Site G 34-007-1007
Site H 42-017-0012
40 •
ST*
S
30
S 20
c
o
10
1=winter
2=spring
3=summer
4~fali
B
C
D
E
F
G
H
12,8
13.5
14.7
12.5
13,5
13.5
13.2
346
331
999
348
539
340
317
7.4
7.7
8.3
7.4
8.0
8,5
8.2
1234 1234 1234 1234 1234 1234 1234 1234
Site!
Site J
Site K
Site L
Site M
Site N
SitoO
AOS Site ID
42-029-0100
42-045-0002
42-091-0013
42-101-0004
42-101-0024
42-101-0047
42-101-0138
sr
E
Mean	14.2
Obs	277
5°	8,3
50 '
40
— 30
c
o
c
*
u
c
o
u
20 -
10
1=winter
2=spring
3=summer
4=falt
J
K
L
M
N
O
150
12.6
13.8
12,8
14.9
13.4
331
307
890
805
596
780
8,1
7,7
8,4
80
8,3
77
1234 12 3 4 1234 1234 12 3 4 1234 12 3 4
Figure A-57. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Philadelphia, PA.
July 2009
A-123
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-8. Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Philadelphia, PA.
ABCDEFGHIJKLMNO
A 1.00	034	056	038	052	056	053	059	035	052	056	096	096	095	0.97
(0.0,0.00) (4.7,0.12) (3.1,0.08) (3.2,0.08) (4.8,0.12) (3.5,0.10) (4.2,0.11) 15.3,0.13) (4.2,0.12) (4.6,0.14) (4.7,0.15) (3.5,0.08) (3.7,0.10) (4.5,0.12) (3.2,0.08)
335	305	282	318	311	312	308	289	247	298	277	283	243	236	236
_B	UN)	055	053	054	092	088	083	050	057	051	051	052	058	0.89
	(0.0,0.00) (4.3,0.12) (6.4,0.15) (3.4,0.11) (5.2,0.14) (6.0,0.15) (6.8,0.17) (6.7,0.17) (6.5,0.18) (5.9,0.18) (6.5,0.14) (5.0,0.14) (7.3,0.17) (5.9,0.13)
	346	288	329	318	313	31_5	293	253	302	285	293	253	238	246
_C	150	056	055	054	058	058	053	058	054	053	053	051	0.93
	(0.0,050) (4.3,0.09) (3.5,0.11) (4.7,0.12) (5.3,0.14) (6.0,0.14) (3.5,0.12) (6.6,0.16) (5.5,0.17) (5.0,0.12) (4.8,0.13) (6.0,0.14) (4.6,0.11)
	331	312	289	292	286	270	242	278	261	281	245	225	237
_D	150	051	054	052	058	054	050	055	055	053	053	0.95
	(0.0,050) (6.5,0.15) (4.9,0.12) (5.0,0.14) (6.3,0.15) (4.1,0.12) (5.3,0.14) (5.8,0.18) (4.3,0.11) (5.6,0.14) (4.2,0.10) (4.5,0.11)
	999	325	490	31_7	297	257	312	287	801	732	540	704
J	150	051	057	053	050	056	056	058	050	057	0.89
	(0.0,050) (5.6,0.14) (6.1,0.15) (6.7,0.16) (6.6,0.16) (7.1,0.19) (5.7,0.15) (6.8,0.15) (5.3,0.13) (7.0,0.18) (5.7,0.13)
	348	320	321	301	255	310	287	296	255	242	254
_F	150	055	050	052	059	057	056	056	055	0.96
	(05,0.00) (3.4,0.09) (5.3,0.13) (5.4,0.14) (5.9,0.16) (4.4,0.15) (3.7,0.10) (3.6,0.10) (4.5,0.13) (3.4,0.09)
	539	31_7	296	261	309	284	466	437	£4	396
_G	150	050	050	057	055	053	057	052	0.96
	(0.0,050) (4.8,0.14) (5.9,0.16) (6.2,0.17) (4.7,0.16) (3.7,0.09) (3.1,0.09) (5.7,0.13) (3.5,0.08)
	340	295	258	305	289	288	251	235	240
_H	150	054	053	059	050	054	057	0.89
	(05,0.00) (5.7,0.16) (8.0,0.19) (4.4,0.13) (5.0,0.13) (4.0,0.12) (5.9,0.17) (4.8,0.13)
	31_7	240	288	275	273	234	21_5	227
J	150	057	051	051	052	050	0.92
	(05,0.00) (5.5,0.17) (5.7,0.17) (4.9,0.14) (5.4,0.15) (5.2,0.16) (5.1,0.14)
	R	277	248	228	235	215	196	195
J	(P90, COD)	150	079	059	059	059	0.91
	N	(0.0,050) (7.4,0.21) (5.8,0.15) (6.4,0.17) (5.7,0.13) (5.0,0.14)
	331	278	282	246	237	231
_K	150	057	055	054	0.86
	(05,0.00) (4.7,0.15) (3.7,0.13) (6.8,0.20) (4.3,0.13)
	307	268	230	211	212
_L	150	058	055	0.97
	(0.0,050) (3.1,0.09) (3.7,0.11) (3.4,0.07)
	890	672	512	630
M	LOO	055	0.96
	(0.0,050) (4.7,0.14) (3.2,0.09)
	805	495	563
J	LOO	0.97
	(0.0,050) (3.5,0.10)
	596	447
_0	1.00
	(0.0,050)
	780
July 2009
A-124
DRAFT-DO NOT CITE OR QUOTE

-------
1
• *%#*»•~~~ ~ ~
~ ~ ~ ~~ ~ ~ ~ ~ ~~
%	~ ~ ~~	A t
\ ~	~ f ~
~ ~
* ^ ~~
~
~ %
0.6
o
o
0.4
0.2
10	20	30	40	50	60	70	80	90	100
Distance Between Samplers (km)
Figure A-58. PM2.5 inter-sampler correlations as a function of distance between monitors for
Philadelphia, PA.
July 2009
A-125
DRAFT-DO NOT CITE OR QUOTE

-------
Q)
• Phoenix PM2.5 Monitors
	 Phoenix Interstates
Phoenix Major Highways
Phoenix
0 10 20 40 60 80 100
~ Kilometers
Figure A-59. PM2,5 monitor distribution and major highways. Phoenix, AZ.
July 2009
A-126
DRAFT-DO NOT CITE OR QUOTE

-------
AGS Site ID
Site A 04-013-0019'
Site B 04-013-4003
SiteC 04-013-9997
Site D 04-021-0001
Site E 04-021-3002
Mean
Obs
SD
50
40
E
cn
c
o
+¦»
to
0>
u
c
o
u
30
20
10
1=winter
2	spring
3	summer q
4=fall
A
12.3
370
7,8
B
12.6
352
7.3
C
9.8
360
5.5
D
8.9
227
4.4
E
5.9
325
2.8
1234 1234 1234 1234 1234
Figure A-60. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Phoenix, AZ.
Table A-9. Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Phoenix, AZ.

A
B
C
D
E
A
1.00
0.87
0.92
0.50
0.12

(0.0,0.00)
(6.4,0.15)
(6.5,0.16)
(10.4,0.25)
(14.4,0.40)

370
345
355
222
321
B

1.00
0.89
0.54
0.23


(0.0,0.00)
(6.8,0.17)
(9.6,0.25)
(13.2,0.40)


352
338
212
307
C


1.00
0.54
0.18



(0.0,0.00)
(7.2,0.20)
(9.3,0.33)
July 2009
A-127
DRAFT-DO NOT CITE OR QUOTE

-------

A
B C
D
E


360
216
315
D


1.00
0.51
(0.0,0.00)	(7.8,0.27)
227	200
1.00
(0.0,0.00)
40	50	60
Distance Between Samplers (km)
100
Figure A-61. PM2.5 inter-sampler correlations as a function of distance between monitors for Phoenix,
AZ.
July 2009
A-128
DRAFT-DO NOT CITE OR QUOTE

-------
q!\
*
• Riverside PM2.5 Monitors
	Riverside Interstates
Riverside Major Highways
Riverside
0 1020 40 60 80 100
^¦=1 Kilometers
Figure A-62. PM2,5 monitor distribution and major highways. Riverside, CA.
July 2009
A-129
DRAFT-DO NOT CITE OR QUOTE

-------
Site A
Site B
SiteC
Site D
Site E
Site F
SiteG
AQS Site ID
06-065-1003
06-065-2002
06-065-8001
06-071-0025
06-071-0306
06-071-2002
06-071-9004
cn
c
o
ro
c
(U
u
c
o
u
A
Mean 17.7
Obs 314
SD 11.6
70 r_
60
50
40
30
20
10
1--winter
2=spring
3=summer g
4—fall
B
9.9
310
4.9
C
19.7
934
12.7
D
18.4
319
10.9
E
9.9
236
4.4
F
18.4
328
11.9
I
f
G
17.7
310
12.2
1234 1234 1234 1234 1234 1234 1234
Figure A-63. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Riverside, CA.
July 2009
A-130
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-10. Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Riverside, CA.

A
B
C
D
E
F
G
A
1.00
0.45
0.96
0.92
0.36
0.94
0.90

(0.0,0.00)
(20.6,0.32)
(5.0,0.10)
(7.2,0.13)
(22.1,0.35)
(6.0,0.12)
(5.7,0.13)

314
269
297
282
191
281
273
B

1.00
0.49
0.49
0.42
0.49
0.50


(0.0,0.00)
(22.7,0.35)
(20.9,0.34)
(8.2,0.25)
(19.7,0.33)
(18.8,0.31)


310
289
270
203
285
266
C


1.00
0.91
0.37
0.92
0.91



(0.0,0.00)
(8.2,0.14)
(26.6,0.37)
(6.9,0.12)
(7.6,0.12)



934
300
227
302
287
D



1.00
0.36
0.93
0.82




(0.0,0.00)
(20.1,0.35)
(6.7,0.14)
(9.6,0.17)




319
195
289
274
E

R


1.00
0.40
0.41


(P90, COD)


(0.0,0.00)
(21.1,0.36)
(21.6,0.34)


N


236
201
190
F





1.00
0.90






(0.0,0.00)
(6.7,0.12)






328
276
G






1.00
(0.0,0.00)
310
~ ~
~
~ ««
0	10	20	30	40	50	60	70	80	90	100
Distance Between Samplers (km)
Figure A-64. PM2.5 inter-sampler correlations as a function of distance between monitors for
Riverside CA.
July 2009
A-131
DRAFT-DO NOT CITE OR QUOTE

-------
Q)
r
• Seattle PM2.5 Monitors
	 Seattle Interstates
Seattle Major Highways
Seattle
0 10 20 40 60 80
100
~ Kilometers
Figure A-65. PM2.5 monitor figudistribution and major highways, Seattle, WA.
July 2009
A-132
DRAFT-DO NOT CITE OR QUOTE

-------
Site A
Site B
SiteC
AGS Site ID
53-033-0024
53-053-0029
53-061-1007

A
B
c
Mean
8.9
10.2
9.2
Obs
352
354
591
SD
7,3
10.1
7.9
1!
c
.2
ro
fin -
50
40
30
c
(L>
U
§ 20
10
1 =winter
2=spring
3=summer 0
4—fali
1234 1234 1234
Figure A-66. Box plots illustrating the seasonal distribution of 24-h avg PM2.5 concentrations for
Seattle, WA.
Table A-11. Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for
Seattle, WA.


A
B
C
A

1.00
0.89
0.86


(0.0,0.00)
(6.3,0.16)
(4.5,0.14)


352
337
331
B
R

1.00
0.80

(P90, COD)

(0.0,0.00)
(7.8,0.20)

N

354
335
C



1.00
(0.0,0.00)
591
July 2009
A-133
DRAFT-DO NOT CITE OR QUOTE

-------
10	20	30	40	50	60
Distance Between Samplers (km)
70
80
90
100
Figure A-67. PM2.5 inter-sampler correlations as a function of distance between monitors for Seattle,
WA.
July 2009
A-134
DRAFT-DO NOT CITE OR QUOTE

-------
Interstates
Louis Major Highways
Louis
100
ZD Kilometers
Figure A-68. FM2.5 monitor distribution and major highways, St. Louis, MO.
July 2009
A-135
DRAFT-DO NOT CITE OR QUOTE

-------
Site A
SiteB
SiteC
SrteD
SiteE
Site F
AGS Site ID
17-083-1001
17-119-1007
17-118-2009
17-119-3007
17-183-0010
17-183-4001
Mean
Obs
SO
50
A
13,2
173
7.9
B
16.5
329
82
C
14.6
163
7.7
D
14.4
349
7.5
E
15.8
186
7.8
F
14.2
34S
7.1
ST
E
c
o
c
0>
u
c
o
u
40
30
20
10
1=winter
2=$prirtg
3=summer
4=fall
0 -
!i
Site G
SiteH
Site I
Site J
Site K
Site L
AOS Site ID
29-099-0012
29-183-1002
29-189-2003
29-510-0007
29-510-0085
29-510-0087
12 3 4
6
Mean 13.9
Obs 1040
SD 74
50 '

-------
Table A-12. Inter-sampler correlation statistics for each pair of PM2.5 monitors reporting to AQS for St.
Louis, MO.
Site
A









K
L
A
1.00
0.85
0.93
0.89
0.88
0.86
0.85
0.93
0.86
0.84
0.84
0.88

(0.0,0.00)
(10.5,0.23)
(4.7,0.17)
(5.0,0.17)
(7.3,0.20)
(6.2,0.18)
(4.8,0.17)
(4.1,0.13)
(4.4,0.16)
(6.0,0.18)
(5.7,0.19)
(5.3,0.17)

173
156
129
162
146
156
167
158
162
168
169
166


1.00
0.89
0.86
0.85
0.82
0.88
0.89
0.88
0.86
0.87
0.89


(0.0,0.00)
(8.6,0.16)
(7.4,0.16)
(7.7,0.16)
(8.6,0.17)
(7.8,0.17)
(8.2,0.18)
(7.9,0.17)
(7.7,0.17)
(7.5,0.16)
(6.8,0.14)


329
135
301
156
306
312
305
318
316
316
315



1.00
0.94
0.91
0.88
0.90
0.96
0.94
0.90
0.89
0.94



(0.0,0.00)
(4.0,0.11)
(6.4,0.13)
(5.7,0.13)
(5.5,0.13)
(3.9,0.11)
(5.3,0.11)
(5.7,0.13)
(5.6,0.14)
(4.4,0.11)



163
139
124
133
158
141
144
158
160
156




1.00
0.89
0.84
0.89
0.94
0.92
0.89
0.88
0.92




(0.0,0.00)
(5.7,0.13)
(6.0,0.15)
(4.9,0.12)
(4.3,0.12)
(4.5,0.11)
(4.7,0.13)
(4.6,0.12)
(3.9,0.11)




349
156
314
331
315
326
335
332
336





1.00
0.90
0.91
0.90
0.91
0.93
0.91
0.95





(0.0,0.00)
(5.5,0.12)
(6.2,0.13)
(5.8,0.16)
(5.3,0.14)
(5.1,0.13)
(4.9,0.13)
(3.7,0.10)





166
152
159
153
157
160
163
160






1.00
0.89
0.86
0.88
0.88
0.85
0.88






(0.0,0.00)
(5.4,0.12)
(6.1,0.16)
(5.4,0.13)
(5.3,0.14)
(5.6,0.14)
(5.4,0.13)






349
333
317
332
337
332
334







1.00
0.93
0.94
0.96
0.93
0.94


(P90, COD)




(0.0,0.00)
(4.3,0.10)
(3.3,0.08)
(2.9,0.08)
(3.9,0.10)
(3.8,0.10)


N




1040
533
586
994
987
992








1.00
0.96
0.95
0.95
0.96








(0.0,0.00)
(3.0,0.08)
(4.1,0.12)
(3.8,0.12)
(4.0,0.11)








566
550
552
546
544









1.00
0.96
0.95
0.96









(0.0,0.00)
(3.1,0.09)
(3.1,0.10)
(3.4,0.09)









619
605
599
598










1.00
0.96
0.97
(0.0,0.00) (2.5,0.09) (2.5,0.08)
	1049	1001	1007
K	UN]	0.97
	(0.0,0.00) (1.9,0.07)
	1038	991
_L	1.00
	(0.0,0.00)
	1046
10	20
40	50	60
Distance Between Samplers (km)
Figure A-70 PM2.5 inter-sampler correlations as a function of distance between monitors for St.
Louis, MO.
July 2009
A-137
DRAFT-DO NOT CITE OR QUOTE

-------
Atlanta PM10 Monitors
A
Atlanta Interstates
Atlanta Major Highways
Atlanta
100
in Kilometers
Figure A-71. PMio monitor distribution and major highways, Atlanta, GA.
July 2009
A-138
DRAFT-DO NOT CITE OR QUOTE

-------
AGS Site ID
Site A	13-089-2001
Site B	13-097-0003
SiteC	13-121-0001
SiteD	13-121-0032
SiteE	13-121-0043
Site F	13-255-0002
50
cn
3
c 40
o
m
c 30
0)
u
c
o
u 20
10
1-winter
2 spring
3-summer*
4—fall
A
Mean 24,7
Obs 172
SD 13.0
70
60
B
21.4
178
9.3
C
23.4
171
9.5
0
26.6
174
11.8
E
25.0
995
11.5

F
21.6
178
9.7
1234 1234 1234 1234 1234 12 3 4
Figure A-72. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Atlanta, GA.
Table A-13. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Atlanta, GA.
Site
A
B
C
D
E
F
A
1.00
0.69
0.74
0.78
0.70
0.59

(0.0,0.00)
(18.0,0.22)
(15.0,0.20)
(13.0,0.20)
(16.0,0.22)
(20.0,0.24)

172
169
162
165
158
164
B

1.00
0.88
0.79
0.71
0.82


(0.0,0.00)
(6.0,0.12)
(14.5,0.17)
(16.0,0.18)
(10.0,0.14)


178
167
170
162
169
C


1.00
0.88
0.84
0.82



(0.0,0.00)
(9.0,0.13)
(10.0,0.13)
(9.0,0.15)



171
162
155
161
D



1.00
0.75
0.74

R


(0.0,0.00)
(12.0,0.15)
(15.0,0.20)

(P90, COD)


174
158
166
E
N



1.00
0.67





(0.0,0.00)
(17.0,0.19)





995
163
F 1.00
(0.0,0.00)
178
July 2009
A-139
DRAFT-DO NOT CITE OR QUOTE

-------
Distance Between Samplers (km)
90	100
Figure A-73. PM10 inter-sampler correlations as a function of distance between monitors for Atlanta,
GA.
July 2009
A-140
DRAFT-DO NOT CITE OR QUOTE

-------
QJ
r
• Birmingham PM10 Monitors
	 Birmingham Interstates
Birmingham Major Highways
Birmingham
0 10 20 40 60
80
100
~ Kilometers
Figure A-74. PMio monitor distribution and major highways, Birmingham, AL.
July 2009
A-141
DRAFT-DO NOT CITE OR QUOTE

-------
Site A
SfeB
Site C
Site 0
SiteE
SieF
SileG
SifeH
Site I
Site J
AOS Site ID
01-073-0002
01-073-0023
01 073-0034
01-073-1003
01-073-1008
01-073-1010
01-073-2003
01-073-6002
01-073-6003
01-07340(14
	1	winter
2	spring
3	-summer
4=fa!l
A
B
C
O
E
F
G
H
i
J
29 J
32.9
26.5
27.1
278
24?
20.0
28.8
38.0
48.2
180
1095
224
183
179
179
1090
181
1087
1080
126
20 0
If 4
11.8
143
11,0
14,0
12.0
29 5
30.0
f H
1234 1234 1234 1234 1234 1234 1234 123 4 1234 1234
Figure A-75. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Birmingham, AL.
Table A-14. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Birmingham, AL.

A
B
C
D
E
F
G
H
I
J
A
1.00
0.80
0.88
0.86
0.78
0.84
0.77
0.78
0.41
0.29

(0.0,0.00)
(23.0,0.16)
(11.0,0.11)
(12.0,0.13)
(12.0,0.14)
(13.0,0.13)
(15.0,0.18)
(14.0,0.15)
(41.0,0.30)
(68.0,0.34)

180
180
174
180
176
171
180
178
179
177
B

1.00
0.82
0.74
0.61
0.73
0.75
0.71
0.26
0.23


(0.0,0.00)
(23.0,0.17)
(25.0,0.21)
(26.0,0.20)
(26.0,0.19)
(25.0,0.20)
(25.0,0.22)
(51.0,0.33)
(57.0,0.36)


1095
224
183
179
179
1090
181
1087
1080
C


1.00
0.84
0.66
0.78
0.74
0.80
0.33
0.41



(0.0,0.00)
(10.0,0.12)
(15.0,0.16)
(12.0,0.14)
(14.0,0.17)
(13.0,0.15)
(43.0,0.32)
(62.0,0.34)



224
175
171
168
224
173
222
221
D



1.00
0.67
0.79
0.76
0.84
0.45
0.41




(0.0,0.00)
(15.0,0.17)
(12.0,0.15)
(14.0,0.17)
(11.0,0.12)
(42.0,0.30)
(65.5,0.34)




183
178
173
183
180
182
180
E




1.00
0.67
0.64
0.56
0.33
0.12





(0.0,0.00)
(16.0,0.15)
(18.0,0.18)
(19.0,0.20)
(45.0,0.32)
(71.0,0.39)





179
169
179
176
178
176
F





1.00
0.75
0.74
0.36
0.21






(0.0,0.00)
(14.0,0.16)
(15.0,0.17)
(43.0,0.32)
(71.0,0.38)
179	179	171	178	177
1.00	076	OJ59	0.15
(0.0,0.00)	(15.0,0.19)	(43.0,0.27)	(63.0,0.39)

1090
181
1083
1075
H

1.00
0.58
0.50


(0.0,0.00)
(38.0,0.27)
(59.0,0.31)


181
180
178
I


1.00
0.05



(0.0,0.00)
(72.0,0.40)



1087
1072
J



1.00
(0.0,0.00)
1080
July 2009
A-142
DRAFT-DO NOT CITE OR QUOTE

-------
Distance Between Samplers (km)
90	100
Figure A-76 PM10 inter-sampler correlations as a function of distance between monitors for
Birmingham, AL.
July 2009
A-143
DRAFT-DO NOT CITE OR QUOTE

-------
cj\
Chicago PM10 Monitors
Chicago Interstates
Chicago Major Highways
Chicago
0 10 20 40 60 80
100
~ Kilometers
Figure A-77. PMio monitor distribution and major highways, Chicago, IL.
July 2009
A-144
DRAFT-DO NOT CITE OR QUOTE

-------

AQS Site ID

A
B
C
D
E
F
G
H
Site A
17-031-0001
Mean
22,4
24,7
30.4
32.4
24.5
27.2
26.9
21.8
Site 8
17-031-0022
Obs
179
1077
176
1058
174
178
174
176
SiteC
17-031-0060
SD
10,9
13,1
15.1
17.3
11,9
13.4
12,3
12.3
Site D 17-031-1016 80 -
SfeE	17-031-1901
Site F	17-031-2001
Site 6	17-031-3301
SteH	17-197-1002
60 -
1 =winter
2=spring _
3=summer 								
4=fa||	1234 1234 1234 1234 1234 1234 1234 1234
AOS Site ID
Site I	18-089-0006 Mean
Site J	18-089-0022
Site K	18-089-0023
Site I	18-089-2010
SiteM 18-127-0022
Site N	18-127-0023
Site 0 18-127-0024
Obs
SD
60 -
70
S-
£
88
50
c
o
2 40
30
20
10
1-- winter
2=ipring
3=5ummer g
4- fall
I
28 2
f 82
138
J
27$
1059
16 1
K
26 4
172
95
18,6
189
91
M
172
173
112
N
24 J
1051
151
0
18 3
'89
107
I

1234 1234 1234 1234 1234 1234 1234
Figure A-78. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Chicago, IL.
July 2009
A-145
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-15. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Chicago, IL.
Site
A B
C
D
E
F
G
H
I
J
K
L
M
N
0
A
1.00 0.78
0.68
0.83
0.93
0.92
0.86
0.79
0.75
0.14
0.69
0.89
0.55
0.27
0.75

(0.0,0.00) (15.0,0.18)
(23.0,0.24)
(25.0,0.22)
(8.0,0.10)
(11.0,0.13)
(12.0,0.17)
(12.0,0.18)
(13.0,0.18)
(22.0,0.28)
(15.0,0.21)
(13.0,0.22)
(21.0,0.30)
(16.0,0.24)
(15.0,0.23)

179 176
173
174
171
173
171
167
179
173
169
166
170
171
166
B
1.00
0.66
0.74
0.76
0.84
0.79
0.74
0.68
0.36
0.73
0.81
0.66
0.33
0.77

(0.0,0.00)
(23.0,0.23)
(23.0,0.21)
(14.0,0.17)
(12.0,0.15)
(13.0,0.18)
(17.0,0.23)
(16.0,0.19)
(22.0,0.24)
(16.0,0.19)
(18.0,0.27)
(23.0,0.31)
(19.0,0.25)
(20.0,0.26)

1077
173
1040
171
173
171
173
179
1041
169
166
170
1033
166
C

1.00
0.63
0.72
0.74
0.64
0.62
0.62
0.19
0.49
0.66
0.39
0.27
0.61


(0.0,0.00)
(26.0,0.23)
(21.0,0.21)
(18.5,0.19)
(19.0,0.21)
(22.0,0.27)
(23.0,0.20)
(26.5,0.28)
(24.0,0.23)
(29.0,0.37)
(33.0,0.40)
(26.0,0.26)
(31.0,0.35)


176
171
169
170
168
164
176
170
166
163
167
168
163
D


1.00
0.79
0.85
0.79
0.74
0.70
0.23
0.69
0.82
0.61
0.29
0.76



(0.0,0.00)
(27.0,0.21)
(19.0,0.17)
(23.0,0.19)
(27.0,0.28)
(20.0,0.19)
(32.0,0.29)
(24.0,0.23)
(31.0,0.36)
(36.0,0.39)
(31.0,0.29)
(31.0,0.33)



1058
169
171
169
171
177
1022
168
166
168
1020
164
E



1.00
0.93
0.84
0.86
0.74
0.17
0.70
0.89
0.53
0.34
0.73




(0.0,0.00)
(9.0,0.10)
(13.0,0.16)
(10.0,0.16)
(13.0,0.16)
(22.0,0.26)
(15.0,0.19)
(15.0,0.25)
(22.0,0.33)
(17.0,0.22)
(18.0,0.25)




174
168
166
163
174
168
164
161
166
166
163
F




1.00
0.84
0.86
0.77
0.21
0.75
0.89
0.62
0.32
0.80





(0.0,0.00)
(12.0,0.15)
(13.0,0.19)
(12.0,0.14)
(23.0,0.25)
(16.0,0.17)
(18.0,0.28)
(25.0,0.34)
(20.0,0.23)
(20.0,0.27)





176
169
165
176
170
166
163
167
168
163
G





1.00
0.77
0.69
0.28
0.74
0.86
0.52
0.33
0.70






(0.0,0.00)
(15.0,0.22)
(14.0,0.18)
(23.0,0.26)
(14.0,0.18)
(19.0,0.31)
(24.0,0.36)
(19.0,0.24)
(22.0,0.30)






174
162
174
168
165
161
165
166
163
H






1.00
0.71
0.18
0.66
0.83
0.59
0.36
0.76







(0.0,0.00)
(16.0,0.23)
(27.0,0.30)
(18.0,0.25)
(13.0,0.23)
(19.0,0.29)
(17.0,0.25)
(14.0,0.22)







176
170
169
161
157
161
168
157
I







1.00
0.24
0.69
0.75
0.50
0.39
0.68








(0.0,0.00)
(22.0,0.24)
(12.0,0.15)
(20.0,0.32)
(26.0,0.37)
(16.0,0.21)
(21.0,0.30)








182
176
172
169
173
174
169
J

R






1.00
0.49
0.38
0.22
0.48
0.22


(P90, COD)






(0.0,0.00)
(15.0,0.20)
(25.0,0.34)
(28.0,0.36)
(22.0,0.21)
(27.0,0.33)


N






1059
166
163
168
1018
164
K









1.00
0.80
0.54
0.49
0.65










(0.0,0.00)
(17.0,0.32)
(24.0,0.35)
(14.0,0.19)
(21.0,0.31)










172
161
165
164
162
L










1.00
0.60
0.33
0.78











(0.0,0.00)
(15.0,0.26)
(19.0,0.31)
(10.0,0.20)











169
161
161
158
M











1.00
0.24
0.84












(0.0,0.00)
(21.0,0.35)
(8.0,0.16)












173
165
161
N












1.00
0.31













(0.0,0.00)
(19.0,0.29)













1051
161
0













1.00
(0.0,0.00)
169
July 2009
A-146
DRAFT-DO NOT CITE OR QUOTE

-------
0.6

~ ~ , ~~	
•	~ ~ ~ «
~t 4	*	~~~ ~	~	~	~	*
l	»A	. A*	~ ^
~ ~"
~
o
o
0.4
0.2
~
~
10	20	30	40	50	60	70	80	90	100
Distance Between Samplers (km)
Figure A-79. PM10 inter-sampler correlations as a function of distance between monitors for Chicago,
IL.
July 2009
A-147
DRAFT-DO NOT CITE OR QUOTE

-------
m
Denver PM10 Monitors
Denver Interstates
Denver Major Highways
Denver
0 10 20 40 60 80
100
~ Kilometers
Figure A-80. PMm monitor distribution and major highways, Denver, CO.
July 2009
A-148
DRAFT-DO NOT CITE OR QUOTE

-------
AQS Site ID
08-001-0006
08-001-3001
SiteC 08-013-0012
SiteD 08-031-0002
08-031-0017
08-123-0006
Site A
Site B
Site E
Site F

A
B
C
D
E
F
Mean
36.0
28.2
19.8
24.2
25.8
22.2
Obs
1043
1074
169
1039
1028
353
SD
18.3
13.2
9.7
10.6
11.5
11.2
90
80
70 -
E 60
01
n
c 50
o
2	40
4->
c
a>
c 30
o
u
20
1=winter
2=spring
3=summer
4=fa 11
10

1234 1234 1234 1234 1234 1234
Figure A-81. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Denver, CO.
Table A-16. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Denver, CO.
Site
A
B
C
D
E
F
A
1.00
0.84
0.43
0.70
0.72
0.67

(0.0,0.00)
(20.0,0.16)
(36.0,0.34)
(29.0,0.24)
(26.0,0.21)
(27.0,0.28)

1043
1022
164
987
980
339
B

1.00
0.57
0.72
0.74
0.72


(0.0,0.00)
(28.0,0.27)
(17.0,0.18)
(15.0,0.16)
(18.0,0.22)


1074
169
1019
1007
348
C


1.00
0.75
0.72
0.51



(0.0,0.00)
(17.0,0.23)
(16.0,0.23)
(16.0,0.23)



169
169
156
164
D
R


1.00
0.89
0.52

(P90, COD)


(0.0,0.00)
(9.0,0.13)
(17.0,0.22)

N


1039
976
341
E




1.00
0.58





(0.0,0.00)
(17.0,0.23)





1028
330
F 1.00
July 2009
A-149
DRAFT-DO NOT CITE OR QUOTE

-------
Site
A B
C
D
E F
(0.0,0.00)
353
~
~
10	20	30	40	50	60
Distance Between Samplers (km)
70	80	90	100
Figure A-82. PM10 inter-sampler correlations as a function of distance between monitors for Denver,
CO.
July 2009
A-150
DRAFT-DO NOT CITE OR QUOTE

-------
• Detroit PM10 Monitors
Detroit Interstates
Detroit Major Highways
Detroit
0 10 20	40	60	80	100
i—i—^^—	i Kilometers
Figure A-83. PMio monitor distribution and major highways, Detroit, Mi.
July 2009
A-151
DRAFT-DO NOT CITE OR QUOTE

-------
AGS Site ID
Site A 26-163-0001
Site B 26-163-0015
SiteC 26-163-0033
10
1=winter
2=spring
3=summer 0
4=fall
1 2 3 4 1 2 3 4 1 2 3 4
Figure A-84. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Detroit, Ml.
Table A-17. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Detroit, Ml.
Site
A
B
C
A
1.00
0.77
0.74

(0.0,0.00)
(14.0,0.18)
(28.0,0.26)

174
169
172
B

1.00
0.79


(0.0,0.00)
(21.0,0.21)


176
174
C
R

1.00

(P90, COD)
N

(0.0,0.00)
1057
July 2009
A-152
DRAFT-DO NOT CITE OR QUOTE

-------
1
0.8	#	
«
~
0.6
o
o
0.4
0.2
10	20	30	40	50	60
Distance Between Samplers (km)
70
80
90
100
Figure A-85. PM10 inter-sampler correlations as a function of distance between monitors for Detroit,
Ml.
July 2009
A-153
DRAFT-DO NOT CITE OR QUOTE

-------
•G

\Pj
q!\
r
Houston PM10 Monitors
Houston Interstates
Houston Major Highways
Houston
0 10 20 40 60 80 100
I Kilometers
Figure A-8S. PM10 monitor distribution and major highways, Houston, TX.
July 2009
A-154
DRAFT-DO NOT CITE OR QUOTE

-------
AQS Site ID
Site A	48-201-0024
SiteB	48-201-0047
Site C	48-201-0082
Site D	48-201-0068
Site E	48-201-0071
Site F	48-201-1035
Site 6	48-201-1039
Mean
Obs
SD
170
160
150
140
130
sr 120
| 110
§L 100
c
o
+-»
<0
*¦*
c
HI
u
c
o
u
1=winter
2=spring
3-summer
4--fall
90
80
70
80
SO
40
30
20
10
0
A
B
C
D
E
F
G
24,5
24.0
23.4
22.3
22,7
54.8
15,8
174
178
174
175
174
359
163
9.7
9.1
9.7
10.0
9.3
35.5
7.7

1234 1234 1234 1234 1234 1234 1234
Figure A-87. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Houston, TX.
Table A-18. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Houston, TX.
SITE
A
B
C
D
E
F
G
A
1.00
0.84
0.78
0.76
0.43
0.56
0.75

(0.0,0.00)
(9.0,0.12)
(11.0,0.16)
(12.0,0.16)
(15.0,0.20)
(77.0,0.37)
(17.0,0.28)

174
163
158
165
167
159
156
B

1.00
0.86
0.86
0.38
0.52
0.79


(0.0,0.00)
(9.0,0.11)
(9.0,0.12)
(15.0,0.19)
(74.0,0.39)
(16.0,0.26)


178
156
160
163
158
152
C


1.00
0.83
0.41
0.38
0.85



(0.0,0.00)
(10.0,0.14)
(17.0,0.19)
(74.0,0.40)
(14.5,0.25)



174
156
159
151
150
D



1.00
0.32
0.43
0.76




(0.0,0.00)
(18.0,0.20)
(81.0,0.43)
(16.0,0.23)




175
163
155
154
E




1.00
0.15
0.38





(0.0,0.00)
(78.0,0.43)
(20.0,0.28)

R



174
158
157
F
(P90, COD)




1.00
0.37

N




(0.0,0.00)
(92.0,0.54)






359
149
G






1.00
(0.0,0.00)
163
July 2009
A-155
DRAFT-DO NOT CITE OR QUOTE

-------
60 -
50 -
Figure A-88. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for New
York City, NY.
Table A-19. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
New York City, NY.
Site A
B
C
A 1.00
0.88
0.82
(0.0,0.00)
(11.0,0.20)
(12.0,0.16)
167
156
164
B
1.00
0.74

(0.0,0.00)
(18.0,0.25)
R
169
166
C (P90, COD)

1.00
N

(0.0,0.00)
178
July 2009
A-156
DRAFT-DO NOT CITE OR QUOTE

-------
1
0.8
0.6
o
o
0.4
0.2
0 -I	1	1	1	1	1	1	1	1	1	
0	10	20	30	40	50	60	70	80	90	100
Distance Between Samplers (km)
Figure A-89. PM10 inter-sampler correlations as a function of distance between monitors for New
York City, NY.
July 2009
A-157
DRAFT-DO NOT CITE OR QUOTE

-------
q!\
• Philadelphia PM10 Monitors
	 Philadelphia Interstates
Philadelphia Major Highways
Philadelphia
0 10 20 40 60
80
100
~ Kilometers
Figure A-90. PM10 monitor distribution and major highways, Philadelphia, PA.
July 2009
A-158
DRAFT-DO NOT CITE OR QUOTE

-------
AQS Site ID
Site A 10-003-2004	Mean
Site B 42-017-0012	Obs
SiteC 42-045-0002	SD
Site D 42-091-0013	60 •
A
B
G
D
22.8
17.1
19.9
17.6
1059
1040
1059
1049
117
9.3
9.4
9.8
50 -
£ 40
§L
c
o
«-<
to
c
(U
u
c
o
u
30
20
10
1-winter
2=spring
3=summer
4=fail
1234 1234 1 2 3 4 1 2 3 4
Figure A-91. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Philadelphia, PA.
Table A-20. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Philadelphia, PA.
Site
A
B
C
D
A
1.00
0.81
0.64
0.84

(0.0,0.00)
(13.0,0.21)
(14.0,0.19)
(12.0,0.20)

1059
1005
1025
1013
B

1.00
0.71
0.93


(0.0,0.00)
(11.0,0.20)
(6.0,0.12)


1040
1006
994
C


1.00
0.73



(0.0,0.00)
(11.0,0.19)

R

1059
1014
D
(P90, COD)


1.00

N


(0.0,0.00)
1049
July 2009
A-159
DRAFT-DO NOT CITE OR QUOTE

-------
1
~
0.6
o
o
0.4
0.2
0 	,	,	,	,	,	,	,	,	,	
0	10	20	30	40	50	60	70	80	90	100
Distance Between Samplers (km)
Figure A-92. PM10 inter-sampler correlations as a function of distance between monitors for
Philadelphia, PA.
July 2009
A-160
DRAFT-DO NOT CITE OR QUOTE

-------
q!\
• Phoenix PM10 Monitors
	 Phoenix interstates
Phoenix Major Highways
Phoenix
0 10 20 40 60 80 100
~ Kilometers
Figure A-93. PMio monitor distribution and major highways, Phoenix, AZ.
July 2009
A-161
DRAFT-DO NOT CITE OR QUOTE

-------
AGS Site ID
Site A	04-013-0019
SiteB	04-013-1003
SiteC	04-013-1004
SiteD	04-013-3002
Site E	04-013-3003
Site F	04-013-3010
Site 6	04-013-4003
SiteB	04-013-4004
sr-
E
1
c
o

A
B
C
D
E
F
G
H
Mean
486
309
32 8
408
32 5
515
568
34 7
Obs
790
179
182
1084
182
780
336
181
SO
23.0
14-5
146
20 0
15.2
23.1
25.8
17.0
240
220
200
180
180
140
120
I 100
y
e 80
u
60 H
40 -
1 -winter
2=$pring
3=summer
4=fall
AQS Site ID
Site I 04-013-4008
Site J 04-013-4009
SiteK 04-013-4010
Site L 04-013-4011
SiteM 04-013-8008
SiteN 04-013-3812
SiteO 04-013-999?
SiteP 04-021-0001
sr
£
20 -
260
240
220
200
180
3 160
c ho
O
2 120
§ 100
u
O 80
u
60
40
1=wir»ter
2—spring
3=symmer
4=fal
20
12 3 4
12 3 4
12 3 4
12 3 4
12 3 4
12 3 4
12 3 4
12 3 4
l
J
K
L
M
N
o
P
556
75 6
32 5
53 0
58.4
65.5
343
49 7
1073
1083
178
1O30
1?4
1080
toe?
4or
30 9
3S5
16 1
27®
WS
34 9
213
54 2
ill
i|
1234 12 3 4 1234 1234 1234 1234 1234 1234
July 2009
A-162
DRAFT-DO NOT CITE OR QUOTE

-------
Site Q	04-021-300
SiteR	04-021-300
SiteS	04-021-300
Site T	04-021-300/
Site U	04-021-3008
SiteV	04-021-3011
SiteW	04-021-3012
SiteX	04-021-7004
Obs
SD
260
240
220
?00
180
160
140
120
u 100
60
40
1=wir>ter 20
2-spring
3=summer o
4=fall
G
R
S
T
U
V
W
X
20.8
38.8
117
27.0
80.8
747
21.5
54.3
172
171
172
175
476
475
170
322
10.0
21.4
8.8
17.1
72.5
137.5
12.5
36.3
1234 1234 1234 1234 1234 1234 1234 1234
Figure A-94. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Phoenix, AZ.
Table A-21. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Phoenix, AZ.
Site
A B
C
D
E
F
G
H
I
J
K
L
M
A
1.00 0.71
0.85
0.85
0.67
0.94
0.86
0.77
0.73
0.83
0.77
0.70
0.87

(0.0,0.00) (38.0,0.25)
(33.0,0.21)
(21.0,0.12)
(38.0,0.23)
(14.0,0.09)
(22.0,0.13)
(34.0,0.21)
(35.0,0.18)
(59.0,0.24)
(34.0,0.24)
(30.0,0.17)
(28.5,0.16)

790 178
181
788
181
779
335
180
772
781
177
789
170
B
1.00
0.84
0.82
0.85
0.67
0.74
0.81
0.67
0.68
0.75
0.60
0.63

(0.0,0.00)
(13.0,0.12)
(23.0,0.19)
(11.0,0.11)
(37.0,0.29)
(47.0,0.30)
(13.0,0.13)
(49.0,0.30)
(84.0,0.43)
(16.0,0.15)
(51.0,0.31)
(56.0,0.32)

179
179
177
179
175
179
178
175
176
175
178
164
C

1.00
0.88
0.81
0.78
0.80
0.81
0.70
0.73
0.81
0.63
0.75


(0.0,0.00)
(20.0,0.16)
(12.0,0.11)
(38.0,0.27)
(44.0,0.28)
(13.0,0.13)
(48.0,0.29)
(84.0,0.41)
(15.0,0.14)
(49.0,0.29)
(55.0, 0.30)


182
180
182
178
182
181
178
179
178
181
167
D


1.00
0.76
0.88
0.81
0.82
0.76
0.78
0.79
0.65
0.83



(0.0,0.00)
(23.0,0.17)
(22.0,0.14)
(29.0,0.16)
(18.0,0.17)
(39.0,0.20)
(71.0,0.31)
(22.0,0.19)
(35.0,0.20)
(42.0,0.21)



1084
180
778
334
179
1062
1072
176
1080
172
E



1.00
0.64
0.68
0.74
0.66
0.59
0.67
0.51
0.61




(0.0,0.00)
(40.0,0.27)
(47.0,0.29)
(16.0,0.14)
(48.0,0.29)
(88.0,0.42)
(15.0,0.15)
(49.0,0.30)
(58.0,0.31)




182
178
182
181
178
179
178
181
167
F




1.00
0.83
0.76
0.75
0.86
0.74
0.69
0.87





(0.0,0.00)
(22.0,0.13)
(36.0,0.25)
(32.0,0.17)
(54.0,0.21)
(41.0,0.28)
(30.0,0.17)
(25.0,0.15)





780
331
177
762
772
175
779
167
G





1.00
0.77
0.65
0.78
0.71
0.65
0.80






(0.0,0.00)
(44.0,0.26)
(38.0,0.19)
(48.0,0.19)
(46.0,0.30)
(36.0,0.19)
(33.0,0.16)






336
181
326
333
178
335
169
H






1.00
0.79
0.81
0.79
0.69
0.72







(0.0,0.00)
(47.0,0.26)
(79.0,0.39)
(16.0,0.14)
(43.0,0.27)
(53.0,0.29)
181	177	178	177	180	167
1.00	079	076	0£9	0.68
(0.0,0.00) (52.0,0.22) (48.0,0.29) (33.0,0.17) (38.0,0.20)

1073
1061
174
1068
171
J

1.00
0.78
0.73
0.80


(0.0,0.00)
(83.0,0.42)
(57.0,0.23)
(51.0,0.22)


1083
175
1078
171
K


1.00
0.72
0.68



(0.0,0.00)
(45.0,0.29)
(56.0,0.32)



178
177
164
L



1.00
0.63




(0.0,0.00)
(42.0,0.20)




1090
173
M




1.00
(0.0,0.00)
174
July 2009
A-163
DRAFT-DO NOT CITE OR QUOTE

-------
Site A	B	C	D	E	F	G	H	I	J	K	L	M
Table A-40, continued

N
0
P
Q
R
S
T
U
V
W
X
A
0.87
0.68
0.47
0.53
0.68
0.40
0.69
0.50
0.27
0.56
0.65

(39.0,0.18)
(28.0,0.17)
(29.0,0.19)
(49.0,0.42)
(34.0,0.27)
(64.0,0.57)
(40.0,0.34)
(82.0,0.31)
(49.0,0.27)
(48.0,0.43)
(31.0,0.20)

784
783
406
171
171
171
174
475
474
169
262
B
0.59
0.75
0.75
0.73
0.63
0.55
0.59
0.53
0.66
0.65
0.64

(67.0,0.37)
(15.0,0.15)
(22.0,0.17)
(23.0,0.27)
(30.0,0.25)
(32.0,0.43)
(21.0,0.24)
(94.0,0.41)
(62.0,0.34)
(24.0,0.30)
(46.0,0.29)

178
179
175
169
168
169
172
172
177
167
155
C
0.70
0.87
0.80
0.70
0.71
0.48
0.64
0.56
0.71
0.62
0.60

(69.0,0.35)
(11.0,0.12)
(19.0,0.15)
(24.0,0.28)
(26.0,0.24)
(36.0,0.44)
(22.0,0.24)
(91.0,0.40)
(59.0,0.32)
(28.0,0.31)
(43.0,0.28)

181
182
178
172
171
172
175
175
180
170
157
D
0.78
0.86
0.73
0.63
0.68
0.49
0.65
0.66
0.45
0.58
0.70

(57.0, 0.25)
(15.0,0.12)
(30.0,0.19)
(38.0,0.38)
(27.0,0.25)
(46.0,0.53)
(31.0,0.31)
(87.0,0.34)
(59.0,0.30)
(38.0,0.39)
(32.0,0.21)

1075
1056
405
170
169
170
173
474
473
168
318
E
0.60
0.73
0.68
0.72
0.64
0.43
0.48
0.42
0.69
0.51
0.52

(67.0,0.35)
(14.0,0.14)
(21.0,0.17)
(21.0,0.28)
(27.0,0.24)
(33.0,0.44)
(21.0,0.25)
(93.0,0.41)
(63.0,0.32)
(25.0,0.32)
(46.0,0.28)

181
182
178
172
171
172
175
175
180
170
157
F
0.91
0.68
0.46
0.48
0.63
0.38
0.63
0.47
0.28
0.42
0.66

(35.0,0.14)
(31.0,0.21)
(30.0,0.22)
(60.0,0.46)
(37.0,0.30)
(68.0,0.60)
(45.0,0.39)
(80.0,0.31)
(50.0,0.27)
(57.0,0.47)
(34.0,0.22)

774
773
403
169
167
168
172
470
469
166
259
G
0.77
0.57
0.47
0.55
0.65
0.46
0.62
0.49
0.44
0.57
0.64

(35.0,0.16)
(41.0,0.25)
(36.5,0.24)
(61.0,0.47)
(41.0,0.30)
(73.0,0.61)
(58.0,0.41)
(78.0,0.28)
(45.0,0.24)
(59.0,0.48)
(32.0,0.22)

332
336
330
172
171
172
175
329
334
170
185
H
0.70
0.75
0.82
0.63
0.74
0.55
0.62
0.60
0.76
0.64
0.76

(66.0,0.33)
(15.0,0.14)
(18.0,0.15)
(29.0,0.31)
(24.5,0.22)
(37.0,0.46)
(24.0,0.25)
(84.0,0.38)
(58.0,0.29)
(30.0,0.33)
(39.0,0.25)

180
181
177
171
170
171
174
174
179
169
156
1
0.76
0.61
0.52
0.57
0.71
0.51
0.58
0.59
0.37
0.51
0.80

(42.0,0.18)
(49.0,0.27)
(39.0,0.22)
(66.0,0.47)
(41.0,0.27)
(77.0,0.60)
(60.0,0.40)
(72.0,0.27)
(46.0,0.23)
(63.0,0.47)
(30.0,0.16)

1064
1045
397
169
168
168
171
461
461
167
314
J
0.91
0.58
0.41
0.48
0.65
0.48
0.65
0.51
0.28
0.46
0.74

(29.0,0.12)
(83.0,0.38)
(68.0,0.31)
(103.0,0.58)
(75.0,0.40)
(115.0,0.69)
(92.0,0.51)
(69.0,0.26)
(59.0,0.27)
(101.0,0.58)
(62.0,0.27)

1074
1055
404
169
168
169
172
473
472
167
319
K
0.69
0.71
0.75
0.52
0.64
0.52
0.62
0.71
0.68
0.55
0.68

(73.0,0.36)
(16.0,0.16)
(19.0,0.18)
(28.0,0.29)
(27.0,0.23)
(34.0,0.44)
(22.0,0.24)
(89.0,0.40)
(59.0,0.33)
(28.0,0.32)
(44.0,0.29)

177
178
174
168
167
168
171
171
176
166
153
L
0.68
0.55
0.51
0.47
0.57
0.48
0.49
0.59
0.33
0.50
0.68

(48.0,0.20)
(44.0,0.26)
(37.0,0.22)
(66.0,0.47)
(44.5,0.29)
(71.0,0.60)
(62.0,0.40)
(75.0,0.27)
(53.0,0.24)
(67.0,0.48)
(29.0,0.18)

1081
1063
406
171
170
171
174
475
474
169
321
M
0.86
0.81
0.75
0.48
0.64
0.37
0.62
0.46
0.65
0.44
0.59

(32.0,0.16)
(53.0,0.29)
(47.0,0.30)
(74.0,0.48)
(51.0,0.32)
(80.0,0.61)
(58.5,0.41)
(62.0,0.31)
(48.0,0.26)
(68.0,0.49)
(42.0,0.24)

173
174
165
157
158
158
160
165
168
156
145
N
1.00
0.58
0.41
0.48
0.67
0.42
0.63
0.42
0.26
0.40
0.60

(0.0,0.00)
(66.0,0.32)
(51.0,0.27)
(88.0,0.53)
(62.5, 0.35)
(98.0,0.65)
(75.0,0.46)
(71.0,0.29)
(55.0, 0.27)
(88.0,0.54)
(48.0,0.24)

1086
1059
403
171
170
171
174
470
469
169
319
0

1.00
0.90
0.61
0.64
0.39
0.60
0.72
0.59
0.55
0.64


(0.0,0.00)
(35.0,0.22)
(28.0,0.31)
(25.0,0.24)
(38.0,0.47)
(22.0,0.26)
(94.0,0.39)
(69.0,0.35)
(29.0,0.33)
(44.0,0.26)


1067
407
172
171
172
175
475
473
170
317
P


1.00
0.67
0.81
0.58
0.78
0.82
0.64
0.71
0.67



(0.0,0.00)
(32.0,0.29)
(22.0,0.19)
(44.0,0.45)
(21.0,0.21)
(80.0,0.30)
(52.0,0.23)
(32.0,0.31)
(39.0,0.24)



407
169
170
169
172
400
404
167
197
Q



1.00
0.72
0.65
0.57
0.36
0.58
0.68
0.47




(0.0,0.00)
(40.0,0.33)
(15.0,0.28)
(23.0,0.24)
(104.0,0.53)
(78.0,0.46)
(15.0,0.22)
(62.0,0.43)




172
162
163
167
165
171
161
148
R




1.00
0.66
0.68
0.53
0.82
0.68
0.68





(0.0,0.00)
(55.0,0.48)
(32.0,0.27)
(75.0, 0.35)
(47.0,0.25)
(40.0,0.34)
(39.0,0.24)





171
162
165
164
171
160
148
S





1.00
0.60
0.46
0.59
0.72
0.52






(0.0,0.00)
(28.0,0.35)
(115.0,0.65)
(86.0,0.59)
(19.0,0.28)
(74.0,0.58)






172
167
165
171
162
149
T






1.00
0.56
0.66
0.68
0.61







(0.0,0.00)
(94.0,0.47)
(71.0,0.39)
(18.0,0.24)
(51.5,0.37)







175
169
174
165
150
U







1.00
0.54
0.52
0.71








(0.0,0.00)
(66.0,0.24)
(101.0,0.53)
(61.0,0.25)








476
464
165
204
V








1.00
0.60
0.64









(0.0,0.00)
(78.0,0.47)
(35.0,0.20)









475
169
206
W









1.00
0.56










(0.0,0.00)
(63.0,0.44)










170
145











1.00











(0.0,0.00)











322
July 2009
A-164
DRAFT-DO NOT CITE OR QUOTE

-------
1
~~ )
~ ~~
~
>%
~ *
~
*
~	- 47 ~ 4 ~ ~	~ ~	"
~*	~•~•~~•	4 ~	A ~~	4
4 4^	*4^44	~	A
14	* #	~	~	>4	444
~	W0# 4	~ ~ ~ 4	~
i ~ ~~~!~ ~ ~ •~ ~ ~ ~ ~ ~~< * * ~
~	~ ~ ~*~~~ ~ ~ ~ % ~ ~~•~% » ~~
40	50	60
Distance Between Samplers (km)
100
Figure A-95. PM10 inter-sampler correlations as a function of distance between monitors for Phoenix,
AZ.
July 2009
A-165
DRAFT-DO NOT CITE OR QUOTE

-------
rJ\
• Riverside PM10 Monitors
	 Riverside Interstates
Riverside Major Highways
Riverside
0 1020 40 60 80 100
^¦=1 Kilometers
Figure A-96. PM10 monitor distribution and major highways. Riverside, CA.
July 2009
A-166
DRAFT-DO NOT CITE OR QUOTE

-------
AGS Site ID
Site A	08-065-0003
SiteB	08-065-2002
SiteC	08-065-5001
SiteD	08-085-6001
SiteE	08-065-8001
Site F	08-071-0013
SiteG	08-071-0025
E
A
Mean 37 6
Obs 174
SO 27 9
1 JO
120
110
100
. 90
SO
70
60
50
40
30
20
1=winter
2=sprlng 10
3=$ummer
4=fal!	U
0
C
D
E
F
G
52 3
280
536
551
24.8
435
315
170
173
358
177
131
2? 4
21 1
319
35 4
20 9
24.8

AGS Site ID
Site H	06-071-0306
Site I	08-071-1234
Site J	08-071-2002
SieK	06-071-4001
SfteL	06-0714003
SiteM	08-071-9004
20 6
1015
154
Ol
ZL
c
o
c
©
o
c
o
o
l=winter
2~spring
3-summer g
4=faII
1234 1234 1234
H	I
Mean 314
Obs 1060
SD 20.2
130
120
110
100
90
80
70
80
SO
40
30
20
10
J
54.4
178
29.5
12 3 4 12 3
K
23,8
173
136
12 3 4
L
36.5
178
20.2
2 3 4
M
477
175
24 3
I
1234 1234 1234 1234 1234 1234
Figure A-97. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Riverside, CA.
July 2009
A-167
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-22. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Riverside, CA.

A
B
C
D
E
F
G
H
I
J
K
L
M
A
1.00
0.09
0.15
0.90
0.94
0.25
0.94
0.24
0.12
0.83
0.27
0.46
0.78

(0.0,0.00)
(50.0,0.31)
(36.0,0.32)
(33.0,0.19)
(37.0,0.24)
(41.0,0.38)
(16.0,0.13)
(25.0,0.22)
(40.0,0.39)
(38.5,0.24)
(30.0,0.23)
(32.0,0.25)
(33.0,0.21)

174
170
155
165
172
169
171
174
173
160
158
169
164
B

1.00
0.86
0.07
0.13
0.31
0.12
0.32
0.29
0.13
0.31
0.35
0.29


(0.0,0.00)
(48.0,0.37)
(47.0,0.28)
(45.0,0.27)
(57.0,0.47)
(49.0,0.26)
(48.0,0.33)
(55.0,0.49)
(51.0,0.25)
(49.0,0.35)
(51.0,0.31)
(44.0,0.24)


315
161
167
298
173
176
309
302
172
163
173
168
C


1.00
0.13
0.21
0.36
0.20
0.34
0.36
0.23
0.38
0.50
0.40



(0.0,0.00)
(49.0,0.37)
(58.0,0.42)
(24.0,0.31)
(40.0,0.35)
(27.0,0.28)
(24.0,0.30)
(57.5,0.41)
(24.0,0.27)
(30.0,0.25)
(41.0,0.34)



170
151
162
156
160
170
168
150
147
159
154
D



1.00
0.93
0.19
0.83
0.11
0.05
0.73
0.13
0.38
0.69




(0.0,0.00)
(29.0,0.17)
(52.0,0.43)
(23.0,0.17)
(38.0,0.27)
(52.0,0.46)
(26.0,0.18)
(43.0,0.30)
(40.0,0.26)
(24.5,0.16)




173
169
167
168
173
172
157
155
165
160
E




1.00
0.23
0.93
0.26
0.16
0.86
0.27
0.57
0.82





(0.0,0.00)
(63.0,0.48)
(27.0,0.17)
(46.0,0.33)
(63.5,0.51)
(18.0,0.13)
(54.0,0.36)
(40.0,0.28)
(26.0,0.15)





358
174
179
351
340
175
165
175
171
F





1.00
0.27
0.73
0.32
0.35
0.43
0.44
0.48






(0.0,0.00)
(44.0,0.41)
(28.0,0.33)
(27.0,0.32)
(57.0,0.46)
(24.5,0.32)
(35.0,0.35)
(46.0,0.43)






177
173
177
176
162
160
170
164
G






1.00
0.27
0.20
0.90
0.35
0.58
0.85







(0.0,0.00)
(30.0,0.25)
(46.5,0.45)
(25.0,0.16)
(34.0,0.27)
(29.0,0.24)
(24.0,0.15)







181
181
180
165
163
174
168
H







1.00
0.26
0.47
0.48
0.40
0.44








(0.0,0.00)
(27.0,0.33)
(45.0,0.32)
(18.0,0.18)
(29.0,0.25)
(34.0,0.26)








1060
983
178
172
178
175
I








1.00
0.20
0.45
0.38
0.35









(0.0,0.00)
(62.0,0.51)
(25.0,0.32)
(41.0,0.39)
(48.0,0.46)









1015
177
172
177
173
J

R







1.00
0.42
0.70
0.85


(P90, COD)







(0.0,0.00)
(49.0,0.35)
(37.0,0.27)
(20.0,0.15)


N







178
155
163
157
K










1.00
0.49
0.48











(0.0,0.00)
(30.0,0.26)
(38.0,0.29)











173
162
157
L











1.00
0.84












(0.0,0.00)
(24.0,0.20)












178
167
M












1.00
(0.0,0.00)
175
July 2009
A-168
DRAFT-DO NOT CITE OR QUOTE

-------
40	50	60
Distance Between Samplers (km)
100
Figure A-98. PM10 inter-sampler correlations as a function of distance between monitors for
Riverside, CA.
July 2009
A-169
DRAFT-DO NOT CITE OR QUOTE

-------
q!\
r
• Seattle PM10 Monitors
	 Seattle Interstates
Seattle Major Highways
Seattle
0 10 20 40 60 80
100
~ Kilometers
Figure A-99. PMio monitor distribution and major highways, Seattle, WA.
July 2009
A-170
DRAFT-DO NOT CITE OR QUOTE

-------

AQS Site ID

A
B
Site A
53-033-0057
Mean
21.9
15.7
Site B
53-033-2004
Obs
1059
1077


SD
9.9
8.6


60 -


50 -
c

-------
1
0.8
0.6
o
o
0.4
0.2
10	20	30
40	50	60
Distance Between Samplers (km)
70	80	90	100
Figure A-101. PM10 inter-sampler correlations as a function of distance between monitors for Seattle,
WA.
July 2009
A-172
DRAFT-DO NOT CITE OR QUOTE

-------
0 10 20 40 60 80 100
I Kilometers
PM10 Monitors
Interstates
Major Highways
St Louis
St Louis
St Louis
St Louis
Figure A-102. PMio monitor distribution and major highways, St. Louis, MO.
July 2009
A-173
DRAFT-DO NOT CITE OR QUOTE

-------

AQS Site ID

A
B
C
O
E
F
G
H
I
Site A
17-117-0002
Mean
22,8
35.4
28,3
34.0
20.5
28.2
22,7
29.5
40.3
Site B
17-119-0010
Obs
171
173
174
178
185
176
179
180
1050
SrteC
17-119-3007
SD
9,4
15.5
12,2
14.7
14,4
11,7
10,4
12.8
28.5
Site D 17-183-0010 130 "
Site E: 29-189-5001 120 _
Site F 29-510-0085
Site G 29-510-0088 110 -
SiteH 29-510-0087
Site I 29-510-0088
ST 90 -
E
"oi 80 -
3
c 70 -
o
<5 60 -
1 - winter
2=spring
3-summer
1234 1234 1234 1234 1234 1234 1234 1234 1234
Figure A-103. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for St.
Louis, M0.
Table A-24. Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for St.
Louis, M0.

A
B
C
D
E
F
G
H
I
A
1.00
0.50
0.75
0.67
0.47
0.65
0.67
0.73
0.55

(0.0,0.00)
(30.0,0.28)
(14.0,0.17)
(23.0,0.24)
(16.0,0.29)
(16.0,0.18)
(13.0,0.17)
(18.0,0.19)
(52.0,0.33)

171
161
158
156
158
163
166
168
164
B

1.00
0.65
0.63
0.46
0.68
0.68
0.64
0.52


(0.0,0.00)
(20.0,0.21)
(20.0,0.19)
(37.0,0.42)
(23.0,0.20)
(28.0,0.28)
(22.0,0.20)
(36.0,0.28)


173
161
158
160
167
169
170
166
C


1.00
0.75
0.57
0.80
0.76
0.82
0.65



(0.0,0.00)
(17.0,0.17)
(23.0,0.33)
(12.0,0.13)
(13.0,0.18)
(12.0,0.13)
(41.0,0.27)



174
157
158
165
169
169
168
D



1.00
0.44
0.82
0.81
0.80
0.59




(0.0,0.00)
(30.0,0.40)
(16.0,0.15)
(21.0,0.24)
(14.0,0.15)
(36.0,0.27)




176
157
163
165
166
169
E




1.00
0.53
0.62
0.56
0.34





(0.0,0.00)
(22.0,0.34)
(17.0,0.26)
(25.0, 0.35)
(55.0,0.42)





185
164
166
167
179
F





1.00
0.89
0.86
0.67






(0.0,0.00)
(11.0,0.16)
(12.0,0.11)
(41.0,0.27)


R



176
173
174
169
G

(P90, COD)




1.00
0.83
0.65


N




(0.0,0.00)
(16.0,0.19)
(47.0,0.32)







179
177
173
H







1.00
0.64








(0.0,0.00)
(41.0,0.27)
180 173
I 1.00
(0.0,0.00)
1050
July 2009
A-174
DRAFT-DO NOT CITE OR QUOTE

-------
1
0.2
0 	,	,	,	,	,	,	,	,	,	
0	10	20	30	40	50	60	70	80	90	100
Distance Between Samplers (km)
Figure A-104. PM10 inter-sampler correlations as a function of distance between monitors for St.
Louis, M0.
Table A-25. Correlation coefficients of hourly and daily average particle number, surface and volume
concentrations in selected particle size ranges.
Size range (nm)


Hourly averages


Oaily avg
All days IN = 5481)
Sundays IN = 701)
Weekdays IN = 3227)
Event days (N = 577)
No events (N = 4904)
All days IN = 263)
3-10
0.40
0.24
0.42
0.73
0.37
0.32
10-30
0.35
0.22
0.31
0.57
0.33
0.27
30-50
0.38
0.42
0.29
0.56
0.36
0.36
50-100
0.46
0.56
0.39
0.57
0.45
0.46
100-500
0.55
0.65
0.49
0.62
0.55
0.55
500-800
0.73
0.75
0.70
0.76
0.72
0.71
10-100
0.31
0.28
0.24
0.52
0.29
0.24
10-800
0.55
0.65
0.49
0.62
0.55
0.55
Total number
0.30
0.24
0.24
0.58
0.28
0.20
Total surface
0.57
0.63
0.51
0.65
0.56
0.57
Total volume
0.66
0.69
0.62
0.73
0.65
0.67
Source: Tuch et al. (2006)
July 2009
A-175
DRAFT-DO NOT CITE OR QUOTE

-------
A.2.3.
Speciation
Atlanta, GA	FRM PMJ S speciation • 4-Season Avg
Atlanta, GA
Atlanta, GA
I	1
Sulfate
OCM
Atlanta, GA	FRKPMzbapeciaion • T_fali	Atlanta, GA	FRK PM1S spectator) - winter

Figure A-105. Seasonally averaged PMLs speciation data for 2005-2007 for a) annual, bj winter,
c) spring, d) summer and e) fall derived using the SANDWICH method in Atlanta, GA.
July 2009
A-176
DRAFT-DO NOT CITE OR QUOTE

-------
Birmingham, AL
FRM PM2.5specia6on • 4-SeasoriAvg
i i Sulfate	Nitrate
I I OCM	Crustal
SANDWICH sulfate and ivtrate include ammonium and water
Birmingham, AL
FRM PM2.5 speciality Spring
Birmingham, AL
FRM PM 2 5 special ion Summer
Birmingham, AL frmpm 2s special-wwer
Sulfate
Figure A-106. Seasonally averaged PM.-* speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Birmingham, AL
FRM PM2.5 speciation • T_Fall
Birmingham, AL
July 2009
A-177
DRAFT-DO NOT CITE OR QUOTE

-------
Boston, MA/NH

Boston, MA/NH
Boston, MA/NH
Boston, MA/NH
Boston, MA/NH
Boston, MA/NH
Figure A-107. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c}
spring, d) summer and e) fall derived using the SANDWICH method in Boston, MA.
July 2009
A-178
DRAFT-DO NOT CITE OR QUOTE

-------
Chicago, IL/IN
frm PM2.5speclallon - 4-SeasonAvg
we
•Jitrate	EC

I	1 OCM M
SANDWICH sulfate and nitrate include ammonium and water
-jo-'
Chicago, IL/IN
FRM PM2.5 speciation - Spring
Chicago, IL/IN
FRM PM 2.5 speciation - Summer
J it ra e ec
SANDWICH sulfate and nitrate include ammonium and water
-rustal
SANDWICH sulfate and titrate include ammonium and water
Chicago, IL/IN
FRM PM2.5 spectatlon - T_Fali
Chicago, IL/IN	FRM PM 2 5 speciation - Winter
] OCM	Cru!
?

18	_ 10
SAICWICH sulfate and nirate include amrromum and water
1=3 Sulfate ¦
I	1 OCM	Crustal
SANDWICH sulfate and nitrate include amrronium and water
Figure A-108. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Chicago, IL,
July 2009
A-179
DRAFT-DO NOT CITE OR QUOTE

-------
Denver, C^^	FRMPM2.5speeiation • 4-SeasonAvg
Denver, GO	FRMPM2.5spectation- spring	DenVeT, GO	FRMPMZSspecaten-Surmw'r
Sulfate
Nitrat
Crustal
i i Sulfate
Nitrate
Denver, CO	FRMPM2.5speciation- T_Fail	DenVeT, CO	FRM PM25ap«iaBon • Winter
Sulfate
OCM
Figure A-109. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, ci
spring, d) summer and e) fall derived using the SANDWICH method in Denver, CO.
July 2009
A-180
DRAFT-DO NOT CITE OR QUOTE

-------
Detroit, Ml
FRM PM2.5 speciatlon - 4-SeasonAvg
SANDWICH sulfate and nitrate include ammonium and water
Detroit, Ml
FRM PM2.5 speciation - Spring
i i Sulfate	Nitrate
I I OCM	Crust;
SANDWICH sulfate and nitrate in
Detroit, Ml
i i Sulfate Nitrate
I	1 OCM	Crust; I
SANDWICH sulfate and nitrate hclutie ammonium
I I OCM	Crustal
SANDWICH sulfate and nitrate include ammonium and water
Detroit, Ml
FRM PM2.5 speciatlon - T_Fall
Detroit, Ml
FRM PM2.5 speciatson - Winter
i i Sulfate	Nitrate
I	1 OCM	Crustal
SANDWICH sutfate and nitrate ttclude ammonium and water
Figure A-110. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Detroit, ML
July 2009
A-181
DRAFT-DO NOT CITE OR QUOTE

-------
Houston, TX
SANDWICH sulfate and nitrate include ammonium and water
Houston, TX
rustal
Houston, TX
Houston, TX
Houston, TX
FRM PM2.5speclation • Winter
Sulfate
Sulfate
OCM
SANDWICH sulfate and nilrate iixlude ammonium and waiei
SANDWICH sulfate and nitrate include
Figure A-111. Seasonally averaged PM2.5 speciation data for 2005-2007 for a) annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Houston, TX.
July 2009
A-182
DRAFT-DO NOT CITE OR QUOTE

-------
LOS Angeles, OA FRMPM2,5speciafion- 4-8easonAv9
Los Angeles, OA frmpm2.5 specie*)- spring
LOS AngeleS, OA	FRMPM2.5;{>eciation- Sumner
Los Angeles, OA
Sulfate
Nitrate
OCM
Figure A-112. Seasonally averaged PM2.5 speciation data for 2005-2007 for a} annual, b) winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Los Angeles, CA.
Los Angeles, CA FRM PM 2.5 speciation - Whter
i 1 Sulfate Nitrate ¦¦ EC
1	1 OCM	Crustal
SANDWICH sulfate and nitrate include amrnorium
July 2009
A-183
DRAFT-DO NOT CITE OR QUOTE

-------
New York, NY/NJ/OT- PM2 5
-------
Philadelphia, PA/N3,MPM25speciauon-4SeasonAV9
rsiiiui nibble"
I I OCM	Crustal
SANDWICH sulfate and nitrate hclude
Philadelphia, PA/NJ
Philadelphia, PA/NJ
I	1 OCM	Crustal
SANDWICH stftate and nitrate irclude ammonium and water
3"'julldlu	I III jLl
I	1 OCM HM Crustal
SANDWICH sulfate ane rttrate Include armronium and voter
Philadelphia, PA/NJ
FRM PM2.5 specialist* • T.Fall
Sulfate
' ' OCM	Crustal
SANDWICH sdfate and nitrate include ammonium and w»er
Philadelphia, PA/NJ
' ' OCM	Crustal
SANDWICH sulfate and ntrate Include
Figure A-114. Seasonally averaged pm** speciation data for 2005-2007 for a) annual, bj winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Philadelphia.
July 2009
A-185
DRAFT-DO NOT CITE OR QUOTE

-------
Phoenix, AZ
FRM pM2,5speciallon - ^-Season Avg
Ni-mte
Crustal
SANDWICH sulfate and nitrate Include ammonium and water
Phoenix, AZ
FRM PM2.5 speciatlon - Spring
Phoenix, AZ
FRM PM 2.5 speclation - Summer

iulfate ¦¦¦ Nitrate MM EC
3 CM MM Crystal
SANDWICH sulfate and lit rate Inslude ammonium aid water
^ Q
~ a rate
Nitrate
~ OCM

81 NDVrtCH sulfate and nitrate nciude
Phoenix, AZ
FRM PM2.5 speciatlon - T_Ftll
Fhoenix, AZ
FRM PM 2.5 speclation - wntei
1 Jr.alP
Crustal
SANDWICH sulfate and nlrate include ammonium and water
IZZD S> Ifate t
ocm i
S# NDWICH sulfate and nitrate include ammonium s
I Crystal
Q
Figure A-115. Seasonally averaged pm». speciation data for 2005-2007 for a) annual, bj winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Phoenix, AZ.
July 2009
A-186
DRAFT-DO NOT CITE OR QUOTE

-------
Pittsburgh PA FRMPM2.5speetation- 4-SeasonAvg
Crustal
SANDWICH sulfate
Pittsburgh, PA
Pittsburgh, PA
FRM PM 2.5 special ion - Summer
I I OCM	Crustal
SANDWICH sulfate and nBrate include ammonium and water
I	1 OCM	Crustal
SANDWICH sulfate and nutate nciude ammortum and valet
Pittsburgh, PA
FRM PM2.5 specialltn - T_Fall
Pittsburgh, PA
FRM PM25 speciatlon - V
Sulfate
Crustal
Crustal
SANDWICH sulfate and nitrate
Figure A-116. Seasonally averaged pm** speciation data for 2005-2007 for a) annual, bj winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Pittsburgh, PA.
July 2009	A-187	DRAFT - DO NOT CITE OR QUOTE

-------
Riverside, OA FRMPM2.5spedation- 4-SeasonAvg
3 Sulfate hh Nitrate '*¦ EC
I I OCM	Crustal
SANDWICH sulfate and nitrate include ammonium and water
Riverside. CA
FRM PM2.5 speciatlon - Spring
Riverside, CA
FRM PM 2.5 spec lation - Summer
i ) Sulfate	Nitrate "¦¦¦ EC
i	1 OCM	Crustal
SANDWCH sulfate and nirate include
S'lNr- L	3 flri.rP MMI EC
i i OCM	Crustal
SANDWICH sulfate and nitrate iiclude arrmorium and water
Riverside. CA
FRM PM2.5 speciatlon - T_Fall
Riverside, CA
FRM PM2.5 stieclation - Winter
i i Sulfate
i	1 OCM	Crustal
SANDWICH sulfate and nirate include
Nitrate -E<
i i Sulfate Nitrate
' 1 OCM	Crustal
SANDWICH sulfate and nitrate incline
Figure A-117. Seasonally averaged pm». speciation data for 2005-2007 for a) annual, bj winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Riverside, OA.
July 2009
A-188
DRAFT-DO NOT CITE OR QUOTE

-------
Seattle, WA	FRMPM25 3ieciatldn . s-irinf;
Seattle, WA
Seattle, WA
FRM PM2.6 speciation - T_Fall
Sulfate
Sulfate
Crustal
Crustai
SANDWICH sufate and i
SANDWICH sulfate ana rflrate
Figure A-118. Seasonally averaged pm** speciation data for 2005-2007 for a) annual, bj winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in Seattle, WA.
July 2009
A-189
DRAFT-DO NOT CITE OR QUOTE

-------
St. Louis, MO/IL FRM PM2.Sspeclation - 4-SeasonAvg
1=1 Sulfate
Crus al
St. Louis, MO/IL
Crus al
St. Louis, MO/IL
Nitra-c
St. Louis, MO/IL
J?-*
St. Louis, MO/IL
FRM PM2.5 speclatlon - T_Fall
Sulfate
Mitra'e
Figure A-119. Seasonally averaged pm** speciation data for 2005-2007 for a) annual, bj winter, c)
spring, d) summer and e) fall derived using the SANDWICH method in St. Louis, MO,
July 2009
A-190
DRAFT-DO NOT CITE OR QUOTE

-------
1 2 3 4 5 6 7 8 9 10 11 12
Figure A-120. Seasonal patterns in pm2.5 chemical composition from city-wide monthly average values
for Atlanta, GA, 2005-2007. The gray line represents the difference in OCM calculated
using material balance and blank corrected OC x 1.4.
Figure A-121. Seasonal patterns in pm2.5 chemical composition from city-wide monthly average values
for Birmingham, AL, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.
July 2009
A-191
~RAFT - DO NOT CITE OR QUOTE

-------
30-
25-
c
c
o
Q.
E
p
"D
Q-
0.0
-5-
4
Chem I I Sulfate mass	Nitrate mass	EC
Figure A-122. Seasonal patterns in pm2.5 chemical composition from city-wide monthly average values
for Boston, MA, 2005-2007. The gray line represents the difference in OCM calculated
using material balance and blank corrected OC x 1.4.
30 -
25 -
jr
i 5.o-BPbi-
i
. •
1 2 3 4 5 6 7 8 9 10 11 12
Figure A-123. Seasonal patterns in pm2.5 chemical composition from city-wide monthly average values
for Chicago, IL, 2005-2007. The gray line represents the difference in OCM calculated
using material balance and blank corrected OC x 1.4.
July 2009
A-192
DRAFT-DO NOT CITE OR QUOTE

-------
2 3
_J Sulfate m
9 10 11 12
Figure A-124. Seasonal patterns in pm2.5 chemical composition from city-wide monthly average values
for Denver, CO, 2005-2007. The gray line represents the difference in OCM calculated
using material balance and blank corrected OC x 1.4.
£ 15-
a	r*-l
/V
HI
1 2 3 4 5 6 7 8 9 10 11 12
Figure A-125. Seasonal patterns in pm2.5 chemical composition from city-wide monthly average values
for Detroit, Ml, 2005-2007. The gray line represents the difference in OCM calculated
using material balance and blank corrected OC x 1.4.
July 2009
A-193
DRAFT-DO NOT CITE OR QUOTE

-------
9 10 11 12
Figure A-126. Seasonal patterns in pm2.5 chemical composition from city-wide monthly average values
for Houston, TX, 2005-2007. The gray line represents the difference in OCM calculated
using material balance and blank corrected OC x 1.4.
9 0	2
Figure A-127. Seasonal patterns in pm2.5 chemical composition from city-wide monthly average values
for Los Angeles, CA, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.
July 2009
A-194
DRAFT-DO NOT CITE OR QUOTE

-------
30
£ 15

2 3
_J Sulfate m
9 10 11 12
Figure A-128. Seasonal patterns in PM2.5 chemical composition from city-wide monthly average values
for New York City, NY, 2005-2007. The gray line represents the difference in 0CM
calculated using material balance and blank corrected 0C x 1.4.
£ 15
ill:
9 10 11 12
Figure A-129. Seasonal patterns in PM2.5 chemical composition from city-wide monthly average values
for Philadelphia, PA, 2005-2007. The gray line represents the difference in 0CM
calculated using material balance and blank corrected 0C x 1.4.
July 2009
A-195
DRAFT-DO NOT CITE OR QUOTE

-------
30
JflJTftt
1 2 3 4 5 6 7
Chem I I Sulfate mass	I \
9 10 11 12
Figure A-130. Seasonal patterns in pm2.5 chemical composition from city-wide monthly average values
for Phoenix, AZ, 2005-2007. The gray line represents the difference in 0CM calculated
using material balance and blank corrected 0C x 1.4.
1 8 9 10 11 12
Figure A-131. Seasonal patterns in pm2.5 chemical composition from city-wide monthly average values
for Pittsburgh, PA, 2005-2007. The gray line represents the difference in 0CM
calculated using material balance and blank corrected 0C x 1.4.
July 2009
A-196
DRAFT-DO NOT CITE OR QUOTE

-------

Figure A-132. Seasonal patterns in pm2.5 chemical composition from city-wide monthly average values
for Riverside, CA, 2005-2007. The gray line represents the difference in OCM calculated
using material balance and blank corrected OC x 1.4.
25-
20-
15-
10-
5.0-
0.0-
1 2 3 4 5 6 7 8 9 10 11 12
Chem	Sulfate mass	Nitrate mass	EC
Figure A-133. Seasonal patterns in pm2.5 chemical composition from city-wide monthly average values
for Seattle, WA, 2005-2007. The gray line represents the difference in OCM calculated
using material balance and blank corrected OC x 1.4.
July 2009
A-197
DRAFT-DO NOT CITE OR QUOTE

-------
30--T
£ 15-


1 2 3 4 5 6 7
Chem I I Sulfate mass	I \
9 10 11 12
Figure A-134. Seasonal patterns in pm2.5 chemical composition from city-wide monthly average values
for St. Louis, MO, 2005-2007. The gray line represents the difference in OCM calculated
using material balance and blank corrected OC x 1.4.
A.2.4. Diel Trends
to
c\i
77 i
58 -
38 -
19
1
Weekday (N = 5156)
T~
6
1 I 1
12
1$T
24
77 -j
58 -
38 -
19 -
1
Weekend (N = 2086)
12
18
24
	Median
	Mean
	90'th & 10 th
95'th & 5'th
Figure A-135. Diel plots generated from all available hourly FRM-like pm2.5 data, stratified by weekday
(left) and weekend (right), in Atlanta, GA. Included are the number of monitor days (N)
and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
July 2009
A-198
DRAFT-DO NOT CITE OR QUOTE

-------
Weekday(N =1975)
in
CN
77
58
38
19 -
1
12
18
24
Weekend (N = 797)
77 -j
58 -
38
19 -j
1
12
18
24
• Median
¦ Mean
-- 90th & 10'th
	 95th & 5th
Figure A-136. Diel plots generated from all available hourly FRM-like pm2.5 data, stratified by weekday
(left) and weekend (right), in Chicago, IL. Included are the number of monitor days (N) and
the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
IT)
csi
11 -j
58 -
38 -
19 -
Weekday(N = 7032)
—T—
6
12
18
24
Weekend (N = 2854)
77 -i
58 -
38 -
19 -

1
~~r^
6
12
18
24
• Median
	Mean
	 90th & 10th
	95th & 5th
Figure A-137. Diel plots generated from all available hourly FRM-like pm2.5 data, stratified by weekday
(left) and weekend (right), in Houston, TX. Included are the number of monitor days (N)
and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
July 2009
A-199
DRAFT-DO NOT CITE OR QUOTE

-------
m
CN
Weekday(N = 3348)
77
58
38
19
—
6
12
18
24
77 -j
58 -
38 -
19 -
1
Weekend (N = 1364)
~ I " '
12
18
24
¦ Median
Mean
	90'th & 10'th
	95'th & 5'th
Figure A-138. Diel plots generated from all available hourly FRM-like pm2.5 data, stratified by weekday
(left) and weekend (right), in New York City, NY. Included are the number of monitor
days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
in
CN
77
58 H
38
19
1
Weekday(N =981)
12
—i—
18
24
Weekend (N =407)
77 -j
58 -
38 -
19 -
1
12
18
24
¦	Median
¦	Mean
	90'th & 10'th
95'th & 5'th
Figure A-139. Diel plots generated from all available hourly FRM-like pm2.5 data, stratified by weekday
(left) and weekend (right), in Pittsburgh, PA. Included are the number of monitor days (N)
and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
July 2009
A-200
DRAFT-DO NOT CITE OR QUOTE

-------
m
csi
77 i
58 ¦
38 -
19 -
Weekday(N = 5775)
—i—
6
—i—
12
—i—
18
—i
24
77 -j
58 -
38 -
19 -
Weekend (N = 2332)
1
—i—
6
—i—
12
—i—
18
24
	Median
	Mean
	90'th & 10'th
	95'th & 5th
Figure A-140. Diel plots generated from all available hourly FRM-like pm2.5 data, stratified by weekday
(left) and weekend (right), in Seattle, WA. Included are the number of monitor days (N)
and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
Lf)
cn
77 n
58 -
38 -
19 -
Weekday (N = 1727)
12
18
24
Weekend (N = 692)
77 n
58 -
38 -
19 -

12
18
24
¦ Median
	Mean
	90'th & 10th
	95'th & 5'th
Figure A-141. Diel plots generated from all available hourly FRM-like PM2.5 data, stratified by weekday
(left) and weekend (right), in St. Louis, MO. Included are the number of monitor days (N)
and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
July 2009
A-201
DRAFT-DO NOT CITE OR QUOTE

-------
Weekday (N = 715)
Weekend (N = 293)
291 i
218
145 ¦
73 ¦
12 18 24
291 i
218
145 -
73 -
0 -P
¦	Median
¦	Mean
90th & 10'th
95th & 5th
1 6 12 18 24
Figure A-142. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Atlanta, GA. Included are the number of monitor days (N)
and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
Weekday(N =1971)
291 n
218
145 -
73 -
» — •_ — *
Weekend (N = 793)
1
12 18
24
291
218 -j
145
73 -
0 -P
1
12 18
24
¦	Median
¦	Mean
	90'th & 10'th
95'th & 5th
Figure A-143. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Chicago, IL. Included are the number of monitor days (N) and
the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
July 2009
A-202
DRAFT-DO NOT CITE OR QUOTE

-------
291 n
218
145 -
73 -
Weekday (N = 1310)
1
12 18 24
Weekend (N = 529)
291 i
218 -
145 -
73 -
1
' I '
6
	Median
	Mean
	90'th & 10'th
	95'th & 5'th
12 18
24
Figure A-144. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Denver, CO. Included are the number of monitor days (N)
and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
291 i
218 -
145 -
73 -
Weekday (N = 756)
Weekend (N = 302)
291 n
218
145 -
73

1
12 18 24
1 6 12 18 24
¦ Median
	Mean
	90'th & 10'th
	95'th & 5'th
Figure A-145. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Detroit, Ml. Included are the number of monitor days (N) and
the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
July 2009
A-203
DRAFT-DO NOT CITE OR QUOTE

-------
Weekday(N =692)
Weekend (N = 277)
291
218
145
291
218
145 1
73 -
0
1 6
12
18
24
¦ Median
	Mean
	90'th & 10'th
95'th & 5th
Figure A-146. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Los Angeles, CA. Included are the number of monitor days
(N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
291 i
218 -
145
73 -
Weekday (N = 742)
o r--r
1 6
12
18
24
291 i
218 -
145
73
Weekend (N = 303)
1
12
18
24
¦ Median
	Mean
	90'th & 10'th
95'th & 5th
Figure A-147. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Philadelphia, PA. Included are the number of monitor days
(N) and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
July 2009
A-204
DRAFT-DO NOT CITE OR QUOTE

-------
Weekday(N =1532)
Weekend (N = 618)
291
218 -
145 -
291
218 ¦
145 -
73 -
1
—I—
6
12
TfT
"* I
24
¦ Median
	Mean
	90th & 10'th
	95'th & 5th
Figure A-148. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Phoenix, AZ. Included are the number of monitor days (N)
and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
291
218 -
145 -
73 ¦
Weekday(N = 6385)
— _ — — *
1
12

—i
24
Weekend (N = 2600)
291 -j
218 -
145
73
1
6
' I 1
12
18
—1
24
	Median
	Mean
	90 th & 10 th
	95'th & 5th
Figure A-149. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Pittsburgh, PA. Included are the number of monitor days (N)
and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
July 2009
A-205
DRAFT-DO NOT CITE OR QUOTE

-------
Weekday (N = 1660)
291
218
145
73
Weekend (N = 673)
291 n
218 ¦
145 ¦
73 -
1
12 18
24
0 P"
	Median
	Mean
	90'th & 10 th
	95'th & 5th
1
12 18 24
Figure A-150. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Riverside, CA. Included are the number of monitor days (N)
and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
291
218
145 ¦
73
0
Weekday (N = 752)
1
12 18 24
291 i
218 -
145 -
73
Weekend (N = 309)
o r.T?r
1 6
12 18 24
¦ Median
	Mean
	90'th & 10'th
	95'th & 5'th
Figure A-151. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Seattle, WA. Included are the number of monitor days (N)
and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
July 2009
A-206
DRAFT-DO NOT CITE OR QUOTE

-------
Weekday(N =1741)
291
218 -
145 -j
73
0
N — — — ,
1
1 I '
6
¦I	I 1
12 18
24
Weekend(N = 706)
291
218 -I
145
73 +
¦ Median
	Mean
	90'th & 10'th
	95'th & 5'th
o r i'
1 6
~X2 18 24
Figure A-152. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in St. Louis, MO. Included are the number of monitor days (N)
and the median, mean, 5th, 10th, 90th and 95th percentiles for each hour.
A.2.5. Copollutant Measurements
Winter
Spring
PM-; j • (daw jivgi
SO; {daly arg)
NO; {daly
O j (daman 8-r*>
' »» ^ * * * + 4 * .
Summer
tr & tp 4> ^	if 
CO (daly wg)
O) (defy max B-hr)
Fall
PM* (daly avg)

* 4
PM* (daly avg)

*
, (daly avg)


PM-*,» (da* avfl)
4

SO; (daly »g)


SO; (daly avg)


NO, (da*y wg)

m
NO; (daly wg)

4 4
CO (daly avg)
4
4
00 (daly wg)

4 4
0] (daly max 8-hr)

49
Oj (daly max 8-hr)

4 * 4
i (correlrton coofcnnO
Figure A-153. Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PMio 2.5, SO2, I\I0 2 and CO
and daily maximum 8-h avg O3 for Atlanta, GA, stratified by season (2005-2007). One
point is included for each available monitor pair.
July 2009
A-207
DRAFT-DO NOT CITE OR QUOTE

-------
Winter
*>»-- **>
SO; «a*y «<$
ms
r!l ill » 1	
> ! > 1 < 1 t
e # #
& >f '? »*
Spring
SO; tm was
NO,
CO

Figure A-154. Correlations between 24-h PM2.5 and co-located 24-h avg PMio, PMio2.5, SO2, NO2 and CO
and daily maximum 8-h avg O3 for Birmingham, AL, stratified by season (2005-2007).
One point is included for each available monitor pair.
Writer	Spring
PU
I 4 $


* M*


PM" . . 1*5a% »BS

m
80, i«Wy
* * • w *


wmm * * »
m i<«r
1 att dsMtt


m *
«5 4^w#
1 * m
co >

** * *
O- «*#, f*»
* *
0. >«B* W» IMv>

m *m *
'
¦' 	 .¦ ¦ '¦¦¦ ¦¦¦' /-
'

„ + ,> * * ,> * * -

Summer

Fall
PM.,
1 ****
pw* im »*$.'



T *
pvwtmtH#

*

~ *| m * *
S& »»

> W ~ *

I « ** *
W, I.Mf »0

* 4i
m '.0* m-
* it *
co fsMy »$<

* «• m
0) >.
i>" i," (.'k K* <¦" fr" !>' '
k
Figure A-155. Correlations between 24-h PWh* and co-located 24-h avg PMio, PM102.5, SO2, I\I02 and CO
and daily maximum 8-h avg O3 for Boston, MA, stratified by season (2005-2007). One
point is included for each available monitor pair.
July 2009
A-208
DRAFT-DO NOT CITE OR QUOTE

-------

r*sf iui1? «»»)
Pits-: i Ntiir »m)
SBJ toil; >TI)
m Nsilj S?|J
U«!lj S.Jf
n«:s? *f *-»' »»§>
¦ >* f? »' f? »fc »? •& |»V ft* »S H?" I? •»* V •
r (correlation coeffSciant)
•	j? „*¦ ^	*> # «?	^ ^ ^ ^ ^
C cwt® I at i on coefficient)
summer
fall
PKSS f< a i If »•»}
Pm»-1 5 (iS • I 5 J Jtg)
sdi { If
£tt Null; jrji
Sj»»« iin i ly #t* l-»r svg>
? V ? >v ? V o" .V V ,» «* ¦
r (correlation coefficient)
,» «« ^ >*
5."* J.' J? S.V	1? »v «V i* A* •.» V	,» .?
r (corre1at i oo coefficient)
Figure A-156. Correlations between 24-h PWh* and co-located 24-h avg PM10, PM102.5, SO2, l\IO 2 and CO
and daily maximum 8-h avg O3 for Chicago, IL, stratified by season (2005-2007). One
point is included for each available monitor pair.
Writer
Spring
Su-nmer
Fall
-S? «»• ^ *» 5? ^
Figure A-157. Correlations between 24-h pm2.5 and co-located 24-h avg PMio, PM102.5, SO2, l\IO 2 and CO
and daily maximum 8-h avg O3 for Denver, CO, stratified by season (2005-2007). One
point is included for each available monitor pair.
July 2009
A-209
DRAFT-DO NOT CITE OR QUOTE

-------
Writer
Spring
Summer
«
Fail

Figure A-158. Correlations between 24-h PWh* and co-located 24-h avg PM10, PM102.5, SO2, l\IO 2 and CO
and daily maximum 8-h avg O3 for Houston, TX, stratified by season (2005-2007). One
point is included for each available monitor pair.
Winter
Spring
SO. !(** m®
-»	»""i	i" 'i" li	t
J
Summer
Fall
(do*

* *m *
PU,t tiesiPf m.yj

# »1 * **



m,-., „ i d&>>


so? am

**
SO;. (csa%- wtfr

* *
m (m <*¦&

("MM# *
»: m* m

****
CO (iMf


GO i


cm®nnfffmis-m


Q-A
-------
Winter
Spring
t.«My







*xj>


SO; f «*|f *><»

*
1
1
a
ser ym*>»

«nr *m*n
a*** ms

mem
id* 


00
*
t m
CO «** wsi 1
*•
G, (Jot, inrn. »-m

—
G.Wttfc «**»»*•) J
1#* »* t * *«

*v # -s* ,
Figure A-160. Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM102.5, SO2, NO2 and CO
and daily maximum 8-h avg O3 for Philadelphia, PA, stratified by season (2005-2007).
One point is included for each available monitor pair.
Winte?
Spring

to* t
-------
Winter
Spring
t.cwy
*m,„ im
SO; f «*|f *><»
mt a*** ms
Qo vm
V ' I	?	S	8	1 "1	8	t	*	
mm, skm
,imt m». s i*)
r«j id* w®
»» ,• #
Wh <«*»
Figure A-162. Correlations between 24-h PM2.5 and co-located 24-h avg PMio, PM102.5, SO2, NO2 and CO
and daily maximum 8-h avg O3 for Pittsburgh, PA, stratified by season (2005-2007). One
point is included for each available monitor pair.
Winter	Spring


* ~ **


m nm*
MK„Ma*»e





SO,****#

: « *


t >* m


* * *


* ** *
co i ti&Qf 1 I
mm*
>> **
Summer	Fall

f >1 m
>a*

*****
PM.c .

< i«My 4»85


SO, EtW
* * #
m>.~ mi

4
no, m. < t
* * *
no, <**«$

* * «
CO (*», * ,<
• * *
CO 50W* w®


£>•
1 * m**
O. ™ *•*>
* *
* >1
.»* ,* ,j» ,p „<>	,*	- „»	<* «'¦
Figure A-163. Correlations between 24-h pm2.5 and co-located 24-h avg PM10, PM102.5, SO2, l\IO 2 and CO
and daily maximum 8-h avg O3 for Riverside, CA, stratified by season (2005-2007). One
point is included for each available monitor pair.
July 2009
A-212
DRAFT-DO NOT CITE OR QUOTE

-------
Winter
(•¥, CO* j
».« , (-itffy #*JJ J
SO, <** w$ I
	r"'"r	
¦¦¦ .v	/. ¦¦ .	..	* # ,y **
Summer
* \ * *
no,*****!
SO «My w#|
o, ema ma* mn
so, km**#.
NO. «**»*.
CO iCWy m®.
Spring
* *
*1* *
"'r"r"r","r'T"T,"i	r,'*-"T"-"r"'"?	rr^T-r-r-r-^
Faff
* *
* * «<

!	i*
Figure A-164. Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PMio2.5, SO2, NO2 and CO
and daily maximum 8-h avg O3 for St. Louis, MO, stratified by season (2005-2007). One
point is included for each available monitor pair.
Winter
Spring
Summer
Fall
a'- V-P	5? 6* ^	»V ,J	^
s (corcslsCon c-aetkfinSj
-o*	# -,P # tP *fc ^	
-------
Winter
Spring
CO {Ciaiy avgi
Summer
NO: (daily avg).
c? ^	e"' v> «"•' c* ^
! (£OKfaatQ«coefte«(iti
•s? o*	«¦> • c®
«ice" niters e:
v,fe	,»•	^	^	*•< #
Figure A-167. Correlations between 24-h PM10 and co-located 24-h avg pm2.5, PM102.5 SO2, l\IO2 and CO
and daily maximum 8-h avg O3 for Boston, MA, stratified by season (2005-2007). One
point is included for each available monitor pair.
July 2009
A-214
DRAFT-DO NOT CITE OR QUOTE

-------
Winter
Spring
PMss (Oaty avg)
PM.b1a (daly avg)
SO? (daily avg)
NO? Maty wg)
CO (ctely 8wg)
Os (daty ma* 8-hr)
Summer
PMi- idssy ava)
(Saty avg)
SO- (daily avg)

CO (daily avs)
Oj (daiy max 6-hr)
PM;. (dafy aval
PMm.= , {daly avg).
SOv (daly avg).
NO; (dafy svg£
CO (daily avgj
Os (cJaity max fi-in i
Figure A-168. Correlations between 24-h PM10 and co-located 24-h avg pm2.5, PMio2.5 SO2, l\IO 2 and CO
and daily maximum 8-h avg O3 for Chicago, IL, stratified by season (2005-2007). One
point is included for each available monitor pair.
Winter
Spring
PM- > idafy-
SO.. iOety swgt
NQi ¦fSfttf?
co i&*t
O. {dnV fnm 8 (rV
PM, •. (ete%r ijagS,
PM- ; . , (W®.
SO; <««* ®*B>.
f©! «•#,
CO tcwy avgi'.
C-, {<*#t ">»


*¥ **
(iMff avffi.
CO
itwy maw
Summer
;dMy iwal
*KS; ?Ka«y ami
4 m
1,0
O, (
-------
Spring
Summer
rtai% awg)
dai^ fwg}
tiaiy aval
daily avg)
daily avg)
PM? 3 (daiy avg!
* Nk
HMio.-s {ttetty swffl
SO; ftKSty avg)

CO idaky avg}
Figure A-170. Correlations between 24-h PM10 and co-located 24-h avg PM^, PM102.5 SO2, NO2 and CO
and daily maximum 8-h avg O3 for Detroit, Ml, stratified by season (2005-2007). One
point is included for each available monitor pair.
Spring
(itoly avg}
PM«,;s fdsily nvg)
SO2 (daily sa-g)
NO? (daily
CO fOaly avg)
O.. (daily ma* frhr)
Q, («>•
SO;
ty augJ
CO tasty avg)
Oj (ctefy ma* 8-lhr)
daty a
Summer
Fall
(dsiy avg) -

*
PM.W. (risiy avg).


SO2 {Bafy avg),
*
*
NO, {etaiy avg).


CO (daily avg).
I**
m *
Q> ftSaiy ma* 8-he) -
*
m
—1—1—1—r~
~i—1—!—1—1—1—r~
Figure A-171. Correlations between 24-h PM10 and co-located 24-h avg pm2.5, PMio2.5 SO2, l\IO 2 and CO
and daily maximum 8-h avg O3 for Houston, TX, stratified by season (2005-2007). One
point is included for each available monitor pair.
July 2009
A-216
DRAFT-DO NOT CITE OR QUOTE

-------
Winter
Spring
Summer
Fall
pm;5
da«y avg}
PM.ms
daty awgj
SO,
cMy avg)
fCh.
eft®? avg)
CO
daily avg)
0. (daiy
max 3-itt)
** * * **
# * * H »W *
PM,< (daily avg) .
PMic^ , (il3*y avg) „
SO-; (Safty avg) .
NO- (daiy avg) ,
CO (daily avg) .
Oj (daily max 8-3ir) •

s*»f #	>~ »* *
H*** H #	>~
t j*e *
r (ceiTe-tsBM coeSc»«(i
—i—i—i—i—i—i—s—i—s—I—?—i—i—i—r—r=
r teorrelslon coe*c«*nn
Figure A-172. Correlations between 24-h PM10 and co-located 24-h avg pm2.5, PMio2.5 SO2, l\IO 2 and CO
and daily maximum 8-h avg O3 for Los Angeles, CA, stratified by season (2005-2007).
One point is included for each available monitor pair.
Winter
PM„
daily avg)

daily avgj
so?
daily swal
NO,-
dafy avg}
co
daty avg)
0; (daty
IT® 8-hr)
Spring
-1—1—1—1—1—1—i—r
PMi, (daty avg
PM.J.H (daily av£
SO; (daily avg
NO? (daily avg
CO {dsSy avg
0;, mat 8-h?
"I	1—I	1—I	1—I	1—I	1
* >* *
* H
1	1—I	1—I	1—I	1	1—I	1—I	1—I	1—I	i	1—i~
Summer
Fail
PMjt,
daily avg)
PM.S...3
daily avgj}
SO:
dafiyavgf
NOi
daHy avgj
CO
dssly avg)
Oj. (daily
ma* 0-lif)
PM;;, (daily avg
PM,j. (daily avg
* *
t>0; i/iaify avg.
NO; (daily SW£
Q (ssty sag
Oi Ma>ly mas 8-hr
f (cotrelason cm*.-»r>r)
1 (r.orrefaton coe1feient>
Figure A-173. Correlations between 24-h PM10 and co-located 24-h avg pm2.5, PMio2.5 SO2, I\I0 2 and CO
and daily maximum 8-h avg O3 for New York City, NY, stratified by season (2005-2007).
One point is included for each available monitor pair.
July 2009
A-217
DRAFT-DO NOT CITE OR QUOTE

-------
Winter
Spring
NO? (daily avcj)
Summer
Fall
Figure A-174. Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PMio2.5 SO2, NO2 and CO
and daily maximum 8-h avg O3 for Philadelphia, PA, stratified by season (2005-2007).
One point is included for each available monitor pair.
Winter
Spring
(daily *vg)'
pm,mi (aafy wg) •
SO: (doty svgs •
NO. fdaty avg) ¦
CO itfstty  (dB«y »
SOi (daSy avg) ¦
UQ, awgi.
CO {daily sifgj -
On (tteJy max S-r*) •
* * *
* * #
#* * * »~ 4* ~*
*»* ** * *** *
Figure A-175. Correlations between 24-h PM10 and co-located 24-h avg pm2.5, PMio2.5 SO2, l\IO2 and CO
and daily maximum 8-h avg O3 for Phoenix, AZ, stratified by season (2005-2007). One
point is included for each available monitor pair.
July 2009
A-218
DRAFT-DO NOT CITE OR QUOTE

-------
Winter
Spring
PM?. ftfesty avg)
PM.'.;;. (daly awg!
SO; {daily swg)
NO? (daily avg}
CO (dsay avg>
Oj (Oai!y max Q+tr>
' 1 I I ! J I ! i	! I I I I I I F
1—I—I—!—I—I—r-
Summer
Fall
PM, s (daily w
PM )5, [, (daily avg) .
SO? W»V swa) •
NO? (dafy avg) ,
CO ( ¦
* ** >1
PM;. I'OMy avg? .
PMi»f j (dedy ava1 .
SOj (d»ty a«g) .
MO; (dials*®
GO (daify avg),
Oj (daily ma* 8-hr).
H * ** **	H
* *t *
** 3* sfc *
9* ## *
i—r—i—i—i—i—I—i
Figure A-176. Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PMio2.5 SO2, l\IO 2 and CO
and daily maximum 8-h avg O3 for Pittsburgh, PA, stratified by season (2005-2007). One
point is included for each available monitor pair.
PM:s ftfefy avg)
PM..;,;V (daily avg)
SO, (dsty avg)
MO-j (daty avai
CO (dally avg)
0? (daily i3!«* 8-lirS
PM.'i (dSlty =3Vg)
PM1 >tm **
Summer
Fall
»* ** »t si#
(6 »We *
>i M H, *
¥¥ ** **>* *
PMj.. (daily cwg) -
PMsw, (dafty avg).
SO; fcJaSy avgf.
NO; (efesfy avg).
CO (daily w
O, (fla% mas 8-!»>.
+H	>~ * *
S * #	5#e
~»* * *
=*	»* *
**	*
Figure A-177. Correlations between 24-h PM10 and co-located 24-h avg pm2.5, PMio2.5 SO2, l\IO 2 and CO
and daily maximum 8-h avg O3 for Riverside, CA, stratified by season (2005-2007). One
point is included for each available monitor pair.
July 2009
A-219
DRAFT-DO NOT CITE OR QUOTE

-------
Winter
Spring
PM-v. {daSy avgi-j
(dafyava)-|
SO; (claiy *
(daty avgj4
CO (daisy svgH
Or. i'cls«y max ft-m-j
~1—J—1—!—f—r
h *t	m
-1—\—i—e—r~
# * **
* H
** * * *
Summer
Fall
SO; (dsiy a
NO;. (daily a*g> i
CO (iiaSy avgsi
O. idaly max 8-hf) -j
~ * **
PE(fl.,s Waits' augi
PM»?> (da% avgl
80^ (da% awft!
NO, (dally a
CO {(SBily a
Gs (daily max S-lir)
* *
**
Figure A-178. Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PMio2.5 SO2, l\IO 2 and CO
and daily maximum 8-h avg O3 for St. Louis, MO, stratified by season (2005-2007). One
point is included for each available monitor pair.
July 2009
A-220
DRAFT-DO NOT CITE OR QUOTE

-------
A.3. Source Apportionment
A.3.1. Type of Receptor Models
Table A-46. Different receptor models used in the Supersite source apportionment studies: chemical
mass balance.
Receptor Model
Description
Strengths and Weaknesses
Effective Variance CMB 42,121
(Note that all models based on eq 1 or 2
are CMB equations. The term CMB used
here rejects the historical solution in
which source proxies are explicitly used
as model input and a single sample
effective variance solution is reported.)
CMB software is currently distributed by
EPA. The most recent version is the CMB
8.2, which is run in the Microsoft
Windows system.
Principle
Ambient chemical concentrations are expressed as the sum of products of species
abundances in source emissions and source contributions (eq 1 or 2). These equations are
solved for the source contribution estimates when ambient concentrations and source
pronles are input. The single-sample effective variance least squares122 is the most
commonly used solution method because it incorporates uncertainties of ambient
concentrations and source proxies in the estimate of source contributions and their
uncertainties. This reduced to the tracer solution when it is assumed that there is one
unique species for each source. Choices of source proxies should avoid collinearity, which
occurs when chemical compositions of various source emissions are not sufficiently
different.121
Data Needs
CMB requires source proxies, which are the mass fractions of particulate or gas species in
source emissions. The species and particle size fraction measured in source emissions
should match those in ambient samples to be apportioned. Several sampling and analysis
methods provide time-integrated speciation of PM2.6 and volatile organic compounds (VOCs)
for CMB. Source proxies are preferably obtained in the same geographical region as the
ambient samples, although using source proxies from different regions is commonly
practiced in the literature. The practitioner needs to decide the source proxies and species
being included in the model, on the basis of the conceptual model and model performance
measures.
Output
Effective variance CMB determines, if converged, source contributions to each sample in
terms of PM or VOC mass. CMB also generates various model performance measures,
including correlation R2, deviation n2, residue/ uncertainty ratio, and MPIN matrix that are
useful for renning the model inputs to obtain the best and most meaningful source
apportionment resolution.
Strengths
Software available providing a
good user interface.
Provides quantitative uncertainties
on source contribution estimates
based on input concentrations,
measurement uncertainties, and
collinearity of source pronles.
Quantities contributions from
source types with single particle
and organic compound
measurements.
Weaknesses
Completely compatible source and
receptor measurements are not
commonly available.
Assumes all observed mass is due
to the sources selected in advance,
which involves some subjectivity.
Chemically similar sources may
result in collinearity without more
specinc chemical markers.
Typically does not apportion
secondary particle constituents to
sources. Must be combined with
pronle aging model to estimate
secondary PM.
2 Hidy and Friedlander (1972,156546)
21 Watson et al. (1997,157121) l22(1984,045693)
Source: Watson et al. (2008,157128)
July 2009
A-221
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-47. Different receptor models used in the Supersites source apportionment studies: factor
analysis.
Receptor Model
Description
Strengths and Weaknesses
PMF123,124
PMFx (PMF2 and PMF3) software
is available from Dr. Pentti Paatero
at the University of Helsinki,
Finland. This software is a
Microsoft DOS application. EPA
distributes EPA PMF76 version 1.1
as a Microsoft Windows
application with better user
interface.
Principle	Strengths
PMFx contains PMF2 and PMF3. PMF2 solves the CMB equations (i.e., eqs 2 and 3) using an Software available,
iterative minimization algorithm. Source proxies F, and contribution Sit are solved simultaneously.
The non-negativity constraint is implemented in the algorithm to decrease the number of possible
solutions (local minimums) in the PMF analyses, because both source pronle and contribution should
not contain negative values. There is rotational ambiguity in all two-way factor analyses (i.e., F; and
Sit matrices may be rotated and still nt the data). PMF2 allows using the FPEAK parameter to
control the rotation. A positive FPEAK value forces the program to search such solutions where
there are many zeros and large values but few intermediate values in the source matrix Fii.Fkgy can
further bind individual elements in Fn to zero. On the basis of a similar algorithm, PMF3 solves a
three-way problem.
PMFx and UNMIX estimate Fijand Sjt by minimizing:
Qorx2
=IIR-1- FiW
(A-1)
Where the weighing factor, ait, represents the magnitude of Eit, PMFx limits solutions of eq 2 to
non-negative Fij and Sjt.
Data Needs
A large number of ambient samples (usually much more than the number of factors in the model) are
required to produce a meaningful solution. Species commonly used in PMF are also those in CMB.
Weighting factors associated with each measurement need to be assigned before analysis. The
practitioner also needs to decide the number of factors, FPEAK, and Fksy in the model.
Output
PMFx reports all the elements in Fij and Sjt matrices (PMF2). It also calculates model performance
measures such as deviation n2 and standard deviation of each matrix element. The practitioner
needs to interpret the results linking them to source proxies and source contributions.
Can handle missing or below-
detection-limit data.
Weights species concentrations
by their analytical precisions.
Downweight outliers in the
robust mode.
Derives source proxies from
ambient measurements as they
would appear at the receptor
(does not require source
measurements).
Weaknesses
Requires large (> 100) ambient
datasets.
Need to determine the number of
retaining factors.
Requires knowledge of source
pronles or existing proxies to
verify the representativeness of
calculated factor proxies and
uncertainties of factor
contributions.
Relies on many parameters/initial
conditions adjustable to model
input; sensitive to the preset
parameters.
ME2125
ME2 code is available from Dr.
Pentti Paatero at the University of
Helsinki, Finland as a Microsoft
DOS application.
Principle
The PMFx algorithm is derived from ME2. Unlike PMFx that is limited to questions in the form of eq
1 or 2, ME2 solves all models in which the data values are ntted by sums of products of unknown
(and known) factor elements. The nrst part of the algorithm interprets instructions from the user
and generates a table that specines the model. The second part solves the model using an iterative
minimization approach. Additional constraints could be programmed into the model to reduce the
ambiguity in source apportionment. These constraints may include known source proxies and/or
contributions (e.g., contributions are known to be zero in some cases).
Data Needs
Data needs are similar to those of PMFx but are more nexible. In theory, any measured or unknown
variables may be included in the model as long as they satisfy linear relationships. The users need
to specify the model structure, the input, and the output.
Output
ME2 calculates and reports all unknown variables in the model.
Strengths
Software available.
Can handle user-specined models.
Possibility to include all measured
variables into the model, such as
speciated concentration over
different time scales, size
distributions, meteorological
variables, and noise parameters.
Weaknesses
Require substantial training to
access the full feature of the
software and develop a model.
Generally requires large ambient
datasets.
Need to assume linear
relationships between all
variables.
Relies on many parameters/initial
conditions adjustable to model
input; sensitive to the preset
parameters.
July 2009
A-222
DRAFT-DO NOT CITE OR QUOTE

-------
Receptor Model
Description
Strengths and Weaknesses
UNMIX 20A,2e
UNMIX code is available from Dr.
Ron Henry at the University of
Southern California as an MatLab
application. A stand alone version
(UNMIX version 6) is also available
from EPA.
Principle
UNMIX views each sample as a data point in a multidimensional space with each dimension
representing a measured species. UNMIX solves eqs 2 and 3 by using a principle component
analysis (PCA) approach to reduce the number of dimensions in the space to the number of factors
that produce the data, followed by an unique "edge detection" technique to identify "edges" denned
by the data points in the space of reduced dimension (e.g., Figures 1 and 3). The number of factors
is estimated by the NUMFACT algorithm in advancel 27, which reports the R2 and signal-to-noise
(S|N) ratio associated with the nrst N principle components (PCs) in the data matrix. The number of
factors should coincide with the number of PCs with SfN ratio > 2. Once the data are plotted on
the reduced space, an edge is actually a hyperplan that signines missing or small contribution from
one or more factors. Therefore, UNMIX searches all the edges and uses them to calculate the
vertices of the simplex, which are then converted back to source composition and contributions.
Geometrical concepts of self-modeling curve resolution are used to ensure that the results obey (to
within error) non-negativity constraints on source compositions and contributions.
Data Needs
A large number of ambient samples (usually much more than the number of factors in the model) are
required to achieve a meaningful solution. Species commonly used in UNMIX are also those in CMB.
The measurement precision is not required. The practitioner needs to specify the number of factors
on the basis of the NUMFACT results.
Output
UNMIX determines all the elements in the factor (Fij) and contribution (Sjt) matrices. It also
calculates the uncertainty associated with the factor elements and model performance measures
including: (1) R2, (2) SfN ratio, and (3) strength.
Strengths
Software available with graphical
user interface.
Does not require source
measurements.
Provide graphical problem
diagnostic tools (e.g., species
scatter plot).
Provide evaluation tools (e.g., R2,
S|N ratio).
Weaknesses
Requires large (> 100) ambient
datasets.
Need to assume or predetermine
number of retained factors.
Does not make explicit use of
errors or uncertainties in ambient
measurements.
Cannot use samples containing
missing data in any species.
Limited to a maximum of 7 or 14
(UNMIX version 6) factors.
Can report multiple or no
solutions.
Requires knowledge of existing
source proxies to evaluate the
solutions.
July 2009
A-223
DRAFT-DO NOT CITE OR QUOTE

-------
Receptor Model
Description
Strengths and Weaknesses
PDRM07	Principle
PDRM was developed under the PDRM estimates contributions from selected stationary sources for a receptor site using high time-
Supersites Program and requires resolution measurements and meteorological data. In PDRM, eq 2 is modined to:
MatLab or equivalent software to
perform the calculation.	ft = 2 I + ^
i	!
(A-2)
where is interpreted as the emission rate of species i from stationary source j and (n/O,t is the
meteorological dispersion factor averaged over the time interval t. Equation 4 is solved for ER(\ and
/n/Q,t simultaneously by a nonlinear nt minimizing the objective function, FUN:
1 " " r /v\™»
iil§) -
11=1 pi l ^
(A-3)
Because the number of solutions for a product of unknowns is innnite, additional constraints are
set up for (n/Q,i on the basis of the Gaussian plume model, thus:
/vA**	/Y\*<
mi. -(i). -«ti
' |1 ' |,1	!,1
(x\m _ 1
U/i(	2Sfj
1 (i + h\2
(A-4)
Eqs 6 and 7 limit the solution of eq 5 within the lower (LB) and upper (UB) bound of those predicted
by the Gaussian plume model using different parameterizations.
Data Needs
PDRM requires speciated measurements at a higher time-resolution than typical CMB or PMF
applications because of the fast-changing meteorological parameters. PDRM also requires data for
eq 7: transport speed (u), lateral and vertical dispersion parameters (cry and ai), and stack height
(A).
Output
PDRM determines emission rates and contributions from each point source considered in the model
at the same time resolution as the measurement.
Strengths
Explicitly include meteorological
information and stack
connguration of stationary
sources into the model.
Do not require source
measurements.
Do not need to interpret the
relations between factors and
sources.
Commercial software (e.g.,
MatLab) available for performing
nonlinear Dt.
Suitable for high time-resolution
measurement.
Weaknesses
Can only handle stationary
sources but not area or mobile
sources.
Need to assume that only
stationary sources are considered
in the model contribute
signincantly for a measurement
at the receptor site.
Do not account for uncertainty in
the measurement.
Meteorological data may not be
always available or accurate.
Gaussian plume model may not be
representative of the actual
atmospheric dispersion.
Sensitive to the imposed
constraints (UB and LB).
July 2009
A-224
DRAFT-DO NOT CITE OR QUOTE

-------
Receptor Model
Description
Strengths and Weaknesses
PLS128
Principle
PLS examines the relationships between a set of predictor (independent) and response (dependent)
variables. It assumes that the predictor and response variables are controlled by independent
"latent variables" less in number than either the predictor or the response variables. In recent
applications,90 PM chemical composition and size distribution are used as predictor [X) and
response (K) variables, respectively. Eq 2 is modined to:
* = 2¥h + f»
«. = E¥i.+ °i
(A-5)
(A-6)
where T and U are matrices of so-called "latent variables," and P and C are loading matrices. If /
and Kare correlated to some degree, Tand //would show some similarity. Equations 8 and 9 are
solved by an iterative algorithm "NIPALS," which attempts to minimize E,D, and the difference
between 7" and f simultaneously. If T and U end up being close enough, the / and Yvariables can
be explained by the same latent variables. These latent variables may then be interpreted as source
or source categories.
Data Needs
Typical applications of PLS require both chemical speciated and size-segregated measurements. The
practitioner needs to decide the number of latent variables on the basis of the correlation of
resulting Tand //matrices.
Output
PLS calculates latent variables, which are common factors best explaining the predictor and
response variables, and the residues from ~tting. R> and Ry,
Strengths
Fit two types of measurements
(e.g., chemistry and size) with
common factors. Provide more
information to identify sources.
Analyze strongly collinear and
noisy dataset.
Do not require source
measurements.
Weaknesses
Requires large (> 100) ambient
datasets.
Difncult to relate latent variables
to any physical quantities.
Do not provide quantitative
source contribution estimates.
Need to decide the number of
latent variables.
Do not explicitly make use of
measurement uncertainties.
R, = 1 - wr(EV»ar(A)
R, = 1 - wwymrfy)
indicate the degree to which variables / and Y are explained by the latent variables.
Oan result in no solution.
(A-7)
(A-8)
9 Henry (1997,020941)
4 Lewis etal. (2003, 088413)
6Ogulei etal. (2006,119975)
1 Park et al. (2005,156844)
23	Paatero (1997, 087001)
24	Paatero et al. (2002,156836)
25	Paatero (1999,156835)
26	Henry (2003,156540)
Source: Watson et al. (2008,157128)
July 2009
A-225
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-48. Different receptor models used in the Supersites source apportionment studies: tracer-
based methods.
Receptor Model
Description
Strengths and Weaknesses
£p 129,130
The EF method may use a MLR
algorithm, which is available in most
statistical and spreadsheet software
Principle
A tracer (or marker) for a particular source or source category is a species enriched heavily
in the source emission against other species and other sources. Using EFs-, concentration
of the ith pollutant at a receptor site at time t (i.e., Ci.t) can be expressed as:
!	i V
(A-7)
where the enrichment factor EFi.pj is the ratio of emission rate of the pollutant of interest
(Fii) and tracer species (Fn) from source j. Cpj.t is the concentration of tracer species for
source j at time t, and Zi,t represents contributions from all other sources (including the
background level). The solution for eq 12 is situation-dependent. EFi,n is usually unknown
but may be estimated from source proxies, edges of a two-way scatter plot (e.g., Figures
1 and 3), or the ratio of Ci.tto Cw,t for a particular period when it is believed that a single
source is dominant. In cases where Zi.t is a constant, EFi,n may be derived from MLR.
Data Needs
The minimum data needs include concentrations of all primary tracers at the receptor site.
Known EFs or background levels are helpful.
Output
The EF method determines contributions to species i from each source considered in the
model.
Strengths
No special software needed.
Indicate presence or absence of particular
emitters.
Provides evidence of secondary PM
formation and changes in source impacts
by changes in ambient composition.
Could use a large (> 100) dataset or a
small (e.g., < 10) dataset.
Weaknesses
Semiquantitative method, not specinc
especially when the EFs are unknown in
advance.
Limited to sources with unique markers.
Tracer species must be exclusively from
the sources or source categories
examined.
Provide very limited error estimates.
More useful for sourcefprocess
identincation than for quantisation.
NNLS
The MatLab Optimization Toolbox
provides a function "Isqnonneg" for
performing the NNLS calculation.
Principle
NNLS also solves the EF equation (eq 12 or equivalent) with known target species and
tracer concentrations. Conventional MLR solutions to eq 12 may lead to negative EFs due
to the uncertainty in measurements or colinearity in source contributions. This is avoided in
the NNLS approach since additional non-negative constraints are built into the algorithm,
i.e.:
(A-8)
Utilizing orthogonal decomposition, a NNLS problem can be reduced to the more familiar
least-distance programming and solved by a set of iterative subroutines developed and
tested by Lawson and Hanson.131 In a more general sense, NNLS linearly relates a response
variable to a set of independent variables with only non-negative coefncients.
Data Needs
When applied to EF or MLR problems, NNLS requires the concentration of target (response)
and tracer (independent) species.
Output
NNLS generates non-negative regression coefncients for an EFfMLR problem and these
coefncients can be related to the source contributions.
Strengths
Implemented by many statistical
software packages.
Generate only non-negative EFs or
regression coefncients.
Do not require source measurements.
Possible to include meteorological or
other (besides chemistry) data into the
model.
Weaknesses
Require a large (> 100) set of ambient
measurements.
Semiquantitative method, not specinc.
Do not explicitly consider measurement
uncertainties.
Tracer species must be exclusively from
the sources or source categories
examined.
Non-negative constraints may not be
appropriate in some cases.
July 2009
A-226
DRAFT-DO NOT CITE OR QUOTE

-------
Receptor Model
Description
Strengths and Weaknesses
FAC1
Principle
FAC provides a simple mean of estimating the SOA production rate using the emission
inventories of primary precursor VOCs. FAC is actually a source-oriented modeling
technique but it does not take into account all the atmospheric processes. FAC is denned
as the fraction of SOA that would result from the reactions of a particular VOC:
1 = V
FAC, x ((VQCJ) x fraction of VOC (feasted)
(A-9)
where [VOCijo is the emission rate of VOCi and [SOA] is the formation rate of SOA.
Equation 14 can be viewed as an extension of eq 12 but concentrations are replaced with
emission rates and EFs are replaced with FACs. FAC and the fraction of VOC reacted under
typical ambient conditions have been developed for a large number of hydrocarbons
>C6111 The most signincant SOA precursors are aromatic compounds (especially toluene,
xylene, and trimethylbenzenes) and terpenes. In most applications, these FACs are used
directly to estimate SOA.
Data Needs
FAC requires the VOC emission inventory in the region of interest. The knowledge of O3
and radiation intensity is also helpful for slight modincations of the FACs.
Output
FAC method estimates the total production rate of SOA.
Strengths
Link SOA to primary VOC emissions so
that SOA can also be treated as primary
particles in the PM modeling.
Simple and inexpensive.
Weaknesses
Ignore the innuence of aerosol
concentration and temperature-dependent
gas-particle partitioning on SOA yield.
Limited by the accuracy of VOC emission
inventory.
Do not directly infer the contribution of
each source to ambient SOA
concentration.
Difncult to verify.
"Grosjean and Seinfeld (1989, 045643)
BDarnsetal. (1970,156379)
30Reimann and De Caritat (2000,013269)
3,Lawson and Hanson (1974)
32 Wang and Hopke (1989)
Source: Watson et al. (2008)
July 2009
A-227
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-49. Different receptor models used in the Supersites source apportionment studies:
meteorology-based methods.
Receptor Model
Description
Strengths and
Weaknesses
CPF 1
Principle
CPF estimates the probability that a given source contribution from a given wind direction will exceed a
predetermined threshold criterion (e.g., upper 25th percentile of the fractional contribution from the source of
interest). The calculation of CPF uses source contributions (i.e., S{t in eq 2) determined for the receptor site and
local wind direction data matching each of the source contributions in time. These data are then segregated to
several sectors according to wind direction and the desired resolution (usually 36 sectors at a 10° resolution).
Data with very low wind speed (e.g., < 0.1 mfsec) are usually excluded from analysis because of the uncertain
wind direction. CPF is then determined by:
CPFi H'i
"A»
'•Ai
(A-10)
where mnn is the number of occurrences in the direction sector ~~~ + ~~ that exceeds the specked threshold,
and nnn is the total number of wind occurrences in that sector. Because wind direction is changing rapidly, high-
time resolution measurements (e.g., minutes to hours) are preferred for a CPF analysis. If the calculated source
contributions represent long-term averages, wind direction needs to be averaged over the same duration. In
addition to source contribution, CPF can be applied directly to pollutant concentration measurements at a
receptor site.
Data Needs
CPF requires the time series of source contributions at a receptor site, which is usually determined by CMB or
factor analysis methods using speciated measurements at the site. CPF also requires wind direction and wind
speed data averaged over the same time resolution as the sampling duration.
Output
CPF reports the probability of "high" contribution from a particular source or factor occurring within each wind
direction sector. The results are often presented in a wind rose plot (e.g., Figure 6a).
Strengths
Infer the direction of
sources or factors
relative to the
receptor site.
Provide vesication
for the source
identincation made by
factor analysis
method.
Easy to implement.
Weaknesses
Criterion for the
threshold is
subjective.
Absolute source
contribution (or
fractional
contribution) may be
innuenced by other
factors besides wind
direction (e.g., wind
speed, mixing height).
Local and near-
surface wind direction
only has a limited
implication for long-
range transport.
Easy to be biased by a
small number of wind
occurrences in a
particular sector.
Work better for
stationary sources
than area or mobile
sources.
July 2009
A-228
DRAFT-DO NOT CITE OR QUOTE

-------
Receptor Model
Description
Strengths and
Weaknesses
NPR 136,137	Principle
NPR calculates the expected (averaged) source contribution as a function of wind direction following:
/ a — Wr\
m= v,fi£ir
ai ;
(A-11)
where Wi is the wind direction for the ith sample and Si is the contribution from a specinc source to that sample,
determined from measurements at the receptor site. K is a weighting function called the kernel estimator. There
are many possible choices for K. Henry et al.130 recommend either Gaussian or Epanechnikov functions. The
most important decision in NPR is the choice of the smoothing parameter ~~. If ~~ is too large, S(n) will be too
smooth and meaningful peaks could be lost. If it is too small, S(n) will have too many small, meaningless peaks.
~~ needs to be chosen according to the project-specinc spatial distribution of sources. NPR also estimates the
conndence intervals of S(n) based on the asymptotic normal distribution of the kernel estimates, thus:
t;-i /8 - W\	,
2, *1—x is, - s#))1
Asw= ' /„ /E-Wiu'	
f^KNr
(A-12)
Data Needs
NPR requires the same data as the CPF method, including the time series of source/factor contributions (or
fractional contributions) at the receptor site and local wind direction data matching the sampling duration in
time.
Output
NPR reports the distribution of source contribution as a function of wind direction and the conndence level
associated with it.
Strengths
Infer the direction of
sources or factors
relative to the
receptor site.
Provide verincation
for the source
identincation made by
factor analysis
method.
Require no
assumption about the
function form of the
relationship between
wind direction and
source contribution.
Provide uncertainty
estimates.
Easy to implement.
Weaknesses
Choices for the kernel
estimator and
smoothing factor are
subjective.
Absolute source
contribution (or
fractional
contribution) may be
innuenced by other
factors besides wind
direction (e.g., wind
speed, mixing height).
Local and near-
surface wind direction
only has a limited
implication for long-
range transport.
Easy to be biased by a
small number of wind
occurrences in a
particular sector.
Work better for
stationary sources
than area or mobile
sources.
July 2009
A-229
DRAFT-DO NOT CITE OR QUOTE

-------
Receptor Model
Description
Strengths and
Weaknesses
TSA 138
TSA requires the calculation of air
parcel back trajectory, which is
often accomplished using the HY-
SPLIT model.115,139 HY-SPLIT
version 4.5 is available at
http://www.arl.noaa.gov/-
ready/hysplit4.html.
Principle
Similar to CPF, TSA clusters the measured pollutant concentration or calculated source contribution according
to the wind pattern. However, air parcel back trajectory, rather than local wind direction, is used. A back
trajectory traces the air parcel backward in time from a receptor. The initial height is often between 200 and
1000 m above ground level where the wind direction could differ from the surface wind direction substantively.
For each sample i, TSA obtains one or more trajectories and calculates their total residence time in the jth
directional sector (ny, i.e., the total number of 1 -h trajectory end points that fall into the sector). The pollutant
concentration or source contribution in the sample, Si, is then linearly apportioned into each directional sector
according to Dm and averaged over all samples to produce the directional dependent pollutant
concentration/source contribution for the period of interest:
«s =
i ^ M |
(A-13)
where N is the number of samples. Compared with CPF and NPR, TSA considers the entire air mass history
rather than just the wind direction at the receptor.
Data Needs
TSA requires the time series of pollutant concentration or source contribution at the receptor site, and back
trajectories initiated over the site during the sampling duration. Trajectory is usually calculated once every hour
so TSA is more suitable for analyzing measurements of > 1-h resolution.
Output
TSA reports the avg pollutant concentration or source contribution as a function of wind direction based on back
trajectory calculations.
Strengths
Infer the direction of
sources or factors
relative to the
sampling site.
Provide verincation
for the source
identincation made by
factor analysis
method.
Account for air mass
transport over
hundreds to
thousands of
kilometers and on the
order of several days.
Can represent plume
spread from vertical
wind shear at
different hours of day
by adjusting the initial
height of back
trajectories.
Weaknesses
Need to generate and
analyze the back
trajectory data.
Uncertainty in back
trajectory calculation
increases with its
length in time.
Source contribution
depends on not only
trajectory residence
time but also
entrainment
efnciency, dispersion,
and deposition.
Difncult to resolve the
direction of more
localized sources.
July 2009
A-230
DRAFT-DO NOT CITE OR QUOTE

-------
Receptor Model
Description
Strengths and
Weaknesses
PSCF 140
PSCF requires the calculation of
air parcel back trajectory, which
is often accomplished using the
HY-SPLIT model. 115,139 HY-
SPLIT version 4.5 is available at
http://www.arl.noaa.gov/-
ready/hysplit4.html.
Principle
Ensemble air parcel trajectory analysis refers to the statistical analysis on a group of trajectories to retrieve
useful patterns regarding the spatial distribution of sources. Uncertainties associated with individual trajectory
calculations largely cancel out for a sufncient number of trajectories or trajectory segments. As a popular
ensemble back trajectory analysis, PSCF estimates the probability that an upwind area contributes to high
pollutant concentration or source contribution. Back trajectories are nrst calculated for each sample at the
receptor site. To determine the PSCF, a study domain containing the receptor site is divided into an array of grid
cells. Trajectory residence time (the time it spends) in each grid cell is calculated for all back trajectories and for
a subset of trajectories corresponding to "high" pollutant concentration or source contribution at the site. PSCF
in cell (i,j) is then denned as:
Sum o( "htah" residence time in ell n i
PSCf 		*		
11 Sum of aN residence time In i h|| ( , i
(A-14)
The criterion for high pollutant concentration or source contribution is critical for the PSCF calculation. The 75th
or 90th percentile of the concentration or factor is often used.,,3',4U42 Residence time can be represented by
the number of trajectory end points in a cell.
Data Needs
Similar to TSA, PSCF calculation requires the time series of pollutant concentration or source contribution at the
receptor site, and back trajectories initiated over the site during the sampling period. Trajectories should be
calculated with 1 -to 3-h segment to reduce the uncertainty from interpolation (if needed).
Output
PSCF reports the probability that an upwind area contributes to high pollutant concentrations or source
contribution at the downwind receptor site. The results are often presented as a contour plot on the map. A high
probability usually suggests potential source region (e.g., Figure 4b).
Strengths
Infer the location of
sources or factors
relative to the
sampling site.
Provide verincation
for the source
identincation made by
factor analysis
method
Account for air mass
transport over
hundreds to
thousands of
kilometers and on the
order of several days.
Resolve the spatial
distribution of source
strength
(qualitatively).
Weaknesses
Need to generate and
analyze the back
trajectory data.
Need to correct for
the central tendency
(residence time
always increases
toward the receptor
site regardless of
source contribution).
Uncertainty in back
trajectory calculation
increases with its
length in time.
Source contribution
depends on not only
trajectory residence
time but also
entrainment
efnciency, dispersion,
and deposition.
Difncult to resolve the
location of more
localized sources.
July 2009
A-231
DRAFT-DO NOT CITE OR QUOTE

-------
Receptor Model
Description
Strengths and
Weaknesses
SQTBA ,17'143
SQTBA requires the calculation of
air parcel back trajectory, which
is often accomplished using the
HY-SPLIT model. 115,139 HY-
SPLIT version 4.5 is available at
http://www.arl.noaa.gov/'
ready/hysplit4.html.
Principle
SQTBA is another type of ensemble air parcel trajectory analysis. The concept of SQTBA is to estimate the
"transport neld" for each trajectory ignoring the effects of chemical reactions and deposition. Back trajectories
are nrst calculated for each sample at the receptor site, and a study domain containing the receptor site is
divided into an array of grid cells. SQTBA assumes that the transition probability that an air parcel at (x',y',t'|,
where x' and y' are spatial coordinates and t' means time, will reach a receptor site at (x,y,t) is approximately
normally distributed along the trajectory with a standard deviation that increases linearly with time
upwind144'146, thus:
Ok. y, ?i*', y',z') ¦
1

-j asp
¦=; 11 —

ut'
-y'a i
i i
(A-15)
where (X,Y) is the coordinate of the grid center, a is the dispersion speed, and x'(t') and x' ft') represent the
trajectory. The probability neld, Q, for a given trajectory is then integrated over the upwind period, ~, to produce
a two-dimensional "natural" (nonweighted) transport neld:
y') =-
Oft, y, !fr", f, z')
dr
(A-16)
After the transport neld for each trajectory is established, they are weighted by the corresponding pollutant
concentration or source contribution at the receptor site and summed to yield the overall SQTBA neld.117
Data Needs
SQTBA requires the time series of pollutant concentration or source contribution at the receptor site, and back
trajectories initiated over the site during the sampling period. Trajectories should be calculated with 1 to 3-h
segment to reduce the uncertainty from interpolation (if needed).
Output
SQTBA put more weight on trajectories associated higher pollutant concentration or source contribution and
therefore the resulting neld may imply the major transport path.
Strengths
Imply the location of
sources or factors
relative to the
sampling site.
Account for air mass
transport over
hundreds to
thousands of
kilometers and on the
order of several days.
Resolve the spatial
distribution of source
strength
(qualitatively).
Weaknesses
Need to generate and
analyze the back
trajectory data.
Need to correct for
the central tendency
(residence time
always increases
toward the receptor
site regardless of
source contribution).
Need to estimate
dispersion velocity.
Involve complicated
calculations.
Physical meaning of
the SQTBA neld is
unclear.
Difncult to resolve the
location of more
localized sources.
July 2009
A-232
DRAFT-DO NOT CITE OR QUOTE

-------
Receptor Model
Description
Strengths and
Weaknesses
RTWC 146
RTWC requires the calculation of
air parcel back trajectory, which
is often accomplished using the
HY-SPLIT model. 115,139 HY-
SPLIT version 4.5 is available at
http://www.arl.noaa.gov/
ready/hysplit4.html
Principle
As an ensemble air parcel trajectory analysis, RTWC requires back trajectories calculated for each sample at the
receptor site, and a study domain containing the receptor site divided into an array of grid cells. RTWC assumes
that no major pollutant sources are located along "clean" (associated with low pollutant concentrations)
trajectories and that "polluted" trajectories picked up emissions along their paths. In practice, RTWC distributes
pollutant concentrations at the receptor to upwind grid cells along the back trajectories according to the
trajectory residence times in those cells.117 ,40
% = st
resident time In cell
aferage residence lime in each ceil
(A-17)
where Sk is the pollutant concentration or source contribution determined upon the arrival of trajectory k and Si>
is the redistributed pollutant concentration or source contribution for cell i upwind.
RTWC is known for the problem of "tailing effect," i.e., spurious source areas can be identined when cells are
crossed by a very small number of trajectories. Although some corrections were proposed147 these approaches
are purely empirical.
Strengths
Imply the location of
sources or factors
relative to the
sampling site.
Account for air mass
transport over
hundreds to
thousands of
kilometers and on the
order of several days.
Resolve the spatial
distribution of source
strength
(qualitatively).
Weaknesses
Need to generate and
analyze the back
trajectory data.
Need to correct for
the central tendency
and tailing effect.
The amount of
emission entrainment
should not be
proportional to the
residence time of
trajectories (so there
is no linear
relationship between
RTWC neld and
source strength).
Physical meaning of
the RTWC neld is
unclear.
Difncult to resolve the
location of more
localized sources.
113 (Pekney et al„ 2006,086115)
117 (Zhou et al„ 2004.157190)
134	(Ashbaugh, 1983.156229)
135	(Ashbaugh et al„ 1984, 045148)
136	(Henry et al.. 2002.136097)
137	(Yu et al., 2004.101779)
138	(Parekhand Husain, 1981.156840)
140 (Hopke et al., 1995.156566)
143	(Keeler and Samson, 1989,156633)
144	(Samson. 1978,156941)
145	(Samson. 1980, 073010)
146	(Stohl, 1996,157014)
147	(Cheng et al.. 1993, 052294)
Source: (Watson et al., 2008,157128)
July 2009
A-233
DRAFT-DO NOT CITE OR QUOTE

-------
A.3.2. Source Profiles
Table A-50. Source Profiles: Part I
Element
Symbol
Motor Vehicle Exhaust ¦ Gasoline
Coal Combustion
Highway Road Dust
Unpaved Road Dust

Refinery
Weight %
Uncertainty
Weight % Uncertainty Weight % Uncertainty Weight % Uncertainty Weight % Uncertainty
Aluminum
Al
0.1
¦99
5.968
0.5247
5.729
0.4058
7.4822
0.9315
8.4853
2.3478
Antimony
Sb
0.01
¦99
0
0.0625
0
0.0335
0
0.1601
0
0.0285
Arsenic
As


0
0.0164
0
0.0123
0
0.0226
0
0.0045
Barium
Ba
0.01
¦99
1.3315
1.0801
0.1377
0.1027
0
0.5473
0
0.0979
Cadmium
Cd


0
0.0341
0
0.019
0
0.0881
0
0.0155
Calcium
Ca
0.42
¦99
3.4536
1.0411
2.5657
0.1388
2.163
1.0444
0.1236
0.056
Chloride ion
CI"
0.39
¦99








Chromium
Cr
0.01
¦99
0.0176
0.0041
0.0271
0.0023
0.0312
0.0161
0.0443
0.0127
Cobalt
Co


0
0.0432
0
0.0668
0
0.0869
0
0.0218
Copper
Cu
0.02
¦99
0.0179
0.0112
0.0219
0.0101
0.0474
0.0307
0.0299
0.0082
Total carbon
TC


4.2763
4.2579
14.3927
2.3449
4.2671
3.7193
0
1.6175
Gallium
Ga


0.014
0.014
0
0.005
0
0.0233
0
0.0059
Gold
Au










Indium
In
0
¦99
0
0.0404
0
0.022
0
0.1041
0
0.0183
Iron
Fe
1.27
¦99
2.916
0.3827
4.5713
0.2661
5.5128
2.1152
1.4708
0.2216
Lanthanum
La
0
¦99
0
0.2462
0
0.1341
0
0.6521
0
0.1146
Lead
Pb
0.08
¦99
0.068
0.0336
0.067
0.0074
0.0288
0.0284
0.0097
0.0063
Magnesium
Mg
0.14
¦99








Manganese
Mn
0.01
¦99
0.0284
0.0139
0.087
0.009
0.1372
0.0509
0.016
0.002
Mercury
Hg
0
¦99
0
0.0154
0
0.0083
0
0.0383
0
0.0073
Molybdenum
Mo


0
0.0134
0
0.0071
0
0.0331
0.0079
0.0088
Nickel
Ni
0.01
¦99
0.0072
0.0019
0.0081
0.0015
0.0091
0.0057
0.04
0.0065
Nitrate
NOs"
0.06
¦99
0
0.2116
0
0.094
0
0.6371
0
0.0772
Organic carbon OC
59.37
¦99
0
2.9263
12.7127
2.1296
4.2671
2.2637
0
1.5288
Palladium
Pd


0
0.0263
0
0.0151
0
0.0701
0
0.0127
Phosphorus
P
0.27
¦99
0.9372
0.6322
0
0.0324
0.1603
0.044
0.0689
0.0144
Potassium
K
0.01
¦99
0.4644
0.0602
2.7161
0.3069
2.8299
0.4949
0.0825
0.0234
Rubidium
Rb


0.0053
0.0043
0.0184
0.0023
0.0184
0.0093
0
0.002
Selenium
Se


0.0406
0.0407
0
0.0024
0
0.0108
0
0.0021
Silicon
Si
1.61
¦99
9.0112
0.5675
17.596
1.4183
24.2969
4.0089
17.9733
5.1834
Silver
Ag


0
0.0312
0
0.0175
0
0.083
0
0.0151
Sodium
Na
0.01
¦99








Strontium
Sr


0.1964
0.0686
0.0395
0.0078
0.0313
0.0112
0.0094
0.0031
July 2009
A-234
DRAFT-DO NOT CITE OR QUOTE

-------


Motor Vehicle Exhaust -
Gasoline Coal Combustion
Highway Road Dust
Unpaved Road Dust

Refinery
Sulfate
sor

10.1716
8.9405
1.1604 0.2003
0.8688 1.3788
2.3243
3.4523
Sulfur
S
0.37 -99
2.948
2.729
0.598 0.0509
0.2808 0.3884
0.6304
0.9627
Thallium
Tl







Tin
Sn

0
0.0527
0 0.0298
0 0.1464
0
0.0254
Titanium
Ti

0.4315
0.0651
0.3612 0.0313
0.5258 0.1289
0.6178
0.0711
Uranium
U







Vanadium
V

0
0.0734
0.0288 0.0074
0 0.0646
0.0432
0.0084
Yttrium
Y

0
0.006
0.0046 0.0012
0 0.0146
0
0.0029
Zinc
Zn
0.49 -99
0.0797
0.0341
0.0932 0.0256
0.0502 0.021
0.0166
0.003
Zirconium
Zr

0.0247
0.0043
0.0128 0.0025
0.0219 0.0168
0.0166
0.0022
Ammonium
NH4+
0.34 -99
0.3476
0.1352
0 0.025
0 0.1317
0.3281
0.5565
Sodium ion
Na+







Carbonate
C03 -







Organic carbon 0C2
II
Organic carbon 0C3
III
Organic carbon 0C4
IV
EC I
EC1







Chlorine atom
cr

0.0629
0.0221
3.4403 0.5505
0.1519 0.0755
0.0186
0.0074
EC III
EC3







EC
EC
16.44 -99
4.2763
3.0931
1.68 0.9817
0 2.9512
0
0.5283
Bromine Atom
Br

0.0147
0.0154
0.0037 0.0011
0 0.0078
0
0.0017
Organic carbon 0C1
I
EC II
EC2







Sulfur dioxide
SO?

7262.6687
7677.5681




Potassium ion
K+

0.1109
0.0571
0.2295 0.1046
0.1263 0.0744
0.0115
0.0059
Source: USA EPA Speciate database http:llwww.epa.govlttnchie1lsoftwarelspeciatelindex.html
Part II
Element
Symbol
Residential Wood Burning
Oil Combustion

DE
Fly Ash
Incinerator
Weight %
Uncert-ainty
Weight %
Uncert-ainty
Weight %
Uncert-ainty
Weight %
Uncert-ainty Weight %
Uncert-ainty
Aluminum
Al
0.0034
0.0103
0
0.05
0
0.01
1.5708
0.4755
1.15
0.83
Antimony
Sb
0.0002
0.0108
0
0.01
0
0.01
0.007
0.0218
0.01
0.15
Arsenic
As
0.0003
0.0016
0.02
0
0
0
0.001
0.0023
0
0.04
Barium
Ba
0.0093
0.0369
0
0.03
0.01
0.04
0.0303
0.0655
0.14
0.55
Cadmium
Cd
0.0013
0.0058
0
0.01
0
0.01
0
0.0154
0.01
0.08
Calcium
Ca
0.0664
0.0165
0
0.04
0.01
0.01
10.1398
1.7825
2.37
0.62
Chloride ion
ci-
0.0028
0.0004




17.5498
1.5419


July 2009
A-235
DRAFT-DO NOT CITE OR QUOTE

-------


Residential Wood Burning
Oil Combustion

DE

Fly Ash
Incinerator
Chromium
Cr
0.0003
0.0012
0.01
0.01
0
0
0.0054
0.001
0.02
0.02
Cobalt
Co
0.0005
0.0005
0.05
0.01
0
0
0.0015
0.0128
0
0.03
Copper
Cu
0.0002
0.0007
0.01
0.01
0
0
0.017
0.0013
0.08
0.1
Total carbon
TC
70.6416
7.1435
3.55
1.0855
98.94
17.859
1.4329
0.2009
55.79
27.5948
Gallium
Ga
0
0.0016
0.01
0
0
0
0.0013
0.0018
0
0.02
Gold
Au






0.0008
0.0033


Indium
In
0.0021
0.0069
0
0.01
0
0.01
0
0.0164
0.01
0.1
Iron
Fe
0.0038
0.0017
0.68
0.1
0
0
0.8306
0.059
1.72
0.31
Lanthanum
La
0.0086
0.0431
0
0.04
0.02
0.05
0.0046
0.0868
8.43
61.15
Lead
Pb
0.0031
0.0018
0
0
0
0
0.0031
0.0031
14.56
11.69
Magnesium
Mg






0.4455
0.0465


Manganese
Mn
0.003
0.0013
0
0
0
0
0.0426
0.0033
0.04
0.01
Mercury
Hg
0.0004
0.0027
0
0
0
0
0.0008
0.0025
27.63
47.27
Molybdenum
Mo
0
0.0024
0
0
0
0
0.0041
0.001
0.01
0.04
Nickel
Ni
0.0002
0.0005
2.36
0.23
0
0
0.0028
0.0004
0.01
0
Nitrate
NOs-
0.2025
0.0156
0
0
0.06
0.01
0
0.2192
5.5
4.55
Organic carbon
OC
49.4961
5.481
1.71
0.56
90.8
14.79
1.4329
0.1592
37.21
18.03
Palladium
Pd
0.0006
0.0047
0
0
0
0
0
0.0126
0.02
0.07
Phosphorus
P
0
0.0051
0
0.65
0.01
0.02
0.5808
0.2447
0.05
0.16
Potassium
K
0.6346
0.1008
0
0
0
0
24.4341
5.0076
1.28
0.86
Rubidium
Rb
0.0007
0.0007
0
0
0
0
0.0351
0.0026
0
0.02
Selenium
Se
0.0001
0.0008
0.03
0
0
0
0.0018
0.0003
0.01
0.01
Silicon
Si
0.0443
0.0167
0
0.09
0.01
0.01
4.0201
1.2886
4.42
1.82
Silver
Ag
0.0023
0.0054
0
0
0
0.01
0
0.0143
0.02
0.08
Sodium
Na






2.8137
0.2174


Strontium
Sr
0.0006
0.0009
0
0
0
0
0.0406
0.0029
0.02
0.01
Sulfate
S042"
0.4553
0.0359
25.29
5.62
0.53
0.07
8.0717
0.6409
10.46
2.6
Sulfur
S
0.1533
0.0173
16.48
1.62
0.59
0.21
2.6349
0.1873
3.16
0.63
Thallium
Tl






0.0011
0.0025


Tin
Sn
0.0006
0.0092
0
0.01
0
0.01
0.0067
0.0198
0.04
0.14
Titanium
Ti
0.001
0.012
0.01
0.01
0
0.01
0.058
0.0093
0.11
0.17
Uranium
U






0.0021
0.0052


Vanadium
V
0.0007
0.005
0.4
0.04
0
0.01
0.0038
0.011
0.01
0.07
Yttrium
Y
0.0001
0.0011
0
0
0
0
0.0013
0.0021
0
0.02
Zinc
Zn
0.0762
0.0054
0.01
0
0.02
0.02
0.031
0.0023
0.57
0.39
Zirconium
Zr
0
0.0014
0
0
0
0
0.0039
0.0008
0
0.02
Ammonium
NH4 +
0.1132
0.014
0.84
0.24
0.03
0.01
0.0234
0.022
7.41
7.81
Sodium ion
Na+


0.11
0.02
0
0.01
4.7518
0.3438
1.81
2.63
Carbonate
C03 -


0
0.0214
0.2577
0.4463




July 2009
A-236
DRAFT-DO NOT CITE OR QUOTE

-------


Residential Wood Burning
Oil Combustion

DE
Fly Ash
Incinerator
Organic carbon II
0C2
7.513
0.6675







Organic carbon III
0C3
8.9627
1.4665







Organic carbon IV
0C4
2.7683
1.1919







EC I
EC1
20.342
2.9324







Chlorine atom
CI'
0.2874
0.0404
0.05
0.01
0.03
0.01
27.5797
8.1193
6.35 10.46
EC III
EC3
2.2878
0.4252







EC
EC
21.1455
4.5813
1.84
0.93
8.14
10.01
0
0.1227
18.58 20.89
Bromine Atom
Br
0.0029
0.0011
0
0
0
0
0.0441
0.0032
0.19 0.3
Organic carbon I
0C1
25.1452
4.6648







EC II
EC2
2.9362
1.2422







Sulfur dioxide
SO?









Potassium ion
K+
0.5208
0.0795
0.01
0.01
0
0.01
14.5473
1.3393
1.01 0.42
Source: U.S. EPA SPECIATE database http:llwww.epa.govlttnchie1lsoftwarelspeciatelindex.html
A.3.3. Receptor Model Results
Table A-51. PM10 receptor model results
% Contribution
Sampling Site
Wood
Smoke
Diesel
Gasoline
Vehicles
Natural Gas
Combustion
Vegetative
Detritus
Tire Wear
Debris
Total %
Allocated
Apline, CA, 1994-1995
15.00
33.19
46.46

5.31
6.91
99.955752
Apline, CA, 1995
9.92
58.78
11.47

19.63
9.43
99.795918
Apline, CA, 1995
10.97
65.64
10.81

12.66
5.31
100.07722
Atascadero, CA, 1994-1995
44.22
22.16
26.44



99.733333
Atascadero, CA, 1995
21.36
38.99
12.41

17.89

100.08772
Atascadero, CA, 1995
73.45
18.11


3.14

100.01241
Lake Arrowhead, CA, 1994-1995
6.86
46.55
33.92
2.73
9.85
20.42
99.896907
Lake Arrowhead, CA, 1995
4.85
65.20
7.40
4.95
17.65

100.04902
Lake Arrowhead, CA, 1995
9.91
38.90
46.70
0.79
3.66
28.01
99.955947
Lake Elsinore, CA, 1994-1995
12.72
44.01
18.61

4.21
9.78
99.967638
Lake Elsinore, CA, 1995
17.13
74.72

0.26
7.81
29.17
99.924528
Lake Elsinore, CA, 19952
6.84
38.48
10.85
0.21
15.55
11.93
99.946809
Lancaster, CA, 1994-1995
22.49
43.14
20.56
0.45
3.73
26.38
100.14006
Lancaster, CA, 1995
3.69
46.18
12.66
0.20
8.21

100.09967
Lancaster, CA, 1995
34.89
37.30
7.33
0.61
7.78

99.839228
Lompoc, CA, 1994-1995

18.16
49.65

5.89
26.00
100.07092
Lompoc, CA, 1995
13.09
51.27
14.73

20.73
14.11
99.818182
July 2009
A-237
DRAFT-DO NOT CITE OR QUOTE

-------
% Contribution
Lompoc, CA, 1995

79.42
10.19

10.87
16.61
100.48077
Long Beach, CA, 1994-1995
10.12
43.24
16.49
0.13
3.97
19.52
99.955423
Long Beach, CA, 1995
2.38
70.25
5.47
0.86
6.79
15.71
99.865643
Long Beach, CA, 1995
14.32
56.80
6.15
0.72
5.34
9.85
99.939832
Mira Loma, CA, 1994-1995
4.68
48.87
18.10

8.82
20.31
100
Mira Loma, CA, 1995
5.20
53.72
6.65

18.79
19.06
100.07092
Mira Loma, CA, 1995
27.97
41.88
8.87

11.50
20.17
100.07519
Riverside, CA, 1994-1995
14.14
46.67
12.03

6.83

99.972222
Riverside, CA, 1995
6.20
52.15
7.93
0.16
14.54
7.85
100.0409
Riverside, CA, 1995
25.28
47.65


6.91
8.15
100
San Dimas, CA, 1995
7.62
71.35
4.87
0.15
8.35
12.78
100.17308
San Dimas, CA, 1995
22.01
61.34
4.48
0.23
3.70
15.05
99.919463
Santa Maria, CA, 1994-1995
18.66
23.99
22.03

5.58
14.70
100.14493
Santa Maria, CA, 1995
12.94
52.57
11.87
0.27
9.63
11.25
100.05348
Santa Maria, CA, 1995
12.24
48.13
10.79
0.47
18.04
9.81
104.71963
Upland, CA, 1994-1995
20.33
46.39
14.08

4.49

100
Upland, CA, 1995
7.33
68.69
3.50
0.17
9.19

100.12891
Upland, CA, 1995
28.10
46.52
4.90
0.33
10.30

99.952774
Source: Manchester-Neesvig et al. (2003, 098102)
Table A-52. PM2.5 receptor model results
Sampling Site
Measured PM2.5
Concentration
Vegetative
Burning
R°aSoMUSt ,NH4'2S°4 NH4N03 NaCL Tailpipe
Brake
Wear
Total %
Allocated
Albany, NY 2000-2004
34.9
7.60
11.70 2.70
4.90
11.70
2.90

118.91
Birmingham, AL, 2000-2004
24.1
3.90
8.40 3.70
2.70
0.10
5.70

101.66
Houston, TX, 2000-2004
17.6
3.10
6.90 1.60
2.50
0.10
3.80

106.25
Long Beach, CA, 2000-2004
46.8
4.60
9.60 2.10
18.90
0.80
6.50
3.50
98.29
Source: Abu-Allaban et al. (2007,098575)
July 2009
A-238
DRAFT-DO NOT CITE OR QUOTE

-------
A.4. Exposure Assessment
A.4.1. Exposure Assessment Study Findings
Table A-53. Exposure Assessment Study Summaries
Abou Chakra et al. (2007,098819)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Experimental, in vitro. HeLa cells incubated with organic extracts of personal PMio and PM2.6 samples
NR
3 French metropolitan areas (Paris, Rouen, Strasbourg) with varying air quality and emission sources
Individuals from urban areas with varying air pollution levels and emissions sources
Children ages 6-13. Ages of adults not given
NR
Harvard multi-pollutant Chempass personal exposure sampler placed in backpacks with BGI pump operating at 1.8 l/min.
PM10, PM2.6
NR
NR
Organic extracts of PM 10, organic extracts of PM2.6
Genotoxic effects were stronger for organic extracts of PM2.6 than for PM 10 and greater in winter than summer. Greater effects for winter
samples may be attributed to elevated winter PAH concentrations.
Abu-Allaban et al. (2004,156187)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Exposure assessment of real world motor vehicle emissions
May 18-22,1999
Tuscarosa Mountain Tunnel, Pennsylvania Turnpike
Highway tunnel
No personal exposure assessment was conducted.
0.01-0.5 fjm
NR
NR
Monitoring sessions with the highest fraction heavy-duty vehicles had the highest particle concentrations; Observed particle size distribution
was a combination of 2 bimodal log-normal distributions a dominant nucleation mode (86% of area under the curve).
Adar et al. (2007,098635)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Periods
Personal Size
Microenvironment Size
Cohort
March 2002-June 2002
St. Louis, Missouri
Senior citizens exposed to traffic-related PM
60
NR
Samples of FeNO were collected between 8:00 and 9:00 a.m. on the mornings before and after each trip. In the hours surrounding these
samples, group-level measurements of particle concentrations also were collected using several continuous instruments installed on two
portable carts. These carts were first positioned in a central location inside the participants' living facilities 24-h before each trip. The carts
remained at the facilities until it was time for the trips, at which point they followed the participants from the health testing room, onto the
bus, to the group activity, and to lunch. After the trip home aboard the bus, the carts were returned to the central location in the living
facility where they remained until the conclusion of the health testing on the following morning. Continuous measurements of ambient
particles and gases also were collected from a central monitoring station in East St. Louis, Illinois.; Specifics-; Two portable carts containing
continuous air pollution monitors were used to measure group-level micro-environmental exposures to traffic related pollutants, including fine
particulate mass (<2.5 |Jm aerodynamic diameter; PM2.6), BC, and size-specific particle counts. PM2.6 concentrations were measured
continuously using a DustTrak aerosol monitor model 8520 with a Nafion diffusion dryer. Integrated samples of PM2.6 mass also were
collected using a Harvard Impactor for daily calibration of the trip and facility (missing items?)
Continuous BC concentrations were measured using a portable aethalometerwith a 2.5-|Jm impaction inlet. Particle counts were measured
using a model CI500 optical particle counter with a modified flow rate of 0.1 cubic feet per minute.
NR
PM2.6
July 2009
A-239
DRAFT-DO NOT CITE OR QUOTE

-------
Ambient Size PM2.6; PMio
Component(s) BC; Pollen and Mold also assessed
Primary Findings Fine particle exposures resulted in increased levels of FeNO in elderly adults, suggestive of increased airway inflammation. These associations
were best assessed by microenvironmental exposure measurements during periods of high personal particle exposures.In pre-trip samples,
both microenvironmental and ambient exposures to fine particles were positively associated with FeNO. For example, an interquartile
increase of 4/yglm3 in the daily microenvironmental PM2.6 concentration was associated with a 13% [95% CI: 2-24) increase in FeNO. After
the trips, however, FeNO concentrations were associated predominantly with microenvironmental exposures, with significant associations
for concentrations measured throughout the whole day. Associations with exposures during the trip also were strong and statistically
significant with a 24% (95% CI: 15-34) increase in FeNO predicted per interquartile increase of 9 //gfnv1 in PM2.6. Although pre-trip findings
were generally robust and the post-trip findings were generally robust, the post-trip findings were sensitive to several influential days.
Adgate et al. (2002, 030676)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Comparison of outdoor, indoor and personal PM2.6 in three communities.
April-June, June-August, September-November, 1999
Battle Creek, East St. Paul, and Phillips, Minnesota, constituting the Minneapolis-St. Paul metropolitan area.
adults in urban areas
mean age 42 ±10, range 24-64 yr
No
Inertial impactors (PEM) in a foam-insulated bag with shoulder strap with the inlet mounted on the front.
PM2.6
PM2.6
PM2.6
NR
The relative level of concentrations report in other studies was duplicated. Outdoor < indoor < personal. On days with paired samples
(n - 29), outdoor concentrations were significantly lower (mean difference 2.9 /yglm3, p - 0.026) than outdoor at home.
Adgate etal. (2007,156196)
Study Design NR
Period 1999-; April 26-June 20, June 21 -August 11, September 23-November 21
Location Minneapolis-St. Paul metropolitan area
Population NR
Indoor Source Cigarette smoke, resuspension of house dust from carpets, furniture and clothes, and emissions from stoves and kerosene heaters (Leaderer
et al., 1993: Ferro et al., 2004).
Personal Method Personal monitoring was conducted using a and consisted of two consecutive days, and was conducted so that the two 24-h averages
matched indoor (I) and personal (P) measurements were collected in concert with 0 samples in each community. Gravimetric concentrations
for P and I were collected using inertial impactor environmental monitoring inlets and air sampling pumps. To obtain I measurements,
monitors were placed inside each residence in a room where the participants reported spending the most waking hours. P measurements
were obtained by carrying personal pumps in small bags.: Outdoor central site samples (0) were collected near the approximate geographic
center of each neighborhood and monitors ran from midnight to midnight for two consecutive 24-h periods, followed by a day to change
filters. Gravimetric 0 PM2.6 concentrations were obtained using a federal reference method sampler.
Personal Size PM2.B-broken down into TE
Microenvironment Size PM2.B-broken down into TE
Ambient Size PM2.B-broken down into TE
Component(s) Ag, A I, Ca, Cd, Co, Cr, Cs, Cu, Fe, K, La, Mg, Mn, Na, Ni, Pb, S, Sb, Sc, Ti, Tl, V, Zn
Primary Findings The relationships among P, I, and 0 concentrations varied across TEs. Unadjusted mixed-model results demonstrated that 0 monitors are
more likely to underestimate than overestimate exposure to many of the TEs that are suspected to play a role in the causation of air
pollution related health effects. These data also support the conclusion that TE exposures are more likely to be underestimated in the lower
income and centrally located PHI community than in the comparatively higher income BC K community. Within the limits of statistical power
for this sample size, the adjusted models indicated clear seasonal and community related effects that should be incorporated in long-term
exposure estimates for this population.
Adgate etal. (2003,040341)
Study Design	Time-series epidemiologic study
Period April-November 1999: spring: 26 April-20 June: summer: 21 June-11 August: fall: 23 September-21 November
Location	Minneapolis-St. Paul, Minnesota
Population	Healthy non-smoking results
Age Groups	24-64 yr (mean age 42 ± 10)
Indoor Source	NR
July 2009
A-240
DRAFT-DO NOT CITE OR QUOTE

-------
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Personal and indoor gravimetric PM concentrations were collected using PM2.6 inertial impactor environmental monitoring inlets and air
sampling pumps. Monitors were placed inside each participant's residence in the room where he/she reported spending the majority of their
waking hours to obtain indoor (I) measurements. Participants also carried personal pumps in small bags to obtain personal (P) measurements.
Start times for indoor and personal monitors were always within a few minutes of each other. Gravimetric outdoor (0) and central site PM2.6
concentrations were obtained using a federal reference method sampler and EPA site requirements for ambient sampling. Gravometric
samples were collected near the approximate geographic center of each neighborhood, and monitors ran from midnight to midnight for 2
consecutive 24-h periods, followed by a day to change filters.
Personal (P) PM2.6 concentrations were higher than indoor (I) concentrations, which were higher than outdoor (0) concentrations. In healthy
non smoking adults, moderate median PI; modest median 10; and minimal median PO longitudinal correlation coefficients were observed for
PM2.6 measurements. A sensitivity analysis indicated that correlations did not increase if the days with exposures to environmental tobacco
smoke or occupational exposures were excluded. In the sample population neither P nor I monitors provided a highly correlated estimate of
exposure to 0 PM2.6 over time. These results suggest that the studies showing relatively strong longitudinal correlation coefficients between
P and 0 PM2.E for individuals sensitive to air pollution health effects do not necessarily predict exposure to PM2.6 in the general population.
Alander etal. (2004,055650)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Exposure assessment, characterization of effects of fuel reformulation, engine design, and exhaust after-treatment on PM emissions
NR
Laboratory
Diesel-powered passenger cars of different engine types and different formulations of diesel fuel
No personal exposure assessment was conducted
NR
Total carbon, organic carbon, elemental carbon, sulfate, nitrate, chloride
Reformulated low sulfur diesel fuel produced 40% less total carbon mass compared to standard fuel. Organic carbon constituted 27-61%
carbon mass from an indirect ignition engine. Low sulfur fuel reduced organic carbon mass by 10-55%, depending on engine.
Allen et al. (2003,053578)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Use of continuous light scattering data to separate indoor PM into indoor- and outdoor-generated components to enhance knowledge of the
outdoor contribution to total indoor and personal PM exposures.
November 1999-May 2001
Seattle, WA
Elderly people and children spending most of their time (up tp 70%) indoors.; The study included healthy elderly subjects, elderly with COPD
and coronary heart disease (CHD), and child subjects with asthma.
Age n; 0-29 25; 30-59 36; > 60 22; unknown 2
Suggested (not identified)
NR. Indoor and outdoor sampling conducted
NR
PM2.6
PM2.6
Sulfur
A recursive mass balance model can be successfully used to attribute indoor PM to its outdoor and indoor components and to estimate an avg
P, a, k, and NH4+ for each residence.
Allen et al. (2007,154226)
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Heating season
Seattle, WA
October-February; Non-heating season March-September; (Year not specified)
Indoor and outdoor PM2.6 was measured using a 10-lfmin single-stage Harvard Impactor (HI) with 37-mm Teflon filters. The relationship
between particle mass concentration and light scattering coefficient (bsp) was also measured on a continuous basis indoors and outdoors
using nephelometers (model 902 and 903).
NR
PM2.6
PM2.6
Sulfur (measured by XRF)
July 2009
A-241
DRAFT-DO NOT CITE OR QUOTE

-------
Primary Findings The authors showed that RM can reliably estimate Fint. Simulation results suggest that the RM Fuestimates are minimally impacted by
measurement error. In addition, the average light scattering response per unit mass concentration was greater indoors than outdoors.
Results show that the RM method is unable to provide satisfactory estimates of the individual components of Fm. Iliu et ndividual homes
vary in their infiltration efficiencies, thereby contributing to exposure misclassification in epidemiologic studies that assign exposures using
ambient monitoring data. This variation across homes indicates the need for home-specific estimation methods, such as RM or sulfur tracer,
instead of techniques that give average estimates of infiltration across homes.
Allen et al. (2007,156207)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Component(s)
Primary Findings
Primarily a study of exposure to indoor PBDE congeners.
Jan-Mar 2006
Greater Boston area, Massachusetts
Urban dwellers
No particulate sampled
No particulate sampled
No particulate sampled
Polybrominated diphenyls (PBDEs), divided into 13 congeners and total BDE (SBDE), which includes both vapor and particulate phase.
Total personal air concentrations (particulate ± vapor) were 469 pg/m3 for non-209 BDEs and 174 pg/m3 for BDE 209, significantly higher
than bedroom and main living room concentrations (p - 0.01). The ratio of personal air to room air increased from 1 for vapor-phase
congeners to 4 for fully particulate-bound congeners, indicating a personal cloud effect.
Andresen et al. (2005,156216)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Residential exposure assessment personal and indoor
June-July 2002 to December 2002
Mysore, India
Women working at home, non-smoking, and primary household cook
18-50 yr old
Cooking fuel source
PM2.6 gravimetric filter measurement
PM2.6
Using kerosene for cooking was associated with higher personal PM2.B exposure in both winter and summer compared to LPG.; Kerosene use
during winter was associated with higher personal PM2.6 compared to summer.; LPG use was associated with comparable personal PM2.6
across both seasons.; Indoor PM2.6 measurements followed similar patterns by fuel-type and season.; Socioeconomic status, age, season,
and income were significant predictors of cooking fuel choice.
Annesi-Maesano et al. (2007,093180)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Population based
March 1999 to October 2000
Bordeaux, France; Clermont-Ferrand, France; Creteil, France; Marseille, France; Strasbourg, France; Reims, France
School children
10.4 ± 0.7 yr
NR
PM2.E was monitored simultaneously in both schoolyards (proximity level) and fixed-site monitoring stations (city level) using 4L/min battery
operated pumps attached to polyethylene filter sampling cartridges.
NR
NR
PM2.6
NR
Results show an increased risk for EIB and flexural dermatitis at the period of the survey, past year atopic asthma and SPT positivity to
indoor allergens in children exposed to high levels of traffic-related air pollution (PM2.6 concentrations exceeding 10 //gfnv1]. Population based
findings are also consistent with experimental data that have demonstrated that inhalation of traffic-related air pollutants either individually
or in combination, can enhance the immune responses and airway response to inhaled allergens, such as pollens or house dust mites, in
atopic subjects.
Balakrishnan et al. (2002,156247)
Study Design	Exposure assessment
Period July-December 1999 (20 weeks)
Location	50 villages, Tamil Nadu, India
Population	Men and women in rural households; children exempt
July 2009
A-242
DRAFT-DO NOT CITE OR QUOTE

-------
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
All, children exempt
Yes
Personal sampler for cooks during cooking time
Respirable Particulates (based on NIOSH protocol)
Respirable Particulates (based on NIOSH protocol)
Respirable Particulates (based on NIOSH protocol)
NR
Fuel type, type and location of the kitchen, and the time spent near the kitchen while cooking are the most important determinants of
exposure across rural households.
Balasubramanian and Lee (2007,156248)
Study Design Case study of 3 rooms of 1 flat on the 8th floor, and "outside the home.'
May 12-23,2004
Singapore
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Findings
Residents of an urban area in a densely populated country.
NR
Time-activity logs identified tobacco smoking, cooking, household cleaning and general resident movements.
NR
NR
PM2.6
PM2.6
Indoor/outdoor ratios (I/O) suggest that chemicals such as chloride, sodium aluminum, cobalt, copper, iron, manganese, titanium vanadium,
zinc, elemental carbon were derived from the migration of outdoor particles (I/O <1 or ~ 1).
Barn etal. (2008,156252)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Measure indoor infiltration factor (FiaC) of PM2.6 from forest fires/wood smoke, effectiveness of high-efficiency particulate air (HEPA) filter
air cleaners in reducing indoor PM2.6, and to analyze the home determinants of Fia and air cleaner effectiveness (ACE).
2004-2005 (summer 2004 and 2005, winter 2004)
British Columbia, Canada
homes affected by either forest fire smoke or residential wood smoke
Pdr (Personal Data Ram) for ambient air sampling
Indoor home PM2.6
NR
Outdoor home PM2.6
NR
Use of HEPA filter air cleaners can dramatically reduce indoor PM2.6 concentrations. Number of windows and season predict indoor
infiltration Fia (p< 0.001).
Baxter et al. (2007, 092726)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Part of a prospective birth cohort study performed by the Asthma Coalition for Community, Environment, and Social Stress
(ACCESS)
2003-2005. Non-heating season- May to October; Heating season- December to March
Boston (urban)
Lower socio-economic status (SES) households
NR
NR
PM2.5 samples were collected with Harvard personal environmental monitors (PEM).; NO concentrations were measured using
Yanagisawa passive filter badges.
NR
PM2.5
PM2.5
EC
The authors' regression models indicated that PM2.5 was influenced less by local traffic but had significant indoor sources, while EC
was associated with local traffic and NO2 was associated with both traffic and indoor sources. However, local traffic was found to be
a larger contributor to indoor NO2 where traffic density is high and windows are opened, whereas indoor sources are a larger
contributor when traffic density is low or windows are closed. Similarly, traffic contributed up to 0.2 mg/m3 to indoor EC for homes
with open windows, with an insignificant contribution for homes where windows were closed.; Comparing models based on p-values
and using a Bayesian approach yielded similar results, with traffic density volume within a 50m buffer of a home and distance from
a designated truck route as important contributors to indoor levels of NO2 and EC, respectively. However, results from the Bayesian
approach also suggested a high degree of uncertainty in selecting the best model. The authors concluded that by utilizing public
databases and focused questionnaire data they could identify important predictors of indoor concentrations for multiple air
pollutants in a high-risk population.
July 2009
A-243
DRAFT-DO NOT CITE OR QUOTE

-------
Baxter et al. (2007, 092725)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Copollutant(s)
Primary Findings
Simultaneous indoor and outdoor samples taken in 43 low SES homes in heating and non-heating seasons. Homes were selected
from a prospective birth cohort study of asthma etiology (n = 25). Non-cohort homes were in similar neighborhoods (n = 18).
2003-2005
Boston, Massachusetts
Lower SES populations in urban areas
Home type, year built, tobacco smoke, opening windows, time spent cooking, use of candles or air freshener, cleaning activities, air
conditioner use.
NR
NR
PM2.5
NR
EC (rrv1 x 10-5); Ca (ng/m3); Fe (ng/m3); K (ng/m3); Si (ng/m3); Na (ng/m3); CI (ng/m3); Zn (ng/m3); S (ng/m3); V (ng/m3)
NO2
The effect of Indoor Sources may be more pronounced in high-density multi-unit dwellings. Cooking times, gas stoves, occupant
density and humidifiers contributed to indoor pollutants.
BeruBeetal. (2004,189731)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
6 homes in Wales and Cornwall were monitored four times per year, inside samples in the living areas and outside the home.
NR but < 2003
Wales and Cornwall, UK
urban, suburban, and rural homes
Tobacco smoke, pets, cleaning, traffic load
NR
NR
PM10 mass
NR
NR
There are greater masses of PM10 indoors, and that the composition of the indoor PM10 is controlled by outdoor sources and to a
lesser extent by indoor anthropogenic activities, except in the presence of tobacco smokers. The indoor and outdoor PM10 collected
was characterized as being a heterogeneous mixture of particles (soot, fibers, sea salt, smelter, gypsum, pollen and fungal spores).
Branis et al. (2005,156290)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Human exposure assessment in a university lecture hall
Oct. 8, 2001-Nov. 11,2001
Prague, Czech Republic
University students
NR
Presence of people identified as a source of course (PM2.5-10) particles; outdoor air identified as a source of indoor fine particles
(PMtoand PM2.5)
Harvard impactors with membrane Teflon filters
PM1, PM2.5, PM10
PM1, PM2.5, PM10
PM10
NR
Presence of people is an important source of coarse particles indoors; Outdoor air may be an important source of fine indoor
particles
Brauer et al. (2006, 090757)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Copollutant(s)
Primary Findings
Cohort study of otitis media and traffic related air pollution
Dec. 1997-Jan 1999
Netherlands and Munich, Germany
Children living near high traffic roads
0-2 yr
Environmental tobacco smoke at home, gas cooking, indoor moulds and dampness, number of siblings, breast-feeding, and pets
indoor.
NR
NR
NR
PM2.5; Light absorbing carbon
Light absorbing carbon
NO2
These findings indicate an association between exposure to traffic-related air pollutants and the incidence of otitis media.
July 2009
A-244
DRAFT-DO NOT CITE OR QUOTE

-------
Brauer et al. (2007, 090691)
Study Design
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Copollutant(s)
Primary Findings
Cohort study from birth to 5 yr Exposure obtained from stationery monitors identified as closest to birth home.
The Netherlands
Children
0-5 yr
NR
NR
NR
PM2.5 and Soot/filter absorbance
NO2
Adjusted odds ratios for wheeze, doctor-diagnosed asthma, ENT infections and flu indicated positive associations with air pollution.
No associations for eczema and bronchitis. Findings at age 4 confirm findings at age 2 in the cohort.
Brunekreef et al. (2005, 090486)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment
Winter and spring 1998-1999
Amsterdam and Helsinki
Elderly
50-84 yr
NR
Amsterdam Gillian with made to fit bags with belt withGK2.05 cyclone samplers 4L/min; Helsinki BGI with shoulder strap or
backpack with GK2.05 cyclone samplers 4L/min
PM2.5
PM2.5
PM2.5
Sulfate
In both cities personal and indoor PM2.5 were lower than highly correlated with outdoor concentrations. For most elements, personal
and indoor concentrations were also highly correlated with outdoor concentrations.
Cao et al. (2005,156321)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Size
Microenvironment Size
Components)
Primary Findings
Case study: 2 roadside homes (RS), 2 urban (UR), 2 rural (RU).
March-April 2004
Hong Kong, China
All
NR
NR
PM2.5
PM2.5
OC, EC
PM2.5 concentrations were roadside >urban >rural. Indoor PM2.5 has an avg of 24.4-36.8% OC and 3.6-6.9% EC.
Chakrabarti etal. (2004,147867)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
This is an evaluation of the active-flow pDR for PM2.5 against the D Attenuation Monitor (BAM) and the gravimetric pDR
NR
Los Angeles, California
NR
NR
NR
NR
NR
NR
PM2.5
NR
The personal pDR can be deployed as a personal monitor. The PM2.5 cyclone prevents larger particles from biasing the results.
Along with a wearable humidity and temperature monitor for correcting the readings, the results correlate highly with other methods.
The samples can be taken every 15 min to provide a more accurate picture of personal exposure in various settings.
Chang etal. (2007,156331)
Study Design	Panel study
Period	2003 to 2005
Location	Taipei County, Taipei
Population	Elderly people
July 2009
A-245
DRAFT-DO NOT CITE OR QUOTE

-------
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
53-75yr (median = 66.2 ± 6.5)
NR
Personal exposures to PM were measured simultaneously with ECG in real-time for twenty-four hours by using a personal dust
monitor (DUSTcheck portable dust monitor, model 1.108) which recorded 1-min mass concentrations of PMi, PM2.5, and PM10, as
well as ambient temperature and relative humidity. To measure subjects' personal PM exposures, all subjects were instructed to
keep the DUSTt check monitor with them at all times.; Details were reported previously (Chuang et al. 2005)
PM10; PM2.5-10; PM2.5; PM1-2.5; PM1
NR
NR
NR
Short-term and medium-term PM exposures were associated with the reduction of HRV in the elderly, with stronger effects found for
coarse particles in comparison with particles of other size ranges. In general, increase was observed with PM for H and the LF/HF
ratio, where the strongest significant effects on H were found at short-term intervals (1 -4 h) for PM2.5-10 and at medium-term duration
(5-8 h) for particles smaller than 2.5 Im in diameter. On the other hand, among the different-sized particles, PM2.5-10 exposures
showed the strongest significant association with decreases in time-domain (SDNN, r- MSSD) and frequency-domain parameters
(LF, HF) in most averaging Periods. Especially for the longer duration of 5-8 h, the strongest association was found for the 6-h
moving average of PM2.5-10 exposures.
Charron et al. (2007,156333)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Copollutant(s)
Primary Flndlng(s)
Environmental PM exposure assessment. In this article, a total of 185 days with daily PM10 concentrations exceeding the limit value
of 50 pg/m3 measured between January 2002 and December 2004 are discussed.
January 2002 and December 2004
Marylebone Road, Westminster, London
NR
NR
NR
NR
NR
PM2.5 PM10
NO3-; SO42-; OC; EC
NOx; CO
The regional background was the largest contributor to PM10 concentrations measured at Marylebone Road between January 2002
and December 2004
Chillrud etal. (2004,
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
054799)
Repeated measures on a cohort of high school students in New York City
Summer and winter of 1999 (eight weeks each)
Manhattan, Bronx, Queens, Brooklyn, NY
Persons traveling the subway
14-18 yr
No
Sampling packs carried by subjects
PM2.5
PM2.5; Home indoor and home outdoor
PM2.5. Urban fixed-site and upwind fixed site operated for three consecutive 48-h periods each week.
Elemental iron, manganese, and chromium are reported in this study out of 28 elements sampled.
Personal samples had significantly higher concentration of iron, manganese, and chromium than home indoor and ambient samples.
The ratios of Fe (ng/pg of PM2.5) vs Mn (pg/pg PM2.5) showed personal samples to be twice the ratio for crustal material. Similarly
for the Cr/Mn ratio. The ratios and strong correlations between pairs of elements suggested steel dust as the source. Time-activity
data suggested subways as a source of the elevated personal metal levels.
Chuang et al. (2005,087989)
Study Design
Period
Location
Population
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Panel Study
Taipei, Taiwan
Individuals with CHD, prehypertension, and hypertension
No
Yes, a technician carrying a DUSTcheck; monitor accompanied each patient
PMo.3-1.Ol PM1.0-2.5; PM2.5-10
NR
NR
NR
HRV reduction in susceptible population was associated with PM0.3-1.0 but was not; associated with either PM1.0-2.5 or PM2.5-10.;
PM0.3-1.0 exposures at 1- to 4-h moving averages were associated with SDNN and r-MSSD in both cardiac and hypertensive;
patients. For an interquartile increase in PM0.3-1.0, there were 1.49-4.88% decreases in SDNN and 2.73-8.25% decreases in r-
July 2009
A-246
DRAFT-DO NOT CITE OR QUOTE

-------
MSSD. PMo.3-1.0 exposures were also associated with decreases in LF; and HF for hypertensive patients at 1- to 3-h moving
averages except for cardiac patients at moving averages of 2 or 3 h.
Cohen at al (2004,056909)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Field evaluation study to test performance of new technology to measure number concentration of acidic ultrafine particles (UFP)
July 1999-September 2000
New York City and nearby suburban location
4 outdoor rural sites and 1 indoor rural site (cafeteria) in Tuxedo, NY. 1 suburban residential site in Newburgh, NY. 1 outdoor urban
site in New York City
NR
NR
No personal exposure assessment was conducted.
NR
Ultrafine (UFP)
Ultrafine (UFP)
Acidic UFP, Hydrogen ions, sulfate ions, ammonium ions
Iron nanofilm detectors can be used with confidence under a range of ambient conditions. Concentrations of UFP determined by
atomic force microscopy analysis of detectors in MOI-EAS and UDM appeared to underestimate number concentrations of total
UFP and [missing items?]
Connell et al. (2005,089458)
Study Design	Times-series
Period	May 2000-May 2002
Location	Steubenville, Ohio = ST; Saint Vincent College, Latrobe, PA (eastern site) = E; Tomlinson Run State Park, New Manchester, WV
(northern site) = N; Hopedale, OH (western site) = W; Jesuit Univ., Wheeling, WV (southern site) = S
Population	NR
Age Groups	No
Indoor Source	NR
Personal Method	NR
Personal Size	NR
Ambient Size	PM10 & PM2.5
Components)	Ammonium, sulfate, nitrate, chloride, and 21 elements, elemental carbon and organic carbon.
Copollutant(s)	SOx, NOx, Co, and O3.
Primary Findings	The average PM2.5 in Steubenville was 18.4 pg/m3, 3.4 pg/m3 above the annual PM2.5 NAAQS.
Conner and Williams (2004,156364)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Finding(s)
This is part of the EPA Baltimore PM Study of the Elderly.
July-August, 1998
Towson, Maryland
65+ adults
65+yr
Personal sampling devices (PEM)
PM2.5
PM2.5
NR
NR
A greater variety of particles was observed in the personal samples compared to the fixed-location apartment samples.
Cortez-Lugo et al. (2008,156368)
Study Design	Cohort
Period	Feb-Nov 2000
Location	Mexico City, Mexico
Population	Ambulatory adults with moderate to severe COPD, active smokers excluded
Age Groups	Adults
Indoor Source	carpeting, aerosol sprays used, boiler use and location, animals, mold, tobacco smoking, windows closed
Personal Method	Personal pumps with 37-mm TefIon filters, flow rate 4 l/min in a bag with shoulder strap. The impactor was near the breathing zone
Personal Size	PM2.5
Microenvironment Size	PM2.5&PM10
Ambient Size	PM2.5&PM10
Components)	NR
Primary Findings	Indoor PM2.5 concentrations explained 40% of the variability of personal exposure.; The best predictors of personal exposure were
indoor contact with animals (12%, 1 -25), mold (27%, 11 -48), being present during cooking (27,12-43), and aerosol use (17%, 4-
31).
July 2009	A-247	DRAFT - DO NOT CITE OR QUOTE

-------
Crist et al. (2008,156372)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Indoor, outdoor, and personal monitoring
January 1999-August 2000
Ohio
Fourth & fifth-grade children
9-11 yr old
Filter, portable pump
Filter, PM2.5
Indoor school; Filter, PM2.5
Outdoor school; Filter, PM2.5
NR
Cyrys etal. (2003, 042232)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment, source apportionment of urban aerosol
September 1,1995-December 21,1998
Erfurt, Germany
Urban Populations
NR
NR
No personal exposure assessment was conducted
NR
NR
Ultrafine (UFP, 0.01-2.5 pm), PM2.5, PM10
Si, Al, Ti, Ca, Fe, Cr, Mg, Na, K, Mn, Ni, V, Co, Sc, Cu, Zn, Pb, Br, S
Low correlation between UFP number concentration and fine particle mass and differences in their diurnal patterns suggest that
different sources contribute to particles in the 2 size ranges. Elements Si, Al, Ti and Ca were highly correlated and had low e
Cyrys et al. (2006,156376)
Study Design
Period
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Exposure assessment, evaluation of sampling methodologies
Sept 1 2000-August 31, 2001
Indoor
No personal monitoring was conducted
NR
NR
NR
Restricted sampling scheme (covering 23% of study period) was able to estimate reliable annual and winter averages in Erfurt,
Germany. Daily PM2.5 means measured by EPA-WINS were higher than those measured by HI, but differences between samplers
were small.
Delfina et al. (2004)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Copollutant(s)
Primary Findings
Panel study with repeated measures
Sep-Oct 1999 or Apr-Jun 2000
Alpine, California
Children
9-19 yr
No
Personal dataRAM (pDR) carried at waist level using a fanny pack, shoulder harness, or vest.
PM2.5 (approximate); 0.1-10 range
PM10 & PM2.5; measured immediately outside the house and in the living room of the home.
PM10
O3 and NO2 measured at central site
Percent predicted FEV1 was inversely associated with personal exposure to fine particles. Also with indoor, outdoor and central site
gravimetric PM2.4, PM10, and with hourly TEOM PM10.
Delfino etal. (2006,090745)
Study Design	Cohort. Measured daily expired NO (FeNO)
Period	Aug-Dec 2003
Location	Riverside and Whittier, California
Population	Children with asthma exacerbations in previous 12 months, non-smokers, non-smoking households
Age Groups	9-18 yr
Indoor Source	No
July 2009
A-248
DRAFT-DO NOT CITE OR QUOTE

-------
Personal Method Wore a backpack during waking hours for PM2.5, EC and OC, NO2, temperature, and relative humidity. Exhaled air collected in Mylar
bags to analyze for NO.
24-h PM2.5; 1 -h max PM2.5; 8-h max PM2.5; 24-h NO2
NR
24-h PM2.5; 24-h PM10; 8-h max O3; 8-h max NO2; 24-h NO2; 8-h max CO
24-h PM2.5 EC; 24-h PM2.5 OC
PM associations with airway inflammation in asthmatics may be missed using ambient particle mass.; The strongest positive
associations were between eNO and 2-day avg pollutant concentrations. Per IQR increases 1.1 ppb FeNO/24 pg/m3 personal
PM2.5.; 0.7 ppb FeNO/O.6 pg/m3 personal EC; 1.6 ppb FeNO /17 ppb personal NO2; Ambient PM2.5 and personal and ambient EC
were significant only when subjects were taking inhaled corticosteroids.; Subjects taking both inhaled steroids and antileukotrienes
had no significant associations.; Distributed lag models showed personal PM2.5 in the preceding 5 h was associated with FeNO.
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Demokritou et al. (2002,156393)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment, evaluation of newly developed personal cascase impactor
NR
Laboratory chamber
Newly developed personal PM sampler
NR
NR
No personal exposure assessment was conducted
NR
9.6-20 pm, 2.6-9.6 pm, 1.0-2.6, 0.5-1.0
NR
NR
The first stage showed excellent separation of particles larger than 9.6 pm from the airstream. In the second stage, for particles
above 4.0 pm, the collection efficiency was greater than 95%. In the third stage, the collection efficiency for particles a
Dermentzoglou et al. (2003,156395)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Sampled rooms in 1 apartment for 2 h and compared to ambient air.
NR, but winter < 2003
Thessaloniki, Greece
Urban apartment dwellers
NR
Woodburing fireplace, cigarette smoking, cooking fish, chicken, sausage & potato.
NR
NR
PM3.0
NR
PAHs, Pb, Cd, Cu, Mn, Ni, VAs
Smoking could be associated with the highest indoor concentration of total and carcinogenic PAHs.The highest level of pyrens, and
phnanthrenes were during fish frying. Smoking and fish frying had significant effect on Cd in indoor air, while woodburning had no
effect of PAH or heavy metal levels.
Diapouli et al. (2007
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
,156397)
Exposure assessment. Sampling of schools, residence, private vehicle
Schools- 11/2003-02/2004 and 10/2004-12/2004.; Residence-10/2004; Vehicle- 10/204-12/2004
Athens, Greece
Primary school children
NR
NR
Handheld portable Condensation Particle Counters (TSI, Model 3007) were used for all sampling locations. Primary schools indoor
measurements were primarily conducted inside classrooms, at table height. However, at three of the schools, rooms of different
uses were selected. These included a teachers' office (where smoking was permitted), a computer day lab (used by students only
part of the day), and a library and gymnasium (where intense activity took place almost all day long). Outdoor measurements took
place in the yard of each school. Residence samples were taken in a bedroom at breathing height and on the terrace, for indoor
and outdoor samples, respectively. In-vehicle samples were taken by placing the CPC 3700 on the passenger seat while the
vehicle drove along predetermined routes.
NR
0.01-1 pm
0.01-1 pm
NR
The results showed that children attending primary schools in the Athens area are exposed to significant PM concentration levels,
both indoors and outdoors, Vehicular emissions seem to be a major contributor to the measured outdoor concentration levels at the
studied sites. Indoor PM concentrations appeared to be influenced by both vehicular emissions and indoor sources including
July 2009
A-249
DRAFT-DO NOT CITE OR QUOTE

-------
cleaning activities, smoking, a high number of people in relation to room volume and furniture material (i.e., carpet.).; UFPs
concentrations diurnal variation, both outside the schools and the residence, supports the close relation of UFPs levels with traffic
density. Indoor concentrations within schools exhibited variability during the school day only when there were significant changes in
room occupancy. 24-h variation of indoor concentrations at the residence followed quite well (R2 = 0.89) the outdoor one, with a
delay of 1 -h or less.
Diapouli et al. (2008,190893)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Indoor, outdoor air monitoring of PM. To determine children exposure in school environment. To evaluate relationship between
indoor and outdoor levels.
Athens, Greece
Primary schools
NR
Indoor PM2.5 and PM10presence of children and activities of children in classrooms PMivehicles
Harvard PEM, Teflon filters Dust Trak Condensation particle counter
NR
Weight concentration	PM2.5, PM10 Number concentration PM1
Weight concentration	PM2.5, PM10 Number concentration PM1
N03-, SO42-
High levels of PM10 and PM2.5 measured indoors and outdoors. PM10 more variable spatially than PM2.5. Indoor/Outdoor ratio for
PM10 and PM2.5 close to 1 at almost all sites. Ratio of PMismaller than 1 in all cases. Vehicular traffic presumed to be the main
source of PM1. Indoor PM2.5 and PM10 levels dependent on the amount of activity in classroom and outdoor levels. Indoor SO42-
concentrations strongly associated with outdoor levels. Result suggests that SO42- can be used as a proper surrogate for indoor PM
of outdoor origin.
Dills et al. (2006,156402)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Dose-response, variability, and applicability of methoxyphenols as biomarkers in a realistic exposure situation mimicking indoor
open fire cooking
August. Year not specified
Seattle, WA
Non-smokers exposed to woodsmoke
20-65 yr
Not required. Subjects exposed to wood smoke. One subject fitted with an integrating nephelometer for a continuous estimate of
particle exposure, and a continuous monitor for CO, CO2, and temperature. For 24-h prior to the exposure, subjects collected all
urine voids at 'will in separate containers for a baseline ofmethoxyphenol excretion. Subjects then collected all urine voids at will for
48 h postexposure for measuring wood smoke biomarker elimination.
Air collected at breathing level using Harvard Personal Environmental Monitor for PM2.5 (HPEM2.5)
PM2.5
NA. No microenvironmental studied
NR
22 methoxyphenols, levoglucosan, and 17 polynuclear hydrocarbons for personal filter samples and urine samples.
According to the authors "A 12-h avg creatinine-adjusted methoxyphenol concentration is a practical metric for the biomarker
exposure to woodsmoke." Propylguaiacol, syringol, methylsyringol, ethylsyringol, and ropylsyringol had peak urinary concentrations
after the woodsmoke exposure.
Dimitroulopoulou et al. (2006, 090302)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment. Development of a model to predict indoor PM2.5 concentrations under various emissions scenarios
1997-1999
5 sites in the UK Harwell, Birmingham East, Bradford, Bloomsbury, Marylebone Rd.
Indoor environments within homes
NR
Smoking, cooking
No personal exposure assessment was conducted.
NR
PM10, PM2.5
PM10, PM2.5
NR
Modeled mean concentrations were most sensitive to variation in outdoor concentrations, air exchange rates, and deposition
velocity. Modeled peak concentrations were most sensitive to variations in emissions rates and room size. Cooking activities incre
Ebelt et al. (2005, 056907)
Study Design Personal exposure assessment related to health outcomes for a sensitive sub-population
Period Summer 1998
Location Vancouver, British Columbia, Canada
July 2009
A-250
DRAFT-DO NOT CITE OR QUOTE

-------
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
16 persons who had COPD
Mean subject age 74 yr, Range 54 to 86
Separated total personal exposure into "ambient" and "non-ambient" based on sulfate results and modeling.
"Subjects wore a PM2.5 sampler that provided 24-h integrated personal PM2.5 exposure data." No other details reported.
PM2.5
"ambient exposure" PM2.5, PM10, PM2.5-10; "non-ambient exposure" PM2.5
PM2.5, PM10, PM2.5-10
Ambient sulfate,; ambient non-sulfate,; personal sulfate,; personal ambient non-sulfate
Ambient exposures and (to a lesser extent) ambient concentrations were associated with health outcomes; total and nonambient
particle exposures were not.
Farmer et al. (2003,089017)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
case control
12 months
Prague, Czech Republic (2 sites); Kosice, Slovak republic; Sofia, Bulgaria
Policeman and Busdrivers usually working through busy streets in 8-1 Oh shifts and a Control Population.
Variable, range not stated
NR
Personal Monitoring Devices; Blood and Urine Samples; Stationary Versatile Air Pollution Samplers (VAPS)
PM10
NR
PM10; PM2.5 (not reported)
EOM; EOM2; B[a]P; c-PAHs
Extractable organic matter (EOM) per PM10 was at least 2-fold higher in winter than in summer, and c-PAHs over 10-fold higher in
winter than in summer. Personal exposure to B[a]P and to total c-PAHs in Prague ca. was 2-fold higher in the exposed group
compared to the control group, in Kosice ca. 3-fold higher, and in Sofia ca. 2.5-fold higher.
Farmer et al. (2003,089017)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
Molecular epidemiology studies of carcinogenic environmental pollutants, particularly PAHs
NR
Prague, Czech Rep.; Kosice, Slovak Rep.; Sofia, Bulgaria
Policemen and bus drivers
NR
No
"Personal monitors for PM10"; Extraction by dichloromethane and analyzed for PAH by HPLC with fluorimetric detection.
PM10
NR
PM10; Extractable organic material (EOM); B[a]P; cPAHs
Benzo[a]pyrene (B[a]P); Carcinogenic polycyclic aromatic hydrocarbons (cPAHs)
Personal exposure to B[a]P and to total carcinogenic PAHs in Prague was twofold higher in the exposed group compared to
controls, in Kosice three fold higher, and in Sofia 2.5-fold higher.
Ferro etal. (2004, 055387)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
Case study, 1 home
Redwood City, California
NR
NR
NR
Co-located real-time particle counters and integrated filter samplers (Met-One Model 237B) were used to measure personal (PEM),
indoor (SIM) and outdoor (SAM) PM concentrations. The PEM was attached to a backpack frame and worn by the investigator
while performing prescribed activities. The SIM was attached to a six foot step-ladder with the intake at breathing height. The SAM
was located under a two-sided roofed shed in the backyard of the home with the filter samplers supported by a metal stand and the
real-time particle counters sitting on a table.
PM5
PM2.5; PMs
PM2.5; PMs
NR
The results of this study indicate that house dust resuspended from a range of human activities increases personal PM
concentrations and this resuspension effect significantly contributes to the personal cloud. The results of this study also suggest
that normal human activities that resuspend house dust may contribute significantly to the strong correlations found between
personal exposure and indoor PM concentrations in previous studies. The PEM/SIM ratios for human activity presented in this
paper are also in the range of those reported by previous studies.
July 2009
A-251
DRAFT-DO NOT CITE OR QUOTE

-------
Ferro et al. (2004, 055676)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Modeling of PM source strengths from human activities
April 2000
Redwood City, CA
Residential home occupants
NR
Yes. Vacuuming resulted in the maximum PM2.5 source strength while two persons walking around and sitting on furniture resulted in
the maximum PM5 source strength.
Met-One Model 237B laser particle counters (2.8 Lpm); AIHL design cyclone samplers with filters (21 and 11 Lpm for PM2.5 and PM5
respectively)
PM2.5, PM5
PM2.5, PM5
PM2.5, PM5
NR
The source strengths were found to be a function of the number of persons performing the activity, the vigor of the activity, and the
type of flooring. Proximity to the source played a large role in the observed differences between indoor concentration and personal
exposure.
Fromme et al. (2007
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Copollutant(s)
Primary Findings
,156453)
Explorative analysis
Winter session December 2004-March 2005; summer session May to July 2005
Munich (and surrounding districts), Germany
Primary and secondary school children
NR
NR
Filter-based measurements of PM2.5 in the classrooms were conducted with a medium volume sampler using a flow controlled pump
(Derenda, Teltow, Germany). The sample inlet was a PM2.5 sampler, having a 50% collection efficiency for particles with a 2.5 mm
aerodynamic diameter. A Munktell 47mm binder free glass fibre filter with a pore size of 2 mm was used. Continuous
measurements of PM (e.g. PM10, PM4, PM2.5) were also done using an optical laser aerosol spectrometer (LAS) (Dust monitor
1.108). A TSI model 3034 scanning mobility particle sizer (SMPS) (TSI Inc., Shoreview, MN, USA) was used to measure particle
number concentrations (PNC) for a discrete size distribution of aerosols. Indoor carbon dioxide was measured using a continuously
monitoring infrared sensor (Testo 445).; The sampling and measuring position in the classroom was opposite to the black board,
about one meter above floor level, the level at which the pupils would normally inhale.
NR
PM2.5; PM4; PM10
PM10, PN (particle number)
CO2
Results clearly showed that exposure to PM in school is high. This study identified parameters correlated with increased
concentrations of PM such as high CO2 concentrations and low class level. Strong seasonal variability was observed, with air
quality being particularly poor in winter. The influence of season on PM concentrations observed has been reported before from the
US (Keeler et al., 2002). This difference is most likely due to the different ventilation practice in summer and winter. Further
parameters correlated with increased concentrations of PM were small room size, high number of occupants, high CO2
concentrations and low class level. No significant differences between PM and values in classrooms with carpets and those with
hard surface floorings were reported. The number of fine and ultra fine particles measured in classrooms was in the same range or
lower as the results from residences or outdoor monitoring sites (reported in similar studies) and show little variation.
Gadkari and Pervez
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
(2007,156459)
Evaluation of relative source contribution estimates of various routes of personal RPM in different urban residential environments.
Summer 2004 (March 15-June 15)
Chattisgarh, India
All likely. Not specified
21 -61 yr, average age 40 ± 15 yr
No
Personal respirable dust samplers (RDS) with GFF
RPM
NR
RPM
Fe, Ca, Mg, Na K, Cd, Hg, Ni, Cr, Zn, As, Pb, Mn and Li
Authors concluded that "(1) indoor activities and poor ventilation qualities are responsible for major portion of high level of indoor
RPM, (2) majority of personal RPM is greatly correlated with residential indoor RPM, (3) time-activity diary of individuals has much
impact on relationship investigations of their personal RPM with their respective indoor and ambient-outdoor RPM levels; as
reported in earlier reports and (4) residential indoors, local road-traffic and soil-borne RPMs are the dominating routes of personal
exposure compared to ambient outdoor RPM levels."
July 2009
A-252
DRAFT-DO NOT CITE OR QUOTE

-------
Gauvin et al. (2002, 034893)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Fine particle exposure assessment for children in French urban environments, part of VESTA study
March 1998-December 2000
Paris, Grenoble, Toulouse, France
Children aged 8-14 yr
Yes-ETS from mother, rodents at home.
SKC pump 4 Lpm with PM2.5 inlet and 37 mm, 2 micron Teflon filter
PM2.5
NR
PM10
NR
The final model explains 36% of the between subjects variance in PM2.5 exposure, with ETS contributing more than a third to this.
Geyh etal. (2004,156467)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Finding(s)
An evaluation of a modified personal monitoring pump (PMASS)
NR
Fresno, CA and Baltimore, Maryland
Persons for personal sampling
NR
PMASS and PEM "adjusted mass measurements downward by 22% to eliminate measurement bias with the Harvard impactor."
Particle mass
Particle mass
NR
EC, OC, nitrate
PMASS measurements of mass showed a significant bias of -24% compared to the reference sampler.; For microenvironmental
sampling the PMASS for mass concentrations again had a bias of -34%, but for EC, OC and nitrate were much closer but still with a
bias of 6.6-17.5%.
Geyh etal. (2005,186949)
Study Design
Period
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Copollutant(s)
Primary Findings
Exposure assessment- representative population (WTC truck drivers) study
October 2001 and April 2002
Indoor
Each driver was given two monitors consisting of small portable pumps and battery packs worn at the waist, and sampling cartridges
worn on the shoulder within the breathing zone. Monitoring was conducted across a work shift on all days of the week during both
day and night shifts (6:00am to 6:00pm, and 6:00pm to 6:00am, respectively). Drivers were asked to wear their monitors at all
times. If they were planning to sleep in their trucks, they were told they could remove the pumps from their belts and place them on
the seat beside them.; Area monitoring was also conducted at the site at four locations around the perimeter of the disaster site on
streets approximately representing the north, east, and south/southwest boundaries of the debris field. In addition, monitoring was
also conducted directly in the debris pile for several days. The set of monitors were hung at head height either from scaffolding,
from a chain link fence, or placed on supports, such as tank cages in the debris pile.; Sampling pumps used for particle sampling
were either SKC Universal pumps(model 223-PCXR4), BGI personal sampling pumps (model 400S0, or ELF personal sampling
pumps (MSA Inc). VOC sampling was conducted with SKC pocket pumps (Personal Packet Pump 210 series).
TD; PM10; PM2.5
NR
TD; PM10; PM2.5
EC; OC
VOC(s)
During October, the median personal exposure to TD was 346 pg/m3. The maximum area concentration 1742 pg/m3, was found in
the middle of the debris. The maximum TD concentration found at the perimeter was 392 pg/m3 implying a strong concentration
gradient from the middle of debris outward. PM2.5 /PM10 ratios ranged from 23% to 100% suggesting significant fire activity during
some of the sampled shifts. During April, the median personal exposure to TD was 144 pg/m3, and the highest area concentration,
195 pg/m3, was found at the perimeter. Although the overall concentrations on PM at the site were significantly lower in April, the
relative contributions of fine particles to the PM10, and EC and OC to the TD were similar. During both months, volatile organic
compounds concentrations were low. Comparison of recorded EC and OC values from October 2001 and April 2002 with previous
studies suggests that the primary source of exposure to EC for the WTC truck drivers was emissions from their own vehicles.
Goyal and Sidhartha (2004,156487)
Study Design	Actual air monitoring measurements are compared with a model.
Period	1998-1999
Location	Delhi, India
Population	Residents near coal-fired power plants (BTPS)
Age Groups	NR
Indoor Source	NR
July 2009
A-253
DRAFT-DO NOT CITE OR QUOTE

-------
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
NR
NR
NR
Suspended PM (SPM)
NR
Measured SPM values are higher during the day than at night. This is because "point sources dominate during the daytime
convective conditions. At night the small depth of the nocturnal boundary layer prevents the dispersion of the pollutants from the
elevated point source to reach the surface. Convective turbulence breaks up the surface-based inversion and the fumigation
process leads to an increased contribution from the point sources."
Graney et al. (2004, 053756)
Study Design
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
The study was designed to assess the trace metal quantification abilities of several analytical methods to measure the total as well
as soluble amounts of metals with PM2.5 collected from indoor and PM samples. (X-ray fluorescence and instrumental neutron
activation analysis)
Retirement facility in Towson, Maryland
Retirement facility with subjects who spent 94% of their time indoors
Mean age = 84 yr
No, this was not the objective of the study
Measured using personal exposure monitors (MSP Inc) with nozzle to remove particles >4 pg/m3
PM2.5
PM2.5
NR
42 elements were analyzed for in the PM2.5 samples collected from personal and well as indoor samples
1) Most of the extractable components of the metals were in a water-soluble form suggesting a high potential for bioavailability of
elements from respiratory exposure to PM2.5. 2) based on comparison of trace metals in central indoor site vs. PE samples, resident
activities result in exposure to higher cone of soluble trace metals.
Guo et al. (2004,156506)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Human exposure assessment
Sept. 2001-Jan. 2002
Hong Kong
Shoppers at food markets
NR
Yes. Elevated concentrations of PM at three markets probably due to outdoor particulates from vehicular exhaust. Poultry stalls in
the markets had higher PM10 due to live chickens.
TSI Dust Trak Model 8520.1 n some locations an Anderson Hi-Vol sampler with filters weighed by electronic microbalance were used
to calibrate the Dust Trak.
PM10
PM10
PM10
NR
Indoor PM10 concentrations at the markets were generally below Hong Kong Indoor Air Quality Objectives. Outdoor sources were
dominant at the five markets, with elevated levels at three markets due to vehicular exhaust.
Hanninen et al. (2004,056812)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
EXPOLIS-human exposure assessment
1996-2000
Athens, Greece; Basle, Switzerland; Helsinki, Finland; Prague, Czech Republic
Residential homes
NR
Yes. Sources identified in statistical analysis wooden building material, use of stove other than electric, PVC floors, attached garage
Pump & filter with gravimetric analysis; Elemental composition using energy dispersive X-ray fluorescence
PM2.5
NR
PM2.5
PM2.5 -bound sulfur
Associated with indoor concentration wooden building material, city, building age, floor of residence (ground, 1st, etc.), and use of
stove other than electric.
Haverinen-Shaughnessy et al. (2007,156526)
Study Design Cross-sectional
Period Winter, year not reported
Location Eastern Sweden
July 2009
A-2 54
DRAFT-DO NOT CITE OR QUOTE

-------
Population
Age Groups
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
Elementary school teachers
NR
Button inhalable aerosol samplers
Particle mass
Particle mass
NR
Absorbance coefficient/m x 10-5; Total fungi (spores/m3); Total bacteria (cells/m3); Viable fungi MEA (CFU/m3); Viable fungi DG18
(CFU/m3); Viable bacteria (CFU/m3)
The recall period of 7 days provided the most reliable data for health effect assessment. Both personal exposure and concentrations
of pollutants at home were more frequently associated with health symptoms than work exposures.
Hazenkamp-von Arx et al. (2003,136487)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
Exposure assessment
November 2000-February 2001
21 European cities Antwerp City, Antwerp South, Albacete, Barcelona, Basel, Erfurt, Galdakao, Grenoble, Goteborg, Huelva,
Ipswich, Norwich, Ovledo, Pavia, Paris, Reykjavik, Tartu, Turin, Umea, Uppsala, Verona
European urban environments
NR
NR
No personal exposure assessment was conducted
NR
NR
NR
NR
Winter mean PM2.5 concentrations were lowest in Iceland and Sweden and highest in Northern Italy (Turin, Verona). Cities also
varied in daily concentrations. Geographical differences may be explained by differences in emissions (proximity of monitor to
traffic.
Henderson et al. (2007, 090675)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Copollutant(s)
Primary Findings
Land use regression was employed to model oxides of nitrogen and fine particulates using two measures of traffic (road length and
vehicle density)
Sampling was conducted from Feb 24 through Mar 14 and Sep 8 through Sep 26, 2003
Vancouver, British Columbia, Canada
NR
NA
NR
Personal monitoring was not conducted. Ambient fine particles were collected on PTFE filters using Harvard Impactors. Flow rate
was 4 L/min Absorption coefficients were also calculated
NR
NR
PM2.5
NO, NO2
Adjusted R2 for the linear regression models predicting NO, NO2, PM2.5, and ABS from fifty-five variables describing each sampling
site ranged from 0.39 to 0.62. The resulting maps show the distribution of NO to be more heterogeneous than that of NO2,
supporting the usefulness of land use regression for assessing spatial patterns of traffic-related pollution
Hertel et al. (2008,156543)
Study Design
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
Exposure assessment
Denmark
Bicycle commuters
NR
No
NR
PM2.5, PM10
NR
NR
NO2
It is possible to significantly reduce the accumulated air pollution exposure during the daily bicycle route between home and work by
following the low exposure route. Travelling outside the rush hour time periods significantly reduced the accumulated air pollution
exposure along the routes through the city.
Ho et al. (2004,056804)
Study Design Human exposure assessment
July 2009
A-255
DRAFT-DO NOT CITE OR QUOTE

-------
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
25 Sept. 2002 to 8 March 2003
Hong Kong
Occupied buildings located near major roadways
NR
Yes. Regression of indoor versus outdoor concentrations of OC and EC revealed an indoor source of OC not present for EC,
presumably due to such activities of cooking, smoking, and cleaning.
Co-located mini-volume samplers (flow rate 5 L/min) and Partisol model 2000 sampler with 2.5 micron inlet. All samples on 47 mm
Whatman quartz microfiber filters, weighed on an electronic microbalance. Analyzed for OC and EC using DRI Model 2001
Thermal/Optical Carbon Analyzer.
PM2.5
NR
PM2.5
OC; EC; OM; TCA
The major source of indoor EC, OC, and PM2.5 appears to be penetration of outdoor air, with a much greater attenuation in
mechanically ventilated buildings.
Hoek etal. (2008,156554)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment, characterizing indoor/outdoor particle relationships
October 2002-March 2004
4 European cities Amsterdam, Athens, Birmingham, Helsinki
Urban populations
NR
Smoking, candle burning, cooking/frying
No personal exposure assessment was conducted
NR
PM10, PM2.5, PMio-2.5, Ultrafine (UFP)
PM10, PM2.5, PMio-2.5, Ultrafine (UFP)
soot, sulfate
Correlation between 24-avg [should this be 24-h avg?] central site and indoor concentrations was lower for UFP than for PM2.5, soot,
or sulfate, probably related to greater losses during infiltration due to smaller particle size. Infiltration factors for UFP and PM2.5 were
low.
Holguin et al. (2003,057326)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Longitudinal analysis (repeated measures) of local PM2.5 and biological markers of cardiovascular dysregulation
3 months (Feb 8-Apr 30, 2000)
Mexico City, Mexico
Elderly residents of a nursing home (non-smokers)
60-96
Sources of indoor PM concentrations may be idling buses parked for a few hours close to living areas at least 3 times per week
Mini-vol portable air samplers operating at 4 l/min used to monitor outdoor and indoor PM2.5 concentrations at a nursing home.
Gravimetric analysis of filters. Personal exposure calculated using time-weighted averages of outdoor and indoor concentration
NR
NR
NR
NR
Increases in personal PM2.5 concentrations were associated with significant decreases in the high-frequency component of heart
rate variability (HRV-hF) among elderly. Associations remained significant after adjusting for ozone; Indoor and outdoor PM2.5
Hopkeetal. (2003,095544)
Study Design Epidemiology-Exposure study
Period 26 July to 22 August 1998
Location Retirement facility in Towson, MD
Population "A potentially susceptible elderly subpopulation"
Age Groups Mean age of 84
Indoor Source Ammonium sulfate and ammonium nitrate, secondary sulfate, OC, and motor vehicle exhaust
Personal Method Personal exposure samples were collected on 37mm Teflon filters using inertial impactor PEM in the breathing zone of the subjects.
A "scalper" (MSP, PEM-019) nozzle was used on the PEM to exclude particles >4mm in order to reduce the potential of
overloading the impactor. Centralized indoor sampling was conducted in an unoccupied apartment on the fifth floor of the retirement
facility (central indoor). The windows of the apartment were kept closed and the front door was kept open to the common hallway
with a small fan providing active air exchange. Residential outdoor sampling at the retirement facility was conducted from the
rooftop of an attached three-story nursing care facility (outdoor). PM2.5 measurements at an ambient site in Towson, MD were made
on the roof of a sampling shelter approximately four meters off the ground (community). Daily community, outdoor, and central
indoor PM2.5 samples were collected with VAPS samplers.
Personal Size PM2.5
July 2009
A-256
DRAFT-DO NOT CITE OR QUOTE

-------
Microenvironment Size PM2.5
Ambient Size PM2.5
Components) NR
Primary Findings VAPS and PEM data from the BPMEES were separately analyzed by different receptor models. These two approaches were
complementary and allowed for evaluation of all of the available data. A three-way analysis of the VAPS data provided four sources
of PM2.5, nitrate-sulfate, sulfate, OC, and MV exhaust. The largest contribution to the community, outdoor, and central indoor
sampling locations was the sulfate source. Infiltration of the sources varied depending on the source and ranged from 38% to 4%
for the Sulfate, and Nitrate-Sulfate sources, respectively. In addition, MV exhaust had a penetration rate similar to Sulfate (32%).
The OC source had little variability compared to the other sources and contributed approximately 8% of the community and outdoor
PM and 18% of the central indoor PM.; The PEM data were analyzed using a complex model with a target for soil that included
factors that are common to all of the types of samples (external factors) and factors that only apply to the data from the individual
and apartment samples (internal factors). From these results, the impact of outdoor sources and indoor sources on indoor
concentrations were assessed. The identified external factors were sulfate, soil, and an unknown factor. Internal factors were
identified as gypsum or wall board, personal care products, and a factor representing variability not explained by the other indoor
sources. The latter factor had a composition similar to outdoor particulate matter and explained 36% of the personal exposure.
External factors contributed 63% to personal exposure with the largest contribution from sulfate (48%).
Jacquemin et al. (2007,156600)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Assessment of relationship between outdoor and personal concentrations of PM25] absorbance and sulfur among post-myocardial
infarction patients
January 2004-June 2004
Barcelona, Spain
Survivors of a myocardial infarction exposed to environmental tobacco smoke (ETS)
n = 38 (males 32 (84%), age over or equal to 65 [something missing here] 15 (39%)
Not identified in this study. Results from other studies discussed.
Personal samplers (BGI GK2.05 cyclones and battery operated BGI AFC400S pumps)
PM2.5
NA
PM2.5
Sulfur (S)
Authors suggest that "outdoor measurements of absorbance and sulfur can be used to estimate both the daily variation and levels of
personal exposures also in Southern European countries, especially when exposure to ETS has been taken into account. For PM2.5,
indoor sources need to be carefully considered."
Janssen et al. (2005, 088692)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Panel Study
Amsterdam 11 /2/1998-6/18/1999; Helsinki 11/1/1998-4/30/1999
Amsterdam, The Netherlands; Helsinki, Finland
Elderly Cardiovascular Patients
50-84 yr
No
Personal PM2.5GK2.05; cyclones; indoor & outdoor Harvard Impactors; Reflectance EEL 43 reflectometers; Elemental Composition
Tracor Spectrace 5000 ED-XRF system
PM2.5
PM2.5
PM2.5
Estimated Elemental Carbon (Abs); Elemental composition of a subset of personal, indoor and outdoor samples
For most elements, personal and indoor; concentrations were lower than and highly correlated with outdoor concentrations. The
highest correlations (median r.0.9) were found for sulfur and particle absorbance (EC), which both represent fine; mode particles
from outdoor origin. Low correlations were observed for elements that represent the coarser part of the PM2.5 particles (Ca, Cu, Si,
CI).
Jansen et al. (2005,082236)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Panel Study
Winter 2002-2003
Seattle, Washington, USA
Elderly Respiratory Disease Patients (asthma/COPD)
71-86 yr
No
Personal PM10MPEMIO; Indoor home and Outdoor home PM2.5, PM10 Single-stage inertial Harvard Impactors and 37-mm Teflon
filters
PM10
PM10, PM2.5, fine particles (-PM1)
PM10, PM2.5
July 2009
A-257
DRAFT-DO NOT CITE OR QUOTE

-------
Components) BC, as an estimate of elemental carbon (EC)
Primary Findings For 7 subjects with asthma, a 10 pg/m3 increase in 24-h avg outdoor PM10; and PM2.5 was associated with a 5.9 [95% CI: 2.9-8.9]
and 4.2 ppb (95% CI: 1.3-7.1) increase in FeNO, respectively. A 1 pg/m3 increase in outdoor, indoor, and personal BC was
associated with increases in FeNO of 2.3 ppb (95% CI: 1.1-3.6), 4.0 ppb (95% CI: 2.0-5.9), and 1.2 ppb (95% CI: 0.2-2.2),
respectively. No significant association was found between; PM or BC measures and changes in spirometry, blood pressure, pulse
rate, or Sa02 in these subjects.
Jaques et al. (2004,155878)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment, field evaluation of continuous PM2.5 monitor in comparison to integrated samplers
February-August 2002
Claremont, California
Continuous PM2.5 sampler, time-integrated PM2.5 samplers
NR
NR
No personal exposure assessment was conducted
NR
NR
PM2.5
NR
PM2.5 mass measurements using the Differential TEOM monitor are consistent with those of the MOUDI and Partisol. Differences
can be explained by loss of ammonium nitrate from reference time-integrated samplers. Partisol underestimates MOUDI measured
mass.
Jedrychowski et al. (2006,156606)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Prospective cohort
11/2000-3/2003
Krakow, Poland
Non-smoking pregnant women
Yes
Personal; Exposure Monitor Sampler (PEMS, Harvard; School of Public Health) with
PM2.5
NR
PM10
NR
The contribution of the background ambient PM10 level was very strong determinant of the total personal exposure to PM2.5 and it
explained about 31 % of variance between the subjects.
Jo and Lee(2006,156613)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Case study
Winter of 2004 and summer of 2005
Daegu, Korea
Residents of high-rise apartment buildings
NR
All of the surveyed apartments were constructed with concrete and iron frames. The apartments used liquid petroleum gas for
cooking and as their primary heating system. The exhaust gas generated from heating or cooking was mechanically vented out of
the apartments.
The PM10 concentrations were measured using real-time light scattering PM10 monitors (HAZDUST Model EPAM-500). The CO
concentrations were measured using a CO dosimeter (CMCD-1 OP) equipped with an activated charcoal-Purafil prefilter.; From each
building, one lower-floor apartment (first or second floor) and one higher-floor apartment (between 10th and 15th floor) were
simultaneously surveyed. The concentrations of CO and PM10 were measured at the breathing height in the main living area where
the participants spent most of their time and from the apartment balconies outdoors.
NR
PM10
PM10
CO
This study found that the outdoor air concentrations of CO and PM10 were higher for lower-floor apartments than for higher-floor
apartments situated in residential areas. In addition, the concentrations were significantly higher in winter and in summer,
regardless of the floor height of the apartments. The indoor concentrations in the lower- and higher-floor apartments, however, were
not consistent with the outdoor concentrations. Proximity to a major roadway was found to increase the indoor and outdoor
concentrations of PM10 in high-rise apartments and therefore cause elevated exposures of the residents during presence at home.
This was not observed for CO. Atmospheric stability was found to elevate indoor and outdoor air pollutant concentrations and was
therefore determined to be another important factor regarding the level of exposure to CO and PM10.
July 2009
A-258
DRAFT-DO NOT CITE OR QUOTE

-------
Johannesson et al.
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
(2007,156614)
Cohort
Spring and fall seasons of 2002 and 2003
Gothenburg, Sweden
General adult population
23-51 yr
NR
Fine particles were measured for 24 h using both personal and stationary monitoring equipment. Personal monitoring of PM2.5 and
PM1 was carried out simultaneously with parallel measurements of PM2.5 and PM1 indoors in living rooms and outside the house on
a balcony, porch, etc. In addition, urban background PM2.5 levels were measured. Personal monitoring was performed in two ways.
The 20 randomly selected subjects carried personal monitoring equipment for PM2.5 only, while the 10 staff members carried two
pieces of personal monitoring equipment at the same time. On the first measuring occasion, the staff members carried one PM2.5
cyclone and one PM1 cyclone. On the second occasion, duplicate monitors for PM2.5 were used. For personal and residential
monitoring, the BGI Personal Sampling Pump was used together with the GK2.05 cyclone for PM2.5 sampling and the Triplex
cyclone SCC1.062 for PMisampling. The personal sampling pump was placed in a small shoulder bag and the cyclone attached to
the shoulder strap near the subject's breathing zone. The personal monitoring equipment was carried by the subject during awake
time. During the night, it was placed in the living room. For indoor monitoring in living rooms, cyclones (PM2.5 and PM1) were placed
at about 1.5 m above the floor. The same setup was used for residential outdoor monitoring. The urban background monitor was
placed on top of a roof somewhat south of the city center but not near any major highway.
PM2.5; PM1
PM2.5; PM1
PM2.5; PM1
BS
Personal exposure of PM2.5 correlated well with indoor levels, and the associations with residential outdoor and urban background
concentrations were also acceptable. Statistically significantly higher personal exposure compared with residential outdoor levels of
PM2.5was found for nonsmokers. PM1 made up a considerable proportion (about 70-80%) of PM2.5. For BS, significantly higher
levels were found outdoors compared with indoors, and levels were higher outdoors during the fall than during spring. There were
relatively low correlations between particle mass and BS. The urban background station provided a good estimate of the residential
outdoor concentrations of both PM2.5 and BS2.5 within the city. The air mass origin affected the outdoor levels of both PM2.5 and
BS2.5; however, no effect was seen on personal exposure or indoor levels.
Jones and Harrison
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
(2006,155886)
NR
January 2002-December 2004
England London Marylebone Road (Located beside arterial road in street canyon carrying approximately 80,000 vehicles per day);
London North Kensington (In grounds of school in west London suburb); Harwell (On western side of business park surrounded by
agricultural land)
NR
NR
NR
No personal monitoring.
NR
NR
PM10
NaCI; Strongly bound H20; Secondary organic material
Using existing co-located and coincident data it has been possible to show that the removal of three natural components—sea salt
(NaCI), strongly bound water and secondary organic matter—would reduce the number of days on which the PM10 concentration
exceeds 50 pg/m3 by about 50%. At each site, amongst the three natural components considered, the strongly bound water makes
the greatest contribution to the mean or median concentrations of PM10, followed by NaCI, and SOC respectively. Strongly bound
water was shown to have the biggest effect on the number of days on which the PM10 concentrations exceeded a value of 50 pg/m3;
however, removal of estimated NaCI and SOC components also had a noteworthy effect on reducing PM10 concentrations.
Therefore, application of this proposed measure would make a very major difference to the likelihood of compliance or otherwise
with the 24-h limit value for PM10.
Jones et al. (2007,156615)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Monitoring living room, child's bedroom, cot, and at 2 heights 1.4 & 0.2 m above the floor
NR but probably 2006
Perth, Australia
Children 0-2 yr
0-2 yr
House age, house type, building material, # of stories, attached garage, main heating fuel, air conditioning
NR
NR
PM10, PM2.5, PMT
July 2009
A-259
DRAFT-DO NOT CITE OR QUOTE

-------
Ambient Size NR
Component(s) NR
Primary Findings No difference between samples at 0.2 and 1.4 m from floor in 3 PM fractions, no difference between living room, child's bedroom,
and crib. Large variability between homes.
Kaur et al. (2005, 086504)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment, evaluation of exposures between modes of transport, routes, timing
April 28-May 23, 2003
Street canyon intersection in Central London, UK
Users of an urban street canyon intersection
NR
NR
PM2.5 measured usnig high-flow gravimetric personal samplers (PM2.5) operating at a flow rate of 16 l/min carried in a backpack with
sampling head positioned in personal breathing zone; Ultrafine particles measured using TSI P-TRAK Ultrafine Particle Counters in
which ambient aerosol mixes with isopropyl alcohol. Alcohol condenses to form droplets that can be easily counted by a
photodetector as they pass through a laser beam.
PM2.5, Ultrafine particles (UFP, 0.02-1 .Oum)
PM2.5, Ultrafine particles (UFP, 0.02-1 .Oum)
PM2.5
NR
Personal exposures to PM2.5 while walking were significantly lower then while riding in a car or taxi, likely a function of greater
distance to roadside. No significant differences in PM2.5 were observed between exposures on the high traffic road compared with
the backroad. Personal exposure levels were lowest during midday measurements for PM2.5 and highest in the early evening.
Personal exposures to ultrafine particles were lowest while walking and highest while riding the bus. Exposures to ultrafine particles
were also significantly higher on the high traffic road and during morning measurements. Exposure to ultrafine particles were
highest in the morning, likely the result of peak traffic density in the morning. Exposure assessment also revealed that the
background and curbside monitoring stations were not representative of the personal exposure of individuals to PM2.5 and CO at
and around a street canyon intersection.
Kaur, et al. (2005,088175)
Study Design	Personal exposure assessment of pedestrians walking along high-traffic urban road
Period	April 19, 2004-June 11, 2004
Location	Central London, UK
Population	Pedestrians
Age Groups	NR (adults, presumably)
Indoor Source	NR
Personal Method	PIVksgravimetric filter measurement; Ultrafine PM (0.02-1 pm) P-TRAK device; Reflectance reflectometer measurement of PM2.5
filter
Personal Size	PM2.5; Ultrafine PM (0.02-1 pm); Reflectance ("blackness") of PM2.5 filter
Microenvironment Size	NR
Ambient Size	PM2.5; Ultrafine PM (0.02-1 pm); Reflectance ("blackness") of PM2.5 filter
Components)	NR
Primary Findings	PM2.5 pedestrian exposure was well correlated with and above background fixed-site monitoring levels. PM pedestrian exposure was
influenced by proximity to curbside and the side of the road walked on.
Kim et al. (2005,156640)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Panel study
8/1999-11/2001
Toronto, Canada
Cardiac-compromised patients
Mean age 64 yr
Yes. Tracer molecules/elements were used to investigate sources of indoor PM, including regional long range transport,
combustion, local crustal materials. All were statistically significantly associated with indoor PM2.5
Rupprecht and Patashnick ChemPass Personal Sampling System
PM2.5
NR
PM2.5
Sulfate, Elemental carbon (EC), Calcium, Magnesium, Potassium, Sodium.
Kim et al. (2005,156640)
Study Design Panel study
Period 8/1999-11/2001
Location Toronto, Canada
July 2009
A-260
DRAFT-DO NOT CITE OR QUOTE

-------
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
Cardiac-compromised patients
Mean age 64 yr
Gas range (68%); indoor grill (11%); outdoor barbeque (30%); Gas heating fuel (68%); Oil heating fuel (7%)
Rupprecht and Patashnick ChemPass Personal Sampling System
PM2.5
NR
PM2.5
NR
Personal PM2.5 exposures were higher than outdoor ambient levels. Personal PM2.5 exposures levels were correlated with ambient
levels, mean r = 0.58
Koenig et al. (2003,156653)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
Comprehensive exposure assessment. "The research was part of an intensive exposure assessment and health effects panel study
of susceptible sub-populations in Seattle from; 1999 through 2002 (Liu et al., 2003, 073841)."
10-day monitoring session winter 2000-2001; 10-day monitoring session spring 2001
Seattle, WA
Asmatic children
6-13 yr
Harvard personal environmental monitors; Continuous PM monitors (nephelometers); Harvard Impactors; TEOM monitors; and
exhaled breath measurements into an NO-inert Mylar balloon
PM2.5
PM2.5
PM2.5
NR
This study found a consistent relationship between daily eNO values in children with asthma and daily PM2.5 measured at fixed sites
and on subjects. As hypothesized, the authors found that the use of ICS therapy modified the association between eNO and PM2.5.
Including ambient NO values for the hour of the home visit from a central site in the model and discarding high NO days (>100 ppb)
attenuated the magnitude but did not alter the association between PM2.5 and eNO in all analyses. Same-day outdoor, indoor,
personal, and central PM2.5 levels were associated with eNO in either analysis. These data suggest ambient PM2.5 exposure in
Seattle is associated with an increase in eNO in children with asthma. Because eNO is a marker of airway inflammation, and PM
has been shown to cause inflammation in animal studies, this result is biologically plausible. This finding also agrees with previous
asthma research in Seattle that showed associations between PM2.5 and lung function decrements in children with asthma.
Koistinen et al. (2004,156655)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
Representative Population-based study
Oct 1996-Dec 1997
Helsinki, Finland
Non-smoking adults not exposed to environmental tobacco smoke.
Adults 25-55 yr
Soil from outdoors, cooking, smoking, aerosol cleaners, sea salt, combustion sources
Aluminum briefcase with personal sampling monitor
PM2.5; BS
PM2.5; BS
PM2.5; BS
% contribution to PM2.5; Outdoor-I ndoor-Work-Personal; CoPM * 35 28 32 33; Secondary" 46 36 37 31; Soil 16 27 27 27;
Detergents 0 6 2 6; Sea Salt 3212
Population exposure assessment of PM2.5, based on outdoor fixed-site monitoring, overestimates exposures to outdoor sources like
traffic and long-range transport and does not account for the contribution of significant indoor sources.
Kousa et al. (2001,025270)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Population based exposure assessment
October 1996 to June 1998
Helsinki, Finland; Basel, Switzerland; Prague, Czech Republic; Athens, Greece
Adult urban populations
25-55 yr
NR, Workplace and residential outdoor samples could not be collected for every participant. The number of non-ETS-exposed
participants was particularly small in Prague (12) and Athens (29), and therefore, these results from these cities should be
interpreted with caution. The population sampling and sample representativity issues are described in detail in Rotko et al. (2000,
0121181. and, for the Basel sample, in Oglesby et al. (2000, 0018321.
The 48 h PM2.5 exposure was measured by a personal exposure monitor (PEM). Two filter holders were provided for each
participant. One 'workday filter' for work and commuting, about 2 8-10 h, and one "leisure time filter" for the remaining time.
Microenvironmental PM2.5 monitors (MEMs) were placed at the participant's home outdoors and indoors and in their workplaces.
The pumps were programmed to run at home outdoors and indoors during the expected leisure time hours and at workplace during
July 2009
A-261
DRAFT-DO NOT CITE OR QUOTE

-------
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
expected working hours of each participant. The personal and microenvironmentai PM2.5 sampling methods and QA/QC results are
presented in detail in Koistinen et al. (1999, 010628).
PM2.5
PM2.5
PM2.5; PM10
NR
In Helsinki the concentration associations are high between the outdoor air concentrations of PM2.5 and PM10 measured
simultaneously at different locations. The highest exposure correlations are observed between the personal exposures and the
respective indoor air concentrations. Correlations between the personal exposures and outdoor/ambient air concentrations are
considerably lower (all centers). Personal exposures during leisure time correlate better with outdoor/ambient concentrations than
during the workday (Helsinki and Prague). Leisure time and workday exposures correlate poorly with each other (all centers).
Removing ETS improved the correlations between personal (indoor) air and ambient (outdoor) air, but decreased the correlations
between personal exposures and indoor air concentrations and also between the personal exposures during workday and leisure
time. In spite of these generalizations, there are considerable differences between the cities.
Koutrakis et al. (2005,095800)
Study Design	Panel study
Period	Baltimore 6/28/98-8/22/98 (summer), 2/1/99-3/16/99 (winter); Boston 6/13/99-7/23/99 (summer), 2/1/00-3/12/00 (winter)
Location	Baltimore, MD Boston, MA
Population	Healthy older adults, children, adults with COPD
Age Groups	Children 9-13 y/o; Seniors 65+y/o; COPD Subjects NR
Indoor Source	No
Personal Method	Personal exposure samples of PM2.5; were collected using a specially designed multipollutant sampler (Demokritou et al. 2001).
PM2.5was collected using personal environmental monitors (PEMs) and 37-mm; Teflon filters (Teflo, Gelman Sciences, Ann Arbor
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
PM2.5
NR
PM2.5
Elemental Carbon (EC); SO42-
Ambient PM2.5 and SO4 are strong predictors of respective personal exposures. Ambient SO4 is a strong predictor of personal
exposure to PM2.5. Because PM2.5 has substantial indoor sources and SO4 does not, the investigators; concluded that personal
exposure to SO4 accurately reflects exposure to ambient PM2.5 and therefore the ambient component of personal exposure to PM2.5
as well.
Kulkarni and Patil (2003,156664)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Personal exposure assessment of toxic metals
NR
Mumbai, India; Two localities or sites, namely, Marol and Sakinaka, denoted as Sites 1 and 2 respectively
Outdoor workers- low-income group population working and residing in industrial areas
NR
low grade cooking fuel and inadequate ventilation
A personal sampler (Cassela/ SKC make), which consists of a diaphragm pump and operates on rechargeable batteries, was used
along with a cyclone to measure personal exposure to respirable PM (RPM). The device was fitted to the waist belt of the
respondent and connected by a flexible tube to the cyclone, which can be clipped to the shirt collar. The inlet of the cyclone was
kept near the breathing level of the respondent. After working hours, the personal sampler was worn by the respondent in his/ her
residence. Before sleeping, the sampler was removed from the waist and kept in the "on" condition as close to the breathing level
as possible.; The RPM in ambient air was measured simultaneously by using high volume sampler (HVS) with a cyclone
attachment for removal of particles with size greater than 10 pm.
PM5
NR
PM5
Lead; Nickel; Cadmium; Copper; Chromium; Potassium; Iron; Manganese
All listed metals were detected in the ambient air where as only Lead, Cadmium, Manganese, and Potassium were detected in
personal exposures. Mean daily exposure to lead exceeds the Indian NAAQS by a factor of 4.2. However, ambient concentration of
lead conforms to this standard. There is a rising trend in the personal exposures and ambient levels of cadmium. However, they are
low and do not pose any major health risk as yet. Personal exposures to toxic metals exceed the corresponding ambient levels by a
large factor ranging from 6.1 to 13.2. Thus, ambient concentrations may underestimate health risk due to personal exposure of toxic
metals. Outdoor exposure to toxic metals is greater than the indoor (ratios ranging from 2.3 to 1.1) except for potassium (ratio 0.77).
However, there is no significant correlation between these two.
Kumar etal. (2006,129347)
Study Design Use of one year's 24-h monitoring data to model exposure to vehicular emissions.
Period Apr 1991-Feb 1992
Location Mumbai, India
July 2009
A-262
DRAFT-DO NOT CITE OR QUOTE

-------
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
exposure to lead at busy intersections
NR
NR
NR
NR
NR
Suspended PM (SPM)
Ai As Ca Cu Cr Fe Hg K Mg Mn Na Ni NO2 Pb O2 SO4 SPM "Concentration in ng/m3, number of samples = 45.
Application of a hybrid, receptor cum dispersion model is one possible way to evaluate effective emission factors for vehicles in
different operating conditions like those at traffic-junctions. The composite approach of receptor and dispersion model gives realistic
effective emission factors and will be useful for air quality management.
Laietal. (2006,090262)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
Population-based assessment of urban adult exposures. Identifying determinants of indoor PM concentrations
1996-2000
Athens, Basel, Helsinki, Milan, Oxford, Prague
Homes of urban adults
NR
Number of people smoking at home, duration of gas stove use. A previous paper is cited for full details on sampling methodology
(Jantunenet al, 1998)
No personal exposure assessment was conducted.
NR
NR
NR
BS
Number of people smoking at home, outdoor PM2.5 conc., wind speed, duration of gas stove use, and outdoor temperature were
significant determinants of indoor PM2.5. City-specific effects included outdoor PM2.5 conc., smoking, and wind speed. Outdoor BS,
Laietal. (2004,056811)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
Personal exposure study
December 1998-February 2000
Oxford, UK
Adults
25-55 yr(avg = 41)
Cooking, active smoking, passive smoking heating by gas heater
Personal exposure monitors (PEM) were carried by the participant for 48-h personal sampling, and microenvironmental monitors
(MEM) were placed inside the participant's home (residential indoor), outside the home (residential outdoor) and in the participant's
workplace (workplace indoor).; The PM2.5 samplers used were GK2.05 cyclones (KTL) with 2um pore Gelman Teflo filters, and
WINS PM2.5 impactors for personal exposure and residential outdoor samples, respectively. VOC sampling was accomplished with
Perkin Elmer Tenax-TA tubes, CO Enhanced Measurer T15s were used for CO samples, and NO2 passive sampling badges were
used to sample NO2. No residential outdoor CO or NO2 samples were taken.
PM2.5
PM2.5
PM2.5
Ag Cr Mn Si; Al Cu Na Sm; As Fe Ni Sn; Ba Ga P Sr; Br Ge Pb Ti; Ca Hg Rb Tl; Cd I S Tm; CI K Sb V; Co Mg Se Zn; Zr
Both the indoor and outdoor environments have sources that elevated the indoor concentrations in a different extent, in turn led to
higher personal exposures to various pollutants.; Geometric mean (GM) of personal and home indoor levels of PM2.5,14 elements,
total VOC (TVOC) and 8 individual compounds were over 20% higher than their GM outdoor levels. Those of NO2, 5 aromatic
VOCs, and 5 other elements were close to their GM outdoor levels. For PM2.5 and TVOC, personal exposures and residential indoor
levels (in GM) were about 2 times higher among the tobacco-smoke exposed group compared to the non-smoke exposed group,
suggesting that smoking is an important determinant of these exposures. Determinants for CO were visualized by real-time
monitoring and that the peak levels of personal exposure to CO were associated with smoking, cooking and transportation
activities. Moderate to good correlations were only found between the personal exposures and residential indoor levels for both
PM2.5 (r = 060; p< 0 001) and NO2 (r = 047; p = 0 003).
Laietal. (2004,156666)
Study Design Longitudinal exposure assessment
Period January 4-14, 2001
Location Taipei, Taiwan; (highway toll station)
Population Highway toll station workers
Age Groups 19.3-43.6 yr; mean = 25.7 ± 5.71
Indoor Source Indirect exposure assessment was based on information on (a) lane-specific traffic density (available for all lanes throughout the
study period), (b) estimated relationships between lane- and shift-specific traffic density and the average PM2.5 concentrations, and
(c) information on time periods spent by individuals in different working environments.
July 2009
A-263
DRAFT-DO NOT CITE OR QUOTE

-------
Personal Method Direct exposure assessments were conducted by installing battery-operated personal PM2.5 monitors (University Research
Glassware Corp.) in the booth near the breathing zone of the workers.
PM2.5
NR
NR
NR
Toll workers on Taipei highways are exposed high concentrations of PM2.5. Mean PM2.5 concentration per vehicle in the truck and
bus lanes was 6.4 and 3.7 times higher, respectively, than that of ticket- or car-payment car lanes. There was a statistically
significant correlation between traffic density and PM2.5 concentrations in car lanes with ticket payments and truck and bus lanes.
Wind speed and humidity had a significant inverse association withPIVks concentration in car lanes with ticket and cash payments.
Bus and truck lane was the strongest determinant of log (PM2.5). The results of this study show that personal exposure to PM2.5 can
be reliably estimated using indirect traffic approaches.
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Larson et al. (2004, 098145)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Time-series epidemiologic study
Sep 26, 2000-May 25, 2001
Seattle, Washington
"Susceptible Populations"
Time-activity diary
Harvard Personal Environmental Monitor
PM2.5
PM2.5 outside subject's residence, and inside residence
PM2.5 at Central outdoor site (downtown Seattle)
Light absorbing carbon (LAC) and trace elements
Five sources of PM2.5 identified vegetative burning, mobile emissions, secondary sulfate, a source rich in chlorine, and crustal-
derived material. The burning of vegetation (in homes) contributed more PM2.5 mass on average than any other sources in all
microenvironments.
Lee et al. (2006,098249)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment for instrument development
11/2003, 5/2004
Boston, MA
NR
No
A new personal respirable particulate sampler (PRPS), operating at 5L/min. Sampler is designed to collect PM0.5, PM1.0, PM2.5,
PM4.5; and PM10 as well as O3, SO2, and NO2. Sampler consists of 5 impaction stages, a backup filter, and two diffusion passive
samplers. Particles are collected onto a polyurethane foam (PUF) substrate.
NR
NR
PM>10, PM10-2.5, PM2.5
NR
In the field, the PM10, PM2.5, and sulfate concentrations measured by PRPS were in a very good agreement with those obtained from
the; reference samplers.; In the lab, the size distributions measured by the PRPS were found to be much closer to those; measured
by the real-time particle sizing instruments than to those measured by the MOI.
Lee et al. (2006,188450)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Cross-sectional
NR, but prior to 2006
Charleston, Ottawa, Clarksville, Ohio
Farmers
NR, but prior to 2006
Hogs, poultry, cattle, feed, bedding
The dust & microorganisms passed thru an optical particle counter and a filter sampler to collect airborne microorganisms.
0.7-1 pm; 1-2 pm; 2-3 pm; 3-5 pm; 5-10 pm; Total dust
NR
NR
Fungal spores and bacteria
The highest contribution of large particles (3-10 pm) in total particles was found during grain harvesting. In animal confinements the
particles were dominated by smaller particles <3 pm. A high proportion of the particles between 2 and 10 pm were fungal spores.
Lewne et al. (2006,189293)
Study Design Personal exposure study to investigate the occurrence of systematic differences in the PE exposure to motor exhaust and to study if
these are influenced by the choice of exposure indicator gaseous or particulate
July 2009
A-264
DRAFT-DO NOT CITE OR QUOTE

-------
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Components)
Primary Findings
Sep 1997 to Oct 1999
Stockholm, Sweden
Taxi, bus, and lorry drivers
NR
NR
PM was measured with a logging instrument Data-RAM, using nephelometric monitoring (Data-RAM measures PM 0.1 to 10 pm)
PM10
NR
PM10
NR
N02
1) Lorry drivers experienced the highest exposure and taxi drivers the lowest with bus drivers in an intermediate position, regardless
of whether NO2 or particles were used as exposure indicator; 2) The levels of both NO2 and particles were higher for bus drivers in
the city than for them driving in the suburbs; 3) Using diesel or petrol as a fuel for taxis had no influence on the exposure for the
drivers, indicating that the taxi drivers' exposure mainly depends on exhaust from surrounding traffic.
Lewne et al. (2007,156690)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
7 groups of occupations defined by common or high exposure to DE
Oct 2002-June 2004
Stockholm, Sweden
Persons exposed to DE
Vehicle exhaust
Pump units and gravitmetric for PM1& PM2.5 and real-time monitoring of elemental carbon and total carbon. Diffusive samplers for
NO2 as an indicator of the gas phase of exhaust.
PM1, PM2.5, and DataRAM (PM0.1-10)
NR
NR
elemental carbon (EC), total carbon (TC)
Tunnel construction workers had the highest levels of exposure for all indicators, followed by diesel-exposed garage workers. The
other 5 groups were significantly lower with no difference between the groups.
Li et al. (2003, 047845)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Primary Findings
Concurrent 10-min avg indoor and outdoor concentrations of PM10 and PM2.5 were recorded for 2 days each in 10 homes with
swamp coolers
Summer 2001
El Paso, Texas
Cooking, cleaning, walking
NR
NR
PM2.5 and PM10; indoor and outdoor; tapered element oscillating microbalance (TEOM) instruments. 2 days were monitored for
PM2.5, and2forPMio.
NR
NR
Evaporative coolers were found to act as PM filters, creating indoor concentrations approximately 40% of outdoor PM10 and 35% of
outdoor PM2.5, regardless of cooler type.
Liao et al. (2006,188451)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Case study
January 18-27, 2003
Changhwa, Central Taiwan
Traditional Taiwanese residences
Chinese style cooking, incense burning, cleaning, and people's moving
A portable laser dust monitor (DM1100) was used to analyze the indoor and outdoor PM characteristics. The DM1100 was placed in
a single indoor location, 1.5 m above the floor, adjacent to areas of the kitchen, altar, and living room where the housing activities
occurred.
NR
PMo.5-5
PMo.5-5
NR
Results indicate that only 2.6-8% of indoor particles are from outdoor sources. Both indoor and outdoor PM concentrations increase
with PM size intervals, as do the deposition rates from cooking events; Authors suggest that "results revealed that cooking and
incense burning events were major contributors to indoor concentrations for the particle sizes 1-5 pm. Results demonstrated the
importance of knowing the time-activity data and the real-time indoor and outdoor particle size distribution information for
July 2009
A-265
DRAFT-DO NOT CITE OR QUOTE

-------
understanding exposure to particles of indoor sources. More importantly, this research illustrates that an exposure assessment
based on PM0.5-5 measured indoors can provide valuable information on the fate of indoor particles and hazard to human health."
Liu etal. (2003,073841)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Part of a larger exposure assessment and health effect panel study
Winter 2000-2001 and spring 2001
Seattle, WA
Children with asthma
6-13 yr
NR
Personal PM2.5 measurements were collected from each subject using the Harvard personal environmental monitors.
PM2.5
PM2.5
PM2.5
NR
The ambient-generated component of PM2.5 exposure was consistently associated with increases in eNO and the indoor-generated
component was less strongly associated with eNO. As a result, the eNO results support the hypothesis that PM2.5 of outdoor origin
could be more potent per unit mass than particles of indoor origin. However, the lung function data indicate that PM2.5 of indoor
origin might be more potent per unit mass in resulting in decrements of lung functions, although the results across functional tests
were not consistent. The authors tentatively concluded that partitioning personal exposure into indoor- versus outdoor-generated
particles is useful in understanding the health effects of sources of personal PM2.5 and that the effects of indoor- versus outdoor-
generated particles differ for different health points.
Liu etal. (2005,156704)
Study Design	Exposure assessment, validation set within a prospective occupational cohort (boiler workers)
Period	NR (healthy working adults)
Population
Personal Method	Yes, A personal environmental monitor (PEM, Model 200, MSP Co, Shoreview, MN) with a pump at 4L/min
Personal Size PM10
Microenvironment Size PM10
Ambient Size NR
Components) Metals Vanadium (V), Nickel (Ni), Iron (Fe), Chromium (Cr), Cadmium (Cd), lead (Pb), Manganese (Mn)
Primary Findings	The validation demonstrated good approximations of actual exposures with differences less than 5% for PM.
Liu etal. (2003,073841)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Findings
Comprehensive exposure assessment
1999-2001
Seattle, WA
High-risk sub populations
Children 6-13 yr; elderly 65-90 yr (one person was below 65 but not specified)
Personal PM2.5 exposures were determined using the Harvard Personal Environmental Monitor for PM2.5 (HPEM2.5). Each subject
carried an HPEM2.sin the breathing zone for 24 h, except while sleeping, showering, or using the restroom. The monitor was
attached to the shoulder strap of either a backpack or a fanny pack that contained the air pump. When the monitor was not worn, it
was placed at an elevation of 3-5 feet (e.g., on a table) close to the subjects.; The indoor and outdoor PM concentrations were
measured with single-stage inertial Harvard Impactors (HI) and 37 mm Teflon filters for PM«and PM2.5. One HI2.5-HI10 pair was
located inside each home in the main activity room and connected to a Medo pump (model vp0935a). Concurrently, one HI2.5-HI10
pair was located outside each home and connected to a Gast pump (model DOA-V191-AA). All HI sampling periods were for 24 h
at a flow rate of 10 L/min. HI2.5, Hho, and HPEM2.5 were also co-located with the federal reference method monitor for PM2.5 (FRM2.5)
at the central Beacon Hill site, which is located in a semiresidential area (elevation, 300 feet) and is maintained by the Washington
State Department of Ecology.
PM2.5; PM10
PM2.5; PM10
PM2.5; PM10
The average personal exposures to PM2.5 were similar to the average outdoor PM2.5 concentrations but significantly higher than the
average indoor concentrations. Indoor and outdoor PM2.5, PM10, and the ratio of PM2.5 to PM10 were significantly higher during the
heating season. The increase in outdoor PM10 in winter was primarily due to an increase in the PM2.5 fraction. A similar seasonal
variation was found for personal PM2.5. The children in the study experienced the highest indoor PM2.5 and PM10 concentrations.
Personal PM2.5 exposures varied by study group, with elderly healthy and CHD subjects having the lowest exposures and asthmatic
children having the highest exposures. Within study groups, the PM2.5 exposure varied depending on residence because of different
particle infiltration efficiencies; PM2.5 exposures among the COPD and CHD subjects can be predicted with relatively good power
with a microenvironmental model composed of three microenvironments. The prediction power is the lowest for the asthmatic
children.
July 2009
A-266
DRAFT-DO NOT CITE OR QUOTE

-------
Lonati etal. (2005,126171)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Comparison sampling of an urban background site, UB during cold season and warm season with no traffic and a vehicle tunnel
(TU) cold season.
Aug 2002-Nov 2003
Milan, Italy
Urban population
NR
NR
NR
NR
NR
PM2.5
EC, OC, Particulate organic matter (OM); Total mass; Chloride, Nitrate, Sulfate, Ammonium, Crustal elements, Metals,
undefined+F12
Higher PM2.5 during the cold season, about twice the warm season. Tunnel data are 7 times the urban background. The vehicle
contribution to PM2.5 is 11 % in the warm season and 6% in the cold season.
Lung etal. (2007,156719)
Period
Location
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Weekdays between Nov 1998 and Feb 1999
6 communities in Taiwan, China 2 in Taipei, 2 in Taichung, and 2 in Kaohsiung. Sites are industrial, commercial, residential and
mixed.
18 to >70
Being in kitchen, park, major boulevard, stadium, incense burning, household work, factory, environmental tobacco smoke, traffic,
ventilation conditions
Personal Environmental Monitor with a SKC personal pump at 2 L/min, 37 mm Teflon filters
PM10
PM10
PM10
None
Outdoor rather than indoor levels contributed significantly to personal exposure.; Important factors include time spend outdoors and
on transportation, riding a motorcycle, passing by factories, cooking or being in the kitchen, incense burning at home.
Magari et al. (2002,034813)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Cross-sectional study of boilermakers
NR
NR
NR, probably metal tradesmen
No
Gil-Air 5 personal pump
PM2.5
NR
NR
V, Nr, Cr, Mn, Cu, Pb
There were statistically significant mean increase in the standard deviation of the normal-to-normal heart rate index (SDNN) of 11.30
msec and 3.98 msec for every 1 pg/m3 increase in the lead and vanadium concentrations after adjusting for mean heart rate, age,
and smoking status.
Maitre etal. (2002,156726)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Personal (occupational) and ambient (in traffic area) PM and particle-bound PAH exposure assessment. This study evaluates
individual airborne exposure to gaseous and particulate carcinogenic pollutants in a group of policemen working close to traffic in
the center of Grenoble, France.
Summer (June); winter (January) (year not indicated)
City of Grenoble, located in the southeast of France
Non-smoking policemen working outdoors on foot
NR
NR
Cyclone and filter with personal sampling pump (SKC, United Kingdom)
Respirable particles (defined in this paper as the mass of particles that pass through a size selective orifice with a 50% collection
efficiency at a cut-off aerodynamic diameter of 4 pm)
NR
Respirable particles (defined in this paper as the mass of particles that pass through a size selective orifice with a 50% collection
efficiency at a cut-off aerodynamic diameter of 4 pm)
July 2009
A-267
DRAFT-DO NOT CITE OR QUOTE

-------
Components) PAH, benxene-toluene-xylenes (BTX), aldehydes; Personal BaP; Personal PAHc; Personal PAH; Personal Benzene; Personal
Toluene; Personal Xylene; Personal BTX; Personal Formaldehyde; Personal Acetaldehyde; Personal Aldehydes; Stationary BaP;
Stationary PAHc; Stationary PAH; Stationary Benzene; Stationary Toluene; Stationary Xylene; Stationary BTX; Stationary
Formaldehyde; Stationary Acetaldehyde; Stationary Aldehydes
Primary Findings The occupational exposure of policemen does not exceed any currently applicable occupational or medical exposure limits.
Individual particulate levels should preferably be monitored in Grenoble in winterto avoid underestimations.
Malm at al. (2005,156729)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Finding(s)
Exposure assessment, characterization of physical and optical properties of carbonaceous aerosol species, and comparison of
several semi-continuous monitoring systems
July 15-September 4, 2002
Yosemite National Park at the Interagency Monitoring of Protected Visual Environments (IMPROVE) monitoring site
Different monitoring instruments to quantify ambient aerosol concentrations
NR
No personal exposure assessment was conducted
NR
NR
NR
Inorganic ions (sulfate, nitrate), organic carbon in PM10 and PM2.5 size ranges, elemental carbon
70% of the organic mass was made up of elemental carbon. 24-h bulk measurements of various aerosol species compared more
favorably with each other than with the semi-continuous data.; Semi-continuous sulfate (PILS) correlated well with 24-h measurem
Mar at al. (2005,087566)
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
1999-2001
Seattle, WA USA
"Older subjects" (< 57 y/o), non-smokers
Age 57+ yr
No
Harvard impactor
PM2.5
PM2.5, PM10
PM2.5, PM10
NR
Healthy subjects; taking no medications had decreases in heart rate associated with; indoor and outdoor PM2.5 and PM10. Healthy
subjects on medication; had small increases in systolic blood pressure associated with indoor; PM2.5 and outdoor PM10.
McCormack, et al. (2007,156745)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Findings
Stratified analysis of subjects in the BIESAK study
NR but < 2003
East Baltimore, Maryland
low-income children with asthma
2-6 yr
sweeping, vacuuming, smoking, stove use, burned food, oven, candles/incense, open windows, space heater
NR
NR
PM10 and PM2.5; in child's bedroom
PM10 and PM2.5; central monitoring site
Indoor concentrations of PM2.5 and PM10 were twice as high as the ambient; Sweeping, smoking, and ambient PM contributed
significantly to higher indoor concentrations. Sweeping (not vacuuming) increased the PM10 by 3-4 pg/m3.
Meng et al. (2005,081194)
Study Design 3 Cohorts, one in New Jersey, 1 in Los Angeles, and 1 in Houston.; Personal, home indoor, and home outdoor samples taken for
PM2.5.
Period Summer, 1999-spring, 2001
Location Houston, Texas; Los Angeles, California; and Elizabeth, New Jersey
Population
Personal Method MSP monitors on the front strap of the sampling bag near the breathing zone. Pump, battery, and motion sensor were on the hip or
back.
Personal Size PM2.5
Microenvironment Size PM2.5
Ambient Size NR
Components) NR
July 2009
A-268
DRAFT-DO NOT CITE OR QUOTE

-------
Primary Findings The median contribution of ambient sources to indoor PM2.5 using the mass balance approach was 56% for all study homes, 63% for
California, 52% for New Jersey, and 33% for Texas.
Meng at al.. (2005, 081194)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Evaluation of the use of central-site PM, rather than actual exposure, in PM epidemiology
Summer 1999-spring 2001
3 cities: Houston (TX), Los Angeles County (CA), and Elizabeth (NJ)
People suffering from cardiovascular and respiratory morbidity likely. Not specified
All age groups possible, not specified
Likely sources mentioned, not identified
NR. Indoor and outdoor sampling conducted
PM2.5
NA
NR
EC, OC, S, Si
Use of central-site PM2.5 as an exposure surrogate underestimates the bandwidth of the distribution of exposures to PM of ambient
origin.
Meng etal. (2005,058595)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
RIOPA study matched indoor home & outdoor exposure assessment
May-October (hot); November-April (cool); (1999-2001)
Los Angeles County, CA; Elizabeth, NJ; Houston, TX
Non-smoking homes
Combustion (primary); atmospheric (secondary); sulfate, organics, nitrates; mechanically (abrasion) generated.
Filter (not specified)
NR
Indoor home.; PM2.5
PM2.5, outdoor home
Organic and elemental carbon; 24 elements (metals).
Mihaltan et al. (2006,156761)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Indoor air monitoring. To assess the effect of smoking on air quality in hospitality venues (restaurant, pubs and bars).
NR
Romania
Restaurant/pubs/bars
NA
Smoking
Personal aerosol monitor
NR
Respirable suspended particles, PM2.5
NR
NR
Hospitality venues allowing indoor smoking are significantly more polluted than indoor smoke-free venues and outdoor air in
Romania.
Miller etal. (2007,156765)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure Assessment, evaluation study of effectiveness and accuracy of a nephelometer (portable, direct reading photometer) to
measure tailpipe emissions of elemental carbon from diesel engines
NR
In laboratory
2 Exposure assessment methods to measure elemental carbon
NR
NR
No personal exposure assessment was conducted.
NR
NR
NR
EC, Total Carbon
EC measurements made with a Thermo Electron Personal DataRAM 1200 direct reading nephelometer showed good correlation
with EC mass concentrations quantified by thermal optical analysis of PM2.5 and PM1 samples collected on quartz filters (reference
NIOS
July 2009
A-269
DRAFT-DO NOT CITE OR QUOTE

-------
Miller etal. (2007,156764)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Comprehensive study of key contaminants
Ottawa, Ontario, Canada
NR
NR
NR
PM2.5 and PM10 filter samples were collected in the living room of each home for 7 days, using SKC sampler model 200 PEM on
tared teflon filters. Concurrent PM2.5 samples were collected on 47-mm Teflo 2- pm filters with MiniVol air samplers mounted on a
tripod ~2 m in front of the house. Particulate samples for analysis of airborne endotoxin, ergosterol, and 31, 3-D-glucan were
collected on a three-piece cartridge equipped with an endotoxin-free polycarbonate filter in the living room sand bedrooms of each
home. In the living room, samplers were located at a height between 1.22 m and 1.83 m from the floor and no closer than 0.5 m to
surfaces. In bedrooms, the samples were collected form as close to the beds as feasible. BC concentrations were continuously
recorded for 7 days with a Magee Scientific Aethalometer in the living room of each house.
NR
PM2.5; PM10
NR
BC; Also assessed: Endotoxin; Ergosterol; Glucan; Dust samples Dust >300; Der p1; Der f1; Fel d1;
Airborne concentrations of the contaminants measured generally were unremarkable, although the mass of settled dust per square
meter was well above that associated with increased asthma and comfort symptoms clinical response, particularly in urban homes.
When co-occurrence of inflammatory agents and dust mite allergen burdens in the houses was considered, three homes had higher
relative amounts of the agents considered. Based on what is known about clinical interactions between, for example, endotoxin and
dust mite allergens, a combined exposure possibly results in an increased relative risk of allergic disease.
Molnaretal. (2005,156772)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Indoor/outdoor exposure assessment related to domestic wood burning
10 February to 12 March 2003
Hagfors, Sweden
Adult residents of Haqfors
NR
NR
Identical sets of equipment were used for both personal exposure and indoor sampling a GK2.05 (KTL) cyclone connected to a BGI
400S Personal Sampling Pump with a flow rate of 4 I min 1. Each person was equipped with an easily carried shoulder bag with the
cyclone and pump attached to it. The cyclone was attached to the shoulder strap and placed near the breathing zone. The personal
sampler was worn all day, and at night it was placed next to the stationary indoor sampler in the living room, owing to the noise of
the pump.; Two different types of impactors were used for the outdoor sampling one Sierra Andersen series 240, dichotomous
virtual impactorthat separates particles into two size ranges, coarse and fine particles (PM10-2.5 and PM2.5, respectively); and one
EPA-WINS impactor (PQ100 EPA-WINS Basel PM2.5 Sampler) for collecting PM2.5 particles. The outdoor measurements were
made at a single location on the roof of a single car garage, belonging to one of the subjects, in the middle of the study area.
PM2.5
PM10-2.5; PM2.5
PM10-2.5; PM2.5
BS; S; CI; K; Ca; Mn; Fe; Cu; Zn; Br; Rb; Pb
Statistically significant contributions of wood burning to personal exposure and indoor concentrations have been shown for K, Ca,
and Zn. Increases of 66-80% were found for these elements, which seem to be good wood-smoke markers. In addition, CI, Mn, Cu,
Rb, Pb, and BS were found to be possible wood-smoke markers, though not always to a statistically significant degree for personal
exposure and indoor concentrations. For some of these elements subgroups of wood burners had clearly higher levels which could
not be explained by the information available. Sulfur, one of the more typical elements mentioned as a wood-smoke marker,
showed no relation to wood smoke in this study due to the large variations in outdoor concentrations from LDT air pollution. This
was also the case for PM2.5 mass. Personal exposures and indoor levels correlated well among the subjects for all investigated
species, and personal exposures were generally higher than indoor levels. The correlations between the outdoor and personal or
ind
Molnaretal. (2006,156773)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Cross-sectional
Autumn and spring in 2002 and 2003
Goteborg, Sweden,
Persons living in urban settings
20 subjects 20-50 yr randomly selected from the population and 10 from departmental colleagues.
Nr
The volunteer subjects had a small shoulder bag with one PM2.5 cyclone and a pump attached. Intake was in the breathing zone.
Pump was carried during the day and placed next to the indoor cyclone during the night.; Ten subjects from their staff wore 2 sets
of sampling equipment near the breathing zone. A GK2.05 cyclone for PM2.5 and a Triplex cyclone for PMiin a small shoulder bag.
PM2.5 and PM1
NR
July 2009
A-270
DRAFT-DO NOT CITE OR QUOTE

-------
Ambient Size NR
Components) S; CI; K; Ca; Ti; V; Mn; Fe; Ni Cu; Zn Br Pb.
Primary Findings PM2.5 personal exposures were significantly higher than both outdoor and urban background for the elements CI, K, Ca, Ti, Fe, and
Cu.; Personal exposure was also higher tan indoor levels of CI, Ca, Ti, Fe, and Br, but lower than outdoor Pb. Residential outdoor
levels were significantly higher than the corresponding indoor levels for Br and Pb, but lower for Ti and Cu. The residential levels
were also significantly higher than the urban background for most elements.
Molnar et al. (2007,
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
156774)
Microenvironmental monitoring of PM and elements in 10 schools, 10 preschools, and 20 non-smoking homes.
1 Dec 2003- 1 July 2004
Stockholm, Sweden
Children
6-11 yr (no pre-school children) but sampling was conducted at 10 preschools.
Smoking, gas stoves,
NR
NR
NR
NR
S; K; Ca; Ti; V; Cr; Mn; Fe; Ni; Cu; Zn; Br; Pb
Significantly lower indoor concentrations of S, Ni, Br and Pb, elements from long-range transported air masses, were found in all
locations. Only Ti was significantly higher indoors in all locations, probably because of Ti02 in paint pigment. Similar differences
were found during both seasons for homes and schools. At preschools the infiltration of the long-range transported elements S, Br
and Pb was lower in winter than in spring, and also the crustal elements Ti, Mn and Fe had higher indoor concentrations during
spring. There were spatial differences outdoors, with significantly lower concentrations of elements of crustal and traffic origin in the
background area community.
Monkkonen et al. (2005,156775)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Human exposure assessment in homes in India
Nagpur Mar to Oct 2002; Mysore Aug to Dec 2002
Nagpurand Mysore, India
Residential homes in India
NR
Yes; cooking w/ kero. and LPG; Toaster; Burning incense; Infiltration of outdoor air; Burning coconut husks
TSI Condensation Particle Counter Model 3007 (CPC counts all particles >10 nm); TSI Model 8520 Dust Trak; PM2.5 Environmental
Monitor with Whatman PTFE membrane filters and gravimetric analysis
PM2.5 for mass (pg/m3); Total PM for counts (particles/cm3)
PM2.5 for mass (pg/m3); Total PM for counts (particles/cm3)
Total PM for counts (particles/cm3)
NR
The maximum concentrations observed in most cases were due to indoor combustion sources. Besides cooking stoves that use
LPG or kerosene as the main fuel, high indoor concentrations can be explained by poor ventilation systems.
Mwaiselage et al. (2006,156789)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Cross-sectional; personal monitoring. To determine the effects of cement exposure on acute respiratory health.
June-August 2001
Dar es Salaam, Tanzania
Cement factory workers
NR
Cement production
Cellulose Acetate Filter, Sidekick pump
Respirable dust, total dust
NR
NR
Ca, Al, Fe, K
Results of Cox Regression analysis showed that prevalence ratios for cough, short breathness and stuffy nose for high exposed
workers in the production department compared to low exposed workers in the low exposed workers working in the maintenance
department and the administration building are 6.7, 4.5 and 1.9 respectively. Cross shift decrease in PEF was more in the higher
among high exposed workers (7.6%) than low exposed workers (2.7%). The observed acute acute respiratory health effects are
most likely related to exposure of workers to high concentrations of irritant cement dust.
Na and Cocker (2005,156790)
Study Design
Human exposure assessment
Period Sept. 2001 -January 2002
July 2009
A-271
DRAFT-DO NOT CITE OR QUOTE

-------
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Mira Loma, CA
Residential homes and a hiqh school
NR
Indoor EC (elemental carbon) concentrations primarily of outside origin; Indoor PM2.5 significantly influenced by indoor OC (organic
carbon) sources, including indoor smoking.
PM2.5 Particle trap impactor with 47 mm Teflo substrates; EC/OC Particle trap impactor with 47 mm QAT Tissuquartz quartz
fiber filter, analysis by thermal.optical carbon aerosol analyzer (NIOSH Method 5040)
PM2.5
NR
PM2.5
EC (Elemental carbon); OC (Organic carbon)
Indoor PM2.5 was significant influenced by indoor OC sources. Indoor EC sources were predominantly of outdoor origin.
Naumova et al. (2002, 026105)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment
6/1999-5/2000
Los Angeles County, CA; Houston, TX; Elizabeth, NJ
US. General population
NR
No
None-area sampling only (in home and outdoors)
NR
PM2.5
PM2.5
PAH (total and specific)
See Component Column. Many of the study findings pertain to combined particle-bound and gas-phase PAHs, and are not
presented here.
Naumova et al. (2003, 089213)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
RIOPA Study-PAH partitioning indoor and outdoor. To evaluate the hypothesis that outdoor air pollution contributed strongly to
indoor air pollution.
July 1999-June 2000
Los Angeles, CA, Houston, TX, Elizabeth, NJ
Houses
NR
NR
Modified MSP Samplers, 37 mm quartz filter
NR
Filter, PM2.5
Filter, PM2.5
Organic Carbon (OC), Elemental Carbon (EC)
Multiple linear regression (MLR) log PAH particulate partition coefficient (kp) vs log vapor pressure coefficient (std) 0.888 (0.009)
fraction of elemental carbon in PM coefficient (std) 3.686 (0.238) fraction of elemental carbon in PM coefficient (std) 0.469 (0.055)
temperature coefficient (std) -0.0456 (0.002) intercept (std) 8.398 (0.604) R2 = 0.845. Both EC and OC carbon are important
predictors of gas/particle partitioning of PAHs, with EC being a better predictor. Because EC is highly correlated with (and is a good
tracer of) primary combustion-generated OC, this result suggests that PAHs more readily sorb on combustion-generated aerosol
containing EC. Enrichment of the indoor aerosol in nin-combustion OC suggests that sorption of PAHs is more important in the
indoor air compared to the outdoor air. The MLR developed in this work will improve prediction of gas/particle partitioning of PAHs
in indoor and outdoor air.
Nerriere et al. (2005,088630)
Study Design
Period
Location
Population
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment
2001 -2003; Lung cancer mortality 1999
4 French Cities (Grenoble, Paris, Rouen, Strasbourg)
6-13 y/o children not exposed to passive smoke; 30-71 y/o adults (average age ~40y/o) not occupational^ exposed
yes, using Harvard Chempass worn in a backpack
PM2.5
NR
PM2.5
NR
Number of cases attributable to PM2.5 exposure (95% CI); attributable Fraction (%) (95% CI) Paris 303 (42-553); 8 (1 -16); Grenoble
12 (3-22); 10 (3-19); Rouen 19 (3-35); 10 (2-19); Strasbourg 43 (7-71); 24 (4-10).
July 2009
A-272
DRAFT-DO NOT CITE OR QUOTE

-------
Nerriere et al. (2005,089481)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Copollutant(s)
Primary Findings
Exposure assessment, stratified sampling of children and adults in 3 environments high traffic emissions, local industrial sources,
and urban background.
"Hot" season May-June and "cold" season Feb-Mar. Grenoble in 2001, Paris in 2002, Rouen in 2002-2003, Strasbourg 2003.
Grenoble, Paris, Rouen, and Strasbourg, France
Persons living, working, or going to school in 3 urban areas one highly exposed to traffic emissions, one influenced by local
industrial sources, and a background urban environment. Industrial sources of pollution were present in each city.
6-13 yr and 20-71 yr. All non-smokers and not exposed to environmental tobacco smoke or industrial air pollution.
Daily activity diaries used to do
Rucksack with Harvard ChemPass
PM2.5, PM10
NR
PM2.5, PM10
NO2
The difference between ambient air concentrations and average total exposure is pollutant specific. PM2.5 and PM10 concentrations
underestimate population exposures across almost all cities, season, and age groups, the opposite is true for NO2.
Ng et al. (2005,155996)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Study is to model the dispersion of the 911 WTC destruction cloud to areas of the city and boroughs using "representative persons."
Input data are from extant monitoring stations throughout the area.
14 Sep., 2001 to 30 Sep., 2001
Lower Manhattan, New York City, NY
NYC residents
NR, but both adults and children
Smoking, cooking
Simulated
PM2.5
PM2.5
PM2.5
NR
Although the outdoor PM2.5 was lower than the NAAQS value, personal exposure levels were higher than outdoor and might be of
concern.
Nikasinovic et al. (2006,156807)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Copollutant(s)
Primary Findings
cross-sectional
Oct 1999-Jun 2002
Paris, France
Asthmatic children
7-14 yr
Presence of pets, smoking in the home, house dust mites, home ventilation frequency, allergies to grass, cats, pollen, gas cooking,
barometric pressure.
Active sampler in a rucksack carried by the children whenever they moved.
PM2.5
NR
PM10
Ozone
Pollutant concentrations did not differ between the 2 groups. In asthmatic children only personal PM2.5 levels were correlated to
nasal markers after adjustment for age, sex, house mites, pollens, cat, tobacco smoke, barometric pressure, and respiratory
infection.
Noullett etal (2006,155999)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Cohort
5 February to 16 March 2001
Prince George, British Columbia
Children
10-12 yr
NR, Each child completed a time activity diary every 30 min on the days that they carried the monitor. A motion sensor (HOBO,
Onset Computer Corporation) was also placed in each pack and data from the sensor was downloaded each morning and then
compared to each child's time activity diary as a quality assurance measure.
PM2.5 Harvard Personal Environment Monitors (HPEM2.5) with a PTFE Teflon filter (Pall Gelman R2PJ037) were used for both the
ambient and personal sampling. At ambient sampling sites, the HPEM2.5 was suspended approximately 4 ft above the school
rooftop (20 ft from the ground at all schools), connected to a large flow controlled pump and situated in an open area on the roof
free of air vents, exhausts or intakes. BGI air sampling pumps and battery packs (BGI-400S and BGI-401) were used for the
July 2009
A-273
DRAFT-DO NOT CITE OR QUOTE

-------
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
personal monitoring and were contained in a child-size backpack. The sampler was attached to the strap of the backpack in the
breathing zone of the child with the inlet facing downwards and protected by a 4-in piece of plastic tubing. Subjects were required to
wear the pack whenever possible and otherwise to keep the pack close to them and as close to their breathing zone as possible
PM2.5
NR
PM2.5
SO4; ABS (light absorbing carbon)
In Prince George, a combination of topography, meteorological conditions and location of ambient sources resulted in episodic
levels of fine PM during the short study period in the winter of 2001. Thermal inversions were moderately associated with both high
ambient levels and personal exposures and were likely responsible for the spatial variation and, in combination with wind, the
temporal variation in ambient concentrations throughout the city. The clear link between thermal inversions and both high ambient
levels and measured personal exposures during PM2.5 episodes support management strategies to reduce ambient sources during
periods of limited dispersion in an effort to reduce exposure levels. Despite the significant spatial variation found in ambient levels
throughout the city for all three measures, there was a high correlation between the outdoor sites suggesting that a single monitor
would represent temporal trends. Similar to the findings in other studies, both sulfate and light
O'Neill et al. (2004, 087429)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Time-series epidemiologic study of PMio-associated mortality, comparison of different samplers
January 1,1994-December 30,1998
5 sites in Mexico City, Mexico
Urban environments
NR
NR
No personal exposure assessment was conducted
NR
NR
PM10, PM2.5
NR
PM10 levels were higher in the winter. PM10 levels measured using different methods were not well correlated with each other. Re-
analysis of associations between PM2.5 and mortality with sensitivity analyses (non-parametric modeling) produced lower eff
Offenberg et al. (2004,156821)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment
6/1999-5/2000
Los Angeles County, CA; Houston, TX; Elizabeth, Nj
US. General population
NR
No
None-area sampling only (in home and outdoors)
NR
PM2.5
PM2.5
Chlordane
Geometric mean particle-bound chlordane concentrations were higher indoors relative to outdoors, suggesting indoor sources.
Ogulei et al. (2006,156823)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure Assessment
11/1999-3/2000
Reston, VA
US homes
NR
Yes. Nine primary sources of PM were identified gas burner; use (boiling water), deep-frying tortillas and miscellaneous; cooking of
dinner, burning of citronella candle, combined gas burner and gas oven use (broiling salmon), sweeping/vacuuming, use of electric
toaster; oven, traffic, wood smoke, and pouring of kitty litter.
None
NR
A range 0.01-20.0 mm
NR
NR
Each particle source identified in the study produces distinct particle size distributions.
Pang etal. (2002,037057)
Study Design Field test of prototype Personal Particulate Organic and Mass Sampler
July 2009
A-274
DRAFT-DO NOT CITE OR QUOTE

-------
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
November 2000-May 2001
Seattle, WA
NR
NR
NR
Outdoor sampling for PM2.5 massThe PPOMS was co-located with two Federal Reference Method (FRM) samplers and a HPEM
sampler at the Beacon Hill EPA Air Quality and Particulate Speciation Monitoring Site. Samples were collected on each sample day
for 24 h, starting at 0:00 PST. Indoor sampling for particulate carbon the PPOMS was co-located with an integrated particle sampler
and a Harvard impactor (HI2.5) inside of two residences. Five samples were collected at each house over the course of several
days.
NR
NR
PM2.5
EC; OC
"This study shows that the PPOMS design provides a 2.5 pm size-selective inlet that also prevents the adsorption of gas-phase
SVOC onto quartz filters, thus eliminating the filter positive artifacts. The PPOMS meets a significant current challenge for indoor
and personal sampling of particulate organic carbon. The PPOMS design can also simplify accurate ambient sampling for PM2.5."
Paschold etal. (2003,156847)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Concurrent 48-h indoor and outdoor concentrations of PM10 and PM2.5 in 10 homes with swamp coolers
Summer of 2001
El Paso, Texas
Homes with evaporative coolers
NR
NR
NR
NR
PM10 and PM2.5
NR
Geologic material; Sodium; Magnesium; Aluminum; Potassium; Calcium; Titanium; Manganese; Iron; Trace metals; Copper; Zinc;
Barium; Lead
Indoor elemental concentrations in PM10 were approximately 50-70% lower than outdoor concentrations in 9 of 10 homes,
consistent with the PM10 indoor/outdoor (I/O) mass concentrations previously reported. PM2.51/ O ratio correlations were not as
strong as for PM10; however, reduced correlations could be attributed to a pattern of recurring outlier data pairs, consisting of the
same 3 or 4 elements in all 10 homes.
Polidori et al. (2007,156877)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
time-series epidemiologic study
Site A (Group 1 [G1 ]); -Phase 1 July 6 to August 20, 2005; -Phase 2 October 19 to December 10, 2005; Site B Group 2 [G2]); -
Phase 1	August 24 to October 15, 2005; -Phase 2 January 4 to February 18, 2006
Los Angeles, California (Two Retirement homes)
Elderly residents of Los Angeles, California retirement homes
NR
No
Two identical sampling stations were installed at each location, one indoors and one outdoors. The indoor sampling station at site A
was located in the recreational area of the first community's main building, adjacent to a construction site where work was ongoing.
The indoor sampling area at site B was situated in the dining room of the second community's main building. At both sites, the
outdoor station, set up inside a movable trailer, was positioned within 300m from the indoor station. Two D-attenuation mass
monitors (BAMS) (Model 1020) were used at each indoor and outdoor sampling station to measure hourly PM2.5 mass
concentrations. Continuous NO, NO2, and CO measurements were taken indoors and outdoors using Thermo Environmental NOX
analyzers (Model 42), and Dasibi CO Analyzers (Model 3008) respectively. O3 concentrations were also monitored at each
sampling station by using API Ozone Analyzers (Model 400A). At both indoor and outdoor sampling areas, a water-based
condensation particle counter (CPC model 3785), and a semicontinuous OC-EC analyzer (Model 3F) were operated side by side. A
modified National Institute for Occupational Safety and Health (NIOSH) analysis protocol was used here to evolve particulate OC
and EC.
NR
PM2.5
PM2.5, PN
OC; EC; OC1;OC2-4
Measured indoor and outdoor concentrations of PM2.5, OC, EC, PN, O3, CO, and NOX were generally comparable, although at G2, a
substantial peak in indoor OC, PN, and PM2.5 (probably from cooking) was typically observed between 6:00 and 9:00 am. The study
average percentage contribution of outdoor SOA to outdoor particulate OC was 40% and varied between 40% and 45% in the
summer (during G1P1) and 32% and 40% in the winter (during G2P2). The low AERs (0.25-0.33 h-1) calculated for G1 and G2 are
consistent with the structural characteristics of the sampling sites, the low number of open windows and doors, and the presence of
central air conditioners. F«estimates were highest for EC and also for OC. Lower lvalues were obtained for PM2.5 and PN,
July 2009
A-275
DRAFT-DO NOT CITE OR QUOTE

-------
because these compounds are composed of both volatile and nonvolatile inorganic and organic compounds. Based on a single
compartment mass balance model, it was found that 13-17% (G2P2) to 16-26% (G1P1) of measured indoor OC was emitted or
formed indoors. Although the G2 indoor site was characterized by higher indoor morning OC peaks because of cooking, the overall
contribution of indoor sources to measured indoor OC was higher at the G1 site. The modeling results also showed that the
measured indoor PM2.5 emitted or formed indoors was highly variable (from 6-21 % at G1P1 to 45-51 % at G1P2). The average
percentage contribution of indoor SOA of outdoor origin to measured indoor OC varied from -35% (at site 1) to -45% (at site 2).
Also, outdoor-generated primary OC composed, on average, 36-44% of measured indoor OC during G2P1 and G1P1 respectively.
Poupard etal. (2005, 074025)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Explore relationships between indoor and outdoor air quality
NR
La Rochelle, France
School buildinqs
NR
No
GRIMM 1.108 analyzer
15 size intervals from 0.3 to 15 microns
NR
15 size intervals from 0.3 to 15 microns
NR
Influence of room occupancy on particle concentrations indoors changes with particle size; Indoor ozone and particle concentrations
are negatively correlated.
Price et al. (2003, 098082)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment, comparison of PM10 samplers
November 2000 to August 2001
Sunderland, England (northeast England), monitoring at curbside
Urban Populations near high traffic areas
NR
NR
No personal exposure assessment was conducted
NR
NR
PM10
NR
Correlation between TEOM and partisol appeared to be seasonal, with strongest correlation in the summer when ambient PM10
concentrations were relatively low. In the winter and spring, when PM10 levels are higher, the Partisol sampler records grea
Ramachandran et al. (2003,190112)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Matched PM2.5 24-h and 15-min averages at 9-10 residences in each of 3 communities and at 3 central sites, in 3 seasons.
The measurements were made over 3 seasons—spring (April 26-June 2), summer (June 20-August 10), and fall (September 23-
November 20) of 1999.
Phillips, East St. Paul, and Battle Creek, Metropolitan Minneapolis-St. Paul, Minnesota
Urban residential communities
23 females, 9 males; mean age 42 ± 10, range 24-64 yr
No
NR
NR
PM2.5 in residences
At 3 central sites, PM2.5
NR
Outdoor PM2.5 concentrations across the Minneapolis- St. Paul area appear to be spatially homogeneous on a 24-h time scale as
well as on a 15-min time scale. Short-term average outdoor PM2.5 concentrations can vary by as much as an order of magnitude
within a day.
Riojas-Rodriguez et al. (2006,156913)
Study Design	Panel Study
Period	12/2001-4/2002
Location	Mexico City, Mexico
Population	Patients with heart disease
Age Groups	Avg Age 55 yr (range 25-76)
Indoor Source	Nub
July 2009
A-276
DRAFT-DO NOT CITE OR QUOTE

-------
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Yes, using nephelometers (personal data ram (PDR) model 1200, Monitoring Instruments for the Environment, Inc) connected to a
4L/min pump
PM2.5
NR
NR
NR
Authors found a decrease in HRV measured as high frequency (Ln) (coefficient =-0.008, 95% confidence interval (CI) to -0.015,
0.0004) for each 10 microg/m3 increase of personal PM2.5 exposure.
Robinson et al. (2007,156054)
Study Design A pollution mapping exercise was undertaken to measure average pollution levels on a number of transects across the New South
Wales Valley and the variation with height and land use was determined. Spatial variation was then used to predict population
exposure to PM2.5 pollution and the effect on health.
Period Pollution measurements were made between 17 July 1996 and 10 September 1996
Location Armidale, New South Wales, Australia
Population Armidale, New South Wales, Australia
Age Groups NA
Indoor Source NR
Personal Method A portable Radiance Research M903 integrating nephelometer was used to measure ambient air pollution at four transects; Ambient
air pollution was also measured using a fixed site Belfort nephelometer
Personal Size NR
Microenvironment Size NR
Ambient Size PM2.5
Components) NR
Primary Findings 1) Considerable variability was observed in winter woodsmoke pollution levels; 2) A small number of badly operated heaters can
have a large influence on local air quality; 3) Pollution was generally higher in residential areas; 4) Annual exposure to PM2.5
pollution in Armidale from woodsmoke was double that from all sources in Sydney.
Rojas-Bracho et al. (2004,054772)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Cohort, repeated measures. 18 COPD patients in non-smoking homes were sampled either in winter 1996 or 1997.16 of these also
were sampled in the summer. All subjects were sampled for 6 consecutive days in winter, and one or two sets of 6 consecutive
days in the summer.
1996-1997
Boston, Massachusetts
COPD patients
Housecleaning, cooking, transport in motor vehicles, low-effort home activities, moderate-effort home activities, activities in public
places, and resting or sleeping.
PEM attached to shoulder strap of a bag (near breathing zone) containing the pump and batteries.
PM2.5, PM10, & PM2.5-10
PM2.5, PM10, & PM2.5-10
NR
NR
During both seasons personal exposures were higher than indoor or outdoor means, except the winter indoor concentrations were
higher than the personal or outdoor.
Rotko et al. (2002,037240)
Study Design European multi-city air pollution study
Period Athens, Greece (A) 26 January 1997-4 June 1998; Basel, Switzerland (B) 3 February 1997-23 January 1998; Milan, Italy (M) 10
March 1997-23 May 1998; Oxford, UK (O) November 1998-7 October 1999; Prague, Czech Republic (P) 3 June 1997-4 June
1998; Helsinki, Finland (H) 26 September 1996-10 December 1997
Location Athens, Greece (A); Basel, Switzerland (B); Milan, Italy (M); Oxford, UK (O); Prague, Czech Republic (P); Helsinki, Finland (H)
Population Adults
Age Groups 25-55 yr
Indoor Source NR
Personal Method Personal PM2.5 exposures were collected on two different filters one for the working hours including commuting (personal work) and
the other for the remaining hours of 48-h measurement period (personal leisure time). In addition to personal exposure monitoring,
PM2.5 concentrations were measured in each home (indoors and outdoors) and workplace (indoors). The PM2.5 concentration
measured at work was the avg of two consecutive workdays and at home of the remaining hours of the 48-h monitoring period.
PM2.5 personal cyclones were used as pre-separators at flow rate of 4 I min 1 [this seems wrong] and the EPA-WINS impactors
were employed at 16.7 I min 1 for the microenvironment measurements with Gelman Teflo filters (37 and 47 mm, respectively).
Personal Size PM2.5
Microenvironment Size PM2.5
Ambient Size PM2.5
Copollutant(s) NO2
July 2009
A-277
DRAFT-DO NOT CITE OR QUOTE

-------
Primary Findings * There was a large variation in the levels of air pollution annoyance between the six studied cities. The highest annoyance levels
were experienced while in traffic. "The significant determinants of air pollution annoyance were the city, self-reported sensitivity to
air pollution and respiratory symptoms, downtown residence and gender of the subject.; * No consistent or significant correlations
were seen between the individual levels of annoyance and exposure concentrations to either PM2.5 or NO2. * High air pollution
annoyance in traffic, however, was significantly associated with home outdoor concentrations of both PM2.5 and NO2 and downtown
living (NO2- model). "When average annoyance levels and concentrations were averaged for each city, the perceived annoyance
levels at home correlated highly with the measured personal 48-h PM2.5 and NO2 exposures and home indoor NO2 concentration,
annoyance at work correlated with personal workday exposure and workplace PM2.5 concentrations, and annoyance in traffic wi
Salma et al. (2005,
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
156937)
2 types of samplers collected aerosols in an urban area. 23 samples were collected with each device separately for day and night.
Spring 2002
Budapest, Hungary
urban dwellers
NR
NR
NR
NR
NR
See direct quote in the note below
Al; Si; Ca; Ti; Fe; CI; Zn; Na; Mn; Ni; Cu; Pb; K; S; Br
The variation in the overall size distributions and RMCs for the various elements indicated the existence of multiple sources,
including vehicular (both combustion and frictional) and industrial emissions, resuspension of road and soil dust, and long-range
transport of air masses. The significant coarse mode for some typical anthropogenic elements (Cu and Zn) and the observed
coarse mode concentration differences between daytime periods and nights (e.g., for Ca) point to the importance of frictional
sources and road dust resuspension in cities, which are both primarily related to road traffic.
Salma etal. (2007,113850)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
examination of aerosol air quality and its temporal variation in the Budapest metro
April 20 and 21,2006
Budapest, Hungary
underground metro commuters
NR
No. Air monitoring equipment consisted of a tapered element oscillating microbalance (TEOM), a wind monitor (Campbell), and a
laboratory-made Gent-type stacked filter unit (SFU) aerosol sampler. OM was equipped with a PM10 inlet facing upwards and was
operated with the filter heated to 401C to prevent moistening. The sampling station was ventilated without filtration by drawing air
from the opposite platform to the roof level of a 12-story building next to the station.
No personal monitoring. In situ aerosol measurement and sample collection at the metro station.
PM10-2.O and PM2.0
NA
NR
30 elements (Na, Mg, Al, Si, P, S, CI-, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, Ga, Ge, As, Se, Br, Rb, Sr, Y, Zr, Nb, Mo, Ba, and Pb)
The concentrations observed in the Astoria underground station were clearly lower (by several orders of magnitude) than the
corresponding workplace limits.
Sanderson and Farant (2004,156942)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Indoor and outdoor air monitoring of PAH. Investigate the relationship between indoor and outdoor PAH.
NR
Canada
Residential homes in neighborhoods around aluminum smelting plant
NR
NR
Indoor quartz filter, XAD-2 Resin Outdoor glass fiber filter
NR
NR
NR
4-6 ring PAHs on indoor particle
Indoor concentration of 4-6-ring PAH linked to outdoor sources in residences without any major indoor source, but with industrial
facility as the main outdoor source. Study suggests that simultaneous measurements of indoor and outdoor concentrations of PAH
>4 rings predominantly associated with fine PM could provide useful estimates of particle infiltration efficiency.
Sarnat et al. (2006, 089166)
Study Design Outdoor-indoor pollutant infiltration, occupied residences
July 2009
A-278
DRAFT-DO NOT CITE OR QUOTE

-------
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
July 28, 2001 -February 25, 2002
Los Anqeles, CA
NR
Yes; cleaning, cooking, home ventilation (open windows/doors), kitchen fans, air conditioner/heating usage, number of occupants,
nearby roadways
NR
NR
PM2.5, Particle number
PM2.5
BC (nonvolatile component); N03 (volatile component)
1) Infiltration rate for PM2.5 was intermediate, while BC was highest, N03 lowest; 2) Infiltration rate varied with particle size, air
exchange rate, outdoor N03; 3) PM2.5 infiltration was lowest for volatile component; 4) Outdoor volatile PM2.5 components may be
less representative of indoor exposure to volatile PM2.5 of ambient origin; 5) Outdoor nonvolatile PM2.5 components may be more
representative of indoor exposure to nonvolatile PM2.5 of ambient origin.
Sarnat et al. (2006, 090489)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Personal and ambient exposure assessment
June 14-August 18 (summer); Sep 24-Dec 15 (fall), 2000
Steubenville, OH
Non-smokinq, older adults
No
Integrated filter gravimetric measurement
PM2.5
NR
PM2.5
S04; EC
1) 24-h ambient measurements more representative of personal particle exposure than gases; 2) ventilation is an important
exposure modifier.
Sarnat et al. (2005, 087531)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Copollutant(s)
Primary Findings
Time-series epidemiologic study
Summer 1999 and winter 2000
Boston, MA. Comparisons to a previous study in Baltimore are made.
School children and seniors
NR
PM2.5
NR
PM2.5
S04,
O3, NO2, SO2
Substantial correlations between ambient PM2.5 concentrations and corresponding personal exposures.; Summertime gaseous
pollutant concentrations may be better surrogates of personal PM2.5 exposures (especially personal exposures to PM2.5 of ambient
origin) than they are surrogates of personal exposures to the gases themselves.
Sax et al. (2006,156950)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Finding(s)
2 Cohorts, one in NYC and 1 in LA.; Personal, home indoor, and home outdoor samples taken for PM2.5, VOCs, and aldehydes.
1999-2000, winter and summer in NYC, winter and fall in LA.
New York City, New York, and Los Angeles, California
13-19 yr
No
Customized backpack
NR
NR
NR
NR
Most VOCs has median upper-bound lifetime cancer risks that exceeded the USEPA benchmark of 1 x 10 6 and were generally
greater that the EPA modeled estimates, more so for compounds with predominant indoor sources, chromium, nickel, and arsenic
had median personal cancer risks above the benchmark with exposures largely from outdoors and other microenvironments. The
EPA model overestimate risks for beryllium and chromium and underestimate risks for nickel and arsenic.
July 2009
A-279
DRAFT-DO NOT CITE OR QUOTE

-------
Scheepers et al. (2003,156955)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Field Study
Pilot Study; March 15-18,1999, and March 22-25,1999 (coal mine); April 12-14,1999 (oil shale mine); Main Study; June 5-22, 2000
Pilot Study; Ostrava, Czech Republic (black coal mine); Kohtla-Jarve, Estonia (oil shale mine); Main Study; Kohtla-Jarve, Estonia (oil
shale mine)
Coal miners with high exposures to Diesel-powered machinery
NR
NR
Personal sampling was accomplished by each individual worker carrying personal air sampling equipment (GSA 200 or Gillian)
during two shifts in the same work week (1 shift for the main study). The air sampling pumps operated in the breathing zones of the
individual workers and operated at an electronically controlled flow rate of 2.0l/min.; Inhalable dust samples were collected using a
sampler head developed by the Institute fur Gefahrstoff Forschung der Bergbau Berufsgenossenschaft (IGF). Respirable dust was
collected using an elutriator pre-separator type MPGII (IGF). Particles were collected on polystyrene membrane filters with a Teflon
coating. All samples were taken at a height of -1.5 m above the floor.
NR
NR
NR
1 -nitropyrene (1-NP)
This study confirms that 1-NP in black coal and oil shale mines is mostly associated with respirable particles and that mining
operations involving diesel-powered engines exposures to DEP may be 3- to 10-fold higher for underground miners than workers
on the surface. Furthermore, measurements of particle-associated 1-NP is a more sensitive and discriminating indicator o exposure
to DEP than inhalable or respirable particles because of the relatively high concentrations of mine dust in mining operations.;
Respirable dust concentration were 2- to 3-fold higher in the breathing zone than at fixed sampling locations while 1-NP
concentrations were found to be 2.5-fold and 10-fold higher in the coal mine and oil shale mine respectively. This is thought to be
due to location of fixed sampling points as well as wind and humidity levels within the mines themselves. For these reasons and
others, personal air sampling is preferred over air sampling at fixed sites.
Shalat et al. (2007,
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
156971)
Indoor home exposure assessment; sampling technology demonstration
Winter heating season
Residential home
Children
Pre-toddler (6- to 12-month-old) children
Mobile robotic and stationary. Filter and real-time nephelometer.
Floor; Filter inhalable particles (approximately < 100 pm)
Indoor home; Nephelometer total suspended particles.
NR
NR
Shao et.al. (2007,156973)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment
July and Winter 2003
Beijing, China
General population
NR
Soot aggregates, coal fly ash, minerals, unknown fine particles
PM10 selective inlet heads 30L/min flow rate with polycarbonate filters
PM,o
PM,o
PM,o
NR
Plasmid scission assay, coupled with the image analysis, can be used to evaluate the relationship between particle physico-
chemistry and toxicity.
Shilton et al. (2002, 049602)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Respirable particulates inside and outside of a building were collected and compared
24-h sampling from 12:45 pm Mondays to Fridays between 9/19/00 to 5/01/01
Wolverhampton city center, University of Wolverhamptom, UK
Building near traffic dominated by heavy-duty diesel vehicles
Outdoor (primary); Mn,AI, N03, CI- (wind-blown dust); Cu and Zn-(traffic emissions)
Active sampling using Casella sampler (filter)-
Respirable PM (inside and outside)
Respirable PM (inside)
July 2009
A-280
DRAFT-DO NOT CITE OR QUOTE

-------
Ambient Size Respirable PM (outside)
Components) Respirable PM, metals (Zn, Cu, Mn, Al), sulfate, nitrate, and chloride
Primary Findings The indoor particulate concentration was driven by ambient concentration; meteorological-induced changes in ambient PM were
detected indoors.
Simons et al. (2007,
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Finding(s)
156982)
NR
Baltimore, Maryland; and surrounding counties
Children with asthma
Inner city-6-12 yr; Surrounding counties 6-17 yr
Gas stoves, cats, dogs, smokers, mold/mildew carpet, outside PM, dryer vents
Indoor air was collected from the child's bedroom with 4 L/min MSP impactors over a72-h Period.
NR
PM2.5; PM10
PM2.5; PM10
Allergens were also assessed Dust mite; Bla/g; Mus/m; Fel/d; Can/ f; Airborne Mus/m
Compared with the homes of suburban children with asthma, the homes of inner city Baltimore children with asthma had higher
levels of airborne pollutants (including PM, NO2 and O3 amongst others) and home characteristics that predispose to greater
asthma morbidity. In the inner city homes, median andGM PM10 levels were almost three times as high and the GM PM2.5 levels
were more than three times higher than in the suburban homes. Median GM NO2 andGM O3 levels were found in similar ratios. It is
important to note that PM10 levels were found to be markedly higher in homes on arterial streets compared to those not on arterial
streets. Although standards specific for home indoor air quality have not been established, the authors found that the inner city
children were exposed to home pollutant levels in excess of the Environmental Protection Agency's National Ambient Air Quality
standards.
Smith et al. (2006,156990)
Study Design
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Location U.S.
Trucking industry
working age
Diesel tractors, cigarette smoking, site pollution
Terminal workers had samplers in a special harness; Drivers had a sampling box placed in the cab.
PM2.5
PM2.5; Area samplers in the offices, freight dock, or shop.
PM2.5; Samplers were located in the yard upwind of the terminal.
Elemental carbon (EC); Organic carbon (OC)
Sorensen et al. (2005, 083053)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
panel study
11/1999, 8/2000
Copenhagen, Denmark
Healthy young adults, nonsmokers
20-33 yr, median age = 24yr
No
International Gravity Bureau (BGI, Toulouse; France) (Kenny and Gussman 1997), aKTL; PM2.5 cyclone (Jantunen et al. 1998), a;
BGI400 pump (BGI Inc., Waltham, MA; USA) (flow 4 L/min)
PM2.5
NR
NR
Transition metals (vanadium; chromium, iron, nickel, copper, and platinum)
The 8-oxodG concentration in; lymphocytes was significantly associated with vanadium and chromium concentrations with a 1.9%
increase in; 8-oxodG per 1 Mg/L increase in vanadium and a 2.2% increase in 8-oxodG per 1 Mg/L increase chromium.
Sorensen et al. (2003, 042700)
Study Design Epidemiologic personal exposure study
Period Autumn- November 1999; Winter- January to February 2000; spring- April to May 2000; summer- August 2000
Location Central Copenhagen
Population University students
Age Groups 20-33 yr (median = 24 yr)
Indoor Source NR
Personal Method Particles were sampled using a KTL PM2.5 cyclone developed for the European EXPOLIS study (17), a International Gravity Bureau
400 pump (4l/min), and a battery pack. The equipment was placed in a backpack, which the subjects carried or placed nearby when
they were indoor. Sampling was done on 37-mm Teflon filters.; Urban background concentrations of PM2.5 were measured on the
rooftop of a building (20 meters above the ground) on the Copenhagen University campus.
July 2009
A-281
DRAFT-DO NOT CITE OR QUOTE

-------
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
PM2.5
NR
PM2.5
BS
Personal PM2.5 exposure was found to be a predictor of 8-oxodG in lymphocyte DNA. No other associations between exposure
markers and biomarkers could be distinguished. ETS was not a predictor of any biomarker in the present study. The current study
suggests that exposure to PM2.5 at modest levels can induce oxidative DNA damage and that the association to oxidative DNA
damage was confined to the personal exposure, whereas the ambient background concentrations showed no significant
association.; For most of the biomarkers and external exposure markers, significant differences between the seasons were found.
Similarly, season was a significant predictor of SBs and PAH adducts, with average outdoor temperature as an additional significant
predictor.
Sorensen et al. (2005, 089428)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Copollutant(s)
Primary Findings
Repeated measures cohort study.
Nov 1999-Aug 2000
Copenhagen, Denmark
residents of downtown Copenhagen
20-33 yr old, all non-smokers
Used a questionnaire to get time exposed to environmental tobacco smoke, burning candles, frying food, open windows
wore a backpack, or placed nearby when indoors.
PM2.5 and BS
Bedroom and front door; PM2.5 and BS
Street monitoring station and roof of a campus building; PM2.5 and BS
NO2
For NO2 there was a significant association between personal exposure and the bedroom, the front door and the background,
whereas for PM2.5 and BS only the bedroom and the front door concentrations, and not the background, were significantly
associated with personal exposure. The bedroom concentration was the strongest predictor of all three pollution measurements.
The association between the bedroom and front door concentrations was significant for all three measurements, and the
association between the front door and the background concentrations was significant for PM2.5 and NO2, but not for BS, indicating
greater spatial variation tbr BS than for PM2.5 and NO2. For NO2, the relationship between the personal exposure and the front door
concentration was dependent upon the "season," with a stronger association in the warm season compared with the cold season,
and for PM2.5 and BS the same tendency was seen. Time exposed to burning candles was a significant predictor of personal PM2.5,
BS and NO2 exposure, and time exposed to ETS only associated with personal PM2.5 exposure. These findings imply that the
personal exposure to PM2.5, BS and NO2 depends on many factors besides the outdoor levels, and that information on season or
outdoor temperature and residence exposure, could improve the accuracy of the personal exposure estimation. Regression
coefficients for personal exposure and front door PM2.5 in warm season was 0.67 *, and in the cold season, 0.28. Front door BS in
warm season was 0.86 *, and in the cold season, 0.45.* Front door NO2 in warn season was 0.68 *, and cold season 0.32.*
Sram etal. (2007,188457)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Exposure-Control study 53 policemen (exposed) and 52 age- and sex-matched healthy volunteers (control) were enrolled. Ambient
and PE PM10, PM2.5, and c-PAHs were monitored and chromosomal aberrations were analyzed.
Feb 6-20, 2001
Prague, Czech Republic
Policemen working outdoors in Prague
NA
Personal monitoring using personal samplers (name of instrument not stated)
PM10 PM2.5
NR
PM10 PM2.5
c-PAHs, B[a]P
Ambient air exposure to c-PAHs increased fluorescent in situ hybridization (FISH) cytogenetic parameters in non-smoking policemen
exposed to ambient PM
Srivasta et al (2007,157004)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Exposure assessment of indoor environment
April 5-June 26, 2000
Laboratory in Delhi, India
Building occupants
NR
Re-entrainment of existing dust on floor and other surfaces
No personal exposure assessment was conducted
NR
Suspended PM (SPM)
NR
July 2009
A-282
DRAFT-DO NOT CITE OR QUOTE

-------
Components) metals Ca, Mg. Cu, Zn, Cd, Pb, Cr, Mn, Fe, Co, Ni
Primary Findings Gravimetric analysis and atomic absorption spectrometry results indicated that the suspended PM (SPM) and metal (lead)
concentrations were higher than the National Ambient Air Quality Standards for Delhi, India and SPM standards for residential and
sensitive areas. The maximum concentrations of SPM were observed to be due to penetration of outdoor particles originating from
wind-blown crustal dust and vehicular pollution. Scanning electron microscopy analysis of particles showed dominance of crystalline
silicon and spherical soot particles in samples.
Stein et al. (2002,157008)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Finding(s)
Exposure assessment, evaluation of an aerodynamic particle sizerto accurately measure size-distributed particle mass from
number concentrations
NR
laboratory
PM monitoring devices
NR
No personal exposure assessment was conducted
NR
1.0-13 pm
NR
NR
Significant errors were observed in APS size-distribution measurements with measured mass median diameters (MMAD) as much
as 17 times higher than from cascade impactors. Analysis of APS-correlated time of flight and light scattering data indicated that th
Strand etal. (2007,157018)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Cohort
Winter of 1999-2000; winter of 2000-2001
Denver, Colorado, USA
Asthmatic Children
No
Modeling/Extrapolation from fixed-site ambient monitoring (multiple methods)
No
NR
PM2.5
NR
Using modeled or extrapolated personal ambient PM exposure results in a deattenuation of decrements in FEV1 associated with PM
exposure, relative to use of fixed-site ambient monitoring PM levels. Associations between FEV1 decrements and the various
estimation procedures (modeling and extrapolation) were similar to each other.
Tang at al. (2007,091269)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Cohort Study
12/2003-2/2005
Sin-Chung City, Taiwan
Asthmatic children
6-12 yr
No
Portable particle monitor; DUSTcheck Portable Dust Monitor, model 1.108, GRIMM Labortechnik Ltd., Germany
PM10, PM2.5, PM1, PM2.5-IO, PM1-2.5
NR
PM10, PM2.5, PM2.5-10
NR
Results of linear mixed-effect model analysis suggested that personal PM data was more suitable for the assessment of change in
children's PEFR than ambient monitoring data.
Tatum etal. (2002,157046)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Methodological in this study, the performance of the gravimetric version of the RespiCon was examined in various forest products
industry facilities. The precision of the RespiCon was assessed and its performance was compared with that of both a respirable
cyclone and an inhalable dust sampler. In addition, some RespiCon samples were examined using scanning electron microscopy to
determine physical particle size distribution.
NA
Various forest products industry facilities
occupational
NA
No
NR
July 2009
A-283
DRAFT-DO NOT CITE OR QUOTE

-------
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
NR
Respirable (< PM4 pm), Thoracic (< PM10 um), and Inhalable fractions (all PM) of airborne PM.
NR
NR
The RespiCon is a useful sampling device for those situations in which it is important to simultaneously collect either personal or
area samples of the respirable, thoracic, and inhalable fractions of airborne PM.
Thomaidis et al. (2003,044193)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment (chemical characterization of PM2.5 aerosols, source apportionment)
March 1995-March 1995
Two sites in Athens, Greece 1) Patisson in Athens city center and mainly affected by local traffic; 2) Rentis located in a semi-urban
industrial area 5 km outside city center and mainly influenced by small industries
Urban Populations
NR
NR
No personal exposure sampling.
NR
NR
PM2.5
Pb, Cd, Ni, As
Pb exhibited higher values during the winter, possibly due to increased diesel oil combustion from central heating and motor
vehicles. No seasonal variation was observed for other metals. Annual mean levels of Pb at both sites were below the European
Union guidelines.
Thornburg et al. (2004,157052)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
PM exposure studies RTP PM Panel Study; Tampa Asthmatic Children's Study
RTP summer 2000-spring 2001; Tampa October-November 2002
Research Triangle Park (RTP), NC and; Tampa, FL
Residential home occupants
NR
Yes. Resuspension of PM10 from a carpet was identified as a major source in one home (a trailer), while cooking was identified as a
source in many homes.
20 Lpm Harvard impactors and 2 Lpm Personal Exposure Monitors both with 37 mmTeflo filters and gravimetric analysis.; Also, MIE
pdrl 000 nepholometer.
PM2.5, PM10
NR
PM2.5, PM10
NR
The association of duty cycle with indoor-outdoor (I/O) ratio was confounded by the short time span of ventilation system operation
and the presence of strong indoor sources.
Toivola et al. (2002,026571)
Study Design
Period
Location
Population
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Random sample of teachers
Nov 1998-Mar 1999 and Nov-Dec 1999
2 cities in eastern Finland
Elementary school teachers
Button inhalable aerosol sampler
Particle Mass; BS
Particle Mass; BS
NR
Total fungi, total bacteria, viable fungi, viable bacteria
Personal BS exposure correlated with both home and work BS exposures. BS concentrations explained best the variation of particle
mass in personal and home concentrations.
Tovalin etal. (2006,091322)
Study Design	Biomarker (DNA damage in blood) exposure assessment
Period	Mexico City and Puebla, Mexico
Location	Occupational^ exposed outdoor workers
Population	18-60 years old
Age Groups	NA
Personal Method	Personal integrated filter gravimetric measurement. Questionnaire
Personal Size
Microenvironment Size
PM2.5
NR
July 2009
A-284
DRAFT-DO NOT CITE OR QUOTE

-------
Ambient Size NR
Components) NR
Primary Findings 1) In Mexico City, outdoor workers had greater DNA damage than indoor (median tail length 46.8 vs. 30.1 pm); 2) In Mexico City,
outdoor workers had a greater proportion of cells with high DNA damage (tail length = 41 pm); 3) In Puebla, outdoor and indoor
workers had similar damage; 4) DNA damage was correlated with PM2.5 and ozone exposure; 5) High DNA damage in workers was
associated with ozone, PM2.5, and 1 -ethyl-2-methyi benzene (VOC) exposure.
Tovalin-Ahumada et
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
al. (2007,190165)
Point study
April and May, 2002
Mexico City (Ne and SE) and Puebla, Mexico
Indoor and outdoor workers in large urban areas
18 years of age and older
NR, The exposures described in this report were monitored as part of a larger study directed at evaluating the association between
personal exposures to PM2.5 and VOCs and genetic damage in outdoor and indoor workers reported elsewhere (Tovalin et al.,
2006).
Personal exposures to PM2.5 were monitored using 37mm Teflon filters (Model 225-9006, SKC Inc.), fitted to a single stage personal
impactor (Model PEM-761-203A, SKC) and personal sampling pumps (Model PCXR4, SKC). Two PM2.5 personal air samples
(occupational and nonoccupational) were obtained during a 24-h period for each worker in Mexico City and for the indoor workers in
Puebla. Only one PM2.5 personal sample could be obtained during a 24-h period (an overall exposure) for the bus drivers in Puebla
because of their work shift (from 4AM to 8 PM) with rotating start and end times. At the beginning of the work shift, each participant
was asked to carry a backpack holding the pump; the impactor was attached to the backpack strap, in the breathing zone. At the
end of the work shift, the impactor and pump were removed, and replaced with a new sampling setup that was worn by the worker
until the beginning of the next day work shift.
PM2.5
NR
NR
Si; S; K; Ca; Cl(e); Ti; V; Cr; Mn; Fe; Co; Ni; Cu; Zn; Mo(e); Cd(e); Se; Br; Rb; Sr(e); As; Pb
The results of this study suggest that outdoor workers in both Mexican cities experienced higher exposures to PM2.5 than indoor
workers, and that these exposures are well above either the 35 pg/m3 US-EPA or the 65 pg/m3 24-h Mexican standards for PM2.5.
Both subgroups experienced higher occupational than non-occupational exposures. Mexico City outdoor workers had higher
exposures to Soil, Fuels and Industrial emission-related elements than Puebla outdoor workers did. However, Mexico City outdoor
workers had half the exposure to soil dust-related elements and fuel related elements than Puebla outdoor workers. However the S
exposure was similar in all groups but high, product of the high vehicles density in the areas, responsible for 60% of the emission in
Mexico City (Secretaria del Medio Ambiente, 2005). This study of Mexico City results correlates well with a previous PM2.5
emissions inventory results, which determined that 81.14% of particles are released from mobile sources
Trenga et al. (2006,
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
155209)
Panel study with repeated measures
3 sampling periods Oct 1999-Aug 2000, Oct 2000-May 2001, Oct 2001 -Feb 2002
Seattle, Washington
Adults with and without COPD and children with asthma
adults ages 56-89 and children ages 6-13
NR
Carrying personal monitor (Harvard Personal Environmental Monitor for PM2.5)
PM2.5
PM2.5
Coarse (PM10-PM2.5) and PM2.5 for residential outdoor, PM2.5 for central site
NR
FEV1 decrements associated with 1 -day lagged central site PM2.5 in adult subjects with COPD. Associations between PM and lung
function decrements were significant only in asthmatic children not receiving anti-inflammatory medication. Same day central s
Turpin et al. (2007,157062)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
RIOPA Study 24-h integrated indoor, outdoor, and personal samples collected in 3 cites.
Summer 1991-spring 2001
Elizabeth, NJ, Houston, TX, and Los Angeles County, CA
309 adults and 118 children (89-18)
NR
PEM on the front strap of a harness near the breathing zone. The bag on the hip contained the pump, battery pack, and motion
sensor
PM2.5
PM2.5, in the main living area (not kitchen)
PM2.5, in the front or back yard
18 volatile organics, 17 carbonyl, PM2.5 mass and >23 PM2.5 species, organic carbon, elemental carbon, and PAHs
July 2009
A-285
DRAFT-DO NOT CITE OR QUOTE

-------
Primary Findings The best estimate of the mean contribution of outdoor to indoor PM2.5 was 73% and the outdoor contribution to personal was 26%.
Urch etal. (2004,055629)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
The study was a crossover design in which each participant received a 2-h exposure to filtered air (FA) and CAP+O3, assigned
randomly and on separate occasions. Study objective is to examine the relationship between total and constituent PM2.5 mass
concentrations and acute vascular response.
2000-2001; not explicitly stated
Downtown Toronto, Canada
24 young (35 ±10 yr) healthy, nonasthmatic, nonsmoking people
35 ± 10years
NR
During exposures a sample was collected immediately upstream to the participant on a 47 mm Gelman Teflon filter with a 2 pm pore
size at an airflow of 15 L/min.
PM2.5
NR
PM2.5
Total carbon (elemental carbon, organic carbon), N03-, SO42-. NIV, K+, CI-, Ca, Fe, Al, Mg, Zn, Mn, Pb, Cu, Ba, Se, Cr, Ni, V, Ar,
Cd all have median, min, and max reported for ambient levels. N03-, SO42-. NIV all have median, min, and max reported for
directly measured personal exposure levels. Total carbon (elemental carbon, organic carbon), Ca, K+, Fe, CI-, Al, Mg, Zn, Pb, Mn,
Cu, Ba, Se, Cr, Ni, V, Ar, Cd all have median, min, and max reported for estimated personal exposure
A significant negative association between both the organic and elemental carbon concentrations and the difference in the post-
exposure change in the BAD between CAP+O3 and FA exposure days.
Vallejo etal. (2006,157081)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
4/2002-8/2002
Mexico City, Mexico
Health young, non-smoking adults
Mean age 27 yr
No
pDR nephelometric method (personal DataRam, pDR1200)
PM2.5
NR
NR
NR
Mean personal PM2.5 level was 74 mg/m3
Vallejo etal. (2006,157081)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Pilot Study
April-August 2002
Mexico City, Mexico
Healthy residents of Mexico City
21-40 yr
NR
The participant carried a personal PM2.5 monitor (DataRAM 1200) during a single 13-h period starting at 9:00 a.m. Indoor situations
included activities at home, at work, at school, or in public places such as theaters, stores, restaurants coffee shops, and subway
transportation. Outdoor activities included walking, standing, or sitting in an open space, driving a car or using public transportation
(bus or taxi).
NR
PM2.5
PM2.5
NR
The descriptive analysis showed that overall outdoor median concentration of PM2.5 was higher than the indoor one. In the indoor
microenvironment, the highest concentrations occurred in the subway followed by the school, and the lowest was at home. The
outdoor microenvironment with the highest concentrations was the public transportation (bus), while the automobile had the lowest.
It was found that PM2.5 concentration levels had a circadian-like behavior probably related to an increase in the population daily
activities during the morning hours, which decrease in the evening, especially at indoor microenvironments. The Center city area
was found to have the highest concentrations of PM2.5.; Multivariate analysis corroborated that PM2.5 concentrations are mainly
determined by geographical locations and hour of the day, but not by the type of microenvironment.
van Roosbroeck et al. (2006, 090773)
Study Design Personal exposure assessment, effect of traffic-related pollutants
Period March-June 2003
Location Amsterdam, The Netherlands
July 2009
A-286
DRAFT-DO NOT CITE OR QUOTE

-------
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
Schoolchildren
9-12 yr
Environmental tobacco smoke, cooking
Integrated filter gravimetric measurement. Light absorbance.
PM2.5 absorbance = "soot"" EC (see Notes)
NR
PM2.5 absorbance = "soot"" EC (see Notes)
NR
Children living near busy roads had 35% higher personal exposure to 'soot' than children living in urban background locations.
van Roosbroeck etal. (2006, 090773)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Copollutant(s)
Primary Findings
Exposure assessment
9 months (no year provided)
Utrecht, The Netherlands
School children
10-12 yr
NR
PM2.5 GK2.05 cyclones 4 L/min in a custom made backpack; NO2 and NOX Ogawa passive samplers
PM2.5
NR
PM2.5
NO2
Increased personal exposure to the traffic-related air pollutants soot and NOX was seen in children at the Freeway school. No
increased personal exposure in any of the studied air pollutants was found for children at Ring School.
van Roosbroeck etal. (2006, 090773)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Copollutant(s)
Primary Findings
Exposure assessment
9 months (no year provided)
Utrecht, The Netherlands
School children
10-12 yr
NR
PM2.5 GK2.05 cyclones 4 L/min in a custom made backpack; NO2 and NOX Ogawa passive samplers
PM2.5
NR
PM2.5
NOX
Increased personal exposure to the traffic-related air pollutants soot and NOX was seen in children at the Freeway school. No
increased personal exposure in any of the studied air pollutants was found for children at Ring School.
Verma et al. (2003,157093)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Mlcroenvlronment Size
Ambient Size
Components)
Primary Findings
Task-based exposure assessment of current occupational exposures to chemical agents of Ontario construction workers
June 2000
Ontario, Canada
Ontario construction workers
NR
known source construction activities
Air samples personal sampling pumps and collection media. Direct-reading particulate monitor DustTrak
respirable, inhalable, total, and silica dust; man-made mineral fibers (MMMF)
NR
NR
NR
Authors state "Ontario construction workers are exposed to potentially hazardous levels of chemical agents."
Vinzents et al. (2005, 087482)
Study Design	Panel Study
Period	3/2003-6/2003
Location	Copenhagen, Denmark
Population	Healthy young adults
Age Groups	Mean age = 25 yr
Indoor Source No
Personal Method
July 2009
A-287
DRAFT-DO NOT CITE OR QUOTE

-------
Personal Size Ultra-fine particles (UFP) condensation particle counters; (TSI 3007; TSI, St. Paul,
Microenvironment Size UFP (10-100 nm)
Ambient Size PM10
Primary Findings	NR
, USA)
Wallace (2005,157102)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure Assessment (Indoor, outdoor air monitoring for concentration of ultrafine particles. To determine indoor source of ultrafine
particles-determine the contribution of vented gas clothes dryer)
Not specifically stated. An 18 month period including 2000.
NR
NR
NR
Vented gas clothes dryer
Scanning mobility particle sizer, differential mobility analyzer, condensation particle counter
NR
Ultrafine (PM 0.01-0.45)
Ultrafine (PM 0.01-0.45)
NR
Vented gas clothes dryer was determined to be a major source of indoor ultrafine particles. It consistently produced an order of
magnitude increase in ultrafine particle concentration compared to times when there was no indoor source.
Wallace and Williams (2005,057485)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Cohort
2000-2001
Raleigh, North Carolina
African-American persons with elevated risk from exposure to particles.
NR
NR
PEM PM2.5 monitor
PM2.5
Indoors PM2.5
Outdoors near residence PM2.5 PM2.5
Sulfur
Using outdoor particles to determine the effect on health is not accurate. The infiltration factor is a good estimator for personal
exposure. Indoor and outdoor measurements of sulfur could be used in the absence of personal exposure measurement to
estimate the contribution of outdoor fine particles to personal exposures.
Wallace etal. (2006,088211)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Time series continuous monitoring of subjects with controlled hypertension or implanted defibrillators were monitored for 7
consecutive days in 4 seasons.
2000-2001
North Carolina, probably near Research Triangle Park
Health-compromised adults, non-smokers
Adults [range not specified]
Cooking, cleaning, personal care, smoking
PEM
PM2.5
PM2.5; Indoor and outdoor
NR
NR
Use of continuous particle measuring instruments allowed more precise identification of sources, frequency and magnitude of short-
term peaks, and more accurate calculation of individual personal clouds.
Wallace etal. (2003,053553)
Study Design	Inner City Air Pollution (ICAP) study- Randomized controlled trial
Period	NR
Location	Bronx, NY; Manhattan, NY; Boston, MA; Chicago, IL; Dallas, TX; Seattle, WA; Tucson, AZ
Population	Asthmatic children and their residences
Age Groups	5-11 yr
Indoor Source	Combustion-related particles smoking, cooking, use of a wood stove or fireplace, use of candles or incense, gas or kerosene space
heaters or stoves.
Personal Method	NR
July 2009
A-288
DRAFT-DO NOT CITE OR QUOTE

-------
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
NR
0.10 pg-5 pg; (see note below)
PM2.5
NR
Geometric mean values of indoor concentrations in the seven locations differed by less than a factor of 2, and the shape of the
distributions was very similar across cities, both for the nominal 2-week averages and for hourly averages. The hourly averages
exceeded 100 pg/m3 for at least 2% of all measurements in all cities, and exceeded 1,000 pg/m3 on at least a few occasions in each
city. The most important particle source in these homes was smoking. A second, less powerful source was cooking, particularly
frying/ sauteing or reporting a smoky cooking event. Use of incense also led to significant increases in particle concentrations.
Dusting frequently also led to higher concentrations, possibly considerably higher than indicated by the pDR because of its lack of
sensitivity for coarse particles. Infiltration of outdoor air added about half of the outdoor air concentration to the concentrations
produced by the indoor sources, a result similar to that found by previous studies. Most of the observed variance in indoor
concentrations was day to day, with roughly similar contributions to the variance from visit to visit and home to home within a city
and only a small contribution made by variance among cities. The small variation among cities and the similarity across cities of the
observed indoor air particle distributions suggest that sources of indoor concentrations do not vary considerably from one city to the
next, and thus that simple models can predict indoor air concentrations in cities having only outdoor measurements. A new finding
from this study was the observation that concentrations of fine particles peak in the late evening in homes with smoking, perhaps
reflecting the influence of after dinner smoking.
Wang et al. (2006,157108)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment, identification of sources of outdoor and indoor PM and trace elements
Aug 4-Sep 10, 2004
Guangzhou, China
4 hospitals
NR
NR
No personal exposure assessment was conducted.
NR
PM10, PM2.5
PM10, PM2.5
Trace elements Na, Al, Ca, Fe, Mg, Mn, Ti, K, V, Cr, Ni, Cu, Zn, Cd, Sn, Pb, As, Se
High correlation between PM2.5 and PM10 suggest that they came from similar emission sources. Outdoor infiltration could lead to
direct transportation of PM indoors. Human activities and ventilation types could also influence indoor PM. levels.
Ward et al. (2007,157112)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Study Design
Period
Location
Population
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Indoor air sampling to determine size fractionated concentrations of PM, OC, EC, and total carbon
Jan-Mar 2005
Libby, Montana
Children exposed to wood-burning stoves in elementary and middle schools
Burning wood in stoves for heating
NR
NR
PM >2.5,1.0-2.5, 0.5-1.0, 0.25-0.5, and < 0.25 pm
PM >2.5,1.0-2.5, 0.5-1.0, 0.25-0.5, and < 0.25 pm
Organic carbon (OC) and elemental carbon (EC) in 5 size fractions >2.5,1.0-2.5, 0.5-1.0, 0.25-0.5, and < 0.25 pm
Total measured PM mass concentrations were much higher inside the elementary schools, with particle size fraction (>2.5, 0.5-1.0,
0.25-0.5, and < 0.25 mm) concentrations between 2 and 5 times higher when compared to the middle school. The 1.0-2.5 mm
fraction had the largest difference between the two sites, with elementary school concentrations nearly 10 times higher than the;
middle school values. Weichenthal et al. (2006)
Cross-sectional survey comparing heating systems
Dec2005-Mar 2006
Montreal, Quebec, Pembroke, Ontario, Canada
NR
Yes, by questionnaire on age/size of home, cleaning frequency, type of stove and other cooking appliances, use of kitchen exhaust
fan, number of smokers, burning candles, use of candles, portable heaters, natural gas clothes dryer.
NR
NR
Ultrafine Particles < 100 nm in diameter, PM4
NR
Ultrafine Particles < 100 nm in diameter
Weiseletal. (2005,157131)
Study Design Matched indoor, outdoor, and personal concentrations in proximity to pollution sources.
Period May 1999-Feb 2001
July 2009
A-289
DRAFT-DO NOT CITE OR QUOTE

-------
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Elizabeth, NJ, Houston, TX, and Los Angeles County, CA
urban children and adults
Child 6-19 yr; adult 17-89 yr
Age of house, recent renovations (< 1 yr), type of home (single, multiple family), attached garage, carpet indoors, local pollution
sources.
PEM on a harness with inlet near breathing zone.
PM2.5
PM2.5
PM2.5
NR
Personal PM2.5 was significantly higher than indoor and outdoor by one-way ANOVA and Sheffe test (p < 0.001).
Wichmann et al. (2005, 086240)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment; Ambient (indoor); Personal
November 29,1993-March 30,1994; October 17,1994-December 22,1994
Amsterdam, The Netherlands
Adults and schoolchildren living near high-traffic or low-traffic roads.
Adults (50-70 yr); Schoolchildren (10-12 yr)
No
Personal impactor
Absorbance coefficient measurements of PM10 filter samples
Absorbance coefficient measurements of PM10 filter samples
Absorbance coefficient measurements
NR
Found tentative support for using type of road as a proxy for indoor and personal exposure to traffic-related absorbance (PM).
Williams etal. (2003,053338)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Cohort study, longitudinal
Summer 2000, fall 2000, winter 2001, and spring 2001
SE Raleigh, North Carolina; Chapel Hill, North Carolina
Elderly persons
50 yr
NR. While no smokers were enrolled into the study, 18 participants occasionally recorded passive exposures to environmental
tobacco smoke. Since this study attempted to determine the effects of ambient PM upon personal and residential settings, and ETS
exposures typically overwhelm ambient contributions, gravimetric values believed to have been heavily influenced by ETS were
excluded from the analysis.
A number of filter-based PM monitors widely used in other PM studies were employed here as described below in Table 1. A nylon
vest, matched to the body size of the participant, was used to support and retain all of the personal monitoring equipment. All of the
personal monitoring equipment was located in the participants breathing zone (chest area) with the exception of the nephelometer
which was secured to the front pocket of the vest with the inlet fully exposed. Each participant was asked to wear the vest at all
times with the exception of sleeping, bathing or the changing of clothes. In those instances, they were asked to secure the vest on
nearby furniture or fixture. A local State of North Carolina AIRS monitoring platform in Raleigh, NC was selected to serve as the
ambient monitoring site.
PM2.5
PM2.5; PM10; PM10-2.5
PM2.5; PM10; PM10-2.5
NR
No statistical difference in personal PM2.5 concentration exposures existed between the two cohorts. Neither seasonality nor
community settings were determined to be critical factors in aggregate personal PM2.5 exposures in the two subpopulations. PM2.5,
and to a lesser extent PM10, mass concentrations were determined to be generally homogeneous across a large spatial area. The
lack of a seasonal effect observed in the RTP was unexpected due to the historically higher PM2.5 levels observed in central North
Carolina during the spring and summertime when automotive traffic is highest and regional power plant demands for electricity are
greatest (and subsequent release of emissions). PM2.5 personal cloud estimates in the current study were in agreement with those
observed in other PM studies involving susceptible subpopulations having more sedentary lifestyles. Mean personal PM2.5
exposures in the current study had a moderate Pearson correlation relative to ambient or residential outdoor mass concentrations i
Wilson and Zawar-Resa (2006,088292)
Study Design	Exposure assessment using Advanced Dispersion Modeling to estimate long-term personal PM exposures in small areas within £
city
Period	July 2003 and June 2004-2 winter months
Location	Christchurch, New Zealand
Population	urban environments
Age Groups	NR
Indoor Source	NR
July 2009
A-290
DRAFT-DO NOT CITE OR QUOTE

-------
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
No personal exposure assessment was conducted
NR
NR
PM,o
NR
Despite the area's high intraurban PM concentration variability and meteorological and topographical complexity, the model
performed satisfactorily overall, except for the Mount Pleasant site. The mean of observed measurements across all sites was close
Wilson and Brauer (2006,088933)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Exposure assessment; Ambient (outdoor); Ambient (indoor infiltrating from outdoors); Non-ambient (indoor from indoor sources and
"personal cloud')
April-September 1998
Vancouver, Canada
Subjects with physician-diagnosed COPD
54-86-years-old
No
Personal integrated filter gravimetric measurement; TEOM outdoor ambient
PM2.5
NR
NR
NR
1) Measured ambient PM2.5 exposure comprised 71 % ambient PM2.5 exposure was 71 % of measured ambient concentration and
44% of measured total personal exposure; 2) Non-ambient exposure was independent of ambient exposure; 3) Pearson
correlations of longitudinal estimated ambient exposure with ambient concentration averaged 0.88 (0.77-0.92).
Wu etal. (2006,157156)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Panel study/exposure assessment
9/3/2002-11/1/2002 (The fall agricultural burning season
Pullman, WA
Asthmatic adults
(mean age = 27y/o, min = 18, max = 52 y/o)
No
Yes, using two co-located Harvard; Personal Environmental Monitors (HPEM2.5; Harvard School of Public Health, Boston, MA),
each connected to its own pump (BGI; AFC 400S, Waltham, MA) operated at 4 L/min
PM2.5
PM2.5
PM2.5
Levoglucosan (LG); Elemental Carbon (EC); Organic Carbon (OC)
Wu etal. (2005,086397)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Panel study with repeated measures
1999-2000
Alpine, CA
Asthmatic children
9-17 yr
No
pDR, continuously and 1 -min concentrations (passive), in a fanny pack.
PM2.5
PM2.5, Home inside & home outside
PM2.5
NR
Personal exposure was higher than those at fixed sites. Subjects received only 45.0% of their exposure indoors at, even though they
spent more than 60% of their time there. In contrast, 29.2% of their exposure was received at school where they spent only 16.4%
of their time. Thus, exposures in microenvironments with high PM levels where less time is spent can make significant contributions
to the total exposure.
Wu etal. (2005,157155)
Study Design
Location
Population
Age Groups
Personal Size
Microenvironment Size
Modeling of individual exposure using ambient data from a 10-yr longitudinal study.
Southern California Lancaster, San Dimas, Upland, Mira Loma, Riverside, Long Beach and Lake Elsinore.
Children
NR
No measurements presented in this study
NR
July 2009
A-291
DRAFT-DO NOT CITE OR QUOTE

-------
Yeh and Small (2002,040077)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Comparative assessment of AME and IES models
1997 (364 days) spring March-May, summer June-August, Fall September-November, winter December-February
Los Angeles County, CA
General population; ETS and Non ETS Homes
NR
Indoor Cooking, ETS, Other sources and unexplained particulates that maybe generated with engaging in various activities
NR
PM10 PM2.5
NR
PM10 PM2.5
NR
Adjusting from outdoor concentrations to personal exposures and correcting dose-response bias are nearly equal. Roughly the
same premature mortalities associated with short-term exposure to both ambient PM2.5 and PM10 are predicted by both models
Yeh and Small (2002,040077)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
December-February
Comparative assessment of AME and IES models
1997 (364 days) spring March-May, summer June-August, Fall September-November, winter
Los Angeles County, CA
general Population; ETS and Non ETS Homes
NR
Indoor Cooking, ETS, Other sources and unexplained particulates that maybe generated with engaging in various activities
NR
PM10 PM2.5
NR
PM10 PM2.5
NR
Adjusting from outdoor concentrations to personal exposures and correcting dose-response bias are nearly equal. Roughly the
same premature mortalities associated with short-term exposure to both ambient PM2.5 and PM10 are predicted by both models
Yeh and Small (2002,040077)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Comparative assessment of AME and IES models
1997 (364 days) spring March-May, summer June-August, Fall September-November, winter December-February
Los Angeles County, CA
General population; ETS and Non ETS Homes
NR
Indoor Cooking, ETS, Other sources and unexplained particulates that maybe generated with engaging in various activities
NR
PM10
NR
PM10
NR
Adjusting from outdoor concentrations to personal exposures and correcting dose-response bias are nearly equal. Roughly the
same premature mortalities associated with short-term exposure to both ambient PM2.5 and PM10 are predicted by both models
Yip etal. (2004,157166)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Periods. Children were stratified into
A panel study with repeated measures with personal & home monitoring for 8 2-week
smoking and non-smoking households.
2000-2001
Detroit, Michigan
School-age children with asthma
7-11 yr
PEM in a backpack
PM10
PM10; indoor at home & indoor at school
PM10
NR
Personal PM concentrations were significantly correlated with home environment (r = 0.38 to 0.70), with the strongest relationships
in home with non-smokers.
Zhang etal. (2005,157185)
Study Design Several co-located instruments were used to simultaneously sample air at the Pittsburgh EPA Supersite for 15 days
July 2009
A-292
DRAFT-DO NOT CITE OR QUOTE

-------
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
7-22 Sep 2002
Pittsburgh, Pennsylvania
urban Population
NR
NR
NR
NR
NR
Nonrefractory-PM1
Sulfate, Ammonium, Nitrate, Organics, Chloride
Reasonably good agreement was observed on particle concentrations, composition, and size distributions between the AMS data
and measurements from co-located instruments (given the difference between the PMiand PM2.5 size cuts), including TEOM,
semicontinuous sulfate, 2-h- and 24-h-averaged organic carbon, SMPS, 4-h-averaged ammonium, and micro-orifice uniform
deposit impactor.
Zhao etal. (2006,156181)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Aerosol source apportionment under four environments (personal, residential indoor, residential outdoor and ambient) to evaluate
the relationship between different environments through exposure analysis, and to demonstrate the utility of the combined receptor
model on air quality studies of various environments.
June 2000 to May 2001
Raleigh and Chapel Hill, NC
NR. People with respiratory ailments most likely.
NR
Yes (4 main sources to residential indoor PM Cu-factor mixed with indoor soil, secondary sulfate, Personal care and activity, ETS
and its mixture)
Personal Exposure Monitors (PEM) and Harvard Impactor monitors (HI)
NR
NR
NR
OC, EC, and elements
As per the authors "Secondary sulfate was the largest source for both residential outdoor and ambient PM. Cooking and personal
care activity were two major internal sources for personal and residential indoor PM samples. In this study, secondary sulfate and
motor-vehicle emission contributed significantly to the personal and residential indoor PM.
Zhao etal. (2007,156182)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Components)
Primary Findings
Comprehensive analysis of the sources of PM15 exposure on children with moderate to severe asthma in urban-poor settings.
Two winter periods (October 2002-March 2003 and October 2003-March 2004)
Elementary school for children with significant asthma, Denver, CO
Schoolchildren in urban-poor settings suffering from moderate to severe asthma
6-13 yr (60% in the range 10-13 yr, rest in the range 6-9 yr)
Yes, House cleaning compounds, and smoking were identified as primary internal sources.
Personal Exposure Monitor (PEM)
PM2.5
PM2.5
PM2.5
EC, CI, Si, N03
Four external sources and three internal sources were resolved in this study. Secondary nitrate and motor vehicle were two major
outdoor PM2.5 sources. Cooking was the largest contributor to the personal and indoor samples. Indoor environmental tobacco
smoking also has an important impact on the composition of the personal exposure samples.
Zhuetal. (2005,157191)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Components)
Primary Findings
4 apartments near the freeway were monitored at 2 times for 6 consecutive days, 24 h per day. Subjects did not enter the bedrooms
where the samplers were, no cooking, cleaning, children, or pets.
Oct. 2003-Dec. 2003 and Dec. 2003-Jan. 2004
Los Angles, CA
Urban Populations near major freeways.
NR
NR
NR
Indoor and Outdoor ultrafine particles (6-220 nm)
NR
CO
The size distributions of indoor aerosols showed less variability than the adjacent outdoor aerosols. Indoor to outdoor ratios for
ultrafine particle concentrations depended strongly on particle size. I/O ratios were dependent on the indoor ventilation mechanisms
July 2009
A-293
DRAFT-DO NOT CITE OR QUOTE

-------
applied. Size-dependent particle penetration factors and deposition rates were predicted from data by fitting a dynamic mass
balance model.
Zollner etal. (2007,157192)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Findings
PM exposure was investigated in and outside of schools
Winter Period of 2005 and 2006
Baden-Wuerttemberg, Germany
School children
NR
No personal monitoring done. PM2.5 was collected with filter device LVS 6.01 and analyzed gravimetrically; fifteen particle fractions
(0.30 pm to >20 pm) were recorded with laser particle counters; >0.02 pm particles were recorded using condensation particle
counters
NR
They only reported concentrations for PM2.5. PM 0.02 to >20 were collected and analyzed but only PM2.5 concentration were
reported.
They only reported concentrations for PM2.5. PM 0.02 to >20 were collected and analyzed but only PM2.5 concentration were
reported.
1) The impaction of PM was strongly influenced by specific weather conditions; 2) Time resolution of measurements in classrooms
showed variation in particle concentration depending on the type of building and indoor activities; 3) Concentrations of very small
particles indoors and in ambient air measured by condensation particle counter were influenced by traffic emissions.
Table A-53. Examples of studies showing developments with UFP sampling methods since the 2004
PM AQCD.
Reference
PM Size
Ranges
PM
Constituents
Instruments
Primary Findings
Biswas et al.
(2005,150694)


CPC (water)
Water-based CPC performance eval
Feldpausch et al.
(2006,155773)
20-100
nm
Carbonaceous
aerosols
DS with CPC, compared with DMA
The DS with CPC compared fairly well with the DMA for particle sizes up to 40 nm
with 20 - 40% underestimation depending on discharge frequency settings. The
DS sampling period is 3 - 5 s in comparison with the 1 min scanning time of the
DMA.
Hering et al.
(2005,155838)


CPC (water)
Water-based CPC performance eval
Hermann et al.
(2007,155840)
3 - 40 nm
Ag, NaCI
CPC (water and butanol)
Roughly 95% collection efficiency for d >5 nm for TSI models 3776 and 3786,
95% efficiency for d > 20 nm for model 3775, near 90% efficiency for d > 20 nm
for model 3785, near 90% efficiency for d > 25 nm for model 3772.
Kinsey et al. 10nm-5 DE
(2006,130654) pm
TE0M, SMPS, CPC, DustTrak, E- TE0M best comparison with gravimetric filter among mass concentration
BAM, ELPI, integrated filter samples analyzers, ELPI and SMPS comparable for differential number distribution but ELPI
not useful for gravimetric analysis because mass is not significant at small end of
distribution.
Kulmala et al.
(2007,097838)

CPC
Changing temperature difference between saturator and condenser within CPC
allowed for differences in cut-off diameters.
Kulmala et al.
(2007,155911)
2-20nm Atmospheric
aerosol, Ag
Battery of CPCs (water, butanol, n-
butanol)
Used the battery to discriminate between water-soluble, water-insoluble, butanol-
soluble, and butanol-insoluble nucleation-mode particles
Ntziachristos 7 nm - 1 Automobile
and Samaras |jm exhaust
(2006,116722)
5 instruments used simultaneously to
reduce uncertainty: Teflon-coated
filter downstream of constant volume
sampling, ELPI with thermodenuder,
CPC, SMPS, diffusion charger
Use of four reduced variables combining output from all instruments (ratio of
particle number concentration from CPC and ELPI, estimated mean geometric
mobility diameter from signal of diffusion charger and number concentration from
CPC, ratio of signal of diffusion charger to constant volume sampler mass, ratio of
constant volume sampler mass to volume collected by ELPI) resulted in
identification of clear outliers and factors related to driving and fuel properties
rather than measurement errors.
Olfert et al. 30 - 100 NaCI, ambient
(2008,156004) nm
FIMS (compared with SMPS)
Particle number concentrations reported by the FIMS were 8 - 23% higher than
the SMPS using an inversion technique designed to correct for particle residence
time in the FIMS, which operates at 0.1 s resolution.
July 2009
A-294
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
PM Size PM
n p » Instruments
Ranges Constituents
Primary Findings
Petaja et al.
(2006,156021)
CPC (water)
Water-based CPC performance eval
Winkler et al. 1.5-4 Tungsten oxide CPC (n-Propanol)	Authors remove excess charge on particles with ion trap to detect particles down
(2008,156160) nm	to — 1 nm (by eliminating electrostatic attraction to agglomerate).
Table A-54. Summary of in-vehicle studies of exposure assessment.
Reference
Study Design
Mode of Transport
Exposures
Primary Findings
Rossneret al.
(2008,156927)
Measured PM2.6 exposure of 50
city bus drivers and 50 controls in
Prague, Czech Republic using
personal samplers (type not
specified) and VOCs using passive
samplers. PM2.6 filters analyzed for
c-PAHs. Focus of study is oxidative
stress biomarkers in drivers. Study
period: winter 2005, summer
2006, winter 2006.
Bus
Units: ng/m3
winter 2005:

Bus
c-PAH
7.1
B[a]P
1.3
summer 2006:


Bus
c-PAH
1.8
B[a]P
0.2
winter 2006:


Bus
c-PAH
CJ1
B[a]P
1.0
c-PAH and B[a]P exposure to bus
drivers was significantly higher in
Winter 2006, but control exposure
was significantly higher in Winter
2005 for c-PAH and B[a]P and in
summer 2006 for c-PAH. No
significant difference in V0C
exposure between bus drivers and
controls was observed. Oxidative
stress markers were significantly
higher in bus drivers than controls
for all seasons.
Fruin et al. On-road zero emissions vehicle
(2008,	driven on 33-mi arterial road and
097183); 75-mi freeway was equipped
Westerdahl et measured UFP (CPCs, SMPS, EAD),
al. (2005, BC (aethalometer), NOx
086502) [Note: (chemiluminescence), PM-bound
same data PAHs (UV-photoionization), CO (~-
presented.] Trak). DVD analysis of traffic
density and car speed. Study
Period: Feb-Apr 2003 for 2- to 4-h
periods.
Car
Arterial range of medians:
UFP (1000p/cm3) 13-43
PM2.6 (jjg/m3) 7.9-45
BC (jjg/m3)	0.74-3.3
Freeway range of medians:
UFP (1000p/cm3) 47-190
PM2.6 (jjg/m3) 25-110
BC (jjglm3)	2.4-13
Measurements of freeway UFP, BC,
PM-bound PAH, and NO
concentrations were roughly one
order of magnitude higher than
ambient measurements. Multiple
regression analysis suggests these
concentrations were a function of
truck density and total truck count.
(Only PM measurements reported
here).
Briggs et al.
(2008,156294)
UFP (P-Trak) and PM10, PM2.6, and
PM1 (OSIRIS light scatter) were
operated in a car while driving or
walking on one of 48 routes in
London. Trips ranged 1.5-15 min by
car and were repeated up to 5
times to improve statistics. Study
Period: Weekdays in May and June
2005.
Car
Units: PM1 - PM10 (jjg/m3), UFP (p cm-3)
Walking Avg Car Exposure:
PM10	5.87 (3.09)
PM2.6	3.01(1.10)
PM1	1.82(1.10)
UFP	21639(14379)
Avg Walking Exposure:
PM10
PM2.6
PM1
UFP
27.56 (13.16)
6.59(3.12)
3.37 (3.40)
30334 (17245)
ln-car concentrations of PM2.6,
PM1, and UFP correlated well with
walking concentrations
(R - 0.806,0.800, 0.799
respectively). Avg walking
concentrations were 1.4 - 4.7
times higher than avg in-car
concentrations. Cumulative walking
exposures (not shown here) were
4.4 - 15.2 times higher than those
in cars, likely resulting from longer
transit times for walking.
July 2009
A-295
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Study Design
Mode of Transport
Exposures

Primary Findings
Gomez-Pe rales
PM2.E (personal filter pump), CO
Bus
Units: PM2.6 mass (|jg|m3), components
(% of mass)
Buses and minibuses had similar
et al. (2007,
(T15 electrochemical cell), and
Minibus
Bus:

concentration levels for PM2.6
138816)
benzene (canister) were measured
Metro

mass, and metro exposures were

on transit routes, and PM2.6 filters
PM2.6
20-58
lower. CO and benzene

were analyzed for mass, OC/EC,

(NH4O3
5-8
concentrations were higher on

SO42', NO3', and trace metals.

(NH4)2S04
10-18
minibuses than buses. OC was the

Study period: 3-h morning and

OC
17-39
largest PM constituent for all

evening rush hour Jan - March

modes of transport. Measured

2003

EC
8-20
concentrations were higher in the



Crustal
15-18
morning than in the evening rush



Non-crustal
2-3
hour periods. Maximum historical



wind speeds (1995-2003) appeared



Unknown
6-24
to be inversely associated with





measured concentration.



Minibus:





PM2.6
25-55




(NH4O3
4-13




(NH4)2S04
7-22




OC
22-37




EC
9-19




Crustal
12-13




Non-crustal
3-3




Unknown
4-26




Metro:





PM2.6
24-41




(NH4O3
5-8




(NH4)2S04
10-21




OC
35-42




EC
9-13




Crustal
10-16




Non-crustal
2-4




Unknown
5-20

Diapouli et al.
UFP (CPC) concentrations were
"In-vehicle" (not specified)
15-min median (1000p/cm3):

In-vehicle UFP concentrations were
(2007,156397)
measured at school, residential,

School indoor
13.6
roughly 3.5 - 7 times higher than

and in-vehicle environments in



school or residence concentrations.

Athens, Greece. Study Period:

School outdoor
16.6
Indoor concentration diel patterns

school hours, Nov 2003 - Feb

Residence indoor
11.2
were also shown to follow outdoor

2004 and Oct - Dec 2004

Residence outdoor
24.0
levels, which suggests that indoor



In-vehicle
78.0
levels are of outdoor origin.



Gulliver and
TSP, PM10, PM2.6, and PM1
Car
Mean cone (jjg/m3):

Walking exposures larger than car
Briggs (2007,
sampled (OSIRIS light-scatter
Walk

Walk
and background, and car exposures
155814)
devices) in a car while driving or
TSP-PMio
19.1 (19.8)
were generally larger than

walking on one of 48 routes in

background except for PM1. Peak

London. Trips ranged 1.5-15 min by

PM10 2.5
22.1 (22.8)
exposures during walking were

car and were repeated up to 4

PM2.6 1
10.9(10.4)
significantly higher than peak in-car

times to improve statistics. Study

PM1
4.8 (3.4)
exposures.

Period: Jan - Mar 2005.




July 2009
A-296
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Study Design
Mode of Transport
Exposures
Primary Findings
Sabin et al. BC (aethalometer), particle-bound
School bus (diesel, diesel with
In-bus mean concentration, Units: BC (jjg/m3)
, PAH (ng/m3)
Mean concentrations on diesel
(2005, 088300) PAH (UV-photoionization), and NO
particle trap (TO), compressed
Windows closed:

buses without newer emissions
(luminol reaction) were measured
gas (CNG))
BC
control technologies were 2 - 4.4
on 3 diesel school buses, 1 diesel


times higher than background. On
school bus with a particle trap, and

BG
2.5
buses with particle traps,
one compressed gas bus during

CNG
2.3
concentrations were 1.2 - 2.5
before- and after-school commutes.

TO
7.1
times higher than background,
Study Period: May - June 2002.

while concentrations on


diesel
11
compressed gas-fueled school
Windows open:
buses were actually lower than
background.

BC
BG
1.9
CNG
1.5
TO
2.3
diesel
3.9
Gulliver and
PMio, PM2.6, and PM1 sampled
Car

Walk
ln-car PM10 concentrations were
Briggs (2004,
(OSIRIS light-scatter devices) in a
Walk
PM10
38.2
elevated compared with walking
053238)
car while driving or walking on
PM2.6
15.1
and background. PM2.6 and PM1

northern corridor of Northhampton

concentrations were comparable

UK. Study Period: 1-h interval of

PM1
7.1
for walking and background.

morning and evening rush hour



Periods of elevated PM2.6 compared

during Winter 1999 - 2000.



with PM10 generally corresponded
to times when S042~ levels were
also high.
Gomez-Pe rales
PM2.E (personal filter pump), CO
Bus
PM2.6 (jjg/m3):

Generally, PM2.6 concentration was
et al. (2004,
(T15 electrochemical cell), and
Minibus
Bus
68
higher in the morning than evening
054418)
benzene (canister) were measured
Metro
rush hour, but variability was

on transit routes, and PM2.6 filters
Minibus
71
higher for minibuses than other

were analyzed for mass, OC/EC,

Metro
61
modes of transport. Wind speed

SO42, NO3', and trace metals.



was found to be associated with

Study period: 3-h morning and



PM2.E concentration on minibuses.
evening rush hour May - June
2002
July 2009
A-297
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-55. Summary of personal PM exposure studies with no indoor source during 2002-2008.
Reference / Location
Personal
Micro
Ambient
SOUTHWEST
Delfino et al. (2006,
090745)
Riverside and Whittier,
California
Method: PEM
Riverside:
n
24-h PM2.6
1-h max PM2.6
8-h max PM2.6
Whittier:
n
24-h PM2.6
1-h max PM2.B
8-h max PM2.6
13
32.78(21.84)
97.94 (70.29)
47.21 (30.0)
32
36.2(21.84)
93.63 (75.19)
51.75(36.88)
Method: FRM
Riverside:
24-h PM2.6
(23.46)
24-h PM10
(29.36)
Whittier:
24-h PM2.6
(12.14)
24-h PM10
(16.6)
36.63
70.82
18.0
35.73
Delfino et al. (2004,
056897)
Alpine, California
Method: pDR
Last 2-h PM2.6
(33.7)
Diurnal PM2.6
(31.6)
Nocturnal PM2.6
(13.6)
1-h max PM2.6
(120.3)
4-h max PM2.6
(55.3)
8-h max PM2.6
(39.0)
24-h PM2.6
(19.9)
34.4
55.7
22.3
151.0
87.5
67.6
37.9
Method: HI
Indoor 24-h PM10
3(11.9)
Indoor 24-h PM2.6
1 (5.4)
Outdoor 24-h PM10
9(10.4)
Outdoor 24-h PM2.6
0 (5.4)
30.
12.
25.
Method: TE0M
Diurnal PM10
5.1 (11.3)
Nocturnal PM10
3.3	(8.4)
1-h max PM10
4.4	(13.8)
4-h max PM10
4.5(12.4)
8-h max PM10
9.8(11.2)
24-h PM10
3.6(9.1)
24-h PM2.B
0.3(5.6)
Wu et al. (2005,157155) Method: pDR
Alpine, CA	n
Avg of 24-h PM2.6
11
11.4 (7.8)
Method: pDR
n
Avg of 24-h PM2.6
Method: HI
n
Avg of 24-h PM2.6
14
Method: pDR
n
5.6 (2.9) Avg of 24-h PM2.1,
(11.4)
14
9.8(2.5)
Method: HI
n
Avg of 24-h PM2.6
(7.8)
14.0
14.3
July 2009
A-298
DRAFT-DO NOT CITE OR QUOTE

-------
Reference / Location
Personal
Micro
Ambient
Turpin et al. (2007,
157062)
Los Angeles County, CA
(and Elizabeth, NJ,
Houston, TX)
Method: PEM
Avg of 48-h PM2.6
Child
Adult
40.2
29.2
Method: HI
Avg of 48-h PM2.6:
16.2
Method: HI
Avg of 48-h PM2.6:
19.2
NORTWEST
Jansen et al. (2005,
082236)
Seattle, Washington, USA
Method: PM
Results
Method: HI
Indoor home:
PM10
PM2.6
Method: HI
PM10
11.93 PM2.6
7.29
18.0
14.0
Outdoor home:
PM10
PM2.6
13.47
10.47
Mar et al. (2005, 087566) Method: HI
Seattle, WA USA	PM2.1,:
Method: HI
PM2.6:
PM10:
Healthy:
7 (7.8)
CVD:
2(11.3)
COPD:
1 (6.6)
Method: HI
PM2.6:
Healthy:
9.3 (8.4)
Healthy:
7.4 Healthy:
9.0
CVD:
10.8(8.4)
(4.8)
(4.6)

COPD:
10.5(7.2)
CVD:
9.5 CVD:


(6.8)

12.


COPD:
8.5 ? (7 9)



(5.1)
COPD:
9.2
12.
16.
14.
(5.1)
PM10:
Healthy:
5 (7.0)
CVD:
0 (9.0)
COPD:
3 (6.8)
14.
18.
14.
Wu et al. (2006,157156) During non-burning times: 13.8 (11.1)
Pullman, WA	During burning episodes: 19.0 (11.8)
July 2009
A-299
DRAFT-DO NOT CITE OR QUOTE

-------
Reference / Location
Personal
Micro
Ambient
Trenga et al. (2006,
155209)
Seattle, Washington
Method: PEM
Median PM2.6
Child 11.3
Adult 8.5
Method: HI
Median PM2.6
Child
Adult
Method: HI
Residential Outdoor
7.5	Median PM2.6
7.6	Child
Adult
Residential Outdoor
Median PMcoarse
Child
Adult
Residential Outdoor
Median PM2.6 central site
Child
Adult
9.6
8.6
4.7
5.0
11.2
10.3
Koenig et al. (2003,
156653)
Seattle, WA
13.4 ± 3.2/yg/m3
Inside homes - 11.1 ±4.9
Outside homes — 13.3 ± 1.4
3 Central-sites - 10.1 ± 5.7
Liu S et al. (2003,
073841)
Seattle, WA
Summary of PM concentrations (//g|m3) between October Summary of PM concentrations (//g|m3)
Summary of PM concentrations (//g|m3)
1999 and May 2001 by study group.
Group Mean ± SD Personal PM2.6
COPD 10.5 ± 7.2 Healthy 9.3 ± 8.4 Asthmatic
13.3 ± 8.2 CHD 10.8 ±8.4
between October 1999 and May 2001 by study between October 1999 and May 2001 by study
group.
Group Mean ± SD
Indoor
PM2.6
COPD 8.5 ±5.1
Healthy 7.4 ± 4.8 Asthmatic 9.2 ± 6.0 CHD
9.5 ± 6.8
PM10
COPD 14.1 ± 6.6 Healthy 12.7 ± 7.8
Asthmatic 19.4 ±11.1 CHD 16.2 ± 11.3
group.
Location Pollutant
Group Mean ± SD
Outdoor PM2.6
COPD 9.2 ± 5.1 Healthy 9.0 ± 4.6 Asthmatic
11.3 ± 6.4 CHD 12.7 ± 7.9 PM10
COPD 14.3 ± 6.8 Healthy 14.5 ± 7.0
Asthmatic 16.4 ± 7.4
CHD 18.0 ± 9.0
SOUTH CENTRAL
Turpin et al. (2007,
Houston
Houston: 17.1
Houston: 14.7
157062)
Child: 36.6


Houston (and Elizabeth,
Adult: 37.2


NJ, and Los Angeles



County, CA)



MIDWEST
Sarnat et al. (2006,
Mean (SD): PM2.6

Mean (SD): PM2.6
089784)
Summer

Summer
Steubenville, OH
n - 169

n - 65

mean (SD) - 19.9 (9.4)

mean (SD) - 20.1 (9.3)

Fall

Fall

mean (SD) - 20.1 (11.6)

mean (SD) - 19.3 (12.2)
July 2009	A-300	DRAFT - DO NOT CITE OR QUOTE

-------
Reference / Location
Personal
Micro
Ambient
Adgate et al. (2002,
Battle Creek
Battle Creek
Battle Creek
030676)
All Seasons: 118, 22.7, (25.71,16.2(2.2)
All Seasons: 108,10.6(6.6), 9.0(1.8)
All Seasons: 88 9.4 (6.21,7.8(1.8)
Battle Creek, East St.
Paul, and Phillips,
Minnesota, constituting
Spring: 41, 26.3 (25.7), 19.4 (2.1)
summer: 31, 28.5 (36.1), 20.3 (2.1)
Spring: 25,12.7 (7.71,11.0(1.7)
summer: 36, 8.9 (3.8), 8.1 (1.5)
Spring: 36,10.5(7.11,8.5(2.0)
summer: 22, 8.7 (4.4), 7.8 (1.6)
the Minneapolis-St. Paul
Fall 46,15.5 (13.4),11.9 (2.1)
Fall: 47,10.9(7.4), 8.8(2.0)
Fall: 30, 8.4(6.21,7.1 (1.7)
metropolitan area.
E. St. Paul
E. St. Paul
E. St. Paul

All Seasons: 107, 30.5(38.71,20.6(2.3)
All Seasons: 97,17.4(20.31,12.2(2.2)
All Seasons: 95,10.8(6.6), 9.3(1.8)

Spring: 44, 33.9 (34.4), 23.9 (2.3)
Spring: 30, 20.7(26.41,13.6(2.4)
Spring: 36,12.0(7.31,10.1 (1.9)

summer: 25, 20.5(15.01,17.2(1.8)
summer: 26,15.8 (11.4), 13.7 (1.6)
summer: 25, 8.5 (3.2), 7.8 (1.6)

Fall: 38,33.1(51.9), 19.5(2.5)
Fall 41 16.019.610.4 2.4
Fall: 34,11.3(7.5), 9.6(1.8)

Phillips
Phillips
Phillips

All Seasons: 107, 26.5 (24.3), 20.9 (2.0)
All Seasons: 89,14.2(13.01,11.3(1.9)
All Seasons: 88,10.0(5.8), 8.7,(1.7)

Spring: 28, 37.5(37.61,30.0(1.8)
Spring: 15,16.9 (14.2), 13.0 (2.1)
Spring: 30 (12.11,7.2(10.5)

summer: 40, 22.7 (15.31,19.2(1.7)
summer: 36,13.2(6.41,11.4(1.7)
summer: 30, 8.6 (3.8), 7.8 (1.6)

Fall: 39,22.7 (16.71,17.6(2.1)
Fall: 38,14.4 (16.7), 10.6(2.0)
Fall: 28, 9.3(5.5), 8.1 (1.7)
Crist et al. (2008,
Athens (rural): 17.61 (17.81)
Indoor
Outdoor
156372)
Koebel (urban): 14.59 (13.05)
Athens (rural): 17.20 (13.56)
Athens (rural): 13.66 (8.91)
Ohio River Valley near
Columbus
New Albany (suburb): 13.93 (12.25)
Koebel (urban): 14.98 (12.30)
New Albany (suburb): 16.52 (13.53)
Koebel (urban): 13.89 (9.29)
New Albany (suburb): 12.72 (8.86)
SOUTHEAST
Wallace and Williams
PM2.6 pers - 23.0 (16.4)
PM2.6 in - 19.4(16.5)
PM2.6 out - 19.5(8.6)18.1 (8.1)
(2005, 057485)
PM2.6 pers/PM2.6 out - 1.31 (0.99)
PM2.6 in/PM2.6 out - 1.08 (1.05)

Raleigh, North Carolina



Williams et al. (2003,
Pooled PM mass concentrations (/L/g|m3) across all
Pooled PM mass concentrations (/L/g|m3) across
Pooled PM mass concentrations (/^g/m3) across
053338)
subjects, residences, seasons, and cohorts
all subjects, residences, seasons, and cohorts
all subjects, residences, seasons, and cohorts
SE Raleigh, North Carolina
Variable N Geo mean Mean RSD(a)
Variable N Geo mean Mean RSD(a)
Variable N Geo mean Mean RSD(a)
Chapel Hill, North Carolina
Personal PM2.6 (b) 712 19.2 23.0 70.1
Indoor PM2.1, (c) 761 15.3 19.1 80.1
Ambient PM2.5 (c) 746 17.3 19.2 44.9

(a) Relative standard deviation of the presented
Outdoor PM2.1, (c) 761 17.5 19.3 43.7
Ambient PMio(b) 752 27.9 31.4 51.5

arithmetic mean.
Indoor PMio(b) 761 23.2 27.7 70.6
Ambient PM102.5(d) 210 8.6 10.0 62.3

(beasured using PEMs.
Outdoor PMio(b) 761 27.5 30.4 46.4
Indoor PM10 2.5(d) 761 6.3 8.6 111.8
Outdoor PM10 2.5(d) 761 8.5 11.1 86.9
(a) Relative standard deviation of the presented
(a) Relative standard deviation of the presented
arithmetic mean.
(beasured using PEMs.
(ceasured using HI samplers.
(deasured by difference in PEM PM10 monitor
arithmetic mean	ant' co'ocated HI PM2.6 mass concentrations.
(beasured using PEMs.
(ceasured using HI samplers.
(deasured by difference in PEM PM10 monitor
and co-located HI PM2.6 mass concentrations.
July 2009
A-301
DRAFT-DO NOT CITE OR QUOTE

-------
Reference / Location
Personal
Micro
Ambient
NORTHEAST
Koutrakis et al. (2005,
PM2.6:
PM2.6:
095800)
(Baltimore, Boston)
(Baltimore, Boston)
Baltimore, MD Boston,
Winter: Seniors: 15.1 (14.6), 14.1 (6.0)
Winter:
MA
Children: 24.0(21.81,18.5(12.8)

COPD: 16.4 (12.7), NR
All: 20.1 (9.41,11.6(6.8)

Summer: Seniors: 22.1 (10.1), 18.8 (9.7)
summer:

Children: 18.6(8.1), 30.3(14.2)
Seniors: 25.2(11.51,12.7(5.4)

COPD: NR, NREC:
Children: 23.2(14.01,17.0(11.5)

(Baltimore, Boston)
COPD: NR, NREC:

Winter: Seniors: NR, 1.4 (0.9)
(Baltimore, Boston)

Children: 2.8(1.81,1.6(1.6)

COPD: 2.0(1.2), NR
Winter:

Summer: Seniors: NR, NR
All: 1.2(0.6)

Children: NR, NR
summer: NR, NRSO4:

COPD: NR, NRSO4:
(Baltimore, Boston)


(Baltimore, Boston)
Winter:

Winter: Seniors: 1.9 (1.1), 1.9 (1.2)

Children: NR, 2.3(1.7)
All: 4.0 (1.7), 3.1 (1.8)

COPD: 1.5(0.8), NR
summer:

Summer: Seniors: 5.7 (3.5), 2.9 (1.9)
Seniors: 10.5 (7.1), 3.1 (1.8)

Children: NR, NR
Children: NR, 6.5 (6.0)

COPD: NR, NR
Turpin et al. (2007,
Elizabeth Elizabeth: 20.1
Elizabeth: 20.4
157062)
Child: 54.0

Elizabeth, NJ, (and
Adult: 44.8

Houston, TX, and Los


Angeles County, CA +


Sarnat et al. (2005,
Winter-Children: NR
Winter:
087531)
PM2.1,: 17.4-25.8
PM2.1,: 6.5-15.5
Boston, Massachusetts.
SO4: 1.6-3.3
SO4:1.7-4.2
Comparisons to a previous
Winter-Seniors:
Summer:
study in Baltimore are
PM2.1,: 10.8-16.2
PM2.6:11.9-21.4
made.
SO4: 1.6-2.6
SO4: 3.6-9.0
Summer-Children
PM2.6: 25.4-32.8
S(k 2.7-3.3
Summer-Seniors
PM2.1,: 17.8-20.5
S(k 2.7-3.3
Table A-56. Summary of PM species exposure studies.
Reference
Particle Sizes Measured
Component
Results
Primary Findings
Gadkari et al. (2007,156459)
Personal: RPM
Micro: NR
Ambient: RPM
Fe, Ca, Mg, Na K, Cd, Hg, Ni,
Cr, Zn, As, Pb, Mn and Li
Source contributions varied
widely among 12 sites.
Indoor: 0-95%
Ambient: 0-26%
Road: 0-94%
Soil: 0-75%
Authors conclude that "(1) indoor
activities and poor ventilation
qualities are responsible for
major portion of high level of
indoor RPM, (2) majority of
personal RPM is greatly
correlated with residential indoor
RPM, (3) time-activity diary of
individuals has much impact on
relationship investigations of
their personal RPM with their
July 2009
A-302
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
respective indoor and ambient-
outdoor RPM levels
as reported in earlier reports and
(4) residential indoors, local road-
traffic and soil-borne RPMs are
the dominating routes of
personal exposure compared to
ambient outdoor RPM levels.
Koistinen et al. (2004,156655)
Personal, Micro, and Ambient:
PM2.6
Black smoke, SO42, Nth-, NH4+,
Al, Ca, CI, Cu, K, Mg, P, S, Si, Zn
% contribution to PM2.6
Outdoor ¦ Indoor ¦ Work ¦ Pers
CoPM * 35 28 32 33
Secondary** 46 36 37 31
Soil 16 27 27 27
Detergents 0 6 2 6
Sea Salt 3 2 12
* CoPM is the difference
between total mass and other
identified components; i.e.,
primary combustion particles,
nonvolatile primary and
secondary organic particles, and
particles from tire wear, water,
etc. ** Secondary particles are
the sum of sulfate, nitrate, and
ammonium.4 factors were
identified for each exposure type
(residential indoor, residential
outdoor, workplace indoor, and
personal). The factors contained
the elements Al, Ca, CI, Cu, K,
Mg, P, S, Si, Zn, and black
smoke, (insert in cell to left after
consolidating PM size)
Population exposure assessment
of PM2.E, based on outdoor fixed-
site monitoring, overestimates
exposures to outdoor sources
like traffic and long-range
transport and does not account
for the contribution of significant
indoor sources.
Turpin et al. (2007,157062)
Personal: PM2.6
Micro: PM2.6, in the main living
area (not kitchen)
Ambient: PM2.6, in the front or
back yard
18 volatile organics, 17
carbonyl, PM2.6 mass and > 23
PM2.E species, organic carbon,
elemental carbon, and PAHs
For Los Angeles
Carbon (/^gC/m3)
EC 1.4
OC 4.1
Elements (ng/m3)
Ag 0.5; Al 24.7; As 0.5; Ba
22.9; Br 5.3; Ca 80.9; Cd 0.4; CI
62.0;
Co ND; Cr 0.6; Cu 5.5; Fe 162.9;
Ga 0.1; Ge 0.1; Hg 0.1; In 0.3;
K 74.1; La 2.3; Mn 2.9; Mo 0.4;
Ni 2.0; Pb 4.7; Pd 0.3; P 0.1; Rb
0.1;
S 1022.9; Sb 2.1; Se 1.4; Si
128.9; Sn 7.9; Sr 1.8; Ti 10.4; V
5.3;
Y 0.1; ;Zn 16.4; Zr 0.5
The best estimate of the mean
contribution of outdoor to indoor
PM2.6 was 73% and the outdoor
contribution to personal was
26%.
Delfino et al. (2006, 090745) Personal: 24-h PM2.6 24-h PM2.6 EC
Mean (SD), units: jjg/m3:
PM associations with airway
1-hmaxPM2.B 24-h PM2.6 OC
Riverside
inflammation in asthmatics may
8-h max PM2.6
be missed using ambient particle
Ambient: 24-h PM2.6
24-h PM2.6 EC - 1.61 (0.78)
mass.
24-hPMio
24-h PM2.6 OC - 6.88(1.86)
The strongest positive
(also 24-h NO2, 8-h max O3, 8-h
Whittier
associations were between eNO
max NO2, 24-h NO2, 8-h max CO)
24-h PM2.6 EC - 0.71 (0.43)
and 2-day avg pollutant
concentrations. Per IQR

24-h PM2.6 OC - 3.93(1.49)
increases: 1.1 ppb
FEN0/24/yg/m3 personal PM2.6.
0.7 ppb FEPJ0/0.6 //g/m3
personal EC
1.6 ppb FENO /17 ppb personal
July 2009
A-303
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
NO?
Ambient PM2.6 and personal and
ambient EC were significant only
when subjects were taking
inhaled corticosteroids.
Subjects taking both inhaled
steroids and antileukotrienes had
no significant associations.
Distributed lag models showed
personal PM2.6 in the preceding 5
h was associated with FENO.
Salma et al. (2007,113852)
Personal: PMio-2.0 and PM2.0
Micro: NA
Ambient: NR
30 elements (Na, Mg, Al, Si, P,
S, CI, K, Ca, Ti, V, Cr, Mn, Fe, Ni,
Cu, Zn, Ga, Ge, As, Se, Br, Rb,
Sr, Y, Zr, Nb, Mo, Ba, and Pb)
Units: ng/m3:
PM10-2 o; PM2 o; Mg 296 130; Al
531 93; Si 2.09 442; S 978
828;
CI 305 104; K 318127; Ca 2.57
413; Ti 47 25; Cr 35 15; Mn
310 148; Fe 33.5 15.5; Ni 29 8;
Cu 496 190; Zn 118 50; Br 13
DL; Ba 145 DL; Pb 47 21; PM
83.6 33.0
The concentrations observed in
the Astoria underground station
were clearly lower (by several
orders of magnitude) than the
corresponding workplace limits.
Lai et al. (2004, 056811)
Personal, Micro, and Ambient:
PM2.6
Ag Cr Mn Si
Al Cu Na Sm
As Fe Ni Sn
Ba Ga P Sr
Br Ge Pb Ti
Ca Hg Rb TI
Cdl STm
CIKSbV
Co Mg Se Zn
Zr
GM (GSD), units: ng/m3
P
RO
Wl I/O
Al
As
Br
Ca
Cd
CI
Cu
Fe
Ga
K
Mg
Mn
Na
Ni
Both the indoor and outdoor
environments have sources that
elevated the indoor
concentrations in a different
extent, in turn led to higher
personal exposures to various
pollutants.
Geometric mean (GM) of
personal and home indoor levels
of PM2.6,14 elements, total VOC
(TVOC) and 8 individual
compounds were over 20%
higher than their GM outdoor
levels. Those of NO2, 5 aromatic
VOCs, and 5 other elements
were close to their GM outdoor
levels. For PM2.6 and TVOC,
personal exposures and
residential indoor levels (in GM)
were about 2 times higher
among the tobacco-smoke
exposed group compared to the
non-smoke exposed group,
suggesting that smoking is an
important determinant ofthese
exposures. Determinants for CO
were visualised by real-time
monitoring, and the authors
showed that the peak levels of
personal exposure to CO were
associated with smoking,
cooking and transportation
activities. Moderate to good
correlations were only found
between the personal exposures
and residential indoor levels for
both PM2.6 (r - 0: 60; p< 0:
001) and NO2 (r - 0: 47; p - 0:
003).
July 2009
A-304
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
P
Pb
S
Se
Si
Sn
Ti
V
Zn
Adgate et al. (2007,156196)
Personal, Micro, and Ambient:
PM2.6 ¦ broken down into TE
Ag, Al, Ca, Cd, Co, Cr, Cs, Cu,
Fe, K, La, Mg, Mn, Na, Ni, Pb, S,
Sb, Sc, Ti, TI, V, Zn
Median, units: ngfm3:
0
I
s
334.4
272.1

351.6; Ca


232.285.0


174.1; Al
96.3

23.3
58.6;
Na
33.1
20.6

31.9;

Fe
12.6
43.1

78.6; Mg
10.9

16.3
27.5;
K
3.2
38.4

47.5; Ti
3.0

0.8
1.4;
Zn
2.7
6.5

9.6; Cu
2.4

1.5
4.9;
Ni
NA
¦0.1

1.8; Pb
1.5

2.4
3.2;
Mn
0.6
1.5

2.3; Sb
0.08

0.21
0.30;
Cd
0.05
0.12

0.14; V
0.05

0.12
0.16;
La
0.02
0.05

0.11; Cs
0.00

0.00
0.00;
Th
0.00
0.00

0.00; Sc
0.00

0.00
0.01;
Ag
0.00
0.07

0.08; Co
NA

0.02
0.07;
Cr
¦0.09
1.2

2.6

The relationships among P, I, and
0 concentrations varied across
TEs. Unadjusted mixed-model
results demonstrated that 0
monitors are more likely to
underestimate than overestimate
exposure to many of the TEs
that are suspected to play a role
in the causation of air pollution
related health effects. These
data also support the conclusion
that TE exposures are more
likely to be underestimated in the
lower income and centrally
located PHI community than in
the comparitively higher income
BC K community. Within the
limits of statistical power for
this sample size, the adjusted
models indicated clear seasonal
and community related effects
that should be incorporated in
long-term exposure estimates for
this population.
Ebelt et al. (2005, 056907)
Personal: PM2.6
Micro: "ambient exposure":
PM2.6, PM10, PM2.6 -10;
"non-ambient exposure": PM2.6
Ambient: PM2.6, PM10, PM2.6 -10
Ambient sulfate,
ambient non-sulfate,
personal sulfate,
personal ambient non-sulfate
Mean (SD), units jjg/m3
Ambient sulfate: 2.0 (1.1),
ambient non-sulfate: 9.3 (3.7),
personal sulfate: 1.5 (0.9),
personal ambient non-sulfate:
6.5 (3.0)
Ambient exposures and (to a
lesser extent) ambient
concentrations were associated
with health outcomes; total and
nonambient particle exposures
were not.
Farmer et al. (2003, 089017)
Personal: PM10
Micro: NR
Ambient: PM10
Extractable organic material
(EOM)
Benzo[a]pyrene (B[a]P)
Carcinogenic polycyclic aromatic
hydrocarbons (cPAHs)
Units: ng/m3:
Exposed, controls:
Prague:
cPAHs - 12.04(11.101,6.17
(3.48)
Personal exposure to B[a]P and
to total carcinogenic PAHs in
Prague was two fold higher in
the exposed group compared to
controls, in Kosice three fold
higher, and in Sofia 2.5 fold
July 2009
A-305
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
B[a]P
B[a]P - 1.79(1.671,0.84(0.60) higher.
cPAHs
Kosice:

cPAHs - 21.72(3.12), 6.39

(1.56)

B[a]P - 2.94(1.441,1.07(0.66)

Sofia:

cPAHs - 93.84 (55.0) police,

94.74 (120.34) bus drivers,

41.65 (33.36)

B[a]P - 4.31 (2.6) police, 5.4

(3.18) bus drivers, 1.96 (1.53)
Jansen et al. (2005, 082236)
Personal: PMio	BC, as an estimate of elemental
Micro: PMio, PM2.6, fine particles car']on (EC)
(— PM1)
Ambient: PMio, PM2.6
Mean (IQ Range), units: jjg/m3:
BC
Indoor: 1.34 (1.12)
Outdoor 2.01 (1.68)
Personal 1.64 (2.05)
For 7 subjects with asthma, a
10/yg/m3 increase in 24-h avg
outdoor PMio
and PM2.E was associated with a
5.9 [95% CI, 2.9-8.9] and 4.2
ppb (95% CI, 1.3-7.1) increase
in FENO, respectively. A 1 //g|m3
increase in outdoor, indoor, and
personal BC was associated
with increases in FENO of 2.3
ppb (95% CI, 1.1-3.6), 4.0 ppb
(95% CI,2.0-5.9), and 1.2 ppb
(95% CI, 0.2-2.2), respectively.
No significant association was
found between PM or BC
measures and changes in
spirometry, blood pressure, pulse
rate, or Sa02 in these subjects.
Sorensen et al. (2003,157000) Personal: PM2.6
Micro: NR
Ambient: PM2.6
BS (black smoke)
Units: 10'e/m
n Median Q25-Q75
All 177 6.8 (5.0-13.2)
Autumn 42 7.1 (6.5-17.2)
Winter 46 8.2(5.1-13.3)
Spring 46 12.6 (5.4-10.4)
Summer 47 8.1 (3.4-9.0)
Personal PM2.6 exposure was
found to be a predictor of 8-
oxodG in lymphocyte DNA. No
other associations between
exposure markers and
biomarkers could be
distinguished. ETS was not a
predictor of any biomarker in the
present study. The current study
suggests that exposure to PM2.6
at modest levels can induce
oxidative DNA damage and that
the association to oxidative DNA
damage was confined to the
personal exposure, whereas the
ambient background
concentrations showed no
significant association.
For most of the biomarkers and
external exposure markers,
significant differences between
the seasons were found.
Similarly, season was a
significant predictor of SBs and
PAH adducts, with avg outdoor
temperature as an additional
significant predictor.
Molnar et al. (2005,156772) Personal: 2.5
Micro and Ambient: PMio-2.5
and PM2.6
BS (black smoke)
S
CI
K
Ca
Median, unit - ng/m3
Wood burners Ref 1 -sided p-
value
BS 0.97 0.74 0.053
Statistically significant
contributions of wood burning to
personal exposure and indoor
concentrations have been shown
for K, Ca, and Zn. Increases of
July 2009
A-306
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
Mn
Fe
Cu
Zn
Br
Rb
Pb
S 880 650 0.500
CI 200 160 0.036
K 240 140 0.024
Ca 76 43 0.033
Mn 4.8 3.5 0.250
Fe 64 49 0.139
Cu 8.9 2.4 0.016
Zn 38 22 0.033
66-80% were found for these
elements, which seem to be good
wood-smoke markers. In
addition, CI, Mn, Cu, Rb, Pb, and
BS were found to be possible
wood-smoke markers, though
not always to a statistically
significant degree for personal
exposure and indoor
concentrations. For some of
these elements subgroups of
wood burners had clearly higher
levels which could not be
explained by the information
available.
Sulphur, one of the more typical
elements mentioned as a wood-
smoke marker, showed no
relation to wood smoke in this
study due to the large variations
in outdoor concentrations from
LDT air pollution. This was also
the case for PM2.6 mass.
Personal exposures and indoor
levels correlated well among the
subjects for all investigated
species, and personal exposures
were generally higher than
indoor levels. The correlations
between the outdoor and
personal or ind
Johannesson et al. (2007,
156614)
Personal, Micro, and Ambient:
PM2.6, PM1
BS- Black Smoke
BS2.5 Mean SD
Personal 0.65 0.47
Exclusively smokers 0.62 0.47
Residential indoor 0.56 0.47
Exclusively smokers 0.52 0.46
Residential outdoor 0.68 0.51
Exclusively smokers 0.71 0.54
Urban background 0.63 0.37
All measurements 0.68 0.40
PM1/BS1
Personal 0.55 0.20
Residential indoor 0.54 0.45
Exclusively smokers 0.49 0.43
Residential outdoor 0.66 0.51
Exclusively smokers 0.68
Personal exposure of PM2.6
correlated well with indoor
levels, and the associations with
residential outdoor and urban
background concentrations were
also acceptable. Statistically
significantly higher personal
exposure compared with
residential outdoor levels of
PM2.6 was found for
nonsmokers. PM1 made up a
considerable proportion (about
70-80%) of PM2.6. For BS,
significantly higher levels were
found outdoors compared with
indoors, and levels were higher
outdoors during the fall than
during spring. There were
relatively low correlations
between particle mass and BS.
The urban background station
provided a good estimate of the
residential outdoor
concentrations of both PM2.6 and
BS2.5 within the city. The air
mass origin affected the outdoor
levels of both PM2.6 and BS2.5;
however, no effect was seen on
personal exposure or indoor
levels.
Sram et al. (2007,192084)
Personal: PM10, PM2.6
Micro: NR
c-PAHs, B[a]P
B[a]P: exposed 1.6 ng/m3,
control 0.8 ng/m3; c-PAHs:
exposed 9.7 nglm3, control 5.8
Ambient air exposure to c-PAHs
increased fluorescent in situ
hybridization (FISH) cytogenetic
July 2009
A-307
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings

Ambient: PM10, PM2.6

ng/m3
parameters in non smoking
policemen exposed to ambient
PM
Na and Cocker (2005,156790)
Personal: PM2.6
Micro: NR
Ambient: PM2.6
EC (Elemental carbon)
OC (Organic carbon)
Mean (SD), units - jjg/m3
Residential homes: EC 2.0 (NR)
OC 14.8 (NR)
High school (EC):
Weekday samples 1.1 (0.9)
Weekend samples 1.0 (0.5)
High school (OC):
Weekday samples 8.8 (4.7)
Weekend samples 7.4 (2.4)
Indoor PM2.6 was significant
influenced by indoor OC sources.
Indoor EC sources were
predominantly of outdoor origin.
Geyh et al. (2005,186949)
Personal: TD, PMio, PM2.6
Micro: NR
Ambient: TD, PM10, PM2.6
EC
OC
VOC also
Mean (SD), units - jjg/m3:
Summary Statistics by Area
Location October 2001:
Albany and West
EC 5.9 (NA) OC 36 (NA)
Liberty and Greenwich
EC 5.3 (59) OC 30 (56)
Park Place and Greenwich
EC 14.5 (5.4) OC 72(26)
Church and Dey
EC 7.9 (3.3) OC 48 (15)
April 2002:
Liberty and West
EC 4.2 (2.1) OC 26(13)
Barclay and Greenwich
EC 4.0 (2.6) OC 18(14)
Church and Dey
EC 4.5(1.9) OC 27 (15)
Middle of the Pile
EC 6.7 (1.0) OC 40 (25)
During October, the median
personal exposure to TD was
346 /yglm3. The maximum area
concentration 1742 //g|m3, was
found in the middle of the debris.
The maximum TD concentration
found at the perimeter was
392/yg/m3 implying a strong
concentration gradient from the
middle of debris outward. PM2.6
/PM10 ratios ranged from 23% to
100% suggesting significant fire
activity during some of the
sampled shifts. During April, the
median personal exposure to TD
was 144/yg/m3, and the highest
area concentration, 195 /yglm3,
was found at the perimeter.
Although the overall
concentrations on PM at the site
were significantly lower in April,
the relative contributions of fine
particles to the PM10, and EC
and OC to the TD were similar.
During both months, volatile
organic compounds
concentrations were low.
Comparison of recorded EC and
OC values from October 2001
and April 2002 with previous
studies suggests that the
primary source of exposure to EC
for the WTC truck drivers was
emissions from their own
vehicles.
Zhao et al. (2007,156182)
Personal, Micro, and Ambient:
PM2.6
EC, CI, Si, NOs
Units - jjg/m3:
Personal: EC: 1.64 NOs: 0.135,
Si: 0.176, CI: 0.116; Indoor: EC:
1.819 NOs: 0.013, Si: 0.051, CI:
0.024; Outdoor: EC: 1.876 NOs:
0.292, Si: 0.115, CI: 0.013
Four external sources and three
internal sources were resolved in
this study. Secondary nitrate and
motor vehicle were two major
outdoor PM2.E sources. Cooking
was the largest contributor to
the personal and indoor samples.
Indoor environmental tobacco
smoking also has an important
impact on the composition of the
personal exposure samples.
Meng et al. (2005, 081194)
Personal: PM2.6
Micro: NA
Ambient: NR
EC, OC, S, Si
Mean (SD), units - ng/m3:
Indoor: EC: 1165.9 (2081.0) OC:
7725.5(9359.3) S: 902.3
(602.2) Si: 124.0 (79.0)
Use of central-site PM2.6 as an
exposure surrogate
underestimates the bandwidth of
the distribution of exposures to
July 2009
A-308
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle Sizes Measured
Component
Results Primary Findings



Outdoor: EC: 1144.1 (968.1) OC: PM of ambient origin.



3777.7 (2520.1) S: 1232.3



(633.2) Si: 141.1 (171.3)
Smith et al. (2006,156990)
Personal: PM2.6
Elemental carbon (EC)
Work Area

Micro: PM2.6
Organic carbon (OC)
Office

Area samplers in the offices,

Dock

freight dock, or shop.

Yard



Ambient: PM2.6

Shop

Samplers were located in the


yard upwind of the terminal.

Non-smokers on-site:



Clerk



Dock worker



Mechanic



Hostler
Non-smokers off-site
Pickup/deliver driver
Long haul driver
Smokers On-Site
Clerk
Dock worker
Mechanic
Hostler
Smokers off-site
Pickup & Delivery drivers
Long haul drivers
Koutrakis et al. (2005, 095800) Personal: PM2.6
Micro: NR
Ambient: PM2.6
Elemental Carbon (EC),
SO42-
Mean (SD) data are provided for
Baltimore and Boston,
units - jjg/m3:
EC:
(Baltimore, Boston)
Winter:
Seniors: NR, 1.4 (0.9)
Children: 2.8 (1.8), 1.6(1.6)
COPD: 2.0(1.2), NR
SO4:
(Baltimore, Boston)
Winter:
Seniors: 1.9 (1.1), 1.9(1.2)
Children: NR, 2.3 (1.7)
COPD: 1.5(0.8), NR
Summer:
Seniors: 5.7 (3.5), 2.9 (1.9)
Ambient PM2.6 and SO4 are
strong predictors of respective
personal exposures. Ambient SO4
is a strong predictor of personal
exposure to PM2.6. Because
PM2.E has substantial indoor
sources and SO4 does not, the
investigators
concluded that personal
exposure to SO4 accurately
reflects exposure to ambient
PM2.E and therefore the ambient
component of personal exposure
to PM2.6 as well.
Chillrud et al. (2004, 054799)
Personal: PM2.6
Micro: PM2.6
Home indoor and home outdoor
Ambient: Urban fixed-site and
upwind fixed site operated for
three consecutive 48-h periods
each week.
Elemental iron, manganese, and
chromium are reported in this
study out of 28 elements
sampled.
Mean of duplicate samples:
PM2.6: 62 jjg/m3
Fe: 26 jjg/m3
Mn: 240 ng/m3
Cr: 84 ng/m3
Variability: 1-15%
Personal samples had
significantly higher concentration
of iron, manganese, and
chromium than home indoor and
ambient samples. The ratios of
Fe (ng//yg of PM2.6) vs Mn
(pgl fjg PM2.E) showed personal
samples to be twice the ratio for
crustal material. Similarly for the
Cr/Mn ratio.
The ratios and strong
correlations between pairs of
elements suggested steel dust as
the source. Time-activity data
suggested subways as a source
July 2009
A-309
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings




of the elevated personal metal




levels.
Jansen et al. (2005, 082236) Personal, Micro, and Ambient: Estimated Elemental Carbon Mean (SD), units - jjg/m3:
PM2.6	(Abs)
Elemental composition of a
subset of personal, indoor and
outdoor samples	PM2.6
Abs
S
Zn
Fe
K
Ca
Cu
Si
CI
For most elements, personal and
indoor
concentrations were lower than
and highly correlated with
outdoor concentrations. The
highest correlations (median
r.0.9) were found for sulfur and
particle absorbance (EC), which
both represent fine
mode particles from outdoor
origin. Low correlations were
observed for elements that
represent the coarser part of the
PM2.6 particles (Ca, Cu, Si, CI).
Molnar et al. (2006,156773)
Personal: PM2.6 and PM1
Micro and Ambient: NR
S
CI
K
Ca
Ti
V
Mn
Fe
Ni
Cu
Zn
Br
Pb
Urban background PM2.6
mean, median, range
S 620 320 95-1900
CI 97 54 25-460
K 55 50 32-130
Ca 21 17 6.6-6.2
Ti 2.1 1.9 1.3-3.8
V 3.4 2.41.0-13
Mn 1.6 1.4 0.67- 3.8
Fe 36 33 7.1-100
Ni 1.6 1.2 0.33- 5.7
Cu 2.1 1.4 0.33-11
Zn 14 11 2.8-38
Br 1.7 1.4 0.47-44.3
Pb 3.3 2.1 0.94-11
Personal PM2.6
mean, median, range
S- < 470 270-1400
CI 270 170 60-920
K 140 96 39-690
Ca 110 80 27-670
Ti 11 9.5 3.7-27
V4.7 4.0 2.7-9.4
Mn ¦ ¦ ¦
Fe 68 69 23-150
Ni 4.2 2.6 0.89-46
Cu 10 6.6 1.1-81
Zn 21 16 6.6-70
Br 2.01.3 0.91-14
Pb 2.9 2.6 0.92-8.3
Personal PM1
S- < 470 240-1200
CI - < 110 54-160
K 80 82 50-130
Ca 32 23 8.4-87
Ti6.5 6.3 3.7-11
V- < 4.2 2.8-8.9
Mn ¦ ¦ ¦
Fe 28 25 7.6-68
Ni 8.2 1.2 0.83-58
Cu 5.0 4.4 1.6-14
Zn 15 14 7.6-37
PM2.E personal exposures were
significantly higher than both
outdoor and urban background
for the elements CI, K, Ca, Ti,
Fe, and Cu.
Personal exposure was also
higher tan indoor levels of CI, Ca,
Ti, Fe, and Br, but lower than
outdoor Pb./ Residential outdoor
levels were significantly higher
than the corresponding indoor
levels for Br and Pb, but lower
for Ti and Cu. The residential
levels were also significantly
higher than the urban
background for most elements.
July 2009
A-310
DRAFT-DO NOT CITE OR QUOTE

-------
Reference	Particle Sizes Measured	Component
Results
Primary Findings
Br 1.6 1.5 0.83-4.4
Pb 3.6 2.8 1.1-11
Residential Outdoor PM2.6
mean, median, range
S 640 460 190-1800
CI 6.3 140 57-840
K 200 78 32-200
Ca 82 28 4.6-85
Ti 34 5.2 3.3-21
V 6.3 3.9 2.1-14
Mn
Fe 5.5 31 8.8-200
Ni 45 < 1.6 0.65-5.5
Cu 2.6 1.3 0.65-17
Zn 22 15 5.5-85
Br 2.0 >450 0.91-51
Pb 4.6 2.6 0.90-20
Residential Outdoor PM1
S - 1.3 24-2000
CI- < 11044-170
K 76 68 34-170
Ca- < 125.1-78
Ti- < 5.0 2.2-9.5
V 5.6 4.47 2.2-14
Mn
Fe 23 14 3.7-140
Ni 3.3 1.4 0.73-28
Cu- < 1.1 0.73-12
Zn 15 14 5.2-30
Br 1.5 1.40.78-4.3
Pb 4.1 1.51.0-17
Kulkarni and Patil (2003, Personal: PM5
Lead
Personal samples:
All listed metals were detected
156664) Micro: NR
Nickel
Cadmium
Mean ± SD
in the ambient air where as only
Lead, Cadmium, Manganese, and
Ambient: PM5
Copper
Type
Potassium were detected in

Chromium
Lead
personal exposures. Mean daily

Potassium
Occupational
exposure to lead exceeds the

Iron
4.384 ± 7.766/yg/m3
Indian NAAQS by a factor of 4.2.

Manganese
Residential
4.093 ± 5.925/yg/m3
24-h integrated
4.205 ± 1.523/yg/m3
Cadmium
However, ambient concentration

of lead conforms to this
standard. There is a rising trend
in the personal exposures and
ambient levels of cadmium.
However, they are low and do


Occupational
0.201 ± 0.158/yg/m3
not pose any major health risk as


yet. Personal exposures to toxic
metals exceed the corresponding


Residential
ambient levels by a large factor


0.111 ± 0.165/yg/m3
ranging from 6.1 to 13.2. Thus,


24-h integrated
ambient concentrations may


0.134 ± 0.140 //g/m3
underestimate health risk due to


Manganese
personal exposure of toxic


metals. Outdoor exposure to


Occupational
toxic metals is greater than the


1.979 ± 7.842/yglm3
indoor (ratios ranging from 2.3 to


Residential
1.1) except for potassium (ratio


0.180 ± 0.261 /yglm3
0.77). However, there is no


24-h integrated
significant correlation between


1.983 ± 6.824/yg/m3
these two.


Potassium



Occupational



3.473 ± 4.691 //gfm3

July 2009
A-311
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
Residential
4.589 ± 4.619/yg/m3
24-h integrated Check
Wu et al. (2006,179950)
Personal: PM2.6
Micro: PM2.6
Ambient: PM2.6
levoglucosan (LG)
Elemental Carbon (EC)
Organic Carbon (OC)
Mean personal exposure:
LG: 0.018 (0.024)
EC: 0.4(0.5)
OC: 8.5 (2.7).
Ambient: check component
Authors "found a significant
between-subject variation
between episodes and non-
episodes in both the Exposure
during agricultural burning
estimates and subjects' activity
During non-burning times: 0.026
(0.030)
During burning episodes: 0.010
(0.012)
patterns. This suggests that the
LG measurements at the central
site may not always represent
individual exposures to
agricultural burning smoke
"Evidence of "Hawthorne
Effect": During declared episodes
(i.e. real and sham), subjects
spent less time indoors at home
and more time in transit or
indoors away from home than
during non-declared episode
periods. The differences
remained even when limited to
weekdays only.
Larson et al. (2004, 098145) Personal: PM2.6
Micro: PM2.6 outside subject's
residence, and inside residence
Ambient: PM2.6 at Central
outdoor site (downtown Seattle)
Light absorbing carbon (LAC) and
trace elements
Personal Rl RO Central Mass
10,500 10,250 12,693 11,970
Al 32 19 21 31
As 1 1 2 2
LAC * 1439 01105 1830 1741
Br 3 2 3 3
Ca 72 46 36 50
CI 248 173 75 78
Cr 2 2 1 2
Cu 3 4 2 3
Fe 63 35 61 95
K 57 54 78 67
Mn 2 2 3 6
Ni
Five sources of PM2.6 identified:
vegetative burning, mobile
emissions, secondary sulfate, a
source rich in chlorine, and
crustal-derived material. The
burning of vegetation (in homes)
contributed more PM2.6 mass on
avg than any other sources in all
microenvironments.
Brunekreef et al. (2005,
090486)
Personal, Micro & Ambient:
PM2.6
Nitrate
Mean (SD), units - ng/m3:
Amsterdam:
Personal 1389(1965)
Indoor 1348(1843)
outdoor 4063(4435)
Helsinki:
Personal 161(202)
Indoor 267(215)
Outdoor 1276(1181)
In both cities personal and indoor
PM2.6 were lower than highly
correlated with outdoor
concentrations. For most
elements, personal and indoor
concentrations were also highly
correlated with outdoor
concentrations.
Sorenson et al. (2005, 089428)
Personal: PM2.6 & Black smoke
(BS)
Micro: PM2.6 & Black smoke (BS)
Ambient: Street monitoring
station and roof of a campus
building PM2.6 81 Black smoke
(BS)
Black Smoke (also NO2)
Mean, IQR, Units - jjg/m3:
Personal:
Cold Season: 10.2 (5.6-14.8)
Warm Season: 7.1 (5.5-11.4)
Micro:
Cold Season
Home Indoor: 6.2 (5.5-11.4)
Home front door: 10.8 (7.4-16.3)
Warm Season
Home Indoor: 6.1 (3.7-7.6)
Home front door: 8.8 (5.6-11.54)
Ambient:
Cold Season: Street Station:
Indoor sources of PM and BS (as
well as NO2) were shown to be
greatly influenced by indoor
sources.
July 2009
A-312
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
31.6 (27.5-34.0)
Urban Background: 7.7 (5.9-
11.0)
Warm Season:
Street Station: 30.6 (24.7-36.0)
Urban Background: 6.8 (4.6-8.6)
Ho et al. (2004, 056804)
Personal: PM2.6
OC
Mean, unit - jjg/m3
The major source of indoor EC,

Micro: NR
EC
Indoors:
OC, and PM2.6 appears to be

Ambient: PM2.6
OM
TCA
OM - 18.1; TCA - 22.9
Outdoors:
OM - 20.1; TCA - 26.5
penetration of outdoor air, with
a much greater attenuation in
mechanically ventilated
buildings.
Maitreetal. (2002,156726)
Personal: PM4
Micro: NR
Ambient: PM4
PAH, benxene-toluene-xylenes
(BTX), aldehydes, BaP
PAHc, formaldehyde,
Median
Resp jjg/m3
BaP ng/m3
PAHc ng/m3
PAH ng/m3
Benzene
Mg/m3
Toluene jjg/m3
Xylene jjg/m3
BTX jjg/m3
Formaldehyde jjg/m3
Acetaldehyde jjg/m3
Aldehyde jjg/m3
The occupational exposure of
policemen does not exceed any
currently applicable occupational
or medical exposure limits.
Individual particulate levels
should preferably be monitored in
Grenoble in winter to avoid
underestimations.
Farmer et al. (2003, 089017)
Personal: PM10
PM10
Prague-SM
Extractable organic matter

Micro: NR
EOM
EOM (/yg/m3)
(EOM) per PM10 was at least 2-


E0M2
fold higher in winter than in

Ambient: PM10
B[a]P
E0M2 (%)
summer, and c-PAHs over 10-

PM2.6 (not reported)
c-PAHsb
B[a]P (/yg/m3)
fold higher in winter than in



c-PAHsb (/yg/m3)
summer. Personal exposure to



Prague-LB
B[a]P and to total c-PAHs in




Prague ca. was 2-fold higher in



EOM (/yg/m3)
the exposed group compared to



E0M2 (%)
the control group, in Kosice ca.



B[a]P (/yg/m3)
3-fold higher, and in Sofia ca.



2.5-fold higher.



c-PAHsb (/yg/m3)



Kosice




EOM (/yg/m3)




E0M2 (%)




B[a]P (/yg/m3)




c-PAHsb (/yg/m3)




Sofia




EOM (/yg/m3)




E0M2 (%)




B[a]P (/yg/m3)




c-PAHsb (/yg/m3)

Hanninen et al. (2004, 056812)
Personal: PM2.6
PM2.E -bound sulphur

Associated with indoor

Micro: NR

Athens
concentration: wooden building




material, city, building age, floor

Ambient: PM2.6

Basel
of residence (i.e. ground, 1st,



Helsinki
etc.), and use of stove other than




electric.
July 2009
A-313
DRAFT-DO NOT CITE OR QUOTE

-------
Reference	Particle Sizes Measured	Component
Results
Primary Findings
Shilton et al. (2002, 049602)
Personal, Micro, and Ambient:
Respirable PM
Respirable PM, metals (Zn, Cu,
Mn, Al), sulphate, nitrate, and
chloride
Prague
Indoor
Outdoor
Zn (ng/m3) 241.1	179.5
Cu (ng/m3) 43.3	24.99
Mn (ng/m3) 15.6	4.18
Al (ng/m3) 305.2	52.90
S04 (ng/m3) 4.72	3.47
CI (ng/m3) 1.08	0.15
NO3 (ng/m3).35	1.08
Measurement Mean s.d.
Ambient SO4 2.72* 3.11
Ambient ABS 1.4** 1.0
Personal SO41.33* 1.47
Personal ABS 1.0** 1.7
* Mean SO4 values reported
in/yg/m3
** Mean ABS values reported in
10'6/m-1
Mean (SD), units - jjg/m3:
Personal
Ambient
S04
Summer
5.9 (4.2)
7.7 (4.8)
Fall
4.4 (3.3)
6.2 (4.7)
EC
Summer
1.1 (0.6)
1.1 (0.5)
Fall
1.2(0.7)
1.1 (0.7)
The indoor particulate cone was
driven by ambient cone;
meteorological-induced changes
in ambient PM were detected
indoors;
SO4 and light absorbing carbon
concentrations had higher
personal-ambient correlations
and less variability. This
indicates that SO4 and ABS were
of outdoor origin, while PM2.B
mass was of varied indoor and
outdoor origin.
High association between
personal and ambient S042~ and
EC, especially for SO42 for
which there is no significant
indoor source.
Noulett et al. (2006,155999) Personal: PM2.6	SO4
iyijcro. [y|fj	ABS (light absorbing carbon)
Ambient: PM2.6
Sarnat et al. (2006, 089784) Personal: PM2.6	SO4
Micro: NR
Ambient: PM2.5
EC
Sarnat et al. (2005 RMID 9171) Personal: PM2.6
(2005,08Z531)	Micm:n)a
Ambient: PM2.6
SO4, O3, NO2, SO2
Correlations between personal
PM2.E and ambient gas
O3 correlated in summer.
Spearman's R ~ 0.4, anti-
correlated in winter, R ~ 0.3-0.1.
NOx somewhat correlated in
summer. R ~ 0.3
Winter, R — 0.2-0.4
SO2 not well correlated in
summer or winter. R ~ 0-0.1.
CO somewhat correlated in
summer. R nO.1-0.3.
Correlated in winter R ~ 0.2-0.3.
No results were significant.
Substantial correlations between
ambient PM2.6 concentrations
and corresponding personal
exposures.
Summertime gaseous pollutant
concentrations may be better
surrogates of personal PM2.6
exposures (especially personal
exposures to PM2.6 of ambient
origin) than they are surrogates
of personal exposures to the
gases themselves.
Brunekreef et al. (2005,	Personal, Micro, and Ambient: SO42, N03-
Mean, units - jjg/m3:
In both cities personal and indoor
July 2009
A-314
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
090486)
PM2.6
SO42-:
Amsterdam
Helsinki
NOs-:
PM2.B were lower than highly
correlated with outdoor
concentrations. For most
elements, personal and indoor
concentrations were also highly
correlated with outdoor
concentrations.
Amsterdam
Helsinki
Kimetal. (2005,156640)
Personal: PM2.6
Micro: NR
Ambient: PM2.6
Sulfate, Elemental carbon (EC),
Calcium, Magnesium, Potassium,
Sodium
Mean (SD), units -
SO42-: 2.7 (3.2)
Ca2+: 0.12(0.12)
Mg2+: 0.02 (0.01)
K: 0.07 (0.08)
Na+: 0.09(0.20)
EC: 0.60 (0.54)
jjg/m3: Traffic-related combustion,
regional, and local crustal
materials were found to
contribute 19% ± 17%,
52% ±22%, and 10% ±7%,
respectively.
Among participants that spent
considerable time indoors,
exposure to outdoor PM2.6
includes a greater relative
contribution from combustion
sources, compared with outdoor
(ambient) PM2.6 measurements.
Wallace and Williams (2005,
Personal: PM2.6
Sulfur
Mean (SD), units - ng/m3:
Generally, infiltration factor
057485)
Indoor Micro: PM2.6
Outdoor Micro: PM2.6

Personal: 1046 (633)
Indoor: 1098 (652)
Outdoor: 1951 (1137)
provides a reliable estimate of
personal exposure. Sulfur can be
used in lieu of personal exposure
to PM because it is derived from
outdoors.
Jacquemin et al. (2007,192372)
Personal: PM2.6
Sulfur
Mean, units - jjg/m3:
Authors suggest that"outdoor

Micro: NA
Ambient: PM2.6

Personal: 1.3 outdoor: 1.2
measurements of absorbance
and sulphur can be used to
estimate both the daily variation
and levels of personal exposures
also in Southern European
countries, especially when
exposure to ETS has been taken
into account. For PM2.6, indoor
sources need to be carefully
considered."
July 2009
A-315
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-57. Summary of personal PM exposure source apportionment studies.
Reference
Study Design
Results
Primary Findings
Hopke et al. (2003, 095544)
Source apportionment of personal (PEM)
and indoor central and apartment (VAPS)
and outdoor (VAPS) PM2.6, Baltimore
retirement home with 10 elderly subjects,
July-Aug 1998.
% contr
External
Secondary
S042"
Unknown
Soil
Internal
Gypsum
Activity
Personal care
63% of personal exposure could be
attributed to outdoor sources (with 46%
from sulfate), and resuspension of indoor
PM during vacuuming, cleaning, or other
activities contributed 36% of personal
exposure.
Larson et al. (2004, 098145)
Source apportionment of personal (PEM)
and residences (HI) and central outdoor
(HI) PM2.E around Seattle with 10 elderly
subjects and 10 asthmatic children, Sep
2000 and May 2001. The purpose of the
article was to compare PMF2 and PMF3
methods.
PMF2:
% contr
Veg burn
Mobile
Fuel oil
S, Mn, Fe
Secondary
CI-rich
Crustal
Crustal2
Results showed that vegetative burning
was the largest contributor to personal
exposure and that was related to outdoor
combustion. Crustal exposures were
related to indoor activities.
PMF3:
% contr
Veg burn
Mobile
Secondary
Crustal
Zhao et al. (2006,156181)
Source apportionment of personal (PEM)
and residential indoor (HI) and residential
outdoor (HI) and central outdoor (HI)
PM2.E, Raleigh and Chapel Hill NC with 38
subjects, summer 2000 and Spring 2001.
% contr
Motor vehicle
Soil
Secondary S042~
Secondary N03-
ETS
Personal care
and activity
CU-factor mix
w indoor soil
Cooking
Secondary sulfate was the largest
ambient source and the largest ambient
contribution to personal exposure.
Cooking produced the largest contribution
to personal and indoor concentrations.
Note that sums over 100% because
multiple sources obscured PMF resolution
July 2009
A-316
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Study Design
Results
Primary Findings
Meng et al. (2007, 091197)
Source apportioned infiltration for
personal (PEM) and residential indoor (HI)
and residential outdoor (HI) and central
outdoor (HI) PM2.E, Los Angeles, Houston,
and Elizabeth, NJ with 100 non-smoking
residences and residents in each city, in
each season between summer 1999 and
spring 2001 (RI0PA).
% contr
(Outdoor
Origin)
Mechanically generated
Primary Combustion
Secondary Formation*
"excludes nitrates
Differential infiltration of the PM2.6
resulted in a reduction of secondary
formation products relative to outdoors.
Reff et al. (2007,156045)
Functional group distinction for personal
(PEM) and residential indoor (HI) and
residential outdoor (HI) and central
outdoor (HI) PM2.E, Los Angeles, Houston,
and Elizabeth, NJ with 100 non-smoking
residences and residents in each city, in
each season between summer 1999 and
spring 2001 (RIOPA). PM2.6 samples from
219 homes were used for this analysis.
SO42':
R
0
1
P
C - 0:
The main finding was that indoor and
personal levels of CH in organic carbons
were found to be substantially higher
than outdoors. This reduced the polarity
of indoor and personal organic carbons
CH:
R
0
1
P
Zhao et al. (2007,156182)
Source apportionment of personal (PEM) % contr
and indoor school (FRM) and outdoor
school (FRM) PM2.6, Denver with 56
asthmatic children, Oct 2002-March
2003 and Oct 2003-March 2004.
% contr
The largest personal exposure was from
Secondary SO42'
cooking (54.8%), but motor vehicle
emissions were the largest outdoor
Soil
contributor (13.3%) to personal exposure.
Secondary NO3'
Secondary nitrate comprised the largest
Motor vehicle
outdoor source but accounted for only
Cl-based cleaning
9.4% of personal exposure.
Cooking

ETS

Strand et al. (2006, 089203)	Using positive matrix factorization and an Estimation method, Mean (SD, range): Similar results were found with each
extrapolation method to estimate PM2.B PMF-7 42 (1 93 3 43-12 89)	technique.
based on SO42 and Fe components.
Extrapolation Method:
Using sulphate: 6.38 (1.60, 3.20 ¦ 10.97)
Using sulphate & iron: 6.50 (1.36, 3.54 ¦
10.12)
Using sulphate 81 iron, temperature
adjusted: 7.02 (1.48, 3.79 - 11.02)
Using sulphate (no gamma): 8.23 (2.06,
4.12-14.14)
July 2009	A-317	DRAFT - DO NOT CITE OR QUOTE

-------
Table A-54. Summary of PM infiltration studies.
Reference
Study Design
Finf
Allen et al. (2003, 053578)
Objective: Enhance knowledge of the
outdoor contribution to total indoor and
personal PM exposures.
Methods: Continuous light scattering
monitoring.
Subjects: Elderly and children spending
most of their time indoors. Healthy
individuals, elderly with COPD or CHD
and children with asthma. 44 residences
measured for 55 10-day sessions.
Seattle, WA.
PM2.6 avg- 0.65 ± 0.21;
Non-heating season- 0.79 ± 0.1 £
Heating season- 0.53 ± 0.16;
Open windows (mean)- 0.69;
Closed windows (mean)- 0.58;
All days (mean)- 0.65
Light scattering (whole peak): 0.75 ±
0.25;
Light scattering (uncensored data): 0.77
± 0.24;
Sulfur concentration (slope): 0.65 ± 0.01
Arhami et al. (2009,190096)
Objective: To examine associations
between size-segregated PM, their
particle components, and gaseous co-
pollutants.
Methods: Data analyzed with linear
mixed-effect models.
Subjects: Four different retirement
communities in San Gabriel Valley, CA
and Riverside, CA. 2005-2007.
PM2.6: 0.38-0.57
EC: 0.64-0.82
OC: 0.60-0.98
n/a
Balasubramanian et al. (2007,156248)
Objective: PM monitoring and
assessment based on analysis of
chemical and physical characteristics or
indoor and outdoor particles.
Methods: Particle number and mass
concentrations measured using real-time
particle counter and low-volume
particulate sampler.
Subjects: 3 residential indoor and 1
residential outdoor environments in Choa
Chu Kang, Singapore. May 12-May 23,
2004.
n/a
PM2.1,: 0.93-1.90
Chemical Species: CI': 0.35-0.45,
NO?-: 2.50-4.13,
NOs": 1.41-5.41,
S042:1.21-1.70,
Na+: 0.43-0.74,
NH4: 1.43-2.39,
EC: 0.75-0.96,
OC: 1.04-1.92,
Al: 1.04-1.92,
Co: 0.86-1.32,
Cr: 1.35-2.90,
Cu: 0.50-0.69,
Fe: 0.30-0.42,
Mn: 0.23-0.42,
Pb: 0.40-2.47,
Zn: 0.59-0.81,
Cd: 0.74-1.75,
Ni: 0.71-1.32,
Ti: 0.73-0.78,
V: 1.01-1.05
Barn et al. (2008,156252) Objective: Measure infiltration factor
PM2.6 (mean)- Summer: HEPA- 0.19 ±
Mean- Summer: HEPA- 0.43, Unfiltered-
from PM2.6 from forest fires and
0.20, Unfiltered- 0.61 ± 0.27
0.77
determine effectiveness of HEPA filter.
Winter: HEPA-0.10 ± 0.08, Unfiltered-
Winter: HEPA-0.21, Unfiltered- 0.36
Methods: pDR for ambient air sampling.
0.28 ± 0.18
Both: HEPA- 0.25, Unfiltered- 0.47
Subjects: Homes affected by forest fire
Both: HEPA-0.13 ± 0.14, Unfiltered-

or residential wood smoke. British
0.42 ± 0.27

Columbia, Canada. 38 homes sampled


(valid samples: 19 winter, 13 summer).


July 2009
A-318
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Study Design
Finf
Baxter et al. (2007, 092726)
Objective: To develop predictive models
of residential indoor air pollutant
concentrations for lower SES, urban
households. Part of ACCESS cohort study
of asthma etiology.
Methods: Regression analysis; mass
balance model; Finf from slope in
univariate regression analyses.
Subjects: Lower SES populations. 43
homes, 23 homes monitored in both
seasons, 15 in the non-heating season
(May-Oct) only, 5 in heating season (Dec-
Mar) only; 2003-2005.
PM2.b: 0.91 ±0.23
Chemical species: EC: 0.72 ± 0.49;
Ca: 0.56 ± 0.30;
Fe: 0.38 ± 0.26;
K: 0.83 ± 0.52;
Si: 0.02 ± 0.00;
Na: 0.46 ± 0.43;
CI: 0.40 ± 0.12;
Zn: 0.85 ± 0.28;
S: 0.95 ± 0.78;
V: 0.60 ± 0.77
PM2.6 (coefficient of variation (CV)): 1.14
(0.71)
Chemical species (coefficient of variation
(CV)): EC (CV): 0.89 (0.64);
Ca (CV): 1.16 (1.90);
Fe (CV): 0.69(1.40);
K (CV): 1.10(0.95);
Si (CV): 1.04(1.31);
Na (CV): 1.05(1.84);
CI (CV): 3.18(3.79);
Zn (CV): 0.83 ± (1.13);
S (CV): 0.76 ± (0.32);
V (CV): 0.76 ± (0.46)
Baxter et al. (2007, 092725)
Objective: To predict residential indoor
concentrations of traffic-related air
pollutants in lower SES urban households.
Part of ACCESS cohort study of asthma
etiology.
Methods: Regression modeling, Bayesian
variable selection I/O is slope from
multivariate model
Subjects: Lower statuses, urban
households in Boston, MA. 43 sites
among 39 households, 66 sampling
sessions, nonheating (May-Oct) and
heating (Dec-Mar) 2003-2005
n/a
PM2.E: Ambient Concentrations X Open
Windows- 0.98;
Ambient Concentrations X Closed
Windows- 0.64
EC: Ambient Concentrations- 0.38
Brown et al. (2008,190894)
Objective: To examine if ambient, home
outdoor, and home indoor particle
concentrations can be used as proxies of
corresponding personal exposure.
Methods: Associations characterized
using univariate mixed effects models
that included a random subject term.
Subjects: 15 participants in Boston, MA
in winter (Nov. 1999-Jan. 2000) and
summer (June-July 2000).
n/a
PM2.6: Winter- Median: 1.2, Range: 0.8-
1.8; Summer- Median: 0.9, Range: 0.6-
1.2
EC: Winter- Median: 1.1, Range: 0.7-4.5;
Summer- Median-1.0, Range: 0.9-1.3
SO42': Winter- Median: 0.5, Range: 0.3-
0.8; Summer- Median: 0.8, Range: 0.4-
1.0
Cao et al. (2005,156321)
Objective: To determine relationships
and distributions of indoor and outdoor
PM2.6, OC, and EC. To determine
indoor/outdoor sources to indoor
carbonaceous aerosol.
Methods: Gravimetric analysis to
determine PM2.6 concentrations. OC and
EC determined by TOR following
IMPROVE protocol.
Subjects: 6 residences in Hong Kong (2
roadside, 2 urban, 2 rural). March 6-April
18,2004.
n/a
20min PM2.b- Roadside: 0.7-4.0, Urban:
0.9-6.7, Rural: 0.5-1.7
24h PM2.b- Roadside: 0.8-1.4, Urban: 1.2-
2.0, Rural: 1.0-1.8
OC (average and range)- Roadside: 1.9
(1.1-2.3), Urban: 2.3 (1.5-4.0), Rural: 1.3
(1.2-2.2)
EC (average and range)- Roadside: 1.0
(0.9-1.1), Urban: 1.1 (0.8-1.3), Rural: 1.1
(0.9-1.8)
July 2009
A-319
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Study Design
Finf
Cortez-Lugo et al. (2008,156368)
Objective: To determine personal PM2.6 n/a
and its relationship with outdoor and
indoor PM2.6and PM10.
Methods: Linear regression model used
to compare personal and indoor PM2.6. I/O
variation studied using analysis of
variance and predictors determined by
generalized estimating equation models.
I/O PM2.6 ratio transformed into natural
logarithm.
Subjects: 38 nonsmoking long time
Mexico residents with COPD. Mexico
City, Mexico. Feb-Nov 2000.
PM2.6: Average- 1.2, Range- 0.05-6.1
Crist et al. (2008,156372)
Objective: To examine correlations
between personal, indoor, and outdoor
PM2.E exposures in children.
Methods: Indoor and personal samples
by Whatman Teflon filters. Ambient
samples taken by TEOMs.
Subjects: Fourth and fifth-grade children.
1 school in Athens, OH and 2 schools
inColumbus, OH. No pre-existing health
conditions. 90 children (30 at each site),
3 of which had personal monitors. 194-
332 days of indoor, outdoor, and personal
samples. N samples taken at schools
range 31-235.
n/a
Mean PM2.6 Mass Concentrations
(|jg/m3)- Athens: school-day- 2.61 ±
5.76, Non-school-day- 0.8 ± 0.7
Koebel: school-day-1.71 ± 3.17, non-
school-day- 1.27 ±1.16
New Albany: school-day- 2.98 ± 5.47,
non-school-day- 0.82 ± 0.6
Diapouli et al. (2008,190893)
Objective: To characterize the PM10, n/a
PM2.6, UFP concentrations at primary
schools. To examine the relationship
between indoor and outdoor
concentrations.
Methods: Chemical analysis of collected
filters. Regressions to examine
correlations between indoor and outdoor
concentrations.
Subjects: 7 primary schools with
different characterizations of urbanization
and traffic density in Athens, Greece. No
ventilation system. Nov. 2003-Feb. 2004
and Oct.-Dec. 2004.
PM10: 0.54-2.46
PM2.5- 0.67-2.77
UFP- 0.33-0.74
Dimitroulopoulou et al. (2006, 090302)
Objective: T0 develop a probabilistic n/a
indoor air model (INDAIR).
Methods: INDAIR predicts frequency
distributions of concentrations of up to 4
pollutants simultaneously (NO2, CO, PM10,
PM2.6). 3 scenarios- no source, gas
cooking, smoking.
Subjects: 5 UK sites- Harwell (rural),
Birmingham East (urban background),
Bradford (urban center), Bloomsbury
(urban center), Marylebone Road
(roadside). Winter (October 1-March 31),
summer (April 1 -September 30), 1997-
1999.
No source- PM10: 0.5-0.65;
PM2.6: 0.6-0.7
Gas cooking- PM10: 0.6-0.9 (bedroom),
1.0-2.0 (lounge), 1.6-4.3 (kitchen);
PM2.6: 0.74-0.9 (bedroom), 0.9-1.6
(lounge), 1.6-2.9 (kitchen)
Smoking- PM10: 0.7-1.1 (bedroom), 1.1-
2.7 (lounge), 1.1-2.5 (kitchen);
PM2.6: 0.8-1.3 (bedroom), 1.3-2.8
(lounge), 1.4-2.6 (kitchen)
July 2009
A-320
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Study Design
Finf
Fromme et al. (2008,155147)
Objective: To characterize the chemical n/a
and morphological properties of PM in
classrooms and in corresponding outdoor
air.
Methods: PM Finf derived from sulfate Finf
and a correction factor that results from
division of B™ (increase of indoor PM per
outdoor PM, linear relationship) by Bsulf
(increase of indoor sulfate per outdoor
sulfate, linear relationship). If no indoor
source, the sulfate Fm is equal to the
sulfate I/O.
Subjects: Primary school in northern
Munich. Densely populated residential
area 160m away from a very busy street.
Classrooms had 21-23 students.
Sampling during teaching hours. Oct.-Nov.
2005.
PMio: Sulfate- 0.3,
Nitrate- 0.1,
CI- 0.6,
Na- 0.9,
Ammonium- 0.1,
Mg- 0.6,
Ca-1.4,
EC- 0.7,
0C- 1.1
PM2.6:
Sulfate- 0.4,
Nitrate- 0.2,
CI- 0.5,
Na- 0.6,
Ammonium- 0.3,
Mg- 0.5,
Ca-1.6
Guoet al. (2004,156506)1
Objective: To investigate pollutant
concentrations at air-conditioned and non-
air-conditioned markets. To compare
indoor air quality with the Hong Kong
standard.
Methods: PM10 concentrations measured
by Hi-Vol sampler correlated with
corresponding levels measured by Dust-
Trak monitor.
Subjects: 3 non-air-conditioned and 2 air-
conditioned markets in Hong Kong. Sept.
2001-Jan. 2002.
n/a
PM10: Non-air-conditioned-
conditioned- —0.98
- 0.7, Air-
Hanninenet al. (2004, 056812)2
Objective: T0 assess indoor PM2.6 by
origin and potential determinants.
Methods: Part of EXPOLIS study. Pump
and filter with gravimetric analysis.
Univariate single and stepwise-multiple
regression analyses.
Subjects: Residential homes in Athens,
Greece; Basle, Switzerland; Helsinki,
Finland; Prague, Czech Republic. Homes
by city: Athens 50, Basle 50, Helsinki
189, Prague 49.
PM2.6 (mean): Athens- 0.70 ± 0.12,Basle-
0.63 ± 0.15,Helsinki- 0.59 ±
0.17,Prague- 0.61 ± 0.14
Sulfur (mean): Athens- 0.82 ± 0.14,
Basle- 0.80 ± 0.19, Helsinki- 0.70 ±
0.20, Prague- 0.72 ± 0.16
PM2.6: Athens- —0.84, Basle- —1.37,
Helsinki- —1.30, Prague- —1.33
Sulfur: Athens- — 0.70, Basle- — 0.80,
Helsinki- —0.74, Prague- —0.77
Ho et al. (2004, 056804)3
Objective: PM2.6, OC, and EC exposure
assessment of occupied buildings located
near major roadways under natural
ventilation (NV) and mechanical
ventilation (MV).
Methods: Co-located mini-volume
samplers and Partisol model 2000
sampler with 2.5 micron inlet. IMPROVE
TOR carbon analysis.
Subjects: Occupants of MV (1 classroom
and office) and NV (3 residences)
buildings located within 10m of major
roadway; Hong Kong, China. Sep. 2002-
Feb. 2003.
PM2.6: 0.42
EC- MV: 0.42, NV: 0.76
OC- MV: 0.66, NV: 0.71
PM2.6 (average)- 0.2-1.6; MV (average)-
< 0.7; NV (average)- 0.6-1.6
EC- Range: 0.5±0.1 - 1.1 ±0.4
OC- Range: 0.6±0.2- 1.5±1.0
July 2009
A-321
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Study Design
Finf
Hoeketal. (2008,156554)
Objective: Exposure assessmentof
indoor/outdoor particle
relationships.RUPIOH study.
Methods: Sampling by condensation
particle counters and Harvard impactors.
Gravimetric analysis and reflectance.
Calculations performed for 24h avg
concentrations. Finf estimated by linear
regression analysis.
Subjects: 4 European cities (Helsinki,
Finland; Athens, Greece; Amsterdam, The
Netherlands; Birmingham, England).
Urban populations. >35yrs. Asthma or
COPD. Non smoking households. Work
< 16h/wk outside home. 153 homes
sampled Oct. 2002-Mar. 2004.
Regression slope for indoor vs. central n/a
site outdoor- PM2.6: 0.30-0.51;
PM10: 0.17-0.41;
PM10 2.B: 0.01-0.17;
Sulfate: 0.59-0.78,;
Soot: 0.43-0.87
Regression slope for indoor vs. residential
outdoor- PM2.6: 0.34-0.48;
PM10: 0.26-0.44;
PMio-2.b: 0.11-0.16;
Soot: 0.63-0.84
Hopke et al. (2003, 095544)
Objective: To use advanced factor
analysis models to identify and quantify
PM sources. 1998 BPMEES data.
Methods: PEM, outdoor and indoor
sampling of unoccupied apartment in
retirement facility. PMF used to derive
source contributions. Multilinear Engine
used to derive joint factors.
Subjects: 10 non-smoking elderly
subjects of mean age 84 who did not
cook.Towson, MD. July 26-Aug. 22,
1998.
Nitrate-sulfate- 0.03;
Sulfate- 0.38;
0C- 0.77;
MV Exhaust- 0.32
n/a
Hystad et al. (2008,190890)
Objective: To explore the feasibility of
modeling residential PM2.6 Finf for occupied
residences using data readily available for
most of North America.
Methods: Fm calculated by recursive
mass balance model where Finf is a
function of penetration efficiency,
particle removal rate, and air exchange.
Subjects: 46 residences in Seattle, WA
1999-2003. 38 nonsmoking residences in
Victoria, British Columbia, Canada 2006.
Heating (Oct.-Feb.) and nonheating
(March-Sept.).
Seattle: Mean (all)- 0.59 ± 0.21, Mean
(detached residences)- 0.60 ± 0.20
Victoria: Mean (all)- 0.62 ± 0.22, Mean
(detached residences)- 0.59 ± 0.22
n/a
Klinmalee et al. (2008,190888)
Objective: T0 monitor indoor and outdoor n/a
pollution in an university campus and
shopping center.
Methods: PM measured by PEM and
quartz filters. Analyzed for mass, water
soluble ions by ion chromatography, and
black carbon by a smokestain
reflectometer. I/O calculated for each
sample pair then average taken.
Subjects: University campus and
shopping center in northern suburb of
Bangkok, Thailand. Dec. 2005-Feb. 2006.
PM2.6: University- Weekdays: 0.6,
Weekends: 0.5; Shopping center-
Weekdays: 1.5, Weekends: 2.0
BC in PM2.6: University- 0.9, Shopping
center- 0.67
July 2009
A-322
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Study Design
Finf
Koistinen et al. (2004,156655)
Objective: To identify PM2.6 sources in
personal exposures with principal
component analysis of the elemental
compositions in residential outdoor,
indoor, and workplace indoor
microenvironments. Part of EXPOLIS
study.
Methods: Principal component analysis
to identify sources of microenvironmental
and personal PM2.6 exposure. Specific
mass contributions of sources calculated
by source reconstruction.
Subjects: Non smoking, 25-55yrs.
Helsinki, Finland. Oct. 1996-Dec. 1997.
n/a
Median seasonal- PM2.6: Winter- 0.77,
Spring- 1.03, Summer- 0.95, Fall- 0.92,
Total- 0.92
Pb: Winter- 0.67, Spring- 0.56, Summer-
0.86, Fall- 0.69, Total- 0.67
S: Winter- 0.60, Spring- 0.63, Summer-
0.90, Fall- 0.75, Total- 0.69
Br: Winter- 0.57, Spring- 0.72, Summer-
0.98, Fall- 0.89, Total- 0.77
Black Smoke: Winter- 0.65, Spring- 0.67,
Summer- 0.91, Fall- 0.88, Total- 0.79
Zn: Winter- 0.58, Spring- 0.75, Summer-
0.66, Fall- 0.75, Total- 0.68
Fe: Winter- 0.52, Spring- 0.96, Summer-
0.90, Fall- 0.95, Total- 0.83
K: Winter- 0.95, Spring- 1.05, Summer-
1.01, Fall- 1.08, Total- 1.05
CI: Winter- 1.01, Spring- 1.24, Summer-
1.37, Fall- 1.74, Total- 1.24
Al: Winter- 1.19, Spring- 1.08, Summer-
1.41, Fall- 2.20, Total-1.27
Li et al. (2003, 047845)
Objective: T0 establish effects of
evaporative coolers on indoor PM
concentrations.
Methods: Concurrent 10min avg indoor
and outdoor concentrations recorded for
2 days.I/O determined by equation based
on mass conversation principles.
Subjects: 10 homes with evaporative
coolers. El Paso, TX. June 22-Aug. 23,
2001.
n/a
PMio- All: 0.60, Cooler On: 0.57, Cooler
Off: 0.66
PM2.b- All: 0.65, Cooler On: 0.63, Cooler
Off: 0.73
Lunden et al. (2008,155949)
Objective: To investigate the
physiochemical processes that influence
the transport and fate of outdoor
particles to the indoor environment.
Methods: I/O calculated from
measurements of aerosols collected on
quartz filters.
Subjects: 3-bedroom single-story
unoccupied house in Clovis, CA. 3
periods: Oct. 9-23, 2000; Dec. 1-19,
2000; Jan. 12-23,2001.
n/a
PM2.6: Oct.- 0.46 ± 0.2, Dec.- 0.39 ±
0.2, Jan.- 0.38 ± 0.3, All periods- 0.41
± 0.2
Carbon: Oct.- 0.50 ± 0.1, Dec.- 0.46 ±
0.1, Jan.- 0.52 ± 0.2, All periods- 0.50
± 0.2
OC: Oct.- 0.48 ± 0.1, Dec.- 0.44 ± 0.1,
Jan.- 0.50 ± 0.2, All periods- 0.47 ± 0.2
Black carbon: Oct.- 0.60 ± 0.2, Dec.-
0.60 ± 0.2, Jan.- 0.65 ± 0.2, All
periods- 0.61 ± 0.2
Macintosh et al. (2009,190887)
Objective: T0 estimate the potential for
residential air cleaning systems to
mitigate exposure to fine particles of
ambient origin.
Methods: Multi-zone indoor air quality
model to examine annual, 24h avg and
diurnal concentrations of outdoor PM2.6 in
residential indoor air.
Subjects: Homes in Cincinnati,
Cleveland, and Columbus, OH that have
natural ventilation, forced air heating and
cooling with conventional in-duct
filtration, or forces air heating and
cooling with high-efficiency in-duct air
cleaning. 2005.
n/a
PM2.6 (range): Natural ventilation- 0.23-
0.97; Forced air - conventional filtration-
0.13-0.94; Forced air - high-efficiency
electrostatic- 0.02-0.80
July 2009
A-323
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Study Design
Finf
Martuzevicius et al. (2008,190886)
Objective: To determine the contribution
of traffic-related PM to the indoor
aerosols.
Methods: Receptor modeling based on a
PARAFAC model.
Subjects: 6 houses 30-300m from a
highway, with conventional windows,
central HVAC, and with smoking and
cooking allowed. Spring: Mar. 30-May
14, 2004. Fall: Sept. 13-0ct. 22, 2004.
Cincinnati, OH.
n/a
¦ PM2.6: Spring- 0.5 ± 0.2 - 2.9 ±
1.2; Fall- 0.7 ± 0.1 -4.7 ± 6.9
EC: Spring-0.3 ± 0.1 -2.2 ± 1.7:
Winter- 0.6 ± 0.1 - 1.3 ± 0.7
OC: Spring-1.0 ± 0.7 - 6.9 ± 3.9:
Winter- 1.2 ± 0.1 - 7.6 ± 10
Si: Spring- 0.4 ± 0.1 - 5.1 ± 3.9:
Winter- 0.5 ± 0.1 - 5.3 ± 4.5
S: Spring- 0.4 ± 0.1 - 0.7 ± 0.1: Winter-
0.5 ± 0.1 - 0.9 ± 0.4
Mn: Spring- 0.3 ± 0.2 - 0.8 ± 0.6:
Winter- 0.3 ± 0.2- 1.0 ± 0.2
Fe: Spring- 0.3 ± 0.0 - 1.3 ± 0.8:
Winter- 0.4 ± 0.1 - 0.9 ± 0.6
Zn: Spring- 0.3 ± 0.1 - 0.7 ± 0.6:
Winter- 0.6 ± 0.1 - 1.1 ± 0.8
Br: Spring- 0.3 ±0.1 -1.0 ± 0.5:
Winter- 0.2 ± 0.1 - 0.9 ± 0.6
Pb: Spring- 0.3 ± 0.3 - 0.9 ± 0.6:
Winter- 0.2 ± 0.2- 1.9 ± 2.3
Meng et al. (2005, 058595)
Objective: Analyses of RIOPA data,
which investigated relationships between
indoor, outdoor, and personal exposure
for several air pollutants.
Methods: PM measured on Teflon filters
collected by PEMs for 48h. The mass
balance model and RCS statistical model
used to estimate indoor and and personal
PM concentrations.
Subjects: 212 nonsmoking homes
sampled. Houston, TX: Los Angeles
County, CA: Elizabeth, NJ. Summer
1999-spring 2001, all 4 seasons.
PM2.5- 0.46
Los Angeles: PM2.b- Mean: 0.84, Median:
0.90:
EC- Mean: 0.93, Median: 0.92:
OC- Mean: 1.32, Median: 1.31
Elizabeth: PM2.b- Mean: 0.99, Median:
0.86:
EC- Mean: 1.0, Median: 0.85:
OC- Mean: 2.4, Median: 1.8
Houston: PM2.b- Mean: 1.16, Median:
1.02:
EC- Mean: 1.0, Median: 0.71:
OC- Mean: 2.25, Median: 2.35
Molnar et al. (2007,156774)	Objective: To characterize and compare
indoor and outdoor PM2.6 trace element
concentrations in difference
microenvironments related to children.
Methods: Elemental concentrations
analyzed using X-ray fluorescence
spectroscopy.
Subjects: 40 sampling sites (10
classrooms in 5 schools, 10 preschools,
20 non-smoking homes). 3 communities in
Stockholm, Sweden. Sampled once during
spring and once during winter. Dec. 1,
2003-July 1,2004.
PM2.E (containing sulfur or lead): 0.4-0.9 Sulfur (median): Both seasons- 0.61
(homes), 0.53 (schools), 0.69
(preschools):
Winter- 0.47 (homes), 0.36 (schools),
0.63 (preschools):
Spring- 0.63 (homes), 0.55 (schools),
0.90 (preschools)
Lead (median): Both seasons- 0.70
(homes), 0.59 (schools), 0.70
(preschools):
Winter- 0.62 (homes), 0.43 (schools),
0.63 (preschools):
Spring- 0.70 (homes), 0.64 (schools),
0.75 (preschools)
July 2009
A-324
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Study Design
Finf
Nn et al. (2005,155996)
Objective: To estimate PM exposures
following the September 11, 2001 attack
in NYC.
Methods: Outdoor PM2.6 interpolated and
used in a deterministic micro-
environmental model (INTAIR) to simulate
analytically concentrations in indoor
micro-environments. Linear regression
equations used.
Subjects: Lower Manhattan residents
divided into representative individuals -
home-maker, office/shop-worker,
student/child. Estimates Sept. 14-31.
n/a
Mean I/O in home simulated with INTAIR-
No Source: 0.6, Smoking: 1.9, Cooking:
1.3, Smoking and Cooking: 2.3:
I/O of micro-environments simulated by
analytical and empirical methods (no
indoor source)- Office/Shop: 0.4,
Classroom: 0.9, Transport Area: 1.9,
Store: 1.2
Paschold et al. (2003,156847)	Objective: To identify PM sources inside
homes with evaporative coolers.
Methods: PM element composition
analysis by ICP-MS.
Subjects: 10 residences. El Paso, TX.
Summer 2001.
n/a	PMio: Na- 0.33, Mg- 0.43, Al- 0.50, K-
0.48, Ca- 0.40, Ti- 0.52, Mn- 0.48, Fe-
0.46, Cu- 0.74, Zn- 0.52, Ba- 0.54, Pb-
0.76
PM2.b: Na- 0.20, Mg- 0.29, Al- 0.34, K-
0.30, Ca- 0.52, Ti- 0.40, Mn- 0.35, Fe-
0.30, Cu- 0.67, Zn- 0.34, Ba- 0.47, Pb-
0.51
Polidori et al. (2007,156877)	Objective: To investigate the	PM2.6: July 6-Aug. 20- 0.71 ±0.10: Aug. Only l/O's n1 considered
relationships of indoor and outdoor PM2.6,	24-0ct. 15- 0.60 ± 0.05: Oct. 19-Dec.
its components, seasonal variations, and	10- 0.59 ± 0.07: Jan. 4-Feb. 18- 0.45 ±
gaseous copollutants.	0.06
Methods: Finf estimated by analysis of	OC: July 6-Aug. 20- 0.86 ± 0.05: Aug.
I/O's and a recursive model technique.	24 0ct. 15- 0.77 ± 0.09: Oct. 19-Dec.
Subjects: 2 retirement facilities in Los
Angeles, CA. July 6-Aug. 20, 2005. Aug.
24-Oct. 15, 2005. Oct. 19-Dec. 10,	EC: July 6-Aug. 20- 0.73 ± 0.07: Aug.
2005. Jan. 4-Feb. 18, 2006.	24-0ct. 15- 0.71 ± 0.05: Oct. 19-Dec.
10-0.82 ± 0.07: Jan. 4-Feb. 18- 0.64 ±
0.10
10-0.77 ± 0.06; Jan. 4-Feb. 18- 0.64 ±
0.10
Ramachandran et al. (2003,188454) Objective: T0 examine variability in
measurements of 24h avg and 15min avg
PM2.6 concentrations.
Methods: Linear regression of
gravimetric measurements.
n/a	24h avg: Mean-1.7, Median-1.3,
Standard deviation- 1.6
15min avg: Mean- 2.7, Median- 1.2,
Standard deviation- 8.7
Subjects: 3 urban residential
neighborhoods in Minneaopolis-St. Paul,
MN. 9-10 nonsmoking residences. Spring
(April 26-June 2), summer (June 20-Aug.
10), fall (Sept. 23-Nov. 20) of 1999.
Rojas-Bracho et al. (2004, 054772) Objective: T0 examine determinants of
personal exposure to PM2.6, PM10, PM2.b-
10.
Methods: 2 sets of mixed models.
Personal exposures modeled as dependent
variables. Subject variability modeled
using random effects. Explanatory
variables and season modeled as fixed
effects.
n/a	PM2.6: Winter- Mean: 1.58, Median: 2.11;
Summer- Mean: 1.08, Median: 0.88
PM10: Winter- Mean: 2.02, Median: 3.77;
Summer-Mean: 1.14, Median: 1.05
PM2.B-10: Winter- Mean: 2.65, Median:
3.59; Summer- Mean: 1.26, Median: 1.39
Subjects: 18 COPD subjects in
nonsmoking households. Boston, MA.
Winters of 1996 and 1997, summer of
1996.
July 2009
A-325
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Study Design
Finf
Sarnat et al. (2006, 089166)
Objective: To assess the ability of
outdoor PM2.E, its volatile and nonvolatile
components and particle sizes to infiltrate
indoors.
Methods: PM2.Bmass contributions
estimated by the mean concentration
ratio between each component and PM2.B.
Indoor and outdoor particle
concentrations relationships examined by
Spearman correlation coefficient. I/O
concentration ratios used during
overnight (nonsource) period to estimate
fraction of ambient particles remaining
airborne indoors (Finf).
Subjects: 17 occupied, nonsmoking Los
Angeles, CA residences. July 28, 2001 -
Feb. 25, 2002.
PM2.6: Median- 0.48, Interquartile range-
0.39-0.57
BC: Median- 0.84, Interquartile range-
0.70-0.96
UFP (0.02-0.03 |jm): Median- 0.50,
Interquartile range- 0.39-0.60
UFP (0.08-0.3 |jm): Median- -0.75
Coarse particles (5-10 Mm): Median-
<0.17
PM2.6: Overnight- 0.40-0.57, Morning-
0.43-0.74, Afternoon- 0.45-0.90,
Evening- 0.42-0.82
BC: Overnight- 0.70-0.97, Morning- 0.67-
0.98, Afternoon- 0.77-1.04, Evening-
0.70-1.01
Stranger et al. (2008,190884)
Objective: To assess indoor air quality
by determining indoor and outdoor PM2.B
mass concentrations, elemental
composition, and gaseous compounds.
Methods: PM mass concentrations
determined gravimetrically.
Subjects: 27 primary schools in city
center and suburbs of Antwerp, Belgium.
Dec. 2002 and June 2003.
n/a
PM2.6: Urban- Range: 0.3-6.9, Average:
1.3: Suburban- Range: 0.2-8.8, Average:
2.3
V, Ni, Zn, Pb, Br, Mn: < 1
CI, Ca, Al, Si, K, Ti, Fe: >1
Black smoke: Urban- Average: Dec.- 0.7
± 0.1, June- 1.1 ± 0.3: Suburban: Dec.-
0.8 ± 0.2, June-1.0 ± 0.4
Stranger et al. (2009,190883)
Objective: To assess indoor air quality in
residences by quantifying various
gaseous pollutants, and PM mass
concentrations, elemental composition,
and water-soluble ionic content.
Methods: PM mass concentrations
gravimetrically determined. Elemental
bulk analysis on filters.
Subjects: 19 residential homes in
Antwerp, Belgium that were a subset of
participants in the ECRHS II study.
n/a
PMi: Houses 1-15- Average: 2.0, Range:
0.3-9.6; Smokers: Average- 3.9, Range-
1.2-9.7; Non-smokers: Average: 0.8,
Range- 0.3-14
PM2.6: Houses 1-15- Average: 1.5, Range:
0.4-5.4, Smokers average: 2.5, Smokers
range: 1.2-5.4, Non-smokers average:
0.8, Non-smokers range: 0.4-1.3; Houses
16-19- Average: 2.6, Range: 0.3-3.9
PM10: Houses 1-15- Average: 1.3, Range:
0.4-4.1, Smokers average: 2.1, Smokers
range: 1.1-4.1, Non-smokers average:
0.8, Non-smokers range: 0.4-1.2
Ca, Ti, V,Cr, Mn, Fe, Ni, Zn, Pb, Si, S, CI:
<1
K, Cu, Br, Al: >1
Turpin et al. (2007,157062)
Objective: To characterize and compare
outdoor, indoor, personal PM2.6 exposure.
Identify indoor and personal PM2.B
sources. Estimate outdoor PM2.6 effect on
indoor and personal PM2.B. RIOPA study.
Methods: Fm calculated in three ways:
RCS model used to obtain constant Finf.
Mass balance model shows Finf varying
with AER. Robust regression uses major
PM2.6 species for home-specific Finf.
Subjects: 309 nonsmoking adults and
118 children with no preexisting
conditions. 219 homes sampled. Elizabeth
NJ, Houston TX, and Los Angeles County
CA.
PM2.b- RCS model: 0.46; Least-Trimmed
Squared Regression- Mean: 0.69, Median-
0.70, SD: 0.23; Mass Balance Model-
-0.08--0.85
Los Angeles: PM2.b- Mean: 0.84, Median:
0.90;
EC- Mean: 0.93, Median: 0.92;
0C- Mean: 1.32, Median: 1.31
Elizabeth: PM2.b- Mean: 0.99, Median:
0.86;
EC- Mean: 1.0, Median: 0.85;
0C- Mean: 2.4, Median: 1.8
Houston: PM2.b- Mean: 1.16, Median:
1.02;
EC- Mean: 1.0, Median: 0.71;
0C- Mean: 2.25, Median: 2.35
July 2009
A-326
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Study Design
Finf
I/O
Wallace and Williams (2005, 057485)
Williams et al. (2003, 053338)
Williams et al. (2008,191201)
Wilson and Brauer (2006, 088933)
Wu et al. (2006,179950)
Objective: To estimate the contribution
of outdoor PM2.E to personal exposure in
high-risk subpopulations.
Methods: Longitudinal regressions of
estimated indoor and outdoor PM2.6for
Finf.
Subjects: 29 African-Americans with
hypertension and 8 with implanted
cardiac defibrillators. Measured
7d/season, 4 seasons in 2000-2001.
Raleigh, NC.
Objective: To estimate ambient PM2.6
contributions to personal and indoor
residential PM mass concentrations.
Methods: Fm estimated from least
squares, regression analysis, and mixed
model slope.
Subjects: Nonsmoking, ambulatory, a
50yrs. 2 cohorts: mostly Caucasian with
implanted cardiac defibrillators in Chapel,
NC; 30 African-Americans with controlled
hypertension in low-to-moderate SES
neighborhoods in Raleigh, NC. 7d/season,
4 seasons in 2000-2001.
Objective: To examine the spatial
variability of PM2.6 and PM10-2.B and their
components to determine the suitability
of conducting health outcome studies
using a central site monitor in a
metropolitan area having multiple source
impacts.
Methods: Gravimetric analysis of PM
mass concentrations. ED-XRF analysis of
PM elements.
Subjects: Non-smoking, ambulatory, and
living in detached homes and non-smoking
households. Detroit, Ml.
Objective: To provide additional insight
into factors affecting exposure to
airborne PM and the resultant health
effects.
Methods: Fm estimated by mass balance
equation.
Subjects: 16 nonsmoking subjects with
COPD. 54-86yrs. Vancouver, British
Columbia. April-Sept. 1998.
Objective: To assess personal PM2.6
exposures from ambient sources and
agriculture burning smoke.
Methods: Fnfestimated by RCS model.
Application of Robust regression
algorithm.
Subjects: 33 adult asthmatics. 18-
52yrs. Pullman, WA. Sept. 3, 2002-Nov.
1,2002.
Range: 0.35-0.87
Regression analysis: 0.43 ± 0.06
Mixed model slope: Mean- 0.45 ± 0.21,
Range- 0.05-0.94
PM2.6: Range- 0.16-6.45, Mean- 0.7 ±
0.33, Median- 0.70 (indicate indoor sulfur
source when Finf > 1)
Sulfate: 0.72
Range: 0.25-0.94
PM2.6: Mean- 1.08 ± 1.05, Median- 0.75,
Range- 0.24-9.48
Sulfur: Mean- 0.59 ± 0.16, Median-
0.58, Range- 0.17-1.06
n/a
n/a
Least squares estimate of indoor filtration n/a
factors: Mean- 0.42 ± 0.38, Range- ¦
0.55 to 1.62
July 2009
A-327
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Study Design
Finf
Yang et al. (2009,190885)
Objective: To characterize the
concentrations of different indoor air
pollutants.
Methods: PMio collected on pall flex
membrane filer using MiniVol portable air
samplers. Arithmetic and geometric
means calculated for indoor
concentrations. Differences in
concentrations measured by Kruskal-
Wallis test.
Subjects: 55 schools in 6 metropolitan
areas in Korea. Samples from a
classroom, laboratory, and computer
classroom. 3 seasons, July-Dec. 2004.
n/a
PMio: Classroom- Summer: 1.98,
Autumn: 2.25, Winter: 2.07, Total: 2.06
Laboratory- Summer: 1.33, Autumn:
1.32, Winter: 1.72, Total: 1.46
Computer classroom: Summer- 0.77,
Autumn: 1.43, Winter: 2.08, Total: 1.43
Zhao and Hopke (2004,100956)	Objective: To characterize airborne n/a	n/a
particle sources that are common to
personal, residential indoor, residential
outdoor, and ambient environments. To
study relationships between PM2.6
exposure and emission sources.
Methods: Expanded receptor model.
Subjects: Nonsmoking, ambulatory,
£50yrs. 2 cohorts: mostly Caucasian
with implanted cardiac defibrillators in
Chapel, NC; 30 African-Americans with
controlled hypertension in low-to-
moderate SES neighborhoods in Raleigh,
NC. 7d/season, 4 seasons in 2000-2001.
Zhu et al. (2005,190081)	Objective: T0 determine penetration
behavior of outdoor ultrafine particles
into indoor environments in areas close to
freeways.
Methods: Dynamic mass balance model.
Subjects: 4 2-bedroom apartments
within 60m from the center of the 405
Freeway in Los Angeles, CA. Non-
smoking tenants. 2 sampling periods (non-
cooking, non-cleaning): Oct.-Dec. 2003
and Dec. 2003-Jan. 2004.
n/a	Highest (largest ultrafine particles- 70-
100nm): 0.6-0.9
Lowest (smallest ultrafine particles-10-
20nm): 0.1-0.4
1.	I/O estimated from Figure 8 in study.
2.	I/O calculated from indoor and outdoor concentrations in Table 1 in study.
3.	Fint measured by coefficient of determination, R2.
4.	RIOPA calculated l/O's.
5.	I/O calculated from mean and median indoor and outdoor concentrations listed in Table 1 of study.
6.1/O's estimated from Figure 3 in study.
7.	Mean and median I/O concentrations calculated from all residences in study.
8.	Fint estimated from Figure 2 in study.
9.	Fm presented in box plot (Figure 8), however data is difficult to deduce. No numeric values reported.
July 2009
A-328
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-55. Summary of PM - copollutant exposure studies.
Reference
PM metric
Copollutant metric
Association between PM
and copollutant
Primary findings
Fruin etal. (2008,097183)
In-vehicle UFP, BC, PM-bound
PAH
In-vehicle NOx, CO
R UFP PM2.6 NO BC
CO CO?
UFP 1 0.71 0.97 0.95 0.63
0.72
PM2.6	1 0.69 0.89
0.66 0.68
NO	1 0.91 0.78
0.85
BC	1 0.65
0.74
CO	1
0.94
CO?
1
Note that these correlations are
computed from data presented
by Fruin et al. (2008, 097183)
for mean concentrations at
different loc ations.
Measurements of freeway UFP,
BC, PM-bound PAH, and NO
concentrations were roughly one
order of magnitude higher than
ambient measurements. Multiple
regression analysis suggests
these concentrations were a
function of truck density and
total truck count.
Schwartz et al. (2007, 090220)
Ambient and personal PM2.6 data
from the Baltimore panel study
Ambient and personal O3 and
NO2 data from the Baltimore
panel study.
Median fB for regressions:
Ambient PM2.6 Ambient O3
Ambient NO2
Personal PM2.B 0.0143
0.0016 0.0115
Personal PM2.6 of ambient origin
0.0183 -0.0037
0.0124
Personal S042~ 0.0051
0.0035 0.0006
Results suggest that ambient O3
exposure may be related to
personal SO42' exposure but not
to personal PM2.6 exposure on
the whole. Ambient NO2
exposure was associated with
personal PM2.6 exposure,
possibly because both have
traffic sources.
Personal O3	0.0014
0.0010	0.0009
Personal NO2	0.0015
0.0009	0.0010
July 2009
A-329
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
PM metric
Copollutant metric
Association between PM
and copollutant
Primary findings
Tolbert et al. (2007,090316)
Ambient PMio, PMio-2.5, PM2.6,
EC, 0C, TC, SO42', water-soluble
metals, oxygenated
hydrocarbons
Ambient O3, NO2, CO, SO2
PMm03
PMc PM2.
NO2
5
CO
SO2
PM10I.O




Os 0.6
1.0



NO2 0.5
0.4
1.0


CO 0.5
0.3
0.7
1.0

SO2 0.2
0.2
0.4
0.3
1.0
PMc 0.7
1.0
0.4
0.5
0.4
0.2
PM2.6
0.2
0.8
0.5
0.6
1.0
0.6
0.4
S042 0.7
0.3
0.6
0.8
0.1
0.1
0.1
EC 0.6
0.5
0.4
0.7
0.6
0.7
0.2
OC 0.7
0.5
0.5
0.7
0.6
0.6
0.2
TC 0.7
0.5
0.5
0.7
0.7
0.6
0.2
Metals
0.1
0.7
0.5
0.4
0.7
0.3
0.4
OHC 0.5
0.4
0.4
0.5
0.2
0.3
0.1
Low correlations were seen
between SO2 and PM
constituents. Components were
used in a multi-pollutant model to
predict emergency department
visits in Atlanta. CO was found
to be the most significant
predictor of cardiovascular
disease visits in one-, two-, and
three-pollutant models, and O3
was the most significant
predictor of respiratory disease
visits in one-, two-, and three-
pollutant models.
S042 EC OC TC
Metals OHC
SO421.0
EC 0.3 1.0
OC 0.3 0.8 1.0
TC 0.3 0.9 1.0 1.0
Metals 0.7 0.5 0.5 0.5
1.0
OHC 0.5 0.4 0.4 0.4 0.5
1.0
July 2009
A-330
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
PM metric
Copollutant metric
Association between PM
and copollutant
Primary findings
Brook et al. (2007, 091153)
Anbient PMio, PMio-2.5, PM2.6,
SO42', and trace metals in 10
Canadian cities.
Ambient NO2, NO
R with NO2 (min, Max)
NO? 1.00(1.00,1.00)
NO 0.67(0.51,0.77)
PM2.6 0.54(0.45,0.71)
PMio-2.5 0.31 (0.04, 0.50)
PMio0.50 (0.23, 0.70)
SO42- 0.33(0.10,0.48)
Fe
Zn
Ni
Mn
As
Al
Cu
Pb
Si
Se
0.44 (0.29, 0.56)
0.39 (0.28, 0.52)
0.20 (0.06, 0.40)
0.51 (0.37, 0.62)
0.21 (0.07, 0.39)
0.07 (-0.17, 0.18)
0.03 (-0.07, 0.15)
0.28(0.16, 0.39)
0.19(0.00, 0.32)
0.14 (-0.04, 0.35)
NO2 showed the strongest
association with mortality, but it
is unclear if this association is
due to health effects of NO2 or
health effects of copollutant PM.
Ito et al. (2007,156594)
Ambient PM2.6
Ambient O3, NO2, SO2, CO
Shown in figure format only.
Authors tested relationship
between meteorological
variables and copollutants to
determine if multi-pollutant
models are impacted by spatial
or temporal variation or by
meteorological conditions.
Multicollinearity varied by
pollutant and season.
Kaur et al. (2005, 086504)
Fixed-site and personal PM2.6,
personal UFP
Fixed site and personal CO
Personal R:

PM2.6
UFP
CO
PM2.
5 1
0.5
0.2
UFP
0.5 1
0.7

CO
0.2 0.7
1

Fairly low correlation was
observed between PM2.6 and CO
and between PM2.6 and UFP,
stronger correlations between
UFP and CO.
Kauretal. (2005,088175)
Fixed-site and personal PM2.6
analyzed post-sample for light
absorbance (as indicator for
carbonaceous aerosol), personal
UFP
Fixed site and personal CO
Personal R:
R
PM2.!

Abs
CO
UFP
PM2J
5
1
0.3
¦0.1
0.0
Abs
0.3
1
0.2
0.7

CO
¦0.1
0.2
1
0.1

UFP
0.0
0.7
0.1
1

Strongest correlation observed
between UFP and absorption,
which is reasonable given that
much absorptive carbonaceous
aerosol is in the ultrafine range.
Sorenson et al. (2005, 089428)
Personal, indoor residential, and
outdoor residential PM2.6 and BC
Personal, indoor residential, and
outdoor residential NO2
Personal exposure regression
coefficients to:
PM2.6 BC NO2
Bedroom 0.72 0.47 0.70
Front door0.46 0.61 0.60
Background 0.29 0.03 0.56
Personal NO2 concentration is
more strongly influenced by
background than PM2.6 or BC.
Sabin et al. (2005, 087728)
BC, particle-bound PAH on a
school bus.
NO2 on a school bus.
BC PB-PAH NO2
BC 1 0.94 0.49
PB-PAH 1 0.37
NO2	1
Note that these correlations are
computed from data presented
by Sabin et al. for mean
concentrations when the test
bus travelled behind different
vehicles.
Less correlation was observed
between NO2 and PM species.
This study was aimed more at
fuel choices and control
technologies for children's
exposures on school buses.
July 2009
A-331
DRAFT-DO NOT CITE OR QUOTE

-------
pollutants shown.
„	Association between PM „ .
Reference PM metric Copollutant metric , . Primary findings
			and copollutant	'	_	
Lai et al. (2004, 0568111	Microenvironmental and personal Microenvironmental and personal R PM2.6 TVOC NO2 The EXPOLIS Oxford study was
PM2.6 and trace elements	VOCs, NO2, and CO.	j^qq g 21	more focused on the indoor-
outdoor exposure relationship,
"¦21	but the correlation results
0.41	showed no important
0 22	relationships between the
NO2 -0.1
¦0.02
¦0.16
0.09-0.11
¦0.01
¦0.23
0.03
CO 0.07
0.07
0.3
Correlation coefficients listed (in
order) for personal exposure,
residential indoor, residential
outdoor, and workplace indoor.
Morning and evening
measurements of PM2.6 were on
avg higher and more variable
than for benzene and CO (in
order). Benzene and CO had
higher and more variable
concentrations for minibuses
than for buses and metros,
respectively, while PM2.6
concentrations were not
substantially different for buses
and minibuses.
Sarnat et al. (2001, 019401) Fixed site and personal PM2.6 Ambient O3, NO2, SO2, and CO R PM2.6
monitors.	CO
PM2.6 1
0.15
Os -0.72
¦0.06
NO2 0.75 0.71 1
0.75
SO2 -0.17 0.41-0.17
1 -0.32
CO 0.69-0.67 0.76-0.12
1
Gomez-Perales et al. (2007,
138816; 2004, 054418)
Microenvironmental PM2.6 with
S042", NOs-, EC, OC.
Microenvironmental CO.
Ratio of Cone PM2.6 CO
Benzene
Minibus/Bus 1.041.54 2.01
1.20 1.40 1.33
Minibus/Metro 1.70 2.02 3.20
1.43 3.03 3.10
O3 NO2 SO2 Strong association between
ambient NO2 and personal PM2.6
0 67 0 37 ¦ suggests that ambient gas may
be a suitable surrogate for
personal exposure.
July 2009
A-332
DRAFT-DO NOT CITE OR QUOTE

-------
Table A-56. Summary of studies relating PM, SES, and mortality and/or morbidity.
Reference
Population Studied
Data
interval
Metrics Used
(health; pollutant; SES
variable)
Study Outcome
Bateson and
Schwartz (2004,
086244)
Residents (>65) of Cook
Co. IL with prior cardiac or
respiratory hospitalization,
1988-1991
Days
All-cause mortality; PM10; median
household income, % with bachelor's
degree, % not speaking English at
home
No significant change in mortality with a 10 |jg|m3 increase in PM10
with SES variables.
Cifuentes et al.
(2000,010351)
Residents (aged 25-64) of
Santiago, Chile, 1988-1996
Days
Non-trauma mortality; PM2.6;
educational level
Relative risks of non-trauma mortality were at or near significance in
the group having only elementary education.
Filleul et al.
(2003, 087403)
Residents (aged >65) of
Bordeaux, France, 1988-
1997
Days
Non-trauma mortality; BC (10th or
90th percentile levels); education
level, previous occupation (domestic,
skilled, intellectual)
No significant effect between BC and non-trauma mortality was
observed for either SES variable.
Filleul et al.
(2004, 087404)
Residents (aged >65) of
Bordeaux, France, 1988-
1997
Days
Non-trauma mortality, cardio-
respiratory mortality; BC;
educational level, previous
occupation (never worked, white-
collar, blue collar)
Blue collar SES group had a significant odds ratio of non-trauma
mortality; high education level had a significant odds ratio for cardio-
respiratory mortality.
Filluel et al.
(2005,087357)
Adults (aged 25-59 at
enrollment) in 7 French
cities, 1974-2000
Years
Non-trauma mortality; BC, TSP;
educational level
No trend as a function of education level.
Finkelstein et al.
(2003, 056117)
Adults (aged >40) in
Hamilton-Burlington,
Canada, 1992-2001
Years
Non-trauma mortality; TSP; mean
household income
Significantly higher relative risk as a function of TSP exposure for
low and high income strata
Finkelstein et al.
(2003, 056117)
Adults (aged >40) in
Hamilton-Burlington,
Canada, 1992-2001
Years
Cardio-vascular mortality; Pollution
index (TSP and SO2) (regional, urban,
near-road), traffic proximity;
deprivation index
No significant relative risk as a function of pollution index or traffic
proximity.
Gouveia 81
Fletcher (2000,
012132)
Residents (aged >65) of
Sao Paulo, Brazil, 1991-
1993
Days
Non-trauma mortality; PM10;
composite SES index
Non-significant results show relative risk of non-trauma mortality as
a function of PM10 slightly higher in advantaged neighborhoods.
Gwynn &
Thurston (2001,
017206)
Residents of NY City, 1988-
1990
Days
Respiratory hospital admissions;
PM10, sulfate; race, insurance
Higher but non-significant relative risk for non-whites than whites
but neither with relative risk significantly different from 1 for PM10;
relative risk significantly higher than 1 for sulfate among non-whites.
Hoel et al. (2002,
042364)
Adults (aged 55-69 at
enrollment) in The
Netherlands, 1992-2000
Years
Non-trauma mortality; BC (regional,
urban, near-road); educational level
No significant difference in relative risk as a function of BC exposure
for education level
Ito and Thurston
(1996, 078841)
Residents of Cook County,
IL, 1985-1990
Days
Mortality; PM10; race, sex
Mortality increased with PM10, effects of sex and race were noted
with black females > white females > black males > white males
Krewski et al.
(2000,012281)
Adults (aged 25-74 at
enrollment) in Six Cities
cohort, 1974-1991
Years
Non-trauma mortality, cardio-
pulmonary mortality; PM2.6, sulfates;
educational level
Relative risk significantly greater than 1 for non-trauma mortality
among those with less than high school education caused by
increased PM2.6 and sulfate exposures
Krewskietal. Adults (aged >30 at
(2000, 012281) enrollment) in American
Cancer Society cohort,
follow-up 1982-1989
Years	Non-trauma mortality, cardio-
pulmonary mortality; PM2.6, sulfates;
educational level
Relative risk significantly greater than 1 for non-trauma and cardio-
pulmonary mortality as a function of PM2.6 exposure for less than
high school and high school education; relative risk significantly
greater than 1 for non-trauma and cardio pulmonary mortality as a
function of sulfate exposure for less than high school education.
Lee et al. (2006, Children (aged < 15) in Days	Hospitalized for asthma; PM10; SES
098248)	Seoul, Korea, 2002	(listed as "high," "medium," or "low"
of monitor site without explanation
of criteria)
PM10 level does not vary linearly with increasing SES. Relative risk
significantly greater than 1 for high and low SES.
July 2009
A-333
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Population Studied
Data
interval
Metrics Used
(health; pollutant; SES
variable)
Study Outcome
Linn et al. (1999,
011680)
Residents of South Coast
Air Basin, CA, 1992-1995
Days
Respiratory and cardiovascular
hospital admissions; PMio; sex,
ethnicity (white, black, Hispanic,
other)
Impact of PMio on cardiovascular effects increased in blacks and
whites relative to Hispanics and others.
Martins et al.
(2004, 087457)
Residents (aged >60) of
six zones of Sao Paulo,
Brazil, 1997-1999
Days
Respiratory mortality; PMio; % with
college education, % families with
monthly income > $3500, % living
in slums
% with college education and % families with monthly income
> $3500 have negative impact of effect of PMio on respiratory
mortality, % people living in slums had positive effect.
Norris et al.
(2000, 087104)
Children (aged < 18) in
Seattle, WA, 1995-1996
Days
Emergency room visits for asthma;
PMio; high vs. low emergency room
use
Relationship between PMio and emergency room visits not
significantly impacted by overall emergency room use.
O'Neill et al.
(O'Neill et al.,
2004, 055597)
Residents (aged >65) of
Mexico City, Mexico, 1996-
1998
Days
Non-trauma mortality; PMio and O3;
% homes with electricity, % homes
with piped water, % literacy, %
indigenous language speakers
PMio not associated with non-trauma mortality (but significant
associations for O3).
Ou et al. (2008,
189955)
Residents (aged >30) of
Hong Kong, 1998
Days
Non-trauma mortality; PMio; housing
type, occupational level, education
level
Housing type and blue-collar caused significantly greater impact of
PMio on mortality compared with single family housing or white-
collar and never employed, respectively.
Pope et al.
(2002, 024689)
Adults (aged >30 at
enrollment) in American
Cancer Society cohort,
follow-up 1982-1989
Years
Mortality; PM2.6; education level
Non-trauma mortality increased with PM2.6 increase; greatest impact
among those with less than high school education.
Romieu (2004,
093074)
Children (1 mo. - 1 yr.) in
Ciudad Juarez, Mexico,
1997-2001
Days
Total mortality, respiratory
mortality; PMio; composite SES
index
No significant association between pollutants and total mortality;
significant odds ratio for respiratory mortality and PMio for lowest
SES; nearly significant association between SES and PMio
Samet et al.
(2000,013132)
Residents (all ages) of 20
US cities, 1987-1994
Days
Non-trauma mortality; PMio (adj for
O3, SO2, NO2, CO); % high school
graduates, % annual income
< $12,675, % annual income
>$100,000
No significant association between PMio-related non-trauma
mortality and SES variables.
Schwartz (2000,
002470)
Residents (all ages) of 10
US cities, 1986-1993
Days
Non-trauma mortality; PMio;
unemployment rate, % below
poverty level; % with college degree
No significant difference in the effect of poverty, college degree, or
unemployment rate on the influence of PM on mortality, but
unemployment rate effect slightly higher.
Tolbert et al.
(2000, 001993)
Children (aged < 16) in
Atlanta, GA, 1993-1995
Days
Emergency room visits for asthma;
PMio; race, Medicaid status, sex
Impact of PMio on asthma emergency room visits was not impacted
by any SES variable.
Villeneuve et al.
(2003, 055051)
Residents (aged >65) of
Vancouver, Canada, 1986-
1999
Days
Non-trauma mortality; TSP, PMio,
and PM2.E; mean family income
Significantly higher non-trauma mortality as a function of TSP for
high and low income.
Wheeler & Ben-
Schlomo (2005,
089860)
Respondents to Health
Survey of England, 1995-
1997
n/a
Decreased lung function, asthma
prevalence; air quality index based
on PMio, NO2, SO2, benzene; social
class, sex
In urban areas, lower SES significantly associated with poor air
quality; in rural areas, higher SES significantly associated with poor
air quality. Lower SES was shown to impact the relationship
between PMio and lung function among men but not women.
Wilson et al.
(2007,157149)
Residents (all ages) of
Phoenix, AZ, 1995-1997
Days; lag 0-
5, 6-day
moving avg
Non-trauma mortality,
cardiovascular mortality; PM2.6,
PMio-2.5; % < HS diploma, % below
poverty level, location within city
The lower SES region of Central Phoenix had higher risk of mortality
as a function of PM2.6 exposure. Modification of the effect of PMio-
2.5 on mortality was observed for the higher SES region.
Wojtyniak et al.
(2001,090372)
Residents (aged 0-70 or
>70) of Cracow, Lodz,
Poznan, and Wrockrw
(Poland), 1990-1996
Days
Non-trauma and cardiovascular
mortality; BC; educational level
Non-trauma and cardiovascular mortality was significantly
associated with BC for those with less than secondary education.
July 2009
A-334
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Population Studied
Data
interval
Metrics Used
(health; pollutant; SES
variable)
Study Outcome
Wong et al.
(2008,157151)
Residents of 209 tertiary
planning units (smallest
classification for a town),
1996-2002
Days
Non-trauma and cardiovascular
mortality; PMio; social deprivation
index
Significant associations between PMio and non-trauma and
cardiovascular mortality for medium and high social deprivation
index.
Zanobetti et al.
(2000,011979)
Residents of 10 US cities,
1985-1994
Days
Respiratory and cardiovascular
hospital admissions; PMio; %
poverty, % non-white
No significant effect of SES factors on relationship between hospital
admissions and PMio.
Zanobetti et al.
(2000,012187)
Medicare recipients in Cook
County, IL, 1985-1994
Days
Respiratory and cardiovascular
hospital admissions; PMio; race, sex
No significant effect of SES factors on relationship between hospital
admissions and PMio.
Zanobetti &
Schwartz (2000,
010198)
Residents (all ages) of
Chicago, Detroit,
Minneapolis-St. Paul, and
Pittsburgh, 1986-1993
Days
Non-trauma mortality; PMio Higher but non-significant % increase in non-trauma mortality with
(excluding days when concentrations 10 jjg/m3 increase in PMio for people with less than high school
exceeded 150 jjg/m3); education education,
below or above high school
Zeka et al.
(2006, 088749)
Residents (all ages) of 20
US cities, 1989-2000
Days
Non-trauma mortality, respiratory
mortality, cardiac mortality,
mortality from infarction, mortality
from stroke; PMio; educational level
No significant relationship between increased mortality (any type)
with 10 jjg/m3 increase in PMio for any SES factors.
Some studies measured constituents other than PM; those metrics and results are not reported here.
Source: Adapted from Laurent et al. (2008,156672) and O'Neill et al. (2003, 090310).
July 2009
A-335	DRAFT-DO NOT CITE OR QUOTE

-------
Annex A References
Abou Chakra OR; Joyeux Mi Nerriere Ei Strub MPi Zmirou-Navier D. (2007). Genotoxicity of organic extracts of
urban airborne particulate matter- an assessment within a personal exposure study. Chemosphere, 66-
1375-81. 098819
Abu-Allaban Mi Gillies JAi Gertler AWi Clayton Ri Proffitt D. (2007). Motor vehicle contributions to ambient
PM10 and PM2.5 at selected urban areas in the USA. Environ Monit Assess, 132: 155-63. 098575
Abu-Allaban Mi Rogers CFi Gertler AW. (2004). A quantitative description of vehicle exhaust particle size
distributions in a highway tunnel. J Air Waste Manag Assoc, 54: 360-366. 156187
Adar SDi Adamkiewicz Gi Gold DRi Schwartz Ji Coull BAi Suh H. (2007). Ambient and microenvironmental
particles and exhaled nitric oxide before and after a group bus trip. Environ Health Perspect, 115: 507"
12. 098635
Adgate JLi Mongin SJi Pratt GCi Zhang Ji Field MPi Ramachandran Gi Sexton K. (2007). Relationships
between personal, indoor, and outdoor exposures to trace elements in PM(2.5). Sci Total Environ, 386: 21-
32.156196
Adgate JLi Ramachandran Gi Pratt GCi Waller LAi Sexton K. (2002). Spatial and temporal variability in
outdoor, indoor, and personal PM25 exposure. Atmos Environ, 36: 3255-3265. 030676
Adgate JLi Ramachandran Gi Pratt GCi Waller LAi Sexton K. (2003). Longitudinal variability in outdoor,
indoor, and personal PM25 exposure in healthy non-smoking adults. Atmos Environ, 37: 993-1002.
040341
Alander TJAi Leskinen APi Raunemaa TMi Rantanen L. (2004). Characterization of diesel particles: effects of
fuel reformulation, exhaust aftertreatment, and engine operation on particle carbon composition and
volatility. Environ Sci Technol, 38: 2707-2714. 055650
Allen JGi Webster TFi McClean MDi Stapleton HMi Nelson JW. (2007). Personal exposure to Polybrominated
Diphenyl Ethers (PBDEs) in residential indoor air. Environ Sci Technol, 41: 4574-4579. 156207
Allen Ri Larson T> Sheppard L> Wallace LLiu L-JS. (2003). Use of real-time light scattering data to estimate the
contribution of infiltrated and indoor-generated particles to indoor air. Environ Sci Technol, 37: 3484"
3492.053578
Allen R; Wallace L> Larson T> Sheppard L> Liu L-JS. (2007). Evaluation of the recursive model approach for
estimating particulate matter infiltration efficiencies using continuous light scattering data. J Expo Sci
Environ Epidemiol, 17: 468-477. 154226
Anderson TL; Ogren JA. (1998). Determining Aerosol Radiative Properties Using the TSI 3563 Integrating
Nephelometer. Aerosol Sci Technol, 29: 57"69. 156213
Andresen PR; Maynard A; Ramachandran G; Pai P. (2005). Women's personal and indoor exposures to PM2.5 in
Mysore, India: Impact of domestic fuel usage. Atmos Environ, 39: 5500-5508. 156216
Annesi-Maesano I; Moreau D; Caillaud D; Lavaud F; Le Moullec Y; Taytard A; Pauli G; Charpin D. (2007).
Residential proximity fine particles related to allergic sensitisation and asthma in primary school
children. Respir Med, 101: 1721-1729. 093180
Arhami M; Kuhn T; Fine PM; Delfino RJ; Sioutas C. (2006). Effects of Sampling Artifacts and Operating
Parameters on the Performance of a Semicontinuous Particulate Elemental Carbon/Organic Carbon
Monitor. Environ Sci Technol, 40: 945-954. 156224
Arhami M; Polidori A; Delfino RJ; Tjoa T; Sioutas C. (2009). Associations between personal, indoor, and
residential outdoor pollutant concentrations: implications for exposure assessment to size-fractionated
particulate matter. , 59: 392-404. 190096
Arnott WP; Moosmuller H; Rogers CF; Jin TJ Bruch R. (1999). Photoacoustic spectrometer for measuring light
absorption by aerosol; instrument description. Atmos Environ, 33: 2845-2852. 020650
Ashbaugh LL. (1983). A statistical trajectory technique for determining air pollution source regions. J Air Pollut
Control Assoc, 33: 1096-1098. 156229
Ashbaugh LL; Myrup LO; Flocchini RG. (1984). A principal component analysis of sulfur concentrations in the
western United States. Atmos Environ, 18: 783-791. 045148
Babich P; Wang PY; Allen G; Sioutas C; Koutrakis P. (2000). Development and Evaluation of a Continuous
Ambient PM2.5 Mass Monitor. Aerosol Sci Technol, 32: 309-324. 156239
Balakrishnan K; Sankar S; Parikh J; Padmavathi R; Srividya K; Venugopal V; Prasad S; Pandey VL. (2002).
Daily average exposures to respirable particulate matter from combustion of biomass fuels in rural
households of southern India. Environ Health Perspect, 110: 1069-1075. 156247
Balasubramanian R; Lee SS. (2007). Characteristics of indoor aerosols in residential homes in urban locations: a
case study in Singapore. J Air Waste Manag Assoc, 57: 981-990. 156248
July 2009
A-336
DRAFT-DO NOT CITE OR QUOTE

-------
Barn P; Larson T> Noullett Mi Kennedy Si Copes Ri Brauer M. (2008). Infiltration of forest fire and residential
wood smoke- an evaluation of air cleaner effectiveness. J Expo Sci Environ Epidemiol, 18- 503-511.
156252
Bateson TFi Schwartz J. (2004). Who is sensitive to the effects of particulate air pollution on mortality? A case-
crossover analysis of effect modifiers. Epidemiology, 15- 143-149. 086244
Baxter LKi Clougherty JEi Laden Fi Levy JI. (2007). Predictors of concentrations of nitrogen dioxide, fine
particulate matter, and particle constituents inside of lower socioeconomic status urban homes. J Expo
Sci Environ Epidemiol, 17: 433-444. 092726
Baxter LKi Clougherty JEi Paciorek CJi Wright RJi Levy JI. (2007). Predicting residential indoor
concentrations of nitrogen dioxide, fine particulate matter, and elemental carbon using questionnaire
and geographic information system based data. Atmos Environ, 41- 6561-6571. 092725
Bein KJi Zhao Yi Wexler ASi Johnston MV. (2005). Speciation of size-resolved individual ultrafine particles in
Pittsburgh, Pennsylvania. J Geophys Res, 110: D07S05. 156265
Birch ME. (1998). Analysis of carbonaceous aerosols' interlaboratory comparison. Analyst, 123- 851-857. 024953
Biswas Si Fine PMi Geller MDi Hering SVi Sioutas C. (2005). Performance evaluation of a recently developed
water-based condensation particle counter. Aerosol Sci Technol, 39- 419-427. 150694
Blanchard Pi Hopper JF. (1997). Concentrations and distributions of PAHs and n-alkanes in atmospheric
aerosols at non-urban sites in Ontario, Canada. Presented at 6th International Conference on
Carbonaceous Particles in the Atmosphere; Vienna, Austria. 157195
Blanchard P; Brook JR; Brazal P. (2002). Chemical characterization of the organic fraction of atmospheric
aerosol at two sites in Ontario, Canada. J Geophys Res, 107- D21. 189737
Blazso M> Janitsek S> Gelencser A; Artaxo P; Graham B; Andreae MO. (2003). Study of tropical organic aerosol
by thermally assisted alkylation-gas chromatography mass spectrometry. J Anal Appl Pyrol, 68- 351-369.
156278
Branis M; Rez cov P; Domasov M. (2005). The effect of outdoor air and indoor human activity on mass
concentrations of PM10, PM2.5, and PMI in a classroom. Environ Res, 99: 143-149. 156290
Brauer Mi Gehring Ui Brunekreef Bi De Jongste Ji Gerritsen Ji Rovers Mi Wichmann H"Ei Wijga Ai Heinrich J.
(2006). Traffic-related air pollution and otitis media. Environ Health Perspect, 114: 1414-1418. 090757
Brauer Mi Hoek Gi Smit HAi De Jongste JCi Gerritsen Ji Postma DSi Kerkhof Mi Brunekreef B. (2007). Air
pollution and development of asthma, allergy and infections in a birth cohort. Eur Respir J, 29: 879-888.
090691
Briggs DJi de Hoogh Ki Morris Ci Gulliver J. (2008). Effects of travel mode on exposures to particulate air
pollution. Environ Int, 34: 12-22. 156294
Brook JRi Burnett RT; Dann TF; Cakmak S; Goldberg MS; Fan X; Wheeler AJ. (2007). Further interpretation of
the acute effect of nitrogen dioxide observed in Canadian time-series studies. J Expo Sci Environ
Epidemiol, 17 : S36-S44. 091153
Brown K; Sarnat J; Suh H; Coull B; Spengler J; Koutrakis P. (2008). Ambient site, home outdoor and home
indoor particulate concentrations as proxies of personal exposures. J Environ Monit, 10: 1041-1051.
190894
Brunekreef B; Janssen N; De Hartog J; Oldenwening M; Meliefste K; Hoek G; Lanki T; Timonen K; Vallius M;
Pekkanen J; Van Grieken R. (2005). Personal, indoor, and outdoor exposures to PM2.5 and its
components for groups of cardiovascular patients in Amsterdam and Helsinki. , 127: 1-70. 090486
BeruBe KA; Sexton KJ; Jones TP; Moreno T; Anderson S; Richards RJ. (2004). The spatial and temporal
variations in PM10 mass from six UK homes. , 324: 41-53. 189731
Cabada JC; Rees S; Takahama S; Khlystov A; Pandis SN; Davidson CI; Robinson AL. (2004). Mass size
distributions and size resolved chemical composition of fine particulate matter at the Pittsburgh
supersite. Atmos Environ, 38: 3127-3141. 148859
Cao JJ; Lee SC; Chow JC; Cheng Y; Ho KF; Fung K; Liu SX; Watson JG. (2005). Indoor/outdoor relationships
for PM2.5 and associated carbonaceous pollutants at residential homes in Hong Kong - case study. Indoor
Air, 15: 197-204. 156321
Chakrabarti B; Fine PM; Delfino R; Sioutas C. (2004). Performance evaluation of the active-flow personal
DataRAM PM25 mass monitor (Thermo Anderson pDR-1200) designed for continuous personal exposure
measurements. Atmos Environ, 38: 3329-3340. 147867
Chang LT; Tang CS; Pan YZ; Chan CC. (2007). Association of Heart Rate Variability of the Elderly with
Personal Exposure to PMI, PM1-2.5, and PM2.5-10. 552-556. 156331
Charron A; Harrison RM; Quincey P. (2007). What are the sources and conditions responsible for exceedences of
the 24 h PM10 limit value (50 μg m-3) at a heavily trafficked London site?. Atmos Environ, 41: 1960"
1975. 156333
July 2009
A-337
DRAFT-DO NOT CITE OR QUOTE

-------
Cheng M-D; Hopke PK; Zeng Y. (1993). A receptor-oriented methodology for determining source regions of
particulate sulfate observed at Dorset, Ontario. J Geophys Res, 98: 16,839-16,849. 052294
Chillrud SN; Epstein D; Ross JM; Sax SN; Pederson D; Spengler JD; Kinney PL. (2004). Elevated airborne
exposures of teenagers to manganese, chromium, and iron from steel dust and New York City's subway
system. Environ Sci Technol, 38: 732-737. 054799
Chow JC. (1995). Measurement methods to determine compliance with ambient air quality standards for
suspended particles. J Air Waste Manag Assoc, 45: 320-382. 077012
Chow JC. (2007). The application of thermal methods for determining chemical composition of carbonaceous
aerosols^ A review. , 42: 1521-1541. 157209
Chow JC; Chen L"WA; Lowenthal DH; Doraiswamy P; Park K; Kohl S; Trimble DL; Watson JG. (2005).
California Regional PM10/PM2.5 Air Quality Study (CRPAQS) - Initial data analysis of field program
measurements. 156348
Chow JC; Doraiswamy P; Watson JG; Chen LW; Ho SS; Sodeman DA. (2008). Advances in integrated and
continuous measurements for particle mass and chemical composition. J Air Waste Manag Assoc, 58:
141-163. 156355
Chow JC; Watson JG; Chen LW; Chang MC; Robinson NF; Trimble D; Kohl S. (2007). The IMPROVE_A
temperature protocol for thermal/optical carbon analysis: maintaining consistency with a long-term
database. J Air Waste Manag Assoc, 57: 1014-1023. 156354
Chow JC; Watson JG; Lowenthal DH; Magliano KL. (2005). Loss of PM2.5 nitrate from filter samples in central
California. J Air Waste Manag Assoc, 55: 1158-68. 099030
Chuang K-J; Chan C"C; Chen N"T; Su T-C; Lin L-Y. (2005). Effects of particle size fractions on reducing heart
rate variability in cardiac and hypertensive patients. Environ Health Perspect, 113: 1693-1697. 087989
Chung A; Chang DP; KLeeman MJ; Perry KD; Cahill TA; Dutcher D; McDougall EM; Stroud K. (2001).
Comparison of real-time instruments used to monitor airborne particulate matter. J Air Waste Manag
Assoc, 51: 109-120. 156357
Cifuentes LA; Vega J; Kopfer K; Lave LB. (2000). Effect of the fine fraction of particulate matter versus the
coarse mass and other pollutants on daily mortality in Santiago, Chile. J Air Waste Manag Assoc, 50:
1287-1298. 010351
Cohen BS; Heikkinen MSA; Hazi Y; Gao H; Peters P; Lippmann M. (2004). Field evaluation of nanofilm
detectors for measuring acidic particles in indoor and outdoor air. , 121: 1-35. 056909
Connell DP; Withum JA; Winter SE; Statnick RM; Bilonick RA. (2005). The Steubenville Comprehensive Air
Monitoring Program (SCAMP): associations among fine particulate matter, co-pollutants, and
meteorological conditions. J Air Waste Manag Assoc, 55: 481-496. 089458
Conner TL; Williams RWE. (2004). Identification of possible sources of particulate matter in the personal cloud
using SEM/EDX. Atmos Environ, 38: 5305-5310. 156364
Cornell SE; Jickells TD. (1999). Water-soluble organic nitrogen in atmospheric aerosol: a comparison of UV and
persulfate oxidation methods. Atmos Environ, 33: 833-840. 156367
Cortez-Lugo M; Moreno-Macias H; Holguin-Molina F; Chow JC; Watson JG; Gutierrez-Avedoy V; Mandujano F;
Hernandez-Avila M; Romieu I. (2008). Relationship between indoor, outdoor, and personal fine particle
concentrations for individuals with COPD and predictors of indoor-outdoor ratio in Mexico city. J Expo
Sci Environ Epidemiol, 18: 109-115. 156368
Crimmins BS; Baker JE. (2006). Improved GC/MS methods for measuring hourly PAH and nitro-PAH
concentrations in urban particulate matter. Atmos Environ, 40: 6764-6779. 097008
Crist KC; Liu B; Kim M; Deshpande SR; John K. (2008). Characterization of fine particulate matter in Ohio:
Indoor, outdoor, and personal exposures. Environ Res, 106: 62-71. 156372
Cyrys J; Pitz M; Hazenkamp-von Arx ME; Kunzli N; Heinrich J. (2006). Evaluation of a sampling strategy for
estimation of long-term PM2.5 exposure for epidemiological studies. Environ Monit Assess, 119: 161-171.
156376
Cyrys J; Stolzel M; Heinrich J; Kreyling WG; Menzel N; Wittmaack K; Tuch T; Wichmann H"E. (2003).
Elemental composition and sources of fine and ultrafine ambient particles in Erfurt, Germany. Sci Total
Environ, 305: 143-156. 042232
Dams R; Rahn KA; Robbins JA; Nifong GD; Winchester JW. (1970). Multi-Element Analysis of Air Pollution
Particulates by Nondestructive Neutron Activation. Presented at . 156379
Decesari S; Facchini MC; Fuzzi S; McFiggans GB; Coe H; Bower KN. (2005). The water-soluble organic
component of size-segregated aerosol, cloud water and wet depositions from Jeju Island during ACE"
Asia. Atmos Environ, 39: 211-222. 144536
July 2009
A-338
DRAFT-DO NOT CITE OR QUOTE

-------
Delfino RJ; Quintana PJE; Floro J; Gastanaga VM; Samimi BS; KLeinman MT; Liu L-JS; Bufalino Ci Wu C-F;
McLaren CE. (2004). Association of FEV1 in asthmatic children with personal and microenvironmental
exposure to airborne particulate matter. Environ Health Perspect, 112- 932-941. 056897
Delfino RJ; Staimer N> Gillen D; Tjoa T> Sioutas C> Fung K George SC; KLeinman MT. (2006). Personal and
ambient air pollution is associated with increased exhaled nitric oxide in children with asthma. Environ
Health Perspect, 114: 1736-1743. 090745
Demokritou P; Gupta T> Ferguson Si Koutrakis P. (2002). Development and laboratory performance evaluation
of a personal cascade impactor. J Air Waste Manag Assoc, 52: 1230-1237. 156393
Dermentzoglou Mi Manoli Ei Voutsa Di Samara C. (2003). Sources and patterns of polycyclic aromatic
hydrocarbons and heavy metals in fine indoor particulate matter of Greek houses. , 12: 1511-1519.
156395
Diapouli Ei Chaloulakou Ai Mihalopoulos Ni Spyrellis N. (2008). Indoor and outdoor PM mass and number
concentrations at schools in the Athens area. , 136: 13-20. 190893
Diapouli Ei Chaloulakou Ai Spyrellis N. (2007). Levels of ultrafine particles in different microenvironments--
implications to children exposure. Sci Total Environ, 388: 128-136. 156397
Dills RLAi Paulsen Mi Ahmad Ji Kalman DAAi Elias FNAi Simpson CD. (2006). Evaluation of urinary
methoxyphenols as biomarkers of woodsmoke exposure. Environ Sci Technol, 40: 2163-2170. 156402
Dimitroulopoulou Ci Ashmore MRi Hill MTRi Byrne MAi Kinnersley R. (2006). INDAIR: a probabilistic model of
indoor air pollution in UK homes. Atmos Environ, 40: 6362-6379. 090302
Dong Yi Hays MDi Dean Smith Ni Kinsey JS. (2004). Inverting cascade impactor data for size-resolved
characterization of fine particulate source emissions. J Aerosol Sci, 35: 1497-1512. 156409
Drewnick F> Schwab JJi Hogrefe Oi Peters Si Husain Li Diamond Di Weber Ri Demerjian KL. (2003).
Intercomparison and evaluation of four semi-continuous PM2.5 sulfate instruments. Atmos Environ, 37:
3335-3350. 099160
Eatough DJi Eatough NL; Obeidi F; Pang Y; Modey W', Long R. (2001). Continuous determination of PM25
mass, including semi-volatile species. Aerosol Sci Technol, 34: 1-8. 010303
Ebelt ST; Wilson WE; Brauer M. (2005). Exposure to ambient and nonambient components of particulate
matter: a comparison of health effects. Epidemiology, 16: 396-405. 056907
Ellis EC; Novakov T. (1982). Application of thermal analysis to the characterization of organic aerosol particles.
Sci Total Environ, 23: 227-238. 156416
Emmenegger C; Reinhardt A; Hueglin C; Zenobi R; Kalberer M. (2007). Evaporative light scattering: A novel
detection method for the quantitative analysis of humic-like substances in aerosols. Environ Sci Technol,
41: 2473-2478. 156418
Engling G; Carrico CM; Kreidenweis SM; Collett JL; Day DE; Malm WC; Lincoln E; Min Hao W; Iinuma Y;
Herrmann H. (2006). Determination of levoglucosan in biomass combustion aerosol by high-performance
anion-exchange chromatography with pulsed amperometric detection. Atmos Environ, 40: 299-311.
156422
Fabbri D; Prati S; Vassura I. (2002). Molecular characterisation of organic material in air fine particles (PM10)
using conventional and reactive pyrolysis-gas chromatography-mass spectrometry. J Environ Monit, 4-
210-215. 156426
Falkovich AH; Rudich Y. (2001). Analysis of semivolatile organic compounds in atmospheric aerosols by direct
sample introduction thermal desorption GC/MS. Environ Sci Technol, 35: 2326-2333. 156427
Falkovich AH; Schkolnik G; Ganor E; Rudich Y. (2004). Adsorption of organic compounds pertinent to urban
environments onto mineral dust particles. J Geophys Res, 109: D02208. 156428
Farmer PB; Singh R; Kaur B; Sram RJ; Binkova B; Kalina I; Popov TA; Garte S; Taioli E; Gabelova A;
Cebulska-Wasilewska A. (2003). Molecular epidemiology studies of carcinogenic environmental
pollutants: effects of polycyclic aromatic hydrocarbons (PAHs) in environmental pollution on exogenous
and oxidative DNA damage. , 544: 397-402. 089017
Fehsenfeld FC; Hastie D; Chow JC; Solomon PA. (2004). Particle and gas measurements. : . 156432
Feldpausch P; Fiebig M; Fritzsche L; Petzold A. (2006). Measurement of ultrafine aerosol size distributions by a
combination of diffusion screen separators and condensation particle counters. J Aerosol Sci, 37: 577-597.
155773
Ferro AR; Kopperud RJ; Hildemann LM. (2004). Elevated personal exposure to particulate matter from human
activities in a residence. J Expo Sci Environ Epidemiol, l: S34-S40. 055676
Ferro AR; Kopperud RJ; Hildemann LM. (2004). Source strengths for indoor human activities that resuspend
particulate matter. Environ Sci Technol, 38: 1759-1764. 055387
Filleul L; Baldi I; Dartigues J-F; Tessier J-F. (2003). Risk factors among elderly for short term deaths related to
high levels of air pollution. Occup Environ Med, 60: 684-688. 087403
July 2009
A-339
DRAFT-DO NOT CITE OR QUOTE

-------
Filleul L; Le Tertre A; Baldi I; Tessier J-F. (2004). Difference in the relation between daily mortality and air
pollution among elderly and all-ages populations in southwestern France. Environ Res, 94- 249-253.
087404
Filleul L; Rondeau V; Vandentorren Si Le Moual Ni Cantagrel Ai Annesi-Maesano Ii Charpin Di Declercq Ci
Neukirch Fi Paris Ci Vervloet Di Brochard Pi Tessier JFi Kauffmann Fi Baldi I. (2005). Twenty five year
mortality and air pollution - results from the French PAARC survey. Occup Environ Med, 62- 453-460.
087357
Fine PMi Jaques PAi Hering SVi Sioutas C. (2003). Performance Evaluation and Use of a Continuous Monitor
for Measuring Size-Fractionated PM 2.5 Nitrate. Aerosol Sci Technol, 37- 342 - 354. 155775
Finkelstein MMi Jerrett Mi DeLuca Pi Finkelstein Ni Verma DKi Chapman Ki Sears MR. (2003). Relation
between income, air pollution and mortality- a cohort study. CMAJ, 169- 397-402. 056117
Fitz Di Chan Mi Cass Gi Lawson Di Ashbaugh L. (1989). A multi-component size-classifying aerosol and gas
sampler for ambient air monitoring. Presented at Presented at- 82nd annual meeting of the Air & Waste
Management Association; June! Anaheim, CA. Pittsburgh, PA- Air & Waste Management Association!
paper no. 89-140.1. 077387
Fraser MP; Yue ZW; Buzcu B. (2003). Source apportionment of fine particulate matter in Houston, TX, using
organic molecular markers. Atmos Environ, 37: 2117-2123. 042231
Fraser MP; Yue ZW; Tropp RJ; Kohl SD; Chow JC. (2002). Molecular composition of organic fine particulate
matter in Houston, TX. Atmos Environ, 36: 5751-5758. 140741
Fromme H; Diemer J; Dietrich S; Cyrys J; Heinrich J; Lang W', Kiranoglu M; Twardella D. (2008). Chemical and
morphological properties of particulate matter (PM10, PM25) in school classrooms and outdoor air.
Atmos Environ, 42(27): 6597-6605. 155147
Fromme H; Heitmann D; Schierl R; Liebl B; RDden H; Twardella D; Dietrich S. (2007). Particulate matter in
the indoor air of classrooms-exploratory results from Munich and surrounding area. Atmos Environ, 41:
854-866. 156453
Fruin S; Westerdahl D; Sax T; Sioutas C; Fine PM. (2008). Measurements and predictors of on-road ultrafine
particle concentrations and associated pollutants in Los Angeles. Atmos Environ, 42: 207-219. 097183
Gadkari NM; Pervez S. (2007). Source investigation of personal particulates in relation to identify major routes
of exposure among urban residentials. Atmos Environ, 41: 7951-7963. 156459
Gauvin S; Reungoat P; Cassadou S; Dechenaux J; Momas I; Just J; Zmirou D. (2002). Contribution of indoor and
outdoor environments to PM2.5 personal exposure of children-'VESTA study. Sci Total Environ, 297:
175-181. 034893
Geyh A; Chillrud S; Wiliams D; Herbstman J; Symons J; Rees K; Ross J; Kim S; Lim H; Turpin B; Breysse P.
(2005). Assessing truck driver exposure at the World Trade Center disaster site: personal and area
monitoring for particulate matter and volatile organic compounds during October 2001 and April 2002. J
Occup Environ Hyg, 2' 179-193. 186949
Geyh AS; Hering S; Kreisberg N> John W. (2004). Evaluation of a personal and microenvironmental aerosol
speciation sampler (PMASS). 1"22. 156467
Gomez-Perales JE; Colvile RN; Fernandez-Bremauntz AA; Gutierrez "Ave doy V; Paramo-Figueroa VH; Bianco-
Jimenez S; Bueno-Lopez E; Bernabe-Cabanillas R; Mandujano F; Hidalgo "Navarro M; Nieuwenhuijsen
MJ. (2007). Bus, minibus, metro inter-comparison of commuters' exposure to air pollution in Mexico City.
Atmos Environ, 41: 890-901. 138816
Gomez-Perales JE; Colvile RN; Nieuwenhuijsen MJ; Fernandez-Bremauntz A; Gutierrez-Avedoy VJ; Paramo-
Figueroa VH; Bianco-Jimenez S; Bueno-Lopez E; Mandujano F; Bernabe-Cabanillas R; Ortiz-Segovia E.
(2004). Commuters' exposure to PM25, CO, and benzene in public transport in the metropolitan area of
Mexico City. Atmos Environ, 38: 1219-1229. 054418
Gouveia N; Fletcher T. (2000). Time series analysis of air pollution and mortality: effects by cause, age and
socioeconomic status. J Epidemiol Community Health, 54: 750-755. 012132
Goyal P; Sidhartha. (2004). Modeling and monitoring of suspended particulate matter from Badarpur thermal
power station, Delhi. , 19: 383-390. 156487
Graham B; Falkovich AH; Rudich Y; Maenhaut W; Guyon P; Andreae MO. (2004). Local and regional
contributions to the atmospheric aerosol over Tel Aviv, Israel: a case study using elemental, ionic and
organic tracers. Atmos Environ, 38: 1593-1604. 156490
Graney JR; Landis MS; Norris GA. (2004). Concentrations and solubility of metals from indoor and personal
exposure PM25 samples. Atmos Environ, 38: 237-247. 053756
Greaves RC; Barkley RM; Sievers RE. (1985). Rapid sampling and analysis of volatile constituents of airborne
particulate matter. Anal Chem, 57: 2807-2815. 156494
July 2009
A-340
DRAFT-DO NOT CITE OR QUOTE

-------
Greaves RC; Barkley RM! Sievers RE; Meglen RR. (1987). Covariations in the concentrations of organic
compounds associated with springtime atmospheric aerosols. Atmos Environ, 21- 2549-2561. 156495
Grosjean D; Seinfeld JH. (1989). Parameterization of the formation potential of secondary organic aerosols.
Atmos Environ, 23: 1733-1747. 045643
Grover BD; KLeinman M; Eatough NL; Eatough DJ; Hopke PK; Long RW; Wilson WE; Meyer MB; Ambs JL.
(2005). Measurement of total PM25 mass (nonvolatile plus semivolatile) with the Filter Dynamic
Measurement System tapered element oscillating microbalance monitor. J Geophys Res, 110D07S03.
090044
Gulliver J; Briggs DJ. (2004). Personal exposure to particulate air pollution in transport microenvironments.
Atmos Environ, 38: 1-8. 053238
Gulliver J; Briggs DJ. (2007). Journey-time exposure to particulate air pollution. Atmos Environ, 41: 7195-7207.
155814
Guo H; Lee SC; Chan LY. (2004). Indoor air quality investigation at air-conditioned and non-air-conditioned
markets in Hong Kong. Sci Total Environ, 323: 87"98. 156506
Gwynn RC; Thurston GD. (2001). The burden of air pollution: impacts among racial minorities. Environ Health
Perspect, 1094: 501-506. 017206
Hall PA; Watson AFR; Garner GV; Hall K; Smith S; Waterman D; Horsfield B. (1999). An investigation of
micro-scale sealed vessel thermal extraction-gas chromatography-mass spectrometry (MSSV"GC"MS#)
and micro-scale sealed vessel pyrolysis-gas chromatography-mass spectrometry applied to a standard
reference material of an urban dust. Sci Total Environ, 235: 269-276. 156512
Hamilton JF; Webb PJ; Lewis AC; Hopkins JR; Smith S; Davy P. (2004). Partially oxidised organic components
in urban aerosol using GCXGC-TOF/MS. , 4- 1279-1290. 156516
Hamilton JF; Webb PJ; Lewis AC; Reviejo MM. (2005). Quantifying small molecules in secondary organic
aerosol formed during the photo-oxidation of toluene with hydroxyl radicals. Atmos Environ, 39: 7263"
7275. 088173
Hanninen OO; Lebret E; Ilacqua V; Katsouyanni K; Kunzli N; Sram RJ; Jantunen M. (2004). Infiltration of
ambient PM2.5 and levels of indoor generated non-ETS PM2.5 in residences of four European cities.
Atmos Environ, 38: 6411-6423. 056812
Haverinen-Shaughnessy U; Toivola M; Aim S; Putus T; Nevalainen A. (2007). Personal and microenvironmental
concentrations of particles and microbial aerosol in relation to health symptoms among teachers. J Expo
Sci Environ Epidemiol, 17: 182-190. 156526
Hays MD; Smith ND; Dong Y. (2004). Nature of unresolved complex mixture in size "distributed emissions from
residential wood combustion as measured by thermal desorption-gas chromatography-mass
spectrometry. J Geophys Res, 109: D16S04. 156530
Hays MD; Smith ND; Kinsey J; Dong Y; Kariher P. (2003). Polycyclic aromatic hydrocarbon size distributions in
aerosols from appliances of residential wood combustion as determined by direct thermal desorption—
GC/MS. J Aerosol Sci, 34: 1061-1084. 156529
Hazenkamp-von Arx ME; Fellmann TG; Oglesby L; Ackermann-Liebrich U; Gislason T; Heinrich J; Jarvis D;
Luczynska C; Manzanera AJ; Modig L; Norback D; Pfeifer A; Poli A; Ponzio M; Soon A; Vermeire P;
Kunzli N. (2003). PM2.5 assessment in 21 European study centers of ECRHS II: Method and first winter
results. J Air Waste Manag Assoc, 53: 617-628. 136487
Helmig D; Bauer A; Mueller J; Klein W. (1990). Analysis of particulate organics in a forest atmosphere by
thermodesorption GC/MS. Atmos Environ, 24: 179-184. 156536
Henderson SB; Beckerman B; Jerrett M; Brauer M. (2007). Application of land use regression to estimate long-
term concentrations of traffic-related nitrogen oxides and fine particulate matter. Environ Sci Technol,
41: 2422-2428. 090675
Henning S; Weingartner E; Schwikowski M; Gaggeler HW; Gehrig R; Hinz KP; Trimborn A; Spengler B;
Baltensperger U. (2003). Seasonal variation of water-soluble ions of the aerosol at the high-alpine site
Jungfraujoch (3580 m asl). J Geophys Res, 108: 4030. 156539
Henry RC. (1997). History and fundamentals of multivariate air quality receptor models. , 37: 37"42. 020941
Henry RC. (2003). Multivariate receptor modeling by N-dimensional edge detection. , 65: 179-189. 156540
Henry RC; Chang YS; Spiegelman CH. (2002). Locating nearby sources of air pollution by nonparametric
regression of atmospheric concentrations on wind direction. Atmos Environ, 36: 2237-2244. 136097
Hering S; Cass G. (1999). The magnitude of bias in the measurement of PM2.5 arising from volatilization of
particulate nitrate from Teflon filters. J Air Waste Manag Assoc, 49: 725-733. 084958
Hering S; Fine PM; Sioutas C; Jaques PA; Ambs JL; Hogrefe O; Demerjian KL. (2004). Field assessment of the
dynamics of particulate nitrate vaporization using differential TEOM® and automated nitrate monitors.
Atmos Environ, 38: 5183-5192. 155837
July 2009
A-341
DRAFT-DO NOT CITE OR QUOTE

-------
Hering SV; Lawson DR; Allegrini I; Febo A; Perrino C; Possanzini M; Sickles JE II; Anlauf KG; Wiebe A; Appel
BR; John W; Ondo J; Wall S; Braman RS; Sutton R; Cass GR; Solomon PA; Eatough DJ; Eatough NL;
Ellis EC; Grosjean D; Hicks BB; Womack JD; Horrocks J; Knapp KT; Ellestad TG; Paur RJ; Mitchell WJ;
Pleasant M; Peake E; MacLean A; Pierson WR; Brachaczek W', Schiff HI; Mackay GI; Spicer CW;. (1988).
The nitric acid shootout- field comparison of measurement methods. Atmos Environ, 22: 1519-1539.
036012
Hering SV; Stolzenburg MR; Quant FR; Oberreit DR; Keady PB. (2005). A Laminar-Flow, Water-Based
Condensation Particle Counter (WCPC). Aerosol Sci Technol, 39- 659-672. 155838
Hermann M; Wehner B; Bischof O; Han HS; Krinke T; Liu W', Zerrath A; Wiedensohler A. (2007). Particle
counting efficiencies of new TSI condensation particle counters. J Aerosol Sci, 38- 674-682. 155840
Hertel O; Storm L; Stausgaard L; Hvidberg M; Ketzel M. (2008). A proper choice of route significantly reduces
air pollution exposure - A study on bicycle and bus trips in urban streets. Sci Total Environ, 389- 58-70.
156543
Hidy GM; Friedlander SK. (1972). The nature of the Los Angeles aerosol. Presented at . 156546
Ho KF; Cao JJ; Harrison RM; Lee SC. (2004). Indoor/outdoor relationships of organic carbon (OC) and elemental
carbon (EC) in PM2.5 in roadside environment of Hong Kong. Atmos Environ, 38- 6327-6335. 056804
Ho KF; Chow JC; Watson JG; Fung K; Lee SC; Cao JJ; Li YS. (2006). Variability of organic and elemental
carbon, water soluble organic carbon, and isotopes in Hong Kong. , 6- 4579-4600. 156552
Ho SSH; Yu JZ. (2004). In-injection port thermal desorption and subsequent gas chromatography—mass
spectrometric analysis of polycyclic aromatic hydrocarbons and n-alkanes in atmospheric aerosol
samples. J Chromatogr, 1059- 121-129. 156551
Hoek G; Brunekreef B; Goldbohm S; Fischer P; Van den Brandt PA. (2002). Association between mortality and
indicators of traffic-related air pollution in the Netherlands' a cohort study. Lancet, 360- 1203-1209.
042364
Hoek G; de Hartog J; Meliefste K; ten Brink H; Katsouyanni K; Karakatsani A; Lianou M; Kotronarou A;
Kavouras I; Pekkanen J; Vallius M; Kulmala M; Puustinen A; Thomas S; Meddings C; Ayres J; van
Wijnen J; Hameri K; Kos G; Harrison R. (2008). Indoor-outdoor relationships of particle number and
mass in four European cities. Atmos Environ, 42- 156-169. 156554
Hogrefe O; Schwab JJ; Drewnick F; Lala GG; Peters S; Demerjian KL; Rhoads K; Felton HD; Rattigan OV;
Husain L; Dutkiewicz VA. (2004). Semicontinuous PM2.5 sulfate and nitrate measurements at an urban
and a rural location in New York- PMTACS-NY summer 2001 and 2002 campaigns. J Air Waste Manag
Assoc, 54: 1040-60. 099003
Holguin F; Tellez-Rojo MM; Hernandez M; Cortez M; Chow JC; Watson JG; Mannino D; Romieu I. (2003). Air
pollution and heart rate variability among the elderly in Mexico City. Epidemiology, 14: 521-527. 057326
Hopke PK; Li CL; Ciszek W; Landsberger S. (1995). The use of bootstrapping to estimate conditional probability
fields for source locations of airborne pollutants. , 30: 69"79. 156566
Hopke PK; Ramadan Z; Paatero P; Norris GA; Landis MS; Williams RW; Lewis CW. (2003). Receptor modeling
of ambient and personal exposure samples: 1998 Baltimore Particulate Matter Epidemiology-Exposure
Study. Atmos Environ, 37: 3289-3302. 095544
Hystad P; Setton E; Allen R; Keller P; Brauer M. (2008). Modeling residential fine particulate matter
infiltration for exposure assessment. J Expo Sci Environ Epidemiol, In Press: In Press. 190890
Ito E Thurston GD. (1996). Daily PMlO/mortality associations: an investigation of at-risk subpopulations. J
Expo Sci Environ Epidemiol, 6: 79-95. 078841
Ito K; Thurston GD; Silverman RA. (2007). Characterization of PM2.5, gaseous pollutants, and meteorological
interactions in the context of time-series health effects models. J Expo Sci Environ Epidemiol, 17 Suppl
2: S45-60. 156594
Jacquemin B; Cabrera L; Querol X; Bellander T; Moreno N; Peters A; Pey J; Pekkanen J; Lanki T; Sunyer J.
(2007). Levels of outdoor PM2.5, absorbance and sulphur as surrogates for personal exposures among
post-myocardial infarction patients in Barcelona, Spain. Atmos Environ, 41: 1539-1549. 156600
Jacquemin B; Sunyer J; Forsberg B; Gotschi T; Bayer-Oglesby L; Ackermann-Liebrich U; de Marco R; Heinrich
J; Jarvis D; Toren K; Kiinzli N. (2007). Annoyance due to air pollution in Europe. Int J Epidemiol, 10: 1-
12.192372
Jansen KL; Larson TV; Koenig JQ; Mar TF; Fields C; Stewart J; Lippmann M. (2005). Associations between
health effects and particulate matter and black carbon in subjects with respiratory disease. Environ
Health Perspect, 113: 1741-1746. 082236
Janssen NA; Lanki T; Hoek G; Vallius M; De Hartog JJ; Van Grieken R; PekkanenJ; Brunekreef B. (2005).
Associations between ambient, personal, and indoor exposure to fine particulate matter constituents in
Dutch and Finnish panels of cardiovascular patients. Occup Environ Med, 62: 868-877. 088692
July 2009
A-342
DRAFT-DO NOT CITE OR QUOTE

-------
Jaques PA; Ambs JL; Grant WL; Sioutas C. (2004). Field evaluation of the differential TEOM monitor for
continuous PM 2.5 mass concentrations. Aerosol Sci Technol, 38- 49"59. 155878
Jedrychowski WA; Perera FP; Pac A; Jacek R; Whyatt RM; Spengler JD; Dumyahn TS; Sochacka-Tatara E.
(2006). Variability of total exposure to PM2.5 related to indoor and outdoor pollution sources Krakow
study in pregnant women. Sci Total Environ, 366- 47"54. 156606
Jeon SJ; Meuzelaar HLC; Sheya SAN; Lighty JS; Jarman WM! Kasteler C> Sarofim AF; Simoneit BRT. (2001).
Exploratory studies of PM10 receptor and source profiling by GC/MS and principal component analysis of
temporally and spatially resolved ambient samples. J Air Waste Manag Assoc, 51- 766-784. 016636
Jo WE Lee JY. (2006). Indoor and outdoor levels of respirable particulates (PM10) and Carbon Monoxide (CO)
in high-rise apartment buildings. Atmos Environ, 40- 6067-6076. 156613
Johannesson Si Gustafson Pi Molnar Pi Barregard Li Sallsten G. (2007). Exposure to fine particles (PM2.5 and
PMl) and black smoke in the general population- personal, indoor, and outdoor levels. J Expo Sci
Environ Epidemiol, 17- 613-624. 156614
John Wi Wall SMi Ondo JL. (1988). A new method for nitric acid and nitrate aerosol measurement using the
dichotomous sampler. Atmos Environ, 22- 1627-1635. 045903
Jones AM; Harrison RM. (2006). Assessment of natural components of PM10 at UK urban and rural sites.
Atmos Environ, 40: 7733-7741. 155886
Jones J; Stick S; Dingle P; Franklin P. (2007). Spatial variability of particulates in homes: implications for
infant exposure. Sci Total Environ, 376: 317-323. 156615
Kaur S; Nieuwenhuijsen MJ Colvile R. (2005). Personal exposure of street canyon intersection users to PM25,
ultrafine particle counts and carbon monoxide in central London, UK. Atmos Environ, 39: 3629-3641.
086504
Kaur S; Nieuwenhuijsen MJ; Colvile RN. (2005). Pedestrian exposure to air pollution along a major road in
Central London, UK. Atmos Environ, 39: 7307-7320. 088175
Keeler GJ; Samson PJ. (1989). Spatial representativeness of trace element ratios. Environ Sci Technol, 23:
1358-1364. 156633
Khlystov A; Stanier CO; Takahama S; Pandis SN. (2005). Water content of ambient aerosol during the
Pittsburgh Air Quality Study. J Geophys Res, 110: D07S10. 156635
Kidwell CB; Ondov JM. (2001). Development and evaluation of a prototype system for collecting sub-hourly
ambient aerosol for chemical analysis. Aerosol Sci Technol, 35: 596-601. 017092
Kidwell CB; Ondov JM. (2004). Elemental Analysis of Sub-Hourly Ambient Aerosol Collections. Aerosol Sci
Technol, 38: 205-218. 155898
Kim D; Sass-Kortsak A; Purdham JT; Dales RE; Brook JR. (2005). Sources of personal exposure to fine particles
in Toronto, Ontario, Canada. J Air Waste Manag Assoc, 55: 1134-1146. 156640
Kinsey JS; Mitchell WA; Squier WC; Linna K; King FG; Logan R; Dong Y; Thompson GJ; Clark NN. (2006).
Evaluation of methods for the determination of diesel-generated fine particulate matter: Physical
characterization results. J Aerosol Sci, 37: 63"87. 130654
Kiss G; Varga B; Galambos I; Ganszky I. (2002). Characterization of water-soluble organic matter isolated from
atmospheric fine aerosol. J Geophys Res, 107: 8339. 156646
Klinmalee A; Srimongkol K; Kim Oanh NT. (2008). Indoor air pollution levels in public buildings in Thailand
and exposure assessment. Environ Monit Assess, In Press: In Press. 190888
Koenig JQ; Jansen K; Mar TF; Lumley T; Kaufman J; Trenga CA; Sullivan J; Liu LJ; Shapiro GG; Larson TV.
(2003). Measurement of offline exhaled nitric oxide in a study of community exposure to air pollution.
Environ Health Perspect, 111: 1625-1629. 156653
Koistinen KJ; Edwards RD; Mathys P; Ruuskanen J; Kunzli N; Jantunen MJ. (2004). Sources of fine particulate
matter in personal exposures and residential indoor, residential outdoor and workplace
microenvironments in the Helsinki phase of the EXPOLIS study. Scand J Work Environ Health, 30
Suppl 2: 36-46. 156655
Koistinen KJ; Kousa A; Tenhola V; Hanninen O; Jantunen MJ; Oglesby L; Kuenzli N; Georgoulis L. (1999). Fine
particle (PM25) measurement methodology, quality assurance procedures and pilot results of the
EXPOLIS study. J Air Waste Manag Assoc, 49: 1212-1220. 010628
Kousa A; Monn C; Rotko T; Aim S; Oblesby L; Jantunen MJ. (2001). Personal exposures to N02 in the
EXPOLIS-study: relation to residential indoor, outdoor and workplace concentrations in Basel, Helsinki
and Prague. Atmos Environ, 35: 3405-3412. 025270
Koutrakis P; Suh HH; Sarnat JA; Brown KW; Coull BA; Schwartz J. (2005). Characterization of particulate and
gas exposures of sensitive subpopulations living in Baltimore and Boston. , 131: 1-65. 095800
July 2009
A-343
DRAFT-DO NOT CITE OR QUOTE

-------
Krewski D; Burnett RT; Goldberg MS; Hoover K; Siemiatycki J; Jerrett MJ Abrahamowicz MJ White WH.
(2000). Reanalysis of the Harvard Six Cities study and the American Cancer Society study of particulate
air pollution and mortality- a special report of the Institute's Particle Epidemiology Reanalysis Project.
Health Effects Institute. Cambridge, MA. 012281
Kulkarni MM; Patil RS. (2003). Personal exposure to toxic metals in an Indian metropolitan region. , 84- 23"29.
156664
Kulmala M; Mordas G; Petaja T; Gronholm T; Aalto PP; Vehkamaki H; Hienola AI; Herrmann E; Sipila M;
Riipinen I. (2007). The condensation particle counter battery (CPCB)' A new tool to investigate the
activation properties of nanoparticles. J Aerosol Sci, 38- 289-304. 155911
Kulmala M; Riipinen I; Sipila M; Manninen HE; Petaja T; Junninen H; Maso MD; Mordas G; Mirme A; Vana M;
Hirsikko A; Laakso L; Harrison RM; Hanson I; Leung C; Lehtinen KE; Kerminen VM. (2007). Toward
direct measurement of atmospheric nucleation. , 318- 89"92. 097838
Kumar R; Elizabeth A; Gawane AG. (2006). Air quality profile of inorganic ionic composition of fine aerosols at
two sites in Mumbai City. Aerosol Sci Technol, 40- 477-489. 129347
Labban R; Veranth JM; Watson JG; Chow JC. (2006). Feasibility of soil dust source apportionment by the
pyrolysis-gas chromatography/mass spectrometry method. J Air Waste Manag Assoc, 56- 1230-1242.
156665
Lai CH; Liou SH; Shih TS; Tsai PJ; Chen HL; Chang YC; Buckley TJ; Strickland P; Jaakkola JJ. (2004).
Exposure to fine particulate matter (PM2.5) among highway toll station workers in taipei- direct and
indirect exposure assessment. , 59- 138-148. 156666
Lai HK; Bayer-Oglesby L; Colvile R; Gotschi T; Jantunen MJ; Kunzli N; Kulinskaya E; Schweizer C;
Nieuwenhuijsen MJ. (2006). Determinants of indoor air concentrations of PM25, black smoke and N02 in
six European cities (EXPOLIS study). Atmos Environ, 40: 1299-1313. 090262
Lai HK; Kendall M; Ferrier H; Lindup I; Aim S; Hanninen 0; Jantunen M; Mathys P; Colvile R; Ashmore MR;
Cullinan P; Nieuwenhuijsen MJ. (2004). Personal exposures and microenvironment concentrations of
PM25, VOC, N02 and CO in Oxford, UK. Atmos Environ, 38: 6399-6410. 056811
Lake DA; Tolocka MP; Johnston MV; Wexler AS. (2003). Mass Spectrometry of Individual Particles between 50
and 750 nm in Diameter at the Baltimore Supersite. Environ Sci Technol, 37: 3268-3274. 156669
Lake DA; Tolocka MP; Johnston MV; Wexler AS. (2004). The character of single particle sulfate in Baltimore.
Atmos Environ, 38: 5311-5320. 088411
Larson T; Gould T; Simpson C; Liu LJ; Claiborn C; Lewtas J. (2004). Source apportionment of indoor, outdoor,
and personal PM2.5 in Seattle, Washington, using positive matrix factorization. J Air Waste Manag
Assoc, 54: 1175-87. 098145
Laurent O; Pedrono G; Segala C; Filleul L; Havard S; Deguen S; Schillinger C; Riviere E; Bard D. (2008). Air
pollution, asthma attacks, and socioeconomic deprivation: a small-area case-crossover study. Am J
Epidemiol, 168: 58-65. 156672
Lee JH; Hopke PK; Holsen TM; Polissar AV. (2005). Evaluation of Continuous and Filter-Based Methods for
Measuring PM 2.5 Mass Concentration. Aerosol Sci Technol, 39: 290-303. 156680
Lee JH; Hopke PK; Holsen TM; Polissar AV; Lee DW; Edgerton ES; Ondov JM; Allen G. (2005). Measurements
of fine particle mass concentrations using continuous and integrated monitors in Eastern US Cities.
Aerosol Sci Technol, 39: 261-275. 128139
Lee JT; Son JY; Kim H; Kim SY. (2006). Effect of air pollution on asthma-related hospital admissions for
children by socioeconomic status associated with area of residence. Arch Environ Occup Health, 61: 123-
30. 098248
Lee SA; Adhikari A; Grinshpun SA; McKay R; Shukla R; Reponen T. (2006). Personal exposure to airborne dust
and microorganisms in agricultural environments. , 3: 118-30. 188450
Lee SJ; Demokritou P; Koutrakis P; Delgado-Saborit JM. (2006). Development and evaluation of personal
respirable particulate sampler (PRPS). Atmos Environ, 40: 212-224. 098249
Lewis CW; Norris GA; Conner TL; Henry RC. (2003). Source apportionment of Phoenix PM2.5 aerosol with the
Unmix Receptor model. J Air Waste Manag Assoc, 53: 325-338. 088413
Lewne M; Nise G; Lind ML; Gustavsson P. (2006). Exposure to particles and nitrogen dioxide among taxi, bus
and lorry drivers. , 79: 220-226. 189293
Lewne M; Plato N> Gustavsson P. (2007). Exposure to particles, elemental carbon and nitrogen dioxide in
workers exposed to motor exhaust. Ann Occup Hyg, 51: 693-701. 156690
Li W-W; Paschold H; Morales H; Chianelli J. (2003). Correlations between short-term indoor and outdoor PM
concentrations at residences with evaporative coolers. Atmos Environ, 37: 2691-2703. 047845
July 2009
A-344
DRAFT-DO NOT CITE OR QUOTE

-------
Li YC; Yu JZ. (2005). Simultaneous Determination of Mono-and Dicarboxylic Acids, o-Oxo-carboxylic Acids,
Midchain Ketocarboxylic Acids, and Aldehydes in Atmospheric Aerosol Samples. Environ Sci Technol, 39-
7616-7624. 156692
Liao CM; Chen SC; Chen JW; Liang HM. (2006). Contributions of Chinese-style cooking and incense burning to
personal exposure and residential PM concentrations in Taiwan region. , 358- 72"84. 188451
Lim HJ; Allen G> Maring H; Solomon P; Turpin BJ; Edgerton Ei Hering SV. (2003). Semicontinuous aerosol
carbon measurements' Comparison of Atlanta supersite measurements. J Geophys Res, 108- SOS-SOS.
156697
Linn WS; Gong H Jr. (1999). The 21st century environment and air quality influences on asthma. Curr Opin
Pulm Med, 5: 21-26. 011680
Lithgow GA; Robinson AL; Buckley SG. (2004). Ambient measurements of metal-containing PM25 in an urban
environment using laser-induced breakdown spectroscopy. Atmos Environ, 38- 3319-3328. 126616
Liu L-J; Box Mi Kalman Di Kaufman Ji Koenig Ji Larson Ti Lumley Ti Sheppard Li Wallace L. (2003). Exposure
assessment of particulate matter for susceptible populations in Seattle. Environ Health Perspect, 11-
909-918. 073841
Liu Yi Woodin MAi Hauser Ri Williams PLi Herrick RFi Christiani DCi Smith TJ. (2005). Estimation of
personal exposures to particulate matter and metals in boiler overhaul work. J Occup Environ Med, 47-
68-78. 156704
Lonati Gi Giugliano Mi Butelli Pi Romele Li Tardivo R. (2005). Major chemical components of PM25 in Milan
(Italy). Atmos Environ, 39- 1925-1934. 126171
Lunden MMi Kirchstetter TWi Thatcher TLi Hering SVi Brown NJ. (2008). Factors affecting the indoor
concentrations of carbonaceous aerosols of outdoor origin. Atmos Environ, 42- 5660-5671. 155949
Lung SCi Mao IF; Liu LJ. (2007). Residents' particle exposures in six different communities in Taiwan. Sci Total
Environ, 377: 81-92. 156719
Macintosh D; Minegishi T; Kaufman M; Baker B; Allen J; Levy J; Myatt T. (2009). The benefits of whole-house
in-duct air cleaning in reducing exposures to fine particulate matter of outdoor origin: A modeling
analysis. J Expo Sci Environ Epidemiol, x: x. 190887
Mader BT; Yu JZ; Xu JH; Li QF; Wu WS; Flagan RC; Seinfeld JH. (2004). Molecular composition of the water-
soluble fraction of atmospheric carbonaceous aerosols collected during ACE"Asia. J Geophys Res, 109:
D06206.156724
Magari SR; Schwartz J; Williams PL; Hauser R; Smith TJ; Christiani DC. (2002). The association of particulate
air metal concentrations with heart rate variability. Environ Health Perspect, 110: 875-880. 034813
Maitre A; Soulat JM; Masclet P; Stoklov M; Marques M; de Gaudemaris R. (2002). Exposure to carcinogenic air
pollutants among policemen working close to traffic in an urban area. Scand J Work Environ Health, 28:
402-410. 156726
Malm WC; Collett Jr; McMeeking G; Lee T; Carrillo J; Schichtel B; Day DE; Carrico C; Kreidenweis SM. (2005).
Intercomparison and closure calculations using measurements of aerosol species and optical properties
during the Yosemite aerosol characterization study. J Geophys Res, 110: 1-21. 156729
Manchester-Neesvig JB; Schauer JJ; Cass GR. (2003). The distribution of particle-phase organic compounds in
the atmosphere and their use for source apportionment during the Southern California Children's Health
Study. J Air Waste Manag Assoc, 53: 1065-79. 098102
Mar TF; Koenig JQ; Jansen K; Sullivan J; Kaufman J; Trenga CA; Siahpush SH; Liu L-JS; Neas L. (2005). Fine
particulate air pollution and cardiorespiratory effects in the elderly. Epidemiology, 16: 681-687. 087566
Martins MCH; Fatigati FL; Vespoli TC; Martins LC; Martins MA; Saldiva PHN; Braga ALF. (2004). Influence of
socioeconomic conditions on air pollution effects in elderly people an analysis of six regions in Sao Paolo,
Brazil. J Epidemiol Community Health, 58: 41-46. 087457
Martuzevicius D; Grinshpun S; Lee T; Hu S; Biswas P; Reponen T; LeMasters G. (2008). Traffic-related PM2. 5
aerosol in residential houses located near major highways: Indoor versus outdoor concentrations. Atmos
Environ, 27: 6575-6585. 190886
Mathai CV; Watson Jr JG; Rogers CF; Chow JC; Tombach I; Zwicker JO; Cahill T; Feeney P; Eldred R. (1990).
Intercomparison of ambient aerosol samplers used in western visibility and air quality studies. Environ
Sci Technol, 24: 1090-1099. 156741
Mayol-Bracero OL; Guyon P; Graham B; Roberts G; Andreae MO; Decesari S; Facchini MC; Fuzzi S; Artaxo P.
(2002). Water-soluble organic compounds in biomass burning aerosols over Amazonia: 2 Apportionment
of the chemical composition and importance of the polyacidic fraction. J Geophys Res, 107. 045010
Mazzoleni LR; Zielinska B; Moosmuller H. (2007). Emissions of Levoglucosan, Methoxy Phenols, and Organic
Acids from Prescribed Burns, Laboratory Combustion of Wildland Fuels, and Residential Wood
Combustion. Environ Sci Technol, 41: 2115-2122. 098038
July 2009
A-345
DRAFT-DO NOT CITE OR QUOTE

-------
McCormack MC; Breysse PN; Hansel NN; Matsui EC; Tonorezos ES; CurtirrBrosnan J; Williams DL; Buckley
TJ; Eggleston PA; Diette GB. (2007). Common household activities are associated with elevated
particulate matter concentrations in bedrooms of inner-city Baltimore pre-school children. Environ Res,
106: 148-155. 156745
Meng QY; Turpin BJ; Korn L; Weisel CP; Morandi M; Colome S; Zhang J; Stock T; Spektor D; Winer A; Zhang
L; Lee JH; Giovanetti R; Cui W; Kwon J; Alimokhtari S; Shendell D; Jones J; Farrar C; Maberti S.
(2005). Influence of ambient (outdoor) sources on residential indoor and personal PM2.5 concentrations:
analyses of RIOPA data. J Expo Sci Environ Epidemiol, 15: 17-28. 058595
Meng QY; Turpin BJ; Lee JH; Polidori A; Weisel CP; Morandi M; Colome S; Zhang JF; Stock T; Winer A. (2007).
How does infiltration behavior modify the composition of ambient PM2.5 in indoor spaces? An analysis of
RIOPA data. Environ Sci Technol, 41: 7315-7321. 091197
Meng QY; Turpin BJ; Polidori A; Lee JH; Weisel C> Morandi M; Colome S; Stock T; Winer A; Zhang J. (2005).
PM2.5 of ambient origin: estimates and exposure errors relevant to PM epidemiology. Environ Sci
Technol, 39: 5105-5112. 081194
Middlebrook AM; Murphy DM; Lee S"H; Thomson DS; Prather KA; Wenzel RJ; Liu D"Y; Phares DJ; Rhoads KP;
Wexler AS; Johnston MV; Jimenez JL; Jayne JT; Worsnop DR; Yourshaw I; Seinfeld JH; Flagan RC.
(2003). A comparison of particle mass spectrometers during the 1999 Atlanta Supersites project. J
Geophys Res, 108D7. 042932
Miguel AH; Eiguren-Fernandez A; Jaques PA; Froines JR; Grant BL; Mayo PR; Sioutas C. (2004). Seasonal
variation of the particle size distribution of polycyclic aromatic hydrocarbons and of major aerosol species
in Claremont, California. Atmos Environ, 38: 3241-3251. 123260
Mihaltan F; Munteanu I; Higbee C; Travers M; Hyland A; Cummings KM; Dresler C. (2006). Global air
monitoring study: a multi-country comparison of levels of indoor air pollution in different workplaces.
Results from Romania, May 2006. Pneumologie, 55: 156-160. 156761
Mikel DK. (2001). Quality assurance final report for the Southern Oxidant Study Atlanta Supersite Field
Experiment August 3 - September 1, 1999. 156762
Miller AL; Habjan MC; Park K. (2007). Real-time estimation of elemental carbon emitted from a diesel engine.
Environ Sci Technol, 41: 5783-5788. 156765
Miller JD; Dugandzic R; Frescura AM; Salares V. (2007). Indoor- and outdoor-derived contaminants in urban
and rural homes in Ottawa, Ontario, Canada. J Air Waste Manag Assoc, 57' 297-302. 156764
Molnar P; Bellander T; Sallsten G; Boman J. (2007). Indoor and outdoor concentrations of PM2.5 trace elements
at homes, preschools and schools in Stockholm, Sweden. J Environ Monit, 9' 348-357. 156774
Molnar P; Gustafson P; Johannesson S; Boman J; Barregard L; Sallsten G. (2005). Domestic wood burning and
PM sub(2.5) trace elements: Personal exposures, indoor and outdoor levels. Atmos Environ, 39: 2643"
2653. 156772
Molnar P; Johannesson S; Boman J; Barregard L; Sallsten G. (2006). Personal exposures and indoor, residential
outdoor, and urban background levels of fine particle trace elements in the general population. J Environ
Monit, 8: 543-551. 156773
Monkkonen P; Pai P; Maynard A; Lehtinen KE; Hameri K; Rechkemmer P; Ramachandran G; Prasad B;
Kulmala M. (2005). Fine particle number and mass concentration measurements in urban Indian
households. Sci Total Environ, 347: 131-147. 156775
Mwaiselage J; Moen B; Bratveit M. (2006). Acute respiratory health effects among cement factory workers in
Tanzania: an evaluation of a simple health surveillance tool. Int Arch Occup Environ Health, 79' 49"56.
156789
Na K; Cocker DR. (2005). Organic and elemental carbon concentrations in fine particulate matter in residences,
schoolrooms, and outdoor air in Mira Loma, California. Atmos Environ, 39: 3325-3333. 156790
Naumova YY; Eisenreich SJ; Turpin BJ; Weisel CP; Morandi MT; Colome SD; Totten LA; Stock TH; Winer AM;
Alimokhtari S; Kwon J; Shendell D; Jones J; Maberti S; Wall SJ. (2002). Polycyclic aromatic
hydrocarbons in the indoor and outdoor air of three cities in the US. Environ Sci Technol, 36: 2552-2559.
026105
Naumova YY; Offenberg JH; Eisenreich SJ; Meng QY; Polidori A; Turpin BJ; Weisel CP; Morandi MT; Colome
SD; Stock TH; Winer AM; Alimokhtari S; Kwon J; Maberti S; Shendell D; Jones J; Farrar C. (2003).
Gas/particle distribution of polycyclic aromatic hydrocarbons in coupled outdoor/indoor atmospheres.
Atmos Environ, 37: 703-719. 089213
Nerriere E; Zmirou-Navier D; Blanchard O; Momas I; Laclner J; Le Moullec Y; Personnaz M"B; Lameloise P;
Delmas V; Target A; Desqueyroux H. (2005). Can we use fixed ambient air monitors to estimate
population long-term exposure to air pollutants? The case of spatial variability in the Genotox ER study.
Environ Res, 97: 32-42. 089481
July 2009
A-346
DRAFT-DO NOT CITE OR QUOTE

-------
Nerriere E> Zmirou-Navier D; Desqueyroux P; Leclerc N> Momas L Czernichow P. (2005). Lung cancer risk
assessment in relation with personal exposure to airborne particles in four French metropolitan areas. J
Occup Environ Med, 47: 1211-1217. 088630
Neususs C> Weise D> Birmili W> Wex Hi Wiedensohler Ai Covert DS. (2000). Size-segregated chemical,
gravimetric and number distribution-derived mass closure of the aerosol in Sagres, Portugal during ACE"
2. Tellus B Chem Phys Meteorol, 52: 169-184. 156804
Ng SPi Kendall Mi Dimitroulopoulou C> Grossinho Ai Chen LC. (2005). PM2.5 exposure assessment of the
population in Lower Manhattan area of New York City after the World Trade Center disaster. Atmos
Environ, 39: 1979-1992. 155996
Nikasinovic Li Just J; Sahraoui Fi Seta Ni Grimfeld Ai Momas I. (2006). Nasal inflammation and personal
exposure to fine particles PM2.5 in asthmatic children. J Allergy Clin Immunol, 117: 1382-1388. 156807
NIOSH. (1996). NIOSH Method 5040 Issue 1 (Interim): Elemental Carbon (diesel exhaust). 156810
NIOSH. (1999). NIOSH Method 5040 Issue 3 (Interim): Elemental Carbon (diesel exhaust). 156811
Norris G> Larson Ti Koenig J; Claiborn C> Sheppard Li Finn D. (2000). Asthma aggravation, combustion, and
stagnant air. Thorax, 55: 466-470. 087104
Noullett Mi Jackson PLi Brauer M. (2006). Winter measurements of children's personal exposure and ambient
fine particle mass, sulphate and light absorbing components in a northern community. Atmos Environ,
40: 1971-1990. 155999
Ntziachristos Li Samaras Z. (2006). Combination of aerosol instrument data into reduced variables to study the
consistency of vehicle exhaust particle measurements. Atmos Environ, 40: 6032-6042. 116722
O'Neill MSi Jerrett Mi Kawachi Ii Levy Jli Cohen AJi Gouveia Ni Wilkinson Pi Fletcher Ti Cifuentes Li
Schwartz Ji with input from participants of the Workshop on Air Pollution and Socioeconomic
Conditions. (2003). Health, wealth, and air pollution: advancing theory and methods. Environ Health
Perspect, 111: 1861-1870. 090310
O'Neill MSi Loomis D; Borja Aburto VH; Gold D; Hertz-Picciotto I; Castillejos M. (2004). Do associations
between airborne particles and daily mortality in Mexico City differ by measurement method, region, or
modeling strategy?. J Expo Sci Environ Epidemiol, 14: 429-439. 087429
O'Neill MS; Loomis D; Borja-Aburto VH. (2004). Ozone, area social conditions, and mortality in Mexico City.
Environ Res, 94: 234-242. 055597
Offenberg JH; Naumova YY; Turpin BJ; Eisenreich SJ; Morandi MT; Stock T; Colome SD; Winer AM; Spektor
DM; Zhang J; Weisel CP. (2004). Chlordanes in the indoor and outdoor air of three U.S. cities. Environ
Sci Technol, 38: 2760-2768. 156821
Oglesby L; Kunzli N; Roosli M; Braun-Fahrlander C; Mathys P; Stern W; Jantunen M; Kousa A. (2000). Validity
of ambient levels of fine particles as surrogate for personal exposure to outdoor air pollution-results of
the European EXPOLIS-EAS study (Swiss Center Basel). J Air Waste Manag Assoc, 50: 1251-1261.
001832
Ogulei D; Hopke PK; Chalupa DC; Utell MJ. (2006). Modeling Source Contributions to Submicron Particle
Number Concentrations Measured in Rochester, New York. Aerosol Sci Technol, 41: 179-201. 119975
Ogulei D; Hopke PK; Wallace LA. (2006). Analysis of indoor particle size distributions in an occupied townhouse
using positive matrix factorization. Indoor Air, 16: 204-215. 156823
Olfert JS; Kulkarni P; Wang J. (2008). Measuring aerosol size distributions with the fast integrated mobility
spectrometer. J Aerosol Sci, 39: 940-956. 156004
Ou CQ; Hedley AJ; Chung RY; Thach TQ; Chau YE Chan KP; Yang U Ho SY; Wong CM; Lam TH. (2008).
Socioeconomic disparities in air pollution-associated mortality. Environ Res, 107: 237-244. 189955
Paatero P. (1997). Least squares foumulation of robust non-negative factor analysis. , 37: 23"35. 087001
Paatero P. (1999). The Multilinear Engine: A Table-Driven, Least Squares Program for Solving Multilinear
Problems, including the n-Way Parallel Factor Analysis Model. , 8: 854-888. 156835
Paatero P; Hopke PK; Song XH; Ramadan Z. (2002). Understanding and controlling rotations in factor analytic
models. , 60: 253-264. 156836
Pancras JP; Ondov JM; Zeisler R. (2005). Multi-element electrothermal AAS determination of 11 marker
elements in fine ambient aerosol slurry samples collected with SEAS-II. Anal Chim Acta, 538: 303-312.
098120
Pang Y; Eatough NL; Modey WKi Eatough DJ. (2002). Evaluation of the RAMS continuous monitor for
determination of PM2.5 mass including semi-volatile material in Philadelphia, PA. J Air Waste Manag
Assoc, 52: 563-572. 030353
Pang Y; Gundel LA; Larson T; Finn D; Liu L-JS; Claiborn CS. (2002). Development and evaluation of a personal
particulate organic and mass sampler. Environ Sci Technol, 36: 5205-5210. 037057
July 2009
A-347
DRAFT-DO NOT CITE OR QUOTE

-------
Parekh PP; Husain L. (1981). Trace element concentrations in summer aerosols at rural sites in New York state
and their possible sources. Atmos Environ, 15- 1717-1725. 156840
Park SS; Harrison D; Pancras JP; Ondov JM. (2005). Highly Time-Resolved Organic and Elemental Carbon
Measurements at the Baltimore Supersite in 2002. J Geophys Res, 110- D07S06. 156843
Park SS; Pancras JP; Ondov J; Poor N. (2005). A new pseudodeterministic multivariate receptor model for
individual source apportionment using highly time-resolved ambient concentration measurements. J
Geophys Res, 110: D07S15. 156844
Paschold H; Maciejewska B; Li WW; Morales H; Pingitore NE. (2003). Elemental analysis of airborne
particulate matter and cooling water in west Texas residences. Atmos Environ, 37- 2681-2690. 156847
Pekney NJ; Davidson CI; Bein KJ; Wexler AS; Johnston MV. (2006). Identification of sources of atmospheric PM
at the Pittsburgh supersite, part I- single particle analysis and filter-based positive matrix factorization.
Atmos Environ, 2' 411-423. 086115
Peters TM; Gussman RA; Kenny LC; Vanderpool RW. (2001). Evaluation of PM25 size selectors used in
speciation samplers. Aerosol Sci Technol, 34: 422-429. 017108
Peterson MR; Richards MH. (2002). Thermal-optical-transmittance analysis for organic, elemental, carbonate,
total carbon, and OCX2 in PM2.5 by the EPA/NIOSH method. Presented at Symposium on Air Quality
Measurement Methods and Technology, Pittsburgh. 156861
Petzold A; Kramer H; Schonlinner M. (2002). Continuous Measurement of Atmospheric Black Carbon Using a
Multi-angle Absorption Photometer. Environ Sci Pollut Res Int, 4- 78"82. 156863
Petaja T. (2006). Detection Efficiency of a Water-Based TSI Condensation Particle Counter 3785. Aerosol Sci
Technol, 40: 1090-1097. 156021
Phares DJ; Rhoads KP; Johnston MV; Wexler AS. (2003). Size -resolved ultrafine particle composition analysis
2. Houston. J Geophys Res, 108: 8420. 156866
Pitchford ML; Chow JC; Watson JG; Moore CT; Campbell DE; Eldred RA; Vanderpool RW; Ouchida P; Hering
SV; Frank NH. (1997). Prototype PM2.5 federal reference method field studies report - An EPA staff
report. 156872
Polidori A; Arhami M; Sioutas C; Delfino RJ; Allen R. (2007). Indoor/Outdoor relationships, trends, and
carbonaceous content of fine particulate matter in retirement homes of the Los Angeles Basin. J Air
Waste Manag Assoc, 57: 366-379. 156877
Poore MW. (2000). Oxalic acid in PM2.5 particulate matter in California. J Air Waste Manag Assoc, 50: 1874"
1875. 012839
Pope CA III; Burnett RT; Thun MJ; Calle EE; Krewski D; Ito K; Thurston GD. (2002). Lung cancer,
cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA, 287: 1132-
1141. 024689
Poupard O; Blondeau P; Iordache V; Allard F. (2005). Statistical analysis of parameters influencing the
relationship between outdoor and indoor air quality in schools. Atmos Environ, 39: 2071-2080. 074025
Price M; Bulpitt S; Meyer MB. (2003). A comparison of PM10 monitors at a Kerbside site in the northeast of
England. Atmos Environ, 37: 4425-4434. 098082
Qin X; Prather KA. (2006). Impact of biomass emissions on particle chemistry during the California Regional
Particulate Air Quality Study. Int J Mass Spectrom, 258: 142-150. 156895
Ramachandran G; Adgate J; Pratt G; Sexton K. (2003). Characterizing indoor and outdoor 15 minute average
PM 2.5 concentrations in urban neighborhoods. Aerosol Sci Technol, 37: 33"45. 188454
Ramachandran G; Adgate JL; Pratt GC; Sexton K. (2003). Characterizing indoor and outdoor 15 minute average
PM 2.5 concentrations in urban neighborhoods. Aerosol Sci Technol, 37: 33"45. 190112
Rattigan OV; Hogrefe O; Felton HD; Schwab JJ; Roychowdhury UK; Husain L; Dutkiewicz VA; Demerjian KL.
(2006). Multi-year urban and rural semi-continuous PM25 sulfate and nitrate measurements in New
York state: Evaluation and comparison with filter based measurements. Atmos Environ, 40: 192-205.
115897
Rees SL; Robinson AL; Khlystov A; Stanier CO; Pandis SNSN. (2004). Mass balance closure and the Federal
Reference Method for PM2.5 in Pittsburgh, Pennsylvania. Atmos Environ, 38: 3305-3318. 097164
Reff A; Weisel CP; Zhang J; Morandi M; Stock T; Colome S; Winer A; Turpin BJ; Offenberg JH. (2007). A
functional group characterization of organic PM2.5 exposure: Results from the RIOPA study. Atmos
Environ, 41: 4585-4598. 156045
Reimann C; De Caritat P. (2000). Intrinsic flaws of element enrichment factors (EFs) in environmental
geochemistry. Environ Sci Technol, 34: 5084-5091. 013269
Rinehart LR; Fujita EM; Chow JC; Magliano K; Zielinska B. (2006). Spatial distribution of PM 2.5 associated
organic compounds in central California. Atmos Environ, 40: 290-303. 115184
July 2009
A-348
DRAFT-DO NOT CITE OR QUOTE

-------
Riojas-Rodriguez Hi Escamilla-Cejudo JA; Gonzalez-Hermosillo JA; Tellez-Rojo MM; Vallejo Mi Santos-Burgoa
Ci Rojas-Bracho L. (2006). Personal PM2.5 and CO exposures and heart rate variability in subjects with
known ischemic heart disease in Mexico City. J Expo Sci Environ Epidemiol, 16- 131-137. 156913
Robinson DLi Monro JMi Campbell EA. (2007). Spatial variability and population exposure to PM2.5 pollution
from woodsmoke in a New South Wales country town. Atmos Environ, 41' 5464-5478. 156054
Rojas-Bracho Li Suh HH; Catalano PJ; Koutrakis P. (2004). Personal exposures to particles and their
relationships with personal activities for chronic obstructive pulmonary disease patients living in Boston.
J Air Waste Manag Assoc, 54: 54:207-217. 054772
Romieu L Ramirez-Aguilar Mi Moreno-Macias Hi Barraza-Villarreal A> Miller Pi Hernandez-Cadena Li
Carbajal"Arroyo LA; Hernandez-Avila M. (2004). Infant mortality and air pollution: modifying effect by
social class. J Occup Environ Hyg, 46: 1210-1216. 093074
Rossner Pi Svecova V; Milcova A> Lnenickova Zi Solansky Ni Sram RJ. (2008). Seasonal variability of oxidative
stress markers in city bus drivers - Part I. Oxidative damage to DNA. , 642: 14-20. 156927
Rotko Ti Oglesby Li Kunzli Ni Carrer Pi Nieuwenhuijsen MJi Jantunen M. (2002). Determinants of perceived
air pollution annoyance and association between annoyance scores and air pollution (PM25, N02)
concentrations in the European EXPOLIS study. Atmos Environ, 36: 4593-4602. 037240
Rotko T> Oglesby Li Kunzli Ni Jantunen MJ. (2000). Population sampling in European air pollution exposure
study, EXPOLIS: comparisons between the cities and representativeness of the samples. J Expo Sci
Environ Epidemiol, 10: 355-364. 012118
Rupprecht & Patashnick Co. (2003). Innovative instrument for an ambient air particulate mass measurement
stadard. 157207
Russell Mi Allen DTi Collins DRi Fraser MP. (2004). Daily, seasonal and spatial trends in PM25 mass and
composition in southeast Texas. Aerosol Sci Technol, l: 14-26. 082453
Sabin LDi Kozawa Ki Behrentz Ei Winer AMi Fitz DRi Pankratz DVi Colome SDi Fruin SA. (2005). Analysis of
real-time variables affecting children's exposure to diesel-related pollutants during school bus commutes
in Los Angeles. Atmos Environ, 39: 5243-5254. 087728
Sabin LDi Lim JHi Stolzenbach KDi Schiff KC. (2005). Contribution of trace metals from atmospheric deposition
to stormwater runoff in a small impervious urban catchment. Water Res, 39: 3929-3937. 088300
Salma Ii Maenhaut Wi Ocskay Ri Raes N. (2005). Fine structure of mass size distributions in an urban
environment. Atmos Environ, 39: 5363-5374. 156937
Salma Ii Ocskay Ri Chi Xi Maenhaut W. (2007). Sampling artefacts, concentration and chemical composition of
fine water-soluble organic carbon and humic-like substances in a continental urban atmospheric
environment. Atmos Environ, 41: 4106-4118. 113852
Salma Ii Weidinger Ti Maenhaut W. (2007). Time-resolved mass concentration, composition and sources of
aerosol particles in a metropolitan underground railway station. Atmos Environ, 41: 8391-8405. 113850
Samet JMi Dominici Fi Curriero FCi Coursac Ii Zeger SL. (2000). Fine particulate air pollution and mortality in
20 US cities, 1987-1994. , 343: 1742-1749. 013132
Samson PJ. (1978). Ensemble trajectory analysis of summertime sulfate concentrations in New York State.
Atmos Environ, 12: 1889-1893. 156941
Samson PJ. (1980). Trajectory analysis of summertime sulfate concentrations in the northeastern United
States. , 19: 1382-1394. 073010
Sanderson EGi Farant JP. (2004). Indoor and outdoor polycyclic aromatic hydrocarbons in residences
surrounding a Soderberg aluminum smelter in Canada. Environ Sci Technol, 38: 5350-5356. 156942
Sarnat JAi Brown KWi Schwartz Ji Coull BAi Koutrakis P. (2005). Ambient gas concentrations and personal
particulate matter exposures: implications for studying the health effects of particles. , 16: 385-395.
087531
Sarnat JAi Schwartz Ji Catalano PJi Suh HH. (2001). Gaseous pollutants in particulate matter epidemiology:
confounders or surrogates?. Environ Health Perspect, 109: 1053-1061. 019401
Sarnat SEi Coull BAi Ruiz PAi Koutrakis Pi Suh HH. (2006). The influences of ambient particle composition and
size on particle infiltration in Los Angeles, CA residences. J Air Waste Manag Assoc, 56: 186-196. 089166
Sarnat SEi Coull BAi Schwartz Ji Gold DRi Suh HH. (2006). Factors affecting the association between ambient
concentrations and personal exposures to particles and gases. Environ Health Perspect, 114: 649-654.
089784
Sarnat SEi Suh HHi Coull BAi Schwartz Ji Stone PHi Gold DR. (2006). Ambient particulate air pollution and
cardiac arrhythmia in a panel of older adults in Steubenville, Ohio. Occup Environ Med, 63: 700-706.
090489
July 2009
A-349
DRAFT-DO NOT CITE OR QUOTE

-------
Sax SN; Bennett DH; Chillrud SN; Ross J; Kinney PL; Spengler JD. (2006). A cancer risk assessment of inner-
city teenagers living in New York City and Los Angeles. Environ Health Perspect, 114- 1558-1566.
156950
Schauer JJ; Rogge WF; Hildemann LMJ Mazurek MA; Cass GR. (1996). Source apportionment of airborne
particulate matter using organic compounds as tracers. Atmos Environ, 30- 3837-3855. 051162
Scheepers PT; Micka V; Muzyka V; Anzion R; Dahmann D; Poole J; Bos RP. (2003). Exposure to dust and
particle-associated 1-nitropyrene of drivers of diesel-powered equipment in underground mining. Ann
Occup Hyg, 47: 379-388. 156955
Schnelle-Rreis J; Sklorz M; Peters A; Cyrys J; Zimmermann R. (2005). Analysis of particle-associated semi-
volatile aromatic and aliphatic hydrocarbons in urban particulate matter on a daily basis. Atmos
Environ, 39: 7702-7714. 112944
Schwartz J. (2000). The distributed lag between air pollution and daily deaths. , 11: 320-326. 002470
Schwartz J; Sarnat JA; Coull BA; Wilson WE. (2007). Effects of exposure measurement error on particle matter
epidemiology: a simulation using data from a panel study in Baltimore, MD. J Expo Sci Environ
Epidemiol, 17: S2-S10. 090220
Shalat SL; Lioy PJ; Schmeelck K; Mainelis G. (2007). Improving estimation of indoor exposure to inhalable
particles for children in the first year of life. J Air Waste Manag Assoc, 57: 934-939. 156971
Shao L; Yang S; Li H; Li W', Jones T; Sexton K; B,ruB, K; Li J; Zhao H. (2007). Associations between particle
physicochemical characteristics and oxidative capacity: An indoor PM10 study in Beijing, China. Atmos
Environ, 41: 5316-5326. 156973
Shilton V; Giess P; Mitchell D; Williams C. (2002). The relationships between indoor and outdoor respirable
particulate matter: meteorology, chemistry and personal exposure. Indoor Built Environ, 11: 266-274.
049602
Sidhu S; Graham J; Striebich R. (2001). Semi-volatile and particulate emissions from the combustion of
alternative diesel fuels. Chemosphere, 42: 681-90. 155202
Simons E; Curtin-Brosnan J; Buckley T; Breysse P; Eggleston PA. (2007). Indoor environmental differences
between inner city and suburban homes of children with asthma. J Urban Health, 84: 577-590. 156982
Smith TJ; Davis ME; Reaser P; Natkin J; Hart JE; Laden F; Heff A; Garshick E. (2006). Overview of particulate
exposures in the US trucking industry. J Environ Monit, 8: 711-720. 156990
Solomon P; Baumann K; Edgerton E; Tanner R; Eatough D; Modey W', Marin H; Savoie D; Natarajan S; Meyer
MB. (2003). Comparison of integrated samplers for mass and composition during the 1999 Atlanta
supersites project. J Geophys Res, 108: 8423. 156994
Solomon PA; Norris G; Landis M; Tolocka M. (2001). Chemical analysis methods for atmospheric aerosol
components. : . 156993
Solomon PA; Sioutas C. (2006). Continuous and semi-continuous methods for PM mass and composition. 17-23.
156995
Sorensen M; Daneshvar B; Hansen M; Dragsted LO; Hertel 0; Knudsen L; Loft S. (2003). Personal PM25
exposure and markers of oxidative stress in blood. Environ Health Perspect, 111: 161-165. 042700
Sorensen M; Schins RPF; Hertel O; Loft S. (2005). Transition metals in personal samples of PM25 and oxidative
stress in human volunteers. Cancer Epidemiol Biomarkers Prev, 14: 1340-1343. 083053
Sram R; Beskid O; Binkova B; Chvatalova I; Lnenickova Z; Milcova A; Solansky I; Tulupova E; Bavorova H;
Ocadlikova D. (2007). Chromosomal aberrations in environmentally exposed population in relation to
metabolic and DNA repair genes polymorphisms. , 620: 22"33. 188457
Sram RJ; Beskid O; Rossnerova A; Rossner P; Lnenickova Z; Milcova A; Solansky I; Binkova B. (2007).
Environmental exposure to carcinogenic polycyclic aromatic hydrocarbons-"the interpretation of
cytogenetic analysis by FISH. Toxicol Lett, 172: 12-20. 192084
Srivastava A; Jain VK. (2007). A study to characterize the suspended particulate matter in an indoor
environment in Delhi, India. , 42: 2046-2052. 157004
Stanier CO; Khlystov AY; Chan WR; Mandiro M; Pandis SN. (2004). A Method for the In Situ Measurement of
Fine Aerosol Water Content of Ambient Aerosols: The Dry-Ambient Aerosol Size Spectrometer (DAASS).
Aerosol Sci Technol, 38: 215 - 228. 095955
Stein SW; Myrdal PB; Beck TJ; Gabrio BJ; Oberreit D; Hairston P. (2002). An evaluation of mass-weighted size
distribution measurements with the Model 3320 aerodynamic particle sizer. Aerosol Sci Technol, 36: 845"
854. 157008
Stohl A. (1996). Trajectory statistics-A new method to establish source-receptor relationships of air pollutants
and its application to the transport of particulate sulfate in Europe. Atmos Environ, 30: 579-587. 157014
July 2009
A-350
DRAFT-DO NOT CITE OR QUOTE

-------
Strand M> Hopke PKi Zhao W> Vedal Si Gelfand Ei Rabinovitch N. (2007). A study of health effect estimates
using competing methods to model personal exposures to ambient PM2.5. J Expo Sci Environ Epidemiol,
17: 549-558. 157018
Strand Mi Vedal Si Rodes Ci Dutton SJi Gelfand EWi Rabinovitch N. (2006). Estimating effects of ambient
PM25 exposure on health using PM2.5 component measurements and regression calibration. J Expo Sci
Environ Epidemiol, 16: 30-38. 089203
Stranger Mi Potgieter-Vermaak Si Van Grieken R. (2008). Characterization of indoor air quality in primary
schools in Antwerp, Belgium. Indoor Air, l: 454-4863. 190884
Stranger Mi Potgieter-Vermaak Si Van Grieken R. (2009). Particulate matter and gaseous pollutants in
residences in Antwerp, Belgium. Sci Total Environ, 1182-1192: 407. 190883
Subbalakshmi Yi Patti AFi Lee GSHi Hooper MA. (2000). Structural characterisation of macromolecular organic
material in air particulate matter using Py-GC-MS and solid state 13 C-NMR. J Environ Monit, 2' 561-
565. 157023
Subramanian Ri Khlystov AYi Cabada JCi Robinson AL. (2004). Positive and negative artifacts in particulate
organic carbon measurements with denuded and undenuded sampler configurations. Aerosol Sci Technol,
l: 27-48. 081203
Sorensen Mi Daneshvar Bi Hansen Mi Dragsted LOi Hertel Oi Knudsen Li Loft S. (2003). Personal PM2.5
exposure and markers of oxidative stress in blood. Environ Health Perspect, 111: 161-166. 157000
Sorensen Mi Loft Si Andersen HVi Raaschou-Nielsen Oi Skovgaard LTi Knudsen LEi Nielsen IVi Hertel O.
(2005). Personal exposure to PM25, black smoke and N02 in Copenhagen: relationship to bedroom and
outdoor concentrations covering seasonal variation. J Expo Sci Environ Epidemiol, 15: 413-422. 089428
Takahama Si Wittig AEi Vayenas DVi Davidson Cli Pandis SN. (2004). Modeling the diurnal variation of
nitrate during the Pittsburgh Air Quality Study. J Geophys Res, 109: D16S06. 157038
Tanaka Si Yasushi Ni Sato Ni Fukasawa Ti Santosa SJi Yamanaka Ki Ootoshi T. (1998). Rapid and
simultaneous multi-element analysis of atmospheric particulate matter using inductively coupled plasma
mass spectrometry with laser ablation sample introduction. J Anal At Spectrom, 13: 135-140. 157041
Tang C-Si Chang L"Ti Lee H"Ci Chan C-C. (2007). Effects of personal particulate matter on peak expiratory flow
rate of asthmatic children. Sci Total Environ, 382: 43-51. 091269
Tatum Vi Ray AEi Rovell-Rixx D. (2002). Performance of the RespiCon personal aerosol sampler in forest
products industry workplaces. AIHA J, 63: 311-316. 157046
Thomaidis NSi Bakeasa EBi Siskos PA. (2003). Characterization of lead, cadmium, arsenic and nickel in PM25
particles in the Athens atmosphere, Greece. Chemosphere, 52: 959-966. 044193
Thornburg JWi Stevens CD; Williams RW; Rodes CE; Lawless PA. (2004). A pilot study of the influence of
residential HAC duty cycle on indoor air quality. Atmos Environ, 38: 1567-1577. 157052
Toivola Mi Aim Si Reponen Ti Kolari Si Nevalainen A. (2002). Personal exposures and microenvironmental
concentrations of particles and bioaerosols. J Environ Monit, 4- 166-174. 026571
Tolbert PEi KLeina Mi Peelb JLi Sarnata SEi Sarnata JA. (2007). Multipollutant modeling issues in a study of
ambient air quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 24: 938"
945. 090316
Tolbert PEi Mulholland JAi Macintosh DLi Xu Fi Daniels Di Devine OJi Carlin BPi Klein Mi Dorley Ji Butler
AJi Nordenberg DFi Frumkin Hi Ryan PBi White MC. (2000). Air quality and pediatric emergency room
visits for asthma in Atlanta, Georgia. Am J Epidemiol, 151: 798-810. 001993
Tovalin Hi Valverde Mi Morandi MTi Blanco Si Whitehead Li Rojas E. (2006). DNA damage in outdoor workers
occupationally exposed to environmental air pollutants. Occup Environ Med, 63: 230-236. 091322
Tovalin"Ahumada Hi Whitehead Li Blanco S. (2007). Personal exposure to PM2.5 and element composition- A
comparison between outdoor and indoor workers from two Mexican cities. Atmos Environ, 41: 7401-7413.
190165
Trenga CAi Sullivan JHi Schildcrout JSi Shepherd KPi Shapiro GGi Liu LJi Kaufman JDi Koenig JQ. (2006).
Effect of particulate air pollution on lung function in adult and pediatric subjects in a Seattle panel
study. Chest, 129: 1614-22. 155209
Turpin BJi Weisel CPi Morandi Mi Colome Si Stock Ti Eisenreich Si Buckley B. (2007). Relationships of Indoor,
Outdoor, and Personal Air (RIOPA): Part II. Analyses of concentrations of particulate matter species.
157062
Tursic Ji Podkrajsek Bi Grgic Ii Ctyroky Pi Berner A> Dusek Ui Hitzenberger R. (2006). Chemical composition
and hygroscopic properties of size-segregated aerosol particles collected at the Adriatic coast of Slovenia.
Chemosphere, 63: 1193-1202. 157063
July 2009
A-351
DRAFT-DO NOT CITE OR QUOTE

-------
Urch B; Brook JR; Wasserstein D; Brook RD; Rajagopalan Si Corey Pi Silverman F. (2004). Relative
contributions of PM2.5 chemical constituents to acute arterial vasoconstriction in humans. Inhal Toxicol,
16: 345-352. 055629
Vallejo Mi Ruiz Si Hermosillo AGi Borja-Aburto VHi Cardenas M. (2006). Ambient fine particles modify heart
rate variability in young healthy adults. J Expo Sci Environ Epidemiol, 16: 125-130. 157081
Van Roosbroeck Si Wichmann Ji Janssen NAHi Hoek Gi Van Wijnen JHi Lebret Ei Brunekreef B. (2006). Long-
term personal exposure to traffic-related air pollution among school children, a validation study. Sci
Total Environ, 368: 565-573. 090773
Vaughn Di O'Brien Ti Roberts PTi Rice J. (2005). Comparison of integrated filter and semi-continuous
measurements of PM2.5 nitrate, sulfate, and carbon aerosols in the Speciation Trends Network (STN).
157089
Veltkamp PRi Hansen KJi Barkley RMi Sievers RE. (1996). Principal component analysis of summertime
organic aerosols at Niwot Ridge, Colorado. J Geophys Res, 101: 19,495-19,504. 081594
Venkatachari Pi Zhou Li Hopke Pi Schwab Ji Demerjian Ki Weimer Si Hogrefe Oi Felton Di Rattigan O. (2006).
An intercomparison of measurement methods for carbonaceous aerosol in the ambient air in New York
City. Aerosol Sci Technol, 40: 788-795. 105918
Verma DKi Kurtz LAi Sahai Di Finkelstein MM. (2003). Current chemical exposures among Ontario
construction workers. Appl Occup Environ Hyg, 18: 1031-1047. 157093
Villeneuve PJi Burnett RTi Shi Yi Krewski Di Goldberg MSi Hertzman Ci Chen Yi Brook J. (2003). Atime-
series study of air pollution, socioeconomic status, and mortality in Vancouver, Canada. J Expo Sci
Environ Epidemiol, 13: 427-435. 055051
Vinzents PSi Moller Pi Sorensen Mi Knudsen LEi Herte LQi Jensen FPi Schibye Bi Loft S. (2005). Personal
exposure to ultrafine particles and oxidative DNA damage. Environ Health Perspect, 113: 1485-1490.
087482
Virkkula Ai Ahlquist NCi Covert DSi Arnott WPi Sheridan PJi Quinn PKi Coffman DJ. (2005). Modification,
Calibration and a Field Test of an Instrument for Measuring Light Absorption by Particles. Aerosol Sci
Technol, 39: 68-83. 157097
Voorhees KJi Schulz WDi Kunen SMi Hendricks LJi Currie LAi Klouda G. (1991). Analysis of Insoluble
Carbonaceous Materials from Airborne Particles Collected in Pristine Regions of Colorado. J Anal Appl
Pyrol, 18: 189-205. 157101
Wallace L. (2005). Ultrafine particles from a vented gas clothes dryer. Atmos Environ, 39: 5777-5786. 157102
Wallace Li Williams R. (2005). Use of personal-indoor-outdoor sulfur concentrations to estimate the infiltration
factor and outdoor exposure factor for individual homes and persons. Environ Sci Technol, 39: 1707-1714.
057485
Wallace Li Williams Ri Rea Ai Croghan C. (2006). Continuous weeklong measurements of personal exposures
and indoor concentrations of fine particles for 37 health-impaired North Carolina residents for up to four
seasons. Atmos Environ, 40: 399-414. 088211
Wallace LAi Mitchell Hi O'Connor GTi Neas Li Lippmann Mi Kattan Mi Koenig Ji Stout JWi Vaughn BJi
Wallace Di Walter Mi Adams Ki Liu L-JS. (2003). Particle concentrations in inner-city homes of children
with asthma: the effect of smoking, cooking, and outdoor pollution. Environ Health Perspect, 111: 1265"
1272. 053553
Wan ECHi Yu JZ. (2006). Determination of sugar compounds in atmospheric aerosols by liquid chromatography
combined with positive electrospray ionization mass spectrometry. J Chromatogr, 1107: 175-181. 157104
Wang X; Fu J; Bi X; Sheng G. (2006). Hospital indoor PM10/PM2.5 and associated trace elements in
Guangzhou, China. Sci Total Environ, 366: 124-135. 157108
Ward TJ; Noonan CW; Hooper K. (2007). Results of an indoor size fractionated PM school sampling program in
Libby, Montana. Environ Monit Assess, 130: 163-171. 157112
Waterman D; Horsfield B; Leistner F; Hall Ki Smith S. (2000). Quantification of polycyclic aromatic
hydrocarbons in the NIST Standard Reference Material (SRM1649A) urban dust using thermal
Desorption GC/MS. Anal Chem, 72: 3563-3567. 157116
Waterman Di Smith Si Green Di Horsfield Bi Hall K. (2001). The application of a thermal desorption GCMS
technique for the organic analysis of airborne particulate matter. Adv Mass Spectrom, 14: 887-888.
157117
Watson JGi Chen LWi Chow JCi Doraiswamy Pi Lowenthal DH. (2008). Source apportionment: findings from
the U.S. Supersites Program. J Air Waste Manag Assoc, 58: 265-288. 157128
Watson JGi Chow JC. (2001). Ambient air sampling. In Baron PA; Willeke K (Ed.),Aerosol measurement:
principles, techniques, and applications (pp. 821-844). New York: John Wiley & Sons Ltd. 157123
July 2009
A-352
DRAFT-DO NOT CITE OR QUOTE

-------
Watson JG; Chow JC. (2002). Comparison and evaluation of in situ and filter carbon measurements at the
Fresno Supersite. J Geophys Res, 107. 037873
Watson JG; Chow JC; Chen LWA. (2005). Summary of organic and elemental carbon/black carbon analysis
methods and intercomparisons. Aerosol Air Qual Res, 5- 69—102. 157125
Watson JG; Chow JC; Frazier CA. (1999). X-ray fluorescence analysis of ambient air samples. In Landsberger,
S.; Creatchman, M. (Ed.),Elemental analysis of airborne particlesAmsterdam, The Netherlands' Gordon
and Breach Science Publishers. 020949
Watson JG; Chow JC; Shah JJ; Pace TG. (1983). The effect of sampling inlets on the PM-10 and PM-15 to TSP
concentration ratios. J Air Waste Manag Assoc, 33- 114-119. 045084
Watson JG; Cooper JA; Huntzicker JJ. (1984). The effective variance weighting for least squares calculations
applied to the mass balance receptor model. Atmos Environ, 18- 1347-1355. 045693
Watson JG; Robinson NF; Lewis CW; Coulter CT; Chow JC; Fujita EM; Lowenthal DH; Corner TL; Henry RC;
Willis RD. (1997). Chemical mass balance receptor model version 8 (CMB) users manual. 157121
Watson JG; Turpin BJ; Chow JC. (1989). The measurement process- Precision, accuracy, and validity. • . 157119
Weber RJ; Orsini D; Daun Y; Lee Y-NJ Kotz PJ; Brechtel F. (2001). A particle-into-liquid collector for rapid
measurement of aerosol bulk chemical composition. Aerosol Sci Technol, 35- 718-727. 024640
Weber RJ; Orsini D; Duan Y; Baumann K; Kiang CS; Chameides W', Lee YN; Brechtel F; Klotz P; Jongejan P.
(2003). Intercomparison of near real time monitors of PM2. 5 nitrate and sulfate at the US
Environmental Protection Agency Atlanta Supersite. J Geophys Res, 108- 8421. 157129
Weisel CP; Zhang J; Turpin BJ; Morandi MT; Colome S; Stock TH; Spektor DM; Korn L; Winer AM; Kwon J;
Meng QY; Zhang L; Harrington R; Liu W', Reff A; Lee JH; Alimokhtari S; Mohan K; Shendell D; Jones J;
Farrar L; Maberti S; Fan T. (2005). Relationships of Indoor, Outdoor, and Personal Air (RIOPA): Part I.
Collection methods and descriptive analyses. , 130- 1-107. 157131
Welthagen W; Schnelle-Rreis J; Zimmermann R. (2003). Search criteria and rules for comprehensive two-
dimensional gas chromatography "time-of-flight mass spectrometry analysis of airborne particulate
matter. J Chromatogr, 1019- 233-249. 104056
Wenzel RJ; Liu DY; Edgerton ES; Prather KA. (2003). Aerosol time-of-flight mass spectrometry during the
Atlanta Supersite Experiment- 2. Scaling procedures. J Geophys Res, 108- 8426. 157139
Westerdahl D; Fruin S; Sax T; Fine PM; Sioutas C. (2005). Mobile platform measurements of ultrafine particles
and associated pollutant concentrations on freeways and residential streets in Los Angeles. Atmos
Environ, 39: 3597-3610. 086502
Wheeler BW; Ben-Shlomo Y. (2005). Environmental equity, air quality, socioeconomic status, and respiratory
health: a linkage analysis of routine data from the Health Survey for England. J Epidemiol Community
Health, 59: 948-954. 089860
Wichmann J; Janssen NAH; Van Der Zee S; Brunekreff B. (2005). Traffic-related differences in indoor and
personal absorption coefficient measurements in Amsterdam, the Netherlands. Atmos Environ, 39: 7384"
7392. 086240
Williams B; Goldstein A; Kreisberg N; Hering S. (2006). An In-Situ Instrument for Speciated Organic
Composition of Atmospheric Aerosols: Thermal Desorption Aerosol GC/MS-FID (TAG). Aerosol Sci
Technol, 40: 627-638. 156157
Williams R; Rea A; Vette A; Croghan C; Whitaker D; Stevens C; McDow S; Fortmann R; Sheldon L; Wilson H;
Thornburg J; Phillips M; Lawless P; Rodes C; Daughtrey H. (2008). The design and field implementation
of the Detroit Exposure and Aerosol Research Study. J Expo Sci Environ Epidemiol, x: 1"17. 191201
Williams R; Suggs J; Rea A; Sheldon L; Rodes C; Thornburg J. (2003). The Research Triangle Park particulate
matter panel study: modeling ambient source contribution to personal and residential PM mass
concentrations. Atmos Environ, 37: 5365-5378. 053338
Wilson JG; Zawar-Reza P. (2006). Intraurban-scale dispersion modelling of particulate matter concentrations:
applications for exposure estimates in cohort studies. Atmos Environ, 40: 1053-1063. 088292
Wilson WE; Brauer M. (2006). Estimation of ambient and non-ambient components of particulate matter
exposure from a personal monitoring panel study. J Expo Sci Environ Epidemiol, 16: 264-274. 088933
Wilson WE; Mar TF; Koenig JQ. (2007). Influence of exposure error and effect modification by socioeconomic
status on the association of acute cardiovascular mortality with particulate matter in Phoenix. J Expo Sci
Environ Epidemiol, 17: S11-S19. 157149
Winkler PM; Steiner G; Vrtala A; Vehkamaki H; Noppel M; Lehtinen KEJ; Reischl GP; Wagner PE; Kulmala M.
(2008). Heterogeneous Nucleation Experiments Bridging the Scale from Molecular Ion Clusters to
Nanoparticles. , 319: 1374. 156160
July 2009
A-353
DRAFT-DO NOT CITE OR QUOTE

-------
Wittig AE; Takahama S; Khlystov AY; Pandis SN; Hering S; Kirby B; Davidson C. (2004). Semi-continuous
PM25 inorganic composition measurements during the Pittsburgh Air Quality Study. Atmos Environ, 38-
3201-3213. 103413
Wojtyniak B; Rabczenko D; Stokwiszewski J. (2001). Does air pollution have respect for the socio-economic
status of people?. 090372
Wong CM; Ou CQ; Chan KP; Chau YE Thach TQ; Yang U Chung RY; Thomas GN; Peins JS; Wong TW;
Hedley AJ; Lam TH. (2008). The effects of air pollution on mortality in socially deprived urban areas in
Hong Kong, China. Environ Health Perspect, 116: 1189-1194. 157151
Wu C-F; Delfino RJ; Floro JN; Quintanai. (2005). Exposure assessment and modeling of particulate matter for
asthmatic children using personal nephelometers. Atmos Environ, 39: 3457-3469. 086397
Wu CF; Delfino RJ; Floro JN; Samimi BS; Quintana PJ; Kleinman MT; Liu LJ. (2005). Evaluation and quality
control of personal nephelometers in indoor, outdoor and personal environments. J Expo Sci Environ
Epidemiol, 15: 99-110. 157155
Wu CF; Jimenez J; Claiborn C; Gould T; Simpson CD; Larson T; Liu LJS. (2006). Agricultural burning smoke in
Eastern Washington: Part II. Exposure assessment. Atmos Environ, 40: 5379-5392. 179950
Wu J; M Winer A; J Delfino R. (2006). Exposure assessment of particulate matter air pollution before, during,
and after the 2003 Southern California wildfires. Atmos Environ, 40: 3333-3348. 157156
Xiao H-Y; Liu C-Q. (2004). Chemical characteristics of water-soluble components in TSP over Guiyang, SW
China, 2003. Atmos Environ, 38: 6297-6306. 056801
Yang H; Jian ZY; Hang Ho SS; Xu J; Wu WS; Chun HW; Wang X; Wang L. (2005). The chemical composition of
inorganic and carbonaceous materials in PM25 in Nanjing, China. Atmos Environ, 39: 3735-3749. 102388
Yang H; Li Q; Yu JZ. (2003). Comparison of two methods for the determination of water-soluble organic carbon
in atmospheric particles. Atmos Environ, 37: 865-870. 156167
Yang W; Sohn J; Kim J; Son B; Park J. (2009). Indoor air quality investigation according to age of the school
buildings in Korea. J Environ Manage, 90: 348-354. 190885
Yao X; Fang M; Chan CK; Ho KF; Lee SC. (2004). Characterization of dicarboxylic acids in PM25 in Hong Kong.
Atmos Environ, 38: 963-970. 102213
Yeh S; Small MJ. (2002). Incorporating exposure models in probabilistic assessment of the risks of premature
mortality from particulate matter. J Expo Sci Environ Epidemiol, 12: 389-403. 040077
Yip FY; Robins TG; Parker EA; Israel BA; Brake field-Caldwell W', Keeler GJ; Dvonch JTE. (2004). Personal
exposures to particulate matter among children with asthma in Detroit, Michigan. Atmos Environ, 38:
5227-5236. 157166
Yu KN; Cheung YP; Cheung T; Henry RC. (2004). Identifying the impact of large urban airports on local air
quality by nonparametric regression. Atmos Environ, 38: 4501-4507. 101779
Zanobetti A; Schwartz J. (2000). Race, gender, and social status as modifiers of the effects of PM10 on mortality.
J Occup Environ Med, 42: 469-474. 010198
Zanobetti A; Schwartz J; Dockery DW. (2000). Airborne particles are a risk factor for hospital admissions for
heart and lung disease. Environ Health Perspect, 108: 1071-1077. 011979
Zanobetti A; Schwartz J; Gold D. (2000). Are there sensitive subgroups for the effects of airborne particles?.
Environ Health Perspect, 108: 841-845. 012187
Zeka A; Zanobetti A; Schwartz J. (2006). Individual-level modifiers of the effects of particulate matter on daily
mortality. Am J Epidemiol, 163: 849-859. 088749
Zhang Q; Anastasio C. (2003). Free and combined amino compounds in atmospheric fine particles (PM2. 5) and
fog waters from Northern California. Atmos Environ, 37: 2247-2258. 157182
Zhang Q; Jimenez JL; Canagaratna MR; Jayne JT; Worsnop DR. (2005). Time- and size-resolved chemical
composition of submicron particles in Pittsburgh: Implications for aerosol sources and processes. J
Geophys Res, 110: 1-19. 157185
Zhao W; Hopke PK. (2004). Source apportionment for ambient particles in the San Gorgonio wilderness. Atmos
Environ, 38: 5901-5910. 100956
Zhao W; Hopke PK; Norris G; Williams R; Paatero P. (2006). Source apportionment and analysis on ambient
and personal exposure samples with a combined receptor model and an adaptive blank estimation
strategy. Atmos Environ, 40: 3788-3801. 156181
Zhao W; Rabinovitch N; Hopke PK; Gelfand EW. (2007). Use of an expanded receptor model for personal
exposure analysis in schoolchildren with asthma. Atmos Environ, 41: 4084-4096. 156182
Zheng M; Cass GR; Schauer JJ; Edgerton ES. (2002). Source apportionment of PM2.5 in the southeastern
United States using solvent-extractable organic compounds as tracers. Environ Sci Technol, 36: 2361"
2371. 026100
July 2009
A-3 54
DRAFT-DO NOT CITE OR QUOTE

-------
Zhou L; Hopke PE Liu W. (2004). Comparison of two trajectory based models for locating particle sources for
two rural New York sites. Atmos Environ, 38: 1955-1963. 157190
Zhu Y> Hinds WC; Krudysz M> Kuhn T> Froines J; Sioutas C. (2005). Penetration of freeway ultrafine particles
into indoor environments. J Aerosol Sci, 36^ 303-322. 190081
Zhu Y> Kuhn T> Mayo P; Hinds WC. (2005). Comparison of Daytime and Nighttime Concentration Profiles and
Size Distributions of Ultrafine Particles near a Major Highway. Environ Sci Technol, 39^ 2531-2536.
157191
Zollner I. (2007). Concentrations of Particulate Matter in Schools in Southwest Germany. Inhal Toxicol, 19^ 245"
249. 157192
July 2009
A-355
DRAFT-DO NOT CITE OR QUOTE

-------
Annex B. Dosimetry
B.1. Ultrafine Disposition
Table B-1. Ultrafine disposition in humans.
Reference Study Group Aerosol
Study Protocol
Observations
Mills et
Healthy
Carbon -
al.(2006,
nonsmokers
99mTc
088770)
(5 M, 5 F;
108 nm

21-24 yrs)
CMD


(CJa - 2.2)


Technegas


Generator
Moller et al.
Healthy
Carbon -
(2008)
nonsmokers
99mTc

(n - 9;
— 100 nm

50 ± 11 yrs)
CMD

Smokers
Technegas

(n - 10;
Generator

51 ± 10 yrs)


COPD patients


(n - 7;


69 ±10 yrs)

Lung activity in the lung was measured at
0,1, and 6 h post aerosol inhalation.
On avg, lung activity decreased 3.2 ± 0.7% during the first h and
1.2 ± 1.7% over the next 5 h. With 95.6% of the particles in the
lungs at 6 h post inhalation and no accumulation of radioactivity
detected over the liver or spleen, findings did not support rapid
translocation from the lungs into systemic circulation.
On two separate occasions, subjects
inhaled 100 mL aerosol boli to target front
depths of 150 and 800 mL into the lungs to
target the airways and alveoli, respectively.
Retention measured at 10 min, 1.5, 5.5, 24
and 48 h post inhalation. Isotope (99mTc)
leaching from particles assessed via filters
in saline, blood, and urine. 81mKr utilized to
assess ventilation.
Shallow airways boli - Total deposition in airways (shallow boli)
similar between groups. Pattern of deposition was significantly
more central in the healthy subjects which was thought due to
non uniform ventilation distribution in smokers and COPD patients
as visualized by gamma-camera scans. Airway retention after 1.5
h was significantly lower in healthy subjects (89 ± 6%) than
smokers (97 ± 3%) or COPD patients (96 ± 6%). At 24 and 48 h,
retention significantly remained higher in COPD patients (86 ± 6%
and 82 ± 6%) than healthy subjects (75 ± 10% and 70 ± 9%).
Deep alveolar boli -Total deposition in alveoli (deep boli)
significantly greater in smokers (64 ±11 %) and COPD patients
(62 ± 5%) than healthy subjects (50 ± 8%). Alveolar retention of
particles similar at all times between groups. For example, at 48 h,
97 ± 3% in healthy subject, 96 ± 3% in smokers, and 96 ± 2% in
COPD patients. Retention at 24 and 48 correlated with isotope
leaching, suggesting that the small amount of clearance primarily
reflected the disassociation of 99mTc from the particles with little
transport of particles from the lungs.
Wiebert et al. Subjects having Carbon- Technegas system was modified to reduce
(2006,
156154)
varied health
status (9M, 6F;
46-74 yrs)
6 healthy
5 asthmatic
4 smokers
99mTc
87 nm
CMD
((Jb - 1.7)
Technegas
Generator
leaching of 99mTc radiolabel from
particles. The avg tidal volume during
aerosol inhalation was 1.8 L (range 0.8 -
3.3). Activity in chest region measured at
0, 2, 24, 46, and 70 h after inhalation.
Leaching assessed in vitro and via urine
collection.
Lung function not significantly different between healthy and
affected lungs. The aerosol deposition fraction was 41 ± 10%.
Lung retention was 99 ± 3%, 99 ± 5%, and 99 ± 10% at 24,46,
and 70 h post inhalation. Cumulative in vitro leaching by 70 h was
2.6 ± 0.96%. Except for radiotracer leaching from particles
(1.0 ± 0.6% of initially deposited activity in urine by 24 h), there
was not significant clearance from the lungs by 70 h. Individual
leaching was not correlated with individual retention.
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database
(Health and Environmental Research Online) at http://epa. gov/hero. HERO is a database of scientific literature
used by U.S. EPA in the process of developing science assessments such as the Integrated Science Assessments
(ISA) and the Integrated Risk Information System (IRIS).
July 2009
B-1
DRAFT-DO NOT CITE OR QUOTE

-------
Reference Study Group Aerosol
Study Protocol
Observations
Wiebert et al.
Healthy
Carbon -
(2006,
subjects (4M,
99mTc
157146)
5F; 56 ± 9 yrs)
34 nm

Asthmatics
CMD

(2M, 3F;
(tJB - 1.5)

59 ± 6 yrs)
Technegas

Control (1M;
Generator

50 yrs)

Slow deep aerosol inhalations with 10s
breath hold. Mean inhalation time of 6 min.
Control subject inhaled aerosol with loosely
bound radiolabel. Retention scans at 10
min, 60 min, 100 min, and 24 h post
inhalation. Leaching assessed in vitro and
via collection of blood and urine.
Avg deposition fraction of 60 ± 17% which was correlated with
tidal volume during aerosol inhalation (p - 0.01). Activity excreted
in urine over 24-h post inhalation was 51 % in the control subject
(high 99mTc disassociation) and 3.6 ± 0.9% of deposited activity.
In the blood of the control subject, activity was 30%, 31 %, and
5% of the deposited activity at 20 min, 80 min, and 24-h
(respectively), whereas it was only 0.9 ± 0.6%, 1.1 ± 0.4%, and
1.5 ± 0.5% the other 13 subjects at these times. Lung retention in
the control subject was 30% at 1-h and 18% at 24 h. In the
remainder of subjects, lung retention was approximately 100%
through 24 h.
Table B- 2. Ultrafine disposition in animals.
Reference Study Group
Aerosol
Study Protocol
Observations
Bermudez et
al. (2004,
189730)
Fischer 344 rats,
females (6 wks)
B3C3F1 mice,
females (6 wks)
Hamsters,
females (6 wks)
Ti02:1.29-1.44/ym
MMAD
(erg - 2.46-3.65), 21
nm primary particles
Animals exposed 6 h/day, 5
day/wk, for 13 weeks to 0.5, 2
and 10 mgfm3. Control animals
exposed to filtered air. Animals
sacrificed at 0, 4,13, 26, and 56
(49 for hamsters) postexposure.
Groups of 25 animals per species
and time point.
HO2 pulmonary retention half-times for the low-, mid-, and high-
exposure groups, respectively: 63,132, and 365 days in rats;
48, 40, and 319 days in mice; and 33, 37, and 39 days in
hamsters.
Burden of HO2 in lymph nodes increase with time postexposure
in mid- and high-dosed rats; in high-dosed mice; but was
unaffected in hamsters at any time or dosage group. In
high-exposure groups of mice, epithelial permeability remained
elevated (— 2 x control groups) out to 52 weeks without signs
of recovery. Epithelial permeability was 3-4x control in high
exposed rats through 4 weeks post exposure, but approached
control by 13 weeks. Epithelial permeability was unaffected in
all groups of hamsters.
Chen et al.
(2006,
087947)
Sprague-Dawley
male rats
(220 ± 20 g)
Polystyrene
1251 radiolabel
Ultrafine: 56.4 nm
Fine: 202 nm
Intratracheal instillation of
particles in healthy rats or those
pretreated with LPS (12 h before
particle instillation). Healthy rats
sacrificed between 0.5-2 h and at
24 or 48 h post-instillation. LPS
treated rats were sacrificed 0.5-2
h post-instillation.
In healthy rats, there were no marked differences in lung
retention or systemic distribution between the ultrafine and
fine particles. Results for healthy animals focused on ultrafine
particles which were primarily retained in lungs (72 ± 10% at
0.5-2 h; 65 ± 1 % at 1 day; 62 ± 5% at 5 days). Initially, there
was rapid particle movement into the blood (2 ± 1% at 0.5-2 h;
0.1 ± 0.1 % at 5 days) and liver (3 ± 2% at 0.5-2 h; 1 ± 0.1 %
at 5 days). At 1 day post-instillation, about 13% of the
particles where in the urine or feces. Following LPS treatment,
ultrafine accessed the blood (5 vs. 2%) and liver (11 vs. 4%) to
a significantly greater extent than fine particles.
Geiseretal. Wistar rats 20 Ti02 (22 nm CMD, 1.7 Rats exposed 1-h via endotracheal Distributions of particles among lung compartments followed
(2005,
087362)
Also included
in vitro study
adult males
(250 ± 10 g)
ag)
Spark generated
0.11 mg/m3
7.3 x 106
particles/cm3
tube while anesthetized and
ventilated at constant rate. Lungs
fixed at 1 or 24-h postexposure.
the volume distribution of compartments and did not differ
significantly between 1 and 24-h post-inhalation. On avg,
79.3 ± 7.6% of particles were on the luminal side of the
airway surfaces, 4.6 ± 2.6% in epithelial or endothelial cells,
4.8 ± 4.5% in connective tissues, and 11.3 ± 3.9% within
capillaries. Particles within cells were not membrane-bound.
Kapp et al.
(2004,
156624)
Charles River
rats
5 young adult
male
(250 ± 10 g)
Ti02 (22 nmCMD, 1.7
ag)
Spark generated
-Rats exposed 1-h via
endotracheal tube while
anesthetized and ventilated at
constant rate. Lungs fixed
immediately postexposure.
Of particles in tissues, 72% were aggregates of 2 or more
particles; 93% of aggregates were in round or oval shape
aggregates, 7% were needle-like. The size distribution of
particles in lung tissues (29 nm CMD, 1.7 erg) was remarkably
similar to the aerosol; the small discrepancy may have been due
to differences sizing techniques. A large 350 nm aggregate was
found in a type II pneumocyte, a 37 nm particle in a capillary
close to the endothelial cells, and a 106 nm particle within the
surface-lining layer close to the alveolar epithelium.
July 2009
B-2
DRAFT-DO NOT CITE OR QUOTE

-------
Table B- 3. In vitro studies of ultrafine disposition.
Reference Animal
Particles
Study Protocol
Observations
Edetsberger
et al. (2005,
155759)
Human cervix
carcinoma cells
(HeLa cells)
Polystyrene spheres
(0.020 |jm)
Cells incubated with polystyrene
particles having negative surface
charges. Cell cultures were naive or
treated with Genistein or Cytochalasin
B (CytB) prior to particle application.
Genistein inhibits endocytotic
processes, expecially caveolae
internalization. CytB inhibits actin
polymerization and phagocytosis.
Particles translocated into cells by first measurement (~ 1
min after particle application) independent of treatment
group. In naive cells, agglomerates of 88-117 nm were
seen by 15-20 minutes and of 253-675 nm by 50-60
minutes after particle application. Intracellular aggregates
thought to be result from particle incorporation into
endosomes or similar structures. In treated cells, only a
small number of agglomerates (161-308 nm) were found
and only by 50-60 minutes. At 50-60 minutes, 90% and
98% of particles were in the 20-40 nm range in naive and
treated cells, respectively. Particles did not translocate into
dead cells, rather they attached to outside of the cell
membrane.
Geiser et al.
(2005,
087362)
Also included
inhalation
study
Porcine lung
macrophages (106
cell/mLuman red
blood cells (RBC; 8
x 106 cells/mL)
Fluorescent
polystyrene spheres
(0.078, 0.2, and
1 Mm)
Gold shheres
(0.025 |jm)
Cells cultured for 4 h with each sized
polystyrene spheres. RBC were
employed as a model of nonphagocytic
cells. Some macrophages cultures were
treated with cytochalasin D (cytD) to
inhibit phagocytosis. In addition, RBC
were also cultured with gold particles.
Of the non-cytD treated macrophages, 77 ± 15%,
21 ±11%, and 56 ± 30% contained 0.078, 0.2, and
1 |jm particles, respectively. CytD treatment of
macrophages effectively blocked the phagocytosis of 1 |jm
particles, but did not alter the uptake of the 0.078 and
0.2 |jm particles. Human RBC were found to contain 0.078
and 0.2 |jm polystyrene spheres as well as the 0.025 |jm
gold particles, which were not membrane bound. In
contrast, the RBC did not contain the larger 1 |jm
polystyrene spheres. Results suggest that ultrafine and fine
(0.078 and 0.2 |jm diameter) particles cross cellular
membranes by a non-endocytic (i.e. not involving vesicle
formation) mechanisms such as adhesive interactions and
diffusion.
Geys et al.
(2006,
155789)
Human alveolar
(A549) and
bronchial (Calu-3)
epithelial cellsRat
primary type II
pneumocytes
Amine- and
carboxyl-modified
fluorescent
polystyrene (46 nm)
Cells cultured in clear polyester
transwells with 0.4 or 3 |jm pores.
Monolayer considered "tight" when
< 1 % sodium fluorescein moved from
apical to basolateral compartment.
Particle translocation assessed in
transwells with and without cells. Cells
incubated with particles for 14-16 h to
assess translocation from apical to
basolateral compartment.
Without cells, 13.5% of carboxyl-modified particles passed
through the 0.4 |jm pores (n - 7) and 67.5% through 3 |jm
pores (n - 3). Movement of the amine-modified particles
was 4.2% through 0.4 |jm pores (n - 7) and 52.7%
through 3 |jm pores (n - 3). The integrity of the monolayer
was insufficient for translocation studies using the A549
cells (0.4 and 4 |jm pore size) and rat pneumocytes
(0.3 |jm pore). Using 0.4 |jm pores, there was no
detectable translocation through either Calu-3 or rat
pneumocyte monolayers. Using 3 |jm pores, —6% of both
particle types passed through the Calu-3 monolayer;
however, results were highly variable with no translocation
in 2 (of 5) and 3 (of 6) trials with carboxyl- or
amine-modified particles, respectively.
July 2009
B-3
DRAFT-DO NOT CITE OR QUOTE

-------
B.2. Olfactory Translocation
Table B-4. Olfactory particle translocation.
Reference Study Group Aerosol
Study Protocol
Observations
DeLorenzo
(1970,
156391)
Squirrel monkeys Silver-coated colloidal
young males gold (50 nm)
(1 kg)
Intranasal instillation of 1 mL particle
suspension. Animals sacrificed at 0.25,
0.5,1, and 24-h after instillation.
Rapid movement (30-60 min) into olfactory bulbs. Within
30 min of being placed on nasal mucosa, particle
aggregates were seen in axoplasm of the fila olfactoria
Within 1 h, particles were in olfactory glomerulus.
Particles in the olfactory bulb were located preferentially
in mitochondria and not free in the cytoplasm.
Dorman et
al. (2001,
055433)
Crl: CD rats
Males (6 wks
old)
Soluble and insoluble
Mn particle types;
MMAD - 1.3-2.1 fjm;
GSD<2
Whole body exposure (6 h/day, 14
consecutive days) to 0, 0.03, 0.3, and
3 mg Mn/m3. Tissues analyzed in six
animals per concentration exposed to
soluble (MnS04) or insoluble (Mn304)
aerosols.
Increased Mn levels in olfactory bulb observed following
MnS04 of a 0.3 mg Mn/m3 and following Mn304 of 3 mg
Mn/m3. At 3 mg Mn/m3, Mn levels were significantly
greater in olfactory bulb (1.4-fold) and striatum (2.7-fold)
following soluble Mn04 than insoluble Mn304. Mn levels in
the cerebellum were unaffected following all exposures.
Dorman et
al. (2004,
155752)
Crl: CD rats
Males (6 wks
old)
Soluble and insoluble
Mn particle types;
MMAD - 1.5-2/ym;
GSD - 1.4-1.6
Whole body exposure (6 h/day, 5
days/wk, 13 wks) to MnS04 at 0, 0.01,
0.1, and 0.5 mg Mn/m3. Compared to
Mn phosphate (as hureaulite) exposure
of 0.1 mg Mn/m3. Brain Mn levels
assessed immediately following 90
days of exposure or 45 days
postexposure.
Relative to air, the insoluble hureaulite was significantly
increased at 90 days of exposure in the olfactory bulb, but
not striatum or cerebellum. The soluble Mn phosphate
showed a dose dependent increase in olfactory bulb Mn
levels at 90 days. At 0.1 mg Mn/m3, Mn levels following
Mn phosphate were significantly increased in the olfactory
bulb and striatum relative to hureaulite and air exposures.
At 45 days postexposure, relative to air, olfactory bulb Mn
levels only remained increased Mn phosphate group at 0.5
mg Mn/m3.
Elder et al.
(2006,
089253)
Fisher 344 rats
Males (200-250
Mn oxide ( — 30 nm
equivalent sphere with
3-8 nm primary
particles)
Spark generated
0.5 mg/m3
18 x 10e particles/cm3
Whole body inhalation exposure to
either filtered air or Mn oxide for 12
days (6 h/day, 5 days/wk) with both
nares open or Mn oxide for 2 days (6
h/day) with right nostril blocked.
Intranasal instillation in left nostril of
Mn oxide particles or soluble MnCl2
suspended in 30 |jL saline. Analyzed
Mn in the lung, liver, olfactory bulb, and
other brain regions.
After 12 day exposure via both nostrils, Mn in the
olfactory bulb increased 3.5-fold, whereas in the lung Mn
concentrations doubled; there were also increases in the
striatum, frontal cortex, and cerebellum. After the 2 days
exposure with the right nostril blocked, Mn accumulated in
the mainly in the left olfactory bulb (— 2.4-fold increase)
in to a lesser extent in the right olfactory bulb (1.2-fold
increase). At 24-h post instillation, the left olfactory bulb
contained similar amounts of the poorly soluble Mn oxide
(8.2 ± 0.7%) and soluble MnCh (8.2 ± 3.6%) as a
percent of the amount instilled.
Oberddrster
et al. (2004,
055639)
Fisher 344 rats
Males (14 wks;
284 ± 9 q)
,3C (36 nmCMD, 1.7
CTg)
Spark generated
Rats (n - 12, 3 per time point) exposed
to 160 jjg/m3 for 6 h in whole-body
chamber and sacrificed at 1, 3, 5, and 7
day postexposure. Lung, olfactory bulb,
cerebrum, and cerebellum removed for
,3C analysis. Tissue ,3C-levels were
determined by isotope ratio mass
spectroscopy and background corrected
for ,3C levels in unexposed controls
(n - 3).
At 1 day postexposure, the lungs of rats exposed to
ultrafine ,3C particles contained 1.34 ± 0.22 jjg of ,3C
(1.39 jjg/g-lung) following background corrected. By 7
days postexposure, the 13C concentration had decreased
to 0.59 jjg/g-lung. There was a significant and persistent
increase in 13C in the olfactory bulb of 0.35 jjg/g on day
1, which increased to 0.43 jjg/g by day 7. Day 1
concentrations of 13C in the cerebrum and cerebellum
were also significantly increased but the increase was
inconsistent, possibly reflecting translocation of particles
from the blood across the blood-brain barrier into brain
regions.
Persson et Sprague-Dawley "'ZnCly dissolved in 0.1 Rats: intransal (0.03//g Zn in 10//L) or Zn uptake in olfactory epithelium and transport along
al. (2003,
051846)
male rats (150 g) M HCI
Freshwater Pike
female (3 kg)
intraperitoneally (0.03 /yg Zn in
100 //L); autoradiography and y spec
at 1 day or 1, 3, or 6 weeks
postexposure.
Pike: instilled (0.12 /yg Zn in 10 //L) in
right or both olfactory chambers,
assayed 2 weeks postexposure
olfactory neurons to olfactory bulb. Zn continued into
interior of olfactory bulb and in rat went into anterior
olfactory cortex. Zn found bound to both cellular
constituents and cytosolic components. Some Zn bound to
metallothionein in olfactory mucosa and olfactory bulb.
July 2009	B-4	DRAFT - DO NOT CITE OR QUOTE

-------
Reference Study Group Aerosol
Study Protocol
Observations
Wang et al.
(2007,
156147)
CD-1 (ICR) mice
Rutile Ti02
21 and 80 nm
Anatase Ti02 155 nm
Twenty mice (n - 5 per group) exposed Rutile particles were observed to be column/fiber shaped,
0 or 0.01 g-Ti02 per mL Dl. Instilled whereas anatase was octahedral. Ti02 particles taken up
25 /jL each day for 5 days, then inhaled by olfactory bulb via the olfactory nerve layer, olfactory
10 //L every other day. Mice sacrificed ventricle, and granular cell layer of the olfactory bulb. Fine
after one month.	Ti02 showed greater entry into the olfactory bulb
presumably due to aggregation of smaller rutile particles
that was not seen for the fine anatase particles.
Yu et al. Sprague-Dawley Stainless steel
(2003, male rats, 6 wks welding-fume
156171) old (218 ± 10 g) <0 5/ym
Whole body exposure 2 h/day for 1,15,
30, or 60 days
Low: 64 ± 4 mg/m3 (1.6 mg/m3 Mn)
High: 107 ± 6 mg/m3 (3.5 mg/m3 Mn)
Significant increases in cerebellum Mn at 15 - 30 days of
exposure.
Slight increases in Mn in substantia nigra, basal ganglia,
temporal cortex, and frontal cortex after 60 days.
Significant increase at 30 days in basal ganglia at low
dose. Authors suggested that pharmacokinetics and
distribution of welding fume Mn differs from pure Mn.
B.3. Clearance and Age
Table B-5. Studies of respiratory tract mucosal and macrophage clearance as a function.
Reference Animal
Particles
Study Protocol
Observed Effect(s)
NASAL AND TRACHEAL CLEARANCE
Ho et al. Human, Not applicable
(2001, males and
156549) females
Ninety subjects (47 M, 43 F; 52 ± 23 yrs)
between 11 and 90 years of age were
recruited to measure nasal saccharine
clearance and ciliary beat frequency.
Ciliary beat frequency (n - 90; r - -0.48, p
< 0.0001) and nasal mucociliary clearance
time (n - 43; r - 0.64,p < 0.001) were
correlated with subject age. Nasal clearance
times were significantly \p < 0.001) faster in
individuals under 40 years of age (9.3 ± 5.2
min) versus older subjects (15.4 ± 5.0
minutes). Results similar between males and
females.
Goodman et
al. (1978,
071130)
Humans,
males and
females
Radiolabed Teflon disks (1 mm
diameter, 0.8 mm thick)
Tracheal mucus velocity following delivery via
bronchoscope to the tracheal mucosa. Ten
young (2 M, 8 F; 23 ± 3 yrs) and ten elderly (2
M, 5 F; 63 ± 5 yrs) nonsmokers served as
control subjects. Measurements were also
made in young smokers, ex-smokers, and
individuals with chronic bronchitis.
Young nonsmokers had a tracheal mucus
velocity of 10.1 ± 3.5 mm/min which was
significantly faster than the velocity of
5.8 ± 2.6 observed in the elderly nonsmokers.
Whaley et al.
(1987,
156153)
Beagle Macroaggregated albumin99mTc
labelled
males and
females
Intratracheal instillation of 10- //I droplet of
labelled albumin in saline. Tracheal clearance
followed 25 minutes. Longitudinal measure
measurements in 5 males and 3 females when
young adults (2.8-3 yr), middle-aged (6.7-6.9
yr), and mature (9.6-9.8 yr). Additional 5
females and 3 males comprised immature
group (9-10 mo) and 4 males and 4 females
used as aged group (13-16 yr).
Tracheal mucus velocity significantly
\p < 0.05) greater in young (9.7 ± 0.6 [SE]
mm/min) and middle-aged (6.9 ± 0.5) groups
than in immature (3.6 ± 0.4), mature
(3.5 ± 0.8), and aged (2.9 ± 0.5) dogs.
July 2009
B-5
DRAFT-DO NOT CITE OR QUOTE

-------
Reference Animal
Particles
Study Protocol
Observed Effect(s)
Yeates et al. Humans, Radioaerosols 99mTc labelled
(1981, males and
095391) females
Tracheal mucus velocities compiled for 74
healthy non-smoking subjects (60 M, 14F;
10-65 yrs, mean 30 yrs) from prior studies.
Forty-two (32 M, 10 F) inhaled albumin in
saline droplets (6.2-6.5//m MMAD), Yeates et
al. (1975); twenty-two (21 M, 1 F) inhaled iron
oxide (4.2 /jm MMAD), Yeates et al. (1981 b);
and ten (7 M, 3 F) inhaled monodisperse iron
oxide aerosol (7.5 /jm MMAD), Leikauf et al.
(1981). Inhalations were via a mouthpiece with
an inspiratory flow of ~ 1 liter/sec.
A lognormal distribution of tracheal mucus
velocities was reported. Age did not appear to
affect velocities, e.g., 4.7 ± 2.5 mm/min in
18-24 yrs olds vs. 4.6 ± 3.2 mm/min in
individuals > 30 yrs of age. However, it
should be noted that only 2 subjects were
greater than 45 yrs of age and that the data
was complied from three studies using
differing experimental techniques. Rather
similar tracheal mucus velocities in males
(4.7 ± 3.0 mm/min) and females (4.9 ± 2.4
mm/min).
BRONCHI AND BRONCHIOLES CLEARANCE
Puchelle et
al. (1979,
006863)
Human,
males
7.4 fm MMAD99mTc labelled
resin
Mucociliary clearance measured for 1 h post
aerosol inhalation in 19 healthy non-smoking
males (21-69 yrs of age). Clearance measure
on two occasions in 16 individuals.
Negative correlation (r - -0.472,/; < 0.05)
between mucociliary clearance and age.
Younger subjects (n - 9; 21-37 yrs) had 1-h
clearance of 34 ± 14% which was
significantly greater than the 22 ± 8% found
in the older subjects (n - 5; > 54 yrs).
Separated by 5.4 wks (on avg), there was a
good correlation between repeated clearance
measurements (r - 0.65,/; <0.001)
Svartengren
et al. (2005,
157034)
Humans,
males and
females
6//m MMAD111ln labelled Teflon
Small airway clearance measured in five age
groups (~ 24 yrs, n - 13; 25-29 yrs, n - 8;
30-49 yrs, n - 7; 50-64, n - 9; > 65 yrs,
n - 9) of healthy subjects. Aerosol inhaled via
mouthpiece at extremely slow rate of 0.05 L/s.
Activity in lungs measured at 1 day, 2 days,
and 1, 2, and 3 wks post-exposure. Under the
presumption that most large airway clearance
was complete by 24 h, retention at 24 h was
normalized to 100%.
Large and small airway clearance slowed with
increasing age. Clearance correlated with age
at all times (r - -0.46 to -0.50, -0.55, -0.66,
and-0.70 at 1 day, 2 days, 1 wk, 2 wks, and
3 wks, respectively). Based on linear
regression, the clearance from 1 to 21 days
post-exposure was 47% in a 20 yr-old versus
23% in an 80 yr-old. Lung function was not a
significant predictor of clearance when age
considered.
Vastag et al. Humans, Monodisperseerythrocytes99mTc
(1985, males and labelled
157088) females
Clearance measured for 1-h post-inhalation in	Clearance significantly associated with age.
eighty healthy (59 M, 21 F; 43 ±17 yrs)	Based on linear regression, total mucociliary
subjects who had never smoked. Smokers and	clearance at 1-h post-exposure was 46% in a
ex-smokers also studied. Aerosol inhalation not	20-yr-old versus 23% in an 80 yr-old. Similar
described.	results for males and females.
ALVEOLAR CLEARANCE
Muhle et al.
(1990,
006853)
Fischer 3.5//m MMAD 85Sr labelled
344 rats polystyrene latex
Control animals compared across several	Typical alveolar clearance half-time of 45
studies. Aerosol inhaled by short-term nose	days in 5-month-old rats compared to 74 days
only exposure. Alveolar clearance determined	in 23-month-old rats. Statistical significance
by exponential fit to thoracic activity measured	of findings not proved,
over 75-100 days excluding the first 15 days
post-exposure.
July 2009	B-6	DRAFT - DO NOT CITE OR QUOTE

-------
Annex B References
Bermudez E; Mangum JB; Wong BA; Asgharian B; Hext PM; Warheit DB; Everitt JI. (2004).
Pulmonary Responses of Mice, Rats, and Hamsters to Subchronic Inhalation of Ultrafine
Titanium Dioxide Particles. , 77- 347-357. 189730
Chen OH; Xirasagar S; Lin H"C. (2006). Seasonality in adult asthma admissions, air pollutant levels,
and climate- a population-based study. J Asthma, 43- 287-292. 087947
DeLorenzo AJD. (1970). The olfactory neuron and the blood-brain barrier. In Taste and Smell in
Vertebrates (pp. 151-175). London- Churchill Livingstone. 156391
Dorman DC; McManus BE; Parkinson CU; Manuel CA; McElveen AM; Everitt JI. (2004). Nasal
Toxicity of Manganese Sulfate and Manganese Phosphate in Young Male Rats Following
Subchronic (13-Week) Inhalation Exposure. Inhal Toxicol, 16: 481-488. 155752
Dorman DC; Struve MF; James RA; Marshall MW; Parkinson CU; Wong BA. (2001). Influence of
particle solubility on the delivery of inhaled manganese to the rat brain: manganese sulfate and
manganese tetroxide pharmacokinetics following repeated (l4"day) exposure. Toxicol Appl
Pharmacol, 170: 79-87. 055433
Edetsberger M; Gaubitzer E; Valic E; Waigmann E; Kohler G. (2005). Detection of nanometer-sized
particles in living cells using modern fluorescence fluctuation methods. Biochem Biophys Res
Commun, 332: 109-116. 155759
Elder A; Gelein R; Silva V; Feikert T; Opanashuk L; Carter J; Potter R; Maynard A; Ito Y; Finkelstein
J; Oberdorster G. (2006). Translocation of inhaled ultrafine manganese oxide particles to the
central nervous system. Environ Health Perspect, 114: 1172-1178. 089253
Geiser M; Rothen-Rutishauser B; Kapp N; Schurch S; Kreyling W', Schulz H; Semmler M; Im Hof V;
Heyder J; Gehr P. (2005). Ultrafine particles cross cellular membranes by nonphagocytic
mechanisms in lungs and in cultured cells. Environ Health Perspect, 113: 1555-1560. 087362
Geys J; Coenegrachts L; Vercammen J; Engelborghs Y; Nemmar A; Nemery B; Hoet PHM. (2006). In
vitro study of the pulmonary translocation of nanoparticles A preliminary study. Toxicol Lett,
160: 218-226. 155789
Goodman RM; Yergin BM; Landa JF; Golinvaux MH; Sackner MA. (1978). Relationship of smoking
history and pulmonary function tests to tracheal mucous velocity in nonsmokers, young
smokers, ex-smokers, and patients with chronic bronchitis. Am Rev Respir Dis, 117: 205-214.
071130
Ho JC; Chan KN; Hu WH; Lam WE Zheng L; Tipoe GL; Sun J; Leung R; Tsang KW. (2001). The effect
of aging on nasal mucociliary clearance, beat frequency, and ultrastructure of respiratory cilia.
Am J Respir Crit Care Med, 163: 983-988. 156549
Kapp N; Kreyling W; Schulz H; Im Hof V; Gehr P; Semmler M; Geiser M. (2004). Electron energy loss
spectroscopy for analysis of inhaled ultrafine particles in rat lungs. Microsc Res Tech, 63: 298"
305. 156624
Mills NL; Amin N; Robinson SD; Anand A; Davies J; Patel D; de la Fuente JM; Cassee FR; Boon NA;
Macnee W; Millar AM; Donaldson K; Newby DE. (2006). Do inhaled carbon nanoparticles
translocate directly into the circulation in humans?. Am J Respir Crit Care Med, 173: 426-431.
088770
Muhle H; Creutzenberg 0; Bellmann B; Heinrich U; Mermelstein R. (1990). Dust overloading of lungs:
investigations of various materials, species differences, and irreversibility of effects. , l: Slll"
S128. 006853
Oberdorster G; Sharp Z; Atudorei V; Elder A; Gelein R; Kreyling W; Cox C. (2004). Translocation of
inhaled ultrafine particles to the brain. Inhal Toxicol, 16: 437-445. 055639
Persson E; Henriksson J; Tallkvist J; Rouleau C; Tjalve H. (2003). Transport and subcellular
distribution of intranasally administered zinc in the olfactory system of rats and pikes.
Toxicology, 191: 97-108. 051846
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database
(Health and Environmental Research Online) at http://epa. gov/hero. HERO is a database of scientific literature
used by U.S. EPA in the process of developing science assessments such as the Integrated Science Assessments
(ISA) and the Integrated Risk Information System (IRIS).
July 2009
B-7
DRAFT-DO NOT CITE OR QUOTE

-------
Puchelle E> Zahm J-M; Bertrand A. (1979). Influence of age on bronchial mucociliary transport. , 60-
307-313. 006863
Svartengren Mi Falk Ri Philipson K. (2005). Long-term clearance from small airways decreases with
age. Eur Respir J, 26- 609-615. 157034
Vastag Ei Matthys Hi Kohler Di Gronbeck Li Daikeler G. (1985). Mucociliary clearance and airways
obstruction in smokers, ex-smokers and normal subjects who never smoked. , 139- 93-100.
157088
Wang C. (2007). Impact of direct radiative forcing of black carbon aerosols on tropical convective
precipitation. Geophys Res Lett, 34' 5709. 156147
Whaley SLi Muggenburg BAi Seiler FA; Wolff RK. (1987). Effect of aging on tracheal mucociliary
clearance in beagle dogs. J Appl Physiol, 62: 1331-1334. 156153
Wiebert P; Sanchez-Crespo A; Falk R; Philipson K> Lundin A; Larsson Si Moller Wi Kreyling Wi
Svartengren M. (2006). No Significant Translocation of Inhaled 35-nm Carbon Particles to the
Circulation in Humans. Inhal Toxicol, 18: 741-747. 156154
Wiebert Pi Sanchez-Crespo Ai Seitz Ji Falk Ri Philipson Ki Kreyling WGi Moller Wi Sommerer Ki
Larsson Si Svartengren M. (2006). Negligible clearance of ultrafine particles retained in healthy
and affected human lungs. Eur Respir J, 28: 286-290. 157146
Yeates DBi Gerrity TRi Garrard CS. (1981). Particle deposition and clearance in the bronchial tree.
Ann Biomed Eng, 9: 577-592. 095391
Yu IJi Park JDi Park ES; Song KS; Han KT; Han JH; Chung YH; Choi BS; Chung KH; Cho MH. (2003).
Manganese Distribution in Brains of Sprague—Dawley Rats After 60 Days of Stainless Steel
Welding-Fume Exposure. , 24: 777-785. 156171
July 2009
B-8	DRAFT-DO NOT CITE OR QUOTE

-------
Annex C. Controlled Human
Exposure Studies
Table C-1.
Cardiovascular Effects

Study
Pollutant Exposure
Findings
Reference:
Barregard et al.
Wood smoke
Particle Size: Session 1:
(2006, 091381) GMD 42 nm; Session 2.
Subjects: 13
healthy adults
Gender: 6 Ml 7
F
Age: 20-56 yrs
GMD 112 nm
Particle Number/Count:
Session 1: 180,000/cm3;
Session 2: 95,000/cm3
Concentration: Session
1: median: 279 //g|m3;
Session 2: median
243/yg/m3
Subjects exposed in two groups for 4 h to
filtered air, followed a wk later by a 4-h exposure
to wood smoke. Exposures conducted with two
25-min periods of light exercise. Other measured
combustion products:
Session 1: NO2 (0.08 ppm), CO (13 ppm),
formaldehyde (114/yg/m3), acetaldehyde
(75 /yglm3), benzene (30 /yglm3), 1,3-butadiene
(6.3/yg/m3);
Session 2: NO2 (0.09 ppm), CO (9.1 ppm),
formaldehyde (64/yg/m3), acetaldehyde
(40 /yglm3), benzene (20 /yglm3), 1,3-butadiene
(3.9/yg/m3).
Time to analysis: Immediately following
exposure as well as 3 and 20 h post-exposure.
Statistically significant increase in plasma factor VIII 20 h
post wood smoke exposure relative to filtered air. The factor
Vlllfvon Willebrand ratio in plasma was increased with wood
smoke relative to filtered air at 0, 3, and 20 h post-exposure.
Wood smoke exposure increased the urinary excretion of free
8-iso-prostaglandin2n relative to clean air 20 h post-exposure
(n - 9). These findings were more pronounced in session 1
than session 2 (similar mass concentration but higher number
concentration in session 1).
Reference:
Beckett et al.
(2005,156261)
Subjects: 12
healthy adults
Gender: 6 M/6
F
Age: 23-52 yrs
Ultrafine and fine zinc
oxide
Particle Size: UF:
< 0.1 /jm; Fine: 0.1-
1.0 /jm
Particle Number/Count:
UF: 4.6 x 107/cm3; Fine:
1.9 x 106|cm3
Concentration:
500/yg/m3
Subjects exposed via mouthpiece for 2 h during
rest to filtered air, ultrafine, and fine zinc oxide in
a randomized crossover study design. Exposures
were separated by at least 3 wks.
Time to analysis: Immediately following
exposure and 3, 6,11, 23, and 24 h after
exposure.
Exposure to ultrafine and fine zinc oxide did not affect HRV
(time and frequency domain parameters) relative to clean air
immediately following exposure, or at 3, 6,11, and 23 h
post-exposure. Exposure did not affect blood pressure
through 24 h post-exposure. No effects of exposure to either
fine or ultrafine zinc oxide observed on factor VII, von
Willebrand factor (vWf), tissue plasminogen activator (t-PA),
or fibrinogen. No effect of exposure observed on peripheral
blood cell counts or levels of pro-inflammatory cytokines.
DE
Concentration:
Reference:
Blomberg et al.
(2005.191991) 300/yglm3
Subjects: 15
older adults
(former
smokers) with
COPD
Age: 56-72 yrs
Subjects exposed for 1 h with intermittent
exercise to DE and filtered air in a randomized
crossover study design.
Time to analysis: 6 and 24 h post-exposure.
DE was not observed to affect blood levels of C-reactive
protein, fibrinogen, D-Dimer, prothrombin factor 1-2, or von
Willebrand factor activity at 6 and 24 h post-exposure.
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database
(Health and Environmental Research Online) at http://epa. gov/hero. HERO is a database of scientific literature
used by U.S. EPA in the process of developing science assessments such as the Integrated Science Assessments
(ISA) and the Integrated Risk Information System (IRIS).
July 2009
C-1
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Findings
Reference:
Brauner et al.
(2007,188507)
Subjects: 29
healthy adults
Gender: 20
M/9 F
Age: 20-40 yrs
Urban traffic particles
Particle Number/Count:
6-700nm: 10,067|cm3
Concentration: PM2.6:
9.7/yg/m3; PM2.6 -10:
12.6/yg/m3
Subjects exposed to urban traffic particles and
filtered air for 24 h with and without two 90-min
periods of light exercise in a randomized
crossover study design. Concentrations of NOx
and NO were low and did not differ between
filtered and unfiltered exposures. CO
concentrations were higher with filtered air (0.35
and 0.41 ppm), while ozone concentrations were
lower with filtered air (12.08 and 4.29 ppb).
Time to analysis: 6 and 24 h after the start of
exposure.
An increase in DNA strand breaks and
formamidopyrimidine-DNA glycosylase sites in peripheral
blood mononuclear cells were observed after 6 and 24 h of
exposure to urban particulates. The particle concentration at
the 57nm mode was shown to be the major contributor to
these effects.
Reference:	Indoor air particles
Brauner et al.	partic|e Number/Count:
(2008,156293)	^0.700 nm. 0 016)cm3
Subjects. 42	Concentration: Coarse:
healthy older	g.4/yg/m3. Fine:
adults (21	12.6/yg/m3
couples)
Age: 60-75 yrs
Exposures consisted of two 48 h periods in the
home of each subject with or without the use of
a HEPA filter (randomized crossover design).
HEPA filters reduced coarse concentration from
9.4 to 4.6/yglm3, and fine concentration from
12.6 to 4.7 /yglm3. Concentrations of NO2 did
not differ between the 2 sessions (20 ppb).
Time to analysis: After the completion of each
48 h session.
The use of HEPA filters significantly improved microvascular
function (p - 0.04) after 48 h (reactive hyperemia-peripheral
arterial tonometry). Microvascular function was assessed
using a scoring system representing the extent of reactive
hyperemia. The reduction in PM concentration through the
use of HEPA filters did not significantly affect blood pressure
following the 48-h exposures. Lowering PM concentration did
not significantly affect inflammatory response markers in
peripheral venous blood (IL-6, TNF-n, C-reactive protein,
plasma amyloid A).
Urban traffic particles
Particle Number/Count:
Reference:
Brauner et al.
(2008,191966) 6.700 nm. 0,067|cm3
Subjects. 29 Concentration: PM2.6:
healthy adults g.7/yg,m3. PM26 ,o:
Gender: 20 M, 12.6/yglm3
9 F
Age: M avg 27
yrs, F avg 26
yrs
Subjects exposed to urban traffic particles and
filtered air for 24 h with and without two 90-min
periods of light exercise in a randomized
crossover study design. Concentrations of NOx
and NO were low and did not differ between
filtered and unfiltered exposures. CO
concentrations were higher with filtered air (0.35
and 0.41 ppm), while ozone concentrations were
lower with filtered air (12.08 and 4.29 ppb).
Time to analysis: 6 and 24 h after the start of
exposure.
Exposure to urban traffic particles was not observed to
affect microvascular function (digital peripheral artery tone)
at 6 or 24 h after the start of exposure. No difference in
various blood markers of coagulation, inflammation, or
protein oxidation (e.g., fibrinogen, platelet count, CRP, IL-6,
TNF- ~) were demonstrated between particle and filtered air
exposure.
DE
2002 Cummins B-series
(2007,155714) diese| engine (6BT5 gG6
Reference:
Carlsten et al.
Subjects: 13
healthy adults
Gender: 11
M/2 F
Age: 20-42 yrs
5.9 L) operating at load
Concentration: Fine PM:
100, 200/yg/m3
Subjects exposed for 2 h at rest to filtered air
and each of the two DEPs concentrations in a
randomized crossover study design. Exposures
were separated by at least 2 wks. Other diesel
emissions measured: NO2 (10-35 ppb), CO (0.7-
1.8 ppm).
Time to analysis: 3, 6, and 22 h after the start
of exposure.
No statistically significant changes in plasminogen activator
inhibitor-1 (PAI-1), vWf, D-dimer, or platelet count observed
3, 6, or 22 h following exposure to DE relative to filtered air.
Non-statistically significant increases in D-dimer, vWf, and
platelet count was observed at 6 h following the start of
exposure (4 h post-exposure). No diesel-induced increase in
C reative protein observed in relative to filtered air in
peripheral venous blood at 1 or 20 h post-exposure.
Reference:
Carlsten et al.
DE
2002 Cummins B-series
(2008,156323) diese| engine (6BT5 gG6
Subjects: 16
adults with
metabolic
syndrome
Gender: 10
M/6 F
Age: 25-48 yrs
5.9 L)
Concentration: Fine PM:
100, 200/yg/m3
Subjects exposed for 2 h at rest to filtered air
and each of the two DE particle concentrations
in a randomized crossover study design.
Exposures were separated by at least 2 wks.
Other diesel emissions measured: NO2 (30 ppb),
NO (1.69 ppm), CO (0.65 ppm).
Time to analysis: 3, 7, and 22 h after the start
of exposure.
At 5 h after the end of diesel exposure (fine particulate
concentration 200 //gfm3), the authors observed a significant
decrease in vWf in peripheral venous blood. No other changes
in thrombotic markers (vWf, D-dimer, PAI-1) were observed at
either concentration between 1 and 20 h post-exposure.
July 2009	C-2	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Findings
Reference:
Danielsen et al.
(2008,156382)
Subjects: 13
healthy adults
Gender: 6 Ml 7
F
Age: 20-56 yrs
Wood smoke
Particle Size: Session 1:
GMD 42 nm; Session 2:
GMD 112 nm
Particle Number/Count:
Session 1: 180,000/cm3;
Session 2: 95,000/cm3
Concentration: Session
1: median: 279 /yg/m3;
Session 2: median
243/yg/m3
Subjects exposed in two groups for 4 h to	Exposure to wood smoke increased the mRNA levels of
filtered air, followed a wk later by a 4-h exposure hOGGI in PBMCs relative to filtered air 20 h after exposure.
to wood smoke. Exposures conducted with two
25-min periods of light exercise. Other measured
combustion products:
Session 1: NO2 (0.08 ppm), CO (13 ppm),
formaldehyde (114/yg/m3), acetaldehyde
(75 /yg/m3), benzene (30 /yg/m3), 1,3-butadiene
(6.3/yg/m3);
Session 2: NO2 (0.09 ppm), CO (9.1 ppm),
formaldehyde (64/yg/m3), acetaldehyde
(40 /yg/m3), benzene (20 /yg/m3), 1,3-butadiene
(3.9/yg/m3).
Time to analysis: 3 and 20 h post-exposure.
DNA strand breaks were shown to decrease in PBMCs 20 h
after wood smoke exposure.
Fine CAPs (Chapel Hill, NC) Exposures conducted for 2 h at rest to filtered CAPs exposure resulted in statistically significant reductions
Reference:
Devlin et al.
(2003, 087348) 40.5/yg/m3, Range: desi9
Concentration: Mean:
air and CAPs in a randomized crossover study
n.
Subjects: 10
healthy older
adults
Gender: 7 Ml 3
F
Age: Avg 66.9
yrs
21.2-80.3 /yg/m3	Time to analysis: Immediately following
exposure and 24 h post-exposure.
(p < 0.05) in time domain (PNN50) and frequency domain
(HF power) parameters relative to clean air immediately
following exposure. These relative decreases were still
apparent 24 h after exposure (p < 0.08).
Reference:
Fakhri et al.
Fine CAPs (Toronto)
Concentration: 127 ±
(2009,191914) 62/yg/m3 with and
Subjects: 50
adults (40
healthy, 10
asthmatic)
Gender: 24
M/26 F
Age: 19-48 yrs
without co-exposure to
ozone (114 ± ppb)
Exposures conducted through a facemask which
covered the subject's nose and mouth. Subjects
were exposed to CAPs, ozone, CAPs + ozone
and filtered air for 2 h at rest in a randomized
crossover study design.
Time to analysis: Every 30 min during
exposure, with the final measurement made
immediately prior to the end of the exposure.
Exposure to CAPs or ozone, alone or in combination, resulted
in no significant changes in HRV or blood pressure relative to
filtered air. However, a negative concentration response
relationship was reported between CAPs concentration with
co-exposure to ozone and SDNN, rMSSD, HF power and LF
power (statistically significant for LF power). Diastolic blood
pressure was observed to increase with exposure to CAPs +
ozone, but not with either pollutant alone. There was no
difference in response between asthmatics and healthy
subjects.
Reference:
Frampton et al.
(2006, 088665)
Subjects: 16
asthmatic
adults, 40
healthy adults
Gender:
Asthmatics: 8
M/8 F, Healthy:
20 M/20 F
Age: 18-40 yrs
Ultrafine elemental carbon
Particle Size: CMD —25
nm
Particle Number/Count:
10/yg/m3: -2.0 x
106/cm3; 25/yg/m3: -7.0
x 106/cm3; 50/yg/m3:
-10.8 x 106/cm3
Concentration: 10, 25,
and 50 /yg/m3
Study conducted using a randomized crossover
design with 2-h exposures. Asthmatics (n - 16)
exposed to filtered air and 10 /yg/m3. 12 healthy
adults exposed to filtered air and 10 /yg/m3 at
rest; 12 healthy adults exposed to filtered air, 10
and 25 /yg/m3 with intermittent exercise: 16
healthy adults exposed to filtered air and
50/yg/m3 with intermittent exercise. Exposures
were conducted via mouthpiece.
Time to analysis: Immediately following
exposure as well as 3.5, 21, and 45 h post-
exposure.
No effect of ultrafine particle exposure on leukocyte counts
or leukocyte expression of adhesion molecules observed in
healthy subjects exposed at rest to 10/yg/m3. Among healthy
adults exposed to ultrafine carbon during exercise, monocyte
expression of adhesion molecules CD54 and CD18 decreased
relative to filtered air immediately following exposure. An
ultrafine particle-induced decrease in PMN expression of
CD18was also observed 0-21 h post-exposure. Expression of
CD11b on monocytes and eosinophils was reduced following
exposure to ultrafine particles in exercising asthmatics
0-21 h post-exposure. A decrease in total leukocyte count
was observed following ultrafine particle exposure in
exercising healthy and asthmatic subjects.
July 2009	C-3	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Findings
Reference:
Gong et al.
(2004, 087964)
Subjects: 13
older adults
with COPD, 6
healthy older
adults
Gender: COPD:
5 Ml 8 F,
Healthy: 2 M/4
F
Age: COPD: avg
68 yrs, Healthy:
avg 73 yrs
Fine CAPs (Los Angeles)
Particle Size: 85% of
mass between 0.1 and
2.5 /jm
Concentration: Mean:
194/yg/m3,
135-229 //g/m3
Exposures to CAPs and filtered air (randomized
crossoveronducted for 2 h with intermittent light
exercise (four 15-min periods). Exposures were
separated by at least 2 wks.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post exposure.
SDNN shown to decrease following CAPs exposure relative
to filtered air in healthy older adults (4-22 h post-exposure).
No CAPs-induced changes in HRV were observed in older
adults with COPD. Ectopic heart beats were observed to
increase slightly with CAPs relative to filtered air among
healthy subjects, but decreased among subjects with COPD.
Exposure to CAPs did not affect platelet or white blood cell
count, or levels of fibrinogen, vWF, or factor VII.
Reference:
Gong et al.
(2004, 055628)
Subjects: 12
adult
asthmatics, 4
healthy adults
Gender:
Asthmatics: 4
M/8 F, Healthy:
2 M/2 F
Coarse CAPs (Los Angeles) Exposures to CAPs and filtered air (randomized SDNN shown to decrease following CAPs exposure relative
Particle Size: 80% of
mass between 2.5 and
10 /jm, 20% of mass
<2.b/jm
Concentration: Mean:
157/yg|m3,
56-218 //g/m3
crossoveronducted for 2 h with intermittent light to filtered air in healthy adults (4-22 h post-exposure).
exercise (four 15-min periods). Exposures were
separated by at least 2 wks.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post-exposure.
Decrease in PNN50 also observed in healthy adults at 4 h
post-exposure. No CAPs-induced decreases in HRV
demonstrated in asthmatics.
Asthmatics: avg
38 yrs, Healthy:
avg 32 yrs
Reference:
Gong et al.
(2008,156483)
Subjects: 14
adult
asthmatics, 17
healthy adults
Gender:
Asthmatics: 9
M/5 F, Healthy:
5 M/12 F
Ultrafine CAPs (Los
Angeles)
Particle Number/Count:
145,000/cm3, Range
39,000-312,0001cm3
Concentration: Mean-
100/yglm3, Range-
13-277 //g/m3
Subjects exposed for 2 h during intermittent
exercise (15-min periods) to both CAPs and
filtered air in random order. The first 7 subjects
underwent whole body exposure, while the
remaining subjects were exposed through a
facemask. Facemask exposures had higher
particle counts but lower particle mass than
whole body exposures. Exposures were
separated by at least 2 wks.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post-exposure.
Relative to filtered air, exposure to ultrafine CAPs resulted in
a transient decrease in LF power 4 h post-exposure. This
effect of CAPs on HRV was not influenced by health status.
CAPs exposure was not observed to affect any other
measures of HRV, blood pressure, or blood markers of
inflammation or coagulation. There were no differences in
response observed between facemask and whole body
exposures.
Asthmatics:
34 ± 12 yrs,
Healthy:
24 ± 8 yrs
Reference:
Graff et al.
(2009,191981)
Subjects: 14
healthy adults
Gender: 8 M/6
F
Age: 20-34 yrs
Coarse CAPs (Chapel Hill,
NC)
Concentration: 89
± 49.5 /yglm3 (estimated
inhaled dose ~ 67% of
measured particle mass)
Subjects exposed for 2 h with intermittent
exercise (15-min periods) to coarse CAPs and
filtered air in a randomized crossover design.
Exposures were separated by at least 2 mos.
Time to analysis: 0-1 and 20 h post-exposure.
At 20 h post-exposure, tPA was observed to decrease by
32.9% from baseline (pre-exposure) per 10/yglm3 increase in
CAPs concentration \p - 0.01). D-dimer concentration
decreased 11.3% per 10/yg/m3, a change of marginal
statistical significance \p - 0.07). No other coarse CAPs-
induced changes in blood biomarkers of coagulation (e.g.,
vWF, factor VII, plasminogen, fibrinogen, or PAI-1) or
inflammation (e.g., CRP) were observed. At 20 h post-
exposure, overall HRV (SDNN) was shown to decrease by
14.4% relative to pre-exposure measurements per 10/yg/m3
increase in CAPs concentration. No other changes in HRV
were observed following exposure to coarse CAPs.
July 2009	C-4	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Findings
Reference:
Fine CAPs (Chapel Hill, NC) Subjects exposed to CAPs (n - 30) or filtered air The increase in blood fibrinogen following exposure to fine
Huang et al. Concentration:
(2003,087377) 23.1-311.1 //g/m3
Subjects: 38
healthy adults
Gender: 36
M/2 F
Age: Avg
26.2 ± 0.7 yrs
(n - 8) for 2 h with intermittent exercise
(subjects did not serve as their own controls).
Component data of CAPs was available for 37 of
the 38 subjects.
Time to analysis: 18 h after exposure.
CAPs reported by Ghio et al. (2000, 012140) was shown to
be associated with copper, zinc, and vanadium content in the
CAPs.
Reference:
Larsson et al.
(2007,189320)
Subjects: 16
healthy adults
Gender: 10
M/6 F
Age: 19-59 yrs
Traffic particles (road
tunnel)
Particle Size: PM2.6,
PMio; PM2.6 mass
constituted —36% of
PM10 mass
Particle Number/Count:
20-1,000 nm: 1.1 x
105/cm3, < 100 nm:
0.85 x 105/cm3
Concentration: PM2 e-
46-81 //g/m3; PMio-
130-206 //g/m3
Exposures were conducted for 2 h with
intermittent exercise in a room adjacent to a
busy road tunnel. Study used a randomized
crossover design with each subject also exposed
to normal air (control). Exposures were separated
by 3-10 wks. No exposures to filtered air were
conducted. Other traffic emissions measured: NO
(874//g/m3), NO? (230 //g/m3), CO (5.8 //g/m3
reported, likely 5.8 mg/m3).
Time to analysis: 14 h post-exposure.
No change in plasma levels of fibrinogen or PAI-1 observed
14 h post exposure.
Reference:
Lucking et al.
(2008,191993)
Subjects: 20
healthy adults
Gender: M
Age: 21 -44 yrs
DE
Protocol 1 (n — 8): idling
Deutz diesel engine
(F3M2011, 2.2 L, 500
rpm) using gas oil
Protocol 2 (n — 12): idling
Volvo diesel engine (TD45,
4.5 L, 4 cylinders, 680
rpm) using Gasoil E10
Particle Number/Count:
Protocol 1: 1.2 x 10e/cm3;
Protocol 2:1.26 x
10e/cm3
Concentration: Protocol
1: 348//g/m3, Protocol 2:
330//g/m3
In both protocols, exposures were conducted
with intermittent exercise (15-min periods) to DE
and filtered air in a randomized crossover design
with exposures separated by at least one wk.
Protocol 1 (n — 8): Exposures conducted for 2 h.
Other diesel emissions measured: NOx (0.58
ppm), NO2 (0.23 ppm), NO (0.36 ppm), CO (3.54
ppm), total hydrocarbon (2.8 //g/m3).
Time to analysis: 6 h post-exposure.
Protocol 2 (n -12): Exposures conducted for 1 h.
Other diesel emissions measured: NOx (2.78
ppm), NO2 (0.62 ppm), NO (2.15 ppm), CO (3.08
ppm), total hydrocarbon (1.58 //g/m3).
Time to analysis: 2 and 6 h post-exposure.
Thrombus formation was observed to increase with diesel 2
and 6 h post-exposure using an ex vivo perfusion chamber.
Both platelet-neutrophil and platelet-monocyte aggregates
increased relative to filtered air 2 h following exposure to
diesel (only evaluated in Protocol 2). Plasma concentrations
of soluble CD40L were also observed to increase with diesel.
Exposure to diesel was not shown to affect total leukocyte,
monocyte, or platelet counts.
DE
Idling Cummins diesel
(2009,191159) engjne (5 g L) using
Reference:
Lund et al.
Subjects: 10
healthy adults
Gender: 4 M/6
F
commercial No. 2 fuel
Particle Size: MMAD
0.10//m
Concentration:
e: 18-40 yrs 100//g/m3
Subjects exposed for 1 h with intermittent
exercise (15-min periods) to DE and filtered air in
a randomized crossover study design. Other
diesel emissions measured: NOx (4.7 ppm), NO2
(0.8 ppm), CO (2.8 ppm), total hydrocarbons (2.4
ppm).
Time to analysis: 30 min and 24 h post-
exposure.
Exposure to diesel resulted in an increase in MMP-9 plasma
concentration and activity as well as an increase in
endothelin-1 plasma concentration at both 30 min and 24 h
post-exposure.
Reference:
Lundback et al.
(2009,191967)
Subjects: 12
healthy adults
Gender: M
Age: 21-30 yrs
DE
Idling Volvo diesel engine
(TD45, 4.5 L, 4 cylinders,
680 rpm) using Gasoil E10
Particle Number/Count:
1.26 x 106/cm3
Concentration:
330//g/m3
Subjects exposed for 1 h with intermittent
exercise (15-min periods) to DE and filtered air in
a randomized crossover study design. Exposures
were separated by at least one wk. Other diesel
emissions measured: NOx (2.78 ppm), NO2 (0.62
ppm), NO (2.15 ppm), CO (3.08 ppm), total
hydrocarbon (1.58//g/m3).
Time to analysis: 10, 20, 30, and 40 min post-
exposure.
Diesel-induced increase in arterial stiffness (increases in
augmentation pressure and augmentation index, as well as
decrease in time to wave reflection) observed at 10 and 20
min post-exposure using radial artery pulse wave analysis. No
effect of diesel observed on carotid-femoral pulse wave
velocity which was assessed 40 min post-exposure, but not
at earlier time points. No effect of diesel observed on blood
pressure 10-30 min post-exposure.
July 2009	C-5	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Findings
Reference:
Mills et al.
(2005,188557)
Subjects: 30
healthy adults
Gender: M
Age: 20-38 yrs
DE
Idling 1991 Volvo diesel
engine (TD45, 4.5 L, 4
cylinders, 680 rpm)
Particle Number/Count:
1.2 x 106/cm3
Concentration:
300 /yglm3
Subjects exposed for 1 h with intermittent
exercise (15-min periods) to DE and filtered air in
a randomized crossover study design. Exposures
were separated by two wks. Other diesel
emissions measured: NO2 (1.6 ppm), NO (4.5
ppm), CO (7.5 ppm), total hydrocarbon (4.3 ppm),
formaldehyde (0.26/yg/m3).
Time to analysis: 2-4 h post exposure for 15
subject; 6-8 h post exposure for the other 15
subjects.
Forearm blood flow increase (induced by bradykinin,
acetylcholine, and sodium nitroprusside) was attenuated by
DE 2 and 6 h post exposure. A 6 mmHg increase in diastolic
blood pressure (p - 0.08) 2 h following exposure to DE was
observed relative to filtered air control. Bradykinin-induced
release of t-PA was attenuated by diesel exposure 6 h
post exposure. DE did not affect the release of t-PA 2 h
post-exposure. No diesel-induced changes in serum IL-6 or
TNF-n observed 6 h post-exposure.
Reference:
Mills et al.
(2007,091206)
Subjects: 20
older adults
with prior
myocardial
infarction
Gender: M
Age: 60 ± 1
yrs
DE
Idling 1991 Volvo diesel
engine (TD45, 4.5 L, 4
cylinders, 680 rpm) using
low sulfur gas-oil E10
Particle Size: Median
particle diameter 54 nm,
Range 20-120 nm
Particle Number/Count:
1.26 x 106/cm3
Concentration:
300/yg/m3
Subjects exposed for 1 h with intermittent
exercise (15-min periods) to DE and filtered air in
a randomized crossover study design. Exposures
were separated by at least two wks. Other diesel
emissions measured: NOx (4.45 ppm), NO2 (1.01
ppm), NO (3.45 ppm), CO (2.9 ppm), total
hydrocarbon (2.8 ppm).
Time to analysis: During exposure and 6-8 h
post-exposure.
A greater increase in exercise induced ST-segment depression
and ischemic burden was observed during exposure to DE
than clean air. No diesel-induced effects on vasomotor
dysfunction observed 6 h post-exposure. Bradykinin-induced
release of t-PA was attenuated by diesel exposure relative to
filtered air 6 h post-exposure. Effect of diesel on t-PA release
was not evaluated at earlier times post-exposure. No
diesel-induced changes in blood leukocyte counts or serum
C-reactive protein 6 h post-exposure.
Reference:
Mills et al.
(2008,156766)
Subjects: 12
adults with
coronary heart
disease, 12
healthy adults
Gender: M
Age: CHD:
59 ± 2 yrs,
Healthy:
54 ± 2 yrs
Fine CAPs (Edinburgh,
Scotland, UK)
Particle Size: Mean
1.23/ym
Particle Number/Count:
99,400/cm3
Concentration:
190 ± 37 /yglm3
Exposures conducted for 2 h with intermittent
exercise. Subjects exposed to CAPs and filtered
air using a randomized crossover design with
exposures separated by at least 2 wks.
Time to analysis: 2, 6-8, and 24 h post-
exposure.
CAPs exposure had no significant effect on vascular function
in healthy adults or adults with coronary heart disease 6-8 h
post-exposure (i.e., no change in forearm blood flow as
assessed using venous occlusion plethysmography). The
authors attributed this lack of response to a low
concentration of combustion-derived particles. Small increase
in blood platelet and monocyte concentration observed
following CAPs exposure. Exposure to CAPs did not affect
serum CRP concentration or total leukocyte or neutrophil
count.
Reference:
Peretz et al.
(2007,189082)
Subjects: 5
healthy adults
Gender: M
Age: 20-31 yrs
DE
2002 Cummins B-series
diesel engine (6BT5.9G6,
5.9 L); operating at 75%
of rated capacity
Concentration: Fine PM
50,100, 200/yg/m3
Subjects exposed for 2 h at rest to filtered air
and each of the three DE particle concentrations
in a randomized crossover study design.
Exposures were separated by at least 2 wks.
Other diesel emissions measured, 200 //gfm3
exposure: NO2 (23 ppb), NO (1.75 ppm), CO (1.58
ppm).
Time to analysis: 6 and 22 h after the start of
exposure.
PBMC expression of 10 genes involved in the inflammatory
response were observed to be significantly affected by
exposure to DE at the highest concentration tested (8
upregulated, 2 downregulated) 6 h after the start of
exposure. The expression of 4 genes (1 upregulated, 3
downregulated) associated with the inflammatory response
showed significant changes 22 h after diesel exposure.
PBMC expression of 5 genes involved in the oxidative stress
pathways showed significant changes at 6 h after the start
of diesel exposure at the highest concentration tested (4
upregulated, 1 downregulated). 7 genes involved in the
oxidative stress pathways showed significant changes at
22 h following exposure (4 upregulated, 3 downregulated).
July 2009	C-6	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Findings
Reference:
Peretz et al.
(2008,156854)
Subjects: 17
adults with
metabolic
syndrome, 10
healthy adults
Gender:
Metabolic
syndrome: 11
M/6 F, Healthy:
8 M/2 F
Age: Metabolic
syndrome: 20-
48 yrs, Healthy:
20-42 yrs
DE
2002 Cummins Bseries
diesel engine (6BT5.9G6,
5.9 L) using No. 2 undyed
on-highway fuel: operating
at 75% of rated capacity
Particle Size: Median
particle diameter 1.04/ym
Concentration: Fine PM
100, 200/yg|m3
Subjects exposed for 2 h at rest to both
concentrations of DE as well as filtered air in a
randomized crossover design. Exposures were
separated by at least 2 wks. Other diesel
emissions measured, 100 //gfm3 exposure: NO2
(16.5 ppb), NO (0.96 ppm), CO (0.51 ppm):
200 /yglm3 exposure: NO2 (24.7 ppb), NO (1.54
ppm), CO (0.89 ppm).
Time to analysis: Immediately following
exposure (within 30 min post exposure) and 3 h
from the start of exposure.
Exposure to 200 //gfm3 elicited a statistically significant
decrease in brachial artery diameter relative to filtered air
immediately following exposure. A smaller decrease in
brachial artery diameter was also observed following
exposure to DE at 100/yg/m3. Plasma levels of endothelin-1
were observed to increase following DE exposure
(200 /yglm3). The observed effects were more pronounced in
healthy subjects than in subjects with metabolic syndrome.
DE did not affect endothelium-dependent flow-mediated
dilatation. No effect of DE on blood pressure was
demonstrated immediately following exposure.
Reference:
Peretz et al.
(2008,156855)
Subjects: 13
adults with
metabolic
syndrome, 3
healthy adults
Gender:
Metabolic
syndrome: 8
M/5 F, Healthy:
3 M/0 F
Age: Metabolic
syndrome:
31-48 yrs,
Healthy: 24-39
yrs
DE
2002 Cummins B series
diesel engine (6BT5.9G6,
5.9 L) using No. 2 undyed
on-highway fuel: operating
at 75% of rated capacity
Concentration: Fine PM
100, 200/yg/m3
Subjects exposed for 2 h at rest to both
concentrations of DE as well as filtered air in a
randomized crossover design. Exposures were
separated by at least 2 wks. Other diesel
emissions measured, 100 //gfm3 exposure: NO2
(20.6 ppb), NO (0.95 ppm), CO (0.47 ppm):
200 /yglm3 exposure: NO2 (28.3 ppb), NO (1.63
ppm), CO (0.74 ppm).
Time to analysis: 1, 3, 6, and 22 h from the
start of exposure.
Exposure to 200 //gfm3 increased HF power and decreased
the LF/HF ratio 1h post-exposure: however, this effect was
not consistent across subjects. No effect of DE was observed
at later time points. Subjects with metabolic syndrome did
not experience greater changes in HRV than healthy subjects.
Reference:
Power et al.
(2008,191982)
Subjects: 5
adults with
mild-to-
moderate
allergic asthma
Gender: 1 M/4
F
Age: 28-51 yrs
Carbon and ammonium
nitrate particles
Concentration:
With co-exposure to
0.2ppm O3: 255/yg/m3,
Without co-exposure to
0.2ppm O3: 313/yg/m3
Subjects exposed for 4 h with intermittent
exercise (30-min periods) to filtered air, particles,
and particles + ozone in a crossover study
design. Exposures were separated by at least 3
wks.
Time to analysis: 3 h 40 min from the start of
exposure.
Time and frequency domain HRV parameters were not
affected by particle exposure relative to filtered air.
However, exposure to particles with ozone resulted in a
significant decrease in SDNN as well high and low frequency
power normalized to the difference between total and very
low frequency power.
July 2009	C-7	DRAFT - DO NOT CITE OR QUOTE

-------
Study	Pollutant
Exposure
Findings
Reference:
Routledge et al.
(2006, 088674)
Subjects: 20
older adults
with coronary
artery disease,
20 healthy older
adults
Gender: CAD:
17M/3 F,
Healthy: 10
M/10F
Age: CAD:
52-74 yrs,
Healthy: 56-75
yrs
Ultrafine carbon
Particle Size: < 10-300
nm; mode at 20-30 nm
Concentration: Ultrafine
carbon: 50 //g|m3; SO2:
200ppb
Exposures conducted (head dome system) to
filtered air, ultrafine carbon, SO2, and ultrafine
carbon + SO7 for 1 h at rest using a randomized
crossover study design.
Time to analysis: Immediately following
exposure as well as 3 and 23 h post-exposure.
No PM-induced changes in HRV observed among subjects
with coronary artery disease. Among healthy subjects, small
increase in HRV (RR, SDNN, rMSSD, and LF power)
demonstrated immediately post-carbon exposure. Relative to
filtered air control, exposure to ultrafine carbon did not
significantly affect blood pressure in healthy adults or adults
with coronary artery disease 0-3 h post-exposure. Exposure
to ultrafine carbon, either alone or with S02, did not affect
plasma levels of fibrinogen or D-dimer at 3 or 23 h
post-exposure. Exposure to ultrafine carbon did not affect
peripheral blood leukocyte count or C-reactive protein levels 3
or 23 h post-exposure.
Reference:
Rundell and
Caviston (2008,
191986)
Subjects: 15
healthy college
athletes
Gender: M
Age: Avg 19.5
yrs
Gasoline emissions
2.5 hp gasoline engine
running 10 s each min
during exposure and in the
min prior to exposure
Particle Size: PM1.0
Particle Number/Count:
Trial 1: 336,730 ±
149,206/cm3;
Trial 2: 396,200 ±
82,564/cm3
Subjects were exposed twice to both clean air
and dilute gasoline exhaust during 6-min periods
of maximal exercise on a cycle ergometer. Clean
air exposures occurred first and were separated
by 3 days. Gasoline exhaust exposures were also
separated by 3 days, with the first occurring 7
days after the second clean air exposure. Other
emissions measured: CO (6.3 ± 3.4 ppm).
Time to analysis: 6 min
There was no difference in total work done (kJ) between the
clean air exposures or between the clean air exposures and
the first exposure to gasoline exhaust. However, the second
gasoline exhaust exposure was demonstrated to significantly
decrease work accumulated over the 6min exercise period
compared with either of the other exposure conditions (p <
0.01).
Reference:
Samet et al.
(2007,156940)
Subjects:
Ultrafine CAPs:
20 healthy
adults, Coarse
CAPs: 14
healthy adults
Gender:
Ultrafine CAPs:
11 M/9 F,
Coarse CAPs: 8
M/6 F
Age: Ultrafine
CAPs: 18-35
yrs, Coarse
CAPs: 18-35
yrs
CAPs (Chapel Hill, NC)
Particle Size: Ultrafine
0.049 ± 0.009 /jm;
Coarse 3.59 ± 0.58/ym
Concentration: Ultrafine
47.0 ± 20.2/yg|m3:
Coarse 89.0 ± 49.5/yg/m3
Preliminary report comparing effects of
controlled exposures to coarse, fine, and
ultrafine CAPs among healthy adults (3 separate
studies). A randomized crossover design was
used in evaluating effects of coarse CAPs
(n — 14) and ultrafine CAPs (n-20) relative to
filtered air following 2-h exposures with
intermittent exercise. Results compared with
previous study of controlled exposure to fine
CAPs (Chapel Hill, NC) where subjects did not
serve as their own controls (Ghio et al., 2000,
012140).
Time to analysis: 0-20 h post-exposure.
Statistically significant decrease in SDNN observed 20 h
following exposure to coarse CAPs relative to filtered air.
Subjects in the high ultrafine CAPs group experienced a
decrease in SDNN based on an analysis of 24 h ambulatory
Holter monitoring relative to filtered air. Fine CAPs did not
significantly affect HRV. Increased levels of D-dimer observed
18 h following exposure to ultrafine CAPs. No CAPs-induced
changes in plasma factor VII, plasminogen, fibrinogen, PAI-1,
vWf, or t-PA. No CAPs-induced changes in C-reactive protein
levels were observed.
July 2009
C-8
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Findings
Reference:
Samet et al.
(2009,191913)
Subjects: 19
healthy adults
Gender: 10
M/9 F
Age: 18-35 yrs
Ultrafine CAPs (Chapel
Hill, NC)
Particle Size: <
0.16/ym
Particle Number/Count:
120,662 ± 48,232
particles/cm3
Concentration: 49.8 ±
20/yg/m3
Subjects exposed for 2 h with intermittent 15
periods of exercise to UF CAPs and filtered air
using a randomized crossover study design.
Time to analysis: Immediately following
exposure and 1 and 18 h post exposure.
UF CAPs exposure resulted in an increase in plasma
concentrations of D-dimer both immediately following
exposure (20.6% increase per 10E particles/cm3) as well as
18 h post-exposure (18.2% increase per 106 particles/cm3).
Plasma concentration of PAI1 also increased with UF CAPs,
although this increase was not statistically significant (24%
increase,/; - 0.1). No UF CAPs-induced changes observed in
plasma concentrations of tPA, vWF, CRP, fibrinogen,
plasminogen, or Factor VII. HF and LF power were both
observed to increase with UF CAPs exposure relative to
filtered air at 18 h post-exposure (41.8% and 36%,
respectively, per 106 particles/cm3 increase in UF CAPs). UF
CAPs expressed as mass concentration was not observed
have a statistically significant effect in HF total power. UF
CAPs was not observed to affect time domain measures of
HRV over 24 h. The QT interval was shown to decrease both
immediately following and at 18 h post exposure (not
statistically significant immediately following exposure).
Reference:
Shah et al.
Ultrafine elemental carbon Exposures conducted via mouthpiece for 2 h with Exposure to ultrafine carbon attenuated peak forearm blood
Particle Number/Count:
(2008,156970) 10.8 ± 1.7 x 10e/cm3
Subjects: 16 Concentration:
50/yg/m3
healthy adults
Age:
26.9 ± 6.9 yrs
intermittent exercise to filtered air and ultrafine
carbon in a randomized crossover study design.
Time to analysis: Immediately following
exposure as well as 3.5, 21, and 45 h post-
exposure.
flow after ischemia relative to filtered air 3.5 h
post-exposure. Venous nitrate levels were significantly lower
at 21 h following exposure to UF carbon compared with
filtered air exposure. PM exposure was not observed to
affect blood pressure relative to filtered air at times 0-45 h
post-exposure.
Reference: DE
Tornqvist et al. |d|ing 1 ggl Volvo diesel
(2007, 091279) engine (TD45i 4 5 L 4
Subjects: 15
healthy adults
Gender: M
Age: 18-38 yrs
cylinders, 680 rpm)
Concentration:
300 /yglm3
Subjects exposed for 1 h with intermittent
exercise (15-min periods) to DE and filtered air in
a randomized crossover study design. Exposures
were separated by at least two wks. Other diesel
emissions measured: NOx (4.44 ppm), NO2 (0.82
ppm), NO (3.62 ppm), total hydrocarbon (2.21
ppm).
Time to analysis: 24 h post-exposure.
DE was observed to significantly attenuate
endothelium-dependent vasodilation 24 h post-exposure.
Endothelium-independent vasodilation was not affected by
diesel exposure. Exposure to DE did not affect blood pressure
relative to filtered air 24 h after exposure. DE significantly
increased plasma levels of IL-6 and TNF-n 24 h following
exposure. Exposure to diesel resulted in an increase in total
antioxidant capacity of plasma relative to filtered air 24 h
post-exposure.
Reference:
Urch et al.
(2004, 055629)
Subjects: 24
healthy adults
Gender: 14
M/10 F
Age: 35 ±10
yrs
Fine CAPs (Toronto)
Concentration:
150 /yglm3 (range
101-257 //g/m3) with 120
ppb ozone
Exposures conducted through a facemask which CAPs + ozone exposure resulted in a significant decrease in
covered the subject's nose and mouth. Subjects
were exposed to CAPs + ozone and filtered air
for 2 h at rest in a randomized crossover study
design. Exposures were separated by at least 2
days.
Time to analysis: Immediately following
exposure.
brachial artery diameter immediately post-exposure (Brook et
al., 2002, 024987), which was demonstrated to be
associated with both the organic and elemental carbon
fractions of the CAPs.
Reference:
Urch et al.
(2005,081080)
Subjects: 23
healthy adults
Gender: 13
M/10 F
Age: 32 ± 10
yrs
Fine CAPs (Toronto);
Concentration:
150 /yglm3 (range
102-214/yg/m3) with 120
ppb ozone
Exposures conducted through a facemask which An increase in diastolic blood pressure of 6 mmHg was
covered the subject's nose and mouth. Subjects
were exposed to CAPs + ozone and filtered air
for 2 h at rest in a randomized crossover study
design.
Time to analysis: Every 30 min during
exposure, with the final measurement made
immediately prior to the end of the exposure.
observed at the end of CAPs + ozone exposure, which was
statistically different from the change in blood pressure
experienced during exposure to filtered air (1 mmHg). This
effect was associated with the organic fraction of PM2.6.
July 2009	C-9	DRAFT - DO NOT CITE OR QUOTE

-------
Study	Pollutant
Exposure
Findings
Reference: Ultrafine elemental carbon
Zareba et al. Particle Size: Count
(2009,190101) me(|jan diameter 25 nm
Subjects: 24 Partic|e Number/Count:
2x10efcm3 (10//g/m3),
7x10efcm3 (25//g/m3)
Concentration: 10 //g/m3;
healthy adults
Gender: 12
M/12F
e: 18-40 yrs 25//g/m
Protocol 1 (n — 12, 6 M/6 F): Subjects exposed to
10 //g/m3 UF carbon and filtered air for 2 h at
rest in a randomized crossover design. Exposures
were separated by at least 2 wks.
Protocol 2 (n — 12, 6 M/6 F): Subjects exposed to
10 //g/m3, 25 //g/m3, and filtered air for 2 h with
intermittent exercise (15-min periods) in a
restricted randomized crossover design (all
subjects exposed to 10 //g/m3 before 25 //g/m3).
Exposures were separated by at least 2 wks.
Time to analysis (both protocols): Immediately
following exposure and 3.5 and 21 h post-
exposure.
Exposure to 10 //g/m3 at rest resulted in no change in HRV
frequency domain parameters relative to filtered air exposure.
Time domain parameters were observed to increase slightly
with UF carbon exposure (10 //g/m3 at rest), however, only
the increase in rMSSD was statistically significant \p -
0.032). Some trends toward less shortening of QT interval,
increase in ST segment, and increase in variability of
repolarization (variability of T wave complexity) were
observed with exposure to 10 //g/m3 at rest, none of which
were statistically significant.
In Protocol 2, exposure to 10 //g/m3 UF carbon was observed
to slightly increase HRV time domain parameters as was
demonstrated in Protocol 1. However, this was not observed
at the higher concentration (25 //g/m3). As with exposure at
rest, exposure to UF carbon during exercise was observed to
affect repolarization (reduction in QT duration and increase in
T-wave amplitude), although this effect was not statistically
significant.
Table C-2. Respiratory effects
Reference
Pollutant
Exposure
Findings
Reference: Alexis et al.
(2006, 088636)
Subjects: 9 healthy
adults
Gender: 3 M/6 F
Age: 18-35 yrs
Coarse fraction particles
(Chapel Hill, NC)
Heat-treated (biologically
inactive) and non-heated
particles
Particle Size: MMAD
5 //m
Concentration: 0.65 mg
per subject
Subjects were administered heat-treated PM2.6 Both heat-treated and non-heated coarse PM were
-io, non-heated PM2.6 -10, and 0.9% saline
(control) via nebulization in a randomized
crossover study design. Exposures were
separated by at least 1 wk.
Time to analysis: 2-3 h post-inhalation.
observed to increase neutrophil counts in induced
sputum 2-3 h post-inhalation. Biologically active PM
(non-heated) induced an increase expression of
macrophage TNF-a mRNA, eotaxin, and immune
surface phenotypes on macrophages (mCD14,
CD11b/CR3, and HLA-DR).
Reference: Barregard et
al. (2008,155675)
Subjects: 13 healthy
adults
Gender: 6 M/ 7 F
Age: 20-56 yrs
Wood smoke
Particle Size:
Session 1: geometric
mean diameter 42 nm,
Session 2: geometric
mean diameter 112 nm
Particle Number/Count:
Session 1: 180,000/cm3;
Session 2: 95,000/cm3
Concentration: Session
1: median 279//g/m3;
Session 2: median
243//g/m3
Subjects exposed in two groups for 4 h to
filtered air, followed a wk later by a 4-h
exposure to wood smoke. Exposures conducted
with two 25-min periods of light exercise.
Other measured combustion products:
Session 1: NO2 (0.08 ppm), CO (13 ppm),
formaldehyde (114//g/m3), acetaldehyde
(75 //g/m3), benzene (30 //g/m3), 1,3-butadiene
(6.3//g/m3);
Session 2: NO2 (0.09 ppm), CO (9.1 ppm),
formaldehyde (64 //g/m3), acetaldehyde
(40 //g/m3), benzene (20 //g/m3), 1,3-butadiene
(3.9//g/m3).
Time to analysis: Immediately following
exposure as well as 3 and 20 h post-exposure.
Relative to filtered air, exposure to wood smoke was
observed to increase levels of eNO 3 h post-exposure.
Serum Clara cell protein increased 20 h after wood
smoke exposure. Wood smoke was observed to
increase levels of malondialdehyde in breath
condensate immediately after as well as 20 h
post-exposure. Effects of wood smoke on eNO and
malondialdehyde levels were similar between the two
sessions of wood smoke exposure. However, serum
Clara cell protein was significantly increased with
wood smoke in session 1 (higher particle count) but
not in session 2.
Reference: Bastain et al.
(2003, 098690)
Subjects: 18 nonsmoking
adults with positive
allergy skin test to short
ragweed
Gender: 7 M/11 F
Age: 18-38 yrs
DEP
Isuzu diesel engine, 4
cylinder, 4JB1
Concentration: 0.3 mg in
200 //I saline
Subjects underwent nasal provocation
challenge (intranasal spray) with allergen and
either DEP or placebo (saline) in a randomized
crossover study design. Challenges were
separated by 30 days. This protocol was then
repeated 30 days after the last exposure.
Time to analysis: 24 h post-exposure and 4
and 8 days after exposure.
DEP significantly increased allergic responses to short
ragweed. Relative to allergen + placebo, allergen +
DEP increased allergen specific IgE 4days following
exposure, and increased IL-4 1 day post-exposure. The
enhancement of allergic response with DEP was
observed to be reproducible within subjects.
July 2009
C-10
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Findings
Reference: Beckett et al.
(2005,156261)
Subjects: 12 healthy
adults
Gender: 6 M/6 F
Age: 23-52 yrs
Ultrafine and fine zinc
oxide
Particle Size: UF:
< 0.1 /jm; Fine: 0.1-
1.0/ym
Particle Number/Count:
UF: 4.6 x 107/cm3; Fine:
1.9 x 10W
Concentration:
500/yg/m3
Subjects exposed via mouthpiece for 2 h during
rest to filtered air, ultrafine, and fine zinc oxide
in a randomized crossover study design.
Exposures were separated by at least 3 wks.
Time to analysis: 11 and 24 h after exposure.
No changes observed in neutrophil count in induced
sputum. No PM (zinc oxide)-induced changes in
respiratory symptoms observed 0-24 h post exposure.
Reference: Behndig et al.
(2006, 088286)
Subjects: 15 healthy
adults
Gender: 8 M/7 F
Age: 21-27 yrs
DE
Idling 1991 Volvo diesel
engine (TD45, 4.5 L, 4
cylinders, 680 rpm)
Particle Size: PMio;
majority of PM mass
made up of particles <
1 //m in diameter
Concentration:
100/yg/m3
Exposures conducted for 2 h with intermittent
exercise to both DE and filtered air in a
randomized crossover design. Exposures were
separated by at least 3 wks. Other diesel
emissions measured: NOx (1.8 ppm), NO2 (0.4
ppm), NO (1.3 ppm), CO (10.4 ppm), total
hydrocarbons (1.3 ppm).
Time to analysis: 18 h post-exposure.
Exposure to DE increased neutrophil and mast cell
numbers in bronchial mucosa at 18 h post-exposure.
Neutrophils, IL-8, and myeloperoxidase observed to
increase in bronchial lavage fluid following exposure
relative to filtered air. No inflammatory response
observed in the alveolar compartment. Exposure to DE
increased urate and reduced glutathione
bronchoalveolar lavage at 18 h post-exposure.
Reference: Blomberg et DE
al. (2005,191991) Concentration:
Subjects: 15 older adults 300 //gfm3
(former smokers) with
COPD
Age: 56-72 yrs
Subjects exposed for 1 h with intermittent
exercise to DE and filtered air in a randomized
crossover study design.
Time to analysis: 6 and 24 h post-exposure.
DE was not observed to affect levels of Clara cell
protein in peripheral blood at 6 and 24 h post-
exposure.
Reference: Bosson et al. DE
(2007,156286)
Subjects: 16 healthy
adults
Gender: 7 M/9 F
Age: 20-28 yrs
Subjects exposed to DE for 1 h followed 5 h The percentage of neutrophils and concentration of
Idling Volvo diesel engine
Concentration: PM
300/yg/m3 followed by
exposure to filtered air or
0.2 ppm ozone
later by a 2-h exposure to either filtered air or
ozone (0.2 ppm) using a randomized crossover
study design. All exposures were conducted
with subjects engaged in intermittent exercise.
Time to analysis: 18 h after second exposure
(filtered air or ozone).
myeloperoxidase in induced sputum (18 h
post-ozone/air exposure) was significantly higher
following diesel + ozone than diesel + air.
Reference: Bosson et al. DE
(2008,156287)
Subjects: 14 healthy
adults
Gender: 9 M|5 F
Age: 21-29 yrs
Idling 1991 Volvo diesel
engine (TD45, 4.5 L, 4
cylinders)
Concentration: PM
300/yg/m3 or filtered air
followed by exposure to
0.2 ppm ozone
Subjects exposed to DE or filtered air for 1 h
followed 5 h later by a 2-h exposure to ozone
(0.2 ppm) using a randomized crossover study
design. All exposures were conducted with
subjects engaged in intermittent exercise.
Other diesel emissions measured: NO2 (0.51
ppm), NO (1.65 ppm), total hydrocarbons (1.18
ppm).
Time to analysis: 24 h after the start of the
initial exposure.
Neutrophil and macrophage numbers in bronchial wash
were significantly increased 16 h following ozone
exposure when preceded by exposure to diesel,
compared to ozone exposure preceded by exposure to
filtered air.
Reference: Brauner et al.
(2009,190244)
Subjects: 29 healthy
adults
Gender: 20 M, 9 F
Age: M avg 27 yrs, F avg
26 yrs
Urban traffic particles
Particle Size: PM2.6,
PM2.6 -10
Particle Number/Count:
6-700 nm: 10,067|cm3
Concentration: PM2.6:
9.7/yglm3, PM2.6 -io:
12.6/yg/m3
Subjects exposed to urban traffic particles and
filtered air for 24 h with and without two
90-min periods of light exercise in a
randomized crossover study design.
Concentrations of NOx and NO were low and
did not differ between filtered and unfiltered
exposures. CO concentrations were higher with
filtered air (0.35 and 0.41 ppm), while ozone
concentrations were lower with filtered air
(12.08 and 4.29 ppb).
Time to analysis: 2.5, 6, and 24 h after the
start of exposure.
Epithelial membrane integrity and blood-gas barrier
permeability, assessed using pulmonary clearance of
"To-labeled diethylenetriamine pentaacetic acid
(DTPA), was observed to increase with exercise, but
was not affected by exposure to urban particles (2.5 h
of exposure). Exposure to urban particles was not
shown to affect plasma or urine concentration of
Clara cell 16 protein at 6 and 24 h after the start of
exposure. No relationship between exposure and
pulmonary function was observed at 2.5 h.
Reference: Gilliland et al.
(2004,156471)
Subjects: 19 adults with
allergic rhinitis and
positive skin test to
ragweed, GSTM1 (14
null, 5 present): GSTT1 (9
null, 10 present): GSTP1
codon 105 variants (13
III, 6 l/V, 0 V/V)
Gender: 7 M/12 F
Age: 20-34 yrs
DEP
Isuzu diesel engine, 4
cylinder, 4JB1
Concentration: 0.3 mg
DEP in 300 //L saline
Subjects were challenged intranasal^ with
allergen and placebo (saline) as well as allergen
plus DEP in saline in a randomized crossover
design. Challenges were separated by at least
6 wks.
Time to analysis: 10 min, 24 h, and 72 h
post-challenge.
Subjects who were GSTM1 null or homozygous for
GSTP1 1105 wild-type allele experienced significantly
greater increase in nasal IgE and histamine with diesel
plus allergen compared to subjects with functional
GSTM1 or who were heterozygous for GSTP1
l/V(105).
July 2009	C-11	DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Findings
Reference: Gong et al.
(2004, 087964)
Subjects: 13 older adults
with COPD, 6 healthy
older adults
Gender: COPD: 5 Ml 8 F,
Healthy: 2 M/4 F
Age: COPD: avg 68 yrs,
Healthy: avg 73 yrs
Fine CAPs (Los Angeles)
Particle Size: 85% of
mass between 0.1 and
2.5/ym
Concentration: Mean:
194/yg/m3,
135-229 //g/m3
Exposures to CAPs and filtered air (randomized
crossover) for 2 h with intermittent light
exercise (four 15-min periods). Exposures were
separated by at least 2 wks.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post exposure.
No CAPs-induced respiratory symptoms observed in
healthy older adults or older adults with COPD at 0,4,
or 22 h post exposure. Exposure to CAPs did not
significantly affect FVC or FEVi. CAPs exposure
caused a decrease in arterial oxygen saturation 4 h
post exposure which was more pronounced in healthy
older adults than in older adults with COPD. Exposure
to CAPs was not observed to affect the levels of
white blood cells, columnar epithelial cells, IL-6, or IL-
8 in induced sputum.
Reference: Gong et al.
(2004, 055628)
Subjects: 12 adult
asthmatics, 4 healthy
adults
Gender: Asthmatic: 4
M/8 F, Healthy: 2 M/2 F
Age: Asthmatic: avg 38
yrs, Healthy: avg 32 yrs
Coarse CAPs (Los
Angeles)
Particle Size: 80% of
mass between 2.5 and
10 /jm, 20% of mass
< 2.5/ym
Concentration: Mean:
157/yg/m3;
56-218 //g/m3
Exposures to CAPs and filtered air (randomized
crossover) for 2 h with intermittent light
exercise (four 15-min periods). Exposures were
separated by at least 2 wks.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post-exposure.
No effect of CAPs exposure on spirometry or arterial
oxygen saturation was observed 0,4, or 22 h
post-exposure. No respiratory symptoms reported
0-22 h post-exposure in either healthy or asthmatic
adults. Sputum cell counts at 22 h post-exposure did
not differ between CAPs and filtered air.
Reference: Gong et al.
(2005,087921)
Fine CAPs (Los Angeles)
Concentration: CAPs:
Subjects: 18 older adults 200 /yg/m3; NO2: 0.4 ppm
with COPD, 6 healthy
older adults
Gender: COPD: 9 M/9 F,
Healthy: 2 M/4 F
Age: COPD: avg 72 yrs,
Healthy: avg 68 yrs
Each subject was exposed to CAPs, NO2, CAPs
+ NO2, and filtered air for 2 h with
intermittent exercise. Exposure order was not
fully counterbalanced as NO2 exposures were
conducted after the majority of the CAPs and
filtered air exposures had been completed.
Exposures were separated by at least 2 wks.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post-exposure.
Exposure to CAPs was observed to decrease maximal
mid-expiratory flow and arterial oxygen saturation
relative to filtered air 4-22 h post-exposure. This
response was more pronounced in healthy older adults
than in older adults with COPD. Concomitant exposure
to NO2 did not enhance the response. No other
respiratory responses (symptoms, spirometry, sputum
cell counts) were affected by exposure to CAPs.
Reference: Gong et al.
(2008,156483)
Subjects: 14 adult
asthmatics, 17 healthy
adults
Gender: Asthmatics: 9
M/5 F, Healthy: 5 M/12 F
Age: Asthmatics:
34 ± 12 yrs,
Healthy:24 ± 8 yrs
Ultrafine CAPs (Los
Angeles)
Particle Number/Count:
145,000/cm3, Range
39,000-312,000/cm3
Concentration: Mean:
100/yg/m3, Range:
13-277 //g/m3
Subjects exposed for 2 h during intermittent
exercise (15-min periods) to both CAPs and
filtered air in random order. The first 7
subjects underwent whole body exposure,
while the remaining subjects were exposed
through a facemask. Facemask exposures had
higher particle counts but lower particle mass
than whole body exposures. Exposures were
separated by at least 2 wks.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post-exposure.
No significant differences in respiratory symptoms
observed between filtered air and ultrafine CAPs
exposures. Individuals exposed to higher particle
counts tended to experience greater symptoms with
CAPs than with filtered air. An ultrafine CAPs-induced
decrease in arterial oxygen saturation (0.5%) was
observed at 0,4, and 22 h post-exposure. A decrease
in FEVi (2%) was also observed 22 h post-exposure
relative to filtered air. Responses were not
significantly different between healthy and asthmatic
adults. CAPs exposure was not observed to affect
total sputum cell counts or cytokine levels. There
were no differences in response observed between
facemask and whole body exposures.
Reference: Graff et al. Coarse CAPs (Chapel I
(2009,191981)	NC)
Subjects: 14 healthy
adults
Gender: 8 M/6 F
Age: 20-34 yrs
Concentration: 89
± 49.5/yg/m3 (estimated
inhaled dose ~ 67% of
measured particle mass)
Subjects exposed for 2 h with intermittent
exercise (15-min periods) to coarse CAPs and
filtered air in a randomized crossover design.
Exposures were separated by at least 2 mos.
Time to analysis: 0-1 and 20 h post-exposure.
Pulmonary function (FVC, FEVi, and carbon moNOxide
diffusing capacity) was not affected by exposure to
coarse CAPs either immediately following exposure or
20 h post-exposure. A significant increase in percent
PMNs (10.7% increase per 10/yg/m3 coarse CAPs)
was observed in BAL fluid 20 h post-exposure.
Percent monocytes in BL fluid were slightly decreased
at 20 h post-exposure (2.0% decrease per 10/yg/m3
CAPs:p - 0.05). Total protein in BAL fluid was also
observed to decrease following CAPs exposure (1.8%
decrease per 10/yg/m3 CAPs). Markers of
inflammation in BAL and BL fluids including IL-6, IL-8,
and PGE2 were not affected by exposure to coarse
CAPs.
Reference: Huang et al.	Fine CAPs (Chapel \
(2003,087377)	NC)
Subjects: 38 healthy	Concentration:
adults	23.1-311.1/yg/m3
Gender: 36 M/2 F
Age: Avg 26.2 ± 0.7 yrs
Subjects exposed to CAPs (n - 30) or filtered
air (n - 8) for 2 h with intermittent exercise
(subjects did not serve as their own controls).
Component data of CAPs was available for 37
of the 38 subjects.
Time to analysis: 18 h after exposure.
The increase in bronchoalveolar lavage fluid
neutrophils observed by Ghio et al. (2000, 012140)
following exposure to fine CAPs was shown to be
associated with iron, selenium, and sulfate content of
the CAPs.
July 2009
C-12
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Findings
Reference: Kongerud et
al. (2006,156656)
Subjects: 17 asthmatic
adults, 46 healthy adults
Gender: Asthmatics- 6
M/11 F, Healthy- 24
MI22F
Age: Asthmatics: avg 23
yrs, Healthy: avg 26 yrs
DEP
NIST 1650, heavy duty
diesel engine
Concentration:
Untreated and treated
with 0.1 ppm ozone (48
h); 300 /yg per nostril
DEP (with and without ozone pre-treatment)
were intranasal^ instilled, using the saline
vehicle as control. Subjects did not serve as
their own controls (not a crossover design).
Time to analysis: 4 and 96 h post-instillation.
Exposure to DEP was not observed to alter markers of
inflammation in nasal lavage fluid (e.g., cell counts,
IL-8, IL-6) at 4 or 96 h post-instillation.
Reference: Larsson et al.
(2007,189320)
Subjects: 16 healthy
adults
Gender: 10 M/6 F
Age: 19-59 yrs
Traffic particles (road
tunnel)
Particle Size: PM2.6,
PMio; PM2.6 mass
constituted —36% of
PM10 mass
Particle Number/Count:
20-1,000 nm: 1.1 x
105/cm3, < 100 nm:
0.85 x 105/cm3
Concentration: PM2 e-
46-81 /yg/m3; PMio-
130-206 //g/m3
Exposures were conducted for 2 h with
intermittent exercise in a room adjacent to a
busy road tunnel. Study used a randomized
crossover design with each subject also
exposed to normal air (control). Exposures
were separated by 3-10wks. No exposures to
filtered air were conducted. Other traffic
emissions measured: NO (874/yg/m3), NO2
(230 /yg/m3), CO (5.8 /yg/m3 reported, likely
5.8 mg/m3).
Time to analysis: 14 h post-exposure.
An increase in bronchoaveolar lavage fluid cell
number, lymphocytes, and alveolar macrophages were
observed 14 h after road tunnel exposure relative to
control. Traffic particulate exposure was not shown
to effect cytokine or adhesion molecule expression in
bronchial tissues. Respiratory symptoms were
reported to increase during exposure to road tunnel air
relative to pre exposure symptom ratings. Exposure to
road tunnel air was not shown to affect lung function.
Reference: Mudway et
al. (2004,180208)
Subjects: 25 healthy
adults
Gender: 16 M/9 F
Age: 19-42 yrs
DE
Idling 1991 Volvo diesel
engine (TD45, 4.5 L, 4
cylinders, 680 rpm)
Concentration: PMio
100 /yg/m3
Subjects exposed to DE and filtered air for 2 h
with intermittent exercise (15-min periods) in a
randomized crossover design. Exposures were
separated by at least 3 wks. Other diesel
emissions measured: NO2 (0.2 ppm), NO (0.6
ppm), CO (1.7 ppm), total hydrocarbons (1.4
ppm), formaldehyde (43.5 /yg/m3).
Time to analysis: 1 h after the start of
exposure, immediately following exposure, and
6 h post-exposure.
DE caused mild throat irritation in some subjects and a
significant increase in airways resistance (Raw) during
or immediately following exposure. No changes in
FEVi or FVC were observed following exposure to
diesel. Neutrophil numbers in the bronchial airways
tended to increase following exposure to DE, however,
this increase was highly variable between subjects
and did not reach statistical significance. Exposure to
DE did not affect levels of SOD or malondialdehyde in
the airways. An increase in levels of ascorbate and
GSH in nasal lavage fluid was observed 6 h following
exposure to DE.
Reference: Pietropaoli et
al. (2004,156025)
Subjects: 16 asthmatic
adults, 40 healthy adults
Gender: Asthmatic: 8
M/8 F, Healthy: 20 M/20
F
Age: 18-40 yrs
Ultrafine elemental carbon
Particle Size: CMD
— 25 nm
Particle Number/Count:
10/yg/m3: —2.0 x
10e/cm3; 25 /yg/m3:
-7.0 x 106/cm3;
50 /yg/m3: -10.8 x
106/cm3
Concentration: 10, 25,
and 50 /yg/m3
Study conducted using a randomized crossover
design with 2-h exposures. Asthmatics
(n - 16) exposed to filtered air and 10 /yg/m3.
12 healthy adults exposed to filtered air and
10 /yg/m3 at rest; 12 healthy adults exposed to
filtered air, 10 and 25/yg/m3 with intermittent
exercise; 16 healthy adults exposed to filtered
air and 50 /yg/m3 with intermittent exercise.
Exposures were conducted via mouthpiece.
Time to analysis: Immediately following
exposure as well as 3.5, 21, and 45 h post-
exposure.
No PM-induced changes in eNO or cell counts, IL-6, or
IL-8 in induced sputum were observed in any of the
protocols 21 h following exposure. Ultrafine carbon
was not observed to increase respiratory symptoms in
any of the study protocols. Healthy adults experienced
an ultrafine PM-induced reduction in maximal
mid-expiratory flow and CO diffusing capacity relative
to filtered air 21 h following exposure.
Reference: Pourazar et
al. (2005, 088305)
Subjects: 15 healthy
adults
Gender: 11 M/4 F
Age: 21-28 yrs
DE
Idling Volvo diesel engine
Particle Number/Count:
4.3 x 106/cm3
Concentration: PMio
300 /yg/m3
Subjects exposed to DE and filtered air for 1 h
with intermittent exercise (randomized
crossover study design). Other diesel emissions
measured: NO2 (1.6 ppm), NO (4.5 ppm), CO
(7.5 ppm), total hydrocarbons (4.3 ppm),
formaldehyde (0.26 mg/m3).
Time to analysis: 6 h post-exposure.
Exposure to DE significantly increased nuclear
translocation of NF-kB, AP-1, phosphorylated p38,
and phosphorylated JNK in bronchial epithelium
6 h post-exposure.
Reference: Pourazar et
al. (2008,156884)
Subjects: 15 healthy
adults
Gender: 11 M/4 F
Age: 21-28 yrs
DE
Idling Volvo diesel engine
Particle Number/Count:
4.3 x 106/cm3
Concentration: PMio
300 /yg/m3
Subjects exposed to DE and filtered air for 1 h
with intermittent exercise (randomized
crossover study design). Other diesel emissions
measured: NO2 (1.6 ppm), NO (4.5 ppm), CO
(7.5 ppm), total hydrocarbons (4.3 ppm),
formaldehyde (0.26 mg/m3).
Time to analysis: 6 h post-exposure.
Exposure to DE observed to enhance epidermal growth
factor receptor (EGFR) expression in bronchial
epithelium 6 h post-exposure.
Reference: Riechelmann
et al. (2004,180120)
Subjects: 30 healthy
adults
Gender: 11 M/19F
Age: 22-32 yrs
Urban dust
NIST SRM 1649a
Concentration: 150,
500/yg/m3
Subjects exposed to both concentrations of An increase in nasal secretion (nasal cytology) of IL-6
urban dust (nose only exposure system) as well and IL-8 were observed 24 h after exposure to
as filtered air for 3h at rest in a randomized 500 /yg/m3 urban dust,
crossover design. Exposures were separated by
at least 1 wk.
Time to analysis: 30 min, 8 h, and 24 h post-
exposure.
July 2009
C-13
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Findings
Reference: Samet et al.
(2007,156940)
Subjects: Ultrafine
CAPs: 20 healthy adults,
Coarse CAPs: 14 healthy
adults
Gender: Ultrafine CAPs:
11 M/9 F, Coarse CAPs:
8 M/6 F
Age: 18-35 yrs
CAPs (Chapel Hill, NC)
Particle Size: Ultrafine:
0.049 ± 0.009 /jm,
Coarse: 3.59 ± 0.58/jm
Concentration: Ultrafine:
47.0 ± 20.2/yg|m3,
Coarse:
89.0 ± 49.5/yg/m3
Preliminary report comparing effects of
controlled exposures to coarse, fine, and
ultrafine CAPs among healthy adults (3
separate studies). A randomized crossover
design was used in evaluating effects of
coarse CAPs (n — 14) and ultrafine CAPs
(n — 20) relative to filtered air following of 2-h
exposures with intermittent exercise. Results
compared with previous study of controlled
exposure to fine CAPs (Chapel Hill, NC) where
subjects did not serve as their own controls
(Ghio et al„ 2000, 012140)
Time to analysis: 0-20 h post-exposure.
As was shown with fine CAPs, exposure to coarse
CAPs increased the percentage of neutrophils in
bronchoalveolar lavage fluid 20 h following exposure.
Unlike fine CAPs, coarse CAPs did not increase the
percent of monocytes in bronchoalveolar lavage fluid.
Ultrafine CAPs were not shown to affect any markers
of pulmonary inflammation in bronchoalveolar lavage
fluid 18 h after exposure. No CAPs-induced changes in
lung function were observed.
Reference: Samet et al.
(2009,191913)
Subjects: 19 healthy
adults
Gender: 10 M/9 F
Age: 18-35 yrs
Ultrafine CAPs (Chapel
Hii.NC)
Particle Size: <
0.16/ym
Particle Number/Count:
120,662 ± 48,232
particles/cm3
Concentration: 49.8 ±
20/yg/m3
Subjects exposed for 2 h with intermittent 15
periods of exercise to UF CAPs and filtered air
using a randomized crossover study design.
Time to analysis: Immediately following
exposure and 1 and 18 h post-exposure.
No effect of UF CAPs observed on pulmonary function
immediately following exposure or 18 h post-exposure.
IL-8 in BAL fluid was observed to increase with UF
CAPs 18 h post exposure. UF CAPs was not shown to
alter any other inflammatory markers in BAL fluid.
Reference: Schaumann
et al. (2004, 087966)
Subjects: 12 healthy
adults
Gender: 4 M/8 F
Age: Avg 27 ± 2.5 yrs
Fine PM
Collected (filter) from
industrialized and
non-industrialized areas in
Germany
Concentration: 100 /yg
per subject
Bronchoscopic instillation of particles collected Particles collected from the industrialized area
from both areas was conducted in
contralateral lung segments for each subject.
Time to analysis: 24 h post-instillation.
(transition metal-rich) increased the percentage of
monocytes and oxidant radical generation in
bronchoalveolar lavage fluid 24 h after exposure
compared with PM containing less metal.
Reference: Stenfors et
al. (2004,157009)
Subjects: 15 asthmatic
adults, 25 healthy adults
Gender: Asthmatic: 10
M/5 F, Healthy: 16 M/9 F
Age: Asthmatic: 22-52
yrs, Healthy: 19-42 yrs
DE
Volvo diesel engine
Concentration: PMii
108/yg/m3
Subjects were exposed for 2 h with
intermittent exercise to DE and filtered air
using a randomized crossover study design.
Other diesel emissions measured: NO2 (0.7
ppm).
Time to analysis: 1 h after the start of
exposure, immediately following exposure, and
6 h post-exposure.
DE was observed to increase neutrophilia and IL-8 in
bronchial lavage fluid among healthy subjects 6 h
after exposure. Among asthmatic subjects, exposure
to DE did not cause an increase in inflammatory
markers. No diesel-induced change in pulmonary
function was observed during or immediately
following exposure.
Reference: Tunnicliffe et Aerosols of ammonium
al. (2003, 088744)
Subjects: 12 asthmatic
adults, 12 healthy adults
Gender: Asthmatics: 7
M/5 F, Healthy: 5 M/7 F
Age: Asthmatics: avg
35.7 yrs, Healthy: avg
34.5 yrs
bisulfate and sulfuric acid
Particle Size: MMD
0.3/ym
Concentration: 200,
2,000 /yg/m3
Subjects were exposed for 1 h at rest to
ammonium bisulfate (low and high
concentrations), sulfuric acid (low and high
concentrations) and filtered air in a randomized
crossover design. Exposures were separated by
at least 2 wks and were conducted using a
head dome exposure system.
Time to analysis: Immediately following
exposure as well as 5.5-6 h post-exposure.
Neither ammonium bisulfate nor aerosolized sulfuric
acid were not observed to affect lung function or
respiratory systems following exposures to 200 or
2,000 /yglm3 among healthy or asthmatic adults.
Exposures to ammonium bisulfate at both
concentrations resulted in a significant increase in
eNO in the asthmatic subjects.
Table C- 3. Central Nervous System Effects
Reference
Pollutant
Exposure
Findings
Reference: Cruts DE
et al. (2008,
156374)
Subjects: 10
healthy adults
Gender: M
Age: 18-39 yrs
Idling Volvo diesel engine
(TD45, 4.5 L, 4
cylinders, 680 rpm)
Particle
Number/Count: 1.2 x
106/cm3
Concentration:
300/yg/m3
Subjects were exposed to DE and filtered air for 1 h at rest in Exposure to DE was observed to significantly
a randomized crossover study design. Exposures were
separated by 2-4 days. Other diesel emissions measured: NO2
(1.6 ppm), NO (4.5 ppm), CO (7.5 ppm), total hydrocarbons
(4.3 ppm).
Time to analysis: From onset of exposure until 1 h post-
exposure.
increase the median power frequency (MPF) in
the frontal cortex during exposure, as well as in
the hour following the completion of the
exposure.
July 2009
C-14
DRAFT-DO NOT CITE OR QUOTE

-------
Annex C References
Alexis NE; Lay JC; Zeman KL; Geiser M; Kapp N; Bennett WD. (2006). In vivo particle uptake by
airway macrophages in healthy volunteers. Am J Respir Cell Mol Biol, 34- 305-313. 088636
Barregard L; Sallsten G; Andersson L; Almstrand AC; Gustafson P; Andersson M; Ohn AC. (2008).
Experimental exposure to wood smoke- effects on airway inflammation and oxidative stress.
Occup Environ Med, 65- 319-324. 155675
Barregard L; Sallsten G; Gustafson P; Andersson L; Johansson L; Basu S; Stigendal L. (2006).
Experimental exposure to wood-smoke particles in health humans' effects on markers of
inflammation, coagulation, and lipid peroxidation. Inhal Toxicol, 18- 845-853. 091381
Bastain TM; Gilliland FD; Li Y-FJ Saxon A; Diaz-Sanchez D. (2003). Intraindividual reproducibility of
nasal allergic responses to diesel exhaust particles indicates a susceptible phenotype. , 109- 130-
136. 098690
Beckett WS; Chalupa DF; Pauly-Brown A; Speers DM; Stewart JC; Frampton MWi Utell MJ; Huang
LS; Cox C; Zareba W; Oberdorster G. (2005). Comparing inhaled ultrafine versus fine zinc oxide
particles in healthy adults - A human inhalation study. Am J Respir Crit Care Med, 171- 1129-
1135. 156261
Behndig AF; Mudway IS; Brown JL; Stenfors N Helleday R Duggan ST; Wilson SJ; Boman C Cassee
FR; Frew AJ; Kelly FJ; Sandstrom T Blomberg A. (2006). Airway antioxidant and inflammatory
responses to diesel exhaust exposure in healthy humans. Eur Respir J, 27- 359-365. 088286
Blomberg A; Tornqvist H; Desmyter L; Deneys V; Hermans C. (2005). Exposure to diesel exhaust
nanoparticles does not induce blood hypercoagulability in an at-risk population. J Thromb
Haemost, 3: 2103-2105. 191991
Bosson J; Barath S; Pourazar J; Behndig AF; Sandstrom T; Blomberg A; Adelroth E. (2008). Diesel
exhaust exposure enhances the ozone-induced airway inflammation in healthy humans. Eur
Respir J, 31: 1234-1240. 156287
Bosson J; Pourazar J; Forsberg B; Adelroth E; Sandstrom T; Blomberg A. (2007). Ozone enhances the
airway inflammation initiated by diesel exhaust. Respir Med, 10L 1140-1146. 156286
Brauner EV; Forchhammer L; Moller P; Barregard L; Gunnarsen L; Afshari A; Wahlin P; Glasius M;
Dragsted LO; Basu S; Raaschou-Nielsen O; Loft S. (2008). Indoor particles affect vascular
function in the aged: an air filtration-based intervention study. Am J Respir Crit Care Med, 177:
419-425. 156293
Brauner EV; Forchhammer L; Moller P; Simonsen J; Glasius M; Wahlin P; Raaschou-Nielsen O; Loft S.
(2007). Exposure to ultrafine particles from ambient air and oxidative stress-induced DNA
damage. , 115: 1177-82. 188507
Brauner EV; Mortensen J; Moller P; Bernard A; Vinzents P; Wahlin P; Glasius M; Loft S. (2009).
Effects of ambient air particulate exposure on blood-gas barrier permeability and lung function. ,
21: 38-47. 190244
Brook RD; Brook JR; Urch B; Vincent R; Rajagopalan S; Silverman F. (2002). Inhalation of fine
particulate air pollution and ozone causes acute arterial vasoconstriction in healthy adults. ,
105: 1534-1536. 024987
Brauner EV; Moller P; Barregard L; Dragsted LO; Glasius M; Wahlin P; Vinzents P; Raaschou-Nielsen
O; Loft S. (2008). Exposure to ambient concentrations of particulate air pollution does not
influence vascular function or inflammatory pathways in young healthy individuals. Part Fibre
Toxicol, 5:13. 191966
Carlsten C; Kaufman JD; Trenga CA; Allen J; Peretz A; Sullivan JH. (2008). Thrombotic markers in
metabolic syndrome subjects exposed to diesel exhaust. Inhal Toxicol, 20: 917-921. 156323
Carlsten C; Kaufman Joel D; Peretz A; Trenga Carol A; Sheppard L; Sullivan Jeffrey H. (2007).
Coagulation markers in healthy human subjects exposed to diesel exhaust. , 120: 849-855.
155714
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database
(Health and Environmental Research Online) at http://epa. gov/hero. HERO is a database of scientific literature
used by U.S. EPA in the process of developing science assessments such as the Integrated Science Assessments
(ISA) and the Integrated Risk Information System (IRIS).
July 2009
C-15
DRAFT-DO NOT CITE OR QUOTE

-------
Cruts B; van Etten L; Tornqvist H; Blomberg A; Sandstrom T; Mills NL; Borm PJ. (2008). Exposure to
diesel exhaust induces changes in EEG in human volunteers. Part Fibre Toxicol, 5- 4. 156374
Danielsen PH; Brauner EV; Barregard L; Sallsten G; Wallin M; Ohnski R; Rozalski R; Moller P; Loft S.
(2008).	Oxidatively damaged DNA and its repair after experimental exposure to wood smoke in
healthy humans. , 642- 37"42. 156382
Devlin RB; Ghio AJ; Kehrl H; Sanders G; Cascio W. (2003). Elderly humans exposed to concentrated air
pollution particles have decreased heart rate variability. Eur Respir J, 40- 76S-80S. 087348
Fakhri AA; Ilic LM; Wellenius GA; Urch B; Silverman F; Gold DR; Mittleman MA. (2009). Autonomic
effects of controlled fine particulate exposure in young healthy adults' Effect modification by
ozone. Environ Health Perspect, In Press- In Press. 191914
Frampton MWi Stewart JC; Oberdorster G; Morrow PE; Chalupa D; Pietropaoli AP; Frasier LM;
Speers DM; Cox C; Huang LS; Utell MJ. (2006). Inhalation of ultrafine particles alters blood
leukocyte expression of adhesion molecules in humans. Environ Health Perspect, 114- 51-58.
088665
Ghio AJ; Kim C; Devlin RB. (2000). Concentrated ambient air particles induce mild pulmonary
inflammation in healthy human volunteers. Am J Respir Crit Care Med, 162- 981-988. 012140
Gilliland FD; Li YF; Saxon A; Diaz-Sanchez D. (2004). Effect of glutathione-S-transferase Ml and PI
genotypes on xenobiotic enhancement of allergic responses' randomised, placebo-controlled
crossover study. Lancet, 363' 119-125. 156471
Gong H Jri Linn WS; Clark KW; Anderson KR; Geller MD! Sioutas C. (2005). Respiratory responses to
exposures with fine particulates and nitrogen dioxide in the elderly with and without COPD.
Inhal Toxicol, 17: 123-132. 087921
Gong H Jr; Linn WS; Clark KW; Anderson KR; Sioutas C; Alexis NE; Cascio WE; Devlin RB. (2008).
Exposures of healthy and asthmatic volunteers to concentrated ambient ultrafine particles in
Los Angeles. Inhal Toxicol, 20: 533-545. 156483
Gong H Jr; Linn WS; Terrell SL; Anderson KR; Clark KW; Sioutas C; Cascio WE; Alexis N> Devlin RB.
(2004). Exposures of elderly volunteers with and without chronic obstructive pulmonary disease
(COPD) to concentrated ambient fine particulate pollution. Inhal Toxicol, 16: 731-744. 087964
Gong H Jr; Linn WS; Terrell SL; Clark KW; Geller MD; Anderson KR; Cascio WE; Sioutas C. (2004).
Altered heart-rate variability in asthmatic and healthy volunteers exposed to concentrated
ambient coarse particles. Inhal Toxicol, 16: 335-343. 055628
Graff D; Cascio W', Rappold A; Zhou H; Huang Y; Devlin R. (2009). Exposure to concentrated coarse air
pollution particles causes mild cardiopulmonary effects in healthy young adults . Environ Health
Perspect, 117: 1089-1094. 191981
Huang Y-CT; Ghio AJ; Stonehuerner J; McGee J; Carter JD; Grambow SC; Devlin RB. (2003). The role
of soluble components in ambient fine particles-induced changes in human lungs and blood.
Inhal Toxicol, 15: 327-342. 087377
Kongerud J; Madden MC; Hazucha M; Peden D. (2006). Nasal responses in asthmatic and
nonasthmatic subjects following exposure to diesel exhaust particles. Inhal Toxicol, 18: 589-594.
156656
Larsson BM; Sehlstedt M; Grunewald J; Skold CM; Lundin A; Blomberg A; Sandstrom T; Eklund A;
Svartengren M. (2007). Road tunnel air pollution induces bronchoalveolar inflammation in
healthy subjects. , 29: 699-705. 189320
Lucking A; Lundback M; Mills N; Faratian D; Barath S; Pourazar J; Cassee F; Donaldson K; Boon N;
Badimon J. (2008). Diesel exhaust inhalation increases thrombus formation in man. Eur Heart
J, 29: 3043. 191993
Lund AK; Lucero J; Lucas S; Madden MC; McDonald JD; Seagrave JC; Knuckles TL; Campen MJ.
(2009).	Vehicular emissions induce vascular MMP-9 expression and activity associated with
endothelin-l-mediatedpathways. , 29: 511-7. 191159
Lundback M; Mills NL; Lucking A; Barath S; Donaldson K; Newby DE; Sandstrom T; Blomberg A.
(2009). Experimental exposure to diesel exhaust increases arterial stiffness in man. Part Fibre
Toxicol, 6:7. 191967
Mills NL; Robinson SD; Fokkens PH; Leseman DL; Miller MR; Anderson D; Freney EJ; Heal MR;
Donovan RJ; Blomberg A; Sandstrom T; MacNee W', Boon NA; Donaldson K; Newby DE; Cassee
FR. (2008). Exposure to concentrated ambient particles does not affect vascular function in
patients with coronary heart disease. Environ Health Perspect, 116: 709-715. 156766
July 2009
C-16
DRAFT-DO NOT CITE OR QUOTE

-------
Mills NL; Tornqvist H; Robinson SD; Gonzalez M; Darnley K; MacNee W; Boon NA; Donaldson K;
Blomberg A; Sandstrom T; Newby DE. (2005). Diesel exhaust inhalation causes vascular
dysfunction and impaired endogenous fibrinolysis. Circulation, 112- 3930-6. 188557
Mills NL; Tornqvist H; Gonzalez MC; Vink E; Robinson SD; Soderberg S; Boon NA; Donaldson K;
Sandstrom T; Blomberg A; Newby DE. (2007). Ischemic and thrombotic effects of dilute diesel"
exhaust inhalation in men with coronary heart disease. , 357- 1075-1082. 091206
Mudway IS; Stenfors N; Duggan ST; Roxborough H; Zielinski H; Marklund SL; Blomberg A; Frew AJ;
Sandstrom T; Kelly FJ. (2004). An in vitro and in vivo investigation of the effects of diesel
exhaust on human airway lining fluid antioxidants. Arch Biochem Biophys, 423- 200-212.
180208
Peretz A,Peck EC,Bammler TK,Beyer RP,Sullivan JH,Trenga CA,Srinouanprachnah S,Farin
FM,Kaufman JD. (2007). Diesel exhaust inhalation and assessment of peripheral blood
mononuclear cell gene transcription effects' an exploratory study of healthy human volunteers. ,
19: 1107-19. 189082
Peretz A; Kaufman JD; Trenga CA; Allen J; Carlsten C', Aulet MR; Adar SD; Sullivan JH. (2008).
Effects of diesel exhaust inhalation on heart rate variability in human volunteers. Environ Res,
107: 178-184. 156855
Peretz A; Sullivan JH; Leotta DF; Trenga CA; Sands FN; Allen J; Carlsten C; Wilkinson CW; Gill EA;
Kaufman JD. (2008). Diesel exhaust inhalation elicits acute vasoconstriction in vivo. Environ
Health Perspect, 116: 937-942. 156854
Pietropaoli AP; Frampton MW; Hyde RW; Morrow PE; Oberdorster G; Cox C; Speers DM; Frasier LM;
Chalupa DC; Huang LS; Utell MJ. (2004). Pulmonary function, diffusing capacity, and
inflammation in healthy and asthmatic subjects exposed to ultrafine particles. Inhal Toxicol, 16:
59-72. 156025
Pourazar J; Blomberg A; Kelly FJ; Davies DE; Wilson SJ; Holgate ST; Sandstrom T. (2008). Diesel
exhaust increases EGFR and phosphorylated C"terminal Tyr 1173 in the bronchial epithelium.
Part Fibre Toxicol, 5: 8. 156884
Pourazar J; Mudway IS; Samet JM; Helleday R; Blomberg A; Wilson SJ; Frew AJ; Kelly FJ; Sandstrom
T. (2005). Diesel exhaust activates redox-sensitive transcription factors and kinases in human
airways. Am J Physiol, 289: L724-L730. 088305
Power K; Balmes J; Solomon C. (2008). Controlled exposure to combined particles and ozone decreases
heart rate variability. J Occup Environ Med, 50: 1253. 191982
Riechelmann H; Rettinger G; Lautebach S; Schmittinger S; Deutschle T. (2004). Short-term exposure to
urban dust alters the mediator release of human nasal mucosa. J Occup Environ Med, 46: 316-
322. 180120
Routledge HC; Manney S; Harrison RM; Ayres JG; Townend JN. (2006). Effect of inhaled sulphur
dioxide and carbon particles on heart rate variability and markers of inflammation and
coagulation in human subjects. Heart, 92: 220-227. 088674
Rundell KW! Caviston R. (2008). Ultrafine and fine particulate matter inhalation decreases exercise
performance in healthy subjects. , 22: 2-5. 191986
Samet JM; Graff D; Berntsen J; Ghio AJ; Huang YC; Devlin RB. (2007). A comparison of studies on the
effects of controlled exposure to fine, coarse and ultrafine ambient particulate matter from a
single location. Inhal Toxicol, 19 Suppl l: 29"32. 156940
Samet JM; Rappold A; Graff D; Cascio WE; Berntsen JH; Huang YC, Herbst M; Bassett M; Montilla T;
Hazucha MJ; Bromberg PA; Devlin RB. (2009). Concentrated ambient ultrafine particle
exposure induces cardiac changes in young healthy volunteers. Am J Respir Crit Care Med, 179:
1034-1042. 191913
Schaumann F; Borm PJA; Herb rich A; Knoch J; Pitz M; Schins RPF; Luettig B; Hohlfeld JM; Heinrich
J; Krug N. (2004). Metal-rich ambient particles (particulate matter 25) cause airway
inflammation in healthy subjects. Am J Respir Crit Care Med, 170: 898-903. 087966
Shah AP; Pietropaoli AP; Frasier LM; Speers DM; Chalupa DC; Delehanty JM; Huang LS; Utell MJ;
Frampton MW. (2008). Effect of inhaled carbon ultrafine particles on reactive hyperemia in
healthy human subjects. Environ Health Perspect, 116: 375-380. 156970
Stenfors N; Nordenhall C; Salvi SS; Mudway I; Soderberg M; Blomberg A; Helleday R; Levin JO;
Holgate ST; Kelly FJ; Frew AJ; Sandstrom T. (2004). Different airway inflammatory responses
in asthmatic and healthy humans exposed to diesel. Eur Respir J, 23: 82"86. 157009
July 2009
C-17
DRAFT-DO NOT CITE OR QUOTE

-------
Tornqvist H; Mills NL; Gonzalez M; Miller MR; Robinson SD; Megson IL; MacNee WJ Donaldson K;
Soderberg S> Newby DE; Sandstrom T> Blomberg A. (2007). Persistent endothelial dysfunction in
humans after diesel exhaust inhalation. Am J Respir Crit Care Med, 176- 395-400. 091279
Tunnicliffe WS; Harrison RM; Kelly FJ; Dunster C; Ayres JG. (2003). The effect of sulphurous air
pollutant exposures on symptoms, lung function, exhaled nitric oxide, and nasal epithelial lining
fluid antioxidant concentrations in normal and asthmatic adults. Occup Environ Med, 60.
088744
Urch B; Brook JR; Wasserstein D; Brook RD; Rajagopalan S> Corey P; Silverman F. (2004). Relative
contributions of PM2.5 chemical constituents to acute arterial vasoconstriction in humans. Inhal
Toxicol, 16: 345-352. 055629
Urch B; Silverman F; Corey P; Brook JR; Lukic KZi Rajagopalan S; Brook RD. (2005). Acute blood
pressure responses in healthy adults during controlled air pollution exposures. Environ Health
Perspect, 113: 1052-1055. 081080
Zareba W', Couderc JP; Oberdorster G; Chalupa D; Cox C; Huang LS; Peters A; Utell MJ; Frampton
MW. (2009). ECG parameters and exposure to carbon ultrafine particles in young healthy
subjects. Inhal Toxicol, 21: 223-233. 190101
July 2009
C-18
DRAFT-DO NOT CITE OR QUOTE

-------
Annex D. Toxicological Studies
Table D-1.
Cardiovascular effects.


Study
Pollutant
Exposure
Effects
Reference: Anselme
et al. (2007, 097084)
Species:
Rat
Gender:
Male
Strain: Wistar
Age: Adult
Weight: 200-225g
DE: monocylinder Diesel engine using
Euro 4 ELF 85A reference gasoline
Particle Size: DE: 10-650 nm (85
nm mean mobility diameter)
Route: Whole-body Inhalation
Dose/Concentration: DE: 0.5 mgfm3; Other
emissions measured: non-methane
hydrocarbons (7.7 ppm), NO2 (1.1 ppm), CO
(4.3 ppm)
Time to Analysis: Experiments started 3
months after L coronary artery ligation. ECG
started at tO and the DE exposure at t30 min
for a 3h period; ventricular premature beats
(VPBs) and RMSSD calculated every 30 min
during clean room air exhaust and PE periods.
Early (t210-300 min) and late (t480-540 min)
PE were analyzed.
Immediate decrease in RMSSD was observed in
both healthy and CHF rats PE. Immediate
increase in VPBs observed in CHF rats only:
which lasted 4-5h after exposure ceased.
Whereas HRV progressively returned to
baseline values within 2.5 h post-exposure (PE),
the proarrhythmic effect persisted as late as 5
h PE termination in CHF rats
Reference:
al. (2004, 055638)
Species:
Rat
Gender:
Male
Strain: SH
Age: 13-15wks
et LPS and EHC-93 (PM): Urban Air Route: IT Instillation
collected at the Health Effects
Institute Ottawa, Canada
Particle Size: EHC-93: 0.8-0.4/ym
(mean) (range: < 3 //m)
Dose/Concentration: PM: 10 mg/kg of bw;
LPS- 350 EU/animal
Time to Analysis: Sacrificed 4 or 24 h post-
instillation
PM and LPS elicited a significant increase in
receptor-dependent vasorelaxation of the aorta
compared to saline-instilled rats.
Reference: Bagate et EHC-93 (PM),
al. (2004, 055638)
Species: Rat
Gender: Male
Strain: SH
Age: 13-15wks
CB-V or CB-Fe, LPS
Particle Size: EHC-93: 0.8-0.4/ym
(mean) (range: < 3/ym)
Route: Aortic Suspension Fluid
Dose/Concentration: Cumulative
concentrations of EHC-93, CB-V and CB-Fe
(10,25, 50, 75,100//g/mL)
CB 1.5-2.0 nm (mean) (range < 5//m)
Time to Analysis: Immediately post-exposure
of aortic rings to cumulative concentrations of
EHC-93, CB-V, CB-Fe and LPS.
CB-V particles induced more relaxation than
CB-Fe particles or EHC-93 in a dose-dependent
manner. PM and LPS had an acute transient
effect on the receptor dependent
vasorelaxation. PM and LPS attenuated ACh-
elicited vasocontraction in denuded aortic rings
(DARs).
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database
(Health and Environmental Research Online) at http://epa. gov/hero. HERO is a database of scientific literature
used by U.S. EPA in the process of developing science assessments such as the integrated Science Assessments (ISA)
and the Integrated Risk Information System (IRIS).
July 2009
D-1
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: E
al. (2004, 055638)
Species:
Rat
Gender:
Male
Strain: Wistar Kyoto
Aqe: 13-15wks
EHC-93 (PM): Urban Air collected at
the Health Effects Institute Ottawa,
Canada.
EHC-93 filtrate (PMF)
Zn2+ and Cu2+ particles (10,000 and
845 //g PM respectively)
Particle Size: PM: 4.6 //m
(GSD - 3.2)
Route: In Vitro
Dose/Concentration: PM Suspensions: 10-
100//glmL; CuS04/ZnS04 1-100//mol; Phe 2
//m; arbacol: 10//m
Time to Analysis: Measured immediately
after maximum response for each cumulative
dose was achieved.
PM-lnduced Contraction: No effect of
suspension or filtrate seen on resting tension of
aorta and SMRA.
PM- and Metal-Induced Vasorelaxation:
Cumulative concentrations (10-100 //g/mL) of
PM suspension and its water soluble
components (PMF) elicited dose-dependent
relaxation in aorta. Relaxation induced by
particle suspension was higher than relaxation
induced by free filtrate. The difference was
significant at 100 //glmL. In SMRA, vasore-
laxation similar to aorta's was observed, and
the activity of the particle suspension was
stronger than the filtrate, with the difference
being significant starting at 30 //g/mL. Both
Zn2+ and Cu2+ in sulfate salts (10-100//mol)
induced relaxation in pre-contracted aortic
rings, with Cu2+ having a greater effect than
Zn at the same concentration. Ions didn't
affect ACh relaxation.
Effect of PM on a-Adrenergic Contraction:
Phenylephrine-induced dose-response
contraction, starting at 1//M with max at 100
//mol. Pretreatment of SMRA did not change
the phenylephrine-induced contraction.
Reference: Bagate et
al. (2006, 097608)
Species:
Rat
Gender:
Male
Strain: Wistar Kyoto
(WKY) and SH
Age: 13-15wks
EHC-93 (PM)
EHC-93 (Filtrate)
Cu2+ and Zn2+ solutions
Particle Size: PM: 4.6 //m
(GSD - 3.2)
Route: In Vitro
Dose/Concentration: PM and PMF
Suspensions: 10-100//g/mL; CuS04 or
ZnS04:10-100 //mol; Phenylephrine: 2 //m;
Carbacol: 10//m
Time to Analysis: Responses evaluated at
maximum of each dose-response.
PM and its soluble components elicited
endothelium-independent vasodilation in rat
aorta rings. This response is a result of the
activation of sGC since its inhibition by
NS2028 practically eliminated relaxation. PM
suspensions stimulated cGMP production in
purified isolated sGC. Neither receptor nor their
signaling pathways played a significant role in
the direct relaxation by PM or metals.
Vasodilation responses were significantly
higher in SH than WKY control rats.
Reference: Bagate et
al. (2006, 096157)
Species:
Rat
Gender: Male
Strain: SH/NHsd
Age: 11-12 wks
Weight: 250-350g
pretreated with PM or LPS the isolated heart
had a reduced ability to recover to baseline
levels after occlusion, in comparison with saline
treated rats. After occlusion was released CF
went back to baseline values. Saline and LPS
treated rats, showed a gradual decrease in CF
noted during the reperfusion period. Isolated
hearts from PM-exposed SH showed a
complete restoration of CF and no gradual
decrease. The increase of Zn2+ elicited a rapid
decrease of LDVP and HR. The impairment of
cardiac function measured by LDVP and HR
started immediately upon Zn2+ infusion and
remained the same during the perfusion period
(no Zn2 + was present in the perfusate).
EHC-93 (PM): Urban Air collected at Route: IT Instillation
the Health Effects Institute Ottawa,
Canada.
EHC-93 (Filtrate),
Zinc (in PM), LPS
Particle Size: PM: 4.6//m
(GSD - 3.2)
Dose/Concentration: PM: 10 mg/kg of bw;
LPS: 350 EU/animal (0.5 mL)
Time to Analysis: 4h post-exposure
Effect of Pretreatment on Baseline
parameters of Isolated Perfused Heart:
After PM exposure a slight increase of baseline
coronary flow (CF) and heart rate (HR) was
noted. In contrast, a significant decrease of left
developing ventricular pressure (LDVP) was
observed in SH. LPS also elicited a non-
significant decrease in LVDP.
Effect of Pretreatment and Ischemia on
Cardiac Function: When SH rats were
Reference: Bagate et
al. (2006, 096157)
Species:
Rat
Strain: H9c2 (EACC),
EHC-93 (PM) Filtrate: Urban Air col-
lected at the Health Effects Institute
Ottawa, Canada,
ZnS04
Particle Size: PM: 4.6 //m
Route: In Vitro
Dose/Concentration: PM: 1, 50,100 //g/mL;
ZnS04: 50 //mol
Time to Analysis: 30min incubation
Effect of EHC-93 filtrate on Ca2+ Uptake
in Cardiomyocytes: Both PMF and Zn2 +
inhibited ATP or ionophore-stimulated Ca2 +
influx in cardiomyocytes.
cardiomyocyte cells (GSD - 3.2); Carbon Particles: 44nm
July 2009
D-2
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Bartoli et
al. (2009,156256)
Species: Dog
Gender: Female
Strain: Mixed breed
Age: 2-12yrs
Weight: Average:
15.7kg, Range: 13.6-
18.2kg
CAPs (Boston: Harvard Ambient
Particle Concentrator)
Particle Size: Diameter: 0.15-2.5
/jm
Route: Permanent Tracheostomy
Dose/Concentration: Concentration range
and mean: CAPs: 94.1-1557(358.1 ±306.7)
/yg/m3, BC: 1.3-32(7.5±6.1)/yg/m3, Particle
count: 3000-
69300(18230 ±13.151 articles/cm3
Time to Analysis: Preanesthetized.
Tracheostomy. 5h exposures separated by
minimum 1wk. Prazosin administered in 8 of
13 dogs 30-60min before exposure. 55
exposure days.
CAPs significantly increased SBP, DBP, mean
arterial pressure, HR and rate-pressure product.
Prazosin (o-adrenergic antagonist) decreased
these CAPs-induced effects. CAPs mass, BC,
particle number concentrations were positively
and significantly associated with each of the
cardiovascular parameters except for pulse
pressure.
Reference: Bartoli et
al. (2009,179904)
Species: Dog
Gender: Female
Strain: Mixed breed
Age: Adult
Weight: 14-18kg
CAPs (Boston: Harvard Ambient
Particle Concentration)
Particle Size: Diameter: ~2.5/ym
Route: Permanent Tracheostomy
Dose/Concentration: Concentration range
and mean: CAPs: 94.1-1556.8 (349±282.6)
/yg/m3, BC: 1.3-32 (7.5±5.6)/yg/m3, Particle
number: 3000-69300 (20381 ±13075)//g/m3
Time to Analysis: Tracheostomy. Minimum
3wk recovery. Acclimatized. Exposed 5h. 2
5min occlusions of LAD coronary artery
separated by 20min rest. Exposure days
separated by 1wk minimum.
During coronary artery occlusion, CAPs
exposure reduced myocardial blood flow and
increased coronary vascular resistance, SBP
and DBP. CAPs effects were greater in
ischemic tissue than nonischemic. Increases in
CAPs mass, particle number and BC
concentrations were significantly associated
with decreased myocardial blood flow and
increased coronary vascular resistance.
Reference: Campen
et al. (2005, 083977)
Species: Mouse
Gender: Male
Strain: C57BL/6J and
Apo E'1'
Age: 10-12wks
High Whole DE (HWDE); Low Whole
DE (LWDE): High PM Filtered (HPMF);
Low PM Filtered (LPMF)
Particle Size: NR
Route: Whole-body Inhalation Chambers and
Ex-vivo Exposures (isolated, pressurized septal
coronary arteries)
Dose/Concentration: HWDE: PM - 3.6
mg/m3: NOx - 102 ppm
LWDE: PM - 0.512 mg/m3; NOx - 19 ppm:
PM - 0.770 mg/m3: NOx - 105 ppm
LPMF: PM - 0.006 mg/m3: NOx - 26 ppm
Time to Analysis: Whole-body Exposures: DE
or PFDE for 6h/day for 3 days, euthanized at
the end of last exposure.
Coronary Vessels Exposure: PSS bubbled with
DE to expose coronary vessels to the soluble
contents of DE. Analysis occurred
immediately post exposure.
Whole-body Exposure on ApoE': During DE
exposure, ApoE'1' mice HR consistently
decreased during high concentration exposures,
compared to the C57BL/6J strain.
Coronary Vascular Effects on ApoE': DE
had no significant effects on the resting
myogenic tone of isolated septal coronary
arteries. Control coronary arteries showed
constrictive responses to ET-1 and dilatory
responses to SNP. DE exposed PSS vessels
responses to ET-1 enhanced compared to
control. SNP-induced dilation blunted in vessels
resting in diesel-exposed saline.
Reference: Campen
et al.(2003, 055626)
Species: Rat
Gender: Male and
Female
Strain: SH
Aqe: 4 m
DE: generated by either of two
Cummins (2000 model) 5.9-L ISB
turbo engines fueled by Number 2
Diesel Certification Fuel.
Particle Size: 0.1-0.2 /jm
aerodynamic diameter
Route: Whole-body exposure
Dose/Concentration: 0, 30,100, 300,1000
/yg/m3
Time to Analysis: 6 h/day for 7 days: ECG
measurements taken 4 days post-exposure.
HR: Significantly higher in exposed animals and
not concentration-dependent. More substantial
results seen in male rats.
ECG: The PQ interval was significantly
prolonged among exposed animals in a
concentration-dependent manner.
Reference: Campen
et al. (2006, 096879)
Species: Mouse
Gender: Male
Strain: ApoE1
Age: 10wks
Road dust from paved surfaces
(Reno, NV)
Gasoline engine emissions, containing
PM, NOx, CO and HC
Particle Size: Road dust: 1.6 //in
(Standard Deviation 2.0)
Gasoline engine emissions: Average
particle diameter of 15 nm
Route: Whole-body inhalation
Dose/Concentration: Road dust: 0.5 and 3.5
mg/m3:
Gasoline engine emissions: 5 to 60 /yg/m3 (at
dilutions of 10:1,15:1, and 90:1)
Mean concentrations of PM: 61 /yg/m3; NOx:
18.8 ppm: CO: 80 ppm.
Time to Analysis: 6h/d for 3d. Sacrificed
18h post-exposure.
ET-1: Gasoline exhaust significantly
upregulated ET-1 in a dose-dependent manner.
ET-1 increased levels in the PM filtered group
and decreased in the low levels of road dust.
ECG: HR consistently decreased from beginning
to end of exposure in all groups. No significant
HR effects on road dust or gasoline exposure
was observed. No significant effects on P-
wave, PQ-interval, QRS-interval, or QT-interval
were observed in either treatment.
T-wave: Mice exposed to whole gasoline
exhaust displayed significant increases in T-
wave morphology from the beginning of
exposures: this effect was consistent on all
exposure days.
July 2009	D-3	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Cascio et
al. (1987, 007583)
Species:
Mouse
Gender:
Male
Strain: ICR
Aqe: 6-10wks
UFPM: Ultra fine PM, EPA Chapel
Hill, NC
Particle Size: <0.1 /ym
Route: IT Instillation
Dose/Concentration: 100 //g in 100 ul
Time to Analysis: 24h post exposure (single
exposure)
UFPM exposure double the size of myocardial
infarction attendant to an episode of ischemia
and reperfusion while increasing post ischemic
oxidant stress. UFPM alters endothelium-
dependent/independent regulation of systemic
vascular tone; increases platelet number,
plasma fibrinogen, and soluble P-selectin levels;
reduces bleeding time.
Reference: Chang et
al. (2007,155720)
Species:
Rat
Gender:
Male
Strain: SH
Aqe: 60d
UfCB: Ultra fine carbon blackFerric
sulfate Fe2(S04>3
Nickel sulfate NiS04
Particle Size: UfCB
Route: IT Instillation
Dose/Concentration: UfCB: 415 and 830 //g
Ferric Sulfate: 105 and 210 /yg
Nickel Sulfate: 263 and 526 /yg
Combined UfCB and ferric sulfate: 830 /yg
UfCB + 105/yg Ferric Sulfate
Combined UfCB with Nickel Sulfate: 830 /yg
UfCB + 263 /yg Nickel Sulfate
Time to Analysis: Single dose,
radiotelemetry readings recorded for 72h post
exposure.
Both high/low-dose UfCB decreased ANN
(normal-to-normal intervals) slightly around the
30th hour, concurrent increases of LnSDNN.
LnRMSSD returned to baseline levels after
small initial increases. Minor effects observed
after low-dose Fe and Ni instilllation; biphasic
changes occurred after high-dose instillations.
Combined exposures of UfCB and either Fe or
Ni resulted in HRV trends different from values
estimated from individual-component effects.
Reference: Chang et
al. (2007,155720)
Species: Rat
Gender: Male
Strain: SH
Aqe: 10wks
CAPs: collected during a dust storm
from Chung-Li, Taipei
Particle Size: PM2.6
Route: Nose-only Inhalation
Dose/Concentration: 315.55 //g/m3
Time to Analysis: 6h
A linear mixed-effects model revealed sigmoid
increases in HR and a sigmoid decrease of QAI
during exposure, after an initial incubation
period.
Reference: Chang et
al. (2004, 055637)
Species:
Rat
Gender: Male
Strain: SH
Aqe: 60d
CAPs collected in Chung-Li, Taipei
(spring and summer periods)
Particle Size: PM2.6
Route: Nose-only Inhalation
Dose/Concentration: Spring exposure: 202.0
± 68.8 /yg/m3; Mean number concentration:
2.30 x 106 particles/cm3 (range: 7.12 x 103 -
8.26 x 106)
Summer exposure: 141.0 ± 54.9/yg/m3;
Mean number concentration: 2.78 x 106
particles/cm3 (range: 7.76 x 103 - 8.87 x 106)
Time to Analysis: 4d of spring exposure and
6d of summer exposure for 5h each exposure.
Parameters measured throughout duration of
exposures.
During spring exposures, the maximum increase
of heart rate (HR) and blood pressure (BP) were
51.6 bpm and 8.5 mmHg respectively. The
maximum decrease of QAI (measures cardiac
contractility) noted at the same time was 1.6
ms. Similar pattern was observed during
summer exposure, however., the responses
were less prominent.
Reference: Chang, et CAPs collected in Chung-Li, Taipei
al. (2005, 097776) Partjc|e Sjze. pM26 (o.1-2.5/ym)
Species: Rat
Gender: Male
Strain: SH
Weight: 200g
Route: Nose-only Inhalation
Dose/Concentration: 202.0 ± 68.8 /yglm3
Time to Analysis: 5hfd for 4d
During the inhalation stage, crude effects of
both LnSDNN and LnRMSSD for exposure and
control groups decreased from the baseline
values. Immediately after the experiments, both
LnSDNN and LNRMSSD decreased due to
stresses produced by release from the exposure
system, then returned to the baseline values.
Reference: Chauhan
et al. (2005,155722)
Tumor Cell Line:
A549 derived from
alveolar type II
epithelial cells
SRM-1879 (Si02) and SRM-154b
(Ti02) from the NIST
EHC-93 from Ontario, Canada
(EHCsol, EHCinsol)
Particle Size: EHC-93 median
physical diameter: 0.4 /ym; Ti02 and
SiCh particle size distribution: 0.3-0.6
/ym
Route: Cell Culture
Dose/Concentration: 0,1, 4, and 8 mg EHC
total equivalent per 5 mL
Time to Analysis: Culture medium was
removed from flasks and replaced wI 5 mL of
the particle suspension media. Plates were
incubated for 24h. After 24h cell culture
supernatants were collected and analyzed.
The decreased expression of preproET-1 in
A549 cells suggests that epithelial cells may
not be the source of higher pulmonary ET -1
spillover in the circulation measured in vivo in
response to inhaled urban particles. However,
higher levels ECE-1 in A549 post-exposure to
particles suggests an increased ability to
process bigET-1 into mature ET-1 peptide,
while increased receptor expression implies
responsiveness. The increased release of II-8
and VEGF by epithelial cells in response to
particles could possibly up regulate ET-1
production in the adjacent pulmonary capillary
endothelial cells, with concomitant increased
ET-1 spillover in the systemic circulation.
July 2009	D-4	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Chen et
al. (2005, 087218)
Species: Mouse
Strain: Normal (C57)
and ApoE"1
CAPs (NYU, NY)
Particle Size: PM2.B
Route: Whole-body Inhalation
Dose/Concentration: 10 x ambient
concentrations
19.7/yglm3 average concentration over 5
months (daily average exposure concentration
was 110/yg/m3)
Time to Analysis: 6hfd, 5 dfwk, for 5m.
Parameters measured continuously
throughout.
Significant decreasing patterns of HR, body
temperature, and physical activity for ApoE'1'
mice, with nonsignificant changes for C57
mice. SDNN and RMSSD in the late afternoon
and overnight for ApoE'1' mice showed a gradual
increase for the first 6 weeks, a decline for
about 12 more weeks, and a slight turn upward
at the end of the study period. For C57 mice,
there were no chronic effect changes in SDNN
or RNSSD in the late afternoon, and a slight
increase after 6 weeks for the overnight period.
Reference: Chen LC
and Nadziejkov C,
(2005,087219)
Species: Mouse
Strain: Normal (C57),
ApoE'1',
Age: 26-28wks (C57),
39-41 wks (ApoE 1),
and 18-20wks (LDLr
[DK])
CAPs (NYU, NY)
Particle Size: PM2.6
Route: Whole-body Inhalation
Dose/Concentration: Mean exposure
concentration: 110/yg/m3
Time to Analysis: 6hfd, 5dfwk for up to 5m.
Sacrificed 3 to 6d after last exposure.
All DK mice developed extensive lesions in the
aortic sinus regions. In male DK mice, the lesion
areas appeared to be enhanced by CAPs
exposure. Plaque cellularity was increased, but
there were no CAPs-associated changes in the
lipid content. ApoE '' and DK mice showed
prominent areas of severe atherosclerosis.
Quantitative measurements showed that CAPs
increased the percentage of aortic intimal
surface covered by grossly discernible
atherosclerotic lesion.
Reference: Corey LM
et al. (2006,15G3GG)
Species: Mouse
Gender: Male
Strain: ApoE1
Age: 11-12m
Weight: 32.84g (avg)
PM collected November - March
between 1996-1999(Seattle, WA)
Silica (U.S. Silica Company, Berkeley
Springs, WV)
Particle Size: PM2.6
Route: Nasal Instillation
Dose/Concentration: PM: 1.5 mg/kg; Saline:
50ul; Silica: Min-u-Sil 5 in 50 ul saline
Time to Analysis: Mice monitored for 1 d
baseline prior to and for 4d following
exposure.
After an initial increase in both HR and activity
in all groups, there was delayed bradycardia
with no change in activity of the animals in the
PM and silica exposed groups. In addition, with
PM and silica exposure, there was a decrease
in HRV parameters.
Reference: Cozzi et
al. (2006, 091380)
Species: Mouse
Strain: ICR
Aqe: 6-10wks
Ultrafine PM (collected continuously
over 7 day periods in Oct 2002 in
Chapel Hill, NC)
Particle Size: < 150nm
Route: IT Instillation
Dose/Concentration: 100 /yg of PM in
vehicle
Time to Analysis: 24h post-exposure
Ischemia-Reperfusion: PM exposure doubled
the relative size of myocardial infarction com-
pared with the vehicle control. No difference
was observed in the percentage of the vehicle
at the risk of ischemia. PM exposure increased
the level of oxidative stress in the myocardium
after l-R. The density of neutrophils in the
reperfused myocardium was increased by PM
exposure, but differences in the numbers of
blood leukocytes, expression of adhesion
molecules on circulating neutrophils, and acti-
vation state of circulating neutrophils, 24 h
after PM exposure, could not be correlated to
the increase l-R injury observed.
Isolated Aortas: Aortas isolated from PM-
exposed animals exhibited a reduced
endothelium-dependent relaxation response to
ACh.
Reference: Dvonch
JT et al. (2004,
055741)
Species: Rat
Gender: Male
Strain: Brown
Norway
CAPs, Detroit, Ml
Particle Size: PM2.6
Route: Whole-body Inhalation Chamber
Dose/Concentration: Average concentration
354 /yglm3
Time to Analysis: 8hfd for 3 consecutive
days; plasma samples collected 24h post-
exposure.
Plasma concentrations of asymmetric
dimethylarginine (ADMA) were significantly
elevated in rats exposed to CAPs versus
filtered air.
July 2009
D-5
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Elder et
al. (2004, 055642)
Species: Rat
Gender: Male
Strain: Fischer 344
and SH
Age: 23m (Fischer);
11 -14m (SH)
Weight: NR
UFP ¦ Ultrafine carbon particles;
LPS (Sigma)
Particle Size: UFP: 36nm (median
size)
Route: Intraperitoneal Injection (ip) for saline
and LPS
Whole-body exposure for inhaled particles
Dose/Concentration: Particles: 150 mgfm3;
LPS: 2mg/kg bw
Time to Analysis: Single 6h exposure to
particles. Sacrificed 24h after ip LPS
exposure.
BAL fluid cells: Neither inhaled UFP nor ip LPS
cause a significant increase in BAL fluid total
cells or the percentage of neutrophils in either
rat strain. No significant exposure-related
alteration in total protein concentration or the
activites of LDH and b-glucuronidase.
Peripheral blood: In both rat strains ip LPS
induced significant increase in the number and
percentage of circulating PMNs. When
combined with inhaled UFP, PMNs decreased,
significantly for F-344 rats. Plasma fibrinogen
increased with ip LPS in both rat strains with
the magnitude of change greater in SH rats.
UFP alone decreased plasma fibrinogen in SH
rats. Combined UFP and LPS response was
blunted, but significantly higher than controls.
Hematocrit was not altered in either rat strain
by any treatment.
TAT complexes: With all exposure groups
averaged, plasma TAT complexes in SH rats
were 6.5 times higher than in F-344 rats. LPS
caused an overall increase in TAT complexes
for F-344 rats that was further augmented by
inhaled UFP. UFP alone decreased response. In
SH rats, UFP alone significant increased
response and LPS decreased response.
ROS in BAL cells: In F-344 rats both UFP and
LPS have independent and significant effects
on DCFD oxidation. Effects were in opposite
directions; particles decreased ROS, LPS
increased ROS.
Reference: Finnerty
et al. (2007,156434)
Species: Mouse
Gender: Male
Strain: C57BL/6
Age: 9 wks
Weight: 22-2g
Coal Fly Ash (U.S. EPA), Analysis:
(PM2.E samples) low unburned carbon
(0.53 wt%), moderate levels of
transition metals, including (in //gig):
Fe (30, 400), Mg (31, 200), Ti (6,
180 ), Mn (907), and V (108).
Particle Size: 1.8 and 2.5/ym
Route: IT Instillation
Dose/Concentration: PM: 200/yg; PM + 10
/yg LPS: 200/yg; PM + 100/yg LPS: 200 //g
Time to Analysis: 18h after IT instillation
Plasma: TNF-a significantly increased in both
PM+LPS10 and PM+LPS100 treatments. For
plasma IL-6, all groups tended to rise with a
significant increase in the PM+LPS100 group.
Reference: Floyd et
al. (2009,190350)
Species: Mouse
Gender: Male
Strain: ApoE'1'
Aqe: 20wks
CAPs: PM2.6 concentrated from
Tuxedo, NY (April-Sept 2003)
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: Avg 120 //gfm3
(n-6/group)
Time to Analysis: 6h/d x 5dfwk * 5m
Gene Expression: Microarry gene expression
identified 395 genes downregulated and 216
genes upregulated in the aortic plaques.
Ontologic analysis identified a list of functional
processes associated with gene expression and
included: inflammation, tissue development,
cellular movement, cellular growth and
proliferation, hematological system
development and function, lipid metabolism,
cardiovascular system function, cellular
assembly and organization, and cell death.
Reference: Floyd et
al. (2009,190350)
Species: Mouse
Gender: NR
Strain: ApoE'
Age: 6wks
Weight: NR
CAPs (Tuxedo, NY) (April-September
2003) (modified VACES system)
Particle Size: NR
Route: Exposure Chamber
Dose/Concentration: Average concentration:
120 /yglm3
Time to Analysis: Exposed 6hfd, 5dfwk, 5m.
CAPs altered 611 genes of which 306 were
significantly related to alterations in molecular
pathways and associated with biological
pathways. The 50 most significant biological
function alterations related to CVD. Further
analysis of recent literature showed that some
CAPs-altered genes were the same as genes
altered by CVD or unstable-human plaque gene
expression.
Reference: Folkmann
et al. (2007, 097344)
Species: Mouse
Gender: Female
Strain: Wild type and
ApoE'1'
Age: 11-13wks
Weight: 21 g (avg)
DEP: SRM2975 (particulate fraction
of exhaust from a filtering system
designed for diesel-powered
forklifts).
Particle Size: DEP: NR
Route: Intraperitoneal Injection
Dose/Concentration: 0, 50, 500, 5,000 /yg
DEP/kg of bw
Time to Analysis: 6 or 24 h post-ip injection
The expression of inducible nitric oxide
synthase (iNOS) mRNA was increased in the
liver 6h post-ip injection. The level of oxidized
purine bases, determined by formami-
dopyrimidine DNA glycosylase sites increased
significantly in the liver after 24 h in in mice
injected w/ 50/yg/kg of bw. There was no
indication of systemic inflammation determined
as the serum concentration of nitric oxide and
iNOS expression, and DNA damage was not
increased in the aorta.
July 2009	D-6	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Furuyama
et al. (2006, 097056)
Species: Rat
Cell Type: Heart
Micro vessel
Endothelial (RHMVE)
Cells
OE-DEP, OE-UFP (from Urawa,
Saitama, Japan)
OE - Organic Extracts
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 0, 5,10, 25//gfmL of
OE-DEP or OE-UFP
Time to Analysis: Exposed for 0, 6,12, 24,
or 36h
The cell monolayer exposed to 10 /yglmL OE-
UFP produced a larger amount of HO-1 than
cells exposed to 10 /yglmL OE-UFP. OE-DEP and
OE-UFP exposure reduced PAI-1 production by
the cells but did not affect the production of
thrombomodulin, tissue-type PA, or urokinase-
type PA. Increased PAI-1 synthesis in response
to treatment with 1 ng/mL TNF-a or 0.5ng/mL
TGF (31 was reduced by OE-DEP exposure.
Suppression of PAI-1 production by OE-DEP
exposure was mediated through oxidative
stress and was independent of HO-1 activity.
Reference: Gerlofs-
Nijland et al. (2009,
190353)
Species: Rat
Gender: Male
Strain: SH
Age: 12wks
Weight: 200-300g
PM (Prague, Czech Republic;
Duisburg, Germany; Barcelona, Spain)
(Prague and Barcelona coarse PM
organic extracts)
Particle Size: Coarse: 2.5-10/ym,
Fine: 0.2-2.5 fjm
Route: IT Instillation
Dose/Concentration: 7mg PM/kg body
weight
Time to Analysis: DTPA added to some PM
samples preinstillation. Instilled with PM.
Necropsy 24h postexposure.
Inflammation (LDH, protein, albumin),
cytotoxicity (NAG, MP0, TNF-a), and
fibrinogen were increased by PM, and were
greatest in the coarse PM fraction. Metal-rich
PM had greater inflammatory and cytotoxic
effects, and increased fibrinogen and vWF and
decreased ACE. PAH content influenced greater
inflammation (including neutrophils),
cytotoxicity, and fibrinogen. Generally, whole
PM and coarse PM were more potent than
organic extracts and fine PM, respectively.
Reference: Ghelfi et
al. (2008,156468)
Species: Rat
Strain: Sprague-
Dawley
Aqe: Adult
CAPs
CPZ (Capsazepine) (Axxora LLC, San
Diego, CA)
Particle Size: PM2.6
Route: CAPs: Inhalation via Whole-body
Exposure; CPZ: IP Injection or Aerosol
Dose/Concentration: CAPs: mean mass
concentration: 218 ±23/yglm3; CPZ:
10mg/kg (ip), 500 /ymol (aerosol)
Time to Analysis: Experiment 1: CPZ ip or
20min aerosol pretreatment immediately prior
to CAPs exposure. Single CAPs exposure for
5h. Parameters measured immediately
following exposure.
Experiment 2: CPZ ip pretreatment prior to
CAPs exposure. Exposed to CAPs for 5h/day
for4mths. Parameters measured throughout
duration of experiment.
CPZ (ip or aerosol) decreased CAPs-induced
chemiluminescence (CL), lipid thiobarbituric acid
reactive substances (TBARS), and edema in the
heart, indicating that blocking TRP receptors,
systemically or locally, decreases heart CL.
CAPs exposure led to significant decreases in
HR and in the length of QT, RT, Pdur and Tpe
intervals. These changes were observed
immediately upon exposure, and were
maintained throughout the 5h period of CAPs
inhalation. Changes in cardiac rhythm and ECG
morphology were prevented by CPZ.
Reference: Gilmouret
al. (2004, 054175)
Species: Rat
Gender: Male
Strain: Wistar (Crl:
Wl) BR
ufCB (Printex 90 from Frankfurt,
Germany)
fCB (Huber 990 from UK)
Particle Size: ufCB: 114 nm
(MMAD); fCB: 268 nm (MMAD).
Route: Whole-body Inhalation
Dose/Concentration: ufCB: 1.66 mgfm3;
fCB: 1,40mg/m3
Time to Analysis: Single exposure for 7h
Sacrified and samples taken at 0,16, and 48h equivalent antioxidant status
post-exposure.
Exposure to ultrafine, but not fine, CB particles
was also associated with significant increases
in the total number of blood leukocytes. Plasma
fibrinogen factor VIII and vWF were unaffected
by particle treatments as was plasma Trolox
Reference: Gilmour
PS et al.(2005,
087410)
Species: Human
Cell Types: Primary
Human Monocyte
Derived Macrophages
(MP); Human Umbilical
Vein Endothelial Cells
(HUVEC); A549 cells;
Human Bronchial
Epithelial Cells
(16HBE)
PM in: (Carbon Black from Degussa
Ltd, Frankfurt, Germany)
Particle Size: PMio
Route: Cell Culture
Dose/Concentration: PMio: 50 and
100/yg/mL
Time to Analysis: 6 and 20h
The culture media from MPs and 16HBE cells
but not A549 cells, exposed to PMio had an
enhanced ability to cause clotting. H202 also
increased clotting activity. Apoptosis was
significantly increased in MPs exposed to PMio
and LPS as shown by annexin V binding. TF
gene expression was enhanced in MPs exposed
to PMio and HUVEC tissue factor. tPA gene
and protein expression were inhibited.
July 2009
D-7
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Gilmour et Zinc Sulfate
al. (2006,156472) (ZnS04 in saline solution)
Species: Rat	Particle Size: NR
Gender: Male
Strain: Wistar Kyoto
Age: 12-14wks
Weight: 280-340g
Route: IT Instillation
Dose/Concentration: 131 /yglkg of bw (2
//mol/kg)
Time to Analysis: 1, 4, 24, 48h
Zinc levels in plasma and tissue: At 1 -24 h
post exposure, zinc plasma levels increased to
nearly 20% above baseline.
mRNA expression: Cardiac tissues demon-
strated similar temporal increases in
expressions of TF, PAI-1 and thrombomodulin
mRNA, following pulmonary instillation of Zn.
Cardiac histopathology: Mild and focal acute,
myocardial lesions developed in a few Zn ex-
posed rats. No changes in fibrin deposition or
troponin disappearance were observed. At 24
and 48h PE to Zn, increases occurred in levels
of systemic fibrinogen an the activated partial
thromboplastin time.
Reference: Gong et
al. (2007, 091155)
Species: Human and
Mouse
Cell Type: Human
Microvascular
Endothelial Cells
(HMEC)
Strain: C57BL/6J
Gender: Male (mouse)
Age: 2m (mouse)
Organic DEP extract: collected from
exhaust in a 4JB1 -type LD, 2.74
liter, 4-cylinder Isuzu diesel engine
(provided by Masaru Sagai, Tsukuba,
Japan)
ox-PAPC: (provided by Judith
Berliner, UCLA, CA)
In vivo validation: Ultrafine (ufp) and
fine (fp) particulate matter
Particle Size: DEP < 1 /jm
(diameter)
Route: Cell culture; In vivo validation via
Whole-body inhalation
Dose/Concentration: ox-PAPC: 10, 20, and
40/yglmL; DEP: 5,15, and 25//gfmL; DEP (5
/yg/mL)+ox-PAPC: 10 or 20/yglmL
In Vivo Valudation: Ufp: 3.24x106/cm3; fp:
2.7x106|cm3
In vivo validation: Ufp: <0.18//m;fp: <2.5
/jm
Time to Analysis: 4h
In vivo validation: Exposed to CAPs for
5h/day, 3d/wk for 8wks. Sacrificed 24h after
last CAPs exposure.
Gene-expression profiling showed that both
DEP extract and ox-PAPC co-regulated a large
number of genes. U sinfg network analysis to
identify co-expressed gene modules, led to the
discovery of three modules that were highly
enriched in genes that were differentially
regulated by the stimuli. These modules were
also enriched in synergistically co-regulated
genes and pathways relevant to vascular
inflammation.
In vivo validation: Results were validated by
demonstrating that hypercholesterolemic mice
exposed to ambient ultrafine particles inhibited
significant upregulation of the module genes in
the liver.
Reference: Goto et al
(2004,088100)
Species: Rabbit
Gender: Female
Strains: NZW
Age: NR
Weight: 2.3 kg
Cytokine release: EHC stimulation increased
the release of GM-CSF, IL-6, IL-1(3, TNF-a, IL-8
and MCP-1. No effect on m-CSF and MIP-1 p.
CC particles induced increases in IL-6 and TNF-
a; other cytokine levels did not differ from
control.
Supernatant instillation: AM + EHC increased
circulating WBC and band cell counts.
Circulating monocyte counts were unaffected.
AM+EHC showed a major increase in fraction
and amount of monocyte released as well as
faster clearance when compared to control.
The BM monocyte pool was similar in all
groups.
Monocyte transit time through BM:
Exposure to EHC, CC only shortened the transit
time of monocytes as compared to controls.
AM+EHC also shortened monocyte transit time
whereas AM+CC had a nonsignificant effect.
EHC-93 (Ottawa, ON,Canada)
CC: Coilloidal Carbon (obtained from
Hamburg, Germany)
Particle Size: EHC-93: PMio; CC: <
1 /jm
Route: Intrabronchial Instillation
Dose/Concentration: AMs incubated with
EHC-93 or CC: 0.6 ml/kg
EHC-93 alone: 1 mL (500//g/ mL)
CC alone: 1mL (1% CC)
Time to Analysis: WBC counts measured 4 -
168h after BrdU injection. Sacrificed 7d post
instillation.
Lung distribution of PMio: PM-containing
AMs were distributed diffusely. PM-containing
AMs were more prevalent in the PM exposed
animals. There was no AM-containing particle
difference between the CC-exposed and EHC-
93-exposed groups.
Monocyte release from Bone Marrow: EHC
exposure increased WBC and band cell counts
from 12hours after instillation. Monocyte count
was not affected. Labeled monocytes peaked
more quickly after DEP exposure (12h vs 16 h
for control). There was no observed change in
BM monocyte pool.
July 2009
D-8
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Gottipolu
RR et al.(2009,
190360)
Species: Rat
Gender: Male
Strain: Wistar Kyoto
(WKY), SH
Age: 14-16wks
Weight: NR
DE (30-kW (40hp) 4-cylinder indirect
injection Deutz diesel engine) (O2-
20%, CO- 1.3-4.8ppm, NO- <2.5-
5.9ppm, N02- < 0.25-1.2ppm, S02-
0.2-0.3ppm, 0C/EC- 0.3±0.03)
Particle Size: Number Median
Diameter: Low- 83±2nm, High-
88.2nm; Volume Median Diameter:
Low- 207±2nm, High- 225±2nm
Route: Exposure Chamber
Dose/Concentration: Low- 507±4//g/m3,
High- 2201 ±14//g/m3
Time to Analysis: Exposed 4hfd, 5d/wk,
4wks. Necropsied 1d postexposure.
DE increased neutrophils in a concentration-
dependent manner, and GGT activity at the
high dose. Particle-laden macrophages were
found in DE-exposed rats. DE dose-dependently
inhibited mitochondrial aconitase activity. DE
caused 377 genes to be differentially
expressed within WKY rats, most of which
were downregulated, but none in SH rats.
However, WKY rats had an expression pattern
shift that mimicked baseline expression of SH
rats without DE. These genes regulated
compensatory response, matrix metabolism,
mitochondrial function, and oxidative stress
response.
Reference: Graff et Zinc (Zn); Vanadium (
al. (2005, 087956) Partjc|e Sjze. NR
Species: Rat
Cell Type: Ventricular
Myocytes
Route: Cell Culture
Dose/Concentration: 0, 6.25,12.5, 25, or
50 //m
Time to Analysis: Toxicity: 24h post
exposure
Beat Rate: 0.5,1, 2, 4, and 24h PE
PCR: 6 and 24h PE
Beat Rate: There were statistically significant
reductions in spontaneous beat rate 4 and 24 h
post-exposure (greater reductions were
observed with Zn).
Inflammation: Exposure to Zn or V (6.25-50
//m) for 6h produced significant increases in IL-
6, IL-a, heat shock protein 70, and connexin 43
(Cx43).
Impulse Conduction: 24 h post-exposure, Zn
induced significant changes in the gene ex-
pression of Kv4.2 and KvQLt, a-1 subunit of L-
type Ca channel, Cx43, IL-6, and IL-1 a. V pro-
duced a greater effect on Cx43 and affected
only KvLQT 1.
Reference: Gunnison
and Chen (2005,
087956)
Species: Mouse
Gender: Male
Strain: F2 generation
DK (ApoE1, LDLr'1')
Age: 18-20wks
CAPs (Tuxedo, NY)
Copollutants measured: O3 and NO2.
Particle Size: 389 ± 2nm
Route: Whole-body Inhalation
Dose/Concentration: CAPs: 131 ±99
//g/m3
(range 13-441 //g/m3)
0:i: 10 ppb
NO2: 4.4 ppb
Time to Analysis: 6h/d, 5d/wk for
approximately 4m. Tissue collection was
performed 3-4d after the last day of
exposure.
Gene expression: In CAPs-exposed heart
tissue, the expression of Limdl and Rex3 were
the most consistently affected genes among
the exposed mice. Limdl was down regulated
by 1.5-fold or greater from moderate baseline
expression. Rex3 showed a relatively small
increase in absolute expression.
Reference: Gurgueira
et al. (2002, 036535)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Weight: 250-300g
CAPs:
Carbon Black (CB from Fisher
Scientific, Pittsburgh, PA): C (85.9 ±
0.2%); 0(13 ± 0.2%); S (1.17 ±
0.02%)
ROFA: obtained from an oil-fired
power plant (Boston, MA)
Particle Size: CAPs size range: 0.1-
2.5 //m; CB and ROFA (PM2.6)
Route: Whole-body Exposure Chamber
Dose/Concentration: CAPs: average mass
concentration: 300 ± 60//g/m3; ROFA: 1.7
mg/m3; CB: 170 //g/m3
Time to Analysis: CAPs: 1, 3, and 5h; ROFA:
30min; CB: 5h
Oxidative stress: Rats breathing CAPs
aerosols for 5 h showed significant oxidative
stress, determined as in situ
chemiluminescence (CL) in the lung, heart, but
not in the liver. ROFA also triggered increases
in oxidant levels but not particle-free air or CB.
Increases in CL showed strong associations
with the CAPs content of Fe, Al, Si and Ti in
the heart. The oxidant stress imposed by 5h
exposure to CAPs was associated with slight,
but significant increases in the lung and heart
water content, with increased serum levels of
lactate dehydrogenase, indicating mild damage
to tissues. CAPs inhalation also led to tissue-
specific increases in the activities of SOD and
catalase.
Reference: Gursinsky
et al. (1976, 015607)
Species: Rat
Cell Type: Fibroblasts
isolated from adult
male Wistar rats
hearts
Fly ash (TAF98)
Particle Size: Nl
Route: In Vitro
Dose/Concentration: TAF98: 0,1, 2 3,10,
25, 50,100, 200 //g/mL
Time to Analysis: 0, 5,10, 30, 60,120min
Brief treatment of fibroblasts with fly ash
triggered the immediate formation of ROS.
Using phosphospecific antibodies the activation
of p38 MAP kinase, p44/42 MAP kinase
(ERK1/2) and p70S6 kinase. Prolonged
incubation with fly ash increased the
expression of collagen 1 and TGF-(31, but
decreased mRNA levels of MMP9 and TNF-a.
Cell proliferation was inhibited at high
concentrations of fly ash. An increase in the
level of advanced glycation end product
modification of various cellular proteins was
observed.
July 2009	D-9	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Hansen et DEP: SRM-2975 (NIST)
al. (2007, 090703)
Species: Mouse
Gender: Female
Strain: ApoEand
C57BL/6J ApoE+l+
Aqe: 11-13wks
Particle Size: DEP: 215nm
(geometric mean diameter)
Route: Intraperitoneal Injection
Dose/Concentration: DEP: 0, 0.5 and 5
mg/kg of bw; Aorta segments incubated with
0,10 and 100 f/g DEP/mL
Time to Analysis: Sacrificed 1h after ip
injection.
Exposure to 0.5 mg/kg DEP caused a decrease
in the endothelium-dependent Ach elicited
vasorelaxation in ApoE'1' mice, whereas the re-
sponse was enhanced in ApoE+/+ mice. No
significant changes were observed after
administration of 5 mg/kg DEP. K+ or
phenylephrine induced constriction was not af-
fected.
Reference: Hansen et
al. (2007, 090703)
Species: Mouse
Gender: Female
Strain: ApoE ' and
C57BL/6J ApoE+l+
Use: Aorta rings used
for in-vitro studies
DEP: SRM-2975 (NIST)
Particle Size: DEP: 215nm
(geometric mean diameter)
Route: Cell Culture
Dose/Concentration: 0,10 and 100/yg
DEP/mL
Time to Analysis: Basal tone measured at 5
different points throughout experiment.
Exposure to 100//g DEP/mL enhanced ACh-
induced relaxation and attenuated
phenylephrine-induced constriction.
Vasodilatation induced by sodium nitroprusside
was not affected by any DEP exposure.
Reference: Harder et
al. (2005, 087371)
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 12-15wks
Carbon UFPs ((generated by Electric
Spark Generator GFG 1000; Palas,
Karlsruhe, Germany)
Particle Size: 37.6 ± 0.7nm (mean)
Route: Whole-body Exposure
ChamberDose/Concentration: 180 /yg/m3
Time to Analysis: Days 1-3: baseline read-
ing,
Day 4: exposure to UFPs or filtered air for 4
or 24h then sacrificed immediately following
exposure period OR
Sacrificed following 1 -3d recovery period.
Cardiovascular Performance: Mild but
consistent increase in heart rate (HR), which
was associated with a significant decrease in
HR variability during exposure (particle-induced
alteration of cardiac autonomic balance,
mediated by a pulmonary receptor activation).
Lung Inflammation and Acute-Phase Re-
sponse: BALF revealed significant but low-
grade pulmonary inflammation.
Effects on Blood: There was no evidence of
an inflammation-mediated increase in blood
coagulability; no changes in plasma fibrinogen
or factor Vila.
Pulmonary and Cardiac Histopathology:
Sporadic accumulation of particle-laden
macrophages found in the alveolar region. No
signs of cardiac inflammation or cardio-
myopathy.
mRNA Expression Levels: No significant
changes in the lung or heart.
Reference: Harder et
al. (2005, 087371)
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 14-17wks
Weight: NR
Carbon UFP
Particle Size: Diameter:
37.6±0.7nm
Route: Exposure Chamber
Dose/Concentration: 180/yg/m3
Time to Analysis: Telemeter implanted into
peritoneal cavity. 10d recovery. 3d baseline
reading. 24h exposure. 3d recovery.
Carbon UFP mildly but significantly elevated HR
compared to the control. SDNN was
significantly decreased during exposure. UFP
induced mild pulmonary inflammation,
significantly increased PMN, and increased the
total protein and albumin concentrations.
Particle-laden macrophages sporadically
accumulated in the alveolar region.
Reference: Hirano et	Organic Extracts of DEP (DEP) and
al. (2003, 097345)	Organic Extracts of Ultra Fine
Species: Rat	Particles (UFP).
Cell Types: Heart	(Urawa CitV' Saitama' JaPan'
Microvessel	Particle Size: DEP and UFP: < 2.0
Endothelial Cells	/jm
(RHMVE)
Route: Cells Culture
Dose/Concentration: NAC effects on
viability: DEP: 25 /yg/ml; UFP: 50 /yg/ml
mRNA levels for DEP and UFP: 0,1,3,10
/yg/ml
cell monolayer exposed to DEP and UFP:
1,10,100 //g/ml
Time to Analysis: mRNA levels measured
after 6h incubation with DEP or UFP. Other
parameters measured after 24h.
Cytotoxicity and Oxidative Stress: LC50
values were 17 and 34/yg/mL for DEP and UFP
respectively. The viability of DEP and UFP
exposed cells was ameliorated by N-acetyl-L-
cysteine (NAC).
mRNA levels: mRNA levels increased dose-
dependently with DEP and H0-1 mRNA showed
the most marked response to DEP. mRNA
levels of antioxidant enzymes and heat shock
protein 72 (HSP72) in DEP-exposed cells were
higher than UFP exposed cells at the same
concentration. The transcription levels of H0-1
and HSP72 in DEP and UFP-exposed cells were
also reduced by NAC.
July 2009	D-10	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Hwang et
al. (2005, 089454)
Species: Mouse
Strain: Normal (C57)
and ApoE"1
CAPs (Tuxedo, NY)
Particle Size: 389 ± 2nm
Route: Whole-body Inhalation
Dose/Concentration: CAPs Range: 5-627
/yg/m3. Mean CAPs Concentration: 133//g/m:
Mean Concentrations of Ozone and Nitrogen
in CAPs: 10 and 4.4 ppb respectively.
Time to Analysis: 6hfday, 5dfwk for 5m.
Long-term Analysis: Significant decreasing
patterns of heart rate (HR), body temperature
^ (T), and physical activity (PA) in ApoE'1' mice.
Nonsignificant changes for C57 mice. The
chronic effect changes for HR, T, and PA for
ApoE'1' mice were maximal in the last three
weeks.
Short-term Analysis: Dose-dependent
relationship for HR variations in ApoE'1' mice.
Heart Rate Fluctuation: HR fluctuations in
Apo E : mice during the period of 3-6 h
increased by 1.35 fold at the end of the
exposure and during a 15 min period increases
by 0.7 fold at the end of the exposure.
Reference: Inoue et DEP (obtained from a 4Jb1-type light- Route: IT Instillation
al. (2006,190142) duty, 4-cylinder, 2.74-L Isuzu diesel
Species: Mouse
Gender: Male
Strain: ICR
Age: 6-7wks
engine)
Washed DEP (carbonaceous nuclei of
DEP after extraction) and DEP-0C
(organic chemicals in DEP extracted
with CH2CI2); Washed DEP + LPS and
DEP-0C+LPS
Particle Size: PM2.6
Dose/Concentration: Washed DEP: 4mgfkg
bw. DEP-0C: 4mg/kg bw. LPS: 2.5mg/kg.
Washed DEP+LPS and DEP-0C +LPS:
respective additions of LPS to each
component prior-experimentation.
Time to Analysis: Sacrificed 24h post single
dose instillation.
Both DEP components exacerbated vascular
permeability. The increased fibrinogen and E-
selectin levels induced by LPS. This
exacerbation was more prominent with washed
DEP than with DEP-0C. Washed DEP+LPS
significantly decreased protein C and
antithrombin-lll and elevated circulatory levels
of IL-6, KC and LPs without significance.
Reference: Ito et al.
(2008, 096823)
Species: Rat
Gender: Male
Strain: Wistar Kyoto
(Specific pathogen-
free)
Age: 13-14 wks
CAPs (f-PM) from Inhalation Facility
in Yokohama City, Japan.
Particle Size: 0.1-2.5 /jm
Route: Whole-body Inhalation
Dose/Concentration: 0.6-1.5 mgfm3
Time to Analysis: Three groups exposed to:
(1) filtered air for 4d, (2) filtered air for 3d
and CAPs for 1 d or (3) CAPs for 4d. All
groups exposed for a maximum of 4.5hfd for 4
consecutive days.
mRNA expression and cardiovascular
function: In samples of heart tissue, the mRNA
of cytochrome P450 (CYP) 1B1, heme
oxygenase-1 (HO-1), and endothelin A (ETa)
receptor were up-regulated by CAPs; their
levels were significantly correlated with the
cumulative weight of CAPs in the exposure
chamber. The up-regulation of ETa receptor
mRNA was significantly correlated with the
increase in H0-1 mRNA and weakly with the
increase in MBP.
Reference: Khandoga
A et al. (2004,
087928)
Species: Mouse
Gender: Female
Strain: C57B1/6
Age: 5-7wks
UFPs: Ultra fine carbon black
particles (Printex 90)
Particle Size: 14nm diameter (60 [
<100nm)
Route: Aortic Infusion
Dose/Concentration: 1x10' and 5x10'
total particles infused
300 m2/g surface area
Time to Analysis: Single exposure, analysis
2h post exposure
Platelet effects: Application of UFPs caused
significantly enhanced platelet accumulation on
endothelium of postsinusodal venules and
sinusoids in healthy mice. UFP-induced platelet
adhesion was not preceded by platelet rolling
but was strongly associated with fibrin deposi-
tion and an increase in vWF expression on the
endothelial surface.
Inflammatory effects: In contrast,
inflammatory parameters such as the number
of rolling/adherent leukocytes, P-selectin
expression/translocation, and the number of
apoptotic cells were not elevated. UFPs did not
affect sinusoidal perfusion and Kupffer cell
function.
Reference: Knuckles
et al. (2007,156652)
Species: Rat
Gender: Female
(Pregnant, purchased
at gestation day 19)
Strain Sprague-
Dawley:
Age: 60-90d
Weight: 300g
Use: RMCs were
harvest from 1d old
neonatal pups using
the neonatal rat
cardiomyocyte
isolation kit
ROFA-L: Leachate
Particle Size: < 0.2/jm
Route: Cell Culture
Dose/Concentration: 3.5 //gfmL
Time to Analysis: 1h
ROFA-L induced alterations to the RCM
transcriptosome: 38 genes were suppressed
and 44 genes were induced PE. Genomic
alterations in pathways related to IGF-1, VEGF,
IL-2, PI3/AKT, CVD, and free radical scavenging
were detected. Global gene expression was
altered in a manner consistent with cardiac
myocyte electrophysiological remodeling,
cellular oxidative stress and apoptosis.
ROFA-L induced alterations to the RCM
transcription factor proteome: ROFA-L
altered the transcription factor proteome by
suppressing activity of 24 and activating 40
transcription factors out of 149.
July 2009	D-11	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Knuckles
et al. (2008,191987)
Species: Mouse
Gender: Male
Strain: C57BL/6
Age: 8-10wks
Weight: NR
DE (single cylinder Yanmar diesel
generator burning #2 certified diesel
fuel (Chevron-Phillips, Borger, TX)
under 100% load)
Particle Size: PM2.6
Route: Whole-body Exposure. Ex Vivo.
Dose/Concentration: In vivo: 350//g/m3; Ex
vivo: PM2.6 concentration 2-3mg/m3 flow rate
500mL/min
Time to Analysis: Exposed 4h. Ex vivo
assays.
Veins: DE increased vascular reactivity to ET-
1. Ex vivo exposed vessels had greater
vasoconstriction. L-NAME (an arginine blocker)
did not promote constriction in DE-exposed rats
but did so in controls.
Arteries: DE did not significantly alter vascular
reactivity. Carbonyls or alkanes alone or with
DE did not alter vasoconstriction.
Reference: Kodavanti
et al. (2008,155907)
Species: Rat
Gender: Male
Strain: Wistar Kyoto
(WKY)
Age: 12-14wks
G1: saline (control); G2: Mount Saint
Helen's ash (SH); G3: whole suspen-
sion of oil combustion PM at high
concentration (PM-HD); G4: whole
suspension of oil combustion PM at
low concentration (PM-LD); G5:
saline-leachable fraction of PM high-
concentration suspension; G6: ZnS
7H20
Particle Size: PM2.6
Route: IT Instillation
Dose/Concentration: Doses (mg/kg/week)
are for 8 and 16 weeks (PM-solid and soluble
Zn) respectively. G1: 0.00-0.00 and 0.00-
0.00; G2: 4.60-0.00 and 2.30-0.00; G3: 4.60-
66.8 and 2.30-33.4; G4: 2.30-33.4 and 1.15-
16.7; G5: 0.00-66.8 and 0.00-33.4; G6: 0.00-
66.8 and 0.00-33.4
Time to Analysis: 1 x/wk for 8 or 16wks;
analyzed 48h after last instillation.
DNA damage (left ventricular tissue): All
groups except MSH caused varying degrees of
damage relative to control. Total cardiac
aconitase activity was inhibited in rats
receiving soluble Zn. Analysis of heart tissue
revealed modest changes in mRNA for genes
involved in signaling, ion channels function, oxi-
dative stress, mitochondrial fatty acid metabo-
lism, and cell cycle regulation in Zn, but not
MSH-exposed rats.
Reference: Kyoso M DE
etal. (2005,186998) pM and N0
x exposures
Species: Rat	Particle Size: NR
Gender: NR
Strain: NR
Age: 15m
Route: Whole-body Inhalation
Dose/Concentration: PM (mgfm3): 0.01,
0.109,0.54,1.09, 0.01 (from 1.09 con-
centration w/o PM)
NOx (ppm): 0.19, 0.59, 2.60, 5.53, 5.47 (w/o
PM)
Time to Analysis: Exposed 16h/d (from 5pm-
9am) for 7m
All of the resting R-R intervals before exposure
were lower at night than during the day, but
few changes were found after exposure.
Reference: Lei YC et
al. (2005, 088660)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Weight: 200-250g
(upon arrival)
Use: ip STZ (60 mg/kg
bw) dissolved in citric
acid buffer admin-
istered to 8 rats to
induce diabetes; ip
citric acid buffer
administered to 8 non-
diabetic rats
CAPs: Hsin-Chuang, Taipei
Particle Size: PM: 0.01 ¦ 2.5/jm
Route: IT Instillation
Dose/Concentration: PM2.6: 200/yg in 0.5
mL saline. Components (/L/g/m3):
Organic Carbon (9.8-SD 2.4)),
Elemental Carbon (3.6-SD 3.2),
Sulfate (4.8-SD 1.2), Nitrate (6.3-SD 3.4)
Time to Analysis: Single dose. Animals
sacrificed 24h post instillation.
Effects of Diabetes: Body weight (bw) of
diabetic (D) rats (397.5g) was lower than non-
diabetic (ND) rats (483.1 g). Mean plasma
glucose level was 163 mg/dL in ND rats and
448.2 mg/dL in D rats. D rats had significant
greater levels of 8-OHdG in plasma compared to
ND rats. D rats had significantly increased
levels of plasma [nitrate+nitrite]. No observ-
able changes in TNF-a for D and ND rats.
Effects of PM Exposure ND Rats: Increase in
plasma levels of 8-OHdG and plasma IL-6, TNF-
a, and serum CRP. Significant reduction of
plasma [nitrate+nitrite]. No significant effect
on plasma ET-1.
Effects of PM Exposure STZ-D Rats:
Significant elevation of plasma ET-1. Decrease
in plasma [nitrate+nitrite] Plasma 8-OHdG and
TNF-a significantly increased. No significant
alterations in IL-6 and CRP.
Reference: Lemos M
et al. (2006, 088594)
Species: Mouse
Gender: NR
Strain: BALB/c
Age: 1d (neonatal)
n: 10
Weight: 4-6g
PM10, CO, NO2, and SO2 from
Universidade de Sao Paulo, Brazil.
Particle Size: PM10
Route: Inhalation via Whole-body exposure
Dose/Concentration: Mean (± SD)
concentrations were: CO2: 2.06 ± 0.08ppm
(8h mean); NO2: 104.75 ± 42.62/yg/m3 (24h
mean); SO2:11.07 ± 5.32/yg/m3 (24h mean);
PM10: 35.52 ± 12.84/yg/m3 (24h mean)
Time to Analysis: 24 h/d, 7d/wk for 4m
Morphometric measurements of the ratio
between the lumen and the wall (L/W) areas
were performed on transverse sections of
renal, pulmonary and coronary arteries. A
significant decrease of L/W with exposure to
air pollution was detected in pulmonary and
coronary arteries, whereas no effects of air
pollution were observed in renal vessels.
July 2009	D-12	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Li et al.
(2005, 088647)
Species: Human and
Rat
Strain: Sprague-
Dawley Rats
Tissues/Cell Types:
Cultured HPAECs;
Pulmonary Artery
Rings (PARs)
Urban Particles (UPs SRM 1648)
Major Constituents (mass fraction in
%): Al (3.4), Fe (3.9), K (1.1).
Minor Constituents (mass fraction in
%): Na (0.43), Pb (0.66), Zn (0.48).
Trace Constituents (ng/mg): As (115),
Cd (75), Cr (403), Cu (609),
Mn (786), Ni (82), Se (27), U (5.5),
V (127).
Particle Size: NR
Route: PARs: In vitro organ model HPAECs:
grown to 80% confluence
Dose/Concentration: PARs and HPAECs: 1
to 100 /yglmL; Losartan treatment: 0.2 /ymol
Captopril treatment: 100 /ymol
Time to Analysis: PARs were exposed to
increasing doses of UPs from 1 to 100/yglmL.
Maximum tension was recorded within 5min
after each UPs dose. HPAECs: exposed to
UPs from 1 to 100 /yglmL for up to 20min
Effects of UPs on the constriction of isolated
rat pulmonary PARs and the activation of
extracellular signal-regulated kinases 1 and 2
(ERK1/2I and p38 mitogen-activated protein
kinases (MAPKs) in HPAECs with or without
Losartan at 1 -100 //gf'niL induced acute
vasoconstriction. UPs also produced a time-
and dose-dependent increase in phosphorylation
of ERK1/2 and p38 MAPK. Losartan pre-treat-
ment inhibited both vasoconstriction and
activation of ERK1/2 and p38. The water
soluble fraction of UPs was sufficient for in-
ducing ERK1/2 and p38 phosphorylation, which
was also inhibited by Losartan. Cu (CuS04) and
V (VOSOi), induced pulmonary vasoconstriction
and phosphorylation of ERK1/2 and p38, but
only phosphorylation of p38 was inhibited by
Losartan. UPs induced activation of ERK1/2
and p38 was attenuated by Captopril.
Reference: Li et al.
(2006,156693)
Species: Rat, Rabbit,
and Human
Tissues/Cell Types:
Pulmonary Artery
Rings (PARs) (rat);
isolated buffer-infused
lungs (rabbits) and
cultured HPAECs
Strain: Sprague
Dawley Rats, New
Zealand White Rabbits
Weight: Rat: 200-
350g; Rabbit: 2.5-
3.0kg
Urban Particles (UPs SRM 1648).
Major Constituents (mass fraction in
%): Al (3.4), Fe (3.9), K (1.1).
Minor Constituents (mass fraction in
%): Na (0.43), Pb (0.66), Zn (0.48).
Trace Constituents (ng/mg): As (115),
Cd (76), Cr (403), Cu (609),
Mg (786), Ni (82), Se (27), U (5.5), V
(127).
Particle Size: NR
Route: In Vitro
Dose/Concentration: PARs and HPAECs: 1
to 100/yg/mL;
Time to Analysis: PARs: treatment given
15min prior to exposure. Exposed to
increasing doses of UPs from 1 to 100/yglmL.
Maximum tension was recorded within 5min
after each UPs dose. HPAECs: exposed to
UPs from 1 to 100 /yglmL for 20 and 120min.
Effects of UP on H2O2 release: Within
minutes after UPs treatment, HPAEC increased
H202 production that could be inhibited by DPI,
AP0, and NaN3. The water soluble fraction of
UPs as well as its two transition metal
components Cu and V, also stimulated H202
production. NaN3 inhibited H202 production
stimulated by Cu and V, whereas DPI and AP0
inhibited only Cu-stimulated H202 production.
Inhibitors of other H202-producing enzymes,
including N-methyl-L-arginine, indomethacin,
allopurinol, cimetidine, rotenone, and antimycin,
had no effects.
Effects of UP-induced H2O2 on MAPK
activation: DPI but not NaN3 attenuated UPs-
induced pulmonary vasoconstriction and
phosphorylation of ERK1/2 and p38 MAPKs.
Knockdown of p47phox gene expression by
small interfering RNA attenuated UPs-induced
H202 production and phosphorylation of
ERK1/2 and p38 MAPKs.
Reference: Lippmann
et al. (2005, 087453)
Species: Mouse
Strain: C57 and
ApoE'1'
(March-September 2003). Chemical
Composition: regional secondary
sulfate (SS) characterized by high S,
Si, and organic C; resuspended soil
(RS) characterized by high
concentrations of Ca, Fe, Al, and Si;
RO-fired powered emissions of the
Eastern U.S. identified by the
presence of V, Ni, and Se; and motor
vehicle (MV) traffic and other
sources. Contributors to Average
Mass: SS (56.1%), RS (11.7%), RO
combustion (1.4%), MV traffic and
other sources (30.9%)
Particle Size: PM2.6
Route: Whole-body Inhalation
Dose/Concentration: PM2.6 concentrated
ten-fold, producing an average of 113/yg/m3
Time to Analysis: 6hfd, 5dfwk for 5m.
Parameters measured daily: during exposure,
the afternoon after exposure, and late at
night
Associations between sources and short-
term Heart Rate changes: There were no
significant associations betweeb SS, RS, R0,
and MV factors and HR in C57 mice at any of
the three intervals. There were significant as-
sociations between PM2.6 and the RS source
factor and decreases in HR for the ApoE'1' mice
during the daily CAPs exposures but no
associations with the other factors. There was
no residual association of HR with PM2.6 or the
RS factor later in the afternoon or late at night.
In the afternoon, there was a significant asso-
ciation between decreases in HR and the SS
factor for the ApoE'1' mice that had not been
present during exposure and did not persist into
the night time period. MV traffic and others
were not significantly associated with HR
during any of these three time periods. For the
C57 mice, there were no significant associa-
tions of HR with PM2.6 or any of its
components during any of the three daily time
periods.
Associations between sources and short-
term HRV changes: Signal noise during
exposures did not permit reliable analyses of
HRV changes during the hours of CAP
exposure.
July 2009	D-13	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Lippmann
et al. (2005, 087453)
Species: Mouse
Gender: NR
Strain: ApoE'1', ApoE'1'
LDLr'1', C57BL/6
Age: NR
Weight: NR
CAPs (Sterling Forest, spring-summer
2003)
Particle Size: PM2.6
Route: Inhalation
Dose/Concentration: PM2.6 average
concentration: 110 uglm3, Long-term
average: 19.7 /yglm
Time to Analysis: Exposed 6hfd, 5dfwk, 5 or
6m. Semicontinuous EKG recordings.
HR increased in ApoE'1'mice but not C57 mice.
HRF increased over the duration of the
experiment. Atherosclerotic plaque deposits
and coronary artery disease lesions occurred in
both CAPs-exposed mice and controls, but
invasive lesions were only present in CAPs-
exposed mice. A gene affecting circadian
rhythm was upregulated in double knockout
mice. CAPs activated NF-kB. No inflammation
occurred in the pulmonary system.
Reference: Lippman
M et al. (2006,
091165)
Species: Mouse
Gender: Male
Strain: ApoE'
Age: 6wks
CAPs from Tuxedo, NY. Component
of interest: Ni.
Particle Size: PM2.6
Route: Whole-body Inhalation
Dose/Concentration: Average daily CAPs:
85.6/yglm3 Average daily Ni: 43 ng/m3
Time to Analysis: 6hfd, 5dfwk, for 6m (July
2004-January 2005). 10-second ECG, HR,
activity, and body temperature data were
sampled every 5min for the duration of the
experiment.
For the CAPs-exposed mice, on 14 days there
were Ni peaks at approximately 175 ngfm3 and
usually low CAPs and V. For those days back-
trajectory analysis identified a remote Ni point
source. ECG measurements on CAPs-exposed
and sham-exposed mice showed Ni to be
significantly associated with acute changes in
HR andHRV.
Reference: Lund et al.
(2007,125741)
Species: Mouse
Gender: Male
Strain: ApoE1
Age: 10wks
Use: Mice were
placed on a high fat at
the beginning of the
exposure.
Varying dilutions of gasoline emis-
sions: (generated using two 1996
model 4.3L General Motors V-6
engines, fueld with conventional,
unleaded, non-oxygenated gasoline,
equipped with stock exhaust
systems).
Composition for Hi, Med, and Lo
dilutions:
PM, NOx, CO, and Total
Hydrocarbons (THC)
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: FA: PM (2 //gfnr1), NOx
(0 ppm), CO (0.1 ppm), HC (0.1 ppm);
Low (1: 90 dilution of exhaust): PM (8 /yglm3),
NOx (2 ppm), CO (9 ppm), HC (0.9 ppm);
Mid (1: 20): PM (39 //gfm3), NOx (12 ppm),
CO (50 ppm), HC (8.4 ppm);
High (1: 12): PM (61 //gfm3), NOx (19 ppm),
CO (80 ppm), HC (12 ppm);
High-filtered (1:12): PM (2 //gfm3), NOx (18
ppm), CO (80 ppm), HC (12.7 ppm).
Time to Analysis: 6hfd, 7dfwk for 7wks.
Mice were sacrificed within 16h PE. During
the study period all animals concurrently
exposed to the following: FA: 8 /yglm3 and 40
/yg/m3; PM Whole Exhaust: 60 //g|m3; or
Filtered Exhaust wf gases matching the 60
/yg/m3 concentration.
Inhalation exposure to gasoline engine
emissions resulted in increased aortic mRNA
expression of matrix metalloproteinase-3
(MMP-3), MMP-7, and MMP-9, tissue inhibitor
of MMP-2, ET-1 and HO-1 in ApoE'1' mice;
increased aortic MMP-9 protein levels were
confirmed through immunochemistry. Elevated
R0S were also observed in arteries from
exposed animals, despite absence of plasma
markers. Similar findings were also observed in
the aortas ApoE'1' mice exposed to particle
filtered atmosphere, implicating the gaseous
components of the whole exhaust in mediating
the expression of markers associated with
vasculopathy.
Reference:Lund AK
et al. (2009,191159)
Species: Mouse
Gender: Male
Strain: ApoE1
Age: 10wks
Weight: NR
GEE (conventional unleaded,
nonoxygenated, nonreformulated
gasoline- ChevronPhillips Specialty
Fuels Division)
Particle Size: PM: MMAD- 0.150
fj m
Route: Inhalation
Dose/Concentration: PM: 60//gfm3, NO2:
2ppm, NO: 16ppm, CO: 80ppm, THC:
12.7ppm
Time to Analysis: Mice fed high-fat diet 30d
before exposure. Exposed 6hfd, 1 or 7d. Some
groups dosed with Tempol or BQ-123. Killed
within 18h of last exposure.
Aorta gelatinase activity increased with GEE
exposure time. MMP-2/9 activity spread
throughout the vasculature by day 7. 7d GEE
exposure significantly increased the aorta
protein expression of MMP-9, MMP-2, TIMP-2,
and plasma MMP-9. Generally, in GEE-exposed
mice, Tempol decreased TBARS and vascular
ET-1, and BQ-123 decreased vascular ROS, ET-
1, MMP-9, and gelatinase activity.
Reference:Lund AK
et al. (2009,191159)
Species: Human
Gender: Male, Female
Strain: -
Age: 18-40yrs
Weight: NR
GEE (Cummins engine- 5.9L, 205hp,
at or near idle conditions,
ChevronPhillips- certified commercial
#2 fuel)
Particle Size: PM: MMAD-
0.10 ±0.02 /jm
Route: Inhalation
Dose/Concentration: PM: 100 //gfm3, NO2:
0.4ppm, NO: 3.5ppm, CO: 9ppm, THC:
0.9ppm
Time to Analysis: Exposed 2h on separate
occasions. 4 cycles of 15min rest then 15min
on exercise bike. Blood collected preexposure
and 30min and 24h postexposure. Plasma
samples collected.
GEE significantly increased plasma MMP-9
activity and concentration uniformly among the
subjects. DE significantly upregulated plasma
ET-1 and NOx.
Reference: Montiel-
Davalos et al.(2007,
156778)
Species: Human
Cell Types: HUVEC
(from primary human
endothelial cells) and
U937 (human
leukemia pro-
monocyte) cell
cultures.
PM2.6 and PM10 from Mexico City
Particle Size: PM2.6, PM10
Route: In Vitro
Dose/Concentration: HUVEC TNF-a (10
ng/mL), and a PM range of 5,10, 20, and 40
/yg/cm2 concentrations.
Time to Analysis: 6 or 24 h (early and late
adhesion molecules respectively)
Results showed that both PM2.6 and PM10
induced the adhesion of U937 cells to HUVEC,
and their maximal effect was observed at 20
/yglcm2. This adhesion was associated wf an in-
crease in the expression of all adhesion
molecules evaluated for PM10, and E-selectin,
P-selectin, and ICAM-1 for PM2.6. In general the
maximum expression of adhesion molecules in-
duced by PM2.6 and PM10 was obtained wI 20
fjglcm2; however PMio-induced expression was
observed from 5/yglcm2. E-selectin and ICAM-1
had the strongest expression in response to
particles.
July 2009	D-14	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Moyer et
al. (2002, 052222)
Species: Mouse
Gender: Male and
Female
Strain: B6C3F1
In phosphide (InP), Co sulfate hep-
tahydrate (C0SO47H2O), Vanadium
pentoxide(V20E) Gallium arsenide
(GaAs), Ni oxide (NiO), Ni subsulfide
(NiiiSy], Ni sulfate hexahydrate
(NiSOi - 6H2O), talc, and Mo trioxide
((M0O3)
Particle Size: MMAD particle size
U/m): InP (1.1-1.3), C0SO47H2O (1.5-
1.8), V2OB: (1.0), GaAs: (1.0)
Route: Inhalation
Dose/Concentration: High-Dose Con-
centration in Chronic Studies, Male (/^glm3):
InP: 0.3, C0SO47H2O: 3.0, V2O1,: 4.0, GaAs:
1.0
High-Dose Concentration in Sub-Chronic
Studies, Male or Female (/^glm3): InP: 100,
C0SO47H2O: 30, V20b: 16, GaAs: 75
Time to Analysis: Phase One: Evaluation of
heart, kidney and lung tissues from all control
and high dose male B6C3F1 mice exposed by
inhalation to 9 particulate compounds for a
2yr period. Phase Two: evaluated heart, lung,
kidney and mesentery tissues of control and
high dose male and female B6C3F1 mice from
the 90d studies of the 4-compounds demon-
strating arteritis after a 2yr period.
Phase One: High-dose males developed signifi-
cantly increased incidences of arteritis over
controls in 2 of the 9 studies (InP and
C0SO47H2O), while marginal increases of
arteritis were detected in 2 additional studies
(V20Eand GaAs). In contrast, arteritis of the
muscular arteries of the lung was not observed.
Morphological features of arteritis in these
studies included an influx of mixed inflamma-
tory cells including neutrophils, lymphocytes,
and macrophages. Partial and complete
effacement of the normal vascular wall archi-
tecture, often with the extension of the
inflammatory process into the periarterial con-
nective tissue, was observed.
Phase Two: Results showed that arteritis did
not develop in the 90-day studies, suggesting
that long-term chronic exposure to lower-dose
metallic PM may be necessary to induce or
exacerbate arteritis.
Reference: Mutlu GM
et al. (2007,121441)
Species: Mouse
Gender: Male
Strain: 57BL/6
(IL6+I+ and IL61)
Age: 6-8wks
Weight: 20-25g
PM in from ambient air in Dusseldorf,
Germany
Particle Size: PM10
Route: IT Instillation
Dose/Concentration: PM10: 10//g;
Clodronate: 120mg
Time to Analysis: For alveolar macro|
depletion, clodronate instilled into mice lungs
following endotracheal intubation 48h prior to
instillation of PM. Parameters measured 24h
post-exposure.
Mice treated with PM10 exhibited a shortened
bleeding time, decreased prothrombin and
partial thromboplastin times (decreased plasma
clothing times), increased levels of fibrinogen,
and increased activity of factors II, VIII, and X.
This prothrombotic tendency was associated
with increased generation of intravascular
thrombin, an acceleration of arterial
thrombosis, and an increase in BALF concentra-
tion of prothrombotic IL-6. IL-6'1' mice were
protected against PM-induced intravascular
thrombin formation and the acceleration of
arterial thrombosis. Depletion of macrophages
by the IT administration of liposomal clodronate
attenuated PM-induced IL-6 production and the
resultant prothrombotic tendency.
Reference: Nadziejko
et al. (2002, 050587)
Species: Rat
Gender: Male
Strain: SH Wistar
Kyoto
Age: 16wks
CAPs (PM2.6) from Tuxedo, NY. (SO2,
NO2 0:i and NH3 were removed prior
to exposure).
H2SO4 (fine and ultrafine)
Particle Size: Ultrafine H2SO4 mass
median diameter: 50-75nm
Route: Nose-only Inhalation
Dose/Concentration: CAPs: 80 and 66
/yglm3 (avg 73); Fine H2SO4: 299, 280,119,
and 203//g|m3 (avg 225); Ultrafine H2SO4:
140, 565,416, 750/yglm3 (avg 468)
Time to Analysis: 4h/exposure
Exposure to CAPs caused a striking decrease in
respiratory rate that was apparent soon after
the start of exposure and stopped when
exposure to CAPs ceased. The decrease in
respiratory rate was accompanied by a
decrease in HR. Exposure of the same animals
to fine-particle-size sulfuric acid aerosol also
caused a significant decrease in respiratory
rate similar to the effect of CAPs. Ultrafine
acid had the opposite effect on respiratory rate
compared to CAPs.
Reference: Naziejko
C et al. (2004,
055632)
Species: Rat
Gender: Male
Strain: F344
Age: 18m
PM/CAPs (Tuxedo, NY)
UFC (lab generated)
SO2
Particle Size: PM (Size Range): 0.5-
2.5um; UFC (MMAD): 30-50nm
Route: Nose-only Inhalation
Dose/Concentration: PM (/^gfm3): 161-200,
avg. 180; UFC (^g|m3): 500-1280, avg. 890;
SO2 (ppm): 1.2,1.2, avg. 1.2
Time to Analysis: A total of 8 exposures
were performed: 2 exposures to CAPs, 2
exposures to UFC, 4 exposures to SO2. All
three pollutants were tested wf a crossover
design so that each group alternated exposure
to air and to pollutant. Exposures lasted 4h
and were performed at least 1wk apart.
Parameters measured throughout duration of
experiment.
Old F344 rats had many spontaneous
arrhythmias. There was a significant increase
in the frequency of irregular and delayed beats
after exposure to CAPs. The same rats were
subsequently exposed to UFC, SO2 or air with
repeated crossover design. In these
experiments there was no significant change in
the frequency of any category of spontaneous
arrhythmia following exposure to UFC or SO2.
Reference: Nemmar DEP (SRM 2975)
et al. (2008, 096566)
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Weight: 440 ± 14g
Particle Size: < 1 fjm
Route: Intravenous via the tail vein
Dose/Concentration: DEP: 0.02mg or 0.1mg
DEP/kg (corresponding to about 8 //g or 44
/yg DEP/rat)
Time to Analysis: 48h following systemic
administration of saline or DEP
Intravenous administration of DEP (0.1 mg/kg)
triggered systemic inflammation characterized
by an increase in monocyte an granulocyte
numbers. Both doses of DEP caused a
reduction of RBC numbers and hemoglobin con-
centration. TEM analysis of RBC safter in vitro
incubation (5 //gj'niL] or in vivo administration
of DEP, revealed the presence of ultrafine-sized
aggregates of DEP within the RBC. Larger
aggregates were also taken up by the RBC. The
myocardial morphology and capillary bed were
not affected by DEP exposure.
July 2009	D-15	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Nemmar A DEP (SRM 2975)
et al. (2007,156800) partjc|e size: NR
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 16wks
Weight: 424 ± 8g
Route: Tail Vein Injection
Dose/Concentration: 8, 42, or 212 //g
DEP/rat (150ul of 0.02, 0.1, or 0.5 mg/kg bw)
Time to Analysis: 24h
Effect of DEP on Blood Pressure: Significant
decrease on BP in DEP-exposed rats at doses
of 0.02 mg/kg bw, compared with mean BP
observed in controls.
Effect of DEP onHR: Doses of 0.02, 0.1, and
0.5 mg/kg bw in rats, resulted in significant
reduction of HR compared to controls.
Effect of DEP on Tail Bleeding Time:
Shortening of tail bleeding time in rats exposed
to 0.02, 0.1, and 0.5 mg/kg bw. The shortening
was significant at the dose of 0.02 and 0.5
mg/kg compared w/ controls. Platelet counts in
blood did not significantly increased post-DEP
administration.
Effect of DEP on WBC and RBC numbers:
No significant effect of DEP at doses of 0.02,
0.1 and 0.5 mg/kg on the numbers of
granulocytes, monocytes, or lymphocytes
compared with control.
Reference: Nemmar
et al. (2003, 096567)
Species: Hamster
Gender: Male and
Female
Weight: 100-110g
DEP (SRM 1650)
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 120 /yl (5, 50, or 500
/yg/animal)
Time to Analysis: ln-vivo: formation and
embolization of thrombus were continuously
monitored for 40min. Ex-vivo: animals were
ITly instilled w/ DEPs (0 or 50 /yg per animal),
and blood was collected 5,15, 30, and 60min
post-instillation. In-vitro: Saline or saline-
containing DEPs (0.1, 0.5,1, and 5/yg/mL)
was added to venous blood from untreated
hamsters, and closure time was measured in
the PFA-100 after 5min/animal.
Doses of 5 - 500 /yg enhanced experimental
arterial and venous platelet-rich thrombus
formation in vivo. Blood samples taken from
hamsters 30 and 60 min after instillation of 50
/yg of DEPs yielded accelerated aperture
closure (platelet activation) ex-vivo, when
analyzed in the PFA-100. The direct addition of
as little as 0.5 /yg/mL DEPs to untreated
hamster blood significantly shortened closure
time in vitro.
Reference: Nemmar
et al. (2004, 087959)
Species: Hamster
Gender: Male and
Female
Weight: 100-110g
DEP (SRM 1650); Positively Charged
Polystyrene Particles (PCPSP)
Particle Size: PCPSP: 400 nm; DEP:
NR
Route: IT Instillation
Dose/Concentration: DEP: 50 /yg/animal, or
PCPSP: 500/yg/animal
Time to Analysis: Pretreatment Phase:
Hamsters were pretreated w/ Dexametasone
IP (5mg/kg) or IT (0.1 or 0.5mg/kg) or Sodium
Cromoglycate given IP (40mg/kg), 1h before
DEP or vehicle instillation. Thrombosis: In-vivo
thrombogenesis assessed 24h post-instillation
of DEP or vehicle.
DEP increased thrombosis without elevating
plasma vWF. The IT instillation of PCPSP
equally produced histamine release and
enhanced thrombosis. Histamine in plasma
resulted from basophil activation. IP
pretreatment with Dexametasone abolished the
DEP-induced histamine increase in BALF and
plasma and abrogated airway inflammation and
thrombogenicity. The IT pretreatment with
Dexametasone showed a partial but parallel
inhibition of all these parameters. Pretreatment
with Sodium Cromoglycate strongly inhibited
thrombogenicity, and histamine release.
Reference: Nemmar
et al. (2003, 097487)
Species: Hamster
Gender: Male and
Female
Weight: 100-110g
Ultrafine Particles: Unmodified
Polystyrene Particles (UPSPs); Nega-
tively Charged Carboxylate-Modified
Polystyrene Particles (NCC-MPSPs);
Positively-Charged Amine Modified
Polystyrene Particles (PCA-MPSPs)
Particle Size: UPSPs: 60nm; NCC-
MPSPs: 60nm; PCA-MPSPs: 60 or
400nm
Route: IT Instillation
Dose/Concentration: 5, 50, and 500
/yg/animal in 120 ul saline
Time to Analysis: 1 h post-instillation
Unmodified and negative UFPs did not modify
thrombosis. Positive UFPs increased
thrombosis at 500 and 50 /yg/animal, but not
at 5 /yg/animal. Positive 400 nm particles (500
/yg/animal) did not affect thrombosis. PFA-100
analysis showed that platelets were activated
by the in-vitro addition of positive UFPs and
400 nm particles to blood.
Reference: Nemmar
et al. (2003, 087931)
Species: Hamster
Weight: 100-110g
DEP (SRM 1650)
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 50 /yg/animal in 120 ul
saline
Time to Analysis: 1, 3, 6, and 24h
At 1, 6, and 24 h after instillation of 50 /yg
DEPs, the mean size of in vivo induced and
quantified venous thrombosis was increased by
480, 770, and 460%, respectively. Platelets
activation in blood was confirmed by a short-
ened closure time in the PFA-100 analyzer. In
plasma, histamine was increased only at 6 and
24 h. Pre-treatment with a H1 receptor
antagonist (diphenhydramine, 30 mg/kg
intraperitoneally) did not affect DEP-induced
thrombosis or platelet activation at 1 h;
however both were markedly reduced at 6 and
24 h.
July 2009	D-16	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Niwa et
al. (2007, 091309)
Species: Mouse
Gender: Male
Strain: LDLR/KO
Age, Use: 6 weeks (n
- 20), IT CB
dispersion; 10-14wks
acute effect of CB
dispersion on
circulating CRP
Carbon Black
Particle Size: 23-470nm (mean size
120.7nm)
Route: IT Dispersion
Dose/Concentration: IT CB Dispersion
Study: 1 mg per animal/week; Acute Effect of
CB Dispersion on Circulating CRP Study:
1 mg/animal (single administration)
Time to Analysis: IT CB Dispersion Study:
1x/wk for 10wks;
Acute Effect of CB Dispersion on Circulating
CRP Study: Single CB administration, blood
samples collected 24h post-administration
IT CB Dispersion Study: Although no
difference in body weight (bw) between the
four groups was observed at baseline, and all
mice experienced an increase in bw with
advancing age, the mice treated with CB
tended to be smaller than those treated with
vehicle (air). No significant differences were
observed in cholesterol and TG levels among
the for groups. Development of aortic lipid-rich
lesions occurred in mice under a 0.51 %
cholesterol diet with or without CB infusion,
but not in the mice fed a 0% cholesterol diet.
Acute Effect of CB Dispersion on
Circulating CRP Study: Circulating levels of
CRP were significantly higher in mice exposed
to CB versus those exposed to air, indicating an
acute inflammatory response. Although the
presence of CB in pulmonary macrophage-like
cells in CB treated mice under 0.51%
cholesterol diet was confirmed, CB was not
detected in aortas, livers, kidneys, or spleens.
Reference: Niwa et
al. (2007, 091309)
Species: Mouse
Cell Types:
Macrophages Cell
Lines (RAW264.7)
Carbon Black (CB); Water-Soluble
Fullerene
(C60(0H)24); Fluoresbrite
Carboxylate Microspheres; 0x-LDL;
Acetylated-LDL
Particle Size: Carbon Black and
C6o(OH)24: 7.1 nm (SD 2.4);
Fluoresbrite Carboxylate
Microspheres: 6nm
Route: Cell Culture
Dose/Concentration: CB: 1,10,100/yg|mL;
C6o(OH)24: 20, 10Ong/mL
Time to Analysis: RAW264.7 +CB
for 24h, 13d, and 50d;
RAW264.7 +C60(0H)24 for 24h or 10d;
RAW264.7 +C60(0H)24 for 8d, then co-
treated wf Ox-LDL for an additional 48h;
RAW264.7 -t-Ox-LDL for 5d, and then co-
cultured wI C60(0H|24 for an additional 48h;
RAW264.7 + 6nm beads: 3d, the Ox-LDL or
acetylated-LDL added for 24h
CB alone had no significant effects on
RAW264.7 cell growth. C60(0H)24 alone or CB
and C60(0H)24 together wI Ox-LDL induced
cytotoxic morphological changes, such as Ox-
LDL uptake-induced foam cell-like formation
and decreased cell growth, in a dose-dependent
manner. C60(0H)24 induced L0X-1 protein
expression, pro-matrix metalloprotease-9
protein secretion, and tissue factor mRNA
expression in lipid-laden macrophages. Although
CB or C60(0H)24 alone did not induce platelet
aggregation, C60(0H)24 facilitated ADP-
induced platelet aggregation. C60(0H)24 also
acted as a competitive inhibitor of ADP
receptor antagonists in ADP-mediated platelet
aggregation.
Reference: Niwa Y et
al. (2008,156812)
Species: Rat
Strain: Sprague-
Dawley
Age: 6wks
CB from Kyoto, Japan
Particle Size: Mean size (nm) ± SD
determined at 1, 8,15, 22, and 29d
post-exposure was 118.1 ± 2.4,
119.1 ± 2.7,122.2 ± 2.0,122.4 ±
2.5 and 121.0 ± 3.6 respectively
Route: Whole-body Inhalation
Dose/Concentration: 15.6 ± 3.5mgfm3
Time to Analysis: 6hfd, 5dfwk, for a total of
4wks. BP and HR were measured by tail-cuff
plethysmography at 1,14, and 28d post ¦
exposure. Sacrificed At 1, 7,14, 28, and 30d
post-exposure
Although the presence of CB was confirmed in
pulmonary macrophages, electron microscopic
survey did not detect CB in other tissues
including, liver, spleen and aorta. CB exposure
raised blood pressure levels in a exposure-time
dependent manner. Levels of circulating
inflammatory marker proteins, including
monocyte chemo attractant protein-1, IL-6,
andCRP, were higher in the CB treated groups
than in control groups.
Reference:
Nurkiewicz TR et al.
(2004, 087968)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age: 7-8wks
R0FA (from Everett, MA) Major metal
contaminants are: Fe, Al, V, Ni, Ca,
and Z. Main soluble metals are: Al,
Ni, and Ca.
Particle Size: R0FA mean count
diameter: 2.2 /jm
Route: IT Instillation
Dose/Concentration: R0FA group: 0.1, 0.25,
1, or 2mgfrat. Vehicle control group: 300 ul
saline. Particle control group: Ti02
0.25mg/rat.
Time to Analysis: After single IT instillation
of a particular dose, all rats recovered for
24h.
Saline Treated Rats: A23187 dilated
arterioles up to 72 ± 7% max.
ROFA and Ti02 Exposed Rats: A23187-
induced dilation was significantly attenuated.
Sensitivity of Arteriolar Smooth Muscle to
NO: Similar in saline treated and ROFA exposed
rats.
Other: Significant increase in venular leuko-
cyte-adhesion and rolling observed in ROFA
exposed rats.
July 2009	D-17	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Nurkiewicz et al.
(2006,088611)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Aqe: 7-8wks
ROFA from Everett, MA
Particle Size: ROFA mean count
diameter: 2.2 /jm; Ti02 mean
diameter: 1.0 /jm
Route: IT Instillation
Dose/Concentration: ROFA group: 0.1 or
0.25 mg/rat. Vehicle control group: 300 ul
saline. Particle control group: Ti02 0.1 or 0.25
mg/rat.
Time to Analysis: After single IT instillation
of a particular dose, all rats recovered for
24h.
ROFA or Ti02 Exposure and Arteriolar
Dilation: Exposure caused a dose-dependent
impairment of endothelium-dependent arteriolar
dilation.
ROFA or Ti02 Exposure and Arteriolar
Constriction: Exposure did not affect
microvascular constriction in response to PHE.
ROFA and Ti02 and Leukocyte Rolling and
Adhesion: Exposure significantly increased
leukocyte rolling and adhesion in aired venules,
and these cells were identified as PMN
leukocytes.
ROFA and Ti02 and MPO: MP0 was found in
PMN leukocytes, adhering to the systemic mi-
crovascular wall. Evidence suggests that some
of this MPO had been deposited in the
microvascular wall. There was also evidence of
oxidative stress in the microvascular wall.
Reference:
Nurkiewicz TR et al.
(2008,156816)
Species: Rat
Gender: Male
Strain: SD
Age: 6-7wks
Weight: NR
Ti02 (DeGussa, Sigma-Aldrich;
powders put through fluidized-bed
aerosol generator)
Particle Size: Fine- 1 fjm, UF- 21 nm
Route: Whole-body Exposure
Dose/Concentration: Concentrations: Fine-
3-16mg/m3; UF- 1.5-12mg/m3; Dose: Fine- 8,
20, 36, 67, 90 //g; UF- 4, 6,10,19, 30 //g
Time to Analysis: Acclimated 5d. Exposed 4-
12h. Sacrificed 24h postexposure.
Particle accumulation within AMs, anuclear
macrophages, particle-laden AMs intimately
associated with the alveolar wall were all
present in exposed rats. Calcium ionophore
impaired arteriolar dilation in a dose-dependent
manner in UF and fine exposed rats. UF
produced greater systemic microvascular
dysfunction. Microvascular dysfunction was
the same for three groups of rats exposed to
30 /yg UF Ti02 under different conditions.
Reference:
Nurkiewicz et al.
(2009,191961)
Species: Rat
Gender: Male
Strain: SD
Age: 7-8wks
Weight: NR
Fine Ti02 (Sigma-Aldrich- (titanium
(IV)) oxide, 224227, St. Louis, M0)
( — 99% rutile)
TiOy nanoparticles (DeGussa-
Aeroxide Ti02 P25, Parsippany, NJ)
(80% anatase, 20% rutile)
Particle Size: Fine Ti02- Primary
size: < 5 /jm, MMAD: 402nm, CMD:
710nm; Nano-Ti02- Primary size:
21nm, MMAD: 138nm, CMD: 100nm
Route: Aerosol Inhalation
Dose/Concentration: 1.5-16mgfm3
Time to Analysis: Acclimated 5d. Exposed
240-720min. Anesthetized 24h postexposure.
Intravital microscopy, NO measurement,
microvascular oxidative stress measurement,
nitrotyrosine staining.
Arteriolar Dilation: Nano-Ti02 significantly
impaired endothelium-dependent arteriolar
dilation. Equivalent levels of arteriolar
dysfunction were found in fine and nano-Ti02.
Arteriolar dilation in response to abluminal
microiontophretic application of SNP was not
different from the controls or between the
exposure groups. Arteriolar dilation was
partially restored by radical scavenging with
TEMP0L and catalase, NADPH oxidase with
apocynin, and MPO inhibition with ABAH.
Microcirculation: R0S increased in both
groups. Nano-Ti02 significantly increased the
area of tissue containing nitrotyrosine in the
lung and spinotrapezius microcirculation.
NO: Fine and nano-Ti02 significantly and dose-
dependently decreased stimulated NO
production in isolated microvessels. NO
production was increased by radical scavenging
with TEMP0L and catalase or NADPH oxidase
with apocynin, and was largest in the fine Ti02
group.
July 2009	D-18	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Okayama
et al. (2006,156824)
Species: Rat
Cell Type: Ventricular
Cardiac Myocytes
from Wistar Rats, ap-
proximately 3d old
DEP (Tsukuba, Japan)
DEPE: 5g of DEP in 5 mL PBS
containing 0.05% Tween 80.
Others: Catalase, LDH, MPG and
SOD.
Particle Size: DEP mass median
diameter: 0.34 /ym.
Route: In Vitro
Dose/ Concentration: DEPE: 0-100 /yg/mL;
MPG: 0-1 mM; SOD: 800 U|mL: Catalase:
500 U/mL
Time to Analysis: Long-Term Exposure to
DEPE: cells were incubated for 24 or 48h.
Short-Term Exposure to DEPE: 1, 2, 4, or 8 h
and then medium containing DEPE was
replaced by serum-free medium, and incubated
for an additional 24 h.
LDH Activity of Supernatant: 24 h post-DEPE
exposure.
SOD, Catalase, MPG on DEPE-induced
Toxicity: SOD, catalase or MPG was added to
cells w/ or w/o DEPE & incubated for 4 or
24h. Medium then replaced wfserum-free &
cells incubated for another 24h to analysis.
Cytotoxic Effects of DEPE on Cardiac
Myocytes: DEPE above 20 /yg/mL damaged
cardiac myocytes in a time and concentration-
dependent manner in both long- and short-term
exposure conditions. However damage was
greater after long-term exposure. LDH activity
showed a concentration-dependent increase at
higher levels of exposure (greater than 20
/yg/mL).
Effects of ROS Scavenging Enzymes and
Antioxidant on DEPE-induced Cell Damage:
SOD or catalase attenuated 50 /yg/mL DEPE-
induced cell damage compared with DEPE-
treated groups lacking antioxidant enzymes.
Co-incubation with SOD and catalase showed
more protective effects towards DEPE-induced
cell damage, although these effects were not
statistically significant from cells treated with
SOD only. MPG attenuated 50 /yg/mL DEPE-
induced cell damage in a concentration-
dependent manner in both long and short-term
exposure conditions. Especially in long-term
exposure MPG showed strong protective
effects against DEPE-induced cell damage. Cell
viability was not affected by SOD, catalase, or
MPG.
Reference: Proctor et
al. (2006, 088480)
Species: Rat
Gender: Male
Age: 12wks
Use: Thoracic Aorta
from cp/cp and +/?
Male Rats
cp/cp - homozygous
for cp gene. Prone to
obesity and insulin
resistant.
+/? - heterozygous
for either +/cp or
+/+. Lean and
metabolically normal.
R0FA from Birmingham, AL
Particle Size: 1.95 ± 0.18/ym
aerodynamic diameter of R0FA
Route: Protocol 1: Used two aorta rings per
each experimental treatment group (4 groups
total). Protocol 2: Used four rings.
Dose/Concentration: Protocol 1: exposed to
12.5/yg/mL R0FA-L (at 10mg/mL).
Protocol 2: exposed to 1.56, 3.25, 6.26,12.5
/yg/mL R0FA-L (at 10mg/mL).
Time to Analysis: Protocol 1: Cells treated
with 12.5/yg/mL R0FA-L and/or 104mol/L L-
NAME for 20min
Protocol 2: Parameters measured after R0FA-
L only treatment
Contractile response to phenylephrine (PE)
was measured
R0FA-L (12.5/yg/mL) increased PE-mediated
contraction in obese, but not in lean rat aortae.
Effect was exacerbated by L-NAME, and it
reduced ACh-mediated relaxation in obese and
lean aortae. Initial exposure of aortae to R0FA-
L caused a small contractile response, which
was markedly greater on second exposure in
the obese aortae but marginal in lean.
Reference: Radomski
A et al. (2005,
091377)
Species: Rat Strain:
Wistar-Kyoto
Use: Vascular
Thrombosis
Carbon Nano Particles (CNPs)
(purchased from SES Research,
Houston, TX): Multiplewall
Nanotubes (MWNT); Singlewall
Nanotubes (SWNT); C60 Fullerenes
(C60CS); Mixed Carbon
Nanoparticles (MCN)
PM: (SRM1648) (NIST)
Particle Size: CNPs: NR;PM:1.4
//m average size
Route: Simultaneous single PM injection into
femoral vein as FeCh injected to induce
carotid thrombosis
Dose/Concentration: 0.5 mL suspension of
50/yg/mL of PM in 0.9% NaCI solution.
Time to Analysis: Blood flow continuously
monitored for 900 seconds.
Vascular Thrombosis: FeCh induced carotid
artery thrombosis and MCN had an amplifying
effect in the development of thrombosis.
Infusions of MCN, SWNT, and MWNT signifi-
cantly accelerated the time and rate of
development of carotid artery thrombosis in
rats. SRM1648 was less effective than CNPs
in inducing thrombosis, while C60CS exerted
no significant effect on the development of
vascular thrombosis.
Reference: Radomski
A et al. (2005,
091377)
Species: Human
Cell Types: Platelets
Use: Human |
aggregation
Carbon Nano Particles (CNPs)
(purchased from SES Research,
Houston, TX): Multiplewall
Nanotubes (MWNT); Singlewall
Nanotubes (SWNT): C60 Fullerenes
(C60CS); Mixed Carbon
Nanoparticles (MCN);
PM (SRM1648)
Particle Size: CNPs: NR;PM:1.4
//ni average size
Route: Cell Culture (2.5X108 platelets/mL)
Dose/Concentration: CNPs: 0.2-300/yg/mL;
PM: 5 ¦ 300 /yg/mL
Time to Analysis: Prostacyclin (PGI2), S-ni-
troso-glutathione (GSN0), aspirin, 2-methyl-
thio-AMP, phenanthroline, EDTA and Go6976
were pre-incubated w/ platelets fori min
before particle addition. Particles added to
platelets and platelet aggregation studied for
8min.
Platelet Aggregation: All CNPs, except
C60CS, stimulated platelet aggregation (MCN
> SWNT > MWNT > SRM1648). All particles
resulted in upregulation of GPIIb/llla in
platelets. In contrast, particles differentially
affected the release of platelet granules, as
well as the activity of thromboxane-, ADP,
matrix metalloproteinase- and protein kinase C-
dependent pathways of aggregation. Particle-
induced aggregation was inhibited by
prostacyclin and GSN0, but not by aspirin.
July 2009	D-19	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Reed MD
et al. (2008,156903)
Species: Rat
Gender: Male, Female
Strain: CDF
(F344)/CrlBR, SHR
Age: NR
Weight: NR
GEE (two 1996 General Motors 4.3-L
V-6 engines; regular, unleaded, non-
oxygenated, non-reformulated
Chevron-Phillips gasoline, U.S.
average consumption for summer
2001 and winter 2001-2002)
Particle Size: MMAD: 150nm
Route: Inhalation Exposure Chamber
Dose/Concentration: PM: Low- 6.6±3.7
/yg/m3, Medium- 30.3±11.8/yg/m3, High-
59.1 ±28.3 //g/m3
Time to Analysis: 2wk quarantine period in
chamber. Exposed 6h/d, 7d/wk, 3d-6m. SHR-
surgery to implant telemeter in peritoneal
cavity. 4wks recovery. ECG data obtained
every 15min beginning 3d preexposure, 7d
exposure, 4d postexposure.
Organ Weight: At 6m exposure, the heart
weights of male and female rats increased and
male rats' seminal vesicle weight decreased.
Histopathology: PM-containing macrophages
increased by 6m.
Serum Chemistry: Serum alanine
aminotransferase, aspartate aminotransferase,
and phosphorus decreased in medium and high-
exposure females.
Hematology, Clotting Factors: Hematocrit,
red blood cell count, and hemoglobin dose-
dependently increased for both genders at both
time points. Plasma fibrinogen increased at
1wk in males.
Lung DNA Damage: Hypermethylation
occurred in medium- and high-exposure male
rats at 6m.
BAL: For both genders in the high-exposure
group, LDH and MIP-2 significantly increased at
6m. R0S decreased at 1wk and 6m. Generally,
the production of hydrogen peroxide and
superoxide decreased in the high-exposure
group and medium- and high-exposure groups,
respectively.
CV effects in SHR: Lipid peroxides were
significantly increased in males in the high
exposure group. TAT complexes decreased in
females in the high exposure group.
Removal of Emission PM: The removal of
emission PM strongly linked PM to increased
seminal vesicle weight, red blood cell counts,
LDH, lipid peroxides, and methylation.
Reference: Rhoden et
al. (2005, 087878)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age:A dult
Weight: 300g
Urban Ambient Particles (UAPs):
SRM-1649; CAPs (from Boston, MA)
Particle Size: NR
Route: UAPs: IT Instillation. CAPs: Inhalation
Dose/Concentration: UAPs: 750 /yg
suspended in 300 ul saline; CAPs: 700 ± 180
//g/m3
Time to Analysis: UAPs: 30min post-instil-
lation. CAPs: immediately after 5h exposure
period
Oxidative Stress and HR Function: UAPs
instillation led to significant increases in heart
oxidants. HRincreased immediately after
exposure and returned to basal levels over the
next 30 min. SDNN was unchanged
immediately after exposure, but significantly in-
creased during the recovery phase.
Role of ROS in Cardiac malfunction: Rats
were treated with 50 mg/kg NAC 1 h prior to
UAPs instillation or CAPs inhalation. NAC
prevented changes in heart rate and SDNN in
UAPs-exposed rats.
Role of the Autonomic Nervous System in
PM-induced Oxidative Stress: Rats were
given 5 mg/kg atenolol, 0.30 mg/kg gly-
copyrrolate, or saline immediately before CAPs
exposure. Both atenolol and glycopyrrolate
effectively prevented CAPS-induced cardiac
oxidative stress.
Reference: Rivero DH
et al. (2005, 088653)
Species: Rat
Gender: Male
Strain: Wistar
Age: 3m
Weight: ~250g
PM2.6, collected from heavy traffic
area in Sao Paulo, Brazil. PM2.6
Composition (%): S (3.05), As (0.30),
Br (0.21), CI (2.09), Co (2.65), Fe
(2.67), La (5.42), Mn (0.64), Sb
(0.21), Sc (3.25), Th (8.14)
Particle Size: PM2.6
Route: IT Instillation
Dose/Concentration: 100 or 500 /yg of
PM2.6.
Time to Analysis: 24h post-instillation
Blood: Total reticulocytes significantly
increased at both PM2.6 doses, while
hematocrit levels increased in the 500 /yg
group. Quantification of segmented neutrophils
and fibrinogen levels showed a significant
decrease, while lymphocytes counting
increased with 100 /yg of PM2.6.
Pulmonary vasculature: Significant dose-de-
pendent decrease of intra-acinar pulmonary
arteriole lumen/wall ratio was observed in both
PM2.6 groups.
Wet-to Dry Weight Ratio: Significant
increase in heart wet-to-dry weight ratio was
observed in the 500 /yg group.
July 2009	D-20	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Rodriquez
Ferreira Rivero DH et
al. (2005, 088659)
Species: Rat
Gender: Male
Strain: Wistar
Age: 3m
Weight: — 250g
PM2.6, collected from heavy traffic
area in Sao Paulo, Brazil. PM2.6
Composition (%): S (3.05), As (0.30),
Br (0.21), CI (2.09), Co (2.65),
Fe (2.67), La (5.42), Mn (0.64),
Sb (0.21), Sc (3.25), Th (8.14)
Particle Size: PM2.6
Route: IT Instillation
Dose/Concentration: 50 and 100/yg of
PM2.6.
Time to Analysis: HR and SDNN were
assessed immediately before instillation, 30
and 60min post-instillation.
HR decreased significantly with time, but no
significant effect of treatment or interaction
between time and treatment was observed. In
contrast, there was a significant SDNN
interaction between time and treatment. The
SDNN decreased 60 min after instillation with
PM2.5 concentration of 50 and 100 /yg.
Reference: Seagrave
et al. (2008,191990)
Species: Rat
Gender: Male
Strain: SD
Age: 10-12wks
Weight: 250-300g
GEE (2 1996 General Motors 4.3-L
V6 gasoline engines; conventional
Chevron Phillips gasoline, U.S.
average composition) (CO, NO, NO2,
SO2, THC) (PM2.6 composition- EC,
OC, S04, NH4, NOs)
Simulated downwind coal emission
atmospheres (SDCAs) (fly ash, gas-
phase pollutants, sulfate aerosols,
NO, NO2, SO2)
Paved Road Dust (RD) (Los Angeles,
CA; New York City, NY; Atlanta, GA)
Particle Size: GEE: MMAD- 150nm,
RD: 2.6±1.7 fjm, SDCA: 0.1-1.0//m
Route: Nose-only Inhalation
Dose/Concentration: GEE: 60 //gfm3,
SDCAs: 317-1072//g/m3, RD: 306-954
/yg/m3; GEE: CO- 104ppm, NO- 16.7ppm, N02-
1.1ppm, S02- 1.0ppm,THC- 12ppm; SDCAs:
CO- < 1 ppm, NO- 0.19-0.62ppm, N02- 0.10-
0.37ppm, S02- 0.07-0.24ppm, THC- < 1ppm
Time to Analysis: Quarantined 2wks. 6h
exposure then ip injected. Cannula ligated into
trachea and connected to rodent ventilator.
Thorax and abdomen opened. Killed after
measurements taken.
GEE produced CL in the lungs, heart, and liver.
RD produced a significant effect in the heart at
the low dose. SDCAs had no effect on CL. RD
significantly increased the heart's oxidative
stress, as demonstrated by the TBARS levels.
GEE did not affect the amount of macrophages
or PMN. SDCAs increased macrophages. The
RD low dose increased macrophages and PMN.
SDCAs increased Pur* values and tidal volumes.
Reference:
Simkhovich et al.
(2007, 096594)
Species: Rat
Gender: Female
Strain: Fischer 344 x
Brown Norway hybrid
Age: 4, 26 m
Use: Study performed
in isolated Langen-
dorff-perfused rat
hearts
Ultra Fine Particles (UFPs) isolated
from industrial diesel reference PM
2975
Particle Size: UFPs ~ 0.1 /jm
Route: Heart Perfusion (ex-vivo)
Dose/Concentration: UFPs 12.5, 25, and
37.5 mg.
Time to Analysis: Hearts perfused wI UFPs
for 30 minutes and analysis conducted every
10min.
Young adult and old hearts demonstrated equal
functional deterioration in response to direct
infusion of UFPs. Developed pressure in young
adult UFPs-treated hearts fell from 101 ± 4 to
68 ± 8 mmHg. In the old UFPs-treated hearts
developed pressure fell by 35%. Positive dP/dt
was equally affected in the young adult and old
UFPs-treated hearts and was decreased by
28% in both groups.
Reference: Sun ~ et
al. (2005,186814)
Species: Mouse
Gender: Male
Strain: ApoE1
Aqe: 16wks
CAPs: PM2.6 from Tuxedo, NY.
HFCD: High Fat Chow Diet
NCD: Normal Chow Diet
Particle Size: PM2.6
Route: Whole-body Inhalation
Dose/Concentration: PM2.6: 85//gfm:l; Daily
concentration: 10.6 (SD 3.4) //g|m3 (mean)
Average exposure over 6 mth period: 15.2
/yg/m3.
Time to Analysis: Study diets fed for at
least 10wk prior to exposure to PM2.6 or FA.
Exposed for 6h/d, 5d/wk for 6m. Sacrificed
15-47d after exposure.
Vasomotor Function: Mice fed HFCD and
exposed to PM2.6 demonstrated an increase in
the half-maximal dose for dilation to ACh with
no changes in peak relaxation compared to the
mice exposed to FA and fed HFCD and NCD.
Atherosclerosis Burden with PM2.5: In vivo
MRI imaging of atherosclerosis burden in the
abdominal aorta revealed significantly in-
creased plaque burden in the mice fed HFCD
compared with the mice fed NCD. Mean (SD)
plaque areas in the mice exposed to PM2.6 and
fed HFCD vs. mice exposed to FA and fed HFCD
were 33 (10) vs. 27 (13) units, respectively.
PM2.5 and Vascular Inflammation: A 2.6-fold
higher inducible NOS content was apparent in
the mice exposed to PM2.6 and fed HFCD com-
pared with the mice exposed to FA and fed
HFCD chow and a 4-fold increase in the mice
exposed to PM2.6 and fed NCD compared with
the mice exposed to FA and fed NCD.
July 2009
D-21
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Sun
(2008,157033)
Species: Mouse
Gender: Male
Strain: ApoE1
Age: 6wks
et al. CAPs PM2.6
Collected from Sterling Forest State
Park, Tuxedo NY (40 miles NW of
Manhattan)
Particle Size: PM2.6
Route: Inhalation Chamber
Dose/Concentration: Average Concentration
of: 85 /yglm3 CAPs in chamber.
Average exposure over 6m - 15.2/yg/m3.
Time to Analysis: 6hfd, 5dfwk for 6m.
Mice received two different diets, high-fat
chow and normal-chow.
Macrophage and Tissue Factor Expression
in Aortic Segments: Tissue Factor (TF)
expression was noted predominantly in the
extracellular matrix surrounding macrophages,
foam cell-rich areas and around smooth muscle
cells.
1.	High-Fat Diet: Increased TF and increased
macrophage infiltration was observed in the
plaques of high-fat chow mice exposed to PM
compared to mice exposed to air and high fat
diet.
2.	Normal Diet: PM-exposed mice saw an
increase in CD68 expression compared to air-
exposed. However TF expression was not
significantly different in PM exposed normal
diet mice compared to control normal diet mice.
Reference: Sun et al.
(2008,157033)
Species: Human
Cell Lines: Human
Bronchial Epithelial
Cells (BEAS-2B);
Vascular Smooth
Muscle Cells (hSMCs);
and
Monocytes (THP-1)
Ambient Particles collected from
Sterling Forest State Park, Tuxedo,
NY (24 hid for4wks)
Particle Size: Particle size ranges:
1. < 0.18/ym
2.1.8 ¦ 2.5 /jm or
3. 2.5 ¦ 10/ym
Route: In vitro
Dose/Concentration: 10-300//gj ml
Time to Analysis: Doses were tested for
durations up to 24 h.
Dose durations tested for up to 24-h did not
indicate detectable effects on cell viability.
Effect of PM on TF Expression and Activity
in hSMCs: In the PM size range of 1-3 //m,
significant increases in TF expression was
observed at doses of 100 and 300/jm fniL. In
the < 0.18 /jm size range, significant increase
in TF expression was observed at all doses. The
particles with sizes 0.18 - 1.0 /jm did not
induce significant change in TF expression.
Effect of PM on TF Expression and Activity
in Monocyte Cells: TF protein expression
increased with < 0.18/ym and the 1- 3/jm
range particles. Expression was increased in
the 0.18-1.0/ym particle range but it was
limited compared to the other PM size ranges.
In general TF expression was higher in
monocytes than in hSMCs cells, but not
significantly.
Effect on TF Expression and Activity in
Bronchial Epithelial Cells: 100 //gfmL of the
1-3/jm and < 0.18/ym particles significantly
increased TF expression.
TF mRNA Expression: TF mRNA was
increased rapidly within the first hour in
response to SRM-1694a PM. The lowest dose
of SRM PM in //cj'niL induced highest levels of
mRNA in hSMCs, no further increase was
observed at higher concentrations.
Reference: Sun et al. PM2.6, Ultra Fine Particle (UFP) or FA Route: Inhalation via Whole-body Exposure
(2008,157032)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age: 500-650g
Particle Size: PM2.6; UFP: < 0.1 fjm Dose/Concentration: Mean PM2.6
concentration: 79.1 ±7.4/yg/m3. Normalized
PM2.6 over 10wk period: 14.1 //g|m3.
Time to Analysis: 6hfday, 5dfweek random
exposure to PM2.6, UFP, or FA for a total of
10 weeks. At the end of week 9 exposure,
rats were infused wI 0.75 mg/kg/d of All for
7 days. PM2.6, UFP, or FA, continued during
All infusion period.
All - angiotensin II
Mean Arterial Pressure (MAP): after All
infusion, MAP was significantly higher in PM2.6
-All vs. FA-AII group. Aortic Vasoconstriction to
PE was potentiated with exaggerated relaxa-
tion to the Rho-kinase (ROCK) inhibitor Y-
27632 and increase in R0CK-1 mRNA levels in
the PM2.E -All group. Superoxide production in
the aorta was increased in the PM2.6. All group
compared to FA-AII group, inhabitable by
apocynin and L-NAME with coordinate
upregulation of NAD(PR oxidase subunits
p22phox and p47phox and depletion of
tetrahydrobipterin.
Reference: Sun et al.
(2008,157032)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age: 500-650g
Use: Primary Rat
Aortic Smooth Muscle
Cells (RASMCs),
passages 4 to 8 used
for the experiment.
PM2.6; Ultra Fine Particle (UFP) or
Filtered Air (FA)
Route: Cell Culture
Dose/Concentration: UFP, PM2.6: 10 or 50
Particle Size: PM2b; UFP: <0.1 /jm /yglmL; All: 100 nmol/L
Time to Analysis: Exposed to UFP or PM2.6
and parameters measured at 0,1, 3, 6, and
15min.
All - angiotensin II
Exposure to UFPs and PM2.6 was associated
with an increase in ROCK activity, phosphory-
lation of myosin light chain, and MYPT1.
Pretreatment with N-Acetylcysteine and the
Rho kinase inhibitors (Fasudil and Y-27632)
prevented MLC and MYPT-1 phosphorylation by
UFPs suggesting a Superoxide-mediated
mechanism for PM2.6 and UFPs effects.
July 2009	D-22	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Sun ~ et
al. (2009,190487)
Species: Mouse
Gender: Male
Strain: C57BL/6, c-
finsm (transgenic,
yellow fluorescent
protein under
monocyte-specific
promoter)
Age: 8,10wks
Weight: NR
PM (concentrated- northeastern
regional background; Tuxedo Park,
NY)
Particle Size: Diameter: 2.5 /jm
Route: Whole-body Exposure. IT Instillation.
Dose/Concentration: Exposure chamber
(mean): 72.7 /yglm3, IT: 1.6mg/kg
Time to Analysis: C57BL/6 mice equilibrated
2wks, fed high-fat chow 10wks. Exposed in
vivo 6h/d, 5d, 128d. fins'" rendered diabetic
or fed normal chow 10wks. IT instilled with
PM 5min, 2X/wk, 10wks.
Metabolic Impairment: PM induced insulin,
homeostasis model assessment indexes,
elevated glucose, and abnormalities in lipid
profile consistent with the IR phenotype.
Vascular Endothelium: PM decreased peak
relaxation and EDboto ACH and peak relaxation
to insulin. Lower levels of NO release were
seen.
Insulin Signaling: PM reduced the
phosphorylation of Akt in intact aorta. PKC-
P11 was the only PKC isoform to increase.
Adipose Inflammation, Visceral Adiposity:
PM significantly increased TNF-a, IL-6, E-
selectin, ICAM-1, plasminogen activator
inhibitor-1, and restin. PM increased visceral
and mesenteric fat mass. F4/80+macrophages
in fat tissue and adipocyte size increased. PM
downregulated IL-10 and glactose-N-
acetylgalactosamine-specific lectin.
YFP Cell Adhesion and Infiltration: PM
increased YFP cells in the adipose tissue, YFP
cell infiltration in the mesenteric fat, and YFP
cell adhesion to endothelium.
Reference:
Tamagawa et al.
(2008,191988)
Species: Rabbit
Gender: Female
Strain: New Zealand
White
Age: 12wks
Weight: Acute
(average)- 2.4±0.2kg,
Chronic (average)-
2.7 ±0.3kg
PM in (urban; Ottawa, Canada)
Particle Size: Mean diameter:
0.8±0.4/ym
Route: Intrapharyngeal Instillation
Dose/Concentration: Acute- 2.6mg/kg,
Chronic- 2mg/kg
Time to Analysis: Acute animals exposed
days 1, 3, 5. Chronic animals exposed 2x
4wks. Killed postexposure.
Inflammation: PMio induced more
macrophages, AMs, positive and activated
AMs, and fewer tissue macrophages. NO, WBC
and PMN were only significantly higher in the
first two weeks and IL-6 in the first week.
Vascular endothelial function: PMio
significantly reduced Ach-stimulated relaxation
and did not alter SNP-stimulated relaxation. A
significant inverse relationship between IL-6
and Ach-induced relaxation occurred at week 1
in the acute model and weeks 1 and 2 in the
chronic model.
AMs: The chronic model had a significant
correlation between IL-6 and both positive and
activated AMs at week 1. A significant inverse
relationship occurred between Ach and both
the volume fraction of positive and activated
AMs.
Reference:
Tankersley et al.
(2008,157043)
Species: Mouse
Gender: Male
Strain: C57BL/6,
C3H/HeJ, B6C3F1
Age: 18, 28m
Weight: NR
Carbon black (CB) (Wright dust feed
particle generator - BGI, Waltham,
MA)
Particle Size: 0.1-1.0/ym
Route: Inhalation Chamber
Dose/Concentration: Average PM2.6
concentration- 401 ±46/yglm3, Average PMio
concentration- 553±49/yg/m3
Time to Analysis: Exposed 3hfd 4d.
Hemodynamics: CB significantly elevated
right atrial and ventricular pressures,
pulmonary arterial pressure and vascular
resistance, all of which were more pronounced
in the 28m-old mice. RV contractility
(specifically, the ejection fraction and maximum
change in pressure over time) reduced in CB-
exposed 28m-old mice.
Heart tissue: CB significantly declined Ca2+-
dependent N0S activity and was more
pronounced in 28m-old mice, who also had
N0S2 upregulated. CB enhanced ROS
generation and NOS-uncoupling and was
greatest in 28m-old mice. CB also increased
MMP2, MMP9, ANP, BNP, which were
greatest in 28m-old mice. CB also reduced
PKG-1 in 28m-old mice.
July 2009	D-23	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Tankersley et al.
(2007,097910)
Species: Mouse
Gender: Male
Strain:C3HfHeJ and
C57BL/6J
Age: 10wks
Weight: 22-26g
Carbon Black (CB) and Filtered Air
(FA)
Particle Size: CB: 2.4/ym (MMAD)
(GSD 2.75 //m).
Route: CB: Whole-body Inhalation Chamber;
Sympathetic (S) & Parasympathetic (PS)
blockade: IP Injection
Dose/Concentration: CB: 159 ± 12//gfm:l;
PS (atropine): 0.5mgfkg; S(propanolol):
1 mg/kg
Time to Analysis:Successive 3h CB and FA
Exposures: conducted from 9 a.m. to 1 p.m.,
or at least 3h after dark-to-light transition
(exposure period selected based on the nadir
in circadian pattern in HR responses).
Subgroups of both strains exposed to PS & S
blockade.
FA Exposure with Saline: A significantly
greater 3 h average response occurred in C3
compared with B6 mice.
PS Blockade: No evident strain difference
between C3 and B6 was observed.
S Blockade: 3 h average HR responses for C3
mice were significantly reduced compared with
saline.
CB Exposure: HR responses were significantly
elevated in C3 compared with B6 mice, but
these HR responses were not different relative
to FA exposure.
S Blockade: HR was significantly elevated in
B6 mice during CB relative to FA, but was not
changed in C3 mice.
Reference:
Tankersley CG et al.
(2004, 094378)
Species: Mice
Strain:AKRfJ
Age: — 180d
Use: Age matched
animals
Carbon Black (CB) and Filtered Air
(FA)
Particle Size: CB: 0.1 to 1um.
Route: Whole-body Inhalation
Dose/Concentration: CB average
concentration: 160 ± 22/yg/m3
Time to Analysis: FA exposure on day 1, CB
exposure 3hfd for 3 consecutive days (days 2-
4)
On day 1, HR was significantly depressed
during FA in terminally senescent mice. By day
4, HR had significantly slowed due to the
effects of 3 days CB exposure. The combined
effects of terminal senescence and CB
exposure acted to depress HR to an average (±
SEM) 445 ± 40 bpm, ~ 80 bpm lower com-
pared to healthy HR responses. The change in
rMSSD was significantly greater on day 1 and
day 4 in terminally senescent mice, compared
to healthy mice. LF/HF ratio was significantly
depressed in terminally senescent mice on day
1. By day 4, significant increases in LF/HF were
evident in healthy mice during CB exposure.
Terminally senescent mice modulated a lower
HR without change in the LH/HF ratio during
CB exposure.
Reference: Thomson
et al. (2005, 087554)
Species: Rat
Gender: Male
Strain: Fischer-344
Weight: 200-250g
Urban Ambient Particles (EHC-93)
from Ottawa, Canada; Ozone
Particle Size: Respirable Modes
(aerodynamic diameter): 1.3 and 3.6
fjm. Non-respirable Mode
(aerodynamic diameter): 15/ym
Route: Nose-only Inhalation
Dose/Concentration: EHC-93: 0, 5, 50
mg/m3;
Ozone: 0, 0.4, 0.8ppm
Time to Analysis: 4h to particles, ozone, or
combination of particles and ozone.
Both pollutants individually increased preproET-
1, ET -1 and endothelia N0S mRNA levels in the
lungs shortly after exposure, consistent wI the
concomitant increase in plasma of ET-1 [1 -21 ].
Prepro ET 1 mRNA remained elevated 24 h
post-exposure to particles but no after ozone.
Both pollutants transiently increased ET-B
receptor mRNA expression, while ozone
decreased ET A receptor mRNA levels.
Coexposure to particles plus ozone increased
lung preproET-1 mRNA but not plasma ET-1 [1 -
21], suggesting alternative processing or
degradations of endothelins. This coincided wf
an increase of MMP-2 in the lungs (this enzyme
cleaves bigET-1 to ET-1[1-32]).
Reference: Thomson
E et al. (2006,
097483)
Species: Rat
Gender: Male
Strain: Fischer-344
Weight: 200-250g
Urban Ambient Particles (EHC-93)
from Ottawa, Canada; Ozone
Particle Size: NR
Route: Nose-only Inhalation
Dose/Concentration: EHC-93: 0, 50 mgfm3;
Ozone: 0, 0.8ppm
Time to Analysis: 4h to particles, ozone, or
combination of particles and ozone. Sacrificed
immediately following exposure or following
24h recovery.
Circulating levels of both ET -1 [1 -21 ] and ET-
3[1 -21 ] were increased immediately after
exposure to PM and ozone. While expression of
preproET-1 mRNA in the lungs increased,
expression of preproET-3 mRNA decreased
immediately after exposure. PreproET-2 mRNA
was not detected in the lungs, and exposure to
either pollutant did not affect plasma ET-2
levels. Coexposure to ozone and particles, while
altering lung preproET-1 and preproET-3 mRNA
levels in a fashion similar to ozone alone, did
not cause changes in the circulating levels of
the two corresponding peptides.
July 2009
D-24
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Totlandsdal et al.
(2008,157056)
Species: Rat
Gender: Male
Strain: WKYINCrl and
Crl: Wl (Han)
Age: Adult
Weight: 220-300 g,
WKYINCrl; 250-300 g,
Crl: Wl (Han)
Use: Isolation of
Primary Rat Epithelial
Lung Cells (PRELCs):
from WKYINCrl rats.
Isolation of Rat
Ventricular
Cardiomyocytes and
Cardiofibroblasts
(RVCMs and RVCFBs)
from Crl: Wl (Han)
rats.
Pigment Black Printex 90 (Frankfurt,
Germany): PM: SRM 1648
Particle Size: Printex 90:12-17nm;
PM: NR
Route: Cell Culture
Dose/Concentration: Printex 90: 0, 50,100,
200 or 400 /yglmL; PM: 0, 200 /yglmL
Time to Analysis: 20h
Lung cell cultures: Both particles induced
release of IL-6 and IL-1 p, whereas TNF a was
only detected upon exposure to PM.
Cardiac Cell Cultures: IL-6 release was
strongly enhanced upon exposure to
conditioned media, and markedly exceeded the
response to direct particle exposure. IL-1, but
not TNF-a, seemed necessary, but not
sufficient, for this enhanced IL-6 release. The
role of IL-1 was demonstrated by use the use
of an IL-1 receptor antagonist that partially
reduced the effect of the conditioned media,
and by a stimulating effect on the cardiac cell
release of IL-6 by exogenous addition of IL-1 a
and IL-1 p.
Reference: Tzeng H P
et al. (2007, 097883)
Species: Rat
Strain: Wistar
Cell Type: Primary
Vascular Smooth
Muscle Cell Culture
(VSMCs): isolated
from thoracic aortas
from 200-250g rats.
which was attenuate by antioxidants (NAC,
PDTC). The level of translocation of NF-
kappaB-p65 in the nuclei of VSMCs was also
increased during MEPE exposure. The
potentiating effect of MEPE in serum-induced
VSMC proliferation was abolished by COX-2
selective inhibitor NS-398, specific ERK
inhibitor PD98059, and antioxidants (NAC,
PTDC).
Motorcycle Exhaust Particulate
Extract (MEPE) collected from a
Yamaha motorcycle with a 50 cm3
two-stroke engine using 95% octane
unleaded gasoline.
Particle Size: PM1, PM2.6, PM10
Route: In vitro
Dose/Concentration: 10-100 //g/mL
Time to Analysis: 3d
Exposure of VSMCs to MEPE (10-100/yglmL),
enhanced serum-induced VSMC proliferation.
The expression of proliferating cell antinuclear
antigen was also enhanced in the presence of
MEPE. VSCMs treated with MEPE induced
increase COX-2 mRNA, protein expression, and
PGE2 production, whereas the level of C0X-1
protein was unchanged. MEPE increased the
production of R0S in VSMCs, in a dose-
dependent manner. MEPE triggered time-
dependent ERK1/2 phosphorylation in VSMCs
Reference: Tzeng H P Motorcycle Exhaust Particulate
et al. (2003, 097247)
Species: Rat
Strain: Wistar
Cell Type: Primary
Vascular Smooth
Muscle Cell Culture
(VSMCs)
Extract (MEPE) collected from a
Yamaha motorcycle with a 50 cm3
two-stroke engine using 95% octane
unleaded gasoline.
Particle Size: NR
Route: In vitro
Dose/Concentration: MEPE: 10 //g/mL;
Nifedipine: 10 //mol; Manganese Acetate:
100/ymol; Staurosporine: 1-2 nM; Chel-
erythrine: 1 /jm
Time to Analysis: 18h
MEPE induced a concentration-dependent
enhancement of vasoconstriction elicited by
phenylephrine in the organ cultures of intact
and endothelium-denuded aortas for 18h.
Nifedipine, manganese acetate, and
staurosporine, but not chelerythrine, inhibited
the enhancement of vasoconstriction by MEPE.
ML-9 inhibited the enhancement of
vasoconstriction by MEPE. MEPE enhanced the
phosphorylation of 20k-Da in rat vascular
smooth muscle cells. N-acetylcysteine
significantly inhibited the enhancement of vaso-
constriction by MEPE. A time-dependent
increase in R0S production by MEPE was also
detected in primary cultures of VSMCs.
July 2009
D-25
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Upadhay Ultrafine Carbon Particles (UFCP)
S et al. (2008,
159345)
Species: Rat
Gender: Male
Strain: SHR
Age: 6m
Weight: NR
Particle Size: Size- 31 ±0.3nm,
MMAD- 46nm, Surface area
concentration- 0.139
m2(particle)/m3(air), Mass specific
surface area- 807m2/g
Route: Whole-body Exposure
Dose/Concentration: 172//gfm:l
Time to Analysis: Acclimatized 2d. 1 d
baseline. 24h exposure. 4d recovery.
Sacrificed 1st or 3,d day of recovery.
Cardiophysiology: The mean arterial BP and
HR increased but returned to baseline levels by
the 4th recovery day. SDNN and HRV
decreased. RMSSD and LF/HF decreased but
were not significant.
Pulmonary Inflammation: UFCP did not
cause pulmonary inflammation.
Pulmonary and Cardiac Tissue: HO-1, ET-1,
ETA, ETB, TF, PAI-1 significantly increased in
the lung on the 3,d recovery day. HO-1 was
repressed in the heart, but the other markers
had slight, nonsignificant increases.
Systemic Responses: Neutrophil and
lymphocyte cell differentials significantly
increased on the 1st recovery day. Other blood
parameters were unaffected. The plasma renin
concentration increased on the first 2 recovery
days. Ang I and II concentrations increased on
the 1st recovery day but was not significant.
Reference:
Wallenborn et al.
(2008,191171)
Species: Rat
Gender: Male
Strain: Wistar Kyoto
(WKY)
Age: 13wks
Weight: NR
Zinc Sulfate (ZnS04, aerosolized)
Particle Size: NR
Route: Nose-only Inhalation
Dose/Concentration: 9.0±2.1 fjg zinc/m3,
35 ±8.1 /yg zinc/m3,123.2 ±29.6 /yg zinc/m3
Time to Analysis: Exposed 5h/d, 3d/wk,
16wks. Half of the rats used for
plasma/serum analysis, other half for isolation
of cardiac mitochondria.
A trend toward increased BALF protein was
seen. Cardiac mitochondrial ferritin had a small,
significant increase. Mitochondrial succinate
dehydrogenase and glutathione peroxidase had
small, significant decreases. Subchronic
exposure to 100 /yg/m3 caused expression
changes of cardiac genes involved with cell
signaling events, ion channels regulation, and
coagulation. No pulmonary-related effects were
seen.
Reference: Wellenius
GA et al. (2003,
055691)
Species: Dog
Gender: Female
Strain: Mixed mongrel
Age: NR
Weight: 14-17kg
CAPs
Particle Size: 0.26±0.04/ym
Route: Permanent Tracheostomy
Dose/Concentration: Median: 285.7 /yg/m3,
Range: 161.3-957.3/yg/m3
Time to Analysis: Thoracotomy and
tracheostomy performed. 5-13wks recovery.
Pairs of subjects: exposed 6h/d either 2nd or
3>d exposure time and filtered air other days.
5min preconditioning occlusion. 20min rest
interval. 5min experimental occlusion. Some
dogs exposed 6h/d, 4d (consecutive), filtered
air on day 4.
CAPs increased the ST-segment elevation and
remained elevated 24h after exposure. This
increase was seen in precordial leads V4and Ve.
Multivariate regression analyses showed that
the mass concentration of Si was significantly
associated with the peak ST-segment elevation
and integrated ST-segment change. Univariate
regression analyses showed Pb to also be
significantly associated with these measures.
CAPs had no effect on peak heart rate during
occlusion or the maximum occlusion-induced
increase in heart rate.
Reference: Wellenius
GA et al. (2004,
087874)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age: Adult
Weight: —250 g
Use: Rat Model for
Acute Myocardial
Infarction (AMI): Left-
ventricular Ml induced
by thermocoagulation.
Animals allowed to
recover for at least
12h after surgery.
CAPs:
CO
Particle Size: CAPs: PM2.6
Route: Whole-body Inhalation Chambers
Dose/Concentration: CO: 35ppm; CAPs
(median concentration): 350.5/yg.m3;
CAPs+CO: (CAPs median concentration):
318.2 /yg/m3
Time to Analysis: 1 h exposure to CAPs or
CAPs+CO for 1h. Exposure to pollutants was
preceded and followed by 1h exposure to FA.
Exposure experiments were performed during
the period of 07/2000 and 01/2003
CO exposure reduced the ventricular premature
beat (VPB) frequency by 60.4% during the
exposure time compared to controls. This
effect was modified by both infarct type and
the number of pre exposure VPBs, and was
mediated through changes in hear rate (HR).
Overall, CAPs exposure increased VPB
frequency during the exposure period, but this
did not reach statistical significance. This ef-
fect was modified by the number of pre-
exposure VPBs. In rats with a high number of
pre exposure VPB, CAPS exposure significantly
decreased VPB frequency (67.1%). Overall,
neither CAPs nor CO had any effect on HR, but
CAPs increased HR in specific subgroups. No
significant interactions were observed between
the effects of CO and CAPs.
July 2009
D-26
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Wellenius
et al. (2006,156152)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age: Adult
Weight: — 250g
Use: Rat Model for
Acute Myocardial
Infarction (AMI): Left-
ventricular Ml induced
by thermocoagulation.
Animals allowed to
recover for at least
12h after surgery.
CAPs: (Boston, MA)
Particle Size: CAPs: PM2.6
Route: Whole-body Inhalation Chambers
Dose/Concentration: CO: 35ppm; CAPs
(median concentration): 645.7//g.m3;
CAPs+CO: 37.9ppm
Time to Analysis: CAPs or CAPs+CO
exposure for 1h. Exposure to pollutants was
preceded and followed by 1h exposure to FA.
Among rats in the CAPs group, the probability
of observing supraventricular arrhythmias
(SVA) decreased from the baseline to exposure
and post-exposure periods. The pattern was
significantly different than that observed for
the FA group during the exposure period. In the
subset with one or more SVA during the
baseline period, the change in SVA rate from
baseline to exposure period was significantly
lower in the CAPs and CO groups only, when
compared to the FA group. No significant
effects were observed in the group si-
multaneously exposed to CAPs and CO.
Reference: Wichers
et al. (2004, 055636)
Species: Rat
Gender: Male
Strain: SH
Aqe: 75d
HP-12 (oil-combustion derived PM
obtained from inside wall of a Boston
power plant stack burning residual oil
(number 6).
Water-leachable constituents
(//gfmg): SO4 (217.3); Zn (11.4); Ni
(6.9); Fe (0.0); V (1.3); Cu (0.2); Pb
(0.0)
1M HCI-leachable constituents
Wmg): SO4 (220.6); Zn (15.5); Ni
(14.8); Fe (15.6); V (32.9); Cu (1.1);
Pb (1.7)
Particle Size: 3.76/ym (MMAD)
(GSD2.16)
Route: IT Instillation
Dose/Concentration: HP-12 (mg/kg): 0.00
(saline control), 0.83 (low), 3.33 (mid), 8.33
(high)
Time to Analysis: Single-dose Sacrificed 96h
or 192h post-IT.
Exposures to mid and high-dose HP-12 induced
large decreases in HR, BP, and body
temperature. The decreases in HR and BP were
most pronounced at night and did not return to
pre-IT values until 72h (HR) and 48h (BP) after
dosing. ECG abnormalities (rhythm
disturbances, bundle branch block) were
observed primarily in the high dose group.
Reference: Wold et UFPs from either ambient air (UFAAs) Route: IV Infusion (in vivo study)
al. (2006, 097028)
Species: Rat
Gender: Female
Strain: Sprague-
Dawley
Use: L jugular vein
and R carotid artery
were canulated.
or diesel engine exhaust (UFDGs);
UFIDs from industrial forklift exhaust
and soluble fraction UFID suspension,
particle free (SF-UFID)
Particle Size: UFAAs diameter ~
150 nm; UFDGs diameter ~ 10Onm
Dose/Concentration: UFDG (50//gjm|
Time to Analysis: Infused wfUFAA or UFDG.
Monitored continuously for 1 h then sacrificed.
Infusion of UFDGs caused ventricular prema-
ture beats (VPBs) in 2 out of 3 rats. Ejection
fraction increased slightly in rats receiving
UFAA and was unchanged in the UFDG and
saline groups.
Reference: Wold LE
et al. (2006, 097028)
Species: Rat
Gender: Female
Strain: Sprague-
Dawley
Use: Heart
Langendorff-perfusion
apparatus
UFPs from either ambient air (UFAAs)
or diesel engine exhaust (UFDGs);
UFIDs from industrial forklift exhaust
and soluble fraction UFID suspension,
particle free (SF-UFID)
Particle Size (Distribution): UFAAs
diameter; n150 nm; UFDGs diameter:
~100nm
Route: Lagendorff Heart Perfusion (in vitro)
Dose/Concentration: UFDG (100 /yg /2ml);
UFID (12.5 /yg/l in perfusate); SF-UFID (12.5
Afg/D
Time to Analysis: Lagendorff 1: Treated
w/UFDG. Lagendorff 2: Treated with UFID &
SFUFID. Both experiments were monitored
continuously for 1h after injection.
UFDGs caused a marked increase in left-
ventricular and end-diastolic pressure (LVEDP)
after 30 min of exposure. UFIDs caused a
significant decrease in left-ventricular systolic
pressure (LVSP) at 30min after the start of
infusion. This effect was absent when SF UFID
was studied.
Reference: Yatera et EHC-93 from Ottawa, Canada
al. (2008,157162) partic|e size: NR
Species: Rabbit
Gender: Female
Strain: WHHL
Age: 42wks
Weight: 3.2 ± 0.1 kg
(avg)
Route: IT Instillation
Dose/Concentration: PMm suspension: 5mg
EHC-93 in 1 ml saline
Time to Analysis: Exposed 2xfwk for 4 wks.
Acute effects observed at 0.5,1, 2, 4, 8,12,
and 24h after initial instillation. Chronic
effects observed 1x/wk for 4 wks.
Exposure to PM10 caused progression of
atherosclerotic lesions in thoracic and
abdominal aorta. It also decreased circulating
monocytes expressing high levels of CD31 and
CD49d, and increased expression of CD54
(ICAM-1) and CD106 (VCAM-1) in plaques.
Exposure to PM10 increased the number of
BrdU-labeled (*) monocytes into plaques and
into smooth muscle underneath plaques.
("onocytes labeled with BrdU in donor rabbits
were transfused to recipient rabbits as whole
blood, and the recruitment of BrdU-labeled cells
into vessel walls and plaques in recipients was
measured by quantitative histological
methodology.
July 2009
D-27
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Ying Z et
al. (2001,019011)
Species: Mice
Gender: Male
Strain: ApoE'1'
Age: 16wks
CAPs: PM2.6, New York City
(Manhattan), NY; May-Sept 2007
Particle Size: NR
Route: Inhalation
Dose/Concentration: 138.4 ± 83.7 //gfm3
Time to Analysis: 6h/d x 5dfwk * 4m
Vascular tone: Significant decrease in PE-
induced maximum contraction of aortic rings in
CAPsexposed mice. No difference in sensitivity
to PE between groups. Treatment with the
soluble guanylate cyclase inhibitor 0DQ
restored the response to PE in CAPs aortic
rings. No significant differences in relaxation
induced by ACh. CAPs abolished the relaxation
induced by Ca ionophore A23187. CAPs
exposure slightly (but significantly) decreased
maximum relaxation induced by SNP.
Protein expression: iNOS mRNA expression
was increased in the aortas of CAPs exposed
mice. eNOS and GTPCH levels were unchanged.
Distribution of inOS protein expression was
limited to plaque in air-exposed mice and was
found in the plaque and media for CAPs-
exposed mice.
Superoxide production: Superoxide levels in
CAPs exposed mice were increased in the aorta
compared to air-exposed mice. The addition of
L-NAME significantly increased superoxide
production. Extensive protein nitration in aortas
of CAPs mice. NADPH subunits Rac1 and p47
phox mRNA expression was increased in aortas
of mice exposed to CAPs.
Atherosclerosis: Significant increase in
area of CAPs exposed mice. Higher levels of
macrophage infiltration, collagen deposition,
and lipid composition of plaques from CAPs-
expsoed mice.
Reference: Ying Z et
al. (2001,019011)
Species: Mouse
Gender: Male
Strain: ApoE1
Age: 6wks
Weight: NR
CAPs (New York, NY) (May-Sept.
2007)
Particle Size: Diameter: < 2.5 /jm
Route: Exposure Chamber
Dose/Concentration: Ambient PM2.6:
23±75.9/yg/m3, Chamber PM2.6: 138.4±
83.7 /yglm3
Time to Analysis: Fed high-fat chow 10wks.
Exposed 6h/d, 4m.
Vascular: Constriction responses to KCI were
similar between CAPs exposed and control
mice. PE-induced maximum contraction
significantly decreased in CAPs mice, which
was attenuated by the soluble guanine cyclase
inhibitor 0QD. CAPs prevented calcium
ionophore A23187-induced relaxation. CAPs
decreased SNP-induced maximal relaxation.
iNOS, Aortic O2 : iNOS mRNA expression
increased. Aortic O2 increased and was further
increased by N0S inhibitor L-NAME.
Atherosclerosis: The only NADPH oxidase
subunits affected were Rac1 and p47pl"™,
which increased. Atherosclerotic burden in the
thoracic aorta, macrophage infiltration,
collagen deposition, and lipid composition in the
aorta all increased.
Reference: Yokota S
et al. (2004, 096516)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley (IGS)
Weight: 345-498.2g
DEP (obtained from the Japan
Automobile Research Institute)
Particle Size: NR
Route: IT Instillation
Dose/Concentration: Group 1: DEP:
1 mg/0.1 ml
Group 2: DEP: 0.2 ml (10, 12.5 or 25 mg/ml)
Group 3: DEP 2.5 or 5mg/0.2ml
Time to Analysis: DEP pre-treatment 24-72h
before ischemia/reperfusion.
DEP effects on myocardial
Ischemia/Reperfusion-induced arrhythmia:
An increased mortality was observed in the
DEP group compared to the vehicle-treated
group. 46% of the animals in DEP died during
the first 3 min reperfusion period. The animals
of other groups were intratracheal^ instilled
with DEP at the beginning of ische-
mia/reperfusion experiment, or were pretreated
with polyethylene glycol-conjugated SOD (1000
lU/kg, iv). In these animals, incidences of both
arrhythmia and mortality were similar to those
in the animals treated with the vehicle.
DEP effects on the biochemical and
hematological parameters: Neutrophil count
was elevated by a higher dose (5 mg) of DEP at
24 h after the IT instillation, and oxygen radical
production, which was induced by 12-0-
tetradecanoylphorbol 13-acetate, was
enhanced at 72h.
July 2009	D-28	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Yokota S
et al. (2005, 096003)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley (IGS)
Weight: 303-472.2g
Use: Rats were used
after a 7-day
acclimation period
DEP from Japan
Particle Size: NF
Route: IT Instillation
Dose/Concentration: Vehicle: 0.2 mLfanimal
DEP: 5 mg/animal
Time to Analysis: Single exposure 0.5,1, 2,
3, 6,12, 24, 48h.
At 12 and 24 h post-instillation, circulatory
neutrophil counts in the 5 mg DEP group were
significantly elevated, and were 2.1 -fold (12h|
and 2.3 fold (24 h) in vehicle treated animals. 1
mg DEP caused an increase of approximately
0.4-fold in CNC at 6h. 12-0-
tetradecanoylphorbol 13-acetate induced
oxyradical production (0RP) in the isolated
neutrophil was enhanced at 12 and 24 h after
instillation with 5 mg DEP. In Serum, a marked
elevation of CINC-1 and a slight elevation of
MIP-2 were also observed, while TNF a was
not detected. GM-CSF was not detected in
serum 24 h post-instillation.
Reference: Yokota S
et al. (2008,190109)
Species: Mouse
Gender: Male
Strain: ddy
Age: NR
Weight: 39.6-46.Og
DEP (DMSC (dichloromethane
soluble-component), RPC (residual
particle-component))
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 5mgfkg, 10mg/kg
Time to Analysis: DMSC and RPC extracted
from DEP. Mice acclimatized 7d. DEP, DMSC,
or RPC instilled. BALF and blood obtained and
G-CSF, GM-CSF, IL-6 measured 2, 4,12, 24h
post instillation.
Inflammation: At 5mgfkg DEP increased the
total cell and macrophage count. DEP or RPC
increased neutrophils at 5 and 10mg/kg.
10mg/kg DEP or RPC increased macrophages
at 4h and decreased at 12h.
Hematology: Compared to 5mg/kg DEP, RPC
increased RBC, WBC, and neutrophils. 10mg/kg
RPC or DEP caused sustained increases in RBC,
WBC, and neutrophils.
Cytokines: 5mgfkg RPC markedly increased G-
CSF and IL-6. Other cytokine increases at this
dose were transient. 10mg/kg DEP increased
IL-6 at 4h, and DEP or RPC increased G-CSF
and IL-6 at 12h. DEP or RPC also increased IL-
1p.
Myocardium: Myocardial MPO activity
significantly increased in 5mg/kg RPC at 12
and 24h. Myocardial MIP-2 increased the most
in 5mg/kg RPC, while LIX tended to be lowered
by RPC.
Table D-2. Respiratory effects: in vitro studies.
Study
Pollutant
Exposure
Effects
Reference: Aam BB
and Fonnum F, (2007,
155123)
Species: Human, Rat
Tissues/Cell Types:
Human-Neutrophil
Granulocytes (NG);
Rat- Alveolar
Macrophages (AM)
DEP: SRM 1975
Particle Size: NR
Route: Cell Culture
Dose/Concentration: NG:
280 /yglmL
AM: 140, 280/yg/mL
Vitamin E - 5//M
Time to Analysis: 1h
ROS of NG: Formation of ROS in NG decreased
with increased doses of DEP. Lucigenin
chemiluminescence of ROS formation diminished
25% at 8.8/yglmL DEP and luminol
chemiluminescence 32% with 17.5/yglmL DEP.
DCF fluorescence required much higher doses of
DEP. Controls without PMA stimulation had highly
reduced lucigenin and luminol with DEP dose of
140/yglmL while DCF increased 116%.
ROS of AM: 280/yglmL of DEP decreased ROS
level by 19% with DCF. DEP with PMA-
unstimulated cells increased 24% with DCF.
Necrosis: NG cell death was DEP dose-dependent.
At 280 /yglmL, cell death increased 5.4% as
compared to control. LDH concentration increased
1.6% with 70/yglmL DEP and 3.9% with 280
/yg/mL after 1h.
July 2009
D-29
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Agopyan
et al. (2003, 056065)
Species: Human
Tissues/Cell Types:
BEAS-2B, NHBE,
SAEC
PC: synthetic carboxylate-modified particles
Particle Size: 2,10//m
Route: Cell Culture
Dose/Concentration:
PC2 - 0.83 glmL or 3.4x109 par-
ticles/mL
PC10 - 0.8 g/mL or 3x10e parti-
cles/mL
Time to Analysis:
PC2 - 12, 24, 8h
PC10 - 2, 6,12, 24h
Calcium Imaging: PC10 induced increase of
Ca2+ concentration in all capsaicin-sensitive cells
100%. Similar reaction observed in cells exposed
to PC2. However, more than 3-PC2s were required
to induce a Ca increase unlike PC10. CPZ (1 Oum)
and amiloride could fully block PC-induced
response.
cAMP: Post 6h, a dose-dependent increase in
cAMP was observed. Again, CPZ blocked increase
by 70-90% depending on cell type: SAEC > NHBE
~ BEAS-2B.
Apoptosis: PC10 and PC2 induced apoptosis
time-dependently. PC2 was slower in induction
than PC10. Post 48h, 80-95% cells were
apoptotic in all cell types. Noncapaisin-sensitive
cells (which did not bind to particles) did not
exhibit apoptosis. CPZ reduced apoptosis by 97%
BEAS-2B, 96% NHBE and 98% SAEC. Amiloride
did not block apoptosis.
Necrosis: Induction of necrosis by PC2 and PC 10
was negligible. A slight increase from 1 % to 2%
was observed at 24-48h in NHBE and SAEC.
BEAS-2B showed slight decrease from 3% to 4%
in same time i
Reference: Agopyan R0FA
et al. (2004,156198) MSHA: Mt St He|en Ash
Species: Human, Particle Size: NR
Mouse
Tissues/Cell Types:
Human-NHBE, SAEC;
Mouse-Wildtype and
TRPV1 (-/-) Terminal
Ganglion Neurons
(TG)
-257 Wildtype
Neurons
-187 TRPV (¦/¦)
Neurons
Apoptosis: R0FA or MSHA induced time-
dependent apoptosis, peaking at 24 h. CPZ again
inhibited this response. Neurons bound to PM
(< 25um) induced apoptosis in TRPV1 (+/+). Cells
without bound PM or bound with PM (> 25 //m)
showed no effect. No apoptosis occurred in the
absence of Ca2+.
Necrosis: Necrosis for any of the cell types was
negligible.
PKA: Inhibition of PKA resulted in 90+%
apoptosis in NHBE and SAEC. Again, no apoptosis
was observed in a Ca2+ free environment.
Route: Cell Culture
Dose/Concentration: 100 //glmL
R0FA or MSHA
Time to Analysis: ROFAfMSHA
in NHBE and SAEC - 2,6,24,
48 h
ROFAfMSHA in TG - 24 h
cAMP measurements with NHBE
and SAEC exposed to
R0FA/MSHA - 6h
Calcium Imaging in NHBE and SAEC: In 100%
of reactive cells, ROFAfMSHA induced an increase
in Ca2+. Levels remained elevated as long as PM
bound to plasma membrane. Washing and
disjoining PM from membrane caused Ca2+ to
slowly decline to baseline. CPZ (or CPZ and
amiloride) reversibly inhibited PM-induced rises in
Ca2+.
Calcium Imaging in TRPV1 (+ /+) and (-/-) mice
sensory neurons: All sensitive neurons in
TRPV1 (+/ +) increased Ca2+ in response to
R0FA. No effect of R0FA in TRPV1 (¦!¦).
cAMP: R0FA and MSHA induced increases in
Ca2+ in NHBE and SAEC cells, which was
completely blocked by cAMP.
Reference: Ahn E-K
et al. (2008,156199)
Species: Human
Tissues/Cell Types:
A549
DEP: (6 cyl, 11L, turbo-charged, heavy-duty
diesel engine, South Korea)
Dex: anti-inflammatory (Sigma, St. Louis, M0)
Particle Size: NR
Route: Cell Culture
Dose/Concentrations: 0,1, 5,
10, 50 and 100 //g/mL of DEP
Some cells pre-treated with 10,
20, 40, 50 pg/mL of Dex.
Time to Analysis: 24h
COX-2 Expression: Cells expressed dose-
dependent increases in C0X-2 expression after
treatment with 10-100 //glmL of DEP. Treatment
of 50 //glmL for 24 h induced statistically
significant C0X-2 expression in both mRNA and
protein levels. Pre-treatment with Dex
significantly reduced expression of C0X-2 mRNA
and protein. Dex treatment induced dose-
dependent suppression of DEP-induced protein
levels.
PGE2 Levels: Levels of the inflammatory
mediator, PGE2, increased when were cells
exposed to 50//glmL of DEP. Pre-treatment with
50/yg fniL Dex completely inhibited DEP-induced
release of PGE2.
July 2009
D-30
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Ahsan
(2005,156200)
Species: Human
Tissues/Cell Types:
T rx-1 -transfected
Clone of Murine L-
929 cells; Control
Clone (L-929-l\leo1);
A549
DEP: provided by Dr. Masaru Sagai, University Route: Cell Culture
of Health and Welfare, Aomori, Japan	Dose/Concentration:
Particle Size: NR
DEP: 50 //g/mL
hTrx-1 ¦ or L-929-Neo1: 40 /yglmL
Pretreatment: rhTrx-1 (10/yg/mL)
or DM-rhT rx-1 (NR)
Time to Analysis: Pretreatment
for 1 h. Parameters measured 3h
post exposure.
ROS: DEP induced significant increases of R0S in
L929-Neo1 cells. hTRx-1 cells showed no affect.
RT-PCR revealed hT rx-1 mRNA expression in
transfected cells but not control L929-Neo1 cells.
Endogenous murine Trx-1 mRNA expression
increased in control cells, but not in hT rx-1 cells.
A549 cells had increased ROS levels but these
levels were suppressed with rhTrx-1 pretreatment.
Pretreatment with DM-rhT rx-1 increased ROS
levels more.
Akt (antiapoptotic molecule): Phosphorylated
Akt prevents apoptosis. DEP induced
phosphorylation of Akt in control cells after 3h
and ^phosphorylation after 5h. In hT rx-1 cells,
Akt remained phosphorylated after 5h. In A549
cells, Akt phosphorylated at 3h and slowly turned
off at 12-24 h. Pre-treatment with rhTrx-1
blocked dephosphorylation. This suggests that
Trx-1 preserves active form of Akt and thereby
protects against cytotoxicity from DEP.
Reference: Alfaro-
Moreno et al. (2002,
156204)
Species: Human,
Mouse, Rat
Strain: Human-A549;
Mouse-J7 74A.1,
BALB-c
Tissues/Cell Types:
HUVEC, Mouse
Fibroblasts, Rat Lung
Fibroblasts (RLF)
PMni: Collected from 3 zones in Mexico City:
North (industrial), Center (business) and South
(residential)
Particle Size: PMio
Route: Cell Culture
15000 cells/cm2 except:
Cytotoxicity: Confluent Cultures
180,000 cells/cm2.
DNA Breakage: 20,000 cells/well.
Cytokine Assays: 180,000
cells/cm2
Dose/Concentration:
Cytotoxicity: 10, 20,40, 80,160
/yg/cm2
Apoptosis: 160/yg/cm2
DNA Breakage: 2.5, 5,10, 20, 40
/yg/cm2
Cytokine Assays: 10, 20, 40, 80
/yg/cm2
E-Selectin Expression: 40 /yglcm2
Time to Analysis: Cytotoxicity:
24, 48, 72h; Apoptosis: 24 h;
DNA Breakage: 72h; Cytokine
Assays: 24 h
Cytotoxicity: Cytotoxic effect exhibited dose-
dependency after 72h in proliferating cells of
J774A.1, BALB-c and RLF cell lines.
Proliferating Cells: Northern particles induced a
statistically larger effect than central or southern
particles. J774A.1 was more susceptible while
BALB-c was less susceptible. A549 was most
resistant to decreased viability during exposure.
No significant variation in viability was observed
when compared to the control. Particles were not
cytotoxic among confluent cell growth for any cell
lines when exposed to 20-160 /yglcm2.
Apoptosis: Overall, particles induced low rates of
cell death via apoptosis. J774A.1 depicted similar
levels of apoptosis when exposed to three PM
zones, ~ 15% apoptotic cells measured. BALB-c
was not reported. Results for the A549 measured
apoptotic cells were: South- 4%, Central- 11 % and
North- 15%. HUVEC cells indicated an increase in
apoptosis with northern particles.
DNA Breakage: PMio from all zones induced DNA
breakage. A dose-dependent relationship was
established with PM2.6 particles at concentrations
of 10 /yglcm2. The Southern zone required a higher
dose of PM (10/yg/cm2) to produce the same
effect as other zones (2.5/yg/cm2).
Cytokines: Particles induced TNF-a and IL-6
secretion in J774A.1 cells dose-dependently. IL-6
increased significantly with central particles.
PGE2 secretion in RLF cells induced by exposure
to PM showed dose-dependent responses. PM
from the central zone induced the most PGE2
secretion. Max secretion was observed at doses
of 40 /yglcm2 from all three PM zones.
E-Selectin Expression: HUVEC cells showed a
25% increase in E-selectin expression after
exposure to 40 /yglcm2 of PM.
July 2009	D-31	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Amakawa
et al. (2003,156211)
Species: Mouse,
Human
Strain: Mouse-ICR
Tissues/Cell Types:
AMs
Gender: Mouse-Male;
Human-Male
Age: Mouse 6-7wks;
Human 20-24yrs
DEP
(obtained from a 4JB1, Isuzu, 1500 rpm, 4cyl
diesel engine)
DEPE - DEP Extract (methanol)
CB - Charcoal (Sigma)
Particle Size: DEP- 0.4/ym, CB- 0.7/jm
Route: Cell Culture
Mouse: 5X106 cells/mL Human:
3X106 cells/mL
Dose/Concentration: DEP - 1 or
10/yg/mL:DEPE - 1 or 10//g/mL;
CB - 1,10,100//g/mL
Time to Analysis: Human cells
pre-treated with LPS 1 /yglmL.
Murine cells pre-treated with SOD
300 lll/mL. Parameters measured
24h post exposure.
Cells: For mice, more than 90% of the cells were
macrophages and over90% were viable. For
humans, 96% of the cells were macrophages, 3%
lymphocytes and 1 % neutrophils: over95% of the
human cells were viable.
DEP Cytotoxicity: None observed
Cytokines: DEP (10//gfmL) suppressed release of
TNF-a and IL-6 for both mice and humans in a
dose-dependent manner. Murine cells pre-treated
with LPS or IFN-y released even less TNF-a and
IL-6. IL-10 was unaffected. Human macrophages
pre-treated with LPS also released lower levels of
TNF-a, IL-6 and IL-8.
ROS: Pre-treatment of SOD on murine cells
partially attenuated the suppressive effect of DEP
as well as decreased the production of ROS
generated by DEP (10/yg/mL).
Carbon: Carbon particles did not suppress TNF-a
or IL-6 release from murine AMs; however,100
/yg/mL of CB stimulated TNF-a production.
Methanol: No cytotoxicity nor cytokine release
effects were observed.
DEPE: DEPE suppressed TNF-a and IL-6 release in
a similar way as DEP.
Reference: Amara et
al. (2007,156212)
Species: Human
Cell Lines: A549,
NCI-H292
DEP - SRM 2975
CSC - cigarette smoke condensates
(collected from Kentucky standard cigarettes,
2R4F; University of Kentucky)
DC - DEP + CSC
CB (Degussa, Frankfurt, Germany)
Particle Size: CB: 95nm; DEP: NR
Route: Cell Culture
Dose/Concentration: DEP -
10/yg/cm2
CB - 10/yg/cm2
CSC - 10/yg/cm2
Time to Analysis: 6 or 24h
Inflammatory Markers: LDH of A549 was
unaffected at either time point with DEP or CB.
LDH increased with CSC at concentrations high
than 10/yg|mL at both time points. DC had no
effect.
Proteases: MMP-1 mRNA expression showed a
dose dependent increase with DEP in A549 cells.
DEP also increased MMP-1 in NCI-H292 cells. CB
and CSC had no effect. MMP-1 mRNA expressions
were inhibited by N-acetylcysteine antioxidant.
Similar inhibition was observed with N0X4
oxidase. DC induced a similar effect to DEP. MMP-
1 protein expression increased post 24 h with
DEP. MMP-2, TIMP-1, TIMP-2 mRNA expression
was unaffected.
TGF: TGF (3 mRNA expression was unaffected.
ROS: DEP and DC increased ROS formation after
1h. DEP effect was inhibited by N-acetylcysteine
antioxidant pre-treatment.
MAP-Kinase: DEPinduced MMP-1 expression
increased ERK1/2 phosphorylation after 10 min,
peaking at 30min, and returning to normal levels
at 60min. Treatment with CBPs did not increase
ERK1/2 phosphorylation whereas treatment with
CSC resulted in phosphorylation. Only inhibitors of
ERK1/2 reduced DEP induced MMP-1 activity. P38
and JNK inhibitors had no effect.
Reference: Anseth et
al. (2005, 088646)
Species: Human
Cell Lines: A549;
A549-p0 (lacking
mitrochondria)
s-ROFA: soluble portion
Particle Size: 1.95 ± 0.18 /jm
Route: Cell Culture (3X10E
cells.i'niL)
Dose/Concentration: 100 //gfmL
Time to Analysis: Experiments
conducted by spreading monolayer
of Infasurf (calf lung surfactant
extract on PBS, PBS+R0FA or
conditioned media from A549
AEC. Parameters measured after
16h incubation period.
Lung Surfactant Gelation: R0FA alone and
A549 conditioned media alone did not significantly
alter Infasurf rheology. However, conditioned
media from A549 AEC at 16h induced a
significant increase in elastic storage and viscous
loss moduli. Inhibiting ROS production lowered
effect, indicating s-ROFA gelation mediated
through ROS.
ROS: ROS mediated through mitochondria as
evidenced by the effect of R0FA-AEC on
surfactant gelation in the presence of
mitochondria ROS inhibitors as well as A549-p0
cells.
July 2009	D-32	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Auger et
al. (2006,156235)
Species: Human
Tissue/Cell Type:
Nasal Epithelial Cells
DEP: SRM1650
PM2.6: (obtained from a highway in Paris,
France)
Particle Size: DEP: 400 nm (mean diameter);
PM2.6
Route: Cell Culture (2-3.5x104
cells/cm2)
Dose/Concentration: 10-80
/yg/cm2
Time to Analysis: Cells treated
on apical side. Parameters
measured 24h following
treatment.
Cytotoxicity (LDH): No cytotoxicity for DEP or
PM2.6 (80/yg/cm2).
Cytokines: In non-stimulated ALI cultures, IL-8
was the most abundantly secreted cytokine,
followed by GM-CSF, TNF-a, and IL-6 in
decreasing levels of production. Amphiregulin was
moderately, but consistently, secreted. After
treatment, both DEP and PM2.6 induced IL-8 and
amphiregulin release in a dose-dependent manner
through the basolateral surface. PM2.6 stimulated
IL-6 and GM-CSF release through the apical
surface.
ICAM-1 expression: No effect from DEP or
PM2.6.
ROS: DEP and PM2.6 both increased R0S
production in a dose-dependent manner.
Reference: Bachoual
et al. (2007,155667)
Species: Mouse
Cell Type: RAW
264.7
PM10 from two Paris, France subway sites:
RER and Metro
CB (Frankfurt, Germany)
Ti02 (Calais, France)
DEP: SRM1650 (NIST)
Particle Size: CB: 95nm; Ti02:150um; DEP:
NR;
RER PM10: 79% < 0.5 fjm, 20% 0.5-1 fim;
Metro PM10: 88% <0.5 fjm, 11 % 0.5-1 fjm.
Route: Cell Culture (40,000
cellsfmL)
Dose/Concentration: All
particles: 0.01, 0.1,1,10/yg/cm2
Time to Analysis: 3, 8, 24h
Cell Viability: No effects from any particulate at
concentrations up to 10 /yglcm2 for 24 h.
Inflammatory Effect: Exposure of cells to
10/yg/cm2 of RER or Metro induced time-
dependent increase in TNF-a and MIP-2 protein
release. This effect was similar to both locations.
No effect was observed at low concentrations of
PM10. No effect of CB, Ti02 or DEP was observed.
GM-CSF or KC production: RER and Metro PM10
did not induce any effect at any concentration.
Effect on Protease mRNA Expression: Exposure
of cells to 10/yglcm2 RER or Metro PM10 did not
modify mRNA expression of MMP-2 or -9 or their
inhibitors TIMP-1 and -2. MMP-12 expression
significantly increased after exposure to RER or
Metro PM10 for 8h.
Effects on HO-1 Protein Expression: Exposure
to 10 /yglcm2 of RER or Metro PM10 for 24 h
induced positive cytoplasmic staining for H0-1.
Reference: Batalha
et al. (2002, 088109)
Species: Rat
Gender: Male
Strain: SD
Age: NR
Weight: 200-250g
CAPs (Harvard Ambient Particle Concentrator)
Particle Size: Mean: 2.7 fjm
Route: Inhalation Chamber
Dose/Concentration: Range:
73.5-733 //g/m3
Time to Analysis: CAPs exposure
5h/d, 3d (consecutive). SO2
exposure to induce CB 5hfd,
5d/wk, 6wks. Killed 24h
postexposure.
Histopathology: CAPs slightly increased the wall
thickness of small pulmonary arteries and edema
in the adventitia and hyperplasia of the terminal
bronchiole and alveolar ducts epithelium.
L/W ratio: The L/W ratio decreased in CAPs-
exposed rats as particle mass, Si, Pb, SO42 , EC
and 0C increased. Univariate analyses showed
significant negative correlations between the L/W
ratio and Si and SO42 in normal rats and Si and 0C
in CB rats. Multivariate analysis showed only Si to
be significant in both groups.
July 2009	D-33	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Baulig et
al. (2007,151733)
Species: Human
Cell Line: 16-
HBE14o-
WUB: Winter Urban Background Particles
(obtained from Vitry-sur-Seine, suburb of
Paris, France)
SUB: summer Urban Background Particles
Vitry-sur-Seine)
WC: Winter Curbside Particles, SRM1648
(obtained from Porte-d'Auteuil, ring road of
Paris, France)
SC: summer Curbside Particles, SRM 1648
(Porte-d'Auteuil)
DEP: SRM 1650a (NIST)
DPL (control)
Particle Size: WUB, SUB: PM? b; WC, SC,
DEP: NR
Route: Cell Culture (20,000
cells/cm2)
Dose/Concentration: 10 //gjcnr
Time to Analysis: 18 or 24h
EGF: All native PM2.6 induced similar AR secretion
by bronchial epithelial cells (in decreasing order
WC, WUB, SC, SUB), but this release was
significantly greater than the release induced by
DEP. Pcellulin increased with SC, WUB and WC.
No data was available forSUB or DEP.
Interleukins: IL-1 a increased significantly with
WUB, WC,SC, DEP, DPL (in decreasing order). No
data was available for SUB. Exposure to WUB
caused IL-1 (3 to increase to induction factor of
over 2. IL-11 R a decreased significantly with
SUB.
Cytokines: Exposure to WUB caused G-CSF to
increase with an induction factor of over 2.Though
not statistically significant, TNF-R1 also
increased.
Proteases: TIMP-2 decreased with WUB but
significantly increased with SUB. Overall, SUB
downregulated integrins and interleukins seen
with other particles while upregulating
neurotrophic factors, chemokine receptors and
adhesion molecules. MMPs were not measured.
Chemokines: CCR-3 significantly increased with
SUB. GRO-y and GRO-a increased with WC at
both 18 and 24h. DEP had no effect with GRO-a.
Removal of metal from particles lowered response
of GRO-a.
Reference: Bayram DEP: (obtained from a 4JB1 -type, light-duty, 4 Route: Cell Culture
et al. (2006, 088439) cyl, 2.74-L Isuzu diesel engine)
Species: Human DEP-FCS: DEP + FCS
Cell Type: A549 DEP-NAC: DEP + N-acetylcystine, antioxidant
DEP-A: DEP +AE0L10113, catalytic
antioxidant
DEP-S: DEP + SP600125, inhibitor of JNK
DEP-N: DEP + SN50, inhibitor of NF-kB
Particle Size: DEP: 0.4/ym (mean diameter)
Dose/Concentration: DEP: 0, 5,
10, 50,100, 200/yg/mL
Time to Analysis: 24,48, 72h
Cell Growth: With 10% FCS (as a positive
control), A549 cells exhibited time dependent
growth. A mixture of FCS and DEP did not affect
cell growth for up to 48h. With DEP alone, cell
growth was prevented from cell number reduction
due to removal of serum at 48 and 72h. A dose of
10//gjmL induced a maximum proliferation effect.
Cell Cycle: DEP increased the percentage of
serum-starved cells in S phase at 48h. DEP
decreased the percentage in G0/1 phase and G2/M
phase.
Apoptosis: DEP prevented the increase in
apoptotic, serum-starved cells.
Protein Expression: p21 CIP1 /WAF1 expression
increased at 48h. DEP dose-dependently
decreased this expression.
NAC: NAC alone, at 33 mM, induced an increase
in cell numbers. DEP-NAC inhibited cell numbers at
48h. DEP-NAC inhibited cell numbers in S phase;
thus, cells in G0/1 phase increased. DEP-NAC
induced a further decrease of cells in G2/M phase.
AEOL10113: DEP-A caused a dose-dependent
decrease in cell numbers.
SP600125: Alone, SP600125 increased cell
numbers at 33mM. DEP-S decreased cell numbers.
Reference: Becheret SPM - suspended PM SRM-1648
al. (2007, 097125)
2007 #8030}
Species: Rat
Strain: CrlfWky
Cell Type: Alveolar
Macrophages,
Alveolar Type II
Gender: Male
Weight: 200g
Particle Size: SPM: 6-8 fjm
Route: Cell Culture (1.5x100
cells/well AM; 6x106 cells/well
Type II)
Dose: 200/yg/mL - 20/yg/cm2
Time to Analysis: 20h
Cytokines in Macrophages: SPM increased TNF-
a and MIP-2. NADPH inhibitor DPI reduced MIP-2
response, whereas iNOS inhibitor 1400W did not
affect either.
Cytokines in Type 2 Cells: SPM increased IL-6
and MIP-2 significantly. This SPM effect was
inhibited by DPI, whereas1400W reduced the IL-6
response significantly.
ROS in Type 2 Cells: SPM significantly increased
R0S formation. DPI largely blocked this SPM
effect.
ROS in Macrophages: No significant increases
were observed.
July 2009	D-34	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Becker et
al. (2005, 088590)
Species: Human
Gender: Male and
Female
Age: 18-35yrs
Cell T ypes: Alveolar
Macrophages, NHBE
PM (Coarse, Fine, Ultrafine): (EPA, Chapel H
NC, Chem Vol Cascade)
Particle Size: PM C: PMib; PM-F: PM0.1;
PM-UF: < 0.1/ym
Route: Cell Culture (0.5-1 x 10!l
cells/well NHBE; 2-3 x 10B/mL AM)
Dose/Concentation: NH BE: 25,
50,100, 250/yg|mLof PM;AMs:
50 /yglmL of DEP or 10ng/mL of
LPS
Time to Analysis: 18h for NHBE;
overnight for AMs
Cytokines: All 3 fractions induced dose-
dependent increases in IL-8 secretion with PM-c,
PM-F, PM-UF (in order of decreasing effects). TLR-
2 antibody blocked these particle induced IL-8
effects.
Inhibitors of Endotoxin effects and TLR-4
activation: No effects were observed in NHBE,
but all 3 fractions repressed the IL-6 release in
AMs.
TLR mRNA Expression: PM did not affect TLR-2
mRNA in NHBEs. PM-C and PM-F induced a slight
increase in TLR-4 mRNA in NHBEs while PM-UF
induced a substantial increase. PM-C increased
TLR-2 mRNA in AMs and decreased TLR-4 mRNA
in AMs.
Induction of Hsp70: PM-C and PM-F induced
Hsp70 in NHBE dose-dependently. Hsp70 was not
induced in AM following particle stimulator.
Reference: Becker et PM (Coarse, Fine, Ultrafine): (EPA, Chapel I
al. (2005, 088592)
Species: Human
Gender: Male
Age: 18-35yrs
Cell Types: Alveolar
NC, Chem Vol Cascade)
R0FA
Fe, Si, Cr Components
Oct 2001, Jan 2002, April 2002, July 2002
Particle Size: PM-C: 2.5-10/ym; PM-F:
Macrophages, NHBE < 0.1 //m; PM-UF: < 0.1 fjm
Route: Cell Culture (3-5x10E
cells/well NHBE; 2-3X106 cellsfmL
AM)
Dose/Concentration: NHBE: 11
/yg/mL of PM; AM: 50 /yglmL of
PM
Time to Analysis: 18-24h NHBE;
18h AM
IL-8 Release in NHBE: PM-C and PM-UF induced
effects. No effects from PM-F (all 4 dates).
IL-6 Release in AM: All 3 fractions induced
increase with later dates having generally lower
effects.
ROS (DCF): NHBE, at lower exposures, were
observed to be more responsive to PM than AMs.
AM exhibited highly variable results over time.
ROS (DHR): NHBE cells were observed to be more
responsive to PM than AMs. AM responsiveness
to PM increased over 4 time periods; this was not
observed in NHBE.
Seasonal Variability: Coarse particles were
more potent than F and UF regardless of the
month, and the potency for PM to induce IL-6/IL-8
production varied significantly. Coarse particles
induced a 5-25 fold change in IL-6 release for AMs
and a 3-6 fold change in IL-8 release for NHBEs.
Metal Correlation to IL-6/8 induction: Fe and Si
were positively associated with IL-6 release in
AMs incubated with the coarse fraction. Cr was
positively associated with IL-8 release in NHBE
cells incubated with F or UF.
Reference: Beck-
Speier et al. (2005,
156262)
Species: Human,
Canine (Beagle)
Cell Types: Human
AMs, Canine AM
(CAM)
DEP - SRM 1650a (NIST)
EC - Ultrafine Elemental Carbon (spark
discharge)
P90 - Printex 90 (Carbon Black, Degussa)
PG - Printex G (Carbon Black, Degussa)
Particle Size: (in diameter) DEP: 20-40nm;
EC: 5-1 Onm; P90:14nm; PG: 51 nm
Route: Cell Culture (1X106
cellsfmL AM)
Dose/Concentration: All
particles: 1 (EC only), 3.2,10,
100/yg/mL
Time to Analysis: 60min
Phagocytosis: All particles were phagocytosed
by CAM within 60 min.
Oxidative Potential: EC showed a very high
32, effect. DEP, P90 and PG had no effect
Formation of Lipid Mediators: DEP, EC P90 and
PG increased arachidonic acid and PGE2/TXB2 in
CAM in a dose-dependent manner. Only EC
increased LTB4 and 8-isoprostane.
ROS Activation: All particles increased activity in
canine macrophages with EC, P90 and PG
increasing activity in a dose-dependent manner.
DEP increased activity in canine macrophages.
Similar results were observed human alveolar
macrophages but only EC and P90 were tested.
Particle Mass vs Particle Surface Area:
PGE2/TXB2 effects were highly correlated with
particle surface area.
July 2009
D-35
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Bitterle
et al. (2006,156276)
Species: Human
Cell Type: A549
C-UFP - ultrafine carbonaceous particles
(obtained from a spark discharge aerosol
generator GFG 1000, Palas, Karlsruhe,
Germany)
Particle Size: 90nm (count median mobility
diameter)
Route: Cell Culture (3 x 107 cells)
Dose/Concentration: 44 ± 4
ng/cm2; 87 ± 23 ng/cm2; 230 ±
70 ng/cm2
(ng/cm2 - total mass of deposited
particles per cm2 cell monolayer
after 6h exposure)
Time to Analysis: 6h
Cell Viability: Exposure to clean air resulted in a
93.7 ± 9.1% viability. Exposure to low, mid nad
high doses of C-UFP resulted in a 94.9 ± 9.5%
viability. Thus C-UFP had no effect on cell
viability.
Interleukins: Clean air controls induced a 2-3 fold
increase in IL-6 and IL-8 production vs submersed
control. U-CFP exposures induced a similar effect
on IL-8 and IL-6 levels.
Antioxidant enzyme HO-1: The mid dose
increased transcription of H0-1 by 2.7 fold. There
was no observed effect at the high dose level
which indicates possible cytotoxcity.
Reference: Blanchet
et al. (2004, 087982)
Species: Human
Cell Type:16HBE
PM2.6
(Vitry-sur-Seine, Paris, France)
DEP - SRM1650a
CB - Carbon Black (Degussa)
Ti02 (Huntsman)
Particle Size: CB: 95 nm; Tith: 150 nm
Route: Cell Culture (45,000
cells/cm2)
Dose/Concentration: All
particles: 0.1,1,10, 30/yg/cm2
Time to Analysis: 6,18, 24, 30h
Amphiregulin Expression: DEP and PM2.6 both
increased AR mRNA expression from 6 to 30h,
with PM2.E inducing higher expression levels than
DEP. Both DEP and PM2.6 increased AR protein
secretion. No observed effect for CB and Ti02.
PM2.E induced protein secretion dose-dependently.
Signal Pathways in AR Secretion: MAP kinase
and tyrosine kinase inhibitors reduced effects of
DEP and PM2.6 but p38MAP kinase inhibitor did
not.
Role of Oxidative Stress: N-Acetylcysteine
blocked AR secretion following PM2.6. Antioxidant
enzyme catalase had no effect.
Cytokines: DEP induced a significantly high
release of GM-CSF, higher than PM2.6. EGFR
antibody reduced GM-CSF release at 0.25/yg/mL
dose.
Reference: Bonvallot DEP: SRM 1650
et al. (2001,
156283)1,
2001 #7852}
Species: Human
Cell Type:
16HBE14o-
0E-DEP: dichloromethane extract (2x) of DEP
nDEP: native DEP
sDEP: nDEP - 0E-DEP
CB: Carbon Black FR103 (Degussa)
BaP: Benzo[a]pyrene
CB: 95nm
NR
Particle Size: CB: 95nm; DEP: NR
Route: Cell Culture (3 x 10e cells)
Dose/Concentration: DEP, sDEP,
nDEP and CB - 10/yg/cm2
0E-DEP - 15/yg/mL
BaP - 0.25, 50 and 250/yg/mL
Time to Analysis: 24h
Proinflammatory Response: At 10/yg/cm2, nDEP
induced GM-CSF release by 4.7 fold. 0E-DEP
increased GM-CSF by 3.7 fold. BaP and sDEP also
induced increases of CN-CSF but had smaller
effect. CB had no effect.
NF-kB Activation: nDEP and 0E-DEP induced
enhanced degradation of IkB at 2-4h and 1 h
respectively. NF-kB DNA binding was enhanced by
0E-DEP (15/yg/mL, peak < 1 h) and nDEP (10
/yg/cm2, peak at 2-h with plateau till 4 h). Both
0E- and nDEP enhanced NF-kB DNA binding levels
were higher than BaP enhanced binding levels.
CYP1A1 mRNA: The CYP1A1 mRNA level was
markedly increased in nDEP and 0E-DEP treated
cells in comparison with their respective controls.
Radical Scavengers (decreased ROS in situ):
Increases of GM-CSF and NF-kB DNA binding by
nDEP and 0E-DEP was attenuated by radical
scavengers.
MAPK Activation: Increases by nDEP and 0E-
DEP of GM-CSF was inhibited by Erk1/2 inhibitor
but not by p38 inhibitors. Both nDEP and 0E-DEP
triggered Erk1/2 and p38 phosphorylation. sDEP
affected p38 phosphorylation only.
Reference: Brown et
al. (2007,156300)
Species: Human,
Mouse
Cell Type: PBMC,
A549 (Human):
J774A.1 (Mouse)
PM10 (London, England)
CM from PMio-treated human monocytes
Particle Size: PM10
Route: Cell Culture (1 x 10e
cells/mL J774A.1; 5X100 cells/mL
PBMC: 5X106 cells/well A549)
Dose/Concentration: PM10: 75//I
(10 /yg/mL); CM: 250 /y|; tBHP:
12.5/ym (in J774); TNF: 0,
500pg, 1 ng, 10ng ,
Time to Analysis: tBHP:1, 2, 4h;
PM: 4h; TNF: 18h
Cytokines: PM10 induced release TNF-a protein
from PBMCs at 10/yg/mL for 4h. Further inhibited
by verapamil and BAPTA-AM. Calmodulin inhibitor
W-7 had no effect. CM increased IL-8 from A549
cells 3 fold. Verapamil, BAPTA-AM and W-7
significantly inhibited IL-8 release induced by CM.
ICAM-1: A549 cells treated with TNF-a showed
dose-dependently effect of TNF-a on ICAM-1
upregulation at 18h. CM also induced
upregulation. Verapamil, BAPTA-AM and W-7 fully
inhibited CM-induced upregulation.
July 2009	D-36	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Calcabrini PM2.6 (Rome, Italy)
et al. (2004, 096865) Partic|e Size: PMis
Species: Human
Cell Type: A549
Route: Cell Culture (5 x 10"
cells/well)
Dose/Concentration: 30, 60
/yg/cm2 (aliquot of 0.1 /yglul)
Time to Analysis: 5, 24, 48, 72h
Particle Characterization: Components
measured include C-rich particles, Ca sulfates,
silica, silicates, Fe-rich particles, metals.
Carbonaceous particles made up majority of PM.
Cell Surface Changes: PM deposited on the cell
surface showed dose and time-dependent
increases in microvilli rearrangement and cell
shape alterations without affecting apoptotic
markers for up to 72 h.
PM internalization: At 24 h with the low dose,
aggregates of PM in cytoplasm or surrounded by
membrane was observed. With the high dose, large
particle aggregates often close to nuclear
envelopes were observed.
Cytoskeleton: At 72h PM induced dose-
dependent alterations from
rearrangementfinterweaving of microtubules to
bundling of microtubules with some
shortening/disruption.
Cell Growth: PM decreased cell growth in a dose
and time -dependently manner
ROS: PM increased R0S at the high dose for 5h
but not at 24 h or with the low low dose.
Cytokines: PM induced TNF-a peaked at 5h at
high dose and 48h at low dose, both ND at 72h.
PM induced IL-6 starting at 24 h thru 72h in time
and dose dependent manner.
Reference: Cao et al.
(2007,156322)
Species: Human
Cell Type: HAEC
NIST-DEP: collected using a diesel forklift and
hot bag filter system. (NIST, Minneapolis, MN)
C-DEP: obtained from a 30-kw (40 hp) four-
cylinder Deutz BF4M1008 diesel engine (U.S.
EPA)
Organic extract fraction of particles
NIST- DEP 2%
C-DEP 20%
Particle Size: NR
Route: Cell Culture (5 x 10E cells)
Dose/Concentration: NIST-DEP,
C-DEP: 0,12.5, 25,50,100, 200
/yg/mL
Time to Analysis: 1-4h
Cell Viability: DEP had no effect.
Stat3: Both DEPs induced time-dependent
phosphorylization of Stat3 in cytoplasm. NIST-
DEP induced phosphorylization dose-dependently
from 12.5 to 50/yglmL but stayed level at 100
and 200//gfmL. p-Stat3 induction was inhibited
by antioxidant BHA though it was reactivated
with exposure to H202. Reaction induced by
H202 was similar to that of DEP.
pStat3 Nuclear Transport: NIST-DEP induced
cytoplasmic pStat3 to move from cytoplasm into
nucleus.
pEGFR Dephosphorylation: After 4h of NIST-
DEP exposure, dephosphorylation was inhibited for
up to 90 min.
Reference: Chang C-
C et al. (2005,
097776)
Species: Human
Cell Type: A540,
THP-1
UfCB (Printex 90, Degussa)
Particle Size: 14nm diameter
Route: Cell Culture (7 x 10E cells)
Dose/Concentration: 100 //gfmL
Time to Analysis: 4h
ROS in THP-1 and A549: UFCB increased ROS.
NAC pretreatment blocked most of the UFCB-
induced ROS production.
VEGF in THP-1: UFCB increased VEG. NAC
decreased the UFCB effects below those of the
control.
VEGF in A549: Produced similar, but less marked,
results as with THP-1.
Reference: Chauhan
et al. (2004, 096682)
Species: Mouse
Strain: BALB/c
Cell Type: RAW
264.7 (leukemia virus
transformed
macrophages);
J774A.1 (from a
tumor); WR19M.1
(leukemia virus
transformed
macrophages)
EHC-T: total EHC-93 (Env Health Ctr, Ottawa,
Canada)
EHC-I: insoluble EHC
EHC-S: soluble EHC
SRM1648: urban particulate St.Louis (NIST)
SRM1649: urban dust/organics Washington
(NIST)
VERP: fine PM2.6 (Vermillion, Ohio)
Cristobalite: SRM 1879 (NIST)
TKk SRM 154b (NIST)
Particle Size: EHC-93: 0.5/jm (median
diameter); Cristobalite, SRM 1648, SRM
1649, Ti02,:NR; VERP: PM2.6
Route: Cell Culture (15000
cells/well)
Dose/Concentration: Particle
suspensions: 20, 50,100 /yg/well
LPS: 0-5 /yglmL
IFN-y: 0-1000 U|mL
Time to Analysis: Particles
added to culture at Oh, LPS and
IFN-y added at 2h. Parameters
measured after 22h incubation
period.
Stimulation with LPS/IFN-y: LPS and IFN-y each
induced NO release. Combination of LPS and IFN-y
produced larger effect in all cell lines. L-NMMA,
N0S inhibitor, suppressed most of the NO
production with 1OOnmol/L.
Cellular Viability and Cytotoxicity: Exposure of
cells to particulates did not result in overt
cytotoxicity or excessive loss of cellular material.
There was no correlation between the cytotoxicity
of the particles in the surviving cells and the loss
of protein mass in monolayers.
Nitrite Production: EHC-T, EH-93-I, SRM1648
and SRM 1649 produced dose-dependent
decreases. Cristobalite only decreased at higher
doses. No effect from EHC-S, VERP or Ti02.
iNOS: EHC-I, EHC-T, Crisobalite and SRM1648
inhibited iNOs expression. Ti02 had no effect. EHC
sol, SRM 1649 and VERP were not tested.
July 2009	D-37	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Chauhan
et al. (2005,155722)
Species: Human
Cell Type: A549
EHC-T: total EHC-93
EHC-I: insoluble EHC
EHC-S: soluble EHC
Cristobalite (Si02): SRM-1879
TiO?: SRM-154b
Particle Size: EHC-93: 0.4/ym (median
physical diameter): Tith, Si02: 0.3-0.6/jm
Route: Cell Culture (150000
cells/flask)
Dose/Concentration: All
particles: 0,1, 4, 8 mg/5ml
Time to Analysis: 24h
Cellular Viability: Decreased after exposure to
EHC-T, EHC-I and cristobalite. Rate of reduction
was not consistent across doses. EHC-S and Ti02
had no effect on viability.
ET-1: Release of ET-1 peptide decreased dose-
dependently for EHC-T, S and -I. Fractions of
EHC-S and EHC-I were more potent than EHC-T.
Ti02 and Cristobalite also reduced ET-1 secretion
although this was not consistent across the dose
range.
Cytokines: Results showed no detectable
amounts of GM-CSF, IL-1 (3 or TNF-a in cell
culture supernatants. IL-8 increased dose-
dependently with EHC-T, EHC-I and cristobalite.
VEGF: VEGF significantly increased dose-
dependently with EHC-T, EHC-S and cristobalite.
EHC-S induced a significant decrease in VEGF.
Gene Expression: mRNA levels for preproET-1
reduced at 24 h for all particle types. EHC-S
induced a significant decrease in ET-1 expression
at this high dose. ECE-1 mRNA expression
increased with EHC-T and EHC-I. Cither particles
had no effect. ETaR mRNA increased with EHC-T,
EHC-S, and Tith in biphasic manner where the
highest expression of mRNA was seen at the
middle dose levels. EHC-S had no effect. ETbR
mRNA increased with a low dose of EHC-T and
decreased with a high dose of EHC-T. EHC-S, EHC-
I and cristobalite induced an increase of ETbR.
Ti02 induced a significant decrease.
Proteases: mRNA levels for MMP-2 reacted
similarly to preproET-1. mRNA levels for TIMP-2
was significantly induced with EHC-I. EHC-T and
EHC-S induced small effects.
Reference: Cheng et
al. (2003,156337)
Species: Human
Cell Type: A549
DEP-h: DEP with high sulfur
DEP-LS: DEP with low sulfur
GEP: gasoline engine exhaust particles
Primed cells pretreated with TNF-a
Particle Size: DEP-h: 15.9nm; DEP-
LS: 17.7nm; GEP: 8.3nm
Route: In Vitro Cellular Exposure
(Exhaust flow-through cell culture
with air-cell-interface, exhaust
diluted 10-15x with 8x105
cellsfmL)
Dose/Concentration: DEP (total):
1.5-3.5x10e particles/cm3 air; GEP
(total): 1-2x100 particles/cm3 air;
TNF-y: 5ml (25 ng/ml)
Time to Analysis: 60-360 min
IL-8: DEP-h induced a 3 fold increase in IL-8 than
that of the control. DEP-LS also induced increases.
Primed cell cases had higher levels (1 Ox) than
unprimed when exposed to DEP-LS. DEP-h induced
higher levels of IL-8 than DEP-LS. This response
lasted for up to 6h. GEP induced a statistically
insignificant increase of IL-8 in unprimed cells.
With primed cells, GEP induced levels of IL-8 that
exceeded those of DEP-h and DEP-LS. This
response lasted for 1-2 h.
Reference: Chin et
al. 2002
Species: Rat, Human
Cell Line/Type: RAW
264.7 (Rat), MHS
Alveolar Macrophage
Cell Line (Rat), A549
(Human)
CB: (N339, with benzo[a]pyrene absorbed on
surface. Manufactured in Cabot, Boston, MA)
BaP
Benzo [a] pyrene 1, 6-quinone: BP-1.6-Q
(obtained from NCI, Kansas City, M0)
Particle Size: CB 0.1 /jm (mean diameter)
Route: Cell Culture
Dose/Concentration:
CB: 1, 2,4/yg/mL
BaP: 2/yg/mL
BP-1.6-Q: 1//M
Time to Analysis: 1-24h
HO-1 mRNA Expression: In RAW264.7, H0-1
mRNA levels increased with 2 and 4//gfmL at 2h.
Increases continued to 8h and declined by 24h.
BaP had no effect. BP-1.6-Q increased H0-1
mRNA after 1h and was maintained until 8h. In
A549 and MHS, H0-1 mRNA increased after 1 h,
peaking at 8h in A549 and 4h in MHS.
HO-1 Protein Expression: An increase of protein
was observed from 4-8h in RAW264.7.
AP-1: Increases in binding activity were observed
in RAW 264.7 cells at 2h.
July 2009	D-38	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Churg A
et al. (2005, 088281)
Species: Rat
Strain: Sprague-
Dawley
Weight: 250g
Cell Type: Epithelial
Cells of Tracheal
Explants
EHC93 (Ottawa Urban Air Particles)
TiFe - Iron-loaded fine Ti02 (obtained from
Aldrich Chemicals, Milwaukee, Wl)
3-4um EHC93
0.12 ± 1.4/ym Ti02
Particle Size: EHC-93: 3-4/ym (MMAD);
TiFe: 0.12 ± 1.4/ym (geometric mean
diameter)
Route: Cell Culture
Dose/Concentration: EHC-93,
TiFe: 500 /yglcm2
Time to Analysis: 1, 24h. Some
experiments (referred to as 2h)
explants transferred to different
dish and incubated for additional
hour. Pre-treated with
InhibitorsfChelators for 2h.
Activation of NF-kB: Both particle types
increased nuclear translocation of NF-kB. TiFe and
EHC-93 increased NF-kB 1.5 fold at 1 h. TiFe
increased NF-kB 3.5 fold at 2h. EHC-93 increased
NF-kB more than 2 fold. Ti02 by itself did not
increase NF-kB at any exposure duration.
Activation of NF-kB into tracheal epithelial
cells:No evidence of dust particles was observed
(EHC-93 or Ti02> in the epithelial cell cytoplasm at
2h. No evidence of morphologic cell damage from
particles was observed.
Colchicine: Treatment with colchicine did not
prevent NF-kB activation.
Inhibitors/Activators: Tetramethylthiourea
(TMTU) (membrane-permeable active oxygen
scavenger), Deferoxamine (redox-inactive metal
chelator), PPS (Src inhibitor) AG1478 (epidermal
growth factor receptor inhibitor) prevented NF-kB
activation in both EHC93 and TiFe exposed-cells.
Iron-containing citrate extract of both dusts
increased NF-kB activation in both EHC93 and
TiFe exposed-cells.
Reference: Courtois
et al. (2008,156369)
Species: Rat
Strain: Wistar
Cell Line: Dissected
intrapulmonary
arteries from rats
used in corresponding
in vivo experiments
PM (SRM 1648)
(63% inorganic carbon, 4-7% organic carbon,
mass fraction > 1 %: Si, S, Al, Fe, K, Na)
UF carbon black (FW2, P60)
Particle Size: SRM 1648 mean diameter 0.4
/ym; ultrafine carbon black: FW2- 13nm, P60-
21nm
Route: Cell Culture
Dose/Concentration: 100, 200
/yg/mL
Time to Analysis: 24h incubation
NO: Generally, Ach-induced relaxation in
intrapulmonary arteries decreased, Ach-induced
cGMP accumulation decreased, and relaxation by
SNP or DEA-N0 also decreased. UF carbon black
did not affect NO responsiveness.
Oxidative Stress, Inflammatory:
Dexamethasone prevented SRM 1648-induced
impairment of the Ach relaxation response but
antioxidants did not. TNF-a, MIP2, IL-8 increased.
R0S was not affected.
Reference: Dagher et
al., (2007, 097566)
Species: Human
Cell Type: L132
(Normal Lung
Epithelial Cells)
LC10, LC50 - PM2.6
(collected Jan-Sept in Dunkerque, France)
Particle Size: cumulative frequency: 0.5/ym:
34%; 1 /ym: 64%; 1.5/ym: 79%; 2/ym: 87%;
2.5/ym: 92%; 5/ym: 98%; 10/ym: 100%
Route: Cell Culture (3 x 10e, 1.5 x
106, 0.75 x 106 cells/20mL)
Dose/Concentration: LC10:19
/yglmL; LC50: 75/yglmL
Time to Analysis: 24, 48 or 72h
p65 Protein: Phosphorylation of p65 increased in
PM-exposed L132 cells in dose-dependent manner.
IkBa Protein: Phosphorylated IkBa protein
concentrations increased in cytoplasm with both
particle types at all time points.
p65 and p50 DNA: p65 DNA binding increased at
24h with LC10 and LC50, at 48h with LC10, and
at 72h with LC10 and LC50. p50 DNA binding
increased at all time points with LC10 and LC50.
Reference: Dai et al
(2003, 087944)
Species: Rat
Strain: Sprague-
Dawley
Weight: 250g
Cell Type: Tracheal
Explants
gene expression but this increase was completely
prevented by SN50 and MAP kinase inhibitors
(SB203580 and PD98059). Neither TMTU or DFX
has any effect.
TGFpi : Treatment of explant with EHC93
approximately doubled gene expression for
TGFpi.Treatment with SN50, TMTU and fetuin
(TGFp antagonist) blocked increase. DFX, MAP
kinase inhibitors (SB203580 and PD98059) had
no effect. DEP roughly doubled TGFpi
expression.SN50 and MAP kinase inhibitors
(SB203580 and PD98059) fully blocked this
effect. TMTU and DFX had no effect.
EHC-93 (Environmental Health Center,
Ottawa)
DEP: SRM 1650a (NIST)
Particle Size: EHC-93: 3-4/ym (MMAD): DEP
1.55 ± 0.04/ym (CMD)
Route: Cell Culture
Dose/Concentration: ECH, DEP:
500 /yglcm2
Time to Analysis: Exposed for
1h. Parameters measured
following a 7d incubation period.
Hydroxyproline: EHC93 induced an almost 3 fold
increase in explant hydroxyproline. DEP increased
tissue hydroxyproline 2.5 fold.
Procollagen: EHC-93 doubled gene expression of
procollagen. Procollagen gene expressioncould be
fully inhibited by SN50, TMTU or treatment of the
PM with DFX. Treatment of explants with p38 or
ERK (inhibitors) had no effect on procollagen
expression. DEP induced an increase in procollagen
July 2009
D-39
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Doherty
SP et al. (2007,
096532)
Species: Rat
Strain: NR8383
Cell Types: AMs
Ratios of: V: Fe; Al: Fe; Mn: Fe
V - sodium vanadate (NaV03)
Al - aluminum chloride hexahydrate (AlCh)
Mn - manganese chloride tetrahydrate
(M11CI2)
Fe - ferric chloride hexahydrate (FeCh)
Ratios based on PM2.6 measurements from
NYC, LA and Seattle
Particle Size: Metals from PM2.6 samples
Route: Cell Culture (2X10E
cells.i'niL)
Dose/Concentration: Fe - 16
/ymol (equivalent to urban NYC
500 /yg PM2.6); V and Mn tested in
molar rations of 0.02 to 0.4
relative to Fe; Al tested in molar
ratios of 0.125 to 8 relative to Fe.
Time to Analysis: 20h
IRP: Addition of V increased IRP activity 5 to 9
fold. Though there was no seeming dose
responsivity, IRP activity remained strongly
elevated over the range of V:Fe ratios tested.
Addition of Mn only resulted in an effect at 0.1
molar ratio (two-fold), not at higher or lower
ratios. Al resulted in peak increases of 5 fold at
molar ratios 2 while declining to 2 fold at molar
ratios 4 and 8.
Cytoxicity: Al was cytotoxic at molar ratios of 4
and 8. All other Al, V, Mn ratios had no effect.
Mixtures: The combination of metals tested at
NYC PM ratios and V drove all the Fe transport
activity. Combinations of V+Mn and V+AI
increased activity more than V:Fe alone.
Reference:
Doornaert, et al.
(2003,156410)
Species: Human
Cell Line/Type:
16HBE14o-; P-HBE
DEP: SRM 1650 (NIST)
CB: (Sigma, France)
DPC: Dipalmitoyl phosphatidylcholine (positive
control)
0.5um
Particle Size: NR
Route: Cell Culture (106-2x106)
Dose/Concentration: DEP and
CB: 1- 100/yg/mL
Time to Analysis: Parameters
measured 24, 48, 72h post
exposure. 1-HBE Cell Deadhesion
Capacity: 24h, evaluation of
detachment performed every 5min
for 40min after. Cell Wound Repair
Capacity: 24 h, repair evaluated
3.5, 7, 24 h after.
Cytotoxicity: DEP was cytotoxic at 100/yg/mL
at all time points in a time-dependent manner. CB
and DPC cytotoxicity was substantially lower but
significant at 72h.
Phagocytosis: 1-HBE cell levels that were in
contact with DEP or CB or have phagocytized
those particles increased in a dose-dependent
manner. DEP induced greater levels of cell contact
and phagocytosis than CB.
F-actin: Only DEPs were engulfed by F-actin
stained cell fragments.
Actin CSK Stiffness: DEP (5, 20,100/yglmL)
induced net dose-dependent decrease in
cytoskeleton stiffness and a dose-dependent
decrease in actin cytoskeleton stiffness. CB
produced no significant decrease.
Adhesion Molecules: DEP induced a concomitant
reduction of both CD49 (a3) and CD29 ((31)
integrin subunits and a decrease in level of CD44
(HBE cell-cell and cell-matrix adhesion molecule) at
both 20 and 100/yg/mL.
Proteases: DEP also induced an isolated decrease
in MMP-1 expression without change in tissue
inhibitor of TIMP-1 or TIMP-2 at 100 /yglmL. CB
produced no change or insignificant results.
1-HBE Cell Deadhesion Capacity: DEP exposure
induced a dose-dependent amplification of cell
detachment at 5min of incubation and onward.
Cell Wound Repair Capacity: DEP inhibited
wound repair/wound closure in a dose-dependent
manner.
Reference: Dostert Asbestos
et al. (2008,155753)
Species: Human
Cell Line/Type:
THP1, monocyte-
derived macrophages
Silica
DEP
CSE: cigarette smoke extract
MSU: monosodium urate crystals
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Asbestos:
0.1, 0.2 mg/mL; Silica: 0.1, 0.2,
0.25, 0.5 mglmL; DEP: 0.2, 0.25,
0.5 mglmL; CSE: 5%, 10% in
solution mg/mL; MSU: 0.1, 0.2
mg/mL
Time to Analysis: 1, 3, 6h
IL-10: Increased levels of IL-1p with asbestos
and silica were observed in THP1 at 6h. CSE and
DEP had no effect. MM also had increased levels
with asbestos, silica and MSU at high dose levels
only.
Caspase-1: Asbestos increased caspase-1
activity.
ROS: Asbestos doses in THP1 exhibited an
increase in ROS formation.
Reference: Doyle, et
al. (2004, 088404)
Species: Human
Cell Type: A549
from non-smoking
adults
BD: 1,3-butadiene, known carcinogen
Acrolein: photochemical and NO product of BD
in atmosphere
Acetaldehyde: photochemical and NO product
of BD in atmosphere
Formaldehyde: photochemical and NO product
of BD and ISO in atmosphere
ISO: isoprene, 2-methyl analog of BD
Methacrolein: photochemical and NO product
of ISO in atmosphere
Methyl vinyl ketone: photochemical and NO
product of ISO in atmosphere
Particle Size: NR
Route: Environmental Irradiation
(smog) Chambers
Dose/Concentration: 50 ppb NO;
200 ppbVISO, BD
Time to Analysis: Exposed to
gases for 5h. Analysis 9h post
exposure.
Cytotoxicity: ISO + NO and BD+NO induced small
increases of LDH in A549. However,
ISO+NO+light and BD+NO+light increased LDH
levels 4-6 fold indicating photochemical products
of ISO and BD are highly cytotoxic. LDH levels of
each combination were equivocal.
IL-8 Protein: Methacrolein, methyl vinyl ketone
and formaldehyde (products of ISO) increased IL-8
protein levels significantly. ISO+NO had no effect.
BD photochemical products (acrolein,
acetaldehyde and formaldehyde) also increased IL-
8 protein, more than doubling the photochemical
products induced by ISO. BD+NO had no effect.
IL-8 mRNA: IL-8 mRNA expression also increased
with photochemical products of ISO and BD but
did not reach a statistically significant level.
July 2009
D-40
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Duvall,
R.M. (2008, 097969)
Norris, G.A.
Dailey, L.A.
2008
Species: Human
Cell Type: Airway
Epithelial Cells
PM-F, -C, -UF
Particles collected from: Seattle, WA (PM-S);
Salt Lake City, LIT (PM-SL); Phoenix, AZ (PM-
P); South Bronx, NY (PM-SB); Hunter College,
NY (PM-HRterling Forest, NY (PM-SF)
Particle Size: Coarse: > 2.5/jm; Fine:
< 2.5 fjm; Ultrafine: < 0.1 fjm
Route: Cell Culture (100,000
cells/cm2)
Dose/Concentration: 5 mgfml
Time to Analysis: 1, 24h post
exposure
Particle Characterization: PM-HR, PM-SL and
PM-S contained the highest UF, F, and C
concentrations. PM-SB and PM-HR had similar F
and C concentrations. Sulfate was highest in PM-F
for all sites except in PM-SB and PM-HR. Wood
combustion was highest in PM-SL, PM-S, PM-P.
Soil dust was highest in PM-SL and PM-S.
IL-8: PM-UF induced a greater increase in IL-8
than other types of PM except PM-P. PM-UF is
associated with vanadium, lead, copper, sulfate.
PM-F-HR caused the greatest increase followed by
PM-SB. PM-F-SF and PM-F-P was least effective.
PM C also caused an increase in IL-8 levels and
was associated with vanadium and elemental
carbon.
COX-2: PM-F-S induced the greatest increase in
COX-2 expression. Other PM-F sites induced
similar increases. UF PM had no effect. PM-C,
associated with elemental carbon, induced
increases.
HO-1: PM-F-SF induced the greatest increase in
H0-1. PM-F-SL was the least effective. UF PM
had no effect. PM-C, associated with copper,
barium and elemental carbon, caused an increase.
Reference: Dybdahl
M et al. (2004,
089013)
Species: Human
Cell Type: A549
DEP: SRM 1650 (NIST)
Particle Size: 90nm (MMAD)
Route: Cell Culture (10E cells/mL)
Dose/Concentration: 0,10, 50,
100, 500 /yglmL
Time to Analysis: 2, 5, or 24h
Cytokines: DEP induced dose-dependent
increases of IL-1a, IL-6, IL-8 and TNF-a at 24 h.
Cytokines increased between 4 and 18 fold at the
highest DEP dose as compared to controlled cells.
DEP also increased IL-6 mRNA expression levels in
a dose and time-dependent manner.lL-6 mRNA
levels increased 14 fold at 24 hours, 8 fold at 5
hours, and 2 fold at 2 hours.
Cell Viability: DEP exposure did not decrease cell
viability at any dose tested.
Reference: Dybdahl
M et al. (2004,
089013)
Species: Mouse
Strain: BALB/CJ
Gender: Female
Age: 10wks
Weight: ~20g
DEP: SRM 1650 (NIST)
Particle Size: 90nm (MMAD)
Route: Nose-only Inhalation
Dose/Concentration: Single
exposure: 20 or 80 mg/m3; 4
exposures: 5 or 20 mg/ m3
Time to Analysis: Single 90min
or four 90min exposures.
Inflammation: DEP induced a dose- and time-
dependent expression of pro-inflammatory
cytokines. A single high dose of DEP induced an
increase in IL-6 mRNA levels persisting for at least
22 hours. However, at the same cumulative dose
given as 4 doses, the IL-6 level was only increased
1 hour after exposure.
Cytokines: Only IL-6 was induced after DEP
exposure. A single 90 minute inhalation (at 20
mgfm3) increase IL-6 gene levels dose-dependently
in the lung and was significantly higher than
control levels at both 1 and 22 hours post
exposure. Repeated exposures at 5 mgfm3 did not
affect IL-6 levels.
BAL Cells: Inhalation of DEP did not decrease
viability of BAL cells. 1 hour after the last dose,
mice exposed to 20 mgfm3 exhibited a 3 fold
increase in total cell numbers as compared to
control. The number of total cells and neutrophils
in BAL fluid were still increased 22 hours after
exposure to 20 mgfm3 DEP. BAL cell population
was not affected by repeated exposures to 5
mg/m3DEP. Lymphocyte numbers were also
unchanged by DEP.
July 2009
D-41
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Fritsch,
S.Diabate, S. (2006,
156452)
Species: Mouse
Cell Type: RAW
264.7
MAF02: incinerator fly ash (collected by
electrostatic precipitation in commercial
municipal waste incinerator facility)
composition representing 12% of total mass
(mg/g):
Fe (9.1); Pb (23.3); Zn (75.7); C (7.5)
Particle Size: 165nm (modal value)
Route: Cell Culture (1 x 10e
cells/well)
Dose/Concentration: 6.3-188
/yg/cm2 for Toxicity; 2.6, 6.5,
13.2/yglcm2 for Arachidonic Acid;
13.2/yglcm2 for MAPK Pathway;
Other doses noted in Effect of
Particles
Time to Analysis: 1, 2.5, 5, 24 h
Toxicity: Viability decreased from 99% to 18% at
62.5-188/yglcm2. Lower doses had no effect.
Arachidonic Acid: At 2.5h, AA level increased 2
fold for 6.5 /yglcm2 and 6 fold for 13.2/yglcm2.
No increase was observed after 5h.
MAPKs: Cells pretreated with PD98059, an
inhibitor of MEK-1, inhibited AA liberation due to
MAF02 treatment of 13.2 /yglcm2
COX-2: A time-dependent increase of COX-2
protein expression was exhibited at 2.5 and 5h.
ROS: A dose-dependent increase in ROS formation
was observed at concentrations greater than 31.3
/yg/cm2 after 3 hours.
GSH: There was an observed increase of
production at 20h. Doses greater than 60 //gfcm2
reduced total glutathione.
HO-1: There was an observed dose-dependent
increase in expression at 4 hours.
Reference: Fujii et al.
(2002, 036478)
Species: Human
Cell Type: HBEC
(from current
smokers), AMs, Co-
Culture: AMs + HBEC
Age: HBEC: 48-70yrs
PMio: EHC-93 (Ottawa, Canada)
Particle Size: PMio
Route: Cell Culture (HBEC: 2.5-3 x
10e cells/well); (AMs: 1.0 x107
total)
Dose/Concentration: 100, 500
/yg/mL
Time to Analysis: 2, 8, 24h
Viability: Over 90% of HBEC were viable after a
24 hour exposure of up to 500 /yglmL of PM. AMs
incubated with and without 100/yg|mL saw no
significant difference in viability.
Cytokine mRNA: TNF-a, GM-CSF, IL-1(3, IL-6,
LIF, 0SM and IL-8 mRNA expression increased in
co-culture with 100/yglmL at 2 and 8h. In AMs,
TNF-a, IL-1 (3, IL-6 mRNA expression increased
with 100/yglmL at 2h. IN HBECs, IL-1 p and LIF
increased with 100//gfmL at 2h. HBECs added to
AMs exposed to PMio, further increase in mRNA
of IL-1 (3, LIF and IL-8.
Cytokine Protein: In co-culture and AMs,
significant increase in protein production of GM-
CSF, IL-8, IL-1(3, IL-6 and TNF-a in dose-
dependent manner. GM-CSF and IL-6 production
significantly higher in co-culture then AM or HBEC
alone.
Bone Marrow: Co-culture instillation of
supernatants increased circulating band cell
counts at 6 and 24 h with 100 /yglmL.
Reference: Fujii, T.
Hayashi, S.
Hogg J.C. (2001,
156455)
2001
Species: Human
Cell Type: HBEC
from current smokers
Age: 48-70yrs
PMio:EHC93 (Ottawa, Canada)99% < 3.0um
Particle Size: PMio( 99% < 3.0/ym)
Route: Cell Culture (2.5-3 x 10e
cells/dish)
Dose/Concentration: 10,100,
500 /yglmL
Time to Analysis: 2, 8, 24h
Phagocytosis: 18.6% of cells engulfed particles
when exposed to 100 /yglmL. Over 90% remained
viable.
Cytokine mRNA: LIF mRNA increased dose-
dependently at 2h but declined at 8 and 24 h. GM-
CSF increased dose-dependently at 8h and peaked
at 24 h. IL-1 a increased at 2h, increased dose-
dependently at 8 h and peaked at 24 h. M-CSF,
MCP-1, IL-8 were unaffected.
Cytokine Protein: LIF, GM-CSF, IL-1 p and IL-8
increased dose-dependently. Soluble fraction of
100 /yglmL PMio did not affect cytokine
production.
July 2009
D-42
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Garcon,
G. Dagher, Z.
Zerimech, F. (2006,
096633)
Species: Human
Cell Type: L132
(Embryonic Lung
Epithelium)
PM2.6 (collected in Dunkerque, France for Elmo,
Jan-Sept)
Particle Size: PM2 6 0-0.5um (33.63%), 0.5-
1.0um (30.61%), 1.0-1.5um (14.33%), 1.5-
2.0um (8.69%), 2.0-2.5um (4.89%), > 2.5um
(7.87 %)
Route: Cell Culture: 3 x 10e
cells/20ml (24 h); 1.5 x 106
cells/20ml (48h); 0.75 x 100
cells/20ml (72h)
Dose/Concentration: 18.84,
37.68, 56.52,75.36,150.72
//g/mL; LC10- 18.84//g/mL; LC50-
75.36 //g/mL
Time to Analysis: 24, 48 or 72h
Cytotoxicity: PM induced dose-dependent
(R2 -.9907) cytotoxic effect in proliferating L132
cells.
LDH: Increase at 72 h with 56.52 and 75.36
//g/mL.
Oxidative Stress: A decrease in MDF activity
was observed at all exposure levels at 24, 48, and
72 h (72-h < 5 % of control). MDA levels showed
increase concentration after 72 h.both LC10 and
LC50. LC10 and LC50 saw an increase in SOD
activity at 24 h; LC50 saw a decrease in activity
after 48 and 72 h. 8-OHdG and PARP exhibited
increases at all time points with LC10 and LC50.
Inflammatory Response: Increases of TNF-a
concentration was exhibited at 24 h at LC50, and
at 48 h and 72 h at LC10 and LC50. ilMOS activity
increase at all time points at LC10 and LC50.N0
concentration exhibited increases at all time points
after exposure to LC10 and LC50.
Reference:
Geng H et al. (2005,
096689)
Species: Rat
Strain: Wistar
Tissue/Cell Type:
Lung macrophages
BPM: Blowing PM2.6
NPM: Non Blowing Normal PM2.B
Particle Size: PM2.6 PM collected from
Wuwei City, Gansu Province, China (Blowing
days correspond to desert storm days)
Route: Cell Culture
Dose/Concentration: 0, 33,100,
300 //g/mL
Time to Analysis: 4h
NOTE: Unless otherwise noted, results are
identical for BPM and NPM.
Cytotoxicity: Dosages greater than 150 //g/mL
decreased cell viability.
Plasma Membrane Fluidity: Dose-dependent
decrease had no effect on membrane lipid
hydrophilic region.
Plasma Membrane Permeability: LDH enzyme
activity and extracellular AP activity increased
dose-dependently, indicating increased membrane
permeability, but this was only statistically
significant at 300 //g/mL dose. NPM may affect
some parameters at 100//g/mL. Overall, NPM
induced a slightly higher increase than BPM.
Intracellular Ca2+: A dose-dependent increase
was observed.
Lipid Peroxidation (TBA): An increase was
observed only at 300 //g/mL.
Antioxidant (GSH): A decrease was observed
only at 300//g/mL.
Reference:
Geng H et al. (2006,
097026)
Species: Rat
Strain: Wistar
Tissue/Cell Type:
Lung macrophages
DPM: dust storm samples
NPM: normal PM
Particle Size: PM2.6 PM collected from
Baotou City, Inner Mongolia, China in March
2004
Route: Cell Culture
Dose/Concentration: 0, 33,100,
300 //g/mL
Time to Analysis: 4h
NOTE: Unless otherwise noted, results are
identical for BPM and NPM
Cytotoxicity: MTT reduction assay revealed a
significant decrease in cell viability at 150 //g/mL
and 300 //g/mL. LDH enzyme activity significantly
increased at 150 and 300 //g/mL.
GSH levels: Significant decreases were seen in
cellular GSH levels and increases in TBARS levels
in both groups with a 300 //g/mL dose.
Plasma Membrane Activity: In the plasma
membrane, Na.KATPase were significantly
inhibited. Ca2+Mg2+ -ATPase were unaffected.
Plasma Membrane Lipid Fluidity: Results
indicate that DPM could increase the surface
fluidity of membrane lipid.
Intracellular Ca2+: A dose-dependent increase in
free intracellular Ca2+ levels was observed.
Reference: Ghio et
al. (2005, 088272)
Species: Human
Cell/Tissue Type:
BEAS-2B
FAC: ferric ammonium citrate
(component of R0FA)
VOSO4: vanadyl sulfate
(component of R0FA)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 100//M
FAC - preexposed before metal
compounds or oil fly ash
50//M V0S04- preexposed before
metal compounds or oil fly ash
100//g/mL R0FA
Time to Analysis: 0-1 h, 4h
IRE DMT1: FAC increased mRNA and protein
expression for -IRE DMT 1. VOSO4 decreased
mRNA and protein expression for -IRE DMT 1.
+IRE DMT1 unaffected by any treatment.
Metal transport: Uptake of iron increased after
pre exposure to FAC and decreased after pre-
exposure to VOSO4. Pre exposure to FAC again
increase the uptake of both iron and vandium.
VOSOi induced opposite effect, decreasing Fe
uptake.
ROS: Increased acetaldehyde, indicating increased
oxidative stress. ROS decreased with FAC
pretreatment. ROS increased with VOSO4
pretreatment.
July 2009
D-43
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Gilmour
et al. (2004, 057420)
Species: Rat
Strain: Sprague-
Dawley
Cell/Tissue Type:
Alveolar macrophage
Coal Fly Ash
Mil - Montana Ultrafine
MF - Montana Fine
MC - Montana Coarse
KF - West Kentucky Fine
KC - West Kentucky Coarse
Coal combustion using a laboratory-scale
down-fired furnace rated at 50kW. Montana
subbituminous coal and western Kentucky
bituminous coal
Particle Size: Coarse: >2.5//m; Fine:
< 2.5/ym; Ultrafine: < 0.2/ym
Route: Cell Culture (2X10E
cells.i'niL)
Dose/Concentration: 125 //gfmL
or 250 /yglmL
Time to Analysis: 4 or 24h
LDH: Mid and high doses of Montana ultrafine
particles showed significant increase after 4 h
exposure vs control. Other particle types had no
effect. After 24 h, LDH level was not statistically
significant between particles tested and control.
Cytokines: Treatment with Montana ultrafine
particles resulted in a significant production
increase of TNF-a. MIP-2 showed increases in all
the fine and ultrafine treatments, with Montana
ultrafine and W. Kentucky fine PM showing the
highest increases. IL-6 increased with Montana
ultrafine particles although there was some
variability and the increases were not statistically
significant.
Reference: Gilmour PMio: Collected from the Marylebone and
et al. (2005, 087410) Bloomsbury monitoring sites in London, UK
Species: Human
Cell/Tissue Types:
monocyte derived
macrophages,
HUVECs, A549,
16HBE
Particle Size: PMio
Route: Cell Culture
Dose/Concentration: 50//gfmL
Time to Analysis: 4h, 6h, 20h
IL-8: PMio at 50//gfmL induced a significant
increase in IL-8mRNA and protein expression in
PMM and 16HBE at 6 and 20h. A less substantial
increase was also observed in A549.
Procoagulant Activity: PMio induced a
significant decrease in macrophage mediated
clotting time in 16HBE. Other cell types were
unaffected.
Annexin V Binding: At 100//gfmL, PMio induced
a significant increase in binding macrophages at 4
and 20h. There was no effect at 50 /yglmL.
Tissue Factor mRNA Expression: Expression
was increased in macrophages at 6 h only.
tPA Expression: mRNA expression decreased at
6h. Protein expression decreased at 4 h and 20 h
in a dose-dependent manner.
TF Expression: TF mRNA expression increased in
a dose-dependent manner at 6 h in HUVECs.
Protein levels also increased at 4 h but declined to
basal levels by 20h.
Reference: Gilmour
et al. (2003, 096959)
Species: Human
Cell/Tissue Type:
A549
PMio: Collected from the Marylebone and
Bloomsbury monitoring sites in London, UK
TSA
H202
NAC
Mannitol
Provided by Sigma Chemical, Poole, UK or
GIBCO-BRL, Paisley, UK
Particle Size: PMio
Route: Cell Culture
Dose/Concentration: PMio: 100
/yg/mL; TSA: 100 nglmL; H202:
200//M; NAC and Mannitol: 5 mM
Time to Analysis: 24h
IL-8: PMio, TSA and H202 treatment induced an
increase of IL-8. Concomitant exposure of TSA
with PMio or H202 significantly increased IL-8
release when compared to PMio or H202 alone. IL-
8 mRNA expression with PMio or H202 exposure
and TSA coincubation caused significant
increases. Silver staining of PCR products
indicated that the IL-8 gene promoter was
associated with acetylated H4 following TSA,
PMio and TNF treatment.
H4: PMio exposure significantly increased
acetylation levels of H4 over controls. Increased
acetylated H4 was mediated by PMio in a dose-
dependent manner. Treatment with PMio and
H202 increased HAT activity associated with H4
by 245% and 166% respectively. Significant
increases in acetylation of H4 following treatment
of cells with TSA, PMio and H202 for 24 h was
observed. PMioinduced HAT activity was
significantly decreased in the presence of NAC
and mannitol. Nuclear presence of HDAC2 protein
was significantly reduced by exposure to both
HDAC inhibitor and PMio. There was a decreasing
trend in HDAC2 gene expression following TSA
and PMio treatment.
NF-I
-------
Study
Pollutant
Exposure
Effects
Reference: Graff DW
et al. (2007,156488)
Species: Human
Cell/Tissue Type:
HAEC
PM
-UF: ultrafine
¦F: fine
¦C: coarse
Particles collected from Seattle, WA (-S), Salt
Lake City, LIT (-SL), Phoenix, AZ (-P), South
Bronx, NY (-SB), Hunter College, NY (¦
H), Sterling Forest, NY (-SF)
Particle Size: UF: <0.1 /jm; F: 0.1 ¦ 2.5/jm;
C: 2.5-10/ym
Route: Cell Culture
Dose/Concentration: 250//gfmL
Time to Analysis: 6h, 24h
Gene Expression: PM-UF, PM-F, and PM C both
upregulated and downregulated genes in the
HAECs though downregulation was far more
common for all the three PM fractions. PM-F
affected the greatest number of transcripts,
followed by the UF and C fractions.
IL-8: mRNA expression increased, with PM-F-S
having the greatest impact. Aluminum, strontium,
manganese and potassium were highly associated
with expression. Wood combustion was
moderately associated.
HOX-1: mRNA expression increased, with PM-F-
SF having the greatest impact. Potassium,
manganese, strontium and wood combustion were
highly associated with expression. Aluminum and
vanadium were moderately associated.
Reference: Gualtieri
M et al. (2005,
097841)
Species: Human
Cell Type: A549
TD: Tire debris extracted in methanol,
constituent of PMio
(generated by spinning a new automotive tire
against abrasive surface)
Particle Size: 10-80/ym
Route: Cell Culture
Dose/Concentration: 10, 50, 60,
75/yg/mL
Time to Analysis: 24,48, 72h
Cytotoxic Effect: Treated cells presented
inhibitory effect on reduction of MTT which
appeared to be dose and time-dependent. A
statistically significant reduction was observed at
48 and 72 h. Trypan blue showed a significant PM
lethality as well as a dose-dependent increase in
mortality.
DNA Damage: At 24 and 72 h, DNA damage
increase dose dependently in damaged and ghost
cells.
Cell Cycle Analysis: At 24 h, TD extract-treated
cells presented a significant increase in the
percentage of cells in G1 phase when compared
with control. This increase was associated with a
decrease in the percentage of cells in S phase. At
48 and 72 h, the increase in percentage of cells in
G1 was associated with a decrease in the
percentage of cells in both S and G2/M phases.
Cells exposed to TD extracts presented changed
morphology. Modifications most obvious at 72 h.
The highest dost produced increased vacuolization
in cytoplasm and apoptotic nuclear images.
Reference: Hetland PMC - Coarse
et al. (2005, 087887) p|yp = pjne
Species: Rat
Gender: Male
Strain: CrlfWky
Cell Type: AMs
¦A - Amsterdam
¦L - Lodz
¦R - Rome
¦0 - Oslo
Coexposures PAH, Fe, Al, Zn, Cu, V
Particle Size: PMC: 2.5-10 fjm; PMF: 0.2-
2.5/ym
Route: Cell Culture (1.5x100
cells/well)
Dose/Concentration: 50,100
/yg/mL PM
Time to Analysis: 20h
IL-6: PMC from all cities exhibited increases in IL-
6 release with spring and summer roughly equal
and both inducing higher levels than the winter
PMC. For the Spring and Summer samples, PMC-L
exhibited the highest IL-6 releases (440% and
460% respectively) followed by Rome,
A'dam/Oslo, and Oslo/A'dam. For the winter
samples, Rome and Amsterdam induced higher IL-6
levels (340% and 300% respectively) than Lodz
and Oslo (165% and 160%). The fine fractions did
not induce any significant cytokine release.
TNF-a: PMC from all cities increased TNF-a
release with 50//g/mL generally inducing a slightly
higher increase than 100 //gfmL.
Constituent Correlation: Levels of Fe, Al, Zn, Cu
and V as well as PAH (total and fractions) showed
no correlation with IL-6 release.
Endotoxin Correlation with IL-6 release: A
confirmatory test revealed no correlation.
July 2009
D-45
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Hetland
RB et al. (2004,
097535)
Species: Rat, Human
Cell Type: Alveolar
Macrophages (Rat),
A549 (Human)
Strain: WkylNHsd
(Rat)
Gender: Male (Rat)
Weight: 180-230g
(Rat)
AMC - Ambient Coarse
AMF - Ambient Fine
AMUF - Ambient Ultrafine
(AM samples taken at a suburban site,
without a dominating PM source, near
Utrecht, Netherlands)
Road PM: PMio, (collected in a road tunnel
with predominating road abrasion due to use
of studded tires in Trondheim, Norway)
Particle Size: AMC: 2.5-10/jm; AMUF:
< 0.1 fjm
Route: Cell Culture (1 x 10e
cells/well)
Dose/Concentration: 0,100,
200, 400, 600, 800,1000 //g/mL
Time to Analysis: 20h (Type 2
cells): 40h (A549 cells)
IL-8: All 3 AM fractions showed dose-dependent
increases in A549 cells until 600//gfmL; at that
concentration, levels declined. AMC showed the
most pronounced decline which correlates with
decreased viability. Road PM showed a near linear
response until 1000/yglmL, whereas DEP
plateaued at 600 //gfmL in A549.
MIP-2: AMC and AMUF had no effect on Type 2
cells. DEP induced increases at 200//gfmL,
whereas Road PM induced the strongest increase,
peaking at 600 //gfmL in Type 2 cells.
IL-6: AMC induced increases at 100/yg|mL in Type
2, but levels declined below normal at 200 //gfmL.
AMUF induced a decline of IL-6 levels. Road PM
induced significant increases in Type 2. DEP had a
slight effect. AM fractions induced increases in
A549 cells, peaking at 600/yglmL with AMF. DEP
and Road PM induced a dose-dependent increase.
Cell Survival: AMC showed major effects at 200
//g/ml_ in Type 2. AMUF showed effects at 400
/yglmL. Road PM and DEP showed a gradual
decline from 75% to 50% at 800 /yglmL in Type
2. All AM fractions induced a decrease in viability
after 600/yg/mL in A549 with AMC inducing a
larger decrease than AMUF and AMF; AMUF and
AMF induced similar levels. Road PM and DEP had
no effect on A549.
Apoptosis: AMC elicited a marked induction of
apoptosis 200 /yglmL in Type 2 cells. AMF
showed a dose-dependent increase in A549. Other
AM fractions showed some slight increases in
both cell types. Statistical significance was
reached for all particles except for Road PM.
Reference: Holder AL DEP: generated from a single cylinder diesel
et al. (2008, 093322) engine using , commercial certified #2 diesel
fuel
Species: Human
Cell Type:
16HBE14o
Copollutants: NOx 7 ppm, CO2 0.1%
Particle Size: Suspension: 223nm (mean
diameter); ALL 122nm (mean diameter)
Route: Suspension (1x106
cells/cm2). Air Liquid Interface
(ALI, 1x10B cells/cm2)
Dose/Concentration: Suspension:
0.13, 0.24,1.88, 2.5, and 12.5
/yg/cm2;ALI: 1 g/cm3 (total number
of particles: 2.3 x 107
particles/cm2)
Time to Analysis: Exposure for
6h. Parameters measured 20h
postexposure.
ALI vs Tracheal Bronchial (TB) Deposition: The
TB region deposition is 1.5 nominally x ALI, but
particle diameter deposited in the TB was 62 nm
(geometric mean diameter) as compared to the
particle deposition in the ALI, measuring 260 nm.
Inflammatory Response: Suspended DEP
decreased viability at concentrations of 2.5
/yg/cm2or higher. IL-8 release (corrected for
viability) increased at concentrations of 1.88
/yglcm or higher in a dose-dependent manner. IL-8
exhibited intermediate levels of secretion between
in vitro levels of 0.25 and 1.88 /yglcm2. No
statistically significant results were observed in
ALI. Viability for ALI was near 100% (75%
uncorrected).
Reference: Huang et
al. (2003, 087376)
Species: Human,
Mouse
Cell Type: BEAS-2B
(human), RAW 264.7
(mouse)
PMC: PM coarse
PMF: PM fine
PMSM: PM submicron
Collected between September-December 2000
from 4 ambient monitoring stations in Taiwan
that represented background, urban, traffic,
and industrial sites
Particle Size: PMC: 2.5-10 fjm; PMF: 1 -2.5
/jm; PMSM: < 1 /jm
Route: Cell Culture (5X10E
cells.i'niL)
Dose/Concentration: All PM: 50,
70,100/yg|mL
Time to Analysis: BEAS-2B: 8h;
RAW 264.7: 16h
Viability: None of the PM fractions affected cell
viability.
IL-8: Only PMSM induced a significant IL-8
increase in BEAS-2B. IL-8 response was
associated with a combination of Mn and Cr
(R2 - 0.28). Response was also correlated with
nitrate, although significance disappeared when 1
extreme nitrate value was removed.
Lipid Peroxidation: Only PMSM enhanced lipid
peroxidation in BEAS-2B, correlating with both
elemental and organic carbon.
TNF-a: In RAW264.7, PMSM increased TNF
production. Polymixin pretreatment significantly
reduced TNF levels for all 3 PMs which indicates
an endotoxin role in macrophage response. TNF
production (after polymixin pretreatment only) was
associated with Cr and Fe content.
July 2009
D-46
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Hutchison PMio: Samples collected for 7d during closure
et al. (2005, 097750) (-C) and reopening of steel plant (-R)
Species: Mouse PMT: PM total (aqueous sonicate)
Cell Line: J774.1A PMS: PM soluble aqueous
PMI: PM insoluble aqueous
Particle Size: PMio
Route: Suspension
Dose/Concentration: 500/j\
(estimated concentrations of 112,
143,156,180, 233, 255 //g/1 ml
water)
Time to Analysis: 4h
Particle Characterization: Reopening of the
plant showed a significant increase in the total
and acid extractable metal content of PM.
Aqueous extractable metal content did not
change. Soluble zinc, copper and manganese also
increased significantly post reopening. Iron was
the most abundant in acid extractable metals and
increased greatly at the reopening.
TNF-a: PMT-R and PMT-C induced a statistically
significant increase. Treatment with chelation
agent reduced effect to control levels.
Reference: Imrich A UAP: SRM 1649 (positive control)
et al. (2007,155859) Ti02. Partic|e contm|
Species: Rat
Gender: Female
Age: 12-14wks
Cell Type: AM
CAPs (Boston, MA)
All cells primed with LPS
Coexposure with NAC, dimethylthiourea
(DMTU), H202 or catalase
Particle Size: CAPs: n2.5/jm; UAP: PM2.6;
TiO2: ~ 1 /jm
Route: Cell Culture (2 X10!l
cells/well)
Dose/Concentration: Caps
100/yg/mL; UAP: 50 or 100
/yglmL; LPS 250 nglmL; NAC,
DMTU: 2,10, 20mM; Catalase: 1,
5,10mM; H202 0-50/ym/hr
Time to Analysis: 18-20h
TNF: DMTU at 20mM reduced TNF in LPS-primed
cells in control and UAPtreated groups. NAC at
20mM reduced TNF release but this was not
statistically significant. Catalase significantly
inhibited TNF in control and UAP treated groups.
CAPs (especially the insoluble portion) significantly
increased TNF unless co-exposed with NAC,
DMTU or catalase. All three reduced levels back to
around basal levels. DMTU was particularly
effective at diminishing TNF release. H202
increased TNF release in CAPs-exposed cells. Ti02
had no increased ability to induced cytokine
release when mixed with H202.
Cell Death: Viability decreased substantially
when exposed to H202 + CAPs. The soluble
fraction of CAPs showed to be more effective
with H202 than the insoluble portion. Ti02 had no
significant effect.
NO: Some CAPs induced slight increases when
mixed with H202. No difference was observed
between soluble and insoluble portions of CAPs.
DFO: DF0 at 0.05mM completely inhibited
oxidation induced with soluble CAPs + H202.
Insoluble CAPs + H202was also DFO-sensitive.
DFO was ineffective against the insoluble CAPs
induction of TNF and MIP-2.
Reference: Ishii et al.
(2004,088103)
Species: Human
Cell Type: A549
(collected from 6
lobectomy or
pneumonectomy
smokers), HBEC
EHC-93:PMio (obtained from Environmental
Health Directorate, Ottawa, Ontario, Canada)
Particle Size: PMio
Route: Cell Culture (1x10E7 cells)
Dose/Concentration: 100 //gfmL
Time to Analysis: 3, 6, 24h
Cytokines: TNF-a, IL-1(3, GM-CSF, IL-6, and IL-8
levels were significantly increased in A549 cells.
mRNA Expression: MCP-1, ICAM-1 and IL-8
mRNA expression increased in untreated AM
supernatants at 3h. Only the MCP-1 levels were
statistically significant at 3h. Levels declined by
6h. When A549 cells were exposed to PMio
exposed AM, levels of RANTES, TNF-a, ICAM-1,
IL-1 p, and LIF increased. Except for RANTES
mRNA, these differences were less in the 6h
samples. VEGF increased as well, but this increase
was not statistically significant.
TNF-a and IL-1 p-neutralizing Antibodies: IL-
1 p antibody alone or in combination with TNF-a
significantly reduced expression of all eight
mRNAs. Combinations for some mRNAs reduced
expression by up to 1/2. This effect was not
observed when A549 was treated with the control
AM.
Transcription Factor Binding Activity: Binding
of AP-1 and Sp1 increased when A549 treated
with supernatants from PMio-exposed AM, but not
from control AM.
July 2009
D-47
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Ishii H et
al. (2005, 096138)
Species: Human
Cell Type: AMs
(obtained from 10
smokers who stopped
smoking 6wks prior),
HBEC
EHC-93: PMio (obtained from Environmental
Health Directorate, Ontario, Canada)
Particle Size: PMio
Route: Cell Culture (HBEC: 2.5-
3.0x106 cells: AM: 1x107 cells: co-
culture of AM/HBEC: 5x100 cells)
Dose/Concentration: 100 //gfmL
Time to Analysis: 2, 24h
mRNA Expression after 2h exposure: AM or
HBEC exhibited no effect. In contrast, co-culture
increased expression of MIP-1 p, GM-CSF, M-CSF,
IL-6, MCP-1 and ICAM-1-mRNA.
mRNA Expression after 24 h exposure: AMs
exhibited no effect. HBEC increased levels of GM-
CSf, LIF and ICAM-1. Co-culture, on the other
hand, increased expression of MIP-1 p, GM-CSF,
M-CSF and ICAM-1 mRNA.
Protein Levels: AM and HBEC both increased
GM-CSF, IL-6 and MIP-1 p release into the
supernatant. Co-culture effect was not additive
but synergistic (i.e., higher than expected). MCP-1
levels did not increase significantly. Co-culture
appeared to decrease protein levels for both the
control and PM values. M-CSF levels increased for
co-culture only.
Surface Expression of ICAM-1: Upon 24h
exposure to PM, HBEC exhibited an increase in
expression. Expression in AMs were not affected
by 2h PM stimulation.
ICAM-1 Inhibitors: IgG or anti-CDUb antibody
was unaffected in co-culture.
Reference: Jalava P
et al. (2005, 088648)
Species: Mouse
Cell Type: RAW
264.7
UPM: SRM1649a (Washington, DC)
DEP: SRM1650 (NIST)
EHC-93: Ottawa dust (Environmental Health
Center, Ottawa, Canada)
HFP-00: Pooled ambient air PM2.6 sample from
Helsinki, Finland
M-UPM: methanol extract of UPM
Particle Size: SRM 1649a, SRM 1650, EHC-
93: NR; HFP-00: PM2.1,
Route: Cell Culture (5 X10!l
cells.i'niL)
Dose/Concentration: 150 //gfmL
Time to Analysis: Methanol
treatment of PM samples: 24h;
Exposure to ambient PM samples:
2,4, 8,16, or 24h.
TNF-a: All the PM samples increased TNF-a.
Cell Viability: SRM1649a exhibited the most
cytotoxicity, followed by HFP-00 and EHC-93.
Methanol significantly affected cytotoxicity of the
EHC-93 sample only.
Cytokines: TNF-a concentrations in the cell
culture medium significantly increased at all time
points between 2 and 24 h. The highest increase
was seen in EHC-93. IL-6 production also
increased at different levels with the highest
increase observed in EHC-93. No response was
observed for IL-10.
Cell Viability: Duration of exposures had no
significant effect on any of the samples. A 2h
exposure time was sufficient to induce the typical
reductions in cell viability.
Reference: Jalava PI
et al. (2006,155872)
Species: Mouse
Cell Type: RAW
264.7
PM: Collected east of Helsinki, Finland
between Aug 23 and Sept 23, 2002
Divided in 12 groups (4 sizes by 3 exposure
types):
¦S: seasonal average
¦W: wildfire
¦M: mixed
¦B: blank
Particle Size: PMio 2.5; PM2.B-1: PM1-0.2;
PM0.2
Route: Cell Culture (5x106
cellsfmL)
Dose/Concentration: 15, 50,
150 and 300/yg/mL
Time to Analysis: 24h
Particulate Mass Concentrations in HVCL Size
Ranges: The largest increase of PM
concentrations was observed in PM1-0.2.
NO: All 12 samples increased NO production when
compared to corresponding unexposed controls.
Peaks were observed at 150 /yglmL, except in
PM1-0.2.
Cytokines: All 12 samples increased TNF-a and
IL-6 production. PMio-2.5 and PM2.B-I produced a
much larger response than PM1-0.2 and PM0.2.
IL-6 production for PM0.2 was not measured.
MIP-2 production also increased with similar
trends.
Cytotoxicity: All 12 samples induced dose-
dependent decreases in cell viability. PMio-2.5
were the least active inducers of apoptosis while
PM0.2 showed the highest activity (4-17% of
apoptotic cells).
July 2009
D-48
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Jalava et
al. (2007, 096950)
Species: Mouse
Cell Type: RAW
264.7
Urban background PM
PMio, PM2.6, and PM0.2 collected from 6
European cities during different times of the
year from October 2002 to July 2003:
D: Duisburg (Fall)
P: Prague (Winter)
A: Amsterdam (Winter)
HR: Helsinki (spring),
B: Barcelona (spring)
AT: Athens (summer)
Particle Size: PM10: 2.5-10 /jm; PM2 b: 0.2-
2.5/jm; PM0.2: <0.2//m
Route: Cell Culture (5 X10!l
cells.i'niL)
Dose/Concentration: 15, 50,
150, 300/yg/mL
Time to Analysis: 24h
PM Characterizations: The highest mass
concentrations of PM10 and PM0.2 were
measured in Athens. Prague had the highest PM2.6
concentrations.
NO: All PM fractions induced statistically
significant NO production in macrophages. PM2.6 -
P and PM2.6 -AT produced significantly larger
responses, though all samples at 150 and 200
/yglmL induced statistically significant production.
When compared to the other PM0.2 samples, -P
and HR produced significantly larger responses.
Cytokines: PM10 showed average cytokine
production to be 7.8 fold and 83 fold for TNF-a,
and 4.4 fold and 530 fold for MIP-2 when
compared to PM2.6 and PM0.2 respectively. PM10
induced statistically significant increases in
production of TNF-a, MIP-2 and IL-6. PM2.6, with
exception of Prague, caused significant increases
in cytokines. PM0.2-A and -AT showed small yet
statistically significant increases in TNF-a. An
increase in MIP-2 was observed with -P and -HR.
IL-6 increased significantly with PM10 and slightly
with PM2.6. In the PM0.2 range, only the -A and -
AT samples caused a small, statistically
significant TNF- a production. MIP-2 production
was only detected from the -P and -HR samples.
PM0.2 effects on IL-6 response were negligible.
Cytotoxicity: The average cytotoxicity of PM10
and PM2.6 were roughly equal, but PM0.2 were
less cytotoxic with the exception of -P. The dose-
response trends for most of the samples were
linearly declining, with PM10 and PM2.6 exhibiting
statistically significant declines in viability.
Reference: Jimenez	PM10: Collected from London and Edinburgh
et al. (2002,156610)	air particulate monitoring stations.
Species: Human	Ti02
Cell Type/Line:	UFTi02 Both fine and ultra fine Ti02 fractions
A549, THP-1, Mono	obtained from Tioxide Europe (London, UK)
Mac 6 (DSMZ)	and Degussa-Huls (Cheshire, UK)
Particle Size: PM10, Titk 200nm; UFTi02:
20nm
Route: Cell Culture (110,625
cells/well)
Dose/Concentration: PM10, TiOy,
UFTi02: 100 //g/mL; TNF-a:
10ng/mL
Time to Analysis: 4h
NF-I
-------
Study
Pollutant
Exposure
Effects
Reference: Jung et
al. (2006,132421)
Species: None
Type: Surrogate Lung
Fluid
Soot Particles: Generated using a co-flow,
laminar, diffusion flame system
CB (Degussa)
PM2.E: Collected using IMPROVE air pollution
samplers
Particle Size: Soot: 185nm; CB: 25nm, PM2 t
Route: Surrogate Lung Fluid
Dose/Concentration: Soot: 0-
30mg: CB: 5-1 Omg: PM2.6: 50 or
100/yg/mL
Time to Analysis: Parameters
measured continuously over 2h.
OH Radical Formation: Formation occurred with
linear dependence on soot mass. Average response
was 0.89nmol OH produced per mg of soot.
Formation also occurred with soot + hydrogen
peroxide. Hydrogen peroxide alone did not form OH
radicals.
Fe: Average Fe concentration in soot particles
was 305 ± 172nM. Observed negative correlation
between amount of Fe and amount of OH radical
formation. DSF inhibited iron-induced increase in
OH radical formation.
Carbon Black: OH radical generation by carbon
black was significantly less than soot. OH
generation by CB was observed to be linearly
proportional to PM mass, but CB was much less
efficient at generating the OH radical.
PM2.5: A high variability in the increase of OH
radicals was observed with PM2.6. Pretreatment
with DSF partially blocked OH radical production,
but a significant level remained. This may be due
to PM2.E containing high levels of Fe and Cu.
Reference: Kafoury
and Madden (2005,
156617)
Species: Mouse
Cell Type: RAW
264.7
DEP: SRM 1975 (purchased from NIST,
Rockville, MD)
BAY11-7082, NF-kB inhibitor (coexposure)
IL-1B: obtained from Santa Cruz
Biotechnology (Santa Cruz, CA)
Particle Size: DEP: 0.3/ym (mean diameter)
Route: Cell Culture (3-4x106 cells) TNF-a: DEP induced a significant release of TNF-a
Dose/Concentration: DEP 25,
100, or 250 /yglmL; IL-1B: 100
ng/mL
Time to Analysis: DEP: 4h pre-
treated with BAY11 -7082 for
1.5h; IL-1B: 4h
at 100 and 250 //gfmL dose-dependently.
Exposure at 25/yg/mL had no effect. IL-1 (3
containing PM samples at 100 //gfmL also
resulted in a significant release of TNF-a.
NF-I
-------
Study
Pollutant
Exposure
Effects
Reference:
Katterman ME et al.
(2007, 096358)
Species: Rat
Cell Type/Line: RLE-
6TN (Alveolar
Epithelial Cell Line)
PM: Oils: OAAF, Oil Q, Oil III, NF2
PM: Coal Germany and Ohio
Diesel particulates: ZODDA (doped with Zn),
ZSDDA (doped with Zn and S): S: PMs
washed in solution; F: Fresh samples;L:
Leached
AI2O3, Fe203, Si02, Ti02,Zn0 also tested
Particle Size: NR
Route: Cell Culture (cytotoxicity:
50,000 cells/well; SEM: 25,000
cells)
Dose/Concentration: Oils
0.2mg/mL; Coals 0.7mg/mL; Diesel
0.01 mg/mL; AI203 0.5mg/mL;
Fe203 0.7mg/mL; Si02 0.7mg/mL;
Ti02 0.7mg/mL; ZnO 0.05mg/mL
Time to Analysis: 24h
Metabolic Activity: For oils comprised of 3/4
fresh and 1|4 leached, metabolism decreased.
Coals (fresh and leached) had no effect. ZODDA-F
and ZSDDA-F both induced decreases in activity.
ZSDDA-L had no effect.
Cellular Morphology: PM S had a minimal effect.
PM-F induced widespread cell damage.
Constituent Differences between PM-F and
PM-L: In oil samples Cu, Ti and Ca salts were
removed upon washing. Fe, Al, Si remained
constant.
Grinding Effects: Coal toxicity increased upon
grinding, whereas diesel PM toxicity decreased
upon grinding.
Metal Oxide Effects: Only Si02, and ZnO (much
higher at lower concentrations than other metal
oxides) decreased metabolic activity. Fresh,
washed and sonicated samples exhibited similar
results. Grinding only affected Ti02 (increase) and
ZnO (decrease).
Reference: Kendall et
al. (2004,156634)
Species: Human
Tissue Type: BALF
(obtained by
bronchoscope from 6
nonsmokers and 3
smokers)
PM2.6 sample sites; 2 schools in Bronx, NY, 6
background urban, 6 urban roadside. Sampling
occurred 24h/day for 12d.
Particle Surface Chemistry: 79-87%
carbonaceous material (Ch, COO, C-(0,N)), 10-
17% 0 (01 s), 1.5-4% N (NH4+, N-C, NOs 2-),
0.6-1% S, and 0.3-2% Si.
Only NO3 - higher in roadside samples.
NH4 and NO3 - correlated with NO and NOx in
air but not NO2.
Particle Size: PM2.6
Route: BALF treatment
Dose/Concentration: 5-10ml of
0.5M NaClor BALF
Time to Analysis: Filters treated
with BALF for 4h
Saline Washing: Removed particles and
decreased NH4, N0:i,0 and S relative to C1.
BALF treatment (XPS): PM2.6 surfaces interacted
strongly with BALF within hours of contact.
Specific surface components of PM2.6 immersed in
BALF were desorbed while biomolecules from
BALF were adsorbed to particles. N-C on the PM
surface increased 3 fold for smokers and 4 fold
for nonsmokers (range 1.4-7.4). This is most likely
related to protein-like adsorption on PM.
Treatment also induced a slight increase in COO
and decreases in NH4, NO3, 0 and S.
ToF-SIMS - Organics: Particle loading and
surface hydrocarbons showed a linear correlation.
Loss of hydrocarbons from PM2.6 surface averaged
55% (10-75) after undergoing saline and BALF
washes. In only 3/12 samples BALF removed less
hydrocarbon. BALF treatment increased the amino
acid and phospholipid content of the PM2.6
surface.
ToF-SIMS - Inorganic: Saline washing appeared
to increase Al and Si but with extreme variability;
this increase was not statistically significant.
Both saline and BALF washing decreased NhU and
Na levels to a similar extent. BALF washing did
not affect Al or Si.
Reference: Kim et al. Zn2 +
(2005, 088454) Particle Size: NA
Species: Human
Cell Type/Line:
BEAS-2B
Route: Cell Culture
Dose/Concentration: 15, 50,
100/ymol
Time to Analysis: 1-20h
Cell Viability: At 50//M for 20h, no apoptosis
was induced.
IL-8: At 12h, IL-8 increased in dose-dependent
manner. At 15 or 50//M, Zn2 + increased protein
1.6 and 4.6 fold respectively. IL-8 mRNA
expression increased dose-dependently, reaching
statistical significance at 2 h and continuing until
4h.
EGFP (adenoviral IL-8 promoter): Levels
increased 2.4 fold with 50//M Zn2 + .
Proteases: With 50//M Zn2 +, phosphorylation of
MAPKs ERK, JNK and p38 increased by 15min
and continued increasing up to 2h. Pre exposure of
inhibitors of MEK, JNK, before Zn2 + exposure
caused inhibition of Zn-induced IL-8 mRNA and
protein production. Inhibitor of p38 had no effect.
Dephosphorylation of ERK and JNK was partially
inhibited with exposure to Zn2+.
July 2009	D-51	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Kleinman
et al. (2003, 087938)
Species: Rat
Strain: Wistar,
Fischer 344
Age: 22-24m, 10wks
Cell Type/Line: AMs
UF1: Utrecht 1 Fine (urban freeway)
UC1: Utrecht 1 Coarse
UF2: Utrecht 2 Fine (urban, freeway, light
industrial)
UC2: Utrecht 2 Coarse
SRM 1650
SRM 1648
Particle Size: UF1: 0.2-2.5 fjm; UC1: 2.5-10
/jm; UF2: 0.2-2.5/jm; UC2: 2.5-10/ym
Route: Cell Culture (10E cells/well
at 10e cells/mL)
Dose/Concentration: 1.2 to
1200 ng/106 cells
Time to Analysis: 4,18h
Macrophage PMA-stimulated respiratory
burst activity: SRM 1648 and 1650 induced
dose-dependent decreases approaching 0 at 50 -
100/yg|10E cells. Large dose-dependent decreases
from old rat AMs exposed to fine PM exposure
were followed by young rat AMs exposed to fine
PM. However, no age-related effects were
statistically significant.
Free radical production: All coarse particles
depressed free radical production in a semi-dose-
dependent manner, with UC2 exhibiting more
potency than UC1. Both fine particles also showed
dose-dependent responses but UF1 and UF2
responses were greater than the control at 3
/yg/10e cells.
PM Characterization: Ratios between coarse
and fine PM were similar for metals tested (Al, Fe,
Mn, Zn). Al was higher in coarse samples and Zn
higher in fine PM, although large variability was
observed. Fe and Mn results were roughly
equivalent for all samples.
Reference: Kocbach,
A.
Namork, E.
Schwarze, P.E.
2008
Species: Human
Cell Type/Line: THP-
1
PMW: Wood smoke particles
Collected from conventional Norwegian wood
stove burning birch
PMT+: Traffic-derived particles
Collected from road tunnel in winter when
studded tires were used
PMT-: Traffic-derived particles
Collected from road tunnel in summer without
studded tires
DEP: SRM2975
Porphyr: fine grain syenite porphyry (prepared
by SINTEF, Trondheim, Norway)
Polymyxin B Sulphate (endotoxin inhibitor)
Particle Size: PMW, PMT, DEP: NR; Porphyr
8 /jm (mean)
Route: Cell Culture (1,000,000
cellsfmL)
Dose/Concentration: 30-280
/yg/mL
Time to Analysis: 2, 5,12h
Particle Characterization: PMT + contained a
high mineral particle content. PMT- contained
carbon aggregates, organic carbon and polycyclic
aromatic hydrocarbons (PAH). PMW and DEP
contained carbon aggregates. PAH content of
PMW was greater than DEP. Porphyr was not
included in the analysis.
Cytokines: PMT± induced releases of TNF-a, IL-
1 (3, and IL-8 with 30 or 70/yglmL. PMW similarly
induced TNF-a and IL-8. DEP induced IL-1 p and
IL-8. Porphyr induced IL-8 increases. IL-4, IL-6 and
IL-10 were unaffected. Overall, the order of
effective cytokine induction from most to least
effective was PMT±, PMW, DEP, and Porphyr.
mRNA expression of TNF-a, IL-1 (3, IL-8, and IL-10
increased with 140 /yglmL of PMT ± and slightly
for PMW.
LDH: PMT ± induced small but statistically
significant increases at low doses. DEP increased
LDH at 280/yg/mL only.
Polymyxin B Sulphate: The endotoxin inhibitor
significantly inhibited LPS-induced cytokine
release by 80-90% and reduced PMT ± induction
by 50-60%.
Organic Extraction: PMT + washed and native
particles showed equivocal induction of cytokine
release. PMT + organic extract had no effect.
PMT- and PMW organic extracts significantly
increased TNF-a and IL-8. Washed particles
induced less signficant increases of IL-8. DEP
organic extract had no effect.
Reference: Kristovich CP: carbon particle (carbonaceous negative
R et al. (2004,
087963)
Species: Human
Cell Type/Line:
HUVEC, HPAEC,
HPMVEC, HPBMC
image of zeolite)
CFE: C/Fe particulate (synthesized)
CFE + : C-Fe/F-AI-Si particulate (synthesized)
CFA: Coal Fly Ash (Coal-fired power plant,
N0S)
DEP: (exhaust pipe of diesel powered truck)
CP, CFE, CFE+ approx 1 /jm (resembling
zeolite)
Particle Characterization (Surface chemistry):
CP - 88% C, 1% Si, 10% 0,1% N.
CFE -80% C, 2% Fe, 2% Si, 16% 0.
CFE+ - 20% C, 6% Al, 3% Si, 50% F, 6% 0,
11 % N, 4% Na. CFA - 25% C, 3% Fe, 13%
Al, 17% Si, 41% 0,1 % N. DEP - 70% C, 3%
Fe, 24% 0,1%N, 2% S.
Particle Size: CP, CFE, CFE+: approximately
1 /jm (resembling zeolite): CFA: < 2 fjm; DEP:
150nm
Route: Cell Culture (4x10E6
cells/well)
Dose/Concentration: CP: 5-50
/yg/cm2: CFE: 2.5-25/yglcm2;
CFE+: 2.5-25/yglcm2; CFA: 10-
100/yglcm2; DEP: 2.5-25/yglcm2
Time to Analysis: 4, 8, or 24h
Cytotoxicity: CP exhibited no effects. DEP and
CFE exhibited intermediate toxicities in the range
of 50-70/yglcm2. No toxicity was apparent when
treated with CFA (up to 200/yglcm ) or
synthesized C particulates.
Endothelial Activation: ICAM-1, VCAM-1, and E-
selectin were activated dose-dependently by DEP,
CFE, and CFE +. No effects observed for CFA or
CP. These effects were not the result of endotoxin
Individual Variability: Donors (humans) showed
variability in responses especially for CFA. 3/9 had
a medium response negated by ND responses in
6|9.
July 2009	D-52	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Kubatova, A.
Dronen, L.C.
Picklo, M.J.
2006
Species: Rat, Human
Cell Type/Line: RAW
264.7 (rat), BEAS-2B
(human)
PMW: Wood Smoke
Collected from airtight wood stove burning
hardwoods
¦P: Polar (fraction extracted from 25-50 C)
¦MP: Mid Polar (fraction extracted from 100-
150 C)
¦NP: Nonpolar (fraction extracted from 200-
300 C)
¦C: P + MP + NP
Particle Size: NR
Route: Cell Culture (RAW 264.7:
10e cells/mL; BEAS-2B: 106
cellsfmL)
Dose/Concentration: 50,100,
200 /yglmL
Time to Analysis: 12h
GSH: PMW-MP and PMW-NP induced GSH
depletion substantially in a dose dependent
manner starting at 50 //gfmL in both cell types.
DMS0 had no effect.
Cytotoxicity: PMW-MP and PMW-NP increased
cytotoxicity at 200/yglmL in RAW 264.7. BEAS-
2B was unaffected.
Particle Characterization: PMW-MP contained
higher concentrations of oxy-PAHs, disyringyls,
syringylguaiacyls and PAHs. oxy-PAHs include 9-
fluorenone, 1 -phenalenone, 9,10-anthraquinone
and hydroxycadalene. PAHs included
phenanthrene, fluoranthene and pyrene.
Effects of Individual Components of PMW-MP
on GSH: 1,8-dihydroxy-9-1 Oanthraquinone and
9,10-phenanthraquinone depleted GSH. 9,10-
anthraquione, anthrone, 1 -hydroxypyrene
increased GSH. Phenanthrene, 1-methylpyrene, 9-
fluorenone and xanthone had no effect.
Reference: Kubatova
et al. (2004, 087986)
Species: Monkey
Cell Type/Line:
African green monkey
kidney cells
designated COS-1
(CV-1 cells with origin
¦defective mutants of
SV40), E coli PQ 37
(SOS Chromotest)
DEP: Obtained from diesel bus
PMW: Wood smoke particulates obtained from
airtight wood stove burning hardwood
HSF: Hot pressure fractionation
¦C: P + MP + NP
¦P: Polar
¦MP: Mid Polar
¦NP: Nonpolar
OE: Organic Extraction
¦HNP: n hexane nonpolar
¦MEP: methanol polar
Particle Size: NR
Route: Cell Culture (10,000
cells/180 ul)
Dose/Concentration: 0, 50,100,
150, 200, 250, 300/yg|mL
Time to Analysis: Cytotoxicity:
24h; Chomotest: 2h SOS
Cytoxicity: PMW induced cytotoxicity in a dose-
dependent manner. PMW-HNP induced low
cytotoxicity, followed by PMW-C (intermediate)
and PMW-MEP (highest). Levels above 25/yg/mL
were cytotoxic. DEP-HNP induced cytotoxicity but
was not dose-dependent. Results similar for all 3
fractions (highly variable). All fractions with
concentrations higher than 100 //g/mL were
cytotoxic.
Extraction Water Temperature Effect: PMW
was cytotoxic at temperatures over 50 C. DEP
was cytotoxic at temperatures higher than 200n
C. At 250n, cytotoxicity between DEP and PMW
was similar. At 300n C, PMW cytotoxicity
declined and DEP stayed high, resulting in DEP
inducing higher cytotoxicity than PMW.
SOS Chromotest: fB-Galactosidase formation
increased, peaked at 200n C with DEP and
declined to control at 300n C. Individual fractions
showed linear dose response from 25-200 //gfmL
with 150n C and 200n C extracts significantly
higher.
Reference: Lee et al.
(2005,156682)
Species: Human
Cell Type/Line: A549
MEP: Motorcycle Exhaust Particles (Yamaha
Cabin engine, 95 octane unleaded
gasoline,150 rpm)
MEPE: MEP Particle Free
MEP 0.5//m
MEPE < 0.2//m
Particle Size: MEP 0.5um; MEPE < 0.2/jm
Route: Cell Culture (1x106
cells/well)
Dose/Concentration: MEP 0.02,
0.2, 0.2, 2, 20/yglmL; MEPE 20
/yg/mL
Time to Analysis: 24h
IL-8: MEP induced IL-8 at concentrations greater
than 0.2 /yglmL. Levels increased 2fold at 24 h
with 20 /ygfmL. MEPE induced similar responses
at 20/ygfmL. Induction of IL-8 mRNA expression
was dose-dependent with MEP and MEPE.
Cytotoxicity: Exposure to particles did not affect
cytotoxicity.
NFkB: MEP (20 //gfl) induced time-dependent
activation for 2h and continued at same level for
up to 6h. Pretreatment of PDTC (1 mM) fully
inhibited MEP induction.
MAP Kinase: MEP induced time-dependent
activation up to 30 min and stayed elevated for at
least 60 min.
ROI: MEP treatment induced a time-dependent
increase in ROI for up to 1 h and then continued
the at same level for up to 6h.
July 2009	D-53	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Lee, C.C.
Kang, J.J.
2002
Species: Mouse
Strain: Peritnoeal
Macrophages, RAW
264.7
Species: Rat
Strain: AMs
MEP Yamaha 2-stroke engine using unleaded
gas)
MEPE(particle-free MEP)
Particle Size: 0.5 /jm
Route: Cell Culture (5 X10!l
cells/mL (Cytotoxicity), 3 X106
cellsfmL (Apoptosis), 2 X106 cells
(MMP and ROD, 1 X107 cells
(GSH)
Dose/Concentraion: 5,10, 50,
100,300,1000/yglmL
Time to Analysis: 6,12,18, 24h
Cytotoxicity: Viability decreased dose and time-
dependently in all cell types at 24h.
Apoptosis: subG1 significantly and dose-
dependently increased at the 300 MEP //gfmL
dose in all cell types, indicating increased
apoptosis. MEPE induced similar results. Inhibition
was successful against MEP-induced apoptosis by
calcium chelators EGTA, BAPTA-AM, cyclosporin
A and antioxidants NAC, GSH, catalase and SOD.
Ca2+: MEP and MEPE increased Ca2+ at 300
/yglmL. BAPTA-AM completely inhibited induction.
ROI: MEP increased R0I in a time-dependent
manner. Calcium chelators and antioxidants
substantially attenuated induction.
GSH: MEP significantly decreased GSH.
MMP: Mitochondria membrane potential
decreased dose-dependently with MEP 100 /yglmL
and 300 //gfmL. Calcium chelators and
antioxidants partially inhibited reduction.
Reference: Li et al.
(2002, 042080)
Species: Mouse
Cell Line: RAW
264.7, THP-1
Species: Murine
Cell Line: RAW
264.7
VACES (Biosampler PMio in Downey, CA -DEP
concentrate in water)
DEPM (DEP methanol extract)
DEPME (DEP methylene chloride extracts)
DEPAL (DEPME aliphatic (hexane))
DEPAR (DEPME aromatic (hexane/methlene
chloride))
DEPP0 (DEPME polar (methylene
chloride/methanol))
Particle Size: NR
Route: Cell Culture (2 X10'!
cells/well Mouse RAW 264.7 and
THP-1; 0.67 X106 cells/well
Murine RAW 264.7)
Dose/Concentration: 10 - 200
/yg/mL
JNK Activation and IL-8
Production: THP-1 cells- 0,10, 25,
50,100/yg|mLDEPM; THP-1
cells- 0,10,25, 50,100 //g/mL of
DEP; RAW264.7 cells- 10-100
DEP/yg/mL
Cytotoxicity: 1,10, 25 (THP-1
cells only), 50,100, 200 /yglmL
GHS/GSSG: 0,10,25, 50,100
/yg/mL
H0-1 Expression: 0, 25, 50,100,
200 /yglmL
Time to Analysis: GHS/GSSG:
DEPM, whole DEP (RAW
264.7only) 8h.
H0-1, MnSOD Expression: RAW
264.7, THP-1 7h. RAW 264.7
cells exposed to whole DEP 16h.
JNK Activation, IL-8 Production:
THP-1 cells 30min, 16h. RAW
264.7 cells 90min.
Cytotoxicity: RAW264.7, THP-1
18h.
GSH/GSSG Ratio: DEPM induced dose-dependent
decrease in GSH/GSSG ratios in both cell lines.
DEP induced decreases at comparable doses to
DEPM.
HO-1 Expression: Cells exhibited dose-dependent
increases in H0-1 expression.
HO-1 Expression in Murine RAW 264.7:
VACES-F consistently induced H0-1 expression
over a 9m period, whereas VACES-C was effective
in inducing H0-1 during fall and winter. H0-1
induction positively correlated to higher 0C and
PAHs that were represented in VACES-F, but also
seen with a rise in PAHs in VACES-C during winter
months.
MnSOD: At doses of 2.5/yglmL, DEPM increased
MnSOD in THP-1 cells.
JNK Activation: DEPM dose-dependently
increased JNK phosphorylation but did so without
a change in the JNK expression level. DEP-exposed
mouse RAW264.7 cells exhibited similar increases
in JNK phosphorylation but without increasing
JNK expression.
IL-8: Exposure to DEPM elicited dose-dependent
increase in IL-8 levels of THP-1 cells.
July 2009
D-54
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Li et al.
(2002,087451)
Species: Human
Cell Line: BEAS-2B,
NHBE, THP-1
macrophages
Species: Murine
Cell Line: RAW
264.7, macrophages
DEPM (DEP methanol extract)
DEPME (DEP methylene chloride extracts)
DEPAL (DEPME aliphatic (hexane))
DEPAR (DEPME aromatic (hexane/methlene
chloride))
DEPPO (DEPME polar (methylene
chloridefmethanol))
Particle Size: 0.05-1 /ym
Route: Cell Culture (10I! cells/mL) ROS: BEAS-2B cells demonstrated increased HE

50,100 /yg/mL
Time to Analysis: 30, 60,
120min

fluorescence, indicating increased ROS formation.
THP-1 cells were unaffected.

GSH/GSSG Ratio: DEPM dose-dependently
decreased GSH/GSSG in THP-1 and BEAS-2B
cells. Similar changes occurred with NHBE cells.
THP-1 cells maintained a higher ratio of
GSH/GSSG than BEAS-2B and NHBE cells.
NAC on GSH/GSSG Ratio: Exposure to DEPM in
the presence of NAC did not affect the GSH/GSSG
ratio in BEAS-2B and NHBE cells. In THP-1 cells,
NAC prevented a decline in the GSH/GSSG ratio.
MnSOD and HO-1: THP-1, BEAS-2B and NHBE
cells showed constitutive MnSOD expression and
dose-dependent expression of HO-1 protein and
mRNA. No change occurred in the expression of fB-
actin.
DEPAL, DEPAR, DEPPO, CoPP on HO-1
Expression: DEPPO was more potent than
DEPAR. DEPAL lacked activity for THP-1 and
BEAS-2B cells. The potency of DEPPO was
sufficient to affect cellular viability and HO-1.
CoPP induction of HO-1 failed in THP-1 cells, but
succeeded in BEAS-2B cells. However, it did not
protect against the oxidizing effects of DEPM.
JNK: JNK activation increased in DEP-exposed
THP-1 and BEAS-2B cells.JIMK isoforms were
observed at doses of a 25 /yg/mL. In BEAS-2B
cells a high rate of cell death diminished this
response at 100/yg/mL. NHBE also showed
increased JNK phosphorylation at doses 50 -100
/yg/mL.
NAC on JNK: NAC led to inhibition of JNK
activation.
IL-8: THP-1 cells showed dose-dependent
increases of IL-8. NHBE cells showed incremental
increases followed by rapid decline at 100/yg/mL
attributed to apoptosis. BEAS-2B cells responded
to 10 /yg/mL with increased IL-8, but cellular
toxicity and cell death led to a drop in IL-8
production at higher doses.
Cytotoxicity: Comparing cytotoxicity at 25
/yg/mL DEP, BEAS-2B cells had a higher rate of
cell death than THP-1 cells. BEAS-2B cells showed
a significant rise in cell death at doses larger than
10/yg/mL. In THP-1 cells, it took doses of 25
/yg/mL or more before significant increases
occurred.
In BEAS-2B, cell death began at 2h. In THP-1,
increases in cell death prolonged for 8h or longer.
NHBE cells also showed increase rates of
cytotoxicity compared to macrophages. NAC in
THP-1 interfered with a generation of
cytotoxicity, but NAC did not have any decreasing
effect on cell death in BEAS-2B or NHBE cells.
July 2009
D-55
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Lindbom
et al. (2007,155934)
Species: Mouse
Cell Line/Type: RAW
264.7
PMio:
-ST: Street
¦S: Subway
¦G: Granite
¦Q: Quartzite
(-G and -~ generated by road simulator at
Swedish National Road and Transport
Research Institute)
Bimodal with peaks around 4-5um and 7-8um.
Particle Size: PMio
Route: Cell Culture (130000
cells/cm2)
Dose/Concentration: 1,10 or
100/yg/mL
Time to Analysis: 18, 24h
Analysis of Arachidonic Release
(AA):
Cells pre-incubated wf 1 uCI
Cellular Viability: Viability was not influenced by
any particle types and in all cases exhibited 90%
or higher viability, except for the combination of
subway particles and NAC where viability dropped
to 20%.
Cytokines: All particles induced TNF-a secretion
in a dose-dependent fashion. PM-S was most
potent at 1 /yglmL. PM-G and PM-ST induced
effects at 10/yg/mL. PM-Q induced increase of
tritium marked for AA and washed TNF-a at 100/yg/mL.PM-ST induced IL-6 release
exposed to 10, 50,100 and 250 at 10/yg/mL. PM-G, PM-Q, PM-S induced IL-6
//g/mL	secretion at 100/yg/mL. DFX inhibited TNF-a in
cells exposed to PM-S and PM-ST. DFX induced
increase of TNF-a with PM-Q. For all PM types
(except PM-ST) DFX inhibited induced IL-6
secretion.
NO: PM-ST and PM-G induced a significant
release of NO, with PM-ST inducing a higher NO
release than PM-G.
NAC: NAC treatment significantly inhibited both
TNF-a and IL-6 secretion with all PM particles.
L-NAME: L-NAME caused a decrease in NO
secretion at 100 /yglmL of PM-ST. L-NAME did
not have an effect on granite-induced NO
secretion at 100/yg/mL.
Cytokine Gene Expression: TNF-a mRNA
showed a trend to increase for -ST, but this did
not reach significance. IL-6 gene expression
increased for PM-Q, PM-ST, PM-S but not for PM-
Q.
AA Release: PM-S exposure at 100 and 250
/yglmL was the only PM to induce AA release.
Lipid Peroxidation: All particle types induced
lipid peroxidation. PM-S and PM-ST induced
significantly higher lipid peroxidation as compared
to PM-Q and PM-G.
ROS: All particle types induced R0S formation.
PM-S and PM-ST induced significantly higher
formation at 10/yg/mL. PM-Q and PM-G induced
small but significant decreases in absorption at
100/yg/mL. Both PM-ST and PM-S had significant
dose responses for all concentrations tested. No
difference was observed between PM-G and PM-
Q. PM-S and PM-ST pretreated with DFX had a
lower ability to induce ROS formation.
Endotoxin Content: Only PM-ST showed positive
results for endotoxin content.
July 2009	D-56	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Liu et al.
(2005, 088304)
2005
Species: Human
Cell Type: HPAECs
SE: Wood Smoke Extract
(Generated using a stainless steel receptacle
containing 10Og of dry wood dust)
Particle Size: NA
Route: Cell Culture
Dose/Concentration: 40 //gfmL
Time to Analysis: 0-4h;
Mitochondrial Membrane
Destabilization: 0-60min; DNA
Defragmentation: 0-6h;
Cytotoxicity: 24h
Viability: SE exposure reduced cell viability dose-
dependently. Reduction reached —38% of
control.
Effect on Oxidative Stress/ Antioxidant
Enzymes: SE caused an increase in ROS levels, in
particular 02- and H202 in a time-dependent
manner. Exposure to SE for up to 4 h caused a
decrease in GSH levels in a time-dependent
manner. Increased expression of Cu/Zn SOD
mRNA and HO-1 mRNA was observed. Catalase or
GPx mRNA expression was unaffected.
Upregulation of Cu/Zn SOD and HO-1 occurred in a
time-dependent manner
Mitochondrial Translocation/ Capsase-
Independent Apoptosis/DNA fragmentation:
Exposure for up to 60 min caused an increase in
the percentage of annexin V-FITC-pos cells but not
Pl-pos cells. At 4h, FDA-pos cells was unaffected.
SE exposure caused a loss of mitochondrial
membrane potential (indicated by the change in
JC-1 fluorescence). Cytosolic bax levels increased
after exposure for 1 or 2h and returned to basal
level at 4 h after exposure. Levels of procaspase-3
and caspase-9 were unaltered by SE exposure
after 4h. Procaspase-3 increased and caspase-9
decreased by H202 exposure. SE exposure
increased levels of AIF and EndoG (exposure up to
4 h). At 6h, increased DNA defragmentation was
observed. Pre-treatment with caspase inhibitors
(CMK and Z-VAD-FMK) failed to suppress SE-
induced apoptosis.
NAC: Treatment with NAC prevented ROS
increase in cells exposed to SE for 60 min. NAC
addition prevented the reduction of GSH by SE.
NAC decreased nuclear levels of AIF and EndoG
and completely reduced DNA-fragmentation. NAC
alleviated the SE-induced reduced viability. GSH
and DNA fragmentation were unaffected by NAC.
Reference: Long et
al. (2005, 087454)
Species: Human
Cell Types: Human,
Peripheral blood
mononuclear cells
(PBMCs)
differentiated into
MDMs (90-95 %
CD14+) and T
lymphocytes
Synthetic C and C/Fe particles (phenol and
paraformaldehyde polymers on a zeolite
template)
C/Fe analysis Al 1.38 %, Si 0.33 %, Fe 0.46%
Particle Size: 1 //m
Route: Cell Culture (5 x 10e cells
(2 mL /well) MDMs
Dose/Concentration: 5 //gjcnr
Time to Analysis: 2-24h
ROSs release: Oxidative burst form C/Fe maxes
out at 20 min with no effect from C particles.
Cellular particulate actions: C particulates were
present within lysosomes with small clumps
forming after 24 h outside of lysosomes with no
evidence of organelle lysis and/or agglomeration.
C/Fe particulates showed similar initial effects
progressing at 24-h total organelle lysis extending
to the outer cell membrane.
T cell effects: No effects from C or C/Fe particles
Medium Effect: Particle agglomeration appears to
be a direct result of serum present within a
cellfree medium
Hydroxyl radical formation: C/Fe particles
showed an order of magnitude of higher hydroxyl
formation as compared to C particles
July 2009
D-57
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Ma et al.
(2004,088417)
Species: Mouse
Cell Line: JB6P +
(Epidermal Cell Line)
DEP: SRM 1975
Particle Size: 0.5 /jm
Route: Cell Culture
Dose/Concentration: Non-
cytotoxic: 5,10, 20 /yglmL;
Cytotoxic: 0,10, 20, 40, 80,100,
160/yg/mL
Time to Analysis: 24,48h;
NF-kB and AP-1:12h
Phosphorylation of Akt: 5- 120
min.
Effect of LY294002 on DEP: Cells
pretreated with LY294002 (0 or
10//M) for 30 min and then
exposed to DEP for 0-60min.
Viability: Below 20/yglmL, DEP had no effect. At
concentrations greater than 20 /yglmL, DEP
caused apoptosis.
NF-kB and AP-1: DEP stimulated NF-kB activity
at 5 and 10 /yglmL. At 20 /yglmL, NF-kB activity
decreased, but was still greater than the control.
DEP had no effect on AP-1 activity.
P13K/Al
-------
Study
Pollutant
Exposure
Effects
Reference: Matsuo,
M.
Shimada, T.
Uenishi, R.
2003
Species: Human
Cell Type: NHBE,
NHPAE, TIG-1, TIG-7
(normal human lung
embryonic fibroblasts)
DEP: prepared at National Institute for
Environmental Studies (Tsukuba, Japan)
RDEP: residual DEP (after sequential
extraction with hexane (NOS), benzene,
dichloromethane, methanol, 1N ammonium
hydroxide)
Particle Size: 0.4//m (MMAD)
Route: Cell Culture (NHBE: 5x10"
cells/cm2; NHPAE: 3x103
cells/cm2; TIG-1 and TIG-7: 3x103
cells/cm2; Apoptosis: 2x106
cells/cm2; ROS/NO: 2x10"
cells/cm2; Cytoxicity Modulating
Agent: 3x10" cells/cm2; GSH:
3x10" cells/cm2)
Dose/Concentration: 25, 50,
100,200,300, 400, 500 //g/mL
Time to Analysis: 1h
Cytoxicity in NHBE: Both DEP and RDEP
exhibited dose-dependent cytotoxicity at
concentrations beginning from 50 //g/mL and
higher. RDEP was less cytotoxic than DEP. DEP
exposure resulted in necrosis, not apoptosis.
Comparative Cytotoxicity: The order of LCeo
values (50% lethal concentration) was: NHBE
(118//g/ml), NHPAE (137 //g/ml), TIG (270
//g/ml). NHBE's susceptibility was higher than the
susceptibility of NHPAE and TIG cells.
ROS/NO: DEP induced dose-
at 25 and 50 //g/mL.
Reduced Glutathione: DEP induced dose-
dependent decreases. At 200 or 300 //g/mL, GSH
levels decreased by 55.2 or 97.3%, respectively.
Antioxidant Effects: Various antioxidants either
decreased DEP cytotoxicity (PMC, Ebselen, EUK-8)
or had no effect on DEP cytotoxicity (SOD,
catalase, GSH, o-tocopherol)
Chelating Agents: DEP became less cytotoxic
when lon-chelating agents were preincubated for
24h. No effect on DEP cytotoxicity was observed
when chelating agents were administered to cells
immediately after sonication.
Endocytosis inhibitors: Decreased DEP toxicity
was observed in a dose-dependent manner.
Reference:
Matsuzaki, T.
Amakawa, K.
Yamaguchi, K.
2006
Species: Human
Cell Type: Peripheral
neutrophils
Gender: Male and
Female
Age: 20-40yrs
DEP: generated from a 4JB1 -type, 4 cyl Isuzu
diesel engine
me-DEP: methanol extract of DEP (40 % of
DEP by dry weight)
Particle Size: 0.4 //m
Route: Cell Suspensions (5 X106
cellsfmL)
Dose/Concentration: all me-DEP
f-actin: 1, 5,10//g/mL
CD11 b: 5,10, 30 //g/mL
IL-8: 5,10, 30 //g/mL
H202:5,10, 30, 60//g/mL
MMP-9, LTB-4: 5,10, 30, 60
//g/mL
Time to Analysis: f-Actin: 15min
CD11 b: 2h
IL-8: 2 or 24h
H202: 30min
MMP-9, LTB-4: 2 or 24h
F-Actin: Treatment with me-DEP showed a dose-
dependent increase in the f-actin content of
neutrophils and this increase was significantly
higher at 5 and 10//g/mL.
CD-11 b: Treatment increased CD-11 b expression
two-fold at 30 //g/mL.
IL-8: Minimal response was observed after 2h. A
significant increase was observed (243%) at 24 h
with 30//g/mL.
LTB-4: At 2h, LTB4 increased to 115% and 119%
with 30 and 60 //g/mL me-DEP respectively. At
24 h with 60 //g/mL me-DEP, LTB-4 increased
to153%.
H202: Exposure to 30 and 60//g/mL of me-DEP
induced large dose-dependent increases of 563%
and 1220%, respectively.
MMP-9: A significant increase at 2 and 24 h were
observed. In both exposure periods, 30 //g/mL
induced larger increases than 60 //g/mL.
Reference: Molinelli,
A.R.
Santacana, G.E.
Madden, M.C.
2006
Species: Human
Cell Type: NHBE,
BEAS-2b
(transformed
bronchoepithelial
cells)
PMH: PMio extracts in hexane
PMA - PMio extracts in acetone of residue
after hexane extraction
¦G: Guaynabo(Urban) and
¦F: Fajardo (Preservation Area),
PR, USA
1999
Particle Size: PMio
PMio extracts collected by Puerto Rico
Environmental Quality Board
Route: Cell Culture (3x103
cells/well)
Dose/Concentration: NHBE
exposed to 0-100//g/mL of PMio
BEAS-2B exposed to 10,100, 250
//g/mL of PMio
Time to Analysis: 48h
Metal analysis: Hexane extracts Cu, V, Ni all
higher in winter than summer. For hexane extracts
within the same season, metal concentrations
were higher in the Fajardo extracts. On the other
hand, the acetone extracts from Guaynabo
generally had higher metal concentrations than
Fajardo.
Cytotoxicity NHBE: The order of most to least
toxic for PM extracted with hexane is: winter-
G, winter-F, summer -G , summer-F. The order of
most to least toxic for PM extracted with acetone
is: summer-G, summer-F, winter-g.
Cytotoxicity BEAS-2: For PM extracted with
hexane, the cytotoxicity order is: winter-G, winter-
F, summer-G, summer-F. The order for acetone
extracted PM is: summer-G , summer-F, winter-F,
winter-G. Effects trend similar to metal levels (no
analysis). Summer extracts showed linear dose-
response curves. Winter extracts exhibited more
equivocal results, especially for Fajardo. Results
suggest that NHBE cells are more sensitive than
the BEAS-2B cells to PM extracts.
July 2009
D-59
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Moller et
al. (2002, 036589)
Species: Canine,
Mouse
Cell Type:
Dog Alveolar
Macrophages (BD-
AM), J774A.1 (from
cell line BALB/c/NIH)
fTi02 (origin NR)
ufTi02 (origin NR)
ufP-G: carbon black (Printex-G, Degussa,
Frankfurt, Germany)
ufP90: carbon black (Printex90, Degussa,
Frankfurt, Germany)
ufEC90: elemental carbon (produced by
electrical spark generator under standardized
conditions with low impurities)
DEP (SRM 1650)
UrbD: Urban Dust (SRM 1649a)
Particle Size: (in diameter) Ti02: 220nm;
ufTi02: 20nm; ufP-G: 51 nm; ufP90:12nm;
ufEC90: 90nm; DEP: 120nm; UrbD: NR
Route: Cell Suspension
Dose/Concentration: 10, 32,
100, 320/yg|mL
Time to Analysis: 24h
Cytoskeleton of J774A.1: At doses of 32//gfmL
or less, the particles did not significantly influence
relaxation and stiffness. fTi02and ufP90 had no
effect at any dose. ufTi02 at 320 /yglmL induced
retarded relaxation and significant stiffening.
ufEC90 induced dose-dependent retardation of
relaxation and increased stiffening. DEP and UrbD
induced similar results.
Cytoskeleton of BD-AM: ufTi02 and fTi02 both
induced some retarded relaxation and increased
stiffening at 100 /yglmL dose. ufTi02 appears to
increase stiffening in a dose-dependent manner.
ufEC90 induced dose-dependent acceleration of
relaxation due to the carbon content of ufEC90.
DEP also induced acceleration of relaxation as
well as a decrease in stiffness.
Phagocytosis: At 24 h, ufTi02 and fTi02
significantly reduced phagocytotic ability in
J774A.1 but not in BD-AM. All carbonaceous
particles induced significant impairment in
J774A.1. All ultrafine carbon particles inhibited
BD-AMs.
Cell Proliferation: ufTi02 significantly inhibited
proliferation compared to the control and fTi02 at
100/yglmL in J774A.1. ufEC90 and ufP90
inhibited proliferation slightly with ufEC90
inducing slightly greater inhibition than ufP90.
UrbD and DEP also significantly reduced
proliferating.
Apoptosis: All particles induced decreased
viability at 100 //gfmL in both cell types. With
ufTi02 inducing greater apootosis than fTi02,
ufEC90 than ufP90 and ufEC90 than ufP-G.
Reference: Mutlu GM
et al. (2006,155994)
Species: Human, Rat
Cell Type: A549,
Primary Alveolar Type
II Epithelial Cells
PMio
(Collected by baghouse from ambient air in
Dusseldorf, Germany) Particle Size: PMio
Route: Cell Culture
Dose/Concentration: 0.05, 0.5,
5. 50/yg/cm2
Time to Analysis: 24h
Na.K-ATPase Plasma Membrane Protein: PMio
induced a decrease of protein in the plasma
membrane of A549 cells. Total Na.K-ATPase
levels were unaffected.
ROS: Pretreatment with EUK-134, superoxide
dismutase and catalase mimetic, inhibited the
decrease of GSH. Furthermore, it attenuated the
decrease of NA.K ATPase in A549 cells.
I\IA, K-ATPase Activity: PMio induced a dose-
dependent decrease in ouabain-sensitive liberation
of 32P from AT32P in primary rat alveolar type II
cells. This effect was inhibited with pretreatment
with EUK-134.
Reference: Nam,
HR.Y.
Choi, B.HR.
Lee, J.Y.
2004
Species: Human
Cell Type: A549
PM2.6
Collected from hospital rooftop, Seoul, South
Korea
Particle Size: PM2.B
Route: Cell Culture
Dose/Concentration: 0.5,1,10,
25, 50 /yglcm2
Time to Analysis: 6, 24h
NFkB/IkBo: 50/yglcm2 DEP induced iKBa
degradation which peaked at 2h and recovered
after 4h. Treatment with increasing amount of
PM2.6 resulted in a dose-dependent decrease in
IKBa. PM2.E increased NFkB in a dose-dependent
manner up to 10 //g/cm2. NFkB induction peaked
at 12h.
IL-8: PM2.6 treatment increased protein level more
than 3 fold with 100/yglcm2 PM2.6. mRNA levels
also increased.
iNOS Inhibitor: PM2.6 induced IL-8 elevation was
completely blocked by iNOS inhibitor. iNOS
inhibitor also negated PM2.6 induction of NFkB
activity. Antioxidants and iNOS inhibitor reduced
PM-induced IkBq degradation.
Reference: Nozaki Jl
et al. (2007, 097862)
Species: Mouse
Cell Line: LA-4
(Alveolar Epithelial
Cells)
PM: PM Rooftop 5story, urban, Japan
PME: dichloromethane extract of PM filtered
P90: Printex 90 (carbon black) (Degussa)
Particle Size: PM: 0.22 /jm; PME: 2.5 /jm;
P90: 14nm
Route: Cell Culture (1.4x 104
cells/cm2)
Dose/Concentration: 1.1 //gfcm2
Time to Analysis: 24, 28, 72h
Cytotoxicity: P90 had no effect. PM and PME
were cytotoxic at similar levels.
Protein Expression: All particles affected protein
expression (no specific protein- 2D gel
electrophoresis).
July 2009	D-60	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Obot et
al. (2002, 042370)
Species: Mouse
Cell Line: BALB/c
Cell Type: AMs
PM: SRM 1648
PM-100: PM heated to 100 C
PM-500: PM heated to 500 C
PM-PH: PM acid digestion
PMAC: Acetone extraction
PMCH: Cyclohexane extraction
PMH20: Water extraction
All extract fraction used as residual particles
Particle Size: NR
Route: Cell Culture (5 X10!l
cells.i'niL)
Dose/Concentration: PM: 200
//g/mL; PM-100: 188//g/mL; PM-
500: 130 mg/l; PM-PH: 94//g/mL;
PMAC: 173//g/mL; PMCH: 171
//g/mL; PMH20: 188 //g/mL
Fraction doses adjusted for mass
loss during fraction treatment
Time to Analysis: 4h
Cytotoxicity: All 7 fractions had cytotoxic
effects. PM had highest cytotoxicity. PM-500,
PM-PH, PMAC less toxic than PM.
Apoptosis: All 7 fractions significantly increased
apoptosis. The PM fractions that induced the
greatest apoptosis in descending order are: PM,
PMH20, PM-100, PM-500, PMAC, PMCH and
PM-PH. PM-induced apoptosis (only PM, PM-500
and PMAC tested) was blocked by poly I or 2F8
antibody (scavenger receptors).
Particle Characterization: Untreated PM and
PM-100 did not have measurable amounts of
transition metals on its surface. Measured
components include carbon, 02, N, S, Si, Ca, A I, P,
CI. PM-PH mostly contained 02 and Si. PM-500
had increased 02, Si compared to PM and
measurable amounts of Na, K„ Zn, Co, Pb, Fe.
Included increased surface density of S, P, Al.
PMCH lacked nonpolar organic compounds.
Reference: Obot et
al. (2004, 095938)
Species: Mouse (7-
9wks), Human
Cell Line: Mouse-
BALB/c
Cell Type: AMs
PM: SRM 1648 (collected by bag-hosue in St.
Louis, M0).
PM-100: PM heated to 100n C
PM-500: PM heated to 500n C
PM-PH: PM acid digestion
PMAC: Acetone extraction
PMCH: Cyclohexane extraction
PMH20: Water extraction
All of the 6 extract fractions from PM1648
PM2.6: Collected in Houston, TX
Particle Size: PM1648: NR; PMz.b
Route: Cell Culture (5 X10E
cellsi'niL)
Dose/Concentration: PM: 200
//g/mL; PM-100: 188//g/mL; PM-
500: 130 mg/l: PM-PH: 94//g/mL;
PMAC: 173//g/mL; PMCH: 171
//g/mL; PMH20: 188//g/mL
Fraction doses adjusted for mass
loss during fraction treatment
PM2.6 - 50,100,150, 200 //g/mL
Time to Analysis: Mouse-4h;
Human-24h.
Human AM Viability: Only PM, PM-100, PMAC
and PMH20 decreased viability.
Human AM Apoptosis: PM, PM-100 and PMH20
increased apoptosis. PM induced greater apoptosis
than PM-100 and PMH20.
Regression Analysis Mouse vs Human:
Although individual fractions differed somewhat,
cell viability and apoptosis of all 7 fractions
showed linear regression
Human and Mouse AM Viability with PM2.5:
Nearly identical dose-dependent decrease was
exhibited starting at 50 //g/mL
Human and Mouse AM Apoptosis with PM2.5:
Nearly identical dose-dependent increases were
exhibited with human AM responses peaking at
150 //g ;inL and declining at 200 //g/mL (no mouse
data for 200//g/mL).
Regression Analysis with PM2.5: Excellent
correlation of mouse and human responses for
viability and apoptosis was exhibited.
Reference: Okeson et CG: Coal ash, Germany
al. (2003, 042292)
Species: Rat
Cell Type: RLE-6TN
(Type II Alveolar
Epithelial Cells)
CU: Coal ash, USA
5C: PM # 5 Oil fly ash coarse
5F: PM #5 Oil fly ash fine
6MSC: PM #6 Oil med sulfur fly ash coarse
6HSC: PM # 6 Oil high sulfur fly ash coarse
6HSF: PM # 6 Oil high sulfur fly ash fine
Particle Size: CG, CU: NR; 5C, 6MSC, 6HSC
> 2.5 //m; 5F, 6HSF < 2.5 //m
Route: Cell Culture
Dose/Concentration: Coal Fly
Ash 12.5, 25,50,125, 250
//g/mL
Oil Fly Ash ¦ 100//g/mL
Time to Analysis: 24h
Oil PM Characterization: Generally, the fine
fractions had higher metal levels than the coarse
fractions except for Zn. High sulfur had a higher
metal content than med sulfur. Carbon percent
weight was stable across all 5 fractions.
Coal Ash Cytotoxicity: CG treatment exhibited
similar cytototoxic results as CU. Cytotoxic
effects were exhibited at concentrations of 12.5
//g/mL and above. Effects remained steady at
concentrations above 50 //g/mL.
Oil Ash Cytotoxicity: Cytotoxic effects were
induced by all. The order of PM fractions inducing
the most cytotoxicity to the least is the following:
5F, 6HSF, 6HSC, 5C, 6MSC.
Correlation of Metal Content and
Cytotoxicity: Fe, V showed a reasonable
correlation. Zn had no correlation.
Cell Metabolism: An inhibitory effect was
observed with 100 //g/mL coal ash after 6h. After
12h of exposure, CU, unlike CG, does not continue
to inhibit cell metabolism. Oil ash was generally
less effective than coal ash. The order of PM
fractions inhibiting metabolism the most to the
least is the following: 5F, 6HSC, 5C, 6MSC. 6HSF
not tested.
July 2009	D-61	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Okeson et
al. (2004, 087961)
Species: Rat
Cell Type: RLE-6TN
(Type II Alveolar
Epithelial Cells)
Zn, V, Fe chloride as salts (valence state not
reported i.e., Fe II or Fe III)
Particle Size: NR
Route: Cell Culture (50000
cells/well)
Dose/Concentration: 0.001,
0.01,0.1,1.0,10 mM
Time to Analysis: 24h
Cytotoxicity: All metals cytotoxic at
concentrations greater than 0.1 mM. V is 5 times
less cytotoxic than Zn, and Fe is 7 times less
cytotoxic than Zn with a ECbo of 3mM and 4mM,
respectively. At 10 mM of each metal, no
surviving cells were present.
NCS: Incubation with NCS (5 or 10 %) decreased
toxicity of Zn, especially at 0.1 mM, but had no
effect on Fe or V toxicity.
Albumin: BSA decreased Zn toxicity at equivalent
concentrations but to a lesser extent than NCS.
Reference: Osornio-
Vargas et al. (2003,
052417)
Species: Mouse
Cell Line/Type:
J774A.1, L929
(Mesenchymal Cells)
PMio
PM2.6
¦N - Northern (industrial)
¦SE - Southeastern (lake basin dust) sites,
both heavy vehicular traffic, Mexico City,
Mexico
Particle Size: PM10; PM2.6
Route: Cell Culture (J774A.1:
15000 cells/cm2; L929: 30,000
cells/well)
Dose/Concentration: 20, 40, 80
/yg/cm2
Time to Analysis: 24-72h
PM Characterization: Elements similar in particle
types with elements in PM10 more abundant.
Northern particles contained more Co, Zn, Ni, Pb.
Endotoxin: All PM samples had detectable
amounts of endotoxin. PM2.6 -N had 22 EU/mg.
PMio-N had 30 EU/mg. PM2.6 -SE had 12 EU/mg.
PM10-SE had 59 EU/mg.
Cytotoxicity (J774A.1): The two northern
samples, PM2.6 and PM10, both induced similar
cytotoxic effects at 40% survival. PMio-SE and
PM2.6 SE induced dose-dependent responses. In
general, the northern samples had a higher
cytotoxic effect than the southern samples.
Apoptosis (J774A.1): Northern samples induced
more apoptosis than did the southeastern samples.
There was no difference between PM10 and PM2.6
induced apoptosis.
TNF-a and IL-6 (J774A.1): TNF-a and IL-6
induced dose-dependent increases. At 80 /yg/cm2,
PM in -SE induced the most production of IL-6
followed by PM2.6 -SE, PMio-N , and PM2.6 -N.
J774A.1 Supernatant Toxicity (L929):
Conditioned medium from J774A.1 pre-exposed to
each PM type reduced cell viability in L929 cells.
This was correlated with TNF-a level in
supernatants.
Reference: Pei,
X.HR. Nakanishi, Y.
Inoue, HR. 2002
Species: Human
Cell Type: A549
B[a]P
1-NP: 1 -nitropyrene
Both purchased from Aldrich Chemical Co.,
Ml, USA.
Particle Size: NR
Route: Cell Culture
Dose/Concentration: B[a]P:
10//M; 1-NP: 5/ymol
Time to Analysis: 4, 8, or 24h
IL-8: IL-8 mRNA increased by B[a]P and 1-NP in a
time-dependent manner with 1-NP increasing
higher levels than B[a]P at half the molarity.
NK-I10um
BDS-W: solvent washed
Graphite
Composition:
< 2.5um - 92%
2.5- 10um - 5%
>10um - 3%
Particle Size: BDS-P1: < 2.5/jm; BDS-P2:
2.5-10//m; BDS-P3: >10//m
Route: Cell Culture (1 -1.5x106
cells)
Dose/Concentration: 3mg BDS
Time to Analysis: 5min-72h
Particle Characterization: By weight, elemental
carbon makes up 94% of BDS, hydrogen 2%,
nitrogen and sulfur 1%, and oxygen less than
0.1%.
PAH Components of BDS: 13 prominent PAHs:
acenaphthylene, fluorene, anthracene,
cyclopentaphenanthrene, fluoranthene,
acephenanthrylene, pyrene, benzofluorenes,
acepyrene, chrysene, benzopyrenes, perylene,
benzoperylene.
BDS Activity: At 60-120min, BDS was observed
in the cells. At 4h, fluorescence observed in
cytoplasmic vesicles and increased during the first
24 h then plateaued for the next 72h. BDS-W
appeared in vesicles sooner than BDS.
July 2009	D-62	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Pozzi et al. (2005,
088610)
Species: Mouse
Cell Type: RAW
264.7
PM: Collected continuously for 15 days, 8-
10m from street, Sept 1999, Rome, Italy
¦F - Fine particulate
¦C - Coarse particulate
CB (Degussa Huber NG90)
Particle Size: PM-F: 0.4-2.5 fjm; PM C: 2.5-
10 /jm; CB: 200-250nm
Route: Cell Culture (1.3x10E
cells/well)
Dose/Concentration: 30 //gfmL;
14/yg/cm2
120/yg/mL; 54/yg/cm2
Time to Analysis: 5, 24h
Cytotoxicity: For 24h, lower levels of PM-F, PM-
C, and CB had no effect on cell viability. Higher
levels of PM C and CB induced a significant
release of LDH.
Arachidonic Acid (AA): Both fractions of PM
increased AA release in a dose-dependent manner
at 5h. CB increased a release only at the higher
concentrations although, in terms of magnitude,
the CB-induced release was much less than the
ambient PM-induced release. Pretreatment with
deferoxamine was not effective in decreasing AA
release.
TNF-a: TNF-a levels increased significantly for
both concentrations and time periods for PM. PM-
C at 24 h was significantly lower than at 5h for
both concentrations. PM-C at 30//gfmL induced a
much greater TNF-a release than PM-F at 5h.
IL-6: PM-F significantly increased at 5h for both
concentrations. Elevated IL-6 levels were exhibited
at both PM-C doses at 24 h. At 5h, only the high
dose elevated IL-6 levels. CB was devoid of an
effect on IL-6. LPS-induced IL-6 response was
similar to coarse PM at the high dose, with the
response being greater at 24 h than at 5h.
Reference: Prophete Ambient PM2.6
et al. (2006,156888) NYC. st and 26 Stj NYc
Species: Rat
Cell Type: NR8383
AMs
LA: San Gabriel foothills, Claremont, CA
SEA: 15th Ave S and S.Charleston, Seattle,
WA
V, Mn, Al, Fe levels in PM
added metals to cells
V: Na3V04
Al: AICb.6H20
Mn: MnCl2.4H20
Fe: FeCh.6H20
Particle Size: PM2.6
Route: Cell Culture (2 X10!l
cells.i'niL)
Dose/Concentration: Fe(lll) 16
/ymol
V, Mn, and Fe(lll) mixtures with V
or Mn in molar ratios 0.02, 0.08,
0.2 and 0.4 X Fe(lll)
Al and Fe(lll) mixtures with Al in
molar ratios 0.37, 0.75, 2, 7.5 X
Fe(lll)
Time to Analysis: 20h
Particle Characterization: Fe and metal to F
ratios based on ratios observed in PM2.6 from LA,
SEA and NYC sites. V: Fe ratios remarkably similar
among sites. Fe levels fixed at NYC level of 16/ym
(highest).
IRP: Coexposure with 3 metals increased IRP
binding activity relative to Fe III alone, by up to
3.5 fold for Al (1.5-3 ratio), 2 fold for Mn (0.08-
0.2 ratio) and 7 fold for V (0.2 ratio). IRP activity
dropped at higher ratios. A drop in IRP activity at
higher ratios may be result of cytotoxicity for Al,
but not for V and Mn.
iNOS: Al induced iNOS expression dose-
dependently. There was no observed effect for Mn
and V.
Induction of Hypoxia-inducible Factor (HIF-
1a): Only V and Al induced HIF-1a.
Activation of ERK1 and -2: V and Al induced
pERK1, but only V induced pERK2. Mn had no
increasing effects, but data indicated a decreasing
induction.
Reference: Ramage
and Guy (2004,
055640)
Species: Human
Cell Type: A549
PM10: Collected in Wolverhampton, UK
ufCB: Ultrafine Carbon Particles
(Origin not reported)
Particle Size: PM10, ufCB: < 100 nm
(diameter)
Route: Cell Culture
Dose/Concentration: 80//gfmL
Time to Analysis: 0, 0.5, 3, 6,
18h
CRP: Treatment with ufCB or PM10 produced an
increase in CRP expression with similar effects
noted after 6h. PM10 induced greater increases
than ufCB. Both the cytoplasm and nucleus
contained CRP.
Hsp70: PM 11 and ufCB induced increased levels at
all time points with ufCB inducing greater levels
than PM10. Hsp70 expression was observed in the
cytoplasm and nucleus.
Antioxidants of CRP and Hsp70: Coincubation
of ufCB with Nacystelin and Trolox caused a small
reduction in CRP and Hsp 70.
Reference: Rao et al.
(2005, 095756)
Species: Rat
Strain: Sprague-
Dawley
Cell Type: AMs and
cultured lung
fibroblasts
DEP: SRM 2975 (NIST)
Particle Size: 0.5/ym
Route: Cell Culture
Dose/Concentration: 200//gfmL
Time to Analysis: Measured
after 4h incubation.
mRNA Expression: No change in IL-1 (3 or iNOS
were observed. Data suggests that the lung
fibroblasts is the main source of IL-6 and MCP-1 in
BAL fluid because of their comparatively high
message levels. Due to the extreme variability in
results, the cause of an increase on co-culture
with AMs and/or DEPs was not assessed.
July 2009
D-63
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Reibman
et al. (2003,156905)
Species: Human
Cell Type: HBEC,
BEAS-2B
UFPM: Ultrafine PM
FPM: Fine PM
IPM: Intermediate PM
CPM: Coarse PM
CB: Carbon black
All PM collected 8th floor, 26th St and 1st
Ave, New York City, NY
Particle Size: UFPM: <0.18 fjm; FPM: 0.18
¦ 1.0/jm; IPM: 1.0 ¦ 3.2/jm; CPM: > 3.2/jm
Route: Cell Culture
Dose/Concentration: 11 //gfcnr;
100/yg/mL
Time to Analysis: 6,18h
Cytotoxicity: After treatment, cells were more
than 90% viable. UFPM and FPM caused no gross
alterations in cell morphology or adhesion.
MIP3a/CCL20 mRNA (6h): Stimulation of mRNA
released by HBEC upon exposure to UFPM
appeared similar to that provided by TNF-a (5
/yg/mL) and IL-1 p (10 mg/mL).
MIP30/CCL2O protein in HBEC |18h): TNF a
and IL-1 (3 induced a dose-dependent increase in
MIP3a/CCL20 protein (0-10 ng/mL), whereas II-4
and IL-13 induced MIP3afCCL20 protein release
that reached maximum levels at 1 ngfmL. No
release of MIP1 a/CCL3 nor RANTES/CCL-5 was
observed upon stimulation with cytokines.
Secretion of MIP3a/CCL20 in response to PM
(18h): All PM fractions less than 2.5 /jm resulted
in the release of MIP3afCCL20 protein in HBEC
roughly equivalent amounts. CB similar in size to
UF/fine PM did not result in the release of
MIP3a/CCL20, nor did LPS (0.01-1.0//gfmL). No
release of MIP1 a/CCL3 nor RANTES/CCL 5 was
observed upon stimulation by PM fractions.
Activation of MAPK (ERK1/2 and p38): ERK1/2
and p38 was activated by TNF-a, IL-1 (3, IL-4 and
IL-13 within 15 min and was sustained for at least
60 min. Erk1f2 and p38 inhibitors reduced
MIP3a/CCL20 release in BEAS-2B cells in
response to cytokines.
Reference: Riley et
al. (2003, 053237)
Species: Rat
Cell Type: RLE-6TN
(Type II Alveolar
Epithelial Cells)
Metabolism Inhibition Time Course Response
(Cu and Zn only): At high (1 mM] concentrations,
Zn toxicity peaked at 36-48h followed by a 2-fold
recovery by 72h. Cu showed a faster, steady
decline plateauing after 36h. At low
concentrations (0.1 mM), Cu showed a steady
slow decline. At 48h, Zn decreased faster to max
activity and returned to control by 72h.
IL-6 Secretion: Zn and Cu both decreased IL-6
secretion. Decreases were very similar for both
metals and concentrations when expressed as
secretion per viable cell ratio except for Zn at 1.0
mM.
Metal Combinations: Zn and Cu gave variable
results. Zn protected against V cytotoxicity. Zn
and Cu had an additive response. Zn did not affect
Fe toxicity.
Zn: ZnCI2
Cu: CuCI2
Fe: FeCI2
V: VCI4
Ni: NiCI2
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 0.01, 0.1,
1.0,10 mM
Time to Analysis: 2, 4, 24, 72h
Cytotoxicity (SDH): All particles were cytotoxic
in a dose-dependent manner. Zn and V were
cytotoxic at 0.05 mM, Cu at 0.5mM, Ni at 0.8mM
and Fe at 2 mM. For Zn, cell death (LDH) had a
biphasic response: a slow logslope until approx
0.1 mM at which point it rapidly accelerated to a
peak at 5mM with a small decline at 10mM. Most
of Zn cytotoxicity was not due to apoptosis. LPS
did not affect either Zn or Cu cytotoxicity.
July 2009
D-64
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Riley et
al. (2005, 096452)
Species: Rat, Human
Cell Type: RLE-6TN,
NR8383 Alveolar
Macrophages, A549
Fe: FeCI2
Ni: NiCI2
Cu: CuC12
V: VCI2
Particle Size: NR
Route: Cell Culture (5 x104
cells/well Alveolar Cells; 1.2x10E
cells/well NR8383)
Dose/Concentration: AMs: 0.02,
0.05, 0.07, 0.08 mM;
RLE-6TN: 0.1,0.2, 0.6,1.0,6.0
mM;
A549: 0.5, 0.8, 4.4, 4.8 mM
Time to Analysis: 2-48h
Relative Sensitivity of Cell Strains to Metal
Chloride: NR8383 was more sensitive than RLE-
6TN and A549 except for V where
NR8383 and RLE-6TN were both more sensitive
than A549.
Relative sensitivity of Cell Strains to Metal
Chloride vs Sulfate: With the exception of Cr,
sulfate was generally more cytotoxic than chloride
(note V valence state).
A549 Cytotoxicity Time Course: Zn cytotoxicity
takes 24 h to develop whereas Cu cytotoxicity
develops within 2 h. LDH release for Cu, however,
develops in 24h.
RLE Cytotoxicity Time Course: Zn starts at 2 h
and develops until 24 h. Cu develops within 2 h
and continues until 24 h where it is less toxic than
Zn. Both release equivalent amounts of LDH after
24 h.
NR8383 Cytotoxicity Time Course: Both Zn and
Cu exhibit time dependent toxicity beginning as
early as 4h. LDH release maximizes at 12 h and
either remains steady or declines.
Reference: Ritz, S.A.
Wan, J.
Diaz-Sanchez, D.
2007
Species: Human
Cell Type: BEAS-2B,
NHBEC
DX: Extract of DEP (generated from a light
duty four-cylinder diesel engine 4JB1 type
Isuzu Automobile)
Particle Size: < 1//m (diameter)
Route: Cell Culture
Dose/Concentration: 0, 20, 50,
100/yg/mL
Time to Analysis: 24h
NQ01 (Sentinel Phase II Enzyme): Cells
transfected with NQ01 reduced induction of IL-8
by DX exposure.
Sulfurophane: Increased gene expression of
phase II enzymes, particularly NQ01, was
observed in both cell types. Gene expression in
BEAS-2B was greater than that of NHBEC.
Sulfurophane did not upregulate GSTM1 in BEAS-
2B but induced a 2-fold increase in NHBEC.
Pretreatment also inhibited DX-induction of IL-8 in
both cell types.
Cytokines: DX induced significant increase of IL-8
in both cell types at concentrations of 10 //gf'niL
or higher. GM-CSF and IL-8 remained unaffected in
BEAS-2B. GM-CSF and IL-8 Increased in NHBECs
and reached statistical significance at 25 //g/mL.
Reference: Rosas
Perez et al. (2007,
097967)
Species: Mouse
Cell Type: J774A.1
PMio
Collected in Mexico City, Mexico from
January-June, 2002
North: Iztacala, manufacturing industry;
Center: Merced, heavy traffic;
South: Ciudad Universitaria, residential
Principal Component Analysis of Air Pollution
Data:
Group 1:S/K/Ca/Ti/Mn/Fe/Zn/Pb (43% of
variance);
Group 2: Cl/Cr/Ni/Cu (16%);
Group 3: Endotoxins/OC/EC (14%).
For all 3 sites: Averages of Group 1 is
statistically different among the center, north
and south sites with the central site producing
the highest values. Group 2 is similar among
the sites and, for Group 3, the north had a
lower average than the center and south sites.
Particle Size: PMio
Route: Cell Culture (1.5x104
cells/cm2)
Dose/Concentration: 20, 40 or
80/yg/cm2
Time to Analysis: 72h
Cytotoxicity: Responses were dose-dependent;
there was no observed site interaction.
Cytotoxicity seems to be a result of of the
following components: S/K/Ca/Ti/Mn/Fe/Zn/Pb.
IL-6: Only the center site at 40 //gfcm2 induced an
increase. Induction of higher IL-6 levels seems to
be related to high values of S/K/Ca/Ti/Mn/Fe/Zn/Pb
and endotoxins/OC/EC.
TNF- a: Production was induced by all samples in
a dose-dependent manner. Similar to IL-6,
induction of higherTNF-a levels seems to be a
result of high values of S/K/Ca/Ti/Mn/Fe/Zn/Pb and
endotoxins/OC/EC.
p53: Only south PM had effect. Induction of p54
seems to depend on high levels of Cl/Cr/Ni/Cu and
low levels of S/K/Ca/Ti/Mn/Fe/Zn/Pb.
July 2009
D-65
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Sakamoto et al.
(2007, 096282)
Species: Human
Age: 58-82y
(Smokers)
Cell Type: HBEC
PMio: EHC-93 (Obtained from Health Canada,
Canata)
Particle Size: PMio
Route: Cell Culture
Dose/Concentration: 100, 300
and 500 /yglmL
Time to Analysis: Calcium
responses: up to 60min; cytokines:
6 or 24h
Intracellular [Ca2+]: [Ca] concentration slowly
increased, elevating after 10 and 30 min for 500
and 300 mg/mL, respectively. The response
plateaued at 35 min for 500 //gfmL.
Extracellular [Ca2*]: Starting at 20 min, the
removal of extracellular Ca decreased the PMio
response significantly. Calcium channel blocker
(10//M or 1mM) LaCI3 and (5mM) NiCI2
significantly blocked the PM-induced intracellular
Ca. Lacl2 administration (1 mM] inhibited the PM-
induced Ca2+ response in a dose-dependent
manner.
Mode of Action: Intracellular Ca induced by ATP
declined more slowly in the cells exposed by PMio.
This indicates that PMio blocks Ca clearance via
the calcium pumps.
Cytokines: PMio induced a dose-dependent
increase in cytokine mRNA levels and cytokines IL-
1 p, LIF, IL-8 and GM-CSF. Cytokine expression
was unaffected by the reduction of extracellular
Ca2+. Preincubation with the calcium chelator
reduced responses for IL-1 (3 and IL-8 but not LIF
or GM-CSF.
Reference: Salnikow
et al. (2004, 087469)
Species: Human
Cell Line: 1HAEo-
FeSOi
FeCIs
N i S 0i
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 0.25mM
Fe exposures also contained 60
/yg/mL apotransferrin
Time to Analysis: 24h
Cytotoxicity: Both Fe had no effect. NiCL2
caused marginal cytotoxicity (75%).
Hypoxic Stress: At 20h, NiSOi (at
concentrations of 0.25 or 0.5mM) induced NDRG-
1fCap43 protein production indicating hypoxic
stress. DFX and DMOG induced a similar effect.
IL-8: NiS04 induced IL-8 time-dependently for up
to 48h. At 48h, the increase was 6 + fold.
Coexposure (Ni + Fe) on Fe uptake: Fe III
uptake was greater than Fe II uptake. NiS04 had
no effect. Ni uptake was greater than Fe uptake
but was decreased by coexposure to Fe.
Coexposure also did not effect hypoxic stress.
Coexposure with Fe did reduce Ni-induced IL-8
production.
Reference: Salonen
et al. 2004 (2004,
187053)
Species: Mouse
Cell Type: RAW
264.7
PMio (urban traffic) Finland
(filter sonicated in water and methanol,
extracted with methanol)
Pooled as winter (W) spring I (SI) or spring II
(Sll) based on componentftime considerations
HVLI slit impactor
Particle Size: PMio: 0.12-10nm
Route: (2x106 cells/well)
Dose/Concentration: 15,50,
150,500,1000/yglmL
Time to Analysis: 0, 24h
Air quality parameters: Winter and spring I did not
differ.
Sll much lower PM2.6
PMio W - 18.6 ± 10.1; SI - 28.0 ± 5.5;
Sll - 20.3 ± 2.4/yg/m3
Metal data equivocal as well as highly variable
resuspension rates
Total PAHs: W - 303; SI - 233; Sll - 204
ngfmg PMio (ppbnflammation (IL-6, TNF, NO)
/Cytotoxicity: A dose-dependent increase was
observed for TNF, IL-6 and NO except for SI. The
IL-6 levels, of those particles exposed to SI,
decreased at 1000 /yglmL.
TNF, IL-6: SI - Sll > >W>control.
NO production: W> -Sl> - Sll
Cell Viability: W - SI - Sll toxic at 500 and
1000/yg/mL
Watersoluble vs Insoluble: TNF and IL-6 were
nearly entirely the result of insoluble components
of PMio. Cytotoxicity was driven by both soluble
and insoluble components.
Metal Chelation: The addition of metal chelators
did not modify IL-6, TNF or cytotoxicity
LPS inhibitor: Treatment with the LPS inhibitor
eliminated the IL-6 response and, perhaps, slightly
reduced the TNF response but not cytotoxicity
Hydroxy radicals: A dose-dependent induction of
hydroxy radicals and induction of hydroxy radical
lesions (at 500 and 1000 //gfm3) in the calf
thymus DNA were observed.
July 2009	D-66	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Samet et
al. (2003,113782)
Species: Human
Cell Type: A431
(Epidermoid Cells)
As: NaAS03
V: V0S04
Zn: ZnS04
(obtained from Sigma, St. Louis, MO)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 500//M
Time to Analysis: 20, 30 or
90min
EGFR Dimerization: Zn, V or As did not induce
EGFR dimerization in a cell free system i.e., no
direct crosslinking. Zn did not induce dimerization
in whole cells either.
Phosphorylation of EGFR: Zn induced
phosphorylation at 3 sites similar to EGF. As and V
had no effect.
EGFR Kinase Inhibitor: While EGF action was
blocked, Zn continued to induce phosphorylation
and was independent of EGFR kinase activity.
c-Src: Blocking of c-Src tyrosine kinase
(transactivator of phosphorylation) negated all Zn-
induced phosphorylation but only had a slight
effect on EGF stimulated cells.
ERK1/2 Phosphorylation: Zn increased levels of
ERK1/2. Pretreatment with EFGR kinase inhibitor
reduced both Zn and EGF effect. This effect was
not blocked by the c-Src blocker.
Reference: Santini et	DEP: PM2.6
al. (2004, 087879)	Collected adjacent to moderate traffic in
Species: Mouse	Rome, Italy
Cell Type: RAW	Particle Size: PM2.6
264.7
Route: Cell Culture (2.5 X10!l
cells.i'niL)
Dose/Concentration: 0.01, 0.1,
1.0/yg/mL
Time to Analysis: 24h
500 MHz results (no 1 /jg/mL): DEP induced a
dose-dependent increase in choline compounds, a-
and fBgamma- glutaminefglutamate (0.01 >0.1
/yg/mL), lactate, and CH2, CH3 moieties of fatty
acids. DEP decreased inositol and
(phosphoreatinine.
700 MHz results (no 1 /jg/mL): DEP induced
similar results, except a-, fBgamma-glutamine
were dose-dependent. Inositol showed no effect.
Taurine slightly increased. Results were confirmed
after eliminating biological interferences via
perchloric acid.
Growth Curves/Cell Cycle Analyses/Cell
Morphology: DEP had no effect.
Cytokines: IL-6 levels increased at 0.1 and 1
/yg/mL. TNF-a was unaffected.
Species: Mouse
Cell Type: RAW
264.7
Reference: Saxena et DEP: SRM 1650
al. (2003, 096986) co. Cmde 0rganic Extract of DEp
Fractionated into
asphaltene (pentane/hexane),
saturated hydrocarbon,
less polar (aromatic) hydrocarbon,
more polar (aromatic) hydrocarbon,
resins, residual (resins)
Particle Size: NR
Route: Cell Culture (2.5 X10'
cells.i'niL)
Dose/Concentration: DEP, CO 5,
10,15,20, 25 //g/mL
IFNy 10 ng/mL
LPS 1 mg/mL
Time to Analysis: 1 -3d
Cytotoxicity: No cytotoxic effects were
observed.
NO: DEP alone induced NO in a dose-dependent
manner which peaked after 1 d and plateaued for
days 2 and 3. IFNy + DEP showed dose- and time-
dependency. LPS + DEP showed no effect at 1 d,
but dose-dependently reduced NO production on
day 2 and 3. Addition of BC G eliminated the
effect of DEP at 2d but showed a dose-dependent
decrease at 3d.
Effectiveness of Particulate Components: The
carbonaceous core of DEP did not affect BC G-
stimulated NO production. CO significantly
inhibited BC G-stimulated NO production. Study
indicated that the extract of aromatic
hydrocarbons and resins caused an inhibitory
effect in a dose-dependent manner.
July 2009	D-67	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Seagrave
et al. (2007, 097549)
Species: Human
Gender: Male (3
donors)
Age: 16, 23yrs
Cell Type:
Differentiated A549
DE: Generated by DE 5500 watt generator
using #2 certification oil performed under
5000w load. Emissions diluted to 3 mgfm3
total particulate matter.
Particle Size: 0.14 - 0.5 /jm
Route: Air Liquid Interface
Dose/Concentration: 8.33
mLfminfwell
Time to Analysis: 3h; 1 or 21 h
postexposure
Particle Deposition: 140 and 500 nm
microspheres demonstrated uniform deposition of
approx. 10%
Transepithelial Electric Resistance: No effect
of DE; rather, more effect was observed from air
controls
Macromolecular permeability: DE caused an
increase 1-h but returned to control at 21 h.
LDH/Cytotoxicity: DE had a highly variablefdonor
specific) effect at 1-h and returned to control
levels at 21 h
Mitochondrial activity (WST): DE reduced
activity at 1-h and possibly increased activity at
21 h (high donor-to-donor variability)
MucusLike Substance Excretion: There was
high donor to donor variability; no overall effects
were observed.
Alkaline Phosphatase (AP): DE decreased at 1-h
and perhaps increased at 21 h
Glutathione: DE caused a large decrease at 1-h
but returned to normal at 21 h.
HO-1: After DE exposure, levels increased but
were still lower than air exposed controls
Cytokines: No differences for IL- 8 or 12, TNFa,
GM-GSF, IL-1 a, or IFNy were observed. IL-4 and 6
were decreased upon DE exposure.
Reference: Seagrave DPM: SRM2975 (NIST)
et al. (2004, 087470) qpi\/|.q: 0p|\/| organic extract (acetone/DCM)
Species: Human CB. Carbon B|ack (E|f,ex.12, Cabot)
Cell Type. A549 Particle Size: CB: 37nm; Stokes diameter
198nm
Route: Cell Culture (1x106
cells/well)
Dose/Concentration: 0.03 -
1,000 /yglcm2
Time to Analysis: 0,18h
IL-8 release: DPM increased semi dose-
dependently (perhaps steady based on error range)
up to 1 //gjcnr after which IL-8 declined dose
dependently to zero (control - 100%) at 300 and
1000 /yglcm2. LDH release was steady which
indicates no cytotoxicity.
DPM interaction with IL-8: DPM depletes IL-8
from solution in a dose-dependent manner
(cellfree). BSA preincubation reduced the slope of
the dose response but not the final result. CB has
no effect.
DPM-0 residuals act identical to DPM.
Increasing NaCI concentrations reduced DPMs
depletion of IL-8
Neutrophil responses: DPM and bound IL-8
together caused a marked aggregation of cells
resulting in spindle shapes. DEM or IL-8 alone did
not cause this aggregation although DEP did
recruit neutrophils
July 2009	D-68	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Seagrave
et al. (2003, 054979)
Species: Human, Rat
Cell Line: F344jCrl
BR (mouse)
Age: 11 wks (mouse)
Weight: 250g
Cell Type: A549,
AMs
PM filter collection
Collected from diesel or gasoline powered
vehicles as follows:
BG: Black Smoke Gasoline
G30: Normal Emitter gasoline (30F)
G: Normal emitter gasoline (72F)
HD: High Emitter Diesel
D30: current technology diesel (30F)
D: current technology diesel (72F)
WG: White Smoke Gasoline
Particle Size: NR
Route: Cell Culture (1x106
cells/well)
Dose/Concentration: 0.03 -
10,000 /yglcm2
Time to Analysis: 16-18h
Cytotoxicity: LDH activity increased in A549
cells. The types of pollutants that are most toxic,
in decreasing order of cytotoxicity, are the
following: BG, G30, and G which are significantly
different from HD, D30, D, WG which are also
significantly different from DS. LDH activity also
increased in rat macrophages. G, G30, and BG
were the most toxic. HD and D30 were
intermediately toxic and D, WG, and DS were the
least toxic. In both cell types, gasoline was more
cytotoxic than diesel.
Cytokines: All particle types except DS increased
IL-8 levels in A549 though not all increases were
statistically significant. Also, many particle
samples at high concentrations produced an
apparent suppression of IL-8 release.
Alkaline Phosphatase: G30 and G were more
potent than the other particle samples in A549.
WG and D30 induced no significant effects. For
A549 cells, activity increased at low
concentrations and was suppressed at higher
concentrations.
Macrophage Peroxide Production: In rat AMs,
peroxide production was often the highest at the
lowest concentrations and the lowest production
caused by the highest concentrations. D30
followed this trend and induced the highest
production as well as the greatest suppression.
Using two different statistical methods, D30 > 6
others which in turn > DS. Using the second
method D30 and D > all other 6. Order of potency
between two methods completely different.
Authors noted that in vitro potency quite different
from in vivo potency (previous paper).
Reference: Seaton. PM2.6 from London
et al. 2005
Species: Human
Cell Type: A549
PMn from Manchester (positive control)
PM from Holland Park, Hampstead and Oxford
Circus stations (HP, HR and OC)
PM2.6 monitoring
Holland Park, Hampstead and Oxford Circus
PM had a median diameter of 0.4/ym. 80% of
the particles had a diameter less than 1 //m.
Particle Size: PM2.6, PM10, Holland Park,
Hampstead and Oxford Circus PM had a
median diameter of 0.4 fjm. 80% of the
particles had a diameter less than 1 //m.
Route: Cell Culture
Dose/Concentration: 1-100
/yg/mL
Time to Analysis: Cytotoxicity:
24h; IL-8: 8h; Generation of
hydroxy radicals: 8h
Cytotoxicity: Dust from all three tunnels (Holland
Park, Hampstead and Oxford Circus) were able to
cause cell death (LDH). The release of LDH
indicated a dose-dependent relationship. The
highest dose of Holland Park PM induced the
— 17% release of LDH, Hampstead triggered —
13% and Oxford Circus — 3% (no different than
control). PM10 from Manchester caused a 7% LDH
release at the highest dose. The negative control
(Ti02) caused no response (2% release at highest
dose).
IL-8: All three tunnel PMs induced a dose-
dependent release of IL-8. At the highest dose, all
three tunnel dusts induced IL-8 stimulation more
so than the control site PM2.6. HP induced a 3 fold
increase. Also, the highest Ti02 concentration
caused the least IL-8 stimulation.
Hydroxy Radical Generation/ DNA Plasmid
assay: The plasmid assay indicated that the
tunnel dusts induce more free radical activity than
the Manchester PM10 and Ti02.
HP nearly doubled the percentage of DNA damage
with intermediate results for HR and OC. Results
for PM10, Ti02 and control were identical
July 2009
D-69
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Singal et
al. 2005
Species: Human,
Mouse
Cell Type/Line:
A549Luc1 lung
adenocarcinoma
epithelial cell line
(human), MLE15Luc1
and MLE15Luc2
(mouse)
All cells contain
human cytokine IL-8
controlling firefly
luciferase
AE2: Aerosil 200, amorphous silica (Degussa)
CI: Carbon iron particles (25 % Fe)
Particle Size: AE2: 12nm surface area
-200 ± 25 m2/g: CI: ~40nm
Route: Cell Culture (5x106
cells/well)
Dose/Concentration: 18 //gfniL,
36 /yglmL, 72 /yglmL all in 1 mL
/well
Time to Analysis: 24h
Luc activity: Luciferase enzyme activity is
significantly less in MLE15Luc2 cells than in
MLE15Luc1 cells. For both cells, luciferase
activity is time- and dose-dependent peaking at 4-
8 hours.
Aerosil 200: AE2 induced dose- and time-
dependent Luc response which peaked at 3 h and
decreased thereafter in a similar way as TNF.
Contrary to TNF, AE2 induced much cytotoxicity
starting at 6h.
Effect of proteasomal inhibitors (MG-132):
Inhibitor reduced AE2 Luc activity to near control
levels. Similarly, LDH-cytotoxicity was halved
A549 human cell response: AE2 acted similarly
to the MLE response. CI particles showed slightly
less activity without peaks. AE2 increased
cytotoxicity after 12 h, whereas CI had no effect.
Contrary to MLE mouse, MG_132 did not affect
Luc activity but PD98059 (selective
noncompetitive inhibitor of the MAP pathway) and
SN50 (NF- kB inhibitor) reduced AE 2 and CI-
induced activity.
Reference: Song et
al. (2008,156093)
Species: Rat
Cell Type: RAW
264.7
DEP collected from a 4JB1 -type, light-duty
(2740 cc), four-cylinder diesel engine operated
using standard diesel fuel at speeds of 1500
rpm under a load of 10 torque.
Particle Size: 0.4 /jm (mean diameter)
Route: Cell Culture (5 X 10E cells
seeded on a 24-well plate)
Dose/Concentration: 50//gfmL
Time to Analysis: Nitrate
production: 72h.
Nitrite Production: 50//gfmL of DEP induced
production when compared to the control. Over
the 72-h period, a general trend was not observed,
but maximal induction of nitrite occurred at 4 h
after stimulation.
Reference:
Steerenberg et al.
(2006, 088249)
Species: Rat, Human
Cell Line: Crl/WKY
(rat)
Cell Type: AMs (rat),
alveolar epithelium
cells (rat), A549
PMC: PM Coarse
PMF: PM fine
Ambient air samples collected from Rome,
Italy: Oslo, Norway: Lodz, Poland: Amsterdam,
the Netherlands; De Zilk, the Netherlands.
Particle Size: PMC: 2.35 ¦ 8.5/jm; PMF:
0.12 ¦ 2.35/ym
Route: Cell Culture
Dose/Concentration: NR
Time to Analysis: 20h
Crustal material (metals and endotoxin but not Ti,
As, Cd, Zn, V, Ni, Se) were positively associated
with on vitro rat macrophage IL-6 and TNFa and
in vitro Type 2 (rat alveolar) MP-2 and II-6. Sea
spray (Na and CI) was also correlated with
Macrophage IL-6
Reference: Tal et al.
(2006,108588)
Species: Human
Cell Type: HAEC
100 mM Zn(ll) or V(IV) stock solutions (N0S)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 500/ymol
Time to Analysis: 5, 20min
Zn-mediated EGFR phosphorylation: EGFR
kinase activity was required but not EFGR ligand
binding. EGFR Kinase inhibition reduced Zn
mediated EGFR activation, (authors NOTE:
complete reverse of results in B82L and A431
cells). Src Kinase is not required. Zn inhibiting Src
kinase was nearly total after 20 min.
EGFR-specific protein tyrase phosphatase
(PTP): Zn inhibited PTPs, similar to V(IV) resulting
in a decrease of exogenous EGFR
dephosphorylation
Reference: Tamaoki
et al. (2004,157040)
Species: Human
Cell Type: HBEC
UFCB: Ultrafine Carbon Black
Japan)
FCB: Fine Carbon Black (Tokai Carbon, Japan)
Particle Size: UFCB: 11 ± 0.5 nm (mean
diameter)
FCB: 250 ± 16nm (mean diameter)
(Tokai Carbon, Route: Cell Culture (10' cells/well)
Dose/Concentration: 6.1,12.3,
18.4,24.5, 30.7/yg|cm2
Time to Analysis: Up to 72h
DNA synthesis/ Protein synthesis: Synthesis
increased by UFCB (30.7) for up to 72 h and
flattened after 48 h. FCB had no effect. UFCB
also showed a dose-dependent response beginning
at 12.3 /yg.cm2 up to 24.5 whereafter the
response plateaued. The addition of Cu/Zn Super
oxide dismutase (SOD) or a NADPH oxidase
inhibitor completely inhibited the UFCB effects.
Similarly, two different EGFR tyrosine kinase
inhibitors, and a Me inhibitor all reduced UFCB
response to control levels.
ERK activation: UFCB caused phosphorylation of
ERK beginning at 2 min, peaking at 5 min and
decreasing at 10 min. ERK activation was
inhibited by EGFR tyrosine kinase inhibitor Cu/Zn
SOD and neutralizing body for HB-EGF but not by
PDGF-R kinase inhibitor.
HB (polyclonal heparin binding)-EGF release:
UFCB induced rapid cell surface loss with recovery
after 20 min and nearly full recovery at 360 min.
Metalloproteinase inhibitor and Cu/Zn SOD both
prevented HB-EGF release.
July 2009
D-70
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Tao and Kobzik
(2002,157044)
Species: Rat
Cell Type: RLE-6TN
(Alveolar Type II
Epithelial Cells), Fetal
Lung Fibroblasts
(RFL), AMs
UAP: Urban Air Particles - SRM 1649
HO2: Titanium dioxide
SiO?
ROFA
Particle Size: Titk ~1//m;Si02: ~1//m;
ROFA: NR
Route: Cell Cuilture (1x10E5 cells
AM
1.4x10E5 cells RLE/RFL)
Dose/Concentration: 1 -50 //cj/mL
Time to Analysis: 24h
Cytokines: TNF-a and MIP-2 in RLE was
unaffected by any particle samples. TNF-a and
MIP-2 in AM significantly increased with 25
/yg/mL UAP. TNF-a and MIP-2 in the co-culture of
AM + RLE increased with each particle. The order
of particles in decreasing order are as follows:
SiO? at 25/yg/mL, UAP at 12.5/yglmL, ROFA at
25 /yglmL, and Ti02 at 50 /yglmL. Except for Si02,
the blocking of effects caused by LPS absorbed on
the particles did not affect the cytokine response.
For Si02, the response was reduced but still above
the control.
Co-culture: Physically separating AM and RLE
cells and adding PM completely negated the co-
culture's response to PMs. This indicates that cell
to cell contact is required for co-culture
potentiation of PM effects.
Inhibitors: Various inhibitors of cell adhesion
molecules (heparin, f3 -1, 2 or 3 integrin) had no
effect on UAP-induced cytokine release.
Reference: Veranth
et al. (2007, 090346)
Species: Human
Cell Type: BEAS-2B,
A549, NHBE
Artificial particles and PMs
N-AI: nano alumina AI2O3
M-AI: Micro AI2O3
N-Ce: nano Ce02
M-Ce: micro Ce02
N-Fe: nano Fe203
M-Fe: micro Fe203
N-Ni: nano NiO
M-Ni: micro NiO
N-Si: nano Si02
M-Si: micro Si02
N-Ti: nano Ti02
M-Ti: micro Ti02
KLN: kaolin
MUS: Min-U-Sil (ground crystalline silica)
DD: desert rural soil Utah PM2.6
JE: Juarez, urban street PM2.6
MNC: Mancos, rural Utah PM2.6
LPS: lipopolysaccharide
V: VOSO4 (soluble) (19 /yglmL)
TNFa (0.01 /yglmL)
Particle Size: Surface mean diameter/
surface
N-AI: 6nm (261 m2|g)
M-AI: 210nm (7.7)
N-Ce: 14nm (71)
M-Ce: 1500nm (0.6)
N-Fe: 5nm (221)
M-Fe: 10Onm (12)
N-Ni: 6nm (145)
M-Ni: 16nm (57)
N-Si: 19nm (127)
M-Si: 440nm (5.4)
N-Ti: 6nm (242)
M-Ti: 41 Onm (3.5)
KLN: 10Onm (24.3)
MUS (NOS <5//m)
DD: 400nm (6.2)
JE (NOS < 3/ym)
MNC: 200nm (13.0)
Route: Cell Culture (35,000
cells/cm2 BEAS; 2500 cells/cm2
NHBE: 20,000 cm2 A549)
Dose/Concentration: 0.53, 5.3
and 53/yg/cm2
(- 1,10,100/yg/mL)
Time to Analysis: 24h
Cell Viability: Except for Ni and V no cytotoxicity
was observed at the highest concentration.
IL-6 secretion in BEAS-2 B cells: Nano and
micro sizes of the same metal showed no
differences in response (high experiment to
experiment variability). In general, the soil-derived
dusts (JE, DD, MNC) were more potent than the
metal and ceramic oxide particles. In KGM media,
BEAS-2B cells are more responsive to vanadium
and other soluble metals and less responsive to
LPS, but this relationship is reversed in LHC-9
media.
IL-8 secretion in BEAS/LHC vs NHBE in BEGM
cells:
Levels were much higher in NHBE cells than BEAS-
2B cells. For BEAS-2B, the nano size Si and both
sizes of Ni induced levels statistically greater than
the control. For NHBE, only Si and Ni (for both
sizes) were statistically greater than control.
IL-6 in NHBE: The nano and micro sized particles
of Al, Ce, Fe and nano sized Si all induced
statistically significant increases. Control levels of
IL-6 were much higher in NHBE cells than in BEAS-
2B cells. Secretion induced by pure oxide particles
was small for both the mid and high concentration
levels (5.3 and 53 /yg/cm2).
BSA/ Bovine serum addition effect: In a fixed
solution nano-Ni, nano-Ti and KLN all reduced the
measured IL-6 by 60+ percent. Addition of BSA or
bovine serum dose dependently reduced the action
of the particles to near control levels.
PM effects (without added protein) on IL-6 in
solution: Increasing metal concentration did not
affect a fixed IL-6 concentrration until the 100 or
316/yg/mL levels.
July 2009	D-71	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Veranth
et al. (2007, 090346)
Species: Human,
mouse, rat
Cell Type: A549,
BEAS-2B (types E and
U), RAW 264.7,
Primary macrophages
S: desert dust (collected from unpaved desert
road in Utah, PM2.6 enriched)
V: vanadium soluble (prepared from VOSO4,
Alfa Aesar, Ward Hill, MA))
C: Coal fly ash (PM2.6 enriched and derived
from commercial power plant burning Utah
bituminous coal)
D: Diesel PM (tail pipe particles collected from
high emitting black smoker on-road light duty
truck)
L: Lipopolysaccharide
T: Titanium dioxide (Alfa Aesar)
K: Kaolin (purchased from Capitol Ceramics,
UT)
Particle Size: BET surface (m2fg)
Route: Cell Culture
Dose/Concentrations: Maximum
concentrations:
S - 100/yg/cm2
V - 100/yg/cm2
C - 100 /yglcm2
D - 32/yg/cm2
L - 1000 EU/mL
T - 100 /yglcm2
K - 100 /yglcm2
Time to Analysis: 24h
6.2 (PM2.6 enriched)
NA
5.4	(PM2.6 enriched)
NR
NA
3.5	(1-2/ym)
K: 24 (< 200 mesh - 74/ym)
Viability: Generally, cell viability was greater than
75% of the control post treatment. Vanadium, at
the highest concentration, induced less than 50%
of control viability whereas kaolin, also at the
highest concentration, induced cell death.
Cytokine IL-6: BEAS-2B E or U in LHC-9 showed
a response to S and L. BEAS-2B (U) was in LHC-9
medium with added serum (FBS). This resulted in a
doubling of response coupled with at least an 8
fold increase in control levels. BEAS-2B (E)
showed response for S and V but not L. A549
showed response to S and K. RAW 264.7 and Rat
macrophages showed responses to Sfvery low)
and L. In general, the IL-6 responses in A549
and RAW 264.7 were similar and significantly
lower than the responses in rat macrophages or
BEAS-2B.
Effect of culture media composition (BEAS-
2B): Varying ratios of LHC-9 and KGM media
resulted in a near 10 fold increase in control rate
once LHC was 33% or more of the media. Upon
Soil Dust (NOS) exposure IL-6 increased linearly
with % LHC-9 in culture/exposure media. Addition
of calf serum (0.1-10 %) raised control IL-6 levels
at least 40 fold. At a steady PM concentration,
the addition of serum resulted in a log-linear
increase in IL-6 release which blocked any PM
effect.
Reversibility of media effect: Changing media
with every passage showed that media effects do
not persist once media are changed.
Culture Well Size: Going from a 6 well to 96
well plate (decreasing well size) increased IL-6
control values about ten fold, while the positive
control (TNF) response increased 3 fold. Hence the
sensitivity of the test (i.e., positive/control
response) declined from 11 fold to 3 fold with
increasing well number I decreasing well size.
Because cell seeding density and the like were
held constant, these changes suggest that edge
effects are the cause of the IL-6 changes.
Reference: Veranth
et al. (2006, 087479)
Species: Human
Cell Type: BEAS-2B
PM2.6 samples from 28 samples from 8
locations in Utah, New Mexico and Texas
(rural, industrial, road side, military)
2 coal fly ash samples (a product of
combustion using Utah bituminous coal and
New Mexico bituminous coal)
Ti02
kaolin clay
Particle Size: PM2.6; Ti02: 1-2/jm
Route: Cell Culture (35,000 cells/
cm2 in KGM media)
Dose/Concentration: 10, 20,40,
80/yg/cm2
Time to Analysis: 24h
Cell assays: In sample soils viability declined dose
dependently while IL-6 increased dose-
dependently. IL-8 was highly variable (peak at 20
/ygll, dose-dependent increase or flat response.)
IL-6 assays for all soil PMs: Soils ranged across
an order of magnitude greater than LPS, coal fly
ash, Ti02 or kaolin samples. One soil even
exceeded the pos V control at equal
concentrations
Correlation with cell viability: Correlation was
strong for Mn (p < 0.001) and weak for EC3, K,
Se, and Hg (0.01 < p < 0.05).
IL-6 -10 //g/cm2: Correlation was medial for 0C-
1 (Organic Carbon) and P at 0.001 < p< 0.01.
IL-6 80 fjgl cm2: Correlation was strong for 0C3,
OP (pyrolized Carbon), 0C, EC1, TC and
intermediate for 0C2, 0C4, Zn and weak for
Ca2+, EC2, Si, Ca, Ca: Al.
IL-8 10 //g/cm2: Correlation was weak for EU
(Endotoxin), C03, Si, and Br.
IL-8 80 //g/cm2: Correlation was medial for C03,
Sr and weak for K +, EC3, Mg, Si.
IL-8 trend (corr over 10-80 range): Correlation
was strong for EC, intermediate for 0C4, EC1,
EC2, EC3, TC, Ni and weak for OP, 0C, Cr, and
Sr. IL-6 and II-8 were not correlated nor were IL-6
and cell viability. Authors noted that weak
correlations (0.01 
-------
Study
Pollutant
Exposure
Effects
Reference: Veranth
et al. (2004, 087480)
Species: Human
Cell Type: BEAS-2B
Veranth et al. 2004
PM2.E enriched soil samples
DD: desert dust, unpaved road, Utah
WM: West Mesa, sandy grazing site, NM
R40: Range 40 gravel soil, TX
UN: Uinta, sandy soil, UT
Particle Size: 0.4 - 3//m
Route: Cell Culture (20,000/cm2)
Dose/Concentration: 10, 20,40,
80,160 /yglcm2
Time to Analysis: 24h
Elemental Analysis of PM's: Major differences
UN generally lower in major minerals but high Fe
content and high EC. High Mn. Low Pb and Zn
Cytotoxicity: UN and WM were the most
cytotoxic at all dose levels, followed by R40 and
DD. All particles showed a dose-dependent
cytotoxic response.
IL-6 Release: DD and R40 (up to the 160//gfcm2)
showed dose-dependent responses and induced an
8-fold increase at the highest concentration levels.
WM peaked at 40//gfcm2 and UN induced similar
responses above 10 /yglcm2.
IL- 8 release: DD induced a dose-dependent
response. WM peaked at 10 /yglcm2. Release
induced by DD and WM seemed to be limited by
toxicity. There was no treatment with R40.
TNF-a: DD, WM and UN induced release was not
detected at the 40 or 80 /yglcm2 concentrations.
LPS: LPS was the primary factor in inducing IL-6
release when exposed to LPS-containing mixtures.
LPS alone induced lesser responses than
treatment to the environmental dust particles.
TiLPS induced a less than two-fold increase in IL-6
versus the over seven-fold increase induced by soil
dust positive control. LPS treatments were less
cytotoxic than DD. Limited IL-6 and IL-8 responses
were observed at 2000 EU/mL compared with DD
at 80 /yglcm2
Endotoxin: Inverse relationship between
endotoxin content and IL-6 release was observed.
Viability vs Physical Modification of Dust
Sample (no UN): Only leaching in a variety of
water based vehicles increased viability minimally
(generally < 25 %). Heat treatment (150-, 300,
550F) and methanol extraction had no effect
IL-6 release vs Physical Modification of Dust
Sample (no UN): One hour thermal treatment at
150n F had no effect on IL-6 response. All other
treatments reduced IL-6 release (heat 350a 500n
and extractions).
Reference: Veronesi
et al. (2002, 024599)
Species: Human
Cell Type: BEAS-2B
Ambient PM
-	St. Louis: Urban particulates
-	Ottawa: Urban particulates
¦MSH: Volcanic dust from Washington state's
Mt. St. Helen
¦Woodstove: Woodstove particles from
conventional fireplace burner
¦CFA: Coal fly ash from western U.S.
powerplant
-OFA: Oil fly ash from Niagara, NY
¦	A: Total Fractions
-	B: Soluble Fractions
¦	C: Washed Fractions
Particle Size: PM > 2.5/ym; PM: 2-10/jm;
PM > 10/ym
Route: Cell Culture
Dose/Concentration: 50//gfmL;
30/yg/cm2
100/yg/mL; 60/yg/cm2
Time to Analysis: 4,16h
Ca: Calcium increased significantly with all
particles types.
IL-6: At 50 and 100//gfmL, IL-6 increased with all
particle types at 4 and 16h. Overall, fraction -A
was the most potent.
Surface charge: Surface charge correlated
strongly with increases in both Ca2+ and IL-6
levels. OFA, however, was unmeasurable due to
technical difficulties.
July 2009	D-73	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Vogel et
al. (2005, 087891)
Species: Human
Cell Type: U937
(ATCC) monocytes
(macrophage
differentiation)
UDP: SRM 1649 (NIST)
UDP-OE: DCM extract of SRM-1649, 0.45 fjm
filter
sLIDP: stripped particles UDP
DEP: SRM 2975 (NIST)
DEP-OE: DCM extract of SRM-2975, 0.45 fjm
filter
sDEP: stripped particles DEP
CB95: Carbon Black (Degussa)
Particle Size: UDP, DEP: NR; CB95: 95nm
Route: Cell Culture (2 X106 - 2
X10e cellsfmL)
Dose/Concentration: DEP, UDP:
2.5,10 or 40 /yglcm2
(eq to 12.5,40, 200 /yglmL)
DEP-OE, UDP- OE: 10/yg/cm2
(particle equivalent)
Time to Analysis: 24h
Effect on mRNA expression (COX-2, TNFa, IL-
6, IL-8, C/EBPP, CRP, CYP1a1): All DEP and
UDP induced dose-dependent increases. IL-6
tended to plateau at 10/yg/cm2. Generally, with
the exception of C0X-2, UDP effects on genes
were stronger than DEP.
Cytotoxicity: Both DEP and UDP were cytotoxic
at 40 /yglcm2
Fractionation and mRNA expression: For COX-
2, TNFa, IL-8 mRNA fractions were much more
active than parent particles and consequently
stripped particles were much less active than
parent particles. CB95 had no effect. The reverse
effect occurred for IL-6 and CRP mRNA
expression. The particles that induced mRNA
expression in decreasing order are: sUDP, UDP,
UDP-OE.
Inhibition of mRNA expression: CRP:
pretreatment with IgG and wortmannin (Fey
receptor binding and ingestion dependent inhibitors
respj blocked the effects of DEP, UDP and sDEP
and sUDP. Luteolin (AhR inhibitor) had no effect.
COX-2: Only luteolin inhibited C0X-2 expression
for DEP, DEP-OE, UDP, and UDP-OE.
CYP1 a 1: Luteolin also inhibited OE-DEP and OE-
DUP effects (only those two particles tested).
Cholesterol accumulation: DEP, UDP and UDP-
OE and DEP-OE at 10 /yg/cm2 all increased
cholesterol accumulation by at least 2 fold
Reference: Wang et
al. (2003,157106)
Species: Rat
Cell Type: Lung
Myofibroblasts
V20e: (Aldrich Chemical Co., Wisconsin)
Particle Size: NR
Route: Cell Culture (1 X10!l
cells/100 mm dish; 3.2 X104
cells/cm2)
Dose/Concentration: 400 fjm
Time to Analysis: 0.5,1, 4, 24h
H2O2 drives STAT-1 activation: Pretreatment
with NAC or catalase reduced V2OB—induced STAT
activation by more than 90% and completely
abolished H202-induced STAT activation. Within 5
min of V2O6 treatment, H2O2 was significantly
decreased in the supernatants of cultured
myofibroblasts and suppression of H2O2 levels
continued for up to 24h post V2O6 treatment. This
supports the findings that myofibroblast-generated
H2O2 is required for V2OB—induced STAT
activation.
Temporal STAT-1 activation: H2O2 induced
rapid activation within minutes whereas activation
by V2OE occurred more slowly (beginning 8h post
treatment).
p38, ERK, EGFR: p38 and EGFR are required for
H2O2- or V2OB—induced STAT-1 activation whereas
ERK is not required
Reference: Whitekus
et al. (2002,157142)
Species: Mouse
Cell Line: RAW
264.7
DEP (light-duty, four-cylinder engine- 4JB1
type, Isuzu Automobile, Japan; standard diesel
fuel) (extracts)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 50//gfmL
Time to Analysis: DEP extracts
suspended in methanol and
sonicated 20min. Centrifuged
10min. Dried DEP resuspended
and stored -20°C. Cell cultures
maintained 37 °C. Exposed to
antioxidants 5h. H0-1 western
blot, determination of cellular
GSH:GSSG ratios, carbonyl protein
content, lipid hydroperoxides
performed.
DEP significantly reduced the GSH:GSSG ratio.
This effect was prevented by adding thiol
antioxidants NAC or BUC. DEP increased lipid
peroxide levels, but the addition of all antioxidants
decreased these levels. DEP increased carbonyl
groups. NAC, BUC, and luteolin reduced HO-1
expression.
July 2009	D-74	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Wilson et CB: Carbon Black, Printex 90 (Degussa)
al. (2007, 097268) FeC|3
Species: Mouse
Cell Type: J774
ZnCI2
Particle Size: CB: 14 nm
Route: Cell Culture (4 X10!l
cells.i'niL at 1 mL/well)
Dose/Concentration: CB 1.9-31
//g/mL; FeCIs, ZnCI2 0.01 ¦ 100
//mo
Time to Analysis: 4h
ROS production in cells: CB alone increased
R0S. Coexposure with ZnCI2 did not affect ROS.
ROS production - cell free: CB induced a
significant increase in ROS. ZnCI2 had no effect.
Coexposure CB/Zn also had no effect.
TNFa production (Fe -Zn 0.01-100 /rniol):
Coexposure of CB over a range of metals gave no
change over CB alone for Fe. For Zn, only at the
concentration of 100 //mol was there a small
interaction between Zn and CB.
Similar results were seen at metal concentrations
between 20 -100 //mol. Synergism was observed
between Zn and CB and no observed effect of Fe.
Macrophage cytoskeleton: CB resulted in black
vacuoles. Co-treatment of cells with Zn and CB
increased the severity of Zn effects. Fe exhibited
no synergism.
Apoptosis /necrosis: No synergism of CB with
either Fe or Zn
Phagocytosis: Only at 31 //mol CB and 50/ymol
Zn did a synergistic effect occur; it resulted in a 4-
fold reduction
Reference: Wottrich
et al. (2004, 094518)
Species: Human
Cell Type: A549,
THP-1, Mono Mac 6
Fe: hematite a-Fe203
Si60: silicasol (SiOy, amorphous silica)
Si100: silicasol
~: crystalline quartz DQ12
Particle Size: Fe: 50 ¦ 90nm; Si60: 60nm;
Si100: 80 -110 nm; Q <5 /jm
Route: Cell Culture (2 X10'
cells/well. Co-culture: 2 X104
A549 and 2 X103 Macrophages)
Dose/Concentration: A549 light
microscopy hematite 100/yg/mL
(23/yg/cm )
TEM hematite 50//g/mL (16
//g/cm2)
Cytotoxicity 10, 50,100 and 200
//g/mL (6.1, 30, 61 and 121
//g/cm2)
Cytokines 50 and 200//g/mL
Time to Analysis: 24 h
Particle Uptake: Hematite agglomeration was
observed in all 3 cell lines. TEM confirmed cytosol
aggregates as well as single particles, which
includes particles transported intracellular^ to
basolateral membrane of epithelial cells.
Cytotoxicity: LDH increased significantly in
A549. In decreasing order, Q, Fe, S60, and
S100 (which exhibited levels similar to controls)
all induced cytotoxicity. THP-1 cells appeared the
most sensitive with Q, Fe, S60, S100, control
inducing cytotoxicity in decreasing order. Mono
Mac 6 cells were the least sensitive with Fe, S60,
Q, S100.
Cytokines: IL-6 and IL-8 released from A549 cells
upon exposure to all particles. No response was
observed in Mono Mac 6 or in THP-1 cells.
Co-cultures: Mix of A549 with either Mono Mac
6 or THP-1 led to a large (ten fold) increase in
response to particles. Ten fold increases were
observed in IL-6 and IL-8 levels with the Mono
Mac 6 co-culture and the THP-1 co-culture,
respectively.
July 2009	D-75	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Wu et al. ZnSOi (Sigma)
(2007, 098412) Particle Size: NR
Species: Human
Cell Line: B82L
Cell Type: B82L- par
(parental fibroblasts),
B82L-wt (wild type
EGFR), B82L-K721M
(kinase defective
EGFR), B82L-c*958
(COOH-terminally
truncated EGFR at
Tyr-958)
Route: Cell Culture
Dose/Concentration: Zn: 500
/ymol
EGF: 10Ong/mL
Time to Analysis: 20min
EGFR mutations: EGFR-wt has a
functional tyrosine kinase domain,intact Src
phosphorylation (Tyr 845) and 5 tyrosine
autophosphorylation sites. EGFR-c'958 lacks all 5
tyrosine autophosphorylation sites. EGFR-
K721M lacks tyrosine kinase (ATP binding). EGFR-
Y845F lacks Src autophosphorylation (Tyr 845)
and, instead, has a receptor at Tyr 845 that is
phosphorylated by nonreceptor Tyrosine kinase
Src.
Zn Induced Ras (MAPK signaling protein): No
effect was observed in B82L-par cells. Zn had an
effect in -wt, -c'958, and -K721M which confirms
the need for EGFR. This indicates that neither
tyrosine kinase nor autophosphorylation sites
were required for Zn effects. No observed
increase for Y845F indicated that EGFR tyrosine
845 (phosphorylated by c-Src) is required for Zn
effects. However, it was not required for EGF
effects.
Src Kinase requirement: Using a Src blocker
drastically reduced Zn effect but not the EGF
effect. Src activation occurred independent of
EGFR Tyr-845.
Zn induced association of EGFR with Src: Zn
induced a physical association in all 4 mutants;
EGF did not.
Zn induced phosphorylation of EGFR at Tyr-
845: Zn induced phosporylation of EGFR at Tyr-
845 in B82L-wt,-c'958 and -K721M. EGF
exhibited similar effects. Src blockers significantly
reduced phosphorylation induced by Zn but not for
EGF. Neither Zn or EGF induced phosphorylation in
B82L-Y845F cells.
Reference: Wu, W. Zinc Ion: Zn2+
Wang, X.	Particle Size: NR
Zhang, W.
2003
Species: Human
Cell Type: BEAS-2B
Route: Cell Culture
Dose/Concentration: 10, 25, 50
/ymol
Time to Analysis: 0-8h
Cytotoxicity: Exposure to 50 /ymol Zn2+ for 8h
did not result in significant alterations in cell
viability.
PTEN Protein Levels: 50 /ymol Zn2 + for 4 and
8h significantly decreased levels in a dose-
dependent manner. Exposure to 50//M vanadyl
sulfate (tyrosine phosphatase inhibitor) had
minimal effects on PTEN. 100 ng/mL of non-
specified EGF receptor ligand for 1-8h did not
exhibit any significant effects on PTEN levels.
P13K/AI
-------
Study
Pollutant
Exposure
Effects
Reference: Wu et al. Zinc Ion: In2*
(2004, 096949) Particle Size: NR
Species: Human
Cell Type: NHBE
Route: Cell Culture	Cell Viability: After 2 h of exposure, Zn2+
Dose/Concentration: 100//mol indu^d e'fe,cuts in NfnE cells at 100 and 200
/ymol levels (but not 50/ymol). Continuing
Time to Analysis: 2h	exposure to 100/ymol Zn2+ for 4 and 6 h did not
significantly alter cell viability. Thus, in all
subsequent studies, NHBE cells were treated with
100 /ymol Zn2+.
Induced EGFR Phosphorylation: Exposure to
100//M Zn2+ for 1-4 h induced phosphorylation
of EGFR in NHBE cells. EGFR kinase inhibitor
PD153035 (to determine ifphosphorylation of
EGFR was the result of autophosphorylation of
activated EGFR tyrosine kinase activity) caused
Zn2 + induced phosphorylation to subside. Zn2 +
activity requires tyrosine kinase activity.
EGFR Phosphorylation Pathway: To test
whether Zn exposure results in ligand release,
which in turn can activate phosphorylation, NHBE
cells were pretreated with LA1 blocking antibody.
Results showed significant suppression of Zn2 +
induced phosphorylation, therefore Zn2 +
phosphorylation might be initiated by the release
of EGFR ligands.
HB-EGF, TGF-a , EGF: To examine the
involvement of specific ligands (HB-EGF. TGF-a
and EGF) in the phosphorylation pathway, cells
were exposed to anti-HB-EGF, anti-TGF-a and
anti-EGF. Results showed that anti-HB-EGF
reduced Zn2+ induced phosphorylation
significantly, anti-TGF-a produced partial
inhibition and anti-EGF had no inhibitory effect.
Exposure with blocking antibody LA1 was tested
to determine if it caused an increase in soluble HB-
EGF. HB-EGF mRNA expression was also elevated
in cells exposed to Zn2+. Previous studies indicate
metalloproteinase (MMP) involvement in cleaving
ligand precursors. It was found that MMP-3
inhibitor partially blocks Zn2+ induced HB-EGF
release. (MMP-2 and MMP-9 did not show similar
inhibition patterns) Zn2+exposure increased the
release of MMP-3 from HNBE cells.
July 2009
D-77
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Wu et al.
(2005, 097350)
Species: Human
Cell Line: Subclone
S6
Cell Type: BEAS-2B
Zinc Ion: Zn2+
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 50/ymol
Zn2+
Time to Analysis: 4 or 8h; EGFR
phosphorylation: 30, 60,120, 240
min
Cell viability: Exposure to 50/ymol Zn2+ for 8 h
did not result in significant alterations in cell
viability (assessed by LDH release).
P13K/AI
-------
Study
Pollutant
Exposure
Effects
Reference: Zhang et
al. (2007,156179)
Species: Human, Rat
Cell Type: A549,
RLE-6TN (Alveolar
Type II Epithelial
Cells)
PM2.E: Collected by baghouse from Dusseldorf,
Germany
Particle Characterization: Carbon 20%,
Hydrogen 1.4%, Nitrogen < 0.5%, Oxygen
14.1%, Sulfur 2.1%, Ash 63.2%.
Particle Size: PM2.6
Route: Cell Culture
Dose/Concentration: 100 //gfcm2
Time to Analysis: 24h
Apoptosis: At 100 //gfmL for 24 h, PM induced a
2.5 fold increase in apoptosis in A549.
Mitochondrial Membrane Potential: A
significant reduction in AEC mitochondrial
membrane potential was observed.
Caspase -3 & -9: Increased activity of both
enzymes in both cell types was observed. More
specifically, a 2- to 2.5-fold increase of caspase -3
and -9 in A549 and an 8-fold increase of caspase-
9 and 4-fold increase of caspase-3 in RLE-6TN
were observed.
BIM: Downregulation of BIM by RNA interference
inhibited PM-induced apoptosis. An inhibited
decrease in mitochondrial membrane potential and
activation of both caspases were observed.
Reference: Zhang et
al. (2004,157183)
Species: Mice
Cell Line/Type: C10
(alveolar type ll-like
epithelial cell line)
~EP: SRM 1650a
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 5 or 25
/yg/mL
Time to Analysis: 30-360min
fra expression: DEP induces fra-1 but not fra-2
expression. mRNA induction peaks around 180
min DEP affects fra-1 mRNA expression at the
transcriptional level.
ERK/JI\IK/p38 MAPK signaling pathways: 3
inhibitors (PD-98059, SB-202190 or SP-600125)
all reduced DEP stimulated fra-1 induction to near
control levels. DEP stimulated phosphorylation of
the MAPKs which peaks at 60 min but stays
elevated at 180 min.
MMP-9 promoter activity: fra-1 upregulation
may play a role in DEP induced increases in MMP-
9 promoter activity as fra-1 appears to bind at the
¦79 TRE sequence of the MMP-9 promoter.
Study 6203 cell lines: BEAS-2B (A) HR. Bronch.
Epith ATCC # CRL-9609 passage 44
BEAS-2B (E) HR. bronch. Epith. U.S. EPA passage
76-87
BEAS-2C (U) HR. bronch. Epith U. Utah sample
passage 89-97
A 549 HR. alveolar epith U. Utah sample starts at
passage 84
RAW 264.7 M. macrophage ATCC # TIB-71
1 macrophages Rat lung lavage primary cells
Study 6203 culture media
LHC-9 Lechner and Laveck medium Invitrogen
KGM Keratinocyte growth medium Lonza
DMEM/F12 50 % Dulbecco's modification of
Eagle's medium, 50% Ham's F12 medium Gibco
FBS Fetal bovine serum (added to media
formulations) Invitrogen
Table D-3. Respiratory effects: in vivo studies.
Reference
Pollutant
Exposure
Effects
Reference:
Adamson et al.
(2003, 087943)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Weight: 150g
PMio: EHC-93W (whole dust)
EHC-93S (soluble)
EHC-93L (leached)
EHC-2KW, -S, -L
Measured components Zn, Mg,
Pb, Fe, Cu, Al
Particle Size: EHC-93W, -93S,
¦93L, -2KW, -2KS, -2KL: PMio
Route: IT Instillation
Dose/Concentration: 5 mg/rat; 33.3 mg/kg
Time to Analysis: 4h, 1 d, 3d, 7d, 14d
BALF Cells: The greatest increase in cell numbers
was observed with EHC-93W. Activity peaked at 1 d
with a return to normal levels by 7d. EHC-93L also
induced an increase in cell numbers, more so than
EHC-93S, but both particles induced statistically
significant increases. However, these increases were
mostly attributable to an increase in the AM and
PMN populations.
BALF Inflammatory/Injury Markers: Metallo-
proteinase (MMP) 2 and 9 both increased, peaking at
1d and 4h respectively. MMP2 activity appears
related to the soluble fraction whereas MMP-9
activity appears to be related to the leachable
fraction.
July 2009
D-79
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Ahn et
al. (2008,156199)
Species: Mouse
Gender: Male
Strain: BALB/C1
Age: 6wks
Weight: 19-24g
DEP: Collected using a turbo-
charged, intercooler, 6-cylinder,
heavy-duty, diesel engine
(model year 2000
DPBS: control
Particle Size: NR
Route: Oropharyngeal Aspiration
Dose/Concentration: 1,10, 25 mg/kg per day;
Those receiving 25 mg/kg DEP also received pre-
treatment of Dex (1, 5 mg/kg) 1 h prior
Time to Analysis: 5 consecutive days; 72h post
final exposure
BALF Inflammatory/Injury Markers: Lung injury
was more severe in mice exposed to 25 mg/kg of
DEP than when compared to mice exposed to 1
mg/kg DEP. However, lung injury caused by exposure
to 25 mg/kg DEP could be completely prevented with
pre-treatment of 5mg/kg Dex. Treatment with 1
mg/kg Dex prior to exposure to 25 mg/kg DEP
depicted partial reduction in lung injury.
BALF Cells: Treatment with DEP over a 5d period
caused an increase in total number of cells
(macrophages, neutrophils and lymphocytes) when
compared to control.
Total Cells: Control ¦ 5.33 ± 0.44 cells
1 mg/kg DEP ¦ 6.26±0.87 cells
10mg/kg DEP ¦ 14.40 ± 1.90 cells
25 mg/kg DEP ¦ 47.20 ±3.40 cells
COX-2 Expression: Exposure to DEP lead to a dose-
dependent increase in C0X-2 levels; specifically,
treatment with 25 mg/kg significantly increased
C0X-2 levels. This effect was completely reduced by
treatment with 5mg/kg of Dex.
Reference: Ahsan DEP: Obtained from Dr. Masaru Route: Intratrachael Instillation
et al. (2005,
156200)
Species: Mouse
Gender: Male and
Female
Strains: hTrx-1-
transgenic and
C57BL/6 (control)
Aqe: 8-8.5wks
Sagai (Amori, Japan)
Particle Size: NR
Dose/Concentration: Lung Damage: 0.1
mg/mouse; Survival Analysis: 0.2 mg/mouse;
ESR: 0.05 mg/mouse
Time to Analysis: 24h
ESR: hTrx-1 induced 0.05 mg generation of hydroxyl
radicals in the lungs (mid thorax ESR spectra)
compared to control.
BALF Inflammatory/Injury Markers: hTrx-1
attenuated cellular damage from 0.1 mg DEP. Control
mice showed massive edema with neutrophilic
infiltration, hemorrhagic alveolar damage and
collapsed air spaces. hTrx-1 mice showed
mild/moderate edema with clear demarcation of air
spaces.
Viability: After 4,12 and 24 h, survival was 32, 24
and 12% respectively as compared to 80, 52 and
40% for hTrx-1 mice.
Reference: Andre
et al. (2006,
091376)
Species: Mouse
Gender: Female
Strain: BALB/cJ
Aqe: 10-12 wks
UFCP: Ultra Fine Carbon
Particles
(electric spark generator, Model
GFG 1000; Palas, Karlsruhe,
Germany)
Measured Component:
UFCP > 96% elemental carbon
Particle Size: 49nm
Route: Whole-body Inhalation
Dose/Concentration: 380//gfm:l
Time to Analysis: 4 and 24h; 0 and 24h post-
exposure
BALF Cells: A small increase in PMN number
suggests a minor inflammatory response after 24 h
exposure. Number of macrophages did not increase.
BALF Inflammatory/Injury Markers: Total protein
concentration significantly increased post 24 h
inhalation. Post 4h, heat shock proteins were
induced. Post 24 h, immunomodulatory proteins
(osteopontin, galectin-3 and lipocalin-2) significantly
increased in alveolar macrophages and septal cells.
236 (1.9%) genes was increased and 307 (2.5%)
genes were decreased with upregulated genes being
primarily related to the inflammatory process.
Reference:
Antonini et al.
(2004,097199)
Species: Rats
Gender: Male
Strain: Sprague-
Dawley Weight:
— 250g
ROFA-P: Precipitator
¦S: Soluble (0.22/ym filter),
Components: Fe, Al, Ni, Ca,
Mg, Zn
¦I: insoluble, Components: Fe,
Al, Ni, Ca, Mg, Zn,V
¦T: total
R0FA-AH: Air Heater
¦S: Soluble (0.22/jm filter),
Components: Fe, V, Ni, AL
-I: Insoluble, Components: Fe,
V, Ni, AL
¦T: Total
Particle Size: mean diameter:
< 3/jm
Route: IT Instillation
Dose/Concentration: 1 mg/1 OOg bw in 300 //I
saline; 60mg/kg
Time to Analysis: 24h; Clearance Experiment:
two single exposures day 0 and 3 observed at
day 6, 8 and 10
ESR: Only ROFA-P contained free radicals, primarily
in R0FA-P-S.
BALF Cells: No effects on alveolar macrop
were observed, but all ROFA-P fractions increased
lung neutrophils. R0FA-P-S and R0FA-P-I effects
combined roughly equaled R0FA-P-T.
BALF Inflammatory/Injury Markers: ROFA-AH-T
and ROFA-AH-I increased LDH. ROFA-P and -AH
increased albumin for T and I fractions.
Pulmonary Clearance (Listeria Monocytogenes):
R0FA-P-T and ROFA-P-S significantly slowed
bacteria clearance from lungs. ROFA-AH and ROFA-P-
I had no effect.
July 2009	D-80	DRAFT - DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Arimoto et al.
(2007,097973)
Species: Mouse
Strain: ICR
Gender: Male
Age: 6wks
Weight: 29-33g
DEP (collected using a 4JB1 4-
cyl, 2.74L Isuzu diesel engine)
DEP-OC: organic chemical
extracts
LPS
DL - DEP + LPS
DOL - DEP-OC + LPS
Particle Size: 0.4/ym
Route: IT Instillation
Dose/Concentration: DEP or DEP-OC: 4mg/kg;
LPS: 2.5 mg/kg; DL or DOL: NR
Time to Analysis: 24h
Cytokines: DEP-OC or DEP alone did not change
levels of MIP-1a, MCP-1 or MIP-2. DL induced
significant increases in MIP-1, MIP-2 and MCP-1.
LPS: LPS and DOL induced increases in MCP-1
though the increase induced by DL was greater. No
effect on MIP-1 a or MIP-2 was observed.
Reference:
Bachoual et al.
(2007,155667)
Species: Mouse
Strain: C5B17
Gender: Male
Age: 7wks
Weight: 22.3 ±
0.73g
RER: PMio
Paris, France subway
CB
TiO?
DEP
Particle Size: RER: 79% <
0.5/ym; 20%: 0.5-1 fjm
CB: 95nm
Ti02: 150 fjm
DEP: NR
Route: IT Instillation
Dose/Concentration: 5, 50,100 //gfmouse,
0.22, 2.2, 4.5 mg/kg
Time to Analysis: 8 or 24h
BALF Cells: 100/yg RER and 100/yg DEP increased
total cell count and neutrophil influx after 8h and
returned to normal by 24 h. Smaller doses of RER
and DEP induced no effect. CB induced no effect.
BALF Inflammatory/Injury Markers: 100/yg RER
increased BALF protein after 8 h. No effect was
observed after 24 h nor with smaller doses of PM.
RER significantly increased MMP-12 mRNA level
after 8h and HO-1 total lung mRNA content. No
effects on MMP-2 or-9 or TIMP-1 or -2 expression
were observed. No effects from CB or DEP were
observed.
Cytokines: 100/yg RER increased BAL, TNF-a and
MIP-2 protein content after 8 h.
Reference: Becher Suspended PM: SRM-1648
et al. (2007, Particle Size: NR
097125)
Species: Mouse
Strain: CrlfWky
(iNOS(-M) and
C57BI/6,
Gender: Male
Age: 8-14wks
Weight: 25g
Route: IT Instillation
Dose/Concentration: 1.6 //cj/kincj: 64 mg/kg
Time to Analysis: 20h
Cytokines: In both wild and KO strains, all particles
caused increases of IL-6, MIP-2 and TNF-a levels.
NADPH-oxidase KO mice showed significantly lower
levels of IL-6 and MIP-2 responses to SPM compara-
tively to wildtype. iNOS KO mice showed
significantly reduced IL-6, TNF-a, MIP-2 responses
to SPM comparatively to wildtype.
Free Radicals: SPM induced significant increases in
free radical formation in alveolar type 2 cells but
could be inhibited by DPI.
Reference:	Douglas Fir Wood Smoke
Bhattacharyya et (generated by burning wood at
al. (2004, 088095) 400nC in crucible oven)
Species: Mouse
Strain: Sprague-
Dawley
Weight: 200-250g
Particle Size: NR
Route: Nose-only Inhalation
Dose/Concentration: 25gfmouse
Time to Analysis: Various exposure periods (0,
5,1 0,15, 20 min). Parameters measured after
24h recovery period.
Biochemical Parameters: Lipid peroxidation
increased after 20 min of woodsmoke inhalation as
did Myeloperoxidase at 20 min. No effects were
observed at other times or for total antioxidant
status, reduced or oxidized glutathione.
Antioxidant Enzyme Activities: No effect was
observed.
Histology: Dose-dependent damage progressing from
loss of cilia (5 min), degeneration of mucosal
epithelium, loss of mucosal epithelium to disrupted
mucosal epithelium with submucosal edema and
inflammation. Changes persisted for up to 4 days.
Reference: Cao et
al. (2007, 097491)
Species: Rat
Strain: SH and
WKY
Aqe: 12wks
PM2.6
(Thermo Anderson G-2.5
sampler, Shanghai, China)
Components: As, Cd, Cr, Cu,
Fe, Ni, Pb, Zn, V, Ba, Se, Mg,
Co, Mn
Particle Size: PM2.6
Route: IT Instillation
Dose/Concentration: 1.6, 8.0 and 40 mg/kg
Time to Analysis: Exposed 1fd for 3d,
sacrificed 24h following last exposure
BALF Inflammatory/Injury Markers: LDH activity
and TBARs increased a in dose-dependent manner.
Notably, activity in SH rats was much higher than
WKY at the same dose exposed for each dose level.
BALF Cells: PM decreased macrophages and
increased neutrophils and lymphocytes in a dose-
dependent manner. For the same exposed dose, WKY
rats had a higher percentage than SH but a smaller
percentage of neutrophils and lymphocytes.
Cytokines: PM induced pro-inflammatory cytokine
release (IL-ip, TNF-a, CD44, MIP-2, TLR-4, OPN).
Again, SH cytokine level was greater than WKY at all
dose levels. PM induced anti-inflammatory cytokines
CC16 and HO-1 in a similar manner but at much
lower rate.
July 2009
D-81
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Carter
et al. (2006,
095936)
Species: Rat,
Mouse, Hamster
Gender: Female
Strain: F-344 (rat),
B6C3F1 (mouse),
Syrian Golden
(hamster)
Age: 7-10wks (all)
CB: Printex 90
Paritcle Size: primary size:
17nm; 1.2-1.6 /ym (aerosol
aerodynamic diameter)
Route: Whole-body Inhalation
Dose/Concentration: 1, 7, 50 mgfm3
Time to Analysis: 6hfday for 5dfwk for 13
wks; 1 d, 3m, 11 m post exposure
Superoxide: Levels rose in all species at 50mg dose.
Hamsters had no increase at 7 and 1mg doses. Mice
also increased at 7mg. Rats significantly increased at
all dose levels. Rats maintained elevation except for
the 50mg dose at 11 mo postexposure; it declined but
was still higher than control. Mice maintained
elevation at 50mg while 7 mg returned to control
levels by 3mo postexposure.
H202: At 50mg, increased levels in all species, with
the highest in rat, were observed. At 7mg, increased
levels in rats and mice were initially seen but levels
returned to baseline by 11 mo. Hamster levels were
not significant. At 1 mg, no significant changes were
observed.
NO: Induced similar reactions as H202. Rat response
continued through the study while mice and hamsters
returned to baseline by 11 mo postexposure. Rats
produced significantly higher levels at all times than
other species.
BALF Cells: CB induced significant increases in
neutrophils at 7 and 50mg for all species. Rats had
the highest and most prolonged PMN response. Mice
and hamsters had very similar reactions.
Cytokines: TNF-a, MIP-2 and IL-10 increased in a
dose-dependent manner in rats and mice. Hamsters
increased for IL-10 only. MIP-2 levels were highest in
rats. TNF-a level were similarl in all three species at
50mg, but hamsters started with a markedly higher
basal level.
Glutathione Peroxidase: Hamsters were the most
responsive with significant increases at all levels.
Rats and mice increased at 50mg and continued to
increase for up to 11 mo. Hamster levels declined
with time but continued to be higher than control.
Glutathione Reductase: Rats increased only at
50mg and remained elevated for up to 11 mo. Mice
increased at 7 and 50mg and remained elevated for
up to 11 mo. Hamsters increased at all levels at
11 mo, but at 50mg, levels only increased post 1 d.
Superoxide Dismutase: All species reacted in a
dose-dependent manner. Rats were the least
responsive. Rat SOD activity increased over time
while rat and mouse activity decreased at 50mg.
Data were consistent with cytokine data.
Summary: Rats appear to produce proinflammatory
responses while mice and hamsters produce
antiinflammatory responses.
Reference: Cassee
et al. (2005,
087962)
Species: Rat
Gender: Male
Strain: Wistar
(SPFHsdCpb: WU)
and SH/NHsd
Age: 7 wks and 8-
12wks
CAPs: PM2.6
Netherland suburban, industrial
and freeway tunnel site
collections
Wistar rats pre-exposed to
ozone
SO4, NO:; and NH4 ions:
54±4% suburban, 53±7%
industrial and 35±5% freeway
site cone of total CAPS mass
Particle Size: PM2.6
(0.15 < PM < 2.5 //m)
Route: Nose-only Inhalation
Dose/Concentration: PM 365-3720//gfm3
(results from 16 different exposures 2000,
2002); O3: 1600 /yglm3 (0.8 ppm)
Time to Analysis: 8h O3 pre exposure; 6h CAPS
exposure; 48h post-exposure
BALF Cells: Wistar exhibited increased protein,
albumin, NAG and decreased ALP activity and
macrophage numbers. Wistar showed increased
PMNs due to ozone, but was not significantly
increased with additional CAPs exposure. SH showed
no effect of CAPS except for the increased PMNs.
BALF Inflammatory/Injury Markers: No effect on
AL, LDH, Glutathione, GSSG, GSH, Uric Acid was
observed.
Cytokines: No effect on IL-6, MIP-2 or TNF-a was
observed. CAPs induced an increase in CC16 plasma
of SH rats.
Hematology: CAPS induced an increase in RBC,
HGB and HCT of Wistar rats and fibrinogen of SH
rats.
Histology: Wistar and SH rats had no obvious lung
abnormalities. Small changes include increased
macrophages and cellularity of centriacinar septa of
ozone-only rats. Both ozone-only and ozone+CAPS
showed bronchial epithelium hypertrophy and
perivascular influx of PMNs.
BrdU Labeling Index of Terminal Bronchiolar
Epithelium: No CAPs effects were observed.
July 2009
D-82
DRAFT-DO NOT CITE OR QUOTE

-------
Reference	Pollutant	
Reference: Chang UFCB: Ultrafine CarbonBlack -
et al. (2005,	Printex 90 (Degussa)
0977761	Particle Size: 14nm
Species: Mouse
Gender: Male
Strain: ICR
Age: 5wks
Weight: 25-30g
	Exposure	
Route: IT Instillation
Dose/Concentration: 200//gf1 OOulfmouse
Time to Analysis: Parameters measured 4,16,
21,42h post single exposure
Effects
BALF Cells: Neutrophil number was at control level
at 4h, increased after 16h, peaked at 21 h and
returned to normal at 42h. No effect was observed
for the macrophage count.
BALF Inflammatory/Injury Markers: UfCB
increased total protein with peak at 21 h.
Cytokines: TNF-a increased at 4h and returned to
normal at 16h.
VEGF (Vascular Endothelial Growth Factor|:
Increased at 4h and peaked at 16h but remained
elevated at 21 and 42h. VEGF and total protein in
BALF were correlated (R2 - 0.7352).
ROS: Pretreatment with NAC (ROS inhibitor)
decreased induction of BALF VEGF and total protein
by UfCB but did not fully block its effect.
Histology: Thickened alveolar walls in lungs of
UfCB-treated mice 16h post-IT was observed.
Reference: Chang
et al. (2007,
097475)
Species: Mouse
Gender: Male
Strain: ICR
Age: 5wks
Weight: 25-30g
UFCB: Ultrafine Carbon Black -
Printex 90 (Degussa)
Particle Size: 14nm diameter
Route: IT Instillation
Dose/Concentration: 200//gjmouse;
8mg/kg
Time to Analysis: Pretreatment with NAC (N-
acetylcysteine) ip 320 mg/kg, 2-h before UFCB
IT. Parameters measured 24h post exposure.
BALF Cells: Increased relative lung weight, total
protein (2 fold), total cells (11 fold) and number of
neutrophils were observed. BALF AM count was not
affected.
BALF Inflammatory/Injury Markers: Of the 33
identified proteins, the following 6 were confirmed
and validated: Cp (ceruluplasmin), albumin, EGFR,
LIFR (leukemia inhibitory factor receptor), a2M and
fB actin. All were increased following UFCB exposure.
The following were also identified: 3 membrane
proteins, 3 intracellular proteins, 10 protease
inhibitors and 6 antioxidants. UfCB increased LIFR
and EGFR in BALF. UfCB significantly reduced EGFR
and LIFR in lung homegenate. UfCB did not affect
EGFR protein but down-regulated LIFR in A549 cells
treated with UfCB.
Antioxidant: Pretreatment with NAC reduced the
intensity of albumin and a2M bands in BALF as well
as most other proteins.
Statistical analysis showed positive correlation
between VEGF and albumin (R2 - 0.796) and VEGF
and a2M (R2 - 0.7331) in BALF.
Reference: Cho et ROFA: Obtained from Power
al. (2005,156344) unit 4, Boston, MA
Species: Mouse Absent of LPS
Gender: Male Particle Size: NR
Strains: DBA/2J,
129P3/J,
C57BL/6J,
BALB/cJ, A/J,
C3H/HeJ,
C3H/He0uJ
Age: 6-8wks
Route: IT Instillation
Dose/Concentration: 6mg/kg bw (150/yg in
50//I/ 25g)
Time to Analysis: 24h; Additional HeJ and OuJ
mice: single: 1.5, 3 and 6h (compare TLR-
mediated molecular events)
BALF Cells: Significant genetic effects on number of
macrophages and PMNs after ROFA challenge. For
PMNs, DBA/2J, C57BL/6J, BALB/cJ, and 129P3/J
all induced increases significantly higher than
C3H/HeJ. For macrophages, only the A/J strain
induced increases significantly higher than
C57BL/6J. Total protein, PMNs and macrophages all
increased with HeOuJ inducing increases
significantly different from HeJ.
BALF Inflammatory/Injury Markers: Significant
genetic effect on mean total protein concentration
was observed. In decreasing order, DBA/2J, 129P3/J
and C57BL/6J all induced increases significantly
higher than C3H/HeJ.
TLR4 mRNA Expression: A significant decrease
was observed in TLR4 transcript level in HeJ- ROFA
exposed mice post 1,5h. Post 6h, TLR4 levels were
greater than the control levels. OuJ expression
increased beginning 1.5h post exposure.
TLR4 Protein Level: Protein level of OuJ mice
significantly exceeded (~ 2-3 fold) HeJ mice at 1.5,
3 and 6h.
Activation of Downstream Signal Molecules:
Greater activation of MYD88, TRAF6, IRAK-1, NF-
KB, MAPK, and AP-1 was observed in OuJ mice than
in HeJ mice before the development of ROFA-
induced pulmonary injury.
Cytokines: IL-1(3, LT (3, IL-1a, IL-7, IL-13, IL-16
increased in both strains (OuJ and HeJ). Levels of all
cytokines above were significantly higher in OuJ than
in HeJ.
July 2009
D-83
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Churg
et al. (2003,
087899)
Species: Human
Gender: Female
(Mexico City); Male,
Female (Vancouver)
Strain: -
Age: 66±9yrs
(Mexico City);
76 ± 11 yrs
(Vancouver)
Weight: NR
PM (Mexico City- high PM
region, Vancouver- low PM
region)
Particle Size: Geometric mean
size of individual particles in
tissue: 0.040-0.067/ym;
Aggregates in tissue: 0.34-0.54
fjm; Mexico City: 2.5,10/ym
Route: Ambient Air Exposure. Autopsy Tissue.
Dose/Concentration: 10- > 1000X106g dry
tissue; Mexico City: PMio: 66/yglm3, Vancouver:
PMio: 25/yg, PM2.6:15/yg
Time to Analysis: Lung samples taken from
deceased lifelong Mexico City residents and
Vancouver residents > 20yrs. Subjects were
never-smokers, did not work in dust occupations
or cook with biomass fuels.
The lungs from Mexico City residents showed
increased muscle and fibrous tissue in the
membranous bronchioles and respiratory bronchioles
compared to the Vancouver residents. Pigmented
dust, lumental distortion and carbonaceous
aggregates of UFPs were present in the Mexico City
lungs.
Reference: Costa
et al. (2006,
088438
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age: 60d
ROFA
FP&L plant #6 oil, 1% sulfur
Particle Size: ~1.95/ym
Route: ITInstillation vs Nose-only Inhalation (
Dose/Concentration: IT -110/yg/rat
IH - 12 mg/m3
Time to Analysis: IT - single
IH - 6h
24,48, 96h
(histopath 24 and 48 only)
ROFA distribution: IH and IT resulted in equivocal
distribution (/yg/g lung tissue) in 5 different lung
lobes.
Airway Hyperactivity: IT resulted in doubled
airway hyperreactivity at 24 h which was sustained
for 96h. IH hyperreactivity did not reach statistically
significant level.
BALF Inflammatory/Injury Markers: IH and IT
showed very similar responses (R2 - 0.98). Time-
dependent increases were observed for protein and
LDH.
BALF Cells: Neutrophils peaked at 24 h and slowly
declined at 48 and 96h.
Lung Pathology: IT showed more alveolitis,
bronchial inflammatory and fibrinous fluid infiltrate.
IH showed relatively more congestion of small
airways and alveolar hemorrhage.
Reference:
Courtois et al.
(2008,156369)
Species: Rat
Gender: Male
Strain: Wistar
Age: 12-14wks
Weight: NR
PM (SRM 1648; 63%
inorganic carbon, 4-7% organic
carbon, > 1 % mass fraction-
Si, S, A I, Fe, K, Na)
Carbon black (FW, P60)
UF, fine Ti02
Particle Size: PM mean
diameter: 0.4/ym; Carbon
black: FW- 13nm, P60- 21 nm;
Ti02 mean diameter: 0.14/ym
Route: IT Instillation
Dose/Concentration: 5mg PM or Tith
Time to Analysis: I.P injected then instilled. 6-
72h recovery period then killed and lungs
removed.
Particles were present in lung parenchyma that was
removed 12 and 72h post instillation. The Ach
relaxation response significantly decreased in the
12h recovery period group but not in the other
groups. Fine Ti02 did not alter Ach relaxation.
Reference: Dick et
al. (2003, 036605)
Species: Mouse
Gender: Female
Strain: CD1
Age: 8-10wks
Weight: 20-25g
CO: PM Coarse
Fl: PM Fine
FU: PM ultrafine
PM collected in RTP, NC
Particle Size: CO: 3.5 - 20
//m; Fl: 1.7- 3.5/ym;FU: <1.7
fj m
Route: IT Instillation
Dose/Concentration: 10 /yg, 50 /yg, 100
/yg/mouse; 0.5, 2.5, 5.0mg/kg
Time to Analysis: DMTU 500 mg/kg bw 30 min
pre exposure for some mice. Parameters
measured 18h post-exposure.
Particle Characteristics: S increased (CO-
33.20/yg/mg, Fl- 49.44/yg/mg FU- 122.79/yg/mg)
with decreasing particle size (mostly in the water-
soluble fraction). Fe and Cu higher in coarse and fine
fractions (mostly present in the insoluble). CO PM
contained more nickel (in both soluble and insoluble)
than Fl or FU particles. Also, endotoxin levels similar
in CO and Fl; much lower in FU (0.165 EU/mg).
BALF Cells: PMN increased with exposure for all 3
fractions except 100 /yg Fl.
BALF Inflammatory/Injury Markers: Albumin
increased only at 100 /yg Fl. No differences in NAG
or LDH observed.
Cytokines: IL-6 increased at 100 /yg dose for all 3
fractions with similar responses. TNF-a increased a
100 /yg dose of fine PM vs control.
Effect of PM After Pre-treatment w/ DMTU:
Systemic administration of DMTU alone depicted a
two-fold increase in total antioxidant capacity.
DMTU halved neutrophil response observed with
PMs alone: No fractions were increased over DMTU
alone which was at least two-fold saline control. IL-6
concentrations were drastically reduced in the DMTU
group for the mice exposed to coarse particles (all
fractions were reduced but only coarse had a
significant response). TNF-a levels were decreased
after treatment with particles and DMTU but
treatment with particles and saline (control)
produced similar results.
July 2009	D-84	DRAFT - DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Dybdahl et al.
(2004,089013)
Species: Mouse
Gender: Female
Strain: BALB/CJ or
transgenic
(MutaMouse)
Age: 9-10wks
Weight: — 20g
DEP: SRM 1650 (NIST)
Particle Size: DEP: NR;
Control: PM 0.13 fjm diameter
Route: Nose-only Inhalation
Dose/Concentration: I: 20 & 80 mgfm3
II: 5 & 20 mgfm3
Time to Analysis: I: single exposure 90min; II:
90 min/day for 4d; I & II: parameters measured
1, 3, or 22h post exposure
Cytokines: A single 90-min DEP exposure increased
IL-6 gene level dose-dependently in the lung. For 80
mgfm3 DEP, significantly higher IL-6 gene level was
observed, both 1 and 22h post exposure. For 20
mg/m3 DEP, a significantly higher IL-6 level was
observed at 1 h post exposure but normalized at 3h.
BALF Cells: Inhalation of DEP did not decrease
viability of BAL cells (see Tablel). For mice exposed
to 20 mg/m3 DEP, at 1 h post exposure in BAL fluid
there was 3 fold increase in total cell number.
DNA Damage: Level of 8-oxodG increased post
single exposure with 80 mgfm3 inducing levels
significantly higher than controls. Repeated
exposures were associated with significantly higher
DNA strand breaks.
Reference: Elder et
al. (2004, 055642)
Species: Rat
Gender: Male
Strain: Fisher344,
SH
Age: 23m (Fisher),
11 -14m (SH)
UFP: argon-filled chamber with
electric arc discharge (TSI, Inc.,
St. Paul, MN)
Particle Size: 36nm
Route: Whole-body Inhalation. Intraperitoneal
(ip) for saline and LPS
Dose/Concentration: UFP: 150/yg/m3 bw; LPS:
2 mg/kg
Time to Analysis: Parameters measured post
single exposure of 6h, 18h
BALF Cells: Neither inhaled UFP nor LPS cause a
significant increase in BAL fluid total cells or
percentage of neutrophils in either rat strain. No
significant exposure-related alteration in total protein
concentration was observed. In both rat strains LPS
induced a significant increase in the amount of
circulating PMNs. When combined with inhaled UFP,
PMNs decreased: for F-344 rats, this decrease was
significant.
BALF Inflammatory/Injury Markers: Plasma
fibrinogen increased with LPS in both rat strains with
the magnitude of change greater in SH rats. UFP
alone decreased plasma fibrinogen in SH rats.
Combined UFP and LPS response was blunted but
significantly higher than controls. Hematocrit was
not altered in either rat strain by any treatment. No
change in activites of LDH and b-glucuronidase was
observed.
TAT complexes: With all exposure groups averaged,
plasma TAT complexes in SH rats were 6.5 times
higher than in F-344 rats. LPS caused an overall
increase in TAT complexes for F-344 rats which was
further augmented by inhaled UFP. UFP alone
induced a decreased response. In SH rats, UFP alone
exhibited a significantly increased response and LPS
exhibited a decreased response.
ROS in BALF: In F-344 rats, both UFP and LPS have
independent and significant effects on DCFD
oxidation. Effects were in opposite directions;
particles decreased ROS whereas LPS increased
ROS.
Reference: Elder et
al. (2004, 087354)
Species: Rat
Gender: Male
Strain: F344
Age: 21m
Freshly generated vehicle
exhaust emissions from I-90
between Rochester and
Buffalo, NY
Particle Size: NR
Route: Whole-body Inhalation; (IT Instillation:
Influenza)
Dose/Concentration: Vehicle exhaust: 0.95-
3.13 x 106 particles/cm3
Endotoxin: 84 EU
Influenza (IV): 10, 000 EID 50 in 250ul
Time to Analysis: 1x6h, 3x6h or both.
Parameters measured 18h post-exposure. 48h
prior to on-road exposures, instilled
intratracheally with IV. Immediate pre exposure
of priming agent endotoxin.
EXPERIMENTS
1: LPS + PM 6h
2: LPS + PM 6h, 3 x 6h
3: IV + PM 6 h
4: IV + PM 6h, 3x 6 h
No departures from normal baseline cellular or
biochemical values were observed, suggesting that
on-road exposures were well tolerated by the rats.
BALF Inflammatory/Injury Markers: Increase in
total protein concentration, LDH and B-glucuronidase
activities were observed.
Specific results according to groups 1-4 are as
follows:
Experiment 1: No endpoints revealed significant
differences between groups of rats exposed to gas
phase only versus the gas-phase/particle mixture.
Experiment 2: Combination of endotoxin and
particles produced greater inflammatory responses
than those treated with saline and particles post 1 d.
After 3 days, no statistically significant changes
were noted.
Experiment 3: Influenza virus significantly increased
ROS release in BAL cells.
Experiment 4: Influenza virus significantly increased
both percentage of PMNs in lavage fluid and BAL cell
ROS release.
July 2009	D-85	DRAFT - DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Elder et
al. (2005, 088194)
Species: Rat,
Mouse, Syrian
Golden Hamster
Gender: Female
Strain: F-344,
B6C3F1, FIB
HSCb: Printex-90 high surface
area carbon black, Deguss-
Huels (Trostberg, Germany).
LSCb: Sterling V, low surface
area carbon black, Cabot
(Boston, MA)
Particle Size: HSCb - 14nm,
LSCb - 70nm
Reference: Whole-body Inhalation
Dose/Concentration: 0,1, 7, 50 mgfm3 HSCb;
50 mg/m3 LSCb (rats only)
Time to Analysis: 6 h/d, 5 dfwk for 13wks.
Parameters measured 1 d, 3mo, 11 mo post-
exposure
Body Weight: Environmental changes pre and post-
exposure affected test subjects' life spans,
particularly hamsters. Hamsters also experienced
significant loss of body weight when exposed to high
doses of HSCb.
Effects of Carbon Black: In rats, lung weight of the
high dose HSCb doubled. After 11 mo, analysis of all
lungs showed no significant difference. Mice had the
highest relative lung burdens at the end of exposure
time but also cleared particles faster at high doses
than rats. However, clearance slowed over the 11 mo
recovery period, especially in high dose mice.
Hamsters showed significant elevations in lung
carbon black burden for all exposures at all time
points. Hamsters exposed to high dose HSCb
exhibited impaired clearance.
BALF Cells: Presence of PMNs was limited to the
mid and high dose groups. Overall maximal response
was reached in mice and hamsters, but not in rats
with increasing mass dose of HSCb.
Reference: Evans
et al. (2006,
097066)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
DEP: collected under dry,
outdoor, ambient conditions
from tractor exhaust pipe
(1985, Japanese ISEK11500
cc tractor) burning Esso 2000
diesel and 20130 mixture of
Esso light engine oil.
10% UF, 90% fine
Cabosil: amorphous silicon
dioxide
16% UF, 84% fine
Particle Size: DEP: 30nm;
Cabosil: 7nm
Route: IT Instillation
Dose/Concentration: 1 mg/rat DEP; 1 mg/rat
Cabosil
Time to Analysis: Pretreatment with 0.5 unit
of bleomycin; IT 3 or 7d after pre treatment;
1wk post-IT
Lung permeability: In bleomycin-treated group,
obvious inflammatory status and edema within the
lung was observed. This was shown by significant
increases in acellular protein and free cells.
Changes in lung: Body weight ratio, lung surface
protein content, free cell counts, and apical surface
protein of rat type I cells were only altered by
bleomycin treatment and not particle exposure.
Reference:
Finnerty et al.
(2007,156434)
Species: Mouse
Gender: Male
Strain: C57BL/61
Age: 12 wks
Weight: 24.3
±0.3g
Coal Fly Ash (generated at U.S
EPA National Risk Management
Research Laboratory by burning
Montana subbituminous coal
under conditions simulating full-
scale utility boiler conditions)
Transition metals of Coal Fly
Ash: Fe, Mg, Ti, Mn, V
Particle Size: >PM2 b
Route: IT Instillation
Dose/Concentration: PM: 200mg PMfmouse;
9.1 mg/kg
PM+LPS10: 200mg PM + 10mg LPS
PM+LPS100: 200mg PM + 100mg LPS
LPS: 100 /yg
Time to Analysis: Parameters measured 18h
post single exposure
BALF Cells: No significant differences in |
concentration or white blood cell count in any groups
were observed. The percentage of neutrophils in-
creased significantly with PM+LPS100. PMN rose in
PM groups and increased further with LPS
treatment. Increases in PM+LPS were groups
statistically significant. More leukocytes were
present in the alveolar space in PM+LPS10
compared to the PM group. The most severe
response was in the PM+LPS100 group.
Cytokines: Plasma TNF-a and IL-6 significantly
increased for the PM+LPS100 group. An additive
effect of LPS and PM for IL-6 was observed. For
saline and PM groups, pulmonary TNF-a was below
detection range. A synergistic effect for TNF-a was
observed. A less than additive effect for IL-6 was
observed. Pulmonary TNF-a significantly increased in
the PM+ LPS100 group. Pulmonary IL-6 significantly
increased in both PM+LPS groups.
Reference:
Fujimaki et al.
(2006,096601)
Species: Mouse
Gender: Male
Strain: IL-6(-/-) and
WT: B6J129Sv
(control)
Age: 5-6wks
DEP: collected from a 4-
cylinder, 2.74 L, Isuzu diesel
engine
Particle Size: 0.4 /jm
Route: Whole-body Inhalation
Dose/Concentration: 1.0, 3.0 mgfm3
Time to Analysis: 12 h/d for 4wks. Parameters
measured 1d post-exposure
BALF Cells: Treatment significantly increased BAL
cells from WT mice at both dose levels. The increase
of macrophages and neutrophils were dose-
dependent. An increase in lymphocytes were present
in WT mice with the low dose. No significant
increase in cells were observed from IL-6 (¦/¦).
Cytokines: TNF-a largely increased in IL-6(-/-) mice
exposed to 3 mg/m3 compared to WT mice. IL-6
production increased in WT mice exposed to 3
mg/m3. CCL3 increased in both WT and IL-6(-/-) at
high dose. IL-1 (3 remained at the control level.
July 2009	D-86	DRAFT - DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Gerlofs-Nijland et
al. (2005, 088652)
Species: Rat
Gender: Male
Strain: SH/NHsd
Age: 11-12wks
Weight: 250-350g
RTD: road tunnel dust
(obtained from a Motorway
tunnel in Hendrik-ldo-Ambacht,
Netherlands)
EHC-93
(Ottawa, Canada)
Particle Size: Coarse: 2.5 - 10
fjm; fine: 0.1- 2.5 /jm
Route:IT Instillation
Dose/Concentration: 0.3,1, 3,10 mg/kg
EHC-93 10 mg/kg
Time to Analysis: Parameters measured
4, 24, 48h post single exposure
BALF Cells: PMN significantly increased in RTD (3
and 10 mg/kg dose) and EHC-93 exposed animals at
24 h and decreased by 48h but remained statistically
significant. AM numbers decreased for 3 mg/kg RTD
group at 4h.
BALF Inflammatory/Injury Markers: Myelop-
eroxidase (measured at 24 h in 1, 3,10 mg/kg RTD
groups) was elevated in a dose-dependent manner.
RTD induced time-dependent increases in LDH
activity at 24 and 48h, although these increases
were less than EHC-93 values at the same time
points. Alkaline phosphatase increased dose-
dependently for RTD at 48h. GSH decreased at 24 h
to approximately the same levels in 0.3,1, and 3
mg/kg RTD dose groups. Uric acid only decreased in
1 mg/kg RTD group at 24 h.
Cytokines: IL-6 levels were elevated only at 10
mg/kg for RTD and EHC-93 at 4 and 24 h; it
remained elevated for EHC-93 at 48h. A dose-
dependent increase in TNF-a at 4h for RTD was
observed. TNF-a levels remained elevated only for
the 10 mg/kg groups at 24 h and returned to control
levels by 48h. A dose-dependent increase in MIP-2
for all RTD dose groups were observed and remained
elevated through 48h for both PM types (although
values were returning to control levels).
Hematology: No significant changes in plasma for
bigET-1 or von Willebrand factor were observed. At
the highest dose, fibrinogen levels significantly
increased at 24 and 48h for both PM types.
Pulmonary histopathology: A dose-dependent
increase in the number of inflammatory foci at 24
and 48h for 3 and 10 mg/kg RTD groups was
observed. The response was even greater for the
EHC-93 exposed group at similar time points.
July 2009
D-87
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Gerlofs-Nijland et
al. (2007, 097840)
Species: Rat
Gender: Male
Strain: SH/NHsd
Age: 13wks
Weight: 250-350g
PM samples collected from:
1.	MOB high traffic density
2.	HIA high traffic density
3.	ROM high traffic density
4.	DOR moderate traffic
density
5	MGH low traffic density
6	LYC low traffic density
Particle Size: Ooarse: 2.5-10
fjm; Fine: 0.1 ¦ 2.5 fjm
Route: IT Instillation
Dose/Concentration: 3,10mg/kg
Time to Analysis: 24h
BALF Inflammatory/Injury Markers: LDH was
significantly increased for all doses of coarse PM and
for the high dose of fine PM. BALF protein
concentration was observed predominantly at the
high dose of coarse PM. Location ROM had evidence
of attenuated responses with fine PM. Ascorbate
concentrations were reduced but were only
significant for rats exposed to the highest dose of
coarse PM fractions from the locations MOD, HIA,
and LYC.
BALF Cells: Pulmonary inflammation was induced in
a significant and dose-dependent manner for both
dose levels. Inflammation in the BALF included
airway neutrophilia, increased macrophage numbers
and mild lymphocytosis. Both coarse and fine PM
caused dose-dependent aveolitis. Fine PM from LYO
(1 Omg/kg dose) also caused some bronchiolitis.
Cytokines: TNF-a concentrations increased for all
coarse samples with the exception of DOR and LYC.
Fine PM induced similar responses for all sites. MIIP-
2 concentrations increased only at certain sites for
coarse but not fine PM.
Hematology: Fibrinogen responses of SH rats
increased significantly at the high dose of both
fractions of all PM samples, except fine PM from
DOR.
Location-related differences: Coarse PM from
MOB, HIA and MGH induced higher LDH responses
than other locations. Coarse PM from HIA produced
BALF protein concentrations higher than LYC and
ROM. MGH induced greater amounts of BALF protein
than ROM. Coarse PM from LYC lowered fibrinogen
values more than PM from location MOB, HIA, and
MGH. Fine PM showed less differences among the
various sites.
Particle Correlation: Fine PM exhibited significant
correlation between zinc content and BALF
cytotoxicity markers protein and LDH - mainly from
HIA. Fine PM also exhibited positive correlations with
copper and barium. Coarse PM showed positive
correlation with barium and copper, mainly from
MOB.
Reference:
Gerlofs-Nijland et
al. (2009,190353)
Species: Rat
Gender: Male
Strain: SH
Age: 12wks
Weight: 200-300g
PM (Prague, Czech Republic;
Duisburg, Germany; Barcelona,
Spain) (Prague and Barcelona
coarse PM organic extracts)
Particle Size: Coarse: 2.5-10
fjm, Fine: 0.2-2.5/ym
Route: IT Instillation
Dose/Concentration: 7mg PM/kg body weight
Time to Analysis: DTPA added to some PM
samples preinstillation. Instilled with PM.
Necropsy 24h postexposure.
Inflammation (LDH, protein, albumin), cytotoxicity
(NAG, MPO, TNF-a), and fibrinogen were increased
by PM, and were greatest in the coarse PM fraction.
Metal-rich PM had greater inflammatory and
cytotoxic effects, and increased fibrinogen and vWF
and decreased ACE. PAH content influenced greater
inflammation (including neutrophils), cytotoxicity, and
fibrinogen. Generally, whole PM and coarse PM were
more potent than organic extracts and fine PM,
respectively.
Reference: Ghio et
al. (2005, 088272)
Species: Rat
Gender: Male
Strain: N8 b/b
Belgrade rats and
N8+ lb Belgrade
controls
Oil Fly Ash (Southern Research
Institute, Birmingham, AL)
Particle Size: 1.95 ± 0.18
/jm (MMAD)
Route: IT Instillation
Dose/Concentration: 500//g/rat; 2mg/kg
Time to Analysis: 24h
BALF Cells: Homozygous Belgrade with mutation
G185R had higher levels of Fe and V 24 h post-
exposure. This may demonstrate a decreased ability
to remove Fe and V from the lower respiratory tract
than heterozygous +lb littermates. This also
indicates that DMT1 is normally responsible for at
least some Fe and V uptake; thus, a defective DMT1
transports less.
BALF Inflammatory/Injury Markers: Increased
protein and LDH concentrations in the homozygous
strain were observed when compared to control
July 2009
D-88
DRAFT-DO NOT CITE OR QUOTE

-------
Reference	Pollutant	
Reference: Ghio et Ferric ammonium citrate (FAC)
al. (2005, 088275) Vanady| su|fate (V0Sq4)
Species: Rat Particle Size: NR
Gender: Male
Strain: Sprague-
Dawley
Age: 60d
Weight: 250-300g
	Exposure	
Route: IT Instillation
Dose/Concentration: 0.5 mL 100//m FAC/rat;
0.5 mL 10 fjm V0S04/rat; 500 /yg oil fly ash; 2
mg/kg
Time to Analysis: Single or double exposure
with 24 h rest period. Parameters measured 15,
30, 60 min, 24h post exposure.
Effects
DMT1 immunohistochemistry and lung injury:
FAC increased and V0S04 decreased -IRE DMT 1
staining. Same exposures had no effect on +IRE
DMT1. -IRE DMT1 expression in macrophages,
airway and alveolar epithelial cells increased with
increased Fe exposure. Vanadium nearly eliminated
staining except in alveolar macrophages. Increased
metal clearance with pre exposure to FAC. Less
metal clearance with pre exposure to VOSO4. Pre-
exposure to iron diminished lung injury whereas pre-
exposure to vandium increased lung injury after oil fly
ash instillation. Lung injury measured by
concentration of protein and LDH in BAL.
Reference: Gilmour
et al. (2007,
096433)
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 10-12wks
Weight: 20-22g
PM-C0, Fl, UF
(obtained from U.S. Seattle (S),
Salt Lake City (SL), South
Bronx (SB), Sterling Forest
(SF))
SB: included 35% sulfate, 22%
gasoline, diesel and brake
wear.
SF: 48% sulfate.
SL: 34% wood combustion and
28% sulfate
S: 39% wood combustion and
29% sulfate
Residual oil combustion and soil
dust less than 5% for all sites.
Particle Size: CO: 2.5-10/ym;
Fl: ~ 2.5/jm; UF: ~ 0.1/ym
Route: Oropharyngeal Aspiration
Dose/Concentration: 25//g or 100/yg PM; 1.25
or 5 mg/kg
Time to Analysis: 18h
BALF Inflammatory/Injury Markers: Seattle CO
fractions showed no dose-dependent effect on
protein concentration. Results for other locations
were distinctly higher with 100 /yg dose than 25 /yg
and saline doses. SL CO high dose induced the most
significant increase. LDH response was weakly dose-
related. Only SB showed a statistically significant
increase for LDH with the high dose UF.
BALF Cells: PMN increased with the high dose of
CO samples from SB, SL, S, but not SF. No
significant increases from Fl were observed, though
the high dose induced increased PMN. UF from SL
caused a highly variable response.
Cytokines: MIP-2 was similar to PMN response. SB
CO induced the most significant response. SL UF was
highly variable.
Particle Characteristics: LPS was higher in S (CO,
Fl, UF) and SL (CO, Fl, UF). Zn levels were highest in
SB (CO, Fl, UF). Fe was higher in all CO and Fl
samples with SB CO inducing the highest.
Reference:
Gilmour et al.
(2004, 057420)
Species: Mouse
Gender: Female
Strain: CD1
Age: 8-10wks
Weight: 20-25g
Coal Fly Ash
MU: Montana Ultrafine
MF: Montana Fine
MC: Montana Coarse
KF: W.Kentucky Fine
KC: W. Kentucky Coarse
Particle Characteristics:
Montana Sulfur 0.83%, Ash
11.72%. Trace amounts of Be
P, Sr, V, Nb, Cd, Se, Ga, Cu.
Depleted in Si, Al, Fe, Mg, Ti.
Kentucky Sulfur 3.11 %, Ash
8.07%
Particle Size: Coarse:
fj m;
Fine: <2.5//m;
Ultrafine: < 0.2 fjm
Route: Oropharyngeal Aspiration
Dose/Concentration: 25ug or 100//gfmouse
Time to Analysis: 18h
>2.5
BALF Cells: PMN highly increased for MU at both
doses. The level was comparable to the positive
control. PMN also increased with KF at high dose.
Coarse particles caused no significant increase in
PMN. Number of macrophages did not change, but
NAG increased significantly with MU for both dose
levels and with KF and MF at high dose level.
BALF Inflammatory/Injury Markers: Total protein
and LDH was not significantly elevated. Albumin
concentration increased significantly after treatment
with the fine high dose of both particle types.
Cytokines: MU particles caused a significant
increase in TNF-a. MIP-2 increased in all fine and
ultrafine PM-instilled animals with the highest in the
MU and KF at both doses. IL-6 was detectable only
in the BALF of MU and KF with substantial
variability. The IL-6 levels were not significant.
July 2009
D-89
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Gilmour PM (collected from precipitator
et al. (2004,
087948)
Species: Rat
Gender: Male
Strain: SH/NQIBR,
WKY
Age: 12wks
Weight: 280-340g
unit of an oil burning power
plant in Boston)
Measured Components of PM:
S, Zn, Ni, V, Al, Cu, Pb, Fe, Ca,
Na, K, Mg, Endotoxin
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 0.0, 0.83, 3.3, and 8.3
mg/kg in SH rats; 0.0 or 3.3 mg/kg in WKY and
SH rats
Time to Analysis: 24h
BALF Inflammatory/Injury Markers: LDH activity
increased in a dose-related manner; this was
observed in SH rats after exposure to 0.83, 3.33 and
8.3 mg/kg PM. SH rats showed greater lung
permeability following PM exposure than WKY rats.
SH rats showed acute lung inflammatory response
after exposure to PM when compared to WKY rats.
BALF Cells: No increase in macrophage number was
observed in either rat strain following saline or PM
exposure at 24 h.
Cytokines: MIP2 mRNA expression increased
significantly in SH PM exposure group only. No
significant differences in TNF-a RNA expression in
either WKY, SH rats or control treatment groups
were observed.
CD14: A significant increase in lung CD14 protein
was observed only in SH rats exposed to PM.
TLR4: A significant increase in TLR4 protein in SH
rats exposed to PM was observed.
NF-kB: A significant increase in NF-kB binding
protein in the nuclei of SH rats exposed to PM was
observed. This effect was not observed in the control
of PM-exposed WKY rats.
Reference: Gilmour ufCB: Ultrafine carbon black
et al. (2004,
054175)
Species: Rat
Gender: Male
Strain: Wistar
Age: 12wks
(Printex 90 (Degussa)
CB: (Huber 990, HR. Haeffner
and Co)
Particle Size: ufCB: 14nm;
CB: 260nm (primary particle
diameter)
Route: Whole-body Inhalation
Dose/Concentration: ufCB: 1.66 mgfm3
fCB: 1.40 mgfm3
Number concentrations
ufCB: 52380 particles/cm3
fCB: 3800 particles/cm3
Time to Analysis: Exposed for 7 h. Sacrificed
0,16 or 48h post exposure.
BAL Cells: Total number of cells increased
significantly in UfCB-exposed rats at 0 and 16 h.
Recruitment of cells did not occur in response to CB
exposure. PMNs increased significantly in the BALF
of ufCB-exposed rats at 16 h. Leukocytes remained
unchanged following CB exposure but increased
significantly at 0 and 48 h post exposure to ufCB.
Cytokine mRNA: A significant increase in BALF
MIP-2 mRNA expression was observed at 48 h. No
differences in MIP-2 mRNA levels were observed in
the whole lung tissue.
Reference:
Godleski et al.
(2002,156478)
Species: Rat
Gender: Male
Strain: SD
Age: NR
Weight: 200-250g
CAPs (Boston; Harvard
Ambient Particle Concentrator)
Particle Size: 0.27±2.3/jm
(diameter)
Route: Inhalation
Dose/Concentration: 73.5-733//gfm3
Time to Analysis: Exposed 5hfd, 3d
(consecutive). BAL 24h postexposure
PMNs significantly increased with CAPs exposure
and also in relation to CAPs mass, Br, SO42', EC, 0C
and Pb. An overall increase in pro-inflammatory
mediators and decrease in immune enhancer and
evidence of vascular endothelial responses occurred
with CAPs exposure.
Reference:
Gottipolu et al.
(2009,190360)
Species: Rat
Gender: Male
Strain: Wistar
Kyoto (WKY), SH
Age: 14-16wks
Weight: NR
DE (30-kW (40hp) 4-cylinder
indirect injection Deutz diesel
engine) (02- 20%, CO-1.3-
4.8ppm, NO- <2.5-5.9ppm,
NO?- < 0.25-1,2ppm, S02- 0.2-
0.3ppm, OC/EC- 0.3±0.03)
Particle Size: Number Median
Diameter: Low- 83±2nm, High-
88.2nm; Volume Median
Diameter: Low- 207 ±2nm,
High- 225±2nm
Route: Exposure Chamber
Dose/Concentration: Low- 507 ±4/yg/m3, High-
2201 ±14 //g/m3
Time to Analysis: Exposed 4hfd, 5dfwk, 4wks.
Necropsied 1d postexposure.
DE increased neutrophils in a concentration-
dependent manner, and GGT activity at the high
dose. Particle-laden macrophages were found in DE-
exposed rats. DE dose-dependently inhibited
mitochondrial aconitase activity. DE caused 377
genes to be differentially expressed within WKY rats,
most of which were downregulated, but none in SH
rats. However, WKY rats had an expression pattern
shift that mimicked baseline expression of SH rats
without DE. These genes regulated compensatory
response, matrix metabolism, mitochondrial function,
and oxidative stress response.
Reference:
Gunnison and Chen
(2005, 087956)
Species: Mouse
Gender: Male
Strain: DK (ApoE
LDLr'1')
Age: 18-20wks
CAPS
(Northeastern regional back-
ground)
Ambient air copollutants
measured O3, NO?
Particle Size: 389 ± 211 m
Route: Whole-body Inhalation
Dose/Concentration: CAPS — 131 ± 99//g/m3
including O3 - 10 ppb and NO2 - 4.4 ppb
Time to Analysis: 6h/d, 5d/wk for 4m (5/12/03-
9/5/03). Sacrified 3-4d post exposure.
Microarray Data: 13 genes in the heart tissue and
47 genes in the lung tissue were identified as
possibly affected. Strict standards (1.5 fold
response, 10 % false discovery rate) resulted in
responses by only 1/13 genes (Rex3 - no known
heart physiology) in the heart tissue and 0/47 genes
in the lung tissue. Using more liberal response
(nonstatistical) standards (1.5 fold only) and
comparison of each CAPS animal with all 3 control
animals (3x3 array) resulted in possible effects on 7
additional genes in the heart tissue and 37 genes in
the lung tissue.
July 2009	D-90	DRAFT - DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Gurgueira et al.
(2002, 036535)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Weight: 250-300g
CAPs
(Harvard Ambient Particle
Concentrator)
CB
(C198 Fischer Scientific, Pitts-
burg, PA USA)
Composed of 85.9± 0.2%
Carbon, 13.0± 0.2% 02,
1.17± 0.2% Sulfur
ROFA
(Boston, MA USA oil-fired
power plant)
Particle Size: CAPs: 1-2.5
/jm; CB: < 2.5/jm; ROFA:
< 2.5 /jm
Route: Whole-body Inhalation
Dose/Concentration: 300 ± 60 //gfm3
Time to Analysis: 1, 3, 5h CAPs Exposure
followed by immediate post-exposure analysis.
5h CB, immediate analysis.
30min ROFA, Immediate analysis.
In situ Chemiluminescence(CL): Data show a
significant increase in lung and heart CL at 5h. Lung
CL increased linearly with time of exposure.
Oxidants: CAPs-initiated oxidative stress was not
detectable in those rats allowed to recover in room
air after the simulated "peak" in particulate air
pollution. Rats breathing particle-free filtered air for
3 days had significantly lower levels of oxidants.
Exposure to inert CB did not exert oxidant effects on
the heart and lung.
BALF Inflammatory/Injury Markers: The water
content of the lung and heart increased significantly
upon exposure to CAPs but not to filtered air and
increased as a function of length of exposure. Rats
breathing CAPs also showed increases in LDH and
CPK as a function of length of exposure.
Antioxidant enzymes: Data showed an increase in
SOD and catalase activities in both the lung and
heart. The pattern of increase was tissue specific.
Reference: Hamoir PSC: Polystyrene particles, Route: IT Instillation
et al. (2003,
096664)
Species: Rabbit
Strain: New
Zealand
Age: 12-16 wks
Weight: 2.8 ±
0.5kg
Carboxylate modified, 3 types
PSA: Polystyrene particles,
Amine modified, 1 type
Particle Size: PSC: 24,110 or
190nm (PSC24, PSC 110,
PSC190); PSA: 190nm
Dose/Concentration: PSC24: 0.04 or 4
mg/rabbit
PSC110, PSC190, PSA190: 4mgfrabbit
Time to Analysis: 0, 30, 60, 90,120min
Capillary Filtration Coefficient: A time-c
increase correlating to total number of
particles/surface area, not particle size, was
observed.PSA induced a significant increase in mi-
crovascular permeability as compared to PSC. This
suggests that the number of particles exposed should
be considered an important parameter for measuring
air quality rather than total particle surface area.
July 2009	D-91	DRAFT - DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Happo
et al. (2007,
096630)
Species: Mouse
Gender: Male
Strain: C57BL/6J
Weight: 19-30g
Age: 10-11 wks
PMC (Coarse)
PMF (Fine)
PMUF (Ultrafine)
Collected in 6 European cities:
Duisburg, Prague, Amsterdam,
Helsinki, Barcelona, Athens
Particle Size: PMC: PMio2.5;
PMF: PM2.B.0.2: PMUF: PM0.2
Route: IT Instillation
Dose/Concentration: PMC: 5.9-29.6//g/m3;
PMF: 8.3-25.2//g/m3; PMUF: 2.7-6.7 //g/m3
Dose-response: 1, 3,10mg/kg
Time course: 10 mg/kg
Time to Analysis: 1. Dose-Response study:
parameters measured 24h post exposure. 2.
Time course study: parameters measured 4,12,
24 h post single exposure (at 10 mg/kg).
Total Cell Numbers: 1. For the dose-response
study, all the PMC samples exhibited dose-dependent
increases of total cell numbers. The 3 and 10 mg/kg
doses of PMC induced statistically significant
increases. At 10 mg/kg, only 2/6 samples induced
statistically significant increases. No PMUF samples
induced effects at any dose. 2. For the time-response
study, no increases in cell numbers were shown at 4
h. Though the levels induced by PMC at 24 h were
lower than at 12 h, both levels were statistically
significant. PMF induced statistically significant
increases only at 12 h for 4/6 samples. PMUF
induced only 1 significant increase at 12 h; the 24 h
time point was not tested.
Total Protein/LDH: 1. The lower doses of 1 and 3
mg/kg did not induce significant increases in any of
the PM samples, except for PMUF-Athens. All 6
samples of PMC, at 10 mg/kg, induced significant
increases. At 10 mg/kg, 4/6 PMF samples induced
significant increases. 2. At 4 h, none of the samples
increased protein concentration. The PMC samples,
excluding Prague, induced significantly higher
concentrations at 12 h. At 24 h, only 3/6 PMC
samples induced significant increases. Only 2 PMF
samples induced significant increases at 12 and 24
h. At 12 h, effects induced by PMUF were minimal
and inconsistent: the 24 h time point was not tested.
Cytokine: 1. Only PMC induced dose-dependent
responses that reached statistical significance at 10
mg/kg. PMF and PMUF induced minimal and
inconsistent responses.
TNF-a levels: 2. TNF-a levels increased
significantly at 4 and 12 h by PMC. At 24 h, TNF-a
levels returned to near control levels. PMF, at 4 h,
induced statistically significant increases for 3/6
samples and significant increases in 2/6 samples at
12 h. No PMUF samples significantly increased TNF-
a levels.
IL-6: 2. PMC induced the highest IL-6 levels at 4 h.
Levels at 12 and 24 h were reduced with 6/6 and 3/6
samples showing statistically significant increases,
respectively. PMF showed a similar trend with 4 h
inducing the highest levels that were reduced at 12
and 24 h. Of the PMUF samples, only the Helsinki
and Duisburg samples induced statistically significant
results at 4 and 12 h. Generally, the PMUF responses
were negligible when compared to PMC and PMF.
KC production: 2. All PMC samples induced the
highest levels at 4 h. At 12 and 24 h, levels were
reduced but 4/6 samples induced statistically
significant levels. PMF showed a similar trend- the
highest levels were induced at 4 h (in 3/6 samples).
PMUF at 4 h showed small, though not significant,
increases. At 12 h, only 2 samples showed
statistically significant differences from the control:
the 24 h time point was not tested.
Reference: Harder Carbon UFP
et al. (2005,
087371)
Species: Rat
Gender: Male
Strain: Wistar
Kyoto
Age: 14-17wks
Weight: NR
Particle Size: Diameter:
37.6±0.7nm
Route: Exposure Chamber
Dose/Concentration: 180/yg/m3
Time to Analysis: Telemeter implanted into
peritoneal cavity. 10d recovery. 3d baseline
reading. 24h exposure. 3d recovery.
Carbon UFP mildly but significantly elevated HR
compared to the control. SDNN was significantly
decreased during exposure. UFP induced mild
pulmonary inflammation, significantly increased
PMN, and increased the total protein and albumin
concentrations. Particle-laden macrophages
sporadically accumulated in the alveolar region.
July 2009	D-92	DRAFT - DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Harkema et al.
(2004, 056842)
Species: Rat
Gender: Male
Strain: F344, BN
Age: 10-12wks
Weight: NR
CAPs (Detroit; July-Sept.
2000; Harvard Ambient Fine
Particle Concentrator)
Particle Size: 2.5 /jm
(diameter)
Route: Inhalation Exposure Chamber. IT
Instillation.
Dose/Concentration: 4d concentration:
676±288//g/m3, 5d concentration: 313±119
/yg/m3, July concentration: 16-185 //g|m3,
September concentration: 81 -755/yglm3; IT
Instillation- 200 /jl (soluble and insoluble)
Time to Analysis: F344 rats sensitized to
endotoxin, BN rats to OVA. Exposed 10h/d 1,4,
5d (consecutive). Another group of rats i.t
instilled. Both groups killed 24h postexposure.
The retention of PM in the airways was enhanced by
allergic sensitization. Recovery of anthropogenic
trace elements was greatest for CAPs-exposed rats.
Temporal increases in these elements were
associated with eosinophil influx, BALF protein
content and increased airway mucosubstances. A
mild pulmonary neutrophilic inflammation was
observed in rats instilled with the insoluble fraction
but instillation of total, soluble or insoluble PM2.6 in
allergic rats did not result in differential effects.
Reference: Elder et
al. (2004, 055642)
Species: Rat
Gender: Male
Strain: Fisher344,
SH
Age: 23m (Fisher),
11 -14m (SH)
UFP: argon-filled chamber with
electric arc discharge (TSI, Inc.,
St. Paul, MN)
Particle Size: 36nm
Route: Whole-body Inhalation. Intraperitoneal
(ip) for saline and LPS
Dose/Concentration: UFP: 150/yg/m3 bw; LPS:
2 mg/kg
Time to Analysis: Parameters measured post
single exposure of 6h, 18h
BALF Cells: Neither inhaled UFP nor LPS cause a
significant increase in BAL fluid total cells or
percentage of neutrophils in either rat strain. No
significant exposure-related alteration in total protein
concentration was observed. In both rat strains LPS
induced a significant increase in the amount of
circulating PMNs. When combined with inhaled UFP,
PMNs decreased; for F-344 rats, this decrease was
significant.
BALF Inflammatory/Injury Markers: Plasma
fibrinogen increased with LPS in both rat strains with
the magnitude of change greater in SH rats. UFP
alone decreased plasma fibrinogen in SH rats.
Combined UFP and LPS response was blunted but
significantly higher than controls. Hematocrit was
not altered in either rat strain by any treatment. No
change in activites of LDH and b-glucuronidase was
observed.
TAT complexes: With all exposure groups averaged,
plasma TAT complexes in SH rats were 6.5 times
higher than in F-344 rats. LPS caused an overall
increase in TAT complexes for F-344 rats which was
further augmented by inhaled UFP. UFP alone
induced a decreased response. In SH rats, UFP alone
exhibited a significantly increased response and LPS
exhibited a decreased response.
ROS in BALF: In F-344 rats, both UFP and LPS have
independent and significant effects on DCFD
oxidation. Effects were in opposite directions;
particles decreased ROS whereas LPS increased
ROS.
Reference: Elder et
al. (2004, 087354)
Species: Rat
Gender: Male
Strain: F344
Age: 21m
Freshly generated vehicle
exhaust emissions from I-90
between Rochester and
Buffalo, NY
Particle Size: NR
Route: Whole-body Inhalation; (IT Instillation:
Influenza)
Dose/Concentration: Vehicle exhaust: 0.95-
3.13 x 106 particles/cm3
Endotoxin: 84 EU
Influenza (IV): 10, 000 EID 50 in 250ul
Time to Analysis: 1x6h, 3x6h or both.
Parameters measured 18h post-exposure. 48h
prior to on-road exposures, instilled
intratracheally with IV. Immediate pre exposure
of priming agent endotoxin.
EXPERIMENTS
1: LPS + PM 6h
2: LPS + PM 6h, 3 x 6h
3: IV + PM 6 h
4: IV + PM 6h, 3x 6 h
No departures from normal baseline cellular or
biochemical values were observed, suggesting that
on-road exposures were well tolerated by the rats.
BALF Inflammatory/Injury Markers: Increase in
total protein concentration, LDH and B-glucuronidase
activities were observed.
Specific results according to groups 1-4 are as
follows:
Experiment 1: No endpoints revealed significant
differences between groups of rats exposed to gas
phase only versus the gas-phase/particle mixture.
Experiment 2: Combination of endotoxin and
particles produced greater inflammatory responses
than those treated with saline and particles post 1 d.
After 3 days, no statistically significant changes
were noted.
Experiment 3: Influenza virus significantly increased
ROS release in BAL cells.
Experiment 4: Influenza virus significantly increased
both percentage of PMNs in lavage fluid and BAL cell
ROS release.
July 2009	D-93	DRAFT - DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Elder et
al. (2005, 088194)
Species: Rat,
Mouse, Syrian
Golden Hamster
Gender: Female
Strain: F-344,
B6C3F1, FIB
HSCb: Printex-90 high surface
area carbon black, Deguss-
Huels (Trostberg, Germany).
LSCb: Sterling V, low surface
area carbon black, Cabot
(Boston, MA)
Particle Size: HSCb - 14nm,
LSCb - 70nm
Reference: Whole-body Inhalation
Dose/Concentration: 0,1, 7, 50 mgfm3 HSCb;
50 mg/m3 LSCb (rats only)
Time to Analysis: 6 h/d, 5 dfwk for 13wks.
Parameters measured 1 d, 3mo, 11 mo post-
exposure
Body Weight: Environmental changes pre and post-
exposure affected test subjects' life spans,
particularly hamsters. Hamsters also experienced
significant loss of body weight when exposed to high
doses of HSCb.
Effects of Carbon Black: In rats, lung weight of the
high dose HSCb doubled. After 11 mo, analysis of all
lungs showed no significant difference. Mice had the
highest relative lung burdens at the end of exposure
time but also cleared particles faster at high doses
than rats. However, clearance slowed over the 11 mo
recovery period, especially in high dose mice.
Hamsters showed significant elevations in lung
carbon black burden for all exposures at all time
points. Hamsters exposed to high dose HSCb
exhibited impaired clearance.
BALF Cells: Presence of PMNs was limited to the
mid and high dose groups. Overall maximal response
was reached in mice and hamsters, but not in rats
with increasing mass dose of HSCb.
Reference: Evans
et al. (2006,
097066)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
DEP: collected under dry,
outdoor, ambient conditions
from tractor exhaust pipe
(1985, Japanese ISEK11500
cc tractor) burning Esso 2000
diesel and 20130 mixture of
Esso light engine oil.
10% UF, 90% fine
Cabosil: amorphous silicon
dioxide
16% UF, 84% fine
Particle Size: DEP: 30nm;
Cabosil: 7nm
Route: IT Instillation
Dose/Concentration: 1 mg/rat DEP; 1 mg/rat
Cabosil
Time to Analysis: Pretreatment with 0.5 unit
of bleomycin; IT 3 or 7d after pre treatment;
1wk post-IT
Lung permeability: In bleomycin-treated group,
obvious inflammatory status and edema within the
lung was observed. This was shown by significant
increases in acellular protein and free cells.
Changes in lung: Body weight ratio, lung surface
protein content, free cell counts, and apical surface
protein of rat type I cells were only altered by
bleomycin treatment and not particle exposure.
Reference:
Finnerty et al.
(2007,156434)
Species: Mouse
Gender: Male
Strain: C57BL/61
Age: 12 wks
Weight: 24.3
±0.3g
Coal Fly Ash (generated at U.S
EPA National Risk Management
Research Laboratory by burning
Montana subbituminous coal
under conditions simulating full-
scale utility boiler conditions)
Transition metals of Coal Fly
Ash: Fe, Mg, Ti, Mn, V
Particle Size: >PM2 b
Route: IT Instillation
Dose/Concentration: PM: 200mg PMfmouse;
9.1 mg/kg
PM+LPS10: 200mg PM + 10mg LPS
PM+LPS100: 200mg PM + 100mg LPS
LPS: 100 /yg
Time to Analysis: Parameters measured 18h
post single exposure
BALF Cells: No significant differences in |
concentration or white blood cell count in any groups
were observed. The percentage of neutrophils in-
creased significantly with PM+LPS100. PMN rose in
PM groups and increased further with LPS
treatment. Increases in PM+LPS were groups
statistically significant. More leukocytes were
present in the alveolar space in PM+LPS10
compared to the PM group. The most severe
response was in the PM+LPS100 group.
Cytokines: Plasma TNF-a and IL-6 significantly
increased for the PM+LPS100 group. An additive
effect of LPS and PM for IL-6 was observed. For
saline and PM groups, pulmonary TNF-a was below
detection range. A synergistic effect for TNF-a was
observed. A less than additive effect for IL-6 was
observed. Pulmonary TNF-a significantly increased in
the PM+ LPS100 group. Pulmonary IL-6 significantly
increased in both PM+LPS groups.
Reference:
Fujimaki et al.
(2006,096601)
Species: Mouse
Gender: Male
Strain: IL-6(-/-) and
WT: B6J129Sv
(control)
Age: 5-6wks
DEP: collected from a 4-
cylinder, 2.74 L, Isuzu diesel
engine
Particle Size: 0.4 /jm
Route: Whole-body Inhalation
Dose/Concentration: 1.0, 3.0 mgfm3
Time to Analysis: 12 h/d for 4wks. Parameters
measured 1d post-exposure
BALF Cells: Treatment significantly increased BAL
cells from WT mice at both dose levels. The increase
of macrophages and neutrophils were dose-
dependent. An increase in lymphocytes were present
in WT mice with the low dose. No significant
increase in cells were observed from IL-6 (¦/¦).
Cytokines: TNF-a largely increased in IL-6(-/-) mice
exposed to 3 mg/m3 compared to WT mice. IL-6
production increased in WT mice exposed to 3
mg/m3. CCL3 increased in both WT and IL-6(-/-) at
high dose. IL-1 (3 remained at the control level.
July 2009	D-94	DRAFT - DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Gerlofs-Nijland et
al. (2005, 088652)
Species: Rat
Gender: Male
Strain: SH/NHsd
Age: 11-12wks
Weight: 250-350g
RTD: road tunnel dust
(obtained from a Motorway
tunnel in Hendrik-ldo-Ambacht,
Netherlands)
EHC-93
(Ottawa, Canada)
Particle Size: Coarse: 2.5 - 10
fjm; fine: 0.1- 2.5 /jm
Route:IT Instillation
Dose/Concentration: 0.3,1, 3,10 mg/kg
EHC-93 10 mg/kg
Time to Analysis: Parameters measured
4, 24, 48h post single exposure
BALF Cells: PMN significantly increased in RTD (3
and 10 mg/kg dose) and EHC-93 exposed animals at
24 h and decreased by 48h but remained statistically
significant. AM numbers decreased for 3 mg/kg RTD
group at 4h.
BALF Inflammatory/Injury Markers: Myelop-
eroxidase (measured at 24 h in 1, 3,10 mg/kg RTD
groups) was elevated in a dose-dependent manner.
RTD induced time-dependent increases in LDH
activity at 24 and 48h, although these increases
were less than EHC-93 values at the same time
points. Alkaline phosphatase increased dose-
dependently for RTD at 48h. GSH decreased at 24 h
to approximately the same levels in 0.3,1, and 3
mg/kg RTD dose groups. Uric acid only decreased in
1 mg/kg RTD group at 24 h.
Cytokines: IL-6 levels were elevated only at 10
mg/kg for RTD and EHC-93 at 4 and 24 h; it
remained elevated for EHC-93 at 48h. A dose-
dependent increase in TNF-a at 4h for RTD was
observed. TNF-a levels remained elevated only for
the 10 mg/kg groups at 24 h and returned to control
levels by 48h. A dose-dependent increase in MIP-2
for all RTD dose groups were observed and remained
elevated through 48h for both PM types (although
values were returning to control levels).
Hematology: No significant changes in plasma for
bigET-1 or von Willebrand factor were observed. At
the highest dose, fibrinogen levels significantly
increased at 24 and 48h for both PM types.
Pulmonary histopathology: A dose-dependent
increase in the number of inflammatory foci at 24
and 48h for 3 and 10 mg/kg RTD groups was
observed. The response was even greater for the
EHC-93 exposed group at similar time points.
July 2009
D-95
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Gerlofs-Nijland et
al. (2007, 097840)
Species: Rat
Gender: Male
Strain: SH/NHsd
Age: 13wks
Weight: 250-350g
PM samples collected from:
1.	MOB high traffic density
2.	HIA high traffic density
3.	ROM high traffic density
4.	DOR moderate traffic
density
5	MGH low traffic density
6	LYC low traffic density
Particle Size: Ooarse: 2.5-10
fjm; Fine: 0.1 ¦ 2.5 fjm
Route: IT Instillation
Dose/Concentration: 3,10mg/kg
Time to Analysis: 24h
BALF Inflammatory/Injury Markers: LDH was
significantly increased for all doses of coarse PM and
for the high dose of fine PM. BALF protein
concentration was observed predominantly at the
high dose of coarse PM. Location ROM had evidence
of attenuated responses with fine PM. Ascorbate
concentrations were reduced but were only
significant for rats exposed to the highest dose of
coarse PM fractions from the locations MOD, HIA,
and LYC.
BALF Cells: Pulmonary inflammation was induced in
a significant and dose-dependent manner for both
dose levels. Inflammation in the BALF included
airway neutrophilia, increased macrophage numbers
and mild lymphocytosis. Both coarse and fine PM
caused dose-dependent aveolitis. Fine PM from LYO
(1 Omg/kg dose) also caused some bronchiolitis.
Cytokines: TNF-a concentrations increased for all
coarse samples with the exception of DOR and LYC.
Fine PM induced similar responses for all sites. MIIP-
2 concentrations increased only at certain sites for
coarse but not fine PM.
Hematology: Fibrinogen responses of SH rats
increased significantly at the high dose of both
fractions of all PM samples, except fine PM from
DOR.
Location-related differences: Coarse PM from
MOB, HIA and MGH induced higher LDH responses
than other locations. Coarse PM from HIA produced
BALF protein concentrations higher than LYC and
ROM. MGH induced greater amounts of BALF protein
than ROM. Coarse PM from LYC lowered fibrinogen
values more than PM from location MOB, HIA, and
MGH. Fine PM showed less differences among the
various sites.
Particle Correlation: Fine PM exhibited significant
correlation between zinc content and BALF
cytotoxicity markers protein and LDH - mainly from
HIA. Fine PM also exhibited positive correlations with
copper and barium. Coarse PM showed positive
correlation with barium and copper, mainly from
MOB.
Reference:
Gerlofs-Nijland et
al. (2009,190353)
Species: Rat
Gender: Male
Strain: SH
Age: 12wks
Weight: 200-300g
PM (Prague, Czech Republic;
Duisburg, Germany; Barcelona,
Spain) (Prague and Barcelona
coarse PM organic extracts)
Particle Size: Coarse: 2.5-10
fjm, Fine: 0.2-2.5/ym
Route: IT Instillation
Dose/Concentration: 7mg PM/kg body weight
Time to Analysis: DTPA added to some PM
samples preinstillation. Instilled with PM.
Necropsy 24h postexposure.
Inflammation (LDH, protein, albumin), cytotoxicity
(NAG, MPO, TNF-a), and fibrinogen were increased
by PM, and were greatest in the coarse PM fraction.
Metal-rich PM had greater inflammatory and
cytotoxic effects, and increased fibrinogen and vWF
and decreased ACE. PAH content influenced greater
inflammation (including neutrophils), cytotoxicity, and
fibrinogen. Generally, whole PM and coarse PM were
more potent than organic extracts and fine PM,
respectively.
Reference: Ghio et
al. (2005, 088272)
Species: Rat
Gender: Male
Strain: N8 b/b
Belgrade rats and
N8+ lb Belgrade
controls
Oil Fly Ash (Southern Research
Institute, Birmingham, AL)
Particle Size: 1.95 ± 0.18
/jm (MMAD)
Route: IT Instillation
Dose/Concentration: 500//g/rat; 2mg/kg
Time to Analysis: 24h
BALF Cells: Homozygous Belgrade with mutation
G185R had higher levels of Fe and V 24 h post-
exposure. This may demonstrate a decreased ability
to remove Fe and V from the lower respiratory tract
than heterozygous +lb littermates. This also
indicates that DMT1 is normally responsible for at
least some Fe and V uptake; thus, a defective DMT1
transports less.
BALF Inflammatory/Injury Markers: Increased
protein and LDH concentrations in the homozygous
strain were observed when compared to control
July 2009
D-96
DRAFT-DO NOT CITE OR QUOTE

-------
Reference	Pollutant	
Reference: Ghio et Ferric ammonium citrate (FAC)
al. (2005, 088275) Vanady| su|fate (V0Sq4)
Species: Rat Particle Size: NR
Gender: Male
Strain: Sprague-
Dawley
Age: 60d
Weight: 250-300g
	Exposure	
Route: IT Instillation
Dose/Concentration: 0.5 mL 100//m FAC/rat;
0.5 mL 10 fjm V0S04/rat; 500 /yg oil fly ash; 2
mg/kg
Time to Analysis: Single or double exposure
with 24 h rest period. Parameters measured 15,
30, 60 min, 24h post exposure.
Effects
DMT1 immunohistochemistry and lung injury:
FAC increased and V0S04 decreased -IRE DMT 1
staining. Same exposures had no effect on +IRE
DMT1. -IRE DMT1 expression in macrophages,
airway and alveolar epithelial cells increased with
increased Fe exposure. Vanadium nearly eliminated
staining except in alveolar macrophages. Increased
metal clearance with pre exposure to FAC. Less
metal clearance with pre exposure to VOSO4. Pre-
exposure to iron diminished lung injury whereas pre-
exposure to vandium increased lung injury after oil fly
ash instillation. Lung injury measured by
concentration of protein and LDH in BAL.
Reference: Gilmour
et al. (2007,
096433)
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 10-12wks
Weight: 20-22g
PM-C0, Fl, UF
(obtained from U.S. Seattle (S),
Salt Lake City (SL), South
Bronx (SB), Sterling Forest
(SF))
SB: included 35% sulfate, 22%
gasoline, diesel and brake
wear.
SF: 48% sulfate.
SL: 34% wood combustion and
28% sulfate
S: 39% wood combustion and
29% sulfate
Residual oil combustion and soil
dust less than 5% for all sites.
Particle Size: CO: 2.5-10/ym;
Fl: ~ 2.5/jm; UF: ~ 0.1/ym
Route: Oropharyngeal Aspiration
Dose/Concentration: 25//g or 100/yg PM; 1.25
or 5 mg/kg
Time to Analysis: 18h
BALF Inflammatory/Injury Markers: Seattle CO
fractions showed no dose-dependent effect on
protein concentration. Results for other locations
were distinctly higher with 100 /yg dose than 25 /yg
and saline doses. SL CO high dose induced the most
significant increase. LDH response was weakly dose-
related. Only SB showed a statistically significant
increase for LDH with the high dose UF.
BALF Cells: PMN increased with the high dose of
CO samples from SB, SL, S, but not SF. No
significant increases from Fl were observed, though
the high dose induced increased PMN. UF from SL
caused a highly variable response.
Cytokines: MIP-2 was similar to PMN response. SB
CO induced the most significant response. SL UF was
highly variable.
Particle Characteristics: LPS was higher in S (CO,
Fl, UF) and SL (CO, Fl, UF). Zn levels were highest in
SB (CO, Fl, UF). Fe was higher in all CO and Fl
samples with SB CO inducing the highest.
Reference:
Gilmour et al.
(2004, 057420)
Species: Mouse
Gender: Female
Strain: CD1
Age: 8-10wks
Weight: 20-25g
Coal Fly Ash
MU: Montana Ultrafine
MF: Montana Fine
MC: Montana Coarse
KF: W.Kentucky Fine
KC: W. Kentucky Coarse
Particle Characteristics:
Montana Sulfur 0.83%, Ash
11.72%. Trace amounts of Be
P, Sr, V, Nb, Cd, Se, Ga, Cu.
Depleted in Si, Al, Fe, Mg, Ti.
Kentucky Sulfur 3.11 %, Ash
8.07%
Particle Size: Coarse:
fj m;
Fine: <2.5//m;
Ultrafine: < 0.2 fjm
Route: Oropharyngeal Aspiration
Dose/Concentration: 25ug or 100//gfmouse
Time to Analysis: 18h
>2.5
BALF Cells: PMN highly increased for MU at both
doses. The level was comparable to the positive
control. PMN also increased with KF at high dose.
Coarse particles caused no significant increase in
PMN. Number of macrophages did not change, but
NAG increased significantly with MU for both dose
levels and with KF and MF at high dose level.
BALF Inflammatory/Injury Markers: Total protein
and LDH was not significantly elevated. Albumin
concentration increased significantly after treatment
with the fine high dose of both particle types.
Cytokines: MU particles caused a significant
increase in TNF-a. MIP-2 increased in all fine and
ultrafine PM-instilled animals with the highest in the
MU and KF at both doses. IL-6 was detectable only
in the BALF of MU and KF with substantial
variability. The IL-6 levels were not significant.
July 2009
D-97
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Gilmour PM (collected from precipitator
et al. (2004,
087948)
Species: Rat
Gender: Male
Strain: SH/NQIBR,
WKY
Age: 12wks
Weight: 280-340g
unit of an oil burning power
plant in Boston)
Measured Components of PM:
S, Zn, Ni, V, Al, Cu, Pb, Fe, Ca,
Na, K, Mg, Endotoxin
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 0.0, 0.83, 3.3, and 8.3
mg/kg in SH rats; 0.0 or 3.3 mg/kg in WKY and
SH rats
Time to Analysis: 24h
BALF Inflammatory/Injury Markers: LDH activity
increased in a dose-related manner; this was
observed in SH rats after exposure to 0.83, 3.33 and
8.3 mg/kg PM. SH rats showed greater lung
permeability following PM exposure than WKY rats.
SH rats showed acute lung inflammatory response
after exposure to PM when compared to WKY rats.
BALF Cells: No increase in macrophage number was
observed in either rat strain following saline or PM
exposure at 24 h.
Cytokines: MIP2 mRNA expression increased
significantly in SH PM exposure group only. No
significant differences in TNF-a RNA expression in
either WKY, SH rats or control treatment groups
were observed.
CD14: A significant increase in lung CD14 protein
was observed only in SH rats exposed to PM.
TLR4: A significant increase in TLR4 protein in SH
rats exposed to PM was observed.
NF-kB: A significant increase in NF-kB binding
protein in the nuclei of SH rats exposed to PM was
observed. This effect was not observed in the control
of PM-exposed WKY rats.
Reference: Gilmour ufCB: Ultrafine carbon black
et al. (2004,
054175)
Species: Rat
Gender: Male
Strain: Wistar
Age: 12wks
(Printex 90 (Degussa)
CB: (Huber 990, HR. Haeffner
and Co)
Particle Size: ufCB: 14nm;
CB: 260nm (primary particle
diameter)
Route: Whole-body Inhalation
Dose/Concentration: ufCB: 1.66 mgfm3
fCB: 1.40 mgfm3
Number concentrations
ufCB: 52380 particles/cm3
fCB: 3800 particles/cm3
Time to Analysis: Exposed for 7 h. Sacrificed
0,16 or 48h post exposure.
BAL Cells: Total number of cells increased
significantly in UfCB-exposed rats at 0 and 16 h.
Recruitment of cells did not occur in response to CB
exposure. PMNs increased significantly in the BALF
of ufCB-exposed rats at 16 h. Leukocytes remained
unchanged following CB exposure but increased
significantly at 0 and 48 h post exposure to ufCB.
Cytokine mRNA: A significant increase in BALF
MIP-2 mRNA expression was observed at 48 h. No
differences in MIP-2 mRNA levels were observed in
the whole lung tissue.
Reference:
Godleski et al.
(2002,156478)
Species: Rat
Gender: Male
Strain: SD
Age: NR
Weight: 200-250g
CAPs (Boston; Harvard
Ambient Particle Concentrator)
Particle Size: 0.27±2.3/jm
(diameter)
Route: Inhalation
Dose/Concentration: 73.5-733//gfm3
Time to Analysis: Exposed 5hfd, 3d
(consecutive). BAL 24h postexposure
PMNs significantly increased with CAPs exposure
and also in relation to CAPs mass, Br, SO42', EC, 0C
and Pb. An overall increase in pro-inflammatory
mediators and decrease in immune enhancer and
evidence of vascular endothelial responses occurred
with CAPs exposure.
Reference:
Gottipolu et al.
(2009,190360)
Species: Rat
Gender: Male
Strain: Wistar
Kyoto (WKY), SH
Age: 14-16wks
Weight: NR
DE (30-kW (40hp) 4-cylinder
indirect injection Deutz diesel
engine) (02- 20%, CO-1.3-
4.8ppm, NO- <2.5-5.9ppm,
NO?- < 0.25-1,2ppm, S02- 0.2-
0.3ppm, OC/EC- 0.3±0.03)
Particle Size: Number Median
Diameter: Low- 83±2nm, High-
88.2nm; Volume Median
Diameter: Low- 207 ±2nm,
High- 225±2nm
Route: Exposure Chamber
Dose/Concentration: Low- 507 ±4/yg/m3, High-
2201 ±14 //g/m3
Time to Analysis: Exposed 4hfd, 5dfwk, 4wks.
Necropsied 1d postexposure.
DE increased neutrophils in a concentration-
dependent manner, and GGT activity at the high
dose. Particle-laden macrophages were found in DE-
exposed rats. DE dose-dependently inhibited
mitochondrial aconitase activity. DE caused 377
genes to be differentially expressed within WKY rats,
most of which were downregulated, but none in SH
rats. However, WKY rats had an expression pattern
shift that mimicked baseline expression of SH rats
without DE. These genes regulated compensatory
response, matrix metabolism, mitochondrial function,
and oxidative stress response.
Reference:
Gunnison and Chen
(2005, 087956)
Species: Mouse
Gender: Male
Strain: DK (ApoE
LDLr'1')
Age: 18-20wks
CAPS
(Northeastern regional back-
ground)
Ambient air copollutants
measured O3, NO?
Particle Size: 389 ± 211 m
Route: Whole-body Inhalation
Dose/Concentration: CAPS — 131 ± 99//g/m3
including O3 - 10 ppb and NO2 - 4.4 ppb
Time to Analysis: 6h/d, 5d/wk for 4m (5/12/03-
9/5/03). Sacrified 3-4d post exposure.
Microarray Data: 13 genes in the heart tissue and
47 genes in the lung tissue were identified as
possibly affected. Strict standards (1.5 fold
response, 10 % false discovery rate) resulted in
responses by only 1/13 genes (Rex3 - no known
heart physiology) in the heart tissue and 0/47 genes
in the lung tissue. Using more liberal response
(nonstatistical) standards (1.5 fold only) and
comparison of each CAPS animal with all 3 control
animals (3x3 array) resulted in possible effects on 7
additional genes in the heart tissue and 37 genes in
the lung tissue.
July 2009	D-98	DRAFT - DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Gurgueira et al.
(2002, 036535)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Weight: 250-300g
CAPs
(Harvard Ambient Particle
Concentrator)
CB
(C198 Fischer Scientific, Pitts-
burg, PA USA)
Composed of 85.9± 0.2%
Carbon, 13.0± 0.2% 02,
1.17± 0.2% Sulfur
ROFA
(Boston, MA USA oil-fired
power plant)
Particle Size: CAPs: 1-2.5
/jm; CB: < 2.5/jm; ROFA:
< 2.5 /jm
Route: Whole-body Inhalation
Dose/Concentration: 300 ± 60 //gfm3
Time to Analysis: 1, 3, 5h CAPs Exposure
followed by immediate post-exposure analysis.
5h CB, immediate analysis.
30min ROFA, Immediate analysis.
In situ Chemiluminescence(CL): Data show a
significant increase in lung and heart CL at 5h. Lung
CL increased linearly with time of exposure.
Oxidants: CAPs-initiated oxidative stress was not
detectable in those rats allowed to recover in room
air after the simulated "peak" in particulate air
pollution. Rats breathing particle-free filtered air for
3 days had significantly lower levels of oxidants.
Exposure to inert CB did not exert oxidant effects on
the heart and lung.
BALF Inflammatory/Injury Markers: The water
content of the lung and heart increased significantly
upon exposure to CAPs but not to filtered air and
increased as a function of length of exposure. Rats
breathing CAPs also showed increases in LDH and
CPK as a function of length of exposure.
Antioxidant enzymes: Data showed an increase in
SOD and catalase activities in both the lung and
heart. The pattern of increase was tissue specific.
Reference: Hamoir PSC: Polystyrene particles, Route: IT Instillation
et al. (2003,
096664)
Species: Rabbit
Strain: New
Zealand
Age: 12-16 wks
Weight: 2.8 ±
0.5kg
Carboxylate modified, 3 types
PSA: Polystyrene particles,
Amine modified, 1 type
Particle Size: PSC: 24,110 or
190nm (PSC24, PSC 110,
PSC190); PSA: 190nm
Dose/Concentration: PSC24: 0.04 or 4
mg/rabbit
PSC110, PSC190, PSA190: 4mgfrabbit
Time to Analysis: 0, 30, 60, 90,120min
Capillary Filtration Coefficient: A time-c
increase correlating to total number of
particles/surface area, not particle size, was
observed.PSA induced a significant increase in mi-
crovascular permeability as compared to PSC. This
suggests that the number of particles exposed should
be considered an important parameter for measuring
air quality rather than total particle surface area.
July 2009	D-99	DRAFT - DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Happo
et al. (2007,
096630)
Species: Mouse
Gender: Male
Strain: C57BL/6J
Weight: 19-30g
Age: 10-11 wks
PMC (Coarse)
PMF (Fine)
PMUF (Ultrafine)
Collected in 6 European cities:
Duisburg, Prague, Amsterdam,
Helsinki, Barcelona, Athens
Particle Size: PMC: PMio2.5;
PMF: PM2.B.0.2: PMUF: PM0.2
Route: IT Instillation
Dose/Concentration: PMC: 5.9-29.6//g/m3;
PMF: 8.3-25.2//g/m3; PMUF: 2.7-6.7 //g/m3
Dose-response: 1, 3,10mg/kg
Time course: 10 mg/kg
Time to Analysis: 1. Dose-Response study:
parameters measured 24h post exposure. 2.
Time course study: parameters measured 4,12,
24 h post single exposure (at 10 mg/kg).
Total Cell Numbers: 1. For the dose-response
study, all the PMC samples exhibited dose-dependent
increases of total cell numbers. The 3 and 10 mg/kg
doses of PMC induced statistically significant
increases. At 10 mg/kg, only 2/6 samples induced
statistically significant increases. No PMUF samples
induced effects at any dose. 2. For the time-response
study, no increases in cell numbers were shown at 4
h. Though the levels induced by PMC at 24 h were
lower than at 12 h, both levels were statistically
significant. PMF induced statistically significant
increases only at 12 h for 4/6 samples. PMUF
induced only 1 significant increase at 12 h; the 24 h
time point was not tested.
Total Protein/LDH: 1. The lower doses of 1 and 3
mg/kg did not induce significant increases in any of
the PM samples, except for PMUF-Athens. All 6
samples of PMC, at 10 mg/kg, induced significant
increases. At 10 mg/kg, 4/6 PMF samples induced
significant increases. 2. At 4 h, none of the samples
increased protein concentration. The PMC samples,
excluding Prague, induced significantly higher
concentrations at 12 h. At 24 h, only 3/6 PMC
samples induced significant increases. Only 2 PMF
samples induced significant increases at 12 and 24
h. At 12 h, effects induced by PMUF were minimal
and inconsistent: the 24 h time point was not tested.
Cytokine: 1. Only PMC induced dose-dependent
responses that reached statistical significance at 10
mg/kg. PMF and PMUF induced minimal and
inconsistent responses.
TNF-a levels: 2. TNF-a levels increased
significantly at 4 and 12 h by PMC. At 24 h, TNF-a
levels returned to near control levels. PMF, at 4 h,
induced statistically significant increases for 3/6
samples and significant increases in 2/6 samples at
12 h. No PMUF samples significantly increased TNF-
a levels.
IL-6: 2. PMC induced the highest IL-6 levels at 4 h.
Levels at 12 and 24 h were reduced with 6/6 and 3/6
samples showing statistically significant increases,
respectively. PMF showed a similar trend with 4 h
inducing the highest levels that were reduced at 12
and 24 h. Of the PMUF samples, only the Helsinki
and Duisburg samples induced statistically significant
results at 4 and 12 h. Generally, the PMUF responses
were negligible when compared to PMC and PMF.
KC production: 2. All PMC samples induced the
highest levels at 4 h. At 12 and 24 h, levels were
reduced but 4/6 samples induced statistically
significant levels. PMF showed a similar trend- the
highest levels were induced at 4 h (in 3/6 samples).
PMUF at 4 h showed small, though not significant,
increases. At 12 h, only 2 samples showed
statistically significant differences from the control:
the 24 h time point was not tested.
Reference: Harder Carbon UFP
et al. (2005,
087371)
Species: Rat
Gender: Male
Strain: Wistar
Kyoto
Age: 14-17wks
Weight: NR
Particle Size: Diameter:
37.6±0.7nm
Route: Exposure Chamber
Dose/Concentration: 180/yg/m3
Time to Analysis: Telemeter implanted into
peritoneal cavity. 10d recovery. 3d baseline
reading. 24h exposure. 3d recovery.
Carbon UFP mildly but significantly elevated HR
compared to the control. SDNN was significantly
decreased during exposure. UFP induced mild
pulmonary inflammation, significantly increased
PMN, and increased the total protein and albumin
concentrations. Particle-laden macrophages
sporadically accumulated in the alveolar region.
July 2009
D-100
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:	CAPs (Detroit; July-Sept.
Harkema et al. 2000; Harvard Ambient Fine
(2004, 056842) Particle Concentrator)
Species: Rat Particle Size: 2.5/jm
Gender: Male (diameter)
Strain: F344, BN
Age: 10-12wks
Weight: NR
Route: Inhalation Exposure Chamber. IT
Instillation.
Dose/Concentration: 4d concentration:
676±288//g/m3, 5d concentration: 313±119
/yg/m3, July concentration: 16-185 //g/m3,
September concentration: 81 -755/yglm3; IT
Instillation- 200 /jl (soluble and insoluble)
Time to Analysis: F344 rats sensitized to
endotoxin, BN rats to OVA. Exposed 10h/d 1,4,
5d (consecutive). Another group of rats i.t
instilled. Both groups killed 24h postexposure.
The retention of PM in the airways was enhanced by
allergic sensitization. Recovery of anthropogenic
trace elements was greatest for CAPs-exposed rats.
Temporal increases in these elements were
associated with eosinophil influx, BALF protein
content and increased airway mucosubstances. A
mild pulmonary neutrophilic inflammation was
observed in rats instilled with the insoluble fraction
but instillation of total, soluble or insoluble PM2.6 in
allergic rats did not result in differential effects.
Reference:	DE: generated by 2369-cc
Hiramatsu et al.	diesel engine (Isuzu) at 1050
(2003,155846)	rpm and 80% load with
Species: Mouse	commercial light oil
Gender: Female
Strain: BALB/c and
C57BL/6
Age: 8wks
Weight: 17-22g
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: DEP: 10O/zg/nv1 or 3
mg/m3; S02< 0.01 ppm; NO2 2.2 ±0.3 or 15
±1.5 ppm; CO 3.5 ± 0.1 or 9.5 ± 0.6 ppm
Time to Analysis: 7hfd, 5dfwk for 4 or 12wks,
Immediate
BALF Cells: Alveolar macrophages (AMs) increased
dose-dependently at 30 and 90d. High DE exposure
resulted in bronchus-associated lymphoid tissue
(BALT) around DEP-AMs; this was less conspicuous
in C57BL/6 than in BALB/c mice. B- and T-cell
populations were found in the BALT with no
significant differences observed between the strains.
Lympohocytes and neutrophils increased time- and
dose-dependently with agreater increase in BALB/c
than C57BL/6 observed. No eosinophils or basophils
were observed. Mac-1 -positive cells exposed to high
DE levels increased in both strains at 1 month
(33.8%) and 3 months (20.3%) vs. low dose group
(5.3 and 7% respectively).
Cytokines: At 30d, TNF-a, IL-12p40, IL-4 and IL-10
mRNA increased, IL1b and iNOS decreased. IFNy
increased in BALB/c but decreased in C57BL/c. IL-6
mRNA was not affected. At 90d, IL-4 and IL-10
mRNA similarly increased in C57BL/6 mice exposed
to low DE level but decreased at high DE level.
Reference:
Hollingsworth et al.
(2004,097816)
Species: Mouse
Gender: Male
Strains:
C57BL//6TLR+I+,
C57BL//6tlri
Aqe: 8-9 wks
R0FA
Particle Size: NR
Route: Oropharyngeal Aspiration
Dose/Concentratin: 50 ul of 1/yg/mL
suspension per mouse
Time to Analysis: Parameters measured post
single exposure of 6 and 24h.
TLR4-Knocl
-------
Reference
Pollutant
Exposure
Effects
Reference:Inoue
et al.
(2006)
Species: Mouse
Gender: Male
Strain: NCfNga
Age: 10wks
DEP (derived from 4 cyl, 2.74I
light duty diesel)
Particle Size: NR
Route: IT Instillation
Dose/Concentration: lOO/zg/mouse
Time to Analysis: 1fwk for 6wks. Parameters
measured 24h after last administration
BALF Cells: DEP significantly increased total cells,
neutrophils and mononuclear cells but did not induce
an effect on eosinophils.
Cytokines: DEP increased IL-4, KC and MIP-1. The
increase in IL-5 was not statistically significant.
Reference:
Ishihara et al.
(2003, 096404)
Species: Rat
Gender: Male
Strain: Wistar
Age: 5wks
DE
(from 2 engines, produced on
site)
¦L - low level DE
¦M - medium level
¦MG - DE w/o particulates
¦HR - high level
Measured Components: NO2,
S04, SO2, CO, CO2, NOx, NO,
HTHC, HCHO, 02
Particle Size: L: 0.33 -0.50
fj m
M: 0.35 ¦ 0.40 fjm
HR: 0.42- 0.45/m
Route: Whole-body Inhalation
Dose/Concentration: L: 0.18- 0.21 mgfm3
M: 0.92-1.18 mglm3
MG: 0.01 mg/m3
HR: 2.57 - 2.94 mglm3
Time to Analysis: 16h/d, 6d/wk, for 6,12,18
& 24 m. Parameters measured immediately
following last exposure.
Morbidity and Mortality: Weight gain in HR group
was less than other groups at 18 and 24 mo. This
indicates a significant difference between the HR
and C group. Mortality during the study was
frequent. C group experienced an 8% mortality rate,
L group 12%, M group 15%, MG group 12% and HR
group 23%.
BALF Inflammatory/Injury Markers: Significant
differences were seen among groups with respect to
number of total cells and percentages of cell
differential, total protein, fucose, sialic acid,
phospholipid and prostoglandin E2. Total protein
increase was observed in both M and HR dose groups
with the HR group increasing time-dependently.
BALF Cells: The HR group showed a significant
increase in total cell count from 6 to 18mo. The
percentage of PMN increased at 6mo in M, MG and
HR group. M group lymphocytes significantly
increased at 6,12, and 24mo of exposure.
Macrophages decreased at 6mo for the M and HR
groups.
Mucus and Surfactant: The HR group showed a
significant increase from 12 to 18mo.
Reference: Jones,
HR.A.
Hamacher, K.
Clark, J.C.
2005
Species: Rabbit
Strain: New
Zealand
Weight: 2.5- 3.5kg
Ml u vvi\o, 1 iui uuiaoio VVCIC O IIII pic OC III.
At 13 wks, active scarring and raised neutrophil
macrophage counts were still present.
ASP: At 15h, neurophils increased. Macrophages
tripled and remained increased for 3wks.
At 4d, macrophages bore particles.
At 13d, neutrophils decreased significantly.
By 25d, silica spheres were gradually removed from
lungs.
PET Scanning: 18F-fluoroproline showed increased
activity beginning at 14d and peaking at 41 -54d (left
lung control vs right lung challenged).
Microautoradiography: 3h-proline at 13 weeks
showed radiolabel localization to fibroblasts in the
challenged lung
ASP: Amorphous silica particles Route: Intrapulmonary Instillation (Right upper MCSP: At 6h, neutrophils increased. Macrophages
(Hypersil)
MCSP: Microcrystalline silica
particles
Particle Size: ASP: 5(m;
MCSP: 5/ym
lobe of lung)
Dose/Concentration: 50mg in 0.5mL saline
Time to Analysis: Parameters measured at
varying times from 6h to 91 d post treatment.
increased 3 fold.
At 60h, neutrophils were pyknotic and the lungs
displayed a thickened interstitium containing silica
particles.
At 5d, collagen deposition appeared.
AT 8d, fibroblastic activity and necrosis were
observed.
At 15d, aggregation of silica particles and necrotic
debris were apparent.
July 2009
D-102
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Kato
and Kagawa (2003,
089563)
Species: Rat
Gender: Male
Strain: Jcl Wistar
Aqe: 5wks
Roadside air
(Prefectural T okyo-Danishi-
Yokohama highway, Yokohama-
Haneda Airport Metropolitan
expressway and Satsukibashi-
Mizuecho city road, Japan)
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: Exposed group: 55.7 ppb
NO2,
62.7/yg|m3 PM;
Control group: 5.1 ppb NO2,
14.3/yg/m3 PM
Time to Analysis: Exposed for 24,48, 60 wks.
Parameters measured immediately following
exposure.
Respiratory tissue: Post 24wks, the lung surface
was light gray with some BC particle deposits. Post
48-60wks, however, the surface was scattered with
particle deposits in addition to its light gray color.
Airway changes: After 60 wks, no remarkable
changes seen in the epithelium. The structure of the
airways remained normal.
Cells: No proliferation or ectopic growth of goblet
cells were noted. Mast cells increased in epithelial
intercellular space. No mast cell degranulation was
observed. Lysosomes increased in cilitated cells post
48wks. Clara cells were unaffected.
Lymph nodes: Deposition of carbon particles were
noted in the trachea and bronchiole-associated lymph
nodes post 24wks.
Alveolar changes: No changes in morphology of
broncho-alveolar junctions were noted. Anthracosis
observed within alveolar walls and pleura post
24wks and became progressively marked with
increased exposure. No change in the number of
alveolar holes between exposure and control groups
were observed.
Reference: Kato,
T.
Yashiro, T.
Murata, Y.
2003
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age: 7wks
Weight: 190-220g
Polystyrene latex suspension of Route: IT Instillation with nebulizer
latex beads (Japan Synthetic
Rubber Co.), uncoated or
coated with lecithin
Particle Size: 240nm
Dose/Concentration: 5ml of 0.2% suspension
administered over 20 min at flow rate of
0.25ml/min
Time to Analysis: Exposed for 20min.
Parameters measured 30min following
treatment.
Alveolar Macrophages: Following treatment, AMs
appeared undamaged. AMs ingested more uncoated
than coated beads, but both were ingested. Ingestion
of beads differed as coated beads were engulfed
individually while uncoated beads were engulfed
individually or in aggregates.
Epithelial Cells: Type I cells incorporated coated
beads within a layer of cytoplasm. Type II cells
incorporated beads in lamellar bodies. Uncoated
beads were not incorporated.
Other: Neither type of beads were incorporated into
endothelial cells, fibroblasts or interstitium of
alveolar wall
Monocytes: Only the coated beads were
incorporated by the monocytes. They were found
inside and outside phagosomes and lysosomes of
monocytes. PMNs did not incorporate any beads.
Reference:
Kleinman et al.
(2003, 053535)
Species: Rat
Gender: NR
Strain: F344n-NIA
Age: 22-24m
0s
CCL: O3 + Ammonium bisulfate
(ABS) + Elemental Carbon (EC)
CCH: 0s + ABS + EC
Purified Air (control)
Particle Size: CCL: 0.30 ±
2.5 fjm
CCH: 0.29 ±2.3//m
Route: Nose-only Inhalation
Dose/Concentration: O3: 0.2ppm
CCL: 50//g/m3 EC + 70//gfm3ABS ¦
0s
0.2 ppm
140 /yglm3 ABS + 0.2
CCH: 100/yg/m3 EC
ppm O3
Time to Analysis: 4hfd, 3 consecutive dfwk for
4wks
DNA Replication: O3 caused a slight effect of 20-
40% increase. CCL and CCH caused between 250 ¦
340% increase for interstitial and epithelial cells.
CCL induced greater reactions than the high dose.
BALF Inflammatory/Injury Markers: Total protein,
mucus glycoprotein and albumin were somewhat
elevated in all exposure groups but only reached
statistically significance for CCL and protein (very
high variability). CCL and CCH both depressed Fc
receptor side binding.No effect for O3 was observed.
BALF Cells: CCL and CCH induced macrophage
respiratory burst activity. The effect induced by O3
was not significant.
Reference:
Kleinman and
Phalen (2006,
088596)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age: 6 wks
Weight: 200g
L03: Low O3
H03: High Os
LS: Low H2SO4
HS: High H2SO4
L0LS:Low O3 + Low H2SO4
LOHS: Low Os + High H2SO4
HOLS: High Os + low H2SO4
HOHS: High Os + high H2SO4
Particle Size: LS - 0.23
/ym±2.3
HS - 0.28/ym±2.1
LOLS - 0.23//m ±2.3
LOHS - 0.28/ym ± 2.1
HOLS - 0.23//m ± 2.3
HOHS - 0.28/ym ± 2.1
Route: Nose-only Inhalation
Dose/Concentration: L03 - 0.30ppm
H03 - 0.61ppm
LS - 0.48mg/m3
HS - I.OOmgfm3
LOLS - 0.31 ppm + 0.41 mg/m3
LOHS - 0.31 ppm +1.04mg/m3
HOLS - 0.60ppm + 0.52mg/m3
HOHS - 0.60ppm + 0.86mg/m3
Time to Analysis: Exposed for 4h. Parameters
measured 42h post-exposure.
Inflammatory Lesions in Lung Parenchyma:
Neither Type 1 or 2 lung lesions were affected by
sulfuric acid alone. H03 doubled Type 1 lesions and
increased Type 2 lesions 25-fold. Additions of H2SO4
to O3 appeared to have a dose-dependent protective
effect for both types of lesions.
DNA Synthesis in Nasal, Tracheal and Lung
Tissue: Increased DNA synthesis was observed at all
high O3 exposures but was not affected by
coexposure to H2SO4.
Macrophage FcR binding: No effects were
observed (no data for L03 and H03).
Macrophage Phagocytosis: All levels of exposure
(no data for L03 and H03) decreased phagocytosis.
July 2009
D-103
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Kodavanti et al.
(2005, 087946)
Species: Rat
Gender: Male
Strain: WKY and
SH/NCrlBR
Aqe: 11-14wks
CAPs (EPA, NC)
Measured components included
Al, Be, Ba, Co, Cu, Zn, Pb, Mn,
Ni, Ag, Ti, As.
Particle Size: 1 d: 1.07-1.19
/jm; 2d: 1.27-1.48 /jm
Route: Whole-body Inhalation
Dose/Concentration: 1d study: 1138-1765
z^g/m3
2d study: 144-2758//g|m3
Time to Analysis: 4hr (SH only); 4hrfday, 2d
(WKY and SH)
Post-exposure: 1d: 3h except study #4,18-20h;
2d: 18-20h
Breathing Parameters: In a paired analysis of
control SH and treated SH, treated SH showed an
increase in expiratory and inspiratory time due to
CAPs. The treated and control groups of WKY rats
did not show significant differences.
BALF Cells: In the 2d study, WKY rats showed
decreases in total cells; this decrease was
associated with decreased macrophages.WKY
showed an increase in neutrophils.
BALF Inflammatory/Injury Markers: Total protein
and albumin in WKY rats decreased whereas SH rats
maintained the same approximate level. LDH activity
lowered slightly in both strains.
Cell Membrane Integrity: SH rats showed
increased GGT (membrane bound enzyme) activity
and plasma fibrinogen for 5/7 exposures but these
increases did not appear to be dose-dependent.
Cytokines: Levels were undetermined in SH rats.
WKY showed slight increases in IL-6, TNF-a, and
MIP-2 but these increases were not statistically
significant.
Reference: Kooter
et al. (2006,
097547)
Species: Rat
Gender: Male
Strain: SH
Age: 12-14wks
CAP-F - fine (Site I)
CAP-UF - fine + ultrafine (Site
II)
(Netherlands)
Some measured Components:
Ammonium, nitrate, sulphate
ions: 56± 16% CAP-F mass,
17 ± 6% CAP-UF mass
Particle Size: 0.15 < CAP-
F < 2.5
0.65-0.75 /jm
CAP-UF< 2.5
0.58-1.41 fjm
Route: Nose-only Inhalation
Dose/Concentration: CAP-F 399- 3613//g/m3
CAP-UF 269-556//g/m3
Time to Analysis: 6hfd for 2d consecutive, 18h
BALF Inflammatory/Injury Markers: Based on
unchanged levels of LDH and ALP, no cytotoxicity
was noted. No significant change in the levels of
total cells were observed. MDA (malondialdehyde)
decreased with CAP-UF. Ho-1 increased with CAP-UF
and CAP-F.
Cytokines: CC16 decreased at 457/yg/m3 of CAP-F
and increased at 3613 /yglm3 of CAP-F.
Hematology: WBC and lymphocytes decreased with
both CAP-F and CAP-UF. MPV and MPC (mean
platelet volume and component) increased with CAP-
UF.
BALF Cells: A decrease in absolute neutrophils as
well as percentages of reticulocytes and percentages
of neutrophils were observed with CAP-F. Increased
percentages of lymphocytes were observed with
CAP-F.
Pathology. No changes were observed.
Reference: Kumar
et al. (2004,
096655)
Species: Rat
Gender: Male
Strain: Wistar
Weight: 150±20g
Fly Ash (Obra Thermal power Route: Whole-body Inhalation
Station, India)
Particle Size: PM < 5 /jm
(90%)
Dose/Concentration: 14.4±1.77 mg/m3
(fluid bed generator)
Time to Analysis: 4hfd for 28d. Parameters
measured immediately following last exposure.
Lung Weight: Lung body weight increased 25.58%
relative to controls. Total body weight slightly
decreased in the treated group.
BALF Inflammatory/Injury Markers: LDH, GGT,
ALP and lavagable protein increased by 140,450,
160 and 50%, respectively.
BALF Cells: Only eosinophils(%) increased 95% over
controls.Congestion and focal infiltration of
monocytes in alveolar area was seen. Fly ash laden
macrophages in alveoli combined with hypertrophy of
epithelial lining cells was observed.
Reference:Lei et
al. (2004, 087999)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Weight: 318 ± 8g
CAPs
(Yaipei, Taiwan)
Particle Size: PM: 0.01- 2.5
fj m
Reference: Nose-only Inhalation
Dose/Concentration: 371 ± 208//gfm:l
Time to Analysis: 6hfd for 3d,
5h post-exposure pulmonary function.
2d post-exposure for BALF collection
Pulmonary hypertension induced 2wks pre-
exposure
Respiratory Effects: Decreased respiratory
frequency and increased tidal volume for both
experimental and control groups were observed.
However, only the experimental group levels were
statistically significant. There was an increase in
airway responsiveness (Penh/methacholine) for CAPs
group when compared to the control.
BALF Cells: A massive increase in total cell number
and percent neutrophils was observed. There were
no changes in percent macrophages, lymphocytes
and eosinophils.
BALF Inflammatory/Injury Markers: Total protein
and LDH increased in the CAPs group.
Cytokines: TNF-a and IL-6 were not affected.
July 2009
D-104
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:Lei et
al. (2004, 087884)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Weight: 300-350g
CAPs from Asian dust storm
(Taiwan)
Measured Components: Si, Al,
S, Ca, K, Mg, Fe, As, Ni, W, V,
organic carbon, elemental
carbon, SO2, NO2, nitrate,
sulfate
Particle Size: PM: 0.01- 2.5
fj m
Route: Nose-only Inhalation
Dose/Concentration: 315.6 //gfm3 (Low)
or 684.5 //g/m3 (High)
Time to Analysis: Low: Exposed for 6h.
Sacrified 36h post exposure
High: Exposed for 4.5h. Sacrified 36h post-
exposure
Pulmonary hypertension induced 2wk pre-
exposure
Hematology: PM induced a dose-dependent increase
in WBCs. No change was seen in RBCs. Platelet
results were highly variable.
BALF Cells: PM induced dose-dependent increases in
total cells and percentage of neutrophils. No change
in macrophages, lymphocytes or eosinphils occurred.
Basophils were highly variable.
BALF Inflammatory/Injury Markers: Dose-
dependent increases were observed for total protein
andLDH.
Cytokines: IL-6 increased dose-dependently.
(control: 33.5 ± 7.5, LOW 165.1 ± 117.2,273.6 ±
62.8 pg/mL)
Reference: Li et al. DEP (2369-cc diesel engine Route: Exposure Chamber
(2007,155929)
Species: Mouse
Gender: Female
Strain: BALB/c,
C57BL/6
Age: 9wks
Weight: NR
manufactured by Isuzu Motor,
operated at 1050 rpm, 80%
load, commercial light oil)
Particle Size: NR
Dose/Concentration: DEP: 103.1 ±9.2//g/m3,
CO: 3.5±0.1ppm, NO2: 2.2±0.3ppm, SO2:
<0.01ppm
Time to Analysis: Protocol 1: Exposed 7hfd,
5d/wk. Sacrificed at day 0, week 1, 4, 8.
Protocol 2: DE alone or DE+NAC 7h/d, 1-5d.
Airway hyperresponsiveness: Penh values
increased in BALB/c mice compared to the control at
day 0, but no significant changes occurred after this
time. Penh values increased in C57BL/6 mice at 1wk
compared to the control but returned to control levels
at 8wks.
BALF: Compared to the other strain, the total
number of cells and macrophages increased
significantly at 1wk in C57BL/6 mice and at 8wks in
BALB/c mice. Neutrophils, lymphocytes, MCP-1, IL-
12, IL-10, IL-4, IL-13 increased significantly for both
strains. No eosinophils were found. IL-1f3 and IFN-y
increased significantly in BALB/c mice compared to
C57BL/6 mice.
HO-1 mRNA and protein: HO-1 mRNA was more
marked in BALB/c mice at 1wk and C57BL/6 mice at
4 and 8wks. H0-1 protein percentage changes from
the control were greater in BALB/c mice at 1 wk and
C57BL/c mice at 8wks.
NAC: NAC inhibited the increased Penh values, total
number of cells and macrophages in C57BL/6 mice at
1wk and neutrophils and lymphocytes in both strains.
Reference: Liu et
al. (2008,156709)
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 11wks
Weight: NR
DEP (5500-watt single-
cylinder diesel engine generator
(Yanmar, Model YDG 5500E),
406 cc displacement air-cooled
engine, Number 2 Diesel
Certification Fuel, 40 weight
motor oil)
Particle Size: —0.1 /jm
(MMAD)
Route: Intranasal Exposure
Dose/Concentration: Average particle
concentration: 1.28mg/m3
Time to Analysis: Four groups: saline+air
control, saline+DEP,A fumigatus+w,
A. fumigatus+U£P. A. fumigatus exposure every
4d for 6 doses. DEP exposure 5h/d for 3wks
concurrent with A. fumigatus exposure.
A.fumigatus+DEP increased IgE, the mean BAL
eosinophil percentage, goblet cell hyperplasia, and
eosinophilic and mononuclear cell inflammatory
infiltrate around the airways and blood vessels
compared to the A. fumigatus or DEP treatments.
A.fumigatus+DEP also caused methylation at the
IFN-y promoter sites CpG'63, CpG'46, and CpG'206.
Reference: Lopes
et al. (2009,
190430)
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 6-8wks
Weight: NR
PM (high density traffic: winter
2004; Sao Paulo, Brazil) (NO2,
CO, SO2)
Particle Size: Diameter: 10
fj m
Route: Open-Top Exposure Chamber
Dose/Concentration: 33.86 ±2.09 /yg/m3
Time to Analysis: Some rats pretreated with
papain. Exposed to UAP or filtered air 24h/d,
7d/wk, 2m.
The papain+UAP treatment increased Lm values,
collagen fibers, and decreased the density of elastin
fibers over the papain+filtered air treatment. The
papin+UAP treatment increased 8-isoprotane more
than any other group.
July 2009
D-105
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Mangum et al.
(2004, 097326)
Species: Rat
Gender: Female
Strain: CDF
(F344)/CrlBR
Aqe: 7wks
TiCb pigment grade (DuPont)
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: 10, 50 or 250mgfm3
Time to Analysis: 6hfd x 5d/wk for 13wks.
Parameters measured 0, 4,13, 26, 52 wks post-
exposure.
OPN (osteopontin) Expression: At 0 wks, 0PN
mRNA expression exhibited a dose-dependent
increase. Low dose induced a 2-fold increase while
the high dose induced an almost 100 -fold increase.
At 4 wks, the mid-dose and high-dose elevated OPN
mRNA levels. At 13wks, the high dose elevated OPN
mRNA levels.No significant elevation with mid dose
level was observed. At 26wks, the mid and high dose
induced elevated OPN mRNA levels. At 52wks, rats
in the low, mid and high dose groups all indicated
elevated levels of OPN mRNA. Specifically, the low,
mid and high doses induced a 3-fold increase, 7-fold
increase and 400-fold increase, respectively.
OPN Protein in BALF: Data was not reported at 0
and 4 wks. At 13 wks, protein increased 9-fold
(— 800 pg/mL OPN) at mid dose and 100 -fold
(~ 8000 pg/mL OPn) at high dose. At 26 wks, the
mid and high dose groups remained elevated. At 52
wks, protein increased by 2.5 fold in low dose, 7-fold
in mid dose and 166-fold in high dose group.
Histopathology: At 52wks, slight OPN
immunoreactivity was observed in control and low
dose group (immunostaining mostly limited to
intraalveolar MACSI.Trichrome-stained lung sections
from control and low dose groups showed no
increase in collagen. Rats exposed to mid or high
dose groups showed areas of lesions.
Reference: Martin UAP-BA: Urban Air particles
et al. (2007,
096366)
Species: Mouse
Gender: Male
Strain: BALB/c
Aqe: 1-2 mo
(Buenos Airs, Argentina)
Particle Size: < 2.5/ym
Route: Intranasal Installation
Dose/Concentration: 0.17 mg/kg bw
Time to Analysis: 3xday, 3dfwk, 2d apart (1, 4,
7d). Parameters measured 1h post-exposure.
Particle Characteristics: 3 types, ultrafines
< 0.2um (inorganics ND), bunched agglomerates of
ultrafines and <40um with aluminum silicates, ions
and trace metals.
Morphometry: Induced focal inflammatory lesions.
Accumulation of refractile material in upper and
lower respiratory tract. PM in phagocytes of
bronchiolar lumen and alveolar space. No evidence of
fibrosis and/or collagen changes.
BALF Cells: Increased amount of phagocytes in
alveolar area, reducing airspace percentage (control
52.9% ± 1.39, UAP-BA 24.7% ± 2.87). Increased
number of PAS positive cells.
Reference: Mauad PM (busy traffic street Sao
et al. (2008, Paulo, Brazil; Aug. 2005-April
156743)	2006) (NO?, SO?, CO)
Species: Mouse
Gender: Male,
Female
Strain: BALB/c
Age: 10d
Weight: Parental:
21,4±4.0 -
26.3 ± 2.8g; 15d-old
offspring: 7.8±1.1
- 9.0 ± 1,0g; 90d-
old offspring:
20.3±2.3 -
27.4±1.8q
Particle Size: 2.5,10/ym
(diameter
Route: Open-Top Chamber
Dose/Concentration: PM2.6: filtered chamber-
2.9±3.0/yglm3, nonfiltered chamber- 16.9±8.3
/yg/m3; Outdoor concentration: PMio- 36.3 ± 15.8
/yglm3, CO- 1.7±0.7ppm, NO- 89.4±31.9//gfm3,
SO?- 8.1 ±4.8//g/m3
Time to Analysis: Nonfiltered exposure 24hfd
4m. Mated at 120d exposure. After birth, 30
females and offspring transferred to filtered or
nonfiltered chamber. Killed 15 or 90d of age.
Mild foci of macrophage accumulations containing
black dots of carbon pigment occurred in the alveolar
areas on 90d-old mice. Surface-to-volume ratio
decreased from 15 to 90d of age and was higher in
mice exposed to air pollution. PM exposure reduced
inspiratory and expiratory volumes at higher levels of
transpulmonary pressure.
Reference:
McDonald et al.
(2004, 087459)
Species: Mouse
Strain: C57BL/6
Aqe: 8-10wks
DEE: high load, No 2, No cat
(620: 1 dilution)
DEE-ER (Control): Emissions
Reduced (high load, low sulfur
ECD1) (same dilution)
(Yanmar diesel generator, 406
cc, 5500 watt load)
Particle Size: DEE: 11 Onm;
DEE-ER: NR
Route: IT Installation
Dose/Concentration: DEE PM: 236 //gfm3
DEE-ER PM: 7/yg|m3
Time to Analysis: DEE: 6hfd for 7d.
DEE-ER: 6h/d for 7d. RSV administered post-
exposure for some: single, 4d. Those not
infected with RSV sacrificed immediately upon
last exposure.
Differences in Exposure Conditions: CO, PM,
elemental carbon, organic carbon, nitrate, alkyne, c2-
c212 alkenes, phenanthrenes, total particle PAHs,
total Oxy-PAHs, benzene, pyrene, benzo(ayrene, zinc
were reduced by 90-100% in the emissions reduction
case. Most other components were reduced by
around 60%.
DEE vs. DEE-ER Effects: DEE increased viral
retention and lung histopathology. DEE-ER increases
were not statistically significant.
Cytokines: DEE increased TNF-a, IL-6, IFN-y and
HO-1. DEE-ER responses were not statistically
significant (significantly higher variability in DEE-ER
controls vs. DEE controls).
July 2009
D-106
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
McQueen et al.
(2007,096266)
Species: Rat
Gender: Male
Strain: Wistar
Weight: 228-500g
DEP: SRM 2975 (NIST)	Route: IT Instillation
Particle Size: NR	Dose/Concentration: 0.5 mL/rat of 1 mgfmL; 1-
2.2 mg/kg
Time to Analysis: Single exposure, sacrificed
6h post exposure. Pre exposure: Vagotomy
(sectioning of vagus nerve) or
Atropine 1 mg/kg i.p. administered 30 min prior, 2
and 4h post.
BALF Cells: A 9-fold increase in neutrophils with
high individual variability in response was observed.
Bilateral vagotomy prior to DEP reduced neutrophil
increase to 3 fold. Vagotomy with saline instillation
had no effect. Atropine reduced neutrophils to levels
similar to saline response. No differences were
observed between DEP response in anesthetized
when compared to conscious animals. Macrophages,
eosinophils and lymphocytes remain unchanged.
Respiratory Response: RMV increased post DEP.
Vagatomy reduced response by one-third. Atropine
pre-treatment did not have effect.
Cardiovascular Response: Blood pressure and
heart rate were unaffected. Average arterial 02
increased after DEP, but not when compared for
each animal. CO2 and pH were not affected
Reference:
Medeiron et al.
(2004,096012)
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 60d
Weight: 20-30g
CP: Carbon particles
PSA: R0FA (solid waste
incinerator hospital Sao Paulo,
Brazil)
PSB: electric precipitator, steel
plant, Brazil)
PSA/PSB Characteristics:
Generally, PSB had greater
component concentrations than
PSA: Br (100+x), Cr (3x), Fe
(10+x), Mn (2x), Rb (60+x), Se
(7x), Zn(4x). PMA>PMB: Ce
(3x), Co (10+x), La (1 OOx), Sb
(15x), V (50x).
Particle Size: CP: 1.7 ± 2.5
/ym (78% < 2.5/ym)
PMA: 1.2 ±2.2/ym(98
% < 2.5 /ym)
PMB: 1.2 ± 2.2/ym
(98% < 2.5 /ym)
Reference: Intranasal Instillation
Dose/Concentration: CP: 10/yg/mouse;
0.5mg/kg
PSA: 0.1,1 or 10/yg/mouse; 0.005, 0.05, 0.5
mg/kg
PSB: 0.1,1 or 10/yg/mouse; 0.005, 0.05, 0.5
mg/kg
Time to Analysis: Single, 24h
Hematology: PSA and PSB decreased leukocyte
count (all 3 doses) and platelet count (2 high doses).
No effect on hemoglobin, erythrocytes and
reticulocytes was observed. Fibrinogen levels
increased for both PSB and PSA with PSB seeing a
higher increase. None of the effects were dose-
dependent.
Bone Marrow: Erythroblasts increased for PSA at
all dose levels and PSB at mid and high dose levels
(high variability).
BALF Cells: No change in BAL cell count was seen.
Quantitative cellular counts increased for
perivascular area for both groups at all dose levels.
Inflammatory cells in alveolar septum area only
increased for PSA.
Reference:
Mutlu et al. (2006,
155994)
Species: Mouse
Strain: C57BL/6
Age: 6-8wks
Weight: 20-25g
PM10
Collected by baghouse from
Dusseldorf, Germany
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 100 ng/mouse; 1
/yg/mouse; 10/yg/mouse; 100/yg/mouse
Time to Analysis: 1-7d
Alveolar Fluid Clearance: At 100 /yg/mouse,
decreased clearance peaked at 24 h and recovered at
7d.
Histology: Evidence of mild lung injury at doses of
100 /yg/mouse or more was seen.
BALF Cells: Significant increase in total cell number
was observed. Neutrophils increased but this was
not statistically significant.
Wet/Dry Ratio: Exposure did not induce any effects.
Na, K-ATPase: At 100/yg/mouse, decreased activity
of Na, K-ATPase in basolateral membranes was
observed.
Reference:
Nadziejko, et al.
(2002, 087460)
Species: Rat
Gender: Male
Strain: SHR
(spontaneously
hypertensive rats)
Aqe: 16wks
CAPs: produced at Tuxedo, NY
laboratory using centrifugal
aerosol concentrator
FA: Fine Particle Sulfuric Acid
Aerosol
UFA: Ultra-Fine Particle
Sulfuric Acid Aerosol
Particle Size: CAPs: PM2.6;
FA: 160nm; UFA: 50-75nm
Route: Nose-only Inhalation (implanted blood
pressure transmitters)
Dose/Concentration: CAPS 80, 66/yg/m3; avg
73/yg/m3
FA 299, 280,119, 203 /yg/m3; avg 225 /yg/m3
UFA 140, 565, 416, 750 /yg/m3; avg 468 /yg/m3
Time to Analysis: 10 exposures of 4h each,
each exposure at least 1 wk apart.
(2 exposures to CAPs, 4 to FA and 4 to UFA)
Respiratory rate: CAPs decreased the respiratory
rate as did FA at all dose levels. However, the FA-
induced respiratory rate was not statistically
significant unless the data was combined. UFA
increased this rate significantly.
Heart Rate: CAPS depressed the heart rate
significantly during exposures (data only significant
when combined) but returned to normal post-
exposure. FA induced a decrease as well with
continuation 7h post-exposure. UFA increased heart
rate.
Diastolic Blood Pressure: Caps and FA induced a
decrease but this was not statistically significant.
UFA induced a slight increase.
July 2009
D-107
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Nemmaret al.
(2007,156800)
Species: Rat
Gender: Male
Strain: Wistar
Kyoto
Age: 16wks
Weight: 424 ± 8
DEP: SRM 2975
Particle Size: < 1/ym
Route: Intravenous Injection
Dose/Concentration: 0.02, 0.1 or 0.5 mg/kg
Time to Analysis: single, 24h
Cardiovascular: DEP depressed blood pressure at all
doses approximately equally. DEP depressed heart
rate at all doses equally.
Hematology: No effect on number of platelets,
granulocytes, monocytes, lymphocytes or RBC s. Tail
bleeding time (associated with platelet activity)
decreased at doses of 0.02 and 0.5mgfkg.
BALF Cells: Marked cellular influx at all dose levels
Was observed. Macrophages increased at the high
dose, but this was not statistically significant. PMN
increased significantly at all dose levels.
Wet/Dry Ratio: All dose levels induced increases.
Reference:
Nemmaret al.
(2003,087931)
Species: Hamster
Gender: Male and
Female
Weight: 100-110g
PS: Polystyrene particles
PSC: Polystyrene particles,
Carboxylate modified
PSA: Polystyrene particles,
Amine modified
Particle Size: PS, PSC, PSA-
60: 60nm; PSA-400: 400nm
Route: IT Instillation
Dose/Concentration: 5, 50 or 500 //gfanimal;
0.05, 0.5, 5mg/kg
Time to Analysis: Single, 10min post-exposure
Rose Bengal administered to induce thrombosis,
immediate study thereafter
Thrombosis: Only PSA-60 at 50 and 500 levels
enhanced thrombus formation but not in a dose-
dependent manner. At 500 /yg, PSA-60 showed
evidence of pulmonary thrombosis. No effect with
PSA-400 was seen.
BALF Cells: Both PSA-60 and PSA-400 (PSA-
60 > PSA-400) induced a massive influx of PMNs.
PSA-60 effect may exhibit some dose-dependency.
BALF Inflammatory/Injury Markers: Small
increases in total protein were seen at 500 fjg level
for both PSA-60 and PSA-400. LDH was increased at
all PSA-60 levels but not for 500ug PSA-400.
Histamine increased for all PSA-60 levels and PSA-
400 but due to high variability only the effect at 500
/yg PSA-60 was statistically significant.
Hematology: PSA-60 and PSA-400 had an effect on
platelet closure time at very low concentrations: 3
and 9//gfl respectively and plateaued thereafter..
Reference:
Nemmaret al.
(2003, 097487)
Species: Hamster
Gender: NR
Weight: 100-110g
DEP: SRM 1650
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 50 //gfanimal
Time to Analysis: Single exposure, parameters
measured 1, 3, 6 or 24h post exposure.
BALF Cells: DEP led to a significant PMN flux at 1 h
(13% of total cell number), 6 h (22%) and 24 h
(37%).
Thrombosis: DEP induced a significant increase in
the cumulative mass of in vivo generated thrombus
when compared to control subjects.
Hematology: No decrease in platelet count was
observed. Consistent (non time-dependent) decrease
in closure time signified increased platelet activation
for DEP-exposed groups.
Histamine: Concentrations in BALF were
consistently elevated starting at 1 h. Plasma
histamine did not increase until 6 h.
Pretreatment with histamine receptor
antagonist: A major decrease in DEP induced PMN
infiltration was seen. Thrombogenicity was
decreased after 6h as was closure time shortening.
No effect on histamine in BALF or plasma was
observed.
Reference:
Nurkiewicz et al.
(2009,191961)
Species: Rat
Gender: Male
Strain: SD
Age: 7-8wks
Weight: NR
Fine Ti02 (Sigma-Aldrich-
(titanium (IV)) oxide, 224227,
St. Louis, M0) (~ 99% rutile)
TiO/ nanoparticles (DeGussa-
Aeroxide Ti02 P25, Parsippany,
NJ) (80% anatase, 20% rutile)
Particle Size: Fine Ti02-
Primary size: < 5 fjm, MMAD:
402nm, CMD: 710nm; Nano-
Ti02- Primary size: 21 nm,
MMAD: 138nm.CMD: 100nm
Route: Aerosol Inhalation
Dose/Concentration: 1.5-16mg/m3
Time to Analysis: Acclimated 5d. Exposed 240-
720min. Anesthetized 24h postexposure.
Intravital microscopy, NO measurement,
microvascular oxidative stress measurement,
nitrotyrosine staining.
Arteriolar Dilation: Nano-Ti02 significantly impaired
endothelium-dependent arteriolar dilation. Equivalent
levels of arteriolar dysfunction were found in fine
and nano-Ti02. Arteriolar dilation in response to
abluminal microiontophretic application of SNP was
not different from the controls or between the
exposure groups. Arteriolar dilation was partially
restored by radical scavenging with TEMP0L and
catalase, NADPH oxidase with apocynin, and MP0
inhibition with ABAH.
Microcirculation: R0S increased in both groups.
Nano-Ti02 significantly increased the area of tissue
containing nitrotyrosine in the lung and
spinotrapezius microcirculation.
NO: Fine and nano-Ti02 significantly and dose-
dependently decreased stimulated NO production in
isolated microvessels. NO production was increased
by radical scavenging with TEMP0L and catalase or
NADPH oxidase with apocynin, and was largest in
the fine Ti02 group.
July 2009
D-108
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Pereira Ambient Particles
et al. (2007,
156019)
Species: Rat
Gender: Male
Strain: Wistar
Aqe: 3m
(Porto Allegre, Brazil)
Particle Size: < 10/ym
Route: Whole-body Inhalation
Dose/Concentration: P-6: 34, 22 or 225 /yglm3
P-20: 139 or 112 //g/m3
P-l: 99/yglm3
Time to Analysis: P-6: single/continuous for 6h
P-20: single/continuous for 20h
P-l: intermittent (5h) periods per day for 4d
consecutively
Parameters measured 0 or 24 h post-exposure
BALF Inflammatory/Injury Markers: An increase
in lipid peroxidation was statistically significant only
for the 20 h continuously exposed group.Leukocytes
also increased at P-20. No change at P-6. Total
protein remained unaffected at all dose levels.
Wet to Dry Ratio (Oh): No effect was observed.
Reference:
Pinkerton et al.
(2004, 087465)
Species: Rat
Gender: Female
(pregnant),
Offspring- NR
Strain: SD
Age: 10d (pups),
Pregnant females-
10-14d of gestation
Weight: NR
PM (Fe and soot from
combustion of acetylene and
ethylene in a laminar diffusion
flame system)
Particle Size: Median
diameter: 72-74nm; size range:
10-50nm
Route: Inhalation
Dose/Concentration: Mean mass
concentration: 243 ±34/yglm3; Average Fe
concentration: 96 /yglm3
Time to Analysis: Exposed 10d postnatal a
6hfd, 3d (consecutive). Bromodeoxyuridine
injected 2h before necropsy.
A significant reduction of cell proliferation occurred
only within the proximal alveolar region of exposed
animals compared to controls. There were no
significant differences between the groups for
alveolar formation and separation within the proximal
alveolar region.
Reference:
Pinkerton et al.
(2002, 087645)
Species: Rat
Gender: Male,
Female
Strain: SD
Age: 11-13wks
(adult male), 10-
12d (neonatal)
Weight: NR
PM (Fe, Soot) (ethylene, iron
pentacarbonyl, acetylene
combined: Fe203; soot: 60%
EC, 40% OC)) (CO, NO,)
Particle Size: Fe (diameter)-
40nm; Soot (primary particles,
diameter)- 20-40nm
Route: Whole-body Exposure
Dose/Concentration: Adult males: Fe- 57, 90
/yglm3, Soot- 250 /yglm3, Fe+Soot- Fe: 45
/yglm3, Total PM: 250 /yglm3; Neonates:
Fe+Soot- Low: Fe- 30 /yglm3, Total PM: 250
/yglm3, High: Fe- 100 /yglm3, Total PM: 250
/yglm3
Time to Analysis: Adult males exposed to Fe,
soot, Fe+Soot, or filtered air. Exposed 6h/d, 3d
(consecutive). BAL performed within 2h
postexposure, lung tissue evaluations 24h
postexposure. Pregnant dams housed from GD 3
to weaning at 21 d-old. Neonatal rats exposed to
Fe+Soot 10-12d-old and 23-25d-old.
Fe: Only the high dose had significant effects. This
dose increased total protein in the lavage fluid,
decreased total antioxidant power, induced GST
activity, and induced a non-significant, increasing
trend of GSH and GSSG. IL-1 (3, intracellular ferritin,
and NF-kB increased.
Fe+Soot, Soot: Fe+Soot significantly reduced the
total antioxidant power in BALF and supernatant
from lung tissue homogenate. Fe+Soot significantly
increased GSSG, IL-1(3, NF-kB, CYP1A1, and
CYP2E1. CYP2B1 increased but was not significant.
Soot alone was not significant for anything.
Neonates: The high-dose significantly decreased cell
viability, increased LDH activity, and increased IL-1 (3
and ferritin. Both doses significantly increased
GSSG, GRR, and GST, and decreased total
antioxidant power.
Reference: Pires-
Neto et al. (2006,
096734)
Species: Mouse
Gender: Male
Strain: Swiss Age:
6d
Ambient Air: PM2.6, NO2 and CB
(Sao Paulo, Brazil)
Particle Size: PM2.6
Route: Whole-body Inhalation
Dose/Concentration: PM2 e: 46.49 /yglm3
Control: 18.62 /yglm3
NO2: 59.52/yglm3
Control: 37.08/yglm3
CB: 12.52/yglm3
Control: 0/yglm3
Time to Analysis: 24h|d, 7d|wk for 5mo
(weaned at 21 d into exposure - mothers
removed)
Nasal Cavity: Increased total mucus and acidic
mucus at proximal and medial areas of cavity.
Nonsecretory epithelium declined. No significant
changes in amount of neutral mucus, volume
proportion of neutral mucus, volume proportion of
total mucus, thickness of epithelium, volume
proportion of nonsecretory epithelium or ratio
between neutral and acidic mucus were observed.
Types of Acidic Mucus Cells: Proximal and medium
cells increased. Effects on distal cells were
equivocal.
Reference:
Pourazar et al.
(2005, 088305)
Species: Human
Gender: Male and
Female (nonatopic
81 nonsmokers)
Age: 21 -28yrs
DEP: generated from idling
Volvo diesel engine
DEP 300 /yglm3 comprised of:
NO21.6ppm
NO 4.5ppm
CO 7.5ppm
Hydrocarbons 4.3ppm
Formaldehyde 0.26mg|m3
Suspended particulates
4.3x10"6|cm3
Particle Size: < 10/ym
Route: Whole-body exposure chamber
Dose/Concentration: DEP 300 /yglm3
Time to Analysis: Single exposure for 1 h.
Parameters measured 6h post exposure.
Transcription Factors: Exposure induced increased
cytoplasmic and nuclear immunoreactivity of phos-
phorylated p38 MAPK in bronchial epithelium.
Increased nuclear translocation of phosphorylated
p38 and JNK MAPK as well as increased nuclear
phosphorylated tyrosine immunoreactivity were
observed. No change in total or nuclear c-fos
immunoreactivity was seen. Exposure induced
increased nuclear translocation of phosphorylated
JNK significantly associated with phosphorylation of
nuclear c-jun and also resulted in an increase in
nuclear p65.
Cytokines: Expression of IL-8 was positively
associated with nuclear phosphorylated p38 post-
exposure.
July 2009
D-109
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Pradhan et al.
(2005,096128)
Species: Rat
Gender: Female
Strain: Wistar
Albino
Weight: 120-180
RSPM: Respirable Suspended
PM
(Lucknow, India)
Quartz dust (positive control)
Particle Size: < 5/ym
Route: IT Instillation
Dose/Concentration: 2.5, 5.0, or 10.Omg/
0.05ml; 20, 42, 83 mg/kg
Time to Analysis: Sacrificed 15d post single
exposure.
Relative Lung Weight: A dose-dependent increase
in total lung weight of RSPM-instilied animals was
observed.
BALF Cells: Exposure induced a dose-dependent
increase in total cells dose-dependent with the low
and mid dose levels. PMNs increased massively at all
dose levels with RSPM inducing less of an increase
than Quartz. Exposure at low dose levels resulted in
an influx of inflammatory cells (predominantly
macrophages into lumen of alveolar ducts and
alveoli). Reaction at the high dose was more intense
than that seen in mid dose-exposed lungs.
BALF Inflammatory/Injury Markers: A significant
dose-dependent increase in LDH and NO was
observed, but the Quartz-induced increase was
greater than the RSPM-induced increase. An increase
in protein was significant at the mid dose level for
RSPM and significant at the high dose level for both
RSPM and Quartz.
Lung biochemistry: An increase in lipid peroxidation
was dose-dependent. Superoxide dismutase (SOD)
enzyme levels showed a dose-dependent decrease.
Reference: Ramos
et al. (2009,
190116)
Species: Guinea
Pig
Gender: NR
Strain: -
Age: NR
Weight: 330-370g
WS (Pine wood)
(C0(<80ppm), 002(0.35%),
02(20.1%), PM2.6, PM10)
Particle Size: PM2.6, PM10
Route: Whole-body Inhalation Chamber
Dose/Concentration: WS: 60g, PM2.6: 363±23
/yg/m3, PM10: 502±34/yg/m3
Time to Analysis: Exposed 3h, 5dfwk for 1, 2,
3, 4, 6, 7m.
WS significantly decreased body weight between 4
and 7m exposure. The concentration of blood
carboxyhemoglobin increased. Recovered BAL cells
were higher in WS-exposed pigs. Macrophages and
neutrophils increased. Inflammation in the lungs was
seen. Pulmonary arterial hypertension and
emphysematous lesions were observed. Macrophage
and lung tissue homogenate elastolysis increased.
Collagenolysis increased. Generally, MMP-2, MMP-9,
and MMP-1 increased. BAL macrophage apoptosis
increased with time.
Reference: Rao et DEP: SRM 2975
al. (2005, 095756) partjc|e Sjze: g 5/ym
Species: Rat
Strain: Sprague-
Dawley
Weight: 175g
Route: IT Instillation
Dose/Concentration: 5, 35, 50mg/kg bw
Time to Analysis: Sacrificed 1, 7, 30d post
single exposure. Cytokines measured after 24h
incubation (in vitro).
BALF Inflammatory/Injury Markers: Increased
albumin at 1 and 30d at all dose levels. Increased
LDH except at low dose at 7d.
BALF Cells: Macrophages unaffected. Increased
PMNs at 1 d for all dose levels, sustained elevation at
7d for mid and high dose and at 30d for all dose
levels.
Cytokines: The high dose induced a significant
increase of mRNA expression for IL-1 (3, iNOS, MCP-
1, and MIP-2 in BAL cells. MCP-1 mRNA sustained
high levels at 7d for mid and high dose and at 30d for
all dose levels. mRNA expression of IL-6, IL-10, TGF-
(31, TNF-a were unaffected. However, IL-6 and
MCP-1 proteins increased significantly in BALF at 1d
for mid and high dose, returning to basal levels at 7d.
MIP-2 increased for all dose levels at all time points.
NO level unaffected.
July 2009
D-110
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Reed et
al. (2006,156043)
Species: Rat,
Mouse
Gender: Male and
Female
Strain: CDF
(F344)/CrlBR (rat),
SH (rat), A/J
(mouse), and
C57BL/6 (mouse)
Age: 6-12wks
HWS (burned mix of hardwood
in noncertified wood stove
using a Pineridge model 27000,
Heating and Energy Systems,
Inc. Clackamas, OR)
Measured Components: EC,
OM, NO3, SO4, NH4, metals
Particle Size: ~0.25//m
Route: Whole-body Inhalation
Dose/Concentration: Low: 30 //gfm3
Mid-low: 100 /yglm3
Mid-high: 300/yg/m3
High: 1000 //g|m3
Time to Analysis: 6hrfd, 7dfwk for 1 wk or 6m.
Immediate post-exposure analysis.
Organ weights: Liver declined in rats of both
genders at 1 wk and female rats at 6m. Lung volume
increased and lung weight decreased in female rats
at 6m. Spleen weight increased in female mice and
rats at 1wk. Thymus weight decreased in male rats
at 1wk.
Clinical Chemistry: Cholesterol decreased at the
high dose for male rats at 1wk and 6m and increased
at mid-low and mid-high doses for female rats at 6m.
ALP decreased for rats of both genders at 1wk and
6m for mid-low, mid-high and high dose levels (14-
38%). AST decreased by 24% in male rats at 1wk
with high dose. No effect on females. Creatinine
serum levels decreased in males at 1wk at mid-high
and high dose by 13%. No effect observed at 6m.
BUNfCre ratio decreased in females at 1wk (25%)
and both genders at 6m at mid-high and high dose
(18-19%).
Hematology: Hemoglobin and hematocrit increased
in 6mo male rats. Bilirubin increased in female rats at
6m at high dose. Platelets increased for male and
female rats at 1wk (21%, 19% respectively). No
effect observed at 6m. WBC increased in males at
1wk.
Cells: Eosinphils decreased and lymphocytes
increased in males at 6m. Neutrophils decreased at
6m in both genders. Minimal increases in alveolar
macrophages and sparse brown-appearing
macrophages in all species.
Bacterial Clearance: Mice instilled with bacteria
were mostly unaffected by exposure, except for a
decline in histopathology summary score after 6m.
Tumorigenesis: No values for exposed groups
differed significantly from controls. There was no
evidence of progressive exposure related trend.
Reference: Reed et DE: generated from two 2000
al. (2004, 055625) model 5.9 L Cummins ISM
turbo diesel engines
Co-expoure to 8 gas and 8 solid
exhaust components measured
Particle Size: 0.10 ¦ 0.15/ym
Species: Rat,
Mouse
Gender: Male and
Female
Strain: CDF
(F344)/CrlBR (rat),
A/J (mouse)
Age: 12wks
Route: Whole-body Inhalation
Dose/Concentration: Low: 30 //gfm3
Mid-low: 100 /yglm3
Mid-high: 300/yg/m3
High: 1000 /yglm3
Time to Analysis: 6hfd, 7dfwk for 1wk or 6m.
Analyzed 1d post-exposure.
Organ weights: Kidney weight increased after 6m
for both males and female rats at the high dose.
Kidney and liver weight increased for female mice at
all dose levels at 6m. Lung weight increased at high
dose at 6m for female mice and male rats. Spleen
weight decreased in male mice at the low and mid-
high levels.
Clinical Chemistry: There was a massive decrease
in cholesterol (24%) for rats of both genders after
1wk and a smaller decrease for male rats at 6m.
GGT significantly increased at 6m for male and
female rats at the mid-high and high dose. ALP
increased in male rats at 1 wk by 10%. AST
decreased at mid-high (15%) and high dose in female
rats at 6m. BUN and BUN/Creatine declined (19%,
17%) in female rats at mid-high and high doses after
6m. BUN increased by 21% at mid-low, mid-high and
high doses in male rats at 1wk.
Hematology: WBC decreased in high females after
6m. Factor VII (blood clotting) decreased in MH and
HR males after 1wk and male and female HR after
6m. Thrombin-antithrombin complex declined
massively but only in males after 1wk.
Cells: Minimal increases in alveolar macrophages and
PM within the macrophages were seen.
Cytokines: TNF-a decreased in female rats after
6m.
Tumorigenesis: No significant effect was observed.
July 2009
D-111
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Reed et al. (2008,
156903)
Species: Mouse
Gender: Male,
Female
Strain: C57BL/6,
A/J, BALB/c
Age: NR
Weight: NR
GEE (2 1996 General Motors
4.3-L V-6 engines; unleaded
gasoline)
Particle Size: 150nm (MMAD)
Route: Whole-body Inhalation
Dose/Concentration: Control: 2.5±2.9, Low-
exposure: 6.6 ± 3.7, Mid-exposure: 30.3 ±11.8,
High-exposure: 59.1 ±28.3, High filtered
exposure: 2.3 ± 2.6 /yglm3
Time to Analysis: Exposed 6hfd, 7dfwk, 3d-6m.
Body and organ weight and histopathology in
A/J: Kidney weight decreased, but no effects
pertaining to weight were significant. No visible
inflammatory changes were seen.
Lung Damage in A/J: No significant effect was
seen, but hypomethylation was seen in females at
1wk, and methylation was reduced in all exposed
female groups.
Bacteria in lungs of C57BL/6: Exposure did not
affect the clearance of bacteria from the lung.
Respiratory allergic response in BALB/c:
Exposure had little effect, but serum total IgE
increased significantly for the high-exposure group.
Increasing trends were seen in OVA-specific serum
IgE and lgG1, as well as neutrophils and eosinophils.
Reference: Reed et GEE (2 1996 General Motors
al. (2008,156903) 4.3-L V-6 engines: unleaded
Species: Rat gasoline)
Gender: Male,
Female
Strain: CDF
(F344)/CrlBR, SH
Age: NR
Weight: NR
Route: Whole-body Inhalation
Dose/Concentration: Control: 2.5±2.9//g/m3.
Low-exposure: 6.6±3.7/yglm3, Mid-exposure:
30.3± 11.8//g/m3. High-exposure: 59.1 ±28.3
/yg/m3, High filtered exposure: 2.3±2.6/yglm3
Particle Size: 150nm (MMAD)
Time to Analysis: Exposed 6hfd, 7dfwk, 3d-6m.
Body and organ weight and histopathology in
F344: There were no significant effects pertaining to
weight, but heart weight increased and seminal
vesicle weight decreased. No visible inflammatory
changes were seen.
Serum chemistry, hematology, clotting factors
in F344: Serum alanine aminotransferase, aspartate
aminotransferase and phosphorous decreased in
females at 6m in the mid- and high-exposure groups.
Hematocrit, red blood cell count and hemoglobin
increased at 1wk and 6m in males and females in the
mid- and high-exposure groups. Plasma fibrinogen
increased in males at 1wk in the mid-exposure group.
Lung DNA Damage in F344: No significant effect
was seen, but hypermethylation occurred in male
rats at 6m in the mid- and high-exposure groups.
BALF in F344: Generally, in the high-exposure
group, increases were seen in LDH and MIP-2, and
decreases were seen in hydrogen peroxide produced
by unstimulated macrophages and superoxide by both
stimulated and unstimulated macrophages.
SH: HR or ECG parameters were not affected. In the
high-exposure group, lipid peroxides increased in
males and TAT decreased in males and females.
Reference: Reed et
al. (2008,156903)
Species: Mouse
Gender: Male,
Female (only
BALB/c)
Strain: C57BL/6,
A/J, BALB/c
Age: NR
Weight: NR
GEE (two 1996 General Motors
4.3-L V-6 engines: regular,
unleaded, non-oxygenated, non-
reformulated gasoline blended
to US average consumption for
summer 2001 and winter
2001-2002-Chevron-Phillips)
Particle Size: MMAD: 150nm
Route: Inhalation Exposure Chamber
Dose/Concentration: PM: Low- 6.6±3.7
/yg/m3, Medium- 30.3 ± 11.8 /yglm3, High-
59.1 ±28.3//g/m3
Time to Analysis: A/J- 2wk quarantine period in
chamber. Exposed 6h/d, 7d/wk, 3d-6m. C57BL/6-
1wk exposure. Instillation of P. aeruginosa.
Killed 18h postinstillation. BALB/c- Conditioned
to exposure chambers and mated. Pregnant
females exposed GD 1 and throughout gestation.
Offspring exposures continued until 4wks-old.
Half of offspring sensitized to OVA. Tested for
airway reactivity by methacholine challenge 48h
post instillation and euthanized.
The kidney weight of female A/J mice decreased at
6m and was strongly related to PM by the removal of
emission PM. PM-containing macrophages increased
by 6m. Hypomethylation occurred in females at 1wk.
The clearance of P. aeruginosa was unaffected by
exposure. Serum total IgE significantly and dose-
dependently increased. OVA-specific IgE and lgG1
gave slight exposure-related evidence but were not
significant.
July 2009
D-112
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Reed et
al. (2008,156903)
Species: Rat
Gender: Male,
Female
Strain: CDF
(F344)/CrlBR, SHR
Age: NR
Weight: NR
GEE (two 1996 General Motors
4.3-L V-6 engines; regular,
unleaded, non-oxygenated, non-
reformulated Chevron-Phillips
gasoline, U.S. average
consumption for summer 2001
and winter 2001-2002)
Particle Size: MMAD: 150nm
Route: Inhalation Exposure Chamber
Dose/Concentration: PM: Low- 6.6±3.7
/yg/m3, Medium- 30.3 ± 11.8 /yg/m3, High-
59.1 ±28.3//g/m3
Time to Analysis: 2wk quarantine period in
chamber. Exposed 6h/d, 7d/wk, 3d-6m. SHR-
surgery to implant telemeter in peritoneal cavity.
4wks recovery. ECG data obtained every 15min
beginning 3d preexposure, 7d exposure, 4d
postexposure.
Organ Weight: At 6m exposure, the heart weights
of male and female rats increased and male rats'
seminal vesicle weight decreased.
Histopathology: PM-containing macrophages
increased by 6m.
Serum Chemistry: Serum alanine aminotransferase,
aspartate aminotransferase, and phosphorus
decreased in medium and high-exposure females.
Hematology, Clotting Factors: Hematocrit, red
blood cell count, and hemoglobin dose-dependently
increased for both genders at both time points.
Plasma fibrinogen increased at 1wk in males.
Lung DNA Damage: Hypermethylation occurred in
medium- and high-exposure male rats at 6m.
BAL: For both genders in the high-exposure group,
LDH and MIP-2 significantly increased at 6m. R0S
decreased at 1wk and 6m. Generally, the production
of hydrogen peroxide and superoxide decreased in the
high-exposure group and medium- and high-exposure
groups, respectively.
CV effects in SHR: Lipid peroxides were
significantly increased in males in the high exposure
group. TAT complexes decreased in females in the
high exposure group.
Removal of Emission PM: The removal of emission
PM strongly linked PM to increased seminal vesicle
weight, red blood cell counts, LDH, lipid peroxides,
and methylation.
Reference:
Rengasamy et al.
(2003,156907)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Weight: — 200g
DEP: SRM1650
CB Elftex-12 furnace black,
Cabot, Boston, MA
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 5,15, or 35 /yg/kg bw
Time to Analysis: single; 1, 3, 5, 7d post
exposure
CYP1A1: DEP at all doses significantly increased
CYP1A1 protein, was maximal at 1d, and normalized
at 5d. CB had no effect.
CYP2B1: DEP and CB at 15 and 35 mg/kg inhibited
activity at 1 d.
Protein level significantly decreased at 1 d with 5,
15 and 35 mg/kg DEP and at 15 and 35 mg/kg CB. A
time dependent decrease was shown at 35mg/kg for
both DEP and CB.
Reference:
Renwick et al.
(2004, 056067)
Species: Rat
Gender: Male
Strain: Wistar
Weight: 370-470g
FCB: Fine Carbon Black (Huber
990)
UCB: Ultrafine Carbon Black
(Printex 90, Degussa)
FT0: Fine Titanium Dioxide
(Tioxide)
UT0: Ultrafine Titanium dioxide
(Degussa)
Particle Size: FCB: 260nm;
UCB: 14nm; FT0: 250nm;
UT0:29nm
Route: IT Instillation
Dose/Concentration: 125 or 500 /yg/rat
Time to Analysis: Single, 24h
BALF Cells: UT0 and UCB induced a large dose-
dependent increase in percent neutrophils (only
statistically significant at 500 /yg for UT0).
BALF Inflammatory/Injury Markers: UT0 and UCB
also increased total protein content only at the 500
/yg dose. UCB induced LDH release at 125 and 500
/yg, UT0 and CB at 500/yg. UT0 and UCB induced
large dose-dependent increases in GGT activity (only
statistically significant at 500 /yg for UT0).
Phagocytosis: All 4 particles decreased but only at
the 500 /yg level.
Chemotaxis: Only UT0 and UCB at 500 /yg/l
increased chemotactic migration.
Reference: Rhoden
et al. (2004,
087969)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Weight: 250-300g
CAPs
(Boston, MA)
Particle Size: CAPS: 0.1-2.5
/ym
Route: Whole-body Inhalation
Dose/Concentration: 1060 ± 300/yg/m3
Time to Analysis: Single exposure for 5h.
Analyzed 24 h post exposure.
(CAPS-NAC - CAPS with 50mg/kg bw NAS (N-
acetylcysteine) pretreatment)
Particle Characteristics: Major components did not
appear to show any correlation to total particle
mass. Included Na, Mg, A I, Si, S, CI, K, Ca, Ti, V, Cr,
Mn, Fe, Ni, Cu, Zn, Br, Ba, Pb. Metals Al, Si and Fe
(somewhat less for Pb, Cu, K) correlated with
TBARS.
Oxidative Stress: CAPS increased TBARS and
oxidized protein by 2 + fold. NAS fully prevented the
increase in TBARS and partially prevented an
increase in protein carbonyl.
Tissue Damage: Wet/dry ratio increased with CAPS
but significantly decreased with NAC.
BALF Cells: CAPS increased PMN 4 fold. NAS
treatment reduced this increase to control levels.
BALF Inflammatory/Injury Markers: LDH and total
protein not affected. Histology confirms slight
inflammation with CAPS and no inflammation with
CAPs-NAC.
July 2009
D-113
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Rhoden
et al. (2008,
190475)
Species: Rat
Gender: Male
Strain: SD
Age: NR
Weight: 300g
Urban Air Particles (UAP) (SRM
1649)
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 1 mg in 100 //L saline
Time to Analysis: Some rats pre-treated with
MnTBAP 2h prior to UAP exposure. Instilled with
UAP. CL analysis 15min postexposure.
Anesthetized 4h postexposure for BAL
measurements.
UAP significantly increased the total cell number,
PMN, MP0 activity, and protein levels. MnTBAP
prevented UAP-induced lung inflammation. UAP
increased oxidants in lung CL, which was prevented
by MnTBAP.
Reference: Rivero Ambient Air
et al. (2005,	(Sao Paulo, Brazil)
088653)
Particle Size: < 2.5/ym
Species: Rat
Gender: Male
Strain: Wistar
Age: 3m
Weight: 250g
Route: IT Instillation
Dose/Concentration: 100 or 500 //gfrat; 0.4 or
2mg/kg
Time to Analysis: single, 24h
Hematology: Reticulocytes increased at both doses.
At the high dose, hematocrit, percent segmented,
percent neutrophils increased and percent
lymphocytes decreased (relative to control or 100 -
very high variability). Fibrinogen decreased at low
dose but not at high dose.
Histopathology: At both doses, acute alveolar
inflammation was observed and was more
pronounced in the 500 /yg group.
Lung Morphometry: Lumen wall ratio values show
a dose-dependent increase in peribronchial as well as
intra-acinar pulmonary arterioles. No effect in
myocardial arterioles were observed.
Tissue Damage: Increase in the heart wet/dry ratio
at high dose was observed. Lung wet/dry ratios were
unaffected.
Technology: Laser capture microdissection of
airway cells were used to analyze results.
Protein: pERK1/2: ERK1/2 ratio increased by 60%
at 6h and 80% at 24 h. NFkB activity increased at
6h but was not statistically significant.
Reference:
Roberts, E.S.
Charboneau, L.
Espina, V.
2004
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age: 60-90d
Weight: 300-350g
R0FA: SRI (cyclone power
plant)
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 0.5 mg/rat; 1.67mgfkg
Time to Analysis: Single, 6 and 24 h
Reference: Saber
et al. (2005,
097865)
Species: Mouse
Gender: Female
Strain: TNF(-I-)
(B6,129S-
Tnftm1Gk1),
C57/BL
Age: 9-10wks
DEP: SRM 2975
CB: Printex 90 (Degussa)
Particle Size: DEP: 215nm;
CB: 90nm
Route: Nose-only Inhalation
Dose/Concentration: DEP: 20 mgfm3; CB: 20
mg/m3
Time to Analysis: 90minfd for 4d
consecutively, 1h
BALF Cells: Neutrophils increased significantly to
15% when compared to control (4%) with DEP
exposure. No response difference was observed
between TNF (+/+) and TNF(-/-). CB did not induce
any changes in neutrophil numbers.
Cytokines: IL-6 increased 2-3 fold in DEP and CB
exposure in both normal and knockout mice. IL-1f3
was unaffected.
mRNA: In TNF (+/+) mice, DEP and CB increased
expression of TNF mRNA 2- fold. IL-6 mRNA
expression was high in DEP-exposed knockout mice
when compared to normal mice.
DNA: DNA strand breaks increased in both strains.
Knockout mice showed a higher response to CB and
DEP exposure. For normal mice, only CB induced a
statistically significant effect.
Reference: Schins Soluble fractions
et al. (2004,
054173)
Species: Rat
Gender: Female
Strain: Wistar
Weight: 350-550g
PMC: PMio-2.5
PMF: PM2.6
¦B: Borken, Germany (rural)
¦D: Duisburg, Germany
(industrialized)
Particle Size: PM10 2.5, PM2.6
Route: IT Instillation
Dose/Concentration: 0.32 ± 0.01 mg/rat;
0.91 ± 0.58 mg/kg
Time to Analysis: single, 18h
Radical Formation: Formation of hydroxl radicals
increased with exposure. Relative intensity is as
follows: PMC-D, PMF-D, PMC-B, PMF-B, and control.
Cytokines: TNF-a and IL-8 increased with PMC
from both sites. PMF induced a slight increase in IL-8
but did not induce an increase in TNF-a.
BALF cells: Both PMC showed a massive increase in
neutrophils. PMC-B induced the greatest increase
followed by PMC-D. Both PMF did not induce a
significant increase.
BALF Inflammatory/Injury Markers: PMC from
both sites induced markedly higher endotoxin
concentration vs PMF as follows in decreasing order:
PMC-B, PMC-D, PMF-B, PMF-D, control. Glutathione
decreased only for PMC-B. LDH and total protein
were unaffected.
July 2009
D-114
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Seagrave et al.
(2005, 088000)
Species: Rat
Gender: Male
Strain: F344/DCrl
BR
Aqe: 11 ±1 wk
PM from 3 sources:
NT: New Technology bus,
Detroit Diesel 50G, exhaust
oxidation catalyst, 216 miles,
2002 model - in use
NE: Normal emitter bus, Detroit
Diesel 50G, no catalyst,
134259 miles, 1997 model ¦ in
use
HE: High Emitter bus, Cummins
L1OG, no catalyst, > 250, 000
miles, 1992, retired
Fuel composition very similar
for 3 vehicles: methane (96-
96.8%), ethane (1.6-1.9%),
carbon dioxide (0.9-1.1%),
nitrogen (0.6-0.8%), traces of
other gases
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 0.25 - 2.2 mg/rat in
0.5mL saline
Time to Analysis: Sacrificed 24h post single
instillation.
Engine Specific Emission data: HE had
significantly higher PM and SV0C recovered emission
rates than NE and NT.
Organic mass in PM: The following PM sources are
listed in decreasing order of percent of total mass:
HE, NE, NT.
Total PAH: The following PM sources are listed in
decreasing order of total mass: HE, NT, Control, NE.
Nitro PAH: The following PM sources are also listed
in decreasing order of total mass: NE, HE, Control,
NT. Authors note confounding technical issues
(mostly technique related) with mostly mild effects.
BALF Inflammatory/Injury Markers: LDH showed
dose-dependent increases withHE inducing higher
increases than NT and NE. Total protein exhibited
dose-dependent increases with HE, NT and the
positive control SRM2975 inducing higher levels than
NE.
Potency Factors Cytotoxicity and Inflammation:
HE was significantly more potent than NT and NE,
with NT also showing significant potency.
Lung Toxicity: The results were highly variable but
the general toxicity levels in increasing order is the
following: NE, NT, HE, Normal gasoline, diesels, and
high gasolines, though individual factors may differ
greatly.
Reference:
Seagrave et al.
(2006,091291)
Species: Rat
Gender: Male
Strain: F344|Crl
BR,
Aqe: 11 ±1 wk
PM2.6 sources: BHM: Birming-
ham, Alabama; urban
JST: Jefferson Street, Atlanta,
Georgia; urban
PNS: Pensacola, Florida; urban/
residential
CTR: Centreville, Alabama;
rural
"smoke" - downwind of forest
fires/burns (NR)
Particle Size: PM2.6
Route: IT Instillation
Dose/Concentration: 0.75,1.5, 3 mg/rat
Time to Analysis: Sacrificed 24h post single
instillation.
BALF Total cells and PMN: In general, the winter
samples induced greater increases in potency than
the summer samples except for PNS. For the winter
samples, the samples that induced the greatest
increases, in descending order, are: JST, BHM, CTR,
PNS and Smoke. For the summer, the samples that
induced increases, in descending order, are: BHM,
JST, PNS, and CTR.
BALF macrophages: For the winter, the BHM and
JST samples significantly increased potency whereas
the PNS sample induced significantly negative
potency. For the summer, only the BHM sample
significantly induced potency.
BALF lymphocytes: Only the JST-W and BHM-W
significantly increased potency. The BHM-S, CTR-S
and PNS-S also significantly increased potency.
Histopathological Inflammation: All the winter
and summer samples, excepting PNS, significantly
induced potency.
Lung weight/body weight ratio: In general, for all
end points, JST-S was significantly less potent than
JST-W. The summer samples of BHM and CTR were
also generally more potent than their winter
counterparts.
July 2009
D-115
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Seagrave et al.
(2005, 088000)
Species: Rat
Gender: Male,
Female
Strain: CDF(F-
344)/CrlBR
Age: 10-12wks
DE:
(Two 6 cyl Cummins ISB turbO)
HWS - hardwood smoke
(mixed black/white oak,
uncertified conventional wood
stove)
DE:
EC - 557
OC - 269/yg/m3
NO - 45 ppm
NO2 - 4 ppm
CO - 30 PPM
THV - 2 ppm
HWS:
EC - 43
OC - 908/yg/m3
NO or NO2 - 0 ppm
CO - 13 ppm
THV - 3 ppm
Particle Size: DE: 0.14 ± 1.8
fj m
Route: Whole-body Inhalation
Dose/Concentration: 30,100, 300,1000
/yg/m3TPM
HWS: 0.36 ± 2.1 fjm (MMAD + GSD
Time to Analysis: 6hfd, 7dfwk for 6m. 1d post-
exposure
Particle Characteristics: Major differences K:
HWS > > DE; Ca DE > > HWS; Zn: DE > > HWS.
BALF Inflammatory/Injury Markers: LDH was
unaffected by DE. Exposure to HWS induced an
increase for males only at 100 and 300 but not at
1000/yglm3. Protein was unaffected by DE. HWS
exposure showed male-only effects at 100 and 300
/yg/m3 but not at 1000. AP was unaffected by DE or
HWS except for slight decline induced by HWS at
1000 /yglm3 for both genders.
Other: P glucose was unaffected by DE. HWS-
exposed females showed decreased fB-glucose at
100 and 300 but not at 1000/yg/m3.
BALF GSH to (GSH+ GSSG): No effects for DE
were observed. HWS significantly decreased the
ratio in both males and females at 1000 //gfm3. The
effect for females was greater than the male effect.
BALF Cells: No effects were observed except for an
increase in macrophages at 30 /yglm3 for HWS males
exposed to HWS.
Cytokines: IL-1 (3 was unaffected by DE or HWS.
MIP-2 decreased for both genders at 1000 HWS.
TNF-a decreased in females with DE exposure. No
TNF-a effects for HWS were observed.
Reference:
Seagrave et al.
(2008,191990)
Species: Rat
Gender: Male
Strain: SD
Age: 10-12wks
Weight: 250-300g
GEE (2 1996 General Motors
4.3-L V6 gasoline engines;
conventional Chevron Phillips
gasoline, U.S. average
composition) (CO, NO, NO2,
SO2, THC) (PM2.6 composition-
EC, 0C, S04, NH4, NOs)
Simulated downwind coal
emission atmospheres (SDCAs)
(fly ash, gas-phase pollutants,
sulfate aerosols, NO, NO2, SO2)
Paved Road Dust (RD) (Los
Angeles, CA; New York City,
NY; Atlanta, GA)
Particle Size: GEE: MMAD-
150nm, RD: 2.6 ± 1.7 fjm,
SDCA: 0.1-1.O^/m
Route: Nose-only Inhalation
Dose/Concentration: GEE: 60//gfm3, SDCAs:
317-1072/yglm3, RD: 306-954//gfm3; GEE: CO-
104ppm, NO- 16.7ppm, N02- 1.1 ppm, S02-
1 .Oppm, THC- 12ppm; SDCAs: CO- < 1 ppm, NO-
0.19-0.62ppm, N02- 0.10-0.37ppm, S02- 0.07-
0.24ppm, THC- < 1ppm
Time to Analysis: Quarantined 2wks. 6h
exposure then ip injected. Cannula ligated into
trachea and connected to rodent ventilator.
Thorax and abdomen opened. Killed after
measurements taken.
GEE produced CL in the lungs, heart, and liver. RD
produced a significant effect in the heart at the low
dose. SDCAs had no effect on CL. RD significantly
increased the heart's oxidative stress, as
demonstrated by the TBARS levels. GEE did not
affect the amount of macrophages or PMN. SDCAs
increased macrophages. The RD low dose increased
macrophages and PMN. SDCAs increased Pur* values
and tidal volumes.
Reference: Singh
et al. (2004,
087472)
Species: Mouse
Gender: Female
Strain: CD-1
Age: 6-8wks
A-DEP
(4cyl light duty 2.7I Isuzu diesel
at 6 kgfni]
DEP: SRM 2975
Particle Size: A-DEP >50
fj m
Route: Oropharyngeal Aspiration
Dose/Concentration: 25 or 100//gfmouse
Time to Analysis: single, 4h
(18h post-exposure measurements taken but NR
due to similar results)
Particle Characteristics: DEP had 60% EC vs 9%
in A-DEP. A-DEP had 50% OC vs 5% in DEP.
Phenanthrene and Fluoranthene fractions were much
more prevalent in PAH from DEP than A-DEP.
BALF Inflammatory/Injury Markers: Microalbumin
increased for both pollutants except DEP induced
increases only at 100/yg. Endotoxin increased
microalbumin. NAG increased with 100 /yg A-DEP.
BALF Cells: PMNs significantly increased dose-
dependently with DEP and remained elevated at 18h.
Endotoxin induced the greatest increases of PMNs.
Macrophages increased with A-DEP and were
unaffected by DEP.
Cytokines: Endotoxin induced massive responses for
IL-6, MIP-2 and TNF-a but no response from IL-5. A-
DEP increased all 4 cytokines but only at the 100//g
dose level. Similarly, DEP only increased IL-6 at the
100/yg dose level.
July 2009
D-116
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Smith CAPs
et al. (2003, (Fresno, CA)
042107)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Aqe: 11-12wks
Particle Size: <2.5//m
Route: Whole-body Inhalation
Dose/Concentration: 6 exp in 2 sets of 3:
FalM - 847 /yglm3
Fall2 - 260/yg|m3
Fall3 - 369/yg|m3
Winterl - 815/yg/m3
Winter2 - 190/yg/m3
Winter3 - 371 /yglm3
Time to Analysis: 4hfd for 3 consecutive days.
Parameters measured immediately following last
exposure.
Particle Characteristics: Nitrate showed the
highest variability near 10 fold, followed by Si, S and
EC. 0C concentration was relatively consistent.
Metals otherwise appeared proportionate to the
concentrations.
BALF Cells: Total cells increased at wk1. Percent of
macrophages reduced in wk2 with CAPs. Number of
neutrophils increased with CAPs, but only achieved
statistical significance during wk1 of the fall and
winter. Lymphocytes increased but were not
statistically significant.
BAL cell permeability: Upon CAPs exposure, the
proportion of nonviable cells were increased up to
242% when compared to controls. The fall of wk2
induced the highest significant increases followed by
fall wk1, fall wk3, and winter wk3.
Reference: Smith CFA: Coal Fly Ash (400 MW, Route: Nose-only Inhalation
et al. (2006,
110864)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age: 8wks
Weight: 260-270g
Wasatch Plateau, Utah)
(aerodynamic separation)
Particle Size: 0.4-2.5 /jm
Dose/Concentration: 1400 //gfm3 PM2.6
including 600/yg/m3 PM1
Time to Analysis: 4hfd for 3 consecutive days.
Parameters measured 18 or 36h post-exposure.
BALF Cells: Percent and total number of neutrophils
in BALF and blood increased significantly at both 18
and 36h. Percent of macrophages decreased slightly
while number of macrophages increased in
bronchiole-alveolar duct regions at both time periods.
Cytokines: MIP-2 and transferrin increased at 18h.
IL-1 (3 increased at 36h.
Hematology: Plasma protein increased at 18h.
Lymphocyte and hematocrit percentage decreased at
36h.
Other: Gamma glutamyl transferase decreased at
36h. Lung antioxidant increased at 18h.
Reference: Song et DEP collected from a 4JB1 -
al. (2008,156093) type, light-duty (2740 cc), four-
Species: Mouse, cV.linder diesel engine operated
using standard diesel fuel at
Gender: Ffemale, speeds of 1500 rpm under a
Strain: BALB/c, load of 10 torque.
e: 5-6wks
Particle Size: 0.4 fjm (mean
diameter)
Route: Intranasal Instillation (days 1-5), Whole-
body Exposure Chamber (days 6-8)
Dose/Concentration: O.Gmg/mL in 50 /jl of
saline (days 1 -5), 6mg/m3 for 1 h/d for 3d (days
6-8).
Time to Analysis: Enhanced Pause (Penh),
measured on day 9. BAL and lung tissues
collected on day 10.
Airway Hyper-Responsiveness: Intranasal
exposure plus aerosolized DEP caused a significant
increase in methacholine-induced Penh over the
control.
BAL Fluid Analysis: There was no significant
increase in IFN-y in the BAL fluid following DEP
treatment but there was a significant increase in IL-4
levels compared to the control. (IL-4 increase could
indicate that DEP modulates Th-2 cytokines in the
mouse model). DEP also induced an increase in total
neutrophils and lymphocytes in the BAL when
compared to the control. The nitrite concentration in
BAL (indicating NO generation) was significantly
greater in the DEP exposed group than the control.
Histological Assessment: Peribronchial and
perivascular infiltrates were more common in the
group exposed to DEP than the control.
Ym1 and Ym2 Expression: (see explanation in
comments section) Ym1 and Ym2 transcripts were
upregulated in response to DEP exposure in mice.
Reference:
Steerenberg et al.
(2006, 088249)
Species: Rat
Strain: Crl/WKY
Ambient air samples
PMC, PMF:
¦I: Rome, Italy
¦N: Oslo, Norway
¦PL: Lodz, Poland
¦NL: Amsterdam, Netherlands
Measured Components: Li, Be,
B, Na, Mg, A I, K, Ca, Ti, V,Cr,
Mn, Fe, Co, Ni, Cu, Zn, As, Se,
Sr, Mo, Cd, Sn, Sb, Ba, Ce, Nd,
Sm, Au, Hg, TI, Pb, Bi, U, Si,
Endotoxins, CI, NO-, SO4
Particle Size: PMC: 2.35-8.5
fjm; PMF: 0.12-2.35 fim
Route: IT Instillation
Dose/Concentration: 1 and 2.5mgfanimal
Time to Analysis: Parameters measured 24h
post single exposure
Particle Characteristics: Concentrations of metals
were highest in Rome. Amsterdam was noted for
high Mg and V. Lodz was noted for high Pb, Zn, PAH.
More of PMC was composed of Fe, Mn, Al, Cr, Cu.
More of PMF, on the other hand, was composed of
Zn, Pb, Ni, V.
BALF Inflammatory/Injury Markers: CC16
decreased substantially. Crustal material (endotoxin,
Na, CI and metals but not Ti, As, Cd, Zn, V, Ni, Se)
was positively associated with short term CC16.
Albumin increased.
Cytokines: MIP-2 increased dose-dependently. TNF-
a also increased.
BALF Cells: PMNs increased.
July 2009
D-117
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Stinn
et al. (2005,
088307)
Species: Rat
Gender: Male and
Female
Strain: Crl: (WIU
BR
Aqe: 40d
DE
(generated from 1.6 L VW
diesel under USFTP 72)
CO: 10, 37 ppm
CO2: 2170, 6540 ppm
NO: 7.0, 22.8 ppm
NOx: 8.6, 28.3 ppm
SO2: 0.83, 3.09 ppm
NH4: ND
Measured Major Components:
NO, SO2,1 -nitropyrene, Zi.
50% by DE weight is elemental
carbon.
Particle Size: 3mg/m3: 0.19
fjm (MMAD)
Route: Nose-only Inhalation
Dose/Concentration: 3 and 10 mgfm3
10 mg/m3: 0.21 /jm
(MMAD)
Time to Analysis: 6hfd, 7dfwk for 24m; 6 m
postexposure
Body Weight: Mean weight increased substantially
during the first few weeks in all groups. Food
consumption decreased in 1-24mo but was recovered
in 24-30mo. Body weight decreased at 23mo in all
categories, but recovered except in high dose males
at 30mo.
Organ Weight: Absolute weight of lungs, larynx and
trachea increased from 0 to 12 to 24mo and stayed
elevated at 30 mo: Low < Hi, male ~ female.
Pulmonary Parameters: Respiratory frequency,
tidal volume, and minute volume were unaffected in
any group measured between 3 and 24mo.
Elemental carbon increased dose-dependently in
exposure groups. No maleffemale difference was
observed, but increases were greater at 24 mo than
at 18 mo.
BALF Cells: PMNs and lymphocytes showed dose
and time-dependent effects at 18 and 24 mo (no
data at 30 mo). Lymphocytes increased 50 fold in
high dose males at 24mo. Peripheral monocytes and
neutrophils increased 3 fold in DE groups at the end
of the study. Particle-filled macrophages in alveolar
lumen and interstitium increased at 12, 24, 30mo in
both genders at all dose levels.
BALF Inflammatory/Injury Markers: LDH
increased in dose and time-dependent manner.
Hematology: Erythrocytes were unaffected (12, 24,
30) except in high dose females at 24 and 30 mo.
Hemoglobin and hematocrit increased dose-
dependently with no gender differences. Leukocytes
increased in a dose- and time-dependent manner.
Nasal Cavity Histopathology: All effects were
resolved at 30 mo. Nasal cavity hyperplasia
increased at the high dose at 12 and 24mo in both
genders. Squamous metaplasia of respiratory
epithelium increased in high dose females (12,
24mo).
Larynx Histopathology: No effects were observed.
Lung Histopathology: Alveolar region hyperplasia of
alveolar epithelium increased at 12, 24, 30mo in both
genders at all dose levels except for 12mo low dose
males and females. Above lung histopathology was
not time-dependent, though perhaps some small
dose-dependence was observed.The following
histopathology findings showed strong dose- and
time-dependent increases that occurred in both
genders (24-30 mo): goblet cell hyperplasia of
bronchial epithelia, cuboidalfcolumnar hyperplasia of
alveolar epithelium, chronic active inflammation and
septal fibrosis.
Tumorigenicity: Lung tumors were more prevalent
in females than males and appeared to be dose-
dependent. The major 3 types of tumors are the
following:: bronchio-alveolar adenoma, bronchiolo-
alveolar adenoma and benign keratinizing cystic cell
tumors. Enhanced effects in females versus males
may be the result of enhanced metabolism (body
volume versus body weight) and increased repiratory
volumefbw for females.
July 2009
D-118
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Sureshkumar et al.
(2005, 088306)
Species: Mouse
Gender: Male
Strain: Swiss
Age: 10-12wks
Weight: 20-25g
GE: Gasoline Exhaust
(Honda generator EBK 1200,
four stroke one cyl)
Including: SO2 - 0.11 mg/m3
NOx - 0.49 mg/m3
CO - 18.7 ppm
Particle Size: GE
>4/ym - 34.1 %
3-4/ym - 15.8%
2-3/ym - 15.8%
1.5-2 /jm - 10.6%
0.5-1.5//m - 5.3%
< 0.5 fjm - 18.4 %
Route: Nose-only Inhalation
Dose/Concentration: 0.635 mgfm3
Time to Analysis: 15min/d 7,14 or 21 d.
Parameters measured
less than 1h post-exposure
Cytokines: GE caused time-dependent increases in
TNF-a and IL-6. IL-10 and IL-1 (3 were unaffected.
BALF Inflammatory/Injury Markers: y-GGT, ALP
and LDH increased after 2wks of GE exposure and
stayed stable at 21 d. Total protein slightly increased
on 14 and 21 d, though these increases were not
statistically significant.
BALF Cells: Neutrophils (%) increased at 7,14 and
21 d (stable). Total cell count, macrophages and
eosinophils were unaffected. Leukocytes and
lymphocytes increased, though not significantly.
Histopathology: Minor changes at 7d, mild edema in
alveolar region at 14d and sloughing of epithelial
cells in bronchiolar region and focal accumulation of
inflammatory cells in alveolar region at 21 d were
observed in a time-dependent manner.
Reference:
Tesfaigzi et al.
(2002, 025575)
Species: Rat
Gender: NR
Strain: Brown-
Norway
Age: 7-8wks
Weight: 310-330g
WS (wood stove- Vogelzang
Boxwood Stove, Model BX-
42E, wood- Pinus edulis) (CO,
NO, N0>, total hydrocarbon)
Particle Size: Smaller size
fraction: 0.405-0.496/jm,
larger size fraction: 6.7-11.7
fj m
Route: Whole-body Exposure
Dose/Concentration: Target concentration
(low, high exposure): 1,10mg/m3; CO-15-
106.4ppm, NO- 2.2-18.9ppm, NOx- 2.4-19.7ppm,
total hydrocarbon- 3.5-13.8ppm
Time to Analysis: Exposed 3hfd, 5dfwk, 4 or
12wks.
Respiratory: Total pulmonary resistance increased
for exposure groups and was significant for the low-
exposure group. In exposed groups, forced expiratory
flows and quasistatic compliance were lower and
dynamic lung compliance higher, the latter being
significant for the high-exposure group. For the high-
exposure group, vital capacity slightly decreased,
residual volume slightly increased, and CO-diffusing
capacity had a slight, significant decrease.
Histopathology: WS caused minimal to mild chronic
inflammation in the epiglottis of the larynx. PAS-
positive cells increased in the 30d high-exposure
group. AMs increased with time and concentration.
Particle-laden macrophages were seen after 90d. AB-
and PAS-positive epithelial cells increased for the
90d low-exposure group.
BALF: Macrophages decreased significantly in the
high-exposure group. Particle-laden macrophages
increased with concentration. Lymphocytes and
neutrophils slightly increased in the high-exposure
group.
Cytokines: LDH increased slightly and protein levels
decreased slightly in the high-exposure group.
Cytokines were below detectable levels.
Reference: Tin-Tin-
Win-Shwe et al.
(2006,088415)
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 7wks
CB14: Printex 90 (Degussa)
CB90: Flammruss 101
(Degussa)
Particle Size: CB14:14nm
CB95: 95 nm
Route: IT Instillation
Dose/Concentration: 25,125, 625 //gfmouse;
approx. 1, 5, 25mg/kg
Time to Analysis: 1fwk for 4wks. mRNA
expression in lungs and mediastinal lymph nodes
measured 4h post exposure
Body weight, thymus, spleens, splenic cell
count: No effects were observed.
BALF Cells: Increased total cell numbers were
observed for 125, 625/yg CB14 (dose-dependent)
and 625/yg CB95. Total cell count was twice as high
for CB14 at 125 and 625 //g compared to CB95. AM
numbers exhibited a dose-dependent response for
both CB14 and CB95 for all doses except 125//g.
Lymphocyte numbers increased at 125 and 625/yg
for CB14 and 625/yg for CB95. PMN numbers
increased at 125 and 625/yg for CB14 and CB95,
but the response was greater with CB14. PMN
numbers were proportional to dose surface area for
both PM sizes.
BALF Cytokines: CB14 and CB95 induced dose-
dependent increases in IL-1 (3. TNF-a increased at
125 and 625/yg dose in CB14 with the 125 dose
inducing a slightly greater increase. CB14 and CB95
induced CCL-3 increases 125 and 625/yg.
Chemokine mRNA in lung and lymph nodes: CCL-
3 mRNA increased for CB14 but not CB95 4h
following the last exposure. CCL-2 was unchanged.
Mediastinal lymph nodes: The number of CB-laden
phagocytes increased in a dose-dependent manner
for CB14 and CB95. CB14 had higher numbers at all
doses compared to CB95.
July 2009
D-119
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Tong et
al. (2006, 097699)
Species: Mouse
Gender: Male
Strain: KP600 CD-
1
Weight: 22-26g
PM2.E (collected from stacked
filter air sampler in Shanghai,
China)
Fe: FeS04
Zn: ZnS04
PMF: PM2.6 + FeS04
PMFZ: PM2.B + FeS04 +
ZnS04
Major Measured Components:
Fe 26ppm, Zn 9ppm, S 61 ppm
Particle Size: PM2.6
Route: IT Instillation
Dose/Concentration: PM: 25mgfmL,
1mg|mouse
Fe: 15mg/mL, 0.6 mg/mouse
Zn: 15mg/mL, 0.6 mg/mouse
PMF: PM 25mg/mL + Fe 15mg/mL,
1.6mg|mouse
PMFZ: PM 25 mglmL + Fe15mg|mL,
1.6mg|mouse
Time to Analysis: Instilled twice at 0 and 24h.
Parameters measured 24h following last
exposure (at 48h).
Synchrotron X-ray imaging (in vivo): PMFZ
showed the greatest increase in alveolar changes. Fe
induced more hemorrhagic changes, whereas Zn
induced more nonuniformity of lung texture. This
suggests that Zn induces PBMC in a dose-dependent
manner which releases IL-1, IL-6, TNF-a, and IFN-y.
Histopathology: PMFZ induced the most severe
changes including serious inflammation/pus in
bronchia and bronchial epidermal cell hyperplasia. For
Fe or PMF hemorrhagic changes predominated but
were less severe than PMFZ.
Reference:
Upadhyay et al.
(2008,159345)
Species: Rat
Gender: Male
Strain: SHR
Age: 6m
Weight: NR
Ultrafine Carbon Particles
(UFCP)
Particle Size: Size-
31 ±0.3nm, MMAD- 46nm,
Surface area concentration-
0.139 m2(particle)/m3(air),
Mass specific surface area-
807m2/g
Route: Whole-body Exposure
Dose/Concentration: 172//gfm:l
Time to Analysis: Acclimatized 2d. 1 d baseline.
24h exposure. 4d recovery. Sacrificed 1st or 3,d
day of recovery.
Cardiophysiology: The mean arterial BP and HR
increased but returned to baseline levels by the 4th
recovery day. SDNN and HRV decreased. RMSSD
and LF/HF decreased but were not significant.
Pulmonary Inflammation: UFCP did not cause
pulmonary inflammation.
Pulmonary and Cardiac Tissue: HO-1, ET-1, ETA,
ETB, TF, PAI-1 significantly increased in the lung on
the 3,d recovery day. HO-1 was repressed in the
heart, but the other markers had slight,
nonsignificant increases.
Systemic Responses: Neutrophil and lymphocyte
cell differentials significantly increased on the 1st
recovery day. Other blood parameters were
unaffected. The plasma renin concentration
increased on the first 2 recovery days. Ang I and II
concentrations increased on the 1st recovery day but
was not significant.
Reference:
Wallenborn et al.
(2007,156144)
Species: Rat
Gender: Male
Strain: WKY,
SHRWKY, and
stroke-prone SH
(SHRSP)
Age: 12-15wks
PM: precipator unit power
plant residual oil combustion
Particle Size: PM: 3.76/ym
(bulk) ± 2.15
Route: IT Instillation
Dose/Concentration: WKY vs SHRSP: 1.11,
3.33, 8.33 mg/kg
SH vs SHRSP: 3.33, 8.33 mg/kg
Time to Analysis: WKY vs SHRSP: single
exposure, parameters measured 24h post-
exposure.
SH vs SHRSP: single exposure, parameters
measured 24h post-exposure.
Note: 4h post exposure study done on WKY vs
SHRSP but not published.
BALF Cells: A dose-dependent increase in total cells
and neutrophils was observed. Equal response for all
3 strains except for SH, for both concentrations was
observed.
BALF inflammation/Injury Markers: LDH exhibited
a dose-dependent increase in equal response for all 3
strains. WKY had higher baseline levels of NAG
activity but, upon PM exposure, SHRSP induced
higher increases than WKY. GGT exhibited a dose-
dependent response for all 3 strains. SHRSP showed
the highest increase followed by WKY and SH.
Protein levels increased at the high dose level with
SHRSP exhibiting the highest increases followed by
SH and WKY. Albumin levels were inconsistent
between experiments.
Oxidative Stress - Lung: (WKY vs SHRSP only):
SOD decreased following increased exposure levels
with SHRSP levels generally higher than WKY.
Ferritin levels declined only in SHRSP.
Oxidative Stress - Cardiac: SOD increased in the
SHRSP vs WKY experiment only. Only SHRSP at
8.33 mg/kg showed a significant increase when
compared to the control.
GPx: No action but SHRSP levels were similar to
SHR and, in the WKY vs SHRSP experiment, SHRSP
exhibited higher activity level than WKY.
Ferritin: Equivocal results were observed. Levels
decreased at the high dose for WKY and SHRSP but
increased at medium doses for SH and SHRSP.
ICDH: Levels increased for WKY and decreased for
SHRSP.
July 2009
D-120
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Wallenborn et al.
(2008,191171)
Species: Rat
Gender: Male
Strain: Wistar
Kyoto (WKY)
Age: 13wks
Weight: NR
Zinc Sulfate (ZnS04,
aerosolized)
Particle Size: NR
Route: Nose-only Inhalation
Dose/Concentration: 9.0 ±2.1 fjg zinc/m3,
35 ±8.1 /yg zinc/m3,123.2 ±29.6 /yg zinc/m3
Time to Analysis: Exposed 5h/d, 3d/wk,
16wks. Half of the rats used for plasma/serum
analysis, other half for isolation of cardiac
mitochondria.
A trend toward increased BALF protein was seen.
Cardiac mitochondrial ferritin had a small, significant
increase. Mitochondrial succinate dehydrogenase and
glutathione peroxidase had small, significant
decreases. Subchronic exposure to 100/yg/m3
caused expression changes of cardiac genes involved
with cell signaling events, ion channels regulation,
and coagulation. No pulmonary-related effects were
seen.
Reference:	Ambient PM2.6 PM10
Wegesser and Last Collected from San Joaquin
(2008,190506) Valley, CA
Species: Mouse Particle Size: PM10 2 E
Gender: Male
Strain: BALB/c
Aqe: 8-10wks
Route: IT Instillation
Dose/Concentration: 25-50 /yg/mouse
Time to Analysis: 3, 6,18, 24, 48, 72h post IT
instillation.
BALF Cells: Increased amount of viable cells found
in PM-exposed mice with dose-response relationship
between dose of PM and number of total cells
recovered in BALF. At 6h, increased numbers of
macrophages at both 25 and 50/yg/mouse. Increased
percentage of neutrophils observed with 50
/yg/mouse PM only. Furthermore, both macrophages
and neutrophils increased with longer time period
from instillation, peaking at 24 h.
MIP-2: At 50/yg/mouse, MIP-2 concentrations
increased, peaking at 3h, though not statistically
significant and returned to basal levels by 6h.
Positive correlation observed between MIP-2
concentration and increased neutrophil counts. No
correlation found between MIP-2 and macrophages.
Reference:
Whitekus et al.
(2002,157142)
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 6-8wks
Weight: NR
DEP (light-duty, four-cylinder
engine- 4JB1 type, Isuzu
Automobile, Japan; standard
diesel fuel) (extracts)
Particle Size: 0.5-4/ym
Route: Inhalation
Dose/Concentration: 200, 600, 2000/yg/m3
Time to Analysis: Exposed 1 h/d 10d. Animals
receiving OVA had 20min OVA exposure after
DEP exposure.
DEP+0VA dose-dependently increased IgE and IgG 1,
being more effective than the OVA-alone treatment.
This effect was significantly suppressed by thiol
antioxidants NAC or BUC. DEP + 0VA increased
carbonyl protein and lipid peroxide over OVA. NAC or
BUC suppressed lipid peroxide and protein oxidation.
No general markers for inflammation were observed.
Reference:
Wichers et al.
(2004, 055636)
Species: Rat
Gender: Male
Strain: SH
Age: 75d
PM (HP-12): inside wall of
stack of Boston, MA power
plant burning # 6 oil.
Particle Size: PM: 3.76 /ym :
2.15
Route: IT Instillation
Dose/Concentration: 0.83, 3.33 or 8.33 mg/kg
Time to Analysis: single, 6h for Whole-body
plethysmographs (WBP) and repeated daily for 4-
7d,
96 or 192h post exposure
non-WBP animals: single,
24, 96,192h post exposure
Tidal Volume: A dose-dependent decrease in tidal
volume (45 % at high dose) was sustained for 1d
with very slow recovery over 7d.
Breathing frequency: Dose-dependent increase
(100 % at high dose) with recovery at 7d was
observed.
Minute ventilation: Small dose-dependent increases
were observed with a return to normal ventilation in
2d.
Penh (enhanced pause): Equivocal results in all
groups were observed (due to major control
variation).
BALF Cells: Dose-dependent increases in total cells
at 24 h, with declined, but still elevated, levels at
192h. Neutrophils increased significantly (10 fold) at
24 h in the mid and high dose groupsand showed
declined, but still elevated, levels at 192h.
Macrophages slowly increased in a dose-dependent
manner at 192h.
BALF Inflammatory/Injury Markers: Protein and
albumin increased at 24 h, returned to relative basal
level at 192h at the mid and high dose levels. NAG
exhibited dose-dependent increases at 24h and
sustained these levels through 192h.
July 2009
D-121
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Wichers LB et al.
(2006,103806)
Species: Rat
Gender: Male
Strain: SH
Age: 71 -73d
Weight: 255-278g
PM (HP-12): inside wall of
stack of Boston, MA power
plant burning # 6 oil.
Particle Size: 1.95/ym ±
3.49
Route: Whole-body Inhalation
Dose/Concentration: 13 mgfm3
Time to Analysis: Phase I: 1 st day, filtered air,
2nd day, 6h of PM
Phase II: 1 st Da, y filtered air, 4 days of 6h PM
each
Immediate post-exposure
Body/ Lung weight: No effects on Phase I rats were
observed. HP-12 exposure increased body weight,
left lung, right intercostal, and right diaphragmatic
lobes in Phase II rats. However, results appeared due
to normal growth apptern in juvenile rats over 4d.
Lung lobe to Body Weight Ratio: No effects at 1
or 4d were observed.
Deposition calculations: V and Co were used to
estimate deposition rates (good correlation between
two metals at R2 - 0.94). Total HP-12 deposition
using Co was 26 and 99/yg (for 1 day and 4 day
experiments) and using V was 31 and 116/yg.
Modeling information estimated HP-12 deposition at
43% in conducting airways and 57% in alveolar
region.
Breathing parameters: No changes were observed
for 1 or 4d studies except for a possible decrease in
frequency for the 1d study.
Reference: Witten
et al. (2005,
087485)
Species: Rat
Gender: Female
Strain: F344
Age: 8wks
Weight: — 175g
DEP (heavy-duty Cummins N14
research engine operated at
75% throttle)
Particle Size: 7.234-
294.27nm
Route: Nose-only Inhalation
Dose/Concentration: Low- 35.3±4.9//g/m3,
High- 632.9±47.61 //g/m3
Time to Analysis: Exposed 4hfd, 5dfwk, 3wks.
Pretreated with saline or capsaicin.
There were no differences for substance P. The low-
exposure group had significantly less NK1. DEP
reduced NEP activity. Plasma extraversion dose-
dependently increased and was greatest in capsaicin
animals. Respiratory permeability dose-dependently
increased. IL-1 (3 was significantly higher for the low-
exposure group. IL-12 was significantly lower in the
capsaicin high-expsoure group. TNF-a increased in
the high-exposure group and capsaicin low-exposure
group. High exposure induced particle-laden AMs in
the lungs, perivascular cuffing consisting of
mononuclear cells, alveolar edema and increased
mast cell number. Neutrophil and eosinophil influx
was not seen.
Reference: Wong
et al. (2003,
097707)
Species: Rat
Gender: Female
Strain: F344/NH
Age: ~4wks
Weight: — 175g
DEP (Cummins N14 research
engine at 75% throttle) (EC-
34.93-601.67 //g/m3, 0C-
1.90-11.25 /yglm3, Sulfates
0.94-17.96/yg|m3,Na- 4.07-
4.78 ng/m3, Mg- 0.60-0.86
ng/m3, Ca- 5.05-10.66 ng/m3,
Fe- 3.17-6.44, Cr- 0.68-1.31
ng/m3, Mn- 0.11-0.22 ng/m3,
Pb- 0.97-1.24 ng/m3)
Particle Size: 7.5-294.3nm
Route: Nose-only Inhalation
Dose/Concentration: Low- 35.3±4.9/yg/m3,
High- 669.3±47.6/yg/m3
Time to Analysis: Exposed 4h/d, 5d/wk, 3wks.
Pretreated with saline or capsaicin.
DEP dose-dependently increased plasma
extraversion, which was further increased by
capsaicin. In the high-exposure group, particle-laden
AMs (which were reduced by capsaicin),
inflammatory cell margination, perivascular cuffing
with subsequent mononuclear cell migration and
dispersal, increased mast cells, and decreased
substance P were all seen. NK-1R was
downregulated in the low-exposure group and
upregulated in the capsaicin-pretreated high-exposure
group. NEP decreased significantly for both groups.
Reference: Wu, W. Zn2 -t
Wang, X.
Zhang, W.
2003
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Aqe: 60d
Particle Size: NA
Route: IT Instillation
Dose/Concentration: 50 /ym/rat
Time to Analysis: Single, 24 h
Cells: Decreased number of airway epithelial cells
shown with PTEN protein immunostaining.
Macrophages were unaffected.
July 2009
D-122
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Route:
Yamamoto et al.
(2006,096671)
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 7wks
Weight: 23g
CB14: Printex 90 (Degussa)
CB95: Flammruss 101
(Degussa)
LTA: Lipoteichoic acid
14CL: CB14 + LTA
95CL: CB95 + LTA
CB14 measured Components: C
96.79%, HR 0.19%, NO.13%,
S 0.11%, Ash 0.05%, 0 2.74%
CB95 measured Components: C
97.98%, HR 0.15%, N 0.28%,
S 0.46%, Ash 0%, 0 1.14%
Particle Size: CB14: 14nm
CB95: 90 nm
Route: IT Instillation
Dose/Concentration: CB14: 0, 25,125, 625
//gfmouse
CB95: 0, 25,125, 625/yg/mouse
LTA: 10 or 50 /yg/mouse
14CL: 125 /yg CB14 + 10 or 50 //g LTA
95CL: 125/yg CB95 + 10 or 50/yg LTA
Time to Analysis: Single, 4 and 24h
BALF Cells: CB95 induced dose-dependent increases
of PMN. CB14 induced an increase in PMNs but the
increases were not dose-dependent. LTA massively
increased PMN. LTA induced dose-dependent
increases in total cells, especially at high dose at
24 h. LTA had massive synergistic effect with CB14
and CB95 for total cells and PMNs. Total cell count
and PMN levels were highest in 14CL with levels at
24h higher than at 4h. Macrophage data were
inconsistent.
Cytokines: CB95 induced dose-dependent increases
in IL-6, TNF-a, CCL2 and CCL3. CB14 induced dose-
dependent increase in CCL2 and CCL3. Exposure
induced increases of IL-6 at the high dose only. Slight
effect on TNF-a was observed. LTA induced dose-
dependent increases of IL-6, TNF-a and CCL3. 14CL
massively induced IL-6 and CCL2. No combination of
CB and LTA affected TNF-a or CCL3.
mRNA Expression: LTA, 14CL and 95CL increased
TLR2 mRNA expression with 95CL and 14CL
inducing higher increases than LTA. No effect on
TLR4 mRNA expression was observed.
Reference:
Yanagisawa et al.
(2003, 087487)
Species: Mouse
Gender: Male
Strain: ICR
Age: 6wks
Weight: 29-33g
DEP:
(4JB1 light duty 4cyc 2, 74
liter Isuzu engine)
LPS
DEP-OC: organic compounds
DL: DEP + LPS
DOL: DEP-OC + LPS
Particle Size: 0.4 /jm
Route: IT Instillation
Dose/Concentration: DEP/DEP-OC: 125
//gfmouse
LPS: 75 /yg/mouse
Time to Analysis: Single, 24 h
BALF Cells: DEP and DEP-OC increased neutrophils
but the increases were not statistically significant.
LPS increased neutrophils significantly. DL and DOL
massively increased neutrophils at greater levels than
LPS alone. Macrophages were unaffected.
Pulmonary Edema: LPS, DEP and DEP-OC increased
edema. DL further increased this effect. DOL had no
effect compared to LPS alone.
Histology: DL elevated neutrophil inflammation
interstitial edema and alveolar hemorrhages. DOL
induced neutrophilic inflammation without the
alveolar hemorrhages.
Cytokines: LPS increased IL-1 (3, MIP-1a, MCP-1
and KC. DEP and DEP-OC had no effect. DL induced
further increases. DOL decreased cytokines
compared to LPS alone. DEP-OC increased IL-1 (3 and
MIP-1a mRNA expression slightly. DEP had no
effect. LPS significantly increased IL-1 (3 and MIP-1a
mRNA expression. DL increased expressions while
DOL did not.
mRNA Expression of TLRs: DEP-OC, DL, DOL and
LPS increased TLR2. DEP had no effect. All particles
increased TLR4 mRNA expression.
July 2009
D-123
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Yokohira et al.
(2007,097976)
Species: Rat
Gender: Male
Strain: F344/
DuCrj
Aqe: 10wks
DQ-12: Quartz dust (Douche
Montan)
HT: Hydrotalcite (Kyoward
500, PL-1686, KYOWA)
POF: Potassium Octatitanate
fiber (TISMO, Otsuka)
PdO: Palladium Oxide
CB: Carbon Black (Mitsubishi
Kasei)
Particle Size: DQ12 < 7/ym
HT: 7.8 ±1.5^m
POF: <50um length: <2//m
width
PdO: 0.54 ±1.11 //m
CB: 28nm
Route: IT Instillation
Dose/Concentration: 4mg/rat in 0.2ml saline
Time to Analysis: Single, 1 and 28d
Lung weight/body weight ratio: DQ-12, HT and
POF induced increases after 1 d. After 28 days, all
samples induced increases in lung weight.
BALF Cells: Neutrophils increased significantly in
walls and alveolar spaces in all groups on 1d except
at HT. AT 28d, this increase was maintained only in
walls with severe and moderate elevations, except
for DQ-12.
Histopathology: DQ-12 caused pulmonary edema
both at 1 and 28d. PdO and CB induced edema at
28d. Fibrosis was observed after 28d with the most
significant increase, in decreasing order, induced by
DQ-12,PdO, POF, HT, CB, and the control. Histiocyte
infiltration was observed after 1 d for DQ-12, POF
and PdO. At 28d, infiltration was observed for DQ-
12, HT, POF and PdO. Restructuring of alveolar walls
and microgranulation was observed for all 5 particles
but only at 28d with DQ 12, PdO, HT, POF, CB and
control.
Immunohistochemistry: BrdU: At 1 d all 5 particles
elevated in both area and number. Activity declined
after 28d but was still higher than the control.
iNOS: At 1d DQ-12, POF and PdO induced increases.
At 28d, DQ-12 and HT induced increases.
MMP-3: DQ-12 induced increases at both 1 and 28d
and PdO at 28d.
Toxicity scoring: The levels of toxicity are, in
decreasing order, as follows: DQ-12, HT/PdO/POF,
and CB.
Reference: Zhao et
al. (2006,100996)
Species: Rat
Strain: Sprague-
Dawley
DEP: SRM 2975
DEPE: SRM 1975
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 35mgfkg bw
Time to Analysis: AG (amino guanidine) group
pre-treated with 100 mg/kg bw.
Single, 1d.
AG group coexposed 30 pre and 3, 6, 9h post
DEP/DEPE
iNOS expression in AMs: Both DEP and DEPE
increased 12 and 6 fold respectively. NO and
peroxynitrite levels increased accordingly. AG had no
effect on iNOS expression but AG attenuated NO for
both DEP and DEPE but peroxynitrite only for DEPE.
DEP induced much higher levels of oxidants than
DEPE. Unlike DEPE, DEP was unaffected by AG.
Role of iNOS in Lung Injury: DEP and DEPE
induced inflammation (PMN), cellular toxicity (LDH)
and lung injury (protein). AG significantly attenuated
the DEPE response but no effect was observed on
the DEP responses.
Cytokines: IL-12 levels were induced by both DEPE
andDEP, with DEPE inducing higher increases than
DEP, and both were significantly attenuated by AG.
DEP and DEPE induced similar increases in IL-10
levels. AG increased DEP effect 3 fold and attenu-
ated DEPE to control.
CYP enzymes: DEP and DEPE induced increases in
CYP1A1 level and activity.AG attenuated CYP1A1
activity for both DEP and DEPE. CYP2B1 level and
activity were slightly decreased by DEP and DEPE.
AG had no effect.
Cytosol phase II enzymes: DEPE had no effect; AG
treatment increased catalase activity. DEP reduced
catalase and GST activities. AG had no effect.
Neither DEP, DEPE nor AG affected QR quinone
reductase.
July 2009
D-124
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Zhou et UFe: Ultrafine Fe particles
al. (2003, 087940) Partic|e Size: 72nm
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age: 10-12wks
Route: Whole-body Inhalation
Dose/Concentration: 57 or 90//gfm3
Time to Analysis: 6hfd for 3d, parameters
measured within 2h post exposure.
BALF Inflammatory/Injury Markers: At the high
dose, total protein increased. No significant changes
were observed in LDH.
BALF Cells: No significant changes observed in total
cell number, cell viability or cell differentials.
Intracellular ferritin: The high dose induced
increases. No significant differences were observed
between the low dose and control.
Oxidative stress: Antioxidant level by FRAP value
decreased at the high dose. GST (glutathiones-
tranferase) activity increased at the high dose. No
effect on intracellular GSH and GSSG (glutathione
disulfide) was observed.
Cytokines: Only at thehigh dose was an increase in
IL-1 p observed. No effect on TNF-alph or NFkB-DNA
binding activity was observed.
Table D-4. Effects related to immunity and allergy.
Study
Pollutant
Exposure
Effects
Reference:
Apicella C et al.
(2006, 096586)
Species: Mouse
Strain: BALB/c
Cell Line: 112D5
hybridoma
Primary
Macrophages:
Peritoneal
Poly OVA (Ovalbumin on polystyrene
beads) Soluble OVA
Particle Size: NR
Route: Cell Culture
Dose/Concentration:
PolyOVA and Soluble OVA: 0.2,1.0 or 5.0//gfmL
Time to Analysis: 48 h
IL-6: Stimulation with PolyOVA higher than
stimulation with soluble OVA
TNFa: Stimulation with PolyOVA higher than
stimulation with soluble OVA.
IL-10: No modifications in levels after
PolyOVA or soluble OVA stimulation.
Viability of Peritoneal macrophages:
Stimulation with PolyOVA led to 33%
decrease in viability. Stimulation with soluble
OVA led to 24% in viability.
Effects of PolyOVA stimulated
macrophages: Culture supernatants from
PolyOVA stimulated macrophages had a
percentage increase of asymmetric IgG;
however, the addition of rmlL-6 at identical
concentrations did not induce a significant
increase. It also decreased the proliferation of
112D5 hybridoma.
Reference:
Arantes-Costa et
al. (2008,
187137)
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 6wks
Weight: NR
ROFA (solid waste incinerator powered
by combustible oil; Sao Paulo, Brazil)
Particle Size: NR
Route: Intranasal^ Instilled
Dose/Concentration: 60 fjg ROFA in 50 fjL
saline
Time to Analysis: OVA sensitized days 1 and
14. OVA-challenged days 22, 24, 26, and 28.
ROFA exposed 1-3h after OVA challenge or
saline. Pulmonary responsiveness measured day
30 then sacrificed. Lungs removed, fixed for 48h.
ROFA increased pulmonary responsiveness
and decreased ciliated cells in nonsensitized
mice, which were both further amplified in the
presence of OVA. ROFA did not affect
eosinophils, macrophages, chronic
inflammation, or neutral or acidic mucus.
July 2009
0-125
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Araujo et al.
(2008,156222)
Species: Mouse
Gender: Male
Strain: ApoE
C57BL/CJ
Age: 6wks
Weight: NR
CAPs (Los Angeles, California freeway;
Nov-Dec 2005)
Particle Size: Fine particles (FP):
< 2.5 /jm; UFP: <0.18 /jm
Route: Whole-body Exposure
Dose/Concentration: FP: ~440/yg/m3, UFP:
— 110 /jm; PM number concentration: FP: 4.56
X 106 particles/cm3, UFP: 5.59 X106
particles/cm3
Time to Analysis: Exposed 5hfd, 3dfwk for total
of 75h. Killed 24 to 48h postexposure.
Composition: UFP had a higher particle
number, surface area, PAH content and
fractional carbon composition than FP.
Atherosclerosis: UFP had significantly
greater aortic atherosclerotic lesions than FP
and the control. The lesions were comprised
of macrophage infiltration with intracellular
lipid accumulation. Plaques were thicker and
more extensive in the UFP group.
HDL: FP had increased plasma total
cholesterol. Plasma HDL from both groups had
decreased protective effects. The anti-
inflammatory effect was lower in the UFP
group.
Oxidative Stress: Lipid peroxidation
increased in the FP group. The UFP group had
increases in hepatic malondialdehyde, Nef2,
catalase, glutathione S-transferase Ya,
NAD(P-quinone oxidoreductase and superoxide
dismutase 2.
Reference:
Archer et al.
(2004, 088097)
Species: Mouse
Strain: BALB/c
D011.10+/ +
transgenic - ova
specific receptor
for OVA peptide
323-339
Age: 4 wks
PM - SRM 1648 (NIST)
Titanium dioxide (Ti02) as a control
particle
Particle Size:
SRM1648: avg 1.4 ym
Ti02: avg 0.3 /yg (sic)
Route: Intranasal instillation
Dose/Concentration: 500/yg PM/30//I sterile
saline (ultrasonic suspension) initial 0-750/yg
range finding
Time to Analysis: Ova challenge at 68h, Meth-
acholine aerosolization/AR at 72h
Airway responsiveness (WBP): AR induced
by OvafMch challenge was significantly and
dose-dependently increased at doses of
SRM1648 £500ug . Ti02/0va exposure was
not significantly different from saline. PM
associated endotoxin did not contribute to en-
hanced AR.
Lung inflammation/pathology: No increases
in BAL macrophages or eosinophils and no
histological alterations after PM exposure.
Both Ti02 and PM increased pulmonary neu-
trophils, indicating particles alone were
responsible for this increase and that the
inflammatory response could occur
independently of AR.
Reference:
Barrett et al.
(2006,155677)
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 8-10wks
n: Groups of 15-
16 incl. controls
HWS (black/white oak)
CO
Total Vapor Hydrocarbon (TVH)
Particle Size: 0.25 ± 3.3, 0.35 :
2.5, 0.35 ± 2.0, 0.36 ± 2.1 /m
(MMAD±GSD)
Route: Whole-body Inhalation
Dose/Concentration: HWS: 30,100 300,
1000/yg/m3
CO: 0.7,1.6, 4.0,13ppm
TVH: 0.3, 0.6,1.3, 3.1ppm
Time to Analysis: Pretreatment: i|) 10 //g OVA
and 2mg aluminum hydroxide
post-OVA. OVA aerosol challenge on day 14,
followed by 3d of HWS.
Pre-OVA received aerosol OVA challenge on day
14, then 3d of HWS on days 26-28 and an
immediate (second) OVA challengeHWS 6 h/d for
3d. Sacrificed 18h post-exposure.
Allergic Inflammation: A statistically
significant increase in eosinophils was
observed at 300 /yglm3 HWS following OVA
challenge as compared to OVA alone. No
changes in macrophages, neurophils and
lymphocytes were observed. Post-OVA HWS
did not significantly alter BAL cytokine or
serum antibody levels, but linear trend
analyses indicated decreases in IL-2, IL-4, and
IFNn in the absence of OVA, as well as a
statistically significant upward trend in OVA-
specific IgE when HWS exposure followed
OVA challenge. HWS exposure pre-OVA (prior
to second OVA challenge) resulted in a
decrease in IL-13 (statistically significant at
the high dose but no evidence of an exposure-
dependent response), an increase in OVA lgG1
(trend significant) and no change in IL-2, IL-4,
IL-5, IFNn, OVA IgE, total IgE or OVA lgG2a.
Reference:
Burchiel et al.
(2005, 088090)
Species: Mouse
Gender: Female
Strain: A/J
Age: 12-14wks
HWS (black/white oak)
HWS particle Mass
BC
Organic Carbon (OC)
CO
Total Vapor Hydrocarbon
29 other minor components PAH and
metals
Particle Size: 0.3±3, 0.4 ±2, 0.4±2,
0.4 ±2fjm (MMAD±GSD)
Route: Inhalation Chambers
Dose/Concentration: HWS: 30,100, 300,
1000/yg/m3
BC: 3,12, 25, 43//gfm3
0C:40,107, 281, 908//g/m3
CO: 1, 2, 4,13 ppm
TVH: ND, 1,1,3 ppm
Time to Analysis: 6 h/d for 6m
Proliferative responses: HWS increased
splenic T cell proliferation at 100 /ygfm3 with
a dose dependent decrease at 300 and
1000 /yglm3 exposures (p < 0.05) HWS
exposure did not affect T (CD3), helper T cell
(Th, CD4), cytotoxic T cell (CTL, CD8),
macrophage (Mac-1), natural killer cell (NK,
CD16) cell markers or B cell proliferative
response to LPS.
July 2009
D-126
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Burchiel et al.
(2004, 055557)
Species: Mouse
Strain: AJ
Age: 10-12wks
DE generated alternatively from two
2000 Cummins ISB Turbo Diesel 5.9 L
engines using no 2 (chevron) oil and
15w/40 oil (Rotella T, Shell) run ac-
cording to USEPA Dynamometer
Schedule for Heavy-Duty Diesel
Engines
18 PAHs quantified at exposure levels
(text mentions 65)
Particle Size: NR
Route: Inhalation Chambers
Dose/Concentration: 30,100, 300,1000
mg/m3 diesel PM
Time to Analysis: 6 h/d, 7 dfwk for 6m
Proliferative responses: DE (
splenic T cell proliferation at all exposure
levels but was not dose-dependent and most
pronounced at the 30 //g|m3 level, (p < 0.05
at all levels) Splenic B cell proliferation was
increased at the 30 //gfm3 level, but not at the
other exposure levels. Little, if any, PAH was
found in DE, and the majority of PAH tested in
vitro enhanced T cell proliferation (below), so
PAH is likely not responsible for the
immunosuppressive effect of DE on murine
spleen cell responses.
Reference:
Burchiel et al.
(2004, 055557)
Species: Mouse
Strain: AJ
Age: 10-12
weeks
Cell Type: spleen
cells
Use: In vitro
Benzo(ayrene (BaP)
Benzo[a]pyrene-r-7,t-8-dihydrodiol-t-
9,10 epoxide(±) ((antiPDE)
Benzo[a]pyrenetrans-7,8'dihydrodiol
(±)(BP-7,8-diol)
Benzo[a]pyrene-1,6-dione (1,6-BPQ)
Benzo[a]pyrene-3,6-dione (3,6-BPQ)
Benzo[a]pyrene-6,12-dione (6,12 BPQ)
Particle Size:NR
Route: Cell Culture
Dose/Concentration: 1x10 6 cellsfmL in
100//I aliquots
0.01, 0.1 and 1 /jm
Time to Analysis: 72h
BaP at the highest concentration was found
to double splenic T cell proliferation. The
BPQs all also increased T cell proliferation at
much lower concentrations but not in a dose
dependent manner.
Splenic B cell was increased by Bp-(7,8)-diol,
and inhibited by BPDE and 3,6 BPQ but only at
the highest level. Authors concluded that due
to low level of PAH in DE and absence of
BPQs these compounds are not responsible for
immunosuppressive effects of DE.
Reference: Chan
(2006,090193)
Species:
Mouse
Strain: D011.10,
BALB/c,/lfr/2-/-
Cell Types:
Primary bone
marrow dendritic
cells and dendritic
cell line (BCD, T
cells (BMDC)
DEP: DE particles
DEP methanol extract:
Particle Size: PM > 50 as
characterized by Singh et al. 2004
Route: Cell Culture
Dose/Concentration: DEPs 10//gf'niL
LPS 5ng/mL
Time to Analysis: 24h, stored at -20°C until
analyzed
Dendritic cell maturation: Organic DEP
chemicals interfered in the expression of
several DC maturation markers. Both DEP and
DEP extracts were found to inhibit CD86
expression and IL-12 production in LPS-
exposed DCs, and intact particles were not as
effective as DEP extract. DEP extract
treatment of BC1 cells reduced their ability to
stimulate co-cultured antigen-specific T cells,
leading to decreased IFN- y and increased IL-
10 without affecting IL-4 or IL-13. DEP
extract also induced oxidative stress and
interfered with DC activation by several other
Toll-like receptor agonists as well as the NF-
kB cascade. Inhibition of IL-12 production by
DEP extract was shown to be mediated by
pro-oxidative chemicals that engage the Nrf2
pathway. Taken together the inhibition of both
IL-12 and IFN- y indicates a suppression of the
Th1 pathway and provides a novel explanation
for the adjuvant effect of DEPs on allergic
inflammation
July 2009
D-127
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: DE: generated from a 30-kW (40 hp), 4-
Ciencewicki et al. cylinder Deutz BF4M1008 diesel
(2007,096557) engine
Species: Mouse Influeza A/Bangkok/1 /79 (H3N2
Gender: Female p,erotVP?) from Dr Melinda Beck of the
University of North Carolina, Chapel
Strain: BALB/c Hill
Age: 10-12wks O2, CO, NO2, NO, SO2
Weight: 17-20g q2: 20.9. 20.5% (Lo, Hi)
CO: 0.9-5.4 ppm
NO2: 0.25-1.13 ppm
NO: 2.5-10.8 ppm
SO2: 0.06-0.32 ppm
H3N2: NR
Particle Size: NR
Route: Inhalation (oropharyngeal aspiration of
virus)
Dose/Concentration: DE: 529 or 2070/yglm3
Time to Analysis: 4 h/d for 5d. Virus exposure
immediately after last DE exposure. Analyzed 18h
post infection
DE exposure on susceptibility to Influenza
infection: Mice exposed to 0.5 mgfm3 had
significantly greater levels of HA mRNA
compared to air-exposed mice. HA levels not
significantly altered in mice exposed to 2.0
mg/m3.
DE exposure on the Influenza-induced
inflammatory response: f IL-6 mRNA levels
were significantly greater when exposed to
0.5 mg/m3 of DE prior to infection compared
to air exposure. Significantly increased
amount of IL-6 protein observed in exposed
mice. Exposure to DE in absence of influenza
infection had no significant effect on IL-6
mRNA or protein levels.
DE exposure on pulmonary injury: Infection
with the influenza virus increases levels of
PMN in BAL fluid. Exposure to either dose of
DE prior to infection showed no significant
effect on PMN levels Exposure to DE alone
had no effect on PMNs in BAL fluid. Neither
exposure to DE nor infection with influenza
significantly increased BAL fluid protein levels
when compared to non-infected, air-exposed.
Other markers of injury, NAG and MIA were
not statistically affected by DE or influenza
exposure.
DE exposure on the influenza induced
interferon response: No significant change
in TFN-a mRNA levels at either dose of DE,
although mice exposed to 0.5 mgfm3 of DE
prior to infection had significantly greater
levels of IFN-B MRNA compared to air
controls. No effect on any of the IFNs
observed in uninfected mice exposed to DE.
DE exposure on surfactant protein
expression: Influenza virus infection alone
significantly increased expression of SP A in
air-exposed. Exposure to 0.5 mg/m3 of DE
prior to infection had significant decreases in
levels of SP A mRNA in the lungs, this effect
was not observed in 2.0 mgfm3 DE exposed.
Decrease seen in expression of SP A protein in
lungs of mice exposed to 0.5 mg/m3 DE prior
to infection. Levels of SP-D mRNA and protein
were significantly decreased in lungs of mice
exposed to 0.5 mg/m3 of DE prior to infection
compared with mice exposed to air or 2.0
mg/m3 DE prior to infection. Exposure to 0.5
mg/m3 of DE prior to infection with influenza
decreased levels of SP-D, especially in
airways. Mice exposed to 2.0 mg/m3 DE prior
to infection showed no significant difference.
Reference: Day
et al. (2008,
190204)
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 8-10wks
Weight: NR
GEE (General Motors 1996 model 4.3-L
V6 engine; regular unleaded fuel) (CO,
NO, NO?, SO2, NHs)
Particle Size: NR
Route: Whole-body Exposure Chamber
Dose/Concentration: Low(L)- 6.6±3.7 PM/m3,
Medium(M)- 30.3 ±11.8 PM/m3, High(H)-
59.1 ±28.3 PM/m3, High-Filtered(HF)-
Time to Analysis: Pre-OVA protocol: OVA or
saline sensitized 7d. OVA challenge day 14. GEE
or air exposed 6h/d on days 26-28. Immediately
after exposure on day 28 challenged with OVA.
Tested for MCh-induced changes 24h
postexposure then sacrificed. Post-OVA protocol:
OVA or saline sensitized 7d. OVA challenge day
14. GEE or air exposed days 15-17. Tested for
MCh-induced changes 24h postexposure then
sacrificed.
Post-OVA: In nonsensitized mice, neutrophils
dose-dependently decreased, IL-4 decreased in
the M group, IL-5 decreased in the HF group,
and IFN-n decreased at all exposures. In OVA-
sensitized mice, IL-13 dose-dependently
decreased.
Pre-OVA: In nonsensitized mice, neutrophils
and IgE decreased in the H group. IL-2
increased in the HF group and was dose-
dependent. Eosinophils dose-dependently
decreased. OVA-specific IgE increased in the H
group, and OVA-specific lgG2adose-
dependently increased. In OVA-sensitized
mice, OVA-specific IgGi increased in the M
group. Airway hyperresponsiveness was lower
in the M and HF groups.
July 2009
D-128
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: de
Haaret al. (2005,
097872)
Species: Mouse
Gender: Female
Strain:
BALB/cANNCrl
Age: 6-8wks
Weight: NR
CBP: Carbon black particles in
phosphate buffered saline, 1: 10 & 1:
100 dilutions (Brunschwich Chemicals,
Amsterdam, The Netherlands)
OVA: Ovalbumin (igma-Aldrich,
Zwijndrecht, The Netherlands)
Particle Size: CBP: 30-50nm
Route: Intranasal Droplet
Dose/Concentration: CBP± OVA 200, 20, 2/yg
(3.3, 0.33, 0.033 mg/ml)
OVA only: 20 /yg (0.5 mg/ml)
Time to Analysis: Droplet applied on days 0,1,
2. Sacrificed on day 4 or challenged with OVA
droplet on days 25, 26, & 27. Sacrificed on day
28
Acute airway damage and inflammation:
Only day 4 had LDH increased in the 200 /yg
CBP+OVA group. The 200//g CBP + OVA
group induced significantly higher numbers of
BAL cells compared to OVA control. Total
protein and TNF-a levels were increased only
in 200/yg CBP+OVA group. RAS, parameter
for phagocytosis, 200 /yg and 20 /yg
CBP+OVA had higher levels than OVA
controls.
Adjuvant activity on PBLN: Total
lymphocytes in PBLN significantly increased
4-5 fold in the 200 /yg CBP+OVA exposed. 20
/yg and 2 /yg exposures did not increase the
number of PBLN cells compared to OVA
control. All CBP+OVA concentrations induced
higher levels of IL-4, IL-5, IL-10, and IL-13,
with 200 /yg concentration having 10-200
times higher levels. IFN-y cytokine was
increased in the 200 /yg dose.
IgE Production: In CBP+OVA, IgE were
significantly increased.
PBLN and Lung Lymphocytes after OVA
challenge: PBLN cell numbers increased in
OVA and CBP+OVA sensitized mice. CD4 and
CD8 populations increased in both groups.
PBLN levels in CBP+OVA and challenged with
PBS were higher than mice treated with OVA
and challenged with PBS, both groups
cytokine production was low, only IL-5 levels
were significant in the CBP+OVA/PBS group.
Higher lung lymphocyte numbers were caused
by higher numbers of CD4 and CD19.
Production of IL-5 and IL-10 was four to five
times higher than in OVA treated mice.
OVA challenge induces asthma like airway
inflammation in CBP+OVA sensitized
mice: Total number of cells in BAL increased
10 fold in CBP+OVA mice challenged with
OVA. Eosinophils exhibited highest increase in
CBP+ OVA/OVA group. Perivascular and
peribronchial infiltrates and globlet cell
hyperplasia in CBP+OVA/OVA was confirmed
by histological examination. Antigen specific
inflammation induced in CBP+OVA mice.
July 2009
D-129
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: de
Haar (2006,
144746)
Species: Mouse
Gender: Female
Strain:
BALB/cANNCr
Age: 6-8wks
Weight: NR
CBP: fine (F) and ultrafine (UF) carbon
black particles (Ken Donaldson Group)
Ti02: fine and ultrafine titanium oxide
particles (Ken Donaldson Group)
OVA: Ovalbumin (Sigma-Aldrich,
Zwijndrecht, the Netherlands)
Particle Size: F CBP: 260.Onm
UF CBP: 14.0nm
F Ti02: 250.0nm
UFTKk 29.Onm
Route: Intranasal Droplet
Dose/Concentration: CBP: 200/yg (3.3 mgfmL)
Ti02: 200/yg (3.3 mg/mL)
OVA: 10 //g
CBP+OVA: 200 +10//g (see note in 3008)
Time to Analysis: Days 0,1,2: Exposed to OVA
or CBP+OVA. Sacrificed on day 8 & analyzed
after 2h, or continued to second group.
Second group: days 25, 26, 27 given OVA
challenge day 28: sacrificed , analyzed 24h post
sacrifice
Ultrafine particles induce lung
inflammation: UF Ti02 and CBP induced a
local inflammatory response in the airways
and showed higher levels of LDH and total
protein as compared to mice exposed to the F
particles. Cytokine levels were much higher in
groups exposed to ultrafine particles.
Histologic analysis of the airways showed
that exposure to ultrafine Ti02 or CBP leads to
peribronchial and perivascular inflammatory
infiltrates (mostly neutrophils). Exposure to
OVA alone, or combined with fine Ti02 and
fine CBP had no effects on lung histology.
Ultrafine stimulate local immune
responses: Ti02 and CBP particles stimulated
the local immune response against co
administered OVA antigen. Fine Ti02 particles
induced a low but significant increase in PBLN
cell number. Both types of ultrafine particles
elicited higher levels of Th-2 associated
cytokines, with UF CBP stimulating a greater
response. IFNy production was low, but
significantly higher than OVA exposures.
Ultrafine Ti02 increase ovalbumin-specific
IgE and IgG 1 levels: Levels of OVA specific
IgE were significantly increased in animals
exposed to the UF Ti02+ OVA compared to F
Tith or OVA-only.. Average IgE level in mice
exposed to ultrafine CBP+OVA was not a
significant increase. OVA-specific lgG2a not
detected in any groups. Ultrafine particles
stimulate allergic airway sensitization
against ovalbumin: At day 28, the PBLN cell
numbers were significantly higher in both
ultrafine and combination with OVA.
Production of OVA specific IL-4, IL-5, IL-10
and IL-13 by PBLN cells was significantly
increased in both ultrafine Ti02 and CBP. IFNy
levels were significantly increased in ultrafine
CBP+OVA treated animals. F Ti02 had low,
but significant, increases in IL-4 and IFNy
compared to OVA only. Allergic airway
inflammation and Influxes of eosinophils,
neutrophils and lymphocytes were only found
in both groups exposed to ultrafine particles.
Reference: de
Haar (2008,
187128)
Species: Mouse
Gender: Female
Strain: BALB/c,
CD80/CD86-
deficient,
~011.10
Age: 6-8wks
Weight: NR
Ultrafine Carbon Black (UFCB)
(Brunschwich Chemicals; Amsterdam,
The Netherlands)
Particle Size: Diameter: 30-50nm
Route: Intranasal Exposure. IP Injection. Tail
Injected.
Dose/Concentration: 20//gfmL
Time to Analysis: Exposed days 1, 2, 3. Kept in
supine position until recovery. OVA challenge
days 25, 26, 27. Spleens and lymph nodes from
D011.10 mice pooled and CD4+ T-cells isolated.
Solution injected into tail veins of BALB/c mice
day 0. CTLA4-lg ip injected days 0, 2. PBLN cell
suspensions plated, restimulated with OVA 4d.
UFCB+OVA induced proliferation of CD4+T-
cells, increased cytokine production.
UFCB+OVA did not induce any effects in
CD80/CD86-deficient mice. UFCB-induced
airway inflammation is dose-dependent.
Reference: de
Haaret al. (2008,
187128)
Species: Mouse
Cell Line: Myeloid
dendritic cells
(mDCs)
Ultrafine Carbon Black (UFCB)
(Brunschwich Chemicals; Amsterdam,
The Netherlands)
Particle Size: Diameter: 30-50nm
Route: Cell Culture
Dose/Concentration: 25//gfmL
Time to Analysis: mDCs cultured from bone
marrow. Exposed 18h. Cells isolated, stained for
flow cytometry.
UFCB+OVA increased mDCs in the
peribronchial lymph nodes, and their
expressions of CD80, CD86, and MHC-11.
July 2009
D-130
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Dong
et al. (2005,
088079)
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BNfCrlBR)
Age: NR
Weight: 200-
225g
DEP: SRM 2975 (NIST, Gaithersburg,
MD)
OVA: Ovalbumin (Sigma ,St Louis, MO)
Particle Size: 0.5 /jm (MMAD)
Route: Inhalation
Dose/Concentration: DEP: 20.6 ± 2.7 mg/m3
OVA 40.5 ± 6.3mg/m3
Time to Analysis: 4 h/d for 5 days + OVA 30
min/d 1 x wk on days 8,15 & 29. Sacrificed on
days 9 or 30.
Lung Inflammation/Injury: Both the BAL
proteins and inflammatory cell counts for DEP
exposure alone were not different from those
of the air exposed control, suggesting that
DEP exposure did not cause lung injury at 9 or
30 days post-exposure. OVA exposure caused
significant increases in neutrophils,
lymphocytes, eosinophils, albumin and LDH
activity in the lung after two exposures. DEP
did show a strong effect on OVA-induced
inflammatory responses.
Alveolar Macrophage (AM) function: OVA
exposure resulted in an increase in NO levels
in the acellular BAL fluid and AM conditioned
media. This increase was significantly
attenuated in rats exposed to DEP. DEP
exposure had no significant effect on the
production of IL-10 or IL-12 by AM recovered
from rats 9 and 30 days post exposure. In
contrast, OVA sensitization elevated both IL-
10 and IL-12 secretion by AM at both time
points.
Lymphocyte population and cytokine
production: DEP exposure was found to
increase the numbers of total lymphocytes, T
cells and their CD4+ and CD8+ subsets in
LDLN. OVA exposure also significantly
increased these cell counts on days 9 and 30.
DEP+OVA exposure showed a significant
reduction in total lymphocytes, T cells, CD4 +
and CD8+ subsets on day 30. Levels of IL-4
and IFN-y in lymphocyte conditioned media
were below detection limit of the ELISA kits.
Intracellular GSH levels in AM and
Lymphocytes: DEP exposure alone slightly
decrease GSH levels in AM, but markedly
reduced GSH concentration in lymphocytes on
days 9 and 30. OVA exposure significantly
decreased intracellular GSH in both cell types.
Combined exposure showed AM and
lymphocytes to have depleted intracellular
GSH.
OVA specific IgE and IgG levels in serum:
In all samples collected on day 9, both serum
IgG and IgE levels were under the detection
limits. On day 30, no measureable IgE levels
were found. The OVA exposure, however,
resulted in elevated IgE levels, and was
enhanced in rats preexposed to DEP. IgE and
IgG levels for DEP+OVA was twO times
higher than OVA alone indicating that DEP has
an adjuvant effect on the production of IgG
and IgE.
Effects of DEP and OVA on Lung iNOS
expression: AM from various exposure
groups did not stain for iNOS. 1 rat at day 9
from the combined DEP+OVA group showed a
slightly positive iNOS staining. On day 30, 2
of 5 rats from combined exposure group and 1
from the OVA group showed a positive airway
staining.
July 2009
D-131
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Dong
et al. (2005,
088083)
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BNfCrlBR)
Age: NR
Weight: 200-
225g
DEP: SRM 2975 Diesel Exhaust
Particles (NIST)
OVA: Ovalbumin (Sigma Chemical
Company, St Louis, MO)
Particle Size: 0.5/ym (MMAD)
Route: Nose-only Directed Flow Inhalation
Dose/Concentration: DEP 22.7 ± 2.5 mg/m3
OVA 42.3 ± 5.7mg/m3
Time to Analysis: Day 1, 8,15: OVA exposure
30min/day
Days 24-28: DEP exposure 4h/day
Day 29: OVA challenge
Day 30: Whole-body plethysmography
Day 31: Sacrifice
Effect of DEP on OVA induced allergic
responses: DEP exposure had a synergistic
effect with OVA on inducing airway hyper-
responsiveness (AHR) in rats. DEP alone had
no effect on IgG production. Levels of OVA-
specific IgG and IgE increased in OVA+DEP
exposure. This indicates that DEP pre-
exposure augments the immune response of
rats to OVA in the production of allergen
specific IgG and IgE.
Effect of DEP on OVA induced cell
differentiation: Neither DEP, OVA nor the
combination induced elevated levels of LDH
activity or albumin content, indicating that the
exposure protocols did not cause significant
lung injury. DEP alone induced moderate but
significant increase of neutrophil numbers.
OVA exposure induced a greater infiltration of
neutrophils than DEP, and infiltration of
eosinophils and lymphocytes. OVA-induced
eosinophil count markedly increased with DEP
exposure. Total lymphocytes, T cells, and their
CD4+ and CD8+ subsets in LDLN from rats
sensitized and challenged by OVA were
significantly higher than those of air-exposed
non sensitized rats. DEP+OVA exposure
resulted in substantial increase in T cells
compared to OVA alone.
Effect of DEP on OVA-induced oxidant
generation and GSH depletion: Exposure to
DEP or OVA alone had no effect on ROS
production by AM. Substantial elevation seen
in ROS for the DEP + OVA exposed group. Both
OVA and DEP exposures resulted in an
increased presence of NO in the acellular BAL
fluid and in AM conditioned media; OVA +DEP
exposure further increased these levels. The
ATII cells from OVA exposed rats exhibited a
higher percentage of cells that produce NO
and superoxide than air exposed, non
sensitized rats. DEP and OVA exposure
resulted in a significant increase in the
percentage of cells that produce NO and
superoxide over the control.
il\IOS expression: Immunohistological
analysis in lung tissues showed no AM
staining in any group. Airway epithelium was
found to be positive in all 5 rats from the
DEP+OVA group and 3 of 5 rats from single
exposure of DEP or OVA and 2 of 5 in air only
exposed rats. iNOS expression was
significantly higher in ATII cells isolated from
rats exposed to combined DEP and OVA .
GSH levels in AM and lymphocytes: Levels
were slightly lowered by DEP or OVA
exposure, though not statistically significant.
DEP+OVA showed a significant reduction in
GSH levels.
July 2009
D-132
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Drela
et al. (2006,
096352)
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 6 wks
Weight: NR
ASM: Air suspended PM from Upper
Silesia (Poland)
1/yg of ASM:
Pb (1.136 ng)
Cu (0.004//g)
Co (0.072 ng)
Mn (0.406 ng)
Fe (0.016 //g)
Cd (0.154 ng)
Cr (0.418 ng)
Ni (0.238 ng)
Particle Size: 0.3-10/ym
Route: Intraperitoneal Injection
Dose/Concentration: 170mg ASM/kg of body
weight
Time to Analysis: One time exposure, sacrificed
72h post exposure
CD28 expression on thymocytes at
different stages of development: ASM
exposure accelerated thymocyte maturation
but did not alter the expression of CD28 on
peripheral CD4 and CD8 T cells isolated from
lymph nodes. A slight but not statistically
significant decrease in the expression of CD28
on spleen T cells from ASM animals was
observed.
Distribution of CD28(low) and CD28(high):
Acute exposure to ASM resulted in the
increase of CD28(low) and decrease of CD28
(high) thymocyte percentages in the total
thymocyte population. The percentages of
CD28 low and high thymocytes did not differ
between intact and PBS controls. Acute ASM
exposure resulted in the increase of the
percentage of CD28(low) and the decrease of
CD28(high) thymocytes in the CD3 low
subset. The percentage of CD28 low and high
positive thymocytes did not differ in CD3 high
thymocyte subset. Natural regulatory CD4+
CD25+ T cells in the thymus: The
development of thymic natural regulatory cells
was unaffected by ASM.
Proliferation of splenocytes and lymph
node lymphocytes: Decreased proliferative
responses were evident in splenocytes from
ASM-exposed animals when cells were
stimulated with low but not high levels of anti-
CD3 mAb.ln contrast, lymph node
lymphocytes from ASM treated mice had
increased proliferative responses independent
of anti-CD3 concentration. Both CD4+ and
CD8+ T cells from ASM treated mice
proliferated more vigorously than from
controls. Almost all CD8+ T cells from ASM
mice were induced to proliferate.
Reference:
Dybing et al.
(2004, 097545)
Species: Mouse
Gender: NR
Strain: BALB/cA
Age: NR
Weight: NR
Assay: PLN,
ELISA
UP: Urban ambient particles collected
in 5 different sites (Amsterdam, Lodz,
Oslo, Rome, Dutch seaside) during four-
week periods in spring, summer, winter
seasons from March 2001 to March
2004.
DEP as reference std: SRM 1650
(NIST)
OVA: Ovalbumin (Sigma Chemical, St.
Louis, MO)
Particle Size: UP: PMioand PM2.6
Route: Injection in hind foot pad
Dose/Concentration: UP: 100/yg- 200/yg
DEP: 50 //g
OVA: 50 //g
Time to Analysis: PLN assay:
Day 0: 1 exposure to OVA alone, OVA
w/particles, particles alone.
Day 6: Lymph nodes harvested
Day 21: 1 OVA w/o particles exposure
Day 26: Antibody assay
Allergy screening: All samples were
immunostimulatory in the popliteal lymph node
assay; activity was weak in the absence of
OVA but statistically significant when injected
with OVA, indicating an adjuvant effect.
Particle adjuvancy was further demonstrated
via significant enhancement of OVA-specific
antibody responses. All ambient particle
fractions from all seasons increased lgG1.
Except for a few coarse samples, all fractions
significantly increased IgE. All fine fractions
and some coarse fractions significantly
increased lgG2a, indicating that most particles
could exert both Th1 and Th2 adjuvancy. In
general, fine particles demonstrated stronger
adjuvant activity than coarse in a pair-wise
comparison of coarse and fine particles from
the same location.
Reference:
Dybing, et al.
(2004, 097545)
Species Rat
Cell Lines:
Primary rat type 2
cells, rat alveolar
macrophages
Use:
Inflammatory
screening
Route: Cell Culture
Dose/Concentration: 0-50//gfml
UP: Urban ambient particles collected
in 5 different sites (Amsterdam, Lodz,
Oslo, Rome, Dutch seaside) during four
week periods in spring, summer, winter Time to Analysis: 20h
seasons from March 2001 to March
2004.
DEP: SRM 2975 (NIST)
OVA: Ovalbumin (Sigma Chemical, St.
Louis, MO)
Particle Size: PM10 and PM2.6
Inflammation screening: The coarse
fractions were more potent than the fine
fractions. Among the samples, the overall
effects of the coarse fractions on the cells
were dependent on the site of collection. High
MIP-2 levels were found using particles from
some spring collections. Coarse particles
collected in summer demonstrated the highest
potency, and samples collected during winter
proved to be less potent but seasonal variation
was not obvious for all sites. Only minor
responses were observed using fine fractions
from urban sites.
July 2009
D-133
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Farraj
et al. (2006,
141730)
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 6 weeks
Weight: NR
DEP: SRM 2975 NIST
OVA: Ovalbumin
Anti-p75: Rabbit anti-mouse p75
neruotrophin receptor polyclonal
antibody (Ohemicon, Temecula, OA)
Anti-trkA: anti-mouse trkA NGF
receptor antibody (Santa Oruz, Santa
Cruz, OA)
Particle Size: DEP: 1.47//m (MMAD),
2.75 GSD
Route: Nose-only Inhalation
Dose/Concentration: DEP: 1.78 to 2.18 mgfm3
Anti-p75: 50 /j\
Anti-trkA: 50 /j\
OVA injection: 20/yg
MCH: 0,16, 32, 64 mg/ml
Time to Analysis: On day 0: ip injection of 20
/yg OVA
Day 14: intranasal instilation of 50 ul anti-p75 or
anti-trkA
Day 14,1-h after 1st exposure: challenged with
OVA aerosol for 1h followed by a h exposure to
DEP
24 h after DEP exposure: MCH challenge
Airways responsiveness: No significant
differences in avg baseline Penh values of any
treatment groups.
Vehicle sensitized mice: exposure to DEP, anti-
p75 or anti-trkA had no effect on MCH-
induced Penh values.
OVA-sensitized DEP-exposed: seen increase of
Penh values. Administration of anti-p75 or
anti-trkA to OVA sensitized mice reversed DEP
induced Penh increases.
Lung function in ventilated mice: Compared
to vehicle sensitized mice, central airway
resistance (Rn) increased 62% in OVA
sensitized mice was not a significant increase.
OVA-sensitized DEP-exposed mice, anti-p75
decreased central airway resistance (Rn) and
anti-trkA did not significantly alter Rn. though
Rn response for anti-p75 was significantly less
than anti-trkA response, Constant phase
model parameter of tissue elastance not
significantly affected by any treatments or by
increasing MCH dose, indicating development
of significant regional ventilation
inhomogeneity during bronchoconstriction.
Airway pathology: OVA-sensitized mice had
small increases in intraepithelial mucus
compared to vehicle-sensitized mice. DEP
exposure did not enhance severity of OVA-
induced airway pathology. Anti-p75 or anti-
trkA administration did not influence airway
morphology.
BAL cells: Vehicle-sensitized DEP-exposed
mice had significantly enhanced macrophage
numbers by 92% compared to air-exposed,
vehicle-sensitized mice. Anti-p75 or Anti-trkA
administration significantly suppressed DEP-
induced macrophage increase to levels similar
to air-exposed, vehicle-sensitized group. DEP
co exposure significantly decreased number of
macrophages in OVA-sensitized mice to control
levels. Anti-trkA or anti-p75 had no effect in
OVA-sensitized, DEP-exposed. Eosinophil
number greater in OVA-sensitized DEP-exposed
mice than in vehicle-sensitized air-exposed
mice. No significant effects of DEP exposure
on neutrophils from vehicle- or OVA-sensitized
mice.
Cytokines: IL4: OVA-sensitized DEP-exposed
had five-fold increase over vehicle-sensitized,
air-exposed mice and anti-trkA or anti-p75
significantly inhibited the DEP-induced
increase.
IL5, IL13: OVA-sensitized DEP-exposed had no
significant change. Anti-p75 or anti-trkA
administration had no significant effect.
Serum IgE: OVA sensitized mice had a 10 fold
increase in IgE levels for air and DEP exposed
mice. Anti-p75, anti-trkA treatment did not
cause significant effects on IgE levels.
July 2009
D-134
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Farraj
et al. (2006,
189604)
Species: Mouse
Gender: Male
Strains: C57/BI6
Aqe: 6wks
DEP: SRM 2975 collected from diesel-
powered industrial forklift filter (NIST)
OVA: Ovalbumin
Anti-p75: Rabbit anti-mouse p75
neurotrophin receptor polyclonal
antibody
Particle Size: 1.47 (MMAD), 2.75
(GSD)
Route: Nose-only Inhalation
Dose/Concentration: DEP: 0.87 mgfm3
MCH: 0,16, 32, 64mg|ml
OVA: 20 /yg ip
Anti-p75: 50 /j\
Time to Analysis: Day 0: OVA in gel vehicle, ip
Day 14: anti-p75 exposure, intranasal instillation
1 h post anti-p75 exposure, OVA aerosol challenge
forih
1h post OVA challenge: DEP exposure for5h
48h post DEP exposure: MOH challenge
Airway responsiveness: No significant
differences in average Penh values among any
vehicle control groups. No significant
differences in treatment groups in OVA-
sensitized mice at baseline 0,16, or 32 mgfmL
of MCH. At 64 mg/mL MCH, OVA-sensitized,
DEP-exposed mice had a 22% increase in Penh
compared to vehicle mice, and a 68% increase
compared to vehicle-sensitized, air-exposed
mice. Instillation of anti-p75 inhibited the DEP
induced increased Penh.
BAL cells: DEP exposure in vehicle-sensitized
mice significantly increased macrophages by
161% compared to air-exposed, vehicle-
sensitized mice, while OVA-sensitized mice
had 69% increase. Anti-p75 administration
significantly suppressed DEP-induced
macrophage increase in vehicle-sensitized
mice. No significant effects of DEP exposure
or anti-p75 treatment in OVA-allergic mice.
OVA-sensitized air-exposed mice had a several
hundred fold increase in the number of
eosinophils. No significant effects of DEP
exposure or anti-p75 treatment on eosinophils
from OVA-sensitized mice. OVA-exposure or
DEP-exposure had no significant effects on
neutrophil or lymphocyte number.
BAL cytokines: No significant effects of DEP
alone or with OVA on IL-4, IL-5, or IL-13.
Serum IgE: OVA sensitization in the presence
or absence of DEP or anti-p75 caused at least
a 3 fold increase in IgE levels. No significant
effects of DEP or anti-p75 treatment on IgE
levels.
Reference:
Finkelman et al.
(2004, 096572)
Species: Mouse
Gender: Female
Strains: BALB/c
and C57BL/6
Aqe: 2-4 m
DEP: 4JB1 type: Isuzu Automobile,
Tokyo, Japan
Particle Size: 2 //m MMAD
Route: Two groups: Groupl: 1 ip injection of 2
mg of DEP.
Group 2: daily ip injections of 2mg of DEP
Dose/Concentration: 2 mg
Time to Analysis: 2-96h
Serum Cytokines: Mice in group 1
demonstrated an increase in serum IL-6
production but no increase in IL-4 or IL-2
production. IFN-n levels were decreased in
group 2. TNF production was not affected.
Spleen Cytokines: When injected before LPS,
DEP had little effect on the LPS-induced TNF-
a and IL-6 response, but resulted in a minor
suppression of INF-n and IL-10. DEP LPS-
induced increase in INF-n mRNA responses in
spleen cells. DEP caused a dose related
suppression of LPS stimulated INF-n. DEP had
little or no effect on the percentage of NK or
NKT cells in the spleenand inhibited LPS-
induced IFN-n production by NK and NKT. DEP
failed to inhibit the IFN-n response by anti-
CD3 mAb-activated NKT cells. Oxidant
activity was not responsible for DEP inhibition
of LPS-induced IFN-n production.
Reference:
Fujimaki and
Kurokawa (2004,
189576)
Species: Mouse
Gender: Male
Strains: BALB/c
Age: 4 wks
Cell Types:
Cervical lymph-
node (CLN) cells
DE±particles: Comparison of exposure
to DE including particles andexposure
to particle-filtered DE
All mice were injected with sugi basic
protein (SBP), a cedar pollen allergen
before exposure to DE
Particle Size: 0.4 /jm MMAD
Route: Whole-body Exposure
Dose/Concentration: Exposure to: 0,1.0 mgfm3
or 1 .Omg/m3 DE gas only (0.04 mg/m3 PM)
Time to Analysis: Exposure for 12h daily for
5wks. Days 14 and 35 challenge with SBP
intranasaly. Evaluation is 24 and 48h after final
SBP injection.
CLN response: Exposure to DE or DE gas did
not affect B1 lymphocyte subpopulations of
CLN. Culture supernatants of CLN cells from
DE exposedfSBP immunized mice showed
significant increase in MCP-1 at 24 and 48 h.
Exposure to DE or DE gas significantly
increased the amount of TARC and MIP-1a in
CLN cells from SBP-immunized mice at 48 h.
Composition of DE and DE gas: DE: 12.09
± 0.15 NOx, 1 99±0.02 NO2,10.02±0.12
NO, 0.18±0.002S02 and 1769.2±13.2 CO?
(all in ppm). DE gas: 11.93 ± 0.13 NOx,
2.93±0.06N02, 8.91 ±0.09 NO,
0.11 ±0.003S02 and 1838.8±15.3 CO? (all in
ppm)
July 2009
D-135
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Fujimaki et al.
(2005,156456)
Species: Mouse
Gender: Male
Strains: C57BL/6
Age: 4 wks
Cell Types
CLN (lymph node)
and Plasma cells
DE generated by 4 cyl 2.74 I Isuzu
diesel
DE gas - DE filtered to remove
particles
Composition of Disel Exhaust: DE
DEP: 1.01 mg/m3
1796ppm CO2
12.09ppm NOx
0.18 ppm SO2
Composition of filtered DE Gas: DEP:
0.04 mg/m3
1839ppm CO2
11.93ppm NOx
0.11 ppm SO2
Sugi Basic Protein (SBP)- allergen
Particle Size Average diameter of
DEP 0.4/ym
Route: Whole-body Exposure Chamber
Dose/Concentration: 1.0mg DEPfm3 or 1.0mg
DEP/m3 DE gas
Time to Analysis: Exposure for 12h daily for
5wks. All mice were injected IP with 100 fjg
SBP before exposure to gas or DE and again
received 50 /yg SBP intranasal^ on days 14 and
35. Evaluation is 1d after final SBP-immunization
(mice are euthanized and CLN and blood samples
are collected)
CLN: Exposure to DE and gas led to a
decrease in total number of CLN cells and
percentage of CD4+ and TCR-B levelsCell
proliferation response to SBP was higher in
gas-exposed mice than in the control group.
The production of MCP-1 increased in CLN
cells when stimulated with SBP (in vitro) but
the difference was not significant at 24 and
48h. SBP-stimulated cells in gas-exposed mice
showed greatly enhanced MIP-1a production
at 24 and 48 h. Exposure to gas increased the
amount of TARC in the culture supernatants
of CLN cells.
PLASMA: Exposure to DE or gas significantly
decreased the anti-SBP lgG1 antibody titers
and increased increased the anti-SBP lgG2a
antibody titers in mouse plasma.
Reference:
Fujimoto et al.
(2005, 096556)
Species: Mouse
Gender: Female
1st day of
pregnancy)
Strains: Sic: IRC
Cell Types
Fetal Cells/ RNA
Spleen Cells
DEP: generated by a 2369-cc diesel
engine operated at 1050 rpm and 80%
load with commercial light oil
Particle Size: 0.4 /jm MMAD
Route: Whole-body Inhalation Chambers
Dose/Concentration: 0.3,1.0 and 3.0 mg
DEP/m3(Groups 1,2,3)
Time to Analysis: Exposure began at 2d
postcoitum and was continued until 13d
postcoitum. Exposure time was 12h daily for
7d/wk. Pregnant females were sacrificed 14d
postcoitum.
mRNA Expression in Placentas: In groups
exposed to DE, the expression of CYP1A1
mRNA decreased to undetectable levels during
placental absorption and INF-n was increased.
Levels of CYP1A1 mRNA in normal placentas
from DE-exposed mice were unchanged.
mRNA levels of inflammatory cytokines IL-2,
IL-5, IL-12a, IL-12B and GM-CSF increased in
placentas of mice exposed to DE.
Reference: Gao
et al. (2004,
087950)
Species: Human
Cell Line: Lung
fibroblasts
infected with
Mycoplasma
fermentans
R0FA: collected near a power plant in
FL burning low sulfur # 6 oil.
(PM from Dusseldorf, volcanic ash for
Mt. St. Helens, PM from Utah used to
compare against R0FA in one
experiment)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Cells were seeded into 6-
well plates (3-4.5X106 cells/3mL/well) or 24-well
plates (0.6-1 X106 cells/1.0 mL/well)
3,10, 20, 40, 50 jug/ml
Time to Analysis: 24,48h
Cytokines: R0FA exposure in combination
with Mycoplasma fermentans infection
synergistically amplifies the induction of IL-6
production in human lung fibroblasts (HLF).
PM from the other souces has little
synergistic effect on IL-6 release. Exposing
HLF cells to,M. fermentans derived
macrophage activating lipopeptide-2 (MALP-2)
and R0FA has the same synergistic effect as
M. fermentans infection and R0FA. MALP-2
and R0FA extract have a similar synergistic
effect that requires more time to appear.
R0FA contains high levels of V, Ni, Fe and Cu.
Exposure of HLF to NiS04alone and NiS04
with MALP-2 produced 10 and 50 fold
increases, respectively, in IL-6 production.
Exposure of HLF to CuS04, V0S04 and
Na3V04, with and without the presence of
MALP-2, did not produce as dramatic results
as seen with Ni. The action of NiS04 and
MALP-2 on IL-6 production was found to be
dose dependent.
July 2009
D-136
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Gavett et al.
(2003,053153)
Species: Mouse
Gender: Female
Strain: BALB/c
Aqe: 7wks
PM2.6 from the German cities of
Hettstedt or Zerbst
PM Composition: samples from
Hettstedt have several-fold higher
levels of zinc, magnesium, lead, copper
and cadmium than samples from
Zerbst.
Particle Size: PM2.6
Route: Oropharyngeal Aspirations
Dose/Concentration: 50-100/yg PM in 50/yL
saline
Time to Analysis: Mice were exposed to one
dose of 100 /yg BAL 18h after exposure.
Sensitization Model: Mice were exposed to 50/yg
PM 2h before being sensitized with 10 fjg OVA,
repeated two days later. On day 14- all mice
were challenged with20 /yg OVA.
Parameters measured on days 2 and 7 after
FINAL exposure to OVA.
Challenge Model: Mice were sensitized IP with
20 /yg OVA or adjuvant only. 14d later mice were
exposed to 100 /yg PM2.6 followed 2h later by 20
/yg OVA. Parameters measured on days 2 and 7
after FINAL exposure to OVA.
BAL Analysis: Hettstedt PM significantly
increased BAL protein and NAG levels. Zerbst
PM did not. Mice exposed to Zerbst had lower
levels of LDH than control groups. Hettstedt
exposed mice had increased levels of IL-1B,
IL-6 and MIP-2 in comparison to control and to
mice exposed to Zerbst PM. PM2.6 at a dose of
100 /yg was not found to be toxic, therefore
used for subsequent studies.
Airway Responsiveness (PenH):: In allergic
mice tested immediately after exposure,
Hettstedt PM increased PenH 190% compared
to baseline, Zerbst increased PenH by 120%
and the Control increased by 44%.:.: Two
days after OVA challenge, no differences in
non-allergic mice from either group. In allergic
mice, Hettstedt PM still caused a significant
response to Mch responsiveness, Zerbst none.
No effects on day seven.
IgE Levels: Serum collected on day 2 showed
antigen-specific IgE was increased by
Hettstedt PM2.6 in both the sensitization and
challenge phases when compared to the
control and exposure to Zerbst. Day 7 serum
indicated no effect.
BAL Cell Counts: In non-allergic mice both
Hettstedt and Zerbst PM increased BAL
neutrophil numbers (3-fold: not statistically
significant) and in allergic mice, only Hettstedt
PM significantly increased neutrophil count.
Eosinophil numbers were increased only in
allergic mice exposed to Hettstedt PM.
Lymphocyte numbers were not different
among groups.
BAL Lung Injury:: At 2 days after both
Hettstedt and Zerbst PM administered in
allergic mice caused significant increases in
protein, LDH and NAG compared to the non-
allergic groups. Both PMs caused an increase
in LDH in allergic mice compared to the
allergic control, but only Hettstedt caused an
increase NAG in allergic mice compared to
control. At 7 days no effect.
BAL Cytokines: Allergic mice had increased
levels of IL-4, IL-5 and IL-13 compared to non-
allergic mice (at 2 days after). IL-5 was
significantly increased by exposure to either
PM in allergic mice compared to non-allergic
mice. Exposure to either PM caused an
increase in TNF-a and IFN-n (by 6-8 fold) in
allergic mice, there was also an increase in
these inflammatory cytokines in the non-
allergic group but was not statistically
significant. No significant effects were
observed in animals that underwent the
sensitization protocol alone for any
measurement or endpoint.
Reference:
Gowdy et al.
(2008, 097226)
Species: Mouse
Gender: Female
Strain: BALB/c
Age: -12-
14wks
Weight: 17-20g
DEP (30kW (40hp) 4-cylinder Deutz
BF4M1008 diesel engine, steady state,
20% full load) (Low dose: 21% O2,
0.4wt ratio OC/EC; High dose: 20.7%
O2, 0.4wt ratio OC/EC) (CO, N0«, SO2)
Particle Size: Diameter: ~240nm
Route: Inhalation Exposure Chamber
Dose/Concentration: Low- 514±3 //gf'nr1, High-
2026 ± 38 /yglm3
Time to Analysis: Exposed 4hfd, 1 or 5d
(consecutive). Necropsied immediately or 18h
postexposure.
Inflammation: Neutrophils and lung injury
dose-dependently increased. ICAM-1 increased
immediately after both exposures and after
18h postexposure in the low dose.
Cytokines: After 1 d exposure, IFN-n and
TNF-a increased immediately at both doses
and the high dose, respectively. Immediately
after 5d exposure TNF-a and IFN-n increased
at both concentrations and IL-6 increased at
the low dose. At 18h postexposure IL-6 and
IFN-n increased at both doses, TNF-a and IL-
13 increased at the low dose, and MIP-2 dose-
dependently increased.
CCSP, Surfactants: CCSP decreased. SP A
and SP-D decreases were only significant
after 5d exposure, 18h postexposure.
July 2009
D-137
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Hamada et al.
(2007,091235)
Species: Mouse
Gender: Female
(Pregnant close to
partruition)
Strain: BALB/c
Use: Pregnant
mice were
exposed and
offspring were
analyzed
ROFA
Route: Nebulized ROFA lechate
ROFA was obtained from a precipitator Dose/Concentration: 50mgfmL dilution
until of a local power plant.
Composition of ROFA (in/yglmL):
341.2 Ni, 323.4 V, 232.2 Zn, 18.3 Co,
15.8 Mn, 8.4 Ca, 6.7 Cu, 6.1 Sr, 5.0
mg, 0.9 Sb, and 0.6 Cd.
Particle Size: NR
Time to Analysis: Pregnant mice exposed to
nebulized ROFA leachate for 30 min/day at days
14,16 and 18 of pregnancy.
Newborns received a single injection (ip) of
0VA(5 /yg) + alum (1 mgO at day followed by
exposure to: 1. aerosolized OVA days 12,13 and
14 (2-week old protocol)
OR
2. aerosolized OVA days 32, 33 and 34 (5wk old
protocol)
Analysis 48h after final allergen exposure
Susceptibility to Asthma:.Exposure of
motherto PBS aerosols during pregnancy did
not result in prominent asthma features in
young. The offspring of the ROFA mothers
revealed increasing AHRand elevated numbers
of eosinophils in the BAL fluid. Similar results
were seen in both the 2-week and 5-week old
groups.
IgE Levels: Histopathology revealed
prominent inflammation in the lungs of the
ROFA neonates and increased allergen-specific
IgE and lgG1 levels in the 5-week group.
Maternal Influence: Breast milk was not
shown to be responsible for the increased
susceptibility to allergy seen in offspring.
IL-4 and IFN-D: IL-4 and IFN-n levels in
maternal mice showed no difference between
PBS exposed or ROFA exposed mice. Cultured
spleen cells from mice born of ROFA-exposed
mothers showed either increased or similar
levels of IL-4 and decreased production of IFN-
~ causing an increase in the ratio of IL-4/IFI\l-n
indicating greater susceptibility to develop
Th2- allergic response.
Eosinophils: Exposure of mothers to Ni levels
similar to those found in ROFA had no
appreciable effect on BAL eosinophil.
Reference: Hao
et al. (2003,
096565)
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 6-7 weeks
DEP
DEP collected from a 4-cylinder diesel
engine under a 10-torque load.
Particle Size: NR
Route: Nebulization
Dose/Concentration: 2 mgfm3 DEP
Time to Analysis: Mild Sensitization- Mice
receive IP OVA alum and are challenge with
aerosolized OVA with and without DEPs. Mice
sacrificed d19.
Post challenge Model- DEPs are delivered to mice
sensitized by IP OVA and alum. Mice sacrificed
d23.
Transgenic Mice: Mice exposed to nebulized
saline or DEPs for 1-h daily for 3 days. Mice
sacrificed d5.
Mild Sensitization: Exposure of previously
OVA sensitized mice to aerosolized DEP and
OVA did not affect OVA-specific IgE
production, BAL eosinophilia or methacholine-
induced AHR. Aerosolized particles induced
inflammation and increased MBP deposition
and MBP positive eosinophils in the mucosa.
IL-5 Transgenic: Exposure to aerosolized DEP
did not change BAL cytokine levels, but did
increase AHR and BAL cell count.
Classic Sensitization, Post-Challenge: Did
not lead to a discernable increase in OVA-
induced AHR. DEP treatment was associated
with increased airway inflammation and mucin
production in larger and intermediary airways.
Particle Size: Diameter: 2.5/ym
Reference: CAPs (Detroit: July-Sept. 2000;
Harkema et al. Harvard Ambient Fine Particle
(2004,056842) Concentrator)
Species: Rat
Gender: Male
Strain: F344, BN
Age: 10-12wks
Weight: NR
Route: Inhalation Exposure Chamber. IT
Instillation.
Dose/Concentration: 4d concentration:
676±288/yg/m3, 5d concentration: 313± 119
/yg/m3, July concentration: 16-185 /yglm3,
September concentration: 81-755//g|m3; IT
Instillation- 200 fjL (soluble and insoluble)
Time to Analysis: F344 rats sensitized to
endotoxin, BN rats to OVA. Exposed 10h/d 1,4,
5d (consecutive). Another group of rats IT
instilled. Both groups killed 24h postexposure.
The retention of PM in the airways was
enhanced by allergic sensitization. Recovery of
anthropogenic trace elements was greatest
for CAPs-exposed rats. Temporal increases in
these elements were associated with
eosinophil influx, BALF protein content and
increased airway mucosubstances. A mild
pulmonary neutrophilic inflammation was
observed in rats instilled with the insoluble
fraction but instillation of total, soluble or
insoluble PM2.6 in allergic rats did not result in
differential effects.
July 2009
D-138
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Harrod et al.
(2003, 097046)
Species: Mouse
Gender: NR
Strains: C57BL/6
Age: 8-10wks
Use: Infected
with RSV
DEE: Diesel Engine Emissions
generated from a 5.9-liter turbo diesel
engine fueled by Number 2 fuel.
DEE Composition:
NOx: 2.0-43.3ppm
CO: 0.94-29.0ppm
SO?: 8.3-364.9ppb
Particle Size: for high 81 low level:
0.1-0.2//m (MMAD)
Route: DEE: Whole-body Inhalation
RSV: IT administration
Dose/Concentration: DEE: 38.8//gfm3 (low
level)
or
10027 /yglm3 (high level)
RSV: 100//I (1 Oepfu)
Time to Analysis: 6 h/d for 7d.
After the final 6h exposure period mice were
infected with RSV.
Parameters measured 4d post infection
Viral Gene Expression: For air+RSV, RSV-F
gene expression was not apparent but RSV-G
gene expression was detectable at very low
levels. In DEE+RSV (for high and low levels),
RSV-F and -G were markedly elevated. B-
Actin mRNA levels not changed in DEE-
exposed compared to air-treated. DEE+RSV
for high and low levels show 10- to 20- fold
induction of RSV-G mRNA levels as compared
to air+RSV.
Cell numbers in BALF: Uninfected low-level
DEE did not induce statistically significant
increase in cell numbers as compared to
air+RSV. High level DEE+RSV caused
increase as compared to air+RSV. Uninfected
high-level DEE had increase as compared to
uninfected air group. For all groups, alveolar
macrophages were predominant cell type and
no substantial changes in infiltrating cell
populations by exposure to DEE were noted.
Lung inflammation & Airway Epithelial
Morphology: Lung sections from air- or DEE-
exposed, uninfected did not exhibit any
observable change. Low level DEE + RSV had
increased inflammatory cell infiltration in
peribronchial regions and loss of normal
cuboidal appearance of Clara cells as
compared to air+RSV. High level DEE+RSV
had more apparent lung-inflammation,
especially surrounding bronchi and
bronchioles, and increased appearance of
pseudostratified, columnar epithelial cell
morphology and apparent airway epithelial cell
sloughing as compared to low level DEE+RSV,
indicating dose-related increase in lung
histopathology to RSV infection by prior DEE
exposure.
Inflammatory Mediators: TNF-a and IFN-n
were significantly increased in RSV-infected
mice exposed to low or high level DEE and not
increased in RSV-infected mice exposed to air.
TNF-a levels elevated to similar levels for low
and high level DEE+RSV. IFN-n exhibited more
dose-related increase with higher levels in high
level DEE+RSV versus low level DEE+RSV.
Mucous Cell Metaplasia: DEE exposure in
uninfected was not altered. Mucous
metaplasia was increased in epithelium of
RSV-infected mice when exposed to DEE in a
dose-dependent manner. Following high level
DEE+RSV, muscous staining of airway
epithelial cells in more distal airways was
occasionally observed.
CCSP Production in Airway Epithelium:
DEE alone did not have an effect CCSP-
producing cells, or Clara cells, decreased in
Low DEE + RSV iand further decreased in high
level DEE+RSV in large and terminal airways.
Surfactant Protein B: proSP-B staining post
RSV alone shows now discernible decrease
when compared to uninfected. Staining levels
in alveolar lung regions decreased when
exposed to low level DEE +RSV, and further
decreased in high level DEE+ RSV. Staining in
airway epithelium following high level
DEE+RSV diminished when compared to RSV
alone or low level DEE+RSV.
SP-A: In alveolar type II cells and airway
epithelial cells for untreated and air +RSV, no
discernible changes in levels. Prior exposure to
low or high level DEE decreased SP A staining
in alveolar type II cells and airways epithelial
cells during RSV infection.
July 2009
D-139
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Pollutant	Exposure	Effects
Reference: DEE (2 2000 model 5.9-1 Cummins ISB
Harrod et al. turbo diesel engines, No. 2 certification
(2005,088144) diesel fuel)
Species: Mouse Particle Size: NR
Gender: Male
Strain: C57B1/6
Age: 10-12wks
Weight: NR
macrophages. Higher doses had a
concentration-dependent increase.
Ciliated, Clara Cells, TTF-1: Generally,
ciliated cells decreased with exposure dose,
were more discernible in inflamed airways,
and higher doses caused effects in small distal
airways. Clara cells decreased equally at all
exposures and were most notable in the distal
airway epithelium. TTF-1 decreased
postinfection.
Route: Exposure Chamber
Dose/Concentration: Low- 30//gfm3 PM, Mid-
Low- 100/yg/m3PM, Mid-High- 300//g/m3PM,
High-1000 //g/m3 PM
Time to Analysis: Exposed 6hfd, 7dfwk, 1wk or
6m. 1wk exposure repeated on separate
occasion. Immediately after exposure, mice
anethetized, IT instilled with Pseudomonas
aeruginosa.
Bacterial Clearance: Lung bacterial
clearance was decreased at all levels after
1wk exposure and was concentration-
dependent 18h postinfection. Bacterial
clearance was not affected at 6m and
bacterial counts were higher.
Inflammation, Particle Deposition: Lung
inflammation and histopathology were
increased in all exposure groups postinfection.
All pynneiiro nrniinc nnQQOQQoH nartirlo.laHon
Reference:
Heidenfelder et al.
(2009,190026)
Species: Rat
Gender: Male
Strain: Brown-
Norway
Age: 10-12wks
Weight: NR
CAPs (Grand Rapids, Ml; July)
Particle Size: Diameter: 0.1-2.5/jm
Route: Whole-body Inhalation Chamber
Dose/Concentration: CAPs: 493±391 //g/m3;
0C: 244±144«g/m3, EC: 10±4/yg/m3, Sulfate:
79± 131 /yglm , Nitrate: 39±67 /yglm3,
Ammonium: 39 ± 59 /yglm3, Urban dust (Fe, Al,
Ca, Si): 18±6/yglm3
Time to Analysis: Sensitized to OVA 3d.
Challenged with OVA or saline 2wks later for 3d.
Exposed to CAPs 8h/d, 13d. OVA or saline
challenge 9d after first challenge. Sacrificed 24h
after last CAPs exposure.
CAPs enhanced the effects of OVA by causing
differential expression in genes primarily
involved in inflammation and airway
remodeling. CAPs exposure alone had no
effect on gene expression. CAPs+OVA also
increased IgE, mucin glycoprotein, and BALF
total protein, and caused a more severe
bronchopneumonia, increased mucus cell
metaplasia/hyperplasia and mucosubstances.
Reference:
Hiramatsu et al.
(2003,155846)
Species: Mouse
Gender: Female
Strains: BALB/c
and C57BL/6
Age: 8wks
Weight: 17-22g
DE -DE (generated by diesel engine and
diluted with filtered clean air)
Particle Size: NR
Route: Exposure Chamber
Dose/Concentration: Low dose -0.1 mgfm3
Highdose ¦ 3mg/m3
Time to Analysis: 1 or 3 months (7 h/d , 5 dfwk
Lung Histopathology: DEP-laden
macrophages accumulated in the alveoli and
peribronchial tissues in a dose- and duration-
dependent manner in both strains.
Lymphocytes and neutrophils increased in both
strains, but were greatest in the BALB/c mice.
BALF and Mac-1 Positive Cells: BALT
formation in DEP-laden AMs was seen at the
high dose group and was greater in the
BALB/c mice. Mac-1 positive cells, a marker
for phagocytic activation of the AMs, was
observed in the high dose groups of both
strains at 1 and 3 months, and in the low dose
group at 1 month in BALB/c mice.
Cytokine and il\IOS mRNA expression^
month of exposure increased TNF-a, IL-
12p40, IL-4 and IL-10 mRNA in a dose-
dependent manner. IL-1 B and iNOS decreased
in a dose-dependent manner. IFN-n mRNA
expression increased in BALB/c mice and
decreased in C57BL/6 mice. Similar results
were seen at 3 months, except IL-4 and IFN-n
mRNA expression decreased in the BALB/c
mice. In C57BL/6 mice, IL-4 and IL-10 mRNA
increased at the low dose but decreased at
the high dose. NF-nB activation occurred after
1 week and 1 month DE exposure and was
more prevalent in BALB/c mice.
July 2009
D-140
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Hiramatsu,
(2005, 088285)
Species: Mouse
Gender: Female
Strains: BALB/c
Age: 8wks
Weight: 17-22g
DE (generated by diesel engine and
diluted with filtered clean air.)
Mycobacterial Infection -
M.tuberculosis (ATCC35812) Kurono
strain
Particle Size: NR
Route: Exposure Chamber
Dose/Concentration: Low DE dose - 0.1 mgfm3
High DE dose ¦ 3mg/m3
Mycobaterial infection: 5mL (nebulized) of a 10e
colony-forming units (CFU) suspension
Time to Analysis: 1, 2 or 6m (7hfd, 5dfwk). 6
mice from each group infected on last day of DE
exposure. CFU evaluation 7wks postinfection.
Histopathological observations: DEP-laden
AMs and DEPs in the alveoli and peribronchial
tissues increased in a time-dependent manner.
DE-exposed mice had a greater number of
mycobacterial lesions, which were
disseminated. Lesions in the control mice had
clear borders and consisted of epithelial cells
and lymphocytes. Tubercle bacilli and DEPs
coexisted in AMs. BALT was seen around
DEPs in the 2 and 6-month exposure groups.
Inflammation cells increased in a time-
dependent manner with respect to DE
exposure.
Granulomatous lesions in lungs: 6-month
DE-exposed mice had a significantly higher
amount of gross lesions than the 6-month
control mice.
Mycobacterial burden: CFU in lungs were
increased in DE-exposed animals but only the
6 month exposure resulted in statistically
significant increases (a — 4-fold increase over
control). CFU in spleen were not significantly
altered by DE exposure.
Cytokines and il\IOS mRNA expression:
Infected DE-exposed mice had time-dependent
increases of TNF-a, IL-1B, IL-12p40, IFN-n
and ilMOS mRNAs compared to the infected
control mice. IL-12 mRNA expression
decreased in infected 6-month DE-exposed
mice.
Reference:
Ichinose, T. et al.
(2003, 041525)
Species: Mouse
Gender: NR
Strains:
BALB/cAnN, ICR,
and C3H/HeN
Age: 6wks
Weight: NR
DE: DE generated by 3059cc 4-cylinder Route: Exposure Chamber;
IT Instillation
diesel engine
Der f: Crude extract of D. farinae
Particle Size: DEP - Mass median
aerodynamic diameter: 0.4/ym
Dose/Concentration: 1. Air
2.	DE only: 3.0mg/m3
3.	Air + Der f: 1mg Derf
4.	DE 3.0mg/m3 + 1mg Derf
Time to Analysis: DE: 8wks (12hfd, 7d/wk)
Der f: 2wk intervals, 6wks
Analyzed 3d after last instillation
Light microscopic observations: DE
exposure caused the proliferation of
nonciliated cells and epithelial cell
hypertrophy. Soot-containing macrophages
were found in the alveolar tissue spaces.
Accumulated lymphocytes were present in the
peribronchiolar lymphoid tissue. Inflammatory
cells and soot-containing macrophages were
found in the submucosal layer and the vessel
instertitium of mice treated with DE +Der f in
all strains. DE+Der f treated C3H/He mice had
desquamated goblet cells.
Eosinophil infiltration: DE treated C3HfHe
mice had a slight eosinophile infiltration in the
submucosal layer. DE+Der f treated mice in
all strains had a slight to moderate eosinophile
infiltration.
Lymphocyte accumulation: Lymphocytes
significantly increased in all strains under the
DE treatment as compared to the air+saline
treatment, and further increased under the
DE+Derf treatment.
Goblet cell proliferation: Little proliferation
was seen in all strains under the DE
treatment. DE+Der f caused a significant
increase in proliferation compared to air+Der f
in ICR mice, but a significant decrease in
C3H/He mice.
Local cytokine and chemokine expression
in lung tissue supernatant: DE -t-saline
significantly increased MIP-1a in all strains.
MCP-1 also increased but not significantly.
DE+Der f increased IL-5, RANTES, eotaxin,
MCP-1 and MIP-1a in all strains as compared
with air+saline and air+Der f. IL-5 decreased
in C3H/He mice treated with DE +Der f
compared to air+Der f. IL-3 decreased in ICR
and C3HfHe mice compared to air+saline.
Der f-specific immunoglobulin production
in plasma: Increased production of lgG1 was
statistically significant in ICR and C3HfHe
mice treated with DE+Der f as compared to
air+Der f. IgE was low in all strains.
July 2009
D-141
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Ichinose et al.
(2004,180367)
Species: Mouse
Gender: NR
Strains: BALB/c,
ICR and C3H/He
Age: 5wks
Weight: NR
DEP: 2740cc 4-cylinder engine
D. farinae-. crude extract
Particle Size: DEP - Mass median
aerodynmaic diamter of 0.4/ym
Route: IT Instillation
Dose/Concentration: 1. D. farinae-. 1 fjm in PBS
2. D. farinae + DEP: 1 /yg in PBS + 50/yg mg
DEP
Time to Analysis: 4 times at 2wk intervals.
Mice examined 3wks after last instillation
Histological changes: Mice in all three
strains treated with DEP+Z7. farinae had a
significant recruitment of eosinophils, more
proliferation of goblet cells, and more eotaxin
positive macrophages in the alveoli than mice
treated with D. farinae alone.
Local cytokine expression in lung tissue
supernatant: DEP +Z7. farinae induced
significant elevation of IL-5 in ICR and C3HfHe
mice as compared to D. farinae alone.
Production levels of IL-4 and RANTES did not
correlate with the manifestations of allergic
airway inflammation induced by the D. farinae
treatment with or without DEP.
Cytokine expression in plasma: IL-5 in
C3H/He mice treated with DEP+Z7. farinae
was significantly higher than D. farinae alone.
RANTES was unaffected by the DEP
treatment in all strains.
D. farinae-sfeaWc immunoglobulin
production in plasma: The adjuvant effect of
DEP on lgG1 production was observed in all
three strains, with C3H/H3 being statistically
significant. The production levels of IgG 1
correlated with the manifestations of
eosinophilic airway inflammation by both
treatments. No adjuvant effect on IgE
production was observed.
Reference:Inoue
et al. (2006,
097815)
Species: Mouse
Gender: Male
Strain: ICR
Age: 6wks
Weight: 29-33g
PM-0C: Urban PM, collected for 1
month during early summer, 2001 in
Urawa city Saitama, Japan
LPS
Particle Size: < 2.0mm, 40mg/m3
Route: IT Instillation
Dose/Concentration: Vehicle group: PBS
PM-OC group: 4mg/kg of PM-OC
LPS group: 2.5mg/kg of LPS
PM-OC +LPS group: combined adminstration of
PM-OC +LPS
Time to Analysis: 24h after IT administration
Effects of PM-OC on LPS related lung
inflammation: PM-OC alone did not
significantly increase the infiltration of
neutrophils, but LPS challenge showed a
marked increase in the number of neutrophils
compared with vehicle. Administration of LPS
combined with PM-OC significantly increased
the infiltration of neutrophils compared with
LPS administration alone.
Effects of PM-OC on histological changes
in the lung: Combined treatment with PM-OC
and LPS resulted in enhanced neutrophilic
inflammation.
Effects of PM-OC on pulmonary edema
related to LPS: LPS group compared with
vehicle group had a significant increase in lung
water. The combined administration of PM-OC
and LPS resulted in further increase in the
lung water compared with LPS administration
alone, however it was not statistically
significant.
Effects of PM-OC on protein expression IL-
1BMIP-1a, MCP-1 and KC: The
concentrations of these molecules were below
the detection limits in the PM-OC group. LPS
treatment significantly increased the protein
levels of these molecules compared with the
vehicle treatment. In the PM-OC + LPS group
all concentrations, particularly KC, were
smaller than in the LPS group.
July 2009
D-142
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:Inoue
et al.(2006,
096720)
Species: Mouse
Gender: Male
Strain: ICR
Age: 6wks
Weight: 29-33g
Carbon black (14 nm PrinteX 90; 56nm
PrinteX 25; Degussa, Dusseldorf,
Germany)
Particle Size: 14nm ¦ 300 m2/g
56nm ¦ 45m2lg
Route: IT Instillation
Dose/Concentration: Vehicle group: PBS at
pH7.4
LPS group: 2.5mg/kg of LPS in vehicle
Nanoparticle groups: 4 mg/kg carbon black
nanoparticles (14nm or 56 nm) in vehicle
LPS + nanoparticle group: combined
adminstration of carbon black and LPS in vehicle
Time to Analysis: 24h after IT administration
Effects of nanoparticles: Nanoparticles
alone increased number of total cells and
neutrophils, but not statistically significant.
LPS exposure significantly increased numbers
for both groups. Nanoparticles and/or LPS
enhance pulmonary edema.
Histological changes: Treatment with
LPS +14 nm nanoparticles marketly enhanced
neutrophil sequestration into the lung
parenchyma compared to LPS alone. LPS+56
nm nanoparticles did not.
Proinflammatory cytokine proteins: IL-1B
level significantly greater for both LPS +
nanoparticles groups. TNF-a was not
significantly altered among the experimental
groups.
Chemokine proteins: Challenge with 14nm
nanoparticles alone elevated the levels of all
chemokines without significance except for
KC. LPS alone and with both nanoparticle
groups caused significant increases in all
chemokines..
Formations of 8-OHdG in lung: LPS plus
nanparticles resulted in intensive expression 8-
OHdG, strongest in LPS + 14nm nanoparticle
Plasma coagulatory changes: PT - no
change for any group. APTT ¦ some change
with LPS and LPS + nanoparticle groups,
fibrinogen level significantly elevated after
LPS and for LPS + 14nm nanoparticle. APC
decrease with LPS (significant) and LPS +
nanoparticle groups. vWF increase with LPS
(significant) and LPS + 14nm (significant).
Reference:Inoue
et al. (2004,
087984)
Species: Mouse
Gender: Male
Strain: ICR
Age: 6wks
Weight: 29-33g
DEPs [4JB-1 type light-duty, four-
cylinder, 2.74 litre Isuzu diesel engine
(Isuzu Automobile Co., Tokyo Japan)]
Washed DEP and DEP-0C ¦ extracted
with dichloromethane
Particle Size: NR
Route: IT instillation
Dose/Concentration: Vehicle group: PBS
Washed DEP group: 4mg/kg of DEP DEP-0C
group: 4mg/kg of DEP-0C
LPS group: 2.5mg/kg of LPS Washed DEP+LPS
group: combined adminstration of washed DEP
+LPS
DEP-0C + LPS group: combined admininstration
of DEP-0C + LPS
Time to Analysis: 4h
COX-1 mRNA: Slightly elevated in both
washed DEP and DEP-0C groups, but slightly
decreased in other groups compared to vehicle
group.
COX-2 mRNA: Slightly increased with DEP-
0C, increased with LPS, washed DEP + LPS
and DEP-0C + LPS groups compared to
vehicle. COX-2 in the DEP-0C + LPS
decreased when compared to the LPS only
group.
Pulmonary Oedema: Washed DEP + LPS
group showed a synergistic enhancement of
pulmonary oedema and local expression of
proinflammatory chemokines (MCP-1, MIP-1a,
KC, IL-1B).
Reference:Inoue
et al. (2006,
190142)
Species: Mouse
Gender: Male
Strain: ICR
Age: 6-7wks
Weight: 29-33g
Carbon black (14 nm PrinteX 90; 56nm
PrinteX 25; Degussa, Dusseldorf,
Germany)
Particle Size: 14 nm ¦ 300 m2/g
56nm ¦ 45 m2/g
Route: IT instillation
Dose/Concentration: Vehicle group: PBS
Ovalbumin (OVA) group: 1mg OVA Nanoparticle
groups: 50mg carbon black nanoparticles (14nm
or 56nm)
OVA + nanoparticle group: combined
adminstration of nanoparticles and OVA
Time to Analysis: Vehicle group - weekly for
6wks
OVA group ¦ biweekly for 6wks
Nanoparticle groups - weekly for 6wks
OVA+Nanoparticle group (same protocol as OVA
and Nanoparticle) studied 24h after last
administration
Nanoparticles: Exposure to carbon
nanoparticles resulted in the lung expression
of TARC, GM-CSF and MIP-1a. The levels wre
higher in the 14nm group compared to the
56nm group.
OVA: In the presence of OVA, nanoparticles
enhanced levels of TARC, GM-CSF, MIP-1a,
IL-2 and IL-10, with the effects seen more
prominently in the 14 nm particles + OVA
group.
July 2009
D-143
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:Inoue
et al. (2005,
188444)
Species: Mouse
Gender: Male
Strain: ICR
Age: 6-7wks
Weight: 29-33g
Carbon black (14 nm PrinteX 90; 56nm
PrinteX 25; Degussa, Dusseldorf,
Germany)
Particle Size: 14 nm ¦ 300 m2/g
56nm ¦ 45 m2lg
Route: IT Instillation
Dose/Concentration: Vehicle group: PBS
Ovalbumin (OVA) group: 1mg OVA Nanoparticle
groups: 50mg carbon black nanoparticles (14nm
or 56 nm)
OVA + nanoparticle group: combined
adminstration of nanoparticles and OVA
Time to Analysis: Vehicle group - weekly for 6
wks
OVA group ¦ biweekly for 6 wks
Nanoparticle groups - weekly for 6 wks
OVA+Nanoparticle group: same protocol as OVA
and Nanoparticle studied 24h after last
adminstration
Nanoparticles + OVA: Nanoparticles given
with OVA enhanced airway inflammation,
characterized by increased eosinophils,
neutrophils, mononuclear cells and goblet
cells. In addition, nanoparticles + OVA
significantly increased local expression of IL-4,
IL-5, eotaxin, IL-13, RANTES, MCP-1 and IL-6.
The formation of 8-OHdG was enhanced by
nanoparticles + OVA.
14nm nanoparticles: All these effects were
more prominent when 14nm nanoparticles
were used. The 14nm nanoparticle + OVA
group significantly raised levels of total IgE
and antigen specific production of lgG1 and
IgE.
Reference:Inoue
et al. (2006,
190142)
Species: Mouse
Gender: Male
Strain: ICR
Age: 6wks
Weight: 29-33g
Whole DE (generated by 4-cylinder,
3.059I, Isuzu diesel engine, Isuzu
automobile, Tokyo, Japan)
LPS
Particle Size: Peak particle size
110nm
Route: Whole-body Exposure Chamber
Dose/Concentration: 0.3mgsootfm3
1.0mgsoot/m3
3.0mg soot/m3
LPS: 125 mg/kg
Time to Analysis: LPS prior to
12h exposure to exhaust
BAL fluid, total cells, neutrophils, protein and
gene levels (MCP-1 and KC) ubcreased
compared to control with LPS, but were
smaller with LPS + DE. Results are suggestive
that short-term exposure to DE does not
exacerbate LPS related lung inflammation.
Reference: Inoue DEPs [4JB-1 type light-duty, four-
et al. (2007,
096702)
Species: Mouse
Gender: Male
Strain: ICR
Age: 6wks
Weight: 29-33 Ig
Cell Type
Splenocytes
cylinder, 2.74 litre Isuzu diesel engine
(Isuzu Automobile Co., Tokyo Japan)]
LPS
Particle Size: PM2.6
Route: Cell Culture
Dose/Concentration: Splenocytes resuspended
to cell density of 1X106 mL-1 and 1000 mL
applied into each of 12-well plate
DEP: 100 mg mL-1
LPS: 1 mg mL-1
LPS(1mg mL-1) + DEP (1,10 or 100 mg mL-1)
Time to Analysis: 72h
Cell viability: No effect.
Mononuclear cell response: Incubation with
DEP alone inhibited basal cytokine production.
LPS significantly increased protein levels of
IFN-g, IL-2, and IL-10 compared to control.
DEP suppressed the LPS-enhanced protein
levels in a dose-dependent manner and
moderately elevated the IL-13 level.
Reference:Inoue
et al. (2007,
096702)
Species: Mouse
Gender: Male
Strain: ICR
Age: 6-7wks
Weight: 20-30g
Carbon nanoparticles (Dusseldorf,
Germany)
OVA (Sigma Chemical, St. Louis, MO)
Particle Size: CB14 (PrinteX
90) - 14nm
CB56 (PrinteX 25) - 56nm
Route: IT Instillation
Dose/Concentration: 50//g and/or 1 fjg OVA in
PBS
Time to Analysis: 1Xfwk for 6wks; sacrifice
24h after last exposure
Lung Responsiveness: Respiratory system
resistance, Newtonian resistance and tissue
dampening were significantly higher in the
nanoparticle + OVA groups. Elastance and
tissue elastance were higher in these groups
but not significantly so. Compliance was
significantly lower in the nanoparticle + OVA
groups compared to the control.
Lung mRNA level for Muc5ac: levels were
significantly higher in nanoparticle + OVA
groups compared to the control.
Reference: Inoue DEP-OC collected from 4JB1 type, Route: IT Instillation
et. al. (2007,
096702)
Species: Mouse
Gender: Male
Strain: ICR
Age: 6-7wks
Weight: 29-34g
light duty, 4 cylinder, 2.74 liter Isuzu
diesel engine, Isuzu Automobile
Company, Tokoyo, Japan)
OVA
Particle Size: DEP - 0.4/jm
Dose/Concentration: 50//g and/or 1 //g OVA in
PBS
Time to Analysis: DEP or DEP-OC w/ or w/o
OVA initially; OVA or vehicle every 2wk for 6wks;
DEP components or vehicle 1X/wk for 6wks;
sacrifice 24h after last instillation
Total respiratory system resistance,
elastance, Newtonian resistance, tissue
damping, tissue elastance displayed general
positive trends and were significantly higher in
OVA and OVA + DEP-OC groups. Compliance
displayed a general negative trend and was
significantly lower in the washed DEP + OVA
group.
Reference: Ito et DEP - generated from 2982-cc common Route: Cell Culture
al. (2006,
088391)
Species: Rat
Cell Line: L2 cells
of alveolar
epithelial cell type
II origin
rail direct injection diesel engine with
oxidation catalyst and exahust gas
recirculation system.
Particle Size: PM2.6
Dose/Concentration: 1X106
1,10 or 30mg/mL
Time to Analysis: 3h
ICAM-1 and LDL receptor mRNA: Lip-
regulation in a dose-dependent manner.
Statistically significant at 30 mgfmL
compared to control.
HO-1 andPAF receptor mRNA: Up-regulation
in dose-dependent manner and statistically
significant at all doses compared to control.
Correlation between HO-1 and ICAM-1,
LDL, and PAF: Significant correlation
between HO-1 and each of these.
July 2009
D-144
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Jang
et al. (2005,
155313)
Species: Mouse
Gender: Female
Strains: BALB/c
Aqe: 5-6wks
DEP -generated from 4JB1 type, light
duty, four-cylinder diesel engine (Isuzu
Automobile, Co,Tokyo Japan)
Ozone - (generated with Sander Model
50 ozonizers,Sander, Eltze Germany)
OVA
Particle Size: NR
Route: Whole-body Exposure Chamber
Dose/Concentration: DEP: 2,000/ygl/yL (sic)
Ozone: 2 ppm (ave 1.98 ± 0.08 ppm)
OVA sensitization: 10 mg
OVA challenge:
Time to Analysis: OVA sensitization
DEP, Ozone and OVA Challenge on d21- 23
Exposed to ozone for 3h and DEP for 1h
AH and BAL measured 1 d after last challenge
Airway responsiveness: OVA + ozone +
DEP exposure group had significantly higher
methacholine-induce Ptnh than sham group or
OVA group.
Total cells, proportion of eosinophils and
neutrophils: the OVA + ozone + DEP group
was significantly higher than than OVA group
and 0VA+ Ozone group
IL-4: OVA + ozone, OVA + DEP and OVA +
ozone + DEP IL-4 level increased compare to
OVA group.
IFN-n: levels significantly decreased in OVA +
DEP and OVA + ozone + DEP compared to
OVA + ozone
Reference:
Jaspers et al.
(2005,088115)
Species: Human
Cell Lines: A549
cells, primary
human bronchial
and nasal
epithelial ells
DEas: aqueous-trapped solution of DE Route: Cell Culture. DEas and virus added to
(emissions from Caterpiller diesel
engine, model 3304)
Influenza: A/Bangkok/1/79 (H3PJ2
serotype)
Particle Size: NR
apical side of cells
Dose/Concentration: Influenza: 3 x 106 cells
infected with 320 hemagglutination units (HAU)
DEas: For A549 cells: 6.25,12.5, 25/yglcm2. For
bronchial and nasal cells: 22 or 44 //gfcm2.
Time to Analysis: 2h incubation with DEas then
virus added.
HA RNA levels analyzed at 0,15, 30, 60 or
120min post infection.
IFN and MxA responses: analyzed 24h post
infection.
Fluorescence: some cells treated with GSH-ET 30
min before DEasexposure. Measured 2 h post-
influenza infection.
A549 cells increased susceptibility: DEas
enhances HA RNA levels in A549 cells in a
dose-dependent manner. 25 mg/cm2
significantly enhanced levels in A549 cells
compared to the influenza-infected controls.
Viral protein levels were increased in A549
cells. Exposure to DEas increased the number
of influenza-infected epithelial cells in A549
cells.
Human nasal and bronchial cells
susceptibility: Exposure to DEas increased HA
RNA levels in the nasal and bronchial cells.
Statistically significant at 22mgfcm2 for nasal
cells and approaching significance at 44
mgfcm2 for bronchial cells. Exposure of both
types to 44 mg/cm2 enhanced viral protein
levels.
Influenza induced IFN response in A549:
Exposure to DEas does not suppress but
enhances IFNBmRNA levels. Treatment
enhanced influenza-induced nuclear levels of
both phospho-STAT-1 and ISFG3g. ISRE-
promoter activity was enhanced, but not
significantly. Treatment enhanced myxovirus
resistance protein (MxA) mRNA levels. This
data suggest that DEas exposure enhances
influenza virus replication without suppressing
production of IFNB or IFNB-inducible genes.
Influenza induced IFN response in human
nasal and bronchial cells: Exposure to DEas
increased IFNB and MxA levels.
Oxidative stress in A549: DEas exposure
dose-dependently increases oxidative stress in
A549 cells within 2-h post-exposure. Add the
antioxidant GSH-ET and it reverses the effect.
Pretreatment with GSH-ET A549 cells
reversed the effects of DEas on the number of
influenza-infected cells, and reduced HA RNA
levels.
Oxidative stress in Human bronchial cells:
The results were the same as A549 cells
pretreated with GSH-ET. Or Pretreatment with
GSH-ET also reversed effects of DEas on HA
RNA levels.
July 2009
D-145
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Kaan
and Hegele (2003,
095753)
Species: Guinea
pig
Gender: Female
Strain: Cam
Hartley
Age: 22-29d
Weight: 250-
300g
Cell Types
Alveolar
macrophages
(AM) obtained by
bronchoalveolar
PMio ¦ EHC-93 obtained (Environmental
Health Canada, Ottawa, ON, Canada)
RSV ¦ Human RSV (long strain/lot18D)
(American Tissue Culture Collection,
Bethesda, MD)
Particle Size: PMio (0.35/ym MMAD)
Route: Cell Culture
Dose/Concentration: PMio: 500//l/well (100
/yg/ml MEM)
RSV exposure:: 1 ml/well (6x106pfu/ml MEM)
Groups: PMio+RSV
RSV+PMio
RSV only
PMio only
negative control
Time to Analysis: PMio - 60min
RSV ¦ 90min
Parameters measured 24h post treatment
Interaction on phagocytic ability of AM:
Not affected by sequential exposure to RSV
and PMio. More than 95% of AM exposed to
PMio engulfed PM. AM exposed to PMio
showed significant increase in mean side
scatter in comparison to negative control and
RSV-infected AM. No significant difference
between AM exposed only to PMio and AM
exposed to both agents. No significant side
mean side scatter difference between AM
exposed to PM only and to both agents.
Interaction on RSV Immunopositivity:
PMio exposure inhibits. All RSV-treated groups
showed significantly greater proportion of
RSV'immunopositive cells compared with
negative control. PMio+RSV showed
significantly smaller proportion of RSV-
immunopositive cells compared with RSV
group. RSV+PMio group similar to RSV group.
Proportion of RSV-immunopositive AM was
influenced by the sequence of exposure to
RSV and PMio.
Interaction on RSV Replication: PM
exposure suppressed RSV replication. AM
exposed to both agents produced 3 to 9 fold
less RSV progeny compared with RSV alone
group. Quantity of RSV progency was not
significantly affected by the sequence of
exposure RSV and PMio. Negative control and
PMio only did not propagate progeny.
Interaction of RSV Yield: RSV alone group
produced the highest RSV yield, those exposed
to both agents, independent of sequence,
showed a 5-fold decrease.
Cytokine production: RSV infection
stimulated all three cytokines measure (IL-6,
IL-8 and TNF-a) compared to negative control.
IL-6: PMio significantly reduced RSV-induced
IL-6 production. IL-6 was affected by the
sequence of exposure to PMio and RSV
(PMio+RSV vs. RSV+ PMio, p < 3x10-6). IL-
8: PMio significantly decreases RSV-induced
IL-8 production and baseline. No affect on
sequence of exposure. TNF-a: production was
increased when exposed to RSV, PMio or a
combination of both agents. No differences
among treatments.
Reference:
Kleinman et al.
(2005, 087880)
Species: Mouse
Gender: Male
Strains: BALB/c
Age: 8-19wks
Weight: NR
CAPS: Concentrated fine (F) and
ultrafine (UF) Ambient Particles using
VACES system (performed a 2 sites in
Los Angeles, CA, on 50-m downwind
and another 150-m downwind from a
complex of three roadways, State
Road CA60, Interstate 10, and
Interstate 5.)
F CAPS in 2001 and 2002, UF CAPS in
2002 only
OVA: Ovalbumin (Sigma, St. Louis, MO)
Particle Size: UF: dp ~ 150 nm
F: dp ~ 2.5 /jm
Route: CAPS: inhalation via Whole-body chamber
exposure
OVA sensitization: nasal instillation
OVA challenge: inhalation
Dose/Concentration: UF at 50 m: 433 mgfm3
-UF at 150 m: 283 mg/m3
F at 50m or 150 m: average 400 /yg /m3
OVA sensitization: 50 /yg/5 /j\
OVA challenge: 30 mg/m3
Time to Analysis: CAPS: 4 h/d, 5 dfwk for
2wks
Sensitization: On morning of each exposure
1st Challenge: week after 10d of treatment
2nd Challenge: one week following 1st challenge
Sacrificed: 24h after 2nd challenge
There were significantly higher concentrations
of IL-5, IgE, IgG 1 and eosinophils in mice
exposed to either CAPS compared to air. Mice
exposed to CAPS at 50-m downwind showed
higher levels of IL-5, lgG1, and eosinophils
than those exposed to CAPS 150-m
downwind.
July 2009
D-146
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Kleinman et al.
(2007,097082)
Species: Mouse
Gender: NR
Strains: BALB/c
Aqe: 6-8wks
CAPS ¦ concentrated fine (F) and
ultrafine (UF) using VACES system ¦
performed a 2 sites in Los Angeles, CA,
on 50 m downwind and another 150 m
downwind from State Road CA60 and
Interstate 5. Fall 2001-summer 2004
OVA
Particle Size: F: PM2.6; UF: PMo.ib
Route: Whole-body Chamber
Dose/Concentration: 50m ¦ F: 394 ± 94 mg/m3
50m-UF: 297 ± 189 mglm3
150m - F: 387 ±68 mglm3
150m ¦ UF: 213 ±95 mglm3
OVA ¦ 50 mg in 5 mL saline
Time to Analysis: 3 2wk exposures (4 h|d,
5d|wk for 2 wks.
OVA the morning of each exposure
50m Site: higher levels and statistically
significant concentration curves of IL-5 and
lgG1 in F-CAP mice at the 50m site.
150m Site: in no cases were responses
greater than the 50m or control groups.
F vs. UF: The study was not able to
differentiate between the effects of F PM and
UF PM exposures.
Reference: Klein-
Patel et al. (2006,
097092)
Species: Cattle
and Human
Cell Types
Bovine tracheal
epithelial cells
(BTE) and Human
A 549 cells
ROFA (U.S. EPA NHEERL, Research
Triangle Park, NC)
V2O1,, V0S04, Si02 TiOz, Fe2(S04)s
NiS04LPS
Recombinant human TNF-a and IL-1B
Particle Size:
MMAD- 1.95mm
Route: Cell Culture
Dose/Concentration: 2x106
ROFA: 0, 2.5, 5,10,15, 20//gfcm2
LPS: 10Ong/mL
V205: 0, 0.15, 0.3, 0.61,1.25, 2.5, 5,10, 20
/yg/cm2
NiS04, Fe2(S04)s, Ti02, Si02: 0,1.23, 2.5, 5,10,
20/yg/cm2
V0S04: 0, 0.145, 0.29, 0.58,1.16, 2.32//gfcm2
Time to Analysis: LPS: 0, 6, or 18h
ROFA: 0, 2,4,6h
V205: 0,0.25,0.5,1,2,4, 6, 8h
NiS04, Fe2(S04)3, Ti02, Si02: 6h
V0S04: 6h
ROFA in BTE: ROFA and ROFA leachate
inhibition of LPS-induced TAP gene expression
increases with exposure time and dose.
Washed particles of ROFA at doses 2.5 to 10
mgfcm2 significantly increased inducible TAP
expression.
Soluble Metals in BTE: V2OE inhibition of
LPS and IL-1 B induced TAP gene expression
increases with exposure time and dose. NiS04
exhibits non-significant dose dependent
suppression of inducible TAP gene expression.
Fe2(S04)3, Ti02 and Si02 were found to have
no effect.
A549: Results with ROFA and V2O6 in BTE
were replicated using the A549 cell line and
IL_1 B to induce hBD2 gene expression.
Cellular Viability: Was not significantly
affected in ROFA doses below 20/yglcm2 and
V2O6/VOSO4 doses below 2.5 /yglcm .
Reference: Koike
and Kobayashi
(2005, 088303)
Species: Rat
Gender: Male
Strains: Wistar
Age: 8-10 wks
Weight: 280-
350g
Cell Types
AM - aveolar
macrophages
PBM - peripheral
blood monocytes
T-cells: antigen
sensitized
Whole DEP: Diesel Exhaust Particles
collected in the dilution tunnel of a
diesel inhalation facility. (Ratio of
organic extract to residual particles in
the whole DEP was 3: 1.)
Organic extract of DEP
Residual particles of DEP
OVA: Ovalbumin (Sigma, St. Louis, MO)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Whole DEP: 10, 30,100
/yg/mL
Organic extract of DEP: 7.5, 22.5, 75/yglmL
Residual particles: 2.5, 7.5, 25/yglmL
Time to Analysis: 24h post exposure
la antigen and costimulatory molecules:
Most control AM did not express these
molecules. Whole DEP did not cause any
increase in expression level. 20% of control
PBM expressed la and 10% B7; expression of
these molecules was significantly increased by
whole DEP. Organic extract significantly
increased the expression of la and B7
molecules on PBM similar to whole DEP.
Residuals caused no effect.
PBM exposure to 100/75/25 //g/mL of
DEP/organic extract/residual particles:
Results showed that the DEP-increased
expression of la and B7 in PBM by DEP was
caused by the organic extract fraction.
Organic extract-induced expression of la
antigen in PBM was reduced by treatment
with NAC.
AP activity of PBM: After exposure to
organic extract, T cell proliferation was
significantly increased by the addition of
control PBM in a cell number-dependent
manner. AP activity of PBM was increased
over control by exposure to 3 mgfmL organic
extract, although higher concentrations
suppressed the activity of PBM.
Ctyokine production: Organic extract
treatment of PBM decrease IFNn production
from T-cells stimulated by PBM. No significant
effect on IL-4 observed.
HO-1 protein level: levels in PBM was
significantly increased by exposure to whole
DEP or organic extract. Levels induced by
organic extract was diminished by NAC
treatment.
July 2009
D-147
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:Last
et al. (2004,
097334)
Species: Mouse
Gender: NR
Strains: BALB/c
Age: 6wks
Weight: 16-20 g
PM - aerosol of soot and iron oxide
OVA (ovalbumin, grade V, 98% pure,)
Particle Size: PMoi- PM2.6
Route: PM Exposure Chamber
OVA - Intraperitoneal Injections; Aerosol Exposure
Dose/Concentration: PM - 235 - 256/yglm3
OVA ¦ 10 /yg/0.1 mL injection
OVA aerosol — 10mL of 10mg/mL (1 %) solution
Time to Analysis: PM: 4hfd, 3 dfwk
OVA: 2 ip injections days 1 and 15. Aerosol on
day 28 after first ip; 60 min 3x/wk
2wl< PM exposure/4wl< OVA aerosol
treatment: The OVA alone group had
significantly more airway collagen than the
PM alone group. Histology showed
significantly more collagen in the treatment
than the air alone group. There was a
significantly greater amount of goblet cells
than the OVA alone group.
4wl< OVA aerosol/2wl< PM treatment: The
treatment had significantly more goblet cells
than the PM alone group.
6wl< concurrent PM and OVA treatment:
Significantly more cells were observed in the
OVA alone group over the treatment. The
treatment had significantly more lymphocytes
and significantly less macrophates than
groups exposed to PM before or after OVA.
Histology showed significantly more collagen
in the treatment than the air or PM alone
groups. The treatment had significantly more
goblet cells than the OVA alone group.
Reference: Li et
al. (2007,
093156)
Species: Mouse
Gender: Female
Strain: BALB/c,
C57BL/6
Age: 9wks
Weight: NR
DEP (2369-cc diesel engine
manufactured by Isuzu Motor, operated
at 1050 rpm, 80% load, commercial
light oil)
Particle Size: NR
Route: Exposure Chamber
Dose/Concentration: DEP: 103.1 ±9.2/yglm3,
CO: 3.5±0.1ppm, NO2: 2.2±0.3ppm, SO2:
< O.OIppm
Time to Analysis: Protocol 1: Exposed 7hfd,
5d/wk. Sacrificed at day 0, week 1,4, 8.
Protocol 2: DE alone or DE+NAC 7h/d, 1 -5d.
Airway hyperresponsiveness: Penh values
increased in BALB/c mice compared to the
control at day 0, but no significant changes
occurred after this time. Penh values
increased in C57BL/6 mice at 1wk compared
to the control but returned to control levels at
8wks.
BALF: Compared to the other strain, the total
number of cells and macrophages increased
significantly at 1wk in C57BL/6 mice and at
8wks in BALB/c mice. Neutrophils,
lymphocytes, MCP-1, IL-12, IL-10, IL-4, IL-13
increased significantly for both strains. No
eosinophils were found. IL-1B and IFN-n
increased significantly in BALB/c mice
compared to C57BL/6 mice.
HO-1 mRNA and protein: HO-1 mRNA was
more marked in BALB/c mice at 1wk and
C57BL/6 mice at 4 and 8wks. HO-1 protein
percentage changes from the control were
greater in BALB/c mice at 1wk and C57BL/c
mice at 8wks.
NAC: NAC inhibited the increased Penh
values, total number of cells and macrophages
in C57BL/6 mice at 1wk and neutrophils and
lymphocytes in both strains.
Reference: Li et
al. (2009,
190424)
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 6-8wks
Weight: NR
CAPs (downtown Los Angeles, CA
from major freeway, traffic mainly
passenger cars and diesel trucks; Jan.
2007 or Sept. 2006)
Ultrafine carbon black (UFCB; used as
control)
Particle Size: Diameter: Fine- < 2.5
/jm, UF- < 0.15 /jm
Route: Intranasal Instillation
Dose/Concentration: 0.5/yg PM in 50 fjL
suspension
Time to Analysis: Day 1 exposed to PM or
saline. Day 2 exposed to PM+OVA or OVA or
saline alone. Repeated on days 4, 7, 9. Different
experiment: NAC ip injected 4h pre-instillation on
days 1, 2, 4, 7, 9. All animals rested and OVA
aerosol challenged 30min on days 21, 22.
Sacrificed day 23.
UFP alone had no effect on the lung.
UFP+OVA significantly increased eosinophils,
and OVA-specific lgG1 and IgE. The induction
of eosinophils and lgG1 were inhibited by
NAC. Generally, UFP+OVA mice had greater
signs of inflammation than the other groups as
determined by pulmonary histopathology and
airway morphometry. UFP had a greater PAH
content than fine particles. UFP significantly
increased IL-5, IL-13, TNF-a, IL-6, KC, MCP-1,
and MIP-1a.
Reference: Li et
al. (2009,
190424)
Species: Mouse
Cell Line:
Macrophage RAW
264.7
CAPs (downtown Los Angeles, CA
from major freeway, traffic mainly
passenger cars and diesel trucks; Jan.
2007 or Sept. 2006)
Ultrafine carbon black (UFCB; used as
control)
Route: Cell Culture
Dose/Concentration: 1, 5, 8.3,10/yg/mL
Time to Analysis: Particle suspensions
reconstituted with Dulbecco's Modified Eagle's
Medium powder. Further replenished with 10%
fetal calf serum and 1:200 dilution of
fjm, UF- < 0.15 fjm
16h. 100/yg of lysate protein for HO-1
immunoblotting.
UFP induced greater HO-1 expression than fine
particles. The higher PAH content of UFP
correlated with HO-1 expression.
Particle Size: Diameter: Fine- < 2.5 penicillin/streptomycin/amphotericin B. Incubated
July 2009
D-148
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Liu et al. (2008,
156709)
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 11wks
Weight: NR
DEP (5500-watt single-cylinder diesel
engine generator (Yanmar, Model YDG
5500E), 406 cc displacement air-
cooled engine, Number 2 Diesel
Certification Fuel, 40 weight motor oil)
Particle Size: MMAD: —0.1 /jm
Route: Intranasal Exposure
Dose/Concentration: Average particle
concentration: 1.28mg/m3
Time to Analysis: Four groups: saline+air
control, saline+DEP, A fumigatus+air,
A. fumigatus+U£P. A. fumigatus exposure every
4d for 6 doses. DEP exposure 5h/d for 3wks
concurrent with A. fumigatus exposure.
A. fumigatus+DEP increased IgE, the mean
BAL eosinophil percentage, goblet cell
hyperplasia, and eosinophilic and mononuclear
cell inflammatory infiltrate around the airways
and blood vessels compared to the A.
fumigatus or DEP treatments.
A.fumigatus+BtP also caused methylation at
the IFN-n promoter sites CpG'63, CpG46, and
CpG'206.
Reference: Liu et
al. (2007,
093093)
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 11wks
DEP: 5500-watt single-cylinder diesel
engine.
Particle Size: NR
Route: Inhalation Exposure
Dose/Concentration: Average particle
concentration 1.28 mg/m3.
Time to Analysis: 1. Aerosol vehicle (saline) +
air
2.	Aerosol vehicle (saline) + DEP
3.	A. fumigatus + air
4.	A. fumigatus + DEP
A. fumigatus-. 62.5 /yg aerosolized protein extract
in 50 fjl PBS; 6 total doses, every 4d.
DEP exposure 5h/d 3wks concurrent with A.
fumigatus.
IgE Production: IgE production increased
withtheA/tf/n/^atetreatmen and increased
further with the A.fumigatus and DEP
treatment.
Histopathology: A. fumigatus with DEP
caused an increase in goblet cell hyperplasia
and eosinophil and mononuclear cell infiltrate
around the airways and blood vessels as
compared to the control and DEP treatments.
Gene Methylation: Greater methylation at
the CpG'63 site of the IFN-n promoter occurred
under the A. fumigatus + DEP treatment
compared to the A. fumigatus or DEP
treatments. The DEP treatment did not induce
methylation. Methylation correlated with
increased IgE and hypomethylation with
decreased IgE. Hypomethylation occurred in
the IL-4 promoter under the A. fumigatus +
DEP treatment.
Reference:
Lundborg et al.
(2007, 096040)
Species: Rat
Gender: Male
Strains:Sprague-
Dawley
Age: NR
Weight: 300-
400g
difference existed with the washed diesel
particles.
Survival of the ATCC strain and clinical
isolates of S. pneumoniae when incubated
with carbon loaded AM or control AM:
Carbon significantly increased the CFU of
opsonized and unopsonized bacteria for the
ATCC strain and clinical isolates.
Ability of carbon or washed diesel loaded
AM, incubated with the ATCC strain of S.
pneumoniae, to induce LPO of lung
surfactant: A 97% increase in the surfactant
LPO occurred after incubation with washed
diesel loaded AM compared to contol AM. The
effect of washed diesel particles was
significantly greater than that of carbon
particles.
LPO by carbon loaded AM incubated with
the ATCC strain or clinical isolates in the
presence of absence of surfactant: LPO
induced by AM increased when incubated with
carbon loaded AM compared to control AM.
Carbon-Black Particles (93% C)
Toluene- DEPs (97% C)
10-fold Cr, Mn, N; 50-100 fold Al, Cd,
Cu, Fe, Mg, Pb, Zn more in DEP
aggregates
Particle Size: Carbon aggregates-
mean diameter - 0.17 ± 0.08um
Diesel
Particles- mean diameter: 0.69 ±
0.46/ym
mean size primary particles: 0.044 ±
0.01 fjm
Route: Cell Culture{Citation}
Dose/Concentration: 2 X 10I!AM
A 2mL suspension was prepared with 20 //gfmL
of carbon or diesel particles added.
surface area: 159 ± 4m2lg
Time to Analysis: 6 different experiments. AM
pre-exposed to carbon or washed DEP. Loaded
with particles. Incubated with S. pneumoniae,
ATCC strain or clinical isolates.
Effect of time on survival of S.
pneumoniae when incubated with carbon
loaded AM: Loading AM with carbon
significantly increased the bacterial survival.
Bactera opsonization decreased bacterial
survival.
Effect of carbon load in AM on survival of
S. pneumoniae-. Bacterial survival increased
in a dose-dependent manner as the carbon
particle load of AM increased.
Survival of S. pneumoniae after
incubation with carbon or washed diesel
loaded AM: Bacterial survival increased in
carbon loaded AM compared to the control. No
July 2009
D-149
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Matsumoto et al.
(2006,189213)
Species: Mouse
Gender: Female
Strains: BALB/c
Age: 6wks
Weight: 15-20g
DE
DE collected from a 2369 cm3 diesel
engine operated at 1050 rpm and 80%
load with commercial light oil. Engine
exhaust passed through a particulate
air filter and charcoal filer. Diluted DE
introduced into the exposure chamber.
Composition of the DE: 3.5±0.1ppm
CO, 2.2±0.3ppm NO2, < 0.01 ppm SO2
and 103.1 ±9.2//g/m3 DEP.
Particle Size: NR
Route: Whole-body exposure chambers after
prior sensitization with OVA through ip injection
Dose/Concentration: 100//gfm:l DE
Time to Analysis: Mice were initially sensitized
wf OVA (20ug absorbed to 2 mg alum diluted
with 0.5 mL saline) via ip injection on day 0, 6
and 7. Two weeks later the mice were challenged
with OVA (0.1 mg in 0.1mL saline) intranasally on
day 21.
DE for 1d or 1,2, 3, 4 or 8wks (at 7h/d for
5d/wk).
Airway Hyper-Responsiveness: Airway
Reactivity (evaluated by an increase in PenH)-
Exposure to DE significantly increased airway
reactivity to methacholine after 1 week in
both 24 and 48 mg/mL Mch and after 4 weeks
in the 48 mg/mL. Airway Sensitivity
(evaluated by PC200)- DE exposure caused an
increase in airway sensitivity after 1 week of
exposure, 4 weeks and 8 weeks of exposure
did not result in a significant increase.
BAL Cell Counts: The total cell count was
increased after 1 week of DE exposure. This
increase was mostly due to an increase in
eosinophils. After 1 week the total cell count
dropped drastically even after continuous
exposure to DE. DE did not effect the number
of CD3, CD4, CD8 or NK1 cells at any point in
time.
Cytokine/Chemokine mRNA Levels: DE
exposure on day 1 caused an increase in
mRNA levels of IL-4, IL-5 and IL-13 when
compared to the control mice but longer
periods of DE exposure failed to cause an
increase. Protein levels of IL-4 were
significantly elevated at compared to control
at day 1, but did not persist with time. mRNA
levels of MDC were increased at 1 week of
exposure (compared to control) but also
decreased at time periods after. mRNA levels
of RANTES were increased at 2 and 3 weeks
after exposure and remained elevated at 4
weeks but not significantly. The level of
RANTES protein increased as the weeks went
along, but increased significantly only at 8
weeks. Lung
Histopathology: OVA sensitization caused an
increase peribronchial and perivascular
infiltration of inflammatory cells which peaked
at 1 week after exposure and decreased
afterward. DE exposure did not causefshow
any additional signs of inflammation.
Reference:
Morishita et al.
(2004, 087979)
Species: Rat
Gender: Male
Strain: Brown-
Norway
Age: 10-12wks
Concentrated Air Particles (CAPs)
CAPs were generated from ambient air
in an urban Detriot community. The
sampling site was chosen because it
has a high percentage of pediatric
asthma and is located near a lot of
industry.
Particle Size: 0.1-2.5 /jm
Route: Whole-body Inhalation Chambers
Dose/Concentration: Air chamber received
CAPs at a flow rate of 50 L/min and at a
pressure of 0.94-0.95 atm.
July 676/yg/m3
September 313/yg/m3
Time to Analysis: First rats were sensitized
(days 1-3) and challenged (days 14-16) with
saline (control) or ovalbumin by intranasal
instillation (5% in saline, 150 //Lfnasal passage).
4 days after the last intranasal challenge, rats
began exposure in the chambers. Exposures were
10 h long. There were two separate exposure
periods in July and September. The July exposure
was for 4 consecutive days. The September
exposure was for 5 consecutive days.
Recovery of Trace Elements in Animal
Lung Tissues: July Exposure- Anthropogenic
trace elements were below limit of detection
in pulmonary tissue of animals exposed to July
CAPs. September Exposure- Several elements
were recovered from pulmonary tissue during
the Sept. exposure. La concentrations were
increased in both control/CAPs exposure and
in the OvafCAPs exposure groups. V
concentration was increased in OvafCAPs
exposed animals but not in rats exposed to
just CAPs. S content was only significant in
animals exposed to OvafCAPs compared to the
non-exposed control.
Particle Characterization: July PM had an
average mass concentration twice as high as
the September mass concentration. S
concentration was four-folds higher in July
PM. In the September PM- the concentration
of La was 12.5 fold higher than in July PM, V
was 2.7 fold higher than in July PM and Mn
was 1.5 fold higher than in July PM.
BALF Analysis: Eosinophil concentration was
not significantly different when comparing
rats exposed to CAPs only in either July or
September (this was explained by the elapsed
time between exposure and BALF collection).
However OVA and CAP exposure in the
September group led to elevated eosinophil
levels. Similarly, the protein content was only
significantly increased in the September
OVA/CAP exposed rats, compared to the
control group.
July 2009
D-150
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Nygaard et al.
(2005, 088655)
Species: Mouse
Gender: Female
Strain: BALB/c
Aqe: 6-7 weeks
Coarse and fine ambient air particles
collected in Rome (spring), Oslo (1-
summer, fine only, 2- following spring,
fine and coarse), Lodz (summer) and
Amsterdam (spring). These represent
areas with high population and
dominance of traffic.
DEP (Standard reference material
1650a) used as a positive control.
Particle Size: Fine PM0.1 ¦ 2.5 /m;
Coarse PM 2.5 ¦ 10/ym
Route: Subcutaneous Injection into mouse
footpads.
Dose/Concentration: 100/yg of particle
Time to Analysis: Animals were in eight groups:
1.	Control- Hank's Balanced Salt Solution
2.	OVA- 50ug 3. OVA (50ug)+ Amsterdam
Coarse PM (100/yg) 4. OVA (50ug)+ Amsterdam
Fine PM (100/yg) 5. OVA (50ug)+ Lodz Coarse
PM (100 //g) 6. OVA (50ug) + Lodz Fine PM (100
/yg) 7. 5. OVA (50ug)+ Oslo Coarse PM (100/yg)
8. OVA (50ug)+ Oslo Fine PM (100/yg)
Analysis 5d after injection
Cell Numbers and Cell Phenotypes in the
Lymph Node: The overall number of B
lymphocytes, lymph node cells, PLN cells, and
the expression of MHC class II, CD86 and
CD23 on B lymphocytes were increased by
coexposure of 0VA + the particles compared
to the OVA or particle groups alone. The OVA
+ particle groups displayed a significant
decrese in T lymphocytes. Particles only
significantly increased the number of lymph
node cells and MHC Class II expression. There
were no differences observed between coarse
and fine PM fractions.
Cytokine Production by Lymph Node (ex
vivo culture of popliteal lymph node cells):
The OVA + particle (DEP and Oslol only)
significantly increased IL-4 and IL-10 levels.
No change was observed in IFIM-d. The particle
groups only increased IL-4 and IL-10. All
coarse and fine particle fractions co-exposed
with OVA significantly increased IL-4 and IL-
10 compared to OVA alone. There was no
significant difference between coarse and fine
particles. IFN-n levels were not significantly
affected by most of the groups, but the fine
fractions of PM consistently produced higher
levels of IFN-n.
Lymph Node Histology: OVA + particle
groups resulted in significantly enlarged lymph
nodes and the formation of germinal centers.
Reference:
Nygaard et al.
(2005, 087980)
Species: Mouse
Gender: Female
Strains: BALB/c
Aqe: 6-8wks
Polysterene Particles (PSP
Particle Size): PSP diameter: 0.1 /ym
Route: Subcutaneous Injection into footpads.
Dose/Concentration: 40/yg PSP (5.94 X 10,G
particles) per injection suspended in HBSS. One
injection per footpad
Time to Analysis: 1. HBSS
2.	OVA (10 /yg per injection)
3.	PSP (40 /yg per injection)
4.	OVA (10/yg per injection) + PSP (40 /yg per
injection).
Antibody experiments: reinjected with 10 /yg
OVA on day 21. Killed on day 26.
Popliteal lymph node cell experiments - animals
injected. Killed 1 to 21 d post-injection.
OVA-specific IgE, IgG 1 and lgG2a
Antibodies: Analysis at day 26 indicated IgE,
IgG 1 levels were significantly higher in mice
exposed to OVA+PSP compared to mice
injected with HBSS, OVA or PSP. No
significant difference was observed for lgG2a
levels.
Number of Particle Containing Cells: There
was no significant difference between PSP
alone and OVA+PSP. Throughout days 0 -21
the number of particle-containing cells in the
PSP or 0VA+ PSP groups were significantly
greater than the HBSS group.
Total Cell Numbers, B and T Lymphocytes
and MHC class II Expression: The total cell
number and B lymphocytes significantly
increased by coexposure to OVA + PSP when
compared to the other groups. Both OVA and
OVA+PSP increased T lymphocytes on Days
1, 3 and 5. MCH class II expression was
significantly higher in the OVA +PSP group on
days 5, 7 and 21 than other groups.
Cell Types and Surface Markers: The
number of CD40 + B Lymphocytes showed a
slight but significant decrease with OVA+PSP
and OVA compared to HBSS and PSP.
CD86+, CD23+ and CD69+ B lymphocytes
were significantly higher in OVA +PSP group
than other groups. PSP alone did not affect
CD86+ or CD23+ levels.
Cytokine Production: IL-4 and IL-10 were
significantly higher in the OVA +PSP group
when compared to the other groups. OVA
alone caused a slight increase compared to
PSP. PSP did not alter IL-4 or IL-10 levels.
July 2009
D-151
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Nygaard et al.
(2004, 058558)
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 6-7wks
Weight: NR
CB (carbon blackfDEP)
Polystrene Particles (PSP)
Particle Size: PSP diameter: 0.0588,
0.202,1.053, 4.64 or11.14//m
Route: Single subcutaneous injection into
footpad
Dose/Concentration: 10//g OVA + 40//g (low
dose) or 200 fjg (high dose) of particles
Time to Analysis: Mice killed 5d after OVA
injection; PLN excised
OVA Specific IgE and Ig2a: OVA with CB,
DEP or PSP of diameters 0.0588 and 0.202
fjm increased IgE compared to OVA alone, as
well as the 1.053, 4.64 and 11.14//m PSP.
OVA with 0.0588 fjm PSP or CB significantly
increased lgG2a compared to OVA alone.
Primary Cellular Response: All OVA and
PSP groups (except the low dose of 11.14 //m
PSP) had more total lymph node cell numbers
than the OVA alone group. The low and high
does groups of 0.202 fjm PSP had the
greatest amount of cell proliferation and
lymphoblasts. The OVA and 0.202 PSP
treatment produced the greatest amounts of B
lymphocytes, IL-4, IL-10 and IFN-n. IL-2 in the
PLN cells was significantly lower in both
dosage groups of OVA and 0.202 PSP than
the OVA control.
Particle mass, size, number and surface
area: Total particle surface area explained
64% of the variance in the IgE levels. 60-80%
variance of the PLN cellular paramters (except
CD23) were explained by total particle surface
area, number and diameter.
Reference: Reed
et al. (2008,
156903)
Species: Mouse
Gender: Male,
Female (only
BALB/c)
Strain: C57BL/6,
A/J, BALB/c
Age: NR
Weight: NR
GEE (two 1996 General Motors 4.3-L
V-6 engines; regular, unleaded, non-
oxygenated, non-reformulated gasoline
blended to US average consumption for
summer 2001 and winter 2001-2002-
Chevron-Phillips)
Particle Size: MMAD: 150nm
Route: Inhalation Exposure Chamber
Dose/Concentration: PM: Low- 6.6±3.7//g|m3,
Medium- 30.3±11.8//gfm3, High- 59.1 ±28.3
Z^g/m3
Time to Analysis: A/J- 2wk quarantine period in
chamber. Exposed 6h/d, 7d/wk, 3d-6m. C57BL/6-
1wk exposure. Instillation of P. aeruginosa. Killed
18h post instillation. BALB/c- Conditioned to
exposure chambers and mated. Pregnant females
exposed GD 1 and throughout gestation.
Offspring exposures continued until 4wks-old.
Half of offspring sensitized to OVA. Tested for
airway reactivity by methacholine challenge 48h
postinstillation and euthanized.
The kidney weight of female A/J mice
decreased at 6m and was strongly related to
PM by the removal of emission PM. PM-
containing macrophages increased by 6m.
Hypomethylation occurred in females at 1wk.
The clearance of P. aeruginosa was
unaffected by exposure. Serum total IgE
significantly and dose-dependently increased.
OVA-specific IgE and lgG1 gave slight
exposure-related evidence but were not
significant.
Reference: Reed
et al. (2008,
156903)
Species: Rat
Gender: Male,
Female
Strain: CDF
(F344)/CrlBR,
SHR
Age: NR
Weight: NR
GEE (two 1996 General Motors 4.3-L
V-6 engines; regular, unleaded, non-
oxygenated, non-reformulated Chevron-
Phillips gasoline, U.S. average
consumption for summer 2001 and
winter 2001-2002)
Particle Size: MMAD: 150nm
Route: Inhalation Exposure Chamber
Dose/Concentration: PM: Low- 6.6±3.7/yglm3,
Medium- 30.3±11.8//gfm3, High- 59.1 ±28.3
Z^g/m3
Time to Analysis: 2wk quarantine period in
chamber. Exposed 6h/d, 7d/wk, 3d-6m. SHR-
surgery to implant telemeter in peritoneal cavity.
4wks recovery. ECG data obtained every 15min
beginning 3d preexposure, 7d exposure, 4d
postexposure.
Organ Weight: At 6m exposure, the heart
weights of male and female rats increased and
male rats' seminal vesicle weight decreased.
Histopathology: PM-containing macrophages
increased by 6m.
Serum Chemistry: Serum alanine
aminotransferase, aspartate
aminotransferase, and phosphorus decreased
in medium and high-exposure females.
Hematology, Clotting Factors: Hematocrit,
red blood cell count, and hemoglobin dose-
dependently increased for both genders at
both time points. Plasma fibrinogen increased
at 1wk in males.
Lung DNA Damage: Hypermethylation
occurred in medium- and high-exposure male
rats at 6m.
BAL: For both genders in the high-exposure
group, LDH and MIP-2 significantly increased
at 6m. R0S decreased at 1wk and 6m.
Generally, the production of hydrogen peroxide
and superoxide decreased in the high-exposure
group and medium- and high-exposure groups,
respectively.
CV effects in SHR: Lipid peroxides were
significantly increased in males in the high
exposure group. TAT complexes decreased in
females in the high exposure group.
Removal of Emission PM: The removal of
emission PM strongly linked PM to increased
seminal vesicle weight, red blood cell counts,
LDH, lipid peroxides, and methylation.
July 2009
D-152
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Roberts et al.
(2007,097623)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age: 10wks
Weight: 250-
300g
R-Total - ROFA (Residual oily fish ash)
Sample
R-Soluble - Soluble fraction of ROFA
R-Chelex - R-Soluble+Chelex
(insoluble resin)
Particle Size: Mean diameter- 2.2//m
Route: IT Instillation
Dose/Concentration: 10mg/kg of body weight
(2.5-3mg)
Time to Analysis: Pre exposure to ROFA
samples on Day 0. Inoculation with 5 x 10'L.
Monocytogenes or saline on day 3. Sacrifice on
days 6, 8,10.
Uninfected groups: Compared to the contols,
the R-total and R-soluble groups had increased
LDH, PMNs, lymphocytes and AMs. The R-
total group had a slight, but significant
increase in IL-6 and the R-soluble group had a
decrease in IL-2.
Infected groups: The R-soluble group had
increased levels of LDH (which also increased
for the R-total group), albumin, BAL cells, NK
cells, PMA-stimulated and zyomason-
stimulated CL compared to all other groups at
various time points. NOx was significantly
elevated in the R-soluble group at early time
points, but in later time points R-soluble and
R-total AMs produced less NOx than the
controls. IL-10 and IL-6 increased in the R-
soluble group, while IL-12, IL-4 and IL-2
decreased. IL-12 also decreased in the R-total
group.
Reference:
Saxena et
al.(2003, 054395)
Species: Mouse
Gender: Female
Strain: C57B1/6J
Age: 18-30wks
Weight: NR
DEPs (standard)
Particle Size: NR
Route: Intrapulmonary Instillation
Dose/Concentration: 10 0 //cj; mouse
Time to Analysis: Pre exposure to 2.5 x 10"
bacillus Calmette-Guerin bacteria (BC G) with or
without coadministration of DEP. Sacrifice 5wks
later.
The BC G + DEP group had four times the BC
G lung load than BC G alone. The load was
significantly greater in other organs in the BC
G + DEP group. Interstitial lymphocytes, T, B
and NK cells were increased in the BC G +
DEP group over the DEP-alone group. DEP
caused no release of NO by AMs, but inhibited
the release of NO in response to IFN-n. Except
for CD8 cells, no increase in IFN-n was seen in
the BC G + DEP group.
Reference:
Schneider et al.
(2005, 088368)
Species: Mouse
Strain: BALB/c
Cell Line: RAW
264.7
macrophage cells
SRM 1648 (greater than 63%
inorganic carbon; 4-7% organic carbon;
Si, S, Fe, Al, K greater than 1 % by
weight; Mg, Pb, Na, Zn, CI, Ti, Cu, As,
Cr, Ba, Br, Mn less than 1%)
Titanium dioxide
Particle Size: Ti02 - 0.3/jm average,
1.0 /jm max
SRM 1648 - 0.4/ym mean diameter
Route: Cell Culture (625,000 cells/cm2 in 96 well
plate)
Dose/Concentration: 62.5//gfcm2
Time to Analysis: Particulate introduction at 0,
7.8,15.6, 31.2, and 62.5 /yglcm2. Measurements
made at 1, 3, 6, and 12h after particulate
introduction.
No significant toxicity was exhibited by SRM
1648. The rate of dye oxidation was
significantly higher in SRM 1648-exposed
cells. SRM 1648 significantly increased
reduced glutathione compared to the control
at the 12-h time point. SRM 1648 increased
GSH and concurrently caused significant PGE2
production compared to the no ester control at
the 6-h and 12-h time points.
Reference:
Schober et al.
(2006,097321)
Species: Human
Gender: Male and
Female
Age: 21 -39yrs,
treatment group;
23-32yrs, control
group
Tissue Type:
Whole blood
samples
PM - organic extracts of airborne
sample
AERex,d - urban aerosol 1 day sample
(total air volume ¦ 1270m3)
AERex6d- urban aerosol 5 day sample
(total air volume ¦ 6230m3)
rBet v 1 (birch pollen allergen 1a,
Biomay, Vienna, Austria)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 100//L heparinized whole
blood
Time to Analysis: Blood stimulated with PBSflL-
3 for 10min. Incubated with rBet v 1 alone or
with AERex for 20min. Ice bath 5min. Incubated
with antibody reagent 20min.
Nine organic compound classes were identified
in AERex,dand AERex6d, with AERex,dhaving
20 times more PAHs. Basophil activation
increased in all treatment groups up to 90%,
with AERex,d being the most pronounced. 5-50
fold lower concentrations of AERex,dwere
needed to achieve the maximal effect on
basophil activation. AERex-induced
enhancement of CD63 upregulation of rBet v
1 in sensitized basophils occurred in a dose-
dependent manner.The AERex-alone treatment
did not affect CD63 expression.
July 2009
D-153
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Shwe
et al.
(2005,111553)
Species: Mouse
Gender: Male
Strains: BALB/c
Age: 8 wks
Weight: NR
CB - carbon black particles (Degussa,
Germany)
CB14:
C: 96.79%
H: 0.19%
N: 0.13%
S: 0.11%
Ash: 0.05%
Others including 0: 2.74%
CB95:
C: 97.98%
H: 0.15%
N: 0.28%
S: 0.46%
Others including 0: 1.14%
Particle Size: CB14 - 14nm primary
particle size (pps)
CB95 - 95nm (pps)
Route: IT instillation
Dose/Concentration: 25,125, or 625/yg in 1
mL saline solution
Time to Analysis: CB14 or CB95 instillation
1/wk for4wks
Sacrifice 24 or 4h after last instillation
BAL cells: In CB14, the total number of BAL
cells increased significantly and dose-
dependently. In CB95, only the 625/yg dose
showed a significant increase.
Cytokine and chemokine: For CB14 and
CB95,125 or 625/yg showed a significant IL-
1B increase in a dose-dependent manner. For
CB14, only the 625 fjg dose showed a
significant IL-6 increase. No difference was
observed in the CB95 group. For CB14, only
larger doses showed a significant TNFa
increase. For CB95, no significant differences
were observed.
In BAL fluid: CCL-2 production was
significantly increased for the 625/yg dose in
both the CB14 and CB95 groups. CCL-3
production was significantly increased for the
larger doses in both the CB14 and CB95
groups.
Splenic Lymphocytes: No significant
differences were detected among the CB14
dosages, except for CD8+. No significant
differences were observed among the various
groups for CB95.
Deposition in lymph nodes: For all dosages,
greater deposition of CB14 than CB95 was
observed.
Chemokine mRNA expression in lungs and
lymph nodes: At 125/yg, significant increases
of CLL-3 mRNA expression was observed for
CB14; for CB95, no differences were
detected.
July 2009
D-154
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:	CAPs: Concentrated Ambient Particles
Sigaud et al.	(Collected from ambient Boston air on
(2007,096100)	Teflon filters.)
Species: Mouse	HO2
Gender: Male	IFNn
Strains: BALB/c	S. pneumoniae (ATCC 6303, American
Age' 8-10wks	Type Culture Collection, Manassas,
Weight: NR
Particle Size: CAPs: <2.5//m
Route: IFNn priming: aerosol
Particle exposure and infection: Intranasal
Instillation
Dose/Concentration: CAPS or TKk 50//g|50
fjl PBS
S. pneumoniae: 106CFU/25/yl saline
Time to Analysis: Primed for 15min
One time particle exposure 3h post priming with
lung RNA analyzed 3, 6, 24h after exposure
Sacrificed 24h after exposure OR one time
infection
Sacrificed 24h infection
Inflammation: Saline-primed and unprimed
mice exposed to CAPs produced a significant
increase in PMNs in the lung (100% more than
mice exposed to Ti02.) Groups primed with
IFNn then exposed to CAPs produced a strong
inflammatory response, a 2.5 increase in
PMNs when compared to the increase caused
by PBS+ CAPs exposure.
Cytokine Levels: IFNn primed and CAPs
exposed groups
lnflammation+ S. pneumo Infection:
Saline-primed and unprimed mice exposed to
CAPs produced a significant increase in PMNs
in the lung (100% more than mice exposed to
Ti02.) Groups primed with IFNy then exposed
to CAPs produced a strong inflammatory
response, a 2.5 increase in PMNs when
compared to the increase caused by
PBS+CAPs exposure.
Cytokine Levels: IFNy primed and CAPs
exposed groups showed a 1.5-fold increase
over the control.
PMNs: Treatment with CAPs enhanced
inflammation, causing a 2-fold increase in
PMN numbers as compared to the infected
control. IFNy+CAPs+S. pneumo produced a
3.5 fold increase compared to the infected
control and a 1.6-fold increase compared to
PBS + CAPs +S.pneumo. Despite increased
numbers of PMNs in the IFNy+CAPs groups,
the lungs were unable to clear the S. pneumo
infection.
Bacterial Load: Control groups showed
efficient clearance of bacteria after infection.
Unprimed, CAPs-treated, infected groups did
not show a decrease in bacterial numbers.
IFNy+CAPs showed a 2.5-fold increase in
bacterial numbers.
Histopathology: Indicated moderate
pneumonia in PBS+CAPs and severe
pneumonia in IFNy+CAPs. The other groups
did not indicate areas of pneumonia.
Bacterial Uptake AM and PMN Cells: In all
the treated groups, the bacterial content in
AMs showed a decrease, with a more marked
decrease in the IFNy+CAPs group, but these
decreases were not statistically significant.
Groups exposed to CAPs showed a
statistically significant decrease in bacterial
uptake by PMNs.
ROS Levels in AM and PMN Cells:
Intracellular ROS significantly increased in AM
cells in the IFNy+CAPs group, approximately
50% greater than controls. In PMNs, iROS
increased 100% in the IFNy+CAPs groups as
compared to the controls.
July 2009
D-155
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Steerenberg et al.
(2004, 087474)
Species: Rat
Gender: Male
Strain: Wistar
Aqe: 6-8wks
Ozone (positive control)
DEP:SRM1650a (NIST, Gaithersburg,
MD
EHC-93: ambient PM (Ottawa, Canada)
L. mono: Listeria monocytogenes
(strain L242/73 type 4B)
Particle Size: Ozone: DEP, EHC-93:
NR
Route: Ozone: Whole-body inhalation chamber
DEP/EHC-93: intranasal droplet:
Dose/Concentration: Ozone: 2mgfm3
DEP/EHC-93: 50/yg (1.0 mg/ml)
L mono: 0.2 or 0.3 ml (5x100PFU/ml) *l have
emailed author regarding correct dose
Time to Analysis: Ozone 24 h/day for 7 days (¦
7d to-1 dR
DEP/EHC-93: 1/day for 7 days (-7d to-1d)
All rats infected on day 0. Sacrificed on days 3,
4, or 5
Body weight: Growth declined for ozone
exposed group while DEP or EHC-93 groups
grew progressively. Exposure to L. mono
caused all groups to decline in weight.
Bacterial Count in the Lung: The number of
bacteria in the lung of those rats exposed to
ozone was significantly greater than those
exposed to saline. No differences in bacteria
number were found for rats exposed to saline,
EHC-93 or DEP at any time.
Bacterial Count in the Spleen: The ozone
exposed group exhibited statistically
significant increases in bacteria numbers
when compared to the saline-treated group.
No differences in bacteria number were found
for rats exposed to saline, EHC-93 or DEP at
any time. Exposure to ozone decreases the
defense of the respiratory tract against L.
mono infection: however, DEP and EHC-93 did
not appear to affect the host defense system
in regards to clearing/fighting L. mono.
Reference: PM: collected from Rome, Oslo, Lodz,
Steerenberg et al. Amsterdam and De Zilk during the
(2005, 088649) spring, summer and winter.
Species: Mouse Rome, Oslo, Lodz and Amsterdam
Gender- Male represent areas with high population
and dominance of traffic. De Zilk,
Strain:	selected as a negative control site, has
BALB/cByJ.ico |ow traffic emissions and natural
Age: 6-8wks allergerns.
EHC-93: used as a positive control (Gift
of Dr. R. Vincent, Ottawa, Canada)
OVA: Ovalbumin (Sigma)
Particle Size: Coarse PM: 2.5-10.0
/jm (MMAD): Fine PM: 0.1 ¦ 2.5/ym
(MMAD): Ultrafine: <0.1 /jm
(MMAD): EHC-93: NR
Route: Intranasal Exposure
OVA challenge: aerosol
Dose/Concentration: PM: 450//g PM (at 0, 3,
or 9 mg/ml)
OVA sensitization: 50 /yg (0.4 mg/ml).
OVA challenge: 20 /yg (0.4 mg/ml)
EHC-93 was administered at
0 - 900/yg to evaluate any dose-response
relationship.
Time to Analysis: Sensitization and PM
exposure on days 0,14
Challenged on days 35, 38, 41 for 20min/day
Sacrificed on day 42
Effects of Coarse and Fine Particles:
Immunoglobulins: 6/13 of the coarse and 9/13
fine PM samples induced an increase in IgE
and IgGlwhen compared to the control. lgG2a
levels were increased in 3/13 of the coarse
and 5/13 of the fine PM. Particles from De
Zilk induced all three immunoglobulins, except
the fine PM did not induce lgG2a. De Zilk was
intended as a negative control (see Table 3).
Analysis among the sites comparing the
subclasses of antibodies indicated a rank as
follows: Lodz > Rome a Oslo.
Histopathology: 9/13 of the coarse PM
samples and 5/13 of the fine PM samples
induced an inflammatory response.
Bronchoalveolar Cells: Lodz (spring/
summer) coarse and fine PM induced a
significant increase in eosinophils, neutrophils
and monocytes. The coarse and fine PM from
Rome (spring) induced an increase in neu-
trophils and the coarse PM induced an
increase in eosinophils. Also both Lodz and
Rome from the coarse PM from the spring
induced an increase in macrophages. Other PM
samples did not have an effect on BAL cell
counts.
Cytokine Production: None of the samples
produced a significant effect on IL-4 levels.
IFNn levels were significantly decreased in
mice exposed to the fine PM fraction (in 8/13
of the samples) when compared to control.
Coarse particle exposure did not appear to
affect IFNn levels. TNFa levels were
significantly increased (in 2 of the 13
samples) when exposed to coarse PM: fine PM
showed similar responses compared to the
OVA only group. IL-5 was significantly
increased in 4/13 of the coarse and fine PM
samples.
Analysis of PM Components: Samples from
Lodz, Oslo and Rome (all spring) were
evaluated and the water-soluble coarse PM
fraction showed increased immunoglobulin and
pathological responses and the water-insoluble
fine PM fraction from Lodz (Spring) showed
increased reactivity. Leukocytes and cytokines
showed no major differences.
July 2009
D-156
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Steerenberg et al.
(2004,087981)
Species: Mouse
Gender: Male
Strain:
BALB/cByJ.ico
Age: 6-8 wks
Treatment:
1.	C.D2-VH6:
Nramp1s and
Nrampl" deficient
2.	B6.129P2:
Nos2tmLau: ilMOS
deficient
3.	BALB/clL4
(tm2Nnt):
deficient in IL-4
4.	BALB/c (wild
type) pretreated
with N-
Acetylcysteine
(NAC)
Coexposure to EHC-93 and ovalbumin
EHC-93: (Gift of Dr. R. Vincent,
Ottawa, Canada)
Particle Size: EHC-93- has a standard
reference size.
Route: Sensitization, Challenge: Intranasal
NAC: IP injection
Dose/Concentration: OVA: 200/yg (0.4 mg/ml)
EHC-93: 150 //g (3 mg/ml)
NAC: 320 mg/kg
Time to Analysis: OVA-only or OVA+EHC-93
sensitization on days 0 and 14.
Some mice received NAC before intranasal
exposure on days 0 and 14
OVA challenge on days 35, 38 and 41
Sacrificed on day 42
Natural-Resistance-Associated
Macrophage Protein 1 (Nrampl): When
exposed to only ovalbumin, Nrampls evoked
less of an antibody responses (IgE, lgG1 and
lgG2a) compared to Nramp1RHowever when
coexposed to OVA and EHC-93, the level of
increased production of antibodies was similar
in both groups. After coexposure, the wild-
type showed increased histopathological
lesions, whereas the macrophage-stimulation-
deficient types showed only a slight increase
(not significant). IL-4, IFN-n, TNF-a and IL-5
levels were similar in wild-type and the
Nrampl strains.
Pretreatment with NAC: lgG2a
concentration was increased further in the
group pretreated with NAC. The wild-type
mice and the NAC pretreated mice showed
similar histopathological lesion patterns. IL-4
levels were similar in wild-type and the NAC
pretreated mice. (IFN-n, TNF-a and IL-5 levels
not reported)
Inducible Nitric Oxide Synthase (iNOS):
The wild-type and the iNOS-deficient mice had
similar levels of increased IgE antibody
production. The lgG1 and lgG2a antibody
response was twice as great in the iNOS-
deficient mice compared to the wild type. The
wild-type and the iNOS-deficient mice showed
similar histopathological lesions. No
differences in BAL cells or cytokines were
observed between the wild-type and iNOS-
deficient mice.
IL-4: The IL-4-deficient mice did not produce
an increase in IgE or lgG1 antibodies, as was
seen in the wild-type mice. The lgG2a
antibody response in the IL-4-deficient mice
was similar to the wild type response resulting
in adjuvant activity for the lgG2a antibodies.
Overall the histological response of the wild-
type mice was greater compared to the IL-4
deficient mice. There was no real difference
between the two strains observed in the BAL
cells, except IL-5 was significantly lower in
the IL-4-deficient mice.
July 2009
D-157
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Stevens et al.
(2008,155363)
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 10-12 wks
Weight: 17-20 g
DE: generated using a 30 kW 4-cylinder
Deutz BF4M1008 diesel engine
connected to a 22.3 kW Saylor Bell air
compressor. The engine was operated
on diesel fuel purchased from a service
station in Research Triangle Park, NC.
The engine was operated at a steady-
state, approx. 20% of engine's full
load.
High composition:
O2: 4.3 ± 0.07 ppm
NO: 9.2 ± 0.30 ppm
NO2: 1.1 ± 0.05 ppm
SO2: 0.2 ± 0.10 ppm
Low composition:
O2.NO, NO7 SO7 below detection limits
Particle Size: NR
Route: Whole-body Inhalation Chambers
(flow rate - 142 L/min)
OVA immunization and challenge: intranasal
Dose/Concentration: High - 2000//gfnrl
Low - 500/yg/m3
Time to Analysis: DE exposure for 4hfd on days
0-4.
OVA immunization 40min after DE exposure on
days 0-2
Challenged on days 18 and 28.
Sacrificed 4h after last exposure of day 4 for
gene set analysis or 18,48, or 96h after the last
challenge
IgE Antibody Production: In the absence
OVA, IgE antibodies were not detected. 18, 48
and 96 hours following OVA, mice exposed to
low and high doses of DE had an increase in
antibodies over time. Mice exposed to high
dose had an increase (non-significant) to the
OVA exposed control at the 48 h time mark
BAL Cell Counts: Cell counts at 18 and 96
hours after OVA treatment did not differ
among treatment groups. At 48 hours the
number of eosinophils, neutrophils and
lymphocytes were significantly increased in
mice exposed to both high and low
concentrations of DE. With DE exposure alone,
only neutrophils were statistically increased in
the high DE concentration. This indicates the
combination exposure of DE and an antigen is
essential to promote the development of
allergic lung disease.
BAL Cytokine Production: IL-6 production
showed a dose-dependent and time-dependent
increase, but was significantly increase in the
high dose group at 96h. The high dose group
saw a non significant increase in IL-10 levels
over time. The greatest increase in IL-10 for
the low dose group occurredl8 hours after
OVA stimulation.
Pulmonary Inflammation and Lung Injury:
No differences among the groups were
observed for macrophage, lymphocyte,
neutrophil and eosinophil counts. Protein and
LDH levels were not found to be increased in
the BALF of any group.
Gene Set Analysis: Pair wise comparisons
revealed significant gene set difference
between the high DE and control groups.
Comparison of the high DE/OVA versus
air/OVA showed significant changes in 23
gene sets, including genes involved in
oxidative stress responses. The high DE/saline
versus the air/saline showed significantly
altered pathways. Altered pathways include
those for cell adhesion, cell cycle control,
apoptosis, growth and differentiation, and
cytokine signaling. The results show that
relatively short exposures to DE cause mild
increases in immunologic sensitization to
en.
July 2009
D-158
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Takizawa et al.
(2003,157039)
Species: Human
Cell Lines:
Normal Small
Airway Epithelial
Cells and
Bronchial
Epithelial Cells
(BET-1A)
Suspended DEP: collected using a
2,300-cc Isuzu diesel engine using
standard diesel fuel at 1,050 rpm
under a load of 6 torque.
DE exposure in vitro (air explosure):
collected using a 2,300 ml Isuzu diesel
engine at 1,050 rpm.
Composition:
Fine particles: 1 mg/m3
CO: 10.6ppm
NO2: 7.3ppm
SO2: 3.3ppm
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Suspended DEP: varying
doses from 0-50 /yglml
IL-13: varying doses from 0-25 ng/ml
DE exposure in vitro (air exposure): 100/yg/m3
Time to Analysis: Cells were exposed to varying
concentrations of suspended DEP for up to 24h.
NF-kB: analyzed at 6h after suspended DEP
exposure
Air exposure at 0, 2,4, 8 or 14h
Preliminary experiments indicated that DEP at
0.1-50 /yglmL had no significant cytotoxicity
to BET-1A cells and human bronchial epithelial
cells (as analyzed by LDH levels).
Eotaxin Production: (Eotaxin is a cc
chemokine that plays a role in eosinophil
accumulation in a variety of allergic disorders)
Epithelial and BET -1A cells treated with
suspended DEP or IL- showed a dose-
dependent stimulatory effect on eotaxin
release or production. Simultaneous exposure
to 25ngfmL IL-13 and DEPdepicted an additive
effect for both cell types.
Eotaxin mRNA: At 25 //gfmL, suspended DEP
showed a time-dependent effect on eotaxin
mRNA levels up to 12 h in both cell types.
Extracted RNA from human bronchial epithelial
cells exposed to varying doses of DEP showed
a dose-dependent effect for both cell types (up
to 25 /yglmL DEP) on eotaxin mRNA levels
after 12h of exposure. IL-13 also induced a
dose-dependent increase on eotaxin mRNA
levels in cells in both cell types. Combination
of IL-13 and DEP showed an additive effect on
mRNA levels in BET-1 A cells. DE exposure in
vitro also showed a time-dependent
stimulatory effect on eotaxin production in
BET-1 A cells.
NF-DB ISTAT6 Activation: (it has been
suggested that NF-nB plays a role in the
transcriptional regulation of eotaxin gene
expression) Cells exposed to 1-25/yglmL DEP
for 6 h increased NF-nB. BET-1A cells treated
with suspended DEP failed to activate STAT6
Effect of NAC and PDTC on Eotaxin mRNA
Levels: (NAC and PDTC are antioxidant
reagents with inhibitory effects on NF-nB
activation) in BET-1 A, both NAC and PDTC
showed a dose-dependent inhibitory effect on
DEP-induced eotaxin production. Both
reagents also blocked DEP-induced eotaxin
mRNA levels in BET -1A cells. NAC and PDTC
did not suppress eotaxin production or eotaxin
mRNA levels in IL-13 stimulated BET -1A cells.
In addition pre-treatment with NAC
attenuated NF-nB activation induced by DEP
but had no effect on STAT6 induction by IL-
13.
These findings suggest that DEP stimulated
eotaxin gene expression via NF-kB dependent,
but STAT6-independent, pathways.
July 2009
D-159
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Tesfaigzi et al.
(2005,156116)
Species: Rat
Gender: Male
Strain: Brown-
Norway
Aqe: 6wks
PM: Wood smoke generated from a
conventional wood stove that has a
0.5m3 firebox and a sliding gate air
intake damper. The stove was
operated over a 3-phase burn cycle
that spanned 6 h. Fire was started
(initiated exposure) with unprinted I
unbleached newspaper and a mix of
black and white oak.
Wood smoke components: organic
material, small amounts of elemental
carbon and metals and associated
analytes.
Particle Size: 0.36 /jm (MMAD)
Route: Whole-body Inhalation
Dose/Concentration: PM: 1000//gfm:l
Time to Analysis: Exposed to wood smoke or
filtered air 6hfday for 70 consecutive days
OVA IP injection immunization on days 2, 9
OVA aerosol exposure 2h/day on days 67-70
following daily exposure to wood smoke or
filtered air
Sacrificed day 70
Body Weight and Respiratory Function: No
difference in clinical signs or body weight was
observed when comparing the two rat groups.
The woodsmoke exposed group had a 45%
lower dynamic lung compliance when
compared to those exposed to the filtered air
groupbefore the methacholine challenge. Chal-
lenging the rats with methacholine caused a
decrease in dynamic lung compliance in both
groups, but the decrease was greater in the
air-exposed group. At the highest dose of
methacholine, the dynamic lung compliance in
controls was similar to the baseline value of
the smoke-exposed group. No significant
differences in total pulmonary resistance were
observed. Wood smoke exposed rats had a
10% increase in functional residual capacity
than the air-exposed group.
BAL Cell Counts and Cytokines: There was
no difference in lymphocyte, eosinophil or
neutrophils in the BALF of either group. There
was an increase, though not statistically
significant, in macrophages the wood smoke
exposed group when compared to the filtered
air group. In the BALF, IFN-n and and IL-1B
levels were significantly decreased, IL-4 and
GRO-d levels were increased in rats exposed
to wood smoke compared to filtered air.
Serum IgE levels experienced a reduction trend
in the wood smoke group, but it did not reach
significance. Both groups showed mild signs
of inflammation. The average eosinophils
present in stained tissue was 21% higher in
the wood smoke exposed group compared to
the air exposed.
Reference:
Tomita et al.
(2006,097827)
Species: Mouse
Gender: Female
Strain:
C57BL/6J; AHR-
deficient;
mEH-deficient;
ARNT floxed (loxP
sequences
inserted in Arnt
gene);
T cell-specific
ARNT-deficient
Age: 7wks
Weight: 20g
DEP: two independent preparations
fractionated into 13 different fractions
based on acidic and basic functionality
(one from light-duty, 4-cylinder diesel
engine using standard diesel fueld from
Dr. Kei Miwa and other generated from
A4JB-type, Isuzu automobile, Japan)
Individual PAH tested (Osaka, Japan):
BbF - benzo[b]fIuoranthene
BeP - Benzo[e]pyrene
IDP - lndeno[1,2,3-cd]pyrene
BpPe - Benzo[ghi]perylene
BaP - Benzo[a]pyrene
BkF - Benzo[k]fluoranthene
Per - Perylene
DBA - Dibenzo[a,h]anthracene
Particle Size: NR
Route: Intraperitoneal Injection
Dose/Concentration: DEP, fractionated DEP or
PAH compounds: 0.5/yg ¦ 10 mg/kg bw in 50 ul
of olive oil
Time to Analysis: Single exposure, sacrificed 3d
post exposure.
Effect on thymus: DEP treatment (10 mg/kg
of body weight) caused severe atrophy of the
thymus while the spleen and lymph nodes
appeared normal. Three days following DEP
treatment showed a marked reduction in
thymus size. The total number of thymocytes
was reduced by more than 70% mostly due to
a massive reduction in DP cells (CD4+CD8+).
DEP induced no significant alterations in the
cell numbers of CD4/CD8 ratios in the spleen
and lymph nodes.
DEP Extracts: Only the WAC (carbonic acid
fraction) and BE (weak basic fraction) did not
produce a significant reduction in thymocyte
numbers in vivo. Among the active fractions, 7
produced a marked selective loss of immature
DP thymocytes, similar to the crude extract of
DEP.
PAH effects: Thymic involution was severely
induced by the N and various other fractions.
7 out of the 8 PAH compounds were
significantly effective in decreasing the
number of thymocytes upon in vivo exposure.
Only BpPe did not have an effect.
AHR/ARNT and mEH deficient mice (BaP
and DEP only): In the absence of AHR, BaP
treatment did not result in a loss of
thymocytes. Like DEP, BaP produced severe
thymic involution in mEH-deficient mice. DEP-
mediated thymic involution was significantly
enhanced in mEH-deficient mice.
July 2009
D-160
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Verstraelen et al.
(2005, 096872)
Species: Human
Tissue/Cell
Types: Monocyte-
derived dendritic
cells (Mo-DC)
Cord blood
samples of seven
women were
collected from
umbilical vessels
of placentas of
normal, full-term
infants.
DEP- SRM 2975
Particle Size: NR
Route: Cell Culture
Dose/Concentration: DEP in varying
concentrations: 0.2, 2, 20, 200, 2000 ng/mL
LPS 100 ng/mL
Time to Analysis: Cells were incubated for 24h
and analyzed via flow cytometry.
Biological Markers: Exposure to DEP alone
did not alter expression levels of HLA-DR,
CD86 or CD83.
Treatment with LPS alone caused a non-
significant increase in all three markers when
compared to the control.
Treatment with DEP+LPS caused a
significant increase in the expression of CD83
and a non-significant increased expression of
HLA DR and CD86. DEP+ LPS induced a bell-
shape dose-response curve on the expression
of all three markers, with a dose of 20 ngfmL
DEP + 100 ng/mL LPS causing the largest
increase in upregulation.
When only the results of the LPS-responsive
donors (5 out of 7 blood cord samples) were
included, the effects described above become
more pronounced.
Reference:
Walczak-
Drzewiecka et al.
(2003, 096784)
Species: Mouse
Cell Line:
C1.MC/C57.1
(C57) Mast Cells
Metal and Transition Metal Ions: Sr2
Ni2+, Cd2+, Al3+, Pb2+
Particle Size: NR
Route: Cell culture,
Dose/Concentration: 0.1-5/ymol
Time to Analysis: 10min - 4h incubation before
analysis
BHex Mediator Release in Mast Cells:
Incubation with SrCb, NiS04, CdCl2 or AlCh
resulted in a 2-5% release of B-hexoaminidase
in mast cells. Incubation with a mixture of all
these compounds induced a greater (11%)
release in B-hexoaminidase, indicating there
might be a additive effect.
Cell Viability: Incubation of cells at
concentrations and incubation time employed
did not result in decrease in cell viability.
Antigen-Mediated Mediator Release in
Mast Cells: Al3+ and Ni2+ enhanced antigen-
mediated release. 10-7 MAICh released 23%
of B-hexoaminidase compared to antigen
alone, which induced 11 % release of B-
hexoaminidase. Cd2+, Sr2+ and Pb2+ enhanced
antigen-mediated release to a lesser extent.
Ni2+, Al3+, Sr2+ and Cd2+ depicted a dose-
dependent relationship with antigen-mediated
B-hexoaminidase release.
Antigen-Induced Protein Phosphorylation:
Addition of the antigen induced the
anticipated phosphorylation of multiple
proteins in C57 mast cells The presence of
Ni2+ and Pb2+ mediated an increase in
phosphorylation of several of the proteins and
Al3+ mediated a decrease in phosphorylation
of multiple proteins (specifically the 56 and 37
kD bands).
Antigen-Mediated Cytokine Secretion (IL-
4): At certain concentrations all tested metal
and transition metal ions were able to induce
IL-4 secretion or enhance antigen-induced IL-4
secretion in mast cells, but no dose-dependent
relationship was established.
Reference: Wan
and Yu (2006,
157104)
Species: Human
Cell Lines:
Human, B cell
lymphocytes
PMBC
(>98.5% B cells-
CD19+CD20+;
<1 % Tcells
(CD3+))
Human
lymphocyte cell
lines - DG75
NQ01 wild type
DEP from 4 cyl Isuzu diesel methanol
extracts (Previously published)
Particle Size: NR
Route: Cell Culture, PMBC - 1 x 106 cell
DG 75 - 3 x 10® cells
IgEPMBC 1 x 106/mL
B-cells 0.5x10e|mL
Dose /Concentration: 2.5, 5,10, 20 /yg DEPX/
plate (text refers to 20 /yg/mL)
IgEDEPX 100 ng/mL
sulfurophane at 0 - 30 /ymol
Time to Analysis: 6h mRNA; 16 h protein
assay. IgE 14d.
Induction of NQ01 by DEPX: In PBMCs and
DG75DEPX dose-dependently inudced NQ01
mRNA expression NQ01 ARE was increased
NAC inhibited NQ01 gene expressiondose
dependently. p38 MAPK and P13K inhibition
partially blocked NQ01 mRNA and ARE
induction by DEPX.
Induction of phase II enzymes: DEPX
induced IgE potentiation was reduced dose
dependently by induced phase II enzymes.
July 2009
D-161
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Whitekus et al.
(2002,157142)
Species: Mouse
Cell Line: RAW
264.7
DEP (light-duty, four-cylinder engine-
4JB1 type, Isuzu Automobile, Japan;
standard diesel fuel) (extracts)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 50//g/mL
Time to Analysis: DEP extracts suspended in
methanol and sonicated 20min. Centrifuged
10min. Dried DEP resuspended and stored -20°C.
Cell cultures maintained 37 °C. Exposed to
antioxidants 5h. HO-1 western blot,
determination of cellular GSH:GSSG ratios,
carbonyl protein content, lipid hydroperoxides
performed.
DEP significantly reduced the GSH:GSSG
ratio. This effect was prevented by adding
thiol antioxidants NAC or BUC. DEP increased
lipid peroxide levels, but the addition of all
antioxidants decreased these levels. DEP
increased carbonyl groups. NAC, BUC, and
luteolin reduced HO-1 expression.
Reference:
Whitekus et al.
(2002,157142)
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 6-8wks
Weight: NR
DEP (light-duty, four-cylinder engine-
4JB1 type, Isuzu Automobile, Japan;
standard diesel fuel) (extracts)
Particle Size: 0.5-4//m
Route: Inhalation
Dose/Concentration: 200, 600, 2000//g/m3
Time to Analysis: Exposed 1 h/d 10d. Animals
receiving OVA had 20min OVA exposure after
DEP exposure.
DEP+OVA dose-dependency increased IgE and
IgG 1, being more effective than the OVA-alone
treatment. This effect was significantly
suppressed by thiol antioxidants NAC or BUC.
DEP+OVA increased carbonyl protein and lipid
peroxide over OVA. NAC or BUC suppressed
lipid peroxide and protein oxidation. No general
markers for inflammation were observed.
Reference:
Witten et al.
(2005, 087485)
Species: Rat
Gender: Female
Strain: F344
Age: 8wks
Weight: — 175g
DEP (heavy-duty Cummins N14
research engine operated at 75%
throttle)
Particle Size: 7.234-294.27nm
Route: Nose-only Inhalation
Dose/Concentration: Low- 35.3±4.9//g/m3,
High- 632.9±47.61 //g/m3
Time to Analysis: Exposed 4hfd, 5dfwk, 3wks.
Pretreated with saline or capsaicin.
There were no differences for substance P.
The low-exposure group had significantly less
NK1. DEP reduced NEP activity. Plasma
extraversion dose-dependently increased and
was greatest in capsaicin animals. Respiratory
permeability dose-dependently increased. IL-
1B was significantly higher for the low-
exposure group. IL-12 was significantly lower
in the capsaicin high-expsoure group. TNF-a
increased in the high-exposure group and
capsaicin low-exposure group. High exposure
induced particle-laden AMs in the lungs,
perivascular cuffing consisting of mononuclear
cells, alveolar edema and increased mast cell
number. Neutrophil and eosinophil influx was
not seen.
Reference: Wong DEP (Cummins N14 research engine at Route: Nose-only Inhalation
et al. (2003,
097707)
Species: Rat
Gender: Female
Strain: F344/NH
Age: ~4wks
Weight: — 175g
75% throttle) (EC- 34.93-601.67
//g/m3, 0C- 1.90-11.25 //g/m3.
Sulfates 0.94-17.96 //g/m3, Na- 4.07-
4.78 ng/m3, Mg- 0.60-0.86 ng/m3, Ca-
5.05-10.66 ng/m3, Fe-3.17-6.44, Cr-
0.68-1.31 ng/m3, Mn- 0.11-0.22 ng/m3,
Pb- 0.97-1.24 ng/m3)
Particle Size: 7.5-294.3nm
Dose/Concentration: Low- 35.3±4.9//g/m3,
High- 669.3±47.6//g/m3
Time to Analysis: Exposed 4h/d, 5d/wk, 3wks.
Pretreated with saline or capsaicin.
DEP dose-dependently increased plasma
extraversion, which was further increased by
capsaicin. In the high-exposure group, particle-
laden AMs (which were reduced by capsaicin),
inflammatory cell margination, perivascular
cuffing with subsequent mononuclear cell
migration and dispersal, increased mast cells,
and decreased substance P were all seen. NK-
1R was downregulated in the low-exposure
group and upregulated in the capsaicin-
pretreated high-exposure group. NEP
decreased significantly for both groups.
July 2009
D-162
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:	Washed DEP (carbonaceous core), DEP-
Yanagisawa et al.	OCfextracted organic chemicals) and
(2006, 096458)	Whole DEP
Species: Mouse	Particles collected from: 4JB1-Type,
Gender' Male	four-cylinder, 2.74 L, Isuzu diesel
engine, while operated on standard
Strain: ICR	diesel fue| at 200g under a load of 10
Age: 5wks	torques.
Weight: 25-28g	Particle Size: MMAD- 0.4/ym
Route: IT Administration
Dose/Concentration: 50/yg/0.1L
Time to Analysis: 1. Control- 0.1mL PBS
2.	DEP-0C- 50 /yg
3.	Washed DEP-50ug
4.	Whole DEP- 50ug DEP-OC + 50ug Washed
DEP5. OVA-1 /yg -
6.	DEP-OC- 1 /yg + OVA
7.	Washed DEP- 50/yg + OVA 8. Whole DEP- 50
/yg DEP-OC + 50/yg Washed DEP + OVA
All groups received OVA or PBS every 2 wks for
6 wks and the PM component or PBS once a
week for 6 wks.
BAL Analysis: DEP-OC + OVA caused a
significant increase in PMN infiltration in the
BAL compared to the control Exposure to
Whole DEP+ OVA caused PMN count to rise
further
Macrophages: OVA alone DEP-OC +0VA,
Washed DEP + OVA and Whole DEP + OVAall
caused a significant increase in macrophages
compared to the control.
Lung Histology: Exposure to OVA, Washed
DEP, DEP-OC and Whole DEP caused a slight
increase in PMNs, mononuclear cells and
goblet cell proliferation. Treatment with all
three DEP groups + OVA caused a significant
increase in mononuclear cells, PMNs and
goblet cell proliferation. Whole DEP + OVA
had the greatest impact.
Thl and Th2 Cytokine Expression: Washed
DEP+OVA caused a significant increase in
IFN-n compared to control, whereas Whole
DEP+OVA caused a significant decrease
compared to control. No significant
differences in IL-2 and IL-4 levels were seen
among groups. DEP-OC + OVA and Whole
DEP+ OVA caused significant increases in IL-5
compared to control and compared to OVA
Whole DEP+OVA caused significant increase
in IL-13 compared to control
Eotaxin and MIP-1a Expression: OVA
increased eotaxin levels and DEP-OC + OVA
caused a more significant increase in eotaxin.
Whole DEP alone caused a significant increase
in MIP-1a and Whole DEP+OVA caused an
even greater increase in MIP-1a.
IgG 1 Levels: Exposure to DEP-OC+OVA
caused an increase in IgG 1 and exposure to
Whole DEP+OVA caused greater elevation in
lgG1 levels.
Reference: Yang DEP-SRM 1650
Particle Size: MMAD- 0.5mm
087886)
Species: Mouse
Gender: Female
Strain: B6C3F1
Age: 6-8wks
Exposure on Spleen IgM AFC: DEP exposure
for 2-weeks induced a dose-dependent
decrease in spleen AFC in response to sRBC
immunization. Mice exposed to 15 /yg/kg
depicted a 35% reduction in total spleen
activity. In the group exposed to DEP for 4
weeks, the decrease in AFC was not
significantly different than the control.
DEP Exposure on Spleen Cell
Number/Lymphocyte Counts: Exposure for 2
or 4 weeks did not affect total number of
nucleated splenocytes. DEP caused a 30%
reduction in total T cells. The number of B
cells were not significantly affected.
DEP Exposure on Spleen T-Cell Function:
(evaluated in 2 week exposure group only)
DEP induced a dose-dependent decrease in
spleen cell proliferation to ConA. DEP did not
affect spleen cell proliferation in response to
anti-CD3 mAb. Production of IL-2 in response
to ConA was reduced in a dose-dependent
manner by DEP exposure. IFN-n production
was decreased by exposure to DEP. IL-4
production was not measuredable.
Route: IT Aspiration
Dose/Concentration: 1, 5, or 15 mg DEP/kg
body weight
Time to Analysis: Mice exposed to 1, 5, or 15
mg DEP/kg of body weight for either 3 times in 2
wks or 6 times in 4 wks.
Toxicity of DEP Exposure: DEP did not have
a significant effect on body, liver or spleen
weight. The highest dose of DEP caused an
increase in lung weight and lung weight
relative to body weight. None of the
hematological parameters were significantly
different in the mice exposed for 2wks; in the
4 wk group there was a significant decrease
in platelet counts in mice exposed to 15mgfkg.
DEP
July 2009
D-163
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Yin et DEP - SRM 2975 (NIST)
al. (2005,	Listeria
MM)	Partjc|e Sjze. NR
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BNfCrlBR)
Age: NR
Weight: 200-
250g
Route: Nose-only inhalation (DEP), NT instillation
tListeria)
Dose/Concentration: 100,000 CFU (Listeria);
21.2 + 2.3 mg/m3 (DEP)
Time to Analysis: DEP exposure for 4 h/d for
5d; infection with Listeria 7d postexposure;
sacrifice 3 and 7d postinfection
Lung deposit: Estimated mean lung deposit of
DEP - 406 + 29//g/rat DEP prolonged
growth of bacteria in lung
Alveolar Macrophage (AM) response: DEP
significantly inhibited Listeria-'mAuwA IL-1B
secretion at d7 and TNF-a snf IL-12 at both
d3 and d7. IL-10 production was enhanced at
d7.
T-Lymphocyte response: DEP significantly
reduced the development of T cells in
response to Listeria infection. These
lymphocytes displayed increased production of
IL-6 at d7, but significantly diminished levels
of IL-10, IL-2 and IFN-n.
Reference: Yin et
al. (2004,
097685)
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BN/CrlBR)
Age: NR
Weight: 200-
250g
DEP - SRM 2975 (NIST)
Listeria
Particle Size: NR
Route: Inhalation Exposure (DEP)
IT instillation (Listeria)
Dose/Concentration: 20.62 + 1.31 mgfm3
(DEP).
100,000 CFU Listeria
Time to Analysis: DEP exposure for 4 h/d for
5d; inoculation with bacteria 2h postexposure;
sacrifice 3, 7,10d postinfection
Lung deposit: Estimated mean lung deposit of
DEP - 389 + 25 //g/rat
Pulmonary responses and bacterial
clearance: DEP significantly augmented
Z/£fe/7d-induced neutrophil infiltration, lung
CFU and recoverable alveolar macrophages
(AM) at all times post-infection. LDH activity
was increased 3d post-infection. Bacterial
count in DEP exposed rats remained
significantly higher through d7.
Cytokine production by AM: DEP exposure
significantly lowered Z/ife/73-induced
production of IL-1 B, TNF-a and IL-12.
Production of IL-10 was strongly augmented.
T-lymphocyte responses: DEP moderately
but not significantly lowered the total number
of lymphocytes, CD4+ cells and lymphocyte
IL-10 production. Listeria-induced T-cell
development was strongly inhibited, as were
the development of CD8+ cells, IL-12
production and IFN-n secretion. DEP and
Listeria exposure showed and increased
production of IL-6 at d3 and d7 post-exposure.
Reference: Yin et
al.
2007
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BN/CrlBR)
Age: NR
Weight: 225-
250g
DEP - SRM 2975
eDEP - organic DEP extract
wDEP - washed DEP
CB - carbon black
Particle Size: DEP: median diameter-
19.4 fjm, surface area- 91 fj2\g; CB:
0.1-0.6 /jm
Route: IT Instillation of Listeria.
Treatment after AM cell isolation
Dose/Concentration: DEP: 10, 50,100//gfmL;
CB: 50/yglmL
Time to Analysis: Sacrifice postinfection or no
infection. AM isolated. DEP or CB treatments for
1,4,16, 24h.
AM Phagocytosis: None of the DEP or CB
treatments were cytotoxic or affected the
number of adherent cells. 10-100 /yglmL. DEP
significantly decreased AM phagocytosis in a
concentration- and time-dependent manner,
with increased concentration and time
decreasing activity.
Bacterial Activity: The inhibition of AM
bactericidal activity by DEP was time- and
concentration-dependent. eDEP and wDEP
inhibitied the AM bactericidal activity but
were less effective than DEP. The CB
treatment was not significant.
Cytokine Secretion by AM: DEP and eDEP
concentration-dependently decreased TNF-a,
IL-1 B and IL-12, but increased IL-10. wDEP
and CB did not show a significant effect.
Cytokine Secretion by Lymphocytes: DEP
and eDEP concentration-dependently
decreased IL-2 and IFN-n. wDEP and CB had
little effect, except high concentrations of
wDEP decreased IFN-n.
July 2009
D-164
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference: Yin et
al. (2004,
087983)
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BNfCrlBR)
Age: NR
Weight: 225-
250g
DEP - SRM2975
eDEP - organic DEP extract
wDEP - washed DEP
CB - carbon black
Particle Size: DEP- NR,
CB- 0.1-0.6/jm
Route: IT Instillation of Listeria.
Treatment after AM cell isolation
Dose/Concentration: 50//gjmL (DEP or CB)
Time to Analysis: Killed 7d post instillation. AM
isolated then incubated. DEP treatments for up to
24h at 37 degrees.
DEP-induced ROS production: R0S was
induced by DEP or eDEP and inhibited by eDEP
with ANF or NAC. eDEP induction of ROS was
time-dependent. wDEP or CB did not induce
ROS.
DEP-induced HO-1 expression: DEP- or
eDEP-induced H0-1 expression was inhibited
by ANF, NAC or SB203580. wDEP or CB did
not induce ROS. DEP or eDEP exposure
resulted in a 2.5- to 3-fold induction of H0-1
expression in uninfected AM.
eDEP-modulated cytokine production:
eDEP exposure resulted in a time-dependent
increase in LPS-stimulated IL-10 orTNF-a
production, and both were inhibitied by ANF or
NAC. wDEP did not affect either. SOD
pretreatment attenuated eDEP-upregulated
H0-1 expression, inhibited IL-10, and reversed
eDEP inhibition of IL-12. Znpp decreased IL-
10.
Reference: Yin et
al.
(2003,096127)
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BN/CrlBR)
Age: NR
Weight: 200-
250g
DEP - SRM 1650a
L. monocytogenes
Particle Size: NR
Route: Nose-only Inhalation (DEP).
IT Instillation (Listeria)
Dose/Concentration: 50 or 100 mgfm3 (DEP);
100,000 bacteria per 500uL sterile saline
(Listeria)
Time to Analysis: DEP exposure for 4h.
Bacterial inoculation. Sacrificed 3, 7d
postexposure.
Lymphocyte population: DEP-alone exposure
increased total lymphocytes, T cells and T-cell
subsets. Elevated cell counts in the combined
exposure were DEP dose-dependent, with the
100 mgfm3 treatment having significant
increases in the cell number and CD8+/CD4+
ratio.
IL-2: DEP exposure in noninfected rats at both
doses increased IL-2 in the 24h culture and
decreased IL-2 in the 48h culture. The
increase in IL-2 at 3 days postinfection was
not significant. DEP exposure increased IL-
2Ra in response to ConA stimulation. DEP-
treated infected rats had increases in ConA-
inducible CD4+|IL-2Ra+ and CD8+/ IL-2Ra+.
IL-6: IL-6 production was dose-dependent in
DEP-treated uninfected rats and infected rats.
The combined exposure produced less IL-6
than the DEP-alone or Listeria a\one
treatments.
IFN-D: DEP decreased IFN-n at 3 days
postexposure, but increased at 7 days
postexposure in a dose-dependent manner.
Uninfected DEP-treated rats did not
substantially respond to HKLM. HKLM-induced
IFN-n production is strongly inhibited at all
tested DEP doses.
Reference:
Zelikoff et al.
(2003, 039009)
Species: Rat
Gender: Male
Strain: Fischer
344
Age: 7-8mo
Weight: NR
CAPS - concentrated ambient PM2.6
from New York City
S.pneumoniae
Particle Size: MMAD- 0.4//m
Route: Nose-only Inhalation (CAPS)
IT Instillation [S.pneumoniae)
Dose/Concentration: CAPS: Study 1- 60-600
^g/rn3,
Mean- 345 /yg/m3;
Study 2- 65-150 //g/m3,
Mean-107 /yglm3;
[S.pneumoniae 2-4 x 107)
Time to Analysis: Study 1: Uninfected rats
exposed to air or CAPS for 3h. Sacrificed 3, 24,
or 72h postexposure or IT instilled 4, 24, 72,
120h and sacrificed 4, 24, 72h postinfection
Study 2: Infection with bacteria. Exposed 48h
later to CAPS or filtered air for 5h. Sacrifice 9,
18, 24, 72,120h postexposure.
Study 1: CAPS did not effect cell numbers,
viability, profiles, lavageable LDH activity,
total protein, or total circulating WBC counts.
Exposure to CAPs prior to infection
significantly increased PMN and decreased
lymphocytes. WBC levels returned to contol
levels by 4h postinfection. CAPS had no effect
on circulating monocyte values. CAPS
significantly increased bacterial burdens at
24h, but thereafter the burden decreased to
below control levels.
Study 2: In CAPS exposed rats, PMN
decreased, Pam increased, and the cytokines
TNF-a, IL-1 B and IL-6 decreased.
Lymphocytes and monocytes were unaffected.
Bacterial burdens in CAP-exposed rats were
about 10% greater than air contols at 9h and
> 300% greater at 18h. CAPS significantly
increased the percent of affected lung area
and severity of infection.
July 2009
D-165
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Pollutant
Exposure
Effects
Reference:
Zelikoff et al.
(2002,037797)
Species: Rat
Gender: Male
Strain: Fischer
344
Age: 7-9m
Weight: NR
Ambient NYC PM
Single transition metals of Fe, Mn, Ni
Streptococcus pneumoniae
Particle Size: NYC PM: PM21,
Fe2+, Mn2+, Ni2+: 0.4//m (MMAD)
Route: NYC PM, single transition metals: Nose-
only inhalation
S. pneumoniae: IT instillation
Dose/Concentration: 15-20 x 106
[S.pneumoniae)-, Single metals/NYC PM: 65-
90/yg/m3
Time to Analysis: Infection/no infection
followed by 5h exposure to NYC PM or single
transition metal. Sacrifice 4, 5, 9,18, 24, and
120h after exposure
CAPs exposure to infected rats significantly
increased pulmonary bacterial burdens of S.
pneumo in a time-dependent manner. At 9h,
18h, 24h, and 5d after CAPs exposure,
bacterial burdens were 10%, 300%, 70% and
30% above controls. Uninfected rats exposed
to the single transition metals showed
significant alterations in PMNs and
lymphocytes values at 1h postexposure.
Exposure to Fe of uninfected rats significantly
increased superoxide anion production by
pulmonary macrophages. Uninfected rats
exposed to inhaled Fe significantly reduced B-
lymphocyte proliferation at 48 h, but did not
affect T-lymphocyte production. Inhaled Ni,
for the uninfected, significantly decreased T-
lymphocyte production at 18h, and did not
affect B-lymphocyte production. Inhalation of
Fe by infected rats facilitated an increase in
bacterial numbers while Ni inhibited bacterial
clearance. Inhaled Fe by infected also
significantly decreased PMNs and lymphocyte
numbers by 35% and increased pulmonary
macrophage numbers by 29% when compared
to the air exposed group. Results
demonstrated that inhalation of Fe altered
innate and adaptive immunity in uninfected
hosts, and both Fe and Ni reduced pulmonary
bacterial clearance in previously infected rats.
Reference: Zhong
et al. (2006,
093264)
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 6-8wks
Weight: NR
Cell Line:
J774A.1
CAPs: Concentrated Air Particles
(Boston, MA)
Urban air particles (UAP) SRM1649
(Washington, DC)
Titanium Oxide (Ti02> (Baker
Chemicals, Phillipsburg, NJ)
Carbon Black (CB) (Sigma, St. Louis,
MO)
Particle Size: UAP - NR;
Ti02/CB - NR; CAPs: nPM?.!,
Streptococcus pneumoniae: strain
ATCC6303
Route: Cell Culture
Dose/Concentration: NR, 100//gfmL
Time to Analysis: Pre-treated CAPs cells
exposed to CAPs for 1h.
All cells, IFNy-primed AMs, unprimed AMs, and
J774A.1, exposed to bacteria for 1h.
Binding measured 15h after bacteria exposure.
Ingestion measured 2h after bacteria exposure.
Rate and number of killed bacteria measured 2h
after bacteria exposure.
Binding, internalization and killing of
bacteria: CAPS significantly increased binding
of bacteria by IFNy-primed AMs, normal AMs
and J774A.1. CAPS decreased internalization
and absolute number of bacteria killed by
macrophages of all types. The rate of killing of
internalized bacteria was similar in the
presence or absence of CAPs; however, CAPs
did cause a decrease in the absolute number
of bacteria killed by all three types of
macrophages, due to the decrease in
internalization.
Effects of other particles: Ti02 and CB had
no effect on J774 binding or internalization of
S. pneumo Ti02 and CB's effects on primed
and unprimed AMs were not reported. Testing
with UAPs, however, showed effects similar
to those observed with CAPs.
Soluble components: The soluble fraction of
CAPs, especially iron, is responsible for
decreased internalization.
July 2009
D-166
DRAFT-DO NOT CITE OR QUOTE

-------
Table D-5. Effects of the central nervous system.
Reference
Pollutant
Exposure
Results
Reference:
Calderon-
Garciduehas et
al. (2003,
156316)
Species: Dog
Gender: Male,
Female
Strain: Mixed
breed
Age: 7d-1 Oyrs
Weight:
349 ± 116g -
20kg
Urban Air (Mexico City-high PM
region, Tlaxcala- low PM region)
(PM, Pb, volatile organic
compounds, formaldehyde,
acetaldehyde, mutagenic PM,
alkane hydrocarbons, Ni, V, Mn,
Cr, peroxyacetyl NO42', LPS,
endotoxins)
Particle Size: PM: 2.5,10 /jm
Route: Ambient Air Exposure
Dose/Concentration: Mexico City: PM10:
78 /yglm3, PM2.6: 21.6 //gfm3, Pb in TSP:
< 0.4/yg/m3
Time to Analysis: Dogs raised in house or
outdoor-indoor kennel. Lifetime exposure.
Mexico City dogs had significantly greater apurinic and
apyrimidic sites in the olfactory bulb and hippocampus.
Histopathological changes in the respiratory and olfactory
epithelium were greatest in Mexico City dogs. Mexico City
dogs also had greater immunoreactivity than the controls
for NFnB, ilMOS, cyclooxygenase-2, glial fibrillatory acidic
protein, ApoE, amyloid precursor product and B-amyloid.
Reference:	CAPs from Los Angeles, lacking
Campbell et al.	reactive organic and H20 soluble
(2005, 087217)	gases, O3, NOx, SOx
Species: Mouse	Particle Size: F+UF: <2.5//m;
Strain: BALB/c	UF: < 0.18//m
Age: 7wks
Route: Whole-body Exposure
Mice were challenged with OVA prior to exposure and 1
Dose/Concentration: 20-fold concentration and 2 weeks following exposure, and then brains were
of near highway ambient air, avg UF
concentration: 282.5//g|m3, avg F
concentration: 441.7 /yglm3
Time to Analysis: 4hfd, 5dfwk for 2wks
assayed. F+UF and UF exposure increased NFnB DNA
binding in brain. TNFa increased with F+UF. IL-1 a
increased with UF and F+UF. This suggests a possible
link between PM exposure and neurodegenerative disease
processes.
Reference: Che
et al. (2007,
096460)
Species: Rat
Strain:
Sprague-Dawley
Gender: Male
and Female
Age: 9wks
Weight: 190-
220g
Gasoline exhaust (collected from
1996 Guangzhou passenger car
with Dongfeng Gasoline Series
155 kw engine and no exhaust
catalytic converter fuelled with
90-octane lead-free gasoline from
China Petroleum).
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 5.6,16.7, or 50.0
L/kg, final volume 0.3 mL/rat
Time to Analysis: 1fwk for 4wks.
Parameters measured 24h post last
instillation.
A dose-dependent increase was observed in brain DNA
damage starting at 5.6 L/kg. Increase in lipid peroxidation
and carbonyl protein was also observed at 50 L/kg.
Decrease in brain SOD occurred at all exposures. GPx
activity was unchanged with exposure. This suggests an
association between gasoline exhaust and oxidative
damage to the brain.
Reference:
Kleinman et al.
(2008,190074)
Species: Mouse
Gender: Male
Strain: ApoE'
Age: 6wks
Weight: NR
CAPs (Los Angeles, CA) (OC, EC -
— 50%; sulfate, nitrate —11%)
Particle Size: NR
Route: Whole-body Exposure Chamber
Dose/Concentration: High dose: Mass
concentration-114.2/yglm3, Concentrated
factor-15.4±3.2; Low dose: Mass
concentration: 30.4/yg/m3, Concentrated
factor- 4.1 ±0.9
Time to Analysis: Exposed 5hfd, 3dfwk,
6wks. Killed 24h postexposure.
Activated AP-1 dose-dependently increased. Activated NF-
~B significantly increased with the high CAPs dose. GFAP
(which represented activated astrocytes) and activated
JNK significantly increased with the low CAPs dose.
Reference: Liu
et al. (2005,
088650)
Species: Rat
Strain: Wistar
Gender: Male
Aqe: 8wks
CAPs from Taiyuan, China
Particle Size: <2.5//m
Route: IT Instillation
Dose/Concentration: 0,1.5, 7.5, or 37.5
mg/kg, final volume 0.2 mL/rat
Time to Analysis: Assayed 24h following
instillation
In the brain, SOD and CAT activity were significantly
decreased at the 2 highest doses; GSH levels were
significantly decreased at the highest dose. This suggests
an association between PM exposure and oxidative
damage mediated by prooxidant/antioxidant imbalance or
high levels of free radicals.
Reference:
Sirivelu et al.
(2006,111151)
Species: Rat
Gender: Male
Strain: Brown
Norway
Aqe: 12-13wks
CAPs from Grand Rapids, Ml
Particle Size: < 2.5 fjm
Route: Whole-body Exposure
Dose/Concentration: 500//gfm3
Time to Analysis: 8h, assayed at 24-h PE
PVN: CAPs alone or with OVA increased NE.
MPA: CAPs increased Da when treated with OVA while
no changes in NE, 5-HIAA and DOPAC were observed.
Arcuate nucleus: OVA sensitization increased NE levels.
OB: CAPs and OVA increased NE levels, but no changes in
Da, DOPAC, or 5-HIAA were observed.
Other areas: No differences in other areas of
hypothalamus, substantia nigra, or cortex were observed.
CAPs alone or with OVA increased serum corticosterone.
These results suggest that CAPs can cause region-
specific modulation of neurotransmitters in brain and that
the stress axis may be activated causing aggravation of
allergic airway disease.
July 2009
D-167
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Results
Reference:
Veronesi et al.
(2001,015977)
Species: Mouse
Strain: ApoE'1'
or C57BI/6
Age: Young
adults
CAPs from Tuxedo, NY
Particle Size: < 2.5 fjm
Route: Whole-body Inhalation
Dose/Concentration: Varied daily, PM2.6
concentrated 10-fold, producing an average
of 113/yg/m3.
Time to Analysis: 6hfd, 5dfwk for 4m
CAPs-exposed ApoE'1' mice had an 29% reduction in TH-
stained neurons and a 8% increase in GFAP staining
compared to air-exposed ApoE1. No differences were see
in C57 mice. The results suggest that ApoE1 mice,
characterized by increased brain oxidative stress, are
susceptible to PM-induced neurodegeneration.
Reference:
Veronesi et al.
(2005,087481)
Species: Mouse
Gender: NR
Strain: ApoE'1',
C57B1/6
Age: NR
Weight: NR
CAPs (Tuxedo, NY)
Particle Size: 2.5 /jm
Route: Whole-body Exposure
Dose/Concentration: Average: 110 //gfnv1
Time to Analysis: Exposed 6hfd, 5dfwk,
4m.
Dopaminergic neurons significantly decreased in CAPs-
exposed mice compared to the controls. CAPs caused
significantly more astrocytes (as measured by GFAP
staining) in the nucleus compacta compared to the
controls.
Reference:
Win-Shwe et al.
(2008,190146)
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 7wks
Weight: NR
DEP (Nanoparticle-rich - NPDE;
81-diesel engine, steady-state
condition, 5h/d, 2000rpm, 0 Nm)
(CO, CO2.NO, NO?, SO?)
Particle Size: Diameter:
26.21 ±1.50nm
Route: Whole-body Exposure Chamber
Dose/Concentration: 148.86 ±8.44 /yglm3
Time to Analysis: Exposed to NPDE or
filtered air 5hfd, 5dfwk, 4wks. Some mice ip
injected with lipoteichoic acid (LTA) 1Xfwk,
4wks. Morris water maze behavioral test:
3d acquisition, 2d probe trial.
Mice in the LTA+NPDE group had significantly longer
mean escape latencies, indicating impaired acquisition of
spatial learning. NPDE directly increased NR1 and TNF-a.
LTA+NPDE increased NR2A, NR2B, and IL-1B, however
LTA was primarily responsible for the increases.
Reference:
Zanchi et al.
(2008,157173)
Species: Rat
Gender: Male
Strain: Wistar
Aqe: 45d
R0FA from Universidade de Sao
Paulo, Brazil
Particle Size: 1.2 ± 2.24/ym
(MMAD)
Route: Intranasal Instillation
Dose/Concentration: 20//gf 10//I saline
Time to Analysis: 30d
Exposed rats had increased lipid peroxidation in striatum
and cerebellum. This could be reversed with N-
acetylcysteine treatment. R0FA treatment altered motor
activity shown by decreased general exploration and
peripheral walking, and was not prevented by NAC.
Results suggests that chronic R0FA induces behavioral
changes and brain oxidative stress.
Table D-6. Reproductive and developmental effects.
Reference
Pollutant
Exposure
Effects
Reference: Fedulov
et al. (2008,
097482)
Species: Mouse
Gender: Female
(pregnant),
Offspring: NR
Strain: BALB/c
Age: NR
Weight: NR
DEP	Route: Intranasal Insulfation
Carbon black (CB) Dose/Concentration: DEP, Ti02: 50/yg in 50
j!q2	/jL, 50 /yg/mouse; CB: 250 /yg in 50 fjL
Particle Size' NR Time t0 Analysis: Particle samples baked 3h.
Protocol 1a: Pregnant mice treated with DEP or
Ti02. Analyzed 19 or 48h later. Protocol 1b:
Pregnant mice DEP, Ti02 or CB treated day 14
of pregnancy. 4d-old offspring i.p. injected with
OVA+alum. 12-14d-old exposed aerosolized
OVA.
DEP increased BAL PMN counts in normal and pregnant
mice. In pregnant mice, DEP and Ti02 increased IL-1 B, TNF-
a, IL-6 and KC compared to nonpregnant controls. Offspring
of DEP, CB or Ti02 exposed mice had increased AHR and
airway inflammation. Ti02 exclusively altered the expression
of 80 genes in pregnant mice.
Reference:
Fujimoto et al.
(2005, 096556)
Species: Mouse
Strain: Slc:ICR
Gender: Females
(pregnant mice and
fetuses)
Age: NR (pregnant
females), 14d of
gestation (fetuses)
DE: generated by 2369
cc diesel engine at 1050
rpm at 80% load with
commercial light oil
Particle Size: 0.4 /jm
(MMAD)
Route: Inhalation Chamber
Dose/Concentration: 0.3,1.0 or 3.0 mg
DEP/m3
Time to Analysis: 12hfd, 7dfwk from 2d post
coitumto 13dpc. Sacrificed 14dpc. mRNA
expression examined in female fetuses.
Significant increase in absorbed placentas were observed in
the 0.3 and 3.0 concentration. A decrease in absorbed
placentas was observed for the 1.0 concentration. Increased
inflammatory cytokine mRNA in placentas from exposed
offspring were observed. An increased number of absorbed
placentas in DE-exposed offspring were seen.
July 2009
D-168
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Hougaard et al.
(2008,156570)
Species: Mouse
Strain: C57BI/6
Gender: Pregnant
females, male and
female offspring
Age: 12 & 16 wks
(female offspring),
13 & 17 wks (male
offspring)
DEP(SRM2975)
Particle Size: 90 m2j
(SA)
Route: Whole-body Inhalation
Dose/Concentration: 20 mg DEP/nr1
Time to Analysis: Exposed 1 h/d from gestation
days 7-19. Mice separated for behavioral
testing on PND 22 (day of delivery is PND 0).
Behavioral testing at 12,16 wks for female
offspring and 13,17 wks for male offspring.
Body wt of exposed Unchanged at birth. Body wt decreased
At weaning. T4
Unchanged dams & Pups @ weaning. At 2 months, exposed
Female pups required Less time to locate
Platform in spatial Reversal task of Morris Water maze.
Reference:
Hougaard et al.
(2008,156570)
Species: Mouse
Gender: Female
(pregnant),
Offspring- male,
female
Strain: C57BL/6
Age: NR
Weight: NR
DEP (SRM 2975)
Particle Size: Surface
area: 90mg2/g, Density:
2.1g/cm3, MMAD:
240nm
Route: Inhalation Chamber
Dose/Concentration: 19.1 ±1.13mg DEP/nr1
Time to Analysis: Pregnant dams exposed GD
7-19,1 h/d. GD 20 named PND 0 for pups.
Weights recorded, 1 pup from each group
sacrificed PND 2. Weights recorded PND 9. PND
22 1 male and female removed from each group
for behavioral testing. Dams and remaining
offspring sacrificed PND 23 or 24.
DEP females gained more weight during gestation.
Generally, DEP pups weighed less. No significant DNA
damage was measured, but DEP caused slightly higher IL-6,
MCP-1, and MIP-2. Plasma thyroxin levels as well as
learning and memory were similar amongst the groups.
Reference: Huang
et al. (2008,
156574)
Species: Rat
Gender: Male
(adults), male and
female (fetuses)
Strain: Wistar
Age: 8 wks (male
adults), 20d of
gestation (fetuses)
ME: Motorcycle Exhaust
(generated from 1992
Yamaha cabin motocycle
with two-stroke 50 cc
engine).
Particle Size: NR
Route: Nose-only Inhalation
Dose/Concentration: 1: 10 and 1: 50 dilutions
Time to Analysis: 2hfd (1 h in morning and 1 h
in afternoon), for 5 consecutive days/wk, for 4
wks (1:50,1:10 dilutions) and 2 wks (1:10
dilution). Male mated with untreated females.
Pregnant females sacrificed on 20d of
gestation. Male and female fetuses observed.
After exposure, decreased body weight and testicular
spermatid number were observed. 1: 10 ME exposure for 4
weeks (no recovery) decreased testicular weight and
increased the inflammatory cytokine mRNA. Glutathione
system and LipidPeroxidation were not affected.
Reference:
Lichtenfels et al.
(2007,097041)
Species: Mouse
Gender: Male and
Female
Strain: Swiss
Age: NR
Ambient air in Sao
Paulo, Brazil
Particle Size: NR
Route: Ambient Air Exposure
Dose/Concentration: NA
Time to Analysis: Males housed in open-top
chambers for 24h/d, everyday for 4m, beginning
10d after birth. Males mated to non-exposed
females immediately following exposure. Males
sacrified immediately following mating.
Pregnant females remain in chamber and
sacrificed on 19d of pregnancy.
Decreased testicular, epididymal sperm counts, decreased
number of germ cells, and decreased elongated spermatids
were observed. Decreased SSR, and a sex ratio shift (fewer
males) also occurred after exposure.
Reference: Mauad PM (busy traffic street
et al. (2008,
156743)
Species: Mouse
Gender: Male,
Female
Strain: BALB/c
Age: 10d
Weight: Parental:
21,4±4.0 -
26.3 ± 2.8g; 15d-old
offspring: 7.8±1.1
- 9.0 ± 1.0g; 90d-
old offspring:
20.3±2.3 -
27.4±1.8q
Sao Paulo, Brazil; Aug.
2005-April 2006) (NO?,
SO?, CO)
Particle Size:
Diameter: 2.5,10 /jm
Route: Ambient Air Exposure
Dose/Concentration: PM2.6: filtered chamber-
2.9±3.0/yglm3, nonfiltered chamber- 16.9±8.3
/yg/m3; Outdoor concentration: PMio-
36.3±15.8/yglm3, CO- 1.7 ±0.7ppm, NO-
89.4±31.9//g/m3, SO?- 8.1 ±4.8//g/m3
Time to Analysis: Nonfiltered exposure 24hfd
4m. Mated at 120d exposure. After birth, 30
females and offspring transferred to filtered or
nonfiltered chamber. Killed 15 or 90d of age.
Mild foci of macrophage accumulations containing black
dots of carbon pigment occurred in the alveolar areas on
90d-old mice. Surface-to-volume ratio decreased from 15 to
90d of age and was higher in mice exposed to air pollution.
PM exposure reduced inspiratory and expiratory volumes at
higher levels of transpulmonary pressure.
July 2009
D-169
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
Mohallem et al.
(2005, 088657)
Species: Mouse
Strain: BALB/c
Gender: Female
Age: 10wks, 10d
Filtered or ambient air in
downtown Sao Paulo
situated at crossroads
with high traffic density
(predominant source of
air pollution is
automotive).
Component
concentrations in
ambient air: CO: 2.2 ±
I.Oppm;	NO2: 107.8 ±
42.3/yg/m3;PMio: 35.5
± 12.8/yg/m3; SO2:
II.2	± 5.3/yg/m3
Component
concentrations in
filtered air: NO2: 19.5 ±
2.8/yg/m3;PMio: 24.1
± 8.1 /yglm3
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: 50 Lfm (polluted
chamber)
Time to Analysis: Exposed for 24 h|7days|wk
for 4 mo. Newborns mated after reaching
reproductive age of 12 wks. All pregnant
females sacrificed between 19th and 20th day of
pregnancy.
No effects in adult exposed animals. Increased implantation
Failure of neonatal Exposed-dams.
Sex ratio, # of Pregnancies,
Resorbtions, Fetal deaths, And fetal placenta Weights
unchanged After neonatal
Ambient air exposure.
Reference: Mori et
DEP:gene
rated by 4-
Route: Dorsal Subcutaneous Instillation
cDNA library screen after
al. (2007, 096564)
cylinder diesel engine
Dose/Concentration: 0.2 ml (of 1.1 mg/ml or
Sub-cut. Injection Found identified activated Clones related
Species: Mouse
Particle Size: NR
0.37 mg/ml)
to Prostanoids 81
Strain: C57/BL


Time to Analysis: 2xfwk for 10 wks.
Arachadonic Acid (Platg2c2c,
Gender: Male


Parameters measured 1wk post last instillation.
AcslG] 81 sperm Production (Stk35). Route of exposure.
Age: 6wks



Unconventional.
Reference: Ono et
al. (2007,156007))
Species: Mouse
Strain: ICR
Gender: Pregnant
females, male
offspring
Age: NR (pregnant
females), 12 wks
(offspring)
DE: generated from 4-
cyl diesel Isuzu engine at
1500 rpm using
standard diesel fuel.
Particle Size: NR
Route: Inhalation (further details NR)
Dose/Concentration:
Time to Analysis: Exposed from 2d post
coitum to 16dpc. Parameters for male offspring
measured on days 8,16, 21, 35, 84 and
sacrificed at 84d.
PND 8 and 16 male repro accessory gl weight decreased.
PND 21 decreased
serum T; PND 84 increased serum T. FSHr, sTAR mRNA
Decreased PND 35 81 84. Rel. testis and epididymal wt
unchanged. Sertoli cell degeneration
Reference: Ono et DE: generated from 4JB- Route: Whole-body Inhalation
al. (2007,156007))
Species: Mouse
Strain: ICR
Gender: Male
offspring, Pregnant
females
Age: 12wks (male
offspring)
2type, light duty 3060
cc 4-cyl Isuzu diesel
engine under 1500 rpm
Particle Size: NR
Dose/Concentration: 1.0 mg DEP/nr
Time to Analysis: Pregnant females
from 2d postcoitum- 16dpc. Without undergoing
further exposure, male offspring sacrificed at
12wks.
Dose-dep increase seminif. Tubule degeneration 81
decreased DSP. 1 mo recovery, DSP recovered at lowest
dose.
Reference:
Pinkerton et al.
(2004, 087465)
Species: Rat
Gender: Female
(pregnant),
Offspring- NR
Strain: SD
Age: 10d (pups),
Pregnant females-
10-14d of gestation
Weight: NR
PM (Fe and soot from
combustion of acetylene
and ethylene in a
laminar diffusion flame
system)
Particle Size: Median
diameter: 72-74nm; size
range: 10-50nm
Route: Inhalation Chamber
Dose/Concentration: Mean mass
concentration: 243±34ug/m3; Average Fe
concentration: 96 /yglm
Time to Analysis: Exposed 10d postnatal a
6hfd, 3d (consecutive). Bromodeoxyuridine
injected 2h before necropsy..
A significant reduction of cell proliferation occurred only
within the proximal alveolar region of exposed animals
compared to controls. There were no significant differences
between the groups for alveolar formation and separation
within the proximal alveolar region
July 2009
D-170
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Silva et
al. (2008,
156981))
Species: Mouse
Strain: Swiss
Gender: Females
(pregnant mice)
Age: 1st, 2nd, 3,d wk
of pregnancy
(females), GD19
(fetuses)
Ambient air Sao	Route: Ambient Air Exposure
Paulo, Brazil	Dose/Concentration:
Particle Size:	Time to Analysis: 1 st week
2nd week
3rd week
or combo of exposure during preg.
Decreased fetal wt with exposure in 1 st week of preg.
Decreased placental Wt with exposure in any of 3 weeks.
Reference: Somers Ambient air
et al. (2002,
078100)
Species: Mouse
Strain: Swiss
Webster
Gender: Male and
Female
Age: 6-8wks (adult
male and females),
5d (pups)
at 2 sites
in Canada (polluted
industrial area 1km
downwind from two
integrated steel mills &
rural location 30km
away)
Particle Size: NR
Route: Ambient Air Exposure
Dose/Concentration: Ambient air
Time to Analysis: Exposed 24hfd, 7dfwk for
10 wks from September 10,1999- November
21,1999. Exposed to clean air for 6wks post-
treatment. Paired with mice within exposure
group. 5d old pups measured.
ESTR germ Line mutations followed .
Heritable mutation rate
increased 1.5 to 2 fold in urban vs. rural site.
Increased freq is paternal line
dependent.
Reference: Somers PM (rural or urban-
et al. (2004,	industrial)
0780981	Particle Size: >0.1
Species: Mouse /jm
Gender: NR
Strain: Sentinal
Lab
Age: NR
Weight: NR
Route: Ambient Air Exposure
Dose/Concentration: Mean TSP: Rural-
16.2±8.3-31.7±13.2/yg|m3, Urban-
Industrial- 38.9± 10.5 - 115.3±25.3//g|m3
Time to Analysis: Exposed 10wks. Bred 9wks
postexposure.
The offspring of urban-industrial mice inherited paternal
ESTR mutations 1.9-2.1 times more than rural or HEPA-
filtered offspring. Maternal ESTR mutations were not
significant.
Reference:
Sugamata et al.
(2006,157025)
Species: Mouse
Strain: ICR
Gender: Pregnant
Females, male and
female offspring
Age: 11 wks
(offspring), NR
(pregnant females)
DE (origins NR)	Route: Inhalation (more specific information NR)
Particle Size: NR Dose/Concentration: 0.3 mg DEP/nr1
Time to Analysis: Pregnant females exposed
from 2d post coitum to 16dpc. Offspring
sacrificed 11 wks after birth.
Exposed pups increased caspase 3 positive cells &
decreased purkinjie cell # (cerebellum). Similar to human
Autism brain phenotype.
Reference: Tozuka
et al. (2004,
090864)
Species: Rat
Strain: Fischer 344
Gender: Pregnant
females, male and
female fetuses
Age: Gestation day
20 (fetuses), NR
(pregnant females)
DE: generated by diesel
engine (309 cc Model
NFAD-50)
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: 1.73mgfm3
Time to Analysis: Exposed 6hfd from GD 7-20
with no exposure on Saturdays or Sundays (4
non-exposure days total). Fetuses and maternal
blood collected on GD20. PAHs: Exposed 6hfd
from GD 7-14 with no exposure on Saturdays or
Sundays. Breast milk collected PND14.
Gest & lactational
Exposure to DE's And PAHs.
7 milk PAHs increased in DE
exposed dams. DE exposure
Can lead to PAH Pup exposure through Breast milk.
Reference: Tsukue DE: generated by 2369 Route: Whole-body Inhalation
et al. (2004,
096643)
Species: Mouse
Strain: Sic: ICR
Gender: Pregnant
females, female
fetuses
Age: Gestation day
14 (fetuses)
cc Isuzu diesel engine
operating at 1050 rpm
with 80% load and using
commercial light oil.
Particle Size: NR
Dose/Concentration: 0.1 mg DEPfm3 (at 1:8
dilution with clean air)
Time to Analysis: Exposed for 8hfd from 2d
postcoitum to 13dpc (with no exposure on days
4, 5,11,12). Sacrificed 14dpc. Only female
fetuses studied.
SF-1 & MIS mRNA no change. Other steroidogenic
Genes unchanged. BMP-15, oocyte Differentiation
mRNA decreased.
July 2009
D-171
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Tsukue
et al. (2002,
030593)
Species: Mouse
Strain: C57/BL
Gender: Females,
male and female
offspring
Age: 6wks, 70d
post natal
(offspring)
Filtered air
DE: generated by light-
duty, 4-cylIsuzu diesel
engine at 1500 rpm.
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: 0.3,1.0 or 3.0
mg DEP/m3
Time to Analysis: Exposed 12hfd, 7dfwk for 4
mo. Some females sacrified immediately
following exposure. Remainder mated with
unexposed males. Parameters measured in
offspring at postnatal day 70.
DE-exposed females had
Decreased uterine wt at 4 months. Offspring decreased
Body weight @ 6 &
8 weeks of age. Decreased rate Of good nest Construction
(3 mg/m3).
AGD decreased In males (30 & 70 days old). Organ weight
Decreased in Females. Female Crown to rump
Length decreased. Relative weights Not taken.
Reference: Ueng et
al. (2004, 096199)
Species: Mouse
Gender: Female
Strain: Wistar
Age: 21 d
Cell Line: MCF-7
ME: generated from a
Yamaha Cabin
motorcycle 2-stroke 50-
cc engine and variable
venture carburetor
Particle Size: NR
Route: Intraperiotenial Instillation. Cell Culture
Dose/Concentration: IP: 1,10, 50//gfml
Cell Culture: 0.01, 0.1,1,10, 50,100//gfml
Time to Analysis: IP: 1fd for 3d and sacrificed
on 24d. Cell Culture: 3, 24, 30, 48h and 2d.
10 mg/kg +E2 induced
anti estrogenic uterine effects & Antiestrogenic with in vitro
(MCF-7 cells) E2 screen.
Reference: Veras
et al. (2008,
190493)
Species: Mouse
Gender: Male,
Female
Strain: BALB/c
Age: 20d,
newborns
Weight: NR
PM (downtown Sao
Paulo, Brazil near
crossroads with high
traffic density, 67%
PM2.6 comprises air
pollution)
Particle Size:
Diameter: 2.5 /jm
Route: Open-Top Chamber
Dose/Concentration: PM2b- 27.5/yg/m3; N02-
101 /yg/m3; CO- 1.81 /yg/m3; S02- 7.66ppm
Time to Analysis: 20d-old mice maintained in
filtered or nonfiltered chamber until 60d-old.
Offspring maintained in respective chambers
until 21 d-old. Offspring mate at 60d-old.
Females euthanized 18th GD.
Fetal weight and maternal blood space volume and surfaces
declined in the groups exposed to nonfiltered air. Fetal
capillary surfaces were greater in nonfiltered air groups.
There was a significant gestational effect on maternaLfetal
surface ratios with values declining significantly in groups
exposed during pregnancy to nonfiltered air. The total
oxygen diffusive conductance of the intervascular barrier
increased significantly during pregestational exposure to
nonfiltered air. Mass-specific conductance increased during
pregestational and gestational periods of exposure to
nonfiltered air.
Reference: Veras
et al. (2009,
190496)
Species: Mouse
Gender: Male,
Female
Strain: BALB/c
Age: 20d
Weight: NR
PM (Sao Paulo, Brazil:
near crossroads with
high traffic density) (Al,
Ca, Cu, Fe, K, Na, Ni, P,
Pb, S, Si, Ti, V, Zn, C)
Particle Size:
Diameter: 2.5 /jm
Route: Open-Top Chamber
Dose/Concentration: Mean: Non-filtered- 27.5
/yg/m3, Filtered- 6.5 /yglm3
Time to Analysis: 20d-old mice maintained in
filtered or non-filtered chamber. Allowed to
mate at 60d. 2 generation model.
Ambient air pollution extended the estrus cycle, which
reduced the number of cycles. Antral follicles decreased.
Mating time increased and fertility and pregnancy indices
decreased. The mean post-implantation loss rate increased,
which was influenced by both pre- and post-gestational
exposure. Fetal weight decreased and was also influenced
by pre- and post-gestational exposure, which exhibited a
significant interaction.
Reference:
Watanabe (2005,
087985)
Species: Rat
Gender: Female
(pregnant),
Offspring- male
Strain: F344/DuCrj
Age: 7d of
gestation -
parturition
(females), 96d
(offspring)
Weight: 240-262g
(offspring)
DE (309cc engine,
Model NFAD50, Yanmar
Diesel Co., Osaka,
Japan, 1800rpm, 45%
load) (PM, NO2)
Particle Size: 90%
<0.5/ym
Route: Inhalation Chamber
Dose/Concentration: High dose total group:
PM- 1.71 /yglm3, N02- 0.79ppm; Low dose total
group: PM- 0.17 /yglm3, N02- 0.10ppm
Time to Analysis: Pregnant rats exposed
gestational day 7 to delivery 6hfd. 5 groups:
high dose total DE, high dose PM, NO2 filtered,
low dose total DE, low dose PM, NO2 filtered,
clean air control. Offspring sacrificed day 96
after birth.
All groups had significantly less daily sperm production than
the control. PM and NO2 in DE decreased spermatogenia but
was not significant, however the high dose PM filtered
group achieved significance. Pachytene cells, spermatids,
and Sertoli cells were lower in all groups compared to the
control.
July 2009
D-172
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Yauk et
al. (2008,157164)
Species: Mouse
Strain: C57BL/6 x
CBA F1 hybrid
Gender: Male
Age: 7-9wks
Hepa-Filtered air (PM
removed) and ambient
air at 2 sites:
¦2km from two
integrated steel mills
¦1km from major
highway on Hamilton
Harbor
Components:
Metals 3.6 ± 0.7 //g/m3
TSP 9.4 + 17 //g/m3
Particle Size: NR
Route: Ambient Air Exposure
Dose/Concentration: Ambient air
Time to Analysis: Parameters measured 3,
10wks, or 10 + 6wks recovery following
exposure.
10+6 weeks exposure induced increased ESTR mutations in
sperm DNA of exposed v filtered. No testicular DNA adducts
seen in exposed males. @ 3wks DNA increased adducts
seen in lungs of exposed males, not in filtered males.
Mutations PM dependent Gas phase independent.
Reference: Yokota DE (2369-cc diesel
et al. (2009,
190518)
Species: Mouse
Gender: Female
(pregnant), Male
(offspring)
Strain: ICR
Age: NR
Weight: NR
engine, Isuzu Motors,
Ltd., Tokyo, Japan;
1050rpm, 80% load,
commercial light oil)
Particle Size: NR
Route: Inhalation Chamber. Pre natal Exposure
Dose/Concentration: DE: I.Omgfm3; CO:
2.67ppm, NO2: 0.23ppm, SO2: < 0.01 ppm
Time to Analysis: Pregnant mice exposed 8h
for 5d from GD 2-17. Mothers and pups kept in
clean room. Pups weaned on PND 21 then
transported to Tokyo University of Science.
2wk acclimation. Exposed 12h light/dark cycle.
Activity monitor with infrared ray sensor
measured spontaneous motor activity (SMA),
10min intervals 2d. After behavioral test, mice
decapitated.
Prenatal DE exposure decreased SMA in the male offspring.
DE decreased locomotor activity during the light phase.
Dopamine levels in the striatum and nucleus accumbens did
not change, but HVA concentrations decreased in DE-
I mice.
Reference: Yoshida DEfgenerated from a 4-
et al. (2006,
156170)
Species: Mouse
Strain: ICR,
C57BI/6J or DDY
Gender: Pregnant
Females, Male
fetuses
Age: 14d of
gestation (fetuses),
2- 13d of gestation
(pregnant females)
cyl., 2300 cc diesel
Isuzu engine at 1050
rpm and 80% load).
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: 0.1 mgDEPfm3
Time to Analysis: Exposure on 2-13d of
gestation. Parameters measured on 14d of
gestation.
Responses to exposure showed strain-related variations
with ICR as the most sensitive followed by C57 and ddY as
the least sensitive. MIS mRNA expression, a factor in male
gonadal differentiation, was significantly decreased in the
ICR and C57 strains. Ad4BP/SF-1 expression was
significantly decreased in the ICR strain only.
Reference: Yoshida
et al. (2006,
097015)
Species: Mouse
Strain: ICR
Gender: Pregnant
females and male
offspring
Age: 2-16d
postcoitum
(pregnant females),
28d (male offspring)
DE: generated by 4Jb1-
type, light duty 4-
cylinder Isuzu diesel
engine using standard
diesel fuel at 1500 rpm.
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: 0.3,1.0 or 3.0 mg
DEP/m3
Time to Analysis: Pregnant females exposed
12h/d, 7d/wk from 2-16d postcoitum. Offspring
sacrificed on postnatal day 28.
NOAEL 0.3 mg DEP/m3 DE exposure induced increased
repro gland weight (two higher doses) male mice. mRNA
decreases aromatase & 3 /y-hD (3.0 mg DEP/m3)
Sex ratio no change. 2 higher doses induced sig increased
repro organ wt. Male pup wt Increased at PND 28. 1.0mg
DEP/m3 pup increased serum T. Serum T positively
correlated with DSP, testis wt, steroid enzyme mRNA.
Reference: Yoshida
et al. 2004 (2004,
097760)
Species: Mouse
Gender: Female
(pregnant),
Offspring- male
Strain: ICR
Age: 4, 6wks
Weight: NR
DE	Route: Inhalation Chamber
Particle Size: NR Dose/Concentration: 6wk-old males, embryos:
0.3,1.0, 3.0mg DEP/m3, Pregnant mice: 0.1,
3.0mg DEP/m3
Time to Analysis: 6wk-old males: Exposed
12h/d, 6m. 1 m clean air exposure. Pregnant
mice: Exposed 2-13p.c. 8h/d. Male embryos:
Exposed 2-16p.c. Examined at 4wks-old.
6wl<-old males: In the seminiferous tubules, DE dose-
dependently caused degenerative and necrotic changes,
desquamation of the seminiferous epithelium, and loss of
spermatozoa. Spermatogenesis was still inhibited after a 1 m
clean air exposure.
Pregnant mice: Ad4BP/5F-1 and MIS mRNA significantly
and dose-dependently decreased in male fetuses exposed to
DE.
4wl<-old male newborns: Tissue weight of the testis and
accessory reproductive glands were significantly greater in
DE-exposed mice. Blood testosterone concentration was 8X
higher than the control at 1 .Omg DEP/m3. No significant
differences occurred for testosterone synthetase mRNA.
July 2009
D-173
DRAFT-DO NOT CITE OR QUOTE

-------
D.1. Carcinogenisis, Mutagenesis, Genotoxicity
Table D-7. Mutagenic/genotoxic effects in bacterial cultures.
Reference
Pollutant
Exposure
Effects
Reference: Binkova
et al. (2007,
156273)
Species:
Salmonella (±S9
(rat liver))
Cell Line: Calf
thymus DNA
PM (Prague, Kosice, Sofia,
Czech Republic; summer,
winter) (organic extracts)
Particle Size: Diameter:
< 10/ym
Route: Cell Culture
Dose/Concentration: 100//g EOM/mL
Time to Analysis: PM collected 24h daily
3m, extracted. 24h incubation BaP, c-PAH,
EOM, with or without S9.32P-Postlabeling
4h. Autoradiography 1 -24h.
DNA adducts in EOM treatments were greater with S9 than
without. Positive correlations were found between the amount
of DNA adducts and the PAH content (notably BaP) in the
EOM treatment.
Reference: Brits et
al. 2007
Species: S.
typhimuriam
Strain: TA98±S9
(Ames); TA104
recN2-4 and
TA104pr1 (Vitotox)
Cell Line: Human
whole blood (Comet,
MN assays)
PM (Flanders, Belgium;
urban, rural, industrial sites)
(organic extracts)
Particle Size: Diameter: 10
/jm
Route: Cell Culture
Dose/Concentration: 2.5, 5,10, 20m3 air
equivalents/mL
Time to Analysis: Air samples extracted.
Ames assay 48h. Vivotox test. Comet
assay 24h. MN assay.
Ames: S9 induced mutagenicity of all extracts from all areas
in a dose-dependent manner. Without S9, only extracts from
the urban and industrial areas were mutagenic at the highest
dose.
Vitotox: Extracts were toxic at the highest dose.
Comet: Significant DNA damage in the extracts was seen and
enhanced by S9.
MN: A dose-response relationship was seen in the urban
extracts for increased micronucleated binuclear cells.
Reference: Brown PM (New Zealand, summer, Route: Cell Culture
et al. (2005,
095919)
Species: S.
typhimuriam
Strain:TA98
Cell Line: Rat
hepatoma H4IIE
winter) (extracts)
Dose/Concentration: 9.7-20.8//gfm3
Particle Size: Diameter: 10 (summer), 21.8-61 /yg/m (winter)
Generally, the mutagenic rate was positively correlated to
PMio, as well as PAH and BaP. PMio levels were higher and
more mutagenic in winter than summer.
fjm
Time to Analysis: Air samples collected
15d, extracted. Ames test: Bacteria
growth 12h, incubated 24h. Hepatoma
bioassay: 24h incubation 2x. ER0D assay.
Reference: Bunger DEP (diesel fuel (DF), low- Route: Cell Culture
et al. (2006,
156303)
Species:
Salmonella
typhimuriam
Strain: TA98, TA
100
sulfur diesel fuel (LSDF),
rapeseed oil methyl ester
(RME), and soybean oil
methyl ester (SME)) (SOF-
soluble organic fractions)
Particle Size: Total
particulate matter (no 0CC)
(gh'): Mean DF- 4.0 ± 0.2;
2.8 ± 0.5; 1.8 ± 0.0; 3.4
± 0.2; 1.2 ± 0.1
Dose/Concentration: Log 2 dilutions of
extracts: 1.0, 0.5, 0.25, 0.125
Time to Analysis: S0F extracted 12h.
Plates incubated 48h.
No OCC: Without oxidation catalytic converter (0CC), DF
extract produced the highest number of revertant colonies at
all load modes in both TA98 and TA100 ± S9. RME, SME, and
LSDF extracts caused lower or no mutagenic effects, seen
especially at partial load modes and idle motion.
OCC: With OCC, all extracts reduced the number of revertant
colonies in TA98 and TA100 ±S9 at partial load modes B, C,
and D. At load mode A (rated power), there was an increase of
the number of revertant colonies in all assays -S9, significant
for extracts from RME (TA98, TA100) and SME (TA98). S9
lowered frequency of mutations. At load mode E (idling),
number of revertant colonies of DF extracts increased ±S9.
Reference: Bunger Diesel engine emissions Route: Cell Culture
et al. (2007,
156305)
Species:
Salmonella
typhimuriam
Strain: TA98, TA
100
(DEE)-rapeseed oil (RSO)
and rapeseed methyl ester
(RME, biodiesel), natural
gas derived synthetic fuel
(GTL), and diesel fuel (DF)
(SOF- soluble organic
fractions)
Particle Size: NR
Dose/Concentraion: Log 2 dilutions of
extracts: 1.0, 0.5, 0.25, 0.125
Time to Analysis: SOF extracted 12h.
Plates incubated 48h.
Compared to DF, RSO significantly increased mutagenic
effects of particle extracts (i.e., revertants) by 9.7-59 in TA98
and by 5.4-22.3 in TA100. (mRSO, RSO with lowered
viscosity and fuel preheating in tank, produced highest number
of revertant colonies in both strains ±S9.) RSO fuels
condensates had 13.5 times stronger mutagenicity than DF.
RME extracts had moderate but significantly higher mutagenic
response in TA98 +S9 and TA100 -S9. Effects of GTL did
not differ significantly from DF.
Reference: de Kok
et al. (2005,
088656)
Species: S.
typhimurium
Strain: TA98 (with
and without rat liver
S9)
Cell Line: Salmon
testis DNA
TSP (Total suspended
particulate, Maastricht,
The Netherlands; PMio and
PM2.6 from 6 urban
locations with different
traffic intensities.)
(organic extracts)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Mutagenicity assay:
2.5, 9, or 18m3 sampled air in 100 //L
DMSO; DNA adduct assay: 5 /jL DMSO
containing PMio or TSP from equivalent
50m3 sampled air. PM2.6 concentration
equivalent to 35m3 sampled air.
Time to Analysis: Mutagenicity assay:
Cells incubated 1h with extracts. DNA
adduct assay: DNA incubated 4h with
extracts.
Overall, the direct mutagenicity and DNA reactivity of PM2.6
extracts were higher compared to PMio and TSP. S9 generally
reduced mutagenic activity in TA98 but increased reactivity to
Salmon testis DNA. Total PAH and total carcinogenic PAH
levels correlated with the mutagenicity of TSP and the S9-
mediated mutagenicity of PM2.6. Neither transition metal
composition nor radical generating capacity of PM correlated
with mutagenic potential. Total PAH and carcinogenic PAH
levels from PMio and PM2.6 correlated with direct and S9-
mediated DNA adducts; for TSP these levels correlated with
direct DNA reactivity only.
July 2009
D-174
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:
DeMarini et al
(2004, 066329)
Species:
Salmonella
Strain: TA98,
TA98NR, TA98/1,
8-DNP6, YG1021,
YG1024, TA100
A-DEPand forklift DEP
(SRM 2975)
DEP (EOM)
Particle Size: Mean
Diameter: 0.4 /jm
Route: Cell Culture
Dose/Concentration: 0, 0.25, 0.5,1.0,
2.0 EOM /yglplate
Time to Analysis: DEPs sonicated 20min.
Centrifuged 10min. Organic material
extracted and concentrated. Ames assay.
Incubated 3d.
A-DEPs were more mutagenic in both TA98 and TA100 than
SRM 2975. There was 22x more PAH-related and 8-45x
more nitroarene-related activity.
Reference: El
Assouli et al. (2007,
186914)
Species: S.
typhimuriam
Strain: TA98
(±S9)
PM (Jeddah, Saudi Arabia;
11 sites, urban, winter)
(organic extracts)
Particle Size: Diameter: 10
/jm
Route: Cell Culture
Dose/Concentration: 2.5, 50,100
/yg/plate; EOM range: 6-40//g|m3
Time to Analysis: 24h air samples,
extracted. Refluxed 18-24h. GC-MS.
Comet assay. 48h incubation. Ames assay.
PAHs varied from 0.83 to 0.18 ng/m3. Only 2 locations of
heavy petrol driven cars showed strong genotoxic responses.
A correlation existed between DNA damage and the amount of
pollutants and PAHs. Toxicity and mutagenicity occurred only
in the presence of S9. Only 3 of the 11 sites exhibited
moderate mutagenic activities.
Reference: Endo et
al. (2003, 097260)
Species: S.
typhimuriam
Strain: YG1024
(±S9)
PM (Tokyo, Japan: winter)
(organic extracts)
Particle Size: Diameter:
>12.1 - 0.06 > /jm;
Bimodal mass
concentration: 1-2 /jm
Route: Cell Culture
Dose/Concentration: 2.5, 5,10//L; 0.30
- 22.76/yglm3
Time to Analysis: Air samples collected,
extracted. 90min pre incubation. 48h
incubation.
Mutagenicity tests showed dose-response relationships that
were higher without S9 and increased with decreasing size.
Reference:
Erdingeret al.
(2005,156423)
Species: S.
typhimurium
Strain: TA98.
TA100, TA98NR
PM (Baden-Wurttemberg,
Germany: urban, 8
locations, glass fiber filters)
(organic extracts)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 0.25, 2.5, 5,12.5,
25 m3/plate
Time to Analysis: Standard Ames test
protocol followed.
Extracts were mutagenic in all strains evaluated. No
significant difference in response with or without metabolic
activation. Activity in TA98NR suggests that the mutagenicity
correlates with concentrations of air pollutants such as NOx.
Reference:Iba et
al. (2006,156582)
Species: S.
typhimuriam
Strain: TA98,
TA100 (±S9 (rat
liver))
PM (wood smoke (WS)
(New Jersey) and cigarette
smoke (CS) (Tobacco
Research and Health
Institute, University of
Kentucky)) (organic
extracts)
Particle Size: 10 //I
aliquots of organic extracts
Route: Cell Culture
Dose/Concentration: 62.5,12.5/yg TPM
equivalent/plate
Time to Analysis: Incubation, shaking
25min. Agar added. 48h incubation. Rat
lung explants incubated 18h. 12h
incubation with treatments.
WS and CS were equally mutagenetic to TA98, but CS was 3-
fold more mutagenetic to TA100 than WS. CS induced
CYP1A1 in the explants, but WS did not.
Reference: Liu et
al. (2005, 097019)
Species: S.
typhimurium
Strain: YG1024,
YG1029
Cell Line: Chinese
hamster lung V79
cells
DEP extract (DP), gasoline
engine exhaust particulate
extract (GP), diesel exhaust
SV0C extract (DSV0C),
gasoline engine SV0C
extract (GSV0C), NIST SRM
1650a
Particle Size: Gasoline PM:
0.554mg extract (mg PM)"1;
Diesel PM: 0.363mg extract
(mg PM)1
Route: Cell Culture
Dose/Concentration: 1.48,4.44,13.3,
40,120, 360,1080 //g/plate
Time to Analysis: 30min preincubation.
48h (YG1029). 66h (YG1024). Overnight
preincubation 20h.
Mutations: All samples induced mutations in both strains.
The increase was highly significant and dose-dependent.
Response with S9 was generally greater than without S9. PM
extract was more mutagenic than SV0C extract.
DP, GP, and GSVOC: Dose-response was seen for DNA
damage and micronuclei induction. GP, GSVOC and SRM
1650a were stronger inducers of micronuclei than DP.
Reference:
Matsumoto et al.
(2007,187020)
Species: S.
typhimuriam
Strain: TA98,
TA100 (±S9)
APM (airborne particulate
matter)
APE (airborne particulate
extracts)
(Hokkaido, Japan;
residential)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Crude APE:
979mg/m3 air (CALUX BaP Equivalent
(BaPEq)), 21mg/m3 air (CALUX TCDD
Equivalent (TCDDEq)); Cleaned APE:
7.87mg/m3air (CALUX BaPEq),
0.614mg|m3 air (CALUX TCDDEq)
Time to Analysis: Air samples collected,
extracted. Preincubation with S.
typhimuriam. 3, 24h exposure in CALUX
assay. RNA extracted from mice 6d after
last application.
Most of the CALUX BaPEq for crude APE was derived from
PAH-like compounds, as suggested by the CALUX BaPEq of
cleaned APE accounting for 0.80% of CALUX BaPEq for crude
APE. CALUX TCDDEq showed TCDD and similar compounds
to have a low contribution. The TA100 strain was more
mutagenic to APE, with and without S9. S9 increased
mutagenecity in both strains.
July 2009
D-175
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference:	PM (EOM) (Plzen, Prague, Route: Cell Culture
«nSn0,rkn°o^8la!- D'" m-"? " Cz8ch	Dose/Concentration: TA98 (4 doses): 20-
(2004,087431)	Republic)	200//g/plate, YG1041 (4 doses): 4-20
Species: 5	Particle Size: Diameter: 10 //gfplate
typhimuriam	m	Time to Analysis: Collected 24h every
Strain: TA98,	18th day, Oct-Mar, 1999-2003. Extracted.
YG1041 (±S9)	Ames assay. 70h incubation.
Significant dose-response effects in mutagenic potency of
EOM occurred. Prague, one of the most polluted cities, had
the highest mutagenicity values. Increasing time-trends were
observed in the TA98±S9 mutagencity and PAH
concentrations.
Reference: Rivedal DEP (SRM 1650)(organic Route: Cell Culture
et al. (2003,
097684)
Species: S.
typhimurim
Strain: TA100,
TA98, TA100NR,
TA98NR, TA98/1,8-
DNPo
extracts) (fractionated into
PAH, nitro-PAH, dinitro-
PAH, aliphatics, polar
fraction)
Particle Size: NR
Dose/Concentration: Ames: 300, 600
DEP/plate; Gap junction: 100, 200 /yglmL
DEP
Time to Analysis: Extracted 16h.
Fractionated. Ames assay. Gap junction
intracellular communication: exposed 1-6h.
Western blot.
TA100 was the most mutagenic without S9 activation. GJIC
was dose- and time-dependently inhibited. The polar fraction
was the most potent inhibitor. Nitro-PAH and dinitro-PAH
were the most responsive fractions in the Ames assay.
Reference:
Seagrave et al.
(2003, 054979)
Species:
Salmonella
Strain: TA98,
TA100
Compressed natural gas
(CNG) emissions (heavy-
duty vehicles): High emitter
(HE), Normal emitter (NE),
New technology (NT)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: PM (mg/mi)- NE-
7.0, NT- 5.0, HE- 406; Recovered PM
(mg/mi)- NE-1.26, NT- 0.71, HE-57.1;
Recovered SV0C- NE- 58, NT- 26.4, HE-
227.5
Time to Analysis: Samples collected in
filters 7x/d over several days. Recovered
PM, recovered SV0C extracts combined.
Ames assay.
All three CNG emissions were mutagenic in both strains.
Mutagenicity was reduced by S9 in TA100 but not in TA98.
Activity ranking in both strains was HE > NE > NT.
Reference: Sharma
et al. (2007,
156975)
Species:.?
typhimurium
Strain: TA98,
YG1041, YG5161
Cell Line: Human
A549 lung epithelial
cells
PM (airborne, 4 sites: an
oven hall and receiving hall
in a waste incineration
plant: heavy-traffic street:
background: Mar-June
2005)
Particle Size: Diameter:
2.5 /jm
Route: Cell Culture
Dose/Concentration: 0.25mgfml
Time to Analysis: Samples taken over 7-
16d. A549 cells incubated 24h. Comet and
microsuspension assays performed.
DNA damage: Samples from all four sites induced DNA
damage in the comet assay with the street samples more
damaging than the oven hall sample.
Mutations: Microsuspension assay was used to assess
mutagenic activity. No mutagenic activity was observed for
any of the non-polar fractions from any sample sites. The
moderately polar fractions were all mutagenic, except for the
oven hall sample, only when S9 was added. Comparatively,
the polar and crude fractions were mutagenic without
metabolic activation, suggesting a direct mutagenic effect.
Reference: Song et
al. (2007,155306)
Species: S.
typhimurium
Strain: TA98,
TA100
Cell Line: Rat
fibrocytes L-929
cells
PM (soluble organic fraction Route: Cell Culture
(S0F) extracts from diesel
engines using fuels blended
with ethanol by volume: E0
¦ base diesel fuel: E5 ¦ 5%:
E10- 10%;E15-15%; E20
¦20%)
Particle Size: Density
(g/cm3): E0- 0.8379; E5-
0.8349; E10-0.8324; E15-
0.8301; E20- 0.8279
Dose/Concentration: Ames Assay:
0.025, 0.05, 0.1 mg/plate; Comet Assay:
0.125, 0.25, 0.5,1.0mg|mL
Time to Analysis: Samples extracted
24h. Ames and comet assays performed
All PM extracts induced higher mutational response in TA98
(3- to 5-fold increase over spontaneous) than in TA100 (2-to
3-fold increase). The highest brake specific revertants (BSR)
±S9 in both strains occurred with E20 and lowest BSR was in
E5 (except in TA98 -S9). E0 and E20 caused more significant
DNA damage (similar in effect) than the other extracts.
Damage was dose-dependent but variable with increasing
ethanol volume.
Reference: Zhang
et al. (2007,
157186)
Species: S.
typhimurium
Strain: TA98,
TA100
Cell Line: Human
lung
adenocarcinoma
A549 cells
Gasoline engine exhaust
(GEE)
Methanol engine exhaust
(MEE)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: MTT Assay- 0.05-
0.8 GEE or MEE L/ml; MN Assay- 0.025,
0.05, 0.1, 0.2 GEE or MEE L/ml; Comet
Assay-0.025, 0.05,0.1,0.2, 0.4 GEE or
MEE L/ml; Ames Assay- GEE: 0.625,1.25,
2.5, 5.0,10, 20 L/plate; MEE: 0.3125,
0.625,1.25, 2.5, 5.0,10, 20 L/plate
Time to Analysis: Organic extracts from
GEE and MEE. MTT assay- 24h incubation,
followed by 2 or 24h incubation, followed
by 4h incubation. MN assay- 24h
incubation. Comet assay. Ames assay- 72h
incubation.
Mutagenicity: GEE was mutagenic in TA98 but not TA100, -
S9 at 10 and 20 L/plate and +S9 at £1.25 L/plate.
Mutagenicity was higher with S9 than without at 0.625-10
L/plate and a dose-response was reported. MEE had no effect
in either strain.
MN: GEE significantly and dose-dependently induced MN. MEE
had no significant effect at any dose.
DNA damage: GEE significantly induced DNA damage at all
doses compared to controls. MEE had no effect at any dose.
July 2009
D-176
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Pollutant
Exposure
Effects
Reference: Zhao et DEP (SRM 2975)
al. (2004,100972)
Species: Rat
Gender: Male
Strain: Sprague-
Dawley
Age: NR
Weight: — 200g
Cell Line: S.
typhimurium
YG1024 (±S9)
DEPE (SRM 1975)
Carbon black (CB) (Elftex-
12 furnace black, Cabot,
Boston, MA)
Particle Size: NR
Route: IT Instilled. Cell Culture.
Dose/Concentration: DEP or CB:
35mg/kg body weight; S9: 25, 50,100,
200 /yg/plate; Cytosolic protein: 20,40,
80,160 /yglplate; Microcosmal protein: 5,
10, 20, 40 /yglplate
Time to Analysis: Rats instilled.
Sacrificed 1, 3, 7d postexposure. S9,
cytosolic, microcosmal fractions prepared
from lung homogenates. Ames assay: 72h
incubation.
DEP and CB-exposed lung S9 time-dependently decreased 2-
aminoanthracene (2-AA) mutagenicity. Metyrapone and a-
napthoflavone inhibited the S9-activation of 2-AA in DEP and
CB exposed rats. Lung S9 increased the mutagenicity of DEPE
but not of DEP or CB. Liver S9 reduced DEPE dose-
dependently. CYP2B1 and CYP1A1 activated DEPE, with
CYP2B1 being more effective.
Reference: Zhao et
al. (2006,100996)
Species: S.
typhimuriam
Strain: YGL024
(±S9)
DEP (SRM 2975)
DEPE (SRM 1975)
Aminoguanidine (AG)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: NR
Time to Analysis: Lung S9 obtained from
rats used in in vivo experiment. Ames test.
Modified microsuspension assay. All
assays in duplicate plates. Repeated 3x.
AG significantly lowered 2-aminoanthracene mutagenic
activity of DEP or DEPE-exposed lung samples, with DEP being
lowered the most.
Table D-8. Mutagenicity and genotoxicity data summary: in vitro studies.
Reference
Particle
Exposure
Effects
Reference: Abou Chakra
et al. (2007, 098819)
Species: Human
Gender: Male, Female
Age: 6-13yrs and Adults
Participant
Characteristics: Non-
smokers
Cell Line: HeLa S3 cells
PM (3 French metropolitan cities:
Urban PM2.6 and PM10 from
"Residential Sector," "Proximity
Sector," "Industrial Sector'!
(organic extracts)
Particle Size: Diameter: 2.5,10 //m
Route: Cell Culture
Dose/Concentration: 261 PM2.6; 76
PMiosamples
Time to Analysis: Cells incubated with
200; fjL organic extract and 20 /jl
aphidicoline for 24h.
Seasonal variation was observed with genotoxic
effects being greater in winter. PM2.6 was more
active than PM10 extracts. Samples from the
"Proximity Sector" (downtown area with heavy
traffic) exhibited the strongest genotoxic
responses.
Reference: Arrieta et al.
(2003,096210)
Species: Rat
Cell Line: Hepatoma
(H4IIE)
Species: Mouse
Cell Line: Hepatoma
H111.1 c2
PM (El Paso, Texas: Juarez,
Chihuahua, Mexico; Sunland Park,
New Mexico) (organic extracts)
Particle Size: Diameter: 10 //in
Route: Cell Culture
Dose/Concentration: EROD test: 0.03,
0.17,0.34,0.50, 0.68, 4.96, 9.93
extract equivalents (m3 air); Luciferase:
0.17,0.51,1.26, 5.01 extract
equivalents (m3 air)
Time to Analysis: Extracts incubated
24h. EROD, luciferase activity, PAH
content determined.
EROD activity declined at higher extract
amounts, but luciferase activity was not
inhibited. Cytotoxicity occurred only at extract
equivalents to 0.47 m3air. PAH concentration
increased with PM mass.
Reference: Bao et al.
(2007,097258)
Cell Line: Human-
hamster hybrid (Al)
DEP (organic extracts) (SRM 2975)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 10, 20, 50,100
/yg/mL
Time to Analysis: Phagocytosis
inhibitors: Exposed 24h with or without
cytochalasin B or ammonium chloride.
Cytotoxicity: 24, 48h incubation.
Mutations: Exposed 24h. 5-7d culture.
Incubated additional 7-8d.
The nucleus of DEP-treated cells was condensed
and shrunken compared to controls. DEPs
accumulated in cells, disrupting the
mitochondrial cristae, and were lodged in large
cytoplasmic vacuoles. DEP produced minimal
toxicity. CD59 locus mutations dose-
dependently increased but decreased when
simultaneously treated with cytochalasin B or
ammonium chloride.
Reference: Carvalho-
Oliveria et al. (2005,
077898)
Species: T. pallida, A.
cepa
PM (Sao Paulo, Brazil; spring, bus
strike and non-strike days) (organic
extracts)
Particle Size: Diameter: 2.5 /jm
Route: Cell Culture
Dose/Concentration: Strike day: 47.32
/yg/m3; Non-strike day: 43.01 /yglm3
Time to Analysis: Air samples from 2d:
bus strike day, bus non-strike day. T.
pallida kept in lab 24h. Exposed 8h. 24h
recovery. Fixed 24h. A. cepa roots
induced 5d. Exposed 30h. Fixed 24h.
Element concentrations, sulfur and BTEX
decreased on the strike day. Micronuclei
decreased in T. pallida during the strike. Toxicity
measured in A. cepa was not significant, but
higher on strike days.
July 2009
D-177
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle
Exposure
Effects
Reference: Dybdahl et al. DEP (SRM 1650)
(2004, 089013)	Particle Size: NR
Species: Human
Cell Line: Lung epithelial
A549
Route: Cell Culture
Dose/Concentration: 10, 50,100, 500
/yg DEP/mL
Time to Analysis: DEP suspended,
sonicated. A549 cells diluted. Fresh
medium added after 24h. After 48h
medium removed, DEP added. 2, 5, 24h
incubation.
DEP induced dose-dependent increases of IL-1 a,
IL-6, IL-8, TNF-a. The cytokines increased 4-18-
fold at the highest dose. Cell viability did not
decrease. Comet tail length increased at 100
and 500 /yglmL for 2, 5, 24h.
Reference: Gabelova et
al. 2006
Species: Human
Cell Line: Hepatoma Hep
G2
PM (winter, summer) (organic
extracts)
Particle Size: Diameter: 10 //in
Route: Cell Culture
Dose/Concentration: 5,10, 20, 50,
100,150/yg|mL
Time to Analysis: 3m sampling periods,
winter, summer. Cells grown 48h.
Exposed 2h. Single cell gel
electrophoresis or cultivated, harvested,
processed 2, 4,16, 24h.
PM, c-PAHs and genotoxicity were higher in
winter air samples than summer. E0M samples
generally had significant dose-dependent
increases in DNA migration. Repair-specific DNA
endonucleases did not increase DNA migration.
8-oxodG was below the steady-state level in
E0M samples.
Reference: Gabelova et
al. 2007
Species: Human
Cell Line: Hepatoma Hep
G2
PM (PRG-SM, PRG-LB, Kosice, Sofia;
winter, summer) (organic extracts)
Particle Size: Diameter: 10 //m
Route: Cell Culture
Dose/Concentration: 5-150//cj/niL
Time to Analysis: Air samples collected
24h intervals, 3m sampling period. Cells
grown 48h. Exposed 2h. 24, 48h
preliminary experiments. Single cell gel
electrophoresis.
Cell viability significantly decreased in the 24,
48h exposure groups compared to the 2h
exposure group. DNA migration significantly
dose-dependently increased at most
concentrations. In general, oxidative DNA
damage did not significantly increase.
Reference: Gabelova et
al. 2007
Species: Human
Cell Line: Hepatoma Hep
G2 cell line
PMio (Prague (Czech Republic),
Ko'sice (Slovak Republic) and Sofia
(Bulgaria); urban, winter, summer)
(organic extracts)
Particle Size: Diameter: 10 //m
Route: Cell Culture
Dose/Concentration: 5 to 150/yglml
E0M from 50/yglml stock solution
Time to Analysis: 24h DNA adduct
formation. 2h Comet assay. Oxidative
DNA damage measured by Fpg-sensitive
sites.
Total DNA adducts ranged from ~ 60 to 200
adducts per 10s nucleotides. Extracts also
produced approximately the same levels of
strand breaks. Results suggested that the
genotoxic potential of ambient air was at least
6-fold greater in the winter compared to
summer. No substantial difference was reported
for oxidative DNA damage induced by summer
vs. winter samples.
Reference: Gong et al.
(2007,091155)
Species: Human
Cell Line: Microvascular
endothelial (HMEC)
DEP (aggregates, exhaust 4JBi-type
LD,274 1,4-cylinder Isuzu diesel
engine, 10 torque load, cyclone
impactor, dilution tunnel constant
volume sampler)
Particle Size: Diameter: < 1 /ym
Route: Cell Culture
Dose/Concentration: 5,15, 25/yglmL
Time to Analysis: Cells treated with
DEP, ox-PAPC (oxidized 1 -palmitoyl-2-
arachidonyl-OT-glycero-3-
phosphorylchlorine), DEP+ox-PAPC.
Analytical tests performed.
H0-1 expression was dose-dependent and
greatest with the DEP+ox-PAPC treatment. DEP
significantly dose-dependently upregulated or
downregulated a number of genes and was
shown to have a synergistic effect with co-
treatment of ox-PAPC. The most varying genes
were significantly enriched for EpRE,
inflammatory response, UPR, immune response,
cell adhesion, lipid metabolism, apoptosis and
protein folding genes.
Reference: Greenwell et
al. (2003, 097478)
Species: Rat
Cell Line: Epithelial fluid;
icosahedral bacteriophage
~X174-RF DNA
PM (South Wales, UK) (urban,
industrial)
Particle Size: Coarse diameter: 10-
2.5 /jm, Fine diameter: 2.5-0.1 /ym
Route: Cell Culture
Dose/Concentration: Urban mean: 18.7
± 4.7mg/day; Industrial mean: 22.6 ±
2.5mg/day
Time to Analysis: 24h air samples 4-
11 d. Substrates vortexed 1 h, suspended
4h, centrifuged 1h. Oxidation assay.
Industrial PM was more bioreactive than urban
PM. Coarse fractions had greater oxidative
potential and bioreactivity than fine fractions.
Reference: Gu et al.
2005
Species: Hamster
Strain: Chinese
Cell Line: Lung fibroblast
(V79)
DPM (1980 model General Motors
5.7-L V-8 engine)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 25, 50,100,150
/yglmL; 10 /yg DPM in 10 /yg in
DPPC/mL; 10 /yg DPM in 10 /yg
DMSO/mL
Time to Analysis: Chromosomal
aberration: 24h incubation. Treated 24h.
Incubated again 24h. MN assay: 24h
treatment. Gene mutation: 24h
treatment. Cells replated. 7d expression
times. Staining at 8,10d.
DPM significantly and dose-dependently
increased aberrant cells at 25-100/yglmL. DPM
increased MN formation dose-dependently.
Mutant frequencies were not significant and
showed no dose-dependent trends. DPM was
toxic to cells at the highest concentration.
Reference: Gualtieri et
al. (2005, 097841)
Species: Human
Cell Line: Alveolar lung
(A549)
TD (Tire debris, generated by rotating
new vehicle wheel against a steel
brush, significant component of PMio)
(organic extracts)
Particle Size: Diameter: 10-80 /ym
Route: Cell Culture
Dose/Concentration: 50, 60, 75/yglmL
Time to Analysis: Particles extracted
6h. Cells subcultured every 3-4d. After
24h, TD treatments 24,48, 72h.
A time- and dose-dependent inhibitory effect on
the reduction of MTT was seen. Mortality
increased dose-dependently and was
significantly greater than the controls. DNA
strand breaks increased significantly in a dose-
dependent manner. A significant cell cycle block
in the G1 phase with a consequent decrease in
the cell number in the S and G2/M phases was
seen. Exposed cells had a modified morphology.
July 2009
D-178
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle
Exposure
Effects
Reference: Gutierrez-
Castillo et al. (2006,
089030)
Species: Human
Cell Line: A549 type II
alveolar epithelial cells
PM2.6 and PM10 (4 monitoring stations
in Mexico City: (1) downtown high
auto traffic, (2) two industrial areas
with high levels of auto traffic and
low vegetation, (3) medium-traffic
residential area) (winter, spring , 4
sampling days in each period)
(aqueous and organic extracts)
Particle Size: Diameter: 2.5 or 10
/jm
Route: Cell Culture
Dose/Concentration: 0.05, 0.07,
0.1 m3/ml equivalents PM2.6; 0.82,1.25,
1.63m3/ml equivalents PM10
Time to Analysis: Cells treated 48h
with water-soluble or organic-soluble PM
extracts.
Higher amounts of water-soluble metals were
found in samples collected during winter. Water-
soluble extracts increased DNA damage 1.7-fold
over the background. Similar results were
observed with organic extracts. In general, PM2.6
extracts had greater genotoxic potential than
PM10 extracts, and water soluble fractions form
both particle sizes were more genotoxic than the
corresponding organic extracts.
Reference: Izawa H et al.
(2007,190387)
Cell Line: NA
DEPE (4JB-1 Isuzu 4-cylinder direct-
injection 2740cc diesel engine;
1500rpm, 10kg/m load)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: DEP: Ah-1
experiment- 111, 55.5, 27.8,13.9, 6.9,
3.5,1.7/yglmL; Foods, polyphenols
experiment- 27.8/yg/mL
Time to Analysis: DEPE incubated 2h
for dioxin toxicity measurement.
Absorbance at 405nm measured. Food,
polyphenol inhibitory effects: food
extract or polyphenol solution added to
cytosol solution, shaken 5min. DEPE
added, shaken 5min. 2h incubation.
Absorbance at 405nm measured.
The dioxin toxicity equivalent was 6,479±58ng
DEQ/g of DEP. The absorbance showed a
sigmoid curve and dose-dependently increased
from 6.9 to 27.8/yg DEP/mL. The Ginkgo bifoba
extract significantly inhibited AhR activation
significantly more than the other foods, and was
followed by green tea, onions, and garlic,
~uercetin and myricetin dose-dependently
inhibited AhR activation. Ginkgolides A and B
had weak inhibitory effects and resveratol was
the weakest.
Reference: Jacobsen et DEP (SRM 1650b)
al. (2008,156597)
Species: Mouse
Cell Line: FE1-MutaTI
lung epithelial cells
Carbon black (CB) (Printex 90)
Particle Size: DEP: 18-30nm; CB:
14nm; Agglomerates in suspensions:
DEP Peaks- 249nm, CB Peaks-
476nm
Route: Cell Culture
Dose/Concentration: 37.5, 75/yg/mL
Time to Analysis: 8 repeated 72h
incubations.
Mutagenicity: The 75/yg/mL dose was
significantly increased compared to the 37.5
/yg/mL dose. Linear regression showed a
significant increasing trend by increasing
exposure. There was no change in the total cell
numbers.
ROS: ROS production increased in DEP-treated
cells after 3h of exposure. CB-treated cells
showed a dose-dependent increase.
Reference: Karlsson et
al. 2004
Species: Human
Cell Line: Fibroblasts;
calf thymus DNA with
human liver microsomes
or rat liver S9
PM (urban dust particles, SRM 1649)
(extracted with DCM, acetone,
DMSO, water) (Fe 3% w/w, Ti 0.32%
w/w, V 0.04% w/w, Mn 0.03% w/w,
Cu 0.025% w/w)
Particle Size: Mean diameter: < 10
/ym
Route: Cell Culture
Dose/Concentration: 0.1,1.0,10,100
/yg/cm2
Time to Analysis: Particles extracted.
Fibroblasts exposed 24h. Comet assay.
Calf thymus incubated 2h with
microsomes or S9.32P-labelled.
DNA damage increased dose-dependently, and a
significant amount of DNA-damaged cells had
particle interactions. DNA damage induced by
the insoluble particle core significantly increased
after each extraction. Native particles were
more genotoxic than those extracted with
DMSO, DCM and water, but not with acetone or
hexane. DMSO extracts had the most adduct-
forming PACs, and water extracts had the most
oxidizing substances.
Reference: Karlsson et
al. (2005, 086392)
Species: Human
Cell Line: Lung epithelial
A549 type II
PM (subway station, urban street)
Subway particles: O2, Fe (Fe from
Fe304) Street particles: Fe from Fe203
Particle Size: Diameter: 10 //in
Route: Cell Culture
Dose/Concentration: Comet: 5,10, 20,
40/yg/cm2; 8-oxodG: 10/yg/cm2
Time to Analysis: Air sampled 24h
daily. Cells grown 24h. Exposed 4h.
Both PM types induced concentration-dependent
DNA damage, but subway particles were more
potent. Subway particles caused more 8-oxodG
formation and oxidation of dG, the latter of
which was inhibited by deferoxaminemesylate.
Oxidation from subway particles was due to
nonsoluble, redox active substances, and soluble
substances from street particles.
Reference: Karlsson et
al. (2006,156625)
Species: Human
Cell Line: Lung epithelium
A549; monocytes from
heparinized whole blood
PM (wood- old, modern boiler; pellets-
pellets burner, electrical ignition; tire-
road simulator studded, friction tires;
Street- busy street, Stockholm;
Subway- platform near street)
Particle Size: Diameter: 2.5,10/ym
Route: Cell Culture
Dose/Concentration: 40/yg/cm2
Time to Analysis: Samples collected.
Blank filter and Teflon filters used. Cells
grown 24h. Comet assay. Monocytes
incubated 10d. Macrophages incubated
18h.
All particles tested caused DNA damage, but
there was no significant difference between the
size fractions. Subway particles were the most
genotoxic. The urban street particles were the
most potent inducers of the cytokines. On the
Teflon filters, PMiowas somewhat more potent
than PM2.6.
Reference: Kubatova et
al. (2004, 087986)
Species: Monkey
Cell Line: African green
kidney COS-1 (CV-1 cells
with origin-defective
SV40 mutants); (±S9)
PM (DE from diesel bus, wood smoke
(WS) from chimney, hardwood smoke)
(organic extracts)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 25, 50,100, 200
/yg/mL; 50mg of each material used for
all experiments
Time to Analysis: DE and WS collected.
Extracted. 24h cytotoxicity. 2h SOS
chromotest.
WS had significantly increased cytotoxicity in
fractions of 25-250°C, and DE in nonpolar
fractions of 250 and 300°C and polar fractions
of 50°C. The cytotoxicity of DE PM nonpolar
fractions corresponded to increased
concentrations of PAHs. WS was not genotoxic
and DE was genotoxic in midpolarity fractions
(50-250°C). Genotoxic response was not
increased after S9 activation.
July 2009
D-179
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle
Exposure
Effects
Reference: Landvik et al.
(2007,096722)
Species: Mouse
Cell Line: Hepatoma
Hepa1c1c7 cells
DEP extracts (DEPE in the paper)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 10, 20, 30, 50,
70/yg/mL
Time to Analysis: Cells exposed 24h.
DNA fragmentation assay.
50 and 70/yglmL DEPE did not induce DNA
fragmentation but did cleave caspase 3 to a
minor extent.
Reference: Mehta et al.
(2008,190440)
Species: Human
Cell Line: Lung
adenocarcinoma (A549)
PM (SRM 1949a)
Particle Size: Diameter: ~ 0.18/ym
Route: Cell Culture
Dose/Concentration: 0, 50,100, 200,
400/yg/mL
Time to Analysis: Cell culture and cell
viability assay: PM treatment 24h. 10d
incubation. Host cell reactivation assay:
pGL3-luciferase plasmid UV irradiated
20min. PM treatment 24h. 16h
transfection. 24h PM incubation. DNA
repair synthesis assay: PM treatment
24h. Proteinase K treatment 30min. supf
mutagenesis assay: PM treatment 24h.
PM culture 60h. DNA extracted.
Overnight incubation of transformed
bacteria.
PM reduced colony-forming ability and repair
synthesis capacity was proportional to the PM
concentration. PM dose-dependently decreased
HCR capacity and decreased more than TSP. PM
induced cyclobutane dimmers and
pyrimidine < 6-4 > pyrimidones mutations in UV-
irradiated supf.
Reference: Meng et al.
2007
Species: Rat
Gender: Male
Strain: Wistar
Age: NR
Weight: Mean: 230g;
Range: 200-250g
Cell Line: AMs from
treated rats
PM (Baotou, Wuwei, China) (normal
weather, dust storms, Mar 1-31)
(organic extracts, water soluble
fractions)
Particle Size: 2.5 /jm
Route: Cell Culture
Dose/Concentration: AM: 0, 33.3,
100, 300/yglmL; Water-soluble: 0, 75,
150, 300/yglmL; Organic extracts: 0,
25, 50,100 /yglmL; Mass concentration
normal day: 68.49±28.83//g|m3; Mass
concentration dust storm day:
221.83 ± 69.89 //g|m3
Time to Analysis: Samples collected
24h after 5pm. Extracted. Rats instilled
24h. Killed, lavaged. Cultures 4h.
0C, NH4+, NO3' were higher in normal weather
PM2.6. SO42', Ca2+were higher in dust storm
PM2.6. Fe, Al, Ca, Mg were 5x higher in dust
storm PM2.E. Cell viability reduced in a
concentration-dependent manner, with normal
weather being slightly more cytotoxic. DNA
damage was dose-dependently induced, with
normal weather and organic extracts showing
the greatest damage.
Reference: Motta et al.
2004
Species: Hamster
Strain: Chinese
Cell Line: Epithelial liver,
ovary
PM (Catania, Sicily; spring) (organic
extracts)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 0.60,1.21, 2.42,
4.85,9.70,19.40/yg/mL; 0.78,1.56,
2.12, 6.25,12.50, 25.00//g/mL
Time to Analysis: 2, 3h air sampling.
Extracts 4h. 24h treatment.
The treatment was only slightly cytotoxic at the
highest dose. DNA damage and aberrant cells
generally increased with dose. No effect was
seen in the Chinese hamster ovary cells without
metabolic activation.
Reference: Oh and Chung
(2006, 088296)
Cell Line: A549 (Comet),
CH0-K1 (CBMN), H4IIE
(EROD-microbiassay)
Crude extract (CE) DEP and fractions
of CE of DEP (organic extracts: F1 ¦
organic bases, F2 - organic acids, F3
- aliphatic, F4 - aromatic, F5 ¦
slightly polar, F6 -moderately polar,
F7 ¦ high polar)
Particle Size: Diameter: <2.5//m,
87.71%, 2.5-10/ym, 3.87%, >10
fjm, 8.42%
Route: Cell Culture
Dose/Concentration: 100 //g/mL
Time to Analysis: DEP generated,
extracted. Comet assay- 24h incubation,
CE, DEP exposed 24h. MN assay-
cultured 24h, 4h treatment, growth
medium incubation 20h. EROD-
microbioassay- 48h.
DNA damage: CE significantly increased the
amount of DNA damage in A549 cells with and
without SKF-525A, a CYP450 inhibitor, and in
CH0-K1 cells. It significantly increased MN
formation ±S9 compared to controls.
Organic Extracts: Organic base (F1) and
neutral (F3-F7) fractions of CE of DEP
significantly induced DNA damage without SKF-
525A compared to controls. Adding SKF-525A
completely inhibited damage caused by F3, F4,
F6 and F7 but kept the effect of F1 similar to
that without SKF and only partially inhibited
that of F5. F2 did not induce DNA damage with
or without SKF. All fractions except F6 induced
MN formation ±S9.
Reference: Poma et al.
(2006, 096903)
Species: Mouse
Cell Line: Macrophage
RAW 264.7
PM (L'Aquila, Italy; urban)	Route: Cell Culture
Carbon black (CB)	Dose/Concentration: 1, 3,10//gfcnr
Particle Size: Diameter: 2.1-0.43 //m Time to Analysis: Air samples collected
weekly basis Jan-Mar 2004. Cells
cultured 48h. Treatment 48h. MN assay: equally-weighted CB
44h incubation, 28h incubation.
PM and CB dose-dependently reduced cell
proliferation and induced micronuclei. PM and
CB also reduced cellular metabolism of the
macrophages and induced significant amounts of
apoptosis. PM produced more micronuclei than
Reference: Poma et al.
2006
Species: Mouse
Cell Line: RAW 264.7
macrophage
PM (urban, winter)
Particle Size: Diameter: 2.5 /jm
Route: Cell Culture
Dose/Concentration: 2.2, 6.6, 22
//g/mL
Time to Analysis: Cells treated with
particulates 44h. 28h incubation after
cytochalasin B. Micronuclei frequency
determined.
Extracts produced a dose-dependent increase in
micronuclei. Fine carbon black particles were
consistently less genotoxic at similar test
concentrations. Results indicate that the
chemicals adsorbed onto the particles were the
primary contributors to the observed genotoxic
effects.
July 2009
D-180
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle
Exposure
Effects
Reference: Roubicek et
al. (2007,156929)
Species: Human
Cell Line: Lung
adenocarcinomaA549 cell
line
PM (Mexico City from an industrial
area with high-traffic and a medium-
traffic residential area)
(aqueous or organic extracts)
Particle Size: Diameter: 10 //in
Route: Cell Culture
Dose/Concentration: 1.25,1.63,
2.5m3/ml equivalents of PMio
Time to Analysis: Cells treated 24h
followed by 48h incubation with
cytochalasin B. Micronuclei frequency
determined.
Water and organic extracts induced a significant
dose-dependent increase in the micronuclei
frequency. After doses of PM from different
regions were normalized for mass differences,
the genotoxic potency was higher for samples
from the industrial area.
Reference: Salonen et al.
(2004,187053)
Species: Mouse
Cell Line: Macrophage
RAW 264.7
PM (Vallila, Finland; busy traffic site;
sping, winter)
Particle Size: Diameter: <10//m
Route: Cell Culture
Dose/Concentration: 15, 50,150, 500,
1000/yglmL of RPMI
Time to Analysis: Air collected 2-7d
periods. Extracted. Exposed 24h.
PAHs decreased from winter to spring. TNF-a
dose-dependently increased and was higher in
spring samples. IL-6 generally increased in spring
but not in winter. NO dose-dependently
increased and was higher in winter. Cell viability
generally decreased but there were no
consistent potency differences between the
samples. Generally, proinflammatory activity,
cytotoxicity and IL-6 were associated with the
insoluble PM fractions. Polymyxin B inhibited IL-
6 and TNF-a. OH and 8-hydroxy-2'-
deoxyguanosine dose-dependently increased and
were higher in the spring and winter,
respectively.
Reference: Seaton et al. PM (3 busy London underground (LU)
2005
Species: Human
Cell Line: Alveolar
epithelial A549
stations and cabs) (LU dust in PM2.6
samples: iron oxide 64-71%,
chromium 0.1-0.2%, manganese 0.5-
1%, copper < 0.1-0.9%; respirable
dust samples: 1-2%)
Particle Size: Diameter: <2.5//m,
10 /jm, Median diameter: 0.4/ym
Route: Cell Culture
Dose/Concentration: Stations: 270-
480/yg/m3 PM2.B, 14000-29000
particles/cm3; Cabs: 130-200/yglm3
PM2.6,17000-23000 particles/cm3;
Assays: 1,10, 50,100/yglmL
Time to Analysis: 3d measurements of
LU and cabs. Air collected for lab tests.
Exposed 8, 24h.
PMio caused less LDH release, IL-8 stimulation
and free radical activity than LU dust particles
that contained PM2.6. Chelation had little effect
on PMio soluble components.
Reference: Sevastyanova PMio (Prague, Czech Republic;
et al. (2007,156969)
Species: Human
Cell Line: HepG2 cell line,
embryonic lung diploid
fibroblasts (HEL), or acute
monocytic leukemia cells
(THP-1)
Ko" sice; Slovak Republic; Sofia,
Bulgaria) (urban, summer, winter)
(organic extracts)
Particle Size: Diameter: 10 //m
Route: Cell Culture
Dose/Concentration: 10-100 //gfml
Time to Analysis: Samples collected
24h daily 3m. HepG2 and THP-1 cells
treated for 24h.
DNA adducts were observed in all cell types
evaluated. Highest adduct levels were observed
in HepG2 cells, followed by HEL and THP-1 cells.
A correlation between DNA adduct levels and
carcinogenic PAH content was observed in
HepG2 cells at 50/yg/ml.
Reference: Shi et al.
(2003, 088248)
Species: Human
Cell Line: Epithelial lung
A549
PM (Dusseldorf, Germany, July-Dec.)
Particle Size: Fine diameter: < 2.5
/jm; Coarse diameter: 10-2.5/jm
Route: Cell Culture
Dose/Concentration: Fine: 0.57-
2.49mg; Coarse: 0.66-1.89mg;
Concentration: 0.57mg/mL
Time to Analysis: Weekly samplings
July-Dec 1999. Electron spin response i
hydrodeoxyguanosine induction,
measurement.
Coarse and fine particles generated OH, but
coarse particles had significantly higher OH
formation as well as 8-hydroxy-2'-
deoxyguanosine formation. 8-hydroxy-2'-
deoxyguanosine and OH had a significant
correlation.
Reference: Skarek et al.
(2007,096814)
Species: Rat
Cell Line: Modified
hepatoma H4IIE.Iuc; SOS:
£ coli PQ37 (±S9)
PM (urban: List! nad Laben, Karvina;
background: Cervenohorske sedlo,
Kosetice - Czech Republic; July)
(organic extracts, TSP) GP (gas
phase)
Particle Size: Diameter: <2.5//m
Route: Cell Culture
Dose/Concentration: SOS: 8, 4, 2,
1m3
ml ; Dioxin: TSP+GP- 8,1.33, 0.22,
0.04m3 ml'1, PM2.6+GP: 4, 0.66, 0.11,
0.02m3 ml-'
Time to Analysis: 24h samples July
2002. Extracted. SOS chromotest: 22h
incubation. Dioxin toxicity test: 24h
exposure.
The urban areas had a much greater level of
carcinogenic PAHs and overall number of PAHs
than the background areas. Significant
genotoxic activity was only detected at
TSP + GP without S9 from urban areas.
PM2.6+GP had lower dioxin activity at the urban
areas, but similar levels of toxicity were seen for
both treatments in the background areas.
Reference: Song et al.
(2007,155306)
Species: S. typhimurium
Strain: TA98, TA100
Cell Line: Rat fibrocytes
L-929 cells
PM (soluble organic fraction (SOF)
extracts from diesel engines using
fuels blended with ethanol by volume:
EO ¦ base diesel fuel; E5 ¦ 5%; E10 ¦
10%;E15-15%;E20-20%)
Particle Size: Density (g/cm3): EO-
0.8379; E5- 0.8349; E10- 0.8324;
E15- 0.8301; E20- 0.8279
Route: Cell Culture
Dose/Concentration: Ames Assay:
0.025, 0.05, 0.1 mg/plate; Comet Assay:
0.125, 0.25, 0.5,1.0mg|mL
Time to Analysis: Samples extracted
24h. Ames and comet assays performed
All PM extracts induced higher mutational
response in TA98 (3- to 5-fold increase over
spontaneous) than in TA100 (2-to 3-fold
increase). The highest brake specific revertants
(BSR) ±S9 in both strains occurred with E20
and lowest BSR was in E5 (except in TA98 -
S9). EO and E20 caused more significant DNA
damage (similar in effect) than the other
extracts. Damage was dose-dependent but
variable with increasing ethanol volume.
July 2009
D-181
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle
Exposure
Effects
Reference: Ueng et al.
(2005, 097054)
Species: Human
Cell Line: Lung epithelium
CL5 (cancerous),
bronchial epithelial BEAS-
2B (noncancerous), WI-38
normal lung fibroblast
MEP (Yamaha cabin motorcycle 2-
strok 50-cc engine)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 1,10,100, 200
/yg/mL
Time to Analysis: Exhaust collected,
extracted. cDNA microarray analysis.
RT-PCR: 2h. ELISA: 12h incubation.
Centrifuged 24h post-treatment.
Bioactivity: 12h incubation. Centrifuged
24h post-treatment. Medium replaced
48h post-incubation. Fibroblasts
determined 96h post-incubation. Time
response studies: 3-48h treatment.
Concentration response studies: 6h
treatment.
Drug metabolism array study: MEP increased
CYP1A1, CYP3A7 and UGT2B.
Cytokine array study: MEP increased
fibroblast growth factor (FGF)-6, FGF-9, IL-1 a,
IL-22 and vascular endothelial growth factor
(VEGF)-D mRNA.
Oncogene, tumor suppressor, estrogen
signaling pathway: MEP increased fra-1, c-src,
SHC, p21, C0X7RP, and decreased p53 and Rb
expression.
RT-PCR: MEP increased CYP1A1, CYP1B1, IL-6,
IL-11, IL-1 a, FGF-6, FGF-9, VEGF-D, fra-1 and
p21.
Concentration and time responses:
Concentration and time-dependent increases
occurred for FGF-9, IL-1a, IL-6, IL-11, but
decreased time-dependently after 6h exposure.
BEAS-2B Cells: MEP had concentration-
dependent increases on CYP1A1 and CYP1B1
but did not affect anything else.
Peroxide, MEP+NAC, WI-38 Cells: MEP
increased peroxide production. The MEP+NAC
treatment reduced MEP-elevated levels of IL-1 a,
IL-6, FGF-9, VEGF-D to control levels. Fibroblasts
increased in WI-38 cells.
Reference: Umbuzeiro et
al. (2008,190491)
Species: Salmonella
typhimurium
Strain: TA98, YG1041
(+/¦ S9)
PM (urban; Sao Paulo, Brazil-
Cerqueira Cesar street station,
Ibirapuera park station) (winter- June
17,18; average temperature: 16°C)
(E0M)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Cerqueira Cesar:
UPM- 156/yg/m3, E0M- 57.7mg/total
UPM; Ibirapuera Park: UPM- 32/yg/m3,
E0M- 41.7mgjtotal UPM; Salmonella
assay-0.5,1,5,10, 50,100 UPM
equiv/plate |//g)
Time to Analysis: Tests performed 20d
after collection date. Organic extraction
20h. PAH fractionation.
Salmonellafmicrosome assay.
The TSP and E0M were similar for both sites.
The PAH fraction had very low mutagenicity for
the Cerqueira Cesar sample in the YG1041
strain and no mutagenicity for the Ibirapuera
sample. Nitro-PAH and oxy-PAH had similar
mutagenetic activities from both samples. S9
decreased mutagenicity in nitro-PAH but was
increased in oxy-PAH. DNA adduct levels were
dose-dependent and not different between the
two sites.
Reference: Upadhyay et
al. (2003, 097370)
Species: Human
Cell Line: Alveolar
epithelial A549
PM (Dusseldorf, Germany) (Particles
contain carbon (19.70%),
hydrogen(1.4%),nitrogen (< .05%),
oxygen(14.12%), sulfur (2.09%), ash
(63.24%)) (lonizable metals
concentrations (ppm): Co(103),
Cu(48),Cr(104),Fe(14,521), Mn(21.3),
Ni(1,519),Ti(131), V(2,767)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 1, 5, 25,100
/yg/cm2; 10, 25, 50,100 /yglcm2
Time to Analysis: Air collected. Cells
grown 24h. PM treatments 1, 4, 8,12,
24h.
PM induced dose- and time-dependent reductions
in ds-DNA due to the formation of DNA-SB. The
soluble component caused higher DNA damage.
Apoptosis and DNA fragmentation increased
dose-dependently. nam decreased dose-
dependently in control cells, but not in cells with
Bcl-xl overexpression. PM caused activation of
caspase 9. Pretreatment with iron chelators or a
free radical scavenger reduced PM-induced DNA-
SB formation, DNA fragmentation, caspase 9
activation, and weakened nam reductions.
Reference: Valavanidis et
al. (2005, 096432)
Cell Line: NR
PM (TSP: high volume pumps, Athens;
DEP: 2.0L engine GM Astra; GEP:
1,6L passenger vehicle Ford; Wood
smoke soot: domestic fireplace
exhaust chimney; PMio: high volume
sampling system, Athens; PM2.6: high
volume cascade impactor (Anderson)
system
Particle Size: Diameter: >10.2-
< 0.41 /jm
Route: Incubation
Dose/Concentration: 20, 40mgf5mL
Time to Analysis: PM collected 30min-
1 h or 24h basis. Incubated with H2O2
and 2' deoxyguanosine (dG). Stored 3-7d
at -20°C. Fenton reaction. EPR analysis.
PM generated OH by a Fenton reaction, which
is increased by the addition of EDTA but
inhibited by deferoxamine. PM dose-dependently
induced dG hydroxylation and 8-hydroxy-2'-
deoxyguanosine formation. Transition metals Ni,
V, Co, Cr that are capable of redox cycling
electron producing ROS were found in the PM
samples.
Reference: Xu and Zhang
(2004, 097231) Species:
Human
Cell Line: Lung epithelial
A549
PM (Taiyuan, Beijing; Nov-Feb)
(Taiyuan: coal-fume pollution; Beijing:
coal-fume and vehicle exhaust)
Particle Size: Diameter: 2.5 /jm
Route: Cell Culture
Dose/Concentration: 5, 50, 200//gfmL
Time to Analysis: Air samples
collected. Cells incubated 12-24h. Comet
assay.
Taiyuan had a significantly higher daily PM2.6
average than Beijing. It was shown that the
smaller the particulate diameter, the higher the
concentration of BaP and Pb. A dose- and time-
response relationship was seen in DNA
fragmentation.
July 2009
D-182
DRAFT-DO NOT CITE OR QUOTE

-------
Table D-9. Mutagenicity and genotoxicity data summary: in vivo studies.
Reference
Particle
Exposure
Effects
Reference: Abou Chakra
et al. (2007, 098819)
Species: Human
Gender: Male, Female
Age: 6-13yrs and Adults
Participant
Characteristics: Non-
smokers
Cell Line: HeLa S3 cells
PM (3 French metropolitan cities:
Urban PM2.6 and PM10 from
"Residential Sector," "Proximity
Sector," "Industrial Sector'!
(organic extracts)
Particle Size: Diameter: 2.5,10 //m
Route: Cell Culture
Dose/Concentration: 261 PM2.6; 76
PMiosamples
Time to Analysis: Cells incubated with
200; fjL organic extract and 20 /jl
aphidicoline for 24h.
Seasonal variation was observed with genotoxic
effects being greater in winter. PM2.6 was more
active than PM10 extracts. Samples from the
"Proximity Sector" (downtown area with heavy
traffic) exhibited the strongest genotoxic
responses.
Reference: Arrieta et al.
(2003,096210)
Species: Rat
Cell Line: Hepatoma
(H4IIE)
Species: Mouse
Cell Line: Hepatoma
H111.1 c2
PM (El Paso, Texas; Juarez,
Chihuahua, Mexico; Sunland Park,
New Mexico) (organic extracts)
Particle Size: Diameter: 10 //in
Route: Cell Culture
Dose/Concentration: EROD test: 0.03,
0.17,0.34,0.50, 0.68, 4.96, 9.93
extract equivalents (m3 air); Luciferase:
0.17,0.51,1.26, 5.01 extract
equivalents (m3 air)
Time to Analysis: Extracts incubated
24h. EROD, luciferase activity, PAH
content determined.
EROD activity declined at higher extract
amounts, but luciferase activity was not
inhibited. Cytotoxicity occurred only at extract
equivalents to 0.47 m3air. PAH concentration
increased with PM mass.
Reference: Bao et al.
(2007,097258)
Cell Line: Human-
hamster hybrid (Al)
DEP (organic extracts) (SRM 2975)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 10, 20, 50,100
/yg/mL
Time to Analysis: Phagocytosis
inhibitors: Exposed 24h with or without
cytochalasin B or ammonium chloride.
Cytotoxicity: 24, 48h incubation.
Mutations: Exposed 24h. 5-7d culture.
Incubated additional 7-8d.
The nucleus of DEP-treated cells was condensed
and shrunken compared to controls. DEPs
accumulated in cells, disrupting the
mitochondrial cristae, and were lodged in large
cytoplasmic vacuoles. DEP produced minimal
toxicity. CD59 locus mutations dose-
dependently increased but decreased when
simultaneously treated with cytochalasin B or
ammonium chloride.
Reference: Carvalho-
Oliveria et al. (2005,
077898)
Species: T. pallida, A.
cepa
PM (Sao Paulo, Brazil; spring, bus
strike and non-strike days) (organic
extracts)
Particle Size: Diameter: 2.5 /jm
Route: Cell Culture
Dose/Concentration: Strike day: 47.32
/yg/m3; Non-strike day: 43.01 /yglm3
Time to Analysis: Air samples from 2d:
bus strike day, bus non-strike day. T.
pallida kept in lab 24h. Exposed 8h. 24h
recovery. Fixed 24h. A. cepa roots
induced 5d. Exposed 30h. Fixed 24h.
Element concentrations, sulfur and BTEX
decreased on the strike day. Micronuclei
decreased in T. pallida during the strike. Toxicity
measured in A. cepa was not significant, but
higher on strike days.
Reference: Dybdahl et al. DEP (SRM 1650)
(2004, 089013)	Particle Size: NR
Species: Human
Cell Line: Lung epithelial
A549
Route: Cell Culture
Dose/Concentration: 10, 50,100, 500
/yg DEP/mL
Time to Analysis: DEP suspended,
sonicated. A549 cells diluted. Fresh
medium added after 24h. After 48h
medium removed, DEP added. 2, 5, 24h
incubation.
DEP induced dose-dependent increases of IL-1a,
IL-6, IL-8, TNF-a. The cytokines increased 4-18-
fold at the highest dose. Cell viability did not
decrease. Comet tail length increased at 100
and 500 /yglmL for 2, 5, 24h.
Reference: Gabelova et
al. 2006
Species: Human
Cell Line: Hepatoma Hep
G2
PM (winter, summer) (organic
extracts)
Particle Size: Diameter: 10 //m
Route: Cell Culture
Dose/Concentration: 5,10, 20, 50,
100,150//gfmL
Time to Analysis: 3m sampling periods,
winter, summer. Cells grown 48h.
Exposed 2h. Single cell gel
electrophoresis or cultivated, harvested,
processed 2, 4,16, 24h.
PM, c-PAHs and genotoxicity were higher in
winter air samples than summer. E0M samples
generally had significant dose-dependent
increases in DNA migration. Repair-specific DNA
endonucleases did not increase DNA migration.
8-oxodG was below the steady-state level in
EOM samples.
Reference: Gabelova et
al. 2007
Species: Human
Cell Line: Hepatoma Hep
G2
PM (PRG-SM, PRG-LB, Kosice, Sofia;
winter, summer) (organic extracts)
Particle Size: Diameter: 10 //m
Route: Cell Culture
Dose/Concentration: 5-150/yglmL
Time to Analysis: Air samples collected
24h intervals, 3m sampling period. Cells
grown 48h. Exposed 2h. 24, 48h
preliminary experiments. Single cell gel
electrophoresis.
Cell viability significantly decreased in the 24,
48h exposure groups compared to the 2h
exposure group. DNA migration significantly
dose-dependently increased at most
concentrations. In general, oxidative DNA
damage did not significantly increase.
July 2009
D-183
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle
Exposure
Effects
Reference: Gabelova et
al. 2007
Species: Human
Cell Line: Hepatoma Hep
G2 cell line
PMio (Prague (Czech Republic),
Ko'sice (Slovak Republic) and Sofia
(Bulgaria); urban, winter, summer)
(organic extracts)
Particle Size: Diameter: 10 //in
Route: Cell Culture
Dose/Concentration: 5 to 150/yglml
EOM from 50/yglml stock solution
Time to Analysis: 24h DNA adduct
formation. 2h Comet assay. Oxidative
DNA damage measured by Fpg-sensitive
sites.
Total DNA adducts ranged from ~ 60 to 200
adducts per 10s nucleotides. Extracts also
produced approximately the same levels of
strand breaks. Results suggested that the
genotoxic potential of ambient air was at least
6-fold greater in the winter compared to
summer. No substantial difference was reported
for oxidative DNA damage induced by summer
vs. winter samples.
Reference: Gong et al.
(2007,091155)
Species: Human
Cell Line: Microvascular
endothelial (HMEC)
DEP (aggregates, exhaust 4JBi-type
LD,274 1,4-cylinder Isuzu diesel
engine, 10 torque load, cyclone
impactor, dilution tunnel constant
volume sampler)
Particle Size: Diameter: < 1 /ym
Route: Cell Culture
Dose/Concentration: 5,15, 25/yglmL
Time to Analysis: Cells treated with
DEP, ox-PAPC (oxidized 1 -palmitoyl-2-
arachidonyl-OT-glycero-3-
phosphorylchlorine), DEP+ox-PAPC.
Analytical tests performed.
H0-1 expression was dose-dependent and
greatest with the DEP+ox-PAPC treatment. DEP
significantly dose-dependently upregulated or
downregulated a number of genes and was
shown to have a synergistic effect with co-
treatment of ox-PAPC. The most varying genes
were significantly enriched for EpRE,
inflammatory response, UPR, immune response,
cell adhesion, lipid metabolism, apoptosis and
protein folding genes.
Reference: Greenwell et
al. (2003, 097478)
Species: Rat
Cell Line: Epithelial fluid;
icosahedral bacteriophage
cpX174-RF DNA
PM (South Wales, UK) (urban,
industrial)
Particle Size: Coarse diameter: 10-
2.5 fjm, Fine diameter: 2.5-0.1 /ym
Route: Cell Culture
Dose/Concentration: Urban mean: 18.7
± 4.7mg/day; Industrial mean: 22.6 ±
2.5mg/day
Time to Analysis: 24h air samples 4-
11 d. Substrates vortexed 1 h, suspended
4h, centrifuged 1h. Oxidation assay.
Industrial PM was more bioreactive than urban
PM. Coarse fractions had greater oxidative
potential and bioreactivity than fine fractions.
Reference: Gu et al.
2005
Species: Hamster
Strain: Chinese
Cell Line: Lung fibroblast
(V79)
DPM (1980 model General Motors
5.7-L V-8 engine)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 25, 50,100,150
/yglmL; 10 /yg DPM in 10 /yg in
DPPCImL; 10 /yg DPM in 10 /yg
DMSO/mL
Time to Analysis: Chromosomal
aberration: 24h incubation. Treated 24h.
Incubated again 24h. MN assay: 24h
treatment. Gene mutation: 24h
treatment. Cells replated. 7d expression
times. Staining at 8,10d.
DPM significantly and dose-dependently
increased aberrant cells at 25-100/yglmL. DPM
increased MN formation dose-dependently.
Mutant frequencies were not significant and
showed no dose-dependent trends. DPM was
toxic to cells at the highest concentration.
Reference: Gualtieri et
al. (2005, 097841)
Species: Human
Cell Line: Alveolar lung
(A549)
TD (Tire debris, generated by rotating
new vehicle wheel against a steel
brush, significant component of PMio)
(organic extracts)
Particle Size: Diameter: 10-80 /ym
Route: Cell Culture
Dose/Concentration: 50, 60, 75/yglmL
Time to Analysis: Particles extracted
6h. Cells subcultured every 3-4d. After
24h, TD treatments 24,48, 72h.
A time- and dose-dependent inhibitory effect on
the reduction of MTT was seen. Mortality
increased dose-dependently and was
significantly greater than the controls. DNA
strand breaks increased significantly in a dose-
dependent manner. A significant cell cycle block
in the G1 phase with a consequent decrease in
the cell number in the S and G2/M phases was
seen. Exposed cells had a modified morphology.
Reference: Gutierrez-
Castillo et al. (2006,
089030)
Species: Human
Cell Line: A549 type II
alveolar epithelial cells
PM2.6 and PMio(4 monitoring stations
in Mexico City: (1) downtown high
auto traffic, (2) two industrial areas
with high levels of auto traffic and
low vegetation, (3) medium-traffic
residential area) (winter, spring , 4
sampling days in each period)
(aqueous and organic extracts)
Particle Size: Diameter: 2.5 or 10
/ym
Route: Cell Culture
Dose/Concentration: 0.05, 0.07,
0.1 m3|ml equivalents PM2.6; 0.82,1.25,
1.63m3/ml equivalents PMio
Time to Analysis: Cells treated 48h
with water-soluble or organic-soluble PM
extracts.
Higher amounts of water-soluble metals were
found in samples collected during winter. Water-
soluble extracts increased DNA damage 1.7-fold
over the background. Similar results were
observed with organic extracts. In general, PM2.6
extracts had greater genotoxic potential than
PMio extracts, and water soluble fractions form
both particle sizes were more genotoxic than the
corresponding organic extracts.
Reference: Izawa H et al.
(2007,190387)
Cell Line: NA
DEPE (4JB-1 Isuzu 4-cylinder direct-
injection 2740cc diesel engine;
1500rpm, 10kg/m load)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: DEP: Ah-1
experiment- 111, 55.5, 27.8,13.9, 6.9,
3.5,1.7/yglmL; Foods, polyphenols
experiment- 27.8/yglmL
Time to Analysis: DEPE incubated 2h
for dioxin toxicity measurement.
Absorbance at 405nm measured. Food,
polyphenol inhibitory effects: food
extract or polyphenol solution added to
cytosol solution, shaken 5min. DEPE
added, shaken 5min. 2h incubation.
Absorbance at 405nm measured.
The dioxin toxicity equivalent was 6,479±58ng
DEQ/g of DEP. The absorbance showed a
sigmoid curve and dose-dependently increased
from 6.9 to 27.8/yg DEPImL. The Ginkgobiloba
extract significantly inhibited AhR activation
significantly more than the other foods, and was
followed by green tea, onions, and garlic,
~uercetin and myricetin dose-dependently
inhibited AhR activation. Ginkgolides A and B
had weak inhibitory effects and resveratol was
the weakest.
July 2009
D-184
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle
Exposure
Effects
Reference: Jacobsen et
al. (2008,156597)
Species: Mouse
Cell Line: FEI Muta™
lung epithelial cells
DEP (SRM 1650b)
Carbon black (CB) (Printex 90)
Particle Size: DEP: 18-30nm; CB:
14nm; Agglomerates in suspensions:
DEP Peaks- 249nm, CB Peaks-
476nm
Route: Cell Culture
Dose/Concentration: 37.5, 75/yg/mL
Time to Analysis: 8 repeated 72h
incubations.
Mutagenicity: The 75/yg/mL dose was
significantly increased compared to the 37.5
/yg/mL dose. Linear regression showed a
significant increasing trend by increasing
exposure. There was no change in the total cell
numbers.
ROS: ROS production increased in DEP-treated
cells after 3h of exposure. CB-treated cells
showed a dose-dependent increase.
Reference: Karlsson et
al. 2004
Species: Human
Cell Line: Fibroblasts;
calf thymus DNA with
human liver microsomes
or rat liver S9
PM (urban dust particles, SRM 1649)
(extracted with DCM, acetone,
DMSO, water) (Fe 3% w/w, Ti 0.32%
w/w, V 0.04% w/w, Mn 0.03% w/w,
Cu 0.025% w/w)
Particle Size: Mean diameter: < 10
/jm
Route: Cell Culture
Dose/Concentration: 0.1,1.0,10,100
/yg/cm2
Time to Analysis: Particles extracted.
Fibroblasts exposed 24h. Comet assay.
Calf thymus incubated 2h with
microsomes or S9.32P-labelled.
DNA damage increased dose-dependently, and a
significant amount of DNA-damaged cells had
particle interactions. DNA damage induced by
the insoluble particle core significantly increased
after each extraction. Native particles were
more genotoxic than those extracted with
DMSO, DCM and water, but not with acetone or
hexane. DMSO extracts had the most adduct-
forming PACs, and water extracts had the most
oxidizing substances.
Reference: Karlsson et
al. (2005, 086392)
Species: Human
Cell Line: Lung epithelial
A549 type II
PM (subway station, urban street)
Subway particles: O2, Fe (Fe from
Fe304) Street particles: Fe from Fe203
Particle Size: Diameter: 10 //in
Route: Cell Culture
Dose/Concentration: Comet: 5,10, 20,
40/yg/cm2; 8-oxodG: 10/yg/cm2
Time to Analysis: Air sampled 24h
daily. Cells grown 24h. Exposed 4h.
Both PM types induced concentration-dependent
DNA damage, but subway particles were more
potent. Subway particles caused more 8-oxodG
formation and oxidation of dG, the latter of
which was inhibited by deferoxaminemesylate.
Oxidation from subway particles was due to
nonsoluble, redox active substances, and soluble
substances from street particles.
Reference: Karlsson et
al. (2006,156625)
Species: Human
Cell Line: Lung epithelium
A549; monocytes from
heparinized whole blood
PM (wood- old, modern boiler; pellets-
pellets burner, electrical ignition; tire-
road simulator studded, friction tires;
Street- busy street, Stockholm;
Subway- platform near street)
Particle Size: Diameter: 2.5,10/ym
Route: Cell Culture
Dose/Concentration: 40/yg/cm2
Time to Analysis: Samples collected.
Blank filter and Teflon filters used. Cells
grown 24h. Comet assay. Monocytes
incubated 10d. Macrophages incubated
18h.
All particles tested caused DNA damage, but
there was no significant difference between the
size fractions. Subway particles were the most
genotoxic. The urban street particles were the
most potent inducers of the cytokines. On the
Teflon filters, PMiowas somewhat more potent
than PM2.6.
Reference: Kubatova et
al. (2004, 087986)
Species: Monkey
Cell Line: African green
kidney COS-1 (CV-1 cells
with origin-defective
SV40 mutants); (±S9)
PM (DE from diesel bus, wood smoke
(WS) from chimney, hardwood smoke)
(organic extracts)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 25, 50,100, 200
/yg/mL; 50mg of each material used for
all experiments
Time to Analysis: DE and WS collected.
Extracted. 24h cytotoxicity. 2h SOS
chromotest.
WS had significantly increased cytotoxicity in
fractions of 25-250°C, and DE in nonpolar
fractions of 250 and 300°C and polar fractions
of 50°C. The cytotoxicity of DE PM nonpolar
fractions corresponded to increased
concentrations of PAHs. WS was not genotoxic
and DE was genotoxic in midpolarity fractions
(50-250°C). Genotoxic response was not
increased after S9 activation.
Reference: Landvik et al.
(2007,096722)
Species: Mouse
Cell Line: Hepatoma
Hepa1c1c7 cells
DEP extracts (DEPE in the paper)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 10, 20, 30, 50,
70/yg/mL
Time to Analysis: Cells exposed 24h.
DNA fragmentation assay.
50 and 70/yg/mL DEPE did not induce DNA
fragmentation but did cleave caspase 3 to a
minor extent.
Reference: Mehta et al.
(2008,190440)
Species: Human
Cell Line: Lung
adenocarcinoma (A549)
PM (SRM 1949a)
Particle Size: Diameter: ~ 0.18/ym
Route: Cell Culture
Dose/Concentration: 0, 50,100, 200,
400/yg/mL
Time to Analysis: Cell culture and cell
viability assay: PM treatment 24h. 10d
incubation. Host cell reactivation assay:
pGL3-luciferase plasmid UV irradiated
20min. PM treatment 24h. 16h
transfection. 24h PM incubation. DNA
repair synthesis assay: PM treatment
24h. Proteinase K treatment 30min. supf
mutagenesis assay: PM treatment 24h.
PM culture 60h. DNA extracted.
Overnight incubation of transformed
bacteria.
PM reduced colony-forming ability and repair
synthesis capacity was proportional to the PM
concentration. PM dose-dependently decreased
HCR capacity and decreased more than TSP. PM
induced cyclobutane dimmers and
pyrimidine < 6-4 > pyrimidones mutations in UV-
irradiated supf.
July 2009
D-185
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle
Exposure
Effects
Reference: Meng et al.
2007
Species: Rat
Gender: Male
Strain: Wistar
Age: NR
Weight: Mean: 230g;
Range: 200-250g
Cell Line: AMs from
treated rats
PM (Baotou, Wuwei, China) (normal Route: Cell Culture
weather, dust storms, Mar 1-31)
(organic extracts, water soluble
fractions)
Particle Size: 2.5 //m
Dose/Concentration: AM: 0, 33.3,
100, 300//g/mL; Water-soluble: 0, 75,
150, 300//g/mL; Organic extracts: 0,
25, 50,100 //g/mL; Mass concentration
normal day: 68.49±28.83//g/m3; Mass
concentration dust storm day:
221.83 ± 69.89 //g/m3
Time to Analysis: Samples collected
24h after 5pm. Extracted. Rats instilled
24h. Killed, lavaged. Cultures 4h.
0C, NH4+, NO3' were higher in normal weather
PM2.6. SO42', Ca2+were higher in dust storm
PM2.6. Fe, Al, Ca, Mg were 5x higher in dust
storm PM2.E. Cell viability reduced in a
concentration-dependent manner, with normal
weather being slightly more cytotoxic. DNA
damage was dose-dependently induced, with
normal weather and organic extracts showing
the greatest damage.
ference: Motta et al.
2004
Species: Hamster
Strain: Chinese
Cell Line: Epithelial liver,
ovary
PM (Catania, Sicily; spring) (organic
extracts)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 0.60,1.21, 2.42,
4.85,9.70,19.40//g/mL; 0.78,1.56,
2.12, 6.25,12.50, 25.00//g/mL
Time to Analysis: 2, 3h air sampling.
Extracts 4h. 24h treatment.
The treatment was only slightly cytotoxic at the
highest dose. DNA damage and aberrant cells
generally increased with dose. No effect was
seen in the Chinese hamster ovary cells without
metabolic activation.
Reference: Oh and Chung Crude extract (CE) DEP and fractions
(2006, 088296)
Cell Line: A549 (Comet),
CH0-K1 (CBMN), H4IIE
(EROD-microbiassay)
of CE of DEP (organic extracts: F1 ¦
organic bases, F2 - organic acids, F3
- aliphatic, F4 - aromatic, F5 ¦
slightly polar, F6 -moderately polar,
F7 ¦ high polar)
Particle Size: Diameter: <2.5//m,
87.71%, 2.5-10//m, 3.87%, >10
//m, 8.42%
Route: Cell Culture
Dose/Concentration: 100 //g/mL
Time to Analysis: DEP generated,
extracted. Comet assay- 24h incubation,
CE, DEP exposed 24h. MN assay-
cultured 24h, 4h treatment, growth
medium incubation 20h. EROD-
microbioassay- 48h.
DNA damage: CE significantly increased the
amount of DNA damage in A549 cells with and
without SKF-525A, a CYP450 inhibitor, and in
CH0-K1 cells. It significantly increased MN
formation ±S9 compared to controls.
Organic Extracts: Organic base (F1) and
neutral (F3-F7) fractions of CE of DEP
significantly induced DNA damage without SKF-
525A compared to controls. Adding SKF-525A
completely inhibited damage caused by F3, F4,
F6 and F7 but kept the effect of F1 similar to
that without SKF and only partially inhibited
that of F5. F2 did not induce DNA damage with
or without SKF. All fractions except F6 induced
MN formation ±S9.
Reference: Poma et al.
(2006, 096903)
Species: Mouse
Cell Line: Macrophage
RAW 264.7
PM (L'Aquila, Italy; urban)	Route: Cell Culture
Carbon black (CB)	Dose/Concentration: 1, 3,10//g/cm2
Particle Size: Diameter: 2.1-0.43 //m Time to Analysis: Air samples collected
weekly basis Jan-Mar 2004. Cells
cultured 48h. Treatment 48h. MN assay: equally-weighted CB.
44h incubation, 28h incubation.
PM and CB dose-dependently reduced cell
proliferation and induced micronuclei. PM and
CB also reduced cellular metabolism of the
macrophages and induced significant amounts of
apoptosis. PM produced more micronuclei than
Reference: Poma et al.
2006
Species: Mouse
Cell Line: RAW 264.7
macrophage
PM (urban, winter)
Particle Size: Diameter: 2.5 //m
Route: Cell Culture
Dose/Concentration: 2.2, 6.6, 22
//g/mL
Time to Analysis: Cells treated with
particulates 44h. 28h incubation after
cytochalasin B. Micronuclei frequency
determined.
Extracts produced a dose-dependent increase in
micronuclei. Fine carbon black particles were
consistently less genotoxic at similar test
concentrations. Results indicate that the
chemicals adsorbed onto the particles were the
primary contributors to the observed genotoxic
effects.
Reference: Roubicek et
al. (2007,156929)
Species: Human
Cell Line: Lung
adenocarcinomaA549 cell
line
PM (Mexico City from an industrial
area with high-traffic and a medium-
traffic residential area)
(aqueous or organic extracts)
Particle Size: Diameter: 10 //in
Route: Cell Culture
Dose/Concentration: 1.25,1.63,
2.5m3|ml equivalents of PM10
Time to Analysis: Cells treated 24h
followed by 48h incubation with
cytochalasin B. Micronuclei frequency
determined.
Water and organic extracts induced a significant
dose-dependent increase in the micronuclei
frequency. After doses of PM from different
regions were normalized for mass differences,
the genotoxic potency was higher for samples
from the industrial area.
Reference: Salonen et al.
(2004,187053)
Species: Mouse
Cell Line: Macrophage
RAW 264.7
PM (Vallila, Finland; busy traffic site;
sping, winter)
Particle Size: Diameter: <10//m
Route: Cell Culture
Dose/Concentration: 15, 50,150, 500,
1000//g/mL of RPMI
Time to Analysis: Air collected 2-7d
periods. Extracted. Exposed 24h.
PAHs decreased from winter to spring. TNF-a
dose-dependently increased and was higher in
spring samples. IL-6 generally increased in spring
but not in winter. NO dose-dependently
increased and was higher in winter. Cell viability
generally decreased but there were no
consistent potency differences between the
samples. Generally, proinflammatory activity,
cytotoxicity and IL-6 were associated with the
insoluble PM fractions. Polymyxin B inhibited IL-
6 and TNF-a. OH and 8-hydroxy-2'-
deoxyguanosine dose-dependently increased and
were higher in the spring and winter,
respectively.
July 2009
D-186
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle
Exposure
Effects
Reference: Seaton et al.
2005
Species: Human
Cell Line: Alveolar
epithelial A549
PM (3 busy London underground (LU)
stations and cabs) (LU dust in PM2.6
samples: iron oxide 64-71%,
chromium 0.1-0.2%, manganese 0.5-
1%, copper < 0.1-0.9%; respirable
dust samples: 1-2%)
Particle Size: Diameter: <2.5//m,
10 /jm, Median diameter: 0.4/ym
Route: Cell Culture
Dose/Concentration: Stations: 270-
480/yg/m3 PM2.B, 14000-29000
particles/cm3: Cabs: 130-200//g|m3
PM2.6,17000-23000 particles/cm3:
Assays: 1,10, 50,100/yglmL
Time to Analysis: 3d measurements of
LU and cabs. Air collected for lab tests.
Exposed 8, 24h.
PM10 caused less LDH release, IL-8 stimulation
and free radical activity than LU dust particles
that contained PM2.6. Chelation had little effect
on PM10 soluble components.
Reference: Sevastyanova
et al. (2007,156969)
Species: Human
Cell Line: HepG2 cell line,
embryonic lung diploid
fibroblasts (HEL), or acute
monocytic leukemia cells
(THP-1)
PM10 (Prague, Czech Republic:
Ko'sice; Slovak Republic: Sofia,
Bulgaria) (urban, summer, winter)
(organic extracts)
Particle Size: Diameter: 10 //in
Route: Cell Culture
Dose/Concentration: 10-100 //gfml
Time to Analysis: Samples collected
24h daily 3m. HepG2 and THP-1 cells
treated for 24h.
DNA adducts were observed in all cell types
evaluated. Highest adduct levels were observed
in HepG2 cells, followed by HEL and THP-1 cells.
A correlation between DNA adduct levels and
carcinogenic PAH content was observed in
HepG2 cells at 50/yg/ml.
Reference: Shi et al.
(2003, 088248)
Species: Human
Cell Line: Epithelial lung
A549
PM (Dusseldorf, Germany, July-Dec.)
Particle Size: Fine diameter: < 2.5
/jm; Coarse diameter: 10-2.5/jm
Route: Cell Culture
Dose/Concentration: Fine: 0.57-
2.49mg: Coarse: 0.66-1.89mg:
Concentration: 0.57mg/mL
Time to Analysis: Weekly samplings
July-Dec 1999. Electron spin response 8-
hydrodeoxyguanosine induction,
measurement.
Coarse and fine particles generated OH, but
coarse particles had significantly higher OH
formation as well as 8-hydroxy-2'-
deoxyguanosine formation. 8-hydroxy-2'-
deoxyguanosine and OH had a significant
correlation.
Reference: Skarek et al.
(2007,096814)
Species: Rat
Cell Line: Modified
hepatoma H4IIE.Iuc; SOS:
£ coli PQ37 (±S9)
PM (urban: List! nad Laben, Karvina;
background: Cervenohorske sedlo,
Kosetice - Czech Republic: July)
(organic extracts, TSP) GP (gas
phase)
Particle Size: Diameter: <2.5//m
Route: Cell Culture
Dose/Concentration: SOS: 8, 4, 2,1m3
ml1; Dioxin: TSP+GP- 8,1.33, 0.22,
0.04m3 ml'1, PM2.6+GP: 4, 0.66, 0.11,
0.02m3 ml-'
Time to Analysis: 24h samples July
2002. Extracted. SOS chromotest: 22h
incubation. Dioxin toxicity test: 24h
exposure.
The urban areas had a much greater level of
carcinogenic PAHs and overall number of PAHs
than the background areas. Significant
genotoxic activity was only detected at
TSP + GP without S9 from urban areas.
PM2.6+GP had lower dioxin activity at the urban
areas, but similar levels of toxicity were seen for
both treatments in the background areas.
Reference: Song et al.
(2007,155306)
Species: S. typhimurium
Strain: TA98, TA100
Cell Line: Rat fibrocytes
L-929 cells
PM (soluble organic fraction (SOF)
extracts from diesel engines using
fuels blended with ethanol by volume:
EO ¦ base diesel fuel: E5 ¦ 5%: E10 ¦
10%;E15-15%;E20-20%)
Particle Size: Density (gfcm3): EO-
0.8379; E5- 0.8349; E10- 0.8324;
E15- 0.8301; E20- 0.8279
Route: Cell Culture
Dose/Concentration: Ames Assay:
0.025, 0.05, 0.1 mg/plate; Comet Assay:
0.125, 0.25, 0.5,1.0mg|mL
Time to Analysis: Samples extracted
24h. Ames and comet assays performed
All PM extracts induced higher mutational
response in TA98 (3- to 5-fold increase over
spontaneous) than in TA100 (2-to 3-fold
increase). The highest brake specific revertants
(BSR) ±S9 in both strains occurred with E20
and lowest BSR was in E5 (except in TA98 -
S9). EO and E20 caused more significant DNA
damage (similar in effect) than the other
extracts. Damage was dose-dependent but
variable with increasing ethanol volume.
Reference: Ueng et al.
(2005, 097054)
Species: Human
Cell Line: Lung epithelium
CL5 (cancerous),
bronchial epithelial BEAS-
2B (noncancerous), WI-38
normal lung fibroblast
MEP (Yamaha cabin motorcycle 2-
strok 50-cc engine)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 1,10,100, 200
/yg/mL
Time to Analysis: Exhaust collected,
extracted. cDNA microarray analysis.
RT-PCR: 2h. ELISA: 12h incubation.
Centrifuged 24h post-treatment.
Bioactivity: 12h incubation. Centrifuged
24h post-treatment. Medium replaced
48h post-incubation. Fibroblasts
determined 96h post-incubation. Time
response studies: 3-48h treatment.
Concentration response studies: 6h
treatment.
Drug metabolism array study: MEP increased
CYP1A1, CYP3A7 and UGT2B.
Cytokine array study: MEP increased
fibroblast growth factor (FGF)-6, FGF-9, IL-1 a,
IL-22 and vascular endothelial growth factor
(VEGF)-D mRNA.
Oncogene, tumor suppressor, estrogen
signaling pathway: MEP increased fra-1, c-src,
SHC, p21, C0X7RP, and decreased p53 and Rb
expression.
RT-PCR: MEP increased CYP1A1, CYP1B1, IL-6,
IL-11, IL-1 a, FGF-6, FGF-9, VEGF-D, fra-1 and
p21.
Concentration and time responses:
Concentration and time-dependent increases
occurred for FGF-9, IL-1a, IL-6, IL-11, but
decreased time-dependently after 6h exposure.
BEAS-2B Cells: MEP had concentration-
dependent increases on CYP1A1 and CYP1B1
but did not affect anything else.
Peroxide, MEP+NAC, WI-38 Cells: MEP
increased peroxide production. The MEP+NAC
treatment reduced MEP-elevated levels of IL-1 a,
IL-6, FGF-9, VEGF-D to control levels. Fibroblasts
increased in WI-38 cells.
July 2009
D-187
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Particle
Exposure
Effects
Reference: Umbuzeiro et
al. (2008,190491)
Species: Salmonella
typhimurium
Strain: TA98, YG1041
(+/¦ S9)
PM (urban; Sao Paulo, Brazil-
Cerqueira Cesar street station,
Ibirapuera park station) (winter- June
17,18; average temperature: 16°C)
(EOM)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Cerqueira Cesar:
UPM- 156/yg/m3, EOM- 57.7mg/total
UPM; Ibirapuera Park: UPM- 32/yg/m3,
EOM- 41.7mg/total UPM; Salmonella
assay-0.5,1,5,10, 50,100 UPM
equiv/plate |//g)
Time to Analysis: Tests performed 20d
after collection date. Organic extraction
20h. PAH fractionation.
Salmonellafmicrosome assay.
The TSP and EOM were similar for both sites.
The PAH fraction had very low mutagenicity for
the Cerqueira Cesar sample in the YG1041
strain and no mutagenicity for the Ibirapuera
sample. Nitro-PAH and oxy-PAH had similar
mutagenetic activities from both samples. S9
decreased mutagenicity in nitro-PAH but was
increased in oxy-PAH. DNA adduct levels were
dose-dependent and not different between the
two sites.
Reference: Upadhyay et
al. (2003, 097370)
Species: Human
Cell Line: Alveolar
epithelial A549
PM (Dusseldorf, Germany) (Particles
contain carbon (19.70%),
hydrogen(1.4%),nitrogen (< .05%),
oxygen(14.12%), sulfur (2.09%), ash
(63.24%)) (lonizable metals
concentrations (ppm): Co(103),
Cu(48),Cr(104),Fe(14,521), Mn(21.3),
Ni(1,519),Ti(131), V(2,767)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 1, 5, 25,100
/yg/cm2; 10, 25, 50,100 /yglcm2
Time to Analysis: Air collected. Cells
grown 24h. PM treatments 1, 4, 8,12,
24h.
PM induced dose- and time-dependent reductions
in ds-DNA due to the formation of DNA-SB. The
soluble component caused higher DNA damage.
Apoptosis and DNA fragmentation increased
dose-dependently. decreased dose-
dependently in control cells, but not in cells with
Bcl-xl overexpression. PM caused activation of
caspase 9. Pretreatment with iron chelators or a
free radical scavenger reduced PM-induced DNA-
SB formation, DNA fragmentation, caspase 9
activation, and weakened AH^m reductions.
Reference: Valavanidis et PM (TSP: high volume pumps, Athens;
al. (2005, 096432)
Cell Line: NR
DEP: 2.0L engine GM Astra; GEP:
1,6L passenger vehicle Ford; Wood
smoke soot: domestic fireplace
exhaust chimney; PMio: high volume
sampling system, Athens; PM2.6: high
volume cascade impactor (Anderson)
system
Particle Size: Diameter: >10.2-
< 0.41 /jm
Route: Incubation
Dose/Concentration: 20, 40mg|5mL
Time to Analysis: PM collected 30min-
1 h or 24h basis. Incubated with H2O2
and 2' deoxyguanosine (dG). Stored 3-7d
at -20°C. Fenton reaction. EPR analysis.
PM generated OH by a Fenton reaction, which
is increased by the addition of EDTA but
inhibited by deferoxamine. PM dose-dependently
induced dG hydroxylation and 8-hydroxy-2'-
deoxyguanosine formation. Transition metals Ni,
V, Co, Cr that are capable of redox cycling
electron producing ROS were found in the PM
samples.
Reference: Xu and Zhang
(2004, 097231) Species:
Human
Cell Line: Lung epithelial
A549
PM (Taiyuan, Beijing; Nov-Feb)
(Taiyuan: coal-fume pollution; Beijing:
coal-fume and vehicle exhaust)
Particle Size: Diameter: 2.5 /jm
Route: Cell Culture	Taiyuan had a significantly higher daily PM2.6
Dose/Concentration: 5, 50, 200/yg/mL avera8e !han Beiiin9- H™38 show" thuatuthe u
smaller the particulate diameter, the higher the
Time to Analysis: Air samples	concentration of BaP and Pb. A dose- and time-
collected. Cells incubated 12-24h. Comet response relationship was seen in DNA
assaV-	fragmentation.
July 2009
D-188
DRAFT-DO NOT CITE OR QUOTE

-------
Annex D References
Aam BBi Fonnum F. (2007). ROS scavenging effects of organic extract of diesel exhaust particles on
human neutrophil granulocytes and rat alveolar macrophages. Toxicology, 230- 207-18. 155123
Abou Chakra OR; Joyeux Mi Nerriere Ei Strub MPi Zmirou-Navier D. (2007). Genotoxicity of organic
extracts of urban airborne particulate matter- an assessment within a personal exposure study.
Chemosphere, 66: 1375-81. 098819
Adamson IYRi Vincent Ri Bakowska J. (2003). Differential production of metalloproteinases after
instilling various urban air particle samples to rat lung. , 29: 375-388. 087943
Agopyan Ni Head Ji Yu Si Simon SA. (2004). TRPV1 receptors mediate particulate matter-induced
apoptosis. Am J Physiol Lung Cell Mol Physiol, 286: 563-572. 156198
Agopyan Ni Li Li Yu Si Simon SA. (2003). Negatively charged 2" and 10-"mu"m particles activate
vanilloid receptors, increase cAMP, and induce cytokine release. Toxicol Appl Pharmacol, 186:
63-76. 056065
Ahn E-Ki Yoon H-Ki Jee Bo Ki Ko H-Ji Lee K-Hi Kim Hyung Ji Lim Y. (2008). COX-2 expression and
inflammatory effects by diesel exhaust particles in vitro and in vivo. Toxicol Lett, 176: 178-187.
156199
Ahsan MKi Nakamura Hi Tanito Mi Yamada Ki Utsumi Hi Yodoi J. (2005). Thioredoxin-1 suppresses
lung injury and apoptosis induced by diesel exhaust particles (DEP) by scavenging reactive
oxygen species and by inhibiting DEP-induced downregulation of Akt. Free Radic Biol Med, 39:
1549-1559. 156200
Alfaro-Moreno Ei Martinez Li Garcia-Cuellar Ci Bonner JCi Murray JCi Rosas Ii Rosales SPi Osornio-
Vargas AR. (2002). Biologic effects induced in vitro by PM10 from three different zones of Mexico
City. Environ Health Perspect, 110: 715-720. 156204
Amakawa Ki Terashima Ti Matsuzaki Ti Matsumaru Ai Sagai Mi Yamaguchi K. (2003). Suppressive
effects of diesel exhaust particles on cytokine release from human and murine alveolar
macrophages. , 29: 149-164. 156211
Amara Ni Bachoual Ri Desmard Mi Golda Si Guichard Ci Lanone Si Aubier Mi Ogier-Denis Ei
Boczkowski J. (2007). Diesel exhaust particles induce matrix metalloprotease-1 in human lung
epithelial cells via a NADP(H) oxidase/NOX4 redox-dependent mechanism. Am J Physiol Lung
Cell Mol Physiol, 293: L170-181. 156212
Andre Ei Stoeger Ti Takenaka Si Bahnweg Mi Ritter Bi Karg Ei Lentner Bi Reinhard Ci Schulz Hi Wjst
M. (2006). Inhalation of ultrafine carbon particles triggers biphasic pro-inflammatory response
in the mouse lung. Eur Respir J, 28: 275-285. 091376
Anselme Fi Loriot Si Henry J-Pi Dionnet Fi Napoleoni J-Gi Thnillez Ci Morin J-P. (2007). Inhalation of
diluted diesel engine emission impacts heart rate variability and arrhythmia occurrence in a rat
model of chronic ischemic heart failure. , 81: 299-307. 097084
Anseth JWi Goffin AJi Fuller GGi Ghio AJi Kao PNi Upadhyay D. (2005). Lung surfactant gelation
induced by epithelial cells exposed to air pollution or oxidative stress. Am J Respir Cell Mol Biol,
33: 161-168. 088646
Antonini JMi Taylor MDi Leonard SSi Lawryk NJi Shi Xi Clarke RWi Roberts JR. (2004). Metal
composition and solubility determine lung toxicity induced by residual oil fly ash collected from
different sites within a power plant. Mol Cell Biochem, 255: 257-265. 097199
Apicella Ci Custidiano Ai Miranda Si Novoa Li Dokmetjian Ji Gentile T. (2006). Differential
macrophage modulation of asymmetric IgG antibody synthesis by soluble or particulate stimuli.
, 103: 177-185. 096586
Arantes-Costa Fi Lopes Fi Toledo Ai Magliarelli-Filho Pi Moriya Hi Carvalho-Oliveira Ri Mauad Ti
Saldiva Pi Martins M. (2008). Effects of residual oil fly ash (ROFA) in mice with chronic allergic
pulmonary inflammation. Toxicol Pathol, 36: 680. 187137
Araujo JAi Barajas Bi KLeinman Mi Wang Xi Bennett BJi Gong KWi Navab Mi Harkema Ji Sioutas Ci
Lusis AJi Nel AE. (2008). Ambient particulate pollutants in the ultrafine range promote early
atherosclerosis and systemic oxidative stress. Circ Res, 102: 589-596. 156222
Archer AJi Cramton JLi Pfau JCi Colasurdo Gi Holian A. (2004). Airway responsiveness after acute
exposure to urban particulate matter 1648 in a DO1110 murine model. Am J Physiol, 286: L337"
L343. 088097
July 2009
D-189
DRAFT-DO NOT CITE OR QUOTE

-------
Arimoto Ti Takano H; Inoue Ki Yanagisawa R; Yoshino Si Yamaki Ki Yoshikawa T. (2007). Pulmonary
exposure to diesel exhaust particle components enhances circulatory chemokines during lung
inflammation. , 20- 197-201. 097973
Arrieta DEi Ontiveros CCi Li W"Wi Garcia JHi Denison MSi McDonald JDi Burchiel SWi Washburn
BS. (2003). Aryl hydrocarbon receptor-mediated activity of particulate organic matter from the
Paso del Norte airshed along the US-Mexico border. ,111- 1299-1305. 096210
Auger Fi Gendron MCi Chamot C> Marano F> Dazy AC. (2006). Responses of well-differentiated nasal
epithelial cells exposed to particles' role of the epithelium in airway inflammation. Toxicol Appl
Pharmacol, 215: 285-294. 156235
Bachoual R; Boczkowski J, Goven Di Amara N> Tabet Li On Di Lecon-Malas V; Aubier Mi Lanone S.
(2007). Biological effects of particles from the paris subway system. , 20: 1426-1433. 155667
Bagate Ki Meiring JJ; Cassee FR; Borm PJA. (2004). The effect of particulate matter on resistance and
conductance vessels in the rat. Inhal Toxicol, 16: 431-436. 055638
Bagate Ki Meiring JJ; Gerlofs-Nijland ME; Cassee FR; Borm PJA. (2006). Signal transduction
pathways involved in particulate matter induced relaxation in rat aorta--spontaneous
hypertensive versus Wistar Kyoto rats. Toxicol In Vitro, 20: 52"62. 097608
Bagate K> Meiring JJ; Gerlofs-Nijland ME; Cassee FR; Wiegand H; Osornio-Vargas A; Borm PJA.
(2006).	Ambient particulate matter affects cardiac recovery in a Langendorff ischemia model. ,
18: 633-643. 096157
Bao L; Chen S; Wu L; Hei Tom K; Wu Y; Yu Z; Xu A. (2007). Mutagenicity of diesel exhaust particles
mediated by cell-particle interaction in mammalian cells. Toxicol Sci, 229: 91-100. 097258
Barrett EG; Henson RD; Seilkop SK; McDonald JD; Reed MD. (2006). Effects of hardwood smoke
exposure on allergic airway inflammation in mice. Inhal Toxicol, 18: 33"43. 155677
Bartoli CR; Wellenius GA; Coull BA; Akiyama I; Diaz EA; Lawrence J; Okabe K; Verrier RL; Godleski
JJ. (2009). Concentrated ambient particles alter myocardial blood flow during acute ischemia in
conscious canines. Environ Health Perspect, 117: 333-337. 179904
Bartoli CR; Wellenius GA; Diaz EA; Lawrence J; Coull BA; Akiyama I; Lee LM; Okabe K; Verrier RL;
Godleski JJ. (2009). Mechanisms of Inhaled Fine Particulate Air Pollution-Induced Arterial
Blood Pressure Changes. Environ Health Perspect, 117: 361-366. 156256
Batalha JR; Saldiva P H; Clarke RW; Coull BA; Stearns RC; Lawrence J; Murthy GG; Koutrakis P;
Godleski JJ. (2002). Concentrated ambient air particles induce vasoconstriction of small
pulmonary arteries in rats. Environ Health Perspect, 110: 1191-1197. 088109
Baulig A; Blanchet S; Rumelhard M; Lacroix G; Marano F; Baeza-Squiban A. (2007). Fine urban
atmospheric particulate matter modulates inflammatory gene and protein expression in human
bronchial epithelial cells. , 12: 771-82. 151733
Bayram H; Ito K; Issa R; Ito M; Sukkar M; Chung KF. (2006). Regulation of human lung epithelial cell
numbers by diesel exhaust particles. Eur Respir J, 27: 705-713. 088439
Becher R; Bucht A; Ovrevik J; Hongslo Jan K; Dahlman Hans J; Samuelsen Jan T; Schwarze Per E.
(2007).	Involvement of NADPH Oxidase andiNOS in Rodent Pulmonary Cytokine Responses to
Urban Air and Mineral Particles. , 19: 645-55. 097125
Beck-Speier I; Dayal N; Karg E; Maier KL; Schumann G; Schulz H; Semmler M; Takenaka S;
Stettmaier K; Bors W', Ghio A; Samet JM; Heyder J. (2005). Oxidative stress and lipid mediators
induced in alveolar macrophages by ultrafine particles. Free Radic Biol Med, 38: 1080-1092.
156262
Becker S; Dailey LA; Soukup JM; Grambow SC; Devlin RB; Huang YC. (2005). Seasonal variations in
air pollution particle-induced inflammatory mediator release and oxidative stress. Environ
Health Perspect, 113: 1032-1038. 088592
Becker S; Mundandhara S; Devlin RB; Madden M. (2005). Regulation of cytokine production in human
alveolar macrophages and airway epithelial cells in response to ambient air pollution particles:
further mechanistic studies. Toxicol Appl Pharmacol, 207: 269-275. 088590
Bhattacharyya SN; Dubick MA; Yantis LD; Enriquez JI; Buchanan KC; Batra SK; Smiley RA. (2004).
In vivo effect of wood smoke on the expression of two mucin genes in rat airways. Inflammation,
28: 67-76. 088095
Binkova B; Chvatalova I; Lnenickova Z; Milcova A; Tulupova E; Farmer PB; Sram RJ. (2007). PAH"
DNA adducts in environmentally exposed population in relation to metabolic and DNA repair
gene polymorphisms. , 620: 49-61. 156273
July 2009
D-190
DRAFT-DO NOT CITE OR QUOTE

-------
Bitterle EJ Karg E; Schroeppel A; Kreyling WG; Tippe A; Ferron GA; Schmid O; Heyder J; Maier KL;
Hofer T. (2006). Dose-controlled exposure of A549 epithelial cells at the air-liquid interface to
airborne ultrafine carbonaceous particles. Chemosphere, 65- 1784-1790. 156276
Blanchet S> Ramgolam K> Baulig A; Marano F> Baeza-Squiban A. (2004). Fine particulate matter
induces amphiregulin secretion by bronchial epithelial cells. Am J Respir Cell Mol Biol, 230:
421-427. 087982
Bonvallot V; Baeza-Squiban A; Baulig A; Brulant Si Boland Si Muzeau Fi Barouki Ri Marano F. (2001).
Organic compounds from diesel exhaust particles ehcit a proinflammatory response in human
airway epithelial cells and induce cytochrome p450 1A1 expression. Am J Respir Cell Mol Biol,
25: 515-521. 156283
Brown DMi Hutchison Li Donaldson Ki Stone V. (2007). The effects of PM10 particles and oxidative
stress on macrophages and lung epithelial cells: modulating effects of calcium-signaling
antagonists. Am J Physiol Lung Cell Mol Physiol, 292: 1444-1451. 156300
Brown LEi Trought KRi Bailey Cli Clemons JH. (2005). 2,3,7,8-TCDD equivalence and mutagenic
activity associated with PM10 from three urban locations in New Zealand. , 349: 161-174.
095919
Bunger Ji Krahl Ji Munack Ai Ruschel Yi Schroder Oi Emmert Bi Westphal Gi Muller Mi Hallier Ei
Bruning T. (2007). Strong mutagenic effects of diesel engine emissions using vegetable oil as
fuel. Arch Toxicol, 81: 599-603. 156305
Bunger Ji Krahl Ji Weigel Ai Schroder Oi Bruning Ti Muller Mi Hallier Ei Westphal G. (2006).
Influence of fuel properties, nitrogen oxides, and exhaust treatment by an oxidation catalytic
converter on the mutagenicity of diesel engine emissions. Arch Toxicol, 80: 540-546. 156303
Burchiel SWi Lauer FTi Dunaway SLi Zawadzki Ji McDonald JDi Reed MD. (2005). Hardwood smoke
alters murine splenic T cell responses to mitogens following a 6-month whole body inhalation
exposure. Toxicol Appl Pharmacol, 202: 229-236. 088090
Burchiel SWi Lauer FTi McDonald JDi Reed MD. (2004). Systemic immunotoxicity in AJ mice following
6-month whole body inhalation exposure to diesel exhaust. Toxicol Appl Pharmacol, 196: 337"
345. 055557
Calcabrini Ai Meschini Si Marra Mi Falzano Li Colone Mi De Berardis Bi Paoletti Li Arancia Gi
Fiorentini C. (2004). Fine environmental particulate engenders alterations in human lung
epithelial A549 cells. , 95: 82-91. 096865
Calderon-Garciduenas Li Maronpot RRi Torres-Jardon Ri Henriquez-Roldan Ci Schoonhoven Ri Acuna-
Ayala Hi Villarreal-Calderon Ai Nakamura Ji Fernando Ri Reed Wi Azzarelli Bi Swenberg JA.
(2003). DNA damage in nasal and brain tissues of canines exposed to air pollutants is associated
with evidence of chronic brain inflammation and neurodegeneration. Toxicol Pathol, 31: 524-538.
156316
Campbell Ai Oldham Mi Becaria Ai Bondy SCi Meacher Di Sioutas Ci Misra Ci Mendez LBi KLeinman
M. (2005). Particulate matter in polluted air may increase biomarkers of inflammation in mouse
brain. Neurotoxicology, 26: 133-140. 087217
Campen MJi Babu NSi Helms GAi Pett Si Wernly Ji Mehran Ri McDonald JD. (2005). Nonparticulate
components of diesel exhaust promote constriction in coronary arteries from ApoE -/- mice.
Toxicol Sci, 88: 95-102. 083977
Campen MJi McDonald JDi Gigliotti APi Seilkop SKi Reed MDi Benson JM. (2003). Cardiovascular
effects of inhaled diesel exhaust in spontaneously hypertensive rats. , 3: 353-361. 055626
Campen MJi McDonald JDi Reed MDi Seagrave J. (2006). Fresh gasoline emissions, not paved road
dust, alter cardiac repolarization in ApoE-/- mice. Cardiovasc Toxicol, 6: 199-210. 096879
Cao Di Tal TLi Graves LMi Gilmour Ii Linak Wi Reed Wi Bromberg PAi Samet JM. (2007). Diesel
exhaust particulate "induced activation of Stat3 requires activities of EGFR and Src in airway
epithelial cells. Am J Physiol Lung Cell Mol Physiol, 292: L422-L429. 156322
Cao Qi Zhang Si Dong Ci Song W. (2007). Pulmonary responses to fine particles: Differences between
the spontaneously hypertensive rats and wistar kyoto rats. , 171: 126-37. 097491
Carter JM; Corson N> Driscoll KE; Elder A; Finkelstein JN; Harkema JN; Gelein R; Wade-Mercer P;
Nguyen K> Oberdorster G. (2006). A comparative dose-related response of several key pro- and
antiinflammatory mediators in the lungs of rats, mice, and hamsters after subchronic inhalation
of carbon black. J Occup Environ Med, 48: 1265-1278. 095936
July 2009
D-191
DRAFT-DO NOT CITE OR QUOTE

-------
Carvalho-Oliveira R; Pozo EMK Lobo DJA; Lichtenfels AJFC; Martins-Junior HA; Bustilho JOWV
Saiki Mi Sato IMi Saldiva PHN. (2005). Diesel emissions significantly influence composition and
mutagenicity of ambient particles' a case study in Sao Paulo, Brazil. Environ Res, 98- 1-7.
077898
Cassee FRi Boere AJFi Fokkens PHBi Leseman DLACi Sioutas Ci Kooter IMi Dormans JAMA. (2005).
Inhalation of concentrated particulate matter produces pulmonary inflammation and systemic
biological effects in compromised rats. J Toxicol Environ Health A, 68- 773-796. 087962
Chan C-Ci Chuang K-Ji Chien LrCi Chen W-Ji Chang W"T. (2006). Urban air pollution and emergency
admissions for cerebrovascular diseases in Taipei, Taiwan. Eur Heart J, 27- 1238-1244. 090193
Chang C"Ci Chen S"Hi Ho S"Hi Yang C-Yi Wang H"Di Tsai M-L. (2007). Proteomic analysis of proteins
from bronchoalveolar lavage fluid reveals the action mechanism of ultrafine carbon black-
induced lung injury in mice. Proteomics, !'¦ 4388-4397. 097475
Chang C-Ci Chiu H-F; Wu Y-S; Li Y-C; Tsai M-L; Shen C-K; Yang C-Y. (2005). The induction of
vascular endothelial growth factor by ultrafine carbon black contributes to the increase of
alveolar-capillary permeability. Environ Health Perspect, 113- 454-460. 097776
Chang C-C; Hwang J-S; Chan C"C; Cheng T-J. (2007). Interaction effects of ultrafine carbon black with
iron and nickel on heart rate variability in spontaneously hypertensive rats. , 115- 1012-1017.
155720
Chang C-C; Hwang J-S; Chan C"C; Wang P-Y; Hu T"H; Cheng T-J. (2004). Effects of concentrated
ambient particles on heart rate, blood pressure, and cardiac contractility in spontaneously
hypertensive rats. Inhal Toxicol, 16- 421-429. 055637
Chauhan V; Breznan D; Goegan P; Nadeau D; Karthikeyan S; Brook JR.; Vincent R. (2004). Effects of
ambient air particles on nitric oxide production in macrophage cell lines. Cell Biol Toxicol, 20-
221-239. 096682
Chauhan V; Breznan D; Thomson E; Karthikeyan S; Vincent R. (2005). Effects of ambient air particles
on the endothelin system in human pulmonary epithelial cells (A549). Cell Biol Toxicol, 21- 191-
205. 155722
Che W; Zhang Z; Zhang H; Wu M; Liang Y; Liu F; Shu Y; Li N. (2007). Compositions and oxidative
damage of condensate, particulate and semivolatile organic compounds from gasoline exhausts.
Environ Toxicol Pharmacol, 24- 11-18. 096460
Chen LC; Hwang JS. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs)
in mice IV Characterization of acute and chronic effects of ambient air fine particulate matter
exposures on heart-rate variability. Inhal Toxicol, 17- 209-216. 087218
Chen LC; Nadziejko C. (2005). Effects of subchronic exposures to concentrated ambient particles
(CAPs) in mice V CAPs exacerbate aortic plaque development in hyperlipidemic mice. Inhal
Toxicol, 17: 217-224. 087219
Cheng MD; Malone B; Storey JM. (2003). Monitoring cellular responses of engine-emitted particles by
using a direct air-cell interface deposition technique. Chemosphere, 53: 237-243. 156337
Cho H-Y; Jedlicka AE; Clarke R; Kleeberger SR. (2005). Role of Toll-like receptor-4 in genetic
susceptibility to lung injury induced by residual oil fly ash. , 22: 108-117. 156344
Churg A; Brauer M; del Carmen Avila-Casado M; Fortoul TI; Wright JL. (2003). Chronic exposure to
high levels of particulate air pollution and small airway remodeling. Environ Health Perspect,
111: 714-718. 087899
Churg A; Xie C; Wang X; Vincent R; Wang RD. (2005). Air pollution particles activate NF-kappaB on
contact with airway epithelial cell surfaces. Toxicol Appl Pharmacol, 208: 37"45. 088281
Ciencewicki J; Gowdy K; Krantz QT; Linak WP; Brighton L; Gilmour MI; Jaspers I. (2007). Diesel
exhaust enhanced susceptibility to influenza infection is associated with decreased surfactant
protein expression. , 19: 1121-33. 096557
Corey LM; Baker C; Luchtel Daniel L. (2006). Heart-rate variability in the apolipoprotein E knockout
transgenic mouse following exposure to Seattle particulate matter. J Toxicol Environ Health A,
69: 953-965. 156366
Costa DL; Lehmann JR; Winsett D; Richards J; Ledbetter AD; Dreher KL. (2006). Comparative
pulmonary toxicological assessment of oil combustion particles following inhalation or
instillation exposure. Toxicol Sci, 91: 237-246. 088438
Courtois A; Andujar P; Ladeiro Y; Baudrimont I; Delannoy E; Leblais V; Begueret H; Galland MAB;
BrochardP; Marano F. (2008). Impairment of NO "Dependent Relaxation in Intralobar
Pulmonary Arteries: Comparison of Urban Particulate Matter and Manufactured Nanoparticles.
Environ Health Perspect, 116: 1294. 156369
July 2009
D-192
DRAFT-DO NOT CITE OR QUOTE

-------
Cozzi E> Hazarika Si Stallings HWi Cascio WEi Devlin RBi Lust RMi Wingard CJi Van Scott MR.
(2006). Ultrafine particulate matter exposure augments ischemia-reperfusion injury in mice. Am
J Physiol, 291: H894-H903. 091380
Dagher Zi Garcon Gi Billet Si Verdin Ai Ledoux Fi Courcot Di Aboukais Ai Shirali P. (2007). Role of
nuclear factor-kappa B activation in the adverse effects induced by air pollution particulate
matter (PM25) in human epithelial lung cells (L132) in culture. , 27: 284-90. 097566
Dai Ji Xie C> Vincent Ri Churg A. (2003). Air pollution particles produce airway wall remodeling in rat
tracheal explants. Am J Respir Cell Mol Biol, 29: 352-358. 087944
Day KCi ReedMD; McDonald JD; Keilkop SK; Barrett EG. (2008). Effects of Gasoline Engine
Emissions on Preexisting Allergic Airway Responses in Mice. Inhal Toxicol, 20: 1145-1155.
190204
de Haar C> Hassing L Bol M> Bleumink R; Pieters R. (2005). Ultrafine carbon black particles cause
early airway inflammation and have adjuvant activity in a mouse allergic airway disease model.
. 87: 409-418. 097872
de Haar Ci Hassing L Bol Mi Bleumink Ri Pieters R. (2006). Ultrafine but not fine particulate matter
causes airway inflammation and allergic airway sensitization to co-administered antigen in
mice. Clin Exp Allergy, 36: 1469-1479. 144746
de Haar Ci Kool Mi Hassing Ii Bol Mi Lambrecht BNi Pieters R. (2008). Lung dendritic cells are
stimulated by ultrafine particles and play a key role in particle adjuvant activity. ,121: 1246-54.
187128
De Kok TMi Hogervorst JG; Briede JJ; Van Herwijnen MHi Maas LMi Moonen EJ; Driece HA;
Kleinjans JC. (2005). Genotoxicity and physicochemical characteristics of traffic-related ambient
particulate matter. Environ Mol Mutagen, 46: 71-80. 088656
DeMarini DM; Brooks LR; Warren SH; Kobayashi T; Gilmour ML Singh P. (2004). Bioassay-directed
fractionation and Salmonella mutagenicity of automobile and forklift diesel exhaust particles.
Environ Health Perspect, 112: 814-819. 066329
Dick CAJ; Brown DM; Donaldson K; Stone V. (2003). The role of free radicals in the toxic and
inflammatory effects of four different ultrafine particle types. Inhal Toxicol, 15: 39"52. 036605
Doherty SP; Prophete C; Maciejczyk P; Salnikow K; Gould T; Larson T; Koenig J; Jaques P; Sioutas C;
Zelikoff JT; Lippmann M; Cohen MD. (2007). Detection of changes in alveolar macrophage iron
status induced by select PM25-associated components using iron-response protein binding
activity. , 19: 553-62. 096532
Dong CC; Yin XJ; Ma JYC; Millecchia L; Barger MW; Roberts JR; Zhang X-D; Antonini JM; Ma JKH.
(2005). Exposure of Brown Norway Rats to diesel exhaust particles Prior to ovalbumin (OVA)
sensitization elicits IgE adjuvant activity but attenuates OVA"induced airway inflammation.
Toxicol Sci, 88: 150-160. 088079
Dong CC; Yin XJ; Ma JYC; Millecchia L; Wu Z-X; Barger MW; Roberts JR; Antonini JM; Dey RD; Ma
JKH. (2005). Effect of diesel exhaust particles on allergic reactions and airway responsiveness in
ovalbumin-sensitized Brown Norway Rats. Toxicol Sci, 88: 202-212. 088083
Doornaert B; Leblond V; Galiacy S; Gras G; Planus E; Laurent V; Isabey D; Lafuma C. (2003). Negative
impact of DEP exposure on human airway epithelial cell adhesion, stiffness, and repair. Am J
Physiol Lung Cell Mol Physiol, 284: L119-132. 156410
Dostert C; Petrilli V; Van Bruggen R; Steele C; Mossman BT; Tschopp J. (2008). Innate Immune
Activation Through Nalp3 Inflammasome Sensing of Asbestos and Silica. , 320: 674. 155753
Doyle M; Sexton KG; Jeffries H; Bridge K; Jaspers I. (2004). Effects of 1,3-butadiene, isoprene, and
their photochemical degradation products on human lung cells. Environ Health Perspect, 112:
1488-1495. 088404
Drela N; Zesko I; Jakubowska M; Biernacka M. (2006). CD28 in thymocyte development and peripheral
T cell activation in mice exposed to suspended particulate matter. Toxicol Appl Pharmacol, 215:
179-188. 096352
Duvall RM; Norris GA; Dailey LA; Burke JM; McGee JK; Gilmour MI; Gordon T; Devlin RB. (2008).
Source apportionment of particulate matter in the US and associations with lung inflammatory
markers. Inhal Toxicol, 20: 671-83. 097969
Dvonch JT; Brook RD; Keeler GJ; Rajagopalan S; D'Alecy LG; Marsik FJ; Morishita M; Yip FY; Brook
JR; Timm EJ; Wagner JG; Harkema JR. (2004). Effects of concentrated fine ambient particles on
rat plasma levels of asymmetric dimethylarginine. Inhal Toxicol, 16: 473-480. 055741
Dybdahl M; Risom L; Bornholdt J; Autrup H; Loft S; Wallin H. (2004). Inflammatory and genotoxic
effects of diesel particles in vitro and in vivo. DNA Repair (Amst), 562: 119-131. 089013
July 2009
D-193
DRAFT-DO NOT CITE OR QUOTE

-------
Dybing Ei Lovdal Ti Hetland RBi Lovik Mi Schwarze PE. (2004). Respiratory allergy adjuvant and
inflammatory effects of urban ambient particles. Toxicol Sci, 198- 307-314. 097545
ElAssouli Si AlQahtani Mi Milaat W. (2007). Genotoxicity of Air Borne Particulates Assessed by Comet
and the Samonella Mutagenicity Test in Jeddah, Saudi Arabia. , 4- 216-223. 186914
Elder Ai Gelein Ri Finkelstein Ji Phipps Ri Frampton Mi Utell Mi Kittelson DBi Watts WFi Hopke Pi
Jeong CHi Kim Ei Liu Wi Zhao Wi Zhuo Li Vincent Ri Kumarathasan Pi Oberdorster G. (2004).
On-road exposure to highway aerosols 2 Exposures of aged, compromised rats. Inhal Toxicol, 16
Suppl l: 41-53. 087354
Elder Ai Gelein Ri Finkelstein JNi Driscoll KEi Harkema Ji Oberdorster G. (2005). Effects of
subchronically inhaled carbon black in three species I Retention kinetics, lung inflammation,
and histopathology. Toxicol Sci, 88- 614-629. 088194
Elder ACPi Gelein Ri Azadniv Mi Frampton Mi Finkelstein Ji Oberdorster G. (2004). Systemic effects of
inhaled ultrafine particles in two compromised, aged rat strains. Inhal Toxicol, 16- 461-471.
055642
Endo Oi Sugita Ki Goto Si Amagai Ti Matsushita H. (2003). Mutagenicity of size-fractioned airborne
particles collected with Andersen low pressure impactor. Eisei Kagaku, 49- 22"27. 097260
Erdinger Li Durr Mi Hopker KA. (2005). Correlations between mutagenic activity of organic extracts of
airborne particulate matter, NOx and sulphur dioxide in southern Germany- results of a two-
year study. Environ Sci Pollut Res Int, 12: 10-20. 156423
Evans S"Ai Al-Mosawi Ai Adams RAi Berube KA. (2006). Inflammation, edema, and peripheral blood
changes in lung-compromised rats after instillation with combustion-derived and manufactured
nanoparticles. , 32: 363-378. 097066
Farraj AKi Haykal-Coates Ni Ledbetter ADi Evansky PAi Gavett SH. (2006). Neurotrophin mediation
of allergic airways responses to inhaled diesel particles in mice. , 94: 183-92. 141730
Farraj AKi Haykal-Coates Ni Ledbetter ADi Evansky PAi Gavett SH. (2006). Neurotrophin mediation
of allergic airways responses to inhaled diesel particles in mice. Toxicol Sci, 94: 183-192. 189604
Fedulov AVi Leme Ai Yang Zi Dahl Mi Lim Ri Mariani TJi Kobzik L. (2008). Pulmonary exposure to
particles during pregnancy causes increased neonatal asthma susceptibility. Am J Respir Cell
Mol Biol, 38: 57-67. 097482
Finkelman FDi Yang Mi Orekhova Ti Clyne Ei Bernstein Ji Whitekus Mi Diaz-Sanchez Di Morris SC.
(2004). Diesel exhaust particles suppress in vivo IFN-gamma production by inhibiting cytokine
effects on NK and NKT cells. , 172: 3808-3813. 096572
Finnerty Ki Choi J-Ei Lau Ai Davis-Gorman Gi Diven Ci Seaver Ni Linak William Pi Witten Mi
McDonagh Paul F. (2007). Instillation of coarse ash particulate matter and lipopolysaccharide
produces a systemic inflammatory response in mice. J Toxicol Environ Health A, 70: 1957-1966.
156434
Floyd HSi Chen LCi Vallanat Bi Dreher K. (2009). Fine ambient air particulate matter exposure
induces molecular alterations associated with vascular disease progression within plaques of
atherosclerotic susceptible mice. Inhal Toxicol, 21: 394-403. 190350
Folkmann JKi Risom Li Hansen CSi Loft Si Moller P. (2007). Oxidatively damaged DNA and
inflammation in the liver of dyslipidemic ApoE-/- mice exposed to diesel exhaust particles.
Toxicology, 237: 134-44. 097344
Fritsch Si Diabate Si Krug HF. (2006). Incinerator fly ash provokes alteration of redox equilibrium and
liberation of arachidonic acid in vitro. Biol Chem, 387: 1421-1428. 156452
Fujii Ti Hayashi Si Hogg JCi Mukae Hi Suwa Ti Goto Yi Vincent Ri Van Eeden SF. (2002). Interaction
of alveolar macrophages and airway epithelial cells following exposure to particulate matter
produces mediators that stimulate the bone marrow. Am J Respir Cell Mol Biol, 27: 34-41.
036478
Fujii Ti Hayashi Si Hogg JCi Vincent Ri Van Eeden SF. (2001). Particulate matter induces cytokine
expression in human bronchial epithelial cells. Am J Respir Cell Mol Biol, 25: 265-271. 156455
Fujimaki Hi Kurokawa Y. (2004). Diesel exhaust-associated gas components enhance chemokine
production by cervical lymph-node cells from mice immunized with sugi basic proteins. Inhal
Toxicol, 16: 61-65. 189576
Fujimaki Hi Kurokawa Yi Yamamoto Si Satoh M. (2006). Distinct requirements for interleukin-6 in
airway inflammation induced by diesel exhaust in mice. Immunopharmacol Immunotoxicol, 28:
703-714. 096601
Fujimaki Hi Yamamoto Si Kurokawa Y. (2005). Effect of diesel exhaust on immune responses in
C57BL/6 mice intranasally immunized with pollen antigen. J UOEH, 27: 11-24. 156456
July 2009
D-194
DRAFT-DO NOT CITE OR QUOTE

-------
Fujimoto A; Tsukue Ni Watanabe Mi Sugawara Ii Yanagisawa Ri Takano Hi Yoshida Si Takeda K.
(2005). Diesel exhaust affects immunological action in the placentas of mice. , 20- 431-440.
096556
Furuyama Ai Hirano Si Koike Ei Kobayashi T. (2006). Induction of oxidative stress and inhibition of
plasminogen activator inhibitor-1 production in endothelial cells following exposure to organic
extracts of diesel exhaust particles and urban fine particles. , 80- 154-162. 097056
Gao Fi Barchowsky Ai Nemec AAi Fabisiak JP. (2004). Microbial stimulation by Mycoplasma
fermentans synergistically amplifies IL"6 release by human lung fibroblasts in response to
residual oil fly ash (ROFA) and nickel. Toxicol Sci, 81: 467-479. 087950
Garcon Gi Dagher Zi Zerimech Fi Ledoux Fi Courcot Di Aboukais Ai Puskaric Ei Shirali P. (2006).
Dunkerque City air pollution particulate matter-induced cytotoxicity, oxidative stress and
inflammation in human epithelial lung cells (L132) in culture. Toxicol In Vitro, 20: 519-528.
096633
Gavett SHi Haykal-Coates Ni Copeland L Bi Heinrich J; Gilmour MI. (2003). Metal composition of
ambient PM2.5 influences severity of allergic airways disease in mice. Environ Health Perspect,
111: 1471-1477. 053153
Geng H; Meng Z; Zhang Q. (2005). Effects of blowing sand fine particles on plasma membrane
permeability and fluidity, and intracellular calcium levels of rat alveolar macrophages. , 157:
129-137. 096689
Geng H; Meng Z; Zhang Q. (2006). In vitro responses of rat alveolar macrophages to particle
suspensions and water-soluble components of dust storm PM(2.5). Toxicol In Vitro, 20: 575-584.
097026
Gerlofs-Nijland ME: Rummelhard Mi Boere AJFi Leseman DLACi Duffin Ri Schins RPFi Borm PJAi
Sillanpaa Mi Salonen ROi Cassee FR. (2009). Particle induced toxicity in relation to transition
metal and polycyclic aromatic hydrocarbon contents . Environ Sci Technol, In Press: 1-8. 190353
Gerlofs-Nijland MEi Boere AJi Leseman DLi Dormans JAi Sandstrom Ti Salonen ROi van Bree Li
Cassee FR. (2005). Effects of particulate matter on the pulmonary and vascular system: time
course in spontaneously hypertensive rats. Part Fibre Toxicol, 24i21: 24i2(l):2. 088652
Gerlofs-Nijland MEi Dormans JAi Bloemen HJi Leseman DLi John Ai Boere Fi Kelly FJi Mudway ISi
Jimenez AAi Donaldson Ki Guastadisegni Ci Janssen NAi Brunekreef Bi Sandstrom Ti van Bree
Li Cassee FR. (2007). Toxicity of coarse and fine particulate matter from sites with contrasting
traffic profiles. , 19: 1055-69. 097840
Ghelfi Ei Rhoden CRi Wellenius GAi Lawrence Ji Gonzalez-Flecha B. (2008). Cardiac oxidative stress
and electrophysiological changes in rats exposed to concentrated air particles are mediated by
TRP-dependent pulmonary reflexes. Toxicol Sci, 102: 328-336. 156468
Ghio AJi Cohen MD. (2005). Disruption of iron homeostasis as a mechanism of biologic effect by
ambient air pollution particles. Inhal Toxicol, 17: 709-716. 088272
Ghio AJi Piantadosi CAi Wang Xi Dailey LAi Stonehuerner JDi Madden MCi Yang Fi Dolan KGi
Garrick MDi Garrick LM. (2005). Divalent metal transporter-1 decreases metal-related injury in
the lung. Am J Physiol, 289: L460-L467. 088275
Gilmour Mli McGee Ji Duvall Rachelle Mi Dailey Li Daniels Mi Boykin Ei Cho S"Hi Doerfler Di Gordon
Ti Devlin Robert B. (2007). Comparative toxicity of size-fractionated airborne particulate matter
obtained from different cities in the United States. , 19 Suppl l: 7-16. 096433
Gilmour Mli O'Connor Si Dick CAJi Miller CAi Linak WP. (2004). Differential pulmonary inflammation
and in vitro cytotoxicity of size-fractionated fly ash particles from pulverized coal combustion. J
Air Waste Manag Assoc, 54: 286-295. 057420
Gilmour PSi Morrison ERi Vickers MAi Ford Ii Ludlam CAi Greaves Mi Donaldson Ki MacNee W.
(2005).	The procoagulant potential of environmental particles (PM10). Occup Environ Med, 62:
164-171. 087410
Gilmour PSi Rahman Ii Donaldson Ki MacNee W. (2003). Histone acetylation regulates epithelial IL"8
release mediated by oxidative stress from environmental particles. Retrieved , from . 096959
Gilmour PSi Schladweiler MCi Nyska Ai McGee JKi Thomas Ri Jaskot RHi Schmid Ji Kodavanti UP.
(2006).	Systemic imbalance of essential metals and cardiac gene expression in rats following
acute pulmonary zinc exposure. J Toxicol Environ Health A, 69: 2011-2032. 156472
Gilmour PSi Schladweiler MCi Richards JHi Ledbetter ADi Kodavanti UP. (2004). Hypertensive rats
are susceptible to TLR4-mediated signaling following exposure to combustion source particulate
matter. Inhal Toxicol, 16 Suppl l: 5" 18. 087948
July 2009
D-195
DRAFT-DO NOT CITE OR QUOTE

-------
Gilmour PS; Ziesenis A; Morrison ER; Vickers MA; Drost EM; Ford I; Karg EJ Mossa CJ Schroeppel A;
Ferron GA; Heyder J; Greaves M; MacNee W; Donaldson K. (2004). Pulmonary and systemic
effects of short-term inhalation exposure to ultrafine carbon black particles. Toxicol Appl
Pharmacol, 195- 35"44. 054175
Godleski JJ; Clarke RW; Coull BA; Saldiva PHN; Jiang NF; Lawrence J; Koutrakis P. (2002).
Composition of inhaled urban air particles determines acute pulmonary responses. Ann Occup
Hyg, 46: 419-424. 156478
Gong KW; Zhao W; Li N> Barajas B; Kleinman M; Sioutas C; Horvath S; Lusis AJ; Nel A; Araujo JA.
(2007). Air pollutant chemicals and oxidized lipids exhibit genome wide synergistic effects on
endothelial cells. , 8- R149. 091155
Goto Y; Hogg JC; Shih CH; Ishii H; Vincent R; van Eeden SF. (2004). Exposure to ambient particles
accelerates monocyte release from bone marrow in atherosclerotic rabbits. Am J Physiol, 287:
L79-L85. 088100
Gottipolu RR; Wallenborn JG; Karoly ED; Schladweiler MC; Ledbetter AD; Krantz T; Linak WP; Nyska
A; Johnson JA; Thomas R; Richards JE; Jaskot RH; Kodavanti UP. (2009). One-month diesel
exhaust inhalation produces hypertensive gene expression pattern in healthy rats. Environ
Health Perspect, 117: 38-46. 190360
Gowdy K; Krantz QT; Daniels M; Linak WP; Jaspers I; Gilmour MI. (2008). Modulation of pulmonary
inflammatory responses and antimicrobial defenses in mice exposed to diesel exhaust. Toxicol
Appl Pharmacol, 229(3): 310-319. 097226
Graff DW; Schmitt MT; Dailey LA; Duvall RM; Karoly ED; Devlin RB. (2007). Assessing the role of
particulate matter size and composition on gene expression in pulmonary cells. Inhal Toxicol, 19
Suppl l: 23-28. 156488
Greenwell LL; Moreno T; Richards RJ. (2003). Pulmonary antioxidants exert differential protective
effects against urban and industrial particulate matter. , 28: 101-107. 097478
Gualtieri M; Rigamonti L; Galeotti V; Camatini M. (2005). Toxicity of tire debris extracts on human
lung cell line A549. Toxicol In Vitro, 19: 1001-1008. 097841
Gunnison A; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs)
in mice VI Gene expression in heart and lung tissue. Inhal Toxicol, 17: 225-233. 087956
Gurgueira SA; Lawrence J; Coull B; Murthy GGK; Gonzalez-Flecha B. (2002). Rapid increases in the
steady-state concentration of reactive oxygen species in the lungs and heart after particulate air
pollution inhalation. Environ Health Perspect, 110: 749-755. 036535
Gutierrez-Castillo ME; Roubicek DA; Cebrian-Garcia ME; De Vizcaya-Ruiz A; Sordo-Cedeno M;
Ostrosky-Wegman P. (2006). Effect of chemical composition on the induction of DNA damage by
urban airborne particulate matter. Environ Mol Mutagen, 47: 199-211. 089030
Hamada K; Suzaki Y; Leme A; Ito T; Miyamoto K; Kobzik L; Kimura H. (2007). Exposure of pregnant
mice to an air pollutant aerosol increases asthma susceptibility in offspring. J Toxicol Environ
Health A, 70: 688-695. 091235
Hamoir J; Nemmar A; Halloy D; Wirth D; Vincke G; Vanderplasschen A; Nemery B; Gustin P. (2003).
Effect of polystyrene particles on lung microvascular permeability in isolated perfused rabbit
lungs: role of size and surface properties. Toxicol Appl Pharmacol, 190: 278-285. 096664
Hansen C; Neller A; Williams G; Simpson R. (2007). Low levels of ambient air pollution during
pregnancy and fetal growth among term neonates in Brisbane, Australia. Environ Res, 103: 383"
389. 090703
Hao M; Cornier S; Wang M; Lee James J; Nel A. (2003). Diesel exhaust particles exert acute effects on
airway inflammation and function in murine allergen provocation models. J Allergy Clin
Immunol, 112: 905-914. 096565
Happo MS; Salonen RO; Halinen AL Jalava PI; Pennanen AS; Kosma VM; Sillanpaa M; Hillamo R;
Brunekreef B; Katsouyanni K; Sunyer J; Hirvonen MR. (2007). Dose and time dependency of
inflammatory responses in the mouse lung to urban air coarse, fine, and ultrafine particles from
six European cities. , 19: 227-246. 096630
Harder V; Gilmour P; Lentner B; Karg E; Takenaka S; Ziesenis A; Stampfl A; Kodavanti U; Heyder J;
Schulz H. (2005). Cardiovascular responses in unrestrained WKY rats to inhaled ultrafine
carbon particles. Inhal Toxicol, 17: 29"42. 087371
Harkema JR; Keeler G; Wagner J; Morishita M; Timm E; Hotchkiss J; Marsik F; Dvonch T; Kaminski
N; Barr E. (2004). Effects of concentrated ambient particles on normal and hypersecretory
airways in rats. 056842
July 2009
D-196
DRAFT-DO NOT CITE OR QUOTE

-------
Harrod KS; Jaramillo RJ; Berger JA; Gigliotti AP; Seilkop SK; Reed MD. (2005). Inhaled diesel engine
emissions reduce bacterial clearance and exacerbate lung disease to Pseudomonas aeruginosa
infection in vivo. Toxicol Sci, 83- 155-165. 088144
Harrod KS; Jaramillo RJ; Rosenberger CL; Wang S"Z; Berger JA; McDonald JD; Reed MD. (2003).
Increased susceptibility to RSV infection by exposure to inhaled diesel engine emissions. , 28-
451-463. 097046
Heidenfelder BL; Reif DM; Harkema JR; Cohen Hubal EA; Hudgens EE; Bramble LA; Wagner JG;
Morishita M; Keeler GJ; Edwards SW; Gallagher JE. (2009). Comparative microarray analysis
and pulmonary changes in brown norway rats exposed to ovalbumin and concentrated air
particulates. Toxicol Sci, 108: 207-221. 190026
Hetland RB; Cassee FR; Lag M; Refsnes M; Dybing E; Schwarze PE. (2005). Cytokine release from
alveolar macrophages exposed to ambient particulate matter: heterogeneity in relation to size,
city and season. Part Fibre Toxicol, http: 2006]. 087887
Hetland RB; Cassee FR; Refsnes M; Schwarze PE; Lag M; Boere AJF; Dybing E. (2004). Release of
inflammatory cytokines, cell toxicity and apoptosis in epithelial lung cells after exposure to
ambient air particles of different size fractions. Toxicol In Vitro, 18: 203-212. 097535
Hiramatsu K; Azuma A; Kudoh S; Desaki M; Takizawa H; Sugawara I. (2003). Inhalation of diesel
exhaust for three months affects major cytokine expression and induces bronchus-associated
lymphoid tissue formation in murine lungs. , 29: 607-622. 155846
Hiramatsu K; Saito Y; Sakakibara K; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2005). The effects
of inhalation of diesel exhaust on murine mycobacterial infection. , 31: 405-415. 088285
Hirano S; Furuyama A; Koike E; Kobayashi T. (2003). Oxidative-stress potency of organic extracts of
diesel exhaust and urban fine particles in rat heart microvessel endothelial cells. Toxicology,
187: 161-170. 097345
Holder AL; Lucas D; Goth-Goldstein R; Koshland CP. (2008). Cellular Response to Diesel Exhaust
Particles Strongly Depends on the Exposure Method. Toxicol Sci, 103: 108-115. 093322
Hollingsworth JW; Cook DN; Brass DM; Walker JKL; Morgan DL; Foster WM; Schwartz DA. (2004).
The role of Toll-like receptor 4 in environmental airway injury in mice. , 170: 126-132. 097816
Hougaard KS; Jensen KA; Nordly P; Taxvig C; Vogel U; Saber AT; Wallin H. (2008). Effects of prenatal
exposure to diesel exhaust particles on postnatal development, behavior, genotoxicity and
inflammation in mice. Part Fibre Toxicol, 5- 3. 156570
Huang JY; Liao JW; Liu YC; Lu SY; Chou CP; Chan WH; Chen SU; Ueng TH. (2008). Motorcycle
exhaust induces reproductive toxicity and testicular interleukin-6 in male rats. Toxicol Sci, 103:
137-148. 156574
Huang SL; Hsu MK; Chan CC. (2003). Effects of submicrometer particle compositions on cytokine
production and lipid peroxidation of human bronchial epithelial cells. Environ Health Perspect,
111: 478-482. 087376
Hutchison GR; Brown DM; Hibbs LR; Heal MR; Donaldson K; Maynard RL; Monaghan M; Nicholl A;
Stone V. (2005). The effect of refurbishing a UK steel plant on PM10 metal composition and
ability to induce inflammation. Respir Res, 6: 43. 097750
Hwang B-F; Lee Y"L; Lin Y-C; Jaakkola JJE Guo YL. (2005). Traffic related air pollution as a
determinant of asthma among Taiwanese school children. Thorax, 60: 467-473. 089454
Iba MM; Fung J; Chung L; Zhao J; Winnik B; Buckley BT; Chen LC; Zelikoff JT; Kou YR. (2006).
Differential inducibility of rat pulmonary CYP1A1 by cigarette smoke and wood smoke. , 606: 1-
11. 156582
Ichinose T; Takano H; Sadakane K; Yanagisawa R; Kawazato H; Sagai M; Shibamoto T. (2003).
Differences in airway-inflammation development by house dust mite and diesel exhaust
inhalation among mouse strains. Toxicol Appl Pharmacol, 187: 29"37. 041525
Ichinose T; Takano H; Sadakane K; Yanagisawa R; Yoshikawa T; Sagai M; Shibamoto T. (2004). Mouse
Strain Differences in Eosinophilic Airway Inflammation Caused by Intratracheal Instillation of
Mite Allergen and Diesel Exhaust Particles. J Appl Toxicol, 24: 69"76. 180367
Imrich A; Ning Y; Lawrence J; Coull B; Gitin E; Knutson M; Kobzik L. (2007). Alveolar macrophage
cytokine response to air pollution particles: oxidant mechanisms. Toxicol Appl Pharmacol, 218:
256-264. 155859
Inoue K-L Takano H; Yanagisawa R; Hirano S; Ichinose T; Shimada A; Yoshikawa T. (2006). The role
of toll-like receptor 4 in airway inflammation induced by diesel exhaust particles. , 80: 275-279.
097815
July 2009
D-197
DRAFT-DO NOT CITE OR QUOTE

-------
Inoue K"L Takano H; Yanagisawa R; Ichinose T> Sakurai M> Yoshikawa T. (2006). Effects of nano
particles on cytokine expression in murine lung in the absence or presence of allergen. , 80- 614"
619. 096720
Inoue K"i; Takano H; Yanagisawa R; Sakurai Mi Ueki Ni Yoshikawa T. (2007). Effects of diesel exhaust
particles on cytokine production by splenocytes stimulated with lipopolysaccharide. , 27- 95-100.
096702
Inoue Ki Takano Hi Yanagisawa Ri Ichinose T> Sadakane K> Yoshino Si Yamaki K> Uchiyama K>
Yoshikawa T. (2004). Components of diesel exhaust particles differentially affect lung expression
of cyclooxygenase-2 related to bacterial endotoxin. J Appl Toxicol, 24- 415-418. 087984
Inoue Ki Takano H; Yanagisawa R; Sakurai M> Ichinose T> Sadakane K> Yoshikawa T. (2005). Effects of
nano particles on antigen-related airway inflammation in mice. , 6- 106. 188444
Inoue E Takano H; Yanagisawa R; Sakurai M> Ueki Ni Yoshikawa T. (2006). Effects of diesel exhaust
on lung inflammation related to bacterial endotoxin in mice. , 99- 346-352. 190142
Ishihara Yi Kagawa J. (2003). Chronic diesel exhaust exposures of rats demonstrate concentration and
time-dependent effects on pulmonary inflammation. Inhal Toxicol, 15- 473-492. 096404
Ishii Hi Fujii Ti Hogg JCi Hayashi Si Mukae Hi Vincent Ri van Eeden SF. (2004). Contribution of IL"1
beta and TNF-alpha to the initiation of the peripheral lung response to atmospheric particulates
(PM10). Am J Physiol, 287: L176-L183. 088103
Ishii Hi Hayashi Si Hogg JCi Fujii Ti Goto Yi Sakamoto Ni Mukae Hi Vincent Ri van Eeden SF. (2005).
Alveolar macrophage-epithelial cell interaction following exposure to atmospheric particles
induces the release of mediators involved in monocyte mobilization and recruitment. Respir Res,
6: 87. 096138
Ito Ki Christensen WFi Eatough DJi Henry RCi Kim Ei Laden Fi Lall Ri Larson TVi Neas Li Hopke PKi
Thurston GD. (2006). PM source apportionment and health effects: 2 An investigation of
intermethod variability in associations between source-apportioned fine particle mass and daily
mortality in Washington, DC. J Expo Sci Environ Epidemiol, 16: 300-310. 088391
Ito Ti Suzuki Ti Tamura Ki Nezu Ti Honda Ki Kobayashi T. (2008). Examination of mRNA expression
in rat hearts and lungs for analysis of effects of exposure to concentrated ambient particles on
cardiovascular function. Toxicol Sci, 243: 271-283. 096823
Izawa Hi Watanabe Gi Taya Ki Sagai M. (2007). Inhibitory effects of foods and polyphenols on
activation of aryl hydrocarbon receptor induced by diesel exhaust particles. , 14: 149-156. 190387
Jacobsen NRi Mrller Pi Cohn CAi Loft Si Vogel Ui Wallin H. (2008). Diesel exhaust particles are
mutagenic in FEl-Muta™ Mouse lung epithelial cells. , 641: 54-57. 156597
Jalava Pi Salonen ROi Halinen Ali Sillanpaa Mi Sandell Ei Hirvonen MR. (2005). Effects of sample
preparation on chemistry, cytotoxicity, and inflammatory responses induced by air particulate
matter. Inhal Toxicol, 17: 107-117. 088648
Jalava PL Salonen ROi Halinen Ali Penttinen Pi Pennanen ASi Sillanpaa Mi Sandell Ei Hillamo Ri
Hirvonen M-R. (2006). In vitro inflammatory and cytotoxic effects of size-segregated particulate
samples collected during long-range transport of wildfire smoke to Helsinki. Toxicol Appl
Pharmacol, 215: 341-353. 155872
Jalava PL Salonen ROi Pennanen ASi Sillanpaa Mi Halinen Ali Happo MSi Hillamo Ri Brunekreef Bi
Katsouyanni Ki Sunyer Ji Hirvonen M-R. (2007). Heterogeneities in inflammatory and cytotoxic
responses of RAW 264.7 macrophage cell line to urban air coarse, fine, and ultrafine particles
from six European sampling campaigns. , 19: 213-225. 096950
Jang A-Si Choi Inseon Si Takizawa Hi Rhim Ti Lee J-Hi Park S"Wi Park C-S. (2005). Additive effect of
diesel exhaust particulates and ozone on airway hyperresponsiveness and inflammation in a
mouse model of asthma. , 20: 759-763. 155313
Jaspers Ii Ciencewicki JMi Zhang Wi Brighton LEi Carson JLi Beck MAi Madden MC. (2005). Diesel
exhaust enhances influenza virus infections in respiratory epithelial cells. Toxicol Sci, 85: 990-
1002.088115
Jimenez LAi Drost EMi Gilmour PSi Rahman Ii Antonicelli Fi Ritchie Hi MacNee Wi Donaldson K.
(2002). PM10-exposed macrophages stimulate a proinflammatory response in lung epithelial
cells via TNF-alpha. Am J Physiol Lung Cell Mol Physiol, 282: L237-248. 156610
Jung Hi Guo Bi Anastasio C> Kennedy IM. (2006). Quantitative measurements of the generation of
hydroxyl radicals by soot particles in a surrogate lung fluid. Atmos Environ, 40: 1043-1052.
132421
Kaan PM; Hegele RG. (2003). Interaction between respiratory syncytial virus and particulate matter in
guinea pig alveolar macrophages. Am J Respir Cell Mol Biol, 28: 697-704. 095753
July 2009
D-198
DRAFT-DO NOT CITE OR QUOTE

-------
Kafoury RM! Madden MC. (2005). Diesel exhaust particles induce the over expression of tumor necrosis
factor-alpha (TNF-alpha) gene in alveolar macrophages and failed to induce apoptosis through
activation of nuclear factor-kappaB (NF-kappaB). Int J Environ Health Res, 2- 107-113. 156617
Karlsson HL; Ljungman AG; Lindbom J; Moller L. (2006). Comparison of genotoxic and inflammatory
effects of particles generated by wood combustion, a road simulator and collected from street and
subway. Toxicol Lett, 165: 203-211. 156625
Karlsson HL; Nilsson L; Moller L. (2005). Subway particles are more genotoxic than street particles
and induce oxidative stress in cultured human lung cells. Chem Res Toxicol, 18: 19-23. 086392
Kato A; Kagawa J. (2003). Morphological effects in rat lungs exposed to urban roadside air. Inhal
Toxicol, 15: 799-818. 089563
Katterman ME; Birchard SJ Seraphin S; Riley Mark R. (2007). Cellular evaluation of the toxicity of
combustion derived particulate matter: influence of particle grinding and washing on cellular
response. Chemosphere, 66: 567-573. 096358
Kendall M; Guntern J; Lockyer NP; Jones FH; Hutton BM; Lippmann M; Tetley TD. (2004). Urban
PM2.5 surface chemistry and interactions with bronchoalveolar lavage fluid. Inhal Toxicol, 16
Suppl l: 115-129. 156634
Khandoga A; Stampfl A; Takenaka S; Schulz H; Radykewicz R; Kreyling W; Krombach F. (2004).
Ultrafine particles exert prothrombotic but not inflammatory effects on the hepatic
microcirculation in healthy mice in vivo. , 109: 1320-1325. 087928
Kim YM; Reed W', Wu W', Bromberg PA; Graves LM; Samet JM. (2005). Zn2+"induced IL"8 expression
involves AP-1, JNK, and ERK activities in human airway epithelial cells. Am J Physiol, 290:
L1028-1035. 088454
Klein-Patel ME; Diamond G; Boniotto M; Saad S; Ryan LK. (2006). Inhibition of beta-defensin gene
expression in airway epithelial cells by low doses of residual oil fly ash is mediated by vanadium.
, 92: 115-125. 097092
Kleinman M; Phalen R. (2006). Toxicological interactions in the respiratory system after inhalation of
ozone and sulfuric acid aerosol mixtures. Inhal Toxicol, 18: 295-303. 088596
Kleinman M; Sioutas C; Stram D; Froines J; Cho A; Chakrabarti B; Hamade A; Meacher D; Oldham M.
(2005). Inhalation of concentrated ambient particulate matter near a heavily trafficked road
stimulates antigen-induced airway responses in mice. J Air Waste Manag Assoc, 55: 1277-1288.
087880
Kleinman MT Sioutas C Chang MC Boere AJ Cassee FR. (2003). Ambient fine and coarse particle
suppression of alveolar macrophage functions. Toxicol Lett, 137: 151-158. 087938
Kleinman MT; Araujo JA; Nel A; Sioutas C; Campbell A; Cong PQ; Li H; Bondy SC. (2008). Inhaled
ultrafine particulate matter affects CNS inflammatory processes and may act via MAP kinase
signaling pathways. Toxicol Lett, 178: 127-130. 190074
Kleinman MT; Hyde DM; Bufalino C; Basbaum C; Bhalla DK; Mautz WJ. (2003). Toxicity of chemical
components of fine particles inhaled by aged rats: effects of concentration. J Air Waste Manag
Assoc, 53: 1080-1087. 053535
Kleinman MT; Sioutas C; Froines JR; Fanning E; Hamade A; Mendez L; Meacher D; Oldham M. (2007).
Inhalation of concentrated ambient particulate matter near a heavily trafficked road stimulates
antigen-induced airway responses in mice. , 19 Suppl l: 117-26. 097082
Knuckles T; Lund A; Lucas S; Campen M. (2008). Diesel exhaust exposure enhances venoconstriction
via uncoupling of eNOS. Toxicol Appl Pharmacol, 230: 346. 191987
Knuckles TL; Dreher KL. (2007). Fine oil combustion particle bioavailable constituents induce
molecular profiles of oxidative stress, altered function, and cellular injury in cardiomyocytes. J
Toxicol Environ Health A, 70: 1824-1837. 156652
Kodavanti UP; Schladweiler MC; Gilmour PS; Wallenborn JG; Mandavilli BS; Ledbetter AD; Christiani
DC; Runge MS; Karoly ED; Costa DL; Peddada S; Jaskot R; Richards JH; Thomas R;
Madamanchi NR; Nyska A. (2008). The role of particulate matter-associated zinc in cardiac
injury in rats. , 116: 13-20. 155907
Kodavanti UP; Schladweiler MC; Ledbetter AD; McGee JK; Walsh L; Gilmour PS; Highfill JW; Davies
D; Pinkerton KE; Richards JH; Crissman K; Andrews D; Costa DL. (2005). Consistent
pulmonary and systemic responses from inhalation of fine concentrated ambient particles: roles
of rat strains used and physicochemical properties. Environ Health Perspect, 113: 1561-1568.
087946
July 2009
D-199
DRAFT-DO NOT CITE OR QUOTE

-------
Koike E> Kobayashi T. (2005). Organic extract of diesel exhaust particles stimulates expression of la
and costimulatory molecules associated with antigen presentation in rat peripheral blood
monocytes but not in alveolar macrophages. Toxicol Appl Pharmacol, 209- 277-285. 088303
Kooter IMJ Boere AJ; Fokkens PH; Leseman DL; Dormans JA; Cassee FR. (2006). Response of
spontaneously hypertensive rats to inhalation of fine and ultrafine particles from traffic-
experimental controlled study. Part Fibre Toxicol, 15- 3-7. 097547
Kristovich R; Knight DA; Long JF; Williams MV; Dutta PK; Waldman WJ. (2004). Macrophage-
mediated endothehal inflammatory responses to airborne particulates' impact of particulate
physicochemical properties. Chem Res Toxicol, 17- 1303-1312. 087963
Kubatova A; Steckler TS; Gallagher JR; Hawthorne SB; Picklo MJSr. (2004). Toxicity of wide-range
polarity fractions from wood smoke and diesel exhaust particulate obtained using hot
pressurized water. Environ Toxicol Chem, 23: 2243-2250. 087986
Kumar VS; Mani U; Prasad AK; Lai K; Gowri V; Gupta A. (2004). Effect of fly ash inhalation on
biochemical and histomorphological changes in rat lungs. , 42: 964-968. 096655
Kyoso M; Narisawa M; Ito E; Ishijima M; Yana K; Kato A; Ito T; Ishihara Y. (2005). Influence of
Exposure to Diesel Emissions in Rats and Distribution Profile for R-R Interval. 186998
Landvik NE; Gorria M; Arlt VM; Asare N; Solhaug A; Lagadic-Gossmann D; Holme JA. (2007). Effects
of nitrated-polycyclic aromatic hydrocarbons and diesel exhaust particle extracts on cell
signalling related to apoptosis: possible implications for their mutagenic and carcinogenic
effects. Toxicology, 231: 159-174. 096722
Last JA; Ward R; Temple L; Pinkerton KE; Kenyon NJ. (2004). Ovalbumin-induced airway
inflammation and fibrosis in mice also exposed to ultrafine particles. , 16: 93-102. 097334
Lee CC; Cheng YW! Kang JJ. (2005). Motorcycle exhaust particles induce IL"8 production through NF-
kappaB activation in human airway epithelial cells. J Toxicol Environ Health A, 68: 1537-1555.
156682
Lei Y-C; Chan C-C; Wang P-Y; Lee C-T; Cheng T-J. (2004). Effects of Asian dust event particles on
inflammation markers in peripheral blood and bronchoalveolar lavage in pulmonary
hypertensive rats. Environ Res, 95: 71-76. 087884
Lei YC; Chen MC; Chan CC; Wang PY; Lee CT; Cheng TJ. (2004). Effects of concentrated ambient
particles on airway responsiveness and pulmonary inflammation in pulmonary hypertensive
rats. Inhal Toxicol, 16: 785-792. 087999
Lei YC; Hwang JS; Chan CC; Lee CT; Cheng TJ. (2005). Enhanced oxidative stress and endothelial
dysfunction in streptozotocin-diabetic rats exposed to fine particles. Environ Res, 99: 335-343.
088660
Lemos M; Mohallen S; Macchione M; Dolhnikoff M; Assunaao J; Godleski J; Saldiva P. (2006). Chronic
exposure to urban air Pollution induces structural alterations in murine pulmonary and
coronary arteries. Inhal Toxicol, 18: 247-253. 088594
Li J; Ghio AJ; Cho SH; Brinckerhoff CE; Simon SA; Liedtke W. (2009). Diesel exhaust particles activate
the matrix-metalloproteinase-1 gene in human bronchial epithelia in a beta-arrestin-dependent
manner via activation of RAS. Environ Health Perspect, 117: 400-409. 190424
Li J; Li Q; Xu J; Li J; Cai X; Liu R; Li Y; Ma J; Li W. (2007). Comparative study on the acute
pulmonary toxicity induced by 3 and 20 nm Ti02 primary particles in mice. Environ Toxicol
Pharmacol, 24: 239-244. 093156
Li N; Kim S; Wang M; Froines JR; Sioutas. (2002). Use of a stratified oxidative stress model to study
the biological effects of ambient concentrated and diesel exhaust particulate matter. Inhal
Toxicol, 14: 459-486. 042080
Li N; Wang M; Oberley TD; Sempf JM; Ne. (2002). Comparison of the pro-oxidative and
proinflammatory effects of organic diesel exhaust particle chemicals in bronchial epithelial cells
and macrophages. J Immunol, 169: 4531-4541. 087451
Li Y-J; Kawada T; Matsumoto A; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2007). Airway
inflammatory responses to oxidative stress induced by low-dose diesel exhaust particle exposure
differ between mouse strains. Exp Lung Res, 33: 227-244. 155929
Li Z; Carter JD; Dailey LA; Huang YC. (2005). Pollutant particles produce vasoconstriction and
enhance MAPK signaling via angiotensin type I receptor. Environ Health Perspect, 11: 1009"
1014. 088647
Li Z; Hyseni X; Carter JD; Soukup JM; Dailey LA; Huang Y-CT. (2006). Pollutant particles enhanced
H202 production from NAD(P)H oxidase and mitochondria in human pulmonary artery
endothelial cells. Am J Physiol Lung Cell Mol Physiol, 291: C357-365. 156693
July 2009
D-200
DRAFT-DO NOT CITE OR QUOTE

-------
Lichtenfels AJFC; Gomes JB; Pieri PC; Miraglia SGEK; Hallak J; Saldiva PHN. (2007). Increased
levels of air pollution and a decrease in the human and mouse male-to-female ration in Sao
Paulo, Brazil. Fertil Steril, 87: 230-232. 097041
Lijinsky W> Reuber MD. (1987). Chronic carcinogenesis studies of acrolein and related compounds.
Toxicol Ind Health, 3: 337-345. 007583
Lindbom J; Gustafsson M> Blomqvist G> Dahl A; Gudmundsson A; Swietlicki E> Ljungman AG. (2007).
Wear particles generated from studded tires and pavement induces inflammatory reactions in
mouse macrophage cells. , 20: 937-946. 155934
Lippmann Mi Hwang Ji Maclejczyk Pi Chen L. (2005). PM source apportionment for short-term cardiac
function changes in ApoE-/- mice. Environ Health Perspect, 113: 1575-1579. 087453
Lippmann Mi Ito Ki Hwang JSi Maciejczyk Pi Chen LC. (2006). Cardiovascular effects of nickel in
ambient air. Environ Health Perspect, 114: 1662-9. 091165
Liu Ji Ballaney Mi Al-Alem U> Quan Ci Jin X> Perera F> Chen LC; Miller RL. (2008). Combined Inhaled
Diesel Exhaust Particles and Allergen Exposure Alter Methylation of T Helper Genes and IgE
Production In Vivo. Toxicol Sci, 102: 76-81. 156709
Liu L-JS; Curjuric L Keidel D> Heldstab J; Kunzli N> Bayer-Oglesby L> Ackermann-Liebrich U>
Schindler Ci SAPALDIA team. (2007). Characterization of source-specific air pollution exposure
for a large population-based Swiss cohort (SAPALDIA). Environ Health Perspect, 115: 1638"
1645. 093093
Liu P-Li Chen Y"Li Chen Y"Hi Lin S-Ji Kou Y-R. (2005). Wood smoke extract induces oxidative stress-
mediated caspase-independent apoptosis in human lung endothelial cells: role of AIF and
EndoG. Am J Physiol, 289: L739-L749. 088304
Liu Xi Meng Z. (2005). Effects of airborne fine particulate matter on antioxidant capacity and lipid
peroxidation in multiple organs of rats. Inhal Toxicol, 17: 467-473. 088650
Liu Y-Qi Keane Mi Ensell Mi Miller Wi Kashon Mi Ong T-mi Mauderly Ji Lawson Di Gautam Mi
Zielinska Bi Whitney Ki Eberhardt Ji Wallace W. (2005). In vitro genotoxicity of exhaust
emissions of diesel and gasoline engine vehicles operated on a unified driving cycle. , !'¦ 60-66.
097019
Long JFi Waldman WJi Kristovich Ri Williams Mi Knight Di Dutta PK. (2005). Comparison of
ultrastructural cytotoxic effects of carbon and carbon/iron particulates on human monocyte-
derived macrophages. Environ Health Perspect, 113: 170-174. 087454
Lopes FDi Pinto TS; Arantes-Costa FM; Moriya HT; Biselli PJ; Ferraz LF; Lichtenfels AJ; Saldiva PH;
Mauad T> Martins MA. (2009). Exposure to ambient levels of particles emitted by traffic worsens
emphysema in mice. Environ Res, 109: 544-551. 190430
Lund AK; Knuckles TL; Obot Akata C> Shohet R; McDonald JD; Gigliotti A; Seagrave JC; Campen MJ.
(2007). Gasoline exhaust emissions induce vascular remodeling pathways involved in
atherosclerosis. Toxicol Sci, 95: 485-94. 125741
Lund AE Lucero J; Lucas S; Madden MC; McDonald JD; Seagrave JC; Knuckles TL; Campen MJ.
(2009). Vehicular emissions induce vascular MMP-9 expression and activity associated with
endothelin-l-mediated pathways. , 29: 511-7. 191159
Lundborg M; Bouhafs R; Gerde P; Ewing P; Camner P; Dahlen SE; Jarstrand C. (2007). Aggregates of
ultrafine particles modulate lipid peroxidation and bacterial killing by alveolar macrophages. ,
104: 250-57. 096040
Ma C; Wang J; Luo J. (2004). Activation of nuclear factor kappa B by diesel exhaust particles in mouse
epidermal cells through phosphatidylinositol 3-kinase/Akt signaling pathway. Biochem
Pharmacol, 67: 1975-1983. 088417
Maciejczyk P; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles
(CAPs) in mice: VIII source-related daily variations in in vitro responses to CAPs. Inhal Toxicol,
17: 243-253. 087456
Mangum JB; Bermudez E; Sar M; Everitt JI. (2004). Osteopontin expression in particle-induced lung
disease. , 30: 585-598. 097326
Martin S; Dawidowski L; Mandalunis P; Cereceda-Balic F; Tasat DR. (2007). Characterization and
biological effect of Buenos Aires urban air particles on mice lungs. , 105: 340-9. 096366
Mason CE. (2002). Gasoline ETBE vapor condensate rat micronucleus test (satellite procedure). 087645
Matsumoto A; Hiramatsu K; Li Y; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2006). Repeated
exposure to low-dose diesel exhaust after allergen challenge exaggerates asthmatic responses in
mice. Clin Immunol, 121: 227-235. 189213
July 2009
D-201
DRAFT-DO NOT CITE OR QUOTE

-------
Matsumoto Si Ishii Hi Tanabe Ti Kawai J. (2007). Chemical state analysis of fine particles using XAFS.
, 93: 132-137. 187020
Mauad Ti Rivero DHi de Oliveira RCi Lichtenfels AJi Guimaraes ETi de Andre PAi Kasahara Dli
Bueno HMi Saldiva PH. (2008). Chronic exposure to ambient levels of urban particles affects
mouse lung development. Am J Respir Crit Care Med, 178: 721-728. 156743
McDonald JDi Harrod KSi Seagrave Ji Seikop SKi Mauderly JL. (2004). Effects of low sulfur fuel
composition and a catalyzed particle trap on the composition and toxicity of diesel emissions.
Environ Health Perspect, 112: 1307-1312. 087459
McQueen DSi Donaldson Ki Bond SMi McNeilly JDi Newman Si Barton NJi Duffin R. (2007). Bilateral
vagotomy or atropine pre "treatment reduces experimental diesel-soot induced lung
inflammation. Toxicol Appl Pharmacol, 219: 62-71. 096266
Medeiros N Jri Rivero DHi Kasahara DIi Saiki Mi Godleski JJi Koutrakis Pi Capelozzi VLi Saldiva PHi
Antonangelo L. (2004). Acute pulmonary and hematological effects of two types of particle
surrogates are influenced by their elemental composition. , 95: 62-70. 096012
Mehta Mi Chen LCi Gordon Ti Rom Wi Tang MS. (2008). Particulate matter inhibits DNA repair and
enhances mutagenesis. , 657: 116-121. 190440
Mohallem SVi de Araujo Lobo DJi Pesquero CRi Assuncao JVi de Andre PAi Saldiva PHi Dolhnikoff M.
(2005). Decreased fertility in mice exposed to environmental air pollution in the city of Sao
Paulo. Environ Res, 98: 196-202. 088657
Moller Wi Hofer Ti Ziesenis Ai Karg Ei Heyder J. (2002). Ultrafine particles cause cytoskeletal
dysfunctions in macrophages. Toxicol Appl Pharmacol, 182: 197-207. 036589
Montiel-Davalos Ai Alfaro-Moreno Ei Lopez-Marure R. (2007). PM2.5 and PM10 induce the expression
of adhesion molecules and the adhesion of monocytic cells to human umbilical vein endothelial
cells. Inhal Toxicol, 19 Suppl l: 91-98. 156778
Mori Ti Watanuki Ti Kashiwagura T. (2007). Diesel exhaust particles disturb gene expression in mouse
testis. , 22: 58-63. 096564
Morishita Mi Keeler Gi Wagner Ji Marsik Fi Timm Ei Dvonch Ji Harkema J. (2004). Pulmonary
retention of particulate matter is associated with airway inflammation in allergic rats exposed
to air pollution in urban Detroit. Inhal Toxicol, 16: 663-674. 087979
Moyer CFi Kodavanti UPi Haseman JKi Costa DLi Nyska A. (2002). Systemic vascular disease in male
B6C3F1 mice exposed to particulate matter by inhalation: studies conducted by the national
toxicology program. Toxicol Pathol, 30: 427-434. 052222
Mutlu GMi Green Di Bellmeyer Ai Baker CMi Burgess Zi Rajamannan Ni Christman JWi Foiles Ni
Kamp DWi Ghio AJi Chandel NSi Dean DAi Sznajder Jli Budinger GR. (2007). Ambient
particulate matter accelerates coagulation via an IL"6"dependent pathway. , 117: 2952-61.
121441
Mutlu GMi Snyder Ci Bellmeyer Ai Wang Hi Hawkins Ki Soberanes Si Welch LCi Ghio AJi Chandel
NSi Kamp Di Sznajder Jacob Ii Budinger GRS. (2006). Airborne particulate matter inhibits
alveolar fluid reabsorption in mice via oxidant generation. , 34: 670-676. 155994
Nadziejko Ci Fang Ki Chen LCi Cohen Bi Karpatkin Mi Nadas A. (2002). Effect of concentrated
ambient particulate matter on blood coagulation parameters in rats. 050587
Nadziejko Ci Fang Ki Nadziejko Ei Narciso SPi Zhong Mi Chen LC. (2002). Immediate effects of
particulate air pollutants on heart rate and respiratory rate in hypertensive rats. , 2' 245-252.
087460
Nadziejko Ci Fang Ki Narciso Si Zhong Mi Su WCi Gordon Ti Nadas Ai Chen LC. (2004). Effect of
particulate and gaseous pollutants on spontaneous arrhythmias in aged rats. Inhal Toxicol, 16:
373-380. 055632
Nemmar Ai Al-Maskari Si Ali Badreldin Hi Al-Amri Issa S. (2007). Cardiovascular and lung
inflammatory effects induced by systemically administered diesel exhaust particles in rats. Am J
Physiol Lung Cell Mol Physiol, 292: L664-L670. 156800
Nemmar Ai Hoet PHi Dinsdale Di Vermylen Ji Hoylaerts MFi Nemery B. (2003). Diesel exhaust
particles in lung acutely enhance experimental peripheral thrombosis. , 107: 1202-8. 096567
Nemmar Ai Hoet PHi Vermylen Ji Nemery Bi Hoylaerts MF. (2004). Pharmacological stabilization of
mast cells abrogates late thrombotic events induced by diesel exhaust particles in hamsters. ,
110: 1670-1677. 087959
Nemmar Ai Hoylaerts MFi Hoet PHMi Vermylen Ji Nemery B. (2003). Size effect of intratr ache ally
instilled particles on pulmonary inflammation and vascular thrombosis. Toxicol Appl
Pharmacol, 186: 38-45. 087931
July 2009
D-202
DRAFT-DO NOT CITE OR QUOTE

-------
Nemmar A; Inuwa IM. (2008). Diesel exhaust particles in blood trigger systemic and pulmonary
morphological alterations. , 176- 20-30. 096566
Nemmar A; Nemery B; Hoet PHM; Vermylen J; Hoylaerts MF. (2003). Pulmonary inflammation and
thrombogenicity caused by diesel particles in hamsters' role of histamine. , 168- 1366-1372.
097487
Niwa Y> Hiura Y> Murayama T> Yokode Mi Iwai N. (2007). Nano-sized carbon black exposure
exacerbates atherosclerosis in LDL-receptor knockout mice. Circ J, 71- 1157-1161. 091309
Niwa Yi Hiura Y> Sawamura Hi Iwai N. (2008). Inhalation exposure to carbon black induces
inflammatory response in rats. Circ J, 72- 144-149. 156812
Nozaki JI; Yamamoto R; Ma Li Shima M. (2007). Trial to evaluate effects of ambient particulate matter
on health- A preliminary study using two-dimensional gel electrophoresis. , 12- 138-42. 097862
Nurkiewicz TR; Porter DW; Barger Mi Castranova V; Boegehold MA. (2004). Particulate matter
exposure impairs systemic microvascular endothelium-dependent dilation. Environ Health
Perspect, 112: 1299-1306. 087968
Nurkiewicz TR; Porter DW; Barger Mi Millecchia Li Rao KM; Marvar PJ; Hubbs AF; Castranova V;
Boegehold MA. (2006). Systemic microvascular dysfunction and inflammation after pulmonary
particulate matter exposure. Environ Health Perspect, 114: 412-419. 088611
Nurkiewicz TR; Porter DW; Hubbs AF; Cumpston JL; Chen BT; Frazer DG; Castranova V. (2008).
Nanoparticle inhalation augments particle-dependent systemic microvascular dysfunction. Part
Fibre Toxicol, 5: 1. 156816
Nurkiewicz TR; Porter DW; Hubbs AF; Stone S; Chen BT; Frazer DG; Boegehold MA; Castranova V.
(2009). Pulmonary nanoparticle exposure disrupts systemic microvascular nitric oxide signaling.
, 110: 191-203. 191961
Nygaard UC; Alberg T; Bleumink R; Aase A; Dybing E; Pieters R; Lovik M. (2005). Ambient air
particles from four European cities increase the primary cellular response to allergen in the
draining lymph node. Toxicology, 207: 241-254. 088655
Nygaard UC; Ormstad H; Aase A; Lovik M. (2005). The IgE adjuvant effect of particles:
characterisation of the primary cellular response in the draining lymph node. Toxicology, 206:
181-193. 087980
Nygaard UC; Samuelsen M; Aase A; Lovik M. (2004). The capacity of particles to increase allergic
sensitization is predicted by particle number and surface area, not by particle mass. Toxicol Sci,
82: 515-524. 058558
Obot CJ; Morandi MT; Beebe TP; Hamilton RF; Holian A. (2002). Surface components of airborne
particulate matter induce macrophage apoptosis through scavenger receptors. Toxicol Appl
Pharmacol, 184: 98-106. 042370
Obot CJ; Morandi MT; Hamilton RF; Holian A. (2004). A comparison of murine and human alveolar
macrophage responses to urban particulate matter. , 16: 69"76. 095938
Oh S-M; Chung K"H. (2006). Identification of mammalian cell genotoxins in respirable diesel exhaust
particles by bioassay-directed chemical analysis. Toxicol Lett, 161: 226-235. 088296
Okayama Y; Kuwahara M; Suzuki AK; Tsubone H. (2006). Role of reactive oxygen species on diesel
exhaust particle-induced cytotoxicity in rat cardiac myocytes. J Toxicol Environ Health A, 69:
1699-1710. 156824
Okeson CD; Riley MR; Fernandez A; Wendt JOL. (2003). Impact of the composition of combustion
generated fine particles on epithelial cell toxicity: influences of metals on metabolism.
Chemosphere, 51: 1121-1128. 042292
Okeson CD; Riley MR; Riley-Saxton E. (2004). In vitro alveolar cytotoxicity of soluble components of
airborne particulate matter: effects of serum on toxicity of transition metals. Toxicol In Vitro, 18:
673-680. 087961
Ono N; Oshio S; Niwata Y; Yoshida S; Tsukue N; Sugawara I; Takano H; Takeda K. (2007). Prenatal
exposure to diesel exhaust impairs mouse spermatogenesis. Inhal Toxicol, 19: 275-281. 156007
Osornio-Vargas AR; Bonner JC; Alfaro-Moreno E; Martinez L; Garcia-Cuellar CJ Rosales SP; Miranda
J; Rosas I. (2003). Proinflammatory and cytotoxic effects of Mexico City air pollution particulate
matter in vitro are dependent on particle size and composition. Environ Health Perspect, 111:
1289-1293. 052417
Pastorkova A; Cerna M; Smid J; Vrbikova V. (2004). Mutagenicity of airborne particulate matter PM10.
Cent Eur J Public Health, 12: S72-S75. 087431
July 2009
D-203
DRAFT-DO NOT CITE OR QUOTE

-------
Penn A; Murphy Gi Barker Si Henk Wi Penn L. (2005). Combustion-derived ultrafine particles
transport organic toxicants to target respiratory cells. Environ Health Perspect, 113- 956-963.
088257
Pereira CELi Heck TGi Saldiva PHNi Rhoden CR. (2007). Ambient particulate air pollution from
vehicles promotes lipid peroxidation and inflammatory responses in rat lung. , 40- 1353-1359.
156019
Pinkerton KEi Zhou Yi Teague SVi Peake JLi Walther RCi Kennedy IMi Leppert VJi Aust AE. (2004).
Reduced lung cell proliferation following short-term exposure to ultrafine soot and iron particles
in neonatal rats' key to impaired lung growth?. Inhal Toxicol, 1- 73-81. 087465
Pires-Neto RCi Lichtenfels AJi Soares SRi Macchione Mi Saldiva PHNi Dolhnikoff M. (2006). Effects of
Sao Paulo air pollution on the upper airways of mice. , 101- 356-361. 096734
Poma Ai Limongi Ti Pisani Ci Granato Vi Picozzi P. (2006). Genotoxicity induced by fine urban air
particulate matter in the macrophages cell line RAW 264.7. Toxicol In Vitro, 20: 1023-1029.
096903
Pourazar Ji Mudway ISi Samet JMi Helleday Ri Blomberg Ai Wilson SJi Frew AJi Kelly FJi Sandstrom
T. (2005). Diesel exhaust activates redox-sensitive transcription factors and kinases in human
airways. Am J Physiol, 289: L724-L730. 088305
Pozzi Ri De Berardis Bi Paoletti Li Guastadisegni C. (2005). Winter urban air particles from Rome
(Italy): effects on the monocytic-macrophagic RAW 2647 cell line. Environ Res, 99: 344-354.
088610
Pradhan A; Waseem Mi Dogra Si Khanna AKi Kaw JL. (2005). Alterations in bronchoalveolar lavage
constituents, oxidant/antioxidant status, and lung histology following intratracheal instillation
of respirable suspended particulate matter. J Environ Pathol Toxicol Oncol, 24: 19-32. 096128
Proctor SDi Dreher KLi Kelly SEi Russell JC. (2006). Hypersensitivity of prediabetic JCR: LA_cp rats to
fine airborne combustion particle-induced direct and noradrenergic-mediated vascular
contraction. Toxicol Sci, 90: 385-391. 088480
Prophete Ci Maciejczyk Pi Salnikow Ki Gould Ti Larson Ti Koenig Ji Jaques Pi Sioutas Ci Lippmann Mi
Cohen M. (2006). Effects of select PM-associated metals on alveolar macrophage phosphorylated
ERK1 and -2 and iNOS expression during ongoing alteration in iron homeostasis. J Toxicol
Environ Health A, 69: 935-951. 156888
Radomski Ai Jurasz Pi Alonso-Escalano Di Drews Mi Morandi Mi Malinski Ti Radomski MW. (2005).
Nanoparticle-induced platelet aggregation and vascular thrombosis. , 146: 882-893. 091377
Ramage Li Guy K. (2004). Expression of C-reactive protein and heat-shock protein-70 in the lung
epithelial cell line A549, in response to PM10 exposure. Inhal Toxicol, 16: 447-452. 055640
Ramos Ci Cisneros Ji Gonzalez-Avila Gi Becerril Ci Ruiz Vi Montano M. (2009). Increase of matrix
metalloproteinases in woodsmoke-induced lung emphysema in guinea pigs. Inhal Toxicol, 21:
119-132. 190116
Rao KM; Ma JYi Meighan Ti Barger MWi Pack Di Vallyathan V. (2005). Time course of gene expression
of inflammatory mediators in rat lung after diesel exhaust particle exposure. Environ Health
Perspect, 113: 612-617. 095756
Reed MDi Barrett EGi Campen MJi Divine KKi Gigliotti APi McDonald JDi Seagrave JCi Mauderly JLi
Seilkop SKi Swenberg JA. (2008). Health effects of subchronic inhalation exposure to gasoline
engine exhaust. Inhal Toxicol, 20: 1125-1143. 156903
Reed MDi Campen MJi Gigliotti APi Harrod KSi McDonald JDi Seagrave JCi Mauderly JLi Seilkop SK.
(2006). Health effects of subchronic exposure to environmental levels of hardwood smoke. Inhal
Toxicol, 18: 523-539. 156043
Reed MDi Gigliotti APi McDonald JDi Seagrave JCi Seilkop SKi Mauderly JL. (2004). Health effects of
subchronic exposure to environmental levels of diesel exhaust. Inhal Toxicol, 16: 177-193.
055625
Reibman Ji Hsu Yi Chen LCi Bleck Bi Gordon T. (2003). Airway epithelial cells release MIP"
3alpha/CCL20 in response to cytokines and ambient particulate matter. Am J Respir Cell Mol
Biol, 28: 648-654. 156905
Rengasamy Ai Barger MWi Kane Ei Ma JKHi Castranova Vi Ma JYC. (2003). Diesel exhaust particle-
induced alterations of pulmonary phase I and phase II enzymes of rats. J Toxicol Environ Health
A, 66: 153-167. 156907
Renwick LCi Brown Di Clouter Ai Donaldson K. (2004). Increased inflammation and altered
macrophage chemotactic responses caused by two ultrafine particle types. Occup Environ Med,
61: 442-447. 056067
July 2009
D-204
DRAFT-DO NOT CITE OR QUOTE

-------
Rhoden CR; Ghelfi E> Gonzalez-Flecha B. (2008). Pulmonary inflammation by ambient air particles is
mediated by superoxide anion. Inhal Toxicol, 20' 11-15. 190475
Rhoden CR; Lawrence J; Godleski JJ; Gonzalez-Flecha B. (2004). N-acetylcysteine prevents lung
inflammation after short-term inhalation exposure to concentrated ambient particles. Toxicol
Sci, 79: 296-303. 087969
Rhoden CR; Wellenius GA; Ghelfi E; Lawrence J; Gonzalez-Flecha B. (2005). PM"induced cardiac
oxidative stress and dysfunction are mediated by autonomic stimulation. Biochim Biophys Acta,
1725: 305-313. 087878
Riley MR; Boesewetter DE; Kim AM; Sirvent FP. (2003). Effects of metals Cu, Fe, Ni, V, and Zn on rat
lung epithelial cells. Toxicology, 190: 171-184. 053237
Riley MR; Boesewetter DE; Turner RA; Kim AM; Collier JM; Hamilton A. (2005). Comparison of the
sensitivity of three lung derived cell lines to metals from combustion derived particulate matter.
Toxicol In Vitro, 19: 411-419. 096452
Rivedal E; Myhre O; Sanner T; Eide I. (2003). Supplemental role of the Ames mutation assay and gap
junction intercellular communication in studies of possible carcinogenic compounds from diesel
exhaust particles. , 77: 533-542. 097684
Rivero DH; Soares SR; Lorenzi-Filho G; Saiki M; Godleski JJ; Antonangelo L; Dolhnikoff M; Saldiva
PH. (2005). Acute cardiopulmonary alterations induced by fine particulate matter of Sao Paulo,
Brazil. Toxicol Sci, 85: 898-905. 088653
Roberts JR; Young S"H; Castranova V; Antonini JM. (2007). Soluble metals in residual oil fly ash alter
innate and adaptive pulmonary immune responses to bacterial infection in rats. Toxicol Appl
Pharmacol, 221: 306-319. 097623
Rodriguez Ferreira Rivero DH; Sassaki C; Lorenzi-Filho G; Nascimento Saldiva PH. (2005). PM2.5
induces acute electrocardiographic alterations in healthy rats. Environ Res, 99: 262-266. 088659
Rosas Perez I; Serrano J; Alfaro-Moreno E; Baumgardner D; Garcia-Cuellar C; Martin Del Campo JM;
Raga GB; Castillejos M; Colin RD; Osornio Vargas AR. (2007). Relations between PM10
composition and cell toxicity: a multivariate and graphical approach. Chemosphere, 67: 1218-28.
097967
Roubicek DA; Gutierrez-Castillo ME; Sordo M; Cebrian-Garcia ME; Ostrosky-Wegman P. (2007).
Micronuclei induced by airborne particulate matter from Mexico City. , 631: 9-15. 156929
Saber AT; Bornholdt J; Dybdahl M; Sharma AK; Loft S; Vogel U; Wallin H. (2005). Tumor necrosis
factor is not required for particle-induced genotoxicity and pulmonary inflammation. , 79: 177-
182. 097865
Sakamoto N> Hayashi S; Gosselink J; Ishii H; Ishimatsu Y; Mukae H; Hogg JC; van Eeden SF. (2007).
Calcium dependent and independent cytokine synthesis by air pollution particle-exposed human
bronchial epithelial cells. Toxicol Appl Pharmacol, 225: 134-141. 096282
Salnikow K; Li X; Lippmann M. (2004). Effect of nickel and iron co-exposure on human lung cells.
Toxicol Appl Pharmacol, 196: 258-265. 087469
Salonen R; Halinen A; Pennanen A; Hirvonen M; Sillanpaa M; Hillamo R; Shi R; Borm P; Sandell E;
Koskentalo T; Aarnio P. (2004). Chemical and in vitro toxicologic characterization of wintertime
and springtime urban-air particles with an aerodynamic diameter below 10 jim in Helsinki. , 30:
80-90. 187053
Samet JM; Dewar BJ; Wu W', Graves LM. (2003). Mechanisms of Zn2+"induced signal initiation
through the epidermal growth factor receptor. Toxicol Appl Pharmacol, 191: 86"93. 113782
Santini MT; Rainaldi G; Ferrante A; Romano R; Clemente S; Motta A; De Berardis B; Balduzzi M;
Paoletti L; Indovina PL. (2004). Environmental fine particulate matter (PM25) activates the
RAW 2647 macrophage cell line even at very low concentrations as revealed by 1H NMR. Chem
Res Toxicol, 17: 63-74. 087879
Saxena QB; Saxena RK; Siegel PD; Lewis DM. (2003). Identification of organic fractions of diesel
exhaust particulate (DEP) which inhibit nitric oxide (NO) production from a murine macrophage
cell line. , 143: 317-322. 096986
Saxena RK; Saxena QB; Weissman DN; Simpson JP; Bledsoe TA; Lewis DM. (2003). Effect of diesel
exhaust particulate on bacillus Calmette-Guerin lung infection in mice and attendant changes in
lung interstitial lymphoid subpopulations and IFN gamma response. Toxicol Sci, 73: 66-71.
054395
Schins RPF; Lightbody JH; Borm PJA; Shi T; Donaldson K; Stone V. (2004). Inflammatory effects of
coarse and fine particulate matter in relation to chemical and biological constituents. Toxicol
Appl Pharmacol, 195: 1-11. 054173
July 2009
D-205
DRAFT-DO NOT CITE OR QUOTE

-------
Schneider J; Hock Ni Weimer S> Borrmann Si Kirchner Ui Vogt Ri Scheer V. (2005). Nucleation
particles in diesel exhaust- composition inferred from in situ mas spectrometric analysis.
Environ Sci Technol, 39- 6153-6161. 088368
Schober Wi Belloni Bi Lubitz Si Eberlein-Konig Bi Bohn Pi Saritas Yi Lintelmann Ji Matuschek Gi
Behrendt Hi Buters J. (2006). Organic extracts of urban aerosol (< or =PM25) enhance rBet v 1-
induced upregulation of CD63 in basophils from birch pollen-allergic individuals. , 90- 377-384.
097321
Seagrave Ji Campen Mi McDonald Ji Mauderly Ji Rohr A. (2008). Oxidative stress, inflammation, and
pulmonary function assessment in rats exposed to laboratory-generated pollutant mixtures. , 71-
1352.191990
Seagrave Ji Dunaway Si McDonald JDi Mauderly JLi Hayden Pi Stidley C. (2007). Responses of
differentiated primary human lung epithelial cells to exposure to diesel exhaust at an air-liquid
interface. , 33: 27-51. 097549
Seagrave Ji Knall Ci McDonald JDi Mauderly JL. (2004). Diesel particulate material binds and
concentrates a proinflammatory cytokine that causes neutrophil migration. Inhal Toxicol, l: 93"
98. 087470
Seagrave Ji Mauderly JLi Seilkop SK. (2003). In vitro relative toxicity screening of combined
particulate and semivolatile organic fractions of gasoline and diesel engine emissions. J Toxicol
Environ Health A, 66: 1113-1132. 054979
Seagrave Ji McDonald JDi Reed MDi Seilkop SKi Mauderly JL. (2005). Responses to subchronic
inhalation of low concentrations of diesel exhaust and hardwood smoke measured in rat
bronchoalveolar lavage fluid. Inhal Toxicol, 17: 657-670. 088000
Seagrave JCi McDonald JDi Bedrick Ei Edgerton ESi Gigliotti APi Jansen JJi Ke Li Naeher LPi Seilkop
SKi Zheng Mi Mauderley JL. (2006). Lung toxicity of ambient particulate matter from
southeastern US sites with different contributing sources: relationships between composition
and effects. Environ Health Perspect, 114: 1387-93. 091291
Sevastyanova Oi Binkova Bi Topinka Ji Sram RJi Kalina Ii Popov Ti Novakova Zi Farmer PB. (2007).
In vitro genotoxicity of PAH mixtures and organic extract from urban air particles part II:
human cell lines. , 620: 123-134. 156969
Sharma AKi Jensen KAi Rank Ji White PAi Lundstedt Si Gagne Ri Jacobsen NRi Kristiansen Ji Vogel
Ui Wallin H. (2007). Genotoxicity, inflammation and physico-chemical properties of fine particle
samples from an incineration energy plant and urban air. , 633: 95-111. 156975
Shi Ti Knaapen AMi Begerow Ji Birmili Wi Borm PJi Schins RP. (2003). Temporal variation of hydroxyl
radical generation and 8"hydroxy-2'-deoxyguanosine formation by coarse and fine particulate
matter. Occup Environ Med, 60: 315-321. 088248
Shwe T-T-Wi Yamamoto Si Kakeyama Mi Kobayashi Ti Fujimaki H. (2005). Effect of intratracheal
instillation of ultrafine carbon black on proinflammatory cytokine and chemokine release and
mRNA expression in lung and lymph nodes of mice. Toxicol Appl Pharmacol, 209: 51-61. 111553
Sigaud Si Goldsmith Carroll-Ann Wi Zhou Hi Yang Zi Fedulov Ai Imrich Ai Kobzik L. (2007). Air
pollution particles diminish bacterial clearance in the primed lungs of mice. Toxicol Appl
Pharmacol, 223: 1-9. 096100
Silva PJi Erupe ME; Price D; Elias J; Malloy QG; Li Q; Warren B; Cocker DR, 3rd. (2008).
Trimethylamine as precursor to secondary organic aerosol formation via nitrate radical reaction
in the atmosphere. Environ Sci Technol, 42: 4689-4696. 156981
Simkhovich BZ; Marjoram P; Kleinman MT; Kloner RA. (2007). Direct and acute cardiotoxicity of
ultrafine particles in young adult and old rat hearts. Basic Res Cardiol, 102: 467-475. 096594
Singh P; DeMarini DM; Dick CA; Tabor DG; Ryan JV; Linak WP; Kobayashi T; Gilmour MI. (2004).
Sample characterization of automobile and forklift diesel exhaust particles and comparative
pulmonary toxicity in mice. Environ Health Perspect, 112: 820-825. 087472
Sirivelu MP; MohanKumar SMJ; Wagner JG; Harkema JR; MohanKumar PS. (2006). Activation of the
stress axis and neurochemical alterations in specific brain areas by concentrated ambient
particle exposure with concomitant allergic airway disease. Environ Health Perspect, 114: 870"
874. 111151
Skarek M; Janosek J; Cupr P; Kohoutek J; Novotna-Rychetska A; Holoubek I. (2007). Evaluation of
genotoxic and non-genotoxic effects of organic air pollution using in vitro bioassays. Environ Int,
33: 859-66. 096814
July 2009
D-206
DRAFT-DO NOT CITE OR QUOTE

-------
Smith KR; Kim Si Recendez JJ; Teague SV; Menache MG; Grubbs DE; Sioutas C> Pinkerton KE.
(2003). Airborne particles of the California Central Valley alter the lungs of healthy adult rats.
Environ Health Perspect, 111: 902-908. 042107
Smith KR; Veranth JM; Kodavanti UP; Aust AE; Pinkerton KE. (2006). Acute pulmonary and systemic
effects of inhaled coal fly ash in rats: comparison to ambient environmental particles. , 93: 390-9.
110864
Somers CM; McCarry BE; Malek F; Quinn JS. (2004). Reduction of particulate air pollution lowers the
risk of heritable mutations in mice. , 304: 1008-1010. 078098
Somers CM; Yauk CL; White PA; Parfett CLJ; Quinn JS. (2002). Air pollution induces heritable DNA
mutations. Proc Natl Acad Sci USA, 99: 15904-15907. 078100
Song CL; Zhou YC; Huang RJ; Wang YQ; Huang QF; Lu G; Liu KM. (2007). Influence of ethanol-diesel
blended fuels on diesel exhaust emissions and mutagenic and genotoxic activities of particulate
extracts. , 149: 355-363. 155306
Song H-M; Jang A"S; Ahn M-H; Takizawa H; Lee S-H; Kwon J-H; Lee Y-M; Rhim T; Park C-S. (2008).
Yml and Ym2 expression in a mouse model exposed to diesel exhaust particles. , 23: 110-116.
156093
Steerenberg P; Verlaan A; De Klerk A; Boere A; Loveren H; Cassee F. (2004). Sensitivity to ozone,
diesel exhaust particles, and standardized ambient particulate matter in rats with a listeria
monocytogenes-induced respiratory infection. Inhal Toxicol, 16: 311-317. 087474
Steerenberg P; Withagen C; Dalen W; Dormans J; Loveren H. (2004). Adjuvant Activity of Ambient
Particulate Matter in Macrophage Activity-Suppressed, N"Acetylcysteine-Treated, iNOS" and
IL-4"Deficient Mice. Inhal Toxicol, 16: 835-843. 087981
Steerenberg PA; Van Amelsvoort L; Lovik M; Hetland RB; Alberg T; Halatek T; Bloemen HJT;
Rydzynski K; Swaen G; Schwarze P; Dybing E; Cassee FR. (2006). Relation between sources of
particulate air pollution and biological effect parameters in samples from four European cities:
an exploratory study. Inhal Toxicol, 18: 333-346. 088249
Steerenberg PA; Withagen CE; van Dalen WJ; Dormans JA; Heisterkamp SH; van Loveren H; Cassee
FR. (2005). Dose dependency of adjuvant activity of particulate matter from five European sites
in three seasons in an ovalbumin-mouse model. Inhal Toxicol, 17: 133-145. 088649
Stevens JP; Zahardis J; MacPherson M; Mossman BT; Petrucci GA. (2008). A new method for
quantifiable and controlled dosage of particulate matter for in vitro studies: the Electrostatic
Particulate Dosage and Exposure System (EPDExS). Toxicol In Vitro, 22(7): 1768-1774. 155363
Stinn W; Teredesai A; Anskeit E; Rustemeier K; Schepers G; Schnell P; Haussmann H-J; Carchman
RA; Coggins CR; Reininghaus W. (2005). Chronic nose-only inhalation study in rats, comparing
room-aged sidestream cigarette smoke and diesel engine exhaust. Inhal Toxicol, 17: 549-576.
088307
Sugamata M; Ihara T; Sugamata M; Takeda K. (2006). Maternal Exposure to Diesel Exhaust Leads to
Pathological Similarity to Autism in Newborns. Eisei Kagaku, 52: 486-488. 157025
Sun Q; Wang A; Jin X; Natanzon A; Duquaine D; Brook RD; Aguinaldo J-GS; Fayad ZA; Fuster V;
Lippmann M; Chen Lung C; Rajagopalan S. (2005). Long-term air pollution exposure and
acceleration of atherosclerosis and vascular inflammation in an animal model. , 294: 3003-3010.
186814
Sun Q; Yue P; Deiuliis JA; Lumeng CN; Kampfrath T; Mikolaj MB; Cai Y; Ostrowski MC; Lu B;
Parthasarathy S; Brook RD; Moffatt-Bruce SD; Chen LC; Rajagopalan S. (2009). Ambient air
pollution exaggerates adipose inflammation and insulin resistance in a mouse model of diet-
induced obesity. Circulation, 119: 538-546. 190487
Sun Q; Yue P; Kirk RI; Wang A; Moatti D; Jin X; Lu B; Schecter AD; Lippmann M; Gordon T; Chen LC;
Rajagopalan S. (2008). Ambient air particulate matter exposure and tissue factor expression in
atherosclerosis. Inhal Toxicol, 20: 127-137. 157033
Sun Q; Yue P; Ying Z; Cardounel AJ; Brook RD; Devlin R; Hwang JS; Zweier JL; Chen LC;
Rajagopalan S. (2008). Air Pollution Exposure Potentiates Hypertension Through Reactive
Oxygen Species-Mediated Activation of Rho/ROCK. Arterioscler Thromb Vase Biol, 28: 1760"
1766.157032
Sureshkumar V; Paul B; Uthirappan M; Pandey R; Sahu AP; Lai K; Prasad AK; Srivastava S; Saxena
A; Mathur N; Gupta YK. (2005). Proinflammatory and anti-inflammatory cytokine balance in
gasoline exhaust induced pulmonary injury in mice. Inhal Toxicol, 17: 161-168. 088306
July 2009
D-207
DRAFT-DO NOT CITE OR QUOTE

-------
Takizawa Hi Abe Si Okazaki Hi Kohyama Ti Sugawara Ii Saito Yi Ohtoshi Ti Kawasaki Si Desaki Mi
Nakahara Ki Yamamoto Ki Matsushima Ki Tanaka Mi Sagai Mi Kudoh S. (2003). Diesel exhaust
particles upregulate eotaxin gene expression in human bronchial epithelial cells via nuclear
factor-kappa B-dependent pathway. Am J Physiol Lung Cell Mol Physiol, 284- L1055-1062.
157039
Tal TLi Graves LMi Silbajoris Ri Bromberg PAi Wu Wi Samet JM. (2006). Inhibition of protein tyrosine
phosphatase activity mediates epidermal growth factor receptor signaling in human airway
epithelial cells exposed to Zn2+. Toxicol Appl Pharmacol, 214- 16-23. 108588
Tamagawa Ei Bai Ni Morimoto Ki Gray Ci Mui Ti Yatera Ki Zhang Xi Xing Li Li Yi Laher I. (2008).
Particulate matter exposure induces persistent lung inflammation and endothelial dysfunction. ,
295: L79. 191988
Tamaoki Ji Isono Ki Takeyama Ki Tagaya Ei Nakata Ji Nagai A. (2004). Ultrafine carbon black
particles stimulate proliferation of human airway epithelium via EGF receptor-mediated
signaling pathway. Am J Physiol Lung Cell Mol Physiol, 287: L1127-1133. 157040
Tankersley CGi Bierman Ai Rabold R. (2007). Variation in heart rate regulation and the effects of
particle exposure in inbred mice. , 19: 621-9. 097910
Tankersley CGi Campen Mi Bierman Ai Flanders SEi Broman KWi Rabold R. (2004). Particle effects on
heart-rate regulation in senescent mice. Inhal Toxicol, 16: 381-390. 094378
Tankersley CGi Champion HCi Takimoto Ei Gabrielson Ki Bedja Di Misra Vi El-Haddad Hi Rabold Ri
Mitzner W. (2008). Exposure to inhaled particulate matter impairs cardiac function in senescent
mice. Am J Physiol Regul Integr Comp Physiol, 295: R252-263. 157043
Tao Fi Kobzik L. (2002). Lung macrophage-epithelial cell interactions amplify particle-mediated
cytokine release. Am J Respir Cell Mol Biol, 26: 499-505. 157044
Tesfaigzi Yi McDonald JDi Reed MDi Singh SPi De Sanctis GTi Eynott PR; Hahn FF; Campen MJi
Mauderly JL. (2005). Low-level subchronic exposure to wood smoke exacerbates inflammatory
responses in allergic rats. Toxicol Sci, 88: 505-513. 156116
Tesfaigzi Y; Singh SP; Foster JE; Kubatko J; Barr EB; Fine PM; McDonald JD; Hahn FF; Mauderly JL.
(2002). Health effects of subchronic exposure to low levels of wood smoke in rats. Toxicol Sci, 65:
115-125. 025575
Thomson E; Kumarathasan P; Goegan P; Aubin RA; Vincent R. (2005). Differential regulation of the
lung endothelin system by urban particulate matter and ozone. Toxicol Sci, 88: 103-113. 087554
Thomson E; Kumarathasan PJ Vincent R. (2006). Pulmonary expression of preproET-1 and preproET"3
mRNAs is altered reciprocally in rats after inhalation of air pollutants. Exp Biol Med
(Maywood), 231: 979-984. 097483
Tin-Tin-Win-Shwe; Yamamoto S; Ahmed S; Kakeyama M; Kobayashi T; Fujimaki H. (2006). Brain
cytokine and chemokine mRNA expression in mice induced by intranasal instillation with
ultrafine carbon black. Toxicol Lett, 63: 153-160. 088415
Tomita Si Maekawa S"Ii Rahman Mi Saito Fi Kizu Ri Tohi Ki Ueno Ti Nakase Hi Gonzalez FJi
Hayakawa Ki Korenaga Ti Takahama Y. (2006). Thymic involution produced by diesel exhaust
particles and their constituents in mice. , 88: 113-124. 097827
Tong Yi Zhang Gi Li Yi Tan Mi Wang Wi Chen Ji Hwu Yi Hsu P-Ci Je JHi Margaritondo Gi Song Wi
Jiang Ri Jiang Z. (2006). Synchrotron microradiography study on acute lung injury of mouse
caused by PM(25) aerosols. , 58: 266-272. 097699
Totlandsdal AIi Refsnes Mi Skomedal Ti Osnes JBi Schwarze PEi Lag M. (2008). Particle-induced
cytokine responses in cardiac cell cultures-"the effect of particles versus soluble mediators
released by particle-exposed lung cells. Toxicol Sci, 106: 233-241. 157056
Tozuka Yi Watanabe Ni Ohsawa Mi Toriba Ai Kizu Ri Hayakawa K. (2004). Transfer of polycyclic
aromatic hydrocarbons to fetuses and breast milk of rats exposed to diesel exhaust. Eisei
Kagaku, 50: 497-502. 090864
Tsukue Ni Tsubone Hi Suzuki AK. (2002). Diesel exhaust affects the abnormal delivery in pregnant
mice and the growth of their young. Inhal Toxicol, 14: 635-651. 030593
Tsukue Ni Yoshida Si Sugawara Ii Taked K. (2004). Effect of diesel exhaust on development of fetal
reproductive function in ICR female mice. Eisei Kagaku, 50: 174-80. 096643
Tzeng H"Pi Yang R-Si Ueng T"Hi Lin-Shiau S"Yi Liu S"H. (2003). Motorcycle exhaust particulates
enhance vasoconstriction in organ culture of rat aortas and involve reactive oxygen species. , 75:
66-73. 097247
July 2009
D-208
DRAFT-DO NOT CITE OR QUOTE

-------
Tzeng H-P; Yang Rong Si Ueng T"Hi Liu S_H. (2007). Upregulation of cyclooxygenase-2 by motorcycle
exhaust particulate "induced reactive oxygen species enhances rat vascular smooth muscle cell
proliferation. , 20: 1170-1176. 097883
U.S. EPA. (1976). Air quality data - 1970 annual statistics. 015607
Ueng T-Hi Hung OCi Kuo M-Li Chan P-Ki Hu S-Hi Yang P-Ci Chang Louis W. (2005). Induction of
fibroblast growth factor-9 and interleukin-lalpha gene expression by motorcycle exhaust
particulate extracts and benzo(a)pyrene in human lung adenocarcinoma cells. , 87: 483-496.
097054
Ueng T-Hi Wang H"Wi Huang Y-Pi Hung C-C. (2004). Antiestrogenic effects of motorcycle exhaust
particulate in MCF-7 human breast cancer cells and immature female rats. , 46: 454-462.
096199
Umbuzeiro GAi Franco Ai Martins MHi Kummrow Fi Carvalho Li Schmeiser HHi Leykauf Ji Stiborova
Mi Claxton LD. (2008). Mutagenicity and DNA adduct formation of PAH, nitro-PAH, and oxy-
PAH fractions of atmospheric particulate matter from Sao Paulo, Brazil. , 652: 72-80. 190491
Upadhyay Di Panduri Vi Ghio Ai Kamp DW. (2003). Particulate matter induces alveolar epithelial cell
DNA damage and apoptosis: role of free radicals and the mitochondria. , 29: 180-187. 097370
Upadhyay Si Stoeger Ti Harder Vi Thomas RFi Schladweiler MCi Semmler-Behnke Mi Takenaka Si
Karg Ei Reitmeir Pi Bader Mi Stampfl Ai Kodovanti Ui Schulz H. (2008). Exposure to ultrafine
carbon particles at levels below detectablepulmonary inflammation affects cardiovascular
performance inspontaneously hypertensive rats. Part Fibre Toxicol, 5- 19. 159345
Valavanidis Ai Vlahoyianni Ti Fiotakis K. (2005). Comparative study of the formation of oxidative
damage marker 8"hydroxy-2'-deoxyguanosine (8"OHdG) adduct from the nucleoside 2'-
deoxyguanosine by transition metals and suspensions of particulate matter in relation to metal
content and redox reactivity. , 39: 1071-1081. 096432
Valentine Ri Himmelstein MW. (2001). Overview of the acute, subchronic, reproductive, developmental
and genetic toxicology of "beta"-chloroprene. Chem Biol Interact, 135: 81-100. 019011
Veranth Ji Kaser Ei Veranth Mi Koch Mi Yost G. (2007). Cytokine responses of human lung cells
(BEAS-2B) treated with micron-sized and nanoparticles of metal oxides compared to soil dusts.
Part Fibre Toxicol, 4- 2. 090346
Veranth JMi Moss TAi Chow JCi Labban Ri Nichols WKi Walton JCi Walton JGi Yost GS. (2006).
Correlation of in vitro cytokine responses with the chemical composition of soil-derived
particulate matter. Environ Health Perspect, 114: 341-349. 087479
Veranth JMi Reilly CAi Veranth MMi Moss TAi Langelier CRi Lanza DLi Yost GS. (2004).
Inflammatory cytokines and cell death in BEAS-2B lung cells treated with soil dust,
lipopolysaccharide, and surface-modified particles. Toxicol Sci, 82: 88"96. 087480
Veras MMi Damaceno-Rodrigues NRi Caldini EGi Maciel Ribeiro AAi Mayhew TMi Saldiva PHi
Dolhnikoff M. (2008). Particulate urban air pollution affects the functional morphology of mouse
placenta. Biol Reprod, 79: 578-584. 190493
Veras MMi Damaceno-Rodrigues NRi Guimaraes Silva RMi Scoriza JNi Saldiva PHi Caldini EGi
Dolhnikoff M. (2009). Chronic exposure to fine particulate matter emitted by traffic affects
reproductive and fetal outcomes in mice. Environ Res, 109: 536-543. 190496
Veronesi Bi De Haar Ci Lee Li Oortgiesen M. (2002). The surface charge of visible particulate matter
predicts biological activation in human bronchial epithelial cells. Toxicol Appl Pharmacol, 178:
144-154. 024599
Veronesi Bi Makwana Oi Pooler Mi Chen LC. (2005). Effects of subchronic exposures to concentrated
ambient particles VII Degeneration of dopaminergic neurons in Apo E7- mice. Inhal Toxicol, 17:
235-241. 087481
Veronesi Bi Oortgiesen M. (2001). Neurogenic inflammation and particulate matter (PM) air pollutants.
Neurotoxicology, 22: 795-810. 015977
Verstraelen Si Van Den Heuvel Ri Nelissen Ii Witters Hi Verheyen Gi Schoeters G. (2005). Flow
cytometric characterisation of antigen presenting dendritic cells after in vitro exposure to diesel
exhaust particles. Toxicol In Vitro, 19: 903-907. 096872
Vogel CFAi Sciullo Ei Wong Pi Kuzmicky Pi Kado Ni Matsumura F. (2005). Induction of
proinflammatory cytokines and C-reactive protein in human macrophage cell line U937 exposed
to air pollution particulates. Environ Health Perspect, 113: 1536-1541. 087891
Walczak-Drzewiecka Ai Wyczolkowska Ji Dastych J. (2003). Environmentally relevant metal and
transition metal ions enhance Fc epsilon RI-mediated mast cell activation. , 111: 708-713.
096784
July 2009
D-209
DRAFT-DO NOT CITE OR QUOTE

-------
Wallenborn JG; Evansky P; Shannahan JH; Vallanat B; Ledbetter AD; Schladweiler MC; Richards JH;
Gottipolu RR; Nyska A; Kodavanti UP. (2008). Subchronic inhalation of zinc sulfate induces
cardiac changes in healthy rats. Toxicol Appl Pharmacol, 232- 69"77. 191171
Wallenborn JG; McGee John K; Schladweiler Mette C; Ledbetter Allen D; Kodavanti Urmila P. (2007).
Systemic translocation of particulate matter-associated metals following a single intratracheal
instillation in rats. J Toxicol Sci, 98- 231-239. 156144
Wan ECH; Yu JZ. (2006). Determination of sugar compounds in atmospheric aerosols by liquid
chromatography combined with positive electrospray ionization mass spectrometry. J
Chromatogr, 1107: 175-181. 157104
Wang Y-Z; Ingram JL; Walters DM; Rice AB; Santos JH; Van Houten B; Bonner JC. (2003). Vanadium-
induced STAT-1 activation in lung myofibroblasts requires H202 and P38 MAP kinase. Free
Radic Biol Med, 35: 845-855. 157106
Watanabe N. (2005). Decreased number of sperms and Sertoli cells in mature rats exposed to diesel
exhaust as fetuses. Toxicol Lett, 155: 51-58. 087985
Wegesser TC; Last JA. (2008). Lung response to coarse PM: bioassay in mice. Toxicol Appl Pharmacol,
230: 159-166. 190506
Wellenius GA; Batalha JRF; Diaz EA; Lawrence J; Coull BA; Katz T; Verrier RL; Godleski JJ. (2004).
Cardiac effects of carbon monoxide and ambient particles in a rat model of myocardial infarction.
Toxicol Sci, 80: 367-376. 087874
Wellenius GA; Coull BA; Batalha JRF; Diaz EA; Lawrence J; Godleski JJ. (2006). Effects of ambient
particles and carbon monoxide on supraventricular arrhythmias in a rat model of myocardial
infarction. Inhal Toxicol, 18: 1077-1082. 156152
Wellenius GA; Coull BA; Godleski JJ; Koutrakis P; Okabe K; Savage ST. (2003). Inhalation of
concentrated ambient air particles exacerbates myocardial ischemia in conscious dogs. Environ
Health Perspect, 111: 402-408. 055691
Whitekus MJ; Li N; Zhang M; Wang M; Horwitz MA; Nelson SK; Horwitz LD; Brechun N; Diaz-
Sanchez D; Nel AE. (2002). Thiol antioxidants inhibit the adjuvant effects of aerosolized diesel
exhaust particles in a murine model for ovalbumin sensitization. , 168: 2560-2567. 157142
Wichers LB; Nolan JP; Winsett DW; Ledbetter AD; Kodavanti UP; Schladweiler MCJ; Costa DL;
Watkinson WP. (2004). Effects of instilled combustion-derived particles in spontaneously
hypertensive rats Part II: pulmonary responses. Inhal Toxicol, 16: 407-419. 055636
Wichers LB; Rowan WH, 3rd; Nolan JP; Ledbetter AD; McGee JK; Costa DL; Watkinson WP. (2006).
Particle deposition in spontaneously hypertensive rats exposed via whole-body inhalation:
measured and estimated dose. , 93: 400-10. 103806
Wilson MR; Foucaud L; Barlow PG; Hutchison GR; Sales J; Simpson RJ; Stone V. (2007). Nanoparticle
interactions with zinc and iron: Implications for toxicology and inflammation. Toxicol Appl
Pharmacol, 225: 80-89. 097268
Win-Shwe TT; Yamamoto S; Fujitani Y; Hirano S; Fujimaki H. (2008). Spatial learning and memory
function-related gene expression in the hippocampus of mouse exposed to nanoparticle-rich
diesel exhaust. Neurotoxicology, 29: 940-947. 190146
Witten ML; Wong SS ; Sun NN; Keith I; Kweon C; Foster DE; Schauer JJ; Sherrill DL. (2005).
Neurogenic responses in rat lungs after nose-only exposure to diesel exhaust. 087485
Wold LE; Simkhovich BZ; Kleinman MT; Nordlie MA; Dow JS; Sioutas C; Kloner RA. (2006). In vivo
and in vitro models to test the hypothesis of particle "induced effects on cardiac function and
arrhythmias. Cardiovasc Toxicol, 6: 69"78. 097028
Wong SS; Sun NN; Keith I; Kweon C"B; Foster DE; Schauer James J; Witten ML. (2003). Tachykinin
substance P signaling involved in diesel exhaust-induced bronchopulmonary neurogenic
inflammation in rats. , 77: 638-650. 097707
Wottrich R; Diabate S; Krug HF. (2004). Biological effects of ultrafine model particles in human
macrophages and epithelial cells in mono- and co-culture. Int J Hyg Environ Health, 207: 353"
361. 094518
Wu W; Samet JM; Silbajoris R; Dailey LA; Sheppard D; Bromberg PA; Graves LM. (2004). Heparin-
binding epidermal growth factor cleavage mediates zinc-induced epidermal growth factor
receptor phosphorylation. Am J Respir Cell Mol Biol, 30: 540-7. 096949
Wu W; Silbajoris RA; Whang YE; Graves LM; Bromberg PA; Samet JM. (2005). p38 and EGF receptor
kinase-mediated activation of the phosphatidylinositol 3-kinase/Akt pathway is required for
Zn2+"induced cyclooxygenase-2 expression. Am J Physiol Lung Cell Mol Physiol, 289: L883-9.
097350
July 2009
D-210
DRAFT-DO NOT CITE OR QUOTE

-------
Wu YH; Vincent JH. (2007). A modified Marple-type cascade impactor for assessing aerosol particle size
distributions in workplaces. J Occup Environ Hyg, 4- 798-807. 098412
Xu D-Q; Zhang W-L. (2004). Monitoring of pollution of air fine particles (PM25) and study on their
genetic toxicity. ,17: 452-458. 097231
Yacobi NR; Phuleria HC; Demaio Li Liang CH; Peng CA; Sioutas C> Borok Z; Kim KJ; Crandall ED.
(2007). Nanoparticle effects on rat alveolar epithelial cell monolayer barrier properties. Toxicol
In Vitro, 21: 1373-1381. 156166
Yamamoto Si Tin Tin Win Si Ahmed Si Kobayashi Ti Fujimaki H. (2006). Effect of ultrafine carbon
black particles on lipoteichoic acid-induced early pulmonary inflammation in BALB/c mice.
Toxicol Appl Pharmacol, 213: 256-266. 096671
Yanagisawa Ri Takano Hi Inoue Ki Ichinose Ti Sadakane Ki Yoshino Si Yamaki Ki Kumagai Yi
Uchiyama Ki Yoshikawa Ti Morita M. (2003). Enhancement of acute lung injury related to
bacterial endotoxin by components of diesel exhaust particles. Thorax, 58: 605-612. 087487
Yanagisawa Ri Takano Hi Inoue Kli Ichinose Ti Sadakane Ki Yoshino Si Yamaki Ki Yoshikawa Ti
Hayakawa K. (2006). Components of diesel exhaust particles differentially affect Thl/Th2
response in a murine model of allergic airway inflammation. , 36: 386-395. 096458
Yang C-Yi Tseng Y"Ti Chang C-C. (2003). Effects of air pollution on birthweight among children born
between 1995 and 1997 in Kaohsiung, Taiwan. J Toxicol Environ Health A, 66: 807-816. 087886
Yatera Ki Hsieh Ji Hogg James Ci Tranfield Ei Suzuki Hi Shih C"Hi Behzad Ali Ri Vincent Ri van
Eeden Stephan F. (2008). Particulate matter air pollution exposure promotes recruitment of
monocytes into atherosclerotic plaques. Am J Physiol Heart Circ Physiol, 294: H944-H953.
157162
Yauk Ci Polyzos A> Rowan-Carroll Ai Somers CM; Godschalk RW! Van Schooten FJ; Berndt ML;
Pogribny IP; Koturbash I; Williams A; Douglas GR; Kovalchuk O. (2008). Germ-line mutations,
DNA damage, and global hypermethylation in mice exposed to particulate air pollution in an
urban/industrial location. , 105: 605-610. 157164
Yin H; Too HP; Chow GM. (2005). The effects of particle size and surface coating on the cytotoxicity of
nickel ferrite. Biomaterials, 26: 5818-5826. 088133
Yin X-J; Schafer R; Ma JYC; Antonini JM; Roberts JR; Weissman DN; Siegel PD; Ma JKH. (2003).
Alteration of pulmonary immunity to Listeria monocytogenes by diesel exhaust particles (DEPs)
II Effects of DEPs on T-cell-mediated immune responses in rats. , 111: 524-530. 096127
Yin XJ; Dong CC; Ma JYC; Antonini JM; Roberts JR; Stanley CF; Schafer R; Ma JKH. (2004).
Suppression of cell-mediated immune responses to listeria infection by repeated exposure to
diesel exhaust particles in brown Norway rats. , 77: 263-271. 097685
Yin XJ; Ma JYC; Antonini JM; Castranova V; Ma JKH. (2004). Roles of reactive oxygen species and
heme oxygenase-1 in modulation of alveolar macrophage-mediated pulmonary immune
responses to listeria monocytogenes by diesel exhaust particles. Toxicol Sci, 82: 143-153. 087983
Yokohira M; Takeuchi H; Yamakawa K; Saoo K; Matsuda Y; Zeng Y; Hosokawa K; Imaida K. (2007).
Bioassay by intratracheal instillation for detection of lung toxicity due to fine particles in F344
male rats. Exp Toxicol Pathol, 58: 211-21. 097976
Yokota S; Furuya M; Seki T; Marumo H; Ohara N; Kato A. (2004). Delayed exacerbation of acute
myocardial ischemia/reperfusion-induced arrhythmia by tracheal instillation of diesel exhaust
particles. , 16: 319-331. 096516
Yokota S; Mizuo K; Moriya N; Oshio S; Sugawara I; Takeda K. (2009). Effect of prenatal exposure to
diesel exhaust on dopaminergic system in mice. Neurosci Lett, 449: 38-41. 190518
Yokota S; Seki T; Furuya M; Ohara N. (2005). Acute functional enhancement of circulatory neutrophils
after intratracheal instillation with diesel exhaust particles in rats. , 17: 671-679. 096003
Yokota S; Seki T; Naito Y; Tachibana S; Hirabayashi N; Nakasaka T; Ohara N; Kobayashi H. (2008).
Tracheal instillation of diesel exhaust particles component causes blood and pulmonary
neutrophilia and enhances myocardial oxidative stress in mice. J Toxicol Sci, 33: 609-620.
190109
Yoshida S; Ono N> Tsukue N> Oshio S; Umeda T; Takano H; Takeda K. (2006). In utero exposure to
diesel exhaust increased accessory reproductive gland weight and serum testosterone
concentration in male mice. , 13: 139-147. 097015
Yoshida S; Takedab K. (2004). The effects of diesel exhaust on murine male reproductive function. Eisei
Kagaku, 50: 210-4. 097760
Yoshida S; Yoshida M; Sugawara I; Takeda K. (2006). Mice strain differences in effects of fetal exposure
to diesel exhaust gas on male gonadal differentiation. , 13: 117-123. 156170
July 2009
D-211
DRAFT-DO NOT CITE OR QUOTE

-------
Yun Y-P; Joo J-D; Lee J-Y; Nam H-Y; Kim Y-H; Lee K-H; Lim C-S; Kim H-J; Lim Y-G; Lim Y. (2005).
Induction of nuclear factor-"kappa"B activation through TAK1 and NIK by diesel exhaust
particles in L2 cell lines. Toxicol Lett, 155- 337-342. 088302
Zanchi AC; Venturini CD; Saiki M; Nascimento Saldiva PH; Tannhauser Barros HM! Rhoden CR.
(2008). Chronic nasal instillation of residual-oil fly ash (ROFA) induces brain lipid peroxidation
and behavioral changes in rats. Inhal Toxicol, 20: 795-800. 157173
Zelikoff JT; Chen LC; Cohen MD! Fang K; Gordon Ti Li Y; Nadziejko C> Schlesinger RB. (2003). Effects
of inhaled ambient particulate matter on pulmonary antimicrobial immune defense. Inhal
Toxicol, 15: 131-150. 039009
Zelikoff JT; Schermerhorn KR; Fang K; Cohen MD; Schlesinger RB. (2002). A role for associated
transition metals in the immunotoxicity of inhaled ambient particulate matter. Environ Health
Perspect, 5: 871-875. 037797
Zhang J; Ghio AJ; Chang W', Kamdar O; Rosen GD; Upadhyay D. (2007). Bim mediates mitochondria-
regulated particulate matter-induced apoptosis in alveolar epithelial cells. FEBS Lett, 581:
4148-4152. 156179
Zhang Q; Kleeberger SR; Reddy SP. (2004). DEP-induced fra-1 expression correlates with a distinct
activation of AP-1 "dependent gene transcription in the lung. Am J Physiol Lung Cell Mol
Physiol, 286: L427-436. 157183
Zhang Z; Che W; Liang Y; Wu M; Li N; Shu Y; Liu F; Wu D. (2007). Comparison of cytotoxicity and
genotoxicity induced by the extracts of methanol and gasoline engine exhausts. Toxicol In Vitro,
21: 1058-1065. 157186
Zhao H; Barger MW; Ma JKH; Castranova V; Ma JYC. (2006). Cooperation of the inducible nitric oxide
synthase and cytochrome P450 1A1 in mediating lung inflammation and mutagenicity induced
by diesel exhaust particles. Environ Health Perspect, 114: 1253-1258. 100996
Zhao L; Wang X; He Q; Wang H; Sheng G; Chan LY; Fu J; Blake DR. (2004). Exposure to hazardous
volatile organic compounds, PM sub(lO) and CO while walking along streets in urban
Guangzhou, China. Atmos Environ, 38: 6177-6184. 100972
Zhong W; Levin L; Reponen T; Hershey GK; Adhikari A; Shukla R; LeMasters G. (2006). Analysis of
short-term influences of ambient aeroallergens on pediatric asthma hospital visits. Sci Total
Environ, 370: 330-336. 093264
Zhou Y-M; Zhong C"Y; Kennedy IM; Pinkerton KE. (2003). Pulmonary responses of acute exposure to
ultrafine iron particles in healthy adult rats. Environ Toxicol, 18: 227-235. 087940
July 2009
D-212
DRAFT-DO NOT CITE OR QUOTE

-------
Annex E. Epidemiologic Studies
E.1. Short-Term Exposure and Cardiovascular Outcomes
E.I.1. Cardiovascular Morbidity Studies
Table E-1 Short-term exposure - cardiovascular morbidity outcomes - PM10
Study
Design & Methods
Concentrations
Effect Estimates (95% CI)
Reference: Baccarelli et al. (2007,
091310)
Outcome: Fasting and postmethionine-
load total homocysteine (tHcy)
Period of Study: Jan 1995 - Aug 2005 Age Groups: 11 -84 yrs
Location: Lombardia region, Italy	Study Design: Cross-sectional I Panel
N: 1,213 participants
Statistical Analyses: Generalized
additive models
Covariates: age, sex, BMI, smoking,
alcohol, hormone use, temperature, day
of the year, and long-term trends
Season: Adjusted for long-term trends to
account for season
Dose-response Investigated? No
Statistical Package: R v2.2.1
Lags Considered: 1 d, 7d moving avg.
Pollutant: PMio (some TSP measures
used to predict PMio)
Averaging Time: 24 h
Mean (SD): NR
Percentiles: 25th: 20.1
50th: 34.1
75th: 52.6
Max: 390.0
Monitoring Stations: 53
Copollutant: CO, NO2, SO2, O3
PM Increment: IQR
Percent Change: [Lower CI, Upper CI]:
Homocysteine, fasting: 0.4 (-2.4, 3.3)
Homocysteine, postmethionine-load: 1.1
(-1.5, 3.7)
Percent Change: per 25.7m3 increase
in 7-day moving avg of PMio
Homocysteine, fasting: 1.0 (-1.9, 3.9)
Homocysteine, postmethionine-load: 2.0
(-0.6, 4.7)
Percent Change: on fasting
homocysteine per IQR increase in 24-
h PMio levels
Among smokers: 6.2 (0.0,12.7)
Among non-smokers: -1.6 (-5.5, 2.5)
Percent Change: on postmethionine-
load homocysteine per IQR increase in
24-h PMio levels: Among smokers: 6.0
(0.5,11.8)
Among non-smokers: -0.1 (-3.6, 3.5)
Reference: Baccarelli et al. (2007,
090733) Period of Study: Jan 1995 -
Aug 2005
Location: Lombardia region Italy
Outcome: Prothrombin time (PT)
Activated partial thromboplastin time
(APTT)
Fibrinogen
Functional antithrombin
Functional protein C
Protein C, antigen
Functional protein S
Free protein S
Age Groups: 11 -84 yrs
Study Design: Cross-sectional I Panel
N: 1,218 participants
Statistical Analyses: Generalized
additive models
Covariates: Age, sex, BMI, smoking,
alcohol, hormone use, temperature, day
of the year, and long-term trends
Season: Adjusted for long-term trends to
account for season
Pollutant: PMio (some TSP measures
used to predict PMio)
Averaging Time: Hourly concentrations
used to calculate lags of same day, 7-
day, 30-day, and h 0-6
Mean (SD): NR
Percentiles:
Sep-Nov: 2
5th: 33.1
50th: 51.2
75th: 76.5
Max: 148.9
Dec-Feb:
25th: 47.9
50th: 68.5
75th: 95.3
Max: 238.3
Mar-May:
25th: 30.0
50th: 64.1
75th: 64.8
Max: 158.5
PM Increment: SD
Effect Estimate [Lower CI, Upper CI]:
Estimated changes in endpoint
PT (international normalized ratio):
At time of blood sample: -0.06 (-0.12,
0.00)
Avg levels 7 days prior: -0.03 (-0.10,
0.04)
Avg levels 30 days prior: -0.08 (-0.14, ¦
0.01)
(Hourly moving Avgs presented in Fig 2)
APTT (ratio to reference plasma):
At time of blood sample: 0.02 (-0.04,
0.08)
Avg levels 7 days prior: 0.00 (-0.07,
0.06)
Avg levels 30 days prior: 0.01 (-0.06,
0.08)
Fibrinogen:
At time of blood sample: 0.01 (-0.05,
0.07)
Avg levels 7 days prior: -0.03 (-0.09,
0.04)
July 2009
E-1
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Dose-response Investigated? No
Statistical Package: R software v2.2.1
Jun-Aug:
25th: 28.0
50th: 44.3
75th: 61.3
Max: 94.7
Monitoring Stations: 53 sites
Copollutant: CO, NO2, SO2, O3
Avg levels 30 days prior: -0.02 (-0.09,
0.05)
Functional antithrombin:
At time of blood sample: -0.02 (-0.09,
0.04)
Avg levels 7 days prior: -0.06 (-0.13,
0.01)
Avg levels 30 days prior: -0.06 (-0.13,
0.02)
Functional protein C:
At time of blood sample: 0.00 (-0.06, 6.1)
Avg levels 7 days prior: -0.06 (-0.12,
0.01)
Avg levels 30 days prior: -0.06 (-0.14,
0.01)
Protein C, antigen:
At time of blood sample: 0.00 (-0.06, 6.0)
Avg levels 7 days prior: -0.04 (-0.10,
0.03)
Avg levels 30 days prior: -0.06 (-0.14,
0.01)
Functional protein S:
At time of blood sample: 0.04 (-0.03,
0.10)
Avg levels 7 days prior: -0.03 (-0.11,
0.06)
Avg levels 30 days prior: -0.14 (-0.23, ¦
0.05)
Free protein S:
At time of blood sample: 0.05 (-0.01,
0.10)
Avg levels 7 days prior: 0.01 (-0.05,
0.07)
Avg levels 30 days prior: -0.01 (-0.08,
0.06)
July 2009
E-2
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Barclay et al.
(2009,179935)
Period of Study: Jan 2003 - May 2005
Location: Aberdeen, Scotland
Outcome: Haematological outcomes,
Heart Rhythm outcomes, & Heart Rate
Variability outcomes
Age Groups: 70.4 (8.9)
Study Design: panel
N: 132 patients wf chronic heart failure
Statistical Analyses: Linear & Mixed
Effects Regression Model
Covariates: age, temperature, humidity,
pressure
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: lags 0-2d
Pollutant: PMio
Averaging Time: daily
Mean (SD): 20.25
Min: 7.375
Max: 68.3
Monitoring Stations: 1
Copollutant: PM2.6, PNC, NO2
Co-pollutant Correlation
NO2 city: 0.294
NO city: 0.112
NO2 personal: 0.055
PNC DE0M: 0.241
PM2.6 total: 0.476*
PM2.6 traffic: 0.882*
PNC total: 0.125
PNC traffic: 0.190
"correlations based on 3-day average
concentrations
PM Increment: NR
Beta (Lower CI, Upper CI):
Haemoglobin: 0.136 (-0.274, 0.546)
Mean corpuscular haemoglobin: 0.030 (¦
0.232,0.291)
Platelets: 0.096 (-0.923,1.115)
Haematocrit: 0.131 (-0.289, 0.551)
White blood cells: 0.034 (-1.175,1.244)
C reactive protein: -4.872 (-12.094,
2.351)
IL-6: 2.207 (-4.995, 9.410)
von Willebrand factor: 0.660 (-2.651,
3.970)
E-selectin:-0.536 (-2.528,1.457)
Fibrinogen:-0.432 (-2.470,1.607)
Factor VII: 0.990 (-1.265, 3.245)
d-dimer: -1.225 (-4.505, 2.055)
All arrhythmias: -3.447 (-11.521,4.627)
Ventricular ectopic beats: -2.110 (-
12.135,7.915)
Ventricular couplets: -1.561 (-10.811,
7.689)
Ventricular runs: -0.709 (-6.677, 5.259)
Supraventricular ectopic beats: 0.033 (¦
9.242, 9.308)
Supraventricular couplets: 0.006 (-8.618,
8.629)
Supraventricular runs: 3.710 (-2.847,
10.266)
Average HR: 0.321 (-0.197,0.838)
24h SDNN: 1.040 (-0.415, 2.494)
24h SDANN: 1.195 (-0.473, 2.863)
24h RMSSD: 0.321 (-0.197,0.838)
24h PNN: 2.837 (-3.791, 9.465)
24h LF power: 0583 (-3.622,4.787)
24h LF normalized:-3.137 (-5.540, -
0.733)*
24h HF power: 0.872 (-4.649, 6.392)
24h HF normalized: -2.223 (-4.952,
0.505)
24h LF/HF ratio: -0.296 (-3.832, 3.240)
*p< 0.05
Notes: LF - low frequency
HF- high frequency
Reference: Briet et al. (2007, 093049)
Period of Study: NR
Location: Paris, France
Outcome: Endothelial Function
Age Groups: 20-40 yrs
Study Design: panel
N: 40 white male nonsmokers
Statistical Analyses: Multiple Robust
Regrssion
Covariates: R53R/R53H genotype, diet,
subject factor, visit, temperature
Dose-response Investigated? No
Statistical Package: NCSS
Lags Considered: 0-5d
Pollutant: PM10
Averaging Time: 24h
5 d Mean (SD): 43 (10)
Monitoring Stations: NR
Co-pollutant: PM2.6, SO2, NO, NO2, CO
Co-pollutant Correlation
n/a
PM Increment: 1 SD
Beta (Lower CI, Upper CI), P, R2:
Flow-mediated brachial artery dilation:
0.07 (-0.62, 0.76), NS, 0.03
Reactive hyperemia:
15.91 (7.74,24.0), <0.001,0.16
July 2009
E-3
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Choi et al (2007, 093196)
Period of Study: 2001-2003
Location: Incheon, South Korea
Outcome: Blood pressure
Study Design: Cross-sectional
N: 10459 subjects with a hospital health
examination
Statistical Analyses: Linear regression
Covariates: Season: Effect modification
by season
Pollutant: PMio
Averaging Time: Measured hourly and
calculated 24-h means
Percentiles: Warm season: Median: 36.7
Cold season: Median: 45.7
Monitoring Stations: 9 stations
Copollutant: NO2, SO2
PM Increment: 10/yg/m3
Effect Estimate [Lower CI, Upper CI]:
Estimate (p-value) for the relationship
between systolic blood pressure (SBP)
and diastolic blood pressure (DBP) and an
increase in PM10 on lag day 1
SBP: Warm season: 0.0798 (p < 0.001)
DBP: Warm season: 0.0240 (p< 0.001)
Note: No evidence of associations
between PM10 and BP during the cold
season
Reference: Chuang et al. (2007,
091063)
Period of Study: Between Apr-Jun 2004
or 2005
Location: Taipei, Taiwan
Outcome: High-sensitivity C-reactive
protein (hs-CRP)
Fibrinogen, plasminogen activator fibrin-
ogen inhibitor-1 (PAI-1), tissue-type plas-
minogen activator (tPA), 8-hydroxy-2'-
deoxyguanosine (8-OHdG), and log-
transformed HRV indices (SDNN - stan-
dard deviation of NN intervals, r-
MSSD - square root of the mean of the
sum of the squares of differences
between adjacent NN intervals, LF - low
frequency [0.04-0.15Hz], and HF - high
frequency [0.15-0.40Hzj)
Age Groups: 18-25 yrs
Study Design: Panel (cross-sectional)
N: 76 students
Statistical Analyses: linear mixed-
effects models
Covariates: Age, sex, BMI, weekday,
temperature of previous day, relative
humidity
Season: Only 1 season of data collection
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM10
Averaging Time: Hourly data used to
calculate averages over 1-3 day periods
Mean (SD): 1 -day avg: 49.2 (18.0)
2-day	avg: 55.3 (18.6)
3-day	avg: 54.9 (18.2)
Range (Min, Max): 1 -day avg: 29.5,
83.4
2-day	avg: 25.5, 85.1
3-day	avg: 22.2, 87.2
Monitoring Stations: 2 sites (each
pollutant measured at one site only)
PM Increment: IQR (1-d avg: 32.7
2-day	avg: 34.5
3-day	avg: 26.0)
Effect Estimate [Lower CI, Upper Cl[:
% change in health endpoint per increase
in IQR of PM10 (1-3 day averaging period
single pollutant models)
hs-CRP: 1-d: 135.8(1.8,269.7)
2-d: 108.2 (-10.9, 227.3)
3 d: 109.6 (2.5, 216.7)
8-OHdG: 1-d: -9.2 (-21.5, 3.2)
2-d:-6.1	(-17.0,4.8)
3-d:	-5.6 (-13.8, 2.6)
Copollutant: PM2.6, Sulfate, Nitrate, 0C,	1-d; 30.0 (12.4,47.7)
EC, NO?, CO, SO2, 0s	2-d: 19.1 (3.6, 34.7)
3 d: 21.2 (9.7, 32.8)
tPA: 1-d: 16.0 (-4.1, 36.2)
2-d: 10.4 (-6.3, 27.2)
3 d: 8.8 (-2.8, 20.5)
Fibrinogen: 1-d: 5.3 (1.5,15.2)
2-d: 1.5 (-4.4, 7.5)
3 d: 3.3 (-1.1, 7.7)
Heart Rate Variability
SDNN: 1-d: -4.9 (-7.8, -2.1)
2-d:	-4.0 (-6.6,-1.4)
3-d:	-4.1 (-6.1, -2.2)
r-MSSD: 1-d: -4.8 (-12.3, 2.7)
2-d:	-2.2 (-9.0, 4.7)
3-d:	-4.0 (-9.0, 0.9)
LF: 1-d:-6.1 (-10.1,-2.1)
2-d: -3.0 (-7.2,1.2)
3 d: -4.3 (-7.0,-1.6)
HF: 1-d: -5.5 (-13.0, 2.1)
2-d:	-2.7 (-9.5, 4.1)
3-d:	-2.0 (-7.2, 3.2)
July 2009
E-4
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ebelt et al. (2005, 056907)
Period of Study: Summer of 1998
Location: Vancouver, Canada
Outcome: CVD
Age Groups: range from 54-86 yrs
mean age - 74 years
Study Design: extended analysis of a
repeated-measures panel study
N: 16 persons with COPD
Statistical Analyses:
Earlier analysis expanded by developing
mixed-effect regression models and by
evaluating additional exposure indicators
Dose-response Investigated? No
Statistical Package: SAS V8
Pollutant: PMio
Averaging Time: 24 h
Mean (SD):
Ambient PMio: 17 ± 6
Exposure to ambient PMio: 10.3 ± 4.6
Range (Min, Max): Ambient PM102E: 7 -
36
Exposure to ambient PMio-2.b: 1.5 ¦ 23.8
Monitoring Stations: 5
Copollutant (correlation):
Ambient concentrations and exposure to
ambient PM were highly correlated for
each respective metric: r a 0.71
PMio-2.b: r 2 0.72
PM2.b: r> 0.92
Note: Total personal fine particle
exposure (T) were dominated by
exposures to non ambient particles which
were not correlated with ambient fine
particle exposure (A) or ambient
concentrations (C). Results for each of
these metrics are listed.
Effect estimates and 95% CI for IQR
range increases in exposure
Increment: C,G: IQR - 7//g|m3
SBP (mm Hg): -2.2 (-4.78-0.38)
DBP (mm Hg):-0.78 (-2.65-1.09)
Ln-SVE (bph): 0.16 (-0.07-0.40)
HR (bpm): 1.02 (-0.79-2.82)
SDNN(ms):-2.14 (-6.94-2.65)
R-MSSD (ms):-2.24 (-4.27-0.21)
Increment: A10: IQR - 6.5/yg/m3
SBP (mm Hg): -2.81 (-5.67-0.05)
DBP (mm Hg):-0.59 (-2.79-1.62)
Ln-SVE (bph): 0.27 (0.03-0.52)
HR (bpm): 0.86 (-1.61-3.33)
SDNN (ms):-3.91 (-9.73-1.91)
R-MSSD (ms): -0.81 (-4.94-3.31)
Reference: Folino et al. (2009,191902) Outcome: HRV & Inflammatory Markers
Period of Study: Jun 2006 - May 2007 Age Groups: 45-65 yrs
Location: Padua, Italy
Study Design: panel
N: 39 patients wf myocardial infarction
Statistical Analyses: Linear Regression
Model, AN0VA
Covariates: temperature, relative
humidity, atmospheric pressure, beta-
blocker, aspirin, or nitrate consumption,
smoking habit
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: NR
Pollutant: PMio
Averaging Time: 24h
Mean (SD):
Summer: 46.4 (16.1)
Winter: 73.0 (30.9)
Spring: 38.3 (15.4)
Monitoring Stations: NR
Copollutant: PM2.6, PM0.26
Co-pollutant Correlation
NR
PM Increment: 1 //g|m3
Beta (SE), p-value:
SDNN: 0.115(0.093), 0.218
SDANN: 0.138 (0.103), 0.182
RMSSD: 0.049 (0.0341,0.146
pH: 0.002 (0.0011,0.033
LTB4: 0.427(0.0279), 0.126
eNO: 0.000 (0.002), 0.851
PTX3:-0.003 (0.001), 0.033
C-reactive protein: -0.006 (0.004), 0.161
CC16:-0.002 (0.002), 0.280
IL-8: 0.000 (0.003), 0.895
Reference: Forbes et al. (2009,190351) Outcome: Inflammation Markers
Period of Study: 1994,1998, 2003 Age Groups: 16+ yrs
Location: England
Study Design: cross-sectional
N: 25,000 white adults wf fibrinogen
measurements & 17,000 white adults w/
C-reactive protein measurements
Statistical Analyses: Multilevel Linear
Regression Models
Covariates: age, sex, BMI, social class,
region, cigarette smoking
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: NR
Pollutant: PMio
PM Increment: 1 //g|m3
Averaging Time: yearly
Percent Change (Lower CI, Upper CI):
1994
Fibrinogen
Median: 19.5
1994 Crude:-0.068 (-0.367,0.231)
Range: 12.5-36.1
1994 Adjusted: 0.080 (-0.164, 0.326)
IQR: 3.7
1998 Crude:-0.592 (-0.902,-0.280)
1998
1998 Adjusted: -0.388 (-0.727, -0.047)
2003 Crude:-0.339 (-0.696, 0.019)
Median: 17.9
2003 Adjusted: -0.069 (-0.458, 0.322)
Range: 12.6-27.0
Combined:-0.077 (-0.254, 0.100)
IQR: 2.7
2003
C-reactive protein
1998 Crude:-0.914 (-2.206, 0.395)
Median: 16.2
1998 Adjusted:-0.266 (-1.782,1.274)
Range: 11.0-22.7
2003 Crude: 0.286 (-1.327,1.925)
IQR: 2.6
2003 Adjusted: 0.661 (-1.068, 2.421)
Monitoring Stations: NR
Combined: 0.140 (-1.003,1.296)
Copollutant: NO2, SO2, O3

Co-pollutant Correlation

n/a
July 2009
E-5
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kaufman (1987,190960)
Period of Study: Nov 2004 ¦ 2005
Location: Isfahan, Iran
Outcome: Inflammation
Age Groups: 10-18 yrs
Study Design: panel
N: 374 children
Statistical Analyses: Linear Regression,
Logistic Regression
Covariates: age, gender, BMI, waist
circumference, healthy eating index,
physical activity level
Dose-response Investigated? No
Statistical Package: SPSS
Lags Considered: 0-7d avg
Pollutant: PMio
Averaging Time: 24h
Mean (SD): 122.08 (33.63)
0": 11.00
25": 86.50
50": 153.0
75": 191.00
Monitoring Stations: 3
Copollutant: 0:i, SO2, NO2, CO
Co-pollutant Correlation
NR
PM Increment: NR
Beta |SE):
CRP: 1.5(0.2)
Ox-LDL: 1.4 (0.1)
MDA: 1.3 (0.1)
CDE: 1.1 (0.1)
H0MA-IR: 1.1 (0.3)
Reference: Liao et al. (2004, 056590)
Period of Study: 1996-1998
Location: ARIC study cohort (Washing-
ton County, MD
Forsyth County, NC
and selected suburbs of Minneapolis,
MN).
The 4th quarter of the ARIC cohort was
sampled exclusively from black residents
of Jackson, MS.
Outcome: 5-min HR, HRV indices (HF, LF,
SDNN)
Study Design: Cross-sectional
Statistical Analyses: Linear regression
Pollutant: PM10
Averaging Time: 24-h
Mean (SD): 24.3(11.5)
Copollutant: O3
CO
SO?
NO?
PM Increment: SD
Effect Estimate [Lower CI, Upper CI]:
Estimate (SE)
HF:-0.06 ms2 (0.018)
SDNN:-1.03 ms (0.31)
H: 0.32 beats/min (0.158)
Reference: Liao et al. (2005, 088677)
Period of Study: 1987-1989 baseline
health exam
Location: 3 centers in the US (Forsyth
County, NC
suburbs of Minneapolis, MN
black residents of Jackson, MS)
Outcome: Fibrinogen, factor VIII co-
agulant activity (Vlll-C), von Willebrand
factor (vWF), white blood cell count
(WBC), and serum albumin
Age Groups: 45-64 yrs
Study Design: Cross-sectional
N: 10,208 participants (7705 for PM)
Statistical Analyses: Multiple linear
regression
Covariates: Age, sex, ethnicity-center,
education, smoking, drinking status, BMI,
history of chronic respiratory disease,
humidity, season, cloud cover, and
temperature
Dose-response Investigated?
Yes, examined higher-ordered terms for
each pollutant
Statistical Package: SAS v8.2
Pollutant: PM10
Averaging Time: 24-h averages (1, 2,
and 3 days prior to the exam)
Mean (SD): 29.9 (29.9)
Mean (SD) within Quartiles:
Q1-3: 24.0(6.96)
Q4: 47.3(10.11)
Copollutant:
CO, SO2, N02,0s
PM Increment: 1 SD (12.8/yg/m3)
Effect Estimate: Adjusted regression
coefficient (SE): Fibrinogen (mg/dl): 0.163
(0.755)
Factor Vlll-C (%): Non-linear association:
P (PM 10) --5.30, p< 0.01
P (PM10)2 - 0.80, p< 0.05
vWF (%): Diabetics: 3.93 (1.80)
Nondiabetics: -0.54 (0.58)
Albumin (g/dl): CVD: -0.006 (0.003)
Non-CVD: 0.029 (0.017)
p < 0.05
July 2009
E-6
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Liao et al. (2007,180272)
Period of Study: 1999-2004
Location: 24 US states
Outcome: Ectopy
Age Groups: women 50-79 yrs
Study Design: panel
N: 57,422
Statistical Analyses: logistic regression
& random effects modeling
Covariates: age, race, center, education,
history of CVD/chronic lung disease, rel.
humidity, temperature, smoking
Dose-response Investigated? No
Statistical Package: SAS, Stata
Lags Considered: lags 0-365d
Monitors used in model for spatial
interpolation of daily PM values.
Pollutant: PMio
Averaging Time: daily
Mean (SD)*:
All: 27.5 (12.1)
No Ectopy: 27.5(12.1)
Any Ectopy: 27.5 (11.9)
5"1, 95"1 percentile*:
All: 12.2, 48.9
No Ectopy: 12.3, 48.8
Any Ectopy: 11.8, 49.3
Monitoring Stations: NR*
Copollutant: PM2.6
Co-pollutant Correlation
NR
PM Increment: 10 /yglm3
Percent Change (Lower CI, Upper CI):
All Ventricular Ectopy
Lag 0:1.01 (0.95,1.07)
Lag 1:1.02 (0.96,1.09)
Lag 2: 0.99 (0.93,1.06)
Current Smoker Ventricular Ectopy
Lag 0:1.21 (0.96,1.53)
Lag 1:1.32 (1.07,1.65)
Lag 2: 1.22 (0.95,1.56)
Nonsmoker Ventricular Ectopy
LagO: 1 (0.93,1.06)
Lag 1: 1.01 (0.94,1.07)
Lag 2: 0.98 (0.92,1.05)
All Supraventricular Ectopy
LagO: 1 (0.95,1.06)
Lag 1:1 (0.95,1.05)
Lag 2: 0.99 (0.94,1.04)
All Ventricular or Supraventricular Ectopy
Lag 0:1 (0.95,1.04)
Lag 1:1 (0.96,1.04)
Lag 2: 0.98 (0.94,1.02)
Reference: Liu et al. (2007,156705)
Period of Study: May 24, 2005-Jul 8,
2005
Location: Windsor, Ontario, Canada
Outcome: Heart rate, blood pressure,
brachial arterial diameter, flow-mediated
vasodilatation (FMD), plasma cytokines,
and thiobarbituric acid reactive
substances (TBARS)
Age Groups: 18-65 yrs
Study Design: Panel
N: 24 nonsmoking subjects with type I or
II diabetes over a 7 week period (2-14
visits for subjects)
170 total vascular measurements and
134 total blood samples collected
Statistical Analyses: Mixed effects
regression models
Covariates: (time-dependent covariates)
Daily temperature, relative humidity,
blood glucose level, also checked for
confounding by ambient air pollutant
concentrations (controlled for ambient
PM2.B)
Season: No adjustment since testing
was completed within a 7 week period
during early summer
Dose-response Investigated? No
Statistical Package: S-Plus
Pollutant: PMio (personal)
Averaging Time: Real-time monitor
measured exposure during 24-h period
prior to clinic measures
Median (5th-95th percentile): 0-24 hrs:
25.5 (9.8-133.0)
0-6hrs: 15.3 (5.3-83.2)
7-12hrs: 17.0 (7.1-186.3)
13-18hrs: 28.5 (11.4-167.0)
19-24 hrs: 30.5 (10.1-148.2)
Monitoring Stations: Personal
monitoring
Copollutant (correlation): Ambient
PM2.6 (r - 0.34)
PM Increment: 10/yg/m3
Effect Estimate [Lower CI, Upper CI]:
"p < 0.05
"p <0.10. Regression coefficients (SE)
End-diastolic basal diameter (/jm): All
subjects (n — 24): -2.52 (3.27)
subjects not taking vasoactive meds
(n-17):-3.93 (3.66)
subjects w/BMI £ 29kg/m2 (n -14): 8.85
(5.85)
End-systolic basal diameter |//m): All
subjects (n — 24): -9.02 (3.58)"
subjects not taking vasoactive meds
(n-17):-10.59 (4.36)"
subjects w/BMI £29kg/m2 (n -14): 3.85
(5.49)
End-diastolic FMD (%): All subjects
(n - 24): 0.20 (0.08)"
subjects not taking vasoactive meds
(n-17): 0.23 (0.09)"
subjects w/BMI £29kg/m2 (n — 14): 0.12
(0.05)"
End-systolic FMD (%): All subjects
(n - 24): 0.38 (0.18)"
subjects not taking vasoactive meds
(n-17): 0.51 (0.22)"
subjects w/BMI £ 29kg/m2 (n -14): 0.18
(0.10)*
Flow (cm/s): All subjects (n — 24): -0.16
(0.19)
subjects not taking vasoactive meds
(n-17):-0.48 (0.21)"
subjects w/BMI £ 29kg/m2 (n -14): -0.39
(0.23)*
Heart rate (bpm): All subjects (n — 24):
0.01 (0.11)
subjects not taking vasoactive meds
(n-17):-0.06 (0.12)
subjects w/BMI £ 29kg/m2 (n -14): 0.15
(0.12)
Diastolic blood pressure (mm Hg): All
subjects (n — 24): 0.19 (0.16)
subjects not taking vasoactive meds
(n-17): 0.40 (0.18)**
July 2009
E-7
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
(0.21)
Systolic blood pressure (mm Hg): All
subjects (n — 24): 0.17 (0.19)
subjects not taking vasoactive meds
(n-17): 0.43 (0.24)"
subjects wI BMI £ 29kg/m2 (n — 14): 0.38
(0.24)
CRP (|jg/mL): All subjects (n — 24): 0.11
(0.07)
subjects not taking vasoactive meds
(n-17): 0.10 (0.09)
subjects wI BMI £ 29kg/m2 (n — 14): 0.02
(0.03)
ET-1 (pg/mL): All subjects (n — 24): 0.00
(0.00)
subjects not taking vasoactive meds
(n-17): 0.00 (0.00)
subjects w/BMI £ 29kg/m2 (n -14): 0.00
(0.01)
IL-6 (pg/mL): All subjects (n — 24): 0.00
(0.05)
subjects not taking vasoactive meds
(n-17): 0.01 (0.05)
subjects w/BMI £ 29kg/m2 (n -14): -0.00
(0.03)
TI\IF-a (pg/mL): All subjects (n — 24):
0.03(0.05)
subjects not taking vasoactive meds
(n-17): 0.02 (0.05)
subjects wI BMI £ 29kg/m2 (n — 14): 0.03
(0.08)
TBARS (pmol/mL) All subjects (n - 24):
16.12(4.00)"
subjects not taking vasoactive meds
(n-17): 8.10 (9.18)
subjects wI BMI £ 29kg/m2 (n — 14): ¦
0.28(6.60)
regression coefficients (SE) among
subjects not taking vasoactive
medications, with lag time
End-diastolic basal diameter (/jm): 0-
6 h: 29.91 (10.64)"
7-12 h: 0.72(3.95)
13-18 h:-3.62 (2.80)
19-24 h: -0.57 (1.7)
End-systolic basal diameter (/-/m): 0-6
h: 28.88 (11.22)"
7-12 h:-0.78 (4.58)
13-18 h: -7.70 (3.30)"
19-24 h: -2.87 (2.05)
End-diastolic FMD (%): 0-6 h:-0.12
(0.10)
7-12 h: 0.04 (0.05)
13-18 h: 0.11 (0.03)"
19-24 h: 0.12(0.04)"
End-systolic FMD (%): 0-6 h: 0.36
(0.08)"
7-12 h: 0.48 (0.32)
13-18 h: 0.19(0.06)"
19-24 h: 0.34 (0.13)"
Flow (cm/s): 0-6 h:-0.34 (0.22)
7-12 h: -0.26 (0.27
13-18 h: -0.27 (0.15)"
19-24 h: -0.30 (0.11)"
Heart rate (bpm): 0-6 h: 0.31 (0.13)"
7-12 h: 0.26(0.12)"
13-18 h: 0.01 (0.09)
19-24 h:-0.08 (0.05)
July 2009
E-8
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lipsett et al. (2006, 088753)
Period of Study: February-May 2000
Location: Coachella Valley, CA
Outcome: HRV parameters: SDNN,
SDANN, r-MSSD, LF, HF, total power,
triangular index (TRII).
Study Design: Panel study
N: 19 non smoking adults with coronary
artery disease
Statistical Analysis: Mixed linear re-
gression models with random effects
parameters
Pollutant: PMio
Averaging Time: 2 h
Mean (range): Indio: 23.2 (6.3-90.4)
Palm Springs: 14 (4.7-52)
Monitoring Stations: 2
Copollutant: 0:i
Diastolic blood pressure (mm Hg): 0-
6 h: -0.29 (0.12)"
7-12 h: 0.24 (0.12)"
13-18 h: 0.46 (0.17)"
19-24 h: 0.18(0.14)
Systolic blood pressure (mm Hg): 0-
6 h: -0.65 (0.18)"
7-12 h: 0.17(0.19)
13-18 h: 0.86 (0.24)"
19-24 h: 0.11 (0.10)
CRP (|jg/mL): 0-6 h: 0.15(0.13)
7-12 h: 0.15(0.13)
13-18 h: 0.03 (0.06)
19-24 h: 0.04 (0.03)
ET-1 (pg/mL): 0-6 h: 0.02(0.00)": 7-12
h: -0.00 (0.00)
13-18 h:-0.00 (0.00)
19-24 h: 0.00 (0.00)
IL-6 (pg/mL): 0-6 h: 0.03 (0.06)
7-12 h: 0.00(0.06)
13-18 h: 0.02 (0.03)
19-24 h: 0.00 (0.02)
TNF-a (pg/mL): 0-6 h: 0.01 (0.07)
7-12 h: 0.09(0.04)"
13-18 h: 0.01 (0.04)
19-24 h: -0.00 (0.03)
TBARS (pmol/mL): 0-6 h: -4.44 (6.72)
7-12 h: 11.94 (5.08)"
13-18 h: 5.06(4.03)
19-24 h: 1.06(4.64)
Note: Adding ambient PM2.6 data as a
covariate in the model yielded similar
regression coefficients for personal PMio
PM Increment: SE*1000
Effect Estimate (change in HRV per
unit increase in PM concentration):
SDNN: -0.71 msec (SE - 0.268)
Notes: Weekly ambulatory 24 h ECG
recordings (once per week for up to 12
weeks), using Holter monitors, were
made. Subjects' residences were within 5
miles of one of two PM monitoring sites.
Regressed HRV parameters against 18:
00-20: 00 mean particulate pollution.
July 2009
E-9
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ljungman et al. (2008,
180266)
Period of Study: Aug 2001 - Dec 2006
Location: Gothenburg & Stockholm,
Sweden
Outcome: Ventricular Arrhythmia
Age Groups: 28-85 yrs
Study Design: case-crossover
N: 88 patients wf implantable
cardioverter defibrillators
Statistical Analyses: conditional
logistic regression
Covariates: temperature, humidity,
pressure, ischemic heart disease, ejection
fraction, heart disease, diabetes, use of
beta-blockers, age, BMI, location at time
of arrhythmia, distance from air pollution
monitor
Dose-response Investigated? No
Statistical Package: Stata, S-plus
Lags Considered: lags 2-24h
Pollutant: PMio
Averaging Time: hourly
Gothenburg, Stockholm
Median:
2h: 18.95,14.62
24h: 19.92,15.23
Min:
2h: 0.00, 0.33
24h: 2.13,3.96
Max:
2h: 203.75,159.79
24h: 78.01,90.50
IQR:
2h: 14.16,11.59
24h: 11.49,9.59
Monitoring Stations: 2
Copollutant: PM2.6, NO2
Co-pollutant Correlation
2h NO?: 0.36
24h NO?: 0.29
PM Increment: Interquartile Range
Odds Ratio (Lower CI, Upper CI):
2h: 1.31 (1.00,1.72)
24h: 1.24 (0.87,1.76)
Notes: OR of ventricular arrhythmia for
an IQR increase of air pollutants in
different subgroups (figure 2)
Reference: Ljungman et al. (2009,
191983)
Period of Study: May 2003 - July 2004
Location: Athens, Greece
Helsinki, Finland
Ausburg, Germany
Barcelona, Spain
Rome, Italy
Stokholm, Sweeden
Outcome: lnterleukin-6 Response
Age Groups: 35-80 yrs
Study Design: panel
N: 955 male myocardial infarction
survivors
Statistical Analyses: Additive Mixed
Models
Covariates: age, sex, BMI, city,
HDL/total cholesterol, smoking, alcohol
intake, HbA1c, NT-proBNP, history of M
heart failure, or diabetes, phlegm
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1d
Pollutant: PM10
Averaging Time: 24h
Mean: 31.6
25": 21.1
75": 38.4
Monitoring Stations: NR
Copollutant: CO, NO2, PNC, PM2.6
Co-pollutant Correlation
PM2.6: 0.81
PM Increment: Interquartile Range (17.4
j"g/m3)
Change of IL-6 (Lower CI, Upper CI), p-
value:
0.0(-1.3,1.3), 1.0
Reference: Mar et al. (2005, 087566)
Period of Study: 1999-2001
Location: Seattle, WA
Outcome: Change in arterial O2 satura-
tion, heart rate, and blood pressure (SBP
and DBP)
Age Groups: >75 years
Study Design: Panel study
N: 88 elderly subjects
Statistical Analysis: GEE
Pollutant: PM10
Averaging Time: 24-hs
Mean (SD): Indoor: 12.6 (7.8)
Outdoor: 14.5 (7.0)
PM Increment: 10 |Jgfm3
Unit change in measure(95% CI):
Among all subjects: Each increase in
outdoor same day PM10 was associated
with: SBP: -0.10 mmHg (95% CI:-1.37,
1.18)
DBP: -0.03 mmHg (95% CI:-0.79, 0.73)
HR: -0.48 beats/min (95% CI:-1.03,
0.06)
Each increase in indoor same day
PM2.6 was associated with: SBP: 0.92
mmHg (95% CI:-0.95, 2.78)
DBP: 0.63 mmHg (95% CI:-0.29,1.56)
HR: 0.02 beats/min (95% CI: -0.54, 0.58)
Notes: Results by health status
presented in Fig 1. Used 2 sessions that
each were 10 consecutive days of
measurement. Used personal, indoor, and
outdoor measures of PM2.6
July 2009
E-10
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Reference: Metzgeret al. (2007,
092856)
Period of Study: January 1993-
December 2002
Location: Atlanta, GA
	Design & Methods	
Outcome: Days with any event recorded
by the ICD, days with ICD
shocksfdefibrillation and days with either
cardiac pacing or defibrillation
Study Design: Repeated measures
N: 884 subjects
Statistical Analysis: Logistic regression
with GEE to account for residual
autocorrelation within subjects
Concentrations'
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): 28.0(12.2)
Median: 26.4
Copollutant: 0:i, NO2, CO, SO2.
Aug1998 Dec2002: Oxygenated
hydrocarbons
Effect Estimates (95% CI)
PM Increment: OR (95% CI):
Outcome - Any event recorded by ICD
OR - 1.00(95% CI: 0.97,1.03)
Reference: Min et al. (2008,191901)
Period of Study: Dec 2003-Jan 2004
Location: Taein Isalnd, South Korea
24h-avg: -13.15 (-23.36,-1.57)**
48-havg:-0.10 (-9.99,10.87)
72-h avg: -7.61 (-17.04,2.88)
HF
6-h avg: -1.07 (-10.43, 9.28)
9-h avg:-3.28 (-13.72, 8.43)
12-h avg: -4.06 (-14.77, 8.00)
24h-avg: -1.22 (-13.96,13.41)
48-havg:-3.55 (-14.01, 8.18)
72-h avg:-3.88 (-14.64, 8.23)
Notes: Percent Change in HRV for air
pollution children, adults, and the elderly
(figure 2)
Percent Change in HRV for PM10 exposure
in all ages (figure 3)
Outcome: Heart Rate Variability
Age Mean (SD): 44.3 (21.9)
Study Design: Panel
N: 1.349 participants
Statistical Analyses: Linear Regression
Covariates: age, sex, BMI, smoking
Dose-response Investigated? No
Statistical Package: SAS, R
Lags Considered: 0-72h
Pollutant: PM10
Averaging Time: 1 h
Mean (SD): 33.244(19.017)
Percentiles:
25th: 18.000
50th: 26.000
75th: 41.000
Range: 187.000. 16.000
Monitoring Stations: 1
Copollutant: NO2, SO2
PM Increment: 1 SD (19 |Jg|m3)
Percent Change: [Lower CI, Upper CI]:
SDNN
6-h avg: -4.34 (-7.99, -0.55)**
9-h avg: -5.48 (-9.61,-1.17)**h"
12-h avg:-6.23 (-10.47,-1.79)***
24h-avg: -4.73 (-9.73, 0.56)-
48-h avg:-1.25 (-5.59, 3.29)
72-h avg:-0.85 (-5.35, 3.86)
LF
6-h avg: -10.32 (-18.05,-1.86)**
9-h avg:-13.79 (-22.26, -4.39)***
12-h avg:-14.48 (-23.18, -4.80)***
Reference: Peters et al. (2009,191992) Outcome: Plasma Fibrinogen
Period of Study: May 2003 - July 2004 Age Groups: 37-81
Study Design: panel
N: 854 adults
Location: Helsinki, Finland
Ausburg, Germany
Barcelona, Spain
Rome, Italy
Stokholm, Sweeden
Statistical Analyses: Additive Mixed
Models
Covariates: age, sex, BMI, city,
HDL/total cholesterol, smoking, HbA1c,
NT-proBNP, history of arrhythmia,
asthma, arthrosis, stroke, bronchitis,
season, apparent temperature, relative
humidity, weekday, hour of visit
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-5d avg
Pollutant: PM10
Averaging Time: 24h
Mean (SD): 30.3
Min: 0
Max: 194
Monitoring Stations: NR
Copollutant: PM2.6, PM102 E
Co-pollutant Correlation
NR
PM Increment: 13.5 /yglm3
Change (Lower CI, Upper CI):
Genotype 1 1
rs2070006:1.22 (0.47,1.96)
rs2070011:1.16 (0.41,1.90)
rs1800790: 0.27 (-0.36, 0.91)
rs2227399: 0.27 (-0.36, 0.91)
rs6056: 0.19 (-0.45, 0.83)
rs4220: 0.19 (-0.45, 0.83)
Haplotype in cluster 2: 0.09 (-0.53, 0.76)
rs1800791: 0.18 (0.21,1.40)
Genotype 1 2
rs2070006: 0.5 (-0.19, 2.15)
rs2070011:0.42 (-0.28,1.13)
rs1800790:1.28 (0.54,2.01)
rs2227399:1.28 (0.55, 2.02)
rs6056: 1.26 (0.49, 2.04)
rs4220: 1.27 (0.49, 2.04)
Haplotype in cluster 2: 1.17 (0.35,1.99)
rs1800791: 0.40 (-0.48,1.28)
Genotype 2 2
rs2070006: 0.11 (-1.94,2.15)
rs2070011:0.08 (-2.08, 2.24)
rs1800790: 2.15 (0.71, 3.60)
S2227399: 2.18(0.73, 3.63)
s6056: 2.24 (0.72, 3.77)
s4220: 2.25 (0.73, 3.78)
Haplotype in cluster 2: 2.16 (0.61, 3.71)
rs1800791:-0.13 (-1.84,1.58)
July 2009
E-11
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Rosenlund et al. (2007,
114679)
Period of Study: 1985-1996
Location: Stockholm County
Outcome: Myocardial Infarction
Age Groups: 15-79 yrs
Study Design: case-control
N: 24,387 first event of myocardial
infarction cases and 276,926 population
based controls
Statistical Analyses: Logistic
Regression
Covariates: age, sex, calendar year, SES
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: 5yr
Pollutant: PMio
Averaging Time: 5yrs
Cases
Median: 2.4
5"'-95th: 0.3-6.2
Controls
Median: 2.2
5"'-95th: 0.3-6.0
Monitoring Stations: NR
Copollutant: NO2, CO
Co-pollutant Correlation
HNR
PM Increment: 5th to 95th percentile
(5/yg/m3)
Odds Ratio (Lower CI, Upper CI):
All Subjects
Controls: 1.0
All Cases: 1.04 (1.00,1.09)
Nonfatal Cases: 0.98 (0.963,1.03)
Fatal Cases: 1.16(1.09,1.24)
ln-hospital death: 1.05 (0.95,1.17)
Out-of-hospital death: 1.23 (1.14,1.33)
Subjects who did not move bit population
censuses
Controls: 1.0
All Cases: 1.11 (1.02,1.21)
Nonfatal Cases: 1.05 (0.96,1.15)
Fatal Cases: 1.56(1.28,1.91)
ln-hospital death: 1.58 (1.13, 2.19)
Out-of-hospital death: 1.56 (1.22,1.98)
Reference: Ruckerl et al. (2007,
156931)
Outcome: lnterleukin-6 (IL-6), fibrinogen,
C-reactive protein (CRP)
Pollutant: PMii
PM Increment: IQR
Period of Study: May 2003—Jul 2004 Age Groups: 35-80 yrs
Location: Athens, Augsburg, Barcelona,
Helsinki, Rome, and Stockholm
Study Design: Repeated measures I
longitudinal
N: 1003 Ml survivors
Statistical Analyses: Mixed-effect
models
Covariates: City-specific confounders
(age, sex, BMI)
long-term time trend and apparent
temperature
RH, time of day, day of week included if
adjustment improved model fit
Season: Long-term time trend
Dose-response Investigated? Used p-
splines to allow for nonparametric
exposure-response functions
Statistical Package: SAS v9.1
Averaging Time: Hourly and 24-h (lag 0- Effect Estimate [Lower CI, Upper CI]:
4, mean of lags 0-4, mean of lags 0-1, % change in mean blood markers per
mean of lags2-3, means of lags 0-3) increase in IQR increase of air pollutant.
IL-6: Lag (IQR): % change in GM (95%CI)
Mean (SD): Presented by city only
Percentiles: NR
Range (Min, Max): NR
Lag 0(17.4)
Lag 1 (17.4)
Lag 2 (17.4)
¦0.34 (-1.66, 0.99)
0.69(-1.95, 0.58)
1.59 (-3.99, 0.88)
5-d avg (13.5):-0.87 (-2.28, 0.55)
Monitoring Stations: Central monitoring .
? .,	0 Fibrinogen:
citoc in oann nti/	°
sites in each city
Copollutant: SO2
03
NO
NO?
Lag (IQR): % change in AM
(95%CI)
Lag 0(17.4): 0.06 (-0.43, 0.55)
Lag 1 (17.4): 0.14 (-0.35, 0.63)
Lag 2 (17.4): 0.24 (-0.24,0.72)
5-d avg (13.5): 0.60(0.10,1.09)
CRP: Lag (IQR): % change in GM (95%CI)
Lag 0(17.4)
Lag 1 (17.4)
Lag 2 (17.4)
¦0.71 (-2.75,1.37)
¦0.63 (-2.61,1.39)
1.42(4.23,1.47)
5-d avg (13.5):-1.35 (-3.45, 0.79)
July 2009
E-12
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ruckerl et al. (2006,
088754)
Period of Study: Oct 2000—Apr 2001
Location: Erfurt, Germany
Outcome: C-reactive protein (CRP)
serum amyloid A (SAA)
E-selectin
vWF
intercellular adhesion molecule-1 (ICAM-
1)
fibrinogen
Factor VII
prothrombin fragment 1 +2
Ddimer
Age Groups: 50+ yrs
Study Design: Panel (12 repeated
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Statistical Analyses: Fixed effects
linear and logistic regression models
Covariates: Models adjusted for differ-
ent factors based on health endpoint
CRP: RH, temperature, trend, ID
ICAM-1: temperature, trend, ID
vWF: air pressure, RH, temperature,
trend, ID
FVII: air pressure, RH, temperature, trend,
ID, weekday
Season: Time trend as covariate
Dose-response Investigated? Sensi-
tivity analyses examined nonlinear
exposure-response functions
Statistical Package: SAS v8.2 and S-
Plus v6.0
Pollutant: PMio
Averaging Time: 24-h
Mean (SD): 20.0(13.0)
Percentiles: 25th: 10.8
50th: 15.6
75th: 26.0
Range (Min, Max): 5.4, 74.5
Monitoring Stations: 1 site
Copollutant: UFPs (ultrafine particles)
AP (accumulation mode particles)
PM2.6
PM10
0C (organic carbon)
EC (elemental carbon)
NO?
CO
PM Increment: IQR (15.2
5-davg: 12.8)
Effect Estimate [Lower CI, Upper CI]:
Effects of air pollution on blood markers
presented as OR (95%CI) for an increase
in the blood marker above the 90th
percentile per increase in IQR air
pollutant.
CRP: Time before draw: 0 to 23 h: 1.2
(0.8,1.9)
24 to 47 h: 2.0(1.1,3.6)
48 to 71 h: 2.2(1.2, 3.8)
5-d mean: 2.0 (1.2, 3.7)
ICAM-1: Time before draw: 0 to 23 h:
1.3(0.9,1.8)
24 to 47 h: 3.1 (2.0, 4.8)
48 to 71 h: 3.4(2.2, 5.2)
5-d mean: 3.4 (2.2, 5.3)
Effects of air pollution on blood markers
presented as % change from the
meanfGM in the blood marker per
increase in IQR air pollutant.
vWF: Time before draw: 0 to 23 h: 4.0
(-0.6, 8.5)
24 to 47 h: 6.0(0.6,11.5)
48 to 71 h: 1.1 (-4.9,7.0)
5-d mean: 6.1 ( 0.6,12.8)
FVII: Time before draw: 0 to 23 h: -6.6
(-10.4 to -2.5)
24 to 47 h:-8.4 (-12.3 to-4.3)
48 to 71 h: -5.9 (-9.6,-2.0)
5-d mean: -8.0 (-12.4, -3.4)
Note: summary of results presented in
figures. SAA results indicate increases in
association with PM (not as strong and
consistent as with CRP)
no association observed between E-
selectin and PM
an increase in prothrombin fragment 1 +2
was consistently observed, particularly
with lag 4
fibrinogen results revealed few significant
associations, potentially due to chance
D dimer results revealed null associations
in linear and logistic analyses
July 2009
E-13
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ruckerl et al. (2007,
091379)
Period of Study: Oct 2000—Apr 2001
Location: Erfurt, Germany
Outcome: Soluble CD40 ligand (sCD40L),
platelets, leukocytes, erythrocytes,
hemoglobin
Age Groups: 50+ yrs
Study Design: Panel (12 re
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Statistical Analyses: Fixed effects
linear regression models
Covariates: Long-term time trend, week-
day of the visit, temperature, RH, baro-
metric pressure
Season: Time trend as covariate
Dose-response Investigated? No
Statistical Package: SAS v8.2 and S-
Plus v6.0
Pollutant: PMio
Averaging Time: 24-h
Mean (SD): 20.0(13.0)
Percentiles: 25: 10.8
50: 15.6
75: 26.0
Range (Min, Max): 5.4, 74.5
Monitoring Stations: 1 site
Copollutant: UFPs (ultrafine particles),
AP (accumulation mode particles), PM2.6,
PM10, NO
PM Increment: IQR (15.2
5-d avg: 12.8)
Effect Estimate [Lower CI, Upper CI]:
Effects of air pollution on blood markers
presented as % change from the
meanfGM in the blood marker per
increase in IQR air pollutant.
sCD40L, % change GM (pg/mL): lagO:
1.6(-3.5, 7.0)
lag 1: 1.1 (-5.4,7.9)
Iag2: -3.5 (-8.9, 2.2)
Iag3: -1.4 (-6.0, 3.4)
5-d mean: -1.2 (-7.8, 5.8)
Platelets, % change mean (103/|Jl):
lagO: -0.4 (-1.9,1.0)
Iag1: 0.4 (-1.4, 2.3)
Iag2: 0.5 (-1.4, 2.3)
Iag3: -0.1 (-1.6,1.4)
5-d mean: 0.0 (2.1, 0.0)
Leukocytes, % change in mean
(103/|Jl): lagO: -1.1 (-2.8,0.7)
Iag1: -0.5 (-2.6,1.5)
Iag2: 0.1 (-2.1,2.4)
Iag3: -0.7 (-2.6,1.2)
5-d mean: -1.1 (-3.6,1.4)
Erythrocytes, % change mean (10E/|Jl):
lagO: 0.0 (-0.4, 0.5)
Iag1: -0.4 (-1.0, 0.1)
Iag2: -0.7 (-1.2, -0.2)
Iag3: -0.4 (-0.8, 0.0)
5-d mean: -0.6 (-1.2, -0.1)
Hemoglobin, % change mean (g/dl):
lagO: -0.1 (-0.7,0.6)
Iag1: -0.4 (-1.2, 0.3)
Iag2: -0.7 (-1.3, 0.0)
Iag3: -0.3 (-0.9, 0.2)
5-d mean: -0.7 (-1.5, 0.1)
Reference: Steinvil et al. (2008,
188893)
Period of Study: 2003-2006
Location: Tel-Aviv, Israel
Outcome: Inflammation
Age Groups:
Mean (SD): 46 (12) years
Study Design: panel
N:3659
Statistical Analyses: Linear Regression
Covariates: age, waist circumference,
BMI, HDL, 0LDL, triglycerides, diastolic 81
systolic BP, alcohol consumption, sports
intensity, medications, smoking status,
family history of CHD, temperature,
humidity, precipitation, season, 81 year
Dose-response Investigated? No
Statistical Package: SPSS
Lags Considered: 0-7d
Pollutant: PM10
Averaging Time: 24h
Mean (SD): 64 (100.8)
25": 33.1
50": 43.0
75": 60.7
Monitoring Stations: NR
Copollutant: SO2, NO2, O3, CO
Co-pollutant Correlation
SO2: 0.043
NO2: 0.082
0s:-0.113
CO: 0.075
PM Increment: Interquartile Range (27.6
j"g/m3)
hs-CRP Relative % Change (Lower CI,
Upper CI):
Men:
Lag 0: -1 (-2,1)
Lag 1:0 (-1,1); Lag 2:-1 (-2,1)
Lag 3: -1 (-2, 0)
Lag 4:0 (-1,1)
Lag 5:0(-1,2)
Lag 6: 1 (0, 2)
Lag 7:1 (0,1)
0-7 avg: -2 (-5,1)
Women:
Lag 0: 0 (-2, 2)
Lag 1:0 (-1,2)
Lag 2: 1 (0, 2)
Lag 3:0 (-1,1)
Lag 4:0 (-1,2)
Lag 5:0(-1,2)
Lag 6: -1 (-3,1)
Lag 7:0 (-2,1)
0-7 avg: 1 (-2, 4)
Fibrinogen Absolute % Change (Lower
CI, Upper CI):
Men:
LagO: 0.7(0.0,1.5); Lag1: 0.4( 0.2, 0.9);
Lag2:-0.1( 0.9, 0.6)
Lag3: -0.1( 0.7, 0.6); Lag4: 0.0( 0.7, 0.7);
Lag5: 0.1 (-0.7,1.0)
Lag6: 0.61-0.1,1.3); Lag7: 0.4(0.0,0.8);
July 2009	E-14	DRAFT - DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
0-7 avg: -0.4(-1.9,1.0)

Women:

LagO: 0.3( 0.6,1.2); Lag1: -0.1( 0.8,
0.7);
Lag2: -0.3( 0.9, 0.3)

Lag3: -0.1( 0.7, 0.5); Lag4: 0.2( 0.4,
0.9):
Lag5: 0.2( 0.7,1.2)

Lag6:-0.3(-1.4, 0.8); Lag7: 0.7( 0.1,
1.5):
0-7 avg: 0.0(-1.5,1.5)

WBC Absolute Change (Lower CI,

Upper CI):

Men:

LagO: 2 (-22, 27)

Lag1: 3 (-14,19)

Lag2: 1 (-22, 24)

Lag3: -7 (-28,14)

Lag4: -22 (-44,-1)

Lag5: -20 (-46, 7)

Lag6: -5 (-27,16)

Lag7: -4(-16, 9)

0-7avg: -11 (-58, 36)

Women:

Lag 0: 20 (-6, 46)

Reference: Su et al. (2006,157022)
Period of Study: February-April 2002
Location: Taipei, Taiwan
Outcome: Total cholesterol, HDL,
tryglycerides, LDL, hs-CRP, IL-6, TNF-a,
tPA, PAI-1, and fibrinogen
Age Groups: 40-75 years
Study Design: Panel study
N: 49 subjects (31 males and 18 females)
with coronary heart disease or multiple
risk factors for CHD
Statistical Analysis: Linear mixed
effects regression
Pollutant: PMio
Averaging Time: 1 h
(High pollution day
to 18: 00 >100)
Copollutant: 0:i
¦ PMio from 08: 00
PM Increment: High vs. Low pollution
days
Effect Estimate [Lower CI, Upper CI]:
CHD patients (n = 23): Pvalue for
paired t-test comparing health endpoint
means on high and low pollution days
hs-CRP: p - 0.568
IL-6: p - 0.856
TNF-a: p - 0.246
PAI-1: p - 0.008
tPA: p - 0.322
Fibrinogen: p - 0.189
P value for health endpoint in mixed-
effects models
PAI-1: p - 0.010
tPA: p - 0.329
Fibrinogen: p - 0.747
Patients with multiple CHD risk
factors (n = 26): P value for paired t-
test comparing health endpoint means on
high and low pollution days
hs-CRP: p - 0.475
IL-6: p - 0.561
TNF-a: p - 0.572
PAI-1: p - 0.098
tPA: p - 0.260
Fibrinogen: p - 0.087
P value for health endpoint in mixed-
effects models
PAI-1: p - 0.891
tPA: p - 0.789
Fibrinogen: p - 0.923
Notes: Subjects had paired fasting blood
samples taken during high and low air
pollution days.
July 2009
E-15
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Vedal et al„ (2004, 055630)
Period of Study: 1997-2000
Location: Vancouver, British Columbia
Outcome: Implantable cardioverter
defibrillator (ICD) discharge
Age Groups: All
Study Design: Time series
(Retrospective, longitudinal panel study)
N: 50 ICD patients with 1 + discharges
(40,328 person-days and 257 arrhythmia
event days)
Statistical Analyses: Multiple logistic
regression with GEE
Covariates: Temperature, relative
humidity, barometric pressure, rainfall,
wind direction and speed
Season: Summer (May-Sep) and winter
(Oct-Apr)
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: -3 days
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max): 12.9 (3.8-49.3)
SD - 5.6
Monitoring Stations: 8
Copollutant (correlation): O3: r - 0.11
SO2: r - 0.70
NO?: r - 0.49
CO: r - 0.43
Other variables: Temp: r - 0.43
Humidity: r - -0.35
Baro Pressure: r - 0.26
Rain: r - -0.63
Wind: r - -0.53
PM Increment: 5.6/yglm3 (SD)
Percent Change [CI]: Values NR
Notes: The author states that significant
negative associations were found for ICD
discharge with same-day lag, and also for
3-day lag with more arrhythmia-prone
patients. All other non-significant percent
change estimates are shown in Fig 3 and
4.
Reference: Vedal et al. (2004, 055630) Outcome: ICD discharges (arrhythmias)
Period of Study: 1997-2000
Location: Vancouver, British Columbia,
Canada
N: 150 patients w/ICD, 4 yrs
Statistical Analysis: Logistic
regression, GEE
Covariates: Temporal trends,
temperature, relative humidity, wind
d, rain
Pollutant: PM10
Mean: 12.9 (SD - 5.6)
Copollutant): O3, SO2, NO2, CO
Increment: 1 SD
Effect Estimates, e.g., % change in the
rate of arrhythmia, were presented in
Figure 3. No association with PM10 was
observed while SO2 was associated with
an increase in the rate of arrhythmia
among 16 patients with at least 2
discharges per year.
Season: Summer, Winter
Dose-response Investigated? No
Lags Considered: 0.1.2.3d
July 2009
E-16
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Whitsel et al. (2009,
191980)
Period of Study: 1993-2004
Location: US
Outcome: Heart Rate Variability
Age Groups: 50-79 yrs
Study Design: panel
N: 4,295 women
Statistical Analyses: Random Effects
Model
Covariates: temperature, humidity
Dose-response Investigated? No
Statistical Package: SUDAAN
Lags Considered: 0
Pollutant: PMio
Averaging Time: 24h
Amsterdam
Mean: 20.0
Min: 3.8
25": 10.4
50": 16.9
75": 23.9
Max: 82.2
Erfurt
Mean: 23.1
Min: 4.5
25": 10.5
50": 16.3
75": 27.4
Max: 118.1
Helsinki
Mean: 12.7
Min: 3.1
25": 8.1
50": 10.6
75": 16.0
Max: 39.8
Monitoring Stations: 3
Copollutant: NR
Co-pollutant Correlation
n/a
PM Increment: 10/yg/m3
Beta (Lower CI, Upper CI):
Supine Position, Amsterdam
Lag 0
|1
I 2
¦0.06 (-0.95, 0.84)
0.18(-0.74,1.10)
0.93 (0.01,1.85)
5d avg: 0.49 (-0.74,1.72)
Supine Position, Erfurt
Lag 0
|1
I 2
0.36 (-0.83, 0.11)
¦0.40 (-0.91,0.11)
¦0.68 (-1.20,-0.17)
5d avg:-0.68 (-1.44, 0.09)
Supine Position, Helsinki
Lag 0:-0.44 (-2.27,1.40)
Lag 1: -0.17 (-1.69,1.3.5)
Lag 2: -1.14 (-2.51, 0.23)
5d avg: -0.59 (-3.08,1.90)
Supine Position, Pooled
Lag 0
|1
I 2
¦0.30 (-0.71,0.11)
¦0.25 (-0.68, 0.18)
0.26 (-1.22, 0.70)*
5d avg: -0.36 (-0.99, 0.27)
Standing Position, Amsterdam
Lag 0
|1
I 2
¦0.44 (-1.6, 0.72)
¦0.61 (-1.8, 0.59)
0.32(-0.88,1.51)
5d avg: -0.55 (-2.15,1.04)
Standing Position, Erfurt
Lag 0
|1
I 2
i: -0.59 (-1.24,0.06)
:-0.70 (-1.42, 0.03)
!: -0.65 (-1.37, 0.07)
5d avg: -0.68 (-1.74, 0.39)
Standing Position, Helsinki
Lag 0:1.17 (-1.46, 3.80)
Lag 1:0.01 (-2.17, 2.19)
Lag 2: -0.63 (-2.60,1.34)
5d avg: -1.96 (-5.51,1.60)
Standing Position, Pooled
Lag 0
|1
I 2
0.48 (-1.03, 0.07)
¦0.62 (-1.21,-0.03)
¦0.41 (-1.00, 0.17)
5d avg: -0.72 (-1.57, 0.14)
*p<0.1
Reference: Yeatts et al. (2007, 091266)
Outcome: Heart Rate Variability
Pollutant: PMio
PM Increment: 1 /yglm3
Period of Study: 12 wk period bit Sept
Age Groups: 21-50 yrs
Averaging Time: 24h
Beta, SE, p-value (Lower CI, Upper CI):
2003 -July 2004


NR
Study Design: panel
Mean (SD): 17.5 (7.8)

Location: Chapel Hill, NC



N: 12 asthmatics
Min: 1.4


Statistical Analyses: Linear Mixed
Max: 45.6


Model




Monitoring Stations: 1


Covariates: temperature, humidity,
Copollutant: PM2.6, PM102 E


pressure


Dose-response Investigated? No
Co-pollutant Correlation


PM2.6 - 0.90*


Statistical Package: SAS
PM10-2.6 - 0.73*


Lags Considered: 1 day
*p < 0.01

'All units expressed in/yglm3 unless otherwise specified.
July 2009
E-17
DRAFT-DO NOT CITE OR QUOTE

-------
Table E-2. Short-term exposure - cardiovascular morbidity studies: PM102.5.
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Chuang et al. (2007,
091063)
Period of Study: Nov 2002 - Mar 2003
Location: Taipei, Taiwan
Outcome: Heart Rate Variability
Age Groups: 52-76 yrs
Study Design: panel
N: 10 CHD & 16 Hypertensive Patients
Statistical Analyses: Linear Mixed
Effects Model
Covariates: age, sex, BMI, time of day,
temperature, humidity, pressure, HRV
Dose-response Investigated? No
Statistical Package: S-PLUS
Lags Considered: 1-4h ma
Pollutant: PM10 2 E
Averaging Time: 1h
among CHD, among hypertensive
Mean (SD): 16.4 (10.7), 14.0(11.1)
IQR: 14.8,11.9
Min: 0.7, 0.3
Max: 59.6, 66.5
Monitoring Stations: 1 personal
monitor each
Copollutant: PMi 02B, PMoti.o
Co-pollutant Correlation
NR
PM Increment: Interquartile range
Percent Change (Lower CI, Upper CI):
Cardiac Patients- SDNN
1h moving
2h moving
3h moving
4h moving
:-1.73 (-3.53, 0.08)
:-1.97 (-4.43, 0.49)
:-1.70 (-4.39, 0.89)
:-1.75 (-5.42,1.92)
Cardiac Patients- r-MSSD
1h moving
2h moving
3h moving
4h moving
: -4.39 (-9.54, 0.03)
:-4.36 (-8.99, 0.27)
:-4.20 (-9.02, 0.61)
: -2.70 (-9.24, 3.84)
Cardiac Patients- LF
1h moving
2h moving
3h moving
4h moving
:-1.85 (-4.33, 0.62)
:-3.87 (-8.22, 0.47
:-2.98 (-6.65, 0.69)
:-3.11 (-8.22,1.99)
Cardiac Patients- HF
1h moving
2h moving
3h moving
4h moving
¦4.46 (-9.23, 0.32)
¦4.41 (-9.55, 0.72)
¦3.80 (-9.12,1.53)
¦3.39 (-10.62, 3.84)
Cardiac Patients- LF: HF ratio
1h moving: 8.45 (-3.48, 20.38)
2h moving: 1.66 (-15.22,18.55)
3h moving: 11.69 (-7.27, 30.64)
4h moving: 8.18 (-17.22, 33.57)
Hypertensive Patients- SDNN
1h moving: -2.64 (-3.93, 0.55
2h moving: -3.51 (-7.87, 0.85)
3h moving: -2.74 (-6.22, 0.74)
4h moving: -2.49 (-6.13,1.15)
Hypertensive Patients- r-MSSD
1h moving: -2.53 (-5.10, 0.04)
2h moving: -5.42 (-10.92, 0.09)
3h moving:-3.15 (-6.32, 0.03)
4h moving: -4.23 (-8.88, 0.42)
Hypertensive Patients- LF
1h moving: -4.38 (-8.78, 0.03)
2h moving: -5.23 (-10.95, 0.05)
3h moving: -3.34 (-1.72, 0.04)
4h moving: -2.96 (-6.63, 0.71)
Hypertensive Patients- HF
1h moving: -4.92 (-9.94, 0.10)
2h moving: -6.07 (-12.28, 0.13)
3h moving: -1.94 (-5.44,1.55)
4h moving: -2.78 (-6.78,1.21)
Hypertensive Patients- LF: HF ratio
1 h moving: 5.94 (-3.27,15.15)
2h moving: 10.70 (-2.19, 23.59)
3h moving: -1.51 (-17.02,14.00)
4h moving: 3.41 (-16.91,23.74)
*p< 0.05
July 2009
E-18
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Reference: Ebelt et al. (2005, 056907)
Period of Study: Summer of 1998
Location: Vancouver, Canada
Design & Methods
Outcome: CVD
Age Groups: range from 54-86 yrs
mean age - 74 years
Study Design: extended analysis of a
repeated-measures panel study
N: 16 persons with COPD
Statistical Analyses:
Earlier analysis expanded by developing
mixed-effect regression models and by
evaluating additional exposure indicators
Dose-response Investigated? No
Statistical Package: SAS V8
Concentrations'
Pollutant: PM10 2 E
Averaging Time: 24 h
Mean (SD):
Ambient PMio-2.5: 5.6 (3.0)
Exposure to ambient PMio-2.5: 2.4 (1.7)
Range (Min, Max): Ambient PM102E: (¦
1.2-11.9)
Exposure to ambient PMkh.b: (-0.4-7.2)
Monitoring Stations: 5
Copollutant (correlation):
Ambient concentrations and exposure to
ambient PM were highly correlated for
each respective metric: r a 0.71
Effect Estimates (95% CI)
Note: Total personal fine particle
exposure (T) were dominated by
exposures to non ambient particles which
were not correlated with ambient fine
particle exposure (A) or ambient
concentrations (C). Results for each of
these metrics are listed.
PM Increment:
Increment: C10-26: IQR - 4.5//gfm3
SBP (mm Hg): -2.12 (-5.07-0.82)
DBP (mm Hg):-0.92 (-3.37-0.36)
Ln-SVE (bph): 0.06 (-0.24-0.36)
HR (bpm): 1.09 (-0.69-2.86)
SDNN(ms): 2.64 (-2.85-8.13)
R-MSSD (ms): -0.33 (-4.49-3.82)
Increment: A10"2 6: IQR -2.4//gfm3
SBP (mm Hg):-2.55 (-6.15-1.05)
DBP (mm Hg):-0.75 (-3.50-2.01)
Ln-SVE (bph): 0.26 (-0.07-0.58)
HR (bpm): 1.04 (-0.95-3.03)
SDNN (ms): 0.68 (-3.07-4.42)
R-MSSD (ms): 1.10 (-3.08-5.28)
Reference: Lipsett et al. (2006, 088753) Outcome: HRV parameters, specifically
SDNN, SDANN, r-MSSD, LF, HF, total
Period of Study: February-May 2000
Location: Coachella Valley, CA
power, triangular index (TRII).
Study Design: Panel study
N: 19 non-smoking adults with coronary
artery disease
Statistical Analysis: Mixed linear
regression models with random effects
parameters
Pollutant: PMio-2.5
Averaging Time: 2 h
Monitoring Stations: 2
Copollutant: 03
PM Increment: SE*1000
Effect Estimate (change in HRV per
unit increase in PM concentration):
SDNN:-0.72 msec (SE - 0.296)
Notes: PM10-2.E calculated by subtracting
PM2.6 concentration from PM10
concentration. Weekly ambulatory 24 h
ECG recordings (once per week for up to
12 weeks), using Holter monitors, were
made. Subjects' residences were within 5
miles of one of two PM monitoring sites.
Regressed HRV parameters against 18:
00-20: 00 mean particulate pollution
Reference: Metzgeret al. (2007,
092856)
Period of Study: August 1998-
December 2002
Location: Atlanta, GA
Outcome: Days with any event recorded
by the ICD, days with ICD
shocksfdefibrillation and days with either
cardiac pacing or defibrillation
Study Design: Repeated measures
N: 884 subjects between 1993 and 2002
Statistical Analysis: Logistic regression
with GEE to account for residual
autocorrelation within subjects
Pollutant: PM10 2.5 (n/cm3)
Averaging Time: 24-hs
Mean (SD): 9.6 (5.4)
Median: 8.7
Copollutant: O3, NO2, CO, SO2,
oxygenated hydrocarbons
PM Increment: OR (95% CI): OR - 1.03
(95% CI: 1.00,1.07)
Reference: Pekkanen et al. (2002,
035050)
Period of Study: Winter 1998 to 1999
Location: Helsinki, Finland
Outcome: ST Segment Depression
(> 0.1 mV)
Study Design: Panel of ULTRA Study
participants
N: 45 subjects, 342 biweekly submaximal
exercise tests, 72 exercise induced ST
Segment Depressions
Statistical Analysis: Logistic regression
I GAM
Pollutant: PM10 2.5 (n/cm3)
Averaging Time: 24 h
Median: 4.8
IQR: 5.5
Monitoring Stations: 1
Copollutant: NO2, CO, PM2.6, PM1, ACP,
ultrafine
PM Increment: IQR
Effect Estimate(s): PM10 2.5: OR - 1.99
(0.70,5.67), lag 2
Notes: The effect was strongest for ACP
and PM2.E, which in two pollutant models
appeared independent. Increases in NO2
and CO were also associated with
increased risk of ST segment depression,
but not with coarse particles.
Reference: Timonen et al. (2006,
088747)
Period of Study: 1998-1999
Location: Amsterdam, Netherlands
Erfurt, Germany
Helsinki, Finland
Outcome: HRV measurements: [LF, HF,
LFHFR, NN interval, SDNN, r-MSSD]
Study Design: Panel study
N: 131 elderly subjects with stable
coronary heart disease
Statistical Analysis: Linear mixed
models
Pollutant: PMio-2.5
Means: Amsterdam: 15.3
Erfurt: 3.7
Helsinki: 6.7
Copollutant: NO2, CO
PM Increment: 10/yglm3
Effect Estimate: SDNN
0.69ms (95% CI:-1.24, 2.63)
HF: 2.9% (95% CI:-7.3,13.1)
LFHFR:-3.3 (95% CI:-12.7, 6.1)
Notes: Followed for 6 months with
biweekly clinic visits
2 day lag. ULTRA Study
July 2009
E-19
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Yeatts et al. (2007, 091266) Outcome: Heart Rate Variability
Period of Study: 12 wk period bit Sept Age Groups: 21 -50 yrs
2003 -July 2004
Study Design: panel
Location: Chapel Hill, NC
N: 12 asthmatics
Statistical Analyses: Linear Mixed
Model
Covariates: temperature, humidity,
pressure
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 1 day
Pollutant: PM10 2 E
Averaging Time: 24h
Mean (SD): 5.3 (2.8)
Min: 0
Max: 14.6
Monitoring Stations: 1
Copollutant: PM2.6, PMio
Co-pollutant Correlation
PM2.6 - 0.46*
PM10 - NR
*p<0.01
PM Increment: 1 //gf'nr1. Beta, SE
(Lower CI, Upper CI), p-value
HRV
Max Heart Rate: -1.95, 0.88 (-3.67, ¦
0.231,0.03
ASDNN5:-0.77, 0.37 (-1.580,-0.04),
0.05
SDANN5: -3.76,1.53 (-6.76,-0.76), 0.02
SDNN24HR(mesc): -3.36,1.38 (-6.06, ¦
0.651,0.02
rMSSD: -0.75, 0.53 (-1.79, 0.28), 0.16
pNN50_24hour: -0.50, 0.27 (-1.03,
0.031,0.07
pNN50_7min: -1.88, 0.55 (-2.95, -0.81),
0.07
Low-frequency power: -0.19, 0.42 (-1.01,
0.631,0.65
Percent low frequency: 0.57,1.08 (-1.55,
2.691,0.60
High-frequency power: -0.46, 0.17 (¦
0.79,-0.14), 0.01
Percent high frequency: -2.14, 0.94 (¦
3.98,-0.30), 0.03
Blood Lipids
Triglycerides: 4.78, 2.02 (0.81, 8.74),
0.02
VLDL: 1.15, 0.44(0.29,2.02), 0.01
Total cholesterol: 0.78, 0.54 (-0.28,
1.841,0.15
Hematologic Factors
Circulating eosinophils: 0.16, 0.06 (0.04,
0.281,0.01
Platelets:-1.71,1.11 (-3.89,0.47), 0.13
Circulating Proteins
Plasminogen: -0.01, 0.01 (-0.02, 0.00),
0.08
Fibrenogen: -0.04, 0.02 (-0.08, 0.00),
0.07
Von Willibrand factor: -1.23, 0.66 (-2.53,
0.06),0.07
Factor VII: -0.90, 0.85 (-2.58, 0.77), 0.29
'All units expressed in/yglm3 unless otherwise specified.
Table E-3. Short-term exposure - cardiovascular morbidity studies: PM2.5 (including PM
components/sources).
Concentrations'
Reference
Design & Methods
Effect Estimates (95% CI)
Reference: Adar et al. (2007, 001458)
Period of Study: Mar-Jun 2002
Location: St. Louis, Missouri
Outcome: Heart rate variability: heart
rate, standard deviation of all normal-to-
normal intervals (SDNN), square root of
the mean squared difference between
adjacent normal-to-normal intervals
(rMSSD), percentage of adjacent normal-
to-normal intervals that differed by more
than 50 ms (pNN50), high frequency
power (HF
in the range of 0.15-0.4Hz), low frequency
power (LF, in the range of 0.04-0.15Hz),
and the ratio of LF/HF
Age Groups: a 60 yrs
Study Design: Panel (4 planned re|
measures surrounding bus trips with a
total of 158 person-trips
35 participating in all 4 trips)
Pollutant: PM2.6 (|Jg|m3)
Averaging Time: Measurements
collected over 48 h period surrounding the
bus trip (during which health endpoints
were measured) used to calculate 5-, 30-,
60-minute, 4-h, 24-h moving averages
Median (IQR): All: 7.7 (6.8)
Facility: 6.8 (5.1)
Bus: 17.2(10.3)
Activity: 8.2 (16.1)
Lunch: 11.2(5.9)
Monitoring Stations: 2 portable carts
Copollutant: PM2.6
BC
Fine particle counts
coarse particle counts
Correlation notes: 24-h mean PM2.6, BC,
_and_^me_£article_count_concentratmn^^^
PM Increment: IQR
Effect Estimate [Lower CI, Upper CI]: %
change (95%CI) in HRV per IQR in the 24-
h moving avg of the microenvironmental
pollutant (IQr - 4.5 |Jg|m3)
Single-pollutant models: SDNN: -5.5
(-6.3, -4.8)
rMSSD: -9.1 (-9.8,-8.4)
pNN50 + 1: -12.2 (-13.3,-11.1). LF: -
10.8 (-12.3,-9.3)
HF: -15.1 (-16.7,-13.7)
LF/HF: 5.1 (3.9, 6.4)
H: 1.0 (0.9, 1.2)
Two-pollutant models (with particle
number count coarse): SDNN: -5.7 (-6.5,
¦4.9)
rMSSD: -9.4(-10.1,-8.6)
pNN50 + 1: -13.1 (-14.3, -11.9). LF: ¦
July 2009
E-20
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
N: 44 participants
Statistical Analyses: Generalized
additive models
Covariates: Subject, weekday, time,
apparent temperature, trip type, activity,
medications, and autoregressive terms
Season: Limited data collection period
Dose-response Investigated? No
Statistical Package: SAS v8.02, R
V2.0.1
ranged from 0.80 to 0.98
r - 0.76 to 0.97 when limited to time
spent on the bus
r - 0.55 to 0.86 when comparing bus
concentrations to 24-h moving averages
r - -0.003 to 0.51 when comparing 5-min
averages and 24-h moving averages
Poor correlations found between coarse
particle count concentrations and all fine
particulate measures during all times
periods
10.7(-12.4,-9.1)
HF: -14.9(-16.5, -13.3); LF/HF: 4.9 (3.6,
6.2)' H: 0.9 (0.7,1.1)
Independent short- and medium-term
associations with HRV across all time
periods
% change per IQR (95%CI)
IQR 5-min means - 6.8 |Jg|m3 and 23: 55-
h means - 4.2 |Jg|m3
SDNN: 5-min mean: -0.5 (-0.8, -0.1)
23: 55-h mean: -4.6 (-5.3, -4.0)
rMSSD: 5-min mean: -0.9 (-1.3, -0.5)
23: 55-h mean: -7.5 (-8.1 to -6.8)
pNN50 + 1
5-min mean: -1.1 (-1.7 to -0.5)
23: 55-h mean: -9.9 (-10.9 to -8.9). LF
5-min mean: 0.4 (-0.5,1.2)
23: 55-h mean: -10.0 (-11.4 to -8.6)
HF
5-min mean: -1.5 (-2.3 to -0.6)
23: 55-h mean: -12.9 (-14.2 to-11.5)
LF/HF
5-min mean: 1.9 (1.3, 2.4)
23: 55-h mean: 3.2(2.1,4.3)
H: 5-min mean: 0.1 (0.1, 0.2)
23: 55-h mean: 0.8 (0.7, 0.9)
Independent associations of short-term
averages (5-min means) of PM with HRV
by bus and nonbus periods
IQR for bus - 10 |jg/m3) and
nonbus - 5.6 |jg/m3)
% change (95%CI)
p-value of interaction
SDNN
Bus: -5.0 (-6.3 to -3.7)
Nonbus: -0.5 (-0.9 to -0.2)
p-value for interaction: < 0.0001. rMSSD
Bus: -4.8 (-6.2 to -3.5)
Nonbus: -0.7 (-1.1 to -0.4. p-value for
interaction: < 0.0001
pNN50 + 1
Bus: -6.3 (-8.4 to -4.2)
Nonbus: -0.8 (-1.4 to -0.3)
p-value for interaction: <0.0001
LF: Bus: -7.0 (-9.8 to -4.1) Nonbus: 0.6
(-0.1,1.4)
p-value for interaction: < 0.0001. HF:
Bus: -10.7 (-13.5 to -7.9)' Nonbus: -0.7
(-1.5, 0.04) p-value for interaction:
< 0.0001. LF/HF: Bus: 3.9(1.7,6.0)
July 2009
E-21
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Adar et al. (2007, 001458)
Period of Study: March-Jun 2002
Location: St. Louis, Missouri
Outcome: Heart rate variability: heart
rate, standard deviation of all normal-to-
normal intervals (SDNN), square root of
the mean squared difference between
adjacent normal-to-normal intervals
(rMSSD), percentage of adjacent normal-
to-normal intervals that differed by more
than 50 ms (pNN50), high frequency
power (HF
in the range of 0.15-0.4Hz), low frequency
power (LF, in the range of 0.04-0.15Hz),
and the ratio of LF/HF
Age Groups: a 60 yrs
Study Design: Panel (4 planned repeated
measures with a total of 158 person-trips
35 participating in all 4 trips)
N: 44 participants
Statistical Analyses: Generalized
additive models
Covariates: Subject, weekday, time,
apparent temperature, trip type, activity,
medications, and autoregressive terms
Season: Limited data collection period
Dose-response Investigated? No
Statistical Package: SAS v8.02, R
V2.0.1
Pollutant: BC (ngfm3)
Averaging Time: Measurements
collected over 48 h period surrounding the
bus trip (during which health endpoints
were measured) used to calculate 5-, 30-,
60-minute, 4-h, 24-h moving averages
Median (IQR): All: 330 (337)
Facility: 285 (270)
Bus: 2911 (2464)
Activity: 482 (1168)
Lunch: 434 (276)
Monitoring Stations: 2 portable carts
Copollutant: PM2.6
BC
Fine particle counts
Coarse particle counts
Correlation notes: 24-h mean PM2.6, BC,
and fine particle count concentrations
ranged from 0.80 to 0.98
r - 0.76 to 0.97 when limited to time
spent on the bus
r - 0.55 to 0.86 when comparing bus
concentrations to 24-h moving averages
r - -0.003 to 0.51 when comparing 5-min
averages and 24-h moving averages
Poor correlations found between coarse
particle count concentrations and all fine
particulate measures during all times
periods
Nonbus: 1.4 (0.8,1.9)
p-value for interaction: 0.39. H: Bus: 0.7
(0.5,1.0)
Nonbus:-0.01 (-0.08,0.1)
p-value for interaction: <0.0001
Note: Exposure to health associations by
all lag periods presented in Figure 2
(magnitude of associations increased with
averaging period, with the largest
associations consistently found for 24-h
moving averages)
PM Increment: IQR
Effect Estimate [Lower CI, Upper CI]: %
change (95%CI) in HRV per IQR in the 24-
h moving avg of the microenvironmental
pollutant (IQr - 459 ng/m3)
Single-pollutant models
SDNN:-5.3 (-6.5 to-4.1)
rMSSD:-10.7 (-11.9 to-9.5)
pNN50 + 1:-13.2 (-15.0 to-11.4)
LF:-11.3 (-13.7 to-8.8)
HF:-18.8 (-21.1 to-16.5)
LF/HF: 9.3(7.2,11.4)
H: 1.0 (0.8, 1.3)
Independent short- and medium-term
associations with HRV across all time
periods
% change per IQR (95%CI)
IQR 5-min means - 337 ng/m3 and 23:
55-h means - 490 ng/m3)
SDNN: 5-min mean: -0.3 (-0.5 to -0.1)
23: 55-h mean: -4.7 (-5.9 to -3.5)
rMSSD: 5-min mean: -0.3 (-0.5 to -0.1)
23: 55-h mean: -9.3 (-10.5 to -8.1)
pNN50 + 1: 5-min mean: -0.3 (-0.6 to ¦
0.1)
23: 55-h mean: -10.5 (-12.3 to -8.7)
LF: 5-min mean: -0.5 (-0.9 to-0.1)
23: 55-h mean: -9.8 (-12.4 to -7.2)
HF: 5-min mean: -0.9 (-1.2 to-0.5)
23: 55-h mean: -15.4 (-17.8 to-12.9)
LF/HF: 5-min mean: 0.3 (0.1, 0.6)
23: 55-h mean: 6.5 (4.5, 8.6)
H: 5-min mean: 0.1 (0.1, 0.2)
23: 55-h mean: 0.4 (0.2, 0.7)
Independent associations of short-term
averages (5-min means) of PM with HRV
by bus and nonbus periods
IQR for bus - 2.6 |Jg/m3) and
nonbus - 0.27 |Jg/m3)
% change (95%CI)
p-value of interaction
SDNN: Bus: -4.6 (-6.1 to -3.0)' Nonbus: -
0.1 (-0.3,0.1)
p-value for interaction: <0.0001
DRAFT-DO NOT CITE OR QUOTE
July 2009
E-22

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Auchincloss et al. (2008,
156234)
Period of Study: Jul 2000-Aug 2002
Location: 6 US communities (Baltimore
City and Baltimore County, Maryland
Chicago, Illinois
Forsyth County, North Carolina
Los Angeles, California
Northern Manhattan and the Bronx, New
York
and St. Paul, Minnesota)
part of MESA (Multi-ethnic Study of
Atherosclerosis)
Outcome: Blood pressure: systolic (SBP),
diastolic (DBP), mean arterial (MAP), pulse
pressure (PP)
Avg of 2nd and 3,d BP measurement used
for analyses
Age Groups: 45-84 years
Study Design: Cross-sectional (Multi-
Ethnic Study of Atherosclerosis baseline
examination)
N: 5,112 persons (free of clinically
apparent cardiovascular disease)
Statistical Analyses: Linear regression
secondary analyses used log binomial
models to fit a binary hypertension
outcome
Covariates: Age, sex, race/ ethnicity, per
capita family income, education, BMI,
diabetes status, cigarette smoking status,
exposure to ETS, high alcohol use, physi-
cal activity, BP medication use,
meteorology variables, and copollutants
examined site as a potential confounder
and effect modifier
heterogeneity of effects also examined by
traffic-related exposures, age, sex, type 2
diabetes, hypertensive status, cigarette
use, by levels of SO2 and CO, and for
weather variables
Season: Adjusted for temperature and
barometric pressure to adjust for
seasonality (because seasons vary by the
study sites)
Also performed sensitivity analyses
adjusting for season to examine the
potential for residual confounding not
accounted for by weather variables
Pollutant: PM2.6
Averaging Time: 5 exposure metrics
constructed: prior day, avg of prior 2
days, prior 7 days, prior 30 days, and prior
60 days
Mean (SD): Prior day: 17.0(10.5)
Prior 2 days: 16.8 (9.3)
Prior 7 days: 17.0 (6.9)
Prior 30 days: 16.8 (5.0)
Prior 60 days: 16.7 (4.4)
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: Used monitor
nearest the participant's residence to
calculate exposure metrics
Copollutant: SO2
NO?
CO
Traffic-related exposures (straight-line
distance to a highway
total road length around a residence)
Correlations with PM2.5 averaged over
prior 30 days:
O3
Cool: r - -0.67
Moderate: r - -0.30
Warm: r - 0.23
CO
Cool: r - 0.20
Moderate: r - 0.71
Warm: r - 0.23
SO2
Cool: r - 0.36
Moderate: r - -0.17
rMSSD: Bus: -2.6 (-4.2 to -0.9): Nonbus: ¦
0.3 (-0.5 to-0.1)
p-value for interaction: 0.64
pNN50 + 1: Bus: -2.0 (-4.5, 0.5): Nonbus:
¦0.5 (-0.8 to-0.1)
p-value for interaction: 0.34
LF: Bus: -6.0 (-9.3 to -2.5): Nonbus: -0.2
(-0.7, 0.3)
p-value for interaction: 0.028
HF: Bus: -5.8 (-9.1 to -2.3)
Nonbus: -0.9 (-1.4 to -0.4)
p-value for interaction: 0.50
LF/HF: Bus: -0.8 (-3.1,1.7)
Nonbus: 0.8(0.5,1.1)
p-value for interaction: <0.0001
H: Bus:-0.5 (-0.8 to-0.2)
Nonbus: 0.3(0.26,0.34)
p-value for interaction: <0.0001
Note: Exposure to health associations by
all lag periods presented in Figure 2
(magnitude of associations increased with
averaging period, with the largest
associations consistently found for 24-h
moving averages)
PM Increment: 10/yg/m3 (approx.
equivalent to difference between 90th and
10th percentile for prior 30 day mean)
Effect Estimate [Lower CI, Upper CI]:
Adjusted mean difference (95% CI) in PP
and SBP (mmHg) per 10 /yglm3 increase in
PM2.6 (averaged for the prior 30 days)
Pulse Pressure
(PM2.6 averaged for prior 30 days)
Adjustment variables:
Person-level Covariates: 1.04 (0.25,1.84),
p - 0.010
Person-level cov., weather: 1.12 (0.28,
1.97), p - 0.009
Person-level cov., weather, gaseous
copollutants: 2.66 (1.61, 3.71), p -
0.000
Person-level cov., study site: 0.93 (-0.04,
1.90), p - 0.060
Person-level cov., study site, weather:
1.11 (0.01, 2.22), p - 0.049
Person-level cov., study site, weather,
gaseous copollutants: 1.34 (0.10, 2.59),
p - 0.035
Systolic Blood Pressure
Adjustment variables:
Person-level Covariates: 0.66 (-0.41,1.74)
,p - 0.226
Person-level cov., weather: 0.99 (-0.15,
2.13), p - 0.089
Person-level cov., weather, gaseous
July 2009
E-23
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Dose-response Investigated? Assessed
nonlinear relationships-no evidence of
strong threshold/nonlinear effects for
PM2.6
Statistical Package: NR
Warm: r - -0.11
NO?
Cool: r - 0.55
Moderate: r - 0.66
Warm: 0.32
copollutants: 2.8 (1.38, 4.22), p - 0.000
Person-level cov., study site: 0.86 (-0.45,
2.17), p - 0.200
Person-level cov., study site, weather:
1.32 (-0.18, 2.82), p - 0.085
Person-level cov., study site, weather,
gaseous copollutants: 1.52 (-0.16, 3.21),
p - 0.077
Additional results: Associations became
stronger with longer averaging periods up
to 30 days. For example: Adjusted
(personal covariates and weather) mean
differences in PP: Prior day: -0.38 (-0.76,
0.00)
Prior 2 days:-0.22 (-0.65, 0.21)
Prior 7 days: 0.52 (-0.08,1.11)
Prior 30 days: 1.12 (0.28,1.97)
Prior 60 days: 1.08 (0.11,2.05)
(Pattern held for additional adjustments
and for SBP results
therefore, only results for 30-day mean
differences were presented)
Additional results (not presented):
None of DBP results were statistically
significant
results for MAP were similar to SBP,
though weaker and generally not
significant
Effect modification: associations
between PM2.6 and BP were stronger for
persons taking medications, with
hypertension, during warmer weather, in
the presence of high NO2, residing ~ 300m
from a highway, and surrounded by a high
density of roads (Fig 1)
associations were not modified by age,
sex, diabetes, cigarette smoking, study
site, high levels of CO or SO2, season , nor
residence ~ 400m fro a highway
Note: supplementary material available
on-line shows results for DBP and MAP,
among others
Reference: Baccarelli et al. (2009,
188183)
Period of Study: Nov 2000 - Jun 2005
Location: Boston, Mass
Outcome: Heart rate variability
Age Groups: elderly
Study Design: Panel
N:549 men
Statistical Analyses: Mixed-effects
model
Covariates: age, past/current CHD, BMI,
mean arterial pressure, fasting blood
glucose, smoking, alcohol consumption,
use of beta-blockers, CA channel blockers,
angiotensin-converting enzyme inhibitors,
room temperature, season, apparent
temperature
Season:no
Dose-response Investigated? No
Pollutant: PM2.6
Averaging Time: 48 h moving average
Geometric Mean (95%CI):
All Visits: 10.5(10.0,10.9)
Visits wI Genotype Data: 10.4 (9.9,11.0)
Visits w/o Genotype Data: 10.5 (9.8,
11.4)
Monitoring Stations: 1
Copollutant: NR
Correlation: n/a
PM Increment: 10/yg/m3
Percent Change [Lower CI, Upper CI],
P:
All Subjects wf Genotype Data
SDNN:-6.0 (-13.5, 2.0), 0.14
HF: -17.1 (-32.3,1.6), 0.07
LF:-8.2 (-22.1, 8.2), 0.31
All Subjects
SDNN: -7.1 (-13.2,-0.6), 0.03
HF: -18.7 (-31.1,-4.0), 0.01
LF:-11.8 (-23.2,-1.3), 0.08
Statistical Package: SAS
July 2009
E-24
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Barclay et al. (2007,192229) Outcome: Haematological outcomes,
Heart Rhythm outcomes, & Heart Rate
Period of Study: Jan 2003 - May 2005 Variability outcomes
Location: Aberdeen, Scotland
Age Groups: 70.4 (8.9)
Study Design: panel
N: 132 patients wf chronic heart failure
Statistical Analyses: Linear & Mixed
Effects Regression Model
Covariates: age, temperature, humidity,
pressure
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: lags 0-2d
Pollutant: PM2.B
Averaging Time: daily
Mean: 7.454
Min: 1.092
Max: 21.97
Monitoring Stations: 0
Copollutant: PMio, PNC, NO2
Co-pollutant Correlation
NO2 city: 0.164
NO city: 0.048
PM10 city: 0.476*
NO2 personal: 0.169
PNC DE0M: 0.115
PM2.6 traffic: 0.522*
PNC total: 0.367*
PNC traffic: 0.234
"correlations based on 3-day average
concentrations
Notes: PM2.6 values model predicted
PM Increment: NR
Beta (Lower CI, Upper CI):
Haemoglobin: -0.509 (-1.560, 0.542)
Mean corpuscular haemoglobin: 0.188 (¦
0.481,0.857)
Platelets: 3.022 (0.403, 5.642)
Haematocrit: -0.813 (-1.892, 0.267)
White blood cells: -1.652 (-4.727,1.424)
C reactive protein: 4.924 (-13.022,
22.869)
IL-6:-5.980 (-23.649,11.690)
von Willebrand factor: 1.363 (-6.561,
9.287)
E-selectin: 2.136 (-2.946, 7.217)
Fibrinogen: -5.579 (-10.403, -0.755)*
Factor VII: 3.747 (-1.959, 9.452)
d-dimer: 5.211 (-2.974,13.397)
All arrhythmias: -7.082 (-28.789,14.626)
Ventricular ectopic beats: -12.203 (¦
39.021,14.615)
Ventricular couplets: -1.255 (-25.678,
23.168)
Ventricular runs: -2.548 (-17.448,12.351)
Supraventricular ectopic beats: 4.898 (¦
19.772,29.568)
Supraventricular couplets: 6.138 (¦
16.242,28.518)
Supraventricular runs: -0.545 (-17.577,
16.487)
Average HR: 0.617 (-0.782, 2.016)
24h SDNN: 3.645 (-0.227, 7.517)
24h SDANN: 4.437 (0.030, 8.844)*
24h RMSSD: 0.617 (-0.782, 2.016)
24h PNN 50%: 11.247 (-6.228, 28.722)
24h LF power: 4.439 (-6.823,15.701)
24h LF normalized: -5.659 (-11.815,
0.497)
24h HF power: 3.800 (-10.863,18.464)
24hHF normalized:-6.597 (-13.724,
0.531)
24h LF/HF ratio: 1.033 (-8.355,10.414)
*p< 0.05
Notes: Estimates also available for PM2.6
traffic
LF - low frequency
HF- high frequency
Reference: Briet et al. (2007, 093049)
Period of Study: NR
Location: Paris, France
Outcome: Endothelial Function
Age Groups: 20-40 yrs
Study Design: panel
N: 40 white male nonsmokers
Statistical Analyses: Multiple Robust
Regrssion
Covariates: R53R/R53H genotype, diet,
subject factor, visit, temperature
Dose-response Investigated? No
Statistical Package: NCSS
Lags Considered: 0-5d
Pollutant: PM2.6
Averaging Time: 24h
5 d Mean (SD): 28 (6)
Monitoring Stations: NR
Co-pollutant: PM10, SO2, NO, NO2, CO
Co-pollutant Correlation
n/a
PM Increment: 1 SD
Beta (Lower CI, Upper CI), P, R2:
Flow-mediated brachial artery dilation:
¦0.32 (-1.10, 0.46), NS, 0.04
Reactive hyperemia:
15.68(7.11,23.30), <0.0001,0.24
Changes in Endothelial function bit visits:
1.98(0.67, 3.259), 0.004,0.44
July 2009
E-25
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Cardenas et al. (2008,
191900)
Period of Study: NR
Location: Mexico City, Mexico
Outcome: Heart Rate Variability
Age Groups: 20-40 yrs
Study Design: panel
N: 54 subjects
Statistical Analyses: Linear GEE models
Covariates: localization, supine position,
gender, age, humidity, heart rate,
orthostatic position, head-up tilt test
result
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: NR
Pollutant: PM2.6
Averaging Time: NR
25", 50", 75" percentile:
Indoor: 14.8,28.3, 47.9
Outdoor: 6.4,10.8,16.8
Monitoring Stations: NR
Co-pollutant: NR
Co-pollutant Correlation
n/a
PM Increment: NR
Mean Difference (Lower CI, Upper CI),
lag:
Ln low frequency
Indoors:-0.028 (-0.0423,-0.0138)
Outdoors: -0.194 (-0.4509, 0.0627)
Ln high frequency
Indoors:-0.019 (-0.0338,-0.0044)
Outdoors:-0.298 (-0.5553,-0.0401)
Ln LF/HF ratio
Indoors:-0.017 (-0.0330,-0.0007)
Outdoors: -0.278 (-0.5540, 0.0030)
Reference: Cavallari et al. (2007,
157425)
Period of Study: 1999-2006
Location: Massachusetts
Outcome: Heart Rate Variability
Age Groups: 22-63
Study Design: panel
N: 36 males
Statistical Analyses: Mixed Effects
Regression Model
Covariates: age, smoking, heart rate at
work
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: lags 0-14h
Pollutant: PM2.6
Averaging Time: hourly
Mean (SD): 1.12 (0.76)
Min: 0.12
Max: 3.99
Monitoring Stations: NR
Copollutant: NR
Co-pollutant Correlation
n/a
PM 1 mg Increment: Im3
Beta (Lower CI, Upper CI):
1.44 (¦
5.33 (¦
6.86 (¦
2.17 (¦
4.73 (¦
3.52 (¦
1.59 (
0.72 (¦
5.55 (¦
¦3.66
¦8.60
¦5.98
¦8.27
¦4.19
7.75, 4.87)
10.97,0.31)*
11.91,-1.81)*
9.33, 4.99)
11.99,2.53)
9.89, 2.84)
7.53, 4.35)
7.63, 6.20)
10.65,-0.45)*
(-8.85,1.53)
(-17.45, 0.24)*
(-14.67, 2.70)
(-17.00, 0.46)*
(-12.71,4.33)
Model 1
Laglh
Lag2h
Lag3h
Lag4h
Lag5h
Lag6h
Lag7h
Lag8h
Lag9h
Lag1Oh
Lag11h
Lag12h
Lag13h
Lag14h
Model 2
Laglh: 4.10 (-0.39, 8.60)*
Lag2h:-3.21, (-8.78, 2.37)
Lag3h: -6.45 (-11.59,-1.31)*
Lag4h: -0.01 (-6.96, 6.94)
Lag5h:-2.03 (-8.27, 4.22)
Lag6h:-1.99 (-8.46, 4.48)
Lag7h: -0.34 (-6.22, 5.54)
Lag8h: 0.72 (-6.35, 7.78)
Lag9h: -5.26 (-10.62, 0.11)*
Lag10h:-3.68 (-9.17,1.80)
Lag11h:-9.41 (-18.60,-0.23)*
Lag12h: -6.45 (-15.62, 2.72)
Lag13h: -7.33 (-16.55,1.89)
Lag14h: -4.75 (-13.81, 4.32)
*p< 0.05, *p<0.10
Notes: Model 1 adjusted for smoking
status & age only. Model 2 adjusted for
smoking status, age, & heart rate during
work.
July 2009
E-26
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Chahine et al. (2007,
156327)
Period of Study: Jan 2000 - Jun 2005
Location: Boston, MA
Outcome: Heart Rate Variability
Age Groups: mean 72.8(6.6) yrs
Study Design: panel
N: 539 white males
Statistical Analyses: Mixed Effects
Model
Covariates: age, BMI, mean arterial
pressure, fasting blood glucose, smoking,
alcohol consumption, use of beta-blockers,
calcium channel blockers, ACE inhibitors,
room temperature, season, outdoor
temperature
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-2d ma
Pollutant: PM2.6
Averaging Time: 1h
Mean (SD): 11.7 (7.8)
Monitoring Stations: 1
Copollutant: PMi.o
Co-pollutant Correlation
n/a
PM Increment: 10/yg/m3
Percent Change (Lower CI, Upper CI),
p-value:
Iog10 SDIMIM
Total:-6.8 (-12.9,-0.2), 0.0436
GSTM1 wildtype:-2.0 (-11.3, 8.3),
0.6908
GSTM1 null:-10.5 (-18.2,-2.2), 0.0150
HM0X-1 <25 repeats: 7.4 (-8.7, 26.2),
0.3891
HM0X-1 >25 repeats: -8.5 (-14.8,-1.8),
0.0137
loglOHF
Total:-17.3 (-30.0,-2.3), 0.0263
GSTM1 wildtype: -4.0 (-24.8, 22.6),
0.7442
GSTM1 null:-24.2 (-39.2,-5.5), 0.0139
HM0X-1 < 25 repeats: 8.9 (-27.1, 62.8),
0.6759
HM0X-1 >25 repeats: -20.1 (-32.9,-5.0),
0.0115
loglOLF
Total: -11.2 (-22.8, 2.2), 0.0986
GSTM1 wildtype:-0.6 (-19.0, 22.0),
0.9545
GSTM1 null:-17.0 (-31.0,-0.2), 0.0478
HM0X-1 <25 repeats: 14.0 (-18.6,
59.5), 0.4465
HM0X-1 >25 repeats: -14.0 (-25.7,-0.5),
0.0430
Reference: Chen and Schwartz (2008,
190106)
Period of Study: 1989-1991
Location: US
Outcome: White Blood Cell count
Age Groups: 20-89 yrs
Study Design: panel
N: 2,978 participants
Statistical Analyses: Mixed Effects
Models
Covariates: age, sex, race, SES, smoking,
alcohol consumption, MS abnormalities,
indoor air pollutants, exercise
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: NR
Pollutant: PM2.6
Averaging Time: 24h
Mean (SD): 36.8 (13.0) Median(range)
for
Qi: 23.1 (14.6-27.8)
02:31.2 (27.9-34.3)
03:38.8 (34.3-43.3)
Qt:53.7 (43.3-78.5)
Monitoring Stations: NR
Copollutant: NR
Co-pollutant Correlation
n/a
PM Increment: quartile, 1 yr avg (36.8
/yg/m3)
Average WBC count(SE) by PM
quartile:
Qi: 6760 (79)
02:6942 (99)
03:6895(84)
0*7109(61)
Beta(Lower CI, Upper CI), p-value:
Crude: 239 (58, 420), 0.01
Model 1:145 (10, 281), 0.035
Model 2:141 (6, 277), 0.041
Model 3:138 (2, 273), 0.046
Model 1: age, sex, race, SES, smoking,
alcohol consumption, MS abnormalities.
Model 2: Model 1 plus indoor air
pollutants, exercise. Model 3: clean areas
(Q1) vs. .other more polluted areas
July 2009
E-27
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Chuang et al. (2007, 091063)
Period of Study: Between Apr-Jim 2004
or 2005
Location: Taipei, Taiwan
Outcome: High-sensitivity C-reactive
protein (hs-CRP)
Fibrinogen, plasminogen activator
fibrinogen inhibitor-1 (PAI-1), tissue-type
plasminogen activator (tPA), 8-hydroxy-2'-
deoxyguanosine (8-OHdG), and log-
transformed HRV indices
(SDNN - standard deviation of NN
intervals, r-MSSD - square root of the
mean of the sum of the squares of
differences between adjacent NN
intervals, LF - low frequency [0.04-
0.15Hz], and HF - high frequency[0.15-
0.40Hz])
Age Groups: 18-25 yrs
Study Design: Panel (cross-sectional)
N: 76 students
Statistical Analyses: linear mixed-
effects models
Covariates: Age, sex, BMI, weekday,
temperature of previous day, relative
humidity
Season: Only 1 season of data collection
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PMio, nitrate, sulfate
Averaging Time: Hourly data used to
calculate averages over 1-3 day periods
Mean (SD): 1 -day avg: 31.8 (10.6)
2-day	avg: 36.4 (12.6)
3-day	avg: 36.5 (12.6)
Range (Min, Max): 1 -day avg: 16.2, 50.1
2-day	avg: 15.0, 53.4
3-day	avg: 12.7, 59.5
Monitoring Stations: 2 sites (each
pollutant measured at one site only)
Copollutant: PMio
Sulfate
Nitrate
0C
EC
NO?
CO
SO?
Os
PM2.5 Increment: IQR (1 -d avg: 20.4
2-day	avg: 25.2
3-day	avg: 20.0)
Effect Estimate [Lower CI, Upper CI]: %
change in health endpoint per increase in
IQR of PM2.E (1-3 day averaging period
single pollutant models)
hs-CRP: 1-d: 90.2 (-10.2,190.1)
2-d:	99.1 (-26.1, 224.3)
3-d:	100.4 (-2.9, 203.7)
8-OHdG: 1-d:-5.0 (-14.3, 4.4)
2-d:-5.5	(-15.6, 4.6)
3-d:	-5.6 (-13.8, 2.6)
PAI-1: 1-d: 20.4 (17.3, 33.5)
2-d:	16.2 (1.9, 30.5)
3-d:	20.0 (18.5, 31.5)
tPA: 1-d: 12.0 (-2.4, 26.3)
2-d: 12.0 (-2.9, 26.9);
3 d: 12.0 (-2.7, 26.6)
Fibrinogen: 1-d: 2.6 (-2.7, 7.8)
2-d:	1.5 (-4.1, 7.1);
3-d:	3.6 (-0.8, 8.1)
Heart Rate Variability
SDNN: 1-d: -4.0 (-6.1 to-1.9)
2-d:-2.5	(-4.6 to -0.4)
3-d:	-3.0 (-5.0 to -1.1)
r-MSSD: 1-d: -3.0 (-8.7, 2.7)
2-d:	-2.0 (-8.4, 4.4);
3-d:-3.6	(-8.8,1.6)
LF: 1-d:-3.1 (-6.1 to-0.1)
2-d:-3.2	(-4.6, 0.1);
3-d:-3.4	(-6.1 to -0.6)
HF: 1-d:-3.7 (-9.4, 2.1)
2-d:	-2.1 (-8.4, 4.3);
3-d:	-4.0 (-9.3,1.2)
July 2009
E-28
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Chuang et al. (2007, 091063)
Period of Study: Between Apr-Jim 2004
or 2005
Location: Taipei, Taiwan
Outcome: High-sensitivity C-reactive
protein (hs-CRP)
Fibrinogen, plasminogen activator
fibrinogen inhibitor-1 (PAI-1), tissue-type
plasminogen activator (tPA), 8-hydroxy-2'-
deoxyguanosine (8-OHdG), and log-
transformed HRV indices
(SDNN - standard deviation of NN
intervals, r-MSSD - square root of the
mean of the sum of the squares of
differences between adjacent NN
intervals, LF - low frequency [0.04-
0.15Hz], and HF - high frequency[0.15-
0.40Hz])
Age Groups: 18-25 yrs
Study Design: Panel (cross-sectional)
N: 76 students
Statistical Analyses: Linear mixed-
effects models
Covariates: Age, sex, BMI, weekday,
temperature of previous day, relative
humidity
Season: Only 1 season of data collection
Dose-response Investigated? No
Statistical Package: NR
Pollutant: Nitrate
Averaging Time: Hourly data used to
calculate averages over 1-3 day periods
Mean (SD): 1 -day avg: 4.5 (2.7)
2-day	avg: 4.7 (2.4)
3-day	avg: 4.4(2.2)
Range (Min, Max): 1-day avg: 0.7,10.6
2-day	avg: 0.7, 8.9
3-day	avg: 0.8, 7.5
Monitoring Stations: 2 sites (each
pollutant measured at one site only)
Copollutant: PMio
Sulfate
PM2.6
0C
EC
NO?
CO
SO?
0s
Nitrate Increment: IQR (1 -d avg: 2.5
2-day	avg: 4.0
3-day	avg: 3.4)
Effect Estimate [Lower CI, Upper CI]: %
change in health endpoint per increase in
IQR of nitrate (1-3 day averaging period
single pollutant models)
hs-CRP: 1-d:-2.1 (-21.9,17.8)
2-d:-11.6	(-58.6, 35.5)
3-d:-18.7	(-69.9, 32.5)
8-OHdG: 1-d: 9.0(4.0,14.1)
2-d:	15.1 (5.9,24.3)
3-d:	15.0 (4.9, 25.0)
PAI-1: 1-d: 4.0 (-2.5,10.4)
2-d:	11.6 (0.1, 23.1)
3-d:	16.9 (4.3, 29.4)
tPA: 1-d: 2.0 (-6.2,10.3)
2-d:	12.9 (-1.6, 27.5)
3-d:	10.0 (-5.8, 25.8)
Fibrinogen: 1-d: 1.6 (-1.3, 4.5)
2-d:	1.3 (-3.9, 6.5)
3-d:	1.0 (-4.6, 6.6)
Heart Rate Variability
SDNN: 1-d: -1.5 (-2.6 to -0.3)
2-d:-2.6	(-4.7 to -0.5)
3-d:-3.0	(-5.3 to -0.7)
r-MSSD: 1-d: -5.5 (-8.7 to-2.2)
2-d:-7.1	(-14.0 to-0.2)
3-d:-8.1	(-14.5 to-1.8)
LF: 1-d: -1.0 (-1.6 to-0.5)
2-d:-2.0	(-5.6,1.6)
3-d:-2.0	(-5.2,1.2)
HF: 1-d: -2.0 (-5.3,14[potential typo,
possibly 1.4])
2-d:	-4.9 (-10.9, 0.9)
3-d:-6.9	(-13.4 to-0.3)
July 2009
E-29
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Chuang et al. (2007, 091063)
Period of Study: Between Apr-Jim 2004
or 2005
Location: Taipei, Taiwan
Period of Study: NR
Location: Boston, MA
Outcome: High-sensitivity C-reactive
protein (hs-CRP)
Fibrinogen, plasminogen activator
fibrinogen inhibitor-1 (PAI-1), tissue-type
plasminogen activator (tPA), 8-hydroxy-2'-
deoxyguanosine (8-OHdG), and log-
transformed HRV indices
(SDNN - standard deviation of NN
intervals, r-MSSD - square root of the
mean of the sum of the squares of
differences between adjacent NN
intervals, LF - low frequency [0.04-
0.15Hz], and HF - high frequency[0.15-
0.40Hz])
Age Groups: 18-25 yrs
Study Design: Panel (cross-sectional)
N: 76 students
Statistical Analyses: Linear mixed-
effects models
Covariates: Age, sex, BMI, weekday,
temperature of previous day, relative
humidity
Season: Only 1 season of data collection
Dose-response Investigated? No
Statistical Package: NR
Age Groups: 43-75 yrs
Study Design: panel
N: 48 coronary artery disease patients
Statistical Analyses: Linear & Mixed
Logistic Regression models
Covariates: participant, day of week,
order of visit, visit date, hour of day,
hourly temperature
Dose-response Investigated? No
Statistical Package: R
Lags Considered: lags 1-72h
Pollutant: Sulfate
Averaging Time: Hourly data used to
calculate averages over 1-3 day periods
Mean (SD): 1 -day avg: 4.1 (3.6)
2-day	avg: 4.1 (3.7)
3-day	avg: 3.9 (3.5)
Range (Min, Max): 1-day avg: 0.4,10.9
2-day	avg: 0.4,11.9
3-day	avg: 0.4,11.5
Monitoring Stations: 2 sites (each
pollutant measured at one site only)
Copollutant: PMio
PM2.6
Nitrate
0C
EC
NO?
CO
SO?
0s
Pollutant: PM2.B
Averaging Time: hourly
25", 50", 75" percentile:
12h avg: 6.18,8.91,13.18
24h avg: 6.38,9.20,13.31
Max:
12h avg: 37.13
24h avg: 40.38
Monitoring Stations: 1
Co-pollutant: BC, CO, O3, NO2, SO2
Co-pollutant Correlation
BC: 0.56
0s: 0.20
NO2: 0.38
SO2: 0.25
Sulfate Increment: IQR (1 -d avg: 3.9
2-day	avg: 4.3
3-day	avg: 3.8)
Effect Estimate [Lower CI, Upper CI]: %
change in health endpoint per increase in
IQR of sulfate (1-3 day averaging period
single pollutant models)
hs-CRP: 1-d: 80.0(9.8,150.2)
2-d: 87.1 (14.9,159.4)
3 d: 71.1 (13.0,129.2)
8-OHdG: 1-d: 1.0(0.3,1.3)
2-d:	-0.4 (-5.4, 4.7)
3-d:-0.3	(-4.3, 3.7)
PAI-1: 1-d: 12.0(5.4,18.7)
2-d: 13.3 (6.6,19.9)
3 d: 11.2 (5.7,16.6)
tPA: 1-d: 2.0 (-4.6, 8.7)
2-d: 3.8 (-2.8,10.3)
3 d: 3.0 (-2.3, 8.2)
Fibrinogen: 1-d: 2.9 (0.2, 5.5)
2-d:	2.8 (0.1, 5.5)
3-d:	2.2 (0.4, 4.7)
Heart Rate Variability
SDNN: 1-d:-3.1 (-4.1 to-2.1)
2-d:	-4.1 (-5.2 to -3.1)
3-d:-2.0	(-2.9 to-1.2)
r-MSSD: 1-d: -5.0 (-8.0 to-2.0)
2-d:-6.0	(-8.9 to -2.9)
3-d:-5.7	(-8.2 to -3.2)
LF: 1-d:-3.4 (-4.9 to-1.8)
2-d:-3.0	(-4.5 to-1.5)
3-d:-3.0	(-4.3 to-1.7)
HF: 1-d: -3.5 (-6.5 to -0.4)
2-d:-3.9	(-7.0 to -0.8)
3-d:-3.0	(-5.5 to -0.5)
PM Increment: Interquartile Increase
Change (Lower CI, Upper CI):
12-hour mean
PM2.1,: -0.022 (-0.032,-0.012)
PM2.B+NO2:-0.023 (-0.034,-0.012)
PM2.B+ SO2:-0.009 (-0.02, 0.001)
PM2.B+ BC:-0.011 (-0.023, 0.001)
24-hour mean
PM2.B:-0.026 (-0.037,-0.015)
PM2.B+ NO2:-0.017 (-0.029, 0.004)
PM2.B+ SO2:-0.014 (-0.025,-0.002)
PM2.B+ BC: -0.012 (-0.026, 0.003)
Relative Risk (Lower CI, Upper CI):
12-hour mean
PM2.6: 1.02 (0.86,1.21)
PM2.6+ NO2: 0.99 (0.82,1.21)
PM2.B+ SO2: 0.87 (0.71,1.05)
PM2.B+ BC: 0.92(0.74,1.14)
24-hour mean
PM2.6: 1.22 (0.99,1.50)
PM2.B+NO2:1.00 (0.80,1.25)
PM2.B+ SO2:1.04 (0.83,1.30)
PM2.B+ BC: 0.87 (0.65, 1.17)
Mean (Lower CI, Upper CI):
12-hour mean
Myocardial Infarction: -0.042 (-0.057, ¦
0.026)
No Myocardial Infarction: -0.012 (-0.023,
0.00)
DRAFT-DO NOT CITE OR QUOTE
July 2009
E-30
Reference: Chuang et al. (2007, 098629) Outcome: ST Segment Depression

-------
Reference	Design & Methods	Concentrations'	Effect Estimates (95% CTT
p- for interaction: 0.002
Visit 1: -0.102 (-0.12,-0.085)
Visits 2-4: 0.006(-0.005, 0.017)
p- for interaction: < 0.001
Diabetic: -0.097 (-0.119,-0.074)
Non-diabetic: -0.009 (-0.019, 0.002)
p- for interaction: < 0.001
Diurnal daytime pattern: -0.032 (-0.043, ¦
0.021)
Diurnal nighttime pattern: -0.006 (-0.018,
0.006)
p- for interaction: < 0.001
24-hour mean
Myocardial Infarction: -0.027 (-0.043, ¦
0.012)
No Myocardial Infarction: -0.025 (-0.038,
0.011)
p- for interaction: 0.787
Visit 1:-0.127 (-0.148,-0.105)
Visits 2-4: 0.001 (-0.011,0.013)
p- for interaction: < 0.001
Diabetic:-0.118 (-0.144,-0.091)
Non-diabetic: -0.13 (-0.024, -0.002)
p- for interaction: < 0.001
Diurnal daytime pattern: -0.031 (-0.043, ¦
0.020)
Diurnal nighttime pattern: -0.018 (-0.030,
¦0.005)
p- for interaction: 0.233
Notes: The effects of PM on half-hour St
segment levels (figure 1)
PM Increment: Interquartile Range
(27.02 /yglm3)
Beta (SE), p-value:
Flow mediated vasodilation (%): -0.016
(0.0072) p-0.03
Heart Rate (beats/min): 0.081 (0.135)
p — 0.55
Diastolic blood pressure (mmHg): 0.088
(0.088) p-0.32
Systolic blood pressure (mmHg): -0.108
(0.006) p-0.48
Reference: Dales et al. (2007,155743)
Period of Study: NR
Location: Ottawa, Canada
Outcome: Vascular Reactivity
Age Groups: 18-50 yrs
Study Design: panel
N: 39 volunteers
Statistical Analyses: Mixed Effects
Model
Covariates: temperature, humidity, wind
speed, time of day testing was done, site
Dose-response Investigated? No
Statistical Package: S-PLUS
Lags Considered: NR
Pollutant: PM2.6
Averaging Time: 2h
Mean (SD):
Downtown: 40 (20)
Tunney's Pasture: 10 (10)
p-value 0.000
Monitoring Stations: NR
Copollutant: PMi.o
Co-pollutant Correlation
n/a
July 2009
E-31
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: de Hartog et al. (2009,
191904)
Period of Study: 1998-1999
Location: Amsterdam, Netherlands
Erfurt, Germany
and Helsinki, Finland
Outcome: Heart Rate Variability
Age Groups: 50 +
Study Design: panel
N: 122 coronary heart disease patients
Statistical Analyses: Linear Regression
Covariates: time trend, temperature,
humidity, pressure
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: lags 0-3d
Pollutant: PM2.6
Averaging Time: daily
p25, p50, p75. p95:
Amsterdam: 10.4,16.7, 23.9, 47.0
Erfurt: 10.8,16.3, 26.7, 62.3
Helsinki: 8.3,10.6,15.9, 25.8
Monitoring Stations: NR
Copollutant: PM <0.1, PMo.i-i.o, NO2, SO2
Co-pollutant Correlation
NR
Note: Correlations are provided for
source-specific PM 2.6 & elements
PM Increment: 1 /yglm3
Beta (Lower CI, Upper CI):
SDNN
Local traffic:-0.12 (-0.36, 0.12)
Long-range transport: -0.04 (-0.14, 0.06)
Oil combustion: -0.29 (-1.04, 0.45)
Industry: 0.03 (-0.12, 0.19)
Crustal: 0.11 (-0.35,0.56)
Salt:-0.19 (-1.92,1.55)
HF
Local traffic: 0.43 (-0.91,1.79)
Long-range transport: 0.19 (-0.38, 0.77)
Oil combustion: 1.05 (-2.70, 4.94)
Industry: 0.62 (-0.34,1.59)
Crustal: 1.57 (-1.28, 4.50)
Salt:-1.43 (-9.86, 7.78)
SDNN
ABS: -0.52 (-1.39, 0.31)
S:-0.51 (-1.36,0.33)
V:-0.66 (-1.73, 0.41)
Zn: 0.12( 0.55, 0.79)
Ca: 0.27 (-0.58,1.11)
CI: 0.14 (-0.39, 0.67)
Fe: 0.15(-1.00,1.30)
Cu:-0.08 (-0.74, 0.57)
SDNN
ABS: 2.91 (-2.54,8.67)
S: 0.25(-4.42, 5.14)
V: 0.73 (-4.74, 6.53)
Zn: 3.85( 0.26, 8.13)
Ca: 3.39(-1.80, 8.86)
CI: 1.13(-1.48, 3.81)
Fe: 6.69(0.11,13.69)
Cu: 3.00 (-0.85, 7.00)
Notes: Estimates provided are for all
subjects at lag 1, estimates are also
available at lags 0, 2, and 3, as well as for
subjects wfo beta-blockers at lags 0-3.
July 2009
E-32
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: DeMeo et al. (2004, 087346) Outcome: Oxygen Saturation
Age Groups: 60.4 to 89.2 years
Period of Study: July through August
1999
Location: Boston, MA
Pollutant: PM2.6
Averaging Time: 6 h, 12 h, 24 h, 48 h
Study Design: Cross-sectional study
N: 28 adult participants
Statistical Analyses: GLM, Natural
Spline Smoothing, Regression Analysis,
Random-effects model
Covariates: Mean temperature, Dew
point temperature, Barometric pressure,
Medication use
Season: Summer
Dose-response Investigated? No
Statistical Package:
S-PLUS, SAS
Lags Considered: Hourly lags between 2
and 7 h
PM Increment: IQR (13.42/yg/m3)
increase
6 h: 13.42/yg/m3
12 h: 10.81 /yglm3
24 h: 10.26/yg/m3
48: 10.57/yg|m3
Overall: 0.172% (-0.313, 0.031) decrease
6-h: -0.769% (-1.21 to -0.327) decrease
B-blocker users: -0.062% (-0.248, 0.123)
Rest: 6 h:-0.173 (-0.345 to-0.001)
12 h:-0.160 (-0.308 to-0.012)
24 h:-0.169 (-0.316 to-0.022)
48 h:-0.153 (-0.304, 0.002)
Exercise: 6 h: -0.005 (-0.215, 0.205)
12 h:-0.014 (-0.196, 0.168)
24 h: 0.001 (-0.180,0.182)
48 h:-0.011 (-0.196, 0.174)
Post exercise Rest: 6 h: -0.173 (-0.332 to
¦0.014)
12 h:-0.128 (-0.266, 0.010)
4 h:-0.113 (-0.250, 0.023)
48 h:-0.157 (-295 to-0.019)
Paced breathing: 6 h: -0.142 (-0.292,
0.007)
12 h:-0.139 (-0.269 to-0.010)
24 h:-0.121 (-0.248, 0.007)
48 h:-0.082 (0.211, 0.047)
Summary over protocol
6 h:-0.131 (-0.247 to-0.015)
12 h:-0.120 (-0.221, 0.020)
24 h: -0.112 (-0.212 to -0.013)
Notes: Figure of the Variation in Oxygen
Saturation during the first rest period
versus individual hourly lag measurements
for PM2.6
July 2009
E-33
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Diez-Roux et al. (2006,
156400)
Period of Study: Baseline data collected
June 2000-Aug 2002
Location: USA (6 field centers: Baltimore,
MD
Chicago, IL
Forsyth Co, NC
Los Angeles, CA
New York, NY
St. Paul, MN
Outcome: C-reactive protein (CRP)
assessed continuously and as a
dichotomous variable (cutpoint, 3 mg/L)
interleukin-6 (IL-6)
Age Groups: 45-84 yrs
Study Design: Cross-sectional
N: 5634 persons
Statistical Analyses: Linear regression
& logistic regression
Covariates: Age, sex, race/ethnicity,
general health status, BMI, diabetes,
cigarette status, secondhand smoke,
physical activity, arthritis flare in last 2
weeks, medications, infections in last 2
weeks (also ran models including site,
copollutants, and weather)
Season: Examined seasonal patterns in
the residuals of fully adjusted models
stratified by season
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM2.6
Averaging Time: Prior day, prior 2 days,
prior week, prior 30 days, and prior 60
days
Mean (SD): Presented in Fig 1 by site
Percentiles: Presented in Fig 1 by site
Range: NR
Monitoring Stations: NR
Long-term exposure to PM estimated
based on residential history reported
retrospectively
all addresses geocoded
ambient AP obtained from US EPA
Copollutant: SO2
NO?
CO
03
PM Increment: 10 |Jgfm3
Effect Estimate [Lower CI, Upper CI]:
Adjusted (all personal-level covariates)
relative difference in CRP (mg/L) per 10
|jg/m3 increase in PM2.6
Prior day: 0.99 (0.96,1.01)
Prior 2 days: 0.99 (0.96,1.01)
Prior 7 days: 1.00 (0.96,1.04)
Prior 30 days: 1.03 (0.98,1.10)
Prior 60 days: 1.04 (0.97,1.11)
Odds Ratios of CRP of a 3 mg/L per 10
|jg/m3 increase in PM2.6 (adjusted for all
personal-level covariates)
Prior day: 0.98 (0.92,1.04)
Prior 2 days: 0.99 (0.93,1.06)
Prior 7 days: 1.05 (0.96,1.15)
Prior 30 days: 1.12 (0.98,1.29)
Prior 60 days: 1.12 (0.96,1.32)
Reference: Dubowsky et al. (2006,
088750)
Period of Study: March-Jun 2002
Location: St. Louis, Missouri
Outcome: White blood cells (WBC), C-
reactive protein (CRP), interleukin-6 (IL-6)
Age Groups: a 60 yrs
Study Design: Panel (4 planned repeated
measures
n - 35 participated in 4 trips)
N: 44 participants
Statistical Analyses: Linear mixed
models
Covariates: Sex, obesity, diabetes,
smoking history, time-varying parameters
(apparent temperature, h, day, trip,
residence, mold, pollen, illness, and juice
intake), medication and vitamin
consumption (day of blood draw)
Season: Limited data collection period
Dose-response Investigated? No
Statistical Package: SAS v8.02
Pollutant: PM2.6 (ambient)
Averaging Time: Hourly data used to
calculate avg concentrations over 1-7
days preceding the blood draw (ambient
PM2.6)
microenvironmental PM2.6 measures were
averaged over the 1-2 days preceding the
blood draw
Mean (SD) (1-day): 16 (6.0)
Percentiles (1 -day): 0: 6.5
25th: 12
75th: 22
100th: 28
Monitoring Stations: 1 ambient monitor
Copollutant: PM2.6 (ambient)
BC (ambient)
PM2.6 (microenvironment)
CO
NO2
SO2
Os
PM Increment: 6.1 /yg/m3 (5-d mean)
Effect Estimate [Lower CI, Upper CI]:
Note: Most results presented in figures.
Selected result in abstract text: % change
in WBC per increase in IQR (5.4 /yg/m3) of
PM2.E averaged over the previous week:
5.5(0.1,11)
Associations (% changes and 95%CI)
between 5-day mean ambient
concentrations and markers of
inflammation per increase (IQR) in
pollutant.
CRP: All participants: 14 (-5.4, 37)
Among those with all 3 conditions
(diabetes, obesity, and hypertension): 81
(21,172)
Among those with at least 2 of the
conditions: 11 (-7.3, 33)
IL-6: All participants: -2.1 (-13,11)
Among those with all 3 conditions
(diabetes, obesity, and hypertension): 23
(-5.3, 59)
Among those with at least 2 of the
conditions: -3.1 (-14, 9.7)
WBC (x109/L): All participants: 3.4 (-1.8,
8.9)
Among those with all 3 conditions
(diabetes, obesity, and hypertension): 0.4
(-8.8,11)
Among those with at least 2 of the
conditions: 3.6 (-1.7, 9.1)
July 2009
E-34
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Dubowsky et al. (2006,
088750)
Period of Study: March-Jun 2002
Location: St. Louis, Missouri
Outcome: White blood cells (WBC), C-
reactive protein (CRP), interleukin-6 (IL-6)
Age Groups: a 60 yrs
Study Design: Panel (4 planned repeated
measures
n - 35 participated in 4 trips)
N: 44 participants
Statistical Analyses: Linear mixed
models
Covariates: Sex, obesity, diabetes,
smoking history, time-varying parameters
(apparent temperature, h, day, trip,
residence, mold, pollen, illness, and juice
intake), medication and vitamin
consumption (day of blood draw)
Season: Limited data collection period
Dose-response Investigated? No
Statistical Package: SAS v8.02
Pollutant: BC (ngfm3) (ambient)
Averaging Time: Hourly data used to
calculate avg concentrations over 1-7
days preceding the blood draw (ambient
PM)
microenvironmental PM2.6 measures were
averaged over the 1-2 days preceding the
blood draw
Mean (SD) (1-day): 900 (280)
Percentiles (1 -day): 0: 290
25th: 730
75th: 1,100
100th: 1,400
Monitoring Stations: 1 ambient monitor
Copollutant: PM2.6 (ambient)
BC (ambient)
PM2.6 (microenvironment)
CO; NO2
SO?
0s
PM Increment: 230 ng/m3 (5-d mean)
Effect Estimate [Lower CI, Upper CI]:
Note: Most results presented in figures.
Associations (% changes and 95%CI)
between 5-day mean ambient
concentrations and markers of
inflammation per increase (IQR) in
pollutant.
CRP: All participants: 13 (-0.34, 28)
Among those with all 3 conditions
(diabetes, obesity, and hypertension): 49
(16, 90)
Among those with at least 2 of the
conditions: 9.0 (-3.8, 24)
IL-6: All participants: -0.8 (-8.9, 8.0)
Among those with all 3 conditions
(diabetes, obesity, and hypertension): 15
(-2.2, 35)
Among those with at least 2 of the
conditions: -2.7 (-11, 6.2)
WBC (x10B/L): All participants: 1.3 (-2.1,
4.8)
Among those with all 3 conditions
(diabetes, obesity, and hypertension): 0.05
(-5.9, 6.3)
Among those with at least 2 of the
conditions: 1.5 (-2.0, 5.1)
Reference: Ebelt et al. (2005, 056907)
Period of Study: Summer of 1998
Location: Vancouver, Canada
Outcome: CVD
Age Groups: range from 54-86 yrs
mean age - 74 years
Study Design: extended analysis of a
repeated-measures panel study
N: 16 persons with C0PD
Statistical Analyses:
Earlier analysis expanded by developing
mixed-effect regression models and by
evaluating additional exposure indicators
Dose-response Investigated? No
Statistical Package: SAS V8
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD):
Ambient PM2.6:11.4 ± 4.6
Exposure to ambient PM2.6: 7.9 ± 3.7
Range (Min, Max): Ambient PM2.6: 4.2 -
28.7
Exposure to ambient PM2.6: 0.9 ¦ 21.3
Monitoring Stations: 5
Copollutant (correlation):
Ambient concentrations and exposure to
ambient PM were highly correlated for
each respective metric: r a 0.71
Note: Total personal fine particle
exposure (T) were dominated by exposures
to non ambient particles which were not
correlated with ambient fine particle
exposure (A) or ambient concentrations
(C). Results for each of these metrics are
listed.
PM Increment:
Increment: C26: IQR - 5.8
SBP (mm Hg):-1.70 (-3.48-0.08)
DBP (mm Hg): -0.58 (-2.02-0.85)
Ln-SVE (bph): 0.20 (0.00-0.40)
HR (bpm): 0.93 (-0.90-2.75)
SDNN (ms): -4.37 (-9.40-0.65)
R-MSSD (ms):-2.79 (-6.16-0.57)
Increment: NS C26: IQR - 4.2
SBP (mm Hg):-1.52 (-2.94- -0.09)
DBP (mm Hg): -0.77 (-1.87-0.32)
Ln-SVE (bph): 0.19 (-0.01-0.38)
HR (bpm): 1.03 (-0.43-2.48)
SDNN (ms): -3.83 (-7.77-0.11)
R-MSSD (ms):-2.90 (-5.55- -0.25)
Increment: S_C26: IQR - 1.5
SBP (mm Hg):-1.10 (-3.48-1.28)
DBP (mm Hg): 0.76 (-1.15-2.68)
Ln-SVE (bph): 0.09 (-0.05-0.23)
HR (bpm):-0.42 (-2.28-1.44)
SDNN (ms):-3.14 (-9.73-3.45)
R-MSSD (ms): 0.24 (-5.14-5.63)
Increment: A26: IQR - 4.4
SBP (mm Hg):-1.90 (-3.66--0.14)
DBP (mm Hg): -0.33 (-1.72-1.06)
Ln-SVE (bph): 0.20 (0.02-0.37)
HR (bpm): 0.57 (-1.34-2.47)
SDNN (ms): -3.91 (-8.79-0.97)
R-MSSD (ms):-1.05 (-4.79-2.17)
Increment: NS A2 6: IQR - 3.4
July 2009
E-35
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
SBP (mm Hg): -1.70 (-3.27- -0.14)
DBP (mm Hg):-0.51 (-1.71-0.70)
Ln-SVE (bph): 0.20 (0.02-0.37)
HR (bpm): 0.69 (-0.96-2.35)
SDNN (ms):-4.18 (-8.51-0.15)
R-MSSD (ms):-1.40 (-4.40-1.60)
Increment: S T2 6: IQR - 0.9
SBP (mm Hg):-1.55 (-3.35-0.26)
DBP (mm Hg): 0.49 (-0.91-1.90)
Ln-SVE (bph): 0.08 (-0.14-0.19)
HR (bpm):-0.24 (-1.75-1.26)
SDNN (ms): -0.68 (-4.74-3.38)
R-MSSD (ms): 0.91 (-3.51-5.33)
Increment: T26: IQR - 10.1
SBP (mm Hg):-1.26 (-2.60-0.08)
DBP (mm Hg): 0.34 (-1.26-1.94)
Ln-SVE (bph): 0.01 (-0.10-0.11)
HR (bpm):-0.23 (-1.09-0.63)
SDNN (ms): -2.11 (-4.90-0.68)
R-MSSD (ms):-0.83 (-3.60-1.94)
Increment: N26: IQR - 8.9
SBP (mm Hg):-0.81 (-2.15-0.53)
DBP (mm Hg): 0.40 (-1.19-1.98)
Ln-SVE (bph):-0.04 (-0.18-0.10)
HR (bpm):-0.35 (-0.85-0.14)
SDNN (ms):-1.10 (-3.10-0.90)
R-MSSD (ms):-0.54 (-2.54-1.46)
Reference: Fan et al. (2008,191979)
Period of Study: Feb - May 2005
Location: Paterson, New Jersey
Outcome: Cardiopulmonary Health (FEV,
FVC, PEF, SDNN, HR)
Age Groups: 61.2 (13.7)
Study Design: panel
N: 11
Statistical Analyses: Mixed Effects
models, Linear Regression models
Covariates: temperature, humidity
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0
Pollutant: PM2.6
Averaging Time: daily
Mean (SD): |
APM2.6 avg
Morning: 35.2 (25.9)
Afternoon: 24.1 (22.1)
APM2.6 peak
Morning: 71.3 (56.1)
Afternoon: 64.3 (43.5)
Range:
APM2.6 avg
Morning: 1.1-87
Afternoon: 1.2-98
APM2.6 peak
Morning: 4.0 ¦ 278
Afternoon: 3.0 ¦ 150
Monitoring Stations: NR
Copollutant: NR
Co-pollutant Correlation
n/a
PM Increment: 10/yg/m3
Beta (SE), p-value:
ASDNN
Morning, APM2.6 avg
15min: -14.5 (6.9), 0.06
2h:-18.9 (4.2), 0.0002
4h: -2.5 (8.6), 0.78
Morning, APM2.6 peak
15min: -9.2 (11.2), 0.43
2h: -5.1 (13.8), 0.72
4h: -7.4 (12.0), 0.55
Afternoon, APM2.6 avg
15min: -2.4 (7.6), 0.77
2h: -20.2 (10.8), 0.10
4h: -0.7 (11.2), 0.95
Afternoon, APM2.6 peak
15min: 0.6 (8.9), 0.95
2h: 19.2(14.6), 0.23
4h: -6.8 (14.1), 0.64
A HR
Morning, APM2.6 avg
15min: 1.2(3.1), 0.71
2h: -5.5 (2.9), 0.08
4h: -3.1 (4.6),0.51
Morning, APM2.6 peak
15min: 0.8 (4.4), 0.86
2h: -7.2 (4.2), 0.11
4h: -7.1 (6.3),0.28
Afternoon, APM2.6 avg
15min: -2.0 (4.0), 0.62
2h: 0.9 (5.4), 0.87
4h: 8.2(5.2), 0.14
Afternoon, APM2.6 peak
15min: -5.6 (5.3), 0.31
2h: 3.1 (8.1), 0.71
4h: 11.1 (8.1),0.20
AFEVi
Morning, APM2.6 avg: 0.02 (0.04), 0.68
Morning, APM2.6 peak: -0.13 (0.08), 0.16
A FVC
JVIomincj^APM^^
July 2009
E-36
DRAFT-DO NOT CITE OR QUOTE

-------
Reference	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Morning, APM2.6 peak: -0.12 (0.17), 0.51
A PEF
Morning, APM2.6 avg: -0.54 (0.62), 0.42
Morning, APM2.6 peak: -1.46 (1.12), 0.24
Notes: Estimates relative to increases in
the average and peak PM2.B
concentrations
Reference: Folino et al. (2009,191902)
Period of Study: Jun 2006 - May 2007
Location: Padua, Italy
Outcome: HRV 81 Inflammatory Markers
Age Groups: 45-65 yrs
Study Design: panel
N: 39 patients wf myocardial infarction
Statistical Analyses: Linear Regression
Model, ANOVA
Covariates: temperature, relative
humidity, atmospheric pressure, beta-
blocker, aspirin, or nitrate consumption,
smoking habit
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: NR
Pollutant: PM2.6
Averaging Time: 24h
Mean (SD):
Summer: 33.9 (12.7)
Winter: 62.1 (27.9)
Spring: 30.8 (14.0)
Monitoring Stations: NR
Copollutant: PM10.PM02E
Co-pollutant Correlation
NR
PM Increment: 1 /yglm3
Beta (SE), p-value:
SDNN: 0.109 (0.1151,0.345
SDANN: 0.127 (0.1261,0.314
RMSSD: 0.045 (0.040), 0.256
pH: 0.002 (0.0011,0.041
LTB4: 0.590 (0.324), 0.069
eNO: -0.002 (0.003), 0.503
PTX3:-0.004 (0.002), 0.013
C-reactive protein: -0.008 (0.005), 0.115
CC16:-0.002 (0.002), 0.410
IL-8: 0.000 (0.003), 0.989
Reference: Folino et al. (2009,191902)
Period of Study: Jun 2006 - May 2007
Location: Padua, Italy
Outcome: HRV 81 Inflammatory Markers
Age Groups: 45-65 yrs
Study Design: panel
N: 39 patients wf myocardial infarction
Statistical Analyses: Linear Regression
Model, ANOVA
Covariates: temperature, relative
humidity, atmospheric pressure, beta-
blocker, aspirin, or nitrate consumption,
smoking habit
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: NR
Pollutant: PM0.26
Averaging Time: 24h
Mean (SD):
Summer: 17.6 (7.5)
Winter: 30.5(17.4)
Spring: 18.8 (10.8)
Monitoring Stations: NR
Copollutant: PM10.PM2 E
Co-pollutant Correlation
NR
PM Increment: 1 /yglm3
Beta (SE), p-value:
SDNN: 0.214(0.2041,0.295
SDANN: 0.214(0.2141,0.316
RMSSD: 0.081 (0.077), 0.291
pH: 0.005 (0.002), 0.004
LTB4: 0.835 (0.5331,0.117
eNO:-0.006 (0.005), 0.182
PTX3:-0.006 (0.003), 0.071
C-reactive protein: -0.011 (0.007), 0.104
CC16: 0.001 (0.004), 0.890
IL-8:-0.004 (0.006), 0.527
Reference: Goldberg et al. (2008,
180380)
Period of Study: Jul 2002 - Oct 2003
Location: Montreal, Canada
Outcome: Oxygen saturation 81 pulse rate
Age Groups: 50-85 yrs
Study Design: panel
N: 31
Statistical Analyses: Mixed Random
Effects Model
Covariates: body temperature,
consumption of salt, intake of fluids, being NO2: 0.62
ill the day before, ambient temperature,
relative humidity, barometric pressure
Dose-response Investigated? No
Statistical Package: Splus
Lags Considered: lags 1d 81 0-2d avg
Pollutant: PM2.6
Averaging Time: daily
IQR: 7.3
Monitoring Stations: 8
Co-pollutant: CO, NO2, SO2, O3
Co-pollutant Correlation
CO: 0.72
PM Increment: Interquartile Range (7.3
//g/m3)
Mean Difference (Lower CI, Upper CI),
lag:
Oxygen Saturation
-0.087 (-0.143, -0.031), lag 0
-0.058 (-0.114, -0.002), I
-0.083 (-0.155, -0.010), I
|1
0.056(-0.117, 0.005), lag 0
0.019(-0.079, 0.041), lag 1
0.039(-0.118, 0.039), lag 0-2d
Unadjusted
Unadjusted
Unadjusted
0-2d avg
Adjusted
Adjusted
Adjusted:
avg
Pu|sg p ate
Unadjusted: 0.226 (-0.037, 0.489), lag 0
Unadjusted: 0.288 (0.022, 0.554), lag 1
Unadjusted: 0.420 (0.067, 0.772), lag 0-
2d avg
Adjusted: 0.158 (-0.136, 0.451), lag 0
Adjusted: 0.246 (-0.040, 0.531), lag 1
Adjusted: 0.353 (-0.034, 0.740), lag 0-2d
avg
July 2009
E-37
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Goldberg et al. (2008,
180380)
Period of Study: Jul 2002 - Oct 2003
Location: Montreal, Canada
Outcome: Shortness of Breath & General
health
Age Groups: 50-85 yrs
Study Design: panel
N: 31
Statistical Analyses: Mixed Random
Effects Model
Covariates: body temperature,
consumption of salt, intake of fluids, being
ill the day before, ambient temperature,
relative humidity, barometric pressure
Dose-response Investigated? No
Statistical Package: Splus
Lags Considered: lags 0-4d & 0-2d avg
Pollutant: PM2.6
Averaging Time: daily
Mean: 9.5
Median: 7.0
Min: 0.8
Max: 50.2
IQR: 7.3
Monitoring Stations: 8
Co-pollutant: CO, NO2, SO2, O3
Co-pollutant Correlation
CO: 0.66
NO?: 0.54
O3: 0.32
SO2: 0.50
PM Increment: Interquartile Range (7.3
j"g/m3)
Mean Difference (Lower CI, Upper CI),
lag:
General Health
Unadjusted: -0.317 (-0.699, 0.064), lag 0
Unadjusted: -0.284 (-0.670, 0.103), lag 1
Unadjusted: -0.048 (-0.427, 0.332), lag 2
Unadjusted: -0.241 (-0.620, 0.139), lag 3
Unadjusted: -0.010 (-0.390, 0.370), lag 4
Unadjusted: -0.482 (-1.053, 0.090), lag 0-
2d avg
Adjusted: -0.125 (-0
Adjusted: -0.167 (-0
Adjusted: -0.081 (-0.
Adjusted: -0.222 (-0
Adjusted: 0.016 (-0.:
Adjusted: -0.281 (-0
avg
Shortness of breath at night
Unadjusted: -0.421 (-0.847, 0.006), lag 0
Unadjusted: -0.278 (-0.711, 0.155), lag 1
Unadjusted: -0.100 (-0.526, 0.327), lag 2
Unadjusted: -0.220 (-0.645, 0.206), lag 3
Unadjusted: -0.206 (-0.632, 0.220), lag 4
Unadjusted: -0.555 (-1.172, 0.063), lag 0-
2d avg
Adjusted:-0.171 (-0.639, 0.297), lag 0
Adjusted:-0.130 (-0.579, 0.319), lag 1
Adjusted:-0.127 (-0.553, 0.299), lag 2
Adjusted:-0.192 (-0.616, 0.231), lag 3
Adjusted:-0.171 (-0.594, 0.253), lag 4
Adjusted:-0.301 (-0.952, 0.350), lag 0-2d
avg
545, 0.295), lag 0
568, 0.234), lag 1
464, 0.302), lag 2
602, 0.157), lag 3
364, 0.396), lag 4
886, 0.325), lag 0-2d
Reference: Ibald-Mulli et al. (2004,
087415)
Period of Study: winter 1998-1999
Location: Helsinki, Finland
Erfurt, Germany
Amsterdam, the Netherlands
Outcome: Blood Pressure 81 Heart Rate
Age Groups: 40-84
Study Design: panel
N: 131 adults wICHD
Statistical Analyses: Linear Regression
Covariates: trend, day of week,
temperature, barometric pressure, relative
humidity, medication use
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-2, 5d avg
Pollutant: PM2.6
Averaging Time: 24h
Mean (SD):
Downtown: 40 (20)
Tunney's Pasture: 10 (10)
p-value 0.000
Monitoring Stations: NR
Copollutant: PM1.0
Co-pollutant Correlation
n/a
PM Increment: Interquartile Range
(27.02 /yglm3)
Beta (SE), p-value:
Flow mediated vasodilation (%):
¦0.016 (0.0072) p-0.03
Heart Rate (beats/min):
0.081 (0.135) p-0.55
Diastolic blood pressure (mmHg):
0.088 (0.088) p-0.32
Systolic blood pressure (mmHg):
¦0.108 (0.006) p-0.48
Reference: Langrish et al. (2009,
191908)
Period of Study: August 2008
Location: Beijing, China
Outcome: Cardiovascular Effects
Age Groups: median 28 yrs
Study Design: panel
N: 15
Statistical Analyses: NR
Covariates: NR
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: NR
Pollutant: PM2.6
Averaging Time: NR
Mean:
W/o mask: 86
Wlmask: 140
Monitoring Stations: NR
Co-pollutant: CO, SO2, NO2
Co-pollutant Correlation:
n/a
PM Increment: NR
Mean (Lower CI, Upper CI):
W/o Mask (Day)
SBP: 100(104,116)
DBP: 73 (69, 76)
MAP: 85(81,88)
Heart Rate: 79 (74, 84)
Avg NN interval: 829 (789, 869)
pNN50: 15.9(10.7,21.0)
RMSSD: 35.1 (29.2,41.0)
SDNN: 61.2(54.9, 67.5)
Triangular index: 12.9 (11.9,13.9)
LF power: 816(628,1004)
HF power: 460 (325, 595)
LFn: 62.8(56.7,68.9)
HFn: 29.2(25.5, 32.8)
HF/LF ratio: 0.738 (0.507, 0.970)
W/ Mask (Day)
SBP: 109(104,114)
DBP: 73 (70-76)
MAP: 85(81,89)	
July 2009
E-38
DRAFT-DO NOT CITE OR QUOTE

-------
Reference	Design & Methods	Concentrations'	Effect Estimates (95% Cl)
Heart Rate: 78 (73, 82)
Avg NN interval: 850 (805, 896)
pNN50: 17.9(14.2,21.6)
RMSSD: 37.1 (32.2,42.0)
SDNN: 65.5(59.0,72.2)*
Triangular index: 13.8 (13.0,14.5)
LF power: 919 (717,1122)*
HF power: 485 (400, 569)
LFn: 64.5 (60.6, 68.4)
HFn: 30.0 (27.0, 33.1)
HF/LF ratio: 0.680 (0.519,0.842)
Wfo Mask (During Walk)
SBP: 121 (115,127)
DBP: 81 (75-87)
MAP: 94 (89, 99)
Heart Rate: 88 (82, 94)
Avg NN interval: 594 (562, 627)
pNN50: 3.3(0.8, 5.7)
RMSSD: 17.2 (13.4, 21.0)
SDNN: 45.8 (36.8, 54.8)
Triangular index: 10.7 (9.1,12.4)
LF power: 313(170, 455)
HF power: 76.5(33.6,120.0)
LFn: 68.2(60.9,75.5)
HFn: 16.1 (11.9, 20.3)
HF/LF ratio: 0.259 (0.173,0.344)
Wf Mask (During Walk)
SBP: 114(108,120)
DBP: 79 (74, 83)
MAP: 90 (86, 94)
Heart Rate: 91 (85, 97)
Avg NN interval: 613 (571, 655)
pNN50: 2.1 (-0.1,4.4)
RMSSD: 20.0 (15.5, 24.6)
SDNN: 54.8 (42.5, 67.0)
Triangular index: 11.4(9.4,13.3)
Wf Mask (During Walk)
LF power: 414(233, 595)
HF power: 116.8(52.6,181.0)
LFn: 67.9(61.9,73.9)
HFn: 16.0(12.5,19.4)
HF/LF ratio: 0.247 (0.180,0.314)
Mean (SD):
W/o Mask (After Walk)
Headache: 2.53 (5.55)
Dizziness: 1.07 (2.22)
Tiredness: 8.47 (12.14)
Sickness: 1.07 (2.22)
Cough: 1.80(4.80)
Difficulty Breathing: 0.67 (0.90)
Eye irritation: 1.40 (3.60)
Throat irritation: 1.47 (4.07)
Nose irritation: 1.53 (3.78)
Unpleasant Smell: 0.93 (1.22)
Bad taste: 0.73 (0.96)
Difficulty walking: 12.53 (13.24)
Perception of Pollution: 19.80 (18.37)
W| Mask (After Walk)
Headache: 0.73 (1.03)
Dizziness: 0.80 (1.57
Tiredness: 7.40 (9.37)
Sickness: 0.87 (1.51)
Cough: 1.00(1.73)
Difficulty Breathing: 3.80 (8.10)
Eye irritation: 1.67 (3.27)
Throat irritation: 1.07 (2.63)
Nose irritation: 1.07 (1.91)
Unpleasant Smell: 0.60 (0.91)
Bad taste: 0.60 (1.18)
Difficulty walking: 15.13 (11.51)
Perception of Pollution: 11.60 (10.44)
*p< 0.05
Notes: Estimates also available for 24
July 2009
E-39
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)



hours, night, before walk, and 24 hours



after walk.
Reference: Lanki et al. (2006, 088412)
Period of Study: Autumn 1998-spring
1999
Location: Helsinki, Finland
Outcome: ST segment depressions (2
endpoints: >0.1 mV regardless of the
direction of the ST slope and >0.1mV
with horizontal or downward slope
[stricter criteria])
Age Groups: Mean - 68.2 (6.5) yrs
Study Design: Panel
N: 45 elderly nonsmoking persons with
stable coronary heart disease
342 total exercise tests for analyses
Statistical Analyses: Generalized
additive models with penalized splines
(logistic regression)
principal components analysis and linear
regression of 13 measured elements used
to apportion PM2.6 mass between different
sources
Covariates: Subject, linear terms for time
trend, temperature, relative humidity,
penalized spline for change in heart rate
during the exercise test
Season: NR
Dose-response Investigated? No
Statistical Package: S-plus 2000 and R
Pollutant: PM2.6 (Analyses conducted for
source specific PM2.6)
Averaging Time: Daily filter samples
Mean: Crustal: 0.6
Long-range transported: 6.4
Oil combustion: 1.6
Salt: 0.9
Local traffic: 2.9
Total: 12.8
Percentiles: Crustal
25: 0.0
50: 0.4
75: 1.1;
Max: 5.3
Long-range transported
25: 2.2
50: 5.5
75: 9.8;
Max: 26.5
Oil combustion
25: 0.6
50: 1.3
75: 2.3;
Max: 12.2
Salt
25: 0.3
50: 0.8
75: 1.2;
Max: 5.9
Local traffic
25: 1.7
50: 2.5
75: 3.4;
Max: 12.0
Total
25: 8.3
50: 10.6
75:15.9;
Max: 39.8
Monitoring Stations: 1 monitor
Copollutant (correlation): Correlations
with PM2.6: Crustal: r - -0.01
Long-range transported: r - 0.82
Oil combustion: r - 0.35
Salt: r - 0.19
Local traffic: r - 0.26
PM Increment: 1 /yglm3
Effect Estimate [Lower CI, Upper CI]:
Adjusted ORs between daily source-
specific PM2.6 concentrations and ST
segment depressions. ST segment
depression defined as >0.1 mV (n - 62)
Crustal
LagO: 0.80 (0.47,1.36)
Lag1:0.66 (0.40,1.10)
Lag2:1.18 (0.68, 2.06)
Lag3:1.87 (0.85, 4.09)
Long-range transport
LagO: 0.94 (0.84,1.05)
Lag1:1.00 (0.92,1.08)
Lag2:1.11 (1.02,1.20)
Lag3:1.06 (0.95,1.18)
Oil combustion
LagO: 0.87 (0.57,1.32)
Lag1:1.04 (0.75,1.45)
Lag2:1.10 (0.83,1.46)
Lag3:1.12 (0.79,1.58)
Salt
LagO: 1.03 (0.57,1.85)
Lag1: 0.72 (0.37,1.40)
Lag2: 0.66 (0.31,1.40)
Lag3:1.55 (0.83, 2.89)
Local traffic
LagO: 0.91 (0.69,1.21)
Lag1:1.22 (0.88,1.69)
Lag2:1.53 (1.19,1.97)
Lag3: 0.98 (0.78,1.23)
ST segment depression defined as >0.1
mV with horizontal or downward slope
(n - 46)
Crustal
LagO: 0.76 (0.42,1.35)
Lag1: 0.41 (0.22, 0.79)
Lag2:1.17 (0.65, 2.09)
Lag3:1.60 (0.72, 3.59)
Long-range transport
LagO: 0.98 (0.86,1.10)
Lag1:1.03 (0.95,1.12)
Lag2:1.11 (1.02,1.21)
Lag3:1.02 (0.95,1.10)
Oil combustion
LagO: 0.95 (0.61,1.49)
Lag1:1.13 (0.76,1.68)
Lag2:1.33 (0.98,1.80)
Lag3:1.29 (0.90,1.86)
Salt
LagO: 1.15 (0.56, 2.38)
Lag1:0.90 (0.44,1.81)
Lag2:1.39 (0.63, 3.08)
Lag3:1.93 (1.00, 3.72)
Local traffic
LagO: 0.89 (0.64,1.23)
Lag1:1.21 (0.86,1.71)
Lag2:1.37 (1.03,1.83)
Lag3:1.03 (0.80,1.32)
Adjusted ORs for the association of
indicator elements of PM2.6 sources and
ST segment depressions in multipollutant
models (models include all 5 indicator
elements). ST segment depression defined
as >0.1 mV (n - 62)
Si (Crustal)	
July 2009
E-40
DRAFT-DO NOT CITE OR QUOTE

-------
Reference	Design & Methods	Concentrations'	Effect Estimates (95% CI)
LagO: 0.73 (0.39,1.38)
Lag1: 0.48 (0.25, 0.93)
Lag2: 0.78 (0.35,1.71)
Lag3:1.95 (0.69, 5.48)
S (Long-range transport)
LagO: 0.70 (0.25,1.95)
Lag1:0.58 (0.23,1.47)
Lag2:1.08 (0.44,2.63)
Lag3:1.60 (0.73, 3.48)
Ni (Oil combustion)
LagO: 0.78 (0.30, 2.04)
Lag1:1.20 (0.58, 2.46)
Lag2:1.15 (0.61, 2.18)
Lag3:1.02 (0.41,2.54)
CI (Salt)
LagO: 1.03 (0.79,1.34)
Lag1:0.88 (0.56,1.38)
Lag2:1.02 (0.62,1.69)
Lag3:1.27 (0.85,1.91)
ABS (Local traffic)
LagO: 0.92 (0.36, 2.37)
Lag1:1.83 (0.73, 4.59)
Lag2: 4.46 (1.69,11.79)
Lag3: 0.92 (0.40, 2.12)
ST segment depression defined as >0.1
mV with horizontal or downward slope
(n - 46)
Si (Crustal)
LagO: 0.67 (0.33,1.36)
Lag1: 0.34 (0.15, 0.81)
Lag2: 0.81 (0.33, 2.00)
Lag3:1.90 (0.64,5.65)
S (Long-range transport)
LagO: 0.84 (0.29, 2.47)
Lag1:0.89 (0.34,2.32)
Lag2:1.36 (0.54,3.45)
Lag3:1.12 (0.53, 2.40)
Ni (Oil combustion)
LagO: 1.10 (0.36, 3.37)
Lag1:1.16 (0.45, 2.96)
Lag2:1.64 (0.84,3.20)
Lag3:1.63 (0.64,4.14)
CI (Salt)
LagO: 1.13 (0.80,1.62)
Lag1:0.99 (0.58,1.68)
Lag2:1.55 (0.87, 2.76)
Lag3:1.45 (0.94,2.25)
ABS (Local traffic)
LagO: 0.74 (0.25, 2.23)
Lag1:1.76 (0.62, 5.00)
Lag2: 4.86 (1.55,15.26)
Lag3: 0.97 (0.39, 2.41)
July 2009
E-41
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lanki et al. (2008,191984)
Period of Study: Jan 1999 - Apr 1999
Location: Helsinki, Finland
Outcome: ST Segment Depress
>0.1 mV
Age Groups: 50 +
Study Design: panel
N: 41 elderly people wf CHD
Statistical Analyses: Logistic
Model
Covariates: long-term time trend,
temperature, humidity, change in heart
rate following exercise test
Dose-response Investigated? No
Statistical Package: R
Lags Considered: lags 0-24h
Pollutant: PM2.6
Averaging Time: hourly
25", 50", 75", Max:
Personal PM2.6
1 h: 6.9,11.2,15.8, 41.5
4h: 5.9,10.0,14.6, 41.3
sslon 8h: 5.0, 7.9,13.0, 34.9
12h: 5.2,7.8,12.1,28.8
22h: 6.6, 9.3,13.0, 30.2
Outdoor PM2.6
1 h: 8.9,12.9,17.8, 42.9
4h: 8.8,12.5,17.6, 40.8
8h: 8.3,12.1,17.2, 39.2
12h: 8.3,11.9,17.0, 37.0
24h: 9.0,12.5,17.7, 30.5
Monitoring Stations: 1
Co-pollutant: PM
-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ljungman et al. (2008,
180266)
Period of Study: Aug 2001 - Dec 2006
Location: Stockholm, Sweden
Outcome: Ventricular Arrhythmia
Age Groups: 28-85 yrs
Study Design: case-crossover
N: 88 patients wf implantable cardioverter
defibrillators
Statistical Analyses: conditional logistic
regression
Covariates: temperature, humidity,
pressure, ischemic heart disease, ejection
fraction, heart disease, diabetes, use of
beta-blockers, age, BMI, location at time
of arrhythmia, distance from air pollution
monitor
Dose-response Investigated? No
Statistical Package: Stata, S-plus
Lags Considered: lags 2-24h
Pollutant: PM2.6
Averaging Time: hourly
Median:
2h: 9.17
24h: 9.49
Min:
2h: 0.15
24h: 2.97
Max:
2h: 99.25
24h: 47.07
IQR:
2h: 6.69
24h: 5.27
Monitoring Stations: 1
Copollutant: PMio, NO2
Co-pollutant Correlation
NR
PM Increment: Interquartile
Odds Ratio (Lower CI, Upper CI):
2h: 1.23 (0.84,1.80)
24h: 1.28(0.90,1.84)
Notes: OR of ventricular arrhythmia for an
IQR increase of air pollutants in different
subgroups (figure 2)
Reference: Ljungman et al. (2009,
191983)
Period of Study: May 2003 - July 2004
Location: Athens, Greece
Helsinki, Finland
Ausburg, Germany
Barcelona, Spain
Rome, Italy
Stokholm, Sweeden
Outcome: lnterleukin-6 Response
Age Groups: 35-80 yrs
Study Design: panel
N: 955 male myocardial infarction
survivors
Statistical Analyses: Additive Mixed
Models
Covariates: age, sex, BMI, city, HDL/total
cholesterol, smoking, alcohol intake,
HbA1c, NT-proBNP, history of Ml, heart
failure, or diabetes, phlegm
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1d
Pollutant: PM2.6
Averaging Time: 24h
Mean: 17.7
25": 10.9
75": 21.9
Monitoring Stations: NR
Copollutant: CO, NO2, PNC, PM2.6
Co-pollutant Correlation
PM10: 0.81
PM Increment: Interquartile Range (11.0
j"g/m3)
Change of IL-6 (Lower CI, Upper CI), p-
value:
0.6 (-0.8, 2.0), 0.40
Reference: Luttman-Gibson et al. (2006,
089794)
Period of Study: June-December 2000
Location: Steubenville, OH
Outcome: Heart rate variability
Age Groups:
Study Design: Panel study
N: 32 participants
Statistical Analysis: Linear mixed
models
Pollutant: PM2.6
Averaging Time: 1 h
24 h
Mean (IQR)
PM2.6: 20.0(15.2)
Sulfate: 6.9 (5.1)
EC: 1.1 (0.6)
Copollutant: NO2, SO2, O3
PM Increment: IQR
Percent change (95% CI): Each 13.4
/yg/m3 increase in 24 hour mean PM2.6
concentration was associated with:
SDNN: -4.0% (95% CI:-7.0% to-0.9%)
r-MSSD: -6.5% (95% CI: -12.1 % to -0.6%)
HF: -11.4% (95% CI:-21.5% to-0.1%)
Each 5.1 /yglm3 increase in sulfates on the
previous day was associated with: SDNN:
¦3.3% (95% CI:-6.0% to-0.5%)
r-MSSD: -5.6% (95% CI:-10.7%, 0.2%)
HF: -10.3% (95% CI:-19.5% to-0.1%)
Notes: The authors conclude that
increases in both traffic related particles
and sulfates may adversely effect
autonomic function.
July 2009
E-43
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Mar et al. (2005, 087566)
Period of Study: 1999-2001
Location: Seattle, WA
Outcome: Change in arterial O2
saturation, heart rate, and blood pressure
(SBP and DBP)
Age Groups: >75 years
Study Design: Panel study
N: 88 elderly subjects
Statistical Analysis: GEE
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD): Personal: 9.3(8.4)
Indoor: 7.4 (4.8)
Outdoor: 9.0 (4.6)
PM Increment: 10 |jg/m3
Unit change in measure (95% CI):
Among all subjects: Each increase in
outdoor same day PM2.6 was associated
with: SBP: -0.81 mmHg (95% CI: -2.34,
0.73)
DBP:-0.46 mmHg (95% CI:-1.49, 0.57)
H: -0.75 beats/min (95% CI: -1.42 to ¦
0.07)
Each increase in indoor same day PM2.6
was associated with: SBP: 0.92 mmHg
(95% CI: -2.04, 3.87)
DBP: 0.38 mmHg (95% CI:-1.43, 2.20)
H: 0.22 beats/min (95% CI:-0.71,1.16)
Each increase in personal same day PM2.6
was associated with: SBP: 0.37 mmHg
(95% CI:-0.93,1.67)
DBP: -0.20 mmHg (95% CI: -0.85, 0.46)
H: 0.44 beats/min (95% CI: 0.04, 0.84)
Notes: Results by health status presented
in Figure 1
Used 2 sessions that each were 10
consecutive days of measurements
Used personal, indoor, and outdoor
measures of PM2.6
Reference: Metzgeret al. (2007,
092856)
Period of Study: August 1998-
December 2002
Location: Atlanta, GA
Outcome: Days with any event recorded
by the ICD, days with ICD
sliocksi'defibrillation and days with either
cardiac pacing or defibrillation
Study Design: Repeated measures
N: 884 subjects between 1993 and 2002
Statistical Analysis: Logistic regression
with GEE to account for residual
autocorrelation within subjects
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD): PM2.b: 17.8 (8.6)
PM2.6 sulfates: 5.0 (3.4)
PM2.6 EC: 1.7 (1.2)
PM2.6 0C: 4.4 (2.4)
PM2.6 water-soluble metals: 0.029 (0.024)
Percentiles: PM2.6: Median: 16.2
PM2.6 sulfates: Median: 4.1
PM2.6 EC: Median: 1.4
PM2.6 0C: Median: 3.9
PM2.E water-soluble metals: Median:
0.022
Copollutant: O3
NO2
CO
SO2
oxygenated hydrocarbons
PM Increment: OR (95% CI):
Outcome - Any event recorded by ICD
PM2.6
OR - 1.00
(95% CI: 0.95,1.04)
PM2.6 EC
OR - 1.01
(95% CI: 0.98,1.05)
PM2.6 oc
OR - 1.01
(95% CI: 0.98,1.03)
PM2.6 Sulfates
OR - 0.99
(95% CI: 0.93,1.06)
PM2.E Water soluble metals
OR - 0.95
(95% CI: 0.90,1.00
Reference: O'Neill et al. (2007, 091362)
Period of Study: May 1998-Dec 2002
Location: Boston, MA
Outcome: Soluble intercellular adhesion
molecule 1 (ICAM-1)
vascular cell adhesion molecule 1 (VCAM-
von Willebrand factor (vWF)
Age Groups: Mean (SD): 56.6 (10.6)
Study Design: Cross-sectional
N: 92 participants (type 2 diabetic
patients)
Statistical Analyses: linear regression
Covariates: Apparent temperature,
season, age, race, sex, glycosylated
hemoglobin, cholesterol, smoking history,
BMI
Dose-response Investigated? No
Pollutant: PM2.6
Averaging Time: 24 h (lagged moving
averages of days 0 to 1, 2, 3, 4, and 5)
Mean (SD): 11.4 (5.9)
descriptive statistics represent entire
study period
Percentiles: IQR range: 7.6
Range (Min, Max): 0.07, 33.7)
Monitoring Stations: 1 site
Copollutant: PM2.6
BC
SO42-
PM Increment: IQR (specific to lag
period)
Effect Estimate [Lower CI, Upper CI]:
change per IQR of PM2.6
ICAM-1 - All subjects
LagO: 2.87 (-4.63,10.95)
2	dma: 2.25 (-5.15,10.22)
3	dma: 1.48 (-5.63, 9.11)
4	dma: 1.80 (-4.98, 9.07)
5	dma: 1.51 (-5.30, 8.80)
6	dma: 2.12 (-4.23, 8.89)
Subjects not known to be taking
statins
LagO: 5.47 (-3.74,15.57)
2	dma: 5.70 (-3.70,16.01)
3	dma: 4.57 (-4.31,14.27)
4	dma: 4.57 (-4.27,14.23)
5	dma: 3.80 (-4.84,13.22)
6	dma: 3.79 (-4.49,12.80)
July 2009
E-44
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Statistical Package: NR
Subjects who report smoking in the
past (but not within 6 monthsl
LagO: 0.9 (-9.56,12.66)
2	dma: 0.40 (-12.08,14.65)
3	dma: 1.34 (-9.23,13.14)
4	dma: 2.29 (-6.84,12.30)
5	dma: 1.09 (-8.30,11.44)
6	dma: 3.08 (-6.30,13.40);
Subjects who did not report smoking
in the past
Lag 0: 0.46 (-8.23, 9.97)
2	dma: 1.37 (-7.96,11.65)
3	dma: -0.96 (-10.01, 9.00)
4	dma: -1.34 (-10.35, 8.58)
5	dma: -0.87 (-10.17, 9.40)
6	dma: -1.78 (-10.64, 7.94)
VCAM-1 - All subjects
LagO: 6.88 (-2.88,17.62)
2	dma: 8.18 (-1.43,18.72)
3	dma: 6.92 (-1.66,16.25)
4	dma: 6.46 (-1.16,14.66)
5	dma: 8.57 (0.05,17.80)
6	dma: 11.76(3.48, 20.70)
Subjects not known to be taking
statins
LagO: 10.26 (-0.64, 22.35)
2	dma: 15.02(3.76, 27.49)
3	dma: 14.59(3.94,26.34)
4	dma: 15.15(4.54,26.84)
5	dma: 16.16(5.77, 27.58)
6	dma: 17.66(7.77, 28.45)
Subjects who report smoking in the
past (but not within 6 months)
LagO: 13.2 (-1.30, 29.72)
2	dma: 18.4 (0.69,39.33)
3	dma: 15.7 (1.19,32.30)
4	dma: 13.1 (0.88,26.78)
5	dma: 13.2(0.49,27.58)
6	dma: 16.2(3.76,30.10)
Subjects who did not report smoking
in the past
Lag 0: -3.12 (-12.41, 7.17)
2	dma: -0.34 (-10.57,11.05)
3	dma:-1.09 (-11.15,10.12)
4	dma:-0.81 (-10.91,10.43)
5	dma: 2.07 (-8.59,13.96)
6	dma: 4.89 (-5.56,16.50)
vWF - All subjects
LagO: 15.16 (-9.79, 47.01)
2	dma: 12.57 (-9.19, 39.55)
3	dma: 25.14 (-9.87, 73.74)
4	dma: 23.42 (-9.47, 68.25)
5	dma: 17.92 (-10.22, 54.87)
6	dma: 20.48 (-8.82, 59.22)
Subjects not known to be taking
statins
LagO: 7.40 (-19.82, 43.88)
2	dma: 7.10 (-19.09, 41.76)
3	dma: 10.78 (-17.92, 49.52)
4	dma: 11.61 (-16.64, 49.42)
5	dma: 9.15 (-20.32, 49.53)
6	dma: 7.91 (-20.70, 46.85)
Subjects who report smoking in the
past (but not within 6 months)
LagO: 19.23 (-24.29, 87.77)
2	dma: 19.92 (-29.65,104.41)
3	dma: 29.54 (-17.24,102.76)
4	dma: 41.98 (-6.95,116.63)
5	dma: 44.05 (-1.23,110.07)
6	dma: 50.39(9.35,106.82)
Subjects who did not report smoking
_iiithejjast^^^^^^_^^^^^^_
July 2009
E-45
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Lag 0
¦14.21 (-53.20, 57.24)
2 dma
¦20.66 (-63.14, 70.77)
3 dma
¦28.89 (-68.43, 60.19)
4 dma
¦23.51 (-55.11,30.34)
5 dma
¦29.18 (-60.08, 25.66)
6 dma
¦30.68 (-55.95, 9.08)
Reference: O'Neill et al. (2007, 091362) Outcome: Soluble intercellular adhesion
molecule 1 (ICAM-1)
Period of Study: May 1998-Dec 2002
Location: Boston, MA
vascular cell adhesion molecule 1 (VCAM-
1)
von Willebrand factor (vWF)
Age Groups: Mean (SD): 56.6 (10.6)
Study Design: Cross-sectional
N: 92 participants (type 2 diabetic
patients)
Statistical Analyses: Linear regression
Covariates: Apparent temperature,
season, age, race, sex, glycosylated
hemoglobin, cholesterol, smoking history,
BMI
Dose-response Investigated? No
Statistical Package: NR
Pollutant: BC
Averaging Time: 24 h (lagged moving
averages of days 0 to 1, 2, 3, 4, and 5)
Mean (SD): 1.1 (0.8)
descriptive statistics represent entire
study period
Percentiles: IQR range: 0.8
Range (Min, Max): 0.2, 5.8
Monitoring Stations: 1 site
Copollutant: PM2.6
BC
S042-
PM Increment: IQR (specific to lag
period)
Effect Estimate [Lower CI, Upper CI]: 9
change perIQR of BC
ICAM-1- -All subjects
LagO: 5.09 (-2.37,13.11)
2	dma: 3.97 (-10.24, 20.42)
3	dma: 5.10 (-10.17, 22.96)
4	dma: 8.38 (-6.46, 25.56)
5	dma: 10.09 (-7.36, 30.83)
6	dma: 10.58 (-5.34, 29.18)
Subjects not known to be taking
statins
LagO: 5.77 (-3.92,16.44)
2	dma: 2.39 (-7.65,13.52)
3	dma: 0.84 (-8.16,10.73)
4	dma: 1.67 (-6.71,10.80)
5	dma: 1.55 (-6.46,10.24)
6	dma: 2.20 (-6.47,11.68)
Subjects who report smoking in the
past (but not within 6 months)
LagO: 5.84 (0.87,11.05)
2	dma: 5.08 (-2.34,13.07)
3	dma: 4.44 (-2.70,12.11)
4	dma: 5.02 (-1.78,12.29)
5	dma: 5.89 (-2.14,14.58)
6	dma: 6.73 (-1.54,15.70)
Subjects who did not report smoking
in the past
LagO: 6.04 (0.87,11.48)
2	dma: 6.54 (-1.64,15.39)
3	dma: 5.86 (-1.90,14.22)
4	dma: 6.11 (-1.18,13.94)
5	dma: 6.89 (-1.42,15.89)
6	dma: 7.86 (-1.35,17.94)
VCAM-1 - All subjects
LagO: 9.26(2.98,15.91)
2	dma: 10.18 (1.93,19.10)
3	dma: 15.45 (2.70, 29.78)
4	dma: 17.97(3.63, 34.30)
5	dma: 23.83(8.41,41.44)
6	dma: 27.51 (11.96, 45.21)
Subjects not known to be taking
statins
LagO: 9.19(3.23,15.49)
2	dma: 14.64 (5.02, 25.14)
3	dma: 14.39(5.30, 24.28)
4	dma: 14.19(5.71,23.36)
5	dma: 19.11 (9.44,29.65)
6	dma: 22.60(11.79, 34.45)
Subjects who report smoking in the
past (but not within 6 months)
LagO: 12.4(2.77, 22.92)
2	dma: 28.5 (8.38, 52.24)
3	dma: 25.14(3.50,51.30)
4	dma: 23.1 (2.70,47.58)
5	dma: 32.0(7.29,62.30)
6	dma: 31.8(9.74,58.26)
Subjects who did not report smoking
in the past
LagO: 5.15 (-5.63,17.17)
2	dma: 2.09 (-9.07 ,14.61)
3	dma: 3.90 (-6.38,15.31)
4	dma: 4.92(4.63,15.43)
5	dma: 7.89 (-1.31,17.95)	
July 2009
E-46
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
6 dma: 10.97(0.98,21.96)
vWF- All subjects
LagO: 7.96(4.34,21.84)
2	dma: 14.87 (-2.85, 35.82)
3	dma: 15.34 (-3.22, 37.45)
4	dma: 15.47 (-7.60, 44.31)
5	dma: 19.50 (-8.89, 56.74)
6	dma: 20.53 (-9.80, 61.05)
Subjects not known to be taking
statins
LagO: 3.23 (-8.91,17.00)
2	dma: 9.82 (-8.39, 31.66)
3	dma: 17.79 (-16.03, 65.21)
4	dma: 13.14 (-18.71, 57.47)
5	dma: 16.14 (-20.43, 69.52)
6	dma: 13.25 (-22.09, 64.62)
Subjects who report smoking in the
past (but not within 6 months)
LagO: 7.63 (-17.01, 39.58)
2	dma: 37.64 (-7.18,104.10)
3	dma: 75.41 (6.16,189.85)
4	dma: 72.05 (-3.34, 206.22)
5	dma: 73.14(6.94,180.32)
6	dma: 71.23(14.00,157.19)
Subjects who did not report smoking
in the past
LagO: 10.22 (-23.14, 58.04)
2	dma: 17.07 (-18.86, 68.91)
3	dma: 6.56 (-42.75, 98.36)
4	dma: -9.20 (-65.79,140.99)
5	dma:-23.86 (-71.05,100.29)
6	dma: -48.69 (-77.75,18.29)
Reference: O'Neill et al. (2005, 088423)
Period of Study:
Baseline period: May 1998—January 2000 n'tro9'Yce™ mediated)
Outcome: Changes in vascular reactivity,
specifically percent change in brachial
artery diameter (flow-mediated and
Time trial: 2000-2002
Location: Boston, MA
N: 270 patients with diabetes or at risk of
diabetes, who participated in non-air pollu-
tion related studies at the Joselyn
Diabetes Center in Boston
Statistical Analysis: Linear regression
Pollutant: PM2.6
Mean (SD): 11.5 (6.4)
Range: 1.1-40.0
Monitoring Stations: 1
Copollutant: Sulfates
BC
Ultrafine particle counts
PM Increment: IQR (value not given)
Percent change (95% CI): PM2.6 6-day
moving avg
Nitroglycerin-mediated reactivity: -7.6%
(95% CI: 12.8% to-2.1%)
Notes: PM2.6 was positively associated
with nitroglycerin-mediated reactivity
an association was also reported with
ultrafine particles. Effect estimates were
larger in type II than type I diabetes. BC
and sulfate increases were associated
with decreased flow-mediated reactivity
among those with diabetes. Although the
largest associations were with the 6-day
moving avg, similar patterns and
quantitatively similar results appear in the
other lags.
Reference: O'Neill et al. (2007, 091362) Outcome: soluble intercellular adhesion
molecule 1 (ICAM-1)
Period of Study: May 1998-Dec 2002
Location: Boston, MA
vascular cell adhesion molecule 1 (VCAM-
1)
von Willebrand factor (vWF)
Mean Age: 56.6 (10.6)
Study Design: Cross-sectional
N: 92 participants (type 2 diabetic
patients)
Statistical Analyses: Linear regression
Covariates: Apparent temperature,
season, age, race, sex, glycosylated
hemoglobin, cholesterol, smoking history,
BMI
Dose-response Investigated? No
Statistical Package: NR
Pollutant: S042-
Averaging Time: 24 h (lagged moving
averages of days 0 to 1, 2, 3, 4, and 5)
Mean (SD): 3.0 (2.0)
descriptive statistics represent entire
study period
Percentiles: IQR range: 2.2
Range (Min, Max): 0.5, 9.6)
Monitoring Stations: 1 site
Copollutant: PM2.6, BC, S042~
PM Increment: IQR (specific to lag
period)
Effect Estimate [Lower CI, Upper CI]:
change per IQR of PM2.6
ICAM-1 All subjects
LagO: 5.30 (-2.60,13.83)
2	dma: 4.02 (-3.26,11.85)
3	dma: 4.03 (-5.34,14.34)
4	dma: -0.79 (-7.30, 6.18)
5	dma: 1.06 (-7.10, 9.93)
6	dma: 3.15 (-5.66,12.78)
Subjects not known to be taking
statins
LagO: 10.14(0.44,20.77)
2	dma: 9.39 (-1.28, 21.20)
3	dma: 10.93 (-2.23, 25.85)
4	dma: -0.24 (-9.66,10.16)
5	dma: 4.03 (-8.66,18.47)
6	dma: 5.66 (-7.52, 20.72)
Subjects who report smoking in the
July 2009
E-47
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
past (but not within 6 months)
Lag 0: -4.00 (-24.79, 22.52)
2	dma: -4.82 (-18.01,10.48)
3	dma: -7.19 (-23.66,12.83)
4	dma:-9.8 (-27.96,12.97)
5	dma:-10.4 (-29.92,14.44)
6	dma: -6.8 (-25.72,17.03)
Subjects who did not report smoking
in the past
LagO: 6.67 (-4.34,18.94)
2	dma: 5.65 (-4.67,17.10)
3	dma: 10.21 (-5.83,28.99)
4	dma: 0.80 (-9.94,12.83)
5	dma: 2.80 (-10.85,18.54)
6	dma: 5.15 (-7.78,19.89)
VCAM-1 All subjects
LagO: -0.04 (-3.75, 3.80)
2	dma: 0.94 (-4.79, 7.01)
3	dma: -0.87 (-3.50,1.82)
4	dma: 0.13 (-2.02, 2.34)
5	dma: -0.47 (-2.67,1.78)
6	dma: -0.46 (-1.99,1.09)
Subjects not known to be taking
statins
Lag 0: -1.34 (-11.23, 9.66)
2	dma:-0.19 (-11.13,12.09)
3	dma: -2.84 (-13.90, 9.64)
4	dma: 4.28 (-6.18,15.90)
5	dma:-0.26 (-13.44,14.93)
6	dma:-3.44 (-16.51,11.67)
Subjects who report smoking in the
past (but not within 6 months)
Lag 0: 0.07 (-23.40, 30.73)
2	dma:-5.62 (-20.77,12.43)
3	dma: -26.92 (-33.31 to-19.91)
4	dma:-3.06 (-28.01,30.56)
5	dma: -6.42 (-30.75, 26.47)
6	dma: -6.46 (-28.55, 22.47)
Subjects who did not report smoking
in the past
Lag 0: -3.28 (-12.66, 7.12)
2	dma:-3.17 (-11.75, 6.23)
3	dma:-9.67 (-22.07, 4.70)
4	dma: -5.51 (-14.28, 4.15)
5	dma:-12.17 (-22.05 to-1.05)
6	dma: -11.77 (-20.95 to -1.52)
vWF (sulfate measures not available)
Reference: Park et al. (2008,156845)
Period of Study: Jan 1995—Jun 2005
Location: Greater Boston area, MA
Outcome: Total homocysteine (tHcy)
Mean Age: 73.6 ± 6.9 yrs
Study Design: Cross-sectional and
longitudinal analyses performed
N:960 men
Statistical Analyses: Generalized
additive models (also hierarchical mixed-
effects regression models to assess
repeated measures of tHcy)
Covariates: Model 1: season, age, long-
term trend, apparent temperature
Model 2: further adjustment for BMI,
systolic blood pressure, smoking status,
pack years of cigarettes, alcohol consump-
tion
Model 3: further adjustment for serum
creatinine, plasma folate, vitamin B6, and
vitamin B12
Dose-response Investigated? Modeled
continuous covariates as penalized splines
to determine if association with tHcy was
Pollutant: PM2.6
Averaging Time: 24 h (moving averages
up to 7 days prior to blood collection)
Mean (SD): 12.0 (6.6)
Median: 10.6
Range (Min, Max): 2.0, 62.0
Monitoring Stations: 1 site
Copollutant: PM2.6
BC (r - 0.51)
0C (r - 0.51)
S042 (r - 0.85)
PM Increment: IQR
Effect Estimate [Lower CI, Upper CI]:
Estimated % change in tHcy per IQR
increase in pollutant.
Lag model
Concurrent day. IQR: 7.66
Model 1:1.32 (-0.83, 3.52)
Model 2:1.55 (-0.77, 3.91)
Model 3:1.57 (-0.38, 3.56)
1-day	previous. IQR: 6.91
Model 1:-1.43 (-3.51, 0.69)
Model 2:-1.41 (-3.53,0.76)
Model 3:-1.28 (-3.12, 0.60)
2-day	moving avg. IQR: 6.47
Model 1:0.04 (-2.13, 2.26)
Model 2:-0.07 (-2.26,2.17)
Model 3: 0.25 (-1.69, 2.22)
3-day	moving avg. IQR: 5.83
Model 1:-0.64 (-2.92,1.69)
Model 2:-0.74 (-3.04,1.61)
Model 3:-0.59 (-2.63,1.49)
July 2009
E-48
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
linear
Statistical Package: R software
4-day	moving avg. IQR: 5.21
Model 1:-0.63 (-2.94,1.72)
Model 2:-0.86 (-3.19,1.52)
Model 3: 0.73 (-2.78,1.37)
5-day	moving avg. IQR: 4.68
Model 1
Model 2
Model 3
0.51 (-2.79,1.83)
¦0.82 (-3.13,1.54)
¦0.84 (-2.85,1.22)
6-day moving avg. IQR: 4.50
Model 1
Model 2
Model 3
:-0.91 (-3.32,1.56)
:-1.32 (-3.76,1.17)
:-1.44 (-3.58, 0.74)
7-day moving avg. IQR: 4.20
Model 1
Model 2
Model 3
:-0.84 (-3.27,1.64)
:-1.19 (-3.64,1.33)
:-1.69 (-3.84, 0.51)
Stratified analyses: No significant
difference in effect of PM2.6 among those
with high and low levels of vitamins
Reference: Park et al. (2008,156845)
Period of Study: Jan 1995—Jun 2005
Location: Greater Boston area, MA
Outcome: Total homocysteine (tHcy)
Mean Age: 73.6 ± 6.9 yrs
Study Design: cross-sectional and
longitudinal analyses performed
N:960 men
Statistical Analyses: Generalized
additive models (also hierarchical mixed-
effects regression models to assess
repeated measures of tHcy)
Covariates: Model 1: season, age, long-
term trend, apparent temperature
Model 2: further adjustment for BMI,
systolic blood pressure, smoking status,
pack years of cigarettes, alcohol
consumption
Model 3: further adjustment for serum
creatinine, plasma folate, vitamin B6, and
vitamin B12
Dose-response Investigated? Modeled
continuous covariates as penalized splines
to determine if association with tHcy was
linear
Statistical Package: R software
Pollutant: BC
Averaging Time: 24 h (moving averages
up to 7 days prior to blood collection)
Mean (SD): 0.99 (0.56)
Median: 0.87
Range (Min, Max): 0.07, 3.7
Monitoring Stations: 1 site
Copollutant
(correlation): PM2.6 (r - 0.51)
BC
OC (r - 0.0.51)
SO42 (r - 0.50)
PM Increment: IQR
Effect Estimate [Lower CI, Upper CI]:
Estimated % change in tHcy per IQR
increase in pollutant.
Lag model Concurrent day. IQR: 0.66
Model 1:2.64 (-0.12, 5.48)
Model 2: 2.62 (-0.17, 5.48)
Model 3: 3.13 (0.76, 5.55)
1 -day previous. IQR: 0.66
Model 1:1.46 (-0.98, 3.96)
Model 2:1.32 (-1.14,3.85)
Model 3: 0.95 (-1.12, 3.05)
2-day	moving avg. IQR: 0.60
Model 1:2.75 (-0.18, 5.76)
Model 2: 2.63 (-0.33, 5.67)
Model 3: 2.59 (0.10, 5.14)
3-day	moving avg. IQR: 0.57
Model 1:2.95 (-0.44,6.46)
Model 2: 2.97 (-0.46, 6.51)
Model 3: 3.12 (0.21, 6.11)
4-day	moving avg. IQR: 0.52
Model 1:3.94 (0.24,7.78)
Model 2: 3.76(0.02,7.64)
Model 3: 3.00 (-0.13, 6.22)
5-day	moving avg. IQR: 0.49
Model 1:3.26 (-0.60,7.27)
Model 2: 2.64 (-1.23, 6.67)
Model 3: 2.38 (-0.89, 5.77)
6-day	moving avg IQR: 0.44
Model 1:1.63 (-1.99, 5.38)
Model 2:1.03 (-2.62, 4.80)
Model 3: 0.93 (-2.15, 4.11)
7-day	moving avg. IQR: 0.44
Model 1:1.38 (-2.45, 5.36)
Model 2: 0.69 (-3.16, 4.70)
Model 3: 0.45 (-2.81,3.83)
% change in tHcy per IQR increase in BC,
24-h avg
Among those with low folate: 5.31 (2.26,
8.42)
Among those with low B12: 5.06 (2.03,
8.17)
nearly null associations among those with
high levels
July 2009
E-49
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Park et al. (2008,156845)
Period of Study: Jan 1995—Jun 2005
Location: Greater Boston area, MA
Outcome: Total homocysteine (tHcy)
Mean Age: 73.6 ± 6.9 yrs
Study Design: Cross-sectional and
longitudinal analyses performed
N:960 men
Statistical Analyses: Generalized
additive models (also hierarchical mixed-
effects regression models to assess
repeated measures of tHcy)
Covariates: Model 1: season, age, long-
term trend, apparent temperature
Model 2: further adjustment for BMI, sys-
tolic blood pressure, smoking status, pack
years of cigarettes, alcohol consumption
Model 3: further adjustment for serum
creatinine, plasma folate, vitamin B6, and
vitamin B12
Dose-response Investigated? Modeled
continuous covariates as penalized splines
to determine if association with tHcy was
linear
Statistical Package: R software
Pollutant: 0C
Averaging Time: 24 h (moving averages
up to 7 days prior to blood collection)
Mean (SD): 3.5(1.8)
Median: 3.1
Range (Min, Max): 0.29,11.8
Monitoring Stations: 1 site
Copollutant (correlation): PM2.6
(r - 0.51)
BC (r - 0.51)
0C
SO42 (r - 0.41)
PM Increment: IQR
Effect Estimate [Lower CI, Upper CI]:
Estimated % change in tHcy per IQR
increase in pollutant.
Lag model
Concurrent day. IQR: NA
Model 1: NA
Model 2: NA
Model 3: NA
1-day	previous. IQR: 2.00
Model 1:2.12 (-0.98, 5.31)
Model 2:1.69 (-1.51,5.00)
Model 3:1.87 (-0.81,4.62)
2-day	moving avg. IQR: 1.93
Model 1
Model 2
Model 3
0.39 (-3.67, 3.01)
¦0.88 (-4.26, 2.61)
1.05 (-1.86, 4.06)
3-day	moving avg. IQR: 1.68
Model 1:0.53 (-2.66, 3.83)
Model 2: 0.14 (-3.15, 3.54)
Model 3:1.32 (-1.44,4.16)
4-day	moving avg. IQR: 1.64
Model 1:1.57 (-1.89, 5.15)
Model 2:1.42 (-2.14,5.12)
Model 3:1.89 (-1.15, 5.03)
5-day	moving avg, IQR: 1.60
Model 1:2.27 (-1.49, 6.16)
Model 2: 2.11 (-1.77, 6.15)
Model 3: 2.12 (-1.29, 5.65)
6-day	moving avg. IQR: 1.43
Model 1:2.83 (-0.74,6.52)
Model 2: 2.78 (-0.90, 6.60)
Model 3: 2.53 (-0.59, 5.74)
7-day	moving avg. IQR: 1.23
Model 1:2.75 (-0.41, 6.02)
Model 2: 2.55 (-0.71,5.92)
Model 3: 2.55 (-0.21,5.39)
% change in tHcy per IQR increase in OC,
7-d avg.
Among those with low B12: 5.23 (1.59,
9.01)
nearly null associations among those with
high levels
July 2009
E-50
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Park et al. (2005, 057331)
Period of Study: November 2000-
October 2003
Location: Greater Boston area, MA
Outcome: Change in HRV (SDNN, HF, LF,
LFHFR)
Mean age: 72.7 years
Study Design: Cross-sectional
N: 497 adult males living in the Greater
Boston, MA area
Pollutant: PM2.6
Averaging Time: 4 h
24 h
48 h
Mean (SD): 11.4 (8.0)
Range: 6.45-62.9
Copollutant: 0:i, Particle number count,
BC, NO?, SO?, CO
PM Increment: 8 |Jgfm3
Percent change (95% CI): 48h mean
PM2.6: 20.8% decrease in HF (95% CI:
4.6%, 34.2%)
18.6% increase in LFHFR (4.1%, 35.2%).
Notes: Subjects were monitored during a
4-min rest period between 8 a.m. and 1
p.m. Modifying effects of hypertension,
IHD, diabetes, and use of cardiac/anti-
hypertensive medications also examined.
Linear regression analyses. This subject
group is from the VA Normative Aging
Study. The 4-h averaging period was most
strongly associated with HRV indices. The
PM effect was robust in models including
O3. The HRV change per IQR increase in
PM2.E were larger in subjects with
hypertension (n - 335) IHD (n - 142),
and diabetes (n - 72). In addition, those
who did not use calcium-channel blockers
had a greater decline in LF associated with
each IQR increase in PM2.6 than did those
who did use calcium channel blockers. IQR
increases in 48h mean BC concentration
were also associated with adverse
changes in HRV, suggesting traffic
pollution may be particularly toxic.
Reference: Park et al. (2006, 091245)
Period of Study: November 2000-
December 2004
Location: Greater Boston area, MA
Outcome: Change in HF
Study Design: Cross-sectional
N: Statistical Analysis: Linear
regression models
Pollutant: PM2.6
Averaging Time: 48 h
Mean (SD): PM2.B: 11.7 (7.8)
Sulfates: 3.3 (3.3)
BC: 0.92 (0.46)
Copollutant: O3
PM Increment: 10 |Jgfm3
Percent change (95% CI): Wild-type HFE
genotype: 31.7% (95% CI: 10.3, 48.1)
Among those with either of the two HFE
variants, there was no association
between 48h PM2.6 and HF (shown in a
graph, —10% non-significant increase).
Notes: Normative Aging Study. Examining
association between PM and HF among
those with and without the wild-type HFE
genotype.
Reference: Pekkanen et al. (2002,
035050)
Period of Study: Winter 1998 to 1999
Location: Helsinki, Finland
Outcome: ST Segment Depression
(> 0.1 mV)
Study Design: Panel of ULTRA Study
participants
N: 45 Subjects, n - 342 biweekly
submaximal exercise tests, 72 exercise
induced ST Segment Depressions
Statistical Analysis: Logistic regression
I GAM
Pollutant: PM2.6
Averaging Time: 24 h
Median: 10.6
IQR: 7.9
Pollutant: PMi
Median: 7.0
IQR: 5.6
Pollutant: ACP (100 to 1000nm) (n/cm3)
Median: 1200
IQR: 760
Copollutant: NO2, CO, PM 10 2.5, ultrafine
PM Increment: IQR
Effect Estimate(s): ACP: OR - 3.29
(1.57, 6.92), lag 2
PM1: OR - 4.56 (1.73,12.03), lag 2
PM2.6: OR - 2.84(1.42, 5.66), lag 2
Notes: The effect was strongest for ACP
and PM2.E, which in two pollutant models
appeared independent. Increases in NO2
and CO were also associated with
increased risk of ST segment depression,
but not with coarse particles.
July 2009	E-51	DRAFT - DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Park et al. (2008,156845)
Period of Study: Jan 1995—Jun 2005
Location: Greater Boston area, MA
Outcome: Total homocysteine (tHcy)
Mean Age: 73.6 ± 6.9 yrs
Study Design: Cross-sectional and
longitudinal analyses performed
N: 960 men
Statistical Analyses: Generalized
additive models (also hierarchical mixed-
effects regression models to assess
repeated measures of tHcy)
Covariates: Model 1: season, age, long-
term trend, apparent temperature
Model 2: further adjustment for BMI,
systolic blood pressure, smoking status,
pack years of cigarettes, alcohol consump-
tion
Model 3: further adjustment for serum
creatinine, plasma folate, vitamin B6, and
vitamin B12
Dose-response Investigated? Modeled
continuous covariates as penalized splines
to determine if association with tHcy was
linear
Statistical Package: R software
Pollutant: S042
Averaging Time: 24 h (moving averages
up to 7 days prior to blood collection)
Mean (SD): 3.2 (3.0)
Median: 2.4
Range (Min, Max): 0.39, 29.0
Monitoring Stations: 1 site
Copollutant (correlation): PM2.6
(r - 0.85)
BC (r - 0.50)
0C (r - 0.41)
SO42-
PM Increment: IQR
Effect Estimate [Lower CI, Upper CI]:
Estimated % change in tHcy per IQR
increase in pollutant.
Lag model
Concurrent day: IQR: NA
Model 1: NA
Model 2: NA
Model 3: NA
1 -day previous: IQR: 2.61
Model 1:0.91 (-0.77, 2.62)
Model 2: 0.99 (-0.94, 2.95)
Model 3: 0.91 (-0.72, 2.57)
2-day moving avg: IQR: 2.10
Model 1
Model 2
Model 3
0.25 (-2.07,1.60)
¦0.29 (-2.35,1.82)
0.05 (-1.74,1.86)
3-day moving avg: IQR: 1.73
Model 1
Model 2
Model 3
0.15(-1.97,1.69)
¦0.17 (-2.23,1.93)
¦0.01 (-1.78,1.80)
4-day moving avg: IQR: 1.64
Model 1
Model 2
Model 3
¦0.69 (-2.74,1.41)
¦0.60 (-2.95,1.81)
¦0.58 (-2.63,1.51)
5-day	moving avg: IQR: 1.60
Model 1:-1.14 (-3.53,1.30)
Model 2:-0.90 (-3.64,1.92)
Model 3:-1.09 (-3.48,1.36)
6-day	moving avg;' IQR: 1.40
Model 1:0.00 (-2.39, 2.44)
Model 2: 0.36 (-2.36, 3.16)
Model 3: 0.41 (-2.01,2.89)
7-day	moving avg
IQR: 1.30
Model 1
Model 2
Model 3
¦0.16 (-2.51, 2.24)
0.30 (-2.37, 3.04)
0.07 (-2.25, 2.43)
Stratified analyses: No significant
difference in effect of S042- among those
with high and low levels of vitamins
Reference: Peters et al. (2005, 095747) Outcome: Myocardial infarction
Also Peters et al, 2005 (2005,156859)
Period of Study: February 1999-July
2001
Location: Augsburg, Germany
Study Design: Case-crossover
N: 691 myocardial infarction patients
Statistical Analysis: Conditional logistic
regression
Dose-response Investigated? No
Pollutant: PM2.6
Averaging Time: 1 h: Median - 14.5
IQR: 9.1
24-h: Median - 14.9
IQR: 7.7
Copollutant: NO2, SO2, CO
Effect Estimate: 2-h lag: OR - 0.93
95% CI: 0,83,1.04
24-h mean, 2-day lag: OR - 1.18
95% CI: 1.03,1.34
Notes: Examined triggering for Ml at
various lags before Ml onset (up to 6 h
before Ml, up to 5 days before Ml). PM2.6
levels 2 days before Ml onset were
associated with increased risk of Ml, but
not on the concurrent day, or lags 1, 3, 4,
or 5. These findings are consistent with
the prior Boston Ml study for a 1 - to 2-day
lagged effect of PM2.6.
July 2009
E-52
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Pope et al. (2004, 055238)
Period of Study: Winter 1999-2000 (in
Wasatch Front, UT). Summer 2000 (in
Hawthorne, UT).
Winter 2000-2001 (in Bountiful, UT and
Lindon, UT)
Location: Utah: Wasatch Front,
Hawthorne, Bountiful, and Lindon
Period of Study: 1994- 2004
Location: Wasatch Front, Utah
Reference: Pope et al. (2004, 055238)
Period of Study: 1999-2001
Location: Wasatch Front, Utah
Reference: Rich et al. (2005, 079620)
Period of Study: July 1995—July 2002
Location: Eastern Massachusetts, USA
Outcome: Change in autonomic function
(measured by changes in HRV), C-reactive
protein (CRP), blood cell counts, platelets,
and blood viscosity associated with short-
term changes in PM2.6
Age Groups: Elderly (specific age range
not given)
Study Design: Panel study
N: 88 elderly subjects
Statistical Analysis: Linear regression
Season: Winter, summer
Dose-response Investigated? No
Study Design: Case-crossover study
(time-stratified control selection)
N: Statistical Analysis: Conditional
logistic regression
Outcome: Heart rate variability (HRV)
C-reactive protein (CRP)
blood cell counts, whole blood viscosity
Age Groups: 54-89 yrs
Study Design: Panel study
N: 88 participants
Statistical Analyses: Linear regression
Covariates: Subject-specific fixed effects
interactive spline smooths for temp, RH
(partial control for H)
Season: Temperature as covariate
Dose-response Investigated?
Yes, also assessed PM by including cubic
smoothing splines with 3 df
Statistical Package: SAS
Outcome: Confirmed ventricular
arrhythmias
Study Design: Case-crossover (time-
stratified control selection)
N: 203 patients with implantable
cardioverter defibrillators
Statistical Analysis: Conditional logistic
regression
Pollutant: PM2.6 (TE0M)
Averaging Time: 24 h
Mean (SD): 18.9 (13.4)
Copollutant: None
Pollutant: PM2.6 (FRM)
Averaging Time: 24 h
Mean (SD): Site 1:10.1
Site 2: 10.8
Site 3: 11.3
Monitoring Stations: 3
Copollutant: PM10 (FRM) measured at 4
monitoring sites
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD): 23.7 (20.2)
Range (Min, Max): 1.7, 74.0
Monitoring Stations: NR
Copollutant: None
Pollutant: PM2.6 (TE0M)
Averaging Time: 1-h avg
24-h avg
Median (IQR): 1-h avg: Median - 9.2
/yg/m3
24-h avg: Median - 9.8 //g|m3
IQr - 7.8
Copollutant: O3, BC, CO, NO2, SO2
PM Increment: 100/yg/m3
Effect Estimate: Each 100 |Jgfm3
increase associated with: -35 (SE - 8)
msec decline in SDNN
0.81 (SE 0.17) mg/dL increase in CRP
0.31 (SE 9.34) k///L increase in platelets
0.07 (SE 0.21) cP increase in blood
viscosity
Notes: The study observed small but
statistically significant adverse
associations between daily mean PM2.6
and HRV and C-reactive protein (CRP). The
authors point out, however, that most of
the variability in the temporal deviation of
these physiological endpoints was not
explained by PM2.6. These observations
therefore suggest that PM2.6 may be one
of multiple factors that influence HRV and
CRP.
PM Increment: 10/yg/m3
Effect Estimate: For same-day increase
in PM2.6: OR - 1.045
95% CI: 1.011,1.080
Notes: Case-crossover study (time-
stratified control selection) triggering of
acute ischemic heart disease by ambient
PM2.E concentrations on the same and
previous 3 days. PM2.6 measured at 3
sites and estimated for missing days.
Effect estimates were larger for those
with angiographically demonstrated
coronary artery disease.
PM Increment: 100/yg/m3
Effect Estimate [Lower CI, Upper CI]:
Regression coefficients (SE) for
associations with concurrent day
pollutant: Mean H: -4.49 (1.73)
SDNN: -34.94 (8.32)
SDANN:-18.98 (8.67)
r-MSSD:-42.25 (10.90)
CRP: 0.81 (0.18)
Whole blood viscosity: 0.07 (0.21)
WBC: -0.07 (0.38)
Granulocytes: 0.02 (0.37)
Lymphocytes: -0.07 (0.14)
Monocytes: 0.12 (0.04)
Basophils: -0.01 (0.01)
Eosinophils: -0.01 (0.02)
RBC: 0.03 (0.06)
Platelets: 0.31 (9.34)
PM Increment: 7.8/yg/m3
Effect Estimate: For mean PM2.6 in the
24 h before ventricular arrhythmia: OR
- 1.19
95% CI: 1.02,1.38
Notes: 794 ventricular arrhythmias
among 84 subjects.
Lag h: 0-2, 0-6, 0-23, 0-47
Reference: Pope et al. (2006, 091246) Outcome: Acute ischemic heart disease
July 2009
E-53
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Rich et al. (2006, 088427)
Period of Study: July 1995—July 2002
Location: Eastern Massachusetts, USA
Outcome: Confirmed episodes of
paroxysmal atrial fibrillation
Study Design: Case-crossover (time-
stratified control selection)
N: 203 patients with implantable
cardioverter defibrillators
Statistical Analysis: Conditional logistic
regression
Pollutant: PM2.6 (TE0M)
Averaging Time: 1 h avg
24-h avg
Median (IQR): 1-h avg: Median - 9.2
^g/m3
24-h avg: Median - 9.8 //g|m3
IQr - 7.8
Copollutant: O3, BC, CO, NO2, SO2
PM Increment: 9.4/yg/m3
Effect Estimate: O h lag: OR 1.41 (0.82,
2.42)
Notes: 91 paroxysmal atrial fibrillation
(PAF) episodes among 29 subjects.
Lag h: 0, 0 ¦ 23
Positive, but not significant increases in
the relative odds of PAF associated with
PM2.E concentrations in the same h and
24-h before PAF episode onset. Authors
note reduced statistical power for PM2.6
analyses due to missing data.
Reference: Rich et al. (2006, 088427)
Period of Study: July 1995—July 2002
Location: Eastern Massachusetts, USA
Outcome: Confirmed episodes of
paroxysmal atrial fibrillation
Study Design: Case-crossover (time-
stratified control selection)
N: 203 patients with implantable
cardioverter defibrillators
Statistical Analysis: Conditional logistic
regression
Pollutant: BC
Averaging Time: 1-h avg
24-h avg
Median (IQR): IQR: 0.91//gjm3
Copollutant: O3, PM2.6, CO, NO2, SO2
PM Increment: 0.91/yg/m3 (IQR)
Effect Estimate: 0- to 23-h lac
OR 1.46 (95% CI: 0.67, 3.17)'
Notes: 91 paroxysmal atrial fibrillation
(PAF) episodes among 29 subjects.
Lag h: 0, 0-23
Positive, but not significant increases in
the relative odds of PAF associated with
BC concentrations in the same h and 24 h
before PAF episode onset. Authors note
reduced statistical power for BC analyses
due to missing data.
Reference: Rich et al. (2006, 089814)
Period of Study: May 2001-December
2002
Location: St. Louis, M0 metropolitan area
Outcome: Confirmed ventricular
arrhythmia
Study Design: Case-crossover design
(time-stratified control selection)
Dose-response Investigated? No
Pollutant: PM2.6 (CAMM)
Averaging Time: 24 h
Median (IQR): 16.2/yglm3 (IQr - 9.7)
Copollutant: NO2, SO2, CO, O3, EC, 0C
PM Increment: 9.7/yglm3 (IQR)
Effect Estimate: OR (PM2 b) - 0.95 (95%
CI: 0.72,1.27)
OR (SO2) - OR - 1.24(95% CI: 1.07,
1.44)
Notes: 139 confirmed ventricular
arrhythmia episodes among 56 subjects.
Lags: 0-2h, 0-6h, 0-11 h, 0-23h, 0-47h
Authors did not find increased relative
odds of VA associated with each IQR
increase in 24-h mean PM2.6, but did find
non-significantly increased relative odds of
VA associated with 24-h EC. Shorter and
longer lag times' relative odds estimates
provided no evidence of immediate
ventricular arrhythmic effects of air
pollution.
Reference: Rich et al. (2004, 055631)
Period of Study: February-December
2000
Location: Vancouver, British Columbia,
Canada
Outcome: ICD discharges (as a proxy for
VT/VF)
Age Groups: 15-85 years
Study Design: Case-crossover design
(ambidirectional control selection ± 7
days)
N: 34 patients with implantable
cardioverter defibrillators
Statistical Analysis: Conditional logistic
regression
Dose-response Investigated? No
Pollutant: PM2.6 (Partisol)
Averaging Time: 1 h
Mean (SD), IQR: Mean:: 8.2//gfm3
(SD - 10.7)
IQr - 5.2
PM Increment: Effect Estimate: Odds
ratios were less than 1.0 at all lags (0,1,
2, 3) for PM2.6.
No consistent association between any of
the air pollutants and implantable
cardioverter defibrillators discharges.
Copollutant: O3, EC, 0C, S0r , CO, NO2,
SO2, PM10
PM10: Mean:: 13.3/yg/m3
(SD - 4.9)
IQr - 7.4
Notes: Same study as Vedal et al. (2004,
055630), except Rich (2004) used data
from a shorter time period so as to
estimate relative odds of ICD discharge
associated with acute increases in more
pollutants than Vedal (2004, 055630).
July 2009
E-54
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Rich et al. (2008,156910)
Period of Study: NR
Location: New Jersey
Outcome: Pulmonary Artery and Right
Ventricular Pressures
Age Groups: 25-68
Study Design: panel
N: 11 subjects
Statistical Analyses:
Measures
Covariates: long-term trends, calendar
month, weekday, apparent temperature
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-6d
Pollutant: PM2.6
Averaging Time: 24h
Mean (SD): NR
Monitoring Stations: NR
Copollutant: NR
Co-pollutant Correlation
n/a
PM Increment: 11.62/yg/m3
Change (Lower CI, Upper CI), p-value:
ePAD: 0.19(0.05, 0.33), 0.01
RV diastolic pressure: 0.23 (0.11, 0.34),
<0.001
RV systolic pressure: 0.12 (-0.07, 0.31),
0.23
MPAP: 0.12 (-0.05, 0.28), 0.161
Reference: Riedikeret al. (2004,
091261)
Period of Study: Fall 2001
Location: Wake County, North Carolina
Outcome: Heart rate variability (measured
10 h after shift): mean cycle length of
normal R-R intervals (MCL), the standard
deviation of normal R-R intervals (SDNN),
and percentage of normal R-R interval
differences greater than 50 msec
(PNN50), low frequency (0.04-0.15Hz),
high frequency (0.15-0.40Hz), the ratio of
low to high frequency.
Blood analysis (measured 15 h after
shift): Uric acid, blood urea nitrogen,
gamma glutamyl transpeptidase, white
blood cell count, red blood cell count,
hematocrit, hemoglobin, mean red blood
cell volume (MCV), neutronphils (count and
%), lymphocytes (count and %), C-reactive
protein, plasminogen, plasminogen
activator inhibitor type 1, von Willebrand
factor (vWF), endovthzelin-1, protein C,
and interleukin-6
Pollutant: In-vehicle PM2.6 components
identified with factor analysis (crustal
material, wear of steel automotive
components, gasoline combustion, speed-
changing traffic with engine emissions and
brake wear
Age Groups: 23-30 yrs
Study Design: Panel
N: 9 healthy male troopers, repeated
measures (36 person-days)
Statistical Analyses: Mixed effects
regression models (principal factor
analysis for classification of exposure)
Covariates: Potential confounders:
temperature, relative humidity, number of
law-enforcement activities during the shift
and the avg speed during the shift
controlling had no effect on effect esti-
mates for "crustal" and "speed-change"
factors
however, confounder inclusion in the
"speed change" and blood urea nitrogen
and vWF reduced the effect estimate and
the CI included zero
Season: Only 1 season included
Dose-response Investigated? No
Statistical Package: S-Plus 6.1
Averaging Time: Exposure
during 3pm to 12am work shifts
Mean: PM2.Bmass - 23.0/yg/m3
Monitoring Stations: Per vehicle
Copollutant (correlation): Correlation to
PM2.6Mass
Benzene: r - 0.50
Aldehydes: r - 0.34
CO: r - 0.52
Aluminum: r - 0.58
Silicon: r - 0.66
Sulfur: r - 0.58
Calcium: r - 0.37
Titanium: r - 0.41
Chromium: r - 0.51
Iron: r - 0.71
Copper: r - 0.16
Selenium: r - 0.38
Tungsten: r - 0.37
PM2.Lightscatter: r - 0.71
PM Increment: 1 SD change in source
factor
Effect Estimate: % change in the health
outcome per 1 SD change in the "speed
change"factor
MCL: 7%
HRV: 16%
supraventricular ectopic beats: 39%
% Neutrophils: 7%
% lymphocytes: -10%
red blood cell volume MCV: 1 %
vWF: 9%
blood urea nitrogen: 7%
protein C: -11%
% change in the health outcome per 1 SD
change in the "crustal" factor
MCL: 3% serum uric acid concentrations:
5%
Note: Results (including CIs) are reported
in figures 2 & 3.
July 2009
E-55
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Riojas-Rodriguez et al. (2006,
156913)
Period of Study: December 2001 —April
2002
Location: Mexico City metropolitan area
Outcome: Heart rate variability (5-minute
periods)
Study Design: Panel study
N: 30 patients from the outpatient clinic
of the National Institute of Cardiology of
Mexico, where each subject had existing
ischemic heart disease.
Statistical Analysis: Mixed models
Pollutant: PM2.B (nephelometry)
Averaging Time: 5 minutes
Mean (SD), Range: 46.8//gfm3
(SD - 1.82)
Range: 0-483 //g|m3
Copollutant: CO
PM Increment: 10/yg/m3
Effect Estimate: Each 20 //gfm3 increase
in 5 minute PM2.6 was associated with a: -
0.008 decrease in the ln(HF)(95% CI: ¦
0.015, 0.0004
Notes: Population of subjects with known
ischemic heart disease (25 men and 5
women who had at least 1 prior Ml [not in
last 6 months])
Each 10 /yglm3 increase in 5 minute mean
PM2.E was associated with non-
significantly decreased HF, and with
similar, but smaller changes in LF and VLF.
Reference: Romieu et al. (2005, 086297) Outcome: Heart rate variability (HF, LF,
VLF, PNN50, SDNN, r-MSSD)
Period of Study: 2000-2001
Location: Mexico City, Mexico
Age Groups: > 60 years of age
Study Design: Double blind randomized
controlled trial
N: 50 elderly residents of a Mexico City
nursing home
Pollutant: PM2.6
Averaging Time: 24 h
Copollutant: O3, NO2, SO2, PM10
PM Increment: 8 /yglm3
Effect Estimate: In the group receiving
the fish oil supplement, each 8 //gf'nv1
change in 24 h mean total exposure PM2.6
was associated with a: a) 54% reduction
(95% CI:-72% to-24%) in HF (log
transformed) in the pre-supplementation
phase
b) 7% reduction (95% CI: -20%, 7%) in the
supplementation phase.
Changes in other HRV parameters were
also smaller in the supplementation phase.
In the group receiving soy oil
supplementation, the % reduction in HF
was also smaller in the supplementation
phase, but the differences were smaller
and not statistically significant.
Notes: Study of the effect of omega-3-
fatty acid supplementation (2 g/day of fish
oil versus 2 g/day of soy oil) to mitigate
the effect of ambient PM2.6 on HRV.
Subjects had no cardiac arrhythmias,
cardiac pacemakers, allergies to omega-3
fatty acids or fish, treatment with oral
anticoagulants, or history of bleeding
diathesis. PM2.6 was measured and
estimated indoors, outdoors, and with
regards to total exposure (the same as
Holguin et al. (2003)).
July 2009
E-56
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Romieu et al. (2008,156922) Outcome: Copper/zinc superoxide
dismutase activity (Cu/Zn SOD)
Period of Study: Sep 2001-Apr 2002
Location: Mexico City, Mexico
lipoperoxidation (LPO)
reduced glutathione (GSH)
Age Groups: 60-96 yrs
Study Design: Intervention (randomly
assigned fish oil or soy oil)
N: 52 participants
Statistical Analyses: Linear mixed
models
Covariates: Time
Dose-response Investigated? Assessed
possible nonlinearity using generalized
additive mixed models with p-splines
Statistical Package: STATA v8.2 and
SAS v9.1
Pollutant: PM2.6 (indoor)
Averaging Time: 24 h (same day)
Mean (SD): 38.7(14.7)
Percentiles: 25th: 30.62
50th: 35.11
75th: 41.10
Range (Min, Max): 14.8, 70.9
Monitoring Stations: Indoor measured
inside nursing home
Copollutant: 0:i
PM Increment: 10/yg/m3
Effect Estimate [Lower CI, Upper CI]:
Regression coefficient (SE
p-value):
Cu/Zn SOD: -0.05 (0.02
0.001)
LPO (square root transformed): 0.08 (0.09
0.381)
GSH (log-transformed
quadratic term for PM): -0.05 (0.01
0.002)
Regression coefficient (SE
p-value) by supplementation groups (same
transformations as above): Cu/Zn SOD
Soy Oil:-0.06 (0.02, <0.001)
Fish Oil: * 0.04 (0.02, 0.009)
LPO
Soy Oil:-0.02 (0.14,0.904)
Fish Oil: * 0.16 (0.07, 0.024)
GSH
Soy Oil: -0.03 (0.04, 0.406)
Fish Oil:-0.09 (0.04, 0.017)
"Quadratic term for PM
Reference: Ruckerl et al. (2007,156931) Outcome: lnterleukin-6 (IL-6), fibrinogen,
C-reactive protein (CRP)
Period of Study: May 2003—Jul 2004
Location: Athens, Augsburg, Barcelona,
Helsinki, Rome, and Stockholm
Age Groups: 35-80 yrs
Study Design: Repeated measures I
longitudinal
N: 1003 Ml survivors
Statistical Analyses: Mixed-effect
models
Covariates: City-specific confounders
(age, sex, BMI)
long-term time trend and apparent
temperature
RH, time of day, day of week included if
adjustment improved model fit
Season: Long-term time trend
Dose-response Investigated? Used p-
splines to allow for nonparametric
exposure-response functions
Statistical Package: SAS v9.1
Pollutant: PM2.6
Averaging Time: Hourly and 24-h (lag 0-
4, mean of lags 0-4, mean of lags 0-1,
mean of lags2-3, means of lags 0-3)
Mean (SD): Presented by city only
Monitoring Stations: Central monitoring
sites in each city
Copollutant: SO2
03
NO
NO?
PM Increment: IQR
Effect Estimate [Lower CI, Upper CI]:
change in mean blood markers per
increase in IQR of air pollutant.
IL-6
Lag (IQR): % change in GM (95%CI)
Lag 0(11.0): 0.46 (-0.89,1.83)
Lag 1 (11.0):-0.39 (-1.69, 0.93)
Lag 2 (11.0):-0.23 (-1.53,1.07)
5-davg (8.6): 0.05 (-1.37,1.50)
Fibrinogen
Lag (IQR): % change in AM (95%CI)
Lag 0(11.0): 0.05 (-0.48, 0.58)
Lag 1 (11.0): 0.17 (-0.35, 0.69)
Lag 2 (11.0): 0.20 (-0.32, 0.71)
5-davg (8.6): 0.38 (-0.21, 0.96)
CRP
Lag (IQR): % change in GM (95%CI)
Lag 0(11.0): 0.11 (-1.95, 2.21)
Lag 1 (11.0):-0.06 (-1.98,1.90)
Lag 2 (11.0): 0.11 (-1.80, 2.06)
5-davg (8.6):-0.13 (-2.15,1.92)
July 2009
E-57
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ruckerl et al. (2006, 088754)
Period of Study: Oct 2000—Apr 2001
Location: Erfurt, Germany
Age Groups: 50 +
Study Design: Panel (12 repeated
measures at 2-wk intervals)
N: 57 male subjects with coronary disease
Statistical Analyses: Fixed effects linear
and logistic regression models
Covariates: Models adjusted for different
factors based on health endpoint
CRP: RH, temperature, trend, ID
ICAM-1: temperature, trend, ID
vWF: air pressure, RH, temperature, trend,
ID
FVII: air pressure, RH, temperature, trend,
ID,weekday
Season: Time trend as covariate
Dose-response Investigated? Sensitivity
analyses examined nonlinear exposure-
response functions
Statistical Package: SAS v8.2 and S-
Plus v6.0
Outcome: C-reactive protein (CRP)
serum amyloid A (SAA)
E-se lectin
von Willebrand Factor (vWF)
intercellular adhesion molecule-1 (ICAM-1)
fibrinogen
Factor VII
prothrombin fragment 1 +2
D-dimer
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD): 20.0(15.0)
Percentiles: 2th5: 9.7
50th: 14.9
75th: 26.1
Range (Min, Max): 2.6, 83.7
Monitoring Stations: 1 site
Copollutant: UFPs (ultrafine particles)
AP (accumulation mode particles)
PM2.6
PM10
0C (organic carbon)
EC (elemental carbon)
NO?
CO
PM Increment: IQR (16.4
5-davg: 12.2)
Effect Estimate [Lower CI, Upper CI]:
Effects of air pollution on blood markers
presented as OR (95%CI) for an increase
in the blood marker above the 90th
percentile per increase in IQR air pollutant.
CRP
Time before draw: 0 to 23 h: 1.1 (0.7,
1.8)
24 to 47 h: 1.5(0.9, 2.5)
48 to 71 h: 1.2(0.8,1.9)
5-d mean: 1.4(0.9, 2.3)
ICAM-1
Time before draw: 0 to 23 h: 0.7 (0.4,
0.9)
24 to 47 h: 1.3 (0.8,1.8)
48 to 71 h: 1.8(1.2, 2.7)
5-d mean: 1.1 (0.8, 1.5)
Effects of air pollution on blood markers
presented as % change from the meanfGM
in the blood marker per increase in IQR air
pollutant.
vWF
Time before draw: 0 to 23 h: 3.9 (-0.3,
8.1)
24 to 47 h: 3.1 (-1.6,7.8)
48 to 71 h: 3.6 (-1.1, 8.3)
5-d mean: 5.6 (0.5,10.8)
FVII
Time before draw: 0 to 23 h: -2.5 (-6.2,
1.4)
24 to 47 h:-2.8 (-6.1, 0.6)
48 to 71 h: -2.3 (-5.0, 0.6)
5-d mean: -3.5 (-6.4 to -0.4)
Note: Summary of results presented in
figures. SAA results indicate increase in
association with PM (not as strong and
consistent as with CRP)
no association observed between E-
selectin and PM
an increase in prothrombin fragment 1 +2
was consistently observed, particularly
with lag 4
fibrinogen results revealed few significant
associations, potentially due to chance
D-dimer results revealed null associations
in linear and logistic analyses
July 2009
E-58
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ruckerl et al. (2006, 088754)
Period of Study: Oct 2000—Apr 2001
Location: Erfurt, Germany
Age Groups: 50+ yrs
Study Design: Panel (12 repeated
measures at 2-wk intervals)
N: 57 male subjects with coronary disease
Statistical Analyses: Fixed effects linear
and logistic regression models
Covariates: Models adjusted for different
factors based on health endpoint
CRP: RH, temperature, trend, ID
ICAM-1: temperature, trend, ID
vWF: air pressure, RH, temperature, trend,
ID
FVII: air pressure, RH, temperature, trend,
ID,weekday
Season: Time trend as covariate
Dose-response Investigated? Sensitivity
analyses examined nonlinear exposure-
response functions
Statistical Package: SAS v8.2 and S-
Plus v6.0
Outcome: C-reactive protein (CRP)
serum amyloid A (SAA)
E-se lectin
von Willebrand Factor (vWF)
intercellular adhesion molecule-1 (ICAM-1)
fibrinogen
Factor VII
prothrombin fragment 1 +2
D-dimer
Pollutant: EC
Averaging Time: 24 h
Mean (SD): 2.6 (2.4)
Percentiles: 25th: 1.0
50th: 1.8
75th: 3.2
Range (Min, Max): 0.2,12.4
Monitoring Stations: 1 site
Copollutant: UFPs (ultrafine particles)
AP (accumulation mode particles)
PM2.6
PM10
0C (organic carbon)
EC (elemental carbon)
NO?
CO
PM Increment: IQR (2.3
5-d avg: 1.8)
Effect Estimate [Lower CI, Upper CI]:
Effects of air pollution on blood markers
presented as OR (95%CI) for an increase
in the blood marker above the 90th
percentile per increase in IQR air pollutant.
CRP
Time before draw: 0 to 23 h: 1.2 (0.7,
2.0)
24 to 47 h: 1.3(0.7, 2.4)
48 to 71 h: 1.6(0.9,2.7)
5-d mean: 1.2(0.7,2.1)
ICAM-1
Time before draw: 0 to 23 h: 1.0 (0.7,
1.6)
24 to 47 h: 2.6(1.7,3.8)
48 to 71 h: 4.0 (2.5, 6.1)
5-d mean: 2.2 (1.4, 3.3)
Effects of air pollution on blood markers
presented as % change from the meanfGM
in the blood marker per increase in IQR air
pollutant.
vWF
Time before draw: 0 to 23 h: 5.0 (0.0,
10.1)
24 to 47 h: 7.6 (1.4,13.7)
48 to 71 h: 1.1 (-5.2,7.4)
5-d mean: 5.7 ( 0.5,12.0)
FVII
Time before draw: 0 to 23 h: -5.7 (-10.5
to -0.7)
24 to 47 h: -6.9 (-11.2 to -2.3)
48 to 71 h: -4.2 (-8.4, 0.2)
5-d mean: -6.0 (-10.5 to -1.2)
Note: Summary of results presented in
figures. SAA results indicate increase in
association with PM (not as strong and
consistent as with CRP)
no association observed between E-
selectin and PM
an increase in prothrombin fragment 1 +2
was consistently observed, particularly
with lag 4
fibrinogen results revealed few significant
associations, potentially due to chance
D-dimer results revealed null associations
in linear and logistic analyses
July 2009
E-59
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ruckerl et al. (2006, 088754) Outcome (ICD9 and ICD10): C-reactive
Period of Study: Oct 2000—Apr 2001 protein (CRP)
Location: Erfurt, Germany	serum amyloid A (SAA)
E-se lectin
von Willebrand Factor (vWF)
intercellular adhesion molecule-1 (ICAM-1)
fibrinogen
Factor VII
prothrombin fragment 1 +2
Ddimer
Age Groups: 50+ yrs
Study Design: Panel (12 repeated
measures at 2-wk intervals)
N: 57 male subjects with coronary disease
Statistical Analyses: Fixed effects linear
and logistic regression models
Covariates: Models adjusted for different
factors based on health endpoint
CRP: RH, temperature, trend, ID
ICAM-1: temperature, trend, ID
vWF: air pressure, RH, temperature, trend,
ID
FVII: air pressure, RH, temperature, trend,
ID,weekday
Season: Time trend as covariate
Dose-response Investigated? Sensitivity
analyses examined nonlinear exposure-
response functions
Statistical Package: SAS v8.2 and S-
Plus v6.0
Pollutant: OC
Averaging Time: 24 h
Mean (SD): 1.5 (0.6)
Percentiles: 25th: 1.1
50th: 1.4
75th: 1.8
Range (Min, Max): 0.3, 3.4
Monitoring Stations: 1 site
Copollutant: UFPs (ultrafine particles)
AP (accumulation mode particles)
PM2.6
PM10
0C (organic carbon)
EC (elemental carbon)
NO?
CO
PM Increment: IQR (0.7
5-d avg: 0.5)
Effect Estimate [Lower CI, Upper CI]:
Effects of air pollution on blood markers
presented as OR (95%CI) for an increase
in the blood marker above the 90th
percentile per increase in IQR air pollutant.
CRP
Time before draw: 0 to 23 h: 1.2 (0.7,
1.9)
24 to 47 h: 1.3(0.8, 2.1)
48 to 71 h: 1.4(0.8,2.4)
5-d mean: 1.2 (0.7,1.8)
ICAM-1
Time before draw: 0 to 23 h: 0.9 (0.6,
1.3)
24 to 47 h: 2.0(1.3,3.2)
48 to 71 h: 3.0(1.8,4.8)
5-d mean: 1.3 (0.8, 2.0)
Effects of air pollution on blood markers
presented as % change from the meanfGM
in the blood marker per increase in IQR air
pollutant.
vWF
Time before draw: 0 to 23 h: 5.5 (0.2,
10.8)
24 to 47 h: 8.0 (2.1,13.9)
48 to 71 h: 3.5 (-2.6, 9.6)
5-d mean: 7.4(2.0,12.8)
FVII
Time before draw: 0 to 23 h: -6.1 (-10.6
to -1.4)
24 to 47 h: -7.2 (-11.4 to -2.8)
48 to 71 h:-3.8 (-8.2, 0.9)
5-d mean: -5.6 (-9.8 to-1.1)
Note: Summary of results presented in
figures. SAA results indicate increase in
association with PM (not as strong and
consistent as with CRP)
no association observed between E-
selectin and PM
an increase in prothrombin fragment 1 +2
was consistently observed, particularly
with lag 4
fibrinogen results revealed few significant
associations, potentially due to chance
D dimer results revealed null associations
in linear and logistic analyses
July 2009
E-60
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ruckerl et al. (2007, 091379) Outcome: Soluble CD40 ligand (sCD40L),
platelets, leukocytes, erythrocytes,
Period of Study: Oct 2000—Apr 2001
Location: Erfurt, Germany
hemoglobin
Age Groups: 50+ yrs
Study Design: Panel (12
measures at 2-wk intervals)
N: 57 male subjects with coronary disease
Statistical Analyses: Fixed effects linear
regression models
Covariates: Long-term time trend,
weekday of the visit, temperature, RH,
barometric pressure
Season: Time trend as covariate
Dose-response Investigated? No
Statistical Package: SAS v8.2 and S-
Plus v6.0
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD): 20.0(15.0)
Percentiles: 25th: 9.7
50th: 14.9
75th: 26.1
Range (Min, Max): 2.6, 83.7
Monitoring Stations: 1 site
Copollutants: UFPs (ultrafine particles)
AP (accumulation mode
particles)
PM2.6
PM10
NO
PM Increment: IQR (16.4
5-davg: 12.2)
Effect Estimate [Lower CI, Upper CI]:
Effects of air pollution on blood markers
presented as % change from the meanfGM
in the blood marker per increase in IQR air
pollutant.
sCD40L, % change GM (pg/mL)
lagO: 1.5 (-4.0, 7.3)
Lag1: 0.2 (-5.4, 6.2)
Lag2: -2.6 (-8.0, 3.1)
Lag3: 0.5 (-3.9, 5.0)
5-d mean: 0.2 (-5.4, 6.2)
Platelets, % change mean (103/|Jl)
LagO:-0.6 (-1.9, 0.7)
Lag1:0.1 (-1.3,1.5)
Lag2: 0.5 (-0.9,1.9)
Lag3: 0.2 (-1.1,1.5)
5-d mean: -0.4 (-1.9,1.2)
Leukocytes, % change in mean (103/|Jl)
LagO:-1.6 (-3.2, 0.0)
Lag1:-0.4 (-2.2,1.4)
Lag2:-0.2 (-2.1,1.7)
Lag3: -0.8 (-2.4, 0.7)
5-d mean: -1.6 (-3.5, 0.3)
Erythrocytes, % change mean (10E/|Jl)
LagO:-0.1 (-0.5, 0.3)
Lag1:-0.3 (-0.7, 0.2)
Lag2: -0.4 (-0.8, 0.0)
Lag3:-0.2 (-0.5, 0.1)
5-d mean: -0.4 (-0.8, 0.0)
Hemoglobin, % change mean (g/dl)
LagO: 0.0 (-0.6, 0.5)
Lag1:-0.2 (-0.8, 0.3)
Lag2:-0.5 (-1.1, 0.0)
Lag3:-0.2 (-0.7, 0.2)
5-d mean: -0.5 (-1.0, 0.1)
Reference: Sarnat et al. (2006, 090489) Outcome: Supraventricular ectopy (SVE)
or ventricular ectopy (VE)
Period of Study: summer and Autumn
2000
Location: Steubenville, OH
N: 32 nonsmoking older adults
Statistical Analysis: Logistic mixed
effects regression
Season: Summer, Autumn
Dose-response Investigated? No
Pollutant: PM2.6
Averaging Time: 5 days
Median (IQR): PM2.6: Median: 19.0//gfm3
IQr - 10.0
Sulfate: Median: 6.1. IQR: 4.2
EC: Median: 0.9. IQR: 0.5
Copollutants: O3, NO2, SO2
PM Increment: IQR
Effect Estimate: PM21,: SVE: OR - 1.42
(95% CI: 0.99, 2.04)
VE: OR - 1.02 (95% CI: 0.63-1.65)
Sulfate: SVE: OR - 1.70(95% CI: 1.12,
2.57)
VE: OR - 1.08(95% CI: 0.65,1.80)
EC: SVE: OR - 1.15 (95% CI: 0.73,1.81)
VE: OR - 1.00(95% CI: 0.57,1.75)
Notes: Longitudinal study of 32
nonsmoking older adults who had ECG
measurements made every week for 24
weeks. PM measured within 1 mile of
subjects' residences, and central site
pollutant measurements were also made.
July 2009
E-61
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Schneider et al. (2008,
191985)
Outcome: Endothelial Function
Parameters
Period of Study: Nov 2004 - Dec 2005 Age Groups: 48-80 yrs
Location: Chapel Hill, NC	Study Design: panel
N: 22 diabetics
Statistical Analyses: Mixed Models
Pollutant: PM2.6
Averaging Time: daily
Mean (SD): 13.6 (7.0)
Min: 2.0
Max: 38.9
Monitoring Stations: 2
Covariates: season, day of the week,
temperature, relative humidity, barometric Copollutant: NR
pressure
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-4d & 5d ma
PM Increment: 10/yg/m3
Percent Change: (Lower CI, Upper CI),
lag:
FMD: *\
¦17.3 (-34.6, 0.0), lagO
¦4.4 (-24.6,15.8), lag 1
¦18.6(44.8,7.6), lag 2
1.6 (-23.6, 26.9), lag 3
18.4(-3.5, 40.3), lag 4
-19.4 (-62.6, 23.8), 5d ma
NTGMD:
2.5(-9.0,13.9), lagO
¦13.6 (-24.5,-2.6), lag 1*
-10.2 (-23.5, 3.0), lag 2
¦8.0 (-22.4, 6.4), lag 3
3.6 (-7.9,15.0), lag 4
-19.4 (-44.3, 5.5), 5d ma
LAEI:
0.4 (-4.2, 5.0), lag 0
¦0.3 (-6.0, 5.4), lag 1
2.5 (-4.3, 9.4), lag 2
¦7.3 (-13.5,-1.1), lag 3*
¦2.3 (-8.0, 3.3), lag 4
-4.6 (-15.3, 6.1), 5d ma
SAEI:
¦3.0 (-13.0, 7.0), lag 0
¦17.0 (-27.5,-6.4), lag 1**
-9.7 (-23.5,4.2), lag 2
¦15.1 (-29.3,-0.9)*, lag 3
¦2.1 (-14.0, 9.7), lag 4
¦25.4 (-45.4,-5.3), 5d ma*
SVR:
¦1.6 (-3.7, 0.4), lagO
1.6(-0.9, 4.1), lag 1
3.5 (0.5, 6.5), lag 2
2.4	(-0.5, 5.3), lag 3
3.2 (0.7, 5.6), lag 4*
4.5	(-0.3, 9.2), 5d ma
*p< 0.05, " p < 0.01
Notes: Percent change (95%CI) per
10/yg/m3 PM2.6by GSTM1 genotype
(figure 3)
July 2009
E-62
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Schwartz et al. (2005,
074317)
Period of Study: 12 weeks during the
summer of 1999
Location: Boston, MA
Outcome: Heart rate variability (HRV),
((SDNN,
r-MSSD, PNN50, LFHFR)
Age Groups: 61-89 years
Study Design: Panel study
N: 28 elderly subjects
Statistical Analysis: Mixed models. To
examine heterogeneity of effects,
hierarchical modeling was used.
Season: Summer
Dose-response Investigated? No
Pollutant: PM2.B
Averaging Time: 1 h
24 h
Median: 24-hs: 10 |Jgfm3
Monitoring Stations: 1
Copollutant: BC, 0:i, CO, SO2, NO2
PM Increment: IQR (not given)
Effect Estimate: 24 h: 2.6 ms decrease
in SDNN (95% CI: 0.8 to-6.0)
10.1 ms decrease in r-MSSD (95% CI: ¦
2.8 to-16.9).
1 h: 3.4 ms decrease in SDNN (95% CI:
0.6 to-7.3)
7.4 ms decrease in r-MSSD (95% CI: 1.6
to-15.5).
Notes: Various log-transformed HRV
parameters were measured for 30 minutes
once a week. The random effects model
indicated that the negative effect of BC
on HRV was not restricted to a few
subjects.
Same study population as Gold et al.
(2005). Boston Elders Study
For each pollutant/averaging time,
similarly sized changes were observed for
PNN50 (%) and LFHFR.
Reference: Schwartz et al. (2005,
074317)
Period of Study: 12 weeks during the
summer of 1999
Location: Boston, MA
Outcome: Heart rate variability (HRV),
((SDNN, r-MSSD, PNN50, LFHFR)
Age Groups: 61-89 years
Study Design: Panel study
N: 28 elderly subjects
Statistical Analysis: Mixed models. To
examine heterogeneity of effects,
hierarchical modeling was used.
Season: Summer
Dose-response Investigated? No
Pollutant: BC
Averaging Time: 24 h
Median: 1.0 |jg/m3
Monitoring Stations: 1
Copollutant: PM2.6, O3, CO, SO2, NO2
PM Increment: IQR
Effect Estimate: 5.1 ms decrease in
SDNN (-1.5 to-8.6)
10.1 ms decrease in r-MSSD (-2.4 to ¦
17.2).
Notes: Various log-transformed HRV
parameters were measured for 30 minutes
once a week. The random effects model
indicated that the negative effect of BC
on HRV was not restricted to a few
subjects. Same study population as Gold
et al. (2005). Boston Elders Study.
Subjects with a prior Ml experienced
greater declines in BC associated HRV. For
each pollutant/averaging time, similarly
sized changes were observed for PNN50
(%) and LFHFR.
Reference: Schwartz et al. (2005,
074317)
Period of Study: 2000
Location: Boston, Massachusetts
Outcome: HF (high frequency component Pollutant: PM2.6
of heart rate variability)
Study Design: Cross-sectional
N: 497 subjects
Statistical Analysis: Linear regression,
controlling for covariates
Averaging Time: 48 h
Mean (SD): 11.4/yg/m3 (8.0)
Copollutant: None
PM Increment: 10/yg/m3
Effect Estimate: 34% decrease in HF
(95% CI: -9% to -52%) in subjects without
the GSTM1 allele. In subjects with the
allele, no effect was noted. Similar
findings for obese subjects and those with
high neutrophil counts.
Notes: Study population: Normative Aging
Study.
Effects of PM2.E appear to be mediated by
R0S.
July 2009
E-63
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Sorensen et al. (2005,
089428)
Period of Study: Nov 1999-Aug 2000
Location: Copenhagen, Denmark
Outcome: 7-Hydro-8-Oxo-2'-
Deoxyguanosine (8-oxodG) (measured in
lymphocytes and urine)
Age Groups: 20-33 yrs
Study Design: Panel (repeated measures)
N: 49 students living and studying in
central Copenhagen
50 students examined each season (66
subjects total
32 participated in each season
total of 98 measurements)
Statistical Analyses: Mixed models
repeated measures
Covariates: PM2.6, season, subject
(random factor)
Dose-response Investigated? No
Statistical Package: SAS v8e
Pollutant: PM2.6
Averaging Time: 48 h
Mean (SD): Autumn: 20.7
Summer: 12.6
Percentiles: IQR Autumn: 13.1-27.7
IQR summer: 9.4-24.3
Range (Min, Max): NR
Monitoring Stations: NA (personal
assessment)
Copollutant (correlation): Spearman
correlations with PM2.6 mass: chromium
(r - 0.22)
copper (r - 0.33)
iron (r - 0.29)
vanadium (p > 0.5)
nickel (p > 0.5)
platinum (p > 0.5)
PM Increment: see below
Effect Estimate [Lower CI, Upper CI]:
Association between 8-oxodG in
lymphocytes and personal exposure to
transition metals in PM2.6.
% increase in 8-oxodG per increase in
metal concentration indicated
Vanadium: 1.9% per 1 |jg/L (0.6, 3.3)
Chromium: 2.2% per 1 |jg/L (0.8, 3.5)
Platinum: 6.1 % per 1 ng/L (-0.6,13.2)
Nickel: 0.8% per 10 (Jg/L (-2.1, 3.7)
Copper: -0.8% per 10 |jg/L (-2.7,1.0)
Iron: 0.6% per 10 |jg/L (-1.4, 2.6)
Note: PM2.B mass was independently
associated with 8-oxodG in 5 of 6
transition metal models (p < 0.02 in
models with vanadium, chromium, nickel,
copper, and iron
p - 0.07 in platinum model). No transition
metals were associated with 8-oxodG
measured in urine
Reference: Sorensen et al. (2003,
042700)
Period of Study: Nov 1999-Aug 2000
Location: Copenhagen, Denmark
Outcome: RBC count, hemoglobin,
platelet count, fibrinogen, PLAAS (2-
aminoadipic semialdehyde in plasma
proteins), HBGGS (~-glutamyl
semialdehyde in hemoglobin), HBAAS (2-
aminoadipic semialdehyde in hemoglobin),
MDA (malondialdehyde)
Age Groups: 20-33 yrs
Study Design: Panel (repeated measures)
N: 50 students living and studying in
central Copenhagen
50 students examined each season (68
subjects total
31 participated in each season
total of 195 measurements)
Statistical Analyses: Mixed model
repeated-measures analysis
Covariates: Season, avg outdoor
temperature, and sex
Season: Repeated measures 4 times
(once per season)
Dose-response Investigated? No
Statistical Package: SAS v8e
Pollutant: PM2.6 (personal)
Averaging Time: 48 h
Median: 16.1 |Jgfm3
Percentiles: Q25-Q75: 10.0-24.5
Copollutant: Urban background PM2.6
Personal PM2.6
PM Increment: 1 /yglm3
Effect Estimate [Lower CI, Upper CI]:
Relationship between exposure and
biomarkers
Estimate (p-value): Platelet count (x 10e/g
protein): 0.0008 (0.37)
Fibrinogen (nmol/g protein): 0.0006 (0.69)
PLAAS (pmol/mg protein): 0.0016 (0.061)
HBGGS (pmol/mg protein): 0.0001 (0.94)
HBAAS (pmol/mg protein): 0.0006 (0.64)
Increase (95%CI) in biomarkers per 10
|jg/m3 increase in PM2.5
RBC
Men: 0% (-1.6,1.6)
Women: 2.3% (0.5, 4.1)
Hemoglobin
Men: 0.0% (-1.7,1.5)
Women: 2.6% (0.8, 4.5)
July 2009
E-64
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Sorensen et al. (2003,
042700)
Period of Study: Nov 1999-Aug 2000
Location: Copenhagen, Denmark
Outcome: RBC count, hemoglobin,
platelet count, fibrinogen, PLAAS (2-
aminoadipic semialdehyde in plasma pro-
teins), HBGGS (~ glutamyl semialdehyde in
hemoglobin), HBAAS (2-aminoadipic semi-
aldehyde in hemoglobin), MDA
(malondialdehyde)
Age Groups: 20-33 yrs
Study Design: Panel (repeated measures)
N: 50 students living and studying in
central Copenhagen
50 students examined each season (68
subjects total
31 participated in each season
total of 195 measurements)
Statistical Analyses: Mixed model
repeated-measures analysis
Covariates: Season, avg outdoor
temperature, and sex
Season: Repeated measures 4 times
(once per season)
Dose-response Investigated? No
Statistical Package: SAS v8e
Pollutant: Personal exposure to black
carbon (10 6/m)
Averaging Time: 48 h
Median: 8.1
Percentiles: Q25-Q75: 5.0-13.2
Copollutant:
Urban background PM2.6
Personal PM2.6
PM Increment: 10 6/m
Effect Estimate [Lower CI, Upper CI]:
Relationship between exposure and
biomarkers
Estimate (p-value): RBC count (x 10B/g
protein): 0.0003 (0.75)
Hemoglobin (|Jmolfg protein): 0.0004
(0.65)
Platelet count (x 106/g protein): 0.0009
(0.51)
Fibrinogen (nmol/g protein): -0.0027 (0.29)
PLAAS (pmol/mg protein): 0.0041
(0.0009)
HBGGS (pmol/mg protein): 0.0024 (0.25)
HBAAS (pmol/mg protein): 0.0022 (0.20)
MDA (pmol/mg protein): 0.0018 (0.30)
Reference: Sorensen et al. (2003,
042700)
Period of Study: Nov 1999-Aug 2000
Location: Copenhagen, Denmark
Outcome: RBC count, hemoglobin,
platelet count, fibrinogen, PLAAS (2-
aminoadipic semialdehyde in plasma pro-
teins), HBGGS (~-glutamyl semialdehyde in
hemoglobin), HBAAS (2-aminoadipic semi-
aldehyde in hemoglobin), MDA
(malondialdehyde)
Age Groups: 20-33 yrs
Study Design: Panel (repeated measures)
N: 50 students living and studying in
central Copenhagen
50 students examined each season (68
subjects total
31 participated in each season
total of 195 measurements)
Statistical Analyses: Mixed model
repeated-measures analysis
Covariates: Season, avg outdoor
temperature, and sex
Season: Repeated measures 4 times
(once per season)
Dose-response Investigated? No
Statistical Package: SAS v8e
Pollutant: PM2.6 (urban background
concentration)
Averaging Time: 48 h
Median: 9.2 |Jg/m3
Percentiles: Q25-Q75: 5.3-14.8
Copollutant: Urban background PM2.6
Personal carbon black
PM Increment: 1 /yg/m3
Effect Estimate [Lower CI, Upper CI]:
Relationship between exposure and
biomarkers
Estimate (p-value): RBC count (x 10B/g
protein): 0.0008 (0.36)
Hemoglobin (|Jmol/g protein): 0.0005
(0.53)
Platelet count (x 10e/g protein): -0.0008
(0.49)
Fibrinogen (nmol/g protein): 0.0004 (0.84)
PLAAS (pmol/mg protein): 0.0004 (0.76)
HBGGS (pmol/mg protein): -0.0020 (0.39)
HBAAS (pmol/mg protein): -0.0021 (0.29)
MDA (pmol/mg protein): 0.0012 (0.52)
July 2009
E-65
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Sullivan et al. (2007,100083) Outcome: Blood CRP, fibrinogen, Ddimer Pollutant: PM2.6
Averaging Time: 24 h
Study Design: Panel study	Median (IQR): 7.7//gfm3 (6.4)
N: 47 elderly subjects	Monitoring Stations: 1
Copollutant: Indoor PM2.6
Period of Study: February 2000-March Age Groups: > 55 years of
2002
Location: Seattle, Washington, USA
PM Increment: 10/yg/m3
Effect Estimate: Among those with CVD,
PM2.61 day earlier: CRP: 1.25 (95% CI:
0.97,1.58)
Fibrinogen: 1.01 (95% CI: 0.97,1.05)
Ddimer: 1.04 (95% CI: 0.93,1.15)
With C0PD: CRP: 0.69 (95% CI: 0.34,
1.42)
Fibrinogen: 1.05 (95% CI: 0.97,1.13)
D-dimer: 1.10 (95% CI: 0.95,1.28)
Healthy: CRP: 1.01 (95% CI: 0.85,1.19)
Fibrinogen: 0.88 (95% CI: 0.81, 0.95)
D-dimer: 1.10 (95% CI: 0.75,1.58)
Notes: Out of 47 subjects, n - 23 with
CVD and n - 24 (n - 16 C0PD and 8
healthy) without CVD. Blood markers were
measured on 2-3 morning over a 5-10 day
period, and outdoor PM2.6 was measured
at a central monitoring site.
These findings are not consistent with and
effect of fine PM on markers of
inflammation and thrombosis in the
elderly.
Reference: Sullivan et al. (2005,109418)
Outcome: Heart rate variability (H, LF,
Pollutant: PM2.6
PM Increment: 10/yg/m3

HF, r-MSSD, SDNN)

Period of Study: February 2000-March
Averaging Time: 1 h
Effect Estimate: 1 h:
2002
Study Design: Panel study

With CVD: HF: (3% increase, 95% CI: -19,

Median (IQR): 10.7 (7.6)
Location: Seattle, Washington, USA
N: 34 elderly subjects with (n - 21) and
Copollutant: CO, NO2
32)
without (n - 13) CVD.



Without CVD: HF(5% decrease, 95% CI: ¦

Statistical Analysis: Linear mixed

34, 36)

effects regression




Similarly, no association was found for 4-



h or 24-h mean PM2.6 concentrations.



Notes: 285 daily 20 minute HRV



measures were made in the homes of



study subjects over a 10-day period.
Reference: Sullivan et al. (2005,109418) Outcome (ICD9 and ICD10): High-
sensitivity C-reactive protein (hs-CRP)
Period of Study: February 2000-March
2002
Location: Seattle area, WA
fibrinogen
Ddimer
endothelin-1 (ET-1)
interleukin-6 (IL-6
interleukin-6 receptor (IL-6r)
tumor necrosis factor-n (TNF-8- ~)
Pollutant: PM2.6
Averaging Time: 24 h
(0-day and 1-day lags)
Mean (SD): NR
Percentiles: For all subject-days: 25th:
5.2
50th: 7.7
75th: 11.5
tumor necrosis factor-receptors (p55, p75) 90th: 19.9
Range (Min, Max): 1.3, 33.9
monocyte chemoattractant protein-1
(MCP-1)
Age Groups: a 55 yrs
Study Design: Panel
Monitoring Stations: NA, measured at
participant's residence
measures) Copollutant: None
N: 47 participants with (23) and without
(10C0PD and 8 healthy) CVD
Statistical Analyses: Mixed models
Covariates: Age, gender, medication use,
meteorological variables (temperature and
RH)
Dose-response Investigated? No
PM Increment: 10/yg/m3
Effect Estimate [Lower CI, Upper CI]:
Multiplicative change in mean outcome
associated with 10 //g|m3 increase in PM
Among those with different disease
status.
CRP Fold-rise |95%CI)
CM
0-d	lag: 1.21 (0.86,1.70)
CM
1-d	lag: 1.25 (0.97,1.58);
C0PD
0-d	lag: 0.93 (0.48,1.80)
C0PD
1-d	lag: 0.69 (0.33,1.46)
Healthy
0-d	lag: 0.98 (0.88,1.08)
Healthy
1-d	lag: 1.01 (0.841.21)
July 2009
E-66
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Statistical Package: SAS v8.02
Fibrinogen Fold-rise (95%CI)
CM
0-d	lag: 1.02 (0.98,1.06)
CM
1-d	lag: 1.0 (0.97,1.03);
C0PD
0-d	lag: 1.0 (0.91,1.09)
C0PD
1-d	lag: 1.08 (0.99,1.17)
Healthy
0-d	lag: 0.94 (0.87,1.01)
Healthy
1-d	lag: 0.99 (0.88,1.17)
D-dimer Fold-rise (95%CI)
CM
0-d	lag: 1.02 (0.88,1.17)
CM
1-d	lag: 1.03 (0.93,1.15):
C0PD
0-d	lag: 1.04 (0.93,1.16)
COPD
1-d	lag: 1.09 (0.94,1.27)
Healthy
0-d	lag: 0.95 (0.79,1.14)
Healthy
1-d	lag: 0.97 (0.71,1.31)
Among those with cardiovascular
disease
MCP-1 Fold-rise |95%CI)
0-d	lag: 1.3 (1.1, 1.7)
1-d	lag: 1.0 (0.9, 1.3)
ET-1 Fold-rise |95%CI)
0-d	lag: 1.1 (0.8, 1.2)
1-d	lag: 1.1 (0.9, 1.2)
Note: TNF-n and IL-6 measures were
below the limit of detection of assays
Reference: Timonen et al. (2006,
088747)
Period of Study: 1998-1999
Location: Amsterdam, Netherlands
Erfurt, Germany
Helsinki, Finland
Outcome: Heart variability (HRV)
measurements: [LF, HF, LFHFR, NN
interval, SDNN, r-MSSD]
Study Design: Panel study
N: 131 elderly subjects with stable
coronary heart disease
Statistical Analysis: Linear mixed
models
Pollutant: PM2.6
Means: Amsterdam: 20.0
Erfurt: 23.3
Helsinki: 12.7
Copollutant: NO2, CO
PM Increment: 10/yg/m3
Effect Estimate: SDNN
¦0.33ms (95% CI:-1.05, 0.38)
HF: -0.3% (95% CI:-10.6, 5.4)
LFHFR: -1.4 (95% CI: -5.9, 8.7)
Notes: Followed for 6 months with
biweekly clinic visits
2-day lag. ULTRA Study
July 2009
E-67
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Vallejo et al. (2006,157081)
Period of Study: April-August 2002
Location: Mexico City metropolitan area
Outcome: Heart rate variability measures
(SDNN, pNN50)
Age Groups: Mean age 27 yrs
Study Design: Panel study
N: 40 young healthy participants (non-
smokers, no meds or history of CVD,
respiratory, neurological, or endocrine
Pollutant: PM2.6
(pDR nephelometric method-DataRAM)
Copollutant: None
PM Increment: 30 /yglm3
Effect Estimate: pNN50: 0 h lag: -0.01 %
(95% CI:-0.03, 0.01)
Statistical Analysis: Linear mixed
effects models
1	h:-0.01% (95% CI
2	h:-0.05% (95% CI
3	h:-0.07% (95% CI
4	h:-0.08% (95% CI
5	h:-0.06% (95% CI
6	h:-0.05% (95% CI
¦0.04, 0.02)
¦0.09,0.00)
¦0.13 to-0.02)
-0.14 to -0.01)
¦0.13,0.02)
¦0.13,0.04)
SDNN: Oh: 0.00% (95% CI: 0.00, 0.01)
1	h: 0.00% (95% CI:-0.01, 0.01)
2	h: 0.00% (95% CI:-0.02, 0.01)
3	h:-0.01% (95% CI:-0.02, 0.00)
4	h:-0.01% (95% CI:-0.02, 0.01)
5	h:-0.01% (95% CI:-0.02, 0.01)
6	h: 0.00% (95% CI: -0.02, 0.02)
Notes: Subjects underwent 13 h of ECG
monitoring and personal PM2.6
measurement. HRV measures were
regressed against different lags of PM2.6
concentration.
Reference: Van Hee et al. (2009,
192110)
Period of Study: Jul 2000-Aug 2002
Location: Baltimore, Maryland
Chicago, Illinois
Winston-Salem, North Carolina
St. Paul
Minnesota
New York, New York
Los Angeles, California
Outcome: Left Ventricular Mass Index and Pollutant: PM2
Ejection Fraction
Age Groups: 45-84 yrs
Study Design: cross-sectional
N: 3,827 participants
Statistical Analyses: Linear
Models
Covariates: age, race, income, sex,
education, medication use, LDL, HDL,
physical activity, alcohol consumption,
smoking, diabetes, systolic BP, diastolic
BP
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: NR
Averaging Time: NR
Mean (SD): figure only
Monitoring Stations: n/a
Interpolation used
Copollutant: NR
Co-pollutant Correlation
n/a
PM Increment: 10/yg/m3
Difference (Lower CI, Upper CI), p-
value:
Left Ventricular Mass Index
Unadjusted: -6.0 (-7.8, -4.2), <0.0001
All covariates except center, BP: -6.1 (¦
7.8,-4.4), <0.0001
All covariates except BP: 3.7 (-6.0,13.4),
0.46
Full model: 4.6(4.7,13.9), 0.33
Full model plus center/race interaction: 3.8
(-6.1,13.7),0.45
Left Ventricular Ejection Fraction
Unadjusted: 3.0 (2.2, 3.8), < 0.0001
All covariates except center, BP: 1.4 (0.5,
2.2),0.001
All covariates except BP: -1.1 (-5.8, 3.7),
0.66
Full model:-1.3 (-6.0, 3.5), 0.60
Full model plus center/race interaction: -
3.0 (-8.0, 2.0), 0.24
July 2009
E-68
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Wellenius et al. (2007,
092830)
Period of Study: February 2002-March
2003
Outcome: Circulating levels of B-type
natriuretic peptide (BNP
measured in whole blood at 0, 6,12
weeks)
Pollutant: PM2.6
Copollutant: NO2, SO2, O3, CO, BC
Location: Boston, Massachusetts, USA Study Design: Panel study
N: 28 subjects (each with chronic stable
HF and impaired systolic function)
Statistical Analysis: Linear mixed
effects models
PM Increment: 10/yg/m3
Effect Estimate: Same day PM2.6: 0.8%
increase in BNP (95% CI:-16.4, 21.5)
Notes: The study found no association
between any pollutant and measures of
BNP at any lag. Further, the within
subject coefficient of variation was large
suggesting the magnitude of effected air
pollutant health effects are small in
relation to within subject variability in
BNP.
Reference: Wellenius et al. (2007,
092830)
Period of Study: February 2002-March
2003
Location: Boston, Massachusetts
Outcome (ICD9 and ICD10): B-type
natriuretic peptide (BNP) (natural-log
transformed)
Age Groups: 33-88 yrs
Study Design: Panel (blood collected at
0, 6, and 12 weeks)
N: 28 patients with chronic stable heart
failure and impaired systolic function
Statistical Analyses: Linear mixed-
effects models
Covariates: Temperature, dew point,
mean dew point over the past 3 days,
calendar month of blood draw, mea-
surement occasion, treatment assignment,
measurement occasion by treatment
assignment interaction
Season: Adjusted for calendar month
Dose-response Investigated? No
Statistical Package: SAS v9.1
Pollutant: PM2.6
Averaging Time: Daily (assessed lags of
0-3 days)
Mean (SD): 10.9 (8.4)
Percentiles: 50th: 8.0/yg/m3
Range (Min, Max): 0.7-50.9//g|m3
Monitoring Stations: 1 monitor
Copollutant (correlation): CO (r - 0.35)
NO2 (r - 0.31)
SO2 (r - 0.18)
O3 (r - 0.35)
BC(r - 0.68)
PM Increment: IQr - 8.1 //g|m3
Effect Estimate [Lower CI, Upper CI]: %
change in BNP per IQR increase in PM2.6
LagO: 1.5 (-18.7,19.2)
Lag1: 2.1 (-20.0, 30.3)
Lag2:1.3 (12.3,17.1)
Lag3: 5.6 (-16.8, 34.0)
Note: No significant associations
observed between any pollutant and BNP
levels at any lags (presented in Fig 2)
Reference: Wheeler et al. (2006,
088453)
Period of Study: Fall 1999 and spring
2000
Location: Atlanta, GA
Outcome: Heart rate variability
Age Groups: 49-76 years
N: 18 subjects with C0PD and 12
subjects with a recent Ml
Statistical Analysis: Linear-mixed effect
model
Season: Fall and spring
Pollutant: PM2.6
Averaging Time:
1 h
4 h
24 h
Mean: 24-hs: 17.8/yglm3
Copollutant: O3, CO, SO2, NO2
PM Increment: 11.65/yglm3 (IQR) in 4 h
PM2.6
Effect Estimate: Among C0PD patients:
8.3% increase in SDNN (95% CI: 1.7,
15.3)
Among Ml patients: 2.9% decrease in
SDNN (95% CI: -7.8, 2.3)
Results for 1 h and 24 h averaging times
were similar.
Notes: Data was collected on 7 days in
the Fall of 1999 or spring of 2000.
Effects were modified by medication use,
baseline pulmonary function, and health
status.
July 2009
E-69
DRAFT-DO NOT CITE OR QUOTE

-------
Reference	Design & Methods	Concentrations'	Effect Estimates (95% CTT
Reference: Yeatts et al. (2007, 091266)
Outcome: Heart Rate Variability
Pollutant: PM2.6
PM Increment: 1 /yglm3
Period of Study: 12 wk period bit Sept
Age Groups: 21-50 yrs
Averaging Time: 24h
Beta, SE (Lower CI, Upper CI), p-value:
2003 -July 2004



Study Design: panel
Mean (SD): 12.5 (6.0)
HRV
Location: Chapel Hill, NC



N: 12 asthmatics
Min: 0.6
Max Heart Rate: 0.40, 0.43 (-0.45,1.24),



0.36

Statistical Analyses: Linear Mixed
Max: 37.1


Model

ASDNN5: -0.07, 0.15 (-0.37, 0.22), 0.63


Monitoring Stations: 1

Covariates: temperature, humidity,
Copollutant: PMwj,, PMio
SDANN5: 1.66,0.65 (0.39, 2.93), 0.02

pressure


Co-pollutant Correlation
SDNN24HR(mesc): 1.16,0.58(0.02,

Dose-response Investigated? No
2.29), 0.06

Statistical Package: SAS
PM 10-2.B - 0.46*
rMSSD: 0.53, 0.20 (0.14,0.91), 0.01

Lags Considered: 1 day
PMio - NR
pNN50 24hour:-0.06, 0.11 (-0.27,0.15),


*p<0.01
0.58



pNN50 7min: 0.47, 0.42 (-0.35,1.29),



0.27



Low-frequency power: -0.23, 0.14 (-0.51,



0.05), 0.11



Percent low frequency: -0.78, 0.41 (-1.59,



0.03), 0.07



High-frequency power: 0.14, 0.07 (-0.01,



0.28), 0.07



Percent high frequency: 0.64, 0.36 (-0.07,



1.34), 0.09



Blood Lipids



Triglycerides: -0.63, 0.84 (-2.29,1.02),



0.46



VLDL:-0.17, 0.22 (-0.61, 0.26), 0.44



Total cholesterol: -0.06, 0.22 (-0.49,



0.36), 0.77



Hematologic Factor



Circulating eosinophils: -0.02, 0.00 (-0.02,



-0.02), 0.27



Platelets: -0.01, 0.45 (-0.88, 0.86), 0.98



Circulating Proteins



Plasminogen: 0.00, 0.00 (-0.01, 0.00),



0.82



Fibrenogen: 0.00, 0.01 (-0.01, 0.02), 0.59



Von Willibrand factor: -0.31, 0.29 (-0.87,



0.25), 0.28



Factor VII: -0.65, 0.33 (-1.29, -0.01), 0.05
July 2009
E-70
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Yue et al. (2007, 097968)
Period of Study: October 2000-April
2001
Location: Erfurt, Germany
Outcome: QT interval and T-wave
amplitude for ECG recordings, and vWF,
CRP from blood samples
Study Design: Panel study
N: 56 patients (male CAD patients with
12 clinical visits)
Statistical Analysis: Linear and logistic
regression models
Dose-response Investigated? No
Pollutant: PM2.6, Particle Number
Concentration (PNC) (n/cm3)
Averaging Time: Mean: Mass
concentrations of PNC (0.1-2.84 n/cm3)
Monitoring Stations: 1
Copollutant: None
PM Increment:. IQR
Effect Estimate: Each IQR increase in 0-
23 h mean traffic particle concentration
was associated with: QT interval: 0.6%
(95% CI:-0.3,1.4)
T wave amplitude: -1.6% (95% CI: -3.3,
0.1)
vWF: 3.2% (95% CI:-0.5, 7.0)
CRP: (OR - 1.5
95% CI 1.0-2.3)
Each IQR increase in 0-23 h mean
combustion-generated particle
concentration was associated with: QT
interval: 0.1 %(-0.3, 0.6)
T wave amplitude: -0.2% (-1.2, 0.7)
vWF: 2.8% (0.8, 4.8)
CRP (OR - 1.0
0.8, 1.2)
Notes: Five sources of particles were
identified (airborne soil, local traffic-
related ultrafine particles, combustion-
generated aerosols, diesel traffic-related
particles, and secondary aerosols).
July 2009
E-71
DRAFT-DO NOT CITE OR QUOTE

-------
	Reference	
Reference: Yue et al. (2007, 097968)
Period of Study: Oct 12, 2000-Apr 27,
2001
Location: Erfurt, Germany
	Design & Methods	
Outcome: QT interval, T wave amplitude,
von Willebrand factor (vWF), C-reactive
protein (CRP
above 90th percentile compared to below)
Age Groups: >50 yrs
Study Design: Panel (12 visits
625 observations for repolarization
parameters and 578 observations for
inflammatory markers)
N: 57 male coronary artery disease
patients
Statistical Analyses: Linear and logistic
fixed-effects regression models
(generalized additive models)
Covariates: Trend, weekday, and
meteorological variables (temperature,
relative humidity, barometric pressure)
Dose-response Investigated? No
Statistical Package: SAS v9.1 and S-
Plus v6.0
Concentrations'
Pollutant: Five particle source factors
(airborne soil, local traffic-related ultrafine
particles, combustion-generated aerosols,
diesel traffic-related particles, and
secondary aerosols)
see below for size fractions (factor scores)
Averaging Time: Used daily factor scores
in analyses
Mean (SD): Factor 1: particles from
airborne soil (1.0-2.8//m): 2390 (1696)
Factor 2: ultrafine particles from local
traffic (0.01-0.1 //m): 9931 (5858)
Factor 3: secondary aerosols from local
fuel combustion (0.1 -0.5 //m): 3770
(6129)
Factor 4: particles from traffic (0.01-
0.5 //m): 6865 (5689)
Factor 5: secondary aerosols from multiple
sources (0.2-1.0/ym): 4732 (3890)
Median: Factor 1: 2053
Factor 2: 8531
Factor 3: 1348
Factor 4: 5045
Factor 5: 3752
IQR (5-day avg): Factor 1: 1110
Factor 2: 5749
Factor 3: 4124
Factor 4: 5000
Factor 5: 3393
Range (Min, Max): Factor 1: 284,12960
Factor 2: 866, 26632
Factor 3: 139, 39097
Factor 4: 283, 27605
Factor 5: 67,20129
Monitoring Stations: 1 monitor
Copollutant: NA
Effect Estimates (95% CI)
PM Increment: IQR
Effect Estimate [Lower CI, Upper CI]:
QT interval, % change (95%CI)
Factor 1: 0-5h: -0.1 (-0.6, 0.6)
6-11h:-0.5 (-1.1, 0.2)
12-17h: 0.1 (-0.4,0.4)
18-23h:-0.2 (-0.7, 0.2)
0-23h: -0.2 (-0.9, 0.4)
1 d: -0.1 (-0.7, 0.6)
2d: -0.3 (-0.9, 0.4)
3d:-0.7 (-1.4, 0.1)
4d: -0.2 (-0.9, 0.5)
0-4d avg: -0.7 (-1.8, 0.3)
Factor 2: 0-5h: 0.2 (-0.4, 0.8)
6-11 h: 0.8(-0.0, 1.7)
12-17h: 0.6 (-0.2,1.4)
18-23h: 0.5 (-0.4,1.4)
0-23h: 0.9 (-0.1, 2.0)
1 d: 1.5(0.3,2.7)
2d: -0.4 (-1.7, 1.0)
3d: 0.5(-0.9,1.9)
4d: 0.1 (-1.2,1.4)
0-4d avg: 1.6 (-0.1, 3.3)
Factor 3: 0-5h: 0.1 (-0.3, 0.5)
6-11 h: 0.2(-0.3, 0.6)
12-17h: 0.2 (-0.3, 0.6)
18-23h: 0.1 (-0.3, 0.4)
0-23h: 0.1 (-0.3, 0.6)
1d: 0.1 (-0.3, 0.4)
2d:-0.1 (-0.4,0.3)
3d: -0.2 (-0.5, 0.2)
4d: -0.1 (-0.5, 0.2)
0-4d avg:-0.1 (-0.7, 0.6)
Factor 4: 0-5h: 0.2 (-0.4, 0.8)
6-11 h: 0.8(0.0, 1.6)
12-17h: 0.5 (-0.2,1.3)
18-23h: 0.5 (-0.2,1.2)
0-23h: 0.6(-0.3, 1.4)
1d: -0.4 (-1.5, 0.7)
2d: -0.9 (-2.0, 0.1)
3d:-0.5 (-1.4, 0.5)
4d: -0.5 (-1.3, 0.2)
0-4d avg:-0.3 (-1.7,1.1)
Factor 5: n0-5h: 1.0( 0.1,2.1)
6-11h: 0.9 (-0.2, 2.0)
12-17h: 0.3 (-0.7,1.4)
18-23h: -0.1 (-1.2, 1.0)
0-23h: 0.7(-0.6,1.9)
PMTl-T P,1 ^)0T CITE 0R QU0TE
2d: -0.2 (-1.5, 1.1)
3d: -0.6 (-1.9, 0.8)
July 2009
E-72

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Zanobetti et al. (2004,
087489)
Period of Study: 1999 to 2001
Location: Boston, Massachusetts, USA
Outcome: Blood pressure (systolic blood
pressure, diastolic blood pressure, mean
arterial blood pressure)
Age Groups: Elderly
Study Design: Panel study
N: 62 elderly subjects with n - 631
repeated visits for cardiac rehabilitation
Statistical Analysis: Linear mixed
effects models
Pollutant: PM2.6
Averaging Time: 24 h
Median (10th—90th percentile)
Median: 8.8/yg/m3
10th-90th: 13.4
Monitoring Stations: 1
Copollutant: SO2, O3, CO, NO2, BC
120-h avg
Median: 0.651
10th-90th: 0.376
PM Increment: .10.4/yg/m3 for 5 day
mean, 13.9 //g|m3 for 2-day mean
Effect Estimate: Each 10.4/yg/m3
increase in 5 day mean PM2.6
concentration was associated with:
Systolic BP: 2.8mmHg (95% CI: 0.1, 5.5)
Diastolic BP: 2.7mmHg (95% CI: 1.2, 4.3)
Mean arterial BP: 2.7mmHg (95% CI: 1.0,
4.5)
Each 13.9/yglm3 increase in 2-day mean
PM2.5, during exercise in person with
H.70bpm
Diastolic: 7.0mmHg (95% CI: 2.3,12.1)
Mean arterial BP: 4.7mmHg (95% CI: 0.5,
9.1)
July 2009
E-73
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Zeka et al. (2006,157177)
Period of Study: Nov 2000-Dec 2004
Location: Greater Boston area
(Massachusetts)
Outcome: White blood cells (WBC), C-
reactive protein (CRP), sediment rate,
fibrinogen
Age Groups: Mean age (SD) - 73.0 (6.7)
Study Design: Cross-sectional
N: 710 subjects
Statistical Analyses: Linear regression
Covariates: Age, BMI, season (also
assessed potential for confounding by
temperature, RH, barometric pressure,
hypertensive or cardiac medications,
hypertension, smoking, alcohol, and
fasting glucose levels)
Dose-response Investigated? No
Pollutant: BC
Averaging Time: Hourly (PN, BC, PM2 e)
and 24-h (S042-) measurements used to
create 48-h, 1-wk, and 4-wk moving
averages
Mean (SD): 0.77 (0.63)
Percentiles: 50th: 0.61
75th: 1.00
90th: 1.51
Monitoring Stations: 2 sites
Units: ngfm3
Copollutant (correlation): PM2.B
(r - 0.52)
BC
PN (r - 0.30)
S042- (r - 0.30)
PM Increment: 1 SD increase
Effect Estimate [Lower CI, Upper CI]: %
increase (95% CI) in biomarker per 1 SD
increase in pollutant.
Fibrinogen
48 h: 0.84 (-0.63, 2.31)
1 wk: 0.60 (-0.95, 2.15)
4wk: 1.78(0.19,3.36)
CRP
48 h: 4.51 (-2.03,11.06)
1wk: 1.07 (-5.55,7.68)
4wk: 5.41 (-1.00,11.81)
Sediment rate
48 h: -4.56 (-25.55,16.43)
1 wk: 1.98(-18.15, 22.11)
4wk: 21.65 (1.48, 41.82)
WBC count
48 h: -0.63 (-2.45,1.19)
1 wk: -0.13 (-1.87,1.60)
4 wk:-0.55 (-2.36,1.26)
Note: No statistically significant
difference was reported for any category
of effect modifiers (age, obesity,
medications, homozygous for the deletion
of GSTM1-null, hypertension)
however, results suggested almost all the
effect of BC on sediment rate was among
the younger group (<78 yrs)
there was a 4-fold difference for the
association between BC and CRP in the
presence of obesity
also evidence for effect modification by
obesity of the association between BC
and sediment rate
there was a suggestive greater effect of
BC on CRP among GSTMI-null subjects
(9.73% [1.48,17.98]) vs. GSTM1-present
subjects (-2.97% [-14.05, 8.10] for
concentrations 4-wk prior)
a stronger effect of BC on sediment rate
was seen among non-users of statins
(36.01% [13.88, 58.13]) vs users
(-12.29% [39.13,14.55])
July 2009
E-74
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Zeka et al. (2006,157177)
Period of Study: Nov 2000-Dec 2004
Location: Greater Boston area
(Massachusetts)
Outcome: White blood cells (WBC), C-
reactive protein (CRP), sediment rate,
fibrinogen
Age Groups: Mean age (SD) - 73.0 (6.7)
Study Design: Cross-sectional
N: 710 subjects
Statistical Analyses: Linear regression
Covariates: Age, BMI, season (also
assessed potential for confounding by
temperature, RH, barometric pressure,
hypertensive or cardiac medications,
hypertension, smoking, alcohol, and
fasting glucose levels)
Dose-response Investigated? No
Pollutant: S042-
Averaging Time: Hourly (PN, BC, PM2 e)
and 24-h (S042-) measurements used to
create 48-h, 1-wk, and 4-wk moving
averages
Mean (SD): 2.29(1.62)
Percentiles: 50th: 1.84
75th: 2.81
90th: 4.10
Monitoring Stations: 2 sites
Copollutant (correlation): PM2.B
(r - 0.50)
BC (r - 0.30)
PN (r - -0.15)
S042-
PM Increment: 1 SD increase
Effect Estimate [Lower CI, Upper CI]: %
increase (95% CI) in biomarker per 1 SD
increase in pollutant.
Fibrinogen: 48 h: 0.60 (-1.23, 2.42)
1 wk: 0.03(-1.93,1.99)
4wk: 1.12(-0.52, 2.77)
CRP: 48 h: 1.57 (-7.13,10.27)
1 wk: 0.21 (-8.27, 8.69)
4wk: 5.29 (-1.91,12.49)
Sediment rate: 48 h: 4.05 (-23.26,
31.36)
1 wk: -5.87 (-32.39, 20.64)
4 wk: -1.60 (-25.24, 22.04)
WBC count: 48 h: -0.12 (-2.35, 2.11)
1 wk: -0.48 (-2.87,1.90)
4wk: 0.75(-1.30, 2.80)
Note: No statistically significant
difference was reported for any category
of effect modifiers (age, obesity,
medications, homozygous for the deletion
of GSTM1-null, hypertension)
Reference: Zeka et al. (2006,157177)
Period of Study: Nov 2000-Dec 2004
Location: Greater Boston area
(Massachusetts)
Outcome (ICD9 and ICD10): White blood Pollutant: PM2 e
cells (WBC), C-reactive protein (CRP),
sediment rate, fibrinogen
Age Groups: Mean age (SD) - 73.0 (6.7)
Study Design: Cross-sectional
N: 710 subjects
Statistical Analyses: Linear regression
Covariates: Age, BMI, season (also
assessed potential for confounding by
temperature, RH, barometric pressure,
hypertensive or cardiac medications,
hypertension, smoking, alcohol, and
fasting glucose levels)
Dose-response Investigated? No
Statistical Package: NR
Averaging Time: Hourly (PN, BC, PM2 e)
and 24-h (S042-) measurements used to
create 48-h, 1-wk, and 4-wk moving
averages
Mean (SD): 11.16 (7.95)
Percentiles: 50th: 9.39
75th: 14.57
90th: 21.48
Monitoring Stations: 2 sites
Copollutant (correlation): PM2 e
BC (r - 0.52)
PN (r - -0.02)
S042- (r - 0.50)
PM Increment: 1 SD increase
Effect Estimate [Lower CI, Upper CI]: %
increase (95% CI) in biomarker per 1 SD
increase in pollutant.
Fibrinogen: 48 h:-0.18 (-1.93,1.57)
1 wk: -1.39 (-3.46, 0.67)
4wk: 1.14 (-0.60, 2.88)
CRP: 48 h:-4.88 (-13.29, 3.53)
1 wk: -1.37 (-10.44,7.71)
4 wk: 4.36 (-3.25,11.96)
Sediment rate: 48 h: -16.91 (-43.66,
9.84)
1 wk:-18.89 (-47.48, 9.70)
4wk: 24.93 (0.68, 49.18)
WBC count: 48 h: -3.18 (-5.39 to -0.97)
1 wk: -0.51 (-3.02, 2.00)
4wk: -0.03 (-2.17, 2.10)
Note: No statistically significant
difference was reported for any category
of effect modifiers (age, obesity,
medications, homozygous for the deletion
of GSTM1-null, hypertension)
Reference: Zhang et al. (2009,191970) Outcome: Myocardial Ischemia
Period of Study: 1999-2003
Location: US
Age Groups: 52-90
Study Design: panel
N: 55,529
Statistical Analyses: Logistic 81 Linear
ession
Covariates: age, race/ethnicity,
education, exam site, BMI, current
Pollutant: PM2.B
Averaging Time: NR
Mean (SD):
Lag 0: 14.1 (8)
Lag 1:13.8 (8)
Lag 2: 13.8(8)
Lag 3: 13.8(8)
PM Increment: 10/yg/m3
Odds Ratio (Lower CI, Upper CI), lag:
Minnesota Codes*
MC4: 1.04 (0.97,1.10), lag 0-2
MC4: 1.04 (0.98,1.11), lag 3-5
MC5: 1.05 (1.00,1.09), lag 0-2
MC5: 1.04 (1.00,1.08), lag 3-5
July 2009
E-75
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
smoking status, history of CHD, diabetes,
hypertension, SBP, chronic lung disease,
or hypercholesterolemia, day of week,
time of day, temperature, dew point,
pressure, season
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-5d
Lag 4: 13.9(8)
Lag 5: 14.1 (8)
Lag 0-2: 13.9(7)
Monitoring Stations: NR*
Co-pollutant: NR
Monitors used in model for spatial
interpolation of daily PM values.
MC 4 or 5:1.04 (1.00,1.09), lag 0-2
MC 4 or 5:1.03 (0.99,1.07), lag 3-5
Change (Lower CI, Upper CI), lag:
ST-segment amplitude
Lead I: -0.07 (-0.36, 0.21), lag 0-2
Lead I: 0.18 (-0.10, 0.46), lag 3-5
Lead II: -0.12 (-0.47, 0.23), lag 0-2
Lead II: 0.16 (-0.18, 0.50), lag 3-5
Lead aVL:-0.01 (-0.25, 0.23), lag 0-2
LeadaVL: 0.11 (-0.12, 0.34), lag 3-5
Lead V1: -0.02 (-0.39, 0.35), lag 0-2
Lead V1: -0.22 (-0.58, 0.14), lag 3-5
Lead V2: 0.07 (-0.57, 0.70), lag 0-2
Lead V2: -0.01 (-0.61, 0.62), lag 3-5
Lead V3: -0.11 (-0.68, 0.47), lag 0-2
Lead V3: -0.02 (-0.58, 0.54), lag 3-5
Lead V4: -0.0.3 (-0.51, 0.45), lag 0-2
Lead V4: 0.24( 0.23, 0.71), lag 3-5
Lead V5: -0.01 (-0.41, 0.39), lag 0-2
Lead V5: 0.35 (-0.04, 0.74), lag 3-5
Lead V6: 0.02 (-0.30, 0.33), lag 0-2
Lead V6: 0.35 (0.04, 0.65), lag 3-5
T-wave amplitude
Lead I: -1.60 (-3.07, -0.13), lag 0-2
Lead I: -0.31 (-1.73,1.11), lag 3-5
Lead II: -0.54 (-1.99,0.92), lag 0-2
Lead II: 0.71 (-0.70,2.13), lag 3-5
Lead aVL:-1.21 (-2.50,0.10), lag 0-2
Lead aVL:-0.55 (-1.18, 0.71), lag 3-5
Lead V1: 1.45( 0.16, 3.06), lag 0-2
Lead V1: 0.03 (-1.53,1.59), lag 3-5
Lead V2: -0.18 (-2.96, 2.60), lag 0-2
Lead V2: 0.57 (-2.12, 3.27), lag 3-5
Lead V3: -2.33 (-5.15, 0.49), lag 0-2
Lead V3: -0.13 (-2.87, 2.60), lag 3-5
Lead V4: -2.03 (-4.69, 0.63), lag 0-2
Lead V4: 0.64 (-1.94, 3.22), lag 3-5
Lead V5: -1.92 (-4.22, 0.38), lag 0-2
Lead V5: 0.55 (-1.69, 2.78), lag 3-5
Lead V6: -0.63 (-2.36,1.10), lag 0-2
Lead V6: 0.82 (-0.86, 2.49), lag 3-5
QRSi'T angles and heart rate (change)
QRS/T angle - spatial (°): 0.19 (-0.21,
0.59), lag 0-2
QRS/T angle - spatial (°): -0.20 (-0.59,
0.19), lag 3-5
July 2009
E-76
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
QRS/T angle - frontal (°): 0.13 (-0.24,
0.50), lag 0-2
QRS/T angle - frontal (°): 0.35 (-0.01,
0.71), lag 3-5
Heart Rate (beats/min): 0.16 (0.02, 0.30),
lag 0-2
Heart Rate (beats/min): 0.04 (-0.10, 0.18),
lag 3-5
"Any ST abnormality (MC 4.1-4.4)
Any T abnormality (MC 5.1-5.4)
'All units expressed in /_yg/m' unless otherwise specified.
Table E-4. Short-term exposure - cardiovascular morbidity studies - other size fractions.
Reference
Design & Methods
Concentrations
Effect Estimates (95% CI)
Reference: Adar et al. (2007, 001458)
Period of Study: March-June 2002
Location: St. Louis, Missouri
Outcome: Heart rate variability: heart
rate, standard deviation of all normal-to-
normal intervals (SDNN), square root of
the mean squared difference between
adjacent normal-to-normal intervals
(rMSSD), percentage of adjacent normal-
to-normal intervals that differed by more
than 50 ms (pNN50), high frequency
power (HF
in the range of 0.15-0.4Hz), low
frequency power (LF, in the range of
0.04-0.15Hz), and the ratio of LF/HF
Age Groups: a 60 yrs
Study Design: Panel (4 planned repeated
measures with a total of 158 person-trips
35 participating in all 4 trips)
N: 44 participants
Statistical Analyses: Generalized
additive models
Covariates: Subject, weekday, time,
apparent temperature, trip type, activity,
medications, and autoregressive terms
Season: Limited data collection period
Dose-response Investigated? No
Statistical Package: SAS v8.02, R
V2.0.1
Pollutant: Particle count fine (PC fine)
(particles/cm3)
Averaging Time: Measurements
collected over 48 h period surrounding
the bus trip (during which health
endpoints were measured) used to
calculate 5-, 30-, 60-minute, 4-h, 24-h
moving averages
Median (IQR): All: 42 (57)
Facility: 36 (45)
Bus: 105(96)
Activity: 50 (133)
Lunch: 69 (48)
Monitoring Stations: 2 portable carts
Copollutant: PM2.6
BC
Fine particle counts
Coarse particle counts
Correlation notes: 24-h mean PM2.6, BC,
and fine particle count concentrations
ranged from 0.80 to 0.98
r - 0.76 to 0.97 when limited to time
spent on the bus
r - 0.55 to 0.86 when comparing bus
concentrations to 24-h moving averages
r - -0.003 to 0.51 when comparing 5-
min averages and 24-h moving averages.
Poor correlations found between coarse
particle count concentrations and all fine
particulate measures during all times
periods
PM Increment: IQR
Effect Estimate [Lower CI, Upper CI]:
% change (95%CI) in HRV per IQR in the
24-h moving avg of the
microenvironmental pollutant (IQr - 39
pt/cm3)
Single-pollutant models
SDNN:-5.1 (-5.8 to-4.4)
rMSSD:-8.0 (-8.7 to-7.2)
pNN50 + 1:-10.2 (-11.3 to -9.0)
LF:-9.9 (-11.4 to-8.4)
HF:-13.7 (-15.1 to-12.2)
LF/HF: 4.3 (3.1, 5.5)
H: 0.9(0.8,1.1)
Note: Exposure to health associations by
all lag periods presented in Figure 2
(magnitude of associations increased
with averaging period, with the largest
associations consistently found for 24-h
moving averages)
Reference: Adar et al. (2007, 001458)
Period of Study: March-June 2002
Location: St. Louis, Missouri
Outcome: Heart rate variability: heart
rate, standard deviation of all normal-to-
normal intervals (SDNN), square root of
the mean squared difference between
adjacent normal-to-normal intervals
(rMSSD), percentage of adjacent normal-
to-normal intervals that differed by more
than 50 ms (pNN50), high frequency
power (HF
in the range of 0.15-0.4Hz), low
frequency power (LF, in the range of
Pollutant: Particle count coarse (PT
coarse) (pt/cm3)
Averaging Time: Measurements
collected over48-h period surrounding
the bus trip (during which health
endpoints were measured) used to
calculate 5-, 30-, 60-minute, 4-h, and 24-
h moving averages
Median (IQR): All: 0.02(0.11)
Facility: 0.01 (0.04)
PM Increment: IQR
Effect Estimate [Lower CI, Upper CI]:
% change (95%CI) in HRV per IQR in the
24-h moving avg of the
microenvironmental pollutant (IQr
- 0.066 pt/cm3)
Single-pollutant models
SDNN: 2.4(1.3, 3.6)
rMSSD: 3.9 (2.6, 5.1)
July 2009
E-77
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0.04-0.15Hz), and the ratio of LF/HF
Age Groups: a 60 yrs
Study Design: Panel (4 planned
measures with a total of 158 person-trips
35 participating in all 4 trips)
N: 44 participants
Statistical Analyses: Generalized
additive models
Covariates: Subject, weekday, time,
apparent temperature, trip type, activity,
medications, and autoregressive terms
Season: Limited data collection period
Dose-response Investigated? No
Statistical Package: SAS v8.02, R
V2.0.1
Bus: 0.16(0.13)
Activity: 0.29 (0.26)
Lunch: 0.16(0.36)
Monitoring Stations:
2 portable carts
Copollutant: PM2.6
BC
Fine particle counts
Coarse particle counts
Correlation notes: 24-h mean PM2.6, BC,
and fine particle count concentrations
ranged from 0.80 to 0.98
r - 0.76 to 0.97 when limited to time
spent on the bus
r - 0.55 to 0.86 when comparing bus
concentrations to 24-h moving averages
r - -0.003 to 0.51 when comparing 5-
min averages and 24-h moving averages.
Poor correlations found between coarse
particle count concentrations and all fine
particulate measures during all times
periods
pNN50 + 1:2.9(1.0, 4.9)
LF: 6.4 (3.7,9.1)
HF: 10.2 (7.4,13.1)
LF/HF: -3.3 (-5.0 to-1.6)
H:-1.1 (-1.3 to-0.8)
Two-pollutant models (with PM2.5):
SDNN: -0.7 (-1.9, 0.6)
rMSSD: -1.3 (-2.6 to -0.05)
pNN50 + 1:-4.3 (-6.3 to-2.4)
LF: 0.2 (-2.5, 3.0)
HF: 1.3 (-1.5,4.1)
LF/HF: -0.9 (-2.7,1.0)
H: -0.6 (-0.9 to-0.4)
Note: Exposure to health associations by
all lag periods presented in Figure 2
(magnitude of associations increased
with averaging period, with the largest
associations consistently found for 24-h
moving averages)
Reference: Delfino et al. (2008,156390) Outcome: C-reactive protein (CRP)
Period of Study: 2005-2006
Location: Los Angeles, California, air
basin
fibrinogen, tumor necrosis factor-n (TNF-
~) and its soluble receptor-ll (TNF-RII)
interleukin-6 (IL-6)
and its soluble receptor (IL-6sR)
fibrin D-dimer
soluble platelet selectin (sP-selectin)
soluble vascular cell adhesion molecule-1
(sVCAM-1)
intracellular adhesion molecule-1 (sICAM-
1)
and myeloperoxidase (MPO)
erythrocyte lysates for glutathione
peroxidase-1 (GPx-1)
copper-zinc superoxide dismutase (cu, Zn-
SOD)
Age Groups: a 65 yrs
Study Design: Panel (biomarkers
measured weekly 12 times)
N: 29 participants (nonsmoking with
history of coronary artery disease)
Statistical Analyses: Mixed models
Covariates: temperature (infectious
illnesses were excluded by excluding
weeks with such observations)
Season: Collected 6 weeks of data
during warm period and 6 weeks of data
during cool period
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM (multiple size fractions
and components)
Averaging Time: 24-h avg preceding the
blood draw (lag 0) and cumulative
averages up to 5 days preceding the draw
Outdoor hourly PM: EC: Mean (SD):
1.61 (0.62)
Median: 1.56
IQR: 0.92
Min, Max: 0.24, 3.94
0C: Mean (SD): 5.94 (2.11)
Median: 5.58
IQR: 2.79
Min-Max: 2.51,13.60
BC: Mean (SD): 2.00(0.77)
Median: 1.89
IQR: 0.96
Min-Max: 0.58, 5.11
OCpri: Mean (SD): 3.37 (1.21)
Median: 3.21
IQR: 1.63
Min-Max: 0.99,7.11
Secondary 0C: Mean (SD): 2.49 (1.50)
Median: 2.10
IQR: 1.86
Min-Max: 0, 8.10
PN (p/cm3): Mean (SD): 16,043 (5886)
Median: 13,968
PM Increment: IQR
Effect Estimate [Lower CI, Upper CI]:
Note: Nearly all results presented in
figures
Results: The authors found significant
positive associations for CRP, IL-6, sTNF-
Rll, and sP-selectin with outdoor and/or
indoor concentrations of quasi-ultrafine
PM ~ 0.25 |Jm in diameter, EC, OCpri, BC,
PN, CO, and nitrogen dioxide from the
current-day and multiday averages. There
were consistent positive but largely
nonsignificant coefficients for TNF-n,
sVCAM-1, and sICAM-1, but not
fibrinogen, IL-6sR, or D-dimer. The
authors found inverse associations for
erythrocyte Cu, Zn-S0D with these
pollutants and other PM size fractions
(0.25-2.5 and 2.5-10 |Jm). Inverse
associations of GPx-1 and MPO with
pollutants were largely nonsignificant.
Indoor associations were often stronger
for estimated indoor EC, OCpri, and PN of
outdoor origin than for uncharacterized
indoor measurements. There was no
evidence for positive associations with
S0A.
July 2009
E-78
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
IQR: 7,386
Min-Max: 6837,31263
Indoor hourly PM EC: Mean (SD): 1.31
(0.52)
Median: 1.30
IQR: 0.70
Min-Max: 0.19, 2.89
EC of outdoor origin: Mean (SD): 1.11
(0.39)
Median: 1.06
IQR: 0.51
Min-Max: 0.41,2.97
0C: Mean (SD): 5.69 (1.51)
Median: 5.60
IQR: 1.96
Min-Max: 2.34,10.79
OCpri of outdoor origin: Mean (SD): 2.18
(0.82)
Median: 2.15
IQR: 1.07
Min-Max: 0.32, 5.21
Secondary 0C of outdoor origin: Mean
(SD): 2.08(1.26)
Median: 1.75
IQR: 1.45
Min-Max: 0, 6.87
PN (particles/cm3): Mean (SD): 14,494
(6770)
Median: 12,341
IQR: 7,337
Min-Max: 1016, 43027
PN of outdoor origin (pfcm3): Mean (SD):
10,108(3108)
Median: 9,580
IQR: 3,684
Min-Max: 1016,17700
Outdoor PM mass PM0.2e: Mean (SD):
9.47 (2.97)
Median: 9.4
IQR: 4.2
Min-Max: 3.31,18.75
PM0.2b-2.b: Mean (SD): 13.53(10.67)
Median: 11.7
IQR: 11.5
Min-Max: 1.29, 66.77
PM2.B io: Mean (SD): 10.04 (4.07)
Median: 9.9
IQR: 5.9
July 2009
E-79
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Min-Max: 1.76, 22.38
Indoor PM mass PM0.2e: Mean (SD):
10.45 (6.77)
Median: 9.5
IQR: 4.5
Min-Max: 1.42, 69.86
PMo.26-2.6 (/yglm3): Mean (SD): 7.36 (4.57)
Median: 6.5
IQR: 5.7
Min-Max: 0.77, 30.86
PM2.B10: Mean (SD): 4.12(4.76)
Median: 2.8
IQR: 3.5
Min-Max: 0.12, 37.63
Copollutant: Outdoor hourly gases (NO2,
CO, O3) and indoor hourly gases (NO2, CO)
Reference: Pekkanen et al. (2002,
035050)
Period of Study: Winter 1998 to 1999
Location: Helsinki, Finland
Outcome: ST Segment Depression
(> 0.1 mV)
Study Design: Panel of ULTRA Study
participants
N: 45 Subjects, n - 342 biweekly
submaximal exercise tests, 72 exercise
induced ST Segment Depressions
Statistical Analysis: Logistic regression
I GAM
Pollutant: Ultrafine NC0.01-0.1 fjm
(n/cm3)
Averaging Time: 24 h
Median: 14,890
IQR: 9830
Monitoring Stations: 1
Copollutant: NO2, CO, PM2.6, PMio-2.5,
PM1.ACP
PM Increment: IQR
Effect Estimate(s): NC0.01 -0.1: OR
- 3.14(1.56,6.32), lag 2
Notes: The effect was strongest for ACP
and PM2.E, which in two pollutant models
appeared independent. Increases in NO2
and CO were also associated with
increased risk of ST segment depression,
but not with coarse particles.
Reference: Peters et al. (2005, 095747)
Also Peters et al, 2005 (2005,156859)
Period of Study: February 1999-July
2001
Location: Augsburg, Germany
Outcome: Myocardial infarction
Study Design: Case-crossover
N: 691 myocardial infarction patients
Statistical Analysis: Conditional logistic
regression
Dose-response investigated (yes/no)?
No
Pollutant: Ultrafine (TNC) (n/cm3)
Averaging Time: 1 h: Median - 10,001
IQR: 7919
24 h: Median - 10,934
IQR: 6276
Copollutant: NO2, SO2, CO
PM Increment: Effect Estimate: 2-h
lag: OR - 0.95
95% CI: 0.84,1.06
24-h mean, 2-day lag: OR - 1.04
95% CI: 0.90,1.20
Notes: Examined triggering for Ml at
various lags before Ml onset (up to 6 h
before Ml, up to 5 days before Ml). No
statistically significant increases in
lagged ultrafine particle concentration
were found.
Reference: Ruckerl et al. (2006,
088754)
Period of Study: Oct 2000—Apr 2001
Location: Erfurt, Germany
Outcome (ICD9 and ICD10): C-reactive
protein (CRP)
serum amyloid A (SAA)
E-selectin
von Willebrand Factor (vWF)
intercellular adhesion molecule-1 (ICAM-
1)
fibrinogen
Factor VII
prothrombin fragment 1 +2
D-dimer
Age Groups: 50+ yrs
Study Design: Panel (12 re
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Pollutant: AP (n/cm3)
Averaging Time: 24 h
Mean (SD): 1593 (1034)
Percentiles:
25: 821
50: 1238
75: 2120
Range (Min, Max): 328,4908
Unit (i.e./yg/m3): n/cm3
Monitoring Stations: 1 site
Copollutant: UFPs (ultrafine particles)
AP (accumulation mode particles)
PM2.6
PM10
PM Increment: IQR (1299
5-davg: 1127)
Effect Estimate [Lower CI, Upper CI]:
Effects of air pollution on blood markers
presented as OR (95%CI) for an increase
in the blood marker above the 90th
percentile per increase in IQR air
pollutant.
CRP Time before draw: 0 to 23 h: 0.7
(0.5, 1.2)
24 to 47 h: 1.5(0.9, 2.6)
48 to 71 h: 3.2(1.7,6.0)
5-d mean: 1.5 (0.8, 3.0)
ICAM-1 Time before draw: 0 to 23 h: 0.6
(0.4, 0.9)
24 to 47 h: 1.8(1.2, 2.8)
48 to 71 h: 1.6(1.0, 2.5)
July 2009
E-80
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Statistical Analyses: Fixed effects
linear and logistic regression models
Covariates: Models adjusted for
different factors based on health
endpoint
CRP: RH, temperature, trend, ID
ICAM-1: temperature, trend, ID
vWF: air pressure, RH, temperature,
trend, ID
FVII: air pressure, RH, temperature, trend,
ID, weekday
Season: Time trend as covariate
Dose-response Investigated?
Sensitivity analyses examined nonlinear
exposure-response functions
Statistical Package: SAS v8.2 and S-
Plus v6.0
OC (organic carbon)
EC (elemental carbon)
NO?
CO
5-d mean: 0.9 (0.6,1.5)
Effects of air pollution on blood markers
presented as % change from the
meanfGM in the blood marker per
increase in IQR air pollutant.
vWF Time before draw: 0 to 23 h: 4.8
(0.2, 9.3)
24 to 47 h: 5.9(0.4,11.5)
48 to 71 h: 7.0(0.7,13.4)
5-d mean: 13.5 (6.3, 20.6)
FVII Time before draw: 0 to 23 h: 0.0
(-2.9, 3.0)
24 to 47 h:-2.9 (-6.1, 0.4)
48 to 71 h: -3.6 (-6.8 to -0.3)
5-d mean: -4.1 (-7.9 to -0.3)
Note: summary of results presented in
figures.
SAA results indicate increase in
association with PM (not as strong and
consistent as with CRP)
no association observed between E-
selectin and PM
an increase in prothrombin fragment 1 +2
was consistently observed, particularly
with lag 4
fibrinogen results revealed few significant
associations, potentially due to chance
D-dimer results revealed null associations
in linear and logistic analyses
Reference: Ruckerl et al. (2006,
088754)
Period of Study: Oct 2000—Apr 2001
Location: Erfurt, Germany
Outcome: Soluble CD40 ligand (sCD40L),
platelets, leukocytes, erythrocytes,
hemoglobin
Age Groups: 50+ yrs
Study Design: Panel (12 re
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Statistical Analyses: Fixed effects
linear regression models
Covariates: Long-term time trend,
weekday of the visit, temperature, RH,
barometric pressure
Season: Time trend as covariate
Dose-response Investigated? No
Statistical Package: SAS v8.2 and S-
Plus v6.0
Pollutant: AP (n/cm3)
Averaging Time: 24 h
Mean (SD): 1593 (1034)
Percentiles: 25th: 821
50th: 1238
75th: 2120
Range (Min, Max): 328,4908
Monitoring Stations: 1 site
Copollutant: UFPs (ultrafine particles)
AP (accumulation mode particles)
PM2.6
PM10
NO
PM Increment: IQR (1299
5-davg: 1127)
Effect Estimate [Lower CI, Upper CI]:
Effects of air pollution on blood markers
presented as % change from the
meanfGM in the blood marker per
increase in IQR air pollutant.
sCD40L, % change GM (pg/mL)
lagO: 6.9(0.5,13.8)
Iag1: -1.1 (-8.0,6.4)
Iag2: -4.9 (-11.9, 2.7)
Iag3: -3.8 (-10.3, 3.2)
5-d mean: -1.3 (-9.9, 8.1)
Platelets, % change mean (103/|Jl)
lagO: -1.0 (-2.5, 0.5)
Iag1: -0.4 (-2.1,1.6)
Iag2: 0.8 (-1.0, 2.4)
Iag3: 0.0 (-1.8,1.7)
5-d mean: -0.9 (-3.0,1.3)
Leukocytes, % change in mean
(103/mD
lagO: -1.9 (-3.8 to -0.1)
Iag1: -0.6 (-2.9,1.6)
Iag2: -0.6 (-3.2, 2.0)
July 2009
E-81
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Iag3: -2.3 (-4.6, 0.1)
5-d mean: -2.7 (-5.5, 0.1)
Erythrocytes, % change mean (10E/|Jl)
lagO: -0.1 (-0.5,0.3)
Iag1: -0.4 (-0.9, 0.2)
Iag2: -0.4 (-0.9, 0.2)
Iag3: -0.4 (-0.6, 0.3)
5-d mean: -0.4 (-1.0, 0.2)
Hemoglobin, % change mean (g/dl)
lagO: -0.2 (-0.7, 0.4)
Iag1: -0.3 (-1.0, 0.4)
Iag2: -0.1 (-0.9,0.7)
Iag3: -0.1 (-0.8,0.6)
5-d mean: -0.2 (-1.1, 0.6)
Reference: Ruckerl et al. (2007,
156931)
Period of Study: May 2003—Jul 2004
Location: Athens, Augsburg, Barcelona,
Helsinki, Rome, and Stockholm
Outcome: lnterleukin-6
(IL-6), fibrinogen, C-reactive protein (CRP)
Age Groups: 35-80 yrs
Study Design: Repeated measures I
longitudinal
N: 1003 Ml survivors
Statistical Analyses: Mixed-effect
models
Covariates: City-specific confounders
(age, sex, BMI)
long-term time trend and apparent
temperature
RH, time of day, day of week included if
adjustment improved model fit
Season: Long-term time trend
Dose-response Investigated? Used p-
splines to allow for nonparametric
exposure-response functions
Statistical Package: SAS v9.1
Pollutant: UFP (n/cm3)
Averaging Time: Hourly and 24 h (lag 0-
4, mean of lags 0-4, mean of lags 0-1,
mean of lags2-3, means of lags 0-3)
Mean (SD): Presented by city only
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: Central monitoring
sites in each city
Copollutant: SO2
0s
NO
NO?
PM Increment: IQR
Effect Estimate [Lower CI, Upper CI]:
% change in mean blood markers per
increase in IQR of air pollutant.
IL-6
Lag (IQR): % change in GM (95%CI)
Lag 0(11852): 1.88 (-0.16, 3.97)
Lag 1 (11852):-0.67 (-2.56,1.25)
Lag 2 (11852): -2.12 (-4.03 to-0.17)
5-d avg (11003): -0.93 (-3.37,1.56)
Fibrinogen
Lag (IQR): % change in AM (95%CI)
Lag 0(11852): 0.40 (-0.40,1.19)
Lag 1 (11852): 0.11 (-0.69, 0.91)
Lag 2 (11852): 0.09 (-0.71, 0.90)
5-d avg (11003): 0.50 (-2.20, 3.20)
CRP
Lag (IQR): % change in GM (95%CI)
Lag 0(11852): 1.33 (-3.05, 5.90)
Lag 1 (11852):-1.52 (-4.39,1.45)
Lag 2 (11852):-1.63 (-6.70, 3.71)
5-d avg (11003):-0.08 (-3.78,3.75)
Reference: Pekkanen et al. (2002,
035050)
Period of Study: Winter 1998 to 1999
Location: Helsinki, Finland
Outcome: ST Segment Depression
(> 0.1 mV)
Age Groups: Study Design: Panel of
ULTRA Study participants
N: 45 Subjects, n - 342 biweekly
submaximal exercise tests, 72 exercise
induced ST Segment Depressions
Statistical Analysis: Logistic regression
I GAM
Pollutant: Ultrafine NC0.01-0.1 fjm
(n/cm3)
Averaging Time: 24 h
Median: 14,890
IQR: 9830
Monitoring Stations: 1
Copollutant: NO2, CO, PM2.6, PMio-2.5,
PM1.ACP
PM Increment: IQR
Effect Estimate(s): NC0.01 -0.1: OR
- 3.14(1.56,6.32), lag 2
Notes: The effect was strongest for ACP
and PM2.E, which in two pollutant models
appeared independent. Increases in NO2
and CO were also associated with
increased risk of ST segment depression,
but not with coarse particles.
July 2009
E-82
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Peters et al. (2005, 095747) Outcome: Myocardial infarction
Also Peters et al, 2005 (2005,156859)
Period of Study: February 1999-July
2001
Location: Augsburg, Germany
Study Design: Case-crossover
N: 691 myocardial infarction patients
Statistical Analysis: Conditional logistic 24-h: Median
regression
Dose-response Investigated? No
Pollutant: Ultrafine (TNC) (n/cm3)
Averaging Time: 1 h: Median - 10,001
IQR: 7919
10,934
IQR: 6276
Copollutant: NO2, SO2, CO
PM Increment: Effect Estimate:
2 h lag: OR - 0.95
95% CI: 0.84,1.06
24-h mean, 2-day lag: OR - 1.04
95% CI: 0.90,1.20
Notes: Examined triggering for Ml at
various lags before Ml onset (up to 6 h
before Ml, up to 5 days before Ml). No
statistically significant increases in
lagged ultrafine particle concentration
were found.
Reference: Ruckerl et al. (2007,
091379)
Period of Study: Oct 2000—Apr 2001
Location: Erfurt, Germany
Outcome (ICD9 and ICD10): Soluble
CD40 ligand (sCD40L), platelets,
leukocytes, erythrocytes, hemoglobin
Age Groups: 50+ yrs
Study Design: Panel (12 re
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Statistical Analyses: Fixed effects
linear regression models
Covariates: Long-term time trend,
weekday of the visit, temperature, RH,
barometric pressure
Season: Time trend as covariate
Dose-response Investigated? No
Statistical Package: SAS v8.2 and S-
Plus v6.0
Pollutant: UFP
Averaging Time: 24 h
Mean (SD): 12,602 (6455)
Percentiles: 25th: 7326
50th: 11,444
75th: 17,332
Range (Min, Max): 328,4908
Monitoring Stations: 1 site
Copollutant:
AP
PM2.6
PM10
NO
PM Increment: IQR (10,005
5-davg: 6,821)
Effect Estimate [Lower CI, Upper CI]:
sCD40L, % change GM (pg/mL)
lagO: 7.1 (0.1,14.5)
lag 1:0.3 (-6.6, 8.6)
lag 2: 0.6 (-5.9, 8.6)
lag 3: -8.5 (-15.8,-0.5)
5-d mean: -0.7 (-7.6, 6.8)
Platelets, % change mean (103/|Jl)
lag 0:-1.8 (-3.4,-0.2)
lag 1:-1.1 (-2.9, 0.6)
lag 2: 1.0 (-2.9, 0.8)
lag 3: -2.4(4.5, -0.3)
5-d mean: -2.2 (-4.0, -0.3)
Leukocytes, [103/|Jl]
lag 0:-2.4 (-4.5, -0.2)
lag 1: -2.1 (-4.4,0.2)
lag 2: -0.2 (-2.4, 2.8)
lag 3:-1.5 (-4.4,1.4)
5-d mean: -1.6 (-4.1, 0.8)
'All units expressed in //gfm3 unless otherwise specified.
July 2009
E-83
DRAFT-DO NOT CITE OR QUOTE

-------
E.I.2. Cardiovascular Emergency Department Visits and Hospital
Admissions
Table E-5. Short-term exposure-cardiovascular-ED/HA
PM
10
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Anderson et al. (2003,
054820)
Period of Study: 1992-1994
Location: London, United Kingdom
Outcome: Ischemic Heart Disease
Age Groups: 0-15,15-64, 65-74, 75-t
Study Design: Time series
N: NR
Statistical Analyses: NR
Covariates: NR
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max): NR
Monitoring Stations: NR
Copollutant: NR
PM Increment: 10th—90th percentile
% Change in Daily IHD Admissions by
Age [CI]: 0-15 yrs: NR
15-64 yrs: 2.6(0.3,5]
65-74 yrs: 2.5(0.1,4.9]
75+ yrs: 2.2(0.2,4.6]
Notes: RRs are presented in graph form
showing little change with increasing age
(PM increment of 10 /yglm3). This article
is primarily a systematic literature review
of other studies.
Reference: Andersen et al. (2008,
189651)
Period of Study: May 2001 - December
2004
Location: Los Angeles and San Dii
counties, California
Outcome (ICD-10): CVD, including angina
pectoris (I20), myocardial infarction (121-
22), other acute ischemic heart diseases
(124),	chronic ischemic heart disease
(125),	pulmonary embolism (I26), cardiac
arrest (I46), cardiac arrhythmias (I48-
48), and heart failure (I50).
Age Groups: > 65 yrs (CVD and RD), 5-
18 years (asthma)
Study Design: Time series
N: NR
Statistical Analyses: Poisson GAM
Covariates: Temperature, dew-point
temperature, long-term trend,
seasonality, influenza, day of the week,
public holidays.
Season: NR
Dose-response Investigated: No
Statistical Package: R (gam procedure,
mgcv package)
Lags Considered: Lag 0 -5 days, 4-day
pollutant avg (lag 0 -3) for CVD.
Pollutant: PMio
Averaging Time: 24 h
Mean (SD
median
IQR
99th percentile): 24 (14
21
16-29
72)
Monitoring Stations: 1
Copollutant (correlation): NCtot:
r - 0.39
NC100: r - 0.28
NCa12:r - 0.02
NCa23: r - -0.12
NCa57: r - 0.45
NCa2i2: r = 0.63
PM2.6: r - 0.80
CO: r - 0.37
NO?: r - 0.35
NOx: r - 0.32
NOx curbside: r - 0.18
0a: r - -0.21
Other variables: Temperature: r - 0.12
Relative humidity: r - 0.05
PM Increment: 13 |Jg|m3 (IQR)
Relative risk (RR) Estimate [CI]:
CVD hospital admissions
(4-day avg, lag 0 -3), age 65+:
One-pollutant model: 1.03 [1.01-1.05]
Adj for NCtot: 1.04(1.02-1.06]
Adj for NCa2i2:1.05 [1.01-1.09]
RD hospital admissions
(5 day avg, lag 0 -4), age 65+:
One-pollutant model: 1.06 [1.02-1.09]
Adj for NCtot: 1.05 [1.01-1.10]
Adj for NCa2i2:1.04 [0.98-1.11]
Asthma hospital admissions
(6-day avg lag 0-5), age 5-18:
One-pollutant model: 1.02 [0.93-1.12]
Adj for NCtot: 1.01 [0.91-1.12]
Adj for NCa2i2: 0.94 [0.81-1.09]
Estimates for individual day lags reported
only in figure form (see notes):
Notes: Figure 2: Relative risks and 95%
confidence intervals per IQR in single day
concentration (0- to 5-day lag).
Summary of Figure 2: CVD: Positive,
marginally or statistically significant
associations at Lag 0—Lag 2.
July 2009
E-84
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Anderson et al. (2007,
156214)
Period of Study: 1199-12 04
Location: Copenhagen, Denmark
Outcome (ICD10): Hospital Admission,
CVD, including angina pectoris (I20),
myocardial infarction (121 - 22), other
acute ischemic heart diseases (I24),
chronic ischaemic heart disease (I25),
pulmonary embolism (I26), cardiac arrest
(I46), cardiac arrhythmias (I48 - 48), and
heart failure (I50).
Age Groups Analyzed: Age > 65
Study Design: Time series
N: 2192 days, 9 Hospitals
Statistical Analyses: Principal
Component Analysis and Constrained
Physical Receptor Model (COPREM),
Poisson regression, GAM,
Covariates: Season, day of the wk,
public holidays, influenza epidemics and
meteorology
Season: All year
Dose-response Investigated? No
Statistical package: R, gamfmgcv
Lags Considered: 0-6 days
Pollutant: Source specific PMio
components
Averaging Time: 24-h
Mean (SD): Percentiles: 25th: 16
50th (Median): NR
75th: 30
Monitoring Stations: 1
Copollutant (correlation): PMio:
Biomass
r - 0.53
Secondary
r - 0.73
Oil
r - 0.57
Crustal
r - 0.37
Sea salt
r - 0.04
Vehicle
r - 0.02
Notes: Correlations between source
specific PMio components presented in
paper
PM Increment: IQR
RR Estimate
Respiratory disease (age > 65)
Single pollutant model:
PMio: 1.027 (1.013,1.042), IQR-14
PMio (other 5 sources): 1.045 (1.016,
1.074), IQR-13
Biomass: 1.040 (0.009,1.072), IQR-5.4
Secondary: 1.050 (1.021,1.081),
IQR-6.1
Oil: 1.035 (1.006,1.065), IQR-2.8
Crustal: 1.054 (1.028,1.081), IQR-1.8
Sea salt: 0.98 (0.947,1.017), IQR-2.2
Vehicle: 0.989 (0.949,1.032), IQR-0.6
Notes: 2 pollutant model results for PMio
with source specific components and
gases also presented in manuscript.
Reference: Baccarelli et al. (2007,
091310)
Period of Study: Jan 1995—Aug 2005
Location: Lombardia region, Italy
Outcome (ICD9 and ICD 10): Fasting and
postmethionine-load total homocysteine
(tHcy)
Age Groups: 11 -84 yrs
Study Design: Cross-sectional/Panel
N: 1,213 participants
Statistical Analyses: Generalized
additive models
Covariates: age, sex, BMI, smoking,
alcohol, hormone use, temperature, day
of the year, and long-term trends
Season: Adjusted for long-term trends to
account for season
Dose-response Investigated? No
Statistical Package: R software v2.2.1
Pollutant: PMio (some TSP measures
used to predict PMio)
Averaging Time: Hourly concentrations
used to calculate 24-h moving averages
and 7-day moving averages
Mean (SD): NR
Percentiles: 25th: 20.1
50th: 34.1
75th: 52.6
Range (Min, Max): Max: 390.0
Monitoring Stations: 53 sites
Copollutant: CO, NO2
SO?
Os
PM Increment: IQR
Effect Estimate [Lower CI, Upper CI]:
Estimates (%) per 32.5/yglm3 increase in
24-h moving avg of PMio
Homocysteine, fasting: 0.4 (-2.4, 3.3)
Homocysteine, postmethionine-load: (-1.5,
3.7)
Estimates (%) per 25.7m3 increase in 7-
day moving avg of PMio
Homocysteine, fasting: 1.0 (-1.9, 3.9)
Homocysteine, postmethionine-load: 2.0
(-0.6,4.7)
Estimates of effect (%) on fasting
homocysteine per IQR increase in 24-h
PMio levels
Among smokers: 6.2 (0.0,12.7)
Among non-smokers: -1.6 (-5.5, 2.5)
Estimates of effect (%) on
postmethionine-load homocysteine per
IQR increase in 24-h PMio levels
Among smokers: 6.0 (0.5,11.8)
Among non-smokers: -0.1 (-3.6, 3.5)
July 2009
E-85
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ballester et al. (2006,
088746)
Period of Study: 1995 ¦ 1999
Location: 5 Spanish cities: Granada,
Huelva, Madrid, Seville, Zaragoza
Outcome (ICD-9): All cardiovascular
disease (390-459), including all heart
diseases (410-414, 427, 428)
Age Groups: All ages
Study Design: Time series
N: NR
Statistical Analyses: Poisson GAMs
Covariates: daily temp, barometric
pressure relative humidity
daily influenza incidence, day of the
week, holidays, unusual events (ex.
medical strikes), seasonal variation, trend
Dose-response Investigated: No
Statistical Package: S-Plus GAM
function
Lags Considered: lag 0 -3 days, lag 0-1
avg
Pollutant: PMio
Averaging Time: 24 h
Mean (10-90th percentile): overall
mean NR.
City specific means
Granada: 43.2 (24.8, 62.6)
Huelva: 38.6(23.1,57.3)
Madrid: 35.7(21.4,54.4)
Seville: 41.9 (27.3, 57.6)
Zaragoza: 32.8 (17.3, 50.3)
Monitoring Stations: At least three
stations/city
(15+)
Copollutant (correlation): Summary of
the correlation coefficients between each
pair of pollutants within cities: BS:
r - 0.48
TSP: N/A
NO2: from r - 0.13 to r - 0.62 (median
r - 0.40)
SO2: from r - 0.20 to r - 0.51 (median
r - 0.46)
CO: from r - 0.34 to r - 0.45 (median
r - 0.37)
O3: from r - -0.07 to r - 0.16 (median
r - 0.11)
PM Increment: 10/yglm3
Relative risk [CI]: Relative risks are
expressed only in the form of figures (see
notes).
Percentage change in risk [CI]: All
cardiovascular diseases (avg of lags 0 -1):
0.91% [0.35,1.47]
Heart disease (avg of lags 0 -1)
1.56% [0.82, 2.31]
Notes: Relative risks for the single
pollutant models are expressed in
Figure 2.
Figure 2: Time sequence of the combined
association between PM10 and hospital
admissions for all CVD (A) and heart
disease (B).
Summary of results: Significant, positive
association of PM10 with both overall
CVD and heart disease hospitalizations at
Lag 0 and Lag 1.
Relative risks for two pollutant
models are expressed in Figure 3:
Figure 3: Combined estimates of the
association between hospital admissions
for heart diseases and air pollutants (avg
of lags 0-1
adjusted for CO, NO2, O3, or SO2)
Summary of results: Significant, positive
association remains after adjusting for
pollutants.
July 2009
E-86
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Bell et al. (2008, 091268)
Period of Study: 1995 ¦ 2002
Location: Taipei, Taiwan
Outcome (ICD-9): Hospital admissions
for ischemic heart disease (410 ,411,
414), cerebrovascular disease (430—
437).
Age Groups: All
Study Design: Time series
N: 6,909 hospital admissions for
ischaemic heart diseases, 11,466 for
cerebrovascular disease.
Statistical Analyses: Poisson regression
Covariates: Day of the week, time,
apparent temperature, long-term trends,
seasonality
Season: All
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: lags 0-3 days, avg of
lags 0-3
Pollutant: PMio
Averaging Time: 24 h
Mean (range
IQR): 49.1 (12.7-215.5
27.6)
Monitoring Stations: Taipei area: 13
monitors
Taipei City: 5 monitors
Monitors with correlations of 0.75 ¦
PMio: 12 monitors
Copollutant: NR
PM Increment: 28/yg/m3 (near IQR)
Percentage increase estimate [95%
CI]: Ischemic heart disease: Taipei area
(13 monitors): L0: 1.91 (-1.25,5.17)
L1: 0.39 (-2.73, 3.61)
L2: 1.80 (-1.33, 5.04)
L3: 2.01 (-1.14,5.26)
L03: 2.91 (-1.52,7.55)
Taipei City (5 monitors): L0: 2.08 (-1.04,
f°r 5.30)
L1: 0.43 (-2.64, 3.60)
L2: 2.17 (-0.92,5.36)
L3: 2.16 (-0.94, 5.36)
L03: 3.40(-1.19, 8.20)
Monitors with > - 0.75 between
monitor correlations (12 monitors): L0:
1.82 (-1.29, 5.03)
L1: 0.35 (-2.72, 3.52)
L2: 1.93(-1.15, 5.10)
L3: 1.93(-1.16, 5.12)
L03: 2.86 (-1.63, 7.54)
Cerebrovascular disease: Taipei area
(13 monitors): L0: -1.41 (-3.80,1.04)
L1:-1.95 (4.31, 0.48)
L2: 0.77 (-1.62,3.23)
L3: 2.64 (0.21,5.12)
L03: 0.01 (-3.33, 3.47)
Taipei City (5 monitors): L0: -1.27 (-3.64,
1.16)
L1:-2.13 (-4.47, 0.27)
L2: 0.85 (-1.52, 3.28)
L3: 2.52(0.13, 4.97)
L03:-0.07 (-3.53, 3.51)
Monitors with > - 0.75 between
monitor correlations (12 monitors): L0: -
1.34 (-3.70,1.07)
L1:-1.98 (-4.31, 0.40)
L2: 0.80 (-1.56, 3.22)
L3: 2.61 (0.22, 5.05)
L03: -0.02 (-3.40, 3.49)
July 2009
E-87
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Chan et al. (2007, 147787)
Period of Study: Apr 1997 - Dec 2002
Location: Boston, MA
Outcome: Cerebrovascular Emergency
Admissions
Age Groups: 50+ yrs
Study Design: time series
Statistical Analyses: GAM Poisosn
Regression
Covariates: year, month, day of week,
temperature, dew point
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-3d
Pollutant: PMio
Averaging Time: 24h
Mean (SD): 50.2(22.1)
Min: 16.0
Max: 325.4
IQR: 25.4
Monitoring Stations: 16
Copollutant: 0:i, CO, SO2, NO2, PM2.6
Co-pollutant Correlation
0s: 0.43
CO: 0.47
SO2: 0.59
NO?: 0.64
PM2.6: 0.61
PM Increment: Interquartile Range (25.4
j"g/m3)
Percent Change (Lower CI, Upper CI),
p-value:
Cerebrovascular Disease
Lag 0:1.001 (0.969,1.033)
Lag 1:0.999 (0.9787,1.020)
Lag 2:1.023 (0.989,1.057)
Lag 3:1.030 (1.011,1.049)
Lag 3 + 0s: 1.018(0.987,1.049)
Lag 3 + CO: 1.019(0.988,1.050)
Lag 3 + 0s + CO: 1.015(0.985,1.045)
Stroke
Lag 0:0.969 (0.897,1.041)
Lag 1:0.992 (0.918,1.066)
Lag 2:1.004 (0.993,1.015)
Lag 3:1.009 (0.988,1.030)
Ischaemic stroke
Lag 0:0.984 (0.932,1.036)
Lag 1:0.993 (0.939,1.047)
Lag 2: 0.989 (0.927,1.041)
Lag 3:1.042 (0.981,1.103)
Haemorrhagic stroke
Lag 0:0.966 (0.884,1.048)
Lag 1:0.990 (0.908,1.072)
Lag 2:1.002 (0.920,1.084)
Lag 3: 0.974 (0.902,1.046)
Reference: Chan et al. (2008, 093297)
Period of Study: 1995 ¦ 2002
Location: Taipei Metropolitan area,
Taiwan
Outcome (ICD-9): Emergency visits for
ischaemic heart diseases (410-411,
414), cerebrovascular diseases
(430-437), and C0PD (493, 496)
Age Groups: All
Study Design: Time series
N: NR
Pollutant: PM10
Averaging Time: 24 h
Mean (SD): High dust events: Pre-dust
periods: 45.5 (17.6)
Asian dust events: 122.7 (24.4)
Low dust events: Pre-dust periods: 59.4
(31.0)
PM Increment: 25.4/yg/m3 (IQR)
OR [95% CI]: In environmental conditions
without dust storms (results only shown
for best-fitting model)
Lag 3 days: 1.023(1.003,1.041)
Statistical Analyses: Poisson regression Asian dust events: 61.1 (17.8)
models
Covariates: Year, month, day of week,
temperature, dew point temperature,
PM2.6, NO2
Season: All
Dose-response Investigated: No
Statistical Package: SAS version 8.0
Lags Considered: 0- to 7-day lags
Monitoring Stations: 1
Copollutant: NR
July 2009
E-88
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Chanq et al. (2007,147621)
Outcome: CVD HA
Pollutant: PMio
PM Increment: Interquartile Range



(24.51 /yglm3)
Period of Study: 1997-2001
Age Groups: NR
Averaging Time: 24h



Odds Ratio (Lower CI, Upper CI):
Location: Taipei, Taiwan
Study Design: case-crossover
Mean: 48.32


>20°C

Statistical Analyses: Conditional
Min: 14.44


Logistic Regression
25": 32.65
PM10: 1.085 (1.061,1.110)

Covariates: temperature, humidity

PM10+ SO2: 1.131 (1.103,1.161)

50": 42.80


Dose-response Investigated? No

PM10+NO2: 10.977 (0.950,1.006)

75": 57.16


Statistical Package: SAS
Max: 234.91
PM10+CO: 1.025 (0.999,1.052)



Lags Considered: 0-2d

PM10+ O3: 1.064 (1.039,1.090)

Monitoring Stations: 6



<20°C


Copollutant: 0:i, CO, SO2, NO2




PM10: 1.142 (1.105,1.180)


Co-pollutant Correlation


NR
PM10+ SO2: 1.235 (1.184,1.288)



PM10+ NO2: 1.148 (1.103,1.194)



PM10+CO: 1.165 (1.121,1.212)



PMio+Os: 1.142 (1.105,1.180)
Reference: D'lppoliti et al. (2003,
Outcome: Myocardial Infarction HA
Pollutant: TSP
PM Increment: Quartiles
074311)



Age Groups: 18+ yrs
Averaging Time: 24h
Odds Ratio (Lower CI, Upper CI):
Period of Study: Jan 1995 - Jun 1997



Study Design: case-crossover
Mean (SD): 66.9(19.7)
Lag 0-2d avg
Location: Rome, Italy

25": 54.7

Statistical Analyses: Conditional
Ql: 1.0 (ref)

Logistic Regression
50": 66.4


011:1.048 (0.957,1.148)

Covariates: temperature, humidity

QUI: 1.105 (1.007,1.214)

75": 78.4

Dose-response Investigated? No
IQR: 23.7
QIV: 1.132(1.023,1.253)


Statistical Package: NR



Monitoring Stations: 3
Various Lags

Lags Considered: 0-4d



Copollutant: CO, SO2, NO2
Lag 0:1.023 (1.004,1.042)


Co-pollutant Correlation
Lag 1:1.015 (0.996,1.034)


CO: 0.35
Lag 2:1.017 (0.999,1.035)


SO2: 0.29
Lag 3: 0.989 (0.974,1.003)


NO2: 0.38
Lag 4:1.001 (0.987,1.016)
Reference: Funq et al., (2005, 093262)
Outcome (ICD-9): Cardiovascular
Pollutant: PM10
PM Increment: 26/yg/m3
diseases

Period of Study: Nov 1,1995-Dec 31,

Averaging Time: 24 h
% Change in Daily Admission [CI]: Age
2000
(410-414,427-428)

<65

Mean (min-max): 38.0 (5-248)

Location: London, Ontario
Age Groups: < 65 yrs, 65+ yrs
SD - 23.5
Current day mean: 2.6 [-2.3,7.7]


Study Design: Time series
Monitoring Stations: 4
2-day mean: -1.2 [-7.2,5.1]


N: 12,947 CVD admissions

3-day mean: -3 [-9.6,4]

Copollutant (correlation): NO2:

Statistical Analyses: GAM with locally
r - 0.30
Age 65 +

weighted regression smoothers (LOESS)
SO2: r - 0.24
Current day mean: 0.9 [-2.3,4.2]


Covariates: Maximum and minimum



temp, humidity, day of the week,
CO: r - 0.21
2-day mean: -0.9 [-4.8,3.2]

seasonal cycles, secular trends
O3: r - 0.53
3-day mean: -0.1 [-4.4,4.5]

Season: NR
C0H: r - 0.29


Dose-response Investigated? No



Statistical Package: S-Plus



Lags Considered: Current to 3-day mean


July 2009
E-89
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Hanigan et al (2008,
156518)
Period of Study: 1996-2005 (April-
November of each year)
Location: Darwin, Australia
Outcome: Daily emergency hospital
admissions for total cardiovascular (ICD-
9: 390-459
ICD-10:100-I99), ischemic heart disease
(ICD-9: 410-414
ICD-10:120-I25).
Age Groups: All
Study Design: Time series
N: 8,279 hospital admissions
Statistical Analyses: Poisson
generalized linear models
Pollutant: PMio
Averaging Time: 24 h
Mean (SD
range): 21.2 (8.2
55.2)
Monitoring Stations: NfA (see notes)
Copollutant: NR
Covariates: Indigenous status, time in
days, temperature, relative humidity, day
of the week, influenza epidemics, change
between ICD editions, holidays, yearly
population
Season: April-November (corresponding
to the dry season)
Dose-response Investigated? No
Statistical Package: R version 2.3.1
Lags Considered: 0-3
PM Increment: 10/yg/m3
Percent change [95% CI]: Overall CVD:
Lag 0 (indigenous): -3.78 [-13.4, 6.91]
Lag 0 (non-indigenous): -3.43 [-9.00,
2.49]
All unstratified associations either
negative or zero and not statistically
significant.
All other results of stratified analysis (by
indigenous status) reported in a figure
(see notes).
Notes: Figure 3: Associations between
hospitalizations for non-indigenous and
indigenous people with estimated ambient
PMio. Summary: Confidence intervals
were wide, but indigenous people
generally had stronger associations with
PMio than non-indigenous people. Daily
PMio exposure levels were estimated for
the population of the city from visibility
data using a previously validated models.
Reference: Hanigan et al (2008,
156518)
Period of Study: 1996-2005 (April-
November of each year)
Location: Darwin, Australia
Outcome: Cardiorespiratory Disease HA
(ICD 9: 390-519
ICD 10:100-99 & J00-99)
Age Groups: NR
Study Design: time series
N: 8279 events
Statistical Analyses: poisson regression
Covariates: indigenous status, time in
days, temperature, relative humidity, day
of the week, influenza epidemics, change
between ICD editions, holidays, yearly
population
Dose-response Investigated? No
Statistical Package: R
Lags Considered: lags 0-3
Pollutant: PMio
Averaging Time: 24h
Mean (SD): 21.2 (8.2)
Range: 55.2
Monitoring Stations: 2 (monitored &
modeled)
Copollutant: NR
Co-pollutant Correlation
n/a
PM Increment: 10 /yglm3
Percent Change (Lower CI, Upper CI),
lag:
Tot. Cardiovascular, Indigenous: -3.43 (¦
9.00, 2.49), lag 0
Tot Cardiovascular, Non-Indigenous: -
3.78(-13.4, 6.91), lag 0
"figure 3. percent change in hospital
admissions per 10//g/m3 increase in PMio
Reference: Henrotin et al. (2007,
093270)
Period of Study: March 1994-
December 2004
Location: Dijon, France
Outcome: Ischemic and hemorrhagic
strokes
Age Groups: All
Study Design: Bi-directional case-
crossover
N: 1487 (ischemic) and 220
(hemorrhagic) stroke patients
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature, relative
humidity, influenza epidemics, holidays
Season: NR
Dose-response Investigated? Yes
Statistical Package: STATA software
v. 8.2
Lags Considered: 0-3 days
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max): 21.1 (2-103)
SD - 11.3
Monitoring Stations: 1
Copollutant: NR
PM Increment: 10/yg/m3
OR Estimate [CI]: Ischemic stroke
Same-day lag: 1.009(0.930,1.094]
1-day	lag: 1.011 [0.998,1.094]
2-day	lag: 0.960 [0.889,1.036]
3-day	lag: 0.990 [0.919,1.066]
Hemorrhagic stroke
Same-day lag: 0.901 [0.730,1.111]
1-day	lag: 1.014(0.828,1.241]
2-day	lag: 1.100 [0.903,1.339]
3-day	lag: 0.991 [0.881,1.212]
Notes: Ischemic stroke ORs were also
categorized into male and female, yielding
similar results (none were significant for
any lag days).
July 2009
E-90
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Issever et al. (2005, 097736) Outcome: Acute coronary syndrome
(ACS)
Period of Study: 1 Jan, 1997-31 Dec,
2001
Location: Istanbul, Turkey
Age Groups: All
Study Design: Time series
N: 2889 ACS admissions
Statistical Analyses: Multiple stepwise
regression, Pearson correlation
Covariates: Humidity, temperature,
pressure
Season: NR
Dose-response Investigated? No
Pollutant: PMio
Averaging Time: 24 h
Mean: NR
Monitoring Stations: 1
Copollutant (correlation): ACS:
r - 0.37 (p - 0.003)
ACS controlled for temp: r - 0.29
(p - 0.02)
PM Increment: NR
RR Estimate [CI]: NR
Notes: This study focused more on the
seasonal change in acute coronary
syndrome admissions.
Reference: Jalaludin et al. (2006,
189416)
Period of Study: 1 Jan, 1997-31 Dec,
2001
Location: Sydney, Australia
Outcome (ICD-9): Cardiovascular disease
(390-459), cardiac disease (390-429),
ischemic heart disease (410-413) and
cerebrovascular disease or stroke (430-
438)
Age Groups: 65+ yrs
Study Design: Time series
N: NR
Statistical Analyses: GAM, GLM
Covariates: Temperature, humidity
Season: Warm (Nov-Apr) and cool (May-
Oct)
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0-3
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max): 16.8 (3.8-103.9)
SD - 7.2
Monitoring Stations: 14
Copollutant (correlation): Warm
BSP: r - 0.82
PM2.6: r - 0.89
O3: r - 0.59
NO?: r - 0.44;
CO: r - 0.31
SO2: r - 0.37
Cool
BSP: r - 0.75
PM2.6: r - 0.88
0s: r - 0.22
NO2: r - 0.67
CO: r - 0.48
SO2: r - 0.46
Other variables: Warm
Temp: r - 0.36
Rel humidity: r - -0.25
Cool
Temp: r - 0.13
Rel humidity: r - 0.05
PM Increment: 7.8/yglm3 (IQR)
Percent Change Estimate [CI]: All CVD
Same-day lag: 0.72 [-0.14,1.60]
Avg 0-1 day lag: 0.25
[-0.61,1.12]
Cool (same-day lag): 1.34(0.08,2.61]
Warm (same-day lag): 0.33 [-0.83,1.50]
Cardiac disease
Same-day lag: 1.15 [0.14,2.18]
Avg 0-1 day lag: 0.97
[-0.07,2.02]
Cool (same-day lag): 1.35 [-0.16,2.89]
Warm (same-day lag): 1.12 [-0.23,2.48]
Ischemic heart disease
Same-day lag: 0.59 [-0.95,2.17]
Avg 0-1 day lag: 0.61
[-0.95,2.20]
Cool (same-day lag): 0.33 [-2.00,2.72]
Warm (same-day lag): 0.79 [-1.23,2.85]
Stroke
Same-day lag: -1.66 [-3.48,0.20]
Avg 0-1 day lag: -2.05
[-3.88,0.20]
Cool (same-day lag): 0.46 [-2.17,3.17]
Warm (same-day lag): -3.49 [-5.97,-0.95]
Notes: All other lag-day ORs were
provided, yet none were significant.
Percent change in ED attendance was
also reported graphically
(Fig 1-5).
July 2009
E-91
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Johnston et al. (2007,
155882)
Period of Study: 2000, 2004, 2005
(April-November of each year)
Location: Darwin, Australia
Outcome (ICD-10): All cardiovascular
conditions (I00-I99), including ischemic
heart disease (I20-I25).
Age Groups: All
Study Design: Case-crossover
N: 2466 emergency admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Weekly influenza rates,
temperature, humidity, days with rainfall
> 5mm, public holidays, school holiday
periods (for respiratory conditions only)
Season: April-November (dry season)
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-3
Pollutant: PMio
Averaging Time: 24 h
Median (IQR, 10th—90th percentile,
range): 17.4 (13.6-22.3
10.3-27.7
1.1-70.0)
Monitoring Stations: 1
Copollutant: NR
PM Increment: 10 /yglm3
OR Estimate [95% CI]: All respiratory
conditions: Ischemic heart disease:
Lag 0:0.82 [0.68-0.98]
Lag 0 (non-indigenous): 0.75 [0.61-0.93]
Lag 3 (indigenous): 1.71 [1.14-2.55]
Notes:
Figure 5: OR and 95% CI for hospital
admissions for cardiovascular conditions.
Summary: Negative associations in
overall study population and in non-
indigenous people. Positive associations
in Indigenous people at Lag 1, Lag 2, and
Lag 3.
Figure 6: OR and 95% CI for hospital
admissions for ischaemic heart disease.
Summary: Negative associations in
overall study population and non-
indigenous people. Positive association in
indigenous people.
Reference: Koken et al. (2003, 049466) Outcome (ICD-9): Acute myocardial
infarction (410.00-410.92), pulmonary
Period of Study: July and August, 1993- heart disease (416.0-416.9), cardiac
1997
Location: Denver, Colorado
dysrhythmias (427.0-427.9), congestive
heart failure (428.0)
Age Groups: 65+ yrs
Study Design: Time series
N: 298 days
Statistical Analyses: GLM, GEE
Covariates: Maximum temp and dew
point temp
Season: NR
Dose-response Investigated: Yes
Statistical Package: SAS (PR0C
GENM0D)
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max): 24.2 (7.0-51.6)
SD - 6.25
Monitoring Stations: 3
Copollutant (correlation): NO2:
r - 0.56
SO2: r - 0.36
O3: r = 0.03
CO: r - 0.25
Other variables: Max temp: r - 0.38
Dew point temp: r - -0.24
PM Increment: 8.0/yglm3 (IQR)
Percent Change Estimate [CI]: No PM
data reported
Lags Considered: 0-4 days
July 2009
E-92
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lanki et al„ (2006, 089788) Outcome (ICD-9): Acute myocardial
infarction
Period of Study: 1992-2000
Location: Augsburg, Barcelona, Helsinki,
Rome, and Stockholm
(410
ICD-10:121,122)
Age Groups: 35+ yrs, < 75 yrs, 75 +
yrs
Study Design: Time series
N: 26,854 hospitalizations
Statistical Analyses: GAM
Covariates: Temperature, barometric
pressure
Season: Warm (April-September) and
cold (October-March)
Dose-response Investigated: No
Statistical Package: R package mgcv
0.9-5
Lags Considered: 0-3 days
Pollutant: PMio
Averaging Time: 24 h
Median: Augsburg: 43.5
Barcelona: 57.4
Helsinki: 21.0
Rome: 48.5
Stockholm: 12.5
Copollutant (correlation): Augsburg
PNC: r - 0.53
CO: r - 0.56
NO?: r - 0.64
0s: r - 0.43
Barcelona: PNC: r - 0.38
CO: r - 0.44
NO?: r - 0.48
0s:r - 0.01
Helsinki: PNC: r - 0.45
CO: r - 0.21
NO?: r - 0.40
O3: r - 0.40
Rome: PNC: r - 0.32
CO: r - 0.41
NO?: r - 0.29
O3: r - 0.59
Stockholm: PNC: r - 0.06
CO: r - 0.41
NO?: r - 0.29
O3: r - 0.59
PM Increment: 10/yg/m3
Pooled Rate Ratio [CI]: All 5 cities (35 +
yrs)
Same-day lag: 1.003 [0.995,1.011 ]
1-day	lag: 1.001 [0.990,1.011]
2-day	lag: 1.002 [0.994,1.010]
3-day	lag: 1.002 [0.991,1.013]
3 cities with hospital discharge register
(35+ yrs)
Same-day lag: 1.003(0.994,1.012]
1-day	lag: 0.997 [0.988,1.006]
2-day	lag: 1.003 [0.995,1.012]
3-day	lag: 1.003 [0.986,1.020]
Warm season (35+ yrs)
Same-day lag: 1.006(0.990,1.022]
1-day	lag: 1.000 [0.985,1.016]
2-day	lag: 1.005 [0.990,1.020]
3-day	lag: 1.010(0.995,1.025]
Cold season (35+ yrs)
Same-day lag: 1.001 [0.991,1.012]
1-day	lag: 0.998 [0.987,1.009]
2-day	lag: 1.001 [0.991,1.012]
3-day	lag: 0.991 [0.981,1.002]
Age >75
Non-fatal
Same-day lag: 1.012 [0.995,1.029]
1-day	lag: 1.000 [0.983,1.017]
2-day	lag: 0.999 [0.982,1.017]
3-day	lag: 1.001 [0.984,1.018]
Fatal
Same-day lag: 1.009 [0.985,1.034]
1-day	lag: 0.998 [0.974,1.023]
2-day	lag: 1.003 [0.978,1.028]
3-day	lag: 1.018 [0.975,1.063]
Notes: Pooled rate ratios were also
provided for groups < 75 yielding similar
results to the overall 3-city data.
July 2009
E-93
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lee et al. (2003, 095552)
Period of Study: 1 Dec, 1997-31 Dec,
1999
Location: Seoul, Korea
Outcome (ICD-10): Angina pectoris (I20),
acute/subsequent myocardial infarction
(121123), other acute ischemic heart
diseases (124)
Age Groups: All ages, 64+ yrs
Study Design: Time series
l\l: 822 days
Statistical Analyses: GAM with LOESS,
Pearson correlation
Covariates: Temperature, relative
humidity, day of the week
Season: Summer (Jun-Aug) and winter
Dose-response Investigated: Yes
Statistical Package: NR
Lags Considered: 0-6 days
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): 64.0(31.8)
Monitoring Stations: 27
Copollutant (correlation): All year
SO?: r - 0.59
NO?: r - 0.74
O3: r - 0.11
CO: r - 0.60
Temp: r - -0.07
Humidity: r - 0.02
Summer
SO2: r - 0.61
NO2: r - 0.73
O3: r - 0.64
CO: r - 0.55
Temp: r - -0.01
Humidity: r - -0.11
PM Increment: 40.4/yg/m3 (IQR)
RR Estimate [CI]: All year
All ages: 0.99 [0.96,1.01]
64+ yrs: 1.05(1.01,1.10]
Summer
All ages: 1.03 [0.97,1.09]
64+ yrs: 1.09(1.00,1.19]
Two-pollutant model
CO (1 ppmlQI): 1.04(0.98,1.11]
0s (21.7 ppb 1011:1.07(1.03,1.11]
NO2 (14.6 ppb IQI):1.09(1.02,1.16]
SO2 (4.4 ppb): 0.98 (0.94,1.03]
Reference: Lee et al. (2008,192076)
Period of Study: 1996-2005
Location: Taipei, Taiwan
Outcome: Congestive Heart Failure HA
(ICD 9: 428)
Age Groups: NR
Study Design: case-crossover
N: 18593 events
Statistical Analyses: conditional
logistic regression
Covariates: temperature, humidity
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: lags 0-2
Pollutant: PM10
Averaging Time: 24h
Mean: 49.94
Min: 11.33
25": 33.37
50": 45.05
75": 60.82
Max: 234.92
Monitoring Stations: 6
Copollutant: SO2, CO, NO2, O3
Co-pollutant Correlation
SO2: 0.52
CO: 0.67
NO2: 0.35
0s: 0.39
PM Increment: Interquartile
(27.45 /yglm3)
Odds Ratio (Lower CI, Upper CI):
Wf Hypertension: 1.23 (1.15,1.32)
W/o Hypertension: 1.20 (1.15,1.25)
W/ Diabetes: 1.20 (1.12,1.40)
W/o Diabetes: 1.21 (1.15,1.26)
Wf Dysrhythmia: 1.17 (1.08,1.27)
W/o Dysrhythmia: 1.22 (1.17,1.27)
WIC0PD: 1.21 (1.07,1.36)
W/oCOPD: 1.21 (1.16,1.25)
July 2009
E-94
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Larrieu et al. (2007, 093031)
Period of Study: 1998 ¦ 2003
Location: 8 French urban area: Bordeaux,
Le Havre, Lille, Lyon, Marseille, Paris,
Rouen, and Toulouse
Outcome (ICD-10): Hospital admissions
for cardiovascular disease (I00-I99),
cardiac disease (I00-I52), ischemic heart
disease (I20-I25), and stroke
(cerebrovascular disease: 160-64 and
transient ischemic attack: G45-G46).
Age Groups: All, and 65 +
Study Design: Time series
N: Statistical Analyses: generalized
additive Poisson regression
Covariates: Temperature, holidays,
influenza epidemic periods, long-term
trend, season, day of the week,
Season: NR
Dose-response Investigated: No
Statistical Package: R 2.2.1
Lags Considered: 0 -1 day lag (mean)
Pollutant: PMio
PM Increment: 10/yg/m3
Averaging Time: 24 h
ERR [95% CI]:
Mean: Bordeaux: 21.0
CVD: All ages: 0.7(0.1,1.2]
Le Havre: 21.7
65+ years: 1.1 [0.5,1.7]
Lille: 22.1
Cardiac diseases: All ages: 0.8 [0.2,1.4]
Lyon: 24.6
65+ years: 1.5 [0.7, 2.2]
Marseille: 28.9
Ischemic heart diseases: All ages: 1.9

[0.8, 3.0]
Paris: 23.1
Rouen: 21.2
65+ years: 2.9 [1.5, 4.3]

Strokes: All ages: 0.2 [-1.6,1.9]
Toulouse: 21.8

65+ years: 0.8 [-0.9, 2.5]
Monitoring Stations: 32
Copollutant: NR

Reference: Le Tertre et al. (2002,
023746)
Period of Study: 1990-1997
Location: Barcelona, Birmingham,
London, Milan, the Netherlands, Paris,
Rome, and Stockholm
Outcome (ICD-9): Cardiac diseases (390-
429), ischemic heart disease (410-413),
and stroke (430-438)
Age Groups: < 65 yrs, 65+ yrs
Study Design: Time series
N: NR
Statistical Analyses: GAM
Covariates: Long term trend, season,
days of the week, holidays, influenza
epidemics, temperature, and humidity
Season: NR
Dose-response Investigated: No
Statistical Package: S-Plus
Lags Considered: 0-3 days
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): Barcelona: 55.7 (18.4)
Birmingham: 24.8 (13.1)
London: 28.4 (12.3)
Milan: 51.5(22.7)
Netherlands: 39.5 (19.9)
Paris: 22.7 (10.8)
Rome: 52.5 (12.9)
Stockholm: 15.5 (7.2)
Monitoring Stations: 1-12
Copollutant: NR
PM Increment: 10/yg/m3
Pooled Percent Increase [CI]: Cardiac
(all ages)
Fixed: 0.5 [0.3,0.7]
Random: 0.5 [0.2,0.8]
Cardiac (over 65)
Fixed: 0.7 [0.4,1.0]
Random: 0.7 [0.4,1.0]
IHD (< 65)
Fixed: 0.3 [-0.1,0.6]
Random: 0.3 [-0.2,0.7]
IHD (over 65)
Fixed: 0.6 [0.3,0.8]; Random: 0.8
[0.3,1.2]
Stroke (over 65)
Fixed: 0.0 [-0.3,0.3]; Random: 0.0 [¦
0.3,0.3]
Deaths: Cardiac: 0.5 [0.2,0.8]; Cardiac
(65+): 0.7(0.4,1.0]
IHD (65+): 0.8 [0.3,1.2]
Notes: Estimated percentage increases
are also provided by city for cardiac
admissions and ischemic heart disease in
Fig 1-3.
July 2009
E-95
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Mann et al. (2002, 036723) Outcome (ICD-9): Ischemic heart
(410-414), secondary congestive heart
Period of Study: 1988-1995
Location: South Coast Air Basin,
California
failure (sCHF) (428), and secondary
arrhythmia (sARR) (426, 427)
Age Groups: All, 40-59 yrs, > 60 yrs
Study Design: Time series
N: 54,863 IHD admissions
Statistical Analyses: GAM
Covariates: Temperature, day of the
week, relative humidity
Season: NR
Dose-response Investigated: No
Statistical Package: S-Plus
Lags Considered: 0-5 days
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max): 43.7 (0.22-251)
SD - 27.7
Monitoring Stations: 20
Copollutant (correlation):
r - 0.28
0s: r - 0.20
NO?: r - 0.36
Region 2: CO: r - 0.15
0s: r - 0.57
NO?: r - 0.53
Region 3: CO: r - 0.36
O3: r = 0.30
NO?: r - 0.46
Region 4: CO: r - 0.27
O3: r = 0.33
NO?: r - 0.50
Region 5: CO: r - 0.40
O3: r - 0.43
NO?: r - 0.53
Region 6: CO: r - 0.33
O3: r - 0.20
NO?: r - 0.42
Region 7: CO: r - 0.28
0s: r - 0.48
NO2: r - 0.60
n 1: CO:
PM Increment: 10/yg/m3
Percent Change in IHD Admissions
[CI]: Secondary ARR
Same-day lag: 0.59 [-0.71,1.91 ]
1-day	lag: 0.46 [-0.86,1.80]
2-day	lag: -0.04 [-1.37,1.31]
Secondary CHF
Same-day lag: -0.62 [-1.77,0.55]
1-day	lag: -0.45 [-1.60,0.71]
2-day	lag:-0.36 [-1.52,0.82]
No secondary diagnosis
Same-day lag: -0.25 [-1.23,0.75]
1-day	lag: 0.04 [-0.97,1.06]
2-day	lag: 0.18 [-0.82,1.20]
All IHD admissions: 0.19 [-0.576,0.955]
Ml admissions: -0.10 [-1.33,1.12]
Other acute IHD admissions: 0.36 [¦
0.87,1.60]
July 2009
E-96
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Metzgeret al. (2004,
044222)
Period of Study: August 1993-August
2000
Location: Atlanta Metropolitan area
(Georgia)
Outcome (ICD-9): Emergency visits for
ischemic heart disease (410-414),
cardiac dysrhythmias (427), cardiac
arrest (427.5), congestive heart failure
(428), peripheral vascular and
cerebrovascular disease (433-437, 440,
443-444,451-453), atherosclerosis
(440), and stroke (436).
Age Groups: All
Study Design: Time series
N: 4,407,535 emergency department
visits
Statistical Analyses: Poisson
generalized linear modeling
Covariates: Day of the week, hospital
entry and exit indicator variables,
federally observed holidays, temporal
trends, temperature, dew point
temperature
Season: All
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: 3-day moving avg,
lags 0 -7
Pollutant: PMio
Averaging Time: 24 h
Median (10% - 90% range): 26.3 (13.2,
44.7)
Monitoring Stations: NR
Copollutant (correlation): O3: r - 0.59
NO?: r - 0.49
CO: r - 0.47
SO2: r - 0.20
PM2.6: r - 0.84
PM 10-2.5: r - 0.59
UFP: r - -0.13
PM2.6 water-sol
metals: r - 0.74
PM2.6 sulfates: r - 0.74
PM2.6 acidity: r - 0.68
PM2.6 organic carbon: r - 0.69
PM2.6 elemental carbon: r - 0.56
oxygenated hydrocarbon: r - 0.58
Other variables: Temperature: r - 0.58
Dew point: r - 0.44
PM Increment: 10 //g/m3 (approximately
1 SD)
RR [95% CI]: For 3-day moving avg: All
CVD: 1.009 [0.998,1.019]
Dysrhythmia: 1.008 [0.989,1.029]
Congestive heart failure: 0.992 [0.968—
1.016]
Ischemic heart disease: 1.011 [0.992—
1.030]
Peripheral vascular and cerebrovascular
disease: 1.020 [0.999-1.043]
Notes: Results for Lags 0-7 expressed in
figures
Figure 1: RR (95% CI) for single-day lag
models for the association of ER visits for
CVD with daily ambient PM10.
Summary: Statistically significant
association at Lag 0. Positive but not
statistically significant association at Lag
1. Negative, statistically significant
association at Lag 7, and negative asso-
ciations at Lag 2 through Lag 6.
Reference: Middleton et al. (2008,
156760)
Period of Study: 1995-1998, 2000 ¦
2004
Location: Nicosia, Cyprus
Outcome: Hospital admissions for all
cardiovascular disease (ICD-10:100-I52).
Age Groups: All, also stratified by age
(< 15 vs. >15 years)
Study Design: Time series
Statistical Analyses: Generalized
additive Poisson models
Covariates: Seasonality, day of the
week, long- and short-term trend,
temperature, relative humidity
Dose-response Investigated: No
Statistical Package: STATA SE 9.0, R
2.2.0
Lags Considered: Lag 0 -2 days
Pollutant: PM10
Averaging Time: 24 h
Mean (SD
median
5%- 95%
range): Cold: 57.6 (52.5
50.8
20.0-103.0
5.0-1370.6)
Warm: 53.4 (50.5
30.7
32.0-77.6
18.4-933.5)
Monitoring Stations: 2
Copollutant: NR
PM Increment: 10 /yglm3, and across
quartiles of increasing levels of PM10
Percentage increase estimate [CI]: All
age/sex groups (Lag 0): All admissions:
0.85(0.55,1.15)
Cardiovascular: 1.18 (-0.01, 2.37)
Nicosia residents (Lag 0):
Cardiovascular: 0.73 (-0.62, 2.09)
Males (Lag 0): All admissions: 0.96
(0.54,1.39)
Cardiovascular: 1.27 (-0.15, 2.72)
Females (Lag 0): All admissions: 0.74
(0.31,1.18)
Cardiovascular: 0.99 (-1.11, 3.14)
Aged < 15 years (Lag 0): All
admissions: 0.47 (-0.13,1.08)
Aged > 15 years (Lag 0): All
admissions: 0.98 (0.63,1.33)
July 2009
E-97
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Peel et al. (2007, 090442)
Period of Study: 1 Jan, 1993-31 Aug,
2000
Location: Atlanta, GA
Outcome (ICD-9): Ischemic heart
(410-414), dysrhythmia (427), congestive
heart failure (428), peripheral vascular
and cerebrovascular disease (433-437,
440,443,444,451-453)
Age Groups: All
Study Design: Case-crossover
N: 4,407,535 ED visits
Statistical Analyses: Conditional
logistic regression
Covariates: Avg temp and dew point
temp
Season: NR
Dose-response Investigated: No
Statistical Package: SAS v. 9.1
Lags Considered: 0-2 days
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): Daily levels: 27.9 (12.3)
Diff in case and control day avgs: 9.1
(7.5)
Monitoring Stations: 1
Copollutant: NR
PM Increment: 10/yg/m3
OR Estimate [CI]: All CVD: 1.010
[1.000,1.020]
IHD: 1.009(0.991,1.027]
Dysrhythmia: 1.011 [0.991,1.031]
PeripheralfCerebrovascular disease:
1.017(0.996,1.039]
CHF: 1.001 [0.978,1.024]
With comorbid hypertension
IHD: 1.003[0.973,1.034]
Dysrhythmia: 1.037 [0.988,1.089]
PeripheralfCerebrovascular disease:
1.024 [0.990,1.060]
CHF: 1.041 [0.999,1.084]
No comorbid hypertension
IHD: 1.013(0.991,1.036]
Dysrhythmia: 1.006 [0.985,1.028]
PeripheralfCerebrovascular disease:
1.013 [0.987,1.040]
CHF: 0.982 [0.955,1.010]
With comorbid diabetes
IHD: 1.022 [0.979,1.067]
Dysrhythmia: 1.049 [0.968,1.137]
PeripheralfCerebrovascular disease:
1.016 [0.965,1.069]
CHF: 1.029 [0.982,1.078]
No comorbid diabetes
IHD: 1.006 [0.987,1.026]
Dysrhythmia: 1.009 [0.989,1.029]
PeripheralfCerebrovascular disease:
1.018 [0.995,1.042]
CHF: 0.992 [0.966,1.019]
With comorbid COPD
IHD: 0.981 [0.921,1.044]
Dysrhythmia: 0.984 [0.889,1.088]
PeripheralfCerebrovascular disease:
1.086 [0.998,1.181]
CHF: 1.010 [0.954,1.069]
No comorbid COPD
IHD: 1.012(0.993,1.031]
Dysrhythmia: 1.012 [0.992,1.032]
PeripheralfCerebrovascular disease:
1.013 [0.991,1.035]
CHF: 0.999 [0.974,1.025]
July 2009
E-98
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Pope et al„ (2006, 091246) Outcome: Myocardial infarction or
unstable angina (ICD codes not reported)
Period of Study: 1994- 2004
Location: Wasatch Front area, Utah
Age Groups: All
Study Design: Case-crossover
N: 12,865 patients who underwent
coronary arteriography
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and dew point
temperature
Season: NR
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: 0- to 3-day lag, 2- to
4-day lagged moving averages
Pollutant: PMio
Averaging Time: 24 h
Mean (SD
maximum): Ogden: 28.5 (16.5
163)
SLC Hawthorne: 27.7 (17.4
162)
Provo/Orem, Lindom: 32.7 (21.1
240)
SLC AMC: 35.9 (20.4
161)
SLC North: 45.1 (25.1
199)
Monitoring Stations: 5
Copollutant: NR
PM Increment: 10/yg/m3
Percent increase in risk [95% CI]:
Results summarized in figure (see notes).
Notes: Figure 1: Percent increase in risk
(and 95% CI) of acute coronary events
associated with 10 /yglm3 of PMio for
different lag structures.
Summary of Figure 1: Positive,
statistically significant or marginally
significant associations between
association seen for Lag 0, Lag 1
and 2-, 3-, and 4-day moving averages.
Non-statistically significant associations
Reference: Santos et al. (2008,192004) Outcome: Cardiac Arrhythmia ER Visits
(ICD 10:145-I49)
Period of Study: Jan 1998 - Aug 1999
Location: Sao Paulo, Bazil
Age Groups: 17 + yrs
Study Design: time series
N: 3251 ER visits
Statistical Analyses: Poisson
Covariates: temperature, humidity,
seasonality
Dose-response Investigated? Yes
Statistical Package: S-Plus
Lags Considered: lags 0-13
Pollutant: PMio
Averaging Time: 24h
Mean (SD): 48.64 (20.34)
Min: 18.68
Max: 137.76
Monitoring Stations: 14
Copollutant: SO2, CO, NO2, O3
Co-pollutant Correlation
SO2: 0.675*
CO: 0.580*
NO?: 0.781*
0s: 0.438*
*p < 0.01
PM Increment: Interquartile Range (22.2
j"g/m3)
Percent Increase (Lower CI, Upper CI):
PM10+ N02.C0: -5.6 (-12.7, 2.1)
PM10+ CO: -1.1 (-7.0,5.1)
PM10+ NO2: -2.4 (-9.4, 5.1)
Figure 1. PMio effects, reported as
percent increase, on arrhythmia ER visits
caused by interquartile range increases,
lags 0-6.
Figure 2. Relative risks and 95% CI for
arrhythmia ER visits according to the
division of air pollutant daily
concentrations in quintiles.
July 2009
E-99
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Tolbert et al. (2007,
090316)
Period of Study: 1993 ¦ 2004
Location: Atlanta Metropolitan area,
Georgia
Outcome (ICD-9): Combined CVD group,
including: Ischemic heart disease (410-
414), cardiac dysrhythmias (427),
congestive heart failure (428), and
peripheral vascular and cardiovascular
disease (433-437,440, 443-445, and
451-453).
Age Groups: All
Study Design: Time series
N: 10,234,490 ER visits (283,360 and
1,072,429 visits included in the CVD and
RD groups, respectively)
Statistical Analyses: Poisson
generalized linear models
Covariates: Long-term temporal trends,
season (for RD outcome), temperature,
dew point, days of week, federal
holidays, hospital entry and exit
Season: All
Dose-response Investigated: No
Statistical Package: SAS version 9.1
Lags Considered: 3-day moving avgflag
0-2)
Pollutant: PMio
Averaging Time: 24 h
Mean (median
IQR, range, 10th—90th percentiles):
26.6 (24.8
17.5-33.8
0.5-98.4
12.3-42.8)
Monitoring Stations: NR
Copollutant (correlation): O3: r - 0.59
NO?: r - 0.53
CO: r - 0.51
SO2: r - 0.21
Coarse PM: r - 0.67
PM2.6: r - 0.84
PM2.6 SO4: r - 0.69
PM2.6 EC: r - 0.61
PM2.6 0C: r - 0.65
PM2.6 TC: r - 0.67
PM2.6 water-sol metals: r - 0.73
0HC: r - 0.53
PM Increment: 16.30/yglm3 (IQR)
Risk ratio [95% CI]: Single pollutant
models: CVD: 1.008 (0.997-1.020)
Reference: Tsai et al. (2003, 080133)
Period of Study: 1997-2000
Location: Kaohsiung, Taiwan
Outcome (ICD-9): Cerebrovascular
diseases (430-438), subarachnoid
hemorrhagic stroke (430), primary
intracerebral hemorrhage (431-432),
ischemic stroke (433-435), and others
(436-438)
Age Groups: All
Study Design: Case-crossover
N: 23,179 admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and humidity
Season: NR
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: Cumulative 0-2 days
Pollutant: PM10
Averaging Time: 24 h
Mean (min-max): 78.82 (20.50-217.33)
Monitoring Stations: 6
Copollutant: NR
PM Increment: 66.33/yg/m3 (IQR)
OR Estimate [CI]: Two-pollutant model
(all stroke admissions)
Primary intracerebral hemorrhage (PIH)
Adj for SO2:1.55(1.31,1.83]
Adj for NO2:1.28 [1.01,1.61];
Adj for CO: 1.45(1.20,1.74]
Adj for 0s: 1.56 [1.27,1.91]
Ischemic stroke (IS)
Adj for SO2:1.46(1.32,1.61]
Adj for NO2:1.16 [1.01,1.34]
Adj for CO: 1.35(1.21,1.51]
Adj for 0s: 1.51 [1.34,1.71]
Single-pollutant model
Temp >20°C
PIH: 1.54[1.31,1.81]
IS: 1.46[1.32,1.61]
Temp < 20°C
PIH: 0.82 [0.48,1.40]
IS: 0.97[0.65,1.44]
July 2009
E-100
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ulirsch et al. (2007, 091332) Outcome (ICD-9): CVD (390-429).
Age Groups: 65 +
Period of Study: November 1994-
March 2000
Location: Pocatello, Idaho and Chubbuck,
Idaho
Study Design: Time series
N: 39,347 admissions/visits
Statistical Analyses: Log-linear
generalized linear models
Covariates: Time, temperature, relative
humidity, influenza, day of the week
Season: All, and separate analyses were
performed for the all-age group for cool
months (October-March) vs. warm
months (April-September).
Dose-response Investigated: No
Statistical Package: S-plus version 6.1
Lags Considered: 0- to 4-day lags, and
mean of days 0 -4
Pollutant: PMio
Averaging Time: 24 h
Mean (range
10th - 90th percentiles): 24.2 (3.0-
183.0
10.5-40.7)
Monitoring Stations: 4
Copollutant (correlation): NO2:
r - 0.47
Other variables: Correlation for PM10
between monitors: r - 0.42-0.87
PM Increment: 50 //g|m3, and
24.3/yglm3 (mean increase in PM10)
Mean percent of change (% change in
the mean number of daily admissions
and visits) [95% CI]:
For 24.3//gfm3 increase in PM10: All-
age RD/CVD: 3.7[1.3, 6.3]
All-age CVD (Lag 0): -0.02 [-5.9, 6.3]
All-age CVD (Lag 1): 1.9 [-4.1, 8.4]
All-age CVD (Lag 2): -3.1 [-9.1,3.4]
All-age CVD (Lag 3): 0.5 [-5.6, 6.9]
All-age CVD (Lag 4):-1.7 [-4.3, 0.9]
Lag 0-4 days: -0.5 [-8.0, 7.6]
For 50 /yglm3 increase in PM10 (single
pollutant models, CIs not given): All-age
respiratory disease: 8.4
All-age RD/CVD: 7.9
18-64 years RD: 7.2
All-age CVD (Lag 3): 1.0
All-age CVD (Lag 4): -3.6
All-age CVD (Lag 0-4):-1.1
Notes: Included urgent care visits as well
as emergency department visits and
hospital admissions.
Reference: Yang et al. (2007, 092847) Outcome: Congestive Heart Failure HA
(ICD 9: 428)
Period of Study: 1996 ¦ 2005
Location: Taipei, Taiwan
Age Groups: NR
Study Design: case-crossover
N: 24,240 events
Statistical Analyses: Poisson
Covariates: temperature, humidity
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: lags 0-3
Pollutant: PM10
Averaging Time: 24h
Mean: 49.47
Min: 14.42
25": 33.08
50": 44.71
75": 60.10
Max: 234.91
Monitoring Stations: 6
Copollutant: NR
Co-pollutant Correlation
n/a
PM Increment: Interquartile
(27.02/yg/m3)
Odds Ratio (Lower CI, Upper CI):
Temp £20°C
PM10: 1.15(1.10-1.21)*
PM10+ SO2: 1.23 (1.17,1.30)*
PM10+NO2: 1.03 (0.97,1.10)
PM10+ CO2:1.09(1.03,1.15)*
PM10+ 03:1.10 (1.04,1.15)*
Temp < 20°C
PM10: 0.99(0.93,1.05)
PM10+SO2: 0.96 (0.89,1.03)
PM10+ NO?: 0.97 (0.90,1.04)
PM10+CO2: 0.96 (0.90,1.03)
PM10+ O3: 1.00 (0.94,1.05)
*p < 0.05
July 2009
E-101
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Yang et al. (2007, 092847) Outcome: Congestive Heart Failure HA
Period of Study: 1996 - 2001
Location: Taipei, Taiwan
Age Groups: NR
Study Design: case-crossover
N: NR
Statistical Analyses: Poisson
Covariates: NR
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: lags 0-3
Pollutant: PMio
Averaging Time: 24h
Mean (SD):
Index days
111.68(38.32)
Comparison days
55.43 (24.66)
Monitoring Stations: 7
Copollutant: NR
Co-pollutant Correlation
n/a
PM Increment: Index (> 125 /yglm3) vs.
Comparison (n125 //g/m3)
Relative Risk (Lower CI, Upper CI),
lag:
0.915(0.805,1.041), lag 0
1.114(0.993,1.250), lag 1
0.983 (0.873,1.106), lag 2
0.974 (0.870,1.090), lag 3
Reference: Villeneuve et al. (2006,
090191)
Period of Study: April, 1992 -March,
2002
Location: Edmonton, Canada
Outcome (ICD-9): Stroke (430-438),
including ischemic stroke (434-436),
hemorrhagic stroke (430,432), and
transient ischemic attacks (TIA) (435).
Age Groups: 65+ yrs
Study Design: Case-crossover
N: 12,422 visits
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and relative
humidity
Season: summer (Apr-Sep), winter (Oct-
Mar)
Dose-response Investigated: No
Statistical Package: SAS (PHREG)
Lags Considered: 0-, 1-, and 3-day
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): All year:
24.2(14.8)
Summer: 25.9 (16.4)
Winter: 22.6(12.9)
Monitoring Stations: 3
Copollutant (correlation):
All year
SO2: r - 0.19
NO?: r - 0.34;
CO: r - 0.30
03-mean: r - 0.07;
03-max: r - 0.22
PM2.6: r - 0.79
Summer
SO2: r - 0.18
NO2: r - 0.57;
CO: r - 0.38
03-mean: r - 0.20;
03-max: r - 0.40
PM2.6: r - 0.85
Winter
SO2: r - 0.27
NO2: r - 0.48;
CO: r - 0.53
03-mean: r - -0.26;
03-max: r - -0.09
PM2.6: r - 0.70
PM Increment: /yglm3 (IQR)
All year: 16.0
Summer: 17.5
Winter: 16.0
Adjusted OR Estimate [CI]: Acute
ischemic stroke
All year
Same-day lag: 0.98 [0.94,1.03]
1-day lag: 1.00 [0.96,1.05]
3-day lag: 0.99 [.93,1.05]
summer
Same-day lag: 0.93 [0.87,1.00]
1-day lag: 1.01 [0.94,1.08]
3-day lag: 0.96 [0.88,1.04]
Winter
Same-day lag: 1.04 [0.97,1.11]
1-day lag: 1.00 [0.94,1.06];
3-day lag: 1.05 [0.95,1.15]
Hemorrhagic stroke
All year
Same-day lag: 1.01 [0.90,1.12]
1-day lag: 1.03 [0.93,1.15]
3-day lag: 1.13(0.98,1.30]
summer
Same-day lag: 1.02 [0.88,1.20]
1-day lag: 1.07 [0.91,1.26]
3-day lag: 1.20 [0.98,1.46]
Winter
Same-day lag: 1.05 [0.90,1.22]
1-day lag: 1.04 [0.91,1.19]
3-day lag: 1.11 [0.90,1.37]
Transient cerebral ischemic attack
All year
Same-day lag: 0.96 [0.90,1.02]
July 2009
E-102
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1-day lag: 0.99 [0.94,1.05]
3-day lag: 0.94 [0.87,1.01]
summer
Same-day lag: 0.97 [0.89,1.09]
1-day lag: 0.99 [0.91,1.08]
3-day lag: 0.94 [0.84,1.04]
Winter
Same-day lag: 0.95 [0.87,1.04]
1-day lag: 0.99 [0.92,1.07]
3-day lag: 0.93 [0.83,1.05]
Notes: Adjusted ORs are provided for an
IQR increase in the 3-day mean in Fig 1-4
for single and two-pollutant models.
July 2009
E-103
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: von Klot et al. (2005,
088070)
Period of Study: 1992-2001
Location: Augsburg, Germany
Barcelona, Spain
Helsinki, Finland
Rome, Italy
Stockholm, Sweden
Outcome (ICD-9): Acute myocardial
infarction (410
ICD-10:121-122), angina pectoris (411,
413
ICD-10:120,124), dysrhythmia (427
ICD-10:146.0,46.9,147-I49, R00.1,
R00.8), heart failure (428
ICD-10: 150)
Age Groups: 35+ yrs
Study Design: Cohort
N: 22,006 Ml survivors
Statistical Analyses: GAM, Spearman
correlation
Covariates: Temperature, dew point
temp, avg barometric pressure, relative
humidity
Season: NR
Dose-response Investigated: No
Statistical Package: R
Lags Considered: 0-3 days
Pollutant: PMio
Averaging Time: 24 h
Mean (5th-95th percentile): Augsburg:
44.7 (16.8-81.4)
Barcelona: 52.2 (25.3-89.2)
Helsinki: 25.3 (9.5-57.6)
Rome: 51.1 (23.3-89.4)
Stockholm: 14.6(6.4-30.0)
Monitoring Stations: NR
Copollutant (correlation): Augsburg
PNC: r - 0.52
CO: r - 0.57;
NO?: r - 0.64
0a: r - -0.32
Barcelona
PNC: r - 0.29
CO: r - 0.39;
NO?: r - 0.36
0a: r - -0.14
Helsinki
PNC: r - 0.46
CO: r - 0.21;
NO?: r - 0.40
0s: r - 0.02
Rome
PNC: r - 0.33
CO: r - 0.31;
NO?: r - 0.48
Oa: r - -0.22
Stockholm
PNC: r - 0.06
CO: r - 0.38;
NO?: r - 0.29
Os:r - 0.15
PM Increment: 10/yg/m3
Pooled RR Estimate [CI]:
All cardiac admissions: 1.021
[1.005,1.048]
Myocardial infarction: 1.026
[0.995,1.058]
Angina pectoris: 1.008 [0.986,1.032]
Notes: Rate ratios for 0-3 day lags are
provided in graphical form (Fig 1). Same-
day levels were significantly associated
with cardiac readmissions.
July 2009
E-104
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Wellenius et al. (2005,
087483)
Period of Study: 1 Jan, 1987-30 Nov,
1999
Location: Pittsburgh, Pennsylvania
Outcome (ICD-9): Congestive heart
failure (428.0-428.1)
Age Groups: 65+ yrs
Study Design: Case-crossover
N: 55,019 patients
Statistical Analyses: Conditional
logistic regression, Pearson's pairwise
correlation
Covariates: Temperature, barometric
pressure, dew point
Season: NR
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: 0-3 days
Pollutant: PMio
Averaging Time: 24 h
Mean (5th-95th percentile): 31.06 (8.89-
70.49)
SD - 20.10
Monitoring Stations: 17
Copollutant (correlation): CO: r - 0.57
NO?: r - 0.64
0s: r - 0.29
SO?: r - 0.51
PM Increment: 24/yg/m3 (IQR)
Percent Increase [CI]: Single-pollutant:
3.07[1.59,4.57]
Adj. for CO:-1.10 [-3.02,0.86]
Adj. for NOz: 0.52 [-1.46,2.53]
Adj. for 0s: 2.80(1.29,4.33]
Adj. for SO2: 2.18[0.37,4.02]
Percent Increase (with 10/yg/m3
increment)
1.27(0.66,1.88]
Reference: Wellenius et al. (2005,
088685)
Period of Study: 1 Jan, 1986-30 Nov,
1999
Location: Birmingham, Chicago,
Cleveland, Detroit, Minneapolis, New
Haven, Pittsburgh, Salt Lake City, Seattle
Outcome (ICD-NR): Ischemic stroke and
hemorrhagic stroke
Age Groups: 65+ yrs
Study Design: Case-crossover (time-
stratified)
N: 115,503 hospital admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and humidity
Season: NR
Dose-response Investigated: No
Statistical Package: SAS (v.9) and R-
statistical package
Lags Considered: 0-2 lags
Pollutant: PM10
Averaging Time: 24 h
Mean (SD): 32.69 (19.75)
Monitoring Stations: NR
(data obtained from the US EPA)
Copollutant (correlation): CO: r - 0.43
NO?: r - 0.53
SO2: r - 0.39
Other variables: Temp: r - 0.22
PM Increment: 22.96//g|m3 (IQR)
Percent Increase [CI]: Ischemic (same-day
lag): 1.03 [0.04,2.04]
Hemorrhagic: -0.58 [-5.48,4.58]
Notes: Percent increase in rate for
ischemic and hemorrhagic stroke are
provided for each city in graphical form
(Fig A and B).
Reference: Wellenius et al.,(2006,
088748)
Period of Study: 1 Jan, 1986-30 Nov,
1999
Location: Birmingham, Chicago,
Cleveland, Detroit, Minneapolis, New
Haven, Pittsburgh, Salt Lake City, Seattle
Outcome (ICD-9): Congestive heart
failure (428)
Age Groups: 65+ yrs
Study Design: Case-crossover (time-
stratified)
N: 292,918 admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and barometric
pressure
Season: NR
Dose-response Investigated: No
Statistical Package: SAS (v.9) and R-
statistical package
Lags Considered: 0-3 days
Pollutant: PM10
Averaging Time: 24 h
Median: Overall: 28.3
Birmingham: 33.0
Chicago: 31.5
Cleveland: 34.5
Detroit: 29.5
Minneapolis: 24.0
New Haven: 22.
Seattle: 25.8
Monitoring Stations: NR
(data obtained from the US EPA)
Copollutant: NR
PM Increment: 10 //g|m3
Percent Increase [CI]: Same-day lag: 0.72
[0.35,1.10]
p-value - 0.0002
Notes: City-specific percent increases are
graphed in Fig 1 for same-day lag
showing a significant association in
Chicago, Detroit, Seattle, and the
summary values.
Percent increase in admission rate s are
provided for lag 0-3 days in Fig 2 where
same-day lag showed a significant
association.
July 2009
E-105
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Yang et al. (2004, 094376)
Period of Study: 1997-2000
Location: Kaohsiung, Taiwan
Outcome (ICD-9): Cardiovascular
diseases (410-429)
Age Groups: All
Study Design: Case-crossover
N: 29,661 admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and humidity
Season: NR
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: Cumulative 0-2 days
Pollutant: PMio
Averaging Time: 24 h
Median (min-max): 78.82 (20.50-217.33)
Monitoring Stations: 6
Copollutant: NR
PM Increment: 66.33/yglm3 (IQR)
OR Estimate [CI]: Temp >25°C: 1.439
[1.316,1.573]
Temp < 25°C: 1.568(1.433,1.715]
Adj for SO2
Temp > 25°C: 1.460(1.333,1.599]
Temp < 25°C: 1.543(1.404,1.696]
Adj for NO2
Temp >25°C: 1.306 [1.154,1.478]
Temp < 25°C: 0.912 (0.809,1.028]
Adj for CO
Temp >25°C: 1.260 [1.144,1.388]
Temp < 25°C: 1.259 [1.128,1.406]
Adj for O3
Temp > 25°C: 1.086 [0.967,1.220]
Temp < 25°C: 1.703 [1.541,1.883]
Reference: Yang et al (2008,157160)
Period of Study: 1996 ¦ 2004
Location: Taipei, Taiwan
Outcome (ICD-9): Congestive heart
failure (428)
Age Groups: All
Study Design: Case-crossover
N: 24,240 CHF hospital admissions
Statistical Analyses: Conditional
logistic regression
Covariates: temperature, humidity
Season: All
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: Cumulative lag 0-2
days
Pollutant: PM10
Averaging Time: 24 h
Mean (median, range, IQR):
49.47 (44.71,14.42-234.91, 33.08-
44.71) Monitoring Stations: 6
Copollutant: NR
PM Increment: 27.02/yg/m3 (IQR)
OR [95% CI]:
Single pollutant models: >20 °C: 1.15
[1.10-1.21]
<20 °C: 0.99 [0.93-1.05]
Adjusted for SO2: > 20 °C: 1.23 [1.17-
1.30]
<20 °C: 0.96 [0.89-1.03]
Adjusted for NO2: > 20 °C: 1.03 [0.97-
1.10]
<20 °C: 0.97 [0.90-1.04]
Adjusted for CO: >20 °C: 1.09 [1.03-
1.15]
<20 °C: 0.96 [0.90-1.03]
Adjusted for 0s: > 20 °C: 1.10 [1.04-
1.15]
<20 °C: 1.00[0.94-1.05]
July 2009
E-106
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Zanobetti and Schwartz
(2002,034821)
Period of Study: 1988-1994
Location: Cook county (Chicago), Illinois
Wayne county (Detroit), Michigan
Allegheny county (Pittsburgh),
Pennsylvania
and King county (Seattle), Washington
Outcome (ICD-9): Cardiovascular disease
(390-429) with/without diabetes (250)
Age Groups: 65-74 and 75+ yrs with
diabetes, 65-74 and 75+ yrs without
diabetes
Study Design: Time series
N: NR
Statistical Analyses: GAM, meta-
regression
Covariates: Temperature, prior day's
temperature, relative humidity,
barometric pressure, day of the week
Season: NR
Dose-response Investigated: No
Pollutant: PMio
Averaging Time: 24 h
Median (25-75th percentile): Chicago: 33
(23-46)
Detroit: 32 (21-49)
Pittsburgh: 30 (19-47)
Seattle: 27 (18-39)
Monitoring Stations: NR (obtained from
USEPA Aerometric Information Retrieval
System)
Copollutant: NR
PM Increment: 10 /yglm3
Percent Change [CI]: All four cities
<	75 (w/ diabetes): 1.6 [1.2,2.0]
75+ (w/ diabetes): 2.0(1.6,2.4]
<	75 (w/o diabetes): 0.9 [0.6,1.1]
75+ (w/o diabetes): 1.3 [1.0,1.5]
Chicago
<	75 (w/ diabetes): 1.9 [1.1,2.7]
75+ (w/ diabetes): 2.0(1.1,3.0]
<	75 (w/o diabetes): 0.7 [0.2,1.2]
75+ (w/o diabetes): 1.2 [0.8,1.7]
Detroit
<	75 (w/ diabetes): 1.3 [0.5,2.2]
75+ (w/ diabetes): 2.1 [1.0,3.1]
<	75 (w/o diabetes): 1.2 [0.7,1.7]
75+ (w/o diabetes): 1.2 [0.7,1.6]
Pittsburgh
<	75 (w/ diabetes): 1.8 [0.9,2.7]
75+ (w/ diabetes): 0.9 [-0.2,2.0]
<	75 (w/o diabetes): 0.6 [0.1,1.2]
75+ (w/o diabetes): 1.6 [1.2,2.1]
Seattle
<	75 (w/ diabetes): 1.9 [0.1,3.7]
75+ (w/ diabetes): 2.7 [0.7,4.8]
<	75 (w/o diabetes): 0.8 [0.0,1.6]
75+ (w/o diabetes): 0.9 [0.2,1.6]
Notes: Overall percent increases were
also provided for each city, yielding
similar results.
July 2009
E-107
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Zanobetti and Schwartz
(2005, 088069)
Period of Study: 1985-1999
Location: 21 U.S. cities (Birmingham,
Alabama
Boulder, Colorado
Canton, Ohio
Chicago, Illinois
Cincinnati, Ohio
Cleveland, Ohio
Colorado Springs, Colorado
Detroit, Michigan
Honolulu, Hawaii
Houston, Texas
Minneapolis-St. Paul, Minnesota
Nashville, Tennessee
New Haven, Connecticut
Pittsburgh, Pennsylvania
Provo-Orem, Utah
Salt Lake City, Utah
Seattle, Washington
Steubenville, Ohio
Youngstown, Ohio)
Outcome (ICD-9): Myocardial infarction
(410)
Age Groups: >65 yrs
Study Design: Case-crossover
N: 302,453 admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature
Season: NR
Dose-response Investigated: Yes
Statistical Package: SAS (PROC
PHREG)
Lags Considered: 0-2 days
Pollutant: PMio
Averaging Time: 24 h
Median: Ranged from 15.5-34.1 Avg
across all cities - 27
PM Increment: 10 /yglm3
Percent Increase [CI]: Ml only: 0.65
[0.3,1]
Previous C0PD admission: 1.3 [-0.1,2.8]
Monitoring Stations: 1 + (data obtained Secondary pneumonia diagnosis: 1.4
from USEPA's Aerometric Information 0.8,3.6]
Retrieval System)
Copollutant: NR
Notes: Figure 1 presents percent change
in Ml per lag day, showing same-day lag
to be significant. Figure 2 shows percent
change with/without other co-morbidities.
'All units expressed in/yglm3 unless otherwise specified.
July 2009
E-108
DRAFT-DO NOT CITE OR QUOTE

-------
Table E-6. Short-term exposure-cardiovascular-ED/HA - PM102.5.
Reference
Design & Methods
Concentrations
Effect Estimates (95% CI)
Reference: Halonen et al. (2009,
180379)
Period of Study: 1998-2004
Location: Helsinki, Finland
Outcome: Cardiovascular
Hospitalizations & Mortality (ICD 10:100-
99)
Age Groups: 65+ yrs
Study Design: time series
N: NR
Statistical Analyses: Poisson, GAM
Covariates: temperature, humidity,
influenza epidemics, high pollen episodes,
holidays
Dose-response Investigated? No
Statistical Package: R
Lags Considered: lags 0-3 & 5d (0-4)
mean
Pollutant: PMio-2.b
Averaging Time: daily
Mean (SD): NR
Min: 0.0
ZS"1 percentile: 4.9
50"1 percentile: 7.5
TB" percentile: 12.1
Max: 101.4
Monitoring Stations: NR
Copollutant: PM <0.03, PMo.03-0.1, PMco.i
PM<0.10.29, PM2.6, CO, NO2
Co-pollutant Correlation
PM<0.03: 0.14
PM0.03-0.1: 0.28
PM <0.1: 0.24
PM<0.10.29: 0.20
PM2.6: 0.25
PM Increment: Interquartile Range
Percent Change (Lower CI, Upper CI):
All Cardiovascular Morality
Lag 0: -0.01 (-1.52,1.53)
Lag 1: -0.26 (-1.69,1.18)
Lag 2: -0.61 (-2.03, 0.83)
Lag 3: -0.57 (-1.98, 0.85)
5-d mean: -0.70 (-2.56,1.20)
Coronary Heart Disease HA
Lag 0:1.12 (-0.28, 2.55)
Lag 1: -0.38 (-1.68, 0.94)
Lag 2: 0.01 (-1.33,1.37)
Lag 3: -0.53 (-1.82, 0.78)
5-d mean: 0.23 (-0.29, 0.75)
Stroke HA
Lag 0:-1.33 (-3.26, 0.63)
Lag 1: -1.90 (-3.82, 0.07)*
Lag 2: -1.09 (-3.04,0.89)
Lag 3: -0.51 (-2.40,1.43)
5-d mean: -2.21 (-4.75,0.39)
Arrhythmia HA
Lag 0:0.57 (-1.33, 2.49)
Lag 1: -0.65 (-2.55,1.29)
Lag 2: 0.02 (-1.93, 2.00)
Lag 3:-1.34 (-3.26, 0.62)
5-d mean: -1.11 (-3.68,1.53)
*p<0.05, *p<0.10
July 2009
E-109
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Host et al. (2008,155852)
Period of Study: 2000 - 2003
Location: Six French cities: Le Havre,
Lille, Marseille, Paris, Rouen, and
Toulouse
Outcome (ICD-10): Daily hospitalizations
for all cardiovascular (I00-I99), cardiac
(I00-I52), and ischemic heart diseases
(I20-I25).
Age Groups: For cardiovascular
diseases: All ages, and restricted to a 65
years
Study Design: Time series
N: NR (Total population of cities:
approximately 10 million)
Statistical Analyses: Poisson regression
Covariates: Seasons, days of the week,
holidays, influenza epidemics, pollen
counts, temperature, and temporal trends
Dose-response Investigated: No
Statistical Package: MGCV package in
R software (R 2.1.1)
Lags Considered: Avg of 0-1 days
Pollutant: PMio2.5
Averaging Time: 24 h
Mean //g|m3 (5th -95th percentile): Le
Havre: 7.3 (2.5-14.0)
Lille: 7.9(2.2-13.7)
Marseille: 11.0(4.5-21.0)
Paris: 8.3(3.2-15.9)
Rouen: 7.0 (3.0-12.5)
Toulouse: 7.7 (3.0-15.0)
Monitoring Stations: 13 total: 1 in
Toulouse
4 in Paris
2 each in other cities
Copollutant (correlation): PM2 e:
Overall: r > 0.6
Ranged between r - 0.28 and r - 0.73
across the six cities.
PM Increment: 10 //g|m3, and an
18.8 //g/m3 increase (corresponding to an
increase in pollutant levels between the
lowest of the 5th percentiles and the
highest of the 95th percentiles of the
cities' distributions)
ERR (excess relative risk) Estimate [CI]:
For all cardiovascular diseases (10 //gf'nr1
increase): All ages: 0.5% [-1.2, 2.3]
> 65 years: 1.0% [-1.0,3.0]
For all cardiovascular diseases (18 //gfnv1
increase): All ages: 1.0% [-2.3,4.3]
a 65 years: 1.9% [-2.0, 5.9]
For cardiac diseases (10 //gfnv1 increase):
All ages: 0.1% [-1.9, 2.1]
a 65 years: 1.6% [-0.8,4.1]
For cardiac diseases (18.8 //g|m3
increase): All ages: 0.1% [-3.6,4.0]
>65 years: 3.1% [-1.5, 7.9]
For ischemic heart diseases (10 //g|m3
increase): All ages: 2.8% [-0.8, 6.6]
>65 years: 6.4% [1.6,11.4]
For ischemic heart diseases (18 /yglm3
increase): All ages: 5.4% [-1.5,12.8]
>65 years: 12.4 [3.1,22.6]
Reference: Metzgeret al. (2004,
044222)
Period of Study: August 1998-August
2000
Location: Atlanta Metropolitan area
(Georgia)
Outcome (ICD-9): Emergency visits for
ischemic heart disease (410-414),
cardiac dysrhythmias (427), cardiac
arrest (427.5), congestive heart failure
(428), peripheral vascular and
cerebrovascular disease (433-437, 440,
443-444,451-453), atherosclerosis
(440), and stroke (436).
Age Groups: All
Study Design: Time series
N: 4,407,535 emergency department
visits between 1993-2000 (data not
reported for 1998 ¦ 2000)
Statistical Analyses: Poisson
generalized linear modeling
Covariates: Day of the week, hospital
entry and exit indicator variables,
federally observed holidays, temporal
trends, temperature, dew point
temperature
Season: All
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: 3-day moving avg,
lags 0 -7
Pollutant: PM10 2.5
Averaging Time: 24 h
Median/yg/m3 (10% ¦ 90% range): 9.1
(4.4,16.2)
Monitoring Stations: 1
Copollutant (correlation): PM10:
r - 0.59
O3: r - 0.35
NO?: r - 0.46
CO: r - 0.32
SO2: r - 0.21
PM2.6: r - 0.43
UFP: r - 0.13
PM2.6 water
soluble metals: r - 0.47
PM2.6 sulfates: r - 0.26
PM2.6 acidity: r - 0.23
PM2.6 organic carbon: r - 0.51
PM2.6 elemental carbon: r - 0.48
PM2.6 oxygenated hydrocarbon: r - 0.31
Other variables: Temperature: r - 0.20
Dew point: r - 0.00
PM Increment: 5 //gfnv1 (approximately
1 SD)
RR [95% CI]: For 3 day moving avg: All
CVD: 1.012(0.985,1.040]
Dysrhythmia: 1.021 [0.974,1.070]
Congestive heart failure: 1.020 [0.964—
1.079]
Ischemic heart disease: 0.994 [0.946—
1.045]
Peripheral vascular and cerebrovascular
disease: 1.022 [0.972-1.074]]
Results for Lags 0-7 expressed in figures
(see notes).
Notes: Figure 1: RR (95% CI) for single-
day lag models for the association of ER
visits for CVD with daily ambient PMio-
2.5.
Summary of Figure 1 results: Positive
association at Lag 0.
July 2009
E-110
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Peng et al. (2008,156850)
Period of Study: January 1,1999-
December 31, 2005
Location: 108 U.S. counties in the
following states: Alabama, Arizona,
California, Colorado, Connecticut, District
of Columbia, Florida, Georgia, Idaho,
Illinois, Indiana, Kentucky, Louisiana,
Maine, Maryland, Massachusetts,
Michigan, Minnesota, Missouri, Nevada,
New Hampshire, New Jersey, New
Mexico, New York, North Carolina, Ohio,
Oklahoma, Oregon, Pennsylvania, Rhode
Island, South Carolina, Tennessee, Texas,
Utah, Virginia, Washington, West Virginia,
Wisconsin
Outcome (ICD-9): Emergency
hospitalizations for: Cardiovascular
disease, including heart failure (428),
heart rhythm disturbances (426-427),
cerebrovascular events (430-438),
ischemic heart disease (410-414,429),
and peripheral vascular disease (440-
448).
Age Groups: 65 + years, 65-74, 75 +
Study Design: Time series
N: approximately 12 million Medicare
enrollees (3.7 million CVD and 1.4 million
RD admissions)
Statistical Analyses: Two-stage
Bayesian hierarchical models:
Overdispersed Poisson models for county-
specific data. Bayesian hierarchical
models to obtain national avg estimate
Covariates: Day of the week, age-
specific intercept, temperature, dew point
temperature, calendar time, indicator for
age of 75 years or older. Some models
were adjusted for PM2.6.
Dose-response Investigated: No
Statistical Package: R version 2.6.2
Lags Considered: 0-2 days
Pollutant: PM10 2 E
Averaging Time: 24 h
Mean //gfm3 (IQR): All counties assessed:
9.8(6.9-15.0)
Counties in Eastern US: 9.1 (6.6-13.1)
Counties in Western US: 15.4 (10.3-
21.8)
Monitoring Stations: At least 1 pair of
co-located monitors (physically located in
the same place) for PM10 and PM2.6 per
county
Copollutant (correlation): PM2 e:
r - 0.12
PM10: r - 0.75
Other variables: Median within-county
correlations between monitors: r - 0.60
PM Increment: 10 //g|m3
Percentage change [95% CI]: CVD: Lag 0
(unadjusted for PM2.6): 0.36 [0.05, 0.68]
Lag 0 (adjusted for PM2.6): 0.25 [-0.11,
0.60]
Notes: Effect estimates for PMio-2.5 (0-
2 day lags) are showing in Figures 2-5.
Figure 2: Percentage change in
emergency hospital admissions for CVD
per 10 /yglm3 increase in PM (single
pollutant model and model adjusted for
PM2.6 concentration)
Figure 4: Percentage change in
emergency hospital admissions rate for
CVD and RD per a 10 //g|m3 increase in
PMio-2.5 (0-2 day lags, Eastern vs.
Western USA)
Figure 5: County-specific log relative risks
of emergency hospital admissions for
CVD per 10/yglm3 increase in PMio-2.5 at
Lag 0 (unadjusted for PM2.6 and plotted
vs percentage of urbanicity)
No significant associations between
PMio-2.5 and cause-specific
cardiovascular disease.
Reference: Tolbert et al. (2007,
090316)
Period of Study: August 1998-
December 2004
Location: Atlanta Metropolitan area,
Georgia
Outcome (ICD-9): Combined CVD group,
including: Ischemic heart disease (410-
414), cardiac dysrhythmias (427),
congestive heart failure (428), and
peripheral vascular and cardiovascular
disease (433-437,440, 443-445, and
451-453)
Age Groups: All
Study Design: Time series
N: NR for 1998-2004. For 1993-2004:
10,234,490 ER visits (283,360 visits).
Statistical Analyses: Poisson
generalized linear models
Covariates: Long-term temporal trends,
temperature, dew point, days of week,
federal holidays, hospital entry and exit
Season: All
Dose-response Investigated: No
Statistical Package: SAS version 9.1
Lags Considered: 3-day moving avg (lag
0-2)
Pollutant: PMio-2.5
Averaging Time: 24 h
Mean (/L/g|m3) (median
IQR, range, 10th—90th percentiles): 9.0
(8.2
5.6-11.5
0.5-50.3
3.6-15.1)
Monitoring Stations: 1
Copollutant (correlation): PM10:
r - 0.67
O3: r = 0.36
NO2: r - 0.48
CO: r - 0.38S02: r - 0.16
PM2.6: r - 0.47
PM2.6 SO4: r - 0.32
PM2.6 EC: r - 0.49
PM2.6 OC: r - 0.49
PM2.6 TC: r - 0.51
PM2.6 water-sol metals: r - 0.50
OHC: r - 0.41
PM Increment: 5.89 /yglm3 (IQR)
Risk ratio [95% CI]: CVD: 1.004 (0.990-
1.019)
'All units expressed in/yglm3 unless otherwise specified.
July 2009
E-111
DRAFT-DO NOT CITE OR QUOTE

-------
Table E-7. Short-term exposure - cardiovascular-ED/HA - PM2.5 (including PM components/sources)
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Andersen et al. (2008,
189651)
Period of Study: May 2001 - December
2004
Location: Los Angeles and San
counties, California
Outcome (ICD-10): CVD, including angina
pectoris (I20), myocardial infarction (121-
22), other acute ischemic heart diseases
(124),	chronic ischaemic heart disease
(125),	pulmonary embolism (I26), cardiac
arrest (I46), cardiac arrhythmias (I48-
48), and heart failure (I50). RD, including
chronic bronchitis (J41-42), emphysema
(J43), other chronic obstructive pul-
monary disease (J44), asthma (J45), and
status asthmaticus (J46). Pediatric
hospital admissions for asthma (J45) and
status asthmaticus (J46).
Age Groups: > 65 yrs (CVD and RD),
5-18 years (asthma)
Study Design: Time series
N: NR
Statistical Analyses: Poisson GAM
Covariates: Temperature, dew-point
temperature, long-term trend,
seasonality, influenza, day of the week,
public holidays, school holidays (only for
5-18 year olds), pollen (only for pediatric
asthma outcome)
Season: NR
Dose-response Investigated: No
Statistical Package: R statistical
software (gam procedure, mgcv package)
Lags Considered: Lag 0-5 days, 4-day
pollutant avg (lag 0-3) for CVD, 5-day avg
(lag 0-4) for RD, and a 6-day avg (lag 0-5)
for asthma.
Pollutant: PM2.6
Averaging Time: 24 h
Mean |jg/m3 (SD
median
IQR
99th percentile): 10 (5
9
7-12
28)
Monitoring Stations: 1
Copollutant (correlation):
NCtot: r - 0.40
NC100: r - 0.29
NCa12:r - 0.07
Nca23: r - -0.25
NCa57:r - 0.51
NCa2i2: r = 0.82
PMio: r - 0.80
CO: r - 0.46
NO?: r - 0.42
NO,: r - 0.40
NOxCurbside: r - 0.28
0a: r - -0.20
Other variables:
Temperature: r - -0.01
Relative humidity: r - 0.21
PM Increment: 5 [Jg/nr1 (IQR)
Relative risk (RR) Estimate [CI]: CVD
hospital admissions (4 day avg, lag 0 -3),
age 65+: One-pollutant model: 1.03
[1.01-1.06]
Adj for NCtot: 1.03(1.01-1.06]
RD hospital admissions (5 day avg, lag 0 -
4), age 65+:
One-pollutant model: 1.00 [0.95-1.00]
Adj for NCtot: 1.00 [0.95-1.06]
Asthma hospital admissions (6 day avg
lag 0-5), age 5-18:
One-pollutant model: 1.15 [1.00-1.32]
Adj for NCtot: 1.13(0.98-1.32]
Estimates for individual day lags reported
only in figure form (see notes):
Notes: Figure 2: Relative risks and 95%
confidence intervals per IQR in single day
concentration (0-5 day lag). Summary:
CVD: Marginally significant association at
Lag 0. RD: No statistically or marginally
significant associations. Positive
associations at Lag 4-5.Asthma: Wide
confidence intervals make interpretation
difficult. Positive associations at Lag 1,
2, 3.
July 2009
E-112
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ballester et al. (2006,
088746)
Period of Study: 1995 ¦ 1999
Location: 6 Spanish cities: Barcelona,
Bilbao, Pamplona, Valencia, Vigo,
Zaragoza
Outcome (ICD-9): The number of daily
emergency admissions with primary diag-
nosis for all cardiovascular disease (390-
459) and heart diseases (410-414, 427,
428)
Age Groups: All ages
Study Design: Time series
N: NR
Statistical Analyses: Poisson GAMs
Covariates: daily temperature,
barometric pressure, and relative
humidity
daily influenza incidence, day of the
week, holidays, unusual events (ex.
medical strikes), seasonal variation, trend
of the series
Season: NR
Dose-response Investigated: No
Statistical Package: S-Plus GAM
function
Lags Considered: 0-3 days, 0- to 1 -day
avg
Pollutant: Black smoke (BS)
Averaging Time: 24 h
Mean //g/m3 (10-90th percentile): Overall
mean NR.
City specific means
Barcelona: 35.0 (19.4, 53.0)
Bilbao: 18.5(8.8,31.0)
Pamplona: 7.4 (2.3,13.0)
Valencia: 40.3 (20.3, 66.4)
Vigo: 79.4(43.9,122.3)
Zaragoza: 40.4 (23.8, 61.3)
Monitoring Stations: NR (at least three
stations per city)
Copollutant (correlation): Summary of
the correlation coefficients between each
pair of pollutants within cities:
PMio: r - 0.48
TSP: from r - 0.16 to r - 0.69
(median r - 0.43)
NO2: from r - 0.23 to r - 0.69
(median r - 0.48)
SO2: from r - 0.09 to r - 0.59
(median r - 0.24)
PM Increment: 10/yglm3
Relative risk [CI]: Relative risks are
expressed only in the form of figures (see
notes).
Percentage change in risk [CI]: All
cardiovascular diseases (avg of lags 0 -1)
0.24% [-0.18, 0.67]
Heart disease (avg of lags 0 -1) 0.71%
[0.13,1.29]
Notes: Relative risks for the single
pollutant models are expressed in Figure
2. Figure 2: Time sequence of the
combined association between BS and
hospital admissions for all CVD (A) and
heart disease (B). Summary: Significant,
positive association of TSP with both
overall CVD and heart disease
hospitalizations at Lag 0.
Relative risks for two pollutant models
are expressed in Figure 3: Combined
estimates of the association between
hospital admissions for heart diseases
and air pollutants (avg of lags 0-1
adjusted for CO, NO2, O3, or SO2).
Summary: Significant, positive
association remains after adjusting for
NO2, O3, and SO2. Association remains
positive but becomes marginally
significant after adjusting for CO.
CO: from r - 0.62 to r - 0.69
(median r - 0.69)
O3: from r - -0.43 to r - -0.06
(median r - -0.16)
July 2009
E-113
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ballester et al. (2006,
088746)
Period of Study: 1993 ¦ 1999
Location: 7 Spanish cities: Barcelona,
Bilbao, Cartagena, Castellon, Gijon,
Oviedo, Valencia
Outcome (ICD-9): The number of daily
emergency admissions with primary diag-
nosis for all cardiovascular disease (390-
459) and heart diseases (410-414, 427,
428)
Age Groups: All ages
Study Design: Time series
N: NR
Statistical Analyses: Poisson GAMs
Covariates: daily temperature,
barometric pressure, and relative
humidity
daily influenza incidence, day of the
week, holidays, unusual events (ex.
medical strikes), seasonal variation, trend
of the series
Season: NR
Dose-response Investigated: No
Statistical Package: S-Plus GAM
function
Lags Considered: 0-3 days, 0- to 1 -day
avg
Pollutant: TSP
Averaging Time: 24 h
Mean /yglm3 (10-90th percentile): overall
mean NR.
City specific means
Barcelona: 51.8 (29.4, 78.8)
Bilbao: 58.3 (30.3, 92.3)
Cartagena: 54.9 (32.5, 79.9)
Castellon: 60.4 (32.0, 92.1)
Gijon: 77.4(47.4,118.3)
Oviedo: 76.0 (48.3,111.8)
Valencia: 61.0(44.1,80.7)
Monitoring Stations: NR (at least three
stations per city)
Copollutant (correlation): Summary of
the correlation coefficients between each
pair of pollutants within cities: BS: from r
- 0.16 to r - 0.69
(median r - 0.43)
PMio: NA
NO2: from r - -0.13 to r - 0.65
(median r - 0.48)
SO2: from r - 0.06 to r - 0.69
(median r - 0.31)
CO: from r - 0.06 to r - 0.59
(median r - 0.47)
O3: from r - -0.27 to r - 0.07
(median r - -0.03)
PM Increment: 10/yglm3
Relative risk [CI]: Relative risks are
expressed only in the form of figures (see
notes).
Percentage change in risk [CI]: All
cardiovascular diseases: 0.07% [-0.23,
0.36]
Heart disease 0.45% [0.04, 0.86]
Notes: Relative risks for the single
pollutant models are expressed in Figure
2.
Figure 2: Time sequence of the combined
association between TSP and hospital
admissions for all CVD (A) and heart
disease (B).
Summary of results: Positive, marginally
significant association of TSP with
overall CVD at Lag 0. Positive,
statistically significant relation between
TSP and heart disease hospitalizations at
Lag 0.
Relative risks for two pollutant models
are expressed in Figure 3:
Figure 3: Combined estimates of the
association between hospital admissions
for heart diseases and air pollutants (avg
of lags 0-1
adjusted for CO, NO2, O3, or SO2).
Summary of results: Small positive
significant or marginally significant
associations between TSP and general
CVD and heart disease hospitalizations
remain constant after adjustment for CO,
NO2, O3, or SO2.
July 2009
E-114
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Bell et al. (2008, 091268)
Period of Study: 1995 ¦ 2002
Location: Taipei, Taiwan
Outcome (ICD-9): Hospital admissions
for ischemic heart disease (410, 411,
414), cerebrovascular disease (430—
437), asthma (493), and pneumonia
(486).
Age Groups: All
Study Design: Time series
N: 6,909 hospital admissions for
ischaemic heart diseases, 11,466 for
cerebrovascular disease, 19,966 for
pneumonia, and 10,231 for asthma
Statistical Analyses: Poisson regression
Covariates: Day of the week, time,
apparent temperature, long-term trends,
seasonality
Season: All
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: lags 0-3 days, mean of
lags 0-3
Pollutant: PM2.6
Averaging Time: 24 h
Mean //g/m3 (range
IQR): 31.6(0.50-355.0
20.2)
Monitoring Stations: 2
Copollutant (correlation): NR
PM Increment: 20/yg/m3 (near IQR)
Percentage increase estimate [95% CI]:
Ischemic heart disease: L0: 3.48 (-0.39,
7.51)
L1: 3.55 (-0.30, 7.56)
L2: 3.32 (-0.50, 7.29)
L3: 2.80 (-1.04, 6.79)
L03: 8.38 (2.28,14.84)
Cerebrovascular disease: L0: -2.22 (¦
50.2, 0.67)
L1:-1.30 (-4.08,1.55)
L2: 0.24 (-2.49, 3.040
L3: 1.21 (-1.41,3.90)
L03:-1.45 (-5.58, 2.87)
Asthma: L0: 0.46 (-2.41, 3.42)
L1:-1.36 (-4.33,1.71)
L2: -0.83 (-3.67, 2.10)
L3: -0.78 (-3.63, 2.16)
L03: -1.75 (-6.21, 2.92)
Pneumonia: L0: 0.06 (-2.74, 2.94)
L1: 0.34(-2.446, 3.20)
L2: -0.59 (-3.38, 2.29)
L3:-0.44 (-3.22, 2.41)
L03: -0.61 (-4.87, 3.85)
Reference: Bell et al. (2008, 091268)
Period of Study: 1999 ¦ 2005
Location: 202 US counties
Outcome (ICD-9): Heart failure (428),
heart rhythm disturbances (426-427),
cerebrovascular events (430-438),
ischemic heart disease (410-414,429),
peripheral vascular disease (440-449),
C0PD (490-492), respiratory tract
infections (464 ¦ 466, 480 ¦ 487)
Age Groups: 65 +
Study Design: Time series
N: NR
Statistical Analyses: Two-stage
Bayesian hierarchical model to find
national avg
First stage: Poisson regression (county-
specific)
Covariates: day of the week,
temperature, dew point temperature,
temporal trends, indicator for persons
75+ years, population size
Season: All, June-August (Summer),
September-November (Fall), December-
February (Winter), March-May (Spring)
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: 0-2 day lags
Pollutant: PM2.6
Averaging Time: 24 h
Mean (/L/g|m3): Descriptive information
presented in Figure S2 (boxplots):
IQR: 8.7 /yglm3
Monitoring Stations: NR
Copollutant (correlation): NR
PM Increment: 10 //g|m3
Percent increase [95% PI]:
Cardiovascular admissions:
Lag 0 (all seasons): 0.80 [0.59-1.01]
Lag 0 (winter, national): 1.49 [1.09-
1.89]
Lag 0 (winter, northeast): 2.01 [1.39-
2.63]
Lag 0 (winter, southeast): 1.06 [-0.07—
2.21]
Lag 0 (winter, northwest): 0.85 [-4.11—
6.07]
Lag 0 (winter, southwest): 0.76 [-0.25—
1.79]
Lag 0 (spring, national): 0.91 [0.47-1.35]
Lag 0 (spring, northeast)
0.95(0.32-1.58]
Lag 0 (spring, southeast): 0.75 [-0.26—
1.78]
Lag 0 (spring, northwest): -0.07 [-12.40—
13.98]
Lag 0 (spring, southwest): 1.78 [-0.87—
4.51]
Lag 0 (summer, national): 0.18 [-0.23—
0.58]
_Lag_OJsummerjiortheastHL^
July 2009
E-115
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
1.02]
Lag 0 (summer, southeast): -0.67 [-1.60—
0.26]
Lag 0 (summer, northwest): -1.55 [¦
15.22-14.31]
Lag 0 (summer, southwest): -1.20 [-4.90—
2.65]
Lag 0 (autumn, national): 0.68 [0.29—
1.07]
Lag 0 (autumn, northeast): 1.03 [0.48—
1.58]
Lag 0 (autumn, southeast): 0.17 [-0.72—
1.07]
Lag 0 (autumn, northwest): -0.67 [-6.96—
6.05]
Lag 0 (autumn, southwest): 0.30 [-0.98—
1.59]
Lag 1 (all seasons): 0.07 [-0.12-0.26]
Lag 1 (winter): 0.56 [0.16-0.96]
Lag 1 (spring): -0.10 [-0.58-0.39]
Lag 1 (summer): -0.16 [-0.54—0.22]
Lag 1 (autumn): 0.04 [-0.28-0.35]
Lag2 (all seasons): [0.06 [-0.12—0.23]
Lag 2 (winter): 0.27 [-0.12-0.65]
Lag 2 (spring): 0.19 [-0.23—0.60]
Lag 2 (summer): -0.12 [-0.50—0.26]
Lag 2 (autumn): 0.02 [-0.30-0.34]
Respiratory admissions: Lag 0 (all
seasons): 0.22 [-0.12-0.56]
Lag 0 (winter, national): 1.05 [0.29—
1.82]
Lag 0 (winter, northeast): 1.76 [0.60-
2.93]
Lag 0 (winter, southeast): 0.59 [-1.35—
2.58]
Lag 0 (winter, northwest): -0.07 [-6.74—
7.08]
Lag 0 (winter, southwest): 0.03 [-1.25—
1.34]
Lag 0 (spring, national): 0.31 [-0.47—
1.11]
Lag 0 (spring, northeast): 0.34 [-0.66—
1.34]
Lag 0 (spring, southeast): -0.06 [-1.77-
1.68]
Lag 0 (spring, northwest): -8.52 [-25.62—
12.51]
Lag 0 (spring, southwest): 1.87 [-2.00—
5.90]
Lag 0 (summer, national): -0.62 [-1.33—
0.09]
Lag 0 (summer, northeast): -0.8 [-1.65-
0.07]
Lag 0 (summer, southeast): -0.15 [-1.88—
1.61]
July 2009
E-116
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Lag 0 (summer, northwest): 0.25 [¦
21.46-27.96]
Lag 0 (summer, southwest): 0.64 [-5.38-
7.04]
Lag 0 (autumn, national): 0.02 [-0.63—
0.67]
Lag 0 (autumn, northeast): -0.01 [-0.87—
0.85]
	Continued on next page	
	Continued from previous page	
Lag 0 (autumn, southeast): -0.58 [-2.06—
0.91]
Lag 0 (autumn, northwest): -1.38 [¦
11.84-10.32]
Lag 0 (autumn, southwest): 1.77 [-0.73—
4.33]
Lag 1 (all seasons): 0.05 [-0.29-0.39]
Lag 1 (winter): 0.50 [-0.27-1.27]
Lag 1 (spring): -0.24 [-1.01 -0.53]
Lag 1 (summer): 0.28 [-0.39-0.95]
Lag 1 (autumn): 0.15 [-0.49-0.79
Lag 2 (all seasons): 0.41 [0.09-0.74]
Lag 2 (winter, national): 0.72 [0.01 —
1.43]
Lag 2 (winter, northeast): 0.79 [-0.21 —
1.80]
Lag 2 (winter, southeast): 0.4 [¦ 1.45,
2.27]
Lag 2 (winter, northwest): -0.06 [-6.52-
6.85]
Lag 2 (winter, southwest): 1.2 [-0.10-
2.52]
Lag 2 (spring, national): 0.35 [-0.29—
0.99]
Lag 2 (spring, northeast): 0.04 [-0.88—
0.97]
Lag 2 (spring, southeast): 0.75 [-0.82—
2.34]
Lag 2 (spring, northwest): 2.29 [-14.26—
22.03]
Lag 2 (spring, southwest): 1.05 [-2.18-
4.39]
Lag 2 (summer, national): 0.57 [-0.07—
1.23]
Lag 2 (summer, northeast): 0.77 [-0.01 —
1.56]
| Lag 2 (summer, southeast): -0.52 [¦
2.07-1.06]
Lag 2 (summer, northwest): 0.74 [¦
18.73-24.86]
Lag 2 (summer, southwest): 2.41 [-2.61 —
7.69]
Lag 2 (autumn, national): 0.39 [-0.22—
1.01]
iLa£^JautumMiorthe^
July 2009
E-117
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1.07]
Lag 2 (autumn, southeast): 0.14 [-1.29—
1.59]
Lag 2 (autumn, northwest): -0.74 [¦
10.08-9.58]
Lag 2 (autumn, southwest): 0.971-1.36-
3.36]
Reference: Bell et al. (2009,191007)
Period of Study: 1999-2005
Location: 168 US Counties
Outcome: CVD hospital admissions
Study Design: Retrospective Cohort
Covariates: socio-economic conditions,
long term temperature
Statistical Analysis: Bayesian
hierarchical model
Age Groups: £65
Pollutant: PM2.6
Averaging Time: 24h
Mean (SD) Unit: NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 20% of the population
acquiring air conditioning
Percent Change (95% CI) in
community-specific PM health effect
estimates for CVD hospital
admissions
Any AC, including window units
Yearly health effect: -4.3 (-72.7-4.2)
Summer health effect: -148 (-327-31.1)
Winter health effect:-80.0 (-182-22.0)
Central AC
Yearly health effect: -42.5(-63.4-21.6)
Summer health effect: -79.5 (-143 -15.7)
Winter health effect: -41.9 (-124-40.0)
July 2009
E-118
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Bell et al. (2009,191997)
Period of Study: 1999-2005
Location: US
Outcome: Cardiovascular HA
Age Groups: 65 +
Study Design: time series
N: NR
Statistical Analyses: Bayesian
Hierarchical Regression
Covariates: time trend, day of week,
seasonality, dew point, temperature
Statistical Package: NR
Lags Considered: 0-2
Pollutant: PM2.6
Averaging Time: daily
Mean:
EC: 0.715
Ni: 0.002
V: 0.003
Min:
EC: 0.309
Ni: 0.003
V: 0.001
Max:
EC: 1.73
Ni: 0.021
V: 0.010
Interquartile Range:
EC: 0.245
Ni: 0.001
V: 0.001
Interquartile Range of Percents:
EC: 1.7
Ni: 0.01
V: 0.01
Monitoring Stations: NR
Copollutant: Al, NH4+, As, Ca, CI, Cu,
EC, 0MC, Fe, Pb, Mg. Ni, NOs-, K, Si,
Na+, S04-, Ti, V, Zn
Co-pollutant Correlation
Ni, V: 0.48
V, EC: 0.33
Ni, EC: 0.30
Note: Pollutant concentrations available
for all fractions of PM2.6
PM Increment: Interquartile Range in
the fraction of PM2.6
Percent Increase in PM Health Effect
(Lower CI, Upper CI), lag
EC: 25.8(4.4,47.2), lag 0
EC + Ni: 14.0 (-7.6, 35.5), lag 0
EC + V: 14.9(-7.8, 37.6), lag 0
EC+ V, HS education: 15.0 (3.3, 26.8),
lag 0
EC+ V, median income: 15.8 (4.1, 27.5),
lag 0
EC+ V, racial composition: 14.2 (2.8,
25.6), lag 0
EC+ V, percent living in urban area: 14.7
(3.1,26.3), lagO
EC + V, population: 13.6 (2.2, 25.0), lag
0
EC + Ni, V: 11.9 (-10.4,43.2), lag 0
Ni: 19.0 (9.9, 28.2), lag 0
Ni + EC: 17.3(7.7, 26.9), lag 0
Ni + V: 15.5 (4.1, 26.9), lag 0
Ni + EC, V: 14.9(3.4, 26.4), lag 0
V: 27.5 (10.6, 44.4), lag 0
V	+ EC: 23.1 (4.9,41.4), lag 0
V+ Ni: 10.9 (-9.6, 31.5), lag 0
V	+ EC, Ni: 8.1 (-13.3,29.5), lag 0
EC: 11.8 (-69.2, 92.8), lag 1
EC: 21.0(-46.6, 88.6), lag 2
Ni: 20.6 (-15.5, 56.7), lag 1
Ni:-2.3 (-32.5, 27.9), lag 2
V: 34.0 (-31.2, 99.1), lag 1
V: 8.0 (-46.8, 62.7), lag 2
Percent HS education: -17.4 (-46.8,
11.9), lagO
Median income: 21.3 (-20.0, 62.5), lag 0
Percent black: 26.9 (-15.8, 69.6), lag 0
Percent living in urban area: 34.4 (-29.0,
97.8), lagO
Population: -4.3 (-13.3, 4.8), lag 0
Notes: Interquartile ranges in percent HS
education, median income, percent black,
percent living in urban area, and
population are 5.2 %, $9,223,17.3%,
11.0%, and 549,283 respectively.
July 2009
E-119
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Chan et al. (2007, 147787)
Period of Study: Apr 1997 - Dec 2002
Location: Boston, MA
Outcome: Cerebrovascular Emergency
Admissions
Age Groups: 50+ yrs
Study Design: time series
Statistical Analyses: GAM Poisosn
Regression
Covariates: year, month, day of week,
temperature, dew point
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-3d
Pollutant: PM2.6
Averaging Time: 24h
Mean (SD): 31.5 (16.0)
Min: 15.6
Max: 200.6
IQR: 19.7
Monitoring Stations: 16
Copollutant: 0:i, CO, SO2, NO2, PM10
Co-pollutant Correlation
0s: 0.33
CO: 0.44
SO2: 0.51
NO2: 0.50
PM10: 0.61
PM Increment: Interquartile Range (19.7
j"g/m3)
Percent Change (Lower CI, Upper CI),
p-value:
Cerebrovascular Disease
Lag 0:1.006 (0.993,1.019)
Lag 1:1.002 (0.990,1.014)
Lag 2:1.015 (0.978,1.052)
Lag 3:1.021 (1.005,1.037)
Lag 3 + 0s: 1.009 (0.987,1.031)
Lag 3 + CO: 1.014(0.993,1.035)
Lag 3 + 0s + CO: 1.009 (0.987,1.031)
Stroke
Lag 0:0.931 (0.831,1.031)
Lag 1:0.936 (0.845,1.027)
Lag 2: 0.931 (0.820,1.042)
Lag 3: 0.991 (0.969,1.013)
Ischaemic stroke
Lag 0:0.981 (0.907,1.055)
Lag 1:0.994 (0.920,1.078)
Lag 2: 0.960 (0.885,1.035)
Lag 3:1.059 (0.984,1.134)
Haemorrhagic stroke
Lag 0:0.870 (0.740,1.010)
Lag 1:0.882 (0.761,1.003)
Lag 2: 0.909 (0.810,1.008)
Lag 3: 0.921 (0.830,1.012)
Reference: Chan et al. (2008, 093297)
Period of Study: 1995 ¦ 2002
Location: Taipei Metropolitan area,
Taiwan
Outcome (ICD-9): Emergency visits for
ischaemic heart diseases (410-411,
414), cerebrovascular diseases (430-
437), and C0PD (493, 496)
Age Groups: All
Study Design: Time series
N: NR
Statistical Analyses: Poisson regression
Covariates: Year, month, day of week,
temperature, dewpoint temperature,
PM10, NO2
Season: All
Dose-response Investigated: No
Statistical Package: SAS version 8.0
Lags Considered: 0- to 7-day lags
Pollutant: PM2.6
Averaging Time: 24 h
Mean//g/m3 (SD): NR
Monitoring Stations: 1
Copollutant (correlation): NR
PM Increment: 19.7//g|m3 (IQR)
OR [95% CI]: In environmental conditions
without dust storms (results only given
for best-fitting model)
Lag 6 days: 1.024(1.004,1.044)
Reference: Delfino et al, (2008,156390)
Outcome: Cardiovascular hospital
Pollutant: PM2.6
Increment: 10 //g/m3

admissions

Period of Study: 10/1/2003-

Averaging Time: Hourly
Relative Rate (Min CI, Max CI)
1111512003
Study Design: Time series


Mean (SD) Unit by county:
All Cardiovascular
Location: Southern California
Statistical Analysis: Poisson regression



with GEE
Los Angeles
All Periods: 0.996 (0.989-1.003)

Age Groups: All
Before Fires: 27.2 (12.4) //g/m3
Pre-Wildfire: 0.992 (0.976-1.009)


During Fires: 54.1 (21.0) //g/m3
Wildfire: 1.008 (0.999-1.018), p - 0.104
July 2009
E-120
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
After Fires: 15.9 (5.5) //g|m3
Orange
Before Fires: 23.2 (9.6)//g/m3
During Fires: 64.3 (26.5) //g|m3
After Fires: 15.5 (10.2) //g|m3
Riverside
Before Fires: 32.7 (14.7) //g/m3
During Fires: 42.1 (25.5) //g/m3
After Fires: 16.9 (10.2) //g|m3
San Bernadino
Before Fires: 35.7 (16.6) //g/m3
During Fires: 45.3 (28.7) //g|m3
After Fires: 18.5 (8.3) //g|m3
San Diego
Before Fires: 18.5 (6.7) //g/m3
During Fires: 76.1 (66.6)//g/m3
After Fires: 14.2 (7.2) /yglm3
Ventura
Before Fires: 18.4 (8.3) /yglm3
During Fires: 50.1 (50.5) //g/m3
After Fires: 12.9 (4.3) /yglm3
Copollutant (correlation): NR
Post-Wildfire: 0.991 (0.964-1.019), p -
0.955
Ischaemic Heart Disease
All Periods: 0.991 (0.980-1.003)
Pre-Wildfire: 0.990 (0.963-1.017)
Wildfire: 0.117 (0.990-1.024), p - 0.313
Post-Wildfire: 0.989 (0.950-1.030), p -
0.976
Congestive Heart Failure
All Periods: 0.989 (0.974-1.004)
Pre-Wildfire: 0.978(0.942-1.015)
Wildfire: 1.016 (0.933-1.039), p - 0.096
Post-Wildfire: 0.969 (0.914-1.027), p -
0.791
Cardiac Dysrhythmia
All Periods: 0.980 (0.962-0.998)
Pre-Wildfire: 0.979(0.935-1.025)
Wildfire: 0.989 (0.961-1.017), p - 0.721
Post-Wildfire: 0.976 (0.912-1.044), p -
0.934
Cerebrovascular Disease and Stroke
All Periods: 1.019(1.004-1.035)
P re-Wildfire: 1.015(0.980-1.052)
Wildfire: 1.016 (0.997-1.036), p - 0.971
Post-Wildfire: 1.044 (0.987-1.104), p -
0.379
Relative Rate (Min CI, Max CI) in
relation to pre-wildfire period (1)
All Cardiovascular: Wildfire, unadjusted
for PM2.b: 0.958 (0.920-0.997)
Wildfire, adjusted for PM2.6: 0.947
(0.902-0.994)
Post-wildfire, unadjusted for PM2.6: 1.061
(1.006-1.119)
Post-wildfire, adjusted for PM2.6:1.053
(0.994-1.114)
Ischaemic Heart Disease: Wildfire,
unadjusted for PM2.6: 0.913 (0.852-
0.978)
Wildfire, adjusted for PM2.6: 0.905
(0.832-0.985)
Post-wildfire, unadjusted for PM2.6: 1.029
(0.943-1.123)
Post-wildfire, adjusted for PM2.6:1.029
(0.936-1.131)
Congestive Heart Failure: Wildfire,
unadjusted for PM2.6: 0.981 (0.817-
0.972)
Wildfire, adjusted for PM2.6: 0.911
(0.819-1.014)
Post-wildfire, unadjusted for PM2.6: 1.113
(0.997-1.242)
_^ostj/vMdfirei_adiusted^or_PM2;BM^1_05__
July 2009
E-121
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
(0.982-1.244)
Cardiac Dysrhythmia: Wildfire, unadjusted
for PMz.b: 0.968 (0.874-1.072)
Wildfire, adjusted for PM2.6: 0.964
(0.851-1.093)
Post-wildfire, unadjusted for PM2.6: 1.089
(0.949-1.251)
Post-wildfire, adjusted for PM2.6:1.057
(0.914-1.223)
Cerebrovascular Disease and Stroke:
Wildfire, unadjusted for PM2.6:1.066
(0.981-1.159)
Wildfire, adjusted for PM2.6:1.017
(0.922-1.123)
Post-wildfire, unadjusted for PM2.6: 1.013
(0.907-1.132)
Post-wildfire, adjusted for PM2.6:1.013
(0.902-1.138)
Reference: Dominici et al. (2006,
088398)
Period of Study: 1999 ¦ 2002
Location: 204 US counties, located in:
Alabama, Alaska, Arizona, Arkansas,
California, Colorado, Connecticut,
Delaware, District of Columbia, Florida,
Georgia, Hawaii, Idaho, Illinois, Indiana,
Iowa, Kansas, Kentucky, Louisiana,
Maine, Maryland, Massachusetts,
Michigan, Minnesota, Mississippi,
Missouri, Nevada, New Hampshire, New
Jersey, New Mexico, New York, North
Carolina, Ohio, Oklahoma, Oregon,
Pennsylvania, Rhode Island, South
Carolina, Tennessee, Texas, Utah,
Virginia, Washington, West Virginia,
Wisconsin
Outcome (ICD-9: Daily counts of hospital
admissions for primary diagnosis of heart
failure (428), heart rhythm disturbances
(426-427), cerebrovascular events (430-
438), ischemic heart disease (410-414,
429), peripheral vascular disease (440-
448), chronic obstructive pulmonary
disease (490-492), and respiratory tract
infections (464-466, 480-487).
Age Groups: >65 years
Study Design: Time series
N: 11.5 million Medicare enrollees
Statistical Analyses: Bayesian 2-stage
hierarchical models.
First stage: Poisson regression (county-
specific)
Second stage: Bayesian hierarchical
models, to produce a national avg
estimate
Covariates: Day of the week,
seasonality, temperature, dew point
temperature, long-term trends
Season: NR
Dose-response Investigated: No
Statistical Package: R statistical
software version 2.2.0
Lags Considered: 0-2 days, avg of days
0-2
Pollutant: PM2.6
Averaging Time: 24 h
Mean (/^g/m3) (IQR): 13.4 (11.3-15.2)
Monitoring Stations: NR
Copollutant (correlation): NR
Other variables: Median of pairwise
correlations among PM2.6 monitors within
the same county for 2000: r - 0.91 (IQR:
0.81-0.95)
PM Increment: 10 //g|m3 (Results in
figures
see notes)
Percent increase in risk [95% PI]:
Cerebrovascular disease (Lag 0): Age
65 + : 0.81 [0.30,1.32]
Age 65-74: 0.91 [0.01,1.82]
Age 75 + : 0.80 [0.21,1.38]
Peripheral vascular disease (Lag 0): Age
65 + : 0.86 [-0.06,1.79]
Age 65-74:1.21 [-0.26, 2.67]
Age 75+: 0.86 [-0.39, 2.11]
Ischemic heart disease (Lag 2): Age 65+:
0.44 [0.02,0.86]
Age 65-74: 0.37 [-0.22, 0.96]
Age 75+: 0.52 [-0.01,1.04]
Heart rhythm disturbances (Lag 0): Age
65 + : 0.57 [-0.01,1.15]
Age 65-74: 0.46 [-0.63,1.54]
Age 75 + : 0.72 [0.02,1.42]
Heart failure (Lag 0): Age 65 + : 1.28
[0.78,1.78]
Age 65-74: 1.21 [0.35, 2.07]
Age 75 + : 1.36 [0.78,1.94]
COPD (Lag 0): Age 65 +: 0.91 [0.91,
1.64]
Age 65-74:0.42 [-0.64,1.48]
Age 75 + : 1.47 [0.54, 2.40]
Respiratory tract infection: Age 65+:
0.92 [0.41,1.43]
Age 65-74: 0.93 [0.04,1.82]
Age 75 + : 0.92 [0.32,1.53]
Annual reduction in admissions
attributable to a 10 //g|m3 reduction in
daily PM2.6 level (95% PI):
Cerebrovascular disease: Annual number
of admissions: 226,641
July 2009
E-122
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Annual reduction in admissions: 1836
[680, 2992]
Peripheral vascular disease: Annual
number of admissions: 70,061
Annual reduction in admissions: 602 [-42,
1254]
Ischemic heart disease: Annual number of
admissions: 346,082
Annual reduction in admissions: 1523
[69, 2976]
Heart rhythm disturbances: Annual
number of admissions: 169,627
Annual reduction in admissions: 967 [-17,
1951]
Heart failure: Annual number of
admissions: 246,598
Annual reduction in admissions: 3156
[1923,4389]
COPD: Annual number of admissions:
108,812
Annual reduction in admissions: 990
[196,1785]
Respiratory tract infections: Annual
number of admissions: 226,620
Annual reduction in admissions: 2085
[929, 3241]
Notes: Figure 2: Point estimates and 95%
posterior intervals of the % change in
admissions rates per 10 //g/m3 (national
avg relative rates) for single lag (0,1, and
2 days) and distributed lag models for 0
to 2 days (total) for all outcomes.
Summary: Positive significant or
marginally significant associations
between PM2.6 and cerebrovascular
disease at Lag 0
peripheral vascular disease at Lags 0 and
2
ischemic heart disease at Lag 2
heart rhythm disturbances at Lag 0
heart failure at Lag 0, Lag 2, and Lags 0 -
2
COPD at Lag 0, Lag 1, and Lags 0-2
and respiratory tract infections at Lag 2
and Lags 0-2.
Figure 3: Point estimates and 95%
posterior intervals of the % change in
admission rates per 10 //gf'nv1 (regional
relative rates). Summary: For
cardiovascular diseases, all estimates in
the Midwestern, Northeastern, and
Southern regions were positive, while
estimates in the other regions (South,
West, Central, Northwest) were close to
0. For respiratory disease, there were
larger effects in the Central,
Southeastern, Southern, and Western
regions than in the other regions.
Figure 4: Point estimates and 95%
posterior intervals of the % change in
admission per 10 /yglm3 (Eastern vs.
Western regions): Summary: All estimates
for cardiovascular outcomes were
July 2009
E-123
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
positive in the US Eastern region but not
in the US Western region. The estimates
for respiratory tract infections were
larger in the Western region than in the
Eastern region. The estimates for CCPD
were positive in the both regions.
Reference: Halonen et al. (2009,
180379)
Period of Study: 1998-2004
Location: Helsinki, Finland
Outcome: Cardiovascular
Hospitalizations & Mortality (ICD 10:100-
99)
Age Groups: 65+ yrs
Study Design: time series
N: NR
Statistical Analyses: Poisson, GAM
Covariates: temperature, humidity,
influenza epidemics, high pollen episodes,
holidays
Dose-response Investigated? No
Statistical Package: R
Lags Considered: lags 0-3 & 5d (0-4)
mean
Pollutant: PM2.6
Averaging Time: daily
Mean (SD): NR
Min: 1.1
25"1 percentile: 5.5
SO"1 percentile: 9.5
75"1 percentile: 11.7
Max: 69.5
Monitoring Stations: NR
Copollutant: PM <0.03, PMo.03-0.1, PMco.i
PM<0.10.29, PM2.B-10, CO, NO2
Co-pollutant Correlation
PM<0.03: 0.14
PM0.03-0.1: 0.48
PM <0.1: 0.35
PM<0.10.29: 0.88
PM2.E-10: 0.25
PM Increment: Interquartile Range
Percent Change (Lower CI, Upper CI):
All Cardiovascular Morality
LagO: 0.73( 0.66, 2.13)
Lag 1: 0.74(-0.63, 2.13)
Lag 2: 0.74( 0.62, 2.11)
Lag 3: 0.06 (-1.29,1.43)
5-d mean: 0.87 (-0.94, 2.70)
Coronary Heart Disease HA
Lag 0:-0.17 (-1.5.0,1.18)
Lag 1: -0.03 (-1.31,1.26)
Lag 2: -0.63 (-1.87, 0.62)
Lag 3: 0.48 (-0.78,1.76)
5-d mean: 0.80 (-0.94, 2.58)
Stroke HA
Lag 0: -0.99 (-2.78, 0.84)
Lag 1:0.02 (-1.74,1.82)
Lag 2: -1.38 (-3.13, 0.40)
Lag 3: -0.17 (-1.92,1.61)
5-d mean: -0.78 (-3.10,1.60)
Arrhythmia HA
LagO: 0.82 (-1.03, 2.68)
Lag 1:0.18 (-1.58,1.97)
Lag 2: -0.09 (-1.82,1.67)
Lag 3: -0.48 (-2.22,1.29)
5-d mean: 0.16 (-2.16, 2.54)
*p<0.05, *p<0.10
July 2009
E-124
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Host et al„ 2008,155852)
Period of Study: 2000 - 2003
Location: Six French cities: Le Havre,
Lille, Marseille, Paris, Rouen, and
Toulouse
Outcome (ICD-10): Daily hospitalizations
for all cardiovascular (I00-I99), cardiac
(I00-I52), and ischemic heart diseases
(I20-I25), all respiratory diseases (J00-
J99), respiratory infections (J10-J22).
Age Groups: For cardiovascular
diseases: All ages, and restricted to a 65
years. For all respiratory diseases: 0-14
years, 15-64 years, and a 65 years. For
respiratory infections: All ages
Study Design: Time series
N: NR (Total population of cities:
approximately 10 million)
Statistical Analyses: Poisson regression
Covariates: Seasons, days of the week,
holidays, influenza epidemics, pollen
counts, temperature, and temporal trends
Season: NR
Dose-response Investigated: No
Statistical Package: MGCV package in
R software (R 2.1.1)
Lags Considered: Avg of 0-1 days
Pollutant: PM2.6
Averaging Time: 24 h
Mean (5th -95th percentile): Le Havre:
13.8(6.0-30.5)
Lille: 15.9(6.9-26.3)
Marseille: 18.8(8.0-33.0)
Paris: 14.7 (6.5-28.8)
Rouen: 14.4(7.5-28.0)
Toulouse: 13.8 (6.0-25.0)
Monitoring Stations: 13 total: 1 in
Toulouse
4 in Paris
2 each in other cities
Copollutant (correlation): PM102 e:
Overall: r > 0.6
Ranged between r - 0.28 and
r - 0.73 across the six cities.
PM Increment: 10//gf'nv1 increase, and a
27 /yglm3 increase (corresponding to the
difference between the lowest of the 5th
percentiles and the highest of the 95th
percentiles of the cities' distributions)
ERR (excess relative risk) Estimate [CI]:
For all cardiovascular diseases (10 //gf'nr1
increase): All ages: 0.9% [0.1,1.8]
a 65 years: 1.9% [0.9, 3.0]
For all cardiovascular diseases (27 //gfm3
increase): All ages: 2.5% [0.2, 4.9]
>65 years: 5.3% [2.6, 8.2]
For ischemic heart diseases (27 //g|m3
increase): All ages: 5.2% [-0.6,11.3]
>65 years: 12.7% [6.3,19.5]
For cardiac diseases (10 //gf'nv1 increase):
All ages: 0.9% [-0.1,2.0]
>65 years: 2.4% [1.2, 3.7]
For cardiac diseases (27 //g|m3 increase):
All ages: 2.5% [-0.3, 5.4]
a 65 years: 6.8% [3.3,10.3]
For ischemic heart diseases (10 /yglm3
increase): All ages: 1.9 % [-0.2, 4.0]
>65 years: 4.5% [2.3, 6.8]
For all respiratory diseases (10 /yglm3
increase): 0-14 years: 0.4% [-1.2, 2.0]
15-64 years: 0.8% [-0.7, 2.3];
a 65 years: 0.5% [-2.0, 3.0]
For all respiratory diseases (27 /yglm3
increase): 0-14 years: 1.1% [-3.1, 5.5]
15-64 years: 2.2% [-1.8, 6.4];
> 65 years: 1.3% [-5.3,8.2]
For respiratory infections (10 /yglm3
increase): All ages: 2.5% [0.1, 4.8]
For respiratory infections (27 /yglm3
increase): All ages: 7.0% [0.7,13.6]
July 2009
E-125
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Jalaludin et al. (2006,
189416)
Period of Study: 1 Jan, 1997-31 Dec,
2001
Location: Sydney, Australia
Outcome (ICD-9): Cardiovascular disease
(390-459), cardiac disease (390-429),
ischemic heart disease (410-413) and
cerebrovascular disease or stroke (430-
438)
Age Groups: 65+ yrs
Study Design: Time series
N: NR
Statistical Analyses: GAM, GLM
Covariates: Temperature, humidity
Season: Warm (Nov-Apr) and cool (May-
Oct)
Dose-response Investigated: No
Statistical Package: S-Plus
Lags Considered: 0-3 days
Pollutant: PM2.6
Averaging Time: 24 h
Mean (min-max): 9.5 (2.4-82.1)
SD - 5.1
Monitoring Stations: 14
Copollutant (correlation): Warm
BSP: r - 0.93
PMio: r - 0.89
0s: r - 0.57
NO?: r - 0.45
CO: r - 0.35
SO?: r - 0.27
Cool
BSP: r - 0.90
PM10: r - 0.88
O3: r - 0.05
NO?: r - 0.68
CO: r - 0.60
SO2: r - 0.46
Other variables: Warm
Temp: r - 0.24
Rel humidity: r - -0.15
Cool
Temp: r - -0.04
Rel humidity: r - 0.20
PM Increment: 4.8 /yglm3 (IQR)
Percent Change Estimate [CI]: All CVD
Same-day lag: 1.26 [0.56,1.96]
Avg 0-1 day lag: 0.85 [0.18,1.52]
Cool (same-day lag): 2.23 [0.98,3.50]
Warm (same-day lag): 0.73
[-0.05,1.52]
Cardiac disease
Same-day lag: 1.55 [0.74,2.38]
Avg 0-1 day lag: 1.33 [0.54,2.13]
Cool (same-day lag): 2.37 [0.87,3.89]
Warm (same-day lag): 1.13 [0.22,2.04]
Ischemic heart disease
Same-day lag: 1.17 [-0.08,2.44]
Avg 0-1 day lag: 1.24(0.04,2.45]
Cool (same-day lag): 0.57 [-1.74,2.94]
Warm (same-day lag): 1.31
[-0.04,2.68]
Stroke
Same-day lag: -0.89 [-2.41,0.65]
Avg 0-1 day lag:-1.08 [-2.54,0.41]
Cool (same-day lag): 1.45 [-1.17,4.15]
Warm (same-day lag): -2.19 [-4.00,-0.36]
Notes: All other lag-day ORs were
provided, yet none were significant.
Percent change in ED attendance was
also reported graphically
(Fig 1-5).
July 2009
E-126
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Jalaludin et al. (2006,
189416)
Period of Study: 1 Jan, 1997-31 Dec,
2001
Location: Sydney, Australia
Outcome (ICD-9): Cardiovascular disease
(390-459), cardiac disease (390-429),
ischemic heart disease (410-413) and
cerebrovascular disease or stroke (430-
438)
Age Groups: 65+ yrs
Study Design: Time series
N: NR
Statistical Analyses: GAM, GLM
Covariates: Temperature, humidity
Season: Warm (Nov-Apr) and cool (May-
Oct)
Dose-response Investigated: No
Statistical Package: S-Plus
Lags Considered: 0-3 days
Pollutant: Bs.p
Averaging Time: 24 h
Mean/ 104/m (min-max): 0.26 (0.04-3.37)
SD - 0.22
Monitoring Stations: 14
Copollutant (correlation): Warm
PM2.6: r - 0.93
PM10: r - 0.82
0s: r - 0.48
NO?: r - 0.35
CO: r - 0.33
SO2: r - 0.21;
Cool
PM2.6: r - 0.90
PM10: r - 0.75
O3: r - -0.08
NO2: r - 0.59
CO: r - 0.62
SO2: r - 0.48
Other variables: Warm
Temp: r - 0.23
Rel humidity: r - -0.04
Cool
Temp: r - -0.09
Rel humidity: r - 0.36
PM Increment: 0.181104/m (IQR)
Percent Change Estimate [CI]: All CVD
Same-day lag: 1.05 [0.44,1.66]
Avg 0-1 day lag: 0.79 [0.20,1.38];
Cool (same-day lag): 2.38 [1.15,3.62]
Warm (same-day lag): 0.45 [-0.18,1.09]
Cardiac disease
Same-day lag: 1.34 [0.63,2.05]
Avg 0-1 day lag: 1.13 [0.44,1.82];
Cool (same-day lag): 2.50(1.04,3.98]
Warm (same-day lag): 0.80 [0.07,1.54]
Ischemic heart disease
Same-day lag: 0.91 [-0.17,2.02]
Avg 0-1 day lag: 0.90 [-0.14,1.95];
Cool (same-day lag): 0.52 [-1.74,2.83]
Warm (same-day lag): 0.93 [-0.15,2.03]
Stroke
Same-day lag: -0.93 [-2.27,0.42]
Avg 0-1 day lag:-0.82 [-2.11,0.49];
Cool (same-day lag): 1.38 [-1.19,4.01 ];
Warm (same-day lag): -1.85 [-3.31,-0.36]
Notes: All other lag-day ORs were
provided, yet none were significant.
Percent change in ED attendance was
also reported graphically (Fig 1-5).
Reference: Lisabeth et al. (2008,
155939)
Period of Study: 2001 ¦ 2005
Location: Nueces County, Texas
Outcome: Ischemic stroke and transient
ischemic attacks (ICD codes not
reported).
Age Groups: 45+ years
Study Design: Time series
N: 3,508 stroke/TIAs (2,350 strokes, and
1,158 TIAs)
Statistical Analyses: Poisson regression
Covariates: Temperature, day of week,
temporal trends
Season: All, but looked at potential
effect modification by season (Summer:
June-September
Non-summer: October-May)
Dose-response Investigated: No
Statistical Package: S-plus 7.0
Lags Considered: Lags 0-5 days, and
averaged lag effect (0-5 days)
Pollutant: PM2.6
Averaging Time: 24 h
Median/yglm3 (IQR): 7.0(4.8-10.0)
Monitoring Stations: 6
Copollutant (correlation): NR
PM Increment: 5.1 /yglm3 (IQR)
RR Estimate [CI]: Lag 0:1.03 (0.99,
1.07)
Lag 1:1.03 (1.00-1.07)
All other lags and avg (lag 0-5) were not
statistically or marginally significant.
Adjusted for O3: Lag 0: 1.03 (0.99,1.07)
Lag 1:1.03 (0.99-1.06)
All other lags and avg (lag 0-5) were not
statistically or marginally significant.
Notes: Figure 3: % change in stroke/TIA
risk associated with an IQR increase in
PM2.5
July 2009
E-127
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Metzgeret al. (2004,
044222)
Period of Study: August 1998-August
2000
Location: Atlanta Metropolitan area
(Georgia)
Outcome (ICD-9): Emergency visits for
ischemic heart disease (410-414),
cardiac dysrhythmias (427), cardiac
arrest (427.5), congestive heart failure
(428), peripheral vascular and
cerebrovascular disease (433-437, 440,
443-444,451-453), atherosclerosis
(440), and stroke (436).
Age Groups: All
Study Design: Time series
N: 4,407,535 emergency department
visits for 1993-2000 (data not reported
for 1998-2000)
Statistical Analyses: Poisson
generalized linear modeling
Covariates: Day of the week, hospital
entry and exit indicator variables,
federally observed holidays, temporal
trends, temperature, dew point
temperature
Season: All
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: 3-day moving avg,
lags 0 -7
Pollutant: PM2.6
Averaging Time: 24 h
Median/yg/m3 (10%-90% range): PM2.6:
17.8(8.9, 32.3)
PM2.6 water soluble metals: 0.021
(0.006-0.061)
PM2.6 acidity: 4.5 (1.9-1.07)
PM2.6 organic carbon: 0.010
(-0.001-0.045)
PM2.6 elemental carbon: 4.1
(2.2-7.1)
Monitoring Stations: 1
Copollutant (correlation):
PM10: r - 0.84
O3: r - 0.65
NO2: r - 0.46
CO: r - 0.44
SO2: r - 0.17
PMio-2.b: r - .43
UFP: r - -0.16
PM2.6 water-sol metals: r - 0.70
PM2.6 sulfates: r - 0.77
PM2.6 acidity: r - 0.58
PM2.6 organic carbon: r - 0.51
PM2.6 elemental carbon: r - 0.48
oxygenated hydrocarbon: r - 31
Other variables:
Temperature: r - 0.20
Dew point: r - 0.00
PM Increment: Approximately 1 SD
increase: PM2.6: 10/yg/m3
PM2.6 water-sol metals: 0.03 //g|m3
PM2.6 sulfates: 5/yg/m3
PM2.6 acidity: 0.02 /yequ/m3
PM2.6 organic carbon: 2/yg/m3
PM2.6 elemental carbon: 1 /yglm3
RR [95% CI]: PM2.6 (3-day moving avg):
All CVD: 1.033[1.010,1.056]
Dysrhythmia: 1.015 [0.976,1.055]
Congestive heart failure: 1.055 [1.006—
1.105]
Ischemic heart disease: 1.023 [0.983—
1.064]
Peripheral vascular and cerebrovascular
disease: 1.050(1.008-1.093]
PM2.E water soluble metals (3-day moving
avg): All CVD: 1.027(0.998,1.056]
Dysrhythmia: 1.031 [0.982,1.082]
Congestive heart failure: 1.040 [0.981 —
1.103]
Ischemic heart disease: 1.000 [0.951 —
1.051]
Peripheral vascular and cerebrovascular
disease: 1.043 [0.991-1.098]
PM2.E sulfates (3-day moving avg): All
CVD: 1.003 [0.968,1.039]
Dysrhythmia: 0.986 [0.926,1.048]
Congestive heart failure: 1.009 [0.938—
1.085]
Ischemic heart disease: 0.997 [0.936—
1.062]
Peripheral vascular and cerebrovascular
disease: 1.025 [0.964-1.090]
PM2.6 acidity (3-day moving avg): All CVD:
0.994 [0.966,1.022]
Dysrhythmia: 0.991 [0.942,1.043]
Congestive heart failure: 0.989 [0.930—
1.052]
Ischemic heart disease: 0.992 [0.944—
1.043]
Peripheral vascular and cerebrovascular
disease: 1.004 [0.955-1.056]
PM2.5 organic carbon (3-day moving avg):
All CVD: 1.026(1.006,1.046]
Dysrhythmia: 1.008 [0.975,1.044]
Congestive heart failure: 1.048 [1.007—
1.091]
Ischemic heart disease: 1.028 [0.994—
1.064]
Peripheral vascular and cerebrovascular
disease: 1.026 [0.990-1.062]
hydrocarbons simultaneously.
PM2.E organic carbon (3-day moving avg):
All CVD: 1.020(1.005,1.036]
Dysrhythmia: 1.011 [0.985,1.037]
Congestive heart failure: 1.035 [1.003—
1-QffiRFT- DO NOT CITE OR QUOTE
Ischemic heart disease: 1.019 [0.992—
1.046]
Peripheral vascular and cerebrovascular
July 2009
E-128

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Peng et al. (2008,156850)
Period of Study: January 1,1999-
December 31, 2005
Location: 108 U.S. counties in the
following states: Alabama, Arizona,
California, Colorado, Connecticut, District
of Columbia, Florida, Georgia, Idaho,
Illinois, Indiana, Kentucky, Louisiana,
Maine, Maryland, Massachusetts,
Michigan, Minnesota, Missouri, Nevada,
New Hampshire, New Jersey, New
Mexico, New York, North Carolina, Ohio,
Oklahoma, Oregon, Pennsylvania, Rhode
Island, South Carolina, Tennessee, Texas,
Utah, Virginia, Washington, West Virginia,
Wisconsin
Outcome (ICD-9): Emergency
hospitalizations for: Cardiovascular
disease, including heart failure (428),
heart rhythm disturbances (426-427),
cerebrovascular events (430-438),
ischemic heart disease (410-414,429),
and peripheral vascular disease (440-
448). Respiratory disease, including
COPD (490-492) and respiratory tract
infections (464-466, 480-487)
Age Groups: 65 + years, 65-74, ,75 +
Study Design: Time series
N: —12 million Medicare enrollees (3.7
million CVD and 1.4 million RD
admissions)
Statistical Analyses: Two-stage
Bayesian hierarchical models:
Overdispersed Poisson models for county-
specific data
Bayesian hierarchical models to obtain
national avg estimate
Covariates: Day of the week, age-
specific intercept, temperature, dew point
temperature, calendar time, indicator for
age of 75 years or older. Some models
were adjusted for PMio-2.5.
Season: NR
Dose-response Investigated: No
Statistical Package: R version 2.6.2
Lags Considered: 0-2 days
Pollutant: PM2.6
Averaging Time: 24 h
Mean //gfm3 (IQR): All counties assessed:
13.5(11.1-15.8)
Counties in Eastern US:
13.8(12.3-15.8)
Counties in Western US:
11.1 (10.1-14.3)
Monitoring Stations: At least 1 pair of
co-located monitors (physically located in
the same place) for PMio and PM2.6 per
county
Other variables: Median within-county
correlations between monitors: r - 0.92
PM Increment: 10 //g|m3
Percentage change [95% CI]: CVD and RD
(unadjusted for PMio-25): Lag 0: 0.71
[0.45, 0.96]
Lag 2: 0.44(0.06, 0.82]
Most values NR (see note)
Notes: Effect estimates for PMio-2.5 (0-
2 day lags) are showing in Figures 2-5.
Figure 2: Percentage change in
emergency hospital admissions for CVD
per 10 /yglm3 increase in PM2.6 (single
pollutant model and model adjusted for
PMio-2.5 concentration)
Figure 3: Percentage change in
emergency hospital admissions for RD per
10 /yglm3 increase in PM2.6 (single
pollutant model and model adjusted for
PMio-2.5 concentration)
No significant associations between
PM2.E and cause-specific cardiovascular
disease.
Reference: Peters et al. (2005,156859)
Period of Study: February 1999-July
31,2001
Location: Germany: City of Augsburg,
County Augsburg, and County Aichach-
Friedlberg
Outcome: Transmural or nontransmural
acute Ml
Age Groups: NR
Study Design: Case-crossover and time
series
N: 851 Ml survivors
Statistical Analyses: Conditional
logistic regression for case-crossover
element. Poisson regression for time
series element.
Covariates: Case-crossover: Season,
temperature, day of the week, time
series: trend, season, influenza, weather,
and day of the week
Season: All
Dose-response Investigated: No
Statistical Package:
SAS, version 8.2
Poisson: R, version 1.7.1
Lags Considered:
Lags 0-6 h, 0-5 days
Poisson: Single lagged days, 5-day, 15-
day, 30-day, and 45-day moving averages
Pollutant: PM2.6
Averaging Time: 1 h and 24 h
Mean //g/m3 (range
IQR
median
IQR): 1-h avg: 16.3 (-6.9-355.2
10.7-19.8
14.5)
24-h avg: 16.3 (6.1-58.5
11.6-19.3
14.9)
Monitoring Stations: 1
Copollutant (correlation):
24-h avg:
TNC: r - 0.37
TSP: r - 0.89
PM10: r- 0.92
CO: r - 0.57
NO2: r - 0.67
NO: r - 0.59
SO2: r - 0.58
O3: r - -0.24
1 hr avg:
PM Increment: 1-h avg: 9.1 /yglm3 (IQR)
24-h avg: 7.7 //gfm3 (IQR)
OR [95% CI]: Case-Crossover (control
selection method (unidirectional with
three control periods): 1-h averages: Lag
0: 0.98(0.88,1.10)
Lag 1: 0.97 (0.87,1.09)
Lag 2: 0.93 (0.83,1.04)
Lag 3: 0.98 (0.88,1.09)
Lag 4: 0.96 (0.86,1.07)
Lag 5: 0.94 (0.84,1.05)
Lag 6: 0.90 (0.80,1.01). 24-h averages:
LagO: 0.95 (0.83,1.080)
Lag 1:1.10 (0.96,1.25)
Lag 2: 1.18(1.03,1.34)
Lag 3: 1.07 (0.94,1.22)
Lag 4: 0.94 (0.83,1.07)
Lag 5: 0.90 (0.79,1.02)
Case-Crossover (control selection
method: bidirectional with 16 control
periods): 24-h averages: Lag 0: 1.03
(0.94,1.12)
Lag 1: 1.07 (0.98,1.16)
Lag 2: 1.08 (0.99,1.17)
Lag 3: 1.01 (0.92,1.10)
Lag 4: 0.96 (0.88,1.04)
July 2009
E-129
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
TNC: r - 0.42
CO: r - 0.52
NO?: r - 0.58
NO: r - 0.50
SO?: r - 0.48
O3: r - -0.35
Other variables:
24-h avg: Temperature: r - 0.05
1-h avg: Temperature: r - -0.01
Lag 5: 0.93 (0.85,1.02)
Lag 0 -4 (IQR - 5.8): 1.03 (0.94,1.14)
Unidirectional: Model 1 (unadjusted):
1.175(1.033,1.337)
Model 2 (adjusted for day of week using
indicator variables): 1.179(1.035,1.343)
Model 3 (adjusted for temperature-
quadratic, linear air pressure): 1.170
(1.028,1.333)
Model 4 (adjusted for temperature-quad-
ratic, linear air pressure, day of week):
1.176(1.031,1.341)
Model 5 (temperature-quadratic, air
pressure-quadratic, relative humidity-
quadratic, day of week using indicator
variables): 1.170 (1.026,1.336)
Model 6 (temperature-penalized spline,
4.4 df, linear air pressure, day of week
using indicator variables): 1.175 (1.030,
1.340
Model 7 (temperature-penalized spline,
4.4 df, linear air pressure, relative
humidity-penalized spline, 7.8 df, day of
week using indicator variables: 1.177
(1.030,1.344)
Bidirectional (16 control periods): Model 1
(unadjusted): 1.077 (0.988,1.174)
Model 2 (adjusted for day of the week
using indicator variables): 1.078 (0.988,
1.175)
Model 3 (adjusted for temperature-
quadratic, linear air pressure): 1.060
(0.970,1.160)
Model 4 (adjusted for temperature-
quadratic, linear air pressure, day of the
week): 1.060 (0.969,1.160)
Model 5
	Continued on next page	
	Continued from previous page	
(temperature-quadratic, air pressure-
quadratic, relative humidity-quadratic,
day of the week using indicator
variables): 1.065 (0.973,1.166)
Model 6 (temperature-penalized spline,
4.4 df, linear air pressure, day of the
week using indicator variables): 1.068
(0.976,1.168)
Model 7 (temperature-penalized spline,
4.4 df, linear air pressure, relative
humidity-penalized spline, 7.8 df, day of
the week using indicator variables: 1.077
(0.983,1.179)
Bidirectional (4 control periods): Model 1
(unadjusted): NR
Model 2 (adjusted for day of the week by
design): 1.049 (0.964,1.141)
Model 3 (adjusted for temperature-
quadratic, linear air pressure): NR
Model 4 (adjusted for temperature-
quadratic, linear air pressure, day of the
week): 1.032 (0.944,1.128)
July 2009
E-130
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Model 5 (temperature-quadratic, air
pressure-quadratic, relative humidity-
quadratic, day of the week by design):
1.033 (0.945,1.130)
Model 6 (temperature-penalized spline,
4.4 df, linear air pressure, day of the
week by design): 1.036 (0.947,1.132)
Model 7 (temperature-penalized spline,
4.4 df, linear air pressure, relative
humidity-penalized spline, 7.8 df, day of
the week by design): 1.039 (0.950,
1.136)
Stratified: Model 1 (unadjusted): NR
Model 2 (adjusted for day of week by
design): 1.059 (0.972,1.154)
Model 3 (adjusted for temperature-
quadratic, linear air pressure): NR
Model 4 (adjusted for temperature-
quadratic, linear air pressure, day of
week): 1.047 (0.957,1.145)
Model 5 (temperature-quadratic, air
pressure-quadratic, relative humidity-
quadratic, day of week by design): 1.045
(0.954,1.144)
Model 6 (temperature-penalized spline,
4.4 df, linear air pressure, day of week by
design): 1.054 (0.964,1.153odel 7
(temperature-penalized spline, 4.4 df,
linear air pressure, relative humidity-
penalized spline, 7.8 df, day of week by
design): 1.056 (0.965,1.156)
RR (95% CI): Time series (24 h avg): Lag
0: 0.97(0.89,1.07)
Lag 1: 1.04 (0.96,1.13)
Lag 2: 1.07 (0.98,1.15)
Lag 3: 1.03 (0.95,1.11)
Lag 4: 0.98 (0.90,1.07)
Lag 5: 0.98 (0.90,1.06)
Lag 0-4: 1.03 (0.94,1.12)
Lag 0-14: 1.03 (0.95,1.13)
Lag 0-29: 1.09 (1.01,1.18)
Lag 0-44: 1.08 (1.00,1.17)
Time series (OR [95% CI]): Model 1
(unadjusted): 1.059 (0.981,1.142)
Model 2 (adjusted for day of week using
indicator variables): 1.056(0.979,1.140)
Model 3 (adjusted for temperature-
quadratic, linear air pressure): 1.062
(0.982,1.148)
Model 4 (adjusted for temperature-
quadratic, linear air pressure, day of
week): 1.059 (0.979,1.146)
Model 5 (temperature-quadratic, air
pressure-quadratic, relative humidity-
quadratic, day of week using indicator
variables): 1.063 (0.981,1.151)
Model 6 (temperature-penalized spline,
4.4 df, linear air pressure, day of week
using indicator variables): 1.065 (0.985,
1.153)
July 2009
E-131
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Model 7 (temperature-penalized spline,
4.4 df, linear air pressure, relative
humidity-penalized spline, 7.8 df, day of
week using indicator variables: 1.069
(0.988,1.157)
Reference: Pope et al.(2006, 091246)
Period of Study: 1994-2004
Location: Wasatch Front area, Utah
Outcome: Myocardial infarction or
unstable angina (ICD codes not reported)
Age Groups: All, < 65, 65 +
Study Design: Case-crossover
N: 12,865 patients who underwent
coronary arteriography
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and dew point
temperature
Season: NR
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: 0- to 3-day lag, 2- to
4-day lagged moving averages
Pollutant: PM2.6
Averaging Time: 24 h
Mean (/L/g|m3) (SD
maximum): Ogden: 10.8 (10.6
108)
SLC Hawthorne: 11.3 (11.9
94)
Provo/Orem, Lindom: 10.1 (9.8
82)
Monitoring Stations: 3
Copollutant (correlation): NR
PM Increment: 10 /yglm3
Percent increase in risk [95% CI]: Same-
day increase in PM2.6 (Lag 0): Index Ml
and unstable angina: 4.81 [0.98-8.79]
Subsequent Ml: 3.23 [-3.87,10.85]
All acute coronary events: 4.46 [1.07-
7.97]
All acute coronary events excluding
observations using imputed PM2.6 data:
4.24(0.33-8.31]
Stable presentation: -2.57 [-5.39, 0.34]
Remaining results summarized in figures
(see notes).
Notes: Figure 1: Percent increase in risk
(and 95% CI) of acute coronary events
associated with 10/yg/m3 of PM2.6 for
different lag structures.
Summary of Figure 1: Positive,
statistically significant association seen
for Lag 0, Lag 1
and 2, 3, and 4 day moving averages.
Positive but non-statistically significant
associations seen for Lags 2 and 3.
Figure 2: Percent increase in risk (and
95% CI) of acute coronary events
associated with 10/yg/m of PM2.6
stratified by various characteristics.
July 2009
E-132
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Pope et al. (2008,191969)
Period of Study: 1994-2006
Location: Ogden, Salt Lake City, &
Provo/Orem, Utah
Outcome: Heart Failure Hospitalizations
Age Groups: NR
Study Design: case-crossover
l\l: 2,618
Statistical Analyses: Conditional
Logistic Regression
Covariates: age, sex, length of stay,
temperature, pressure, clearing index, day
of the week, seasonality, and long-term
trends
Season: Adjusted for long-term trends to
account for season
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-28d moving avg.
Pollutant: PM2.6
Averaging Time: NR
Mean (SD):
Ogden: 10.6(9.9)
SLC, Hawthorne: 11.1 (11.2)
Provo/Orem, Lindon: 10.1 (9.3)
Max:
Ogden: 108
SLC, Hawthorne: 94 Provo/Orem, Lindon:
82
Monitoring Stations: NR
Copollutant: PMio
PM Increment: 10/yg/m3
Percent Increase: (Lower CI, Upper
CI):
All HF Admissions
All: 13.1 (1.3, 26.2)*
Men: 13.4 (-1.7, 30.7)*
Women: 12.7 (-5.1, 33.9)
Age < 65 yrs: 3.5 (-13.5, 23.8)
Age £65 yrs: 19.6 (4.0, 37.5)*
Length of stay 0-2 days: 24.4 (-0.8, 56.0)
*
Length of stay 3-7 days: 10.8 (-4.6, 28.7)
Length of stay 8+ days: 6.5 (-15.9, 34.8)
First HF Admissions: 2.1 (-11.3,17.5)
Subsequent HF Admits: 32.4 (10.7, 58.4)
All HF Admissions
All: 32.4 (10.7, 58.4)*
Men: 29.2(2.7, 62.6)*
Women: 41.5 (5.4, 89.9)*
Age <65 yrs:-3.1 (-26.5, 27.8)
Age >65 yrs: 64.1 (28.6,109)*
Length of stay 0-2 days: 68.9 (12.5,
154)*
Length of stay 3-7 days: 35.7 (5.9,
73.9)*
Length of stay 8+ days: 2.6 (-28.5,47.1)
*p< 0.05, * p< 0.01, *p< 0.10
July 2009
E-133
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Sarnat et al. (2008, 097972)
Period of Study: November 1998-
December 2002
Location: Atlanta (Georgia) metropolitan
area
Age Groups: All
Study Design: Time series
N: >4.5 million emergency department
visits
Statistical Analyses: Poisson
generalized linear models
Covariates: Day of the week, holidays,
hospital, long-term trends, temperature,
dew point temperature
Season: All, warm season (April 15-
October 14), and cool season (October
15-April 14).
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: 0-day lag
Outcome (ICD-9): Cardiovascular disease
ED visits: Ischemic heart disease (410—
414), cardiac dysrhythmias (427),
congestive heart failure (428), and
peripheral vascular and cerebrovascular
disease (433-437,440, 443-444,451-
453)
Pollutant: PM2.6
Averaging Time: 24 h
Mean l//g/m3) (median
10th-90th percentile): Total PM2.6: Cool
season: 15.8 (14.3
7.5-25.5).	Warm season: 18.2 (17.0
9.1-29.0)
PM2.6 elemental carbon: Cool: 1.7 (1.4
0.6-3.3). Warm: 1.4 (1.3
0.6-2.5)
PM2.B Zn (ng/m3): Cool: 15.7 (11.7
4.6-30.2)
Warm: 10.9 (8.5
3.3-20.2)
PM2.6 K (ng/m3): Cool: 63.0 (53.9
24.3-114.2) Warm: 52.7 (43.3
23.2-93.5)
PM2.6 Si (ng/m3): Cool: 67.7 (54.1
24.3-123.5).	Warm: 110.9(89.0
32.9-186.3)
PM2.6 SO42': Cool: 3.4 (0.6
1.5-5.8). Warm: 6.0 (5.2
2.3-10.8)
PM2.6 N03-: Cool: 1.4 (1.2
0.5-2.6). Warm: 0.7 (2.9
0.3-1.2)
PM2.6 Se (ng/m3): Cool: 1.4 (1.1
0.4-3.0). Warm: 1.2 (0.9
0.4-2.7)
PM2.6 0C: Cool: 4.6 (3.9
1.9-8.0)
Warm: 4.0 (3.7
2.1-6.4)
Monitoring Stations: 1
Copollutants: NR
PM Increment: IQR (specific values not
given)
Risk ratio [95% CI]: CVD (Lag 0): All
seasons: Total PM2.6: 1.022 [1.007,
1.038]
PM2.6 elemental carbon: 1.02 [1.013—
1.037]
PM2.6 zinc: 1.013 [1.005-1.022]
PM2.6 potassium: 1.030 [1.018-1.042]
PM2.6 silicon: 1.008(1.00-1.016]
PM2.6 sulfate: 1.007 [0.994-1.019]
PM2.6 nitrate: 1.002 [0.990-1.014]
PM2.6 selenium: 1.002 [0.991 -1.012]
PM2.6 organic carbon: 1.024 [1.013-
1.035]
Cool season: Total PM2.6: 1.028 [1.012-
1.044]
PM2.6 EC: 1.029 [1.015-1.044]
PM2.6 Zinc: 1.012(1.002-1.022]
PM2.6 K: 1.037(1.021-1.054]
PM2.6 Si: 1.022(1.002-1.043]
PM2.6 sulfate: 1.014 [0.991-1.037]
PM2.6 nitrate: 1.006 [0.993-1.019]
PM2.6 Se: 1.012 [0.997-1.027]
PM2.6 organic carbon: 1.027 [1.013-
1.040]
Warm season: Total PM2.6: 1.006
[0.990-1.022]
PM2.6 EC: 1.021 [1.000-1.043]
PM2.6 Zinc: 1.017 [1.002-1.033]
PM2.6 K: 1.024 [1.007-1.041]
PM2.6 Si: 1.005 [0.996-1.014]
PM2.6 sulfate: 1.001 [0.988-1.015]
PM2.6 nitrate: 1.000 [0.969-1.033]
PM2.6 Se: 0.996 [0.981-1.011]
PM2.B organic carbon: 1.027 [1.004-
1.051]
July 2009
E-134
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Schreuder et al. (2006,
097959)
Period of Study: Sept 1995 - May
2002
Location: Spokane, Washington
Outcome: Cardiac HA
Age Groups: NR
Study Design: time series
Statistical Analyses: GAM Poisosn
Regression
Covariates: season, temperature,
relative humidity, day of week
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0-1 d
Pollutant: PM2.6 (ng/m3)
Averaging Time: 24h
Arithmetic Mean: 10,580
Geometric Mean: 8,790
Min: 930
Max: 43,230
IQR:
entire period: 7.7 //g|m3
heating season: 10.1/yg/m3
non-heating season: 5.5/yg/m3
Monitoring Stations: NR
Copollutant: NR
Co-pollutant Correlation
NR
PM Increment: Interquartile I
Relative Risk (Lower CI, Upper CI):
Entire Period, Lag 0: 1.008 (0.985,
1.032)
Entire Period, Lag 1: 1.000 (0.978,
1.023)
Heating Season, Lag 1:1.015 (0.968,
1.063)
Non-Heating Season, Lag 1: 0.995
(0.969,1.021)
Reference: Sullivan et al. (2005,
109418)
Period of Study: 1988-1994
Location: King County, Washington
Outcome: Acute Ml
Age Groups: All, < 50, 50-59, 70 +
Study Design: Case-crossover
N: 5793 cases of acute Ml (5793 case
days and 20,134 referent exposure days
from these case individuals)
Statistical Analyses: Conditional
logistic regression
Covariates: Relative humidity,
temperature, season, day of week
Season: All, and also conducted
stratified analysis by season of event
(heating season: November-February
nonheating season: March-October)
Dose-response Investigated: No
Statistical Package: SAS version 8.0
and SPSS version 10
Lags Considered: Lag 1 and Lag 2 for
24-h avg
Pollutant: PM2.6
Averaging Time: 1 h, 2 h, 4 h, and 24 h
Summary of PM2.61 h before Ml onset:
Mean (/L/g|m3) (median
IQR, 90th percentile
range): 12.8 (8.6
5.3-15.9
27.3
2.0-147)
Monitoring Stations: 3
Copollutant (correlation): 1-h avg:
PM10: r - 0.78
CO: r - 0.47
SO2: r - 0.16
PM Increment: 10 /yglm3
Odds ratio [95% CI]:
1-h	Averaging Time: 1.01 [0.98,1.05]
2-h	Averaging Time: 1.01 [0.97,1.05]
4-h Averaging Time: 1.02 [0.98,1.04]
24-h Averaging Time: 1.02 [0.98,1.07]
Association between PM2.6 (24 h) lagged
1 or 2 days non-significant (data not
shown)
Season (1-h avg): Heating: 1.01 [0.98—
1.05]
Nonheating: 0.99 [0.91-1.09]
Age (1-h avg): < 50 years: 1.04 [0.95,
1.14]
50-60 years: 0.99 [0.94,1.05]
70+ years: 1.03 [0.98,1.08]
Age (24-h avg): < 50 years: 1.07 [0.98,
1.19]
50-69 years: 0.99 [0.93,1.06]
70+ years: 1.04 [0.99,1.11]
Sex (1-h avg): Men: 1.02[0.98,1.06]
Women: 1.00 [0.95,1.06]
Sex (24-h avg): Men: 1.03 [0.99,1.08]
Women: 1.00 [0.94,1.07]
Race (1-h avg): White: 1.01 [0.97,1.04]
Nonwhite: 1.06 [0.97,1.17]
Race (24-h avg): White: 1.01 [0.97,1.06]
Nonwhite: 1.10[0.99,1.23]
Smoking status (1-h avg): Current: 0.99
[0.93,1.06]
Nonsmoker: 1.03[0.97,1.08]
Smoking status (24-h avg): Current: 0.99
[0.95,1.14]
July 2009
E-135
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Nonsmoker: 1.03 [0.98,1.09]
Survivor of Ml * (1 -h avg): Yes: 1.02
[0.98,1.06]
No: 0.96(0.86,1.08]
Survivor of Ml * (24-h avg): Yes: 1.03
[0.98,1.07]
No: 0.97(0.85,1.10]
Previous congestive heart failure (1 h
avg): Yes: 1.06 [0.97,1.16]
No: 1.00(0.97,1.04]
Previous congestive heart failure (24-h
avg): Yes: 1.08 [0.97,1.2]
No: 1.00(0.97,1.04]
Previous Ml (1 -h avg): Yes: 1.03 [0.97,
1.1]
No: 1.01 [0.96,1.06]
Previous Ml (24-h avg): Yes: 1.04 [0.97,
1.17]
No: 1.02(0.98,1.08]
Hypertension (1-h avg): Yes: 1.02 [0.97,
1.07]
No: 1.01 [0.96,1.06]
Hypertension (24-h avg): Yes: 1.02 [0.97,
1.07]
No: 1.02(0.97,1.08]
Diabetes mellitus (1 -h avg): Yes: 1.06
[0.98,1.14]
No: 1.01 [0.97,1.05]
Diabetes mellitus (24-h avg): Yes: 1.04
[0.95,1.14]
No: 1.01 [0.97,1.06]
"Compares those who survive
hospitalization (yes) with those who died
in hospital from complications of Ml.
July 2009
E-136
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Symons et al. (2006,
091258)
Period of Study: Apr-Dec, 2002
Location: Baltimore, Maryland
Outcome: Congestive heart failure
Age Groups: All
Study Design: Case-crossover
N: 125 patients
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and humidity
Season: NR
Dose-response Investigated: Yes
Statistical Package: SAS and S-Plus
Lags Considered: 0-3 days (single and
cumulative)
Pollutant: PM2.6
Averaging Time: 8 & 24 h
Mean (min-max):
8 h
17.0(0.1-111.9)
SD - 12.7
24 h
16.0 (3.5-69.2)
SD - 10.0
Monitoring Stations: 8
Copollutant (correlation): NR
PM Increment: 9.2 /yglm3 (IQR)
RR Estimate [CI]:
8 h (participant's onset period)
Same-day lag: 0.87 [0.69,1.09]
1-day	lag: 0.96 [0.78,1.18]
2-day	lag: 1.09 [0.91,1.30]
3-day	lag: 0.99 [0.79,1.23]
Cumulative 1-day lag: 0.89 [0.67,1.16]
Cumulative 2-day lag: 0.99 [0.74,1.33]
Cumulative 3-day lag: 0.98 [0.70,1.36]
24 h avg
Same-day lag: 0.81 [0.65,1.01]
1-day	lag: 0.90 [0.74,1.11]
2-day	lag: 0.85 [0.68,1.07]
3-day	lag: 0.86 [0.70,1.05]
Cumulative 1-day lag: 0.82 [0.64,1.04]
Cumulative 2-day lag: 0.76 [0.57,1.01]
Cumulative 3-day lag: 0.70 [0.51,0.97]
Notes: (3 coefficients presented in Fig 5
July 2009
E-137
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Tolbert et al. (2007,
090316)
Period of Study: August 1998-
December 2004
Location: Atlanta Metropolitan area,
Georgia
Outcome (ICD-9):
Combined CVD group, including:
Ischemic heart disease (410-414),
cardiac dysrhythmias (427), congestive
heart failure (428), and peripheral
vascular and cardiovascular disease
(433-437,440,443-445, and 451 -
453)
Age Groups: All
Study Design: Time series
l\l:NR for 1998-2004.
For 1993-2004:10,234,490 ER visits
(283,360 and 1,072,429 visits included
in the CVD and RD groups, respectively)
Statistical Analyses: Poisson
generalized linear models
Covariates: long-term temporal trends,
season (for RD outcome), temperature,
dew point, days of week, federal
holidays, hospital entry and exit
Season: All
Dose-response Investigated: No
Statistical Package: SAS version 9.1
Lags Considered: 3-day moving
0-2)
July 2009
Pollutant: PM2.6
Averaging Time: 24 h
Mean l//g/m3) (median
IQR, range, 10th -90th percentiles): PM2.6:
17.1 (15.6
11.0-21.9
0.8-65.8
7.9-28.8). PM2.6 sulfate: 4.9 (3.9
2.4-6.2
0.5-21.9
1.7-9.5). PM2.6 organic carbon: 4.4 (3.8
2.7-5.3
0.4-25.9
2.1-7.2). PM2.6 elemental carbon: 1.6
(1.3
0.9-2.0
0.1-11.9
0.6-3.0). PM2.6 water-soluble metals:
0.030 (0.023
0.014-0.039
0.003-0.202
0.009-0.059)
Monitoring Stations: 1
Copollutant (correlation): Between
PM2.6 and:
PM10: r - 0.84
O3: r - 0.62
NO2: r - 0.47
CO: r - 0.47
SO2: r - 0.17
PMio-2.5: r - 0.47
PM2.6 SO4: r - 0.76
PM2.6 EC: r - 0.65
PM2.6 0C: r - 0.70
PM2.6 TC: r - 0.71
PM2.6 water-sol metals: r - 0.69
0HC: r - 0.50
Between PM2.6 SO4 and: PM10: r - 0.69
O3: r - 0.56
NO2: r - 0.14
CO: r - 0.14
SO2: r - 0.09
PMio-2.5: r - 0.32
PM2.6: r - 0.76
PM2.6 EC: r - 0.32
PM2.6 0C: r - 0.33
PM2.6 TC: r - 0.34
PM2.6 water-sol metals: r - 0.65
OHC: r - 0.47
-1 Sfitween PM2.6 elemental carbon and:
PM10: r - 0.61
O3: r - 0.40
PM Increment:
PM2.6:10.96/yglm3 (IQR)
PM2.6 sulfate: 3.82/yg/m3 (IQR)
PM2.6 total carbon: 3.63//g|m3 (IQR)
PM2.6 organic carbon: 2.61 //g|m3 (IQR)
PM2.6 elemental carbon: 1.15 /yg/m3 (IQR)
PM2.6 water-soluble metals: 0.03 /yglm3
(IQR)
Risk ratio [95% CI] (single pollutant
models):
PM2.6:
CVD: 1.005 [0.993-1.017]
PM2.6 sulfate:
CVD: 0.999 [0.987-1.011]
PM2.6 total carbon:
CVD: 1.016(1.005-1.026]
PM2.E organic carbon:
CVD: 1.015(1.005-1.026]
PM2.E elemental carbon:
CVD: 1.015 [1.005-1.025]
PM2.E water-soluble metals:
CVD: 1.009 [0.997-1.021]
Notes: Results of selected multi-pollutant
models for cardiovascular disease are
presented in Figure 1.
Figure 1: PM2.6 total carbon adjusted for
CO, NO2, or NO2+CO
Summary of results: PM2.6 total carbon
continued to have a positive, statistically
significant association with CVD after
adjustment for NO2 but not after
adjustment
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Villeneuve et al. (2006,
090191)
Period of Study: 1 Apr, 1992-31 Mar,
2002
Location: Edmonton, Canada
Outcome (ICD-9): Stroke (430-438),
including ischemic stroke (434-436),
hemorrhagic stroke (430,432), and
transient ischemic attacks (TIA) (435).
Age Groups: 65+ yrs
Study Design: Case-crossover
N: 12,422 visits
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and relative
humidity
Season: Summer (Apr-Sep), winter (Oct-
Mar)
Dose-response Investigated: No
Statistical Package: SAS (PHREG)
Lags Considered: 0,1, and 3-day
Pollutant: PM2.6
Averaging Time: 24 h
Mean /yglm3 (SD):
All year: 8.5(6.2)
Summer: 8.7 (7.1)
Winter: 8.3 (5.2)
Monitoring Stations: 3
Copollutant (correlation):
All year
SO?: r - 0.22
NO?: r - 0.41
CO: r - 0.43
03-mean: r - -0.07
03-max: r - 0.07
PM10: r - 0.79
Summer
SO2: r - 0.20
NO2: r - 0.52
CO: r - 0.42
03-mean: r - 0.11
03-max: r - 0.34
PM10: r - 0.85
Winter
SO2: r - 0.28
NO2: r - 0.57
CO: r - 0.71
03-mean: r - -0.45
03-max: r - -0.35
PM10: r - 0.70
PM Increment:/yg/m3 (IQR)
All year: 6.3
Summer: 6.5
Winter: 6.0
Adjusted OR Estimate [CI]:
Acute ischemic stroke
All year: Same-day lag: 1.00 [0.96,1.04]
1-day lag: 1.00 [0.96,1.05]
3-day lag: 1.01 [0.96,1.06]
Summer: Same-day lag: 0.96 [0.90,1.03]
1-day lag: 1.01 [0.94,1.07]
3-day lag: 0.98 [0.89 [1.07]
Winter: Same-day lag: 1.04 [0.99,1.10]
1-day lag: 1.01 [0.96,1.07]
3-day lag: 1.05 [0.98,1.13]
Hemorrhagic stroke
All year: Same-day lag: 0.99 [0.90,1.08]
1-day lag: 1.07 [0.98,1.16]
3-day lag: 1.05 [0.93,1.19]
Summer: Same-day lag: 0.99 [0.86,1.15]
1-day lag: 1.12(0.97,1.30]
3-day lag: 1.08 [0.88,1.31]
Winter: Same-day lag: 1.04 [0.92,1.18]
1-day lag: 1.08 [0.97,1.20]
3-day lag: 1.11 [0.94,1.31]
Transient cerebral ischemic attack
All year: Same-day lag: 0.98 [0.93,1.03]
1-day lag: 0.99 [0.95,1.04]
3-day lag: 0.96 [0.90,1.03]
Summer: Same-day lag: 1.00 [0.92,1.08]
1-day lag: 1.03 [0.95,1.12]
3-day lag: 0.98 [0.88,1.09]
Winter: Same-day lag: 0.97 [0.90,1.05]
1-day lag: 0.97 [0.91,1.04]
3-day lag: 0.94 [0.86,1.03]
Notes: Adjusted ORs are provided for an
IQR increase in the 3-day mean in Fig 1-4
for single and two-pollutant models.
July 2009
E-139
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Zanobetti and Schwartz
Outcome (ICD-9): Myocardial infarction
Pollutant: PM2.6
PM Increment: Difference between the
(2006,090195)
(410) or pneumonia (480-487)
Averaging Time: 24 h
90th and 10th percentile for PM2.6
Period of Study: 1995-1999
Age Groups: 65 + years
Median (/L/g|m3) (IQR
Myocardial infarction cohort (Lag 0):
17.17/yg|m3
Location: Boston Metropolitan area
Study Design: Case-crossover
5th-95th percentile):
Myocardial infarction cohort (Lag 0-1):

N: 15,578 patients admitted for Ml and
25,857 admitted for pneumonia
11.1 (7.23-16.14
16.32 /yglm3


3.87-26.31)
Pneumonia cohort (Lag 0): 17.14/yglm3

Statistical Analyses: conditional

logistic regression
Monitoring Stations: 1
Pneumonia cohort (Lag 0): 16.32 /yglm3

Covariates: temperature, day of the
week.
Copollutant (correlation):
Percentage (%) increase in risk [95% CI]:

BC: r - 0.66
Myocardial infarction cohort:

Season: All, and also tested for
LagO: 8.50 (1.89-14.43)

interaction by warm (April-September)
NO?: r - 0.55

vs. cold season
CO: r - 0.52
Lag 0-1:8.65(1.22-15.38)

Dose-response Investigated: No
0s: r - 0.20
Pneumonia cohort:

Statistical Package: SAS version 8.2
(PR0C PHREG)
PM non-traffic: r - 0.74
LagO: 6.48 (1.13-11.43)

Lags Considered: lag 0 , and mean of

Lag 0-1: 5.56 (-0.45,11.27)

lags 0 -1

Notes: Assessed for effect modification
by season. Results are reported in Figure
2. Summary of results: PM2.6 is
associated with pneumonia
hospitalization in the cold season but not
the hot season. PM2.6 is associated with
Ml hospitalization in the hot season but
not the cold season.
Reference: Zanobetti and Schwartz
(2006,090195)
Period of Study: 1995-1999
Location: Boston Metropolitan area
Outcome (ICD-9): Myocardial infarction
(410) or pneumonia (480-487)
Age Groups: 65 + years
Study Design: Case-crossover
N: 15,578 patients admitted for Ml and
25,857 admitted for pneumonia
Statistical Analyses: conditional
logistic regression
Covariates: temperature, day of the
week.
Season: All, and also assessed for
interaction by hot (April-September) vs.
cold season
Dose-response Investigated: No
Statistical Package: SAS Software
Release 8.2
Lags Considered: lag 0 , and mean of
lags 0 -1
Pollutant: BC
Averaging Time: 24 h
Median l//g/m3) (IQR
5th-95th percentiles): 1.15 (0.74-1.72
0.42-2.83)
Monitoring Stations: 1
Copollutant (correlation):
PM2.6: r - 0.66
NO2: r - 0.70
CO: r - 0.82
0a: r - -0.25
PM non-traffic: r - -0.01
PM Increment: Difference between the
90th and 10th percentile for BC
Myocardial infarction cohort (Lag 0): 2.01
Z^g/m3
Myocardial infarction cohort (Lag 0-1):
1.69/yg/m3
Pneumonia cohort (Lag 0): 2.05 //g|m3
Pneumonia cohort (Lag 0 -1): 1.69 /yglm3
Percentage (%) increase in risk [95% CI]:
Myocardial infarction cohort:
Lag 0:6.98 (-0.27-13.76)
Lag 0-1:8.34 (0.21-15.82)
Pneumonia cohort:
| Lag 0: 10.76(4.54-15.89)
Lag 0-1: 11.71 (4.79,17.36)
Notes: Assessed for effect modification
by season. Results are reported in Figure
2. Summary of results: PM2.BC is
associated with pneumonia
hospitalization in the cold season but not
the hot season. BC had a stronger
positive association with Ml
hospitalization in the cold season, but the
confidence interval was wide.
'All units expressed in //gfm3 unless otherwise specified.
Table E-8. Short-term exposure-cardiovascular-ED/HA-other size fractions.
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Reference: Andersen et al, (2008, Outcome (ICD-10): CVD, including angina Pollutant: Total number concentration of PM Increment: IQR increase in pollutant
189651) pectoris (I20), myocardial infarction (121- ultrafine and accumulation mode particles level: Nctot: 3907 particles/cm3 (IQR)
	22), other acute ischemic heart diseases	
July 2009
E-140
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
(NCtot) (particles/cm3)
Averaging Time: 24 h
Mean (SD
median
IQR
99th percentile:
NCtot*: 8116(3502
7358
5738-9645,19,895)
NCa12: 493 (315
463
308-650
1463)
Nca23: 2253(1364
2057
1280-3066
6096)
NCa57: 5104 (2687
4562
3248-6274
14,410)
NC100: 6847 (2864
6243
4959-8218
16189)
NCa2i2: 392 (441
89
246-584
2248)
*NC, number concentration
tot, total (all particles 6-700 in diameter)
a12, size mode with mean diameter of 12
nm
a23, size mode with median diameter of
23 nm
a57, size mode with median diameter of
57 nm
a212
size mode with median diameter of 212
nm
NC100 - a12+a23 + 0.797*a57+0.084
*a212.
Monitoring Stations: 1
Copollutant (correlation): Correlation of
NCtot with:
PMio: r - 0.39
PM2.6: r - 0.40
NO2: r - 0.68
Effect Estimates (95% CI)
Period of Study: May 2001-December
2004
Location: Los Angeles and San
counties, California
(124),	chronic ischaemic heart disease
(125),	pulmonary embolism (I26), cardiac
arrest (I46), cardiac arrhythmias (I48-
48), and heart failure (I50).
RD, including chronic bronchitis (J41-
42), emphysema (J43), other chronic
obstructive pulmonary disease (J44),
asthma (J45), and status asthmaticus
(J46).
Pediatric hospital admissions for asthma
(J45) and status asthmaticus (J46).
Age Groups: > 65 yrs (CVD and RD), 5-
18 years (asthma)
Study Design: Time series
l\l:NR
Statistical Analyses: Poisson GAM
Covariates: temperature, dew-point
temperature, long-term trend,
seasonality, influenza, day of the week,
public holidays, school holidays (only for
5-18 year olds), pollen (only for pediatric
asthma outcome)
Season: NR
Dose-response Investigated: No
Statistical Package: R statistical
software (gam procedure, mgcv package)
Lags Considered: Lag 0 -5 days, 4-day
pollutant avg (lag 0 -3) for CVD, 5-day
avg (lag 0-4) for RD, and a 6-day avg (lag
0-5) for asthma.
Nca12: 342 particles/cm3 (IQR)
Nca23: 1786 particles/cm3 (IQR)
Nca57: 3026 particles/cm3 (IQR)
NC100: 3259 particles/cm3 (IQR)
Nca212: 495 particles/cm3 (IQR)
Relative risk (RR) Estimate [CI]: CVD
hospital admissions (4 day avg, lag 0 -3),
age 65 +
One-pollutant model (NCtot): 1.00 [0.99—
1.02]
Adj for PM10: 0.98 [0.96-1.01]
Adj for PM2.6: 0.99 [0.95-1.03]
Adj for CO: 0.99 [0.97-1.02]
Adj for NO2: 1.01 [0.98-1.03]
Adj for 0s: 1.01 [0.96-1.06]
One-pollutant model (NC100): 1.00
[0.98-1.02]
One pollutant model (Nca12): 0.99 [.97-
1.01]
Adj for other size fractions: 0.99 [0.97-
1.02]
One pollutant model (Nca23): 0.99 [0.96—
1.01]
Adj for other size fractions: 0.99 [0.96-
1.02]
One pollutant model (NCa57): 1.01
[0.98-1.02]
Adj for other size fractions: 0.99 [0.97-
1.02]
One pollutant model (Nca212): 1.02
[1.00-1.04]
Adj for other size fractions: 1.02 [1.00-
1.05]
Adj for PM10: 0.98 [0.95-1.01]
RD hospital admissions (5 day avg, lag 0 -
4), age 65+: One-pollutant model: 1.04
[1.00-1.07]
Adj for PM10:1.00 [0.96-1.05]
Adj for PM2.6: 0.97 [0.89-1.05]
Adj for CO: 1.03 [0.98-1.07]
Adj for NO2: 1.00(0.95-1.05]
Adj for 0a: 0.95(0.87-1.04]
One pollutant model (NC100): 1.03
[0.99-1.07]
One pollutant model (Nca12): 1.01 [0.98—
1.05]
Adj for other size fractions: 1.01 [0.97-
1.05]
One pollutant model (Nca23): 0.99 [0.94—
1.03]
Adj for other size fractions: 0.98 [0.94-
1.03]
One pollutant model (Nca57): 1.04 [1.00—
July 2009
E-141
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
NOx: r - 0.66
NC100: r - 0.98
NCa12: r - 0.31
NCa23: r - 0.57
NCa57: r - 0.87
NCa2i2: r - 0.29
CO: r - 0.54
NOx curbside: r - 0.36
Os: r - -0.12
Other variables: Temperature: r - -0.06
Relative humidity: r - -0.04
1.08]
Adj for other size fractions: 1.02 [0.97—
1.06]
One pollutant model (Nca212): 1.04
[1.01-1.08]
Adj for other size fractions: 1.03 [0.99—
1.07]
Adj for PMio: 1.01 [0.96-1.07]
Asthma hospital admissions (6 day avg
lag 0-5), age 5-18: One-pollutant model:
1.07 [0.98-1.17]
Adj for PMio: 1.03 [0.92-1.15]
Adj for PM2.6: 1.04 [0.85-1.28]
Adj for CO: 1.09 [0.99-1.21]
Adj for NO2: 1.07(0.96-1.19]
Adj for O3: 1.08 [0.87-1.35]
One pollutant model (NC100): 1.06
[0.97-1.16]
One pollutant model (Nca212): 1.08
[0.99-1.18]
Adj for other size fractions: 1.07 [0.97—
1.19]
One pollutant model (Nca23): 1.09 [0.98—
1.21]
Adj for other size fractions: 1.08 [0.97—
1.21]
One pollutant model (Nca57): 1.02 [0.94—
1.12]
Adj for other size fractions: 0.93 [0.83—
1.04]
One pollutant model (Nca212): 1.08
[1.00-1.17]
Adj for other size fractions: 1.12 [1.02-
1.23]
Adj for PMio: 1.10(0.96-1.13]
Notes: Figure 2: Relative risks and 95%
confidence intervals per IQR in single day
concentration (0-5 day lag).
Summary of Figure 2: CVD: Positive,
marginally or statistically significant
associations at Lag 2 (Nctot, Nca57,
Nca212), Lag 3 (Nca212), and Lag 1
(Nca212). RD: Positive, statistically or
marginally significant associations at Lag
4 (Nctot, Nca57, NCa212) and Lag 5
(Nctot, Nca57, Nca212), and to a lesser
extent Lag 2 (Nctot, Nca212) and Lag 3
(Nctot, Nca212). Asthma: Wide
confidence intervals make interpretation
difficult. Positive, significant association
for Nca212 at Lag 1.
July 2009
E-142
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lanki et al. (2006, 089788) Outcome (ICD-9): Acute myocardial
infarction (410
Period of Study: 1992-2000
Location: Augsburg, Barcelona, Helsinki,
Rome, and Stockholm
ICD-10:121,122)
Age Groups: 35+ yrs,
< 75 yrs, 75+ yrs
Study Design: Time series
N: 26,854 hospitalizations
Statistical Analyses: GAM
Covariates: Temperature, barometric
pressure
Season: Warm (Apr-Sep) and cold (Oct-
Mar)
Dose-response Investigated: No
Statistical Package: R package mgcv
0.9-5
Lags Considered: 0-3 days
Pollutant: UFP (PNC)
Averaging Time: 24 h
Median particles/cm3: Augsburg: 12,400
Barcelona: 76,300
Helsinki: 13,600
Rome: 46,000
Stockholm: 11,800
Copollutant (correlation): Augsburg
PMio: r - 0.53
CO: r - 0.63
NO2: r - 0.65
0s: r - 0.26
Barcelona: PMio: r - 0.38
CO: r - 0.80
NO?: r - 0.49
O3: r - -0.35
Helsinki: PMio: r - 0.45
CO: r - 0.48
NO2: r - 0.82
Os: r - 0.01
Rome: PMio: r - 0.32
CO: r - 0.83
NO2: r - 0.68
O3: r = 0.03
Stockholm: PMio: r - 0.06
CO: r - 0.56
NO2: r - 0.83
Os: r - -0.01
PM Increment: 10,000 particles/cm3
Pooled Rate Ratio [CI]: All 5 cities (35 +
yrs)
Same-day lag: 1.005 [0.996,1.015]
1-day	lag: 0.997 [0.982,1.012]
2-day	lag: 0.999 [0.990,1.008]
3-day	lag: 0.998 [0.979,1.017]
3 cities with hospital discharge register
(35+ yrs)
Same-day lag: 1.013 [1.000,1.026]
1-day	lag: 0.995 [0.953,1.039]
2-day	lag: 1.001 [0.989,1.014]
3-day	lag: 1.009 [0.974,1.046]
Warm season (35+ yrs)
Same-day lag: 1.009 [0.972,1.048]
1-day	lag: 1.023(0.988,1.060];
2-day	lag: 1.050 [1.016,1.085]
3-day	lag: 1.022 [0.987,1.058]
Cold season (35+ yrs)
Same-day lag: 1.014(1.001,1.028]
1-day	lag: 1.001 [0.956,1.048]
2-day	lag: 1.001 [0.989,1.014]
3-day	lag: 1.009 [0.971,1.049]
Age >75Non-fatal
Same-day lag: 1.032 [1.008,1.056]
1-day	lag: 1.009 [0.985,1.032]
2-day	lag: 0.989 [0.966,1.013]
3-day	lag: 1.009 [0.969,1.051]
Fatal
Same-day lag: 1.016 [0.978,1.055]
1-day	lag: 1.001 [0.966,1.038]
2-day	lag: 1.005 [0.969,1.041]
3-day	lag: 0.984 [0.948,1.021]
Notes: Rate ratios for PNC are given for
0-5 lag days in graph form (Fig 1) for
each city. Pooled rate ratios were also
provided for groups < 75 yielding similar
results to the overall 3-city data.
July 2009
E-143
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Metzgeret al. (2004,
044222)
Period of Study: August 1998-August
2000
Location: Atlanta Metropolitan area
(Georgia)
Outcome (ICD-9): Emergency visits for
ischemic heart disease (410-414),
cardiac dysrhythmias (427), cardiac
arrest (427.5), congestive heart failure
(428), peripheral vascular and
cerebrovascular disease (433-437, 440,
443-444, 451-453), atherosclerosis
(440), and stroke (436).
Age Groups: All
Study Design: Time series
N: 4,407,535 emergency department
visits between 1993-2000 (data not
reported for 1998-2000)
Statistical Analyses: Poisson
generalized linear modeling
Covariates: Day of the week, hospital
entry and exit indicator variables,
federally observed holidays, temporal
trends, temperature, dew point
temperature
Season: All
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: 3-day moving avg,
lags 0-7
Pollutant: UFP (10-100 nm particle
count) (no/cm3)
Averaging Time: 24 h
Median (10%-90% range): 25,900
(11,500-74,600)
Monitoring Stations: 1
Copollutant (correlation): PMio: r - -
0.13
0s: r - -0.13
NO2: r - 0.26
CO: r - 0.10
SO2: r - 0.24
PM2.6: r - -0.16
PM2.6 water soluble metals: r - -0.27
PM2.6 sulfates: r - -0.31;
PM2.6 acidity: r - -0.39;
PM2.6 organic carbon: r - 0.08;
PM2.6 elemental carbon: r - 0.08;
PM2.6 oxygenated hydrocarbon: r - 0.05
Other variables: Temperature: r - -0.33
Dew point: r - -0.41
PM Increment: 30,000 no/cm3
(approximately 1 SD|3
RR [95% CI]: For 3 day moving avg: All
CVD: 0.985 [0.965,1.005]
Dysrhythmia: 0.972 [0.937,1.008]
Congestive heart failure: 0.983 [0.943-
1.025]
Ischemic heart disease: 0.989 [0.953—
1.026]
Peripheral vascular and cerebrovascular
disease: 0.998 [0.960-1.039]
Results for Lags 0-7 expressed in figures
(see notes).
Notes: Figure 1: RR (95% CI) for single-
day lag models for the association of ER
visits for CVD with daily ambient UFP.
Summary of Figure 1 results: Null or
negative associations.
July 2009
E-144
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: von Klot et al. (2005,
088070)
Period of Study: 1992-2001
Location: Augsburg, Germany
Barcelona, Spain
Helsinki, Finland
Rome, Italy
Stockholm, Sweden
Outcome (ICD-9): Acute myocardial
infarction (410
ICD-10:121-I22), angina pectoris (411,
413
ICD-10:120,124), dysrhythmia (427
ICD-10:146.0, 46.9,147-I49, R00.1,
R00.8), heart failure (428
ICD-10: 150)
Age Groups: 35+ yrs
Study Design: Cohort
N: 22,006 Ml survivors
Statistical Analyses: GAM, Spearman
correlation
Covariates: Temperature, dew point
temp, avg barometric pressure, relative
humidity
Season: NR
Dose-response Investigated: No
Statistical Package: R-software with
"mgcv"package
Lags Considered: 0-3 days
'All units expressed in //gfm3 unless otherwise specified.
Pollutant: UFP (PNC)	PM Increment: 10,000 particles/cm3
Averaging Time: 24 h
Mean particle/cm3 (5th-95th percentile): Pooled RR Estimate [CI]:
Augsburg:	All cardiac admissions: 1.026
[1.005,1.048]
Barcelona:
Myocardial infarction: 1.039
Helsinki:	[0.998.1.082]
R°me:	Angina pectoris: 1.020 [0.992,1.048]
Stockholm:
Monitoring Stations: NR
Copollutant (correlation):
Augsburg
PMio: r - 0.52
CO: r - 0.63
NO?: r - 0.64
O3: r —-0.32Barcelona
PM10: r - 0.29
CO: r - 0.71;
NO?: r - 0.44
O3: r — -0.55
Helsinki
PM10: r - 0.46
CO: r - 0.47;
NO2: r - 0.83
0s: r —-0.16
Rome
PM10: r = 0.33
CO: r - 0.80;
NO2: r - 0.71
0s: r — -0.47
Stockholm
PM10: r - 0.06
CO: r - 0.54;
NO2: r - 0.80
0s: r — -0.17
July 2009
E-145
DRAFT-DO NOT CITE OR QUOTE

-------
E.2. Short-Term Exposure and Respiratory Outcomes
E.2.1. Respiratory Morbidity Studies
Table E-9. Short-term exposure-respiratory morbidity outcomes -PM10
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Aekplakorn, et al. (2003,
089908)
Period of Study: 107 days, from
October 1,1997 to January 15,1998
Location: Mae Mo district, Lampang
Province, North Thailand
Outcome: Upper respiratory symptoms,
lower respiratory symptoms, cough
Age Groups: 6-14 years old
Study Design: Logistic regression
l\l: 98 asthmatic school children, 98 non-
asthmatic school children
Statistical Analyses: GEE, stratified
analysis, PR0C GENM0D
Covariates: Temperature and relative
humidity
Season: winter
Dose-response Investigated? No
Statistical Package: SAS v 8.1
Pollutant: PMio
Averaging Time: daily
Mean (SD):
Sob Pad station: 31.92
Sob Mo station: 33.64
Hua Fai station: 37.45
Range (Min, Max):
Sob Pad: 6.63,153.25
Sob Mo: 4.23,121.80
Hua Fai: 6.98,113.30
Monitoring Stations: 3
Copollutant: PM2.6, SO2
PM Increment: 10 /yglm3
Odds Ratios [Lower CI, Upper CI]
lag:
Asthmatics: URS: 1.03 (0.99,1.07)
lag 0
LRS: 1.04 (0.99,1.09)
lag 0
Cough: 1.04 (1.00,1.07)
lag 0
Non-Asthmatics: URS: 1.04 (0.99,1.08)
lag 0
LRS: 1.0(0.93,1.07)
lag 0
Cough: 0.99 (0.94,1.05)
lag 0
PM10 + SO2
Asthmatics: URS: 1.03 (0.99,1.07)
lag 0
LRS: 1.03 (0.98,1.09)
lag 0
Cough: 1.04 (1.00,1.08)
lag 0
Non-Asthmatics: URS: 1.04 (0.99,1.08)
lag 0
LRS: 1.0(0.93,1.07)
lag 0
Cough: 0.99 (0.95,1.05)
lag 0
July 2009
E-146
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Andersen et al. (2008,
189651)
Outcome: Daily symptoms (prospective
daily recording of symptoms via diary)
Period of Study: Dec 12,1998-Dec 19, Age Groups: 0-3 yrs
2004
Location: Copenhagen
Study Design: Panel study of children
with genetic susceptibility to asthma
(mothers had asthma)
N: 205 children (living within a 15km
radius of the central monitor during the
first 3 yrs of life)
born between Aug 2,1998 and Dec 12,
2001
Statistical Analyses: logistic regression
model (GEE)
Covariates: temperature, season,
gender, age, exposure to smoking, and
paternal history of asthma
Effect modification: gender, medication
use, and paternal history of asthma
Statistical Package: SAS v9.1
Lag: 0,1,2,3,4,2-4
Pollutant: PMio
Mean: 25.1
SD: 16.7
Percentiles:
25th: 15.7
75th: 30.2
IQR: 14.5
Copollutant (correlation): PM2.6
(r - 0.79)
Number concentration of ultrafine
particles,
UFP (r - 0.37)
NO? (r - 0.43)
NOx (r - 0.40)
CO (r - 0.45)
Os (r - -0.32)
Temp (r - 0.25)
PM Increment: IQR (14.5 |Jg|m )
increase
Odds Ratios (95%CI) for incident
wheezing symptoms
Age 0-1
LO: 1.05(0.88,1.25)
L1: 1.00(0.82,1.22)
L2: 1.01 (0.83,1.23)
L3: 1.20(0.98,1.46)
L4: 1.23(1.02,1.48)
L2-4:1.21 (0.99,1.48)
Age 1-2
LO: 1.00(0.86,1.15)
L1: 1.02(0.87,1.19)
L2: 1.05(0.93,1.19)
L3: 0.96(0.84,1.09)
L4: 1.04 (0.90,1.21)
L2-4:1.03 (0.88,1.22)
Age 2-3
LO: 0.87(0.72,1.06)
L1: 0.95(0.78,1.15)
L2: 0.99(0.82,1.17)
L3: 1.03(0.84,1.25)
L4: 0.89(0.74,1.09)
L2-4: 0.94 (0.74,1.19)
Age 0-3
LO: 0.97(0.87,1.08)
L1: 0.99(0.89,1.10)
L2: 1.01 (0.92,1.12)
L3: 1.03(0.93,1.14)
L4: 1.04 (0.94,1.15)
L2-4:1.04 (0.92,1.17)
Two pollutant models (lag 2-4)
1 -pollutant model: 1.21 (0.99,1.48)
2-pollutant (adj for NO2): 1.13 (0.88,
1.45)
2-pollutant (adj for NOx): 1.16 (0.90,
1.48)
2-pollutant (adj for CO): 1.23 (0.96,1.57)
110 children living within 5km radius
from monitor (sensitivity analysis): Age 0-
1: 1.32(0.95,1.82)
Age 1-2: 1.20(0.87,1.67)
Age 2-3: 0.78(0.52,1.16)
Age 0-3: 1.11 (0.88,1.39)
Reference: Boezen et al. (2005,
087396)
Period of Study: Two consecutive
Outcome: FEVi, airway
hyperresponsiveness (AHR), serum total
IgE and daily data on lower respiratory
symptoms (LRS), upper respiratory
Pollutant: PM10
Averaging Time: 24 h
PM Increment: 10/yg/m3
Effect Estimate [Lower CI, Upper CI]:
AHR-|lgE-
July 2009
E-147
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
winters (winter 1993-winter 1995)
Location: rural (Meppel, Nunspeet) and
urban (Amsterdam) areas in the
Netherlands
symptoms (URS), cough and morning and
evening peak expiratory flow
Age Groups: 50-70 years
Study Design: Case-control study
N: 327 patients
Statistical Analyses: Logistic
regression
Covariates: daily minimum temperature,
linear, quadratic and cubic time trend,
weekend/holidays, and influenza
incidence for the rural and urban areas
and two winters separately
Season: winter
Dose-response Investigated? No
Lags Considered: 0,1, 2, and 5-day
mean
Mean (SD):
Winter 93194 Urban: 41.5
Winter 93194 Rural: 44.1
Winter 94195 Urban: 31.1
Winter 94195 Rural: 26.6
Percentiles: 50th(Median):
Winter 93194 Urban: 34.6
Winter 93194 Rural: 30.4
Winter 94195 Urban: 28.9
Winter 94195 Rural: 23.7
Range (Min, Max):
93194 Urban: (12.1-112.7)
93194 Rural: (7.9-242.2)
94195 Urban: (8.8-89.9)
94195 Rural: (7.1-96.9)
Copollutant:
SO?
NO?
Black Smoke
Upper Respiratory Symptoms
Lag 0: OR - 0.99 (0.97-1.01)
Lag 1: OR - 1.01 (0.99-1.03)
Lag 2: OR - 1.00 (0.96-1.02)
5-day mean: OR - 1.00 (0.96-1.04)
Cough
Lag 0: OR - 1.00(0.99-1.02)
Lag 1: OR - 0.99(0.98-1.01)
Lag 2: OR - 1.00(0.98-1.01)
5-day mean: OR - 0.98 (0.95-1.01)
>10% fall in morning peak expiratory
flow
Lag 1: OR - 1.01 (0.98-1.04)
Lag 2: OR - 0.97 (0.94-1.00)
5-day mean: OR - 0.97 (0.92-1.02)
AHR-/lgE +
Upper Respiratory Symptoms
Lag 0: OR - 1.01 (0.99-1.03)
Lag 1: OR - 1.02(1.00-1.04)
Lag 2: OR - 1.01 (0.99-1.03)
5-day mean: OR - 1.08 (1.04-1.11)
Cough
Lag 0: OR - 1.01 (0.99-1.03)
Lag 1: OR - 0.99(0.98-1.01)
Lag 2: OR - 1.00 (0.98-1.02)
5-day mean: OR - 1.01 (0.97-1.05)
>10% fall in morning peak expiratory
flow
Lag 1: OR - 0.99 (0.97-1.02)
Lag 2: OR - 0.99 (0.97-1.02)
5-day mean: OR - 0.97 (0.93-1.01)
AHR+|lgE-
Upper Respiratory Symptoms
Lag 0: OR - 0.99 (0.95-1.03)
Lag 1: OR - 1.01 (0.97-1.05)
Lag 2: OR - 0.99 (0.96-1.03)
5-day mean: OR - 0.98 (0.91-1.06)
Cough
Lag 0: OR - 1.00 (0.97-1.02)
Lag 1: OR - 1.01 (0.98-1.03)
Lag 2: OR - 0.99 (0.96-1.02)
5-day mean: OR - 1.02 (0.96-1.08)
>10% fall in morning peak expiratory
flow
Lag 1: OR - 0.99 (0.95-1.03)
Lag 2: OR - 0.99 (0.95-1.03)
July 2009
E-148
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
5-day mean: OR - 0.99 (0.93-1.06)
AHR+/lgE +
Upper Respiratory Symptoms
Lag 0: OR - 1.01 (0.98-1.04)
Lag 1: OR - 1.03(1.00-1.05)
Lag 2: OR - 1.02 (0.99-1.05)
5-day mean: OR - 1.06 (1.00-1.11)
Oough
Lag 0: OR - 1.03(1.01-1.06)
Lag 1: OR - 1.00(0.98-1.02)
Lag 2: OR - 0.99 (0.97-1.01)
5-day mean: OR - 0.99 (0.95-1.04)
Lag 2: OR - 0.99 (0.96-1.03)
5-day mean: OR - 0.99 (0.92-1.05)
>10% fall in morning peak expiratory
flow
Lag 1: OR - 1.04(1.00-1.07)
Lag 2: OR - 1.03 (0.99-1.06)
5-day mean: OR - 1.05 (0.99-1.11)
Increment: 100 /yglm3
Odds Ratio (Lower CI, Upper CI)
lag: OR for respiratory symptoms and
exposure to PMio in children with BHR
and high serum total IgE
Lower Respiratory Symptoms
1.32(1.07,1.63)
0
1.36(1.13,1.64)
1
1.36(1.13,1.65)
2
2.39(1.71,3.35)
0-5 avg.
Upper Respiratory Symptoms
1.13(0.97,1.32)
0
1.00(0.87,1.16)
1
0.96(0.84,1.11)
2
0.91 (0.70,1.18)
0-5 avg
>10% morning peak expiratory flow
(PEF) decrease
1.10(0.92,1.33)
0
1.08(0.90,1.28)
Reference: Boezen et al. (1999,
040410)
Periods of Study: 3 Winters (1992-
1995)
Location: Urban and rural areas of the
Netherlands
Outcome: Respiratory symptoms
Pollutant: PMio
Lower respiratory symptoms (wheeze,
Averaging Time: 24-h avg
attacks of wheezing, shortness of breath)
Mean (SD):
Upper respiratory symptoms (sore throat,
Winter 1992-93
runny or blocked nose)
Bronchial hyperresponsiveness (BHR)
Urban: 54.8
Study Design: Time-series
Rural: 44.7
Statistical Analyses: Logistic
Winter 1993-94
regression (PROC model)
Urban: 41.5 3
Age Groups: 7-11
Rural: 44.1

Winter 1994-95

Urban: 31.1

Rural: 26.6

Range (Min, Max):

Winter 1992-93

Urban: (4.7,145.6)

Rural: (4.8,103.8)

Winter 1993-94

Urban: (12.1,112.7)

Rural: (7.9, 242.2)

Winter 1994-95

Urban: (8.8, 89.9)

Rural: (7.1,96.9)

Copollutants:
BS
SO?
NO?
July 2009
E-149
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1
1.03(0.87,1.23)
2
1.10(0.83,1.46)
0-5 avg
> 10% evening peak expiratory flow
(PEF) increase
1.37 (1.16,1.63)
0
1.09(0.92,1.29)
1
1.16(0.98. 1.36)
2
1.35(1.04,1.77)
0-5 avg.
OR for respiratory symptoms and
exposure to PMio in children without BHR
and low serum total IgE
Lower Respiratory Symptoms
1.08(0.75,1.57)
0
1.04 (0.70,1.53)
1
0.98(0.69,1.39)
2
1.15(0.61,2.15)
0-5 avg.
Upper Respiratory Symptoms
1.12(0.99,1.28)
0
1.01 (0.89,1.15)
1
1.01 (0.89,1.15)
2
0.93(0.67,1.28)
0-5 avg
>10% morning PEF decrease
1.07 (0.93,1.23)
0
0.86 (0.75, 0.99)
1
0.97 (0.85,1.11)
2
0.99(0.79,1.23)
0-5 avg
>10% evening PEF decrease
July 2009
E-150
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1.13(0.98,1.30)
0
1.05(0.91,1.21)
1
0.99(0.87,1.14)
2
0.94 (0.75,1.17)
0-5 avg
OR for respiratory symptoms and
exposure to PMio in children with BHR
and low serum total IgE
Lower Respiratory Symptoms
0.77 (0.48,1.24)
0
1.34 (0.94,1.93)
1
1.24 (0.86,1.81)
2
1.92(0.84,4.41)
0-5 avg
Upper Respiratory Symptoms
1.13(0.92,1.40)
0
0.98(0.79,1.22)
1
0.97 (0.79,1.20)
2
0.83(0.54,1.25)
0-5 avg
>10% morning PEF decrease
1.04 (0.78,1.38)
0
0.86(0.66,1.12)
1
0.91 (0.71,1.17)
2
0.78(0.51,1.20)
0-5 avg
>10% evening PEF decrease
1.07 (0.82,1.41)
0
0.98(0.76,1.26)
1
0.93(0.73,1.19)
2
0.83(0.55,1.26)
July 2009
E-151
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0-5 avg
OR for respiratory symptoms and
exposure to PMio in children without BHR
and high serum total IgE
Lower Respiratory Symptoms
1.04 (0.80,1.35)
0
1.21 (0.98,1.51)
1
1.18(0.96,1.45)
2
1.35(0.89, 2.04)
0-5 avg
Upper Respiratory Symptoms
1.01 (0.85,1.20)
0
0.95(0.81,1.12)
1
0.93(0.80,1.09)
2
0.93(0.69,1.25)
0-5 avg
>10% morning PEF decrease
0.97 (0.80,1.17)
0
1.09(0.91,1.30)
1
1.02(0.85,1.21)
2
0.95(0.71,1.28)
0-5 avg
>10% evening PEF decrease
1.02(0.85,1.22)
0
1.06(0.90,1.25)
1
1.08(0.93,1.27)
2
1.04 (0.80,1.34)
0-5 avg.
Reference: Chattopadhyay et al. (2007,
Outcome: pulmonary function tests
Pollutant: PMio
PM Increment: NR
147471)
(respiratory impairments)
Averaging Time: 8 h
Respiratory impairments (SD): North

Period of Study: NR
Age Groups: All ages
Mean (SD):
Kolkata

Location: Three different points in
Study Design: Cross-sectional
North Kolkata: 535.9
Male (n - 137)
Kolkata, India: North, South, and Central
N: 505 people studied for PFT

Restrictive: 4 (2.92)


Central Kolkata: 1114.5

total population of Kolkata not given
South Kolkata: 909.2
Obstructive: 5 (3.64)

July 2009
E-152
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Statistical Analyses: Frequencies
Covariates: Meteorologic data (i.e.
temperature, wind direction, wind speed,
and humidity)
Dose-response Investigated? No
Monitoring Stations: 1
Copollutant:
PM < 10-3.3
PM< 3.3-0.4
Combined Res. And Obs.: 6 (4.37)
Total: 15(10.95)
Female (n - 152)
Restrictive: 3 (1.97)
Obstructive: 5 (3.28)
Combined Res. And Obs.: n/a
Total: 8 (5.26)
Total (n - 289)
Restrictive: 7 (2.42)
Obstructive: 10 (3.46)
Combined Res. And Obs.: 6 (2.07)
Total: 23 (7.96)
Central Kolkata
Male (n - 44)
Restrictive: 6 (13.63)
Obstructive: 1 (2.27)
Combined Res. And Obs.: 1 (2.27)
Total: 8(18.18)
Female (n - 50)
Restrictive: 3 (6.00)
Obstructive: 2 (4.00)
Combined Res. And Obs.: n/a
Total: 5(10.00)
Total (n - 94)
Restrictive: 9 (9.57)
Obstructive: 3 (3.19)
Combined Res. And Obs.: 1 (1.06)
Total: 13(13.82)
South Kolkata
Male (n - 52)
Restrictive: 1 (1.92)
Obstructive: 2 (3.84)
Combined Res. And Obs.: 3 (5.76)
Total: 6(11.53)
Female (n - 70)
Restrictive: 2 (2.85)
Obstructive: 1 (1.42)
Combined Res. And Obs.: n/a
Total: 3 (4.28)
Total (n - 122)
Restrictive: 3 (2.45)
Obstructive: 3 (2.45)
Combined Res. And Obs.: 3 (2.45)
Total: 9 (7.37)
July 2009
E-153
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Dales et al. (2006, 090744)
Period of Study: 1/1/1986-12/31/2000
Location: 11 Canadian Cities: Calgary,
Edmonton, Halifax, London, Hamilton,
Ottawa, St. John, Toronto, Vancouver,
Windsor, Winnipeg
Health Outcome: Respiratory Illness:
Asphyxia (799)
Respiratory failure (799.1)
Dyspnea and respiratory abnormalities
(786)
Respiratory distress syndrome (769)
Unspecified birth asphyxia in live-born
infant (768.9)
Other respiratory problems after birth
(770.8)
Pneumonia (486)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups: 0-27 days
Pollutant: PMio
Averaging Time: 24-h avg
Copollutants (correlation):
O3: r - -0.29 to 0.41
IMO2: r - -0.26 to 0.69
SO2: r - -0.09 to 0.61
CO: r--0.13 to 0.71
Increment: 10/yg/m3
% Increase (Lower CI, Upper CI)
Lag
In respiratory illness and exposure to
PM10 in neonates
PM10 alone: 2.13 (-0.50,4.76)
Multipollutant model
PM10: 1.45 (-1.90,4.80)
PM10, O3: 2.67 (0.98, 4.39)
PM10, NO?: 2.48 (1.18,3.80)
PM10, SO2:1.41 (0.35,2.47)
PM10, CO: 1.30 (0.13, 2.49)
Reference: de Hartog et al. (2003,
001061)
Period of Study: winter of 1998-1999
(in Amsterdam, from November 2,1998
to June 18,1999
in Erfurt, from October 12,1998 to April
4,1999
and in Helsinki, from November 2,1998
to April 30,1999.)
Location: Amsterdam, the Netherlands
Erfurt, Germany
and Helsinki, Finland
Outcome: Respiratory symptoms
Age Groups: a 50 yrs
Study Design: panel
N: 131 subjects with history of coronary
heart disease
Statistical Analyses: Logistic
regression
Covariates: ambient temperature,
relative humidity, atmospheric pressure,
incidence of influenza-like illness
Season: Winter
Dose-response Investigated? No
Statistical Package: S-PLUS 2000
Lags Considered: 0,1, 2, 3, and 5-day
avg
Pollutant: PM10
Averaging Time: 24 h
Mean (SD): Amsterdam, the Netherlands:
36.5
Erfurt, Germany: 27.1
Helsinki, Finland: 19.6
Range (Min, Max):
Amsterdam, the Netherlands: (13.6-
112.0)
Erfurt, Germany: (5.2-104.2)
Helsinki, Finland: (6.4-67.4)
Monitoring Stations: 1
Copollutant: PM2.6
NCo.010.1
CO
NO?
SO2
'There was a tendency toward positive
associations between avoidance of
activities and both particulate air
pollution (PM10) and gases, but none of
the associations were statistically
significant....In both incidence analyses
and prevalence analyses, odds ratios for
PM10 were generally similar to the
corresponding odds ratios for PM2.6, but
were somewhat less significant.'
July 2009
E-154
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Delfino et al. (1998, 051406) Outcome: asthma symptom severity
Period of Study: August 1 -October 30, Age Groups: 9-17
1995
Location: Alpine, CA
Study Design: Panel Study
N: 24 non smoking pediatric asthmatics
Statistical Analyses: GEE
Covariates: day of week, temperature,
humidity, wind speed
Statistical Package: SAS
Lags Considered: 0-5, 0, 0-4
Pollutant: PMio
Averaging Time: 24 h
Mean (SD):
31 (8)
90th: 42
Range (Min, Max): 16, 54
Copollutant (correlation):
0s (r - 0.32)
PM Increment: 42/yglm3 (90th
percentile increase)
Asthma symptoms:
Everyone: 1.47 (0.90, 2.39) lag 0
Everyone: 1.73 (1.03, 2.89) lag 0-4
Less symptomatic: 2.47 (1.23-4.95) lag 0
Less symptomatic: 4.03 (1.22,13.33) lag
0-4
More symptomatic: 1.50 (0.80, 2.80) lag
0
More symptomatic: 1.95 (1.12, 3.43) lag
0-4
PMio + 03
Asthma symptoms: 1.31 (0.84, 2.06) lag
0
1.65(1.03, 2.66) lag 0-4
Less symptomatic: 2.08 (1.12-3.83) lag 0
Less symptomatic: 3.35 (1.06,10.51) lag
0-4
More symptomatic: 1.40 (0.77, 2.53) lag
0
More symptomatic: 1.87 (1.11, 3.13) lag
0-4
July 2009
E-155
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Delfino et al. (2002, 093740) Outcome: Asthma symptoms that
interfere with daily activities
Period of Study: March 1 through April
30,1996	Age Groups: 9-19 yrs
Location: Alpine, California (a semi-rural Study Design: Daily panel study
area)
N: 22 asthmatic children
Statistical Analyses: GEE
Covariates: temperature, relative
humidity, day-of-week trends, linear time
trend across the 61 days, and upper or
lower respiratory infection
Season: "early spring season" of March
through April
Dose-response Investigated? Yes
Statistical Package: SAS, version 8
Lags Considered: 0,1, 2, 3,4, 5, 3-day
mov avg
Pollutant: PMio
Averaging Time: 1 h max
Mean (SD): 38(15)
Percentiles: 90th: 63
Range (Min, Max): (12-69)
Averaging Time: 8 h max
Mean (SD): 28(12)
Percentiles: 90th: 46
Range (Min, Max): (8-57)
Averaging Time: 24 h
Mean (SD): 20(9)
Percentiles: 90th: 32
Range (Min, Max): (7-42)
Copollutant (correlation): 1 h max
PMio: 8 h max PMio: r - 0.93
24 h PMio: r - 0.84
1 h max O3: r - 0.68
8 h max O3: r - 0.95
1 h max NO2: r - 0.49
8 h max NO2: r - 0.55
8 h max PMio: 1 h max PMio: r - 0.93
24 h PMio: r- 0.95
1 h max O3: r - 0.72
8 h max O3: r - 0.65
1 h max NO2: r - 0.48
8 h max NO2: r - 0.55
24 h PMio: 1 h max PMio: r - 0.84
8 h max PMio: r - 0.95
1 h max O3: r - 0.74
8 h max O3: r - 0.71
1 h max NO2: r - 0.37
8 h max NO2: r - 0.44
PM Increment: 90th percentile increase
Effect Estimate [Lower CI, Upper CI]:
ORs for risk of asthma symptoms in
those who report a respiratory infection
compared to those who do not have a
respiratory infection
1 h max PMio lag 0: 4.88 (1.31-18.2)
8 h max PMio lag 0: 6.78 (1.38-33.3)
24 h mean PMio lag 0: 4.68 (0.71-30.7)
3-day mov avg 1 h max PMio: 11.1 (1.10-
112)
3-day mov avg 8 h max PMio: 10.1 (1.42-
72.0)
3-day mov avg 24 h PMio: 2.67 (0.60-
11.8)
Effect modification by anti-inflammatory
medication use on the relationship of
asthma symptoms in children
1	h max PMio lag 0: 1.41 (0.87-2.30)
On medication: 0.96 (0.25-3.69)
Not on medication: 1.92 (1.22-3.02)
8 h max PMio lag 0: 1.19(0.74-1.94)
On medication: 0.75 (0.18-3.04)
Not on medication: 1.68 (0.91-3.09)
24 h mean PMio lag 0:1.08 (0.73-1.61)
On medication: 0.80 (0.24-2.69)
Not on medication: 1.35 (0.82-2.22)
3-day mov avg 1 h max PMio: 1.45 (0.76-
2.76)
On medication: 1.01 (0.14-7.02)
Not on medication: 1.92 (0.99-3.71)
3-day mov avg 8 h max PMio: 1.32 (0.76-
2.29)
On medication: 0.82 (0.17-3.94)
Not on medication: 1.89 (1.10-3.24)
3-day mov avg 24 h PMio: 1.22 (0.84-
1.77)
On medication: 0.75 (0.26-2.14)
Not on medication: 1.75(1.15-2.68)
Dose-response results are found in Figure
2	and not quantitatively reported
elsewhere.
July 2009
E-156
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Delfino et al. (2003, 090941)
Period of Study: November 1999 to
January 2000
Location: Huntington Park, Los Angeles
Outcome: Asthma severity scale
Peak Expiratory Flow Rate (PEF)
Age Groups: Ages 10 to 16
Study Design: Longitudinal study panel
N: 22 children
Statistical Analyses: Regression
analysis (GEE, GLM)
multivariate regression models
Covariates: Day of the week, Maximum
Temperature, Respiratory Infections
Season: Winter
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0,1
Pollutant: PMio
Mean (SD): 59.9 (24.7)
Range (Min, Max):
20-126
IQR: 37
90th: 86.0
Monitoring Stations: 1
Copollutant (correlation):
8-h max NO2 - 0.38
8-h max O3 - -0.16
8-h max CO - 0.50
8-h max SO2 - 0.73
PM Increment: IQR 37.0 /yglm3
OR Estimate [Lower CI, Upper CI]
lag:
Lag 0
Symptom Scores > 1: 1.45 (1.11,1.90)
Symptom Scores >2: NR
Lag 1
Symptom Scores >1:1.07 (0.64,1.77)
Symptom Scores >2: NR
Reference: Delfino et al. (2004, 056897)
Period of Study: September-October
1999
April-June 2000
Location: Alpine, California
Outcome: FEVi
Age Groups: 9-19 years old
Study Design: Panel study
N: 24 children
Statistical Analyses: GLM
Akaike's information criterion and
Bayesian information criterion
Covariates: Day of week, Personal
temperature and relative humidity, time
of FEVi maneuver (morning, afternoon, or
evening), Season (fall 1999 or spring
2000)
As-needed medication use
Presence or absence of upper or lower
respiratory infections
Season: Spring, Fall
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: Lag 0-4
Pollutant: PM10
Averaging Time: 4 h, 8 h, 12 h, 24-h
Personal Monitor
1-h max personal PM last
24-h
Mean (SD): 151.0(12.03)
90th: 292.4
Range (Min, Max): (9.1, 996.8)
Mean personal PM last 24-h
Mean (SD): 37.9(19.9)
90th: 65.1
Range (Min, Max):
(3.9,113.8)
Central outdoor stationary-site PM
1-h Maximum TE0M
PM10 last 24-h
Mean (SD): 54.4(13.8)
90th: 71.0
Range (Min, Max): (24.4, 95.4)
Mean TE0M PM10 last 24-h
Mean (SD): 29.7 (8.6)
90th: 40.9
Range (Min, Max): (12.9, 50.7)
24-h mean PM10
Mean (SD): 23.6 (9.1)
90th: 34.6
Range (Min, Max): (3.2, 48.0)
Copollutant (correlation): 8-h max
personal PM
8-h max O3 - 0.03
8-h Max NO2 - 0.26
24-h Mean Personal
Results presented graphically: Percent
predicted FEVi was inversely associated
with personal exposure to fine particles.
¦ Inverse associations of FEVi with
stationary-site indoor, outdoor and
central-site gravimetric PM2.6 and PM10,
and with hourly TE0M PM10
July 2009
E-157
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
PM - 0.94
8-h Max TEOM PMio - 0.38
24-h Mean TEOM PMio - 0.40
24-h Central HI PMio - 0.37
24-h Central HI PM2.1, - 0.38
24-h Outdoor HI PMio - 0.32
24-h Outdoor HI PM2 b - 0.39
24-h Indoor HI PMio - 0.23
24-h Indoor HI PM2 b - 0.37
24-h mean personal PM
8-h max O3 - 0.01
8-h Max NO2 - 0.27
8-h Max Personal PM - 0.94
8-h Max TEOM PMio - 0.36
24-h Mean TEOM PMio - 0.39
24-h Central HI PMio - 0.36
24-h Central HI PM2.1, - 0.43
24-h Outdoor HI PMio - 0.34
24-h Outdoor HI PM2 b - 0.44
24-h Indoor HI PMio - 0.29
24-h Indoor HI PM2 b - 0.46
24-h Mean TEOM PMio
8-h max O3 - 0.41
8-h Max NO2 - 0.58
8-h Max Personal PM - 0.40
24-h Mean Personal PM - 0.39
8-h Max TEOM PMio - 0.92
24-h Central HI PMio - 0.86
24-h Central HI PM2.1, - 0.78
24-h Outdoor HI PMio - 0.79
24-h Outdoor HI PM2 b - 0.78
24-h Indoor HI PMio - 0.36
24-h Indoor HI PM2 b - 0.59
July 2009
E-158
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Delfino et al. (2006, 090745)
Period of Study: Region 1: August to
Mid December 2003. Region 2: July
through November 2004
Location: Region 1: Riverside, CA.
Region 2: Whittier, CA
Outcome: Fractional Concentration of
Nitric Oxide in exhaled air (FENO)
Age Groups: 9 through 18
Study Design: Longitudinal Panel Study
N: 45 children
Statistical Analyses: Linear mixed-
effects models
Two-stage hierarchical model
Empirical Variograms
Fourth-order polynomial distributed lag
mixed-effects model
Covariates: Personal temperature,
Personal Rel. Humid., 10-day exposure
run, Respiratory infections, Region of
study, Sex, Cumulative daily use of as-
needed B-agonist inhalers
Dose-response Investigated? No
Lags Considered: Lag 0, Lag 1, 2-day
moving avg
Pollutant: PMio
Central Site
Averaging Time: 24- h
Riverside
Mean (SD): 70.82 (29.36) 50th(Median):
65.96
Range (Min, Max): (30.75,
154.05) //g/m3
Whittier
Mean (SD): 35.73 (16.6) 50th(Median):
34.65
Range (Min, Max): (5.86,105.46) //g|m3
Monitoring Stations: 48 personal
nephelometers, 2 central sites
PM Increment: IQR increase (Riverside:
28.41 /yglm3, Whittier 21.87 /yg/m3)
Coefficient [Lower CI, Upper CI]
lag: Lag - 2-day moving avg
Stratified by Medication Use
Not Taking Anti-lnflamm. Medication
Central 0.76 (-1.54, 3.07)
Taking Anti-lnflamm. Medication
Central 0.53 (-0.83,1.90)
Inhaled Corticosteroids
Central 1.28 (-0.01, 2.58)
Antileukotrienes +- inhaled
corticosteroids
Central-2.10 (-5.33,1.12)
Notes: Figure of Estimated lag effect of
hourly personal PM2.6 on FENO.
Figure of the Estimated lag effect of
hourly personal PM2.6 on FENO by use of
medications.
Figure of One- and two-pollutant models
for change in FENO using 2-day Moving
Averages personal and central-site
pollutant measurements.
July 2009
E-159
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Desqueyroux et al. (2002,
026052)
Period of Study: Nov 1995-Nov 1996
Location: Paris, France
Outcome: Asthma attacks
Age Groups: Adults.
Study Design: Panel study
N: 60 moderate to severe adult
asthmatics
Statistical Analyses: Marginal logistic
regression
Covariates: FEVi, smoking, allergy, oral
steroid treatment, mean daily
temperature, relative humidity, pollen
counts, season, holiday period
Season: winter, summer
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 1, 2, 3, 4, 5, 3-5
Pollutant: PMio
Averaging Time: 24 h
Mean (SD):
Summer: 23 (9)
Winter: 28 (14)
Range (Min, Max):
Summer: 6, 63
Winter: 9, 84
Monitoring Stations: 7
Copollutant: SO2, NO2, O3
PM Increment: 10 /yglm3
OR Estimate [Lower CI, Upper CI]
lag: 0.87 [0.71,1.06] lag 1
0.93(0.80,1.08] lag 2
1.11(0.98,1.26] lag 3
1.17(1.03,1.33] lag 4
1.16(1.01,1.34] lag 5
1.21 (1.01,1.34]lag 3-5
vs seasons alone: Winter: 1.41 (1.16,
1.71] lag 3-5
summer: 1.03 (0.72,1.47] lag 3-5
vs link to explanatory factors: No link:
(1.71 (1.20, 2.43] lag 3-5
Link: 1.27 (1.06,1.52] lag 3-5
vs occurrence of infection: Without
infection: 1.52 (1.16, 2.00] lag 3-5
With infection: 1.30 (1.03,1.65] lag 3-5
vs baseline pulmonary function: FEVi
>1-68% predicted: 1.38(1.06,1.79]
lag 3-5
FEV <68% predicted: 1.45(1.11,1.90]
lag 3-5
vs smoking habits: Nonsmokers: 1.53
(1.18,1.98] lag 3-5
Current & ex-smokers: 1.18(0.90,1.54]
lag 3-5
vs allergy: Non-allergic: 1.29 (0.94,1.77]
lag 3-5
Allergic: 1.49(1.17,1.90] lag 3-5
vs regular oral steroid treatment: No:
1.41 (1.15,1.73] lag 3-5
Yes: 1.41 (0.88, 2.25] lag 3-5
Multipollutant model: PM10 + NO2: 1.43
[1.16,1.76] Lag 3-5
PM10 + SO2:1.51 [1.20,1.90] Lag 3-5
PM10 + 03:1.09 [0.71,1.67] Lag 3-5
Reference: Diette et al. (2007,156399) Outcome: Asthma in the last 12 months Pollutant: PM10
(493.x)
Period of Study: 912001-1212003
Location: East Baltimore, MD
Age Groups: 2 to 6 years old
Study Design: Prospective cohort
N: 150 with asthma
150 without asthma
Statistical Analyses: Student's two-
tailed t-test
Kruskal-Wallis test
Pearson's chi square
Fisher's exact test
Covariates: Season of collection
Dose-response Investigated? No
Statistical Package: STATASE 8.0
Averaging Time: 72
50th(Median): 43.7
IQR: (29-70)
Notes: "Pollutant concentrations in the
homes of asthmatic and control children
who lived in the same home for their
whole life were not different compared
with those who had moved at least
July 2009
E-160
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ebelt et al. (2005, 056907)
Period of Study: summer of 1998
Location: Vancouver, Canada
Outcome: spirometry
Age Groups: range from 54-86 yrs
mean age - 74 years
Study Design: extended analysis of a
repeated-measures panel study
N: 16 persons with COPD
Statistical Analyses: Earlier analysis
expanded by developing mixed-effect
regression models and by evaluating
additional exposure indicators
Dose-response Investigated? No
Statistical Package: SAS V8
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): Ambient PMio: 17 (6)
Exposure to ambient PMio: 10.3 (4.6)
Range (Min, Max): Ambient PMio: (7-36)
Exposure to ambient PMio: (1.5-23.8)
Monitoring Stations: 5
Copollutant (correlation): Ambient
PM 10-2.5: r - 0.69
Ambient PM2.6: r - 0.78
Exposure to Ambient PMio: r - 0.71
PM Increment: Ambient PMio: 7 (IQR)
Exposure to ambient PMio: 6.5 (IQR)
Notes: Effect estimates are presented in
Figure 2 and Electronic Appendix Table 1
(only available with electronic version of
article) and not provided quantitatively
elsewhere.
Reference: Fischer et al. (2007,
156435)
Period of Study: 7 weeks (dates not
specified)
Location: Netherlands
Outcome: Respiratory Symptoms, Sore
throat, Runny nose, Cold, Sick at home
Study Design: Prospective cohort
N: 68
Statistical Analyses: Linear regression
model (PR0C mixed)
Age Groups: 10-11
Lag: 1-2
Statistical Package: SAS v 6.11
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): 56 /yglm3
IQ (25th, 75th):
(21,187)/yg/m3
Copollutants:
BS
NO?
CO
NO
Increment: 10/yg/m3
% Increase in eNO and PMio and change
in spirometric lung function
eNO and PMio only
6.5(0.9,12.4)
1
7.8(-11.3, 31.0)
2
FVC mean (SEM)
0.4 (0.5)
1
0.6(1.6)
2
FEVi mean (SEM)
¦0.3(0.5)
1
¦2.1 (1.9)
2
PEF mean (SEM)
¦2.8(3.3)
1
7.1 (12.0)
2
MMEF mean (SEM1-0.5 (1.7)
1
¦2.5(5.9)
2
July 2009
E-161
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Forsberg et al. (1998,	Outcome: Respiratory Symptoms,	Pollutant: PMio
051714)	Shortness of breath
Averaging Time: 24-h avg
Period of Study: 1/3/1994-3/27/1994 Wheeze, Asthma attacks, Recent asthma,
Dry cough, Doctor-diagnosed asthma,	Mean (SD):
Location: Urban and rural areas of Umea, Recently treated for asthma, Early chest	,, , 1Q/1 ,3
Sweden	j||ness	Urban: 13.4/yg/m
Study Design: Cohort panel	"ura''
Statistical Analyses: Logistic linear Range (Min, Max),
regression	Urban: (0, 40.5)/yg/m3
Age Groups: 6-12	Rura|. (1 6, 29.0)/yg/m3
Copollutants (correlation):
BS: r - 0.73
July 2009
E-162
Increment: 10/yg/m3
OR between prevalence of acute
respiratory symptoms and PM10 exposure
for urban and rural children
lag
Urban children - Cough: 1.031 (0.957,
1.112)
0
0.997 (0.923,1.077)
1
1.018(0.940,1.1031:2
1.094 (0.895,1.338)
0-6 avg
Phlegm: 0.998 (0.899,1.108)
0
1.035 (0.928,1.154)
1
1.121 (1.013,1.240)
2
1.043 (0.822,1.324)
0-6 avg
Upper respiratory symptoms: 1.004
(0.949,1.063)
0
0.975(0.922,1.031)
1
0.951 (0.895,1.010)
2
0.849 (0.687,1.050)
0-6 avg
Lower respiratory symptoms: 0.984
(0.872,1.110)
0
0.919(0.812,1.039)
1
0.894 (0.771,1.036)
2
0.800 (0.617,1.038)
0-6 avg
Rural children (control)
Cough: 0.997 (0.900,1.105)
0
1.003 (0.906,1.112)
1
0.997 (0.891,1.116)
2
0.855 (0.655,1.115)
0-6 avg
Phlegm: 1.024 (0.880,1.192)
° DRAFT-DO NOT CITE OR QUOTE
0.995 (0.853,1.160)
1
1.117(0.956,1.305)

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Goncalves et al. (2005,
089884)
Period of Study: Dec-Mar 1992/93.
Dec-Mar 1993194
Location: Sao Paulo
Outcome: Respiratory
morbidity/admissions
Age Groups: Children < 13 yrs
Study Design: Time series
Statistical Analyses: Principal
component analysis
Covariates: Daily mean temperature,
daily mean water vapor density, solar
radiation
Season: summer
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: Lag 3
Pollutant: PMio
Averaging Time: 24 h
Copollutant: SO2,03
PCA coefficients: PC1, PC2, PC3:
Summer 199211993: PM10: 0.69, 0.45,
0.13
Solar Radiation: -0.04, 0.94 to -0.12
Mean Temperature: 0.62, 0.44 to -0.47
Mean Water Vapor Density: 0.73 to -0.46
to-0.26
SO2: 0.78 to-0.03, 0.33
03:0.18,0.63,0.37
Respiratory Mortality: 0.05 to -0.02,
0.81
Variations explained by Principal
Component: PC 1: 0.29
PC2: 0.27
PC3: 0.17
Summer 1993/1994: PM10: 0.38, 0.80 to
¦0.23
Solar Radiation: 0.02, 0.09 to -0.97
Mean Temperature: 0.71, 0.40 to -0.37
Mean Water Vapor Density: 0.88, 0.25,
0.09
SO2: 0.01, 0.92, 0.00
O3: 0.47 to -0.06 to -0.35
Respiratory Mortality: -0.73, 0.11, 0.08
Variations explained by Principal
Component: PC 1: 0.31
PC2: 0.25
PC3: 0.18
Notes: Association between respiratory
morbidity and air pollution more likely
during summer with smaller contrasts in
synoptic weather condition (summer
1992/93) but respiratory morbidity more
related to weather variables during
summer with larger contrasts (summer
1993/94).
Reference: Gordian and Choudhury
(2003, 054842)
Period of Study: 1994-Dec 1996
Location: Anchorage, Alaska
Outcome: Asthma medication among
school children
Age Groups: Elementary school children
(kindergarten-6th grade)
Study Design: Time series
Statistical Analyses: Time series
regression model
Covariates: Day of the week, month,
time trend, temperature
Season: All seasons
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 1, 2, 7,14, 21, 28
Pollutant: PM10
Averaging Time: 24 h
Mean (SD): 36.11 (30.46)
Range (Min, Max): 2.96, 210.0
Monitoring Stations: 1
Model regression slope coefficient for
PM10 (estimated SE)
7.25(2.88)
lag 21
RR: 1.075 (1.016,1.138)
Notes: PM10 coefficients for other lags
were also statistically significant but not
reported.
July 2009
E-163
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Harre et al. (1997, 095726) Outcome: Respiratory symptoms, Cough, Pollutant: PMio
Wheeze, Chest tightness, Shortness of
breath, Change in sputum volume, Nose,
throat, or eye irritation, PEFR
Period of Study: 6/1994-8/1994
Location: Christchurch, New Zealand
Study Design: Prospective cohort
Statistical Analyses: Poisson, log linear
regression
Age Groups: >55
Averaging Time: 24-h avg
Copollutants:
CO
SO?
NO?
Increment: 35.04//g/m3
Relative Risk (Lower CI, Upper CI)
lag:
Chest symptoms: 1.38 (1.07,1.78)
1
Wheeze: 0.97 (0.75,1.26)
1
Nebulizer Use: 0.71 (0.42,1.18)
1
Inhaler Use: 0.94 (0.78,1.13)
1
Reference: Hastings and Jardine (2002,
030344)
Period of Study: 1997-1998
Location: Bosnia (US military camps)
Outcome: Weekly rates of upper
respiratory disease (URD), reported by
the medical treatment facility in each
military camp
Age Groups: US soldiers
Study Design: Ecologic (at level of
military camp)
N: 5 camps
Statistical Analyses: I.Pearson
correlations between weekly URD rates
and weekly PMio (avg and max)
2.Kruskal	Wallace test to compare URD
rates in the 4 exposure quartiles
3.	Mann Whitney test to compare
dichotomized exposure groups (above and
below 50th percentile)
Dose-response Investigated? Yes
Lags Considered: Weekly rates of URD
disease were related to avg weekly PM
levels in the same week
Pollutant: PMio
Mean (SD):
PMio avg: 75.5
PMio max: 92.9
Percentiles: PMio max:
25th: 58.57
50th: 74.55
75th: 107.56
PMio avg:
25th: 42.19
50th: 64.17
75th: 81.75
Range (Min, Max):
PMio avg: 25.0, 338.7
PMio max: 25.0, 338.7
Monitoring Stations: at least one in
each of the 5 camps
PM max Quartiles (combining all camps):
Q1: < 58.7 //g/m3
Q2: 60.1 to < 75.54//g/m3
Q3: 78.56 to < 107.56 //g/m3
04: >107.56//g/m3
For dichotomous analysis
cutoff - 74.55 //g/m3
PM avg Quartiles (combining all camps):
Q1: < 42.19 //g/m3
Q2: 42.19 to 64.17 //g/m3
Q3: 64.17to81.75/yg|m3
04: > 81.75 //g/m3
For dichotomous analysis
cutoff - 64.17 //g/m
Pearson correlation coefficients between
URD rate and PM category [p-value]:
PMio max: quartiles of PM*URD rates
All camps 0.203 [0.041]
Blue Factory camp 0.277 [0.095]
Comanche 0.165 [0.237]
Demi 0.639 [0.123]
McGovern 0.535 [0.177]
Tuzla Main 0.107 [0.327]
PMio max: dichotomous PM*URD rates:
All camps 0.283 [0.007]
Blue Factory camp 0.038 [0.430]
Comanche 0.282 [0.107]
Demi 0.927 [0.012]
McGovern 0.853 [0.033]
Tuzla Main 0.155 [0.258]
PMio avg: quartiles of PM*URD rates: All
camps 0.149 [0.101]
Blue Factory camp 0.301 [0.077]
Comanche 0.246 [0.141]
Demi 0.437 [0.231]
McGovern 0.853 [0.033]
Tuzla Main 0.182 [0.222]
July 2009
E-164
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
PMio avg: dichotomous PM*URD rates:
All camps 0.060 [0.305]
Blue Factory camp -0.075 [0.365]
Comanche 0.143 [0.268]
Demi N/A*
McGovern N/A*
Tuzla Main 0.123 [0.303]
Kruskal Wallace p-value comparing URD
rates across exposure quartiles: PMio
max
All camps 0.047
Blue Factory camp 0.321
Comanche 0.556
Demi 0.165
McGovern 0.202
Tuzla Main 0.554
PMio avg
All camps 0.672
Blue Factory camp 0.809
Comanche 0.658
Demi 0.564
McGovern 0.157
Tuzla Main 0.891
Mann-Whitney p-value comparing URD
rates between upper and lower 50th
percentile of PM: PMio max
All camps 0.034
Blue Factory camp 0.173
Comanche 0.314
Demi 0.083
McGovern 0.401
Tuzla Main 0.481
PMio avg
All camps 0.824
Blue Factory camp 0.682
Comanche 0.508
Demi N/A*
McGovern N/A*
Tuzla Main 0.656
Notes: * there were no days that fell in
the upper 50%ile for PM avg in these
camps
¦Rates of URD by PM quartiles for each
camp presented in figures. Authors state,
"Generally the avg URD rate increased
with quartile of maximum exposure...the
trend was not as clear for quartiles of
PMio avg exposure"
Reference: Hong et al. (2007, 091347) Outcome: Peak expiratory flow rate
(PEFR)
Period of Study: March 23-May 3, 2004
J\t|eJ3n>u|jsjj3rcno_^^
Pollutant: PMio
Averaging Time: 24 h
Effect Estimate: Regression coefficients
of morning and daily mean PEFR on PMio
and metal components using linear mixed-
July 2009
E-165
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Location: School on the Dukjeok Island
near Incheon City, Korea
	Design & Methods
age - 9.6 yrs)
Study Design: panel study
N: 43 schoolchildren
Statistical Analyses: Mixed linear
regression
Covariates: age, sex, height, weight,
asthma history, and passive smoking
exposure at home
Dose-response Investigated? No
Lags Considered: 0,1, 2, 3, 4, 5
Concentrations'
Mean (SD): 35.30 (23.48)
50th (Median): 29.36
Range (Min, Max):
(12.24-124.87)
PM Component:
Fe: mean - 0.208 (0.203)//g/m3
Median - 0.112
Range (Min, Max): (0.061-0.806)
Mn: mean - 0.008 (0.005) //g/m3
Median - 0.007
Range (Min, Max): (0.000-0.019)
Pb: mean - 0.051 (0.031) //g/m3
Median - 0.051
Range (Min, Max): (0.011-0.155)
Zn: mean - 0.021 (0.021) //g/m3
Median - 0.013
Range (Min, Max): (0.006-0.112)
Al: mean - 0.085 (0.100)//g/m3
Median - 0.031
Range (Min, Max): (0.017-0.344)
Copollutant: PM2.6
Effect Estimates (95% CI)
effects regression
Lag 1 (PMio)
Morning PEFR
Crude: B - -0.00, p - 0.99
Adjusted: B - -0.04, p - 0.37
Mean PEFR
Crude: B - 0.00, p - 0.93
Adjusted: B - -0.05, p - 0.12
Lag 1 (logFe)
Morning PEFR
Crude: B - -1.26, p - 0.31
Adjusted: B - -3.24, p - 0.13
Mean PEFR
Crude: B - -1.20, p - 0.20
Adjusted: B - -2.37, p - 0.15
Lag 1 (logMn)
Morning PEFR
Crude: B - -4.40, p < 0.01
Adjusted: B - -9.82, p < 0.01
Mean PEFR
Crude: B - -4.05, p < 0.01
Adjusted: B - -8.44, p < 0.01
Lag 1 (logPb)
Morning PEFR
Crude: B - -6.79, p < 0.01
Adjusted: B - -6.83, p < 0.01
Mean PEFR
Crude: B - -6.23, p < 0.01
Adjusted: B - -6.37, p < 0.01
Lag 1 (logZn)
Morning PEFR
Crude: B - -0.55, p - 0.71
Adjusted: B - -0.98, p - 0.59
Mean PEFR
Crude: B - 1.33, p - 0.24
Adjusted: B - 1.53, p - 0.28
Lag1 (logAI)
Morning PEFR
Crude: B - -0.58, p - 0.57
Adjusted: B - -2.22, p - 0.25
Mean PEFR
Crude: B - -0.59, p - 0.45
Adjusted: B - -1.48, p - 0.32
Regression coefficients of morning and
daily mean PEFR on metal components of
PMio and GSTM1 and GSTT1 genotype
July 2009	E-166	DRAFT - DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
using linear mixed-effects regression
Lag 1 (logPb)
Morning PEFR: B - -7.26, p < 0.01
Mean PEFR: B - -6.43, p< 0.01
GSTM1
Morning PEFR: B - 21.19, p - 0.23
Mean PEFR: B - 20.09, p - 0.25
Lag 1 (logMn)
Morning PEFR: B - -10.31, p < 0.01
Mean PEFR: B--8.66, p< 0.01
GSTM1
Morning PEFR: B - 21.02, p - 0.23
Mean PEFR: B - 19.84, p - 0.25
Lag 1 (logPb)
Morning PEFR: B - -7.26, p < 0.01
Mean PEFR: B--6.43, p< 0.01
GSTT1
Morning PEFR: B - 2.07, p - 0.90
Mean PEFR: B--2.39, p< 0.88
Lag 1 (logMn]
Morning PEFR: B - -10.32, p < 0.01
Mean PEFR: B--8.67, p< 0.01
GSTT1
Morning PEFR: B - 2.02, p - 0.90
Mean PEFR: B - 2.33, p - 0.88
July 2009
E-167
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Hwang et al. (2006, 088971) Outcome: Allergic rhinitis
Period of Study: 2001	Study Design: Cross-sectional
Location: Taiwan	Statistical Analyses: Two-stage
hierarchical models
Age Groups: 6-15
Pollutant: PMio
Averaging Time: 1-h avg
Mean (SD): 55.58 (16.57)
Range (Min, Max):
(29.36, 99.58)
Copollutants (correlation):
CO: r - 0.27
NOx: r - 0.34
0s: r - 0.28
SO?: r - 0.58
Increment: 10/yg/m3
Odds Ratio (Lower CI, Upper CI)
lag:
PM10 alone: 1.00(0.99,1.02)
NOx, PM10: 0.99 (0.97,1.00)
CO, PM10:1.00 (0.99,1.01)
03, PM10:1.00 (0.99,1.02)
Gender
Male: 1.02(0.99,1.04)
Female: 0.99(0.97,1.02)
Parental atopy*
Yes: 1.00 (0.98,1.03)
No: 1.01 (0.99,1.03)
Parental education
< 6 years: 1.05 (0.96,1.14)
6-8 years: 1.03(0.98,1.07)
9-11 years: 1.00(0.98,1.03)
12+years: 0.99(0.97,1.02)
Environmental tobacco smoke
Yes: 1.01 (0.99,1.03)
No: 1.00 (0.98,1.03)
Visible mold**
Yes: 1.02 (0.99,1.06)
No: 1.00 (0.98,1.02)
* Parental atopy was a measure of
genetic predisposition and was defined as
the father or the mother of the index child
ever having been diagnosed as having
asthma, allergic rhinitis, or atopic
eczema.
** Visible mold found in the home.
Reference: Jalaludin et al. (2004,
Outcome: Respiratory symptoms,
Pollutant: PM10
Increment: 10/yg/m3
056595)
Wheeze, Dry cough, Wet cough



Averaging Time: 24-h avg
Odds Ratio (Lower CI, Upper CI)
Period of Study: 2/1/1994-12/31/1994
Study Design: Longitudinal study pani
;l
Mean (SD): 22.8(13.8)
Lag


Location: Western and southwestern
Statistical Analyses: Logistic



Sydney, Australia
regression model (GEE)
IQ Range (25th,75th): (12.00,122.8)
Wheeze


Age Groups: 9-11
Copollutants (correlation): O3: r — 0.13
1.01 (0.99,1
1.03)


NO?: r - 0.26
0




1.01 (0.97,1
1
1.04)



1
0.99(0.96,1
9
1.03)



1.02(0.98,1
1.06)



0-2 avg




1.04 (0.99,1
1.10)



0-5 avg




Dry Cough




1.00(0.98,1
1.03)
July 2009
E-168
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0
1.00(0.97,1.03)
1
1.00(0.97,1.02)
2
1.00(0.97,1.03)
0-2 avg
1.03(0.98,1.08)
0-5 avg
Wet Cough
1.01 (0.99,1.04)
0
0.99(0.97,1.01)
1
1.00(0.97,1.03)
2
0.99(0.96,1.02)
0-2 avg
0.99(0.94,1.04)
0-5 avg
Inhaled B2-agonist Use
0.99(0.98,1.01)
0
1.00(0.98,1.03)
1
0.99(0.97,1.01)
2
1.00(0.97,1.02)
0-2 avg
1.02(0.98,1.06)
0-5 avg
Inhaled Corticosteroid Use
1.00(0.99,1.01)
0
1.00(0.99,1.02)
1
1.00(0.99,1.02)
2
1.00(0.98,1.02)
0-2 avg
1.00(0.97,1.02)
0-5 avg
Doctor Visit for Asthma
1.11 (1.04,1.19)
0
July 2009
E-169
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1.10(1.02,1.19)
1
1.15(1.06,1.24)
2
1.11 (1.03,1.20)
0-2 avg
1.14(0.98,1.31)
0-5 avg
OR for respiratory symptoms and
PMio exposure by different groups
All children
Wheeze: 1.01 (0.99,1.04)
Dry Cough: 1.00 (0.97,1.02)
Wet Cough: 1.01 (0.98,1.04)
Inhaled B2 agonist Use: 1.00 (0.98,1.02)
Inhaled Corticosteroid Use: 0.99 (0.98,
1.01)
Doctor Visit for asthma: 1.11 (1.03,
1.19)
Group 1*
Wheeze: 1.01 (0.98,1.04)
Dry Cough: 0.97 (0.94, 0.99)
Wet Cough: 1.00 (0.97,1.03)
Inhaled B2-agonist use: 1.00 (0.98,1.02)
Inhaled Corticosteroid Use: 1.00 (0.98,
1.01)
Doctor Visit for asthma: 1.09 (0.98,
1.21)
Group 2**
Wheeze: 1.01 (0.97,1.05)
Dry Cough: 1.02 (0.98,1.06)
Wet Cough: 1.01 (0.96,1.06)
Inhaled B2 agonist use: 0.99 (0.94,1.05)
Inhaled Corticosteroid Use: 0.99 (0.97,
1.01)
Doctor Visit for asthma: 1.12 (1.02,
1.23)
Group 3***
Wheeze: 1.08 (0.90,1.31)
Dry Cough: 1.01 (0.91,1.11)
Wet Cough: 1.02 (0.94,1.11)
Inhaled B2-agonist use: 0.98 (0.84,1.11)
Inhaled Corticosteroid Use: 1.27 (1.08,
1.49)
Doctor Visit for asthma: NR
Group 1 consists of children with a
history of wheeze in the past 12 months,
positive histamine challenge, and doctor
diagnosed asthma.
J^Gjjou£_2_cons^sts_o^childrenj/vjth_a__
July 2009
E-170
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
history of wheeze in the past 12 months
and doctor diagnosed asthma.
""Group 3 consists of children only
with a history y of wheeze in the past 12
months.
Reference: Jansen, et al. (2005,
082236)
Period of Study: 1987-2000
Location: Seattle, WA
Outcome: FEN0: fractional exhaled
nitrogen oxide, Spirometry, Blood
pressure, Sa02: oxygen saturation, Pulse
rate
Age Groups: 60-86-years-old
Study Design: short-term cross-sectional
case series
N: 16 subjects diagnosed with C0PD,
asthma, or both
Statistical Analyses: linear mixed
effects model with random intercepts
Covariates: age, relative humidity,
temperature, medication use
Season: winter 2002-2003
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PMio
Averaging Time: 24-h
Mean (SD): Fixed-site Monitor: 18.0
All Subjects (N - 16)
Indoor, home: 11.93
Outdoor, home: 13.47
Personal: 23.34
Asthmatic Subjects (N - 7)
Indoor, home: 12.54
Outdoor, home: 11.86
Personal: 26.88
C0PD Subjects (N - 9)
Indoor, home: 11.45
Outdoor, home: 14.76
Personal: 19.91
Range (Min, Max): Fixed-site Monitor
2.5, 51
IQR: All Subjects
Indoor, home: 6.93
Outdoor, home: 9.53
Personal: 20.72
Asthmatic Subjects
Indoor, home: 10.19
Outdoor, home: 8.77
Personal: 20.08
C0PD Subjects
Indoor, home: 4.56
Outdoor, home: 6.14
Personal: 19.94
PM Increment: 10 (xglm3
Slope [95% CI]: dependence of FEN0
concentration [ppb] on PMio
Asthmatic Subjects
Indoor, home: 3.81 [-0.86: 8.50]
Outdoor, home: 5.87 [2.87: 8.88]*
Personal: 0.66 [-0.56: 1.88]
C0PD Subjects
Indoor, home: 2.19 [-3.48: 7.87]
Outdoor, home: 4.45 [-1.11:10.01]
Personal: 0.17 [-1.61: 1.96]
Results indicate that FENO may be a
more sensitive biomarker of PM exposure
than other traditional health endpoints.
Reference: Johnston, et al. (2006,
091386)
Period of Study: 7 months. (April 7
through November 7, 2004)
Location: Darwin, Australia
Outcome: Asthma symptoms
Age Groups: all ages
Study Design: Time-series
N: 251 people (130 adults, 121 children
Statistical Analyses: Logistic
regression model
Covariates: minimum air temperature,
doctor visits for influenza and the
prevalence of asthma symptoms and, the
fungal spore count and both onset of
asthma symptoms and commencement of
reliever medication
Season: "dry season"-specific months
NR, note Southern Hemisphere
Dose-response Investigated? No
Statistical Package: STATA8
Pollutant: PMio
Averaging Time: daily
Mean (SD): 20 (6.4)
Range (Min, Max): 2.6-43.3
PM Component: Vegetation fire smoke
(95%) and motor vehicle emissions (5%)
Monitoring Stations: 1
Correlation: PM 2.6 r - 0.90
PM Increment: 10/yg/m3
RR Estimate [Lower CI, Upper CI]
Symptoms attributable to asthma
Overall-1.010 (0.98,1.04)
Adults-1.027 (0.987,1.068)
Children-0.930 (0.966,1.060)
Using preventer- 1.022 (0.985,1.060)
Became symptomatic
Overall- 1.240(1.106,1.39)
Adults-1.277 (1.084,1.504)
Children- 1.247 (1.058,1.468)
Using preventer—1.317 (1.124,1.543)
Used Reliever
July 2009
E-171
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Lags Considered: 0-5 days
Overall-1.010 (0.99,1.04)
Adults-1.026(0.990,1.063)
Children- 1.006 (0.960,1.055)
Using preventer—1.035 (1.004,1.060)
Commenced Reliever
Overall- 1.132(0.99,1.29)
Adults-1.199(0.994,1.446)
Children- 1.093 (0.906,1.319)
Using preventer—1.194 (0.996,1.432)
Commenced Oral Steroids
Overall- 1.540(1.01,2.34)
Adults-1.752 (1.008, 3.045)
Children- 1.292 (0.682, 2.448)
Using preventer—1.430 (0.888, 2.304)
Asthma Attack
Overall- 1.030(0.95,1.12)
Adults-1.08 (0.976,1.202)
Children- 0.861 (0.710,1.044)
Using preventer—1.051 (0.939,1.175)
Exercise induced asthma
Overall- 0.980(0.92,1.05)
Adults-0.988 (0.902,1.081)
Children- 0.972(0.844,1.119)
Using preventer—1.026 (0.928,1.134)
Saw a health professional for asthma
Overall- 1.030(0.85,1.26)
Adults-1.064 (0.794,1.424)
Children- 0.998 (0.749,1.328)
Using preventer-0.924 (0.731,1.169)
Missed school or work due to asthma
Overall- 1.102(0.941,1.290)
Adults-1.135(0.897,1.435)
Children- 1.073(0.862,1.333)
Using preventer—1.025 (0.857,1.228)
Mean daily number of asthma symptoms
Overall- 1.020(1.001,1.031)
Adults-1.027 (1.005,1.049)
Children- 1.016(0.986,1.047)
Using preventer—1.034 (1.011,1.058)
Mean Daily number of applications of
reliever
Overall- 1.020(1.00,1.030)
Adults-1.032(1.008,1.057)
Children- 1.002 (0.969,1.034)
Using preventer—1.022 (1.001,1.043)
July 2009
E-172
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Just et al. (2002, 035429)
Period of Study: 4/1/1996-6/30/1996
Location: Paris, France
Outcome: Incident and prevalent
episodes of asthma attacks, nocturnal
cough, wheeze, symptoms of irritation,
respiratory infections, supplementary use
of (32-agonists, Ztransformed peak
expiratory flow (PEF), daily PEF variability
Age Groups: 7-15 years old
Study Design: Cohort
N: 82 children
Statistical Analyses: Linear regression,
logistic regression, GEE
Covariates: Effects of time trend, day of
the week, weather, pollen levels
Season: Spring/summer
Lags Considered: 0, 0-2 mean, 0-4 mean
Pollutant: PMio
Averaging Time: Daily
Mean (SD): 23.5 (8.4)
Range (Min, Max): 9.0, 44.0
Monitoring Stations: 5
Copollutant (correlation):
BS: 0.59
SO?: 0.70
NO?: 0.54
O3: 0.21
temp: 0.04
humid: -0.41
PM Increment: 10/yg/m3 for binary
responses data (results that use odds
ratios [ORs])
Incident episodes of
1)	Asthma
a)	lag 0:1.06 (0.61,1.83)
b)	0-2 mean: 1.09 (0.48, 2.49)
c)	0-4 mean: 1.07 (0.44, 2.65)
2)	Nocturnal cough
a)	lag 0:1.10 (0.88,1.37)
b)	0-2 mean: 1.03 (0.77,1.37)
c)	0-4 mean: 1.11 (0.86,1.42)
3)	Respiratory infections
a)	lag 0:0.64 (0.35,1.15)
b)	0-2 mean: 0.74 (0.38,1.43)
c)	0-4 mean: 0.99 (0.58,1.68)
Prevalent episodes of
1)	Asthma
a)	lag 0:1.07 (0.72,1.59)
b)	0-2 mean: 1.18(0.64, 2.17)
c)	0-4 mean: 1.16(0.63, 2.13)
2)	Nocturnal cough
a)	lag 0:1.05 (0.83,1.34)
b)	0-2 mean: 1.10 (0.81,1.50)
c)	0-4 mean: 1.09(0.79,1.52)
3)	Respiratory infections
a)	lag 0:1.17 (0.68, 2.03)
b)	0-2 mean: 1.31 (0.51,3.36)
c)	0-4 mean: 1.71 (0.71,4.12)
4)	Eye irritation
a)	lag 0:1.18 (1.01,1.39)
b)	0-2 mean: 1.28 (1.03,1.59)
c)	0-4 mean: 1.42(1.12,1.80)
Analysis restricted to days with no
steroid use
Incident episodes of
1)	Eye irritation
a)	lag 0:1.07 (0.66,1.71)
b)	0-2 mean: 0.83 (0.45,1.53)
c)	0-4 mean: 0.92(0.46,1.83)
2)	Throat irritation
a)	lag 0:1.33 (0.66, 2.69)
b)	0-2 mean: 1.28 (0.58, 2.80)
c)	0-4 mean: 1.06 (0.38, 2.95)
3)	Nose irritation
a) lag 0:0.74 (0.48,1.13)
July 2009
E-173
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
b)	0-2 mean: 0.76(0.42,1.36)
c)	0-4 mean: 0.96 (0.53,1.73)
Prevalent episodes of
1)	Eye irritation
a)	lag 0:1.20 (0.88,1.65)
b)	0-2 mean: 1.71 (0.97,3.01)
c)	0-4 mean: 1.97(1.03, 3.76)
2)	Throat irritation
a)	lag 0:1.23 (0.83,1.82)
b)	0-2 mean: 1.08 (0.68,1.73)
c)	0-4 mean: 0.91 (0.47,1.73)
3)	Nose irritation
a)	lag 0:1.20 (0.91,1.58)
b)	0-2 mean: 1.09 (0.78,1.52)
c)	0-4 mean: 1.09(0.73,1.61)
Notes: The authors noted that incident or
prevalent wheeze was not correlated
with levels of any type of pollutant
also, they state no relationship was
observed between PEF variables and
levels of PM.
The authors also note that in a
multipollutant model assessing
independent effects of PM and 03 on
prevalent episodes of eye irritation (mean
0-4), the PM parameter decreased and
was not significant (p - 0.19).
PM Increment: 1.0 /yglm3
% Change [Lower CI, Upper CI]:
Single pollutant model:
FEVi: -4.3 [-8.5, 0.2] p - 0.04
R2 - 0.06
Single pollutant model:
FVC:-1.2 [-5.6,3.2] p - 0.59
R2 - 0.005
Single pollutant model:
FEF2B-7B:-8.6 [-17.3, 0.1] p - 0.05
R2 - 0.06
2 pollutant model with Macrophage
Carbon:
FEVi: PM10 -2.9 [-6.9,1.2]
p - 0.17
(FVC): PM10 0.1 [-4.4, 4.6]
p - 0.96
FEF2b-7b: PM10 -5.5 [-14.2, 3.1]
p - 0.21
Reference: Kulkarni et al (2006,
089257)
Period of Study: 11/2002-12/2003
Location: Leicester, United Kingdom
Outcome: Lung function by spirometry:
FVC, FEVi, FEVi: FVC, FEF25-75
Age Groups: 8-15
Study Design: Cross-sectional
N: 114 children, 64 provided sputum for
assessment of carbon content of
macrophages.
Statistical Analyses: Linear
regressions, Spearman rank correlations.
Mann-Whitney, Chi-square and unpaired t
tests were used to compare results
between asthmatic and non asthmatic
children
Covariates: BMI, sex, exercise, traffic
PM10
Dose-response Investigated? Yes
Statistical Package: SPSS
Pollutant: Primary PM10 (/yglm3)
concentration was modeled, and was
considered a covariate for carbon content
of macrophages. Carbon content of
alveolar macrophages was the primary
variable of interest.
Averaging Time: 1 Yr
50th(Median): Children without asthma,
1.21
Children with asthma, 1.81
Range (Min, Max): Children without
asthma, 0.10, 2.17
Children with asthma, 0.17, 2.13
PM Component: Carbon content in
alveolar macrophages
Monitoring Stations: NR.
Copollutant (correlation): vs carbon
content in macrophages (increment,
coefficient [range])—1.0/yglm3, 0.1 [0.01 ¦
0.18]
July 2009
E-174
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kuo, et al. (2002, 036310)
Period of Study: 1 -yr period (year not
specified)
Location: Central Taiwan
Outcome: Asthma (yes/no)
Age Groups: 13-16 years
Study Design: Cohort
N: 12926 total children
775 asthmatic children
8 junior high schools
Statistical Analyses: Pearson
correlation coefficients
Logistic regression
Pollutant: PMio
Averaging Time: 1-h
Mean (SD):
School A: 59.7
School B: 65.3
School C: 84.3
School D: 59.2
School E: 75.3
School F: 60.2
Covariates: Gender, age, residential
area, level of parental education, number School G: 54.1
cigarettes smoked by family members,
incense burning in the home, frequency of
physical activities
School H: 69.0
Dose-response Investigated? No
Statistical Package: SAS 6.12
Lags Considered: Monthly averages at
each school
Monitoring Stations: 8 (1 for each
school)
PM Increment: Dichotomized annual
avg:
<	65.9 /yg/m3
a 65.9/yg/m3
OR Estimate [Lower CI, Upper CI]
lag:
Crude (outcome - asthma, yes/no)
<	65.9 /yg/m3: 1 (ref)
a 65.9/yg/m3: 0.837 [NR]
Adjusted (outcome - asthma, yes/no)
<	65.9 /yg/m3: 1 (ref)
a 65.9/yg/m3: 0.947 [0.640,1.401]
Notes: asthma prevalence was highest in
urban areas and lowest in rural areas
Pearson correlation between annual PM
levels at each school and asthma
prevalence at each school: 0.214
[p > 0.05]
Reference: Lagorio et al. (2006,
089800)
Period of Study: 5/24/1999 to
6/24/1999 and 11/181999 to
12/22/1999
Location: Rome, Italy
Outcome: Lung function of subjects (FVC
and FEVi) with COPD, Asthma
Age Groups: COPD 50 to 80 yrs
Asthma 18 to 64 yrs
Study Design: Time series panel
N: COPD N - 11
Asthma N - 11
Statistical Analyses: Non-parametric
Spearman correlation
GEE
Covariates: COPD and IHD: daily mean
temperature, season variable (spring or
winter), relative humidity, day of week
Asthma: season variable, temperature,
humidity, and (3-2-agonist use
Season: Spring and winter
Dose-response Investigated? Yes
Statistical Package: STATA
Lags Considered: 1-3 days
Pollutant: PMio
Averaging Time: 24-h
Mean (SD):
Overall: 42.8 (21.8)
Spring: 36.9 (10.8)
Winter: 49.0(28.1)
Range (Min, Max): (7.9,123)
PM Component: NR
Monitoring Stations: Two fixed sites:
(Villa Ada and Istituto superior di Sanita)
Copollutant (correlation): NO2 r - 0.45
O3 r = -0.36
COr - 0.55
SO? r - 0.21
PMio-2.5 r - 0.61
PM2.6 r - 0.93
PM Increment: 1 /yg/m3
They observed negative association
between ambient PMio and respiratory
function (FVC and FEVi) in the COPD
panel. The effect on FVC was seen at lag
24 h, 48 h, and 72 h. The effect on FEVi
was evident at lag 72 h. There was no
statistically significant effect of PMio on
FVC and FEVi in the asthmatic and IHD
(3 Coefficient (SE)
COPD
FVC(%) 24 h-0.66 (0.30)
48-h-0.75 (0.35)
72-h -0.94 (0.47)
FEVi(%) 24 h -0.37 (0.27)
48-h-0.58 (0.31)
72-h -0.87 (0.43)
Asthma
FVC(%) 24 h-0.12 (0.24)
48-h -0.09 (0.29)
72-h -0.08 (0.36)
FEVi(%) 24 h-0.28 (0.28)
48-h -0.40 (0.34)
72-h -0.40 (0.43)
July 2009
E-175
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lee, et al. (2007, 093042)
Period of Study: 2000-2001
Location: South-Western Seoul
Metropolitan area, Seoul, South Korea
Reference: Lewis, et al (2005, 081079)
Period of Study: winter 2001 -spring
2002
Location: Detroit, Michigan, USA
Outcome: PEFR (peak expiratory flow
rate), lower respiratory symptoms (cold,
cough, wheeze)
Age Groups: 61-89 years of age (77.8
mean age)
Study Design: longitudinal panel survey
N: 61 adults
Statistical Analyses: Logistic
regression model
Covariates: Temperature (Celsius),
relative humidity, age, season
Dose-response Investigated? No
Statistical Package: SAS 8.0
Lags Considered: 0-4 days
Outcome: Poorer lung function (increased
diurnal variability and decreased forced
expiratory volume)
Age Groups: 7-11 years old
N: 86 children
Statistical Analyses: descriptive
statistics and bivariate analyses of
exposures, multivariable regression
models that included interaction terms
between exposure measures and CS use
or, alternatively, presence of a URI,
multivariate analog of linear regression.
Covariates: sex, home location, annual
family income, presence of one or more
smokers in household, race, season
(entered as dummy variables), and
parameters to account for intervention
group effect.
Season: Winter 2001 (February 10-23),
spring 2001 (May 5-18), summer 2001
(July 14-27), fall 2001 (September 22-
October 5), winter 2002 (January 18-
31), and spring 2002 (May 18-31).
Dose-response Investigated? No
Lags Considered: 1-2 days
3-5 days
Pollutant: PMio
Averaging Time: 24-h
Mean (SD): 71.40 (30.69)
Percentiles: 25th: 43.47
50th(Median): 74.92
75th: 87.54
Range (Min, Max):
26.23,148.34
Monitoring Stations: 2
Pollutant: PMio
Averaging Time: 2 weeks
Mean (SD): Eastside 23.0 (13.5)
Range (Min, Max): 2.9, 70.9
PM Component: ("likely" in southwest
site) carbon and diesel emissions
Monitoring Stations: 2
Copollutant:
PM2.6 0.93
O3 Daily mean 0.59
O3 8-h peak 0.57
PM Increment: 10 /yglm3
Effect Estimate [Lower CI, Upper CI]
lag:
PEFR (peak expiratory flow rate)
¦0.39 (-0.63 to-0.14)
1 day
relative odds of a lower respiratory
symptom (cold, cough, wheeze)
1.015(0.900,1.144)
1 day
PM Increment: 19.1 /yglm3
Lung function among children
reporting use of maintenance CSs
Diurnal variability FEVi
Lag 1:1.53 [-0.85, 3.90]
Lag 1:2.94 [-1.07, 6.96] PMio + 0s
Lag 2: 5.32 [0.32,10.33]
Lag 2:13.73(8.23,19.23] PMio + 0s
Lag 3-5: 1.46 [-2.21,5.13]
Lag 3-5: 3.30 [0.58,6.02] PMio + 0s
Lowest daily value FEVi
Lag 1: -0.28 [-2.34,1.77]
Lag 1: -6.25 [-11.15 to-1.36] PMio + 0s
Lag 2:-2.21 [-3.97 to -0.46]
Lag 2: -5.97 [-11.06 to -0.87] PMio + 0s
Lag 3-5: -2.58 [-7.65, 2.49]
Lag 3-5: 1.98[-0.38, 4.33] PMio + 0s
Lung function among children
reporting presence of URI on day of
lung function assessment
Diurnal variability FEVi
Lag 1: 3.51 [-4.52,11.55]
Lag 1: 3.21 [-1.28,7.71] PMio + 0s
Lag 2: 1.12 [-4.62, 6.86]
Lag 2: 5.40 [-0.82,11.62] PMio + 0s
Lag 3-5: 3.90 [0.34, 7.47]
Lag 3-5: 6.27 [0.07,12.47] PMio + 0s
Lowest daily value FEVi
Lag 1: -2.72 [-9.47, 4.03]
Lag 1: -13.11 [-21.59 to-4.62] PMio +
0s
Lag 2: 0.24 [-5.10, 4.63]
Lag 2: -3.32 [-6.83, 0.18] PMio + 0s
Lag 3-5: -4.48 [-8.36, 0.60]
Lag 3-5: -3.17 [-5.82 to-0.51] PM10 + 0s
Study Design: longitudinal cohort study Southwest 28.2 (16.1)
July 2009
E-176
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Maret al. (2004, 057309)
Period of Study: 1997-1999
Location: Spokane, Washington
Outcome: Respiratory symptoms
Age Groups: Adults: Ages 20-51 yrs
Children: Ages 7-12 yrs
Study Design: Time-series
N: 25 people
Statistical Analyses: Logistic
regression
Covariates: Temperature, relative
humidity, day of-the-wk
Statistical Package: STATA 6
Lags Considered: 0-2 days
Pollutant: PMio
Mean (SD):
1997: 24.5(18.5)
1998: 20.6(12.3)
1999: 16.8(8.0)
Monitoring Stations:
1 station
Copollutant (correlation): PMio
PMi
r - 0.48
PM2.6
r - 0.61
PMio-2.5
r - 0.93
July 2009
E-177
PM Increment: 10 /yglm3
OR Estimate [Lower CI, Upper CI]
lag:
Adult Respiratory symptoms: Wheeze:
1.01(0.93,1.09]
lag 0
0.98(0.91,1.06]
lag 1
0.99(0.92,1.06]
lag 2
Breath: 1.02(0.96,1.08]
lag 0
1.01(0.97,1.06]
lag 1
1.02(0.97,1.06]
lag 2
Cough: 0.96(0.88,1.05]
lag 0
0.97(0.90,1.04]
lag 1
0.98(0.92,1.05]
lag 2
Sputum: 1.01(0.92,1.12]
lag 0
0.99(0.91,1.08]
lag 1
1.00(0.93,1.08]
lag 2
Runny Nose: 0.98(0.93,1.04]
lag 0
0.97(0.93,1.02]
lag 1:0.97(0.94,1.01]
lag 2
Eye Irritation: 0.97(0.87,1.08]
lag 0
0.97(0.88,1.06]
lag 1
0.97(0.91,1.04]
lag 2
Lower Symptoms: 0.96(0.91,1.02]
lag 0
0.95(0.89,1.00]
lag 1
0.95(0.90,1.00]
lag 2
Any Symptoms: 0.97(0.93,1.02]
lag 0
0.96(0.91,1.00]
^RAFT-DO NOT CITE OR QUOTE
0.95(0.91,0.99]
lag 2

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Mar et al. (2005, 087566)
Period of Study: 1999-2001
Location: Seattle, Washington
Study Design: Time series
Outcome: Pulmonary function (arterial
oxygen saturation) and cardiac function
(heart rate and blood pressure)
Pollutant: PMio
Averaging Time: 24-h avg
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
Lag
N: 88
Indoor
Statistical Analyses: Linear logistic
regression
Age Groups: >57
Systolic: 0.92 (-0.95, 2.78)
0
Diastolic: 0.63 (-0.29,1.56)
0
Outdoor
Systolic:-0.10 (-1.37,1.18)
0
Diastolic: -0.03 (-0.79, 0.73)
0
Nephelometer
Systolic: 0.35 (-0.91,1.61)
0
Diastolic: -0.12 (-0.91, 0.67)
0
% Increase between heart rate and
PMio exposure for people >57
PMio
Indoor: 0.02 (-0.54, 0.58)
0
Outdoor: -0.48 (-1.03, 0.06)
0
Nephelometer: -0.31 (-0.76, 0.14)
0
July 2009
E-178
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Reference: McCormack et al, 2009
Outcome: Asthma symptoms
Pollutant: PM10 2 B, PM2.6
Increment: 10 //g/m3
Period of Study: September 2001- April
Study Design: Panel
Averaging Time: 3d
Relative Risk (Min CI, Max CI)
2004




Statistical Analysis: Chi-square,
Mean (SD) Unit:
Lag
Location: East Baltimore, Maryland
Student t-test. Negative binomial


regression models with GEE, Logistic
PM102.B: 17.4 ± 21.2/yg|m3
Bivariate Models, PM10-2.B

regression with GEE
PM2.6: 40.3 ± 35.4/yg/m3
Cough, wheezing, chest tightness: 1.05

Statistical Package: StataSE
Range (Min, Max): NR
(0.99-1.10), p - 0.08

Age Groups: Asthmatic children aged 2-
R
Copollutant (correlation): NR
Slow down: 1.08 (1.03-1.13). p < 0.01

0

Symptoms with running: 1.03 (0.97-



1.09). p - 0.39



Nocturnal symptoms: 1.06 (1.01-1.11), p



- 0.03



Limited speech: 1.11 (1.05-1.18), p <



0.01



Rescue medication use: 1.06 (1.02-1.11),



p < 0.01



Bivariate Models, PM2.6



Cough, wheezing, chest tightness: 1.01



(0.98-1.05), p - 0.41



Slow down: 1.00 (0.97-1.04), p - 0.85



Symptoms with running: 1.04 (1.01-



1.07), p - 0.14



Nocturnal symptoms: 1.02 (0.98-1.05), p



- 0.37



Limited speech: 1.01 (0.95-1.07), p -



0.33



Rescue medication use: 1.03 (1.00-1.60),



p - 0.06



Multivariate Models, PM10-2.E



Cough, wheezing, chest tightness: 1.06



(1.01-1.12), p - 0.02



Slow down: 1.08 (1.02-1.14), p - 0.01



Symptoms with running: 1.00 (0.94-



1.08), p - 0.81



Nocturnal symptoms: 1.08 (1.01-1.14), p



- 0.02



Limited speech: 1.11 (1.03-1.19), p <



0.01



Rescue medication use: 1.06 (1.01-1.10),



p - 0.02



Multivariate Models, PM2.6



Cough, wheezing, chest tightness: 1.03



(0.99-1.07), p - 0.18



Slow down: 1.04 (1.00-1.09), p - 0.06



Symptoms with running: 1.07 (1.02-



1.11), p < 0.01



Nocturnal symptoms: 1.06 (1.01-1.10), p



- 0.01



Limited speech: 1.07 (1.00-1.14), p -



0.04



Rescue medication use: 1.04 (1.01-1.08),



p - 0.04
July 2009
E-179
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Mortimer et al. (2008,
187280)
Period of Study: 1989-2000
Location: Joaquin Valley, California
Outcome: Respiratory Symptoms,
Decreased lung function
Study Design: Time series
Statistical Analyses:
Deletion/Substitution/ Addition algorithm "0: r " " 05
(GEE)
Pollutant: PMio
Averaging Time: 24-h avg
Copollutants (correlation):
Logistic linear regression
Age Groups: 6-11
NO?: r - 0.30
O3: r = 0.39
Increment: NR
P (SE):
FVC: PM10 (age 0-3 yrs): 0.0121
(0.0037)
FEVi: PM10 (age 0-3 yrs): 0.0102
(0.0034)
PEF: PM10 (Mother smoked during
pregnancy):
¦0.0102(0.0039)
Reference: Mortimer et al. (2002,
Outcome: peak expiratory flow rate
Pollutant: PM10
PM Increment: 20/yg/m3
030281)
(PEFR) and symptoms


Averaging Time: 24 h
Effect Estimate [Lower CI, Upper CI]:
Period of Study: June-August 1993
Age Groups: 4-9 yrs
Mean (SD): 53
(RR estimates are odds ratios for
Location: Eight urban areas of the US:
Study Design: Cohort study
Monitoring Stations: NR
incidence of morning asthma symptoms
Bronx and East Harlem, NY
using the avg of lag 1-2)
N: 846 children with a history of asthma

Baltimore, MD
Copollutant (correlation): 8-h avg
3 urban areas (DE, CL, CH)
Statistical Analyses: Mixed linear
ozone: r - 0.51
Washington, DC
models and GEE

Single pollutant: OR - 1.26 (1.00-1.59)
Detroit, Ml
Covariates: day of study, previous 12-h

Ozone+PM10: OR - 1.25(0.97-1.61)

mean temperature, urban area, diary


Cleveland, OH
number, rain in the past 24 h

Ozone+S02 +NO2+PM10: OR - 1.14
Chicago, IL

(0.80-1.48)
Season: Summer

and St. Louis, M0.
Dose-response Investigated? No



Statistical Package: SAS



Lags Considered: 0,1, 2, 3, 4, 5, 6,1-5



avg, 1-4 avg, 0-4 avg, 0-3 avg


Reference: Moshammer and
(2003, 041956)
Period of Study: 2000-2001
Location: Linz, Austria
Outcome: Lung Function: FVC, FEVi,
MEF26, MEFbo, MEFjb, PEF, LQ Signal,
PAS Signal
Age Groups: Ages 7 to 10
Study Design: Case-crossover
N: 161 children
1898-2120 "half-h means"
Statistical Analyses: Correlations
Regression Analysis
Covariates: Morning, Evening, Night
Season: Spring, summer, Winter, Fall
Dose-response Investigated? No
Pollutant: PM10
Averaging Time: 8 h
Daily Means
Mean (SD): 23.13(20.08)
Range (Min, Max): (NR, 190.79)
Monitoring Stations: 1
Copollutant (correlation): LQ - 0.751
PAS - 0.406
Notes: "Acute effects of 'active particle
surface' as measured by diffusion
charging were found on pulmonary
function (FVC, FEVi, MEF50) of
elementary school children and on
asthma-like symptoms of children who
had been classified as sensitive."
July 2009
E-180
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Moshammeret al. (2006,
Outcome: Respiratory symptoms and
Pollutant: PM10
PM Increment: 10/yglm3
090771)
decreased lung function


Averaging Time: 8-h
% change in Lung Function per
Period of Study: 2000-2001
Age Groups: Children ages 7-10
Mean (SD): Maximum 24 h: 76.39
10/yg/m3
Location: Linz, Austria
Study Design: Time-series
Annual avg: 19.06
FEV: 0.11



N: 163 children

FVC: 0.06


Percentiles: 8-h mean


Statistical Analyses: GEE model
25th: 14.39
FEVob: -0.19



Covariates: Sex, age, height, weight

MEFjb*: -0.30

8-h mean 50th(Median): 24.85


Dose-response Investigated? NR
8-h mean 75th: 38.82
MEFbo%: -0.36



Statistical Package: NR

MEF2b%: 0.41

Monitoring Stations: 1 station


Lags Considered: 1

PEF: 0.22

Copollutant (correlation): PMi



r - 0.91
% change in Lung Function per IQR





FEV:-0.27


PM2.6




FVC: -0.07


r - 0.93




FEVob: -0.47


NO?




MEFjb*: -0.74


r - 0.62




CD
CO
O
s?
uf
LLI
2



MEF2b%: 0.98



PEF:-0.54
Reference: Neuberger et al. (2004,
093249)
Period of Study: Sept 1999-March
2000
Location: Vienna, Austria
Outcome: Ratio measure: Time to peak Pollutant: PMio
tidal expiratory flow divided by total
expiration time (i.e., tidal lung function, a Averaging Time: 24-h
surrogate for bronchial obstruction)
Age Groups: 3.0-5.9 years (preschool
children)
Study Design: Longitudinal prospective
cohort
N: 56 children
Statistical Analyses: mixed models
linear regression, with autoregressive
correlation structure
Covariates: Age, sex, respiratory rate,
phase angle, temperature, kindergarten,
parental education, observer (also in
sensitivity analyses: height, weight,
cold/sneeze on same day, heating with
fossil fuels, hair cotinine, number of tidal
slopes used to measure tidal lung
function)
Dose-response Investigated? No
Statistical Package: SAS 8.0
Lags Considered: 0
Copollutant (correlation): PM2.B
(r - 0.94) in Vienna
PM Increment: Interquartile range (NR)
Change in mean associated with an IQR
increase in PM (p-value)
¦1.067 (0.241)
lag 0
July 2009
E-181
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Neuberger et al. (2004,
093249)
Period of Study: Oct. 2000-May 2001
Location: Linz, Austria
Outcome: Forced oscillatory resistance
(at zero Hz), FVC, FEVi, MEF26, MEFbo,
MEFjb, PEF
Age Groups: 7-10 years
Study Design: Longitudinal prospective
cohort
N: 164 children
Statistical Analyses: mixed models
linear regression with autoregressive
correlation structure
Covariates: sex, time and individual
Season: Oct-May
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-7
Pollutant: PM10
Averaging Time: 24-h
Monitoring Stations: 1
PM Increment: 1 //g|m3
Notes: No significant associations
between PM10 and the metrics of lung
function were reported. The authors
state they only reported significant
associations, so results are assumed to
be null.
Reference: Odajima et al. (2008,
192005)
Period of Study: April 2003-March
2004
Location: Fukuoka, Japan
Outcome: Peak Expiratory Flow (PEF)
Study Design: Panel/Field
Statistical Analysis: GEE
Statistical Package: SAS
Covariates: Age, Sex, Growth Index,
Temperature, NO2, O3
Age Groups: Asthmatic children, 4-11
years old
Pollutant: PM10
Increment: 10 //g/m3
Averaging Time: 3h
Relative Risk (Min CI, Max CI)
Mean (SD) Unit:
Lag
Warmer Months, 5-8am
April-September, Morning Sample, Multi-

pollutant
SPM: 40.7 /yglm3

SPM, 5am-8am: -0.6 (-1.228, 0.028)
NO2:15.2 ppb

SPM, 2am-5am: -0.78 (-1.399,-0.161)
O3: 17.7 ppb

SPM, 11pm-2am: -0.612 (-1.180, -0.045)
Warmer Months, 7-10pm

SPM, 8pm-11am: -0.732 (-1.318, -0.145)
SPM: 41.5/yg/m3

O3, 5am-8am: -0.575 (-1.569, 0.419)
NO2: 20.0 ppb

O3, 2am-5am: -0.052 (-0.997, 0.893)
O3: 28.1 ppb

0s, 11 pm-2am: -0.305 (-1.269, 0.658)
Colder Months, 5-8am

03,8pm-11am:-0.416 (-1.283, 0.451)
SPM: 32.6//gfm3

NO2, 5am-8am:-0.3 (-2.246,1.645)
NO2: 20.5 ppb

NO2, 2am-5am: 0.265 (-1.354,1.885)
O3: 17.5 ppb

NO2,11pm-2am: -0.187 (-1.447,1.073)
Colder Months, 7-10pm
NO2, 8pm-11am: 0.432 (-0.689,1.553)
SPM: 34.7 //gfm3

Single-Pollutant Model
NO2: 28.0 ppb

SPM, 5am-8am:-0.67 (-1.236,-0.104)
O3: 19.4 ppb

SPM, 2am-5am: -0.761 (-1.328, -0.194)
Range (Min, Max):

SPM, 11 pm-2am:-0.661 (-1.159,-0.163)
Warmer Months, 5-8am

SPM, 8pm-11 am: -0.714 (-1.212, -0.215)
SPM: (11.0,126.0)
NO2: (1.3,44.7)
Evening Sample, Multi-pollutant Model
0s: (0.3, 52.3)
SPM, 7pm-10pm: -0.449 (-1.071, 0.174)

SPM, 4pm-7pm: -0.434 (-1.122, 0.254)
Warmer Months, 7-10pm
SPM
SPM: (8.3,191.3)

NO2: (3.0,51.3)
1pm-4pm: -0.415 (-1.015, 0.184)

SPM
0s: (1.3, 71.3)


10am-1pm:-0.522 (-1.199, 0.155)
Colder Months, 5-8am
SPM: (9.0,160.0)
O3, 7pm-10pm: -0.22 (-1.171,0.731)
NO2: (1.3,44.0)
0s, 4pm-7pm: -0.118 (-0.809, 0.574)

0s
O3: (0.6, 48.7)
July 2009
E-182
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Colder Months, 7-1 Opm
SPM: (10.3,131.0)
N(k (3.6,49.0)
0s: (1.0, 60.0)
Copollutant (correlation):
Warmer Months (24h Mean):
0s
r - 0.32
NO?
r - 0.30
Colder Months (24h Mean):
0s
r - -0.02
NO?
r - 0.45
1pm-4pm:-1.086 (-0.888, 0.516)
0s
10am-1pm:-0.315 (-1.123, 0.493)
NO2, 7pm-10pm: 0.296 (-0.806,1.397)
NO2,4pm-7pm: 0.220 (-0.818,1.258)
NO2
1pm-4pm: 0.438 (-0.568,1.444)
NO2
10am-1 pm: 0.536 (-0.546,1.617)
Single-pollutant Model
SPM, 7pm-10pm: -0.449 (-0.956, 0.058)
SPM, 4pm-7pm: -0.449 (-1.029, 0.131)
SPM
1pm-4pm:-0.414 (-0.943, 0.115)
SPM
10am-1pm:-0.486 (-1.051, 0.079)
October-March, Morning Sample, Multi-
pollutant
SPM, 5am-8am: 0.290 (-0.279, 0.859)
SPM, 2am-5am: 0.431 (-0.173,1.036)
SPM, 11pm-2am: 0.304 (-0.311, 0.919)
SPM, 8pm-11am: 0.010 (-0.523, 0.543)
O3, 5am-8am: -0.415 (-1.568, 0.738)
03, 2am-5am: -0.046 (-1.245,1.153)
O3,11pm-2am: 0.004 (-1.265,1.273)
0a, 8pm-11am: -0.470 (-2.017,1.077)
NO2, 5am-8am:-0.319 (-2.269,1.631)
NO2, 2am-5am: 0.262 (-1.777, 2.300)
NO2,11pm-2am: 0.609 (-1.132, 2.350)
NO2, 8pm-11am: 0.155 (-1.545,1.856)
Single-Pollutant Model
SPM, 5am-8am: 0.308 (-0.189, 0.805)
SPM, 2am-5am: 0.485 (-0.026, 0.996)
SPM, 11pm-2am: 0.486 (-0.049,1.022)
SPM, 8pm-11am: 0.100 (-0.414, 0.613)
Evening Sample, Multi-pollutant Model
SPM, 7pm-10pm: 0.059 (-0.397, 0.515)
SPM, 4pm-7pm: 0.360 (-0.093, 0.812)
SPM
1pm-4pm: 0.357 (-0.157, 0.871)
SPM
10am-1pm: 0.169 (-0.394, 0.731)
0a, 7pm-10pm: -0.656 (-2.394,1.083)
0s, 4pm-7pm: 0.046 (-1.140,1.232)
0s
July 2009
E-183
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1pm-4pm: 0.164 (-1.038,1.365)
0s
10am-1pm: 0.665 (-0.613,1.942)
N02,7pm-10pm:-0.415 (-2.444,1.613)
NO2,4pm-7pm:-0.144 (-1.490,1.202)
NO2
1pm-4pm:-0.181 (-1.821,1.459)
NO2
10am-1pm: 0.194 (-1.503,1.890)
Single-pollutant Model
SPM, 7pm-10pm: 0.071 (-0.388, 0.529)
SPM, 4pm-7pm: 0.318 (-0.123, 0.758)
SPM
1pm-4pm: 0.317 (-0.171, 0.804)
SPM
10am-1 pm: 0.112 (-0.412, 0.636)
July 2009
E-184
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Peacock et al. (2003,	Outcome: Reduced peak expiratory flow
042026)	rate (PEFR)
Period of Study: 11/1/1996-2/14/1997 Age Groups: 7-13 years of age
Location: Southern England	Study Design: Time-series
N: 179
Statistical Analyses: GEE, multiple
regression
Covariates: Day of the week, 24-h mean
outside temperature.
Season: winter
Dose-response Investigated? No
Statistical Package: STATA
Lags Considered: Same day, lag 1, lag
2, five day moving avg
Pollutant: PMio
Averaging Time: daily
Mean (SD): Rural (nationally validated)
21.2(11.3)
Rural (locally validated) 18.7 (11.3)
Urban 1 18.4(9.8)
Urban 2 22.7 (10.6)
Percentiles: 10th
Rural (nationally validated) 11.0
Rural (locally validated) 9.0
Urban 1 10.5
Urban 2 12.5
90th
Rural (nationally validated) 33.0
Rural (locally validated) 32.5
Urban 1 32.0
Urban 2 36.0
Range (Min, Max): Rural (nationally
validated) 7.0, 82.0
Rural (locally validated) 6.6, 87.9
Urban 1 4.7, 62.8
Urban 2 6.7, 63.7
Monitoring Stations: 3Copollutants:
NO?
0s
SO?
S042"
Increment: 10/yg/m3
Odds Ratio (Lower CI, Upper CI)
Lag
Change in PEFR
Community
¦0.04 (-0.11,0.03)
0
0.03 (-0.04, 0.05)
1
¦0.01 (-0.07, 0.05)
2
¦0.10 (-0.25, 0.05)
0-4 avg
Local
¦0.01 (-0.06, 0.03)
0
0.04 (0.01,0.08)
1
0.01 (-0.04,0.05)
2
0.04 (-0.05,0.13)
0-4 avg
20% decrease in PEFR
All children
1.012(0.992,1.031)
0
1.016(0.995,1.036)
1
1.013(1.000,1.025)
2
1.037 (0.992,1.084)
0-4 avg
Wheezy Children Only
1.016(0.986,1.047)
0
1.030 (1.001,1.060)
1
1.018(0.995,1.041)
2
1.114(1.057,1.174)
0-4 avg
July 2009
E-185
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Peled, et al (2005,156015)
Period of Study: 5-6 weeks between
March-June 1999 and September-
December 1999.
Location: Ashdod, Ashkelon and Sderot,
Israel
Outcome: Reduced peak expiratory flow
(PEF)
Age Groups: 7-10 years
Study Design: Nested cohort study
N: 285
Statistical Analyses: Time series
analysis, generalized linear model, GEE,
one-way ANOVA
Covariates: seasonal changes,
meteorological conditions and personal
physiological, clinical and socioeconomic
measurements
Season: spring, autumn
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PMio
Averaging Time: daily
Mean:
Ashkelon: 67.1
Sderot: 52.9
Ashdod: 31.0
PM Component: Local industrial
emissions, desert dust, vehicle emissions
and emissions from two electric power
plants
Monitoring Stations: 6
Copollutant: PM2.6
PM Increment: 1 /yglm3
(3 coefficient (SE) [95% CI]
Sderot:
PMio MAX: -0.34 (0.41)1-1.16, 0.46]
PMio MAX x sin(~ 2 day): 0.84 (0.22)
[0.405,1.28]
PMio MAX x cos (n1 day): -1.61 (0.41) [¦
2.43, 0.79]
PMio MAX x sin (n1 day): 0.44 (0.120) [¦
0.68-0.21]
In Sderot, an interaction between PMio
and the sequential day were significantly
associated with PEF.
Reference: Pitard, et al (2004, 087433) Outcome: Respiratory drug sales
Period of Study: 732 days (July 1998-
June 2000)
Location: City of Rouen, France
Age Groups: 0-14,15-64, 65-74, over
75 years
Study Design: Ecological time-series
N: 106,592
Statistical Analyses: Generalized
additive model
Covariates: Days of the weeks, trend,
seasonal variations, influenza epidemics,
meteorological variables, holidays
Dose-response Investigated? No
Statistical Package: S-plus
Lags Considered: 0 to 10 days
Pollutant: PMio
Averaging Time: daily
Mean (SD): 16.7 (13.3)
Percentiles:
25th: 8.00
50th(Median): 13.0
75th: 20
Range (Min, Max): 2.00,126
Monitoring Stations: 2
Copollutant (correlation): SO2 (0.39)
NO? (0.61)
PM Increment: 10/yg/m3
Percent increase in sales of anti-
asthmatics and bronchodilators (Lower
CI, Upper CI)
6.2(2.4,10.1)
lag 10 days
Percent increase in sales of cough and
cold preparation for children under 15
years of age (Lower CI, Upper CI)
9.2(5.9,12.6)
10 days
July 2009
E-186
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Preutthipan et al. (2004,
055598)
Period of Study: 31 days (school days)
from January 14 to February 26,1999
Location: Mae Pra Fatima School,
central Bangkok, Thailand
Outcome: Decreases in peak expiratory
flow rates (PEFR), respiratory symptoms
including wheeze, shortness of breath,
runny/stuffed nose, sneezing, cough,
phlegm, and sore throat
Age Groups: Third to ninth grade
Study Design: Time- Series
N: 133 children (93 asthmatics, 40
nonasthmatics)
Statistical Analyses: For continuous
data, an unpaired t-test or Mann-Whitney
U test was used. For categorical data,
the chi-square test or Fisher's exact test
was used. One way analysis of
covariance (ANCOVA) was used to
compare avg daily reported respiratory
symptoms, diurnal PEFR variability, and
the prevalence of PEFR decrements
between groups of days.
Covariates: Age, sex, weight, height,
parents smoking, person smoking in
home, daily number of household
cigarettes, air-conditioned bedroom, fuel
used for cooking (charcoal, gas), distance
from home to main road
Dose-response Investigated? No
Lags Considered: Up to 5 days
Pollutant: PMio
Averaging Time: daily
Mean (SD): 111.0 (39)
Range (Min, Max): 46, 201
Monitoring Stations: 1
Copollutant:
SO?
CO
Os
PM Increment: Authors classified
exposure according to High and Low PMio
days:
High - > 120 /yglm3
Low - < 120 /yglm3
Daily reported respiratory symptoms and
diurnal PEFR variability as classified by
concurrent days with high vs. low PMio
Mean % reporting (SEM)
Asthmatics: High PMio
Wheeze/shortness of breath - 21.3 (1.4)
Runny/stuffed nose or sneezing - 42.3
(1.8)
Cough - 59.9 (1.9)
Phlegm - 60.5 (2.3)
Sore throat - 23.7 (1.5)
Any respiratory symptoms - 72.2 (3.2)
Diurnal PEFR variability - 3.0 (0.4)
Asthmatics: Low PMio
Wheeze/shortness of breath - 19.3 (1.3)
Runny/stuffed nose or sneezing - 35.8
(1.6)
Cough - 59.1 (1.6)
Phlegm - 58.6 (2.0)
Sore throat - 21.0 (1.4)
Any respiratory symptoms - 63.8 (2.8)
Diurnal PEFR variability - 2.8 (0.3)
Nonasthmatics: High PMio
Wheeze/shortness of breath - 11.7(1.4)
Runny/stuffed nose or sneezing - 40.9
(2.5)
Cough - 50.4 (2.6)
Phlegm - 50.2 (2.5)
Sore throat - 27.1 (1.7)
Any respiratory symptoms - 67.8 (3.7)
Diurnal PEFR variability - 2.4 (0.4)
Nonasthmatics: Low PMio
Wheeze/shortness of breath - 9.3 (1.2)
Runny/stuffed nose or sneezing - 33.1
(2.2)
Cough - 54.0 (2.2)
Phlegm - 49.9 (2.2)
Sore throat - 23.9 (1.5)
Any respiratory symptoms - 56.4 (3.2)
Diurnal PEFR variability - 2.1 (0.4)
Notes: None of the daily reported
respiratory symptoms had significant
direct correlations with daily PMio levels,
according to the authors.
July 2009
E-187
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Rabinovitch et al. (2004,
096753)
Periods of Study: 11/15/1999-
311512000
11/13/2000-3/23/2001
11/15/2001-3/22/2002
Location: Denver, Colorado
Outcome: Respiratory symptoms,
Asthma symptoms (cough and wheeze),
Upper respiratory symptoms
Study Design: Time-series panel
Statistical Analyses: Logistic linear
regression
Age Groups: 6-12
Pollutants: PMio
Averaging Time:
24-h avg
Mean (SD): 28.1 (13.2)
Range (Min, Max):
(6.0,102.0)
Copollutant:
CO
NO?
SO?
0s
Increment: 1 //gf'nr1
PISE)
AM:-0.010 (0.008)
PM:-0.011 (0.010)
Odds Ratio (Lower CI, Upper CI)
Lag
1.016(0.911,1.133)
0-3 avg.
OR for respiratory symptoms and PMio
exposure for children age 6-12
Asthma exacerbation: 1.00 (0.75,1.25)
0-3 avg
Medication: 0.85 (0.75, 0.95)
0-3 avg
Previous night's symptoms: 1.10 (1.00,
1.20)
0-3 avg
Current day's symptoms: 1.00 (0.90,
1.10)
0-3 avg
% Increase (Lower CI, Upper CI)
Lag
% Increase in FEVi or PEF and PMio
exposure for children age 6-12
AM FEVi:-0.01 (-0.02,0.01)
0-3 avg
PM FEVi:-0.02 (-0.03, 0.02)
0-3 avg
AM PEF: -0.025 (-0.035,0.02)
0-3 avg
PM PEF: 0.00(-0.03, 0.03)
0-3 avg.
Reference: Renzetti et al, 2009
Outcome: Airway inflammation and
Pollutant: PMio
All results are presented in figure format.

function

Period of Study:

Averaging Time: daily


Study Design: Panel
Covariates: NR


6/1/06-7/31/06
Mean (SD) Unit:

Location: Pescara and Ovindoli, Italy

Urban: 56.9 ± 13.1 /yg/m3
Rural: 13.8 ± 5.6/yg/m3
Copollutant (correlation): NR

Statistical Analysis: Student T-test,
Pearson's correlation coefficients
Statistical Package: StatView
Age Groups: Children, mean age 9.9
years

Reference: Rojas-Martinez et al. (2007,
Outcome: Lung function: FEVi, FVC,
Pollutant: PMio
PM Increment: IQR
091064)
FEF25-75%



Averaging Time: 24-h, 6-mo
PMio, 6-LC: 36.4
Period of Study: 1996-1999
Age Groups: Children 8 years old at time
Mean (SD): 24-h averaging
GIRLS
of cohort recruitment
Location: Mexico City, Mexico



Study Design: school-based "dynamic"
Tlalnepantla: 66.7 (35.6)
One-pollutant model

cohort study
Xalostoc: 96.7 (49.4)
FVC: -39 [-47:-31]

N: 3170 children
Merced: 79.3 (40.8)
FEV: -29 [-36: -21]

14,545 observations
Pedregal: 53.4 (31.9)
FEF2e-7e«:-17 [-36: 1]
July 2009
E-188
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Statistical Analyses: Three-level
generalized linear mixed models with
unstructured variance-covariance matrix
Covariates: age, body mass index,
height, height by age, weekday spent
outdoors, environmental tobacco smoke,
previous-day mean air pollutant
concentration, time since first test
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1 days
Cerro de la Estrella: 69.6 (35.3)
6-mo averaging
Mean: 75.6
Percentiles: 6-mo averaging
25th: 55.8
50th(Median): 67.5
75th: 92.2
Monitoring Stations: 5 sites for PMio,
10 for other pollutants
Copollutant: 0:i
NO?
FEVi/FVC: 0.12(0.07: 0.17]
Two-pollutant model
PMio, 6-LC & 0s
FVC: -30 [-39:-22]
FEV: -24 [-31:-16]
FEF2b-7b%: -9 [-26: 9]
FEVi/FVC: 0.10 [0.06: 0.15]
PMio, 6-LC & NO?
FVC:-21 [-30:-13]
FEV: -17 [-25: -8]
FEF2B-7B*: -23 [-43: -4]
FEVi/FVC: 0.07 [0.02: 0.13]
Multipollutant model
PMio, 6-LC, Os, & NO?
FVC:-14 [-23:-5]
FEV:-11 [-20: -3]
FEF2B7B*: -7 [-27: 12]
FEVi/FVC: 0.08(0.03: 0.13]
BOYS
One-pollutant model
FVC:-33 [-41:-25]
FEV: -27 [-34: -19]
FEF2b-7b%:-18 [-34: -2]
FEVi/FVC: 0.04 [-0.01:0.09]
Two-pollutant model
PMio, 6-LC & Os
FVC: -28 [-36:-19]
FEV: -22 [-30:-15]
FEF2b-7b%:-10 [-27: 7]
FEVi/FVC: 0.04 [-0.01:0.09]
PMio, 6-LC & NO?
FVC: -16 [-26:-7]
FEV: -19 [-27:-10]
FEF2b-7b%: -26 [-44: -9]
FEVi/FVC: 0.005 [-0.06: 0.05]
Multipollutant model
PMio, 6-LC, Os, & NO?
FVC: -12 [-22: -3]
FEV: -15 [-23: -6]
FEF2b-7b%:-12 [-30: 6]
FEVi/FVC:-0.002 [-0.06: 0.05]
Long-term exposure to O3, PMio, and NO2
is associated with decrements in FVC and
FEVi growth in Mexico City
schoolchildren. In a multipollutant model,
PMio (-12%), Os (-9%), and NOz (-41%)
each contribute independently and
_^tatisticaM^si2nificantl^^o_dmi^mshed__
July 2009
E-189
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
FVC growth. For FEVi, however, the
multipollutant model indicates that only
PMio (-15%) and NO2 (-25%) each
contribute independently and statistically
significantly to diminished FEVi growth.
Reference: Sahsuvaroglu et al, 2009
Period of Study: 1994-1995
Location: Hamilton, Canada
Outcome: Asthma symptoms
Study Design: Panel
Covariates: Neighborhood income,
dwelling value, state of housing,
deprivation index, smoking
Statistical Analysis: Logistic
regressions
Statistical Package: SPSS
N:6388
Age Groups: Children in grades 1 and 8
Pollutant: PM10
Averaging Time: 3 year averages
Average:
All Subjects: 20.90/yglm3
Boys: 20.88 /yglm3
Girls: 20.92/yglm3
Range:
All Subjects: 26.98
Boys: 26.98
Girls: 20.10
Copollutant (correlation!:
NOxTheissen: 0.083
S02Theissen: -0.021
OsTheissen: -0.251
IMOzKriged: 0.126
NO2LUR: 0.072
Increment: NR
Odds Ratio (95%CI) for copollutant
model PMioSpline and NO2LUR
All Girls: 1.063 (0.969-1.666)
Older Girls: 1.058 (0.918-1.219)
Odds Ratio (95%CI) for copollutant
model PMioSpline and NO2LUR,
SOzTheissen and OsTheissen
All Girls: 1.045 (0.943-1.158)
Older Girls: 1.044(0.891-1.225)
Regression coefficients (95%CI)
between non-allergic asthma and
PMioSpline exposure
All Children: 1.043 (0.996-1.092)
Younger Children: 1.011 (0.929-1.100)
Older Children: 1.073(1.013-1.136)
All Girls: 1.069 (0.999-1.144)
All Boys: 1.024 (0.962-1.091)
Younger Girls: 1.065 (0.943-1.203)
Younger Boys: 0.962 (0.853-1.085)
Older Girls: 1.072 (0.984-1.169)
Older Boys: 1.075 (0.995-1.160)
Reference: Sanchez-Carrillo et al. (2003, Outcome: Upper respiratory symptom
indicator (wet cough, sore throat,
hoarseness, nose dryness, and head cold)
098428)
Period of Study: 1996-1997
Location: metropolitan Mexico City,
Mexico
Lower respiratory symptom indicator (dry
cough, lack of air, and chest sounds)
and Ocular symptom indicator (eye
irritation, eye itch, eye burning, teary
eyes, red eyes, and eye infection)
Age Groups: All ages
Study Design: Cohort
N: 151,418 interviews
Statistical Analyses: Logistic
regression models
Covariates: sex, age, education,
cigarette smoking, season, emergency
episode mass media report, temperature,
and relative humidity
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: 1
Pollutant: PM10
Averaging Time: 24 h
Mean (SD):
Northeast: 132 (52)
Northwest: 87 (46)
Central: 85 (37)
Southeast: 79 (35)
Southwest: 55 (28)
Range (Min, Max):
Northeast: (34-269)
Northwest: (10-275)
Central: (9-319)
Southeast: (14-225)
Southwest: (12-264)
Monitoring Stations: Up to 32
Copollutant (correlation):
0s: r - 0.067
0s 8: 00-18: 00 h: r - 0.075
SO2: r - 0.265
NO2: r - 0.265
Effect Estimate [Lower CI, Upper CI]:
PM10 quartiles 10.04-52.62 (ref) 52.63-
73.58
Upper respiratory indicator: 1.02 (0.99-
1.06)
Lower respiratory indicator: 1.04 (0.99-
1.09)
Ocular indicator: 0.99 (0.95-1.03) 73.59-
101.91
Upper respiratory indicator: 1.07 (1.03-
1.10)
Lower respiratory indicator: 1.09 (1.04-
1.14)
Ocular indicator: 0.89 (0.86-
0.92)101.92-318.80
Upper respiratory indicator: 0.93 (0.90-
0.97)
Lower respiratory indicator: 1.03 (0.98-
1.08)
Ocular indicator: 0.84 (0.81-0.87)
Northeast - 2nd quartile
Upper respiratory indicator: 0.354
(0.112-1.222)
Lower respiratory indicator: 0.215
(0.040-1.160)
Ocular indicator: 1.080 (0.915-1.274)
3rd quartile
July 2009
E-190
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Upper respiratory indicator: 0.118
(0.039-0.356)
Lower respiratory indicator: 0.126
(0.023-0.690)
Ocular indicator: 1.228 (0.720-2.095)
4th quartile
Upper respiratory indicator: 0.095
(0.034-0.267)
Lower respiratory indicator: 0.119
(0.026-0.549)
Ocular indicator: 0.878 (0.619-1.246)
Northwest - 2nd quartile
Upper respiratory indicator: 0.990
(0.898-1.090)
Lower respiratory indicator: 1.246
(1.087-1.429)
Ocular indicator: 1.218 (0.808-1.834)
3rd quartile
Upper respiratory indicator: 1.133
(0.974-1.317)
Lower respiratory indicator: 1.202
(1.044-1.385)
Ocular indicator: 0.345 (0.125-0.951)
4th quartile
Upper respiratory indicator: 1.019
(0.904-1.149)
Lower respiratory indicator: 1.344
(1.137-1.589)
Ocular indicator: 1.949 (1.416-2.683)
Central - 2nd quartile
Upper respiratory indicator: 1.088
(1.002-1.183)
Lower respiratory indicator: 1.046
(0.930-1.176)
Ocular indicator: 1.220 (1.115-1.335)
3rd quartile
Upper respiratory indicator: 1.054
(0.977-1.137)
Lower respiratory indicator: 1.055
(0.948-1.175)
Ocular indicator: 1.049 (0.965-1.142)
4th quartile
Upper respiratory indicator: 0.899
(0.826-0.979)
Lower respiratory indicator: 0.952
(0.845-1.073)
Ocular indicator: 0.875 (0.796-0.963)
Southeast - 2nd quartile
Upper respiratory indicator: 0.778
(0.575-1.052)
Lower respiratory indicator: 1.047
(0.916-1.196)
Ocular indicator: 0.460 (0.299-0.708)
July 2009
E-191
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
3rd quartile
Upper respiratory indicator: 1.297
(1.127-1.491)
Lower respiratory indicator: 1.391
(1.131-1.711)
Ocular indicator: 0.474 (0.314-0.715)
4th quartile
Upper respiratory indicator: 0.893
(0.812-0.983)
Lower respiratory indicator: 0.937
(0.818-1.073)
Ocular indicator: 0.314 (0.182-0.542)
Southwest - 2nd quartile
Upper respiratory indicator: 0.987
(0.913-1.066)
Lower respiratory indicator: 2.181
(1.177-4.040)
Ocular indicator: 1.026 (0.928-1.135)
3rd quartile
Upper respiratory indicator: 0.673
(0.673-1.886)
Lower respiratory indicator: 0.899
(0.790-1.024)
Ocular indicator: 1.017 (0.862-1.200)
4th quartile
Upper respiratory indicator: 0.524
(0.524-1.787)
Lower respiratory indicator: 4.346
(0.917-20.606)
Ocular indicator: 0.187 (0.090-0.387)
Reference: Schildcrout et al. (2006,
089812)
Period of Study: November 1993 to
September 1995
Location: Albuquerque, New Mexico
Baltimore, Maryland
Boston, Massachusetts
Denver, Colorado
San Diego, California
Seattle, Washington
St. Louis, Missouri
Toronto, Ontario, Canada
Outcome: Asthma Symptoms, Rescue
Inhaler Uses
Age Groups: 5 to 12 year olds
Study Design: Meta-analysis of CAMP
l\l: 990 children
Statistical Analyses: "Working
independence covariance structure"
Logistic Regression
Poisson Regression
"GEE Procedure"
Covariates: Season, age, race-ethnicity,
annual family income, day of the week
Dose-response Investigated?
Statistical Package: SAS 8.2
R
Lags Considered: 0 day lag, 1 day lag, 2
day lag, 3-day moving sum
Pollutant: PMio
Averaging Time: 24-h averages
Seattle: Daily
Albuquerque: Daily
Baltimore: 50% of study days measured
Boston: 23% of study days measured
Denver: 37% of study days measured
San Diego: 24% of study days measured
St. Louis: 19% of study days measured
Toronto: 47% of study days measured
Percentiles: 10th: 6.8-14.0
25th: 12.0-22.4
50th(Median): 17.7-32.4
75th: 26.2-42.7
90th: 32.5-53.9
Monitoring Stations: 1-12
Copollutant (correlation): NO2
r - 0.26-0.64
SO2 r - 0.31-0.65
Os r - 0.03-0.73
PM Increment: 25/yg/m3
One-pollutant Model
Asthma Symptoms: 1.02 [0.94,1.11]
0
1.01 [0.97,1.06]
1
1.02(0.98,1.07]
2
1.01 [0.98,1.05]
3-day moving sum
Rescue Inhaler Uses: [0.97,1.05]
0
[0.97,1.05]
1
1.00(0.97,1.03]
2
1.01 [0.98,1.03]
3-day moving sum
Two-pollutant Model
July 2009
E-192
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
CO r - 0.24-0.88
Asthma Symptoms:
CO-PMio
1.08 [1.01,1.15]
0
1.06[0.99,1.14]
1
1.08(1.02,1.14]
2
1.05(1.01,1.08]
3-day moving sum
NO2' PM10
1.06(0.99,1.13]
0
1.04 (0.97,1.11]
1
1.08(1.02,1.15]
2
1.04(1.00,1.07]
3-day moving sum
S02-PMio
1.05(0.98,1.13]; 0
1.04 (0.96,1.14]
1
1.05(0.98,1.12]
2
1.04 (0.99,1.08]
3-day moving sum
Rescue Inhaler Uses:
CO-PMio
1.06(0.99,1.13]
0
1.05(0.99,1.11]
1;
1.05(1.01,1.09]
2
1.03(1.00,1.07]
3-day moving sum
NO2' PM10
1.03(0.97,1.08]
0
1.03(0.98,1.08]
1
1.04(1.00,1.09]
2
1.02(1.00,1.05]
July 2009
E-193
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CTT
3-day moving sum
S02-PMio
1.01 [0.95,1.07]
0
1.02[0.97,1.07]
1
1.03[0.98,1.09]
2
1.02(0.98,1.05]
3-day moving sum
'All units expressed in //gfm3 unless otherwise specified.
Table E-10. Short-term exposure - respiratory morbidity outcomes - PM102.5.
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Aekplakorn et al.
(2003, 089908)
Period of Study: 107 days,
from October 1,1997 to
January 15,1998
Location: Mae Mo district,
Lampang Province, north
Thailand
Outcome: Upper respiratory symptoms,
lower respiratory symptoms, cough
Age Groups: 6-14 years old
Study Design: Logistic regression
N: 98 asthmatic school children
Statistical Analyses: Generalized
Estimating Equations, stratified
analysis, PR0C GENM0D
Covariates: Temperature and relative
humidity
Season: Winter
Dose-response Investigated? No
Statistical Package: SAS v 8.1
Pollutant: PMio-2 b
Averaging Time: daily
Mean (SD): NR
Range (Min, Max]: NR
Monitoring Stations: 3
Copollutant: PMio, SO2
PM Increment: 10/yg/m3
Odds Ratios [Lower CI, Upper CI]
lag:
Asthmatics:
URS: 1.04 (0.93,1.17)
lag 0
LRS: 1.09 (0.95,1.26)
lag 0
Cough: 1.08 (0.96,1.21)
lag 0
Non-Asthmatics:
URS: 1.05 (0.99,1.19)
lag 0
LRS: 0.90 (0.72,1.11)
lag 0
Cough: 0.95 (0.81,1.11)
lag 0
Reference: Bourotte et al.
(2007,150040)
Period of Study: 13 May
2002,19 July 2002
Location: Sao Paulo, Brazil
Outcome: Peak expiratory flow (PEF)
Age Groups: Avg age 39.8 +/¦ 12.3
Study Design: Cross-sectional
N: 33 patients
Statistical Analyses: Linear mixed-
effects model
Covariates: Gender, Age, BMI, Air
Pollutants, Ambient temperature,
Relative Humidity
Season: Winter
Dose-response Investigated? No
Statistical Package: S-plus
Lags Considered: 2 day lag, 3 day lag
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD): 21.7 (12.9)//g/m3
Range (Min, Max): (4.13, 62.0)
Components:
Na+
K+
Mg2+
Ca2+
Finf
CI-
NOs"
PM Increment: NR
Effect [Lower CI, Upper CI]
lag:
Morning PEF
Na+ concurrent day - -0.454 (-1.605, 0.697)
Na+ 2-day lag - -0.907 (-2.288, 0.474)
Na+ 3-day lag - -1.361 (-2.972, 0.251)
K+ concurrent day - 1.685 (-0.492, 3.862)
K+ 2-day lag - 1.838 (-1.272, 4.984)
K+ 3-day lag - 2.604 (-0.812, 6.025)
Mg2+ concurrent day - 2.265* (-0.427, 4.956)
Mg2+2-day lag - 1.271 (-1.869,4.410)
July 2009
E-194
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
SO42-
Monitoring Stations: 1
Mg2+ 3-day lag - 0.939 (-2.425,4.303)
Ca2+ concurrent day - 5.491 * (2.558, 8.424)
Ca2+ 2-day lag - 6.358* (2.251,10.465)
Ca2+3-day lag - 6.069 (1.962,10.176)
Fint concurrent day - 1.572 (-0.792, 3.935)
Fint 2-day lag - 1.630 (-1.679, 4.939)
Fint 3-day lag - 2.736* (-1.754, 7.226)
CI concurrent day - -0.951 (-2.238, 0.336)
CI 2-day lag --1.871 (-3.242 to-0.4997)
CI" 3-day lag - -2.286* (-3.934 to -0.638)1
NO3' concurrent day - 4.195* (-0.063, 8.452)
NOs"2-day lag - 6.292* (2.034,10.55)
NOs" 3-day lag - 7.341* (3.083,11.60)
S042~ concurrent day - 3.528 (-0.053, 7.110)
SO42- 2-day lag - 4.411 * (0.829, 7.991) |
SO42- 3-day lag - 6.175* (2.593, 9.756)
Evening PEF
Na+ concurrent day - -0.680 (-1.831, 0.471)
Na+ 2-day lag --1.90 (-3.316 to-0.494)
Na+ 3-day lag - -2.336* (-3.878 to -0.794)
K+ concurrent day - 0.613 (-1.564, 2.790)
K+ 2-day lag - 0.613 (-2.497, 3.723)
K+ 3-day lag - 0.000 (-3.421, 3.421) |
Mg2+ concurrent day - -0.718 (-3.522, 2.085)
Mg2+ 2-day lag - -1.933 (-5.073,1.206)
Mg2+ 3-day lag - -3.591 (-7.056 to -0.126)
Ca2+ concurrent day - 2.312* (-1.208, 5.832)
Ca2+ 2-day lag - 2.023 (-2.084, 6.130)
Ca2+3-day lag - 0.578 (-3.530,4.685)
Fint concurrent day - -1.281 (-3.644,1.083)
Fint 2-day lag - -2.503 (-5.930, 0.924)
Fint 3-day lag - -4.540 (-9.149, 0.068)
CI' concurrent day - -0.317 (-1.604, 0.970)
CI 2-day lag --1.268 (-2.556, 0.019)
CI 3-day lag --1.902 (-3.589 to-0.216)
NO3' concurrent day - 3.146 (-1.112, 7.404)
NOs" 2-day lag - 3.146 (-1.112,7.404)
NOs" 3-day lag - 1.049 (-3.209, 5.306)
S042~ concurrent day - 1.764 (-1.817, 5.346)
SO42- 2-day lag - 2.646 (-0.935, 6.228)
SO42- 3-day lag - 1.764 (-1.817, 5.346)
July 2009
E-195
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ebelt et al. (2005,
056907)
Check Ref: 2006,1st page: 396
Period of Study: Summer of
1998
Location: Vancouver, Canada
Outcome: spirometry
Age Groups: range from 54-86 yrs
mean age - 74 years
Study Design: extended analysis of e
repeated-measures panel study
N: 16 persons with C0PD
Statistical Analyses:
Earlier analysis expanded by developing
mixed-effect regression models and by
evaluating additional exposure indicators
Dose-response Investigated? No
Statistical Package: SAS V8
Pollutant: PM102 E
Averaging Time: 24 h
Mean (SD):
Ambient PMio-2.5: 5.6 (3.0)
Exposure to ambient PMio-2.5:
2.4(1.7)
Range (Min, Max): Ambient
PM102.B: (-1.2-11.9)
Exposure to ambient PMio-2.b: (¦
0.4-7.2)
Monitoring Stations: 5
Copollutant (correlation):
Ambient PMio: r- 0.69
Ambient PM2.6: r- 0.15
Nonsulfate Ambient PM2.6: r-
0.14
Exposure to Ambient PMio-2.b: r-
0.73
PM Increment: Ambient PMio-2.5: 4.5 (IQR)
Exposure to ambient PMio-2.5: 2.4 (IQR)
Notes: Effect estimates are presented in Figure 2 and
Electronic Appendix Table 1 (only available with electronic
version of article) and not provided quantitatively elsewhere.
Reference: Lagorio et
al.(2006, 089800)
Period of Study: 5/24/1999
to 6/24/1999 and 11/181999
to 12/22/1999
Location: Rome, Italy
Outcome: Lung function of subjects
(FVC and FEVi) with C0PD, Asthma
Age Groups: C0PD 50 to 80 yrs
Asthma 18 to 64 yrs
Study Design: Time series
N: C0PD N - 11
Asthma N - 11
Statistical Analyses: Non-parametric
Spearman correlation
GEE
Covariates: C0PD: daily mean
temperature, season variable (spring or
winter), relative humidity, day of week
Asthma: season variable, temperature,
humidity, and ~-2-agonist use
Season: Spring and winter
Dose-response Investigated? Yes
Statistical Package: STATA
Lags Considered: 1-3 days
PM Size: PM10-2.B
Averaging Time: 24 h
Mean (SD): Overall: 15.6(7.2)
Spring: 18.7 (7.4)
Winter: 12.3 (5.4)
Range (Min, Max): (3.4, 39.6)
PM Component:
Cd: 0.46±0.40 ng/m3
Cr: 1.9±1.7 ng/m3
Fe: 283 ± 167 ng/m3
Ni: 4.8±6.5 ng/m3
Pb: 30.6±19.0 ng/m3
Pt: 5.0±8.6 pg/m3
V: 1.8 ± 1.4 ng/m3
Zn: 45.8±33.1 ng/m3
Monitoring Stations: Two fixed
sites: (Villa Ada and Istituto
superior di Sanita)
Copollutant (correlation):
NO2 r - 0.51
O3 r - 0.31
CO r - -0.09
SO2 r - -0.16
PMior - 0.61
PM2.6 r - 0.34
PM Increment: 1 /yg/m3
They observed no statistically significant effect of PM 10-2.6
on FVC and FEVi on any of the panels (C0PD, Asthma).
fB Coefficient (SE)
C0PD
FVC(%)
24 h-1.32 (1.06)^1
48-h-1.46 (1.31)
72-h-1.38 (1.53)
FEVi(%)
24 h-0.59 (0.95)
48-h-1.01 (1.19)
72-h-0.90 (1.42)
Asthma
FVC(%)
24 h-0.17 (0.75)
48-h-0.36 (0.91)
72-h-0.24 (1.07)
FEVi(%)
24 h-0.67 (0.89)
48-h-1.19 (1.07)
72-h-0.51 (1.26)
July 2009
E-196
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Laurent et al.,
2008,156672)
Period of Study: 12/21/1003-
1213112004
Location: Strasbourg, France
Outcome: sales of short acting (3-
agonists
Study Design: Case-crossover
Covariates: NR
Statistical Analysis: Conditional
logistic regression
Age Groups: 0-39 years
Pollutant: PMio
Averaging Time: NR
Mean (SD) Unit: 20.8(10.2)
Z^g/m3
Range (Min, Max): NR
Copollutant (correlation): NO2,
0 i, correlations NR
Increment: 10//g/m3
Percent Increase in Short Acting p-agonists sold
Per increment increase in ambient PM10 at lags 4-7, a 7.5%
increase (95% CI: 4-11.2%) was seen in SABA sales.
All other results were given in figures 1 and 2
Reference: Tang et al. (2007,
091269)
Period of Study: Dec 2003 to
Feb 2005
Location: Sin-Chung City,
Taipei County, Taiwan
Outcome: Peak expiratory flow rate
(PEFR) of asthmatic children
Age Groups: 6-12 years
Study Design: Panel study
N: 30 children
Statistical Analyses:
Linear mixed-effect models were used to
estimate the effect of PM exposure on
PEFR
Covariates: Gender, age, BMI, history
of respiratory or atopic disease in
family, SHS, acute asthmatic
exacerbation in past 12 months,
ambient temperature and relative
humidity, presence of indoor pollutants,
and presence of outdoor pollutants,
Dose-response Investigated? yes
Statistical Package: S-Plus 2000
Lags Considered: 0-2
Pollutant: PM10-2.B
Averaging Time: 1 h
Mean (SD):
Personal: 17.8 (19.6)
Ambient: 17.0 (10.6)
Range (Min, Max):
Personal: 0.3-195.7
Ambient: 0.1-80.2
Monitoring Stations: 1
PM Increment: 15.9/yg/m3
RR Estimate [Lower CI, Upper CI]
lag:
Change in morning PEFR:
¦20.55 (-45.83, 4.73) lag 0
¦39.05 (-104.16, 26.06) lag 1
¦39.56 (-79.56, 0.44) lag 2
¦37.15 (-105.01, 30.7) 2-day mean
¦35.47 (-27.32, 56.38) 3-day mean
Change in evening PEFR:
¦1.68 (-19.13,15.78) lag 0
1.59 (-14.32,17.5) lag 1
0.86 (-30.84, 32.57) lag 2
5.97 (-15.57, 27.5) 2-day mean
29.75 (-1.69, 61.18) 3-day mean
Reference: Trenga et al.,
(2006,155209)
Period of Study: 1999-2002
Location: Seattle, WA
Outcome: Lung function: FEVi, PEF,
MMEF (maximal midexpiratory flow
assessed only for children)
Age Groups: Adults (56-89-years-old)
healthy & with C0PD
asthmatic children 6-13-years-old
Study Design: adult and pediatric panel
study over three years with 1
monitoring period ("session") per year
N: 57 adults (33 healthy, 24 with
C0PD) - 692 subject-days - 207
study-days
17 asthmatic children - 319 subject-
days - 98 study-days
Statistical Analyses: mixed effects,
longitudinal regression models, with the
effects of pollutant decomposed into
each subject's a) overall mean
b)	difference between their session-
specific mean and overall mean
c)	difference between their daily values
and session-specific mean
Covariates: gender, age, ventral site
temperature and relative humidity, CO,
NO?
Season: NR
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1 days
Pollutant: PM10-2.B (coarse)
Averaging Time: 24-h
Percentiles:
Subject-specific exposure
PMio-PM2.b
Outdoor
25th: 3.3
50th (Median): 4.7
75th: 6.9
Adults
Outdoor
25th: 3.3
50th (Median): 5.0
75th: 7.1
Range (Min, Max):
Subject-specific exposure
Children
Outdoor (0.0, 25.3)
Adults
Outdoor (0.0, 25.7)
Monitoring Stations: 2
also subject-specific local
outdoors (i.e., at each home),
indoor, and personal
PM Increment: 10/yg/m3
Adult
Outdoor Home PMio-PM2.b
FEVi
Overall: Lag 0 -27.9 [-87.5: 31.8]
Lag 1 47.1 [-5.1:99.4]
No-COPD: Lag 0 -49.2 [-22.3: 23.9]
Lag 1 74.3 [6.8: 141.8]
C0PD: Lag 0 7.3 [-84.7: 99.4]
Lag 1 11.5[-65.4: 88.3]
PEF
Overall: Lag 0 5.3 [-5.1: 15.7]
Lag 1 -2.5 [-11.6: 6.5]
No-COPD: Lag 0 5.1 [-7.7: 17.8]
Lag 1 -5.8 [-17.5: 5.9]
C0PD: Lag 0 5.7 [-10.3: 21.6]
Lag 1 1.7[-11.5:14.9]
Pediatric
FEVi
Outdoor Home PMio-PM2.b
Overall
Lag 0 -7.43 [-69.41: 54.55]
Lag 1 -25.61 [-88.16: 36.94]
No Anti-inflam. Medication
July 2009
E-197
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Copollutant (correlation!:
CO
NO?
PM2.6
Lag 0 -63.87 [-199.58: 71.84]
Lag 1 -96.48 [-232.48: 39.52]
Anti-inflam. Medication
Lag 0 6.57 [-96.90: 110.04]
Lag 1 -8.63 [-217.39: 200.14]
PEF
Outdoor Home PMio-PM2.b
Overall
Lag 0 4.53 [-6.60: 15.67]
Lag 1 -3.35 [-14.31: 7.62]
No Anti-inflam. Medication
Lag 0 2.05 [-22.36: 26.45]
Lag 1 -6.56 [-30.90: 17.78]
Anti-inflam. Medication
Lag 0 5.15 [-7.90: 18.19]
Lag 1 -2.58 [-15.35: 10.19]
MMEF
Outdoor Home PMio-PM2.b
Overall
Lag 0 -0.01 [-7.29: 7.28]
Lag 1 -2.07 [-9.25: 5.12]
No Anti-inflam. Medication
Lag 0-7.14 [-23.16: 8.87]
Lag 1 -14.39 [-30.11: 1.32]
Anti-inflam. Medication
Lag 0 1.76 [-6.78: 10.30]
Lag 1 0.89 [-7.56: 9.33]
'All units expressed in //gfm3 unless otherwise specified.
Table E-11. Short-term exposure - respiratory morbidity outcomes - PM2.5 (including
components/sources).
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Adamkiewicz et al. (2004,
087925)
Period of Study: August-December
2000
Location: Steubenville, Ohio
Outcome: FENO
Age Groups: ranged 53.5-90.6 years
Study Design: prospective cohort
N: total of 294 breaths from 29 subjects
Statistical Analyses: Fixed effect
models, ANOVA, GLM procedure
Covariates: Subject, week of study, day
of the week, h of the day, ambient
barometric pressure, temperature, and
relative humidity
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: Hourly lags, 0-48 h
Pollutant: PM2.6
Averaging Time: 1 h
Mean (SD): 19.5
Percentiles: 25th: 7.6
75th: 25.5
Range (Min, Max]: NR, 105.!
Monitoring Stations: 1
Averaging Time: 24 h
Mean (SD): 19.7
Percentiles: 25th: 9.7
75th: 27.4
Range (Min, Max]: NR, 57.8
PM Increment: 17.9/yg/m3
Effect Estimate [Lower CI, Upper CI]:
1-h Single pollutant models: 0.36 (0.58-
2.14)
PM Increment: 17.7
Effect Estimate [Lower CI, Upper CI]:
24 h moving avg: 1.45 (0.33-2.57)
Multipollutant models for PM2.6, ambient
NO and room NO and estimated change in
FENO (ppb) for an IQR in pollutant
measure
Model 1 1.95(0.47-3.43)
Model 2 1.38 (0.26-2.51)
July 2009
E-198
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Monitoring Stations: 1
Copollutant (correlation!: Ambient NO
Indoor NO
NO?;
Os
SO?
Model 4 1.97 (0.48-3.46)
Notes: Association of FENO with PM2.6 at
different lags presented in Figure 1 are
not presented quantitatively elsewhere.
Reference: Adar et al. (2007, 098635)
Period of Study: March-June 2002
Location: St. Louis, M0
Outcome: FENO
Age Groups: 60 +
Study Design: Panel Study
N: 44 non-smoking seniors
Statistical Analyses: mixed models
containing random subject effects
Covariates: Day of week, trip type,
FENO collection device, current illness,
use of vitamins, antihistamines, statins,
steroids, and asthma medications,
temperature, pollen, mold, NO
concentration in testing room
Statistical Package: SAS
Lags Considered: 0
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD): Pretrip: 14.8
Post-trip: 16.5
Percentiles:
25th (pretrip): 11.2
75th (pretrip): 20.1
25th (post-trip): 11.7
75th (post-trip): 21.6
Monitoring Stations: 1
Copollutant (correlation): B0
CO
NO2
SO2
Os
PM Increment: 9.8 /yglm3
Effect Estimate [Lower CI, Upper CI]:
Pre-trip % change: 21.9 (6.7, 39.4)
Post-trip % change: -4.7 (-17.1, 9.6)
Reference: Aekplakorn et al (2003,
089908)
Period of Study: 107 days, from
October 1,1997 to January 15,1998
Location: Mae Mo district, Lampang
Province, north Thailand
Outcome: Upper respiratory symptoms,
lower respiratory symptoms, cough
Age Groups: 6-14 years old
Study Design: Logistic regression
N: 98 asthmatic school children
Statistical Analyses: Generalized
Estimating Equations, stratified analysis,
PROC GENMOD
Covariates: Temperature and relative
humidity
Season: Winter
Dose-response Investigated? No
Statistical Package: SAS v 8.1
Pollutant: PM2.6
Averaging Time: daily
Mean (SD):
Sob Pad station: 24.77
Sob Mo station: 24.89
Hua Fai station: 26.27
Range (Min, Max):
Sob Pad: 4.52, 24.77
Sob Mo: 3.13, 24.89
Hua Fai: 3.67,26.27
Monitoring Stations: 3
Copollutant:
PM10
SO2
PM Increment: 10 /yglm3
Odds Ratios [Lower CI, Upper CI]
lag: Asthmatics: URS: 1.04 (0.99,1.09)
lag 0
LRS: 1.05 (0.98,1.2)
lag 0
Cough: 1.05 (0.99,1.10)
lag 0
Non-Asthmatics: URS: 1.03 (0.96,1.09)
lag 0
LRS: 1.02 (0.93,1.10)
lag 0
Cough: 1.00 (0.93,1.07)
lag 0
PM10 + SO2
Asthmatics: URS: 1.04 (0.99,1.10)
lag 0
LRS: 1.05 (0.98,1.10)
lag 0
Cough: 1.05 (0.99,1.11)
lag 0
Non-Asthmatics: URS: 1.03 (0.97,1.09)
lag 0
LRS: 1.02 (0.93,1.11)
July 2009
E-199
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
lag 0
Cough: 1.00 (0.93,1.07)
lag 0
Reference: Allen et al (2008,156208)
Period of Study: 1999-2002 (additional
PM composition data collected Dec 2000
and May 2001)
Location: Seattle, USA
Outcome: daily changes in exhaled nitric
oxide (FENO) and 4 lung function
measures, midexpiratory flow (MEF),
peak expiratory flow (PEF), forced
expiratory volume in one second (FEVi),
and forced vital capacity (FVC)
Age Groups: 6-13 yrs
Study Design: Panel study
N: 19 children with asthma
Statistical Analyses: linear mixed
effects model with random intercept to
test for within participant associations
Covariates: Temperature, relative
humidity, BMI, age, and, in the case of
FENO, ambient NO measured at a
centrally located monitoring site
models also included a term for within-
participant, within-session effects, and a
term for participant between-session
effects
Effect modification: Decided a priori to
include interaction term for PM2.6
exposure and inhaled corticosteroids
Pollutant: PM2.6
Mean (SD): 11.23 (6.48)
Range (Min, Max):
2.76-40.38
25th: 6.38
75th: 14.73
Copollutant (correlation):
Ambient LAC* r—0.83
Ambient LG**r-0.84
Personal PM2.6: r - 0.34
Personal LAC: r-0.54
Ambient-generated PM2.6:
r—0.87
Nonambient-generated PM2.6: r--0.06
* LAC Light-absorbing carbon
** LG: Leroglucosan (a marker of wood
smoke)
Health effect estimates presented in
graphic form (Fig 1). Summary from text
is as follows:
Personal LAC, personal PM2.6, and
ambient-generated PM2.6 were associated
with (p < 0.05) and ambient PM2.6 was
marginally associated (p — 0.09) with
increased FENO. Neither of the ambient
combustion markers (LAC, LG) nor
nonambient-generated PM2.6 was
associated with FENO changes.
All of the ambient concentrations were
associated with decrements in PEF and
MEF while ambient-generated PM2.6 was
marginally associated (p < 0.10).
Only ambient LG was associated with a
decrease in FEVi and there were no
associations between exposure metrics
and FVC.
Reference: Barraza-Villarreal et al.(2008, Outcome: Respiratory Symptoms,
156254)	Coughing, Wheezing, Airway
inflammation, Asthma
Period of Study: 6/2003-6/2005
Study Design: Prospective cohort
Location: Mexico City
Statistical Analyses: Bivarate analysis
Age Groups: 6-14
Pollutant: PM2.6
Averaging Time: Maximum 8-h avg
Mean (SD) unit:
28.9 (2.8)
Range (Min, Max):
(4.2,102.8)
Copollutants (correlation):
0s
NO2
Increment: 17.5/yg/m3
% Increase (Lower CI, Upper CI)
lag:
Asthmatic children
Inflammatory Marker: FENO: 1.08 (1.01,
1.16)
0
IL-8: 1.08 (0.98,1.19)
0
phEBC:-0.03 (-0.09, 0.03)
0
Lung Function: FEVi: -16.0 (-31.0 to ¦
0.13)
0-4 avg
FVC:-23.0 (-42.0 to-5.21)
0-4 avg
FEV2B.7B: -11.0 (-42.0, 20.3)
0-4 avg
Nonasthmatic children
Inflammatory Marker: FENO: 0.89 (0.78,
1.01)
0
IL-8:1.16 (1.00,1.36)
0;
phEBC:-0.05 (-0.14, 0.04)
0
Lung Function: FEVi: -21.0 (-42.3, 0.38)
July 2009
E-200
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0-4 avg
FVC: -29.0 (-52.8 to -4.35)
0-4 avg
FEV25-75: -20.0 (-69.0, 29.0)
0-4 avg
All children age 6-14
Respiratory Symptom: Cough: 1.11 (1.06,
1.17)
Wheezing: 1.06 (0.99,1.13)
Reference: Bennett et al. (2007,
156268)
Period of Study: 1992-2005
Location: Melbourne, Australia
Outcome: Adverse respiratory symptoms Pollutant: PM2.6
(wheeze, shortness of breath on waking,
cough in the morning, phlegm in the
morning, cough with phlegm in the
morning, asthma attack)
Age Groups: All ages with a mean of
37.2 yrs
Study Design: cohort study
N: 1446 persons
Statistical Analyses: Logistic
regression models
Averaging Time: 24 h
Mean (SD): 6.8
Range (Min, Max):
(1.8-73.3)
Monitoring Stations: 1
Covariates: Age, gender, current
smoking status, medication use (B2-
agonist and inhaled steroid), atopy
Dose-response Investigated? No
Statistical Package: STATA statistical
software, version 9 (Statcorp, 2005)
PM Increment: 1 /yglm3
Effect Estimate [Lower CI, Upper CI]:
Within-person (longitudinal effects)
Wheeze: OR-1.08 (0.79-1.48)
SOB on waking: OR-1.34 (0.84-2.16)
Cough in the morning: OR -0.74 (0.47-
1.15)
Phlegm in the morning: OR -1.55 (0.95-
2.53)
Cough w/ phlegm morning: OR -1.28
(0.70-2.33)
Asthma attack: OR — 0.91 (0.55-1.49)
Betweenperson (cross-sectional) effects
Wheeze: OR-1.32 (0.82-2.10)
SOB on waking: OR -1.29 (0.46-3.60)
Cough in the morning: OR -0.21 (0.07-
0.62)
Phlegm in the morning: 0R-0.49 (0.16-
1.44)
Cough wf phlegm morning: OR -0.28
(0.08-0.97)
Asthma attack: OR — 0.52 (0.17-1.59)
Reference: Bourotte et al. (2007,
150040)
Period of Study: 13 May 2002-19 July
2002
Location: Sao Paulo, Brazil
Outcome: Peak expiratory flow (PEF)
Age Groups: Avg age 39.8 +/¦ 12.3
Study Design: Cross-sectional
N: 33 patients
Statistical Analyses:
Linear mixed-effects model
Covariates: Gender, Age, BMI, Air
Pollutants, Ambient temperature, Relative
Humidity
Season: Winter
Dose-response Investigated? No
Statistical Package: S-plus
Lags Considered: 2 day lag, 3 day lag
Pollutant: PM2.6 (Fine)
Averaging Time: 24 h
Mean (SD): 11.9 (5.12)
Range (Min, Max):
(2.82, 26.6)
Components:
K+
Mg2+
Ca2+
Finf
CI-
NOs-
S042-
Monitoring Stations: 1
PM Increment: NR
Effect [Lower CI, Upper CI]
lag:
Morning PEF
Na+ concurrent day - -0.409 (-2.485,
1.667)
Na+ 2-day lag --0.818 (-4.139, 2.503)
Na+ 3-day lag - -0.205 (-4.356, 3.974)
K+ concurrent day - -0.211 (-2.778,
2.357)
K+ 2-day lag - -0.843 (-4.695, 3.008)
K+ 3-day lag - 0.843 (-4.292, 5.978)
Mg2+ concurrent day - -1.750 (-5.302,
1.802)
Mg2+ 2-day I
Mg2+ 3-day I
-5.016 (-10.79, 0.762)
¦3.850 (-10.15, 2.449)
Ca2+ concurrent day - 3.192* (-0.599,
6.943)
Ca2+ 2-day lag - 5.880(1.105,10.65)
July 2009
E-201
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Ca2+3-day lag - 7.560* (2.103,13.02)
Fint concurrent day - 2.218* (-0.033,
4.470)
Fint 2-day lag - 3.697* (1.446, 5.949)
Fint 3-day lag -4.067* (1.065, 7.069)
CI' concurrent day --1.010 (-3.469,
1.450)
CI 2-day lag --1.615 (-5.714, 2.483)
CI 3-day lag --1.615 (-6.534, 3.303)
NO3' concurrent day - 3.144 (0.409,
5.878)
NOs" 2-day lag - 3.593 (0.858, 6.328)
NOs" 3-day lag - 4.491 (1.756,7.226)
S042~concurrent day - 2.210 (-0.032,
4.272)
SO42- 2-day lag - 3.180 (1.028, 5.332)
SO42- 3-day lag - 3.180 (1.028, 5.332)
Evening PEF
Na+ concurrent day - -1.636 (-3.712,
0.440)
Na+ 2-day lag --0.205 (-3.256, 3.117)
Na+3-day lag --1.023 (-5.174, 3.129)
K+ concurrent day - -1.897 (-4.465,
0.670)
K+ 2-day lag --1.686 (-5.966, 2.592)
K+ 3-day lag --1.054 (-6.189, 4.081)
Mg2+ concurrent day - -2.753 (-6.400,
0.894)
Mg2+ 2-day lag - -2.567 (-8.534, 3.401)
Mg2+3-day lag --4.876 (-11.36,1.612)
Ca2+ concurrent day - 2.184 (-1.567,
5.935)
Ca2+ 2-day lag - 5.040 (0.265,9.815)
Ca2+3-day lag - 5.040 (-0.417,10.50)
Fint concurrent day - 1.479 (-0.773,
3.730)
Fint 2-day lag - 1.819 (-0.403, 4.100)
Fint 3-day lag - 2.958 (-0.044, 5.960)
CI' concurrent day - -0.404 (-2.863,
2.055)
CI 2-day lag - 0.000 (-4.099, 4.099)
CI 3-day lag -0.202 (-4.716, 5.120)
NO3' concurrent day - 1.796 (-0.939,
4.531)
NO3" 2-day lag - 2.695 (-0.040, 5.430)
NO3" 3-day lag - 3.144(0.409,5.878)
S042~ concurrent day - 2.120 (-0.032,
4.272)
SO42- 2-day lag - 2.120 (-0.032,4.272)
SO42- 3-day lag -2.120 (-0.032, 4.272)
July 2009
E-202
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: de Hartog et al. (2003,
001061)
Period of Study: winter of 1998-1999
(in Amsterdam, from November 2,1998
to June 18,1999
in Erfurt, from October 12,1998 to April
4,1999
and in Helsinki, from November 2,1998
to April 30,1999.)
Location: Amsterdam, the Netherlands
Erfurt, Germany
and Helsinki, Finland
Outcome: Respiratory symptoms
Age Groups: a 50 yrs
Study Design: Cohort
N: 131 subjects with history of coronary
heart disease
Statistical Analyses: Logistic
regression
Covariates: Ambient temperature,
relative humidity, atmospheric pressure,
incidence of influenza-like illness
Season: Winter
Dose-response Investigated? No
Statistical Package: S-PLUS 2000
Lags Considered: 0,1, 2, 3, and 5-day
avg
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD):
Amsterdam, the Netherlands: 20.0
Erfurt, Germany: 23.4
Helsinki, Finland: 12.8
Range (Min, Max):
Amsterdam, the Netherlands: (3.8-82.2)
Erfurt, Germany: (4.5-118.1)
Helsinki, Finland: (3.1-39.8)
Unit (i.e./yg/m3): //gfm3
Monitoring Stations: 1
Copollutant:
PMio
NCo.oi-0.1
CO
NO?
SO?
PM Increment: 10 //gfm3
Effect Estimate [Lower CI, Upper CI]:
Association of air pollution and incidence
of symptoms in three panels of elderly
subjects
LagO
Chest pain w) physical exertion: 1.04
(0.96-1.13)
Shortness of breath: 1.04 (0.96-1.12)
Awakened, breathing problems: NA
Avoidance of activities: 1.04 (0.96-1.14)
Phlegm: 1.03(0.93-1.13)
Lag 1
Chest pain wf physical exertion: 1.01
(0.93-1.09)
Shortness of breath: 1.06 (0.99-1.14)
Awakened, breathing problems: 1.09
(1.00-1.20)
Avoidance of activities: 1.03 (0.95-1.12)
Phlegm: 1.10(1.01-1.19)
Lag 2
Chest pain wf physical exertion: 0.98
(0.90-1.05)
Shortness of breath: 1.05 (0.98-1.12)
Awakened, breathing problems: 1.04
(0.95-1.14)
Avoidance of activities: 1.05 (0.97-1.14)
Phlegm: 1.08(1.00-1.18)
Lag 3
Chest pain wf physical exertion: 1.00
(0.93-1.08)
Shortness of breath: 1.08 (1.01-1.15)
Awakened, breathing problems: 0.99
(0.91-1.08)
Avoidance of activities: 1.06 (0.98-1.14)
Phlegm: 1.10(1.01-1.19)
5-day
Chest pain wf physical exertion: 1.02
(0.91-1.13)
Shortness of breath: 1.12 (1.02-1.24)
Awakened, breathing problems: 1.03
(0.90-1.18)
Avoidance of activities: OR - 1.09 (0.97-
1.22)
Phlegm: OR- 1.16 (1.03-1.32)
Reference: Delfino et al. (2004, 056897) Outcome: FEVi
Age Groups: 9-19 years old
Study Design: Panel study
N: 24 children
Statistical Analyses: GLM
Period of Study: September-October
1999
April-June 2000
Location: Alpine, California
Pollutant: PM2.6
Averaging Time: 24-h avg 1-h max
personal PM last 24 h
Mean (SD): 151.0 (12.03) 90th: 292.4
Range (Min, Max): (9.1, 996.8)
Results presented graphically; -Percent
predicted FEVi was inversely associated
with personal exposure to fine particles.
¦ Inverse associations of FEVi with
stationary-site indoor, outdoor and
central-site gravimetric PM2.6 and PM10,
and with hourly TE0M PM10
July 2009
E-203
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Akaike's information criterion and
Bayesian information criterion
Covariates: Day of week, Personal
temperature and relative humidity, time
of FEVi maneuver (morning, afternoon, or
evening), Season (fall 1999 or spring
2000), As-needed medication use,
Presence or absence of upper or lower
respiratory infections
Season: Spring, Fall
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-4
Mean personal PM last 24 h
Mean (SD): 37.9(19.9)
90th: 65.1
Range (Min, Max): 3.9,113.8
Home stationary-site PM
24-h Mean indoor PM2.6
Mean (SD): 12.1 (5.4)
90th: 20.2
Range (Min, Max): 2.8, 35.3
24-h Mean outdoor PM2.6
Mean (SD): 11.0 (5.4)
90th: 18.4
Range (Min, Max): 1.8, 31.0
Central outdoor stationary-site PM
24-h Mean PM2.6
Mean (SD): 10.3 (5.6)
90th: 18.4
Range (Min, Max): 1.7, 29.1
Copollutant (correlation):
24-h Central HI PM2.1,
8-h max O3 - 0.24
8-h Max NO2 - 0.73
8-h Max Personal PM - 0.38
24-h Mean Personal PM - 0.43
8-h Max TE0M PM10 - 0.71
24-h Mean TEOMPM10 - 0.78
24-h Central HI PM10 - 0.90
24-h Outdoor HI PM2.1, - 0.89
24-h Outdoor HI PM10 - 0.72
24-h Indoor HI PM10- 0.40
24-h Indoor HI PM2 b - 0.73
Reference: Delfino et al. (2006, 090745) Outcome: Fractional Concentration of
Nitric Oxide in exhaled air (FEN0)
Period of Study: Region 1: August to
Mid December 2003. Region 2: July
through November 2004
Location: Region 1: Riverside, CA.
Region 2: Whittier, CA
Age Groups: 9 through 18
Study Design: Longitudinal Panel Study
N: 45 children
Riverside children
32 Whittier children
Statistical Analyses: Linear mixed-
effects models
Two-stage hierarchical model
Empirical Variograms
Fourth-order polynomial distributed lag
mixed-effects model
Covariates: Personal temperature,
Personal Rel. Humid., 10-day exposure
run, Respiratory infections, Region of
study, Sex, Cumulative daily use of as-
Pollutant: PM2.6
Personal Exposure
Averaging Time: 24 h
Riverside
Mean (SD): 32.78(21.84)
50th(Median): 28.14
Range (Min, Max): 7.27, 98.43
Whittier
Mean (SD): 36.2 (25.46) 50th(Median):
29.07
Range (Min, Max): 7.55,197.05
Personal Exposure
Averaging Time: 1 h
Riverside
Mean (SD): 97.94 (70.29) 50th(Median):
PM Increment: IQR increase (Riverside:
28.41 /yglm3, Whittier 21.87 /yglm3)
Coefficient [Lower CI, Upper CI]
lag:
Mixed-model estimates of the association
between personal and central-site air
pollutant exposure and FEN0
Lag 0
Personal 0.42 (-0.15, 0.99)
Central 0.03 (-0.68, 0.74)
Lag 1
Personal 0.51 (-0.10,1.12)
Central 0.44 (-0.28,1.16)
2-day MA
Personal 1.01 (0.14,1.88)
July 2009
E-204
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
needed B-agonist inhalers
Dose-response Investigated? No
Lags Considered: 0,1, 2, MA day
83.7
Range (Min, Max): 14.9, 431.8
Whittier
Mean (SD): 93.63 (75.19) 50th(Median):
71.95
Range (Min, Max): 5.8, 572.9
Personal Exposure
Averaging Time: 8 h
Riverside
Mean (SD): 47.21 (30.9) 50th(Median):
38.5
Range (Min, Max): 8.9,132.1
Whittier
Mean (SD): 51.75 (36.88) 50th(Median):
40.15
Range (Min, Max): 8.7, 254.1
Central Site
Averaging Time: 24 h
Riverside
Mean (SD): 36.63 (23.46) 50th(Median):
29.26
Range (Min, Max): (9.52, 87.22)
Whittier
Mean (SD): 18 (12.14) 50th(Median):
16.3
Range (Min, Max): 2.7, 77.09
Monitoring Stations: 48 personal
nephelometers
2 central sites
Copollutant (correlation):
Personal
24-h personal PM2.61.00
24-h personal EC 0.18
24-h personal 0C 0.15
24-h personal NO2 0.33
24-h central PM2.6 0.64
24-h central EC 0.12
24-h central 0C 0.21
24-h central NO2 0.22
Central
24-h personal PM2.6 0.64
24-h personal EC 0.00
24-h personal 0C -0.11
24-h personal NO2 0.12
24-h central PM2.61.00
24-h central EC 0.55
24-h central 0C 0.66
24-h central NO2 0.25
Central 0.52 (-0.43,1.47)
Stratified by Medication Use
Lag - 2-day moving avg
Not Taking Anti-lnflamm. Medication
Personal 1.11 (-1.39,3.60)
Central 0.44 (-1.65,2.53)
Taking Anti-lnflamm. Medication
Personal 1.01 (0.19,1.84)
Central 0.55 (-0.47,1.57)
Inhaled Corticosteroids
Personal 1.58(0.72, 2.43)
Central 1.16 (0.11, 2.20)
Antileukotrienes +- inhaled
corticosteroids
Personal -0.89 (-2.73, 0.95)
Central-0.75 (-2.83,1.32)
Notes:
Figure of Estimated lag effect of hourly
personal PM2.6 on FEN0.
Figure of the Estimated lag effect of
hourly personal PM2.6 on FEN0 by use of
medications.
Figure of One- and two-pollutant models
for change in FEN0 using 2-day Moving
Averages personal and central-site
pollutant measurements.
July 2009
E-205
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Delfino et al. (2006, 090745) Outcome: Fractional Concentration of
Nitric Oxide in exhaled air (FENO)
Period of Study: Region 1: August to
Mid December 2003. Region 2: July
through November 2004
Location: Region 1: Riverside, CA.
Region 2: Whittier, CA
Age Groups: 9 through 18
Study Design: Longitudinal Panel Study
N: 45 children
Statistical Analyses: Linear mixed-
effects models
Two-stage hierarchical model
Empirical Variograms
Fourth-order polynomial distributed lag
mixed-effects model
Covariates: Personal temperature,
personal rel. humid., 10-day exposure run,
respiratory infections, region of study,
sex, cumulative daily use of as-needed B-
agonist inhalers
Dose-response Investigated? No
Lags Considered: Lag 0, Lag 1, 2-day
moving avg
Pollutant: PM2.6
PM Component: Elemental carbon
Personal Exposure
Averaging Time: 24 h
Riverside
Mean (SD): 0.42 (0.69) 50th(Median):
0.34/yg/m3
Range (Min, Max): 0.01, 6.94
Whittier
Mean (SD): 0.78(1.42)
50th(Median): 0.47
Range (Min, Max): 0,17.2
Central Site
Averaging Time: 24 h
Riverside
Mean (SD): 1.61 (0.78) 50th(Median):
1.35
Range (Min, Max): 0.52, 3.64
Whittier
Mean (SD): 0.71 (0.43) 50th(Median):
0.63
Range (Min, Max): 0.14, 2.95
Monitoring Stations: 48 personal
nephelometers,
2 central sites
Copollutant (correlation):
Personal
24-h personal PM2.6 0.18
24-h personal EC 1.00
24-h personal 0C 0.41
24-h personal NO2 0.0.21
24-h central PM2.6 0.00
24-h central EC 0.04
24-h central 0C -0.01
24-h central NO2 0.23
Central
24-h personal PM2.6 0.12
24-h personal EC 0.04
24-h personal 0C 0.03
24-h personal NO2 0.19
24-h central PM2.6 0.55
24-h central EC 1.00
24-h central 0C 0.87
24-h central NO2 0.70
PM Increment: IQR increase (Riverside:
28.41 /yglm3, Whittier 21.87 /yg/m3)
Coefficient [Lower CI, Upper CI]
Mixed-model estimates of the association
between personal and central-site air
pollutant exposure and FENO
Lag 0
Personal 0.29 (0.10, 0.48)
Central 0.10 (-0.65, 0.85)
Lag 1
Personal-0.01 (-0.23,0.21)
Central 0.99 (0.27,1.71)
2-day MA
Personal 0.72(0.32,1.12)
Central 1.38(0.15, 2.61)
Stratified by Medication Use
Lag - 2-day moving avg
Not Taking Anti-lnflamm. Medication
Personal 0.84 (0.08,1.60)
Central 1.02 (-2.55,4.60)
Taking Anti-lnflamm. Medication
Personal 0.71 (0.28,1.15)
Central 1.42 (0.25, 2.60)
Inhaled Corticosteroids
Personal 0.67(0.28,1.07)
Central 1.28(0.07, 2.49)
Antileukotrienes +- inhaled
corticosteroids
Personal 0.03 (-3.29, 3.35)
Central 1.15 (-1.58, 3.88)
Notes:
Figure of Estimated lag effect of hourly
personal PM2.6 on FENO.
Figure of the Estimated lag effect of
hourly personal PM2.6 on FENO by use of
medications.
Figure of One- and two-pollutant models
for change in FENO using 2-day Moving
Averages personal and central-site
pollutant measurements.
Reference: Delfino et al. (2006, 090745) Outcome: Fractional Concentration of
Nitric Oxide in exhaled air (FENO)
Period of Study: Region 1: August to
Mid December 2003. Region 2: July Age Groups: 9 through 18
Pollutant: PM2.6
PM Component: Organic carbon
Personal Exposure
PM Increment: IQR increase (Riverside:
28.41 /yglm3, Whittier 21.87 /yglm3)
Mixed-model estimates of the association
between personal and central-site air
July 2009
E-206
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
through November 2004
Location: Region 1: Riverside, CA.
Region 2: Whittier, CA
Study Design: Longitudinal Panel Study
N: 45 children
Statistical Analyses: Linear mixed-
effects models
Two-stage hierarchical model
Empirical Variograms
Fourth-order polynomial distributed lag
mixed-effects model
Covariates: Personal temperature,
personal rel. humid., 10-day exposure run,
respiratory infections, region of study,
sex, cumulative daily use of as-needed B-
agonist inhalers
Dose-response Investigated? No
Lags Considered: Lag 0, Lag 1, 2-day
moving avg
Averaging Time: 24 h
Riverside
Mean |SD): 5.63 (2.59) 50th(Median):
4.98
Range (Min, Max): 1.94,12.38
Whittier
Mean (SD): 6.81 (3.45) 50th(Median):
6.43
Range (Min, Max): 2.18, 31.5
Central Site
Averaging Time: 24 h
Riverside
Mean (SD): 6.88(1.86)
Percentiles: 50th
Median: 6.07
Range (Min, Max): 4.11,11.62
Whittier
Mean (SD): 3.93 (1.49) 50th(Median):
3.76
Range (Min, Max): 1.64, 8.82
Monitoring Stations: 48 personal
nephelometers,
2 central sites
Copollutant (correlation):
Personal
24-h personal PM2.6 0.15
24-h personal EC 0.41
24-h personal 0C 1.00
24-h personal NO2 0.20
24-h central PM2.6 -0.11
24-h central EC 0.03
24-h central 0C -0.02
24-h central NO2 0.21
Central
24-h personal PM2.6 0.21
24-h personal EC -0.01
24-h personal 0C -0.02
24-h personal NO2 0.17
24-h central PM2.6 0.66
24-h central EC 0.87
24-h central 0C 1.00
24-h central NO2 0.62
pollutant exposure and FEN0
Lag 0
Personal 0.51 (-0.28,1.30)
Central 0.93 (-0.20, 2.06)
Lag 1
Personal 0.13 (-0.77,1.03)
Central0.51 (-0.64,1.66)
2-day MA
Personal 0.94 (-0.47, 2.35)
Central 1.6 (-0.17, 3.37)
Stratified by Medication Use
Lag - 2-day moving avg.
Not Taking Anti-lnflamm. Medication
Personal 0.88 (-1.62, 3.38)
Central 0.36 (-4.07,4.79)
Taking Anti-lnflamm. Medication
Personal 0.87 (-0.79, 2.53)
Central 2.05 (0.24, 3.86)
Inhaled Corticosteroids
Personal 2.47 (0.30, 4.64)
Central 1.96 (0.14, 3.78)
Antileukotrienes +- inhaled
corticosteroids
Personal 0.52 (-1.99, 3.02)
Central 1.29 (-2.58, 5.15)
Notes:
Figure of Estimated lag effect of hourly
personal PM2.6 on FEN0.
Figure of the Estimated lag effect of
hourly personal PM2.6 on FEN0 by use of
medications.
Figure of One- and two-pollutant models
for change in FEN0 using 2-day Moving
Averages personal and central-site
pollutant measurements
Reference: Dubowsky et al (2006,
088750)
Period of Study: 3/2002-6/2002
Location: St. Louis, Missouri
Outcome: Chronic inflammation,
Diabetes, Obesity, Hypertension, Cardiac
Risk
Study Design:
Prospective Cohort
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD) unit: 16 (6.0)
Range (Min, Max): 6.5, 28
Copollutants:
Increment: 5.4/yg/m3
% Increase (Lower CI, Upper CI)
Lag
% increase in inflammatory response and
exposure to PM2.6 in people a 60
July 2009
E-207
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Statistical Analyses:
BC
Inflammatory Marker:
Poisson, LOESS
CO
IL-6: -8 (-16, 8)
Age Groups:
NO?
1: -6 (-10, 5)
> 60
SO?
2: -5 (-11, 6)

Os
3: -3 (-9, 6)


4: -4 (-12,10)


5: -5 (-13, 8)


6:-6 (-14, 9)


7


CRP: -2 (-22,15)


1: 3 (-8,17)


2: 4 (-9, 20)


3: 9 (-4, 27)


4: 11 (-5, 35)


5: 8 (-9, 29)


6: 5 (-12, 26)


7


WBC: 0 (-2, 4)


1: 1 (-1,2)


2:2(-1,3)


3: 1 (-2, 5)


4: 3 (-1,10)


5: 5(0,12)


6: 8(0,14)


7


% Increase in inflammatory responses


and exposure to ambient PM2.6


concentrations in people a 60


Inflammatory Marker:


CRP


All conditions*: 14 (-5.4, 37)


0-5 avg


3 conditions met*: 81 (21,172)


0-5 avg


2 conditions met*: 11 (-7.3, 33)


0-5 avg


IL-6


All conditions*: -2.1 (-13,11)


0-5 avg


3 conditions met*: 23 (-5.3, 59)


0-5 avg


2 conditions met*: -3.1 (-14, 9.7)


0-5 avg


WBC


All conditions*: 3.4 (-1.8, 8.9)


0-5 avg
July 2009
E-208
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
3 conditions met*: 0.4 (-8.8,11)
0-5 avg
2 conditions met*: 3.6 (-1.7, 9.1)
0-5 avg
* All conditions met means model is
adjusted for sex, obesity, diabetes,
smoking history, ambient and
microenvironmental apparent
temperature, mold, pollen, trip, h, and
vitamins.
Three conditions met means model is
adjusted for three of the variables.
Two conditions met means model is
adjusted for two of the variables.
Reference: Ebelt et al. (2005, 056907)
Period of Study: Summer of 1998
Location: Vancouver, Canada
Outcome: spirometry,
Age Groups: range from 54-86 yrs
mean age - 74 years
Study Design: extended analysis of a
repeated-measures panel study
N: 16 persons with COPD
Statistical Analyses: Earlier analysis
expanded by developing mixed-effect
regression models and by evaluating
additional exposure indicators
Dose-response Investigated? No
Statistical Package: SAS V8
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD):
Ambient PM2.6:11.4 (4.6)
Exposure to ambient PM2.6: 7.9 (3.7)
Nonsulfate ambient PM2.6: 9.3 (3.7)
Exposure to nonsulfate ambient PM2.6:
6.5 (3.0)
Total exposure to PM2.6: 18.5 (14.9)
Exposure to nonambient PM2.6:10.6
(14.5)
Range (Min, Max):
Ambient PM2.6: (4.2-28.7)
Exposure to ambient PM2.6: (0.9-21.3)
Nonsulfate ambient PM2.6: (3.3-23.3)
Exposure to nonsulfate ambient PM2.6:
(0.7-16.9)
Total exposure to PM2.6: (2.2-90.9)
Exposure to nonambient PM2.6: (-2.6-
85.0)
Monitoring Stations: 5
Copollutant (correlation):
Ambient PM10: r- 0.78
Ambient PMio-2.b: r- 0.15
Ambient Sulfate- 0.82
Nonsufate Ambient PM2.6: r- 0.98
PM Increment: Ambient PM2.6: 5.8 (IQR)
Exposure to ambient PM2.6: 4.4 (IQR)
Nonsulfate ambient PM2.6: 4.2 (IQR)
Exposure to nonsulfate ambient PM2.6:
3.4 (IQR)
Total exposure to PM2.6: 10.1 (IQR)
Exposure to nonambient PM2.6: 8.9 (IQR)
Notes: Effect estimates are presented in
Figure 2 and Electronic Appendix Table 1
(only available with electronic version of
article) and not provided quantitatively
elsewhere.
Reference: Ebelt et al. (2005, 056907) Outcome: spirometry
Period of Study: Summer of 1998
Location: Vancouver, Canada
Age Groups: Range from 54-86 yrs
mean age - 74 years
Study Design: extended analysis of a
repeated-measures panel study
N: 16 persons with COPD
Statistical Analyses: Earlier analysis
expanded by developing mixed-effect
regression models and by evaluating
additional exposure indicators
Dose-response Investigated? No
Pollutant: Sulfate (SO4)
Averaging Time: 24 h
Mean (SD):
Ambient Sulfate: 2.0 (1.1)
Exposure to Ambient Sulfate: 0.2 (4.7)
Range (Min, Max):
Ambient Sulfate: (0.4-5.4)
Exposure to ambient Sulfate: (0.2-4.7)
Monitoring Stations: 5
Copollutant (correlation):
PM Increment: Ambient Sulfate: 1.5
(IQR)
Exposure to Ambient Sulfate: 0.9 (IQR)
Notes: Effect estimates are presented in
Figure 2 and Electronic Appendix Table 1
(only available with electronic version of
article) and not provided quantitatively
elsewhere.
July 2009
E-209
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)

Statistical Package: SAS V8
Ambient PM2.6: r- 0.82



Nonsulfate Ambient PM2.6: r- 0.74



Exposure to Ambient Sulfate: r- 0.82

Reference: Ferdinands et al. (2008,
Outcome: Respiratory Symptoms, airway
Pollutant: PM2.6
The study presents results qualitatively
156433)
inflammation
Averaging Time: 24-h avg
not quantitatively.


Period of Study: 8/16/2004-8/31/2004
Study Design: Prospective cohort
Mean (SD) unit: 27.2 (11.9)


Location: Atlanta, Georgia
Statistical Analyses: Pearson
Range (Min, Max): 21.7, 34.7

Correlation Analysis


Age Groups: 14-18
Copollutants (correlation):



O3: r- 0.8-0.9

Reference: Gent et al. (2003, 052885)
Outcome: Respiratory symptoms
Pollutant: PM2.6
PM Increment: 12 /yg/m3 same day

including: Wheeze, persistent cough,

19/yg/m3 previous day
Period of Study: April 1 through
chest tightness, shortness of breath
Averaging Time: 24 h
September 30, 2001


Age Groups: Infants
Mean (SD): 13.1 (7.9)
Model 5 (same day)
Location: Connecticut


Springfield, MA
Study Design: 1 -year prospective cohort
Percentiles: 20th: 6.9
Wheeze <6.9 - 1.00
study
40th: 9.0
6.9-8.9 - 0.95 (0.83,1.10)

N: 1002 infants
50th(Median): 10.3
9.0-12.0 - 1.04 (0.89,1.20)

17160 observations
60th: 12.1
12.1-18.9 - 1.05(0.92,1.20)

Statistical Analyses: Logistic
80th: 19.0
> 19.0 - 0.93(0.78,1.11)

regression analysis


GEEs
Range (Min, Max): 3.7, 44.2
Persistent Cough <6.9 — 1.00

Tests for linear trend
Monitoring Stations: 4 sites
6.9-8.9 - 0.95 (0.87,1.04)

Test for goodness of fit
Copollutant (correlation):
9.0-12.0 - 0.96 (0.87,1.06)

Hosmer-Lemeshow statistic for
Temperature: 0.58
12.1-18.9 - 1.00(0.91,1.09)

regression

> 19.0 - 0.95(0.83,1.09)

Covariates: Temperature

Chest Tightness <6.9 - 1.00

Dose-response Investigated? No

6.9-8.9 - 1.01 (0.86,1.19)

Statistical Package: SAS

9.0-12.0 - 1.06 (0.89,1.26)

Lags Considered: 1-day lag

12.1-18.9 - 1.24(1.06,1.45)



> 19.0 - 1.05(0.84,1.33)



Shortness of Breath < 6.9 - 1.00



6.9-8.9 - 1.01 (0.87,1.17)



9.0-12.0 - 1.03 (0.87,1.22)



12.1-18.9 - 1.07 (0.91,1.25)



> 19.0 - 1.03(0.83,1.28)



Bronchodilator <6.9 - 1.00



6.9-8.9 - 1.04 (0.99,1.09)



9.0-12.0 - 1.02 (0.96,1.08)



12.1-18.9 - 1.04(0.99,1.09)



> 19.0 - 1.02(0.97,1.08)



Model 6 (previous day)



Wheeze <6.9 - 1.00



6.9-8.9 - 1.06 (0.95,1.20)



9.0-12.0 - 1.09 (0.94,1.28)



12.1-18.9 - 1.03(0.89,1.19)



> 19.0 - 1.14(0.97,1.34)



Persistent Cough <6.9 — 1.00



6.9-8.9 - 1.04 (0.94,1.14)
July 2009
E-210
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
9.0-12.0 - 1.05 (0.94,1.17)
12.1-18.9 - 1.03(0.94,1.14)
>	19.0 - 1.12(1.02 1.24)
Chest Tightness <6.9 - 1.00
6.9-8.9 - 1.03 (0.87,1.23)
9.0-12.0 - 1.04 (0.85,1.27)
12.1-18.9 - 1.00(0.84,1.19)
>	19.0 - 1.21 (1.00,1.46)
Shortness of Breath <6.9 - 1.00
6.9-8.9 - 1.00 (0.84,1.19)
9.0-12.0 - 1.09 (0.90,1.31)
12.1-18.9 - 1.09(0.90,1.31)
>	19.0 - 1.26(1.02,1.54)
Bronchodilator <6.9 - 1.00
6.9-8.9 - 0.98 (0.94,1.03)
9.0-12.0 - 0.99 (0.95,1.03)
12.1-18.9 - 0.97 (0.94,1.01)
>	19.0 - 0.99(0.95,1.04)
PM2.5 + O3: Medication Users: Same-
day
Wheeze < 6.9 - 1.00
6.9-8.9 - 0.89 (0.75,1.29)
9.0-12.0 - 1.02 (0.87,1.19)
12.1-18.9 - 0.94 (0.77,1.15)
>	19.0 - 0.83(0.65,1.06)
Persistent Cough <6.9 - 1.00
6.9-8.9 - 0.95 (0.84,1.06)
9.0-12.0 - 0.97 (0.86,1.10)
12.1-18.9 - 0.94 (0.77,1.15)
>	19.0 - 0.83(0.65,1.06)
Chest Tightness <6.9 - 1.00
6.9-8.9 - 0.90 (0.74,1.09)
9.0-12.0 - 0.97 (0.79,1.18)
12.1-18.9 - 0.97 (0.76,1.25)
>	19.0 - 0.76(0.54,1.05)
Shortness of Breath <6.9 - 1.00
6.9-8.9 - 0.95 (0.80,1.12)
9.0-12.0 - 1.00 (0.82,1.21)
12.1-18.9 - 0.90(0.73,1.12)
>	19.0 - 0.87(0.65,1.17)
Bronchodilator <6.9 - 1.00
6.9-8.9 - 1.03 (0.98,1.08)
9.0-12.0 - 1.01 (0.96,1.07)
12.1-18.9 - 1.02(0.95,1.08)
>	19.0 - 0.99(0.91,1.07)
July 2009
E-211
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Previous Day
Wheeze < 6.9 - 1.00
6.9-8.9 - 1.03 (0.89,1.18)
9.0-12.0 - 1.05 (0.88,1.24)
12.1-18.9 - 0.98(0.82,1.17)
>	19.0 - 1.05(0.85,1.29)
Persistent Cough <6.9 - 1.00
6.9-8.9 - 0.99 (0.89,1.11)
9.0-12.0 - 0.98 (0.86,1.10)
12.1-18.9 - 0.95(0.83,1.10)
>	19.0 - 1.00(0.88,1.15)
Chest Tightness <6.9 - 1.00
6.9-8.9 - 0.89 (0.72,1.10)
9.0-12.0 - 0.90 (0.70,1.16)
12.1-18.9 - 0.81 (0.63,1.03)
>19.0 - 0.91 (0.71,1.17)
Shortness of Breath <6.9 - 1.00
6.9-8.9 - 0.96 (0.78,1.18)
9.0-12.0 - 1.00 (0.81,1.25)
12.1-18.9 - 0.96(0.74,1.24)
>	19.0 - 1.20(0.94,1.52)
Bronchodilator <6.9 - 1.00
6.9-8.9 - 0.99 (0.94,1.04)
9.0-12.0 - 0.97 (0.93,1.02)
12.1-18.9 - 0.96(0.91,1.02)
>	19.0 - 0.97(0.89,1.04)
PM2.5 + O3: Non-users: Same-day
Wheeze < 6.9 - 1.00
6.9-8.9 - 0.92 (0.72,1.17)
9.0-12.0 - 1.08 (0.85,1.36)
12.1-18.9 - 0.94 (0.73,1.22)
>	19.0 - 1.15(0.75,1.75)
Persistent Cough <6.9 - 1.00
6.9-8.9 - 0.96 (0.83,1.12)
9.0-12.0 - 1.02 (0.89,1.18)
12.1-18.9 - 0.93(0.78,1.12)
>	19.0 - 1.07(0.85,1.34)
Chest Tightness <6.9 - 1.00
6.9-8.9 - 0.84 (0.54,1.31)
9.0-12.0 - 1.09 (0.74,1.61)
12.1-18.9 - 0.78(0.47,1.30)
>	19.0 - 0.71 (0.36,1.39)
Shortness of Breath < 6.9 - 1.00
6.9-8.9 - 0.61 (0.39,0.95)
9.0-12.0 - 1.13(0.85,1.50)
July 2009
E-212
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
12.1-18.9 - 0.72(0.42,1.23)
>	19.0 - 1.17(0.72,1.90)
Bronchodilator Use: <6.9 - 1.00
6.9-8.9 - 0.95 (0.78,1.15)
9.0-12.0 - 0.95 (0.78,1.16)
12.1-18.9 - 0.85(0.69,1.06)
>	19.0 - 0.99(0.76,1.30)
Previous-day
Wheeze < 6.9 - 1.00
6.9-8.9 - 1.01 (0.78,1.31)
9.0-12.0 - 1.15(0.88,1.51)
12.1-18.9 - 1.08(0.78,1.51)
>19.0 - 1.18(0.71,1.97)
Persistent Cough <6.9 - 1.00
6.9-8.9 - 1.07 (0.94,1.22)
9.0-12.0 - 1.13(0.97,1.32)
12.1-18.9 - 1.03(0.87,1.22)
>	19.0 - 1.14(0.88,1.46)
Chest Tightness <6.9 - 1.00
6.9-8.9 - 1.44(0.90,2.30)
9.0-12.0 - 1.50 (0.97, 2.33)
12.1-18.9 - 1.56(0.91,2.66)
>	19.0 - 1.76(0.83, 3.73)
Shortness of Breath < 6.9 - 1.00
6.9-8.9 - 0.99 (0.75,1.30)
9.0-12.0 - 1.30 (0.88,1.91)
12.1-18.9 - 0.84 (0.57,1.24)
>	19.0 - 1.48(0.94,2.34)
Bronchodilator Use <6.9 - 1.00
6.9-8.9 - 1.05 (0.85,1.34)
9.0-12.0 - 1.28(1.01,1.62)
12.1-18.9 - 1.05(0.80,1.37)
>19.0 - 1.19(0.83,1.71)
Notes: Line graphs of daily levels of
ozone and PM2.6 and daily temperature
with daily prevalence of respiratory
symptoms for users of asthma
maintenance medication
Reference: Gent et al, (2009,180399)
Period of Study: 2000-2003
Location: New Haven County CT
Outcome: Increased asthma symptoms
and medication use
Study Design: Panel
Covariates: Season, day of the week,
date
Statistical Analysis: Logistic regression
Statistical Package: SAS
Age Groups: Children aged 4-12
Pollutant: PM2 Eand components
Averaging Time: Daily
Mean: (estimated sources, /yg/m3)
Motor Vehicle: 6.6
Road Dust: 2.3
Sulfur: 5.5
Biomass Burning: 0.9
Oil: 0.8
Sea Salt: 0.5
Odds Patio and p-value for sources
and components of PM2.5. Lags are 0,
1 or 2 days, and the mean of days 0-2
(L02).
Source: Motor Vehicle
Elemental Carbon, Increment = 1QQQ
ng/tri3
Wheeze
L0: 1.04, p - 0.04
L1: 1.01, p - 0.70
July 2009
E-213
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Range (Min, Max): NR	L2: 1.00, p - 0.99
Copollutant (correlation): NR	L02: 1.07, p - 0.06
Persistent Cough
L0: 1.01, p - 0.42
L1: 1.01, p - 0.38
L2: 0.99, p - 0.44
L02: 1.03, p - 0.23
Shortness of Breath
L0: 1.06, p - 0.001
L1: 1.01, p - 0.65
L2: 1.01, p - 0.63
L02: 1.12, p - 0.01
Chest Tightness
L0: 1.03, p - 0.20
L1: 1.02, p - 0.24
L2: 1.01, p - 0.59
L02: 1.10, p - 0.04
Inhaler Use
L0: 1.01, p - 0.15
L1: 1.00, p - 0.72
L2: 1.00, p - 0.75
L02: 1.02, p - 0.40
Zn, Increment = 10 ng/ni3
Wheeze
L0: 1.00, p - 0.69
L1: 0.99, p - 0.54
L2: 1.00, p - 0.89
L02: 1.00, p - 0.98
Persistent Cough
L0: 1.00, p - 0.60
L1: 1.00, p - 0.77
L2: 0.99, p - 0.24
L02: 1.00, p - 0.94
Shortness of Breath
L0: 1.02, p - 0.001
L1: 1.00, p - 0.57
L2: 1.01, p - 0.49
L02: 1.04, p - 0.06
Chest Tightness
L0: 1.00, p - 0.72
L1: 1.00, p - 0.96
L2: 1.01, p - 0.38
L02: 1.03, p - 0.13
Inhaler Use
L0: 1.00, p - 0.41
July 2009
E-214
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
L1: 1.00, p - 0.44
L2: 1.00, p - 0.52
L02: 1.01, p - 0.53
Pb, Increment = 5 ng/nf
Wheeze
L0: 1.02, p - 0.31
L1: 1.00, p - 0.91
L2: 1.01, p - 0.62
L02: 1.07, p - 0.13
Persistent Cough
L0: 1.02, p - 0.25
L1: 1.00, p - 0.88
L2: 1.00, p - 0.87
L02: 1.05, p - 0.12
Shortness of Breath
L0: 1.03, p - 0.11
L1: 0.98, p - 0.51
L2: 1.03, p - 0.05
L02: 1.12, p - 0.01
Chest Tightness
L0: 1.02, p - 0.31
L1: 0.99, p - 0.79
L2: 1.03, p - 0.13
L02: 1.10, p - 0.02
Inhaler Use
L0: 1.01, p - 0.06
L1: 0.98, p - 0.11
L2: 1.02, p - 0.04
L02: 1.04, p - 0.10
Cu, Increment = 5 ng/nf
Wheeze
L0: 1.01, p - 0.59
L1: 0.99, p - 0.55
L2: 0.99, p - 0.82
L02: 1.02, p - 0.67
Persistent Cough
L0: 1.02, p - 0.13
L1: 1.02, p - 0.21
L2: 0.98, p - 0.26
L02: 1.05, p - 0.04
Shortness of Breath
L0: 1.06, p - 0.01
L1: 1.01, p - 0.74
L2: 0.96, p - 0.10
L02: 1.06, p - 0.21
July 2009
E-215
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Chest Tightness
LO: 10.3, p - 0.23
L1: 1.02, p - 0.42
L2: 0.97, p - 0.17
L02: 1.04, p - 0.39
Inhaler Use
LO: 1.01, p - 0.22
L1: 0.99, p - 0.37
L2: 1.00, p - 0.70
L02: 1.01, p - 0.46
Se, Increment = 7 ng/ni3
Wheeze
L0: 1.00, p - 0.97
L1: 0.99, p - 0.52
L2: 1.00, p - 0.91
L02: 1.02, p - 0.71
Persistent Cough
L0: 1.00, p - 0.84
L1: 0.99, p - 0.32
L2: 1.00, p - 0.93
L02: 0.98, p - 0.43
Shortness of Breath
L0: 1.02, p - 0.40
L1: 0.97, p - 0.10
L2: 1.01, p - 0.55
L02: 1.02, p - 0.67
Chest Tightness
LO: 1.00, p - 0.79
L1: 0.97, p - 0.13
L2: 1.01, p - 0.72
L02: 0.98, p - 0.61
Inhaler Use
LO: 0.99, p - 0.20
L1: 1.01, p - 0.02
L2: 0.99, p - 0.32
L02: 0.99, p - 0.75
Source: Road Dust
Si, Increment = lOOng/rri3
Wheeze
LO: 1.03, p - 0.03
L1: 1.00, p - 0.99
L2: 1.02, p - 0.26
L02: 1.07, p - 0.04
Persistent Cough
LO: 1.02, p - 0.01
July 2009
E-216
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
L1: 1.00, p - 0.78
L2: 1.01, p - 0.60
L02: 1.05, p - 0.02
Shortness of Breatl .04, p - 0.01 h
L0: 1.04, p - 0.01
L1: 1.01, p - 0.60
L2: 1.01, p - 0.63
L02: 1.08, p - 0.02
Chest Tightness
LO: 1.02, p - 0.20
L1: 1.02, p - 0.17
L2: 1.00, p - 0.88
L02: 1.06, p - 0.10
Inhaler Use
LO: 1.02, p - 0.004
L1: 0.99, p - 0.18
L2: 1.01, p - 0.45
L02: 1.03, p - 0.09
fe, Increment = 100 ng/nf
Wheeze
L0: 1.04, p - 0.02
L1: 1.00, p - 0.80
L2: 1.00, p - 0.87
L02: 1.07, p - 0.05
Persistent Cough
L0: 1.02, p - 0.06
L1: 1.01, p - 0.52
L2: 0.99, p - 0.52
L02: 1.04, p - 0.04
Shortness of Breath
L0: 1.06, p - 0.002
L1: 1.01, p - 0.65
L2: 0.98, p - 0.27
L02: 1.08, p - 0.04
Chest Tightness
LO: 1.01, p - 0.47
L1: 1.02, p - 0.22
L2: 0.98, p - 0.35
L02: 1.05, p - 0.21
Inhaler Use
LO: 1.02, p - 0.004
L1: 0.99, p - 0.44
L2: 1.00, p - 0.91
L02: 1.03, p - 0.08
At, Increment = 50 ng/m3
July 2009
E-217
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Wheeze
LO: 1.02, p - 0.17
L1: 1.01, p - 0.73
L2: 1.02, p - 0.30
L02: 1.07, p - 0.03
Persistent Cough
L0: 1.03, p - 0.001
L1: 1.00, p - 0.96
L2: 1.00, p - 0.68
L02: 1.06, p - 0.01
Shortness of Breath
LO: 1.05, p - 0.002
L1: 1.01, p - 0.63
L2: 1.01, p - 0.59
L02: 1.09, p - 0.004
Chest Tightness
LO: 1.02, p - 0.21
L1: 1.02, p - 0.18
L2: 1.00, p - 0.94
L02: 1.07, p - 0.04
Inhaler Use
LO: 1.02, p - 0.02
L1: 0.99, p - 0.27
L2: 1.01, p - 0.50
L02: 1.02, p - 0.11
Ca, Increment = 50 ng/m3
Wheeze
L0: 1.07, p - 0.02
L1: 1.00, p - 0.97
L2: 1.01, p - 0.74
L02: 1.14, p - 0.04
Persistent Cough
L0: 1.05, p - 0.01
L1: 0.99, p - 0.64
L2: 1.00, p - 0.90
L02: 1.09, p - 0.03
Shortness of Breath
LO: 1.10, p - 0.002
L1: 1.02, p - 0.66
L2: 1.00, p - 0.89
L02: 1.18, p - 0.01
Chest Tightness
LO: 1.04, p - 0.26
L1: 1.03, p - 0.43
L2: 1.00, p - 0.93
July 2009
E-218
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
L02: 1.14, p - 0.07
Inhaler Use
L0: 1.04, p - 0.01
L1: 0.97, p - 0.06
L2: 1.01, p - 0.44
L02: 1.04, p - 0.17
Ba, Increment = Wng/m3
Wheeze
L0: 0.99, p - 0.57
L1: 1.00, p - 0.92
L2: 0.99, p - 0.48
L02: 0.99, p - 0.81
Persistent Cough
L0: 1.00, p - 0.83
L1: 1.01, p - 0.38
L2: 0.99, p - 0.32
L02: 1.00, p - 0.81
Shortness of Breath
L0: 1.04, p - 0.02
L1: 1.00, p - 0.96
L2: 0.96, p - 0.05
L02: 1.03, p - 0.38
Chest Tightness
L0: 1.01, p - 0.63
L1: 1.00, p - 0.88
L2: 0.98, p - 0.30
L02: 1.02, p - 0.51
Inhaler Use
L0: 1.01, p - 0.08
L1: 0.99, p - 0.19
L2: 1.00, p - 0.92
L02: 1.01, p - 0.36
Ti, Increment = 5 ng/m3
Wheeze
L0: 1.00, p - 0.59
L1: 0.99, p - 0.49
L2: 1.01, p - 0.34
L02: 1.01, p - 0.56
Persistent Cough
L0: 1.00, p - 0.57
L1: 1.00, p - 0.55
L2: 1.00, p - 0.30
L02: 1.01, p - 0.29
Shortness of Breath
L0: 1.01, p - 0.01
July 2009
E-219
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
L1: 1.00, p - 0.56
L2: 1.00, p - 0.60
L02: 1.03, p - 0.05
Chest Tightness
L0: 1.00, p - 0.34
L1: 1.00, p - 0.55
L2: 0.99, p - 0.49
L02: 1.01, p - 0.52
Inhaler Use
LO: 1.00, p - 0.72
L1: 1.00, p - 0.30
L2: 1.00, p - 0.67
L02: 1.00, p - 0.66
Source: Sulfur
S, Increment = lOOOngltr?
Wheeze
L0: 0.98, p - 0.43
L1: 0.99, p - 0.62
L2: 1.02, p - 0.29
L02: 1.00, p - 0.99
Persistent Cough
LO: 1.00, p - 0.84
L1: 1.00, p - 0.69
L2: 1.02, p - 0.21
L02: 1.02, p - 0.27
Shortness of Breath
LO: 1.01, p - 0.63
L1: 0.99, p - 0.71
L2: 1.01, p - 0.55
L02: 1.01, p - 0.79
Chest Tightness
LO: 0.99, p - 0.80
L1: 1.01, p - 0.62
L2: 1.01, p - 0.81
L02: 1.02, p - 0.68
Inhaler Use
LO: 0.99, p - 0.13
L1: 1.00, p - 0.81
L2: 1.02, p - 0.04
L02: 1.00, p - 0.81
P, Increment = 50 ng/rri3
Wheeze
LO: 0.98, p - 0.39
L1: 0.98, p - 0.48
L2: 1.02, p - 0.38
July 2009
E-220
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
L02: 0.99, p - 0.89
Persistent Cough
L0: 1.00, p - 0.75
L1: 0.99, p - 0.69
L2: 1.01, p - 0.38
L02: 1.03, p - 0.30
Shortness of Breath
L0: 1.01, p - 0.61
L1: 0.99, p - 0.71
L2: 1.01, p - 0.67
L02: 1.01, p - 0.78
Chest Tightness
LO: 1.00, p - 0.88
L1: 1.01, p - 0.72
L2: 1.00, p - 0.87
L02: 1.02, p - 0.67
Inhaler Use
LO: 0.98, p - 0.15
L1: 1.00, p - 0.83
L2: 1.01, p - 0.11
L02: 1.00, p - 0.99
Source: Biomass Burning
K, Increment = 50 ng/ni3
Wheeze
LO: 0.98, p - 0.06
L1: 0.99, p - 0.43
L2: 1.00, p - 0.85
L02: 0.96, p - 0.04
Persistent Cough
LO: 1.00, p - 0.64
L1: 1.00, p - 0.83
L2: 1.00, p - 0.46
L02: 1.00, p - 0.86
Shortness of Breath
LO: 1.01, p - 0.01
L1: 0.98, p - 0.09
L2: 1.00, p - 0.38
L02: 1.00, p - 0.79
Chest Tightness
LO: 1.00, p - 0.02
L1: 0.99, p - 0.24
L2: 0.98, p - 0.07
L02: 0.99, p - 0.67
Inhaler Use
LO: 1.00, p - 0.68
July 2009
E-221
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
L1: 0.99, p - 0.05
L2: 1.00, p - 0.59
L02: 0.99, p - 0.28
Source: Oil
V, Increment = lOng/ni3
Wheeze
L0: 0.99, p - 0.73
L1: 0.96, p - 0.03
L2: 0.99, p - 0.56
L02: 0.93, p - 0.04
Persistent Cough
L0: 1.01, p - 0.56
L1: 0.99, p - 0.24
L2: 0.98, p - 0.01
L02: 0.96, p - 0.05
Shortness of Breath
L0: 1.01, p - 0.46
L1: 0.98, p - 0.24
L2: 1.00, p - 0.83
L02: 0.98, p - 0.58
Chest Tightness
LO: 0.99, p - 0.71
L1: 0.98, p - 0.32
L2: 0.98, p - 0.23
L02: 0.94, p - 0.12
Inhaler Use
LO: 0.98, p - 0.12
L1: 1.00, p - 0.68
L2: 0.99, p - 0.22
L02: 0.96, p - 0.03
Ni, Increment = 5 ng/nf
Wheeze
L0: 1.01, p - 0.59
L1: 0.97, p - 0.09
L2: 1.00, p - 0.76
L02: 0.99, p - 0.72
Persistent Cough
LO: 1.01, p - 0.21
L1: 0.99, p - 0.57
L2: 0.99, p - 0.23
L02: 1.00, p - 0.99
Shortness of Breath
LO: 1.04, p - 0.05
L1: 0.98, p - 0.36
L2: 1.00, p - 0.81
July 2009
E-222
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
L02: 1.04, p - 0.32
Chest Tightness
L0: 1.01, p - 0.58
L1: 1.00, p - 0.89
L2: 0.98, p - 0.27
L02: 1.01, p - 0.84
Inhaler Use
LO: 1.01, p - 0.48
L1: 1.00, p - 0.83
L2: 0.99, p - 0.51
L02: 1.01, p - 0.48
Source: Sea Salt
Na, Increment = 100 ng/nf
Wheeze
L0: 0.98, p - 0.23
L1: 1.00, p - 0.80
L2: 1.00, p - 0.88
L02: 0.97, p - 0.29
Persistent Cough
L0: 1.00, p - 0.58
L1: 0.99, p - 0.19
L2: 1.00, p - 0.61
L02: 0.98, p - 0.21
Shortness of Breath
L0: 1.00, p - 0.94
L1: 0.99, p - 0.46
L2: 1.01, p - 0.63
L02: 0.99, p - 0.74
Chest Tightness
LO: 0.99, p - 0.43
L1: 0.99, p - 0.75
L2: 1.00, p - 0.88
L02: 0.98, p - 0.61
Inhaler Use
LO: 0.99, p - 0.35
L1: 1.00, p - 0.61
L2: 1.00, p - 0.85
L02: 0.99, p - 0.37
CI, Increment = 10 ng/nf
Wheeze
L0: 1.00, p - 0.89
L1: 1.00, p - 0.88
L2: 1.00, p - 0.38
L02: 1.00, p - 0.81
Persistent Cough
July 2009
E-223
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
LO: 1.00, p - 0.31
L1: 1.00, p - 0.31
L2: 1.00, p - 0.51
L02: 1.00, p - 0.06
Shortness of Breath
L0: 1.00, p - 0.89
L1: 1.00, p - 0.94
L2: 1.00, p - 0.70
L02: 1.00, p - 0.80
Chest Tightness
LO: 1.00, p - 0.24
L1: 1.00, p - 0.28
L2: 1.00, p - 0.52
L02: 1.00, p - 0.65
Inhaler Use
LO: 1.00, p - 0.69
L1: 1.00, p - 0.52
L2: 1.00, p - 0.51
L02: 1.00, p - 0.83
Odds Ratio (95%CI) from repeated
measures logistic regression models
of respiratory symptoms and daily
source concentrations of PM2.5.
Lag 0 Model
Wheeze, p - 0.23
Motor Vehicle: 1.05 (0.99-1.10)
Road Dust: 1.10(1.01-1.19)
Sulfur: 0.97 (0.94-1.00)
Biomass Burning: 0.80 (0.66-0.98)
Oil: 1.02(0.86-1.20)
Sea Salt: 0.96 (0.86-1.07)
Persistent Cough, p < 0.001
Motor Vehicle: 1.02 (0.99-1.04)
Road Dust: 1.06(1.01-1.11)
Sulfur: 1.00 (0.98-1.01)
Biomass Burning: 0.97 (0.92-1.03)
Oil: 1.02(0.95-1.10)
Sea Salt: 0.99 (0.92-1.07)
Shortness of Breath, p < 0.001
Motor Vehicle: 1.06 (1.01-1.11)
Road Dust: 1.12 (1.02-1.22)
Sulfur: 0.98 (0.94-1.02)
Biomass Burning: 1.05 (0.95-1.17)
Oil: 1.07 (0.92-1.26)
Sea Salt: 1.01 (0.92-1.12)
Chest Tightness, p < 0.001
July 2009
E-224
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Motor Vehicle: 1.02 (0.97-1.08)
Road Dust: 1.04 (0.95-1.15)
Sulfur: 0.99 (0.94-1.03)
Biomass Burning: 1.06 (0.95-1.18)
Oil: 0.99(0.82-1.18)
Sea Salt: 0.95 (0.84-1.08)
Inhaler Use, p < 0.001
Motor Vehicle: 1.02 (1.00-1.05)
Road Dust: 1.06(1.02-1.11)
Sulfur: 0.98 (0.97-1.00)
Biomass Burning: 1.00 (0.96-1.03)
Oil: 0.98(0.91-1.05)
Sea Salt: 0.99 (0.94-1.04)
Lag 02 Model
Wheeze, p - 0.86
Motor Vehicle: 1.10(1.01-1.19)
Road Dust: 1.26 (1.05-1.51)
Sulfur: 0.98 (0.92-1.04)
Biomass Burning: 0.64 (0.46-0.88)
Oil: 0.80(0.56-1.08)
Sea Salt: 0.91 (0.82-1.16)
Persistent Cough, p < 0.001
Motor Vehicle: 1.03 (0.98-1.09)
Road Dust: 1.16 (1.02-1.32)
Sulfur: 1.01 (0.98-1.05)
Biomass Burning: 0.93 (0.81-1.06)
Oil: 0.84 (0.71-1.00)
Sea Salt: 0.88 (0.77-1.01)
Shortness of Breath, p - 0.006
Motor Vehicle: 1.12(1.01-1.24)
Road Dust: 1.28 (1.05-1.55)
Sulfur: 0.97 (0.90-1.04)
Biomass Burning: 0.78 (0.52-1.18)
Oil: 0.94 (0.69-1.29)
Sea Salt: 1.01 (0.79-1.29)
Chest Tightness, p - 0.39
Motor Vehicle: 1.08 (0.98-1.20)
Road Dust: 1.20(0.97-1.49)
Sulfur: 1.00 (0.92-1.08)
Biomass Burning: 0.87 (0.62-1.22)
Oil: 0.80(0.58-1.10)
Sea Salt: 0.95 (0.71-1.27)
Inhaler Use, p < 0.001
Motor Vehicle: 1.03 (0.98-1.08)
Road Dust: 1.09(1.00-1.19)
July 2009
E-225
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Sulfur: 1.00 (0.97-1.03)
Biomass Burning: 0.95 (0.87-1.04)
Oil: 0.92(0.81-1.05)
Sea Salt: 0.97 (0.88-1.07)
Odds Ratio (95%CI) from repeated
measures logistic regression models
of respiratory symptoms and daily
source concentrations of PM2.5 when
copollutants are included.
Wheeze
Motor Vehicle
NO?: 1.03 (0.98-1.08)
CO: 1.05 (0.99-1.11)
SO2:1.04 (0.99-1.09)
0a: 1.06 (0.97-1.16)
Road Dust
NO?: 1.11 (1.02-1.20)
CO: 1.10(1.01-1.19)
SO2:1.10 (1.01-1.19)
O3:1.11 (1.01-1.23)
Sulfur
NO2: 0.96 (0.92-0.99)
CO: 0.97 (0.94-1.01)
SO2: 0.97 (0.93-1.00)
0a: 0.95 (0.91-1.00)
Biomass Burning
NO2: 0.79 (0.65-0.98)
CO: 0.80 (0.66-0.98)
SO2: 0.79 (0.64-0.98)
0a: 0.74 (0.57-0.97)
Oil
NO2: 1.02 (0.87-1.21)
CO: 1.02 (0.86-1.20)
SO2:1.01 (0.86-1.19)
0a: 0.92 (0.62-1.39)
Sea Salt
NO2: 0.96 (0.85-1.07)
CO: 0.96 (0.86-1.08)
SO2: 0.95 (0.85-1.07)
O3:1.01 (0.72-1.40)
Inhaler Use
Motor Vehicle
NO2: 1.02 (0.99-1.04)
CO: 1.02 (0.99-1.05)
SO2:1.02 (0.99-1.04)
0a: 1.02 (0.98-1.07)
Road Dust
July 2009
E-226
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
NO?: 1.06 (1.02-1.10)
CO: 1.06 (1.02-1.11)
SO2:1.06 (1.02-1.11)
03:1.06 (1.00-1.13)
Sulfur
NO?: 0.98 (0.96-1.00)
CO: 0.98 (0.96-1.00)
SO2: 0.98 (0.96-1.00)
Oa: 0.97 (0.95-1.00)
Biomass Burning
NO2: 1.00 (0.96-1.03)
CO: 0.99 (0.96-1.03)
SO2: 0.99 (0.96-1.03)
03:0.99 (0.95-1.03)
Oil
NO2: 0.98 (0.91-1.05)
CO: 0.97 (0.91-1.04)
SO2: 0.97 (0.91-1.04)
Os: 1.03 (0.88-1.22)
Sea Salt
NO2: 0.99 (0.94-1.04)
CO: 0.99 (0.94-1.04)
SO2: 0.99 (0.94-1.04)
Os: 1.01 (0.88-1.15)
Reference: Girardot et al. (2006,
088271)
Period of Study: 10 August 2002-16
October 2002
17 June 2003-27 August 2003
Location: Charlies Bunion Trail (portion
of Appalachia Trail)
Outcome: Pulmonary
function/spirometry-FVC, FEVi, PEF,
FVC/FEVi, FEF2B-7B
Age Groups: 18-82 yrs
Study Design: Cohort
N: 354 hikers
Statistical Analyses: Multiple linear
regression
Covariates: Age, h hiked, mean	Estimated personal:
temperature, sex, smoking status, history	r
of asthma or wheeze symptoms, carriage 9 21, 41 g
of backpack, whether reaching summit or
not
Pollutant: PM2.6
Averaging Time: 24 h
Mean:
Trail: 13.9 +1-8.2
Estimated personal: 15.0 +/¦ 7.4
Range (Min, Max):
Trail: 1.6, 38.4
Season: Fall 2002, Summer 2003
Dose-response Investigated? No
Statistical Package: SAS
Copollutant (correlation): O3 (r—0.67,
for estimated personal exposure)
PM Increment: 1 /yglm3
% Change +/¦ CI
p value
Univariate: FVC: 0.023 +/0.035
0.51
FEVi: 0.015 +/¦ 0.029
0.607
PEF: 0.185 +1- 0.091
0.043
FVC/FEVi: 0.003 +/¦ 0.023
0.905
FEF2B-7B*: 0.052 +/¦ 0.093
0.578
Adjusted: FVC: 0.007 +| 0.040
0.966
FEVi: 0.003 +/¦ 0.033
0.937
PEF: 0.258 +1- 0.103
0.013
FVC/FEVi:-0.011 +1-0.027
0.676
July 2009
E-227
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
FEF2B7B*: - 0.041 +1-0.109
0.707
Spirometry result for each quintile +/¦ CI
Quintile 1 (6.0 /yglm3): FVC (L): Prehike:
4.32 +/¦ 0.13
Posthike: 4.33 +1-0.12
FEVi (L): Prehike: 3.39 +1-0.10
Posthike: 3.40 +/¦ 0.10
FEVi/FVC (%): Prehike: 78.66 +/¦ 0.86
Posthike: 78.63 +/¦ 0.81
FEF2b-7b% (L/sec): Prehike: 3.27 +/¦ 0.14
Posthike: 3.26 +/¦ 0.14
PEF (Usee): Prehike: 7.91 +/ 0.22
Posthike: 7.58 +/¦ 0.22
Quintile 2 (10.4 //g/m3): FVC (L):
Prehike: 4.30 +/¦ 0.11
Posthike: 4.30 +/¦ 0.11
FEVi (L): Prehike: 3.42 +/¦ 0.09
Posthike: 3.43 +/¦ 0.09
FEVi/FVC (%): Prehike: 79.37 +/¦ 0.71
Posthike: 79.55 +/¦ 0.69
FEF2b-7b% (L/sec): Prehike: 3.39 +/¦ 0.14
Posthike: 3.38 +/¦ 0.14
PEF (Usee): Prehike: 8.37 +/ 0.23
Posthike: 8.26 +/¦ 0.25
Quintile 3 (14.8 //g/m3): FVC (L):
Prehike: 4.34 +/¦ 0.12
Posthike: 4.33 +1-0.12
FEVi (L): Prehike: 3.42 +1-0.10
Posthike: 3.40 +/¦ 0.09
FEVi/FVC (%): Prehike: 79.20 +/¦ 0.81
Posthike: 78.83 +/¦ 0.80
FEF2b-7b% (L/sec): Prehike: 3.19 +/¦ 0.13
Posthike: 3.21 +/-0.13
PEF (L/sec): Prehike: 8.12 +/0.25
Posthike: 7.89 +/¦ 0.25
Quintile 4 (17.9 /yg/m3):
FVC (L): Prehike: 4.23 +/-0.11
Posthike: 4.23 +/-0.11
FEVi (L): Prehike: 3.36 +/-0.10
Posthike: 3.36 +/¦ 0.10
FEVi/FVC (%): Prehike: 79.18 +/¦ 0.81
Posthike: 79.26 +/¦ 0.79
FEF2B-76* (L/sec): Prehike: 3.34 +/¦ 0.15
Posthike: 3.30 +/¦ 0.15
PEF (L/sec): Prehike: 7.75 +/ 0.25
July 2009
E-228
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Posthike: 7.73 +/¦ 0.26
Quintile 5 (25.6 //gfm3): FVC (L):
Prehike: 4.15 +/¦ 0.11
Posthike: 4.18 +1-0.12
FEVi (L): Prehike: 3.31 +/¦ 0.09
Posthike: 3.33 +/¦ 0.10
FEVi/FVC (%): Prehike: 79.73 +/¦ 0.66
Posthike: 79.55 +/¦ 0.64
FEF2b-7b% (L/sec): Prehike: 3.22 +/¦ 0.14
Posthike: 3.24 +/¦ 0.14
PEF (Usee): Prehike: 7.72 +f 0.22
Posthike: 7.77 +/¦ 0.23
Overall (15.0//gfm3): FVC (L): Prehike:
4.27 +/¦ 0.05
Posthike: 4.27 +/¦ 0.05
FEVi (L): Prehike: 3.38 +/¦ 0.04
Posthike: 3.38 +/¦ 0.04
FEVi/FVC (%): Prehike: 79.2 +/¦ 0.34
Posthike: 79.2 +/¦ 0.33
FEF2b-7b% (L/sec): Prehike: 3.28 +/¦ 0.06
Posthike: 3.28 +/¦ 0.06
PEF (Usee): Prehike: 7.97 +10.11
Posthike: 7.97 +1-0.11
Reference: Hertz-Picciotta et al. (2007,
135917)
Period of Study: 1994-2003
Location: Tepliee and Praehatiee, Czeeh
Republic
Outcome: Lower respiratory illness-
croup (J05, J04), acute bronchitis (J20),
acute bronchiolitis (J21)
Age Groups: Neonates followed for 2 to
4.5 yrs
Study Design: Cohort
N: 1133 children
Statistical Analyses: Generalized linear
longitudinal models
Covariates: District, mother's age,
mother's education, mother or adult
smoke, child's sex, season, day of the
week, fuel for heating and/or cooking,
breastfeeding category, number of other
children, temperature
Season: Winter, spring, summer and fall
Dose-response Investigated? No
Statistical Package: SUDAAN version 8
Lags Considered: 1-3,1-7,1-14,1-30,
1-45
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD):
PAH: 22.3 (SD-16 for 3-day avg and 11
for 45-day avg)
PM Increment: 25/yg/m3
RR Estimate [Lower CI, Upper CI]
lag:
Birth—23 months:
1.30[1.08,1.58] lag 1-30
2-4.5 yrs:
1.23(0.94,1.62] lag 1-30
RR Estimate for categories of exposure
[Lower CI, Upper CI]
Crude RR:
Birth—23 months:
>	50 //g/m3: 2.26 [1.81, 2.82] I
25-50//g/m3: 1.48 [1.32,1.65] li
<	25/yg/m3: Reference
2-4.5 yrs:
>	50 //g/m3: 3.66 [2.07, 6.48] I
25-50//g/m3: 1.60 [1.41,1.82] li
<	25/yg/m3: Reference
31-30
11-30
31-30
11-30
Reference: Hertz- Picciotta et al. (2007, Outcome: Lower respiratory illness- Pollutant: PM2.6
135917)	croup (J05, J04), acute bronchitis (J20),
acute bronchiolitis (J21)	Averaging Time: 24 h
Period of Study: 1994-2003
Age Groups: Neonates followed for 2 to Mean (SD):
Location: Tepliee and Praehatiee, Czech 4 5 yrs
Republic
PAH:
PAH Increment: 100 ng/m3
RR Estimate [Lower CI, Upper
lag:
Birth—23 months:
CI]
July 2009
E-229
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Study Design: Cohort
N: 1133 children
Statistical Analyses: Generalized linear
longitudinal models
Covariates: District, mother's age,
mother's education, mother or adult
smoke, child's sex, season, day of the
week, fuel for heating and/or cooking,
breastfeeding category, number of other
children, temperature
Dose-response Investigated? No
Statistical Package: SUDAAN version 8
Lags Considered: 1-3,1-7,1-14,1-30,
1-45
52.5 ng/m3 (SD-57 ng/m3 for 3-day avg
and 46 ng/m3 for 45-day avg)
1.29[1.07,1.54] lag 1-30
2-4.5 yrs:
1.5611.22, 2.00] lag 1-30
RR Estimate for categories of exposure
[Lower CI, Upper CI]
lag:
Crude RR:
Birth—23 months:
>	100 ng/m3: 2.52 (2.22, 2.87] lag 1-30
40-100 ng/m3:1.87 [1.65, 2.13] lag 1-30
<	40 ng/m3: Reference
2-4.5 yrs:
>	100 ng/m3: 2.26 (1.93, 2.65] lag 1-30
40-100 ng/m3:1.40 [1.20,1.64] lag 1-30
<	40 ng/m3: Reference
Reference: Hogervorst, et al (2006,
189460)
Period of Study: 2002
Location: Maastricht, the Netherlands
(six schools selected)
Outcome: Decreased lung function
Age Groups: 8-13 years old
Study Design: Multivariate linear
regression (enter method) analysis
l\l: 342 children
Statistical Analyses: AN0VA, chi
square
Covariates: Independent variables: /
height, gender, smoking at home by
parents, pets, use of ventilation hoods
during cooking, presence of unvented
geysers, tapestry in the home,
indoor/outdoor time, education level of
parents.
Dependent variables: lung function indices
Dose-response Investigated? No
Pollutant: PM2.6
Averaging Time: Daily
Mean (SD): 19.0 (3.2)
Monitoring Stations: 6
Copollutant:
PMio
Total Suspended Particles (TSP)
PM Increment: 10 /yg/m3
RR Estimate [Lower CI, Upper CI]
lag:
FEV: 3.62 [0.50,7.63]
lag NR
FVC: 1.80[-2.10, 5.80]
lag NR
FEF: 5.93[-2.34,14.89]
lag NR
Reference: Holguin et al, (2007,
099000)
Period of Study: NR???
Location: Ciudad Juarez, Mexico
Outcome: FeNO, FEVi
Study Design: Panel
Covariates: sex, age, body mass index,
day of week, season, years of maternal
and paternal education, passive smoking
Statistical Analysis: linear and
nonlinear mixed effects models
Age Groups: 6-12 years
Pollutant: PM2.6
Averaging Time: 48h
Mean (SD) Unit: 17.5 (8.9)/yg/m3
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: NR
Relative Risk (Min CI, Max CI)
Lag
Results not given in table form, but
abstract states that no significant
associations with PM2.6 were observed.
Reference: Hong et al. (2007, 091347)
Period of Study: March 23-May3, 2004
Location: School on the Dukjeok Island
near Incheon City, Korea
Outcome: Peak expiratory flow rate
(PEFR)
Age Groups: 3rd to 6th grade (mean
age-9.6 yrs)
Study Design: Panel study
l\l: 43 schoolchildren
Statistical Analyses: Mixed linear
regression
Covariates: age, sex, height, weight,
asthma history, and passive smoking
exposure at home
Dose-response Investigated? No
Lags Considered: 0,1, 2, 3, 4, 5
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD): 20.27 (8.23)
50th(Median): 22.07
Range (Min, Max): 5.94-36.28
Copollutant: PM10
Components of PM10 (Fe, Mn, Pb, Zn, Al)
Effect Estimate:
Regression coefficients of morning and
daily mean PEFR on PM2.6
Lag 1 (PM2.6)
Morning PEFR
Crude: B--0.14, p-0.12
Adjusted: B- -0.54, p,0.01
Mean PEFR
Crude: B- -0.15, p — 0.02
Adjusted: B- -0.54, p,0.01
Regression coefficients of morning and
_daM^jTiean_PEFR_on_PM2;B_and_GSTI\/n_
July 2009
E-230
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
and GSTT1 genotype using linear mixed-
effects regression
Lag 1 (PM2.B)
Morning PEFR: B- -0.57, p< 0.01
Mean PEFR: B- -0.56, p< 0.01
GSTM1
Morning PEFR: B- 20.04, p — 0.25
Mean PEFR: B- 18.75, p-0.28
GSTT1
Morning PEFR: B- 2.31, p-0.89
Mean PEFR: B- 1.75, p-0.91
Reference: Jansen, et al. (2005,
082236)
Period of Study: 1987-2000
Location: Seattle, WA
Outcome: FENO: fractional exhaled
nitrogen oxide, Spirometry, Blood
pressure, Sa02: oxygen saturation, Pulse
rate
Age Groups: 60-86-years-old
Study Design: Short-term cross-
sectional case series
N: 16 subjects diagnosed with COPD,
asthma, or both
Statistical Analyses: Linear mixed
effects model with random intercepts
Covariates: Age, relative humidity,
temperature, medication use
Season: Winter 2002-2003
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD):
Fixed-Site Monitor: 14.0
All Subjects (N-16)
Indoor, home: 7.29
Outdoor, home: 10.47
Asthmatic Subjects (N-7)
Indoor, home: 7.25
Outdoor, home: 8.99
COPD Subjects (N-9)
Indoor, home: 7.33
Outdoor, home: 11.66
Range (Min, Max):
Fixed-Site Monitor: 1.3, 44
IQR
All Subjects
Indoor, home: 4.05
Outdoor, home: 8.87
Asthmatic Subjects
Indoor, home: 5.72
Outdoor, home: 7.55
COPD Subjects
Indoor, home: (3.18
Outdoor, home: 6.71
PM Increment: PM2.6: 10 (xglm3
Slope [95% CI]: dependence of FENO
concentration [ppb] on PM2.6
Asthmatic Subjects
Indoor, home: 3.69 [-0.74: 8.12]
Outdoor, home: 4.23 [1.33: 7.13]*
Copd Subjects
Indoor, home: -0.35 [-7.45: 6.75]
Outdoor, home: 3.83 [-1.84: 9.49]
Results indicate that FENO may be a
more sensitive biomarker of PM exposure
than other traditional health endpoints.
Reference: Johnston, et al. (2006,
091386)
Period of Study: 7 months (April 7
through November 7, 2004)
Location: Darwin, Australia
Outcome: Asthma symptoms
Age Groups: All Ages
Study Design: Time-series
N: 251 i
(130 adults, 121 children
Statistical Analyses: Logistic
regression model
Covariates: Minimum air temperature,
doctor visits for influenza and the
prevalence of asthma symptoms and, the
fungal spore count and both onset of
_asthma_s^m£toms_an£^
Pollutant: PM2.6
Averaging Time: Daily
Mean (SD): 11.1 (5.4)
Range (Min, Max): 2.2, 36.5
PM Component: Vegetation fire smoke
(95%) and motor vehicle emissions (5%)
Monitoring Stations: 1
PM Increment: 5 /yglm3
RR Estimate [Lower CI, Upper CI]
lag:
Symptoms attributable to asthma
Overall: 1.000 (0.98,1.01)
Adults: 1.000(0.976,1.026)
Children: 1.008 (0.980,1.037)
Using preventer: 1.013 (0.990,1.037)
Became symptomatic
July 2009
E-231
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
reliever medication
Season: "Dry season" note Southern
Hemisphere
Dose-response Investigated? No
Statistical Package: STATA8
Lags Considered: 0-5 days
Overall: 1.150(1.07,1.23)
Adults: 1.165(1.058,1.284)
Children: 1.148 (1.042,1.264)
Using preventer: 1.181 (1.076,1.296)
Used Reliever
Overall: 1.000(0.98,1.02)
Adults: 1.007 (0.980,1.035)
Children: 1.002 (0.972,1.034)
Using preventer: 1.020 (1.000,1.042)
Commenced Reliever
Overall: 1.120(1.03,1.210)
Adults: 1.141 (1.021,1.275)
Children: 1.112(0.994,1.243)
Using preventer: 1.129 (1.013,1.257)
Commenced Oral Steroids
Overall: 1.310(1.03,1.66)
Adults: 1.601 (1.192,2.150)
Children: 0.995 (0.625,1.459)
Using preventer: 1.350 (1.040,1.752)
Asthma Attack
Overall: 0.980(0.94,1.04)
Adults: 1.026(0.962,1.095)
Children: 0.832 (0.731,0.946)
Using preventer: 1.002 (0.934,1.075)
Exercise induced asthma
Overall: 0.990(0.95,1.03)
Adults: 0.998 (0.943,1.056)
Children: 0.982 (0.899,1.071)
Using preventer: 1.002 (0.942,1.067)
Saw a health professional for asthma
Overall: 1.030(0.91,1.16)
Adults: 1.079(0.899,1.296)
Children: 1.003 (0.841,1.195)
Using preventer: 0.980 (0.847,1.133)
Missed school or work due to asthma
Overall: 1.025(0.9284,1.131)
Adults: 1.077 (0.923,1.247)
Children: 1.000 (0.873,1.458)
Using preventer: 1.005 (0.897,1.124)
Mean daily number of asthma
symptoms
Overall: 1.003(0.99,1.01)
Adults: 0.998 (0.984,1.012)
Children: 1.004 (0.985,1.023)
Using preventer: 1.013 (0.999,1.028)
JV[eanJ)aN^jiumbej_d_a££Ncat!onsj)f_
July 2009
E-232
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
reliever
Overall: 1.002(0.993,1.010)
Adults: 1.001 (0.986,1.016)
Children: 1.000 (0.980,1.021)
Using preventer: 1.005 (0.994,1.017)
Reference: Koenig et al. (2003,156653) Outcome: Exhaled NO (elMO)
Age Groups: 6-13 years old
Period of Study: Winter 2000-2001,
Spring 2001
Location: Seattle, WA
Study Design: Cohort
N: 19 children
Statistical Analyses: Linear mixed-
effects regression
Covariates: Medication use, ambient NO
reading for specific individual on specific
day of session, mean ambient NO for
subject during session, mean ambient NO
for subject during all sessions
Season: Winter, Spring
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PM2.6
Averaging Time: 10 consecutive days
Mean (SD): Outdoor: 13.3 (1.4)
Indoor: 11.1 (4.9)
Personal: 13.4 (3.2)
Central-site: 10.1 (5.7)
Range (Min, Max): Outdoor: Max: 40.4
Indoor: Max: 36.3
Personal: Max: 49.4
Central-site: NR
Monitoring Stations: Outdoor: NR
Indoor: NR
Personal: NR
Central-site: 3
Copollutant (correlation): Outdoor PM-
central-site NO: 0.50
For NO values < 100 ppb, outdoor PM-
central-site NO: 0.04
PM Increment: 10 /yglm3
Results presented as change in eNO (95%
CI)
Among ICS* nonuser
Personal monitor 4.48 (1.02, 7.93)
Outdoor monitor 4.28 (1.38, 7.17)
Indoor monitor 4.21 (1.02,7.41)
Central site 3.82 (1.22, 6.43)
Among ICS* user
Personal monitor -0.09 (-2.39, 2.21)
Outdoor monitor 0.74 (-2.28, 3.76)
Indoor monitor -1.11 (-5.08, 2.87)
Central site 1.28 (-1.23, 3.79)
* ICS: Inhaled corticosteroid
Reference: Koenig et al. (2003,156653) Outcome: Increased exhaled nitric oxide
(eNO)
Period of Study: Winter 2000-2001,
spring 2001
Location: Seattle, WA
Age Groups: 6-13 years of age
Study Design: Combined recursive and
predictive model
N: 19 children with asthma
Statistical Analyses: Linear mixed
effects model
Covariates: Residence type, air cleaner,
avg outdoor temperature, avg daily
rainfall
Season: Winter, Spring
Dose-response Investigated? No
Statistical Package:
STATA 7.0 for health analyses, SAS 8.0
Pollutant: PM2.6
Averaging Time: Daily
Mean: Home indoor 9.5
Home outdoor 11.1
Recursive model Eag: 7.0
Recursive model Eig: 2.1
Predictive model Eag: 6.0
Predictive model Eig: 4.0
Combined model Eag: 6.4
Combined model Eig: 3.2
25th: Home indoor 5.7
Home outdoor 6.3
Recursive model Eag: 4.2
Recursive model Eig: 0.0
Predictive model Eag: 3.4
Predictive model Eig: 0.9
Combined model Eag: 3.7
Combined model Eig: 0.5
50th(Median): Home indoor 7.6
Home outdoor 9.5
Recursive model Eag: 5.9
Recursive model Eig: 1.2
Predictive model Eaq: 5.0
PM Increment: 10-|jg/m3
RR Estimate [Lower CI, Upper CI]
Eag - ambient-generated personal
exposure
Eig- indoor-generated personal exposure
eNO - exhaled nitric oxide
Recursive model with 8 children, Eag was
marginally associated with increases in
eNO [5.6 ppb [-0.6,11.9],
Eig was not associated with eNO (-0.19
ppb).
For those combined estimates, only Eag
was significantly associated with an
increase in eNO:
Eag: 5.0 ppb [0.3, 9.7]
Eig: 3.3 ppb [1.1, 7.7]
Notes: Effects were seen only in children
who were not using corticosteroid
therapy
July 2009
E-233
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kongtip et al. (2006,
096920)
Period of Study: September 1-October
31,2004
Location: Dindang district, Bangkok
metropolitan, Thailand
Predictive model Eig: 2.2
Combined model Eag: 5.5
Combined model Eig: 1.7
75th: Home indoor 10.8
Home outdoor 14.6
Recursive model Eag: 9.2
Recursive model Eig: 2.3
Predictive model Eag: 7.5
Predictive model Eig: 4.9
Combined model Eag: 7.8
Combined model Eig: 4.2
Range (Min, Max): Home indoor 2.3,
36.3
Home outdoor 2.8, 40.4
Recursive Eag: 1.8,22.6
Recursive Eig: 0.0,17.2
Predictive Eag: 1.3,22.6
Predictive Eig: 0.0,33.0
Combined Eag: 1.3,22.6
Combined Eig: 0.0,33.0
Monitoring Stations: 19 personal
environmental monitors
PM Increment: 1 /yglm3
Effect Estimate [Lower CI, Upper CI]:
Model 1
Headache: 1.011 (0.999-1.022)
Nose congestion: 1.006 (0.997-1.015)
Sore throat: 1.000 (0.991-1.008)
Cold: 1.006(0.995-1.017)
Cough: 0.989 (0.980-0.998)
Phlegm: 0.998 (0.992-1.003)
Chest tightness: 0.995 (0.955-1.036)
Fever: 1.008 (0.993-1.024)
Eye irritation: 1.022 (1.011-1.033)
Dizziness: 1.027 (1.013-1.041)
Weakness: 0.996 (0.983-1.008)
Upper respiratory symptom: 1.001
(0.994-1.008)
Lower respiratory symptom: 0.997
(0.992-1.002)
Model 2
Headache: 1.004 (0.996-1.013)
Nose congestion: 1.003 (0.996-1.010)
Sore throat: 0.995 (0.989-1.001)
Cold: 0.996(0.988-1.004)
Cough: 0.990 (0.983-0.996)
Phlegm: 0.995 (0.991-0.999)
Outcome: respiratory and other
Outcomes reported
Age Groups: Age range 15 to 55 yrs
Study Design: panel study
N: 77 street vendors
Statistical Analyses: Binary logistic
regression
Covariates: Gender, age, type of fuel
used, working duration (months)
Dose-response Investigated? No
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD): 70.94
Percentiles: 50th(Median): 72.05
Range (Min, Max): 23.20-120.00
Monitoring Stations: 1
Copollutant (correlation):
SO?
NO?
Os
VOCs
CO
July 2009
E-234
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Chest tightness: 0.997 (0.970-1.025)
Fever: 1.010(0.998-1.022)
Eye irritation: 1.019 (1.010-1.028)
Dizziness: 1.020 (1.009-1.032)
Weakness: 1.003(0.994-1.012)
Upper respiratory symptom: 0.995
(0.990-1.000)
Lower respiratory symptom: 0.995
(0.991-0.999)
Reference: Lagorio et al. (2006,
089800)
Period of Study: 5/24/1999 to
6/24/1999 and 11/181999 to
12/22/1999
Location: Rome, Italy
Outcome: Lung function (FVC and FEVi)
of subjects with COPD, Asthma
Age Groups: COPD 50 to 80 yrs
Asthma 18 to 64 yrs
Study Design: Time series
N:COPD - 11
Asthma - 11
Statistical Analyses: Non-parametric
Spearman correlation
GEE
Covariates: COPD and IHD: daily mean
temperature, season variable (spring or
winter), relative humidity, day of week
Asthma: season variable, temperature,
humidity, and fB-2-agonist use
Season: Spring and Winter
Dose-response Investigated? Yes
Statistical Package: STATA
Lags Considered: 1-3 days
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD): Overall: 27.2(19.4)
Spring: 18.2 (5.0)
Winter: 36.7 (24.1)
Range (Min, Max): 4.5,100
PM Component: Cd: 0.46±0.40 ng/m3
Cr: 1.9± 1.7 ng/m3
Fe: 283 ± 167 ng/m3
Ni: 4.8±6.5 ng/m3
Pb: 30.6 ± 19.0 ng/m3
Pt: 5.0±8.6 pg/m3
V: 1.8 ± 1.4 ng/m3
Zn: 45.8±33.1 ng/m3
Monitoring Stations: 2 fixed sites: (Villa
Ada and Istituto superior di Sanita)
Copollutant (correlation): NO2 r - 0.43
O3 r - -0.51
COr - 0.67
SO? r - 0.34
PM 10-2.6 r - 0.34
PMior - 0.93
PM Increment: 1 /yg/m3
They observed negative association
between ambient PM2.6 and respiratory
function (FVC and FEVi) in the COPD
panel. The effect on FVC was seen at lag
24 h, 48 h, and 72 h. The effect on FEVi
was evident at lag 72 h. There was no
statistically significant effect of PM2.6 on
FVC and FEVi in the asthmatic and IHD
fB Coefficient (SE)
COPD
FVC(%)
24 h -0.80 (0.36)
48-h -0.89 (0.41)
72-h-1.10 (0.55)
FEVi(%)
24 h -0.47 (0.33)
48-h-0.69 (0.37)
72-h-1.06 (0.50)
Asthma
FVC(%)
24 h-0.14 (0.29)
48-h -0.07 (0.33)
72-h -0.06 (0.39)
FEVi(%)
24 h -0.30 (0.34)
48-h -0.36 (0.39)
72-h -0.40 (0.46)
Reference: Lee et al. (2007, 093042)
Outcome: PEFR (peak expiratory flow
Pollutant: PM2.6
PM Increment: 10 /yg/m3
Period of Study: 2000-2001
rate), lower respiratory symptoms (cold,
cough, wheeze)
Averaging Time: 24 h
Effect Estimate [Lower CI, Upper CI]
Location: South-Western Seoul
Metropolitan area, Seoul, South Korea
Age Groups: 61-89 years of age (77.8
mean age)
Mean (SD): 51.15 (19.94)
Percentiles:
lag:
PEFR (peak expiratory flow rate)

Study Design: longitudinal panel survey
25th: 33.00
¦0.54 (-0.89,-0.19)

N: 61 adults
50th(Median): 53.20
1 day

Statistical Analyses: SAS MIXED,
logistic regression model
Covariates: Temperature (Celsius),
relative humidity, age,
75th: 87.54
Range (Min, Max):
17.94,92.71
relative odds of a lower respiratory
symptom (cold, cough, wheeze)
0.976(0.849,1.121)

Dose-response Investigated? No
Monitoring Stations: 2
1 day

Statistical Package: SAS 8.0


July 2009
E-235
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Lags Considered: 0-4 days
Reference: Lewis et al. (2005, 081079) Outcome: Poorer lung function (increased
diurnal variability and decreased forced
Period of Study: winter 2001 spring
2002
Location: Detroit, Michigan, USA
expiratory volume)
Age Groups: 7-11 years old
Study Design: Longitudinal cohort study
N: 86 children
Statistical Analyses: Descriptive
statistics and bivariate analyses of
exposures, multivariable regression
multivariate analog of linear regression.
Covariates: Sex, home location, annual
family income, presence of one or more
smokers in household, race, season
(entered as dummy variables), and
parameters to account for intervention
group effect.
Season: Winter 2001 (February 10-23),
Spring 2001 (May 5-18), Summer 2001
(July 14-27), Fall 2001 (September 22-
October 5), Winter 2002 (January 18-
31), and Spring 2002 (May 18-31)].
Dose-response Investigated? No
Lags Considered: 1 to 2 days, 3-5 days
Pollutant: PM2.6
Averaging Time: 2 weeks
Mean (SD):
Eastside
15.7 (10.6)
Southwest
17.5(12.2)
Range (Min, Max): 1.0, 56.1
Monitoring Stations: 2
Copollutant (correlation):
PMio 0.93
O3 Daily mean 0.57
O3 8-h peak 0.53
PM Increment: 12.5/yg/m3
RR Estimate [Lower CI, Upper CI]
lag:
Lung function among children reporting
use of maintenance CSs
Diurnal variability FEVi
Lag 1:1.61 [-0.5,3.72]
Lag 1:0.99 [-5.64,7.62] PMz.b + 0s
Lag 2: 2.96 [-1.74,7.66]
Lag 2: 4.62 [-4.31,13.54] PM2.1, + 0s
Lag 3-5: 1.37 [-1.49,4.22]
Lag 3-5: 2.70 [1.0, 4.40] PM2.6 + O3
Lowest daily value FEVi
Lag 1: -2.23 [-6.99,2.53]
Lag 1:3.36 [-3.92,10.63] PM2.1, + 0s
Lag 2: -0.21 [-4.09,3.68]
Lag 2: 0.88 [-8.69,10.46] PM2.1, + 0s
Lag 3-5: -0.76 [-5.00, 3.49]
Lag 3-5: -2.78 [-4.87 to-0.70] PM2.6 +
0s
Lung function among children reporting
presence of URI on day of lung function
assessment
Diurnal variability FEVi
Lag 1:4.08 [-1.78, 9.94]
Lag 1:3.99 [-2.76,10.74] PM2.1, + 0s
Lag 2: 7.62 [-0.49,15.73]
Lag 2: 4.10 [-1.41, 9.60] PM2.1, + 0s
Lag 3-5:1.47 [-7.73,10.67]
Lag 3-5: 3.81 [-1.83, 9.45] PM2.1, + 0s
Lowest daily value FEVi
Lag 1:-1.21 [5.62,3.21]
Lag 1:-0.74 [-4.14, 2.65] PM2.1, + 0s
Lag 2:-0.10(4.36,4.16]
Lag 2: -1.67 [-5.09,1.75] PM2.1, + 0s
Lag 3-5: -2.88 [-5.46 to-0.30]
Lag 3-5: -2.78 [-4.79 to-0.77] PM2.1, +
0s
Reference: Liu et al. (2009,192003)
Outcome: Decreased lung function
Pollutant: PM2.6
Increment: 5.4/yg/m3
Period of Study: 4wks in 2005
Study Design: Panel
Averaging Time: 1, 2 & 3 days
Percent Change (Min CI, Max CI)
Location: Windsor, Ontario, Canada
Statistical Analysis: mixed-effects
Mean (SD) Unit (Id): 6.5//g/m3
Lag

regression models



Range (Min, Max): 2.0-19.0
FEV1

Statistical Package: S-PLUS



Copollutant (correlation):
Same Day: -0.5 (-1.3-0.3)

Age Groups: Asthmatic children, 9-14



yrs.
SO2: 0.56
Lag 1 Day: -0.5 (-1.1-0.5)


NO2: 0.71
2-Day Average: -0.6 (-1.5-0.4)


0s:-0.41
3-Day Average: -1.1 (-3.1-0.9)
July 2009
E-236
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
FEF 25%-75%
Same Day: -1.9 (-3.5--0.3)
Lag 1 Day: -1.2 (-2.8-0.3)
2-Day	Average: -2.0 (-3.8 -0.2)
3-Day	Average: -3.3 (-7.2-0.8)
FeNO
Same Day: 5.3 (-3.6-15)
Lag 1 Day: 1.7 (-6.3-15)
2-Day	Average: 4.3 (-5.4-15.1)
3-Day	Average: -17.3 (-33.5-2.9)
TBARS
Same Day: 16.9 (2.2-33.6)
Lag 1 Day: 14.6(0.8-30.4)
2-Day	Average: 22.0 (4.8-42.1)
3-Day	Average: 69.1 (20.1-138.2)
8-lsoprostane
Same Day: 5.1 (-3.6-14.5)
Lag 1 Day:-3.8 (-12.1-5.3)
2-Day	Average: 0.1 (-9.8-11.1)
3-Day	Average: 5.8 (-15.8-33.0)/
Reference: Maret al. (2004, 057309)
Outcome: Respiratory Symptoms
Pollutant: 2.5
PM Increment: 10/yglm3
Period of Study: 1997-1999
Age Groups: Adults: Ages 20-51 yrs
Mean (SD):
OR Estimate [Lower CI, Upper CI]
Location: Spokane, Washington
Children: Ages 7-12 yrs
1997: 11.0(5.9)
lag:

N: 25 people
1998: 10.3(5.4)
Adult Respiratory symptoms: Wheeze:



1.04(0.86,1.26]

Statistical Analyses: Logistic
1999: 8.1 (3.8)

regression
Unit (i.e./yg/m3):
lag 0



Covariates: Temperature, relative
Monitoring Stations: 1 station
1.00(0.83,1.19]
|gn 1

humidity, day of-the-wk

Statistical Package: STATA 6
Copollutant (correlation):
lag i

PM2.6
0.99(0.84,1.17]

Lags Considered: 0-2 days
Inn 9


PMi
lag z


r - 0.92
Breath: 0.97(0.87,1.08]


lag 0


PM10


r - 0.61
0.98(0.87,1.10]




PM 10-2.6
lag 1


r - 0.28
0.95(0.80,1.13]


lag 2
Cough: 0.86(0.62,1.21]
lag 0
0.87(0.63,1.20]
lag 1
0.89[0.66,1.20]
lag 2
Sputum: 0.94(0.63,1.41]
lag 0
0.90(0.62,1.31]
July 2009
E-237
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
lag 1
0.92[0.66,1.27]
lag 2
Runny Nose: 0.98[0.83,1.15]
lag 0
0.95[0.82,1.10]
lag 1
0.93(0.80,1.08]
lag 2
Eye Irritation: 0.91 [0.70,1.20]
lag 0
0.89(0.70,1.13]
lag 1
0.86(0.68,1.08]
lag 2
Lower Symptoms: 0.91(0.73,1.13]
lag 0
0.89(0.72,1.10]
lag 1
0.89(0.72,1.10]
lag 2
Any Symptoms: 0.92(0.80,1.07]
lag 0
0.89(0.76,1.04]
lag 1
0.89(0.75,1.05]
lag 2
Children Respiratory symptoms:
Wheeze: 0.55(0.26,1.19]
lag 0
0.53(0.18,1.58]
lag 1
0.55(0.19,1.64]
lag 2
Breath: 1.13(0.86,1.48]
lag 0
1.12(0.86,1.44]
lag 1
1.10(0.82,1.48]
lag 2
Cough: 1.17(0.98,1.40]
lag 0
1.21(1.00,1.47]
lag 1
1.18(0.99,1.42]
July 2009
E-238
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
lag 2
Sputum: 1.06[0.92,1.22]
lag 0
1.10[0.91,1.34]
lag 1
1.09(0.92,1.30]
lag 2
Runny Nose: 1.09(0.85,1.39]
lag 0
1.12(0.89,1.41]
lag 1
1.16(0.94,1.42]
lag 2
Eye Irritation: 0.93(0.53,1.64]
lag 0
0.75(0.45,1.27]
lag 1
0.77(0.65, 0.91]
lag 2
Lower Symptoms: 1.18(1.00,1.38]
lag 0
1.21(1.00,1.46]
lag 1
1.17(0.96,1.43]
lag 2
Any Symptoms: 1.17(1.03,1.34]
lag 0
1.22(1.04,1.43]
lag 1
1.23(1.07,1.42]
lag 2
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI]
Lag
Personal: Systolic: 0.37 ( 0.93,1.67)
0
Diastolic: -0.20 (-0.85, 0.46)
0
Indoor: Systolic: 0.92 (-2.04, 3.87)
0
Diastolic: 0.38 (-1.43, 2.20)
0
Outdoor: Systolic: -0.81 (-2.34, 0.73)
0
Diastolic: -0.46 (-1.49, 0.57)
Reference: Mar et al. (2005, 087566)
Period of Study: 1999-2001
Location: Seattle, Washington
Outcome: Pulmonary function (arterial
oxygen saturation) and cardiac function
(heart rate and blood pressure)
Study Design: Time series
Statistical Analyses: Linear logistic
regression
Age Groups: > 57
Pollutant: PM2.6
Averaging Time: 24-h avg
July 2009
E-239
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
% Increase between heart rate and PM2.6
exposure for people > 57
PM2.5: Personal: 0.44 (0.04, 0.84)
0
Indoor: 0.22 (-0.71,1.16)
0
Outdoor: -0.75 (-1.42 to-0.07)
0
Reference: Mar et al. (2005, 088759)
Period of Study: 1999-2002
Location: Seattle, Washington
Outcome: Respiratory Symptoms
Age Groups: 6-13 years
Study Design: Time-Series
N: 19 children
Statistical Analyses: Polynomial
distributed lag model, Poisson regression
Covariates: Age, ambient NO levels,
temperature, relative humidity,
modification of use of inhaled
corticosteroids
Season: Winter, Spring
Dose-response Investigated? No
Statistical Package: STATA
Lags Considered: 0-8 h
Pollutant: PM2.6
Averaging Time: 24-h
Mean (SD):
Results presented in Figure 1.
Monitoring Stations: 3 Stations
PM Increment: 10 /yglm3
Change in FE(NO) (exhaled NO
concentration) with air pollution
[Lower CI, Upper CI]
lag:
Medication use:
Nomeds: 6.99(3.43,10.55]
lag 1-h
Meds:-0.18[-3.33, 2.97]
lag 1-h
No meds: 6.30(2.64, 9.97]
lag 4-h
Meds: -0.77(-4.58, 3.04]
lag 4-h
No meds: 0.46(-1.18, 2.11]
lag 8-h
Meds: 0.40(-1.94, 2.74]
lag 8-h
Reference: McCreanoret al. (2007,
092841)
Period of Study: 2003-2005
Location: London, England
Outcome: Decreased Lung Function
Age Groups: Adults
Study Design: Crossover study
N: 60 adults
Pollutant: PM2.6
Averaging Time: 1 h
Mean (SD): NR
50th(Median): Oxford St: 28.3
Statistical Analyses: Linear regression Hyde Park: 11.9
Covariates: Temperature, relative
humidity, age, sex, bod-mass index, and
race or ethnic group
Range (Min, Max): Oxford St: (13.9,
76.1)
Hyde Park: (3, 55.9)
% changes in FEV and FVC are presented
in figures 1-3. Results are not presented
quantitatively in text or tables. The
authors did not find any significant
differences in respiratory symptoms
between the two locations. Also, there
were no significant differences in sputum
eosinophil counts or eosinophil cationic
protein levels.
Reference: Moshammer and Neuberger
(2003, 041956)
Period of Study: 2000-2001
Location: Linz, Austria
Outcome: Lung Function: FVC, FEVi,
MEF26, MEFbo, MEFjb, PEF, LQ Signal,
PAS Signal
Age Groups: Ages 7 to 10
Study Design: Case-crossover
N: 161 children
1898-2120 "half-h means"
Statistical Analyses: Correlations
Regression Analysis
Covariates: Morning, evening, night
Season: Spring, Summer, Winter, Fall
Pollutant: PM2.6
Averaging Time: 8 h means 81 Daily
Means
Mean (SD): 14.61 (10.83)
Range (Min, Max):
(NR, 119.92)
Monitoring Stations: 1
Copollutant (correlation):
LQ - 0.751
PAS - 0.354
Notes: "Acute effects of 'active particle
surface' as measured by diffusion
charging were found on pulmonary
function (FVC, FEVi, MEF50) of
elementary school children and on
asthma-like symptoms of children who
had been classified as sensitive."
July 2009
E-240
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Dose-response Investigated? No
Reference: Moshammeret al. (2006,
090771)
Period of Study: 2000-2001
Location: Linz, Austria
Outcome: Respiratory symptoms and
decreased lung function
Age Groups: Children ages 7-10
Study Design: Time-series
N: 163 children
Statistical Analyses: Generalized
estimating equations model
Covariates: Sex, age, height, weight
Dose-response Investigated? NR
Statistical Package: NR
Lags Considered: 1
Pollutant: PM2.B
Averaging Time: 8 h
Mean (SD):
Maximum 24 h: 76.39
Annual avg: 19.06
Percentiles: 8-h mean 25th: 8.64
8-h mean 50th(Median): 15.70
8-h mean 75th: 25.82
Monitoring Stations: 1 station
Copollutant (correlation): PMi
r - 0.95
PMio
r - 0.93
NO?
r - 0.54
PM Increment: 10//g|m3
% change in Lung Function per 10
z^g/m3
FEV: 0.23
FVC: 0.08
FEVob: 0.33
MEFjb*: -0.49
MEFbo%:-0.58
MEF2b%: -0.83
PEF: 0.41
% change in Lung Function per IQR
FEV:-0.59
FVC: -0.2
FEVob: 0.85
MEF?b%: -1.25
MEFbo%: -1.48
MEF2b%: -2.14
PEF:-1.06
Multiple pollutant model
FEV: 0.10
FVC: 0.21
FEVob: 0.06
MEF7b%: -0.15
MEF50%: 0.04
MEF25%: -0.21
PEF:-0.18
% change in Lung Function per IQR
FEV: 0.27
FVC: 0.54
FEVob: 0.15
MEFjb*: -0.39
MEFbo%: 0.11
MEF2b%: 0.54
PEF: 0.015:-0.47
Reference: Murata et al. (2007,
156787)
Period of Study: Nov 2nd-12th 2004
Location: Tokyo, Japan
Outcome: Exhaled nitric oxide levels,
(eN0), a marker of airway inflammation
Age Groups: 5-10 years
Study Design: Cohort/Panel study
N: 19 schoolchildren*
Statistical Analyses: Linear regression
Covariates: None
Season: November (fall)
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: Lag h 1-24, 8-h
Pollutant: PM2.B
Averaging Time:
Hourly, 24-h
Mean (SD):
39.0 (16.9)/yglm3 (daily mean)
Range (Min, Max):
10,120 (range of hourly values)
Monitoring Stations: 1, on the street
where the children lived
PM Increment: IQR 110/yglm3
Mean [Lower CI, Upper CI]
lag:
0.145 [0.62, 0.228] ppb eNO
8 h moving avg
Notes:
Associations for lag h 1-24 presented in
figures. Authors state "Individual hourly
lag models showed a consistent
association between the eNO value and
PM2.B for exposure in the previous 24 h"
"The trend on the graphs strongly
_suc|[|esUhaH]uctUc^^
July 2009
E-241
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
moving avg, 7-h moving avg, 6-h moving
avg, 24-h moving avg
affected by changes in air pollutants over
at least the previous 8-h period"
PM2.E, black carbon, and NOx were all
highly correlated (shown in figures), so
effects are difficult to separate
Pollutant concentrations peaked in the
morning and evening h during traffic
peaks
Reference: Neuberger et al. (2004,
093249)
Period of Study: 6/1999-6/2000
Location: Austria (Vienna and a rural
area near Linz)
Outcome: Questionnaire derived asthma
score, and a 1-5 point respiratory health
rating by parent
Age Groups: 7-10 years
Study Design: Cross-sectional survey
N: about 2000 children
Statistical Analyses: mixed models
linear regression-used factor analysis to
develop the "asthma score"
Covariates: Pre-existing respiratory
conditions, temperature, rainy days, #
smokers in household, heavy traffic on
residential street, gas stove or heating,
molds, sex, age of child, allergies of child,
asthma in other family members
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 4 week avg (preceding
interview)
Pollutant: PM2.6
Averaging Time: 24 h
Copollutant (correlation):
PM10 (r-0.94)
in Vienna
PM Increment: 10 /yglm3
Change in mean associated unit
increase in PM
(p-value)
lag
Respiratory Health score
Vienna: 0.016 (p > 0.05)
lag 4 week avg
Rural area: 0.022 (p < 0.05)
lag 4 week avg
Asthma score
Vienna: 0.006 (p > 0.05)
lag 4 week avg
Rural area: 0.004 (p > 0.05)
lag 4 week avg
Reference: Neuberger et al. (2004,
093249)
Period of Study: Sept 1999-March
2000
Location: Vienna, Austria
Outcome: Ratio measure: Time to peak
tidal expiratory flow divided by total
expiration time (i.e., tidal lung function, a
surrogate for bronchial obstruction)
Age Groups: 3.0-5.9 years (preschool
children)
Study Design: Longitudinal prospective
cohort
N: 56 children
Statistical Analyses: mixed models
linear regression, with autoregressive
correlation structure
Covariates: Age, sex, respiratory rate,
phase angle, temperature, kindergarten,
parental education, observer (also in
sensitivity analyses: height, weight,
cold/sneeze on same day, heating with
fossil fuels, hair cotinine, number of tidal
slopes used to measure tidal lung
function)
Dose-response Investigated? No
Statistical Package: SAS 8.0
Lags Considered: Lag 0
Pollutant: PM2.6
Averaging Time: 24 h
PM Component: Total carbon
Elemental carbon
Organic Carbon
Copollutant (correlation): PM10
(r-0.94) in Vienna
PM Increment: Interquartile range (NR)
Change in mean associated with an
IQR increase in PM (p-value)
lag
PM2.6 mass: -0.987 (0.091)
lag 0
Total carbon: -0.815 (0.041)
lag 0
Elemental carbon: -0.657 (0.126)
lag 0
Organic carbon: -0.942 (0.025)
lag 0
Reference: Neuberger et al. (2004,
Outcome: Forced oscillatory resistance
Pollutant: PM2.6
PM Increment: 1 /yglm3
093249)
(at zero Hz), FVC, FEVi, MEF26, MEFbo,


MEFjb, PEF
Averaging Time: 24 h
Notes: Authors report increased
Period of Study: Oct. 2000-May 2001
Monitoring Stations: 1
oscillatory resistance significantly
Age Groups: 7-10 years
associated with PM2.6 (lag 0)
Location: Linz, Austria


Study Design: Longitudinal prospective



cohort



N: 164 children



Statistical Analyses: Mixed models



linear regression with autoregressive



correlation structure


July 2009
E-242
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Covariates: Sex, time and individual
Season: October-May
Dose-response Investigated? No
Statistical Package: NR
Lags Considered:
Lag 0-7
Location: Boston, the Bronx, Chicago, (ICAS)-Panelfcohort study
Dallas, New York, Seattle, Tucson
Reference: O'Connor et al. (2008,
156818)
Period of Study: August 1998-July Age Groups: 5-12 years
2001
Outcome: Pulmonary function and	Pollutant: PM2.6
respiratory symptoms
Study Design: Inner-City Asthma Study
Averaging Time: 24-h
Mean (SD): 14
Range (Min, Max):
5-35 (estimated from figure)
Copollutant (correlation):
NO? (r-0.59)
SO? (r — 0.37)
CO (r — 0.44)
O3 (r—-0.02)
PM Increment: 13.2/yg/m3 90th-10th
percentile
Change in pulmonary function
lag
FEVi: -1.47 (-2.00 to-0.94)
N: 861 children
lag 0-4
Statistical Analyses: Mixed effects
models
PEFR:-1.10 (-1.65 to -0.56)
Lags Considered: Lag 0-6, 0-4
lag 0-4
PM2.5+O3+NO2
FEVi: -0.73 (-1.33 to-0.12)
lag 0-4
PEFR: -0.25 (-0.88, 0.38)
lag 0-4
Risk of Respiratory Symptoms
lag
Wheeze: 0.98 (0.88,1.09)
lag 0-4
Nighttime asthma: 1.11 (0.94,1.30)
lag 0-4
Slow play: 1.01 (0.89,1.15)
lag 0-4
Missed school: 1.33 (1.06,1.66)
lag 0-4
PM2.5+O3+NO2
Wheeze: 0.92 (0.81,1.05)
lag 0-4
Nighttime asthma: 1.03 (0.86,1.23)
lag 0-4
Slow play: 0.92 (0.79,1.06)
lag 0-4
Missed school: 1.13 (0.87,1.45)
lag 0-4
July 2009
E-243
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Peacock et al. (2003,
Outcome: Reduced peak expiratory flow
Pollutant: Sulfate (SO42 )
Sulfate (SO42) Increment: 1.3/yg/m3
042026)
rate (PEFR)


Averaging Time: Daily avg
Odds ratio [Lower CI, Upper CI]
Period of Study: November 1,1996 to
Age Groups: 7-13 years of age
Mean (SD): Urban 2
lag:
14 February 1997
Study Design: Time Series
24 h avg: 1.3(1.1)
1.090 [0.898,1.322]
Location: northern Kent, UK
N: 179
Percentiles:
5 days

Statistical Analyses: generalized
10th: Urban 2 0.5


estimating equations


Covariates: Day of the week, 24-h mean
90th: Urban 2 2.4


outside temperature.
Range (Min, Max):


Season: Winter
Urban 2 0.3, 6.7


Dose-response Investigated? No
Unit (i.e./yg/m3):/yg/m3


Statistical Package: STATA
Monitoring Stations: 3


Lags Considered: Same day, lag 1, lag



2, five day moving avg


Reference: Peled, et al. (2005,156015)
Outcome: Reduced peak expiratory flow
(PEF)
Pollutant: PM2.6
PM Increment: 1 //g|m3
Period of Study: 5-6 weeks between
Averaging Time: Daily
(3 coefficient (SE) [95% CI]
March-June 1999 and September-
Age Groups: 7-10 years
Mean:
Ashkelon:
December 1999.

Study Design: Nested cohort study


Location: Ashdod, Ashkelon and Sderot,
Ashkelon: 24.0
PM2.6 MAX:-0.144 (0.12) [-0.38-0.09]
Israel
N: 285
Sderot: 29.2
Ashdod:
Statistical Analyses: Time series
analysis
Generalized linear model, generalized
estimating equations, one way ANOVA,
generalized linear model
Covariates: Seasonal changes,
meteorological conditions and personal
physiological, clinical and socioeconomic
measurements
Season: Spring, Autumn
Dose-response Investigated? No
Statistical Package: STATA
Ashdod: 23.9
PM Component: Local industrial
emissions, desert dust, vehicle emissions
and emissions from two electric power
plants
Monitoring Stations: 6
Copollutant: PMio
PM2.6 MAX: -2.74 (0.61) [-3.95-1.53]
PM2.6 MAX xTMAX: 0.11 (0.02) [0.06-
0.16]
In Ashdod, PM2.6 and an interaction
between PM2.6 and temperature were
significantly associated.
July 2009
E-244
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Penttinen et al. (2006,
087988)
Period of Study: 11/1996-4/1997
Location: Helsinki, Finland
Outcome: Decreased lung function and
respiratory symptoms
Age Groups: Adults, mean age 53 years
Study Design: Time Series
N: 78 people
Statistical Analyses: Generalized least
squares autoregressive model
Covariates: Temperature, relative
humidity, day of study, day of study
squared, binary dummy variable for
weekends
Season: Winter, Spring
Dose-response Investigated? NR
Statistical Package: SAS version 6
Lags Considered: 0-3
Pollutant: PM2.6
PM Component: Soil, heavy fuel oil, sea
salt
Averaging Time: 24 h
Percentiles: 25th: Long range transport:
2.44
Local combustion: 1.75
Soil: 0.14
Heavy fuel oil: -0.13
Sea Salt: 0.22
Unidentifiable: -1.41
All sources: 6.47
50th(Median): Long range transport:
4.15
Local combustion: 2.41
Soil: 0.64
Heavy fuel oil: 0.10
Sea Salt: 0.27
Unidentifiable: 0.02
All sources: 8.37
75th: Long range transport: 7.33
Local combustion: 3.05
Soil: 1.46
Heavy fuel oil: 0.52
Sea Salt: 0.42
Unidentifiable: 0.74
All sources: 11.15
Range (Min, Max): Long range transport:
(¦0.89, 28.31)
Local combustion: (0.83, 6.51)
Soil: (-1.13,6.43)
Heavy fuel oil: (-0.67, 4.74)
Sea Salt: (0.09, 0.98)
Unidentifiable: (-4.40,4.77)
All sources: (4.11, 33.53)
Monitoring Stations: 1 site
July 2009
E-245
PM Increment: 1.3 /yglm3
PM2.5, long range: PEF Morning: 0.37[-
0.59,1.34]
lag 0
¦1.04[-1.88 to -0.19]
lag 1
-0.82[-1.81,0.16]
lag 2
0.221-0.64,1.08]
lag 3
-0.241-1.12,0.64]
5 day mean. PEF Afternoon: 0.201-0.67,
1.06]
lag 0
-0.201-1.24,0.83]
lag 1
-0.301-1.14,0.53]
lag 2
0.451-0.57,1.47]
lag 3
0.031-0.79, 0.85]
5 day mean. PEF Evening: -0.331-1.30,
0.64]
lag 0
-0.291-1.13,0.55]
lag 1
-0.411-1.46,0.64]
lag 2
0.391-0.47,1.24]
lag 3
0.071-0.81,0.95]
5 day mean
PM2.5, local combustion: PEF Morning:
-0.731-1.69, 0.23]
lag 0
-0.461-1.24,0.32]
lag 1
-0.431-1.49,0.63]
lag 2
0.341-0.47,1.15]
lag 3
-0.251-1.03,0.53]
5 day mean. PEF Afternoon: -0.21[-
1.07, 0.65]
lag 0
¦0.81 [-1.77, 0.16]
lag 1
-0.831-1.74,0.09]
lag 2
"¦^H^-^'not CITE OR QUOTE
lag 3
-0.871-1.63 to -0.12]
5 day mean. PEF Evening: -0.511-1.48,

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Pino et al. (2004, 050220)
Period of Study: 411995-1011996
Location: Santiago, Chile
Outcome: Respiratory Symptoms,
Wheezing bronchitis
Study Design: Time-series
Statistical Analyses: Bayesian
hierarchical analysis, cubic spline
Age Groups: 4 months-2 years old
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD) unit: 52.0 (31.6)
Range (5th, 95th): 17.0,114.0
Copollutants (correlation):
SO2: r- 0.73
NO2: r- 0.85
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
% increase in wheezing bronchitis and
PM2.6 exposure for infants 4 months to 2
years old
4.75(1.25,8.25)
1
3.85 (0.45, 7.75)
2
2.25 (-1.00, 6.00)
3
1.75 (-2.20, 5.75)
4
4.00 (0.25, 8.00)
5
5.00(1.00,8.50)
6
7.00(3.50,11.00)
7
8.10(4.00,11.25)
9.00(6.00,12.00)
9
8.75(5.75,12.00)
10
1.50(-3.50,4.75)
11
0.25 (-3.75,4.25)
12
0.00 (-4.00,4.00)
13
1.00(-3.50,4.50)
14
1.50(-3.50,4.50)
15
OR for wheezing bronchitis and PM2.6
exposure in infants 4 months to 2 years
old according to family history of asthma
Yes to family history of asthma
1.09(1.00,1.19)
1
1.10(1.02,1.20)
2
1.11 (1.02,1.22)
3
No to family history of asthma
July 2009
E-246
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1.04 (1.00,1.08)
1
1.02(0.98,1.06)
2
1.01 (0.96,1.05)
3
Reference: Rabinovitch et al„ (2006,
088031)
Period of Study: 2001-2003 (two
winters 2001-2002 and 2002-2003)
Location: Denver, CO
Outcome: Bronchodilator doser
activations (daily) and urinary leukotriene
E4 (daily)
Age Groups: Children 6-13 years old
Study Design: School-based cohort
study
N: 73 children
Statistical Analyses: Doser activation:
Poisson regression with GEE with AR1
working covariance
Urinary leukotriene E4: linear mixed model
with spatial exponential covariance
Covariates: Temperature, pressure,
humidity, time trend, Friday indicator,
upper respiratory infection (URI), height
(leukotriene E4 only).
Season: Winter
Dose-response Investigated? NR
Statistical Package: SAS
Lags Considered: 0-2 days
Pollutant: PM2.6
Averaging Time: Morning (midnight to
11:00 AM) mean
Morning (midnight to 11: 00 AM)
maximum
24-h mean
Mean (SD): 24-h mean, TE0M
Year 1, N: 55 days
6.5 (3.2)
Year 2, N: 128 days
8.2 (3.7)
24-h mean, FRM
Year 1,N: 55 days: 11.8 (7.2)
Year 2, l\l: 122 days: 11.2 (5.5)
Morning mean, TE0M
Year 1, N: 71 days: 7.4 (4.7)
Year 2, N: 127 days: 9.1 (5.0)
Morning maximum, TE0M
Year 1, N: 71 days: 15.5 (9.5)
Year 2, N: 127 days: 18.4 (9.6)
Percentiles: 24-h mean, TE0M
Year 1
25th: 4.4
50th(Median): 6.2
75th: 7.9
Year 2
25th: 55
50th(Median): 7.3
75th: 9.9
24-h mean, FRM
Year 1
25th: 7.8
50th(Median): 10.1
75th: 14.1
Year 2
25th: 7.5
50th(Median): 9.3
75th: 13.3
Morning mean, TE0M
Year 1
PM Increment: IQR (over current and
previous day)
Doser Activation
Morning avg PM2.5 TEOM
Year 1: Pet Increase: 3.0 [-0.5: 6.6] p -
0.10
Year 2: Pet Increase: 2.7 [1.1: 4.4] p -
0.006
Aggregated years: 2.2 [0.7: 3.6] p -
0.005
Morning max PM2.5 TEOM
Year 1 Pet Increase: 4.0 [0.5: 7.6] p -
0.02
Year 2 Pet Increase: 2.3 [0.7: 4.0] p -
0.009
Aggregated years 2.6 [0.9: 4.2] p-
0.002
24-h PM2.6
TEOM
Lag 0: 0.4 [-0.7:1.6] p-value - 0.45
Lag 1: 0.9 [-0.7: 2.4] p-value - 0.27
Lag 2: -0.4 [-1.7: 0.9] p-value - 0.59
Lag 0-2 Avg: 0.6 [-1.0: 2.2] p-value -
0.43
FRM
Lag 0:0.2 [-1.2:1.6] p-value - 0.81
Lag 1: 0.9 [-0.9: 2.6] p-value - 0.31
Lag 2: -0.2 [-2.2:1.8] p-value -0.88
Lag 0-2 Avg: 1.2 [-0.6: 2.9] p-value -
0.20
Morning avg PM2.5
TEOM
URI not adjusted
Mild/Moderate Asthmatics: 1.5 [-0.5: 3.4]
p - 0.14
Severe Asthmatics: 3.7 [1.6: 5.8] p- -
0.0006
Difference between severity groups, p -
0.12
Aggregated severity group: 2.2 [0.7: 3.6]
p- 0.005
URI adjusted
Mild/Moderate Asthmatics: 1.0 [-1.9:
3.9]p — 0.50
Severe Asthmatics: 6.0 [1.8: 10.1] p -
July 2009
E-247
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
25th: 4.0
50th(Median): 5.9
75th: 9.6
Year 2 25th: 5.2
50th (Median): 8.5
75th: 11.6
Morning maximum, TEOM
Year 1 25th: 8
50th (Median): 13
75th: 20
Year 2 25th: 11
50th (Median): 16
75th: 23
Range (Min, Max): 24-h mean, TEOM
Year 1 (2.1,23.7)
Year 2 (1.7, 20.5)
24-h mean, FRM
Year 1 (4.3, 53.5)
Year 2 (3.4, 26.3)
Morning mean, TEOM
Year 1 (1.4,22.7)
Year 2 (1.6, 30.2)
Morning maximum, TEOM
Year 1 (4,42)
Year 2 (4,46)
Monitoring Stations: 2 (1 TEOM and 1
Federal Reference Monitor [FRM])
0.006
Difference between severity groups, p -
0.08
Aggregated severity groups: 2.7 [-0.1:
5.4] p- 0.06
Morning maximum PM2.5
TEOM
URI not adjusted
Mild/Moderate Asthmatics: 1.9 [-0.2: 4.1]
p- 0.07
Severe Asthmatics: 3.9 [1.1: 6.8] p -
0.006
Difference between severity groups, p -
0.29
Aggregated severity groups: 2.6 [0.9:
4.2] p- 0.002
URI adjusted
Mild/Moderate Asthmatics: 1.6 [-2.2: 5.4]
p - 0.41
Severe Asthmatics: 8.1 [2.9: 13.4] p -
0.003
Difference between severity groups, p -
0.03
Aggregated severity groups: 3.8 [0.2:
7.4] p - 0.04
Leukotriene E4
24-h PM2.5
TEOM
Lag 0:3.3 [-0.7: 7.2] p - 0.09
Lag 1: -1.6[-5.7: 2.5] p - 0.40
Lag 2: 1.1 [-2.8: 5.1] p- 0.64
Lag 0-2 Avg: 2.3 [-4.0: 8.6] p - 0.45
FRM
Lag 0: 2.7 [1.1: 6.5] p - 0.12
Lag 1: -0.8 [-4.9: 3.3] p - 0.65
Lag 2: -0.8 [-4.9: 3.3] p - 0.71
Lag 0-2 Avg: 2.6 [-2.3: 7.5] p - 0.27
Leukotriene E4
Morning avg PM2.5 TEOM
Height 25%ile: 8.9 [3.0: 14.7] p- 0.004
Height 50%ile: 5.9 [1.4: 10.4] p - 0.01
Height 75%ile: 1.9 [-3.4: 7.3] p - 0.47
Model wfo Height x Pollutant: 5.6 [1.0:
10.2] p - 0.02
Morning maximum PM2.5
TEOM
Height 25%ile: 8.3 [3.4: 13.2] p - 0.001
Height 50%ile: 6.1 [2.1: 10.2] p- 0.004
Height 75%ile: 3.2 [-2.0: 8.4] p- 0.23
J\/lodelj«/o_Hei2ht_^_Pollutant^_6;2_[1_;9^
July 2009
E-248
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
10.5] p - 0.006
Reference: Rabinovitch et al. (2004,
096753)
Periods of Study: 11/15/1999-
311512000
11/13/2000-3/23/2001
11/15/2001-3/22/2002
Location: Denver, Colorado
Outcome: Respiratory symptoms,
Asthma symptoms (cough and wheeze),
Upper respiratory symptoms
Study Design: Time-series
Statistical Analyses: Logistic linear
regression, PR0C Mixed, PR0C Genmod
Age Groups: 6-12
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD): 10.8 (7.1)
Range (Min, Max): (1.8, 53.5)
Copollutant (correlation):
CO
NO?
SO?
0s
PM Increment: 1 /yg/m3
PISE)
AM:-0.003 (0.009)
PM: 0.004 (0.011)
Odds Ratio (Lower CI, Upper CI)
Lag
0.971 (0.843,1.118)
0-3 avg.
Reference: Ranzi et al. (2004, 089500)
Period of Study: February-May 1999
Location: Emilia-Romagmna, Italy (urban-
industrial and rural area)
Outcome: respiratory symptoms, PEF
measurements, drug consumption and
daily activity
Age Groups: Children, mean age - (7.2-
7.9 yrs)
Study Design: Panel study
N: 120 children
Statistical Analyses: Ecological
analysis and Panel analysis
Covariates: Temperature, humidity,
gender, medicinal use, symptomatic
status of previous day
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0,1, 2, 3, 0-3 mov
avg
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD):
Urban - 53.07
Rural- 29.11
Monitoring Stations: 3
Copollutant (correlation):
TSP: r—0.613
daily air pollution concentrations: r-
0.658
PM Increment: 10 /yg/m3
Effect Estimate:
Urban-industrial panel
Cough and Phlegm: RR -1.0044 (1.0011-
1.0077)
Reference: Rodriguez et al. (2007,
092842)
Period of Study: 1996-2003
Location: Perth, Australia
Outcome: Body temperature, cough,
runny/ blocked nose, wheeze/ rattle chest
(daily)
Age Groups: Children 0-5 years old
Study Design: hospital-based cohort
study
N: 198-263 children
Statistical Analyses: Logistic
regression with GEE and AR (order not
specified) working covariance
Covariates: temperature, humidity
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-5 days
Pollutant: PM2.6
Averaging Time: 1-h and 24-h
Mean (SD): 1-h averaging, 20.767
24-h averaging, 8.534
Range (Min, Max): 1-h averaging
(0.012: 93.433)
24-h averaging
(0.004: 39.404)
Monitoring Stations: 10 total, usually
3-5 sites for each pollutant
Copollutant (correlation):
0s
N0 +
C0
PM Increment: NR
[Lower CI, Upper CI]
lag: NR
LAG: 0 day
PM2.5, 1-h
Body temperature: 1.004 [0.998: 1.011]
Cough: 1.006(1.000: 1.012]
Wheeze/rattle chest: 1.004 [0.998:
1.010]
Runny/blocked nose: 0.997 [0.983:
1.010]
PM2.5, 24-h
Body temperature: 1.005 [0.986: 1.024]
Cough: 1.019(0.999: 1.040]
Wheeze/rattle chest: 0.990 [0.969:
1.012]
Runny/blocked nose: 0.968 [0.926:
1.013]
LAG: 5 days
PM2.5, 1-h
Body temperature: 1.005 [0.999: 1.040]
Cough: 1.003(0.995: 1.010]
Wheeze/rattle chest: 1.005 [0.998:
July 2009
E-249
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
10.12]
Runny/blocked nose: 1.015 [1.000:
1.030]
PM2.5, 24-h
Body temperature: 1.020 [0.998: 1.011]
Cough: 1.006(0.984: 1.011]
Wheeze/rattle chest: 1.018 [0.997:
1.040]
Runny/blocked nose: 1.039 [0.990:
1.089]
LAG: 0-5 days
PM2.5, 1-h
Body temperature: 1.000 [0.998: 1.002]
Cough: 1.001 [0.999: 1.003]
Wheeze/rattle chest: 1.002 [1.000:
1.004]
Runny/blocked nose: 1.001 [0.997:
1.006]
PM2.5, 24-h
Body temperature: 1.000 [0.994: 1.005]
Cough: 1.004(0.997: 1.011]
Wheeze/rattle chest: 1.001 [0.995:
1.007]
Runny/blocked nose: 0.998 [0.985:
1.011]
Effect Estimate:
Multiple regression analysis between
inhaled factors in Antarctica
Total leukocyte
Cigarette smoking- 0.211, p< 0.001
Support staff- 0.139, p-0.024
Total PM- 0.168, p-0.004
Segmented PMN
Cigarette smoking- 0.015, p — 0.805
Support staff- 0.097, p —0.119
Total PM- 0.272, p< 0.001
Band-formed PMN
Cigarette smoking- 0.035, p — 0.543
Support staff- 0.010, p-0.864
Total PM - 0.470, p< 0.001
Monocyte
Cigarette smoking- 0.081, p —0.187
Support staff- 0.019, p-0.759
Total PM- 0.328, p< 0.001
G-CSF
Cigarette smoking- 0.131, p< 0.038
Support staff- 0.176, p-0.005
Total PM- 0.078, p-0.186
IL-6
Reference: Sakai et al. (2004, 087435)
Period of Study: November 14,1999-
March 28, 2001
Location: Diesel-powered ship from
Tokyo, Japan to Showa Station on Ongul
Island, Antarctica for 366 days (from
February 1, 2000) and then heading back
to Japan on February 1, 2001
Outcome: circulating leukocyte counts
and serum inflammatory cytokine levels
Age Groups: 24-57 yrs, mean-36.1 ±
4.7 yrs
Study Design: cohort
N: 39 members of 41 st Japanese
Antarctic Research Expedition (JARE-41)
Statistical Analyses: ANOVA
Covariates: Smoking history,
occupational pollutant exposure
Dose-response Investigated? No
Statistical Package: SPSS 11.5J
Pollutant: PMB 02 0
Averaging Time: 24 h
Unit (i.e. /yg/m3): particles/L
PM Component: organic and inorganic
substances, including microorganisms
Copollutant (correlation):
PM2.00.3
PMlO-5.0
July 2009
E-250
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Outcome: circulating leukocyte counts
and serum inflammatory cytokine levels
Age Groups: 24-57 yrs, mean-36.1 ±
4.7 yrs
Study Design: cohort
N: 39 members of 41 st Japanese
Antarctic Research Expedition (JARE-41)
Statistical Analyses: ANOVA
Covariates: Smoking history,
occupational pollutant exposure
Dose-response Investigated? No
Statistical Package: SPSS 11.5J
Pollutant: PMioe o
Averaging Time: 24-h
Unit (i.e./yg/m3): particles/L
Monitoring Stations: NR
Copollutant (correlation!:
PM2.00.3
PMlO-5.0
Reference: Silkoff et al. (2005, 087471) Outcome: Lung function: FEVi, PEF
Reference: Sakai et al. (2004, 087435)
Period of Study: November 14,1999-
March 28, 2001
Location: Diesel-powered ship from
Tokyo, Japan to Showa Station on Ongul
Island, Antarctica for 366 days (from
February 1, 2000) and then heading back
to Japan on February 1, 2001
Period of Study: Winter 1999-2000,
Winter 2000-2001
Location: Denver, CO
Age Groups: Adults (>40 years-old)
with C0PD, as well as > 10 pack-years
tobacco use, FEVi < 70%, FEV1/FVC <
60%, and no other lung disease
Study Design: C0PD patient panel study
(2 independent panels
one for each winter)
N: 34 subjects (16 1st winter, 18 second
winter)
Statistical Analyses: mixed effects
models with first-order, autoregressive,
moving avg variance-covariance
binary outcomes (rescue medication use,
total symptom score) assessed using
Poisson regression with GEE and first-
order, auto-regressive variance-
covariance
Covariates: temperature, relative
humidity, barometric pressure
Pollutant: PM2.6
Averaging Time: 24-h
Mean (SD):
Winter 1999-2000: 9.0(5.2)
Winter 2000-2001: 14.3(9.6)
Percentiles:
Winter 1999-2000
25th 5.4
50th(Median): 7.7
75th: 11.3
Winter 2000-2001
25th 7.6
50th(Median): 11.7
75th: 17.2
Range (Min, Max): Winter 1999-2000
Cigarette smoking- 0.182, p — 0.004
Support staff- 0.076, p-0.228
Total PM- 0.158, p-0.008
Effect Estimate:
Multiple regression analysis between
inhaled factors in Antarctica
Total leukocyte
Cigarette smoking- 0.211, p< 0.001
Support staff- 0.139, p-0.024
Total PM- 0.168, p-0.004
Segmented PMN
Cigarette smoking- 0.015, p — 0.805
Support staff- 0.097, p —0.119
Total PM- 0.272, p< 0.001
Band-formed PMN
Cigarette smoking- 0.035, p — 0.543
Support staff- 0.010, p-0.864
Total PM - 0.470, p< 0.001
Monocyte
Cigarette smoking- 0.081, p —0.187
Support staff- -0.019, p-0.759
Total PM- 0.328, p< 0.001
G-CSF
Cigarette smoking- 0.131, p< 0.038
Support staff- 0.176, p-0.005
Total PM- 0.078, p-0.186
IL-6
Cigarette smoking- 0.182, p-0.004
Support staff- 0.076, p-0.228
Total PM- 0.158, p-0.008
PM Increment: SD
Winter 1999-2000: 5.2
Winter 2000-2001: 9.6
Model results reported graphically only.
No quantitative results reported.
Direction of slope (+I-) and statistical
significance (SIG: yes
NS: no) inferred from graphs.
Among subjects with severe COPD
observed in Winter 1999-2000,
statistically significant, but marginal,
improvements in PEF associated with
morning lag 0 PM2.6.
There were no statistically significant
associations between rescue medication
use and symptom score with PM.
July 2009
E-251
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)

analysis run separately for each winter
(1.8, 36.6)


Season: Winter
Winter 2000-2001


Dose-response Investigated? No
(3.4, 59.6)


Statistical Package: SAS
Monitoring Stations: multiple sites


Lags Considered: 0-2 days
Copollutant (correlation):
CO
NO2






PM10

Reference: Sivacoumar et al. (2006,
111115)
Period of Study: 4/1998-5/1998
Outcome: Respiratory symptoms,
Decreased pulmonary function
Study Design: Case-control
Pollutant: PM2.6
Averaging Time: 24-h avg
The study does not present quantitative
results of association.
911998-1011998
Statistical Analyses: Poisson


Location: Pammal, India
Age Groups: > 18


Reference: Slaughter et al. (2003,
086294)
Period of Study: 1994
Location: Seattle, WA
Outcome: Asthma attacks, asthma
severity, medication use
Age Groups: 5.1 to 13.1 years old
Study Design: Cross-sectional study
N: 133 children
Statistical Analyses: Ordinal Logistic
Regression
Poisson Modeling
Pollutant: PM2.6
Averaging Time:
Daily Averages
25th: 5.0
50th(Median): 7.33
75th: 11.3
Monitoring Stations: 3
PM Increment: 10/yg/m3 increase
RR Estimate [Lower CI, Upper CI]
lag:
Inhaler use:
1-day lag: 1.04 (0.98,1.10)
OR Estimate [Lower CI, Upper CI]
lag:

Covariates: Temperature, Day of the
Week, Seasonality
Copollutant (correlation):
PM10 - 0.75
Asthma Attack:
1-day lag: 1.20 (1.05,1.37)

Dose-response Investigated? No
CO - 0.82
Previous day: 1.13 (1.03,1.23)

Statistical Package: STATA

Medication Use

Lags Considered: 1, 2, 3 day lag

Nontransition model:
Previous Day: 1.08 (1.01,1.15)
Notes: Figures of estimated odds ratios
for having a more serious asthma attack
for short-term, within-subject increases in
PM2.6, PM10, and CO. Transition models
additionally control for the previous day's
severity.
Figures of estimated relative risks for
having inhaler use for short-term, within-
subject increases in PM2.6, PM10, and CO.
Transition models additionally control for
the previous day's severity.
PM Increment: 10//g|m3
Effects Estimate:
Using the estimated slope for the
validation study model [Lower CI, Upper
CI]
lag:
2.2 percent decrease in FEVi per 10
/yg/m3 increase in ambient PM2.6 [0.0, 4.3
decrease]
1 day
Reference: Strand et al (2006, 089203)
Period of Study: 2002-2004
Location: Denver, Colorado, United
States
Outcome: Reduced forced expiratory
volume (FEVi)
Age Groups: 6-12 years old
Study Design: Mixed model analysis
(using the default restricted maximum
likelihood (REML) estimators)
N: 50 children
Statistical Analyses: least squares
regression, SAS "Output Delivery
System" (ODS)
Season: Autumn and Winter
Dose-response Investigated? Yes
Statistical Package: SAS
Pollutant: PM2.6
Averaging Time: daily
Mean (SD):
Outdoor: 12.699(6.426)
Indoor: 8.148(4.348)
Sulfate/PM2.6/outdoor: 0.079 (0.067)
Sulfate/Plfc/indoor: 0.074 (0.060)
Range (Min, Max):
Mean Personal: (0, 3.035)
Outdoor: (0, 6.303)
Indoor: (0, 2.759)
PM Component: elemental carbon,
sulfate, nitrate and ETS.
July 2009
E-252
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Monitoring Stations: 2 fixed monitors
and up to 10 personal monitors on a given
day.
Copollutant (correlation!: Sulfate (0.63)
Reference: Tang et al. (2007, 091269)
Period of Study: Dec 2003 to Feb 2005
Location: Sin-Chung City, Taipei County,
Taiwan
Outcome: Peak expiratory flow rate
(PEFR) of asthmatic children
Age Groups: 6-12 years
Study Design: Panel study
N: 30 children
Statistical Analyses: Linear mixed-
effect models were used to estimate the
effect of PM exposure on PEFR
Covariates: Gender, age, BMI, history of
respiratory or atopic disease in family,
SHS, acute asthmatic exacerbation in
past 12 months, ambient temp and
relative humidity, presence of indoor
pollutants, and presence of outdoor
pollutants,
Dose-response Investigated? yes
Statistical Package: S-Plus 2000
Lags Considered: 0-2
Pollutant: PM2.B
Averaging Time: 1 h
Mean (SD):
Personal: 27.8 (25.3)
Range (Min, Max):
Personal: 1.4-263.4
Monitoring Stations: 1
PM Increment: 24.5/yg/m3
RR Estimate [Lower CI, Upper CI]
lag:
Change in morning PEFR:
¦6.00 (-29.85,17.85) lag 0
¦12.52 (-77.93, 52.9) lag 1
¦24.87 (-71.49, 21.74) lag 2
¦45.67 (-117.09, 25.74) 2-day mean
¦5.69 (-105.96, 94.59) 3-day mean
Change in evening PEFR:
0.50(-18.82,19.82) lag 0
16.66 (-7.59, 40.9) lag 1
11.60 (-11.1, 34.31) lag 2
39.97 (7.1,72.85) 2-day mean
¦3.32 (-66.14, 59.5) 3-day mean
Reference: Timonen et al. (2004,
087915)
Period of Study: Oct 1998 to April
1999
Location: Amsterdam, Netherlands
Erfurt, Germany
Helsinki, Finland
Outcome: Urinary concentration of Clara
cell protein CC16 of subjects with
coronary heart disease
Age Groups: 50 +
Study Design: Longitudinal cohort study
(panel)
N: 37 (Amsterdam)
47 (Erfurt)
47 (Helsinki)
Statistical Analyses: The response of
interest was log transformed, creatinine
adjusted CC16. Mixed-effect model was
used to investigate the association
between CC16 and air pollutants.
Covariates: Subjects, long term time
trend, temperature (lags 0-3), relative
humidity (lags 0-3), barometric pressure
(lags 0-3), and weekday of visit.
Dose-response Investigated? yes
Statistical Package: S-Plus and SAS
Lags Considered: 0-3
Pollutant: PM2.6
PM Increment: 10/yglm3
Averaging Time: 24 h
RR Estimate [Lower CI, Upper CI]
Mean (SD):
lag:
Amsterdam: 20.0/yg/m3
Pooled estimate;
Erfurt: 23.1 //g|m3
2.8 (-1.1-6.7) lagO
Helsinki: 12.7 //g|m3
2.9 (-0.6-6.5) lag 1
Range (Min, Max):
5.0 (-2.4-12.4) lag 2
Amsterdam: 3.8-82.2
1.6 (-4.7-7.9) lag 3
Erfurt: 4.5-118.1
9.7 (-6.0—25.4) 5-day mean
Helsinki: 3.1-39.8
CC16 was not associated to PM2.6
Monitoring Stations: 3
Copollutant (correlation):
Spearman Correlation:
NC 0.01-0.1: Amsterdam -0.15
Erfurt 0.62
Helsinki 0.14
NCo.i-i.o: Amsterdam 0.80
Erfurt 0.84
in the pooled analysis but CC16 was
significantly associated to PM2.6
in Helsinki:
23.3 (6.3-40.3) lag 0
6.4 (-8.2-21.1) lag 1
20.2 (6.9-33.5) lag 2
17.6 (4.3-30.9) lag 3
38.8 (15.8-61.8) 5-day mean
Helsinki 0.80

NO2: Amsterdam 0.49

Erfurt 0.82

Helsinki 0.35

CO: Amsterdam 0.58

Erfurt 0.77

Helsinki 0.40

Pollutant: PM2.6
PM Increment: 10/yg/m3
Averaging Time: 24-h
ADULT Personal PM2.5 - FEVi
Reference: Trenga et al. (2006,155209) Outcome: Lung function: FEVi, PEF,
MMEF (maximal midexpiratory flow
Period of Study: 1999-2002
July 2009
E-253
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Location: Seattle, WA
assessed only for children)
Age Groups: Adults (56-89-years-old)
healthy & with COPD
asthmatic children 6-13-years-old
Study Design: adult and pediatric panel
study over three years with 1 monitoring
period ("session") per year
N: 57 adults (33 healthy, 24 with COPD)
- 692 subject-days - 207 study-days
17 asthmatic children - 319 subject-
days - 98 study-days
Statistical Analyses: mixed effects,
longitudinal regression models, with the
effects of pollutant decomposed into
each subject's a) overall mean
b)	difference between their session-
specific mean and overall mean
c)	difference between their daily values
and session-specific mean
Covariates: gender, age, ventral site
temperature and relative humidity, 00,
NO?
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1 days
Percentiles:
Children, Personal
25,h: 8.1
50th(Median): 11.3
75th: 16.3
Indoor
25th: 5.7
50th(Median): 7.5
75th: 10.2
Local outdoor
25th: 6.4
50th(Median): 9.6
75th: 14.8
Adults, Personal
25th: 5.9
50th(Median): 8.5
75th: 12.4
Indoor
25th: 5.1
50th(Median): 7.6
75th: 10.8
Local outdoor
25th: 6
50th(Median): 8.6
75th: 13.1
Range (Min, Max):
Children, Personal 1.0, 49.4
Indoor (2.2, 36.3)
Local outdoor (2.8,40.4)
Adults, Personal 1.3, 66.6
lndoor(1.6, 65.3)
Local outdoor(0.0,41.5)
Monitoring Stations: 2
also subject-specific local outdoors (i.e.,
at each home), indoor, and personal
Copollutant (correlation):
CO
NO?
PM2.6
PM10-2.B (coarse)
Overall: Lag 0 -6.0 [-29.1: 17.2]
Lag 1 12.0 [-12.9: 36.9]
No-COPD: Lag 0 -4.6 [-31.0: 21.9]
Lag 1 19.3 [-8.2: 46.7]
COPD: Lag 0 -10.2 [-55.8: 35.4]
Lag 1 -19.0 [-74.1: 36.2]
PEF: Lag 01.5 [-2.2: 5.2]
Lag 1 2.1 [-1.9: 6.1]
No-COPD: Lag 0 3.4 [-0.9: 7.6]
Lag 1 1.9[-2.5: 6.3]
COPD: Lag 0-4.3 [-11.5: 3.0]
Lag 1 2.6 [-6.3: 11.5]
Indoor PM2.5 - FEVi Overall: Lag 0 -12.8
[-44.5:19.0]
Lag 1 19.4[-11.3:50.1]
No-COPD: Lag 0-15.8 [-50.0: 18.4]
Lag 1 28.4[-4.6: 61.3]
COPD: Lag 0 2.6 [-71.7: 76.8]
Lag 1 -29.7 [-102.9: 43.5]
PEF Overall: Lag 0 -0.5 [-5.6: 4.6]
Lag 1 2.3 [-3.3: 7.8]
No-COPD: Lag 0 0.1 [-5.4: 5.6]
Lag 1 2.5 [-3.5: 8.4]
COPD: Lag 0 -3.2 [-15.1: 8.7]
Lag 1 1.1 [-12.0: 14.3]
Outdoor Home PM2.5 - FEVi Overall: Lag
0 -1.4 [-35.6: 32.7]
Lag 1 -2.4 [-37.6: 32.7], No-COPD: Lag 0
1.5 [-36.1: 39.2]
Lag 1 10.7[-26.9:48.4]
COPD: Lag 0-8.9 [-62.2: 44.4]
Lag 1 -45.2 [-102.6: 12.1]
PEF Overall: Lag 0 2.3 [-3.3: 7.9]
Lag 1 0.4[-5.6: 6.4]
No-COPD: Lag 0 4.0 [-2.2:10.1]
Lag 1 2.0 [-4.4: 8.4]
COPD: Lag 0 -1.8 [-10.6: 6.9]
Lag 1 -4.8 [-14.6: 4.9]
Central Sites PM2.5 - FEVi Overall: Lag
0-35.5 [-70.0:-1.0]
Lag 1 -40.4 [-71.1: -9.6], No-COPD: Lag 0
-32.6 [-69.5: 4.3]
Lag 1 -29.0 [-62.5: 4.5]
COPD: Lag 0 -43.6 [-95.0: 7.8]
Lag 1 -70.8 [-118.4: 23.1]
PEF Overall: Lag 0 1.5 [-4.2: 7.1]
Lag 1 -2.3 [-7.4: 2.9]
July 2009
E-254
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
No-COPD: Lag 0 2.5 [-3.5: 8.6]
Lag 1 -0.5 [-6.1: 5.0]
COPD: Lag 0-1.5 [-9.9: 6.9]
Lag 1 -7.1 [-15.0:0.9]
PEDIATRIC FEVi Personal PM2.5
Overall: Lag 0 -13.08 [-38.26: 12.10]
Lag 1 -16.12 [-42.61: 10.37], No Anti-
inflam. Medication: Lag 0 -41.73 [-94.31:
10.84]
Lag 1 -30.99 [-82.17: 20.19], Anti-inflam.
Medication: Lag 0 -4.61 [-34.49: 25.28]
Lag 1 -10.87 [-45.01: 23.27]
Indoor PM2.5 Overall: Lag 0 -45.90 [¦
89.92:1.88]
Lag 1 -64.78 [-111.27: 18.28]
No Anti-inflam. Medication: Lag 0 -75.92
[-145.16: 6.67]
Lag 1 -65.08 [-136.98: 6.82], Anti-inflam.
Medication: Lag 0 -28.50 [-94.72: 37.71]
Lag 1 -64.60-147.23:18.04]
Outdoor Home PM2.5 Overall: Lag 0 -
13.11 [-57.41:31.19]
Lag 1 -9.37 [-54.73: 36.00], No Anti-
inflam. Medication: Lag 0 -24.42 [-81.22:
32.38]
Lag 1 16.52 [-45.76: 78.80], Anti-inflam.
Medication: Lag 0 -3.59 [-75.88: 68.70]
Lag 1-26.76 [-89.53: 36.01]
Central Sites PM2.5. Overall: Lag 0 -
12.32 [-53.21:28.56]
Lag 1 5.75 [-33.27: 44.76], No Anti-
inflam. Medication: Lag 0 -33.59 [-89.99:
22.82]
Lag 1 31.30 [-29.91: 92.51]Anti-inflam.
Medication: Lag 0 -2.13 [-71.99: 67.73]
Lag 1 -3.53 [-67.32: 60.27]
PEF: Personal PM2.5 Overall: Lag 0 0.31
[-4.02: 4.64]
Lag 1 -2.19 [-6.49: 2.12]
No Anti-inflam. Medication: Lag 0 0.22 [¦
8.85: 9.29]
Lag 1 -10.48 [-18.68: 2.28]
Anti-inflam. Medication: Lag 0 0.34 [¦
4.67: 5.35]
Lag 1 0.74[-4.21:5.69]
Indoor PM2.5 Overall: Lag 0 -8.68 [¦
16.64: -0.72]
Lag 1 -9.22 [-17.51: -0.93]
No Anti-inflam. Medication: Lag 0 -13.34
[-25.90: -0.79]
Lag 1 -17.13 [-29.86: 4.41], Anti-inflam.
Medication: Lag 0 -5.98 [-15.85: 3.89]
Lag 1 -4.19 [-14.59: 6.20]
JlutdooiJJomeJ^M^^^
July 2009
E-255
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
6.27 [-14.07: 1.53]
Lag 1 -5.64 [-13.73: 2.44], NoAnti-
inflam. Medication: Lag 0-7.52 [-17.56:
2.51]
Lag 1 -6.92 [-18.03: 4.19], Anti-inflam.
Medication: Lag 0 -5.22 [-14.77: 4.34]
Lag 1 -4.78 [-14.42: 4.86]
Central Sites PM2.5
Overall: Lag 0 -5.62 [-12.86: 1.62]
Lag 1 -2.45 [-9.34: 4.43], No Anti-inflam.
Medication: Lag 0 -6.32 [-16.31: 3.68]
Lag 1 -0.83 [-11.60: 9.95]
Anti-inflam. Medication: Lag 0-5.29 [¦
13.42:2.85]
Lag 1 -3.04 [-10.76: 4.67]
MMEF - Personal PM2.5
Overall: Lag 0 -0.99 [-3.96:1.98]
Lag 1 -1.08 [-4.05: 1.88], No Anti-inflam.
Medication: Lag 0 -3.32 [-9.52: 2.88]
Lag 1 -2.49 [-8.23: 3.25], Anti-inflam.
Medication: Lag 0 -0.31 [-3.77: 3.16]
Lag 1 -0.59 [-4.06: 2.89]
Indoor PM2.5
Overall: Lag 0-3.29 [-8.52:1.94]
Lag 1 -11.08 [-16.26: 5.90], No Anti-
inflam. Medication: Lag 0-12.65 [-20.74:
¦4.56]' Lag 1 -13.84 [-21.82: 5.85]/. Anti-
inflam. Medication: Lag 0 2.14 [-4.17:
8.45]
Lag 1 -9.33 [-15.89:-2.78]
Outdoor Home PM2.5 Overall: Lag 0 -
4.13[-9.28: 1.01]
Lag 1 -0.73 [-6.02: 4.56]
No Anti-inflam. Medication: Lag 0 -8.23 [¦
14.77:1.69]
Lag 1 -1.19 [-8.45: 6.07]
Anti-inflam. Medication: Lag 0-0.68 [¦
6.87: 5.50]
Lag 1 -0.42 [-6.72: 5.87]
Central Sites PM2.5. Overall: Lag 0 -
2.10 [-6.99: 2.79]
Lag 1 -0.12 [-4.67: 4.42]
No Anti-inflam. Medication: Lag 0 -8.21 [¦
14.79:1.62]
Lag 1 -0.22 [-7.34: 6.90]
Anti-inflam. Medication: Lag 0 0.82 [¦
4.48: 6.12], Lag 1-0.09 [-5.19: 5.01]
Reference: Tang et al. (2007, 091269) Outcome: Peak expiratory flow rate Pollutant: PM2 e 1	No quantitative effects reported.
(PEFR) of asthmatic children
Period of Study: Dec 2003 to Feb 2005	Averaging Time: 1 h
Age Groups: 6-12 years
Location: Sin-Chung City, Taipei County,	Mean (SD):
Taiwan	Study Design: Panel study
Personal: 6.2 (4.8)
l\l: 30 children
July 2009
E-256
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Statistical Analyses: Linear mixed-
effect models were used to estimate the
effect of PM exposure on PEFR
Covariates: Gender, age, BMI, history of
respiratory or atopic disease in family,
SHS, acute asthmatic exacerbation in
past 12 months, ambient temp and
relative humidity, presence of indoor
pollutants, and presence of outdoor
pollutants,
Dose-response Investigated? yes
Statistical Package: S-Plus 2000
Lags Considered: 0-2
Range (Min, Max):
Personal: 0.3-86.8
Monitoring Stations: 1
Reference: Tang et al. (2007, 091269)
Period of Study: Dec 2003 to Feb 2005
Location: Sin-Chung City, Taipei County,
Taiwan
Outcome: Peak expiratory flow rate
(PEFR) of asthmatic children
Age Groups: 6-12 years
Study Design: Panel study
N: 30 children
Statistical Analyses: Linear mixed-
effect models were used to estimate the
effect of PM exposure on PEFR
Covariates: Gender, age, BMI, history of
respiratory or atopic disease in family,
SHS, acute asthmatic exacerbation in
past 12 months, ambient temp and
relative humidity, presence of indoor
pollutants, and presence of outdoor
pollutants,
Dose-response Investigated? yes
Statistical Package:
S-Plus 2000
Lags Considered: 0-2
Pollutant: PMi
Averaging Time: 1 h
Mean (SD):
Personal: 34.0 (28.9)
Ambient: 31.4 (18.8)
Range (Min, Max):
Personal: 1.8-284.6
Ambient: 0.1-128.4
Unit (i.e. //g/m3): //g/m3
Monitoring Stations: 1
PM Increment: 27.6//g/m3
RR Estimate [Lower CI, Upper CI]
lag:
Change in morning PEFR:
¦6.44 (-30.18,17.29) lag 0
¦12.26 (-77.6, 53.09) lag 1
¦4.38 (-54.79, 46.03) lag 2
¦44.06 (-113.79, 25.67) 2-day mean
¦6.01 (-101.48, 89.46) 3-day mean
Change in evening PEFR:
I.17	(-17.79, 20.13) lagO
¦4.98 (-27.77,17.81) lag 1
II.30	(-11.55, 34.16) lag 2
41.74(11.36,72.13) 2-day mean
28.21 (-19.08, 75.5) 3-day mean
Reference: Timonen et al. (2004,
087915)
Period of Study: Oct 1998 to April 1999
Location: Amsterdam, Netherlands
Erfurt, Germany
Helsinki, Finland
Outcome: Urinary concentration of Clara
cell protein CC16 of subjects with
coronary heart disease
Pollutant: IMC 0.01-0.1
Averaging Time: 24 h
PM Increment: 10,000 /cm3
RR Estimate [Lower CI, Upper CI]
Age Groups: 50 +
Mean (SD):
lag:
Study Design: Longitudinal cohort study
Amsterdam: 17338 /cm3
Pooled estimate;
(panel)
Erfurt: 21124 /cm3
1.7 (-4.4-7.8) lag 0
N:
Helsinki: 17041 /cm3
¦1.8 (-8.3-4.6) lag 1
l\l - 37 (Amsterdam)
Range (Min, Max):
1.5 (-5.6-8.6) lag 2
N-47 (Erfurt)
Amsterdam: 5699-37195
2.3 (-4.8-9.3) lag 3
N-47 (Helsinki)
Erfurt: 3867-96678
1.8 (-9.4-13.0) 5-day mean
Statistical Analyses: The response of
interest was log transformed, creatinine
adjusted CC16. Mixed-effect model was
used to investigate the association
between CC16 and air pollutants.
Helsinki: 2305-50306
Unit (i.e. //g/m3): 1/cm3
Monitoring Stations: 3
There was no association between NC 0.01
11 and CC16 in the pooled analysis.
Covariates: Subjects, long term time
trend, temperature (lags 0-3), relative
humidity (lags 0-3), barometric pressure
(lags 0-3), and weekday of visit.
PM2.6:
Amsterdam -0.15
Erfurt 0.62

Dose-response Investigated? yes
Helsinki 0.14

Statistical Package:
NO?:

S-Plus and SAS
Amsterdam 0.49

Lags Considered: 0-3
Erfurt 0.82

July 2009
E-257
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)

Helsinki 0.72


CO:


Amsterdam 0.22


Erfurt 0.72


Helsinki 0.35

Outcome: Urinary concentration of Clara
cell protein CC16 of subjects with
coronary heart disease
Pollutant: IMC 10-0.1
Averaging Time: 24 h
PM Increment: 1000 /cm3
RR Estimate [Lower CI, Upper CI]
Age Groups: 50 +
Mean (SD):
lag:
Study Design: Longitudinal cohort study
Amsterdam: 2131 /cm3
Pooled estimate;
(panel)
Erfurt: 1829 / cm3
4.3 (-1.4-10.0) lagO
N:
Helsinki: 1390 /cm3
5.1 (-0.6-10.7) lag 1
l\l - 37 (Amsterdam)
Range (Min, Max):
4.5 (-0.5-9.6) lag 2
N-47 (Erfurt)
Amsterdam: 413-6413
1.6 (-3.5-6.7) lag 3
N-47 (Helsinki)
Erfurt: 303-6848
13.1 (-4.3—30.5) 5-day mean
Statistical Analyses: The response of
interest was log transformed, creatinine
adjusted CC16. Mixed-effect model was
used to investigate the association
between CC16 and air pollutants.
Helsinki: 344-3782
Unit (i.e./yg/m3): 1/cm3
Monitoring Stations: 3
CC16 was not associated to NC 0.1-1.0
in the pooled analysis but CC16 was
significantly associated to NC 0.1-1.0 in
Helsinki:
Covariates: Subjects, long term time
trend, temperature (lags 0-3), relative
humidity (lags 0-3), barometric pressure
(lags 0-3), and weekday of visit.
Copollutant (correlation):
Spearman Correlation:
NC 0.1-0.01:
15.5 (0.001-30.9) lag 0
10.8 (-4.2-25.8) lag 1
10.5 9-4.1-25.1) lag 2
Dose-response Investigated? yes
Amsterdam 0.16
17.4 (3.4-31.4) lag 3
Statistical Package: S-Plus and SAS
Erfurt 0.67
43.2 (17.4-69.0) 5-day mean
Lags Considered: 0-3
Helsinki 0.53
PM2.6:
Amsterdam 0.80
Erfurt 0.84
Helsinki 0.80
NO?:
Amsterdam 0.67
Erfurt 0.82
Helsinki 0.72

Reference: Timonen et al. (2004,
087915)
Period of Study: Oct 1998 to April 1999
Location: Amsterdam, Netherlands
Erfurt, Germany
Helsinki, Finland
CO:
Amsterdam 0.60
Erfurt 0.78
Helsinki 0.51
Reference: von Klot et al. (2002,
034706)
Period of Study: September 1996 to
March 1997 (winter)
Location: Erfurt, Germany
Outcome: Asthma symptoms (wheezing,
shortness of breath at rest, waking up
with breathing problems, or coughing
without having a cold) and Asthma
medication (inhaled short-acting B2-
agonists, inhaled long-acting B2- agonists,
inhaled corticosteroids, cromolyn sodium,
theophylline, oral corticosteroids, and N-
acetylcysteine)
Age Groups: Adults, mean-59.0 yrs and
range —37-77 yrs
Study Design: panel study
N: 53 adult asthmatics
Statistical Analyses: Logistic regression
Pollutant: MCob-o.i
Averaging Time: 10 min intervals
Mean (SD): 24.8
Percentiles:
25th: 11.4
50th(Median): 19.6
75th: 33.1
Range (Min, Max): (2.4-108.3)
Copollutant (correlation):
PMio-2.b: r- 0.51
NC Increment: 1 IQR
Effect Estimate [Lower CI, Upper CI]:
Association between the prevalence of
inhaled B2- agonist use and MC0.1-0.5
Same day, IQR - 21, OR - 0.98(0.92-
1.04)
5-day mean, IQR- 21 OR- 1.11 (1.02-
1.20)
14-day mean IQR - 17, OR - 1.01 (0.93-
1.10)
Association between the prevalence of
inhaled corticosteroid use and MC0.1-0.5
July 2009
E-258
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
models
NCo.i-o.oi: r- 0.45
Covariates: seasonal variation in
NCo.b-o. 1: r- 0.95
medication use or symptom prevalence,
meteorological factors (relative humidity,
NC2.B0.B: r- 0.92
temperature), weekend, Christmas

holidays
MC2.B0.01: r- 1.00
Season: winter
PM10: r- 0.91
Dose-response Investigated? No
NO2: r- 0.69
Statistical Package: NR
CO: r- 0.66
Lags Considered: 0,1, 2, 3, 4, 5, 6, 7, 8,
SO2: r- 0.60
9,10, mov avg calculated from same day

and preceding days

Same day, IQR- 2, OR - 1.09 (1.02-1.17)
5-day mean IQR- 21, OR- 1.28(1.18-
1.39)
14-day mean, IQR - 17, OR- 1.49(1.38-
1.61)
Association between the prevalence of
wheezing and MC0.1-0.5
Same day, IQR - 21, OR - 1.01 (0.94-
1.08)
5-day mean, IQR - 21, OR - 1.08 (0.99-
1.17)
14-day mean, IQR - 17, OR- 1.05 (0.96-
1.15)
Reference: von Klot et al. (2002,
034706)
Period of Study: September 1996 to
March 1997 (winter)
Location: Erfurt, Germany
Outcome: Asthma symptoms (wheezing,
shortness of breath at rest, waking up
with breathing problems, or coughing
without having a cold) and Asthma
medication (inhaled short-acting B2-
agonists, inhaled long-acting B2- agonists,
inhaled corticosteroids, cromolyn sodium,
theophylline, oral corticosteroids, and N-
acetylcysteine)
Age Groups: Adults, mean-59.0 yrs and
range —37-77 yrs
Study Design: panel study
N: 53 adult asthmatics
Statistical Analyses: Logistic regression
models
Covariates: seasonal variation in
medication use or symptom prevalence,
meteorological factors (relative humidity,
temperature), weekend, Ohristmas
holidays
Season: Winter
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0,1, 2, 3, 4, 5, 6, 7, 8,
9,10, mov avg calculated from same day
and preceding days
Pollutant: MC2.B0.01
Averaging Time: 10 min intervals
Mean (SD): 30.3
Percentiles:
25th: 13.5
50th(Median): 24.6
75th: 41.3
Range (Min, Max): (3.6-133.8)
Copollutant (correlation):
PMio-2.b: r- 0.52
NCo.e-o.i: r- 0.45
NC2.B0.B: r- 0.94
MCo.b-o.i: r- 1.00
NCo.i-o.oi: r- 0.45
PM10: r- 0.94
NO?: r- 0.68
CO: r- 0.65
SO2: r- 0.62
NC Increment: 1 IQR
Effect Estimate [Lower CI, Upper CI]:
Association between the prevalence of
inhaled B2- agonist use and MC0.01-2.5
Same day, IQR - 28, OR - 0.96 (0.90-
1.04)
5-day mean, IQR- 26, OR- 1.10(1.01-
1.20)
14-day mean, IQR - 20, OR- 1.03 (0.95-
1.12)
Reference: von Klot et al. (2002,
034706)
Period of Study: September 1996 to
March 1997 (winter)
Location: Erfurt, Germany
Outcome: Asthma symptoms (wheezing,
shortness of breath at rest, waking up
with breathing problems, or coughing
without having a cold) and Asthma
medication (inhaled short-acting B2-
agonists, inhaled long-acting B2- agonists,
inhaled corticosteroids, cromolyn sodium,
theophylline, oral corticosteroids, and N-
acetylcysteine)
Age Groups: Adults, mean-59.0 yrs and
range —37-77 yrs
Study Design: panel study
N: 53 adult asthmatics
Statistical Analyses: Logistic regression
models
Covariates: seasonal variation in
medication use or symptom prevalence,
meteorological factors (relative humidity,
temperature), weekend, Christmas
holidays
Pollutant: NCo.1 -0.01
Averaging Time: 10 min intervals
Mean (SD): 173001cm3
Percentiles:
25th: 9286
50th(Median): 16940
75th: 24484
Range (Min, Max): (3272-46195)
Unit (i.e./yg/m3): 1/cm3
Copollutant (correlation):
PMio-2.b: r- 0.41
NCo.e-o.i: r- 0.55
NC2.B0.B: r- 0.34
MCob-o i: r- 0.45
NC Increment: 1 IQR
Effect Estimate [Lower CI, Upper CI]:
Association between the prevalence of
inhaled B2- agonist use and NC0.01-0.1
Same day, IQR- 15000, OR - 0.97 (0.90-
1.04)
5-day mean, IQR - 10000, OR - 1.11
(1.01-1.21)
14-day mean, IQR - 7700, OR - 1.08
(0.96-1.21)
Association between two pollutants,
jointly in one model, and the Outcomes
Inhaled short-acting B2- agonist use
NCo.i-o.oi OR- 1.07 (0.97-1.18)
MCob-oi: OR- 1.07 (0.98-1.18)
Inhaled corticosteroid use
July 2009
E-259
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Season: winter
MC2.B0.01: r- 0.45
NCo.i-o.oi OR- 1.01 (0.87-1.18)
Dose-response Investigated? No
PM10: r- 0.51
MCob-oi: OR- 1.53 (1.39-1.69)
Statistical Package: NR
NO?: r- 0.66
Wheezing
Lags Considered: 0,1, 2, 3, 4, 5, 6, 7, 8,
CO: r- 0.66
NCo.i-o.oi OR - 1.12(1.01-1.24)
9,10, mov avg calculated from same day


and preceding days
SO2: r- 0.36
MCob-oi: OR- 1.02 (0.92-1.12)


Association between the prevalence of


inhaled corticosteroid use and NCO.01-0.1


Same day, IQR- 15000, OR- 1.07 (1.00-


1.15)


5-day mean, IQR - 10000, OR - 1.22


(1.12-1.33)


14-day mean, IQR - 7700, OR - 1.45


(1.29-1.63)


Association between the prevalence of


wheezing and NCo.i-o.oi


Same day, IQR- 15000, OR - 0.94 (0.86-


1.01)


5-day mean, IQR - 10000, OR - 1.13


(1.03-1.24)


14-day mean, IQR - 7700, OR - 1.27


(1.13-1.43)


Association between the prevalence of


respiratory symptoms and NCo.i-o.oi


Attack of shortness of breath and


wheezing


Same day, IQR- 15000, OR- 1.01 (0.91-


1.12)


5-day mean, IQR - 10000, OR - 1.08


(0.96-1.21)


14-day mean, IQR - 7700, OR - 1.26


(1.08-1.48)


Walking up with breathing problems


Same day, IQR- 15000, OR- 1.04 (0.96-


1.13)


5-day mean, IQR - 10000, OR - 1.09


(0.99-1.19)


14-day mean, IQR - 7700, OR - 1.26


(1.13-1.41)


Shortness of breath


Same day, IQR- 15000, OR - 0.98 (0.90-


1.06)


5-day mean, IQR - 10000, OR - 1.09


(0.99-1.19)


14-day mean, IQR - 7700, OR - 1.24


(1.11-1.40)


Phlegm


Same day, IQR- 15000, OR- 1.01 (0.94-


1.09)


5-day mean, IQR - 10000, OR - 1.11


(1.02-1.21)


14-day mean, IQR - 7700, OR - 1.11


(0.99-1.25)


Cough


Same day, IQR- 15000, OR- 1.07 (0.98-


1.16)
July 2009
E-260
DRAFT-DO NOT CITE OR QUOTE

-------
Study
& Methods
Concentrations'
Effect Estimates (95% CI)
5-day mean, IQR- 10000, OR - 1.17
(1.07-1.28)
14-day mean, IQR - 7700, OR - 1.20
(1.06-1.35)
Reference: von Klot et al. (2002,
034706)
Period of Study: September 1996 to
March 1997 (winter)
Location: Erfurt, Germany
Outcome: Asthma symptoms (wheezing,
shortness of breath at rest, waking up
with breathing problems, or coughing
without having a cold) and Asthma
medication (inhaled short-acting B2-
agonists, inhaled long-acting B2- agonists,
inhaled corticosteroids, cromolyn sodium,
theophylline, oral corticosteroids, and N-
acetylcysteine)
Age Groups: Adults, mean-59.0 yrs and
range —37-77 yrs
Study Design: panel study
N: 53 adult asthmatics
Statistical Analyses: Logistic regression
models
Covariates: seasonal variation in
medication use or symptom prevalence,
meteorological factors (relative humidity,
temperature), weekend, Christmas
holidays
Season: winter
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0,1, 2, 3, 4, 5, 6, 7, 8,
9,10, mov avg calculated from same day
and preceding days
Pollutant: NCo.bo i
Averaging Time: 10 min intervals
Mean (SD): 2005 I cm3
Percentiles:
25th: 958
50th(Median): 1610
75th: 2767
Range (Min, Max): (291-6700)
Unit (i.e./yg/m3): 1/cm3
Copollutant (correlation):
PMio-2.b: r- 0.50
NCo.i-o.oi: r- 0.55
NC2.b-o.b: r- 0.76
MCob-o.i: r- 0.95
MC2.B0.01: r- 0.93
PM10: r- 0.85
NO2: r- 0.75
CO: r- 0.79
SO2: r- 0.51
NC Increment: 1 IQR
Effect Estimate [Lower CI, Upper CI]:
Association between the prevalence of
inhaled B2- agonist use and NCobo.i
Same day, IQR - 1800, OR - 0.99 (0.92-
1.05)
5-day mean, IQR- 1500, OR- 1.10
(1.03-1.19)
14-day mean, IQR- 1450, OR- 0.95
(0.86-1.05)
Association between the prevalence of
inhaled corticosteroid use and NCo.b-oi
Same day, IQR- 1800, OR- 1.06 (0.99-
1.14)
5-day mean, IQR - 1500, OR - 1.23
(1.14-1.32)
14-day mean, IQR- 1450, OR- 1.51
(1.37-1.67)
Association between the prevalence of
wheezing and NCobo.i
Same day, IQR - 1800, OR- 1.00 (0.93-
1.07)
5-day mean, IQR - 1500, OR - 1.08
(1.00-1.17)
14-day mean, IQR- 1450, OR- 1.11
(1.00-1.24)
Reference: von Klot et al. (2002,
034706)
Period of Study: September 1996 to
March 1997 (winter)
Location: Erfurt, Germany
Outcome: Asthma symptoms (wheezing,
shortness of breath at rest, waking up
with breathing problems, or coughing
without having a cold) and Asthma
medication (inhaled short-acting B2-
agonists, inhaled long-acting B2- agonists,
inhaled corticosteroids, cromolyn sodium,
theophylline, oral corticosteroids, and N-
acetylcysteine)
Age Groups: Adults, mean-59.0 yrs and
range —37-77 yrs
Study Design: panel study
N: 53 adult asthmatics
Pollutant: NC2.B0.B
Averaging Time: 10 min intervals
Mean (SD): 21.41 cm3
Percentiles:
25th: 5.6
50th(Median): 13.0
75th: 31.6
Range (Min, Max): (0.9-127.6)
Unit (i.e./yg/m3): 1/cm3
NC Increment: 1 IQR
Effect Estimate [Lower CI, Upper CI]:
Association between the prevalence of
inhaled B2- agonist use and NC2.E-0.E
Same day, IQR - 26, OR - 0.99 (0.93-
1.05)
5-day mean, IQR - 22, OR- 1.09 (1.01-
1.17)
14-day mean, IQR - 17, OR- 1.08(1.02-
1.15)
Association between the prevalence of
inhaled corticosteroid use and NC2.E-0.E
Same day, IQR - 26, OR- 1.13(1.06-
1.21)
5-day mean, IQR - 22, OR- 1.28 (1.19-
1.37)
14-day mean, IQR - 17, OR- 1.44(1.36-
1.53)
Association between the prevalence of
wheezing and NC2.E-0.E
Same day, IQR - 26, OR - 1.03 (0.95-
1.10)
5-day mean, IQR - 22, OR - 1.05 (0.97-
1.13)
J4^da^eanJQR^J7;J)R^_J;03J0;S^
Statistical Analyses: Logistic regression Copollutant (correlation):
models
PM10-2.B: r- 0.48
NCo.i-o.oi: r- 0.34
Covariates: seasonal variation in
medication use or symptom prevalence,
meteorological factors (relative humidity, |\|Cob-o 1: r- 0.76
temperature), weekend, Christmas
holidays	MCob-o i: r- 0.92
Season: winter	MC2.B-0.01: r— 0.94
Dose-response Investigated? No	PM10: r- 0.88
Statistical Package: NR	NO2: r- 0.54
Lags Considered: 0,1, 2, 3, 4, 5, 6, 7, 8, CO: r- 0.46
9,10, mov avg calculated from same day
and preceding days	r" "-66
July 2009
E-261
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1.10)
Reference: von Klot et al. (2002,
034706)
Period of Study: September 1996 to
March 1997 (winter)
Location: Erfurt, Germany
Outcome: Asthma symptoms (wheezing,
shortness of breath at rest, waking up
with breathing problems, or coughing
without having a cold) and Asthma
medication (inhaled short-acting B2-
agonists, inhaled long-acting B2- agonists,
inhaled corticosteroids, cromolyn sodium,
theophylline, oral corticosteroids, and N-
acetylcysteine)
Age Groups: Adults, mean-59.0 yrs and
range —37-77 yrs
Study Design: panel study
N: 53 adult asthmatics
Statistical Analyses: Logistic regression
models
Covariates: seasonal variation in
medication use or symptom prevalence,
meteorological factors (relative humidity,
temperature), weekend, Christmas
holidays
Season: winter
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0,1, 2, 3, 4, 5, 6, 7, 8,
9,10, mov avg calculated from same day
and preceding days
Pollutant: PM102 E
Averaging Time: 24 h
Mean (SD): 10.3
Percentiles:
25th: 2.9
50th(Median): 6.9
75th: 14.6
Range (Min, Max): (-8.7-64.3)
Copollutant (correlation):
NCo.i-o.oi: r- 0.41
NCo.5-0.1: r- 0.50
NC2.b-o.b: r- 0.48
MCobo.i: r- 0.51
MC2.B0.01: r- 0.52
PM10: r- 0.67
NO2: r- 0.45
CO: r- 0.42
SO2: r- 0.28
PM Increment: 1 IQR
Effect Estimate [Lower CI, Upper CI]:
Association between the prevalence of
inhaled B2- agonist use and PM 10-2.6
Same day, IQR- 12, OR - 1.01 (0.95-
1.06)
5-day mean, IQR - 11, OR - 1.01 (0.94-
1.09)
14-day mean, IQR - 6.7, OR - 0.92 (0.86-
1.00)
Association between the prevalence of
inhaled corticosteroid use and PM10-2.E
Same day, IQR- 12, OR - 1.03(0.98-
1.08)
5-day mean,IQR- 11,OR- 1.12(1.04-
1.20)
14-day mean, IQR - 6.7, OR - 1.27 (1.18-
1.37)
Association between the prevalence of
wheezing and PM10 2.5
Same day, IQR- 12, OR- 0.97(0.91-
1.02)
5-day mean, IQR - 11, OR - 1.06 (0.98-
1.15)
14-day mean, IQR - 6.7, OR - 1.05 (0.96-
1.15)
Reference: Ward et al. (2002, 025839)
Period of Study: 1997 (two 8-week
periods)
Outcome: Change in PEF (peak
expiratory flow), self reported respiratory
symptoms (same day cough, illness, short
of breath, waking up at night with cough
Location: Birmingham and Sandwell, UK or w'leeze' wheeze)
Age Groups: 9 year olds
Study Design:
Time-series panel study
N: 162 children from 5 schools
Statistical Analyses: Linear regression
(PEF),
Logistic regression (respiratory
symptoms)
Covariates: Trend, temperature,
schooldays (yes/no)
Season: Winter (Jan 13-Mar 10)
Summer (May 19- July 14)
Dose-response Investigated? No
Statistical Package: Nr
Lags Considered: Lag 0, lag 1, lag 2, lag
3, 7-day moving avg
Pollutant: PM2.6
PM Increment:
Averaging Time: 24-h
Winter: 12.3 //g|m3
Mean (SD):
Summer: 6.3 //g|m3
Winter: 12.7 //g|m3
Mean (PEF l/min) [Lo
Summer: 12.3 //g|m3
lag:
Range (Min, Max):
Winter morning:
Winter: 4, 37
0.80 [-1.97, 3.67]
Summer: 5, 28
lagO
PM Component:
0.62 [-2.22, 3.54]
Total mass
lag 1
Monitoring Stations:
¦0.86 [-4.32, 2.47]
5 stations near the 5 schools
lag 2
Copollutant (correlation):
¦2.47 [-5.30, 0.36]
Winter:
lag 3
PMio(r—0.93)
¦4.07 [-10.60, 2.42]
NO2 (r—0.88)
7-day mean
0s (r—-0.83)
Winter afternoon:
Summer:
0.95 [-2.22,4.23]
HNOs (r — 0.81)
lagO
¦0.99 [-4.69, 2.72]
, Upper CI]
July 2009
E-262
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
¦1.60 [-5.18, 2.01]
lag 2
¦3.45 [-6.53 to -0.25]
lag 3
1.00 [-11.47,13.56]
7-day mean
Summer morning:
-1.49 [-3.65, 0.67]
lag 0
0.21 [-2.12,2.55]
Iag1
2.50(0.28, 4.72]
Iag2
3.41 [1.40,5.44]
Iag3
3.901-2.53,10.33]
7-day mean
Summer afternoon:
¦0.49 [-2.43,1.45]
lag 0
-0.78 [-2.72,1.16]
lag 1
0.57 [-1.35, 2.49]
lag 2
0.16 [-1.85, 2.17]
lag 3
¦0.08 [-5.43, 5.27]
7-day mean
Winter morning in atopy/recent
wheezing subgroup:
-0.0721-0.527,0.383]
lag 0
-0.271 [-0.701,0.159]
lag 1
0.127 [-0.354,0.608]
lag 2
0.055 [-0.391,0.501]
lag 3
Winter morning in no atopy or recent
wheezing subgroup:
0.126 [-0.413,0.666]
lag 0
0.193 [-0.340,0.728] lag 1
-0.170 [-0.788, 0.447]; Iag2
-0.314[-0.846,0.216]
lag 3
July 2009
E-263
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Winter morning in subgroup with
parental atopy/recent wheezing:
0.187 [-0.008,0.382]
lag 0
-0.006 [-0.207,0.195]
lag 1
-0.011 [-0.226,0.204]
lag 2
-0.037 [-0.228,0.154]
lag 3
Winter morning in subgroup without
parental atopy/recent wheezing:
0.026 [-0.341 ,0.395]
lag 0
0.068 [-0.307,0.444]
lag 1
-0.099 [-0.535,0.335]
lag 2
-0.252 [-0.615,0.110]
lag 3
RR Estimate [Lower CI, Upper CI]
lag:
Cough:
Winter: 0.98 [0.80,1.18]
lag 0
0.95[0.77,1.17]
lag 1
1.02[0.83,1.24]
lag 2
1.01 [0.83,1.23]
lag 3
1.31 [0.82,2.09]
7-day mean
Summer: 1.13 [1.04,1.22]
lag 0
1.04 [0.94,1.13]
lag 1
0.94 [0.87,1.02]
lag 2
0.89(0.82,0.96]
lag 3
0.81 [0.62,1.06]
7 day mean
Illness:
Winter: 1.17 [1.05,1.32]
lag 0
July 2009
E-264
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1.07[0.95,1.23]
lag 1
1.16[1.01,1.35]
lag 2
1.01 [0.90,1.16]
lag 3
1.57[1.15, 2.13]
7-day mean
Summer: 1.02 [0.91,1.13]
lag 0
1.00(0.89,1.13]
lag 1
0.96(0.85,1.07]
lag 2
0.97(0.86,1.09]
lag 3
0.68(0.41,1.13]
7-day mean
Shortness of breath:
Winter: 1.07 [0.94,1.24]
lag 0
0.98[0.84,1.13]
lag 1
0.96(0.82,1.13]
Iag2
0.91 (0.79,1.07]
lag 3
0.82(0.58,1.18]
7-day mean
Summer: 1.04 (0.90,1.20]
lag 0
1.08(0.93,1.25]
lag 1
0.97(0.84,1.13]
lag 2
0.93(0.81,1.08]
lag 3
1.16(0.76,1.77]
7-day mean
Wake at night with cough/wheeze:
Winter: 1.10 (0.96,1.26]
lag 0
1.05(0.90,1.22]
lag 1
0.98(0.83,1.13]; lag 2
July 2009
E-265
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0.94[0.81,1.09]; lag 3
0.93[0.66,1.32]
7-day mean
Summer: 0.93 [0.78,1.10]
lag 0
0.81 [0.67, 0.98]
lag 1
0.91 [0.77,1.09]
lag 2
0.97(0.83,1.13]
lag 3
1.04 [0.57,1.90]
7-day mean
Wheeze:
Winter: 0.98 [0.83,1.16]
lag 0
0.90(0.75,1.05]
lag 1
1.00(0.83,1.20]
lag 2
1.13(0.95,1.35]
lag 3
1.02 [0.68,1.57]; 7-day mean
Summer: 1.02 [0.88,1.19]
lag 0
0.98(0.84,1.16]
lag 1
0.87(0.74,1.02]
lag 2
0.85[0.72,0.99]
lag 3
0.96(0.51,1.81]
7-day mean
Reference: Ward et al. (2002, 025839)
Outcome: Change in PEF (peak
expiratory flow), self reported respiratory
Pollutant: Sulfate
PM Increment:
Period of Study: 1997 (two 8-week
symptoms (same day cough, illness, short
Averaging Time: 24-h
Winter: 4.8 //g|m3
periods)
Location: Birmingham and Sandwell, UK
of breath, waking up at night with cough
or wheeze, wheeze)
Mean (SD):
Summer: 3.1 //g|m3
Age Groups: 9 year olds
Winter: 2.4/yg/m3
Mean (PEF l/min) [Lower CI, Upper CI]

Study Design:
Summer: 3.8/yg/m3
la

Time-series panel study
Range (Min, Max):
Winter morning:

N: 162 children from 5 schools
Winter: 0.8,14.9
¦1.75 [-4.00, 0.50]

Statistical Analyses: Linear regression
Summer: 1.1, 7.8
lagO

(PEF),
PM Component:
¦0.91 [-3.44,1.62]

Logistic regression (respiratory
S04
lag 1

symptoms)




Monitoring Stations:
¦0.62 [-3.16,1.91]

Covariates: Trend, temperature,


schooldays (yes/no)
2 stations
lag 2
July 2009
E-266
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Season: Winter (Jan 13-Mar 10)
Summer (May 19- July 14)
Dose-response Investigated? No
Statistical Package: Nr
Lags Considered: Lag 0, lag 1, lag
3, 7-day moving avg
¦1.82 [-4.27, 0.64]
lag 3
¦3.22 [-8.03,1.58]
7-day mean
Winter afternoon:
0.99 [-1.58, 3.55]
0.79 [-2.42,4.00]
lag 1
¦1.89 [-4.99,1.21]
lag 2
-1.73 [-4.69,1.23]
lag 3
¦1.96 [-13.35, 9.42]
7-day mean
Summer morning:
-0.72 [-3.27,1.82]
lag 0
¦1.69 [-4.28, 0.90]
1.35 [-1.27, 3.97]
Iag2
3.38 [1.03, 5.72]
Iag3
2.98[-4.17,10.13]
7-day mean
Summer afternoon:
¦0.32 [-2.81, 2.17]
lag 0
0.84 [-1.63, 3.30]
lag 1
¦0.08 [-2.61,2.44]
lag 2
¦0.25 [-2.69, 2.19]
lag 3
¦2.20 [-9.51,5.12]
7-day mean
Winter morning in atopy/recent
wheezing subgroup:
0.200 [-0.755,1.156]
lag 0
-0.219[-1.318,0.881]
lag 1
-0.431 [-1.526,0.664]; lag 2
1.200 [0.095, 2.305]
July 2009
E-267
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
lag 3
Winter morning in no atopy or recent
wheezing subgroup:
-0.6131-1.714,0.488]
lag 0
-0.174 [-1.423,1.075]
lag 1
0.006 [-1.243,1.253]
lag 2
-1.080 [-2.308,0.148]
lag 3
Winter morning in subgroup with
parental atopy/recent wheezing:
0.457 [0.003,0.910]
lag 0
0.0781-0.503,0.660]
lag 1
-0.102 [-0.656,0.452]
lag 2
0.002 [-0.609,0.613]
lag 3
Winter morning in subgroup without
parental atopy/recent wheezing:
-0.622 [-1.379,0.136]
lag 0
-0.272[-1.147,0.602]
lag 1
-0.138 [-1.005,0.728]
lag 2
-0.496 [-1.359,0.367]
lag 3
RR Estimate [Lower CI, Upper CI]
lag:
Cough:
Winter: 1.01 [0.84,1.20]
lag 0
1.02[0.85,1.24]
lag 1
0.99[0.82,1.20]
lag 2
0.86(0.71,1.05]
lag 3
0.78(0.53,1.14]
7-day mean
Summer: 1.08 [0.98,1.20]
lag 0
July 2009
E-268
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1.03[0.93,1.15]
lag 1
0.97[0.88,1.07]
lag 2' 0.90(0.82, 0.99]
lag 3
0.73(0.54,0.97]
7 day mean
Illness:
Winter: 1.06 [0.96,1.17]
lag 0
1.15(1.03,1.28]
lag 1
1.14(1.00,1.28]
lag 2
1.04 (0.92,1.18]
lag 3
1.30(1.00,1.66]
7-day mean
Summer: 0.98 (0.86,1.11]
lag 0
0.97(0.84,1.12]
lag 1
1.01 (0.88,1.16]
lag 2' 0.95(0.84,1.09]
lag 3
0.72(0.46,1.12]
7-day mean
Shortness of breath:
Winter: 0.96 [0.85,1.07]
lagO: 0.98 [0.86,1.12]
lag 1
0.94 [0.82,1.07]
Iag2
0.93 [0.81,1.08]
lag 3
0.80(0.59,1.07]
7-day mean
Summer: 0.95 [0.80,1.14]
lag 0
1.07[0.89,1.28]
lag 1
1.04 [0.87,1.24]
lag 2
0.94 [0.80,1.12]
lag 3
July 2009
E-269
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
10.58 [0.33,1.04]
7-day mean
Wake at night with cough/wheeze:
Winter: 0.97 [0.87,1.08]
lag 0
1.01 [0.89,1.15]
lag 1
1.00(0.88,1.14]; lag 2
0.93[0.82,1.07]; lag 3
0.79(0.59,1.05]
7-day mean
Summer: 0.95 [0.78,1.16]
lag 0
0.81 [0.67, 0.99]
lag 1
0.93(0.76,1.13]
lag 2
0.87(0.72,1.05]
lag 3
0.77[0.41,1.48]
7-day mean
Wheeze:
Winter: 1.00 [0.87,1.15]
lag 0
0.96[0.82,1.13]
lag 1
0.88(0.75,1.04]
lag 2
1.12(0.95,1.32]
lag 3
0.83 [0.58,1.20]; 7-day mean
Summer: 0.97 [0.80,1.17]
lag 0
.09 [0.89,1.32]
lag 1
1.00[0.82,1.22]
lag 2
0.81 [0.69, 0.97]
lag 3
1.30[0.68, 2.50]
7-day mean
PM Increment: Winter: 6.7 /yglm3
Summer: 3.7 /yglm3
Mean (PEF l/min) [Lower CI, Upper CI]
lag:
Reference: Ward et al. (2002, 025839)
Period of Study: 1997 (two 8-week
periods)
Location: Birmingham and Sandwell, UK
Outcome: Change in PEF (peak
expiratory flow), self reported respiratory
symptoms (same day cough, illness, short
of breath, waking up at night with cough
or wheeze, wheeze)
Age Groups: 9 year olds
Pollutant: NO3
Averaging Time: 24-h
Mean (SD):
Winter: 3.6/yg/m3
July 2009
E-270
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Study Design: Time-series panel study
N: 162 children from 5 schools
Summer: 3.5/yg/m3
Range (Min, Max):
Statistical Analyses: Linear regression Winter: 0.1, 29.9
(PEF),
Logistic regression (respiratory
symptoms)
Covariates: Trend, temperature,
schooldays (yes/no)
Season: Winter (Jan 13-Mar 10)
Summer (May 19- July 14)
Dose-response Investigated? No
Statistical Package: Nr
Lags Considered: Lag 0, lag 1, lag
3, 7-day moving avg
Summer: 0.7,13.2
Monitoring Stations:
2 stations
Winter morning:
¦2.08 [-4.02 to-0.15]
lagO
¦0.64 [-2.87,1.59]
lag 1
0.71 [-1.69,3.11]
lag 2
¦1.38 [-3.61,0.84]
lag 3
¦0.92 [-5.32, 3.47]
7-day mean
Winter afternoon:
0.24 [-1.89,2.38]
-0.72 [-3.87, 2.43]
lag 1
¦1.37 [-5.11, 2.38]
lag 2
¦2.54 [-5.74,0.66]
lag 3
0.21 [-7.67,8.11]
7-day mean
Summer morning:
¦0.80 [-2.74,1.15]
lag 0
0.68[-1.31,2.67]
1.42 [-0.73,3.58]
Iag2
2.54 [0.48, 4.59]
Iag3
1.74 [-2.66, 6.13]
7-day mean
Summer afternoon:
-0.72 [-2.47,1.03]
lag 0
¦0.59 [-2.36,1.18]
lag 1
¦0.33 [-2.11,1.45]
lag 2
0.66 [-1.26, 2.58]
lag 3
0.47 [-3.36,4.29]
7-day mean
J/Vintenrioniint^^
July 2009
E-271
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
wheezing subgroup:
-0.036 [-0.627,0.555]
lag 0
0.142 [-0.573,0.857]
lag 1
0.000 [-0.760,0.759]
lag 2
0.689 [-0.061,1.439]
lag 3
Winter morning in no atopy or recent
wheezing subgroup:
-0.434 [-1.116,0.248]
lag 0
-0.201 [-1.002,0.600]
lag 1
0.154 [-0.703,1.010]
lag 2
-0.605 [-1.422,0.210]
lag 3
Winter morning in subgroup with
parental atopy/recent wheezing:
0.228 [-0.054,0.511]
lag 0
0.476 [0.060, 0.892]
lag 1
0.196 [-0.202,0.594]
lag 2
0.083 [-0.321,0.487]
lag 3
Winter morning in subgroup without
parental atopy/recent wheezing:
-0.482 [-0.952,-0.012]
lag 0
-0.276 [-0.846, 0.294]
lag 1
0.078[-0.520,0.675]
lag 2
-0.298 [-0.864, 0.268]
lag 3
RR Estimate [Lower CI, Upper CI]
lag:
Cough: Winter:
0.92 [0.80,1.07]
lag 0
0.91 [0.77, 1.07]
lag 1
July 2009
E-272
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0.99[0.83,1.17]
lag 2
0.87[0.73,1.03]
lag 3
0.71 [0.52,0.97]
7-day mean
Summer:
1.05[0.97,1.13]
lag 0
1.01 [0.93,1.10]
lag 1
0.95(0.88,1.03]
lag 2
0.89(0.83,0.96]
lag 3
0.81 [0.68, 0.97]
7 day mean
Illness: Winter:
1.05[0.97,1.14]
lag 0
1.11 [1.01,1.22]
lag 1
1.13 [1.01,1.26]
lag 2
1.13 [1.04,1.26]
lag 3
1.13(0.92,1.38]
7-day mean
Summer:
0.97(0.87,1.09]
lag 0
0.98[0.87,1.10]
lag 1
0.95(0.85,1.06]
lag 2
0.94 [0.85,1.05]
lag 3
0.74[0.54,1.03]
7-day mean
Shortness of breath: Winter:
0.99(0.90,1.10]
lag 0
1.01 [0.90,1.13]
lag 1
0.93[0.82,1.05]
July 2009
E-273
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Iag2
0.98[0.86,1.13]
lag 3
0.85[0.67,1.08]
7-day mean
Summer:
1.04 [0.90,1.18]
lag 0
1.12[0.98,1.28]
lag 1
1.04 [0.90,1.20]
lag 2
0.90(0.79,1.03]
lag 3
1.06(0.78,1.43]
7-day mean
Wake at night with cough/wheeze:
Winter:
0.98(0.89,1.08]
lag 0
1.05(0.94,1.16]
lag 1
0.99(0.88,1.12]; lag 2
0.99(0.87,1.12]; lag 3
0.84 (0.67,1.05]
7-day mean
Summer:
0.94 (0.80,1.09]
lag 0
0.86(0.72,1.01]
lag 1
0.94 (0.79,1.11]
lag 2
0.92(0.79,1.07]
lag 3
0.95(0.62,1.47]
7-day mean
Wheeze: Winter:
0.98(0.87,1.10]
lag 0
1.00(0.87,1.14]
lag 1
0.89(0.77,1.03]
lag 2
1.11 (0.95,1.30]
July 2009
E-274
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
lag 3
0.80[0.61,1.07]
7-day mean
Summer:
1.01 [0.87,1.17]
lag 0
0.96(0.83,1.11]
lag 1
0.95(0.82,1.10]
lag 2
0.87(0.75,1.01]
lag 3
1.04 (0.67,1.60]
7-day mean
'All units expressed in //gfm3 unless otherwise specified.
July 2009
E-275
DRAFT-DO NOT CITE OR QUOTE

-------
E.2.2. Respiratory Emergency Department Visits and Hospital
Admissions
Table E-12. Short-term exposure-respiratory-ED/HA-PMio
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Andersen et al. (2008,
189651)
1st page: 458
Period of Study: May 2001 - December
2004
Location: Los Angeles and San Dii
counties, California
Hospital Admissions/ED visits
Outcome (ICD-10):
RD, including chronic bronchitis (J41 -
42), emphysema (J43), other chronic
obstructive pulmonary disease (J44),
asthma (J45), and status asthmaticus
(J46).
Pediatric hospital admissions for
asthma (J45) and status asthmaticus
(J46).
Age Groups Analyzed: > 65 yrs (RD
combined), 5-18 years (asthma)
Study Design: Time series
N: NR
Statistical Analyses: Poisson GAM
Covariates: temperature, dew-point
temperature, long-term trend, seasonality,
influenza, day of the week, public
holidays, school holidays (only for 5 - 18
year olds), pollen (only for pediatric
asthma outcome)
Season: NR
Dose-response Investigated: No
Statistical package: R statistical
software (gam procedure, mgcv package)
Lags Considered: Lag 0 -5 days, 5-day
average (lag 0-4) for RD, and a 6-day
average (lag 0-5) for asthma.
Pollutant: PMio (/L/g|m3)
Averaging Time: 24 h
Mean (SD
median
IQR
99th percentile: 24 (14
21
16-29
72)
Monitoring Stations: 1
Copollutant (correlation): NCtot:
r - 0.39
NC100: r - 0.28
NCa12: r - 0.02
Nca23: r - -0.12
NCa57: r - 0.45
Nca212:r - 0.63
PM2.6: r - 0.80
CO: r - 0.37
NO2: r - 0.35
NOx: r - 0.32
NOx curbside: r - 0.18
O3: r - -0.21
Other variables: Temperature: r - 0.12
Relative humidity: r - 0.05
PM Increment: 13 |Jg|m3 3 (IQR)
Relative risk (RR) Estimate [CI]:
RD hospital admissions (5 day
average, lag 0 -4), age 65+: One-
pollutant model: 1.06 [1.02 - 1.09]
Adj for NCtot: 1.05(1.01 - 1.10]
Adj for NCa212: 1.04 [0.98-1.11]
Asthma hospital admissions (6 day
avg lag 0 - 5), age 5 -18: One-pollutant
model: 1.02(0.93- 1.12]
Adj for NCtot: 1.01 [0.91 - 1.12]
Adj for NCa212: 0.94 [0.81 - 1.09]
Estimates for individual day lags reported
only in figure form (see notes):
Notes: Figure 2: Relative risks and 95%
confidence intervals per IQR in single day
concentration (0-5 day lag).
Summary of Figure 2: RD: Positive,
statistically or marginally significant
associations at Lag 2-5. Asthma: Wide
confidence intervals make interpretation
difficult. Positive associations at Lag 1,
2, 3, and 5.
Reference: Cheng et al. (2007, 093034)
Period of Study: 1996-2004
Location: Kaohsiung, Taiwan
Outcome (ICD-9: 480-486): Pneumonia
Age Groups: NR
Study Design: Case-crossover
N: 82,587 pneumonia hospital admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and humidity
on the same day
Season: NR
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: Cumulative lag period
up to 2 previous days
Pollutant: PM10
Averaging Time: 24 h
Mean (min-max):
77.01 (16.7-232)
Percentiles: 25%: 42.12
50%: 75.27
75%: 104.65
Monitoring Stations: 6
Copollutant: NR
PM Increment: 62.53/yg/m3 (IQR)
OR Estimate [CI]: Single Pollutant Model:
Temp>25°C: 1.21 [1.15,1.28]
Temp< 25°C: 1.57(1.50,1.65]
Two-Pollutant Model: Temp>25°C
Adj. for SO2: 1.21 [1.14,1.28]
Adj. for NO2:1.15 [1.07,1.24]
Adj. for CO: 1.10 [1.03,1.17]
Adj. for 0s: 0.96 [0.89,1.03]
Temp<25°C
Adj. for SO2: 1.56[1.48,1.65]
Adj. for NO2:1.09 [1.02,1.16]
Adj. for CO: 1.30 [1.22,1.39]
Adj. for 0s: 1.56 [1.48,1.65]
July 2009
E-276
DRAFT-DO NOT CITE OR QUOTE

-------
	Reference	
Reference: Chimonas and Gessner
(2007,093261)
Period of Study: January 1,1999-June
30, 2003
Location: Anchorage, Alaska
	Design & Methods	
Outcome (ICD-9): Asthma (493.0-493.9)
Lower respiratory illness-LRI (466.1,
466.0, 480-487, 490,510-511)
Inhaled quick-relief medication
Steroid medication
Age Groups: < 20 years old
Study Design: Time series
N: 42,667 admissions
Statistical Analyses: GEE for
multivariable modeling
Covariates: Season, serial correlation,
year, weekend, temperature,
precipitation, and wind speed
Season: NR
Dose-response Investigated? No
Statistical Package: SPSS (dataset),
SAS (analysis)
Lags Considered: 1 day and 1 week
Concentrations'
Pollutant: PMio
Averaging Time: 24-hs and 1 week
Mean (min-max):
Daily: 27.6(2-421)
Weekly: 25.3 (5.0-116.0)
Monitoring Stations: NR
Copollutant: Daily PM2.6
~	- 0.25 (p< 0.01)
Weekly PM2.6
~	- 0.08 (p - 0.21)
Effect Estimates (95% CI)
PM Increment: 10/yg/m3
RR Estimate [CI]:
Same Day
Outpatient Asthma: 1.006 [1.001,1.013]
Outpatient LRI: 1.001 [0.987,1.015]
Inpatient Asthma: 1.003 [0.922,1.091]
Inpatient LRI: 1.015(0.978,1.053]
Inhaled Steroid Prescriptions: 1.006
[0.996,1.011]
Quick-relief Medication: 1.018
[1.006,1.030]
Weekly (median increase)
Outpatient Asthma: 1.021 [1.004,1.038]
Outpatient LRI: 1.013(0.978,1.049]
Inpatient Asthma: 1.023 [0.948,1.104]
Inpatient LRI: 1.025 [0.981,1.072]
Inhaled Steroid Prescriptions: 0.989
[0.969,1.010]
Quick-relief Medication: 1.057
[1.037,1.077]
Reference: Chiu et al. (2008,191989)
Period of Study: 1996-2001
Location: Taipei, Taiwan
Outcome: Hospital admissions for C0PD
Study Design: Time-series
Covariates: Temperature, humidity,
PMioand O3
Statistical Analysis: Poisson regression
Statistical Package: SAS
Age Groups: All
Pollutant: PM10
Averaging Time: 24h
Mean (SD) Unit:
Index Days: 111.68 ± 38.32//g|m3
Comparison Days: 55.43 ± 24.66//g|m3
Range (Min, Max): NR
Copollutant (correlation): NR
All results refer to "dust storm days" and
can be found in Table 3
Reference: Chiu et al. (2009,190249)
Period of Study: 1996-2004
Location: Taipei, Taiwan
Outcome: Hospital admissions for
pneumonia (ICD-9 480-486)
Study Design: Time-series
Covariates: Weather variables, day of
the week, seasonality, long-term time
trends
Statistical Analysis: Conditional logistic
regression
Statistical Package: SAS
Age Groups: All
Pollutant: PM10
Averaging Time: 24h
Mean Unit: 49.47 //g|m3
Range (Min, Max): 14.42, 234.91
Copollutant (correlation):
SO2: 0.50
NO?: 0.58
CO: 0.34
0s: 0.31
Increment: IQR
Odds Ratio (95% CI)
Temperature > 23° C: 1.11 (1.08-1.14)
Temperature < 23° C: 1.09 (1.07-1.11)
Adjusted for SO2
Temperature a 23° C: 1.10 (1.08-1.13)
Temperature < 23° C: 1.19 (1.17-1.22)
Adjusted for NO2
Temperature a 23° C: 0.90 (0.88-0.93)
Temperature < 23° C: 1.09 (1.07-1.12)
Adjusted for CO
Temperature a 23° C: 1.03 (1.00-1.05)
Temperature < 23° C: 1.07 (1.05-1.10)
Adjusted for O3
Temperature a 23° C: 1.05 (1.03-1.08)
Temperature < 23° C: 1.09 (1.07-1.11)
July 2009
E-277
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Erbas et al. (2005, 073849)
Period of Study: Jan 2000-Dec 2001
Location: Melbourne, Australia
Hospital Admissions
Outcome (ICD-10): Asthma (J45, J46)
Age Groups: M5yrs
Study Design: Time series
N: 8955 asthma cases
Statistical Analyses: GAM, GEE (if
autocorrelation was present in residuals)
Covariates: Temp and humidity
Season: NR
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0,1, 2 days
Pollutant: PMio
Averaging Time: 1 h
Mean (SD):
Western: 2.99 (2.11)
10th percentile: 13.67
90th percentile: 48.00
Inner Melbourne: 4.54 (2.65)
10th percentile: 15.63
90th percentile: 59.73
South/Southeastern: 1.13 (1.18)
10th percentile: 12.00
90th percentile: 36.05
Eastern: 3.61 (2.39)
10th percentile: 16.00
90th percentile: 51.05
Combined: 30.07 (10.55-112.33)
SD - 15.27
10th percentile: 16.00
90th percentile: 50.51
Monitoring Stations: Data obtained
from an air quality simulation model
(TAPM) by CSIR0 Atmospheric Research
Copollutant: NR
PM Increment: Increase from 10th to
90th percentile
RR Estimate [CI]:
Same day lag
Western: NR
Inner Melbourne: 1.17 [1.05,1.31]
South/Southeastern: 1.14 [0.95,1.33]
Eastern: 1.09(1.01,1.18]
Notes: All other lags NR
July 2009
E-278
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Farhat et al. (2005, 089461)
Period of Study: Aug 1996—Aug 1997
Location: Sao Paulo, Brazil
Hospital Admissions and Emergency
Room Visits
Outcome (ICD-9): Lower respiratory
tract diseases (466, 480-519) including
pneumonia or bronchopneumonia (480-
486), asthma (493), bronchiolitis (466)
Age Groups: < 13 yrs
Study Design: Time series
N: 43,635
Statistical Analyses: GAM, Poisson
regression, Pearson correlation
Covariates: Time, temperature, humidity,
weekday
Season: NR
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0-7 days
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max):
62.6 (25.5-186.3)
SD - 26.6
IQr - 30
N - 396
Monitoring Stations: 13
Copollutant (correlation): SO2: r - 0.69
NO2: r - 0.83
O3: r - 0.35
CO: r - 0.72
(all p < 0.05)
Additional correlations:
Rel humidity: r - -0.55
Min temp: r - -0.44
(both p < 0.05)
PM Increment: 30/yg/m3 (IQR)
RR Estimate [CI]:
Lower respiratory tract disease
5-day	moving avg
Copollutant model:
NO?: 2.1 [-7.1,11.3]
SO2: 16.5(10.5,22.6]
O3: 10.1 [5.0,15.2]
CO: 14.1 [8.1,20.2]
Multipollutant model: 5.2 [-4.6,15.1]
Pneumonia or bronchopneumonia
6-day	moving avg
Copollutant model:
NO2: 14.81-3.8,33.4]
SO2: 14.8 [-0.3,30.0]; 0s: 16.2 [1.0,31.3]
CO: 17.6(0.4,34.8]
Multipollutant model: 5.23 [-16.2,26.6]
Asthma or bronchiolitis
2-day moving avg
Copollutant model:
NO2: -11.04 [-50.0,28.0]
SO2: 15.8[-7.8,39.3]
0s: 11.7[-10.4, 33.9]
CO: 12.4 [-14.8,39.7]
Multipollutant model: -15.5 [-61.2,30.2]
July 2009
E-279
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Fung et al. (2006, 089789)
Period of Study: 6/1/95-3/31/99
Location: Vancouver, Canada
Hospital Admission/ED
Outcome: Respiratory diseases (460-
519)
Age Groups: Age > 65
Study Design: Time series
N: 26,275 individuals admitted
Statistical Analyses: Poisson regression
(spline 12 knots), case-crossover
(controls +/7 d days from case date),
Dewanji and Moolgavkar (DM) method
Covariates: Long-term trends, day-of-
the-week effect, weather
Season: All year
Dose-response Investigated? No
Statistical Package: SPIus, R
Lags Considered: 0-7 d
Pollutant: PMio
Averaging Time: 24-h Avg
Mean (SD): 13.31 (6.13)//g/m3
Range (Min, Max): (3.77,52.17)
Monitoring Stations: NR
Copollutant (correlation): PMio:
PM2.6
r - 0.80
PMio 2.5
r - -0.11
CO
r - 0.46
Coh
r - 0.61
0s
r - -0.08
NO?
r - 0.54
SO?
r - 0.61
PM Increment:: 7.9 /yg/m3
Rr Estimate (65+ Years)
Dm Method:
1.014[0.998,1.029]
Lag 0
1.016[0.998,1.034]
3 D Avg
0.988(0.970,1.006]
5 D Avg
0.983(0.963,1.004]
7 D Avg
Time Series:
1.016[0.999,1.033]
Lag 0
1.015(0.996,1.035]
3 D Avg
1.009(0.987,1.032]
5 D Avg
1.009(0.983,1.036]
7 D Avg
Case-Crossover:
1.017(0.998,1.036]
Lag 0
1.015(0.993,1.037]
3 D Avg
1.008(0.984,1.033]
5 D Avg
1.003(0.976,1.031]
7 D Avg
Reference: Fung al. (2005, 093262) Hospital Admissions
Period of Study: Nov 1,1995-Dec 31,
2000
Location: London, Ontario
Outcome (ICD-9): Asthma (493) and all
other respiratory diseases (460-519)
Age Groups: < 65 yrs
65+ yrs
Study Design: Time series
N: 5574 respiratory admissions
Statistical Analyses: GAM with locally
weighted regression smoothers (LOESS)
Covariates: Maximum and minimum
temp, humidity, day of the week,
seasonal cycles, secular trends
Season: NR
Dose-response Investigated? No
Statistical Package: S PIus
Lags Considered: Current to 3-day mean
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max):
38.0 (5-248)
SD - 23.5
Monitoring Stations: 4
Copollutant (correlation): NO2:
r - 0.30
SO2: r - 0.24
CO: r - 0.21
O3: r - 0.53
COH: r - 0.29
PM Increment: 26/yg/m3
% Change in Daily Admission [CI]:
Age < 65
Current day mean: -0.9 [-6.8,5.4]
2-day	mean: -1.3 [-8.5,6.6]
3-day	mean: 1.9 [-6.5,11]
Age 65 +
Current day mean: 3.3 [-1.7,8.6]
2-day	mean: 5 [-1.5,11.9]
3-day	mean: 1.2 [-6.1,9.1 ]
July 2009
E-280
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Galan et al. (2003, 087408)
Period of Study: 1995-1998
Location: Madrid, Spain
Hospital Admissions
Outcome (ICD): Asthma (493)
Age Groups: all ages
Study Design: Time series
l\l: 555,153 at-risk
Statistical Analyses: GAM,
autoregressive Poisson regression
Covariates: temperature, relative
humidity, pollen, year, day of the week,
public holiday
Season: NR
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0,1, 2, 3, and 4-day
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max): 32.1 (11.2-108.6)
SD - 12.1
Monitoring Stations: 13
Copollutant (correlation): SO2:
r - 0.581
NO2: r - 0.717
O3: r - -0.188
Other variables: O.europaea\ r - -0.066
Plantagosp\ r - -0.202
Poaceae: r - -0.132
Urticaceae: r - -0.104
Temp: r - -0.122
Humidity: r - 0.119
PM Increment: 10/yg/m3
RR Estimate [CI]:
Single-pollutant
Current-day lag: 1.011 (0.980-1.042)
1-day	lag: 1.006 (0.976-1.037)
2-day	lag: 1.008 (0.978-1.038)
3-day	lag: 1.039 (1.010-1.068)
4-day	lag: 1.027 (0.999-1.056)
Adjustment for pollen (PM10 3-day lag)
O. europaea\ 1.041 (1.011-1.071)
Plantago sp:. 1.046 (1.017-1.076)
Poaceae: 1.043 (1.015-1.073)
Urticaceae: 1.038 (1.009-1.068)
All four: 1.045(1.016-1.074)
Reference: Hajat et al. (2002, 030358)
Period of Study: 1/1992-12-1994
Location: London, England
Family Practice consultations
Outcome: Upper Resp Disease (excluding
allergic rhinitis) (460-3), (465), (470-5),
(478)
Age Groups: 0-14,
15-64, >65 yrs
Study Design: Time series
N: 268,718-295,740 registered patients
Statistical Analyses: Poisson
regression, GAM, LOESS smoothers,
default convergence criteria
Covariates: long term trends, pollen
counts, flu, meteorological variables
Season: All year
Dose-response Investigated? No
Statistical Package: SPLUS
Lags Considered: 2-3
Pollutant: PM10
Averaging Time: 24-H
Mean (SD): 28.5 (13.7)//g/m3
Percentiles: 10th: 15.8
90th: 46.5
Monitoring Stations: 1
Copollutant: NR
PM Increment: All Year: 18
Warm Season: 15
Cold Season: 20
% Change, Single Pollutant Models:
All Year: Ages 0-14: 2.0[-0.2, 4.2] Lag 3
Ages 15-64: 5.7(2.9, 8.6]
Lag 2
Ages >65:10.2[5.3,15.3] Lag 2
Warm Season: Ages 0-14: 1.1 [-2.4,4.8]
Lag 3
Ages 15-64: 6.0(2.7, 9.4]
Lag 2
Ages >65: 0.1 [-7.7, 8.5] Lag 2
Cold Season: Ages 0-14: 2.7(0.1, 5.5]
Lag 3
Ages 15-64: 3.6(1.0,6.4]
Lag 2
Ages >65: 18.9(11.7, 26.7] Lag 2
% Change, 2 Pollutant Models: 0-14
Yrs
PM10 wINO?: 3.8(1.6, 6.1]
PM10 wfOs: 1.8(-0.4, 3.9]
PM10 wI SO2: 2.0[-0.6,4.6]
15-65 Yrs
PM10 WINO2: 2.8(0.7, 4.9]
PM10 wfOs: 4.8(2.6,7.0]
PM10 wI SO2: 4.8(2.2,7.5]
> 65 Yrs
PM10 w| NO2: 4.6(0.5, 8.8]
PM10 wfOs: 10.7(5.7,16.0]
PM10 wI SO2: 10.6(4.5,17.1]
July 2009
E-281
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Hanigan et al (2008,
156518)
Period of Study: 1996-2005 (April-
November of each year)
Location: Darwin, Australia
Outcome: Cardiorespiratory Disease HA
(ICD 9: 390-519
Pollutant: PMio
Averaging Time: 24h
Mean (SD): 21.2 (8.2)
Range: 55.2
ICD 10:100-99 & J00-99)
Age Groups: NR
Study Design: time series
N: 8279 events
Statistical Analyses: poisson regression Copollutant: NR
Covariates: indigenous status,
Dose-response Investigated? No
Statistical Package: R
Lags Considered: lags 0-3
Monitoring Stations: 2 (monitored &
modeled)
Co-pollutant Correlation
n/a
PM Increment: 10/yg/m3
Percent Change (Lower CI, Upper CI),
lag:
Total Respiratory: 4.81 (-1.04,11.01),
lag 0
Total Resp., Indigenous: 9.40 (1.04,
18.46), lag 0
Total Resp., Non-Indigenous: 3.14 (-2.99,
9.66), lag
Resp. Infection, Indigenous: 15.02 (3.73,
27.54), lag 3
Resp. Infection, Non-Indigenous: 0.67 (¦
7.55,9.61), lag 3
Asthma Indigenous: 16.27 (3.55, 40.17),
lag 1
Asthma Non-Indigenous: 8.54 (-5.60,
24.80), lag 1
figure 3. percent change in hospital
admissions per 10//gfm increase in F
PMio
Reference: Hanigan et al (2008,
156518)
Period of Study: 1996-2005 (April-
November of each year)
Location: Darwin, Australia
Hospital Admissions/ED visits
Outcome (ICD-9 or ICD-10):
Daily emergency hospital admissions for
total respiratory (ICD-9: 460 - 519
ICD-10: J00 - J99), asthma (ICD-9: 493
ICD-10: J45 - J47), C0PD (ICD-9: 490 -
492, 494 - 496
ICD-10: J40-J44, J47.J67), and
respiratory infections (ICD-9: 461 - 466, n,, ¦ ,.
480 -487,514	Other variables:
ICD-10: J00-J22).
Age Groups Analyzed: All
Study Design: Time series
N: 8,279 hospital admissions
Statistical Analyses: Poisson
generalized linear models
Covariates: indigenous status, time in
days, temperature, relative humidity, day
of the week, influenza epidemics, change
between ICD editions, holidays, yearly
population
Season: April - November (corresponding
to the dry season)
Dose-response Investigated? No
Statistical package: R version 2.3.1
Lags Considered: lag 0 -3
Pollutant: PMio
Averaging Time: 24 h
Mean (SD
range): 21.2 (8.2
55.2)
Monitoring Stations: l\l/A (see notes)
Copollutant (correlation): NR
PM Increment: 10/yg/m3
Percent change [95% CI]:
Overall respiratory disease:
LagO: 4.81 [-1.04,11.01]
Lag 0 (indigenous people): 9.40 [1.04,
18.46]
Lag 0 (non-indigenous people): 3.14 [¦
2.99, 9.66]
In unstratified analyses, the subgroups of
respiratory infections, asthma, and C0PD
all had positive associations with PMio
Lag 0.
Asthma:
Lag 1 (indigenous people): 16.27
[-3.55
40.17]
Lag 1 (non-indigenous people): 8.54
[-5.60, 24.80]
Respiratory infections:
Lag 3 (indigenous people): 15.02 [3.73,
27.54]
Lag 3 (non-indigenous people)
0.67
[-7.55,9.61]
Notes:
Figure 3: Associations between
hospitalizations for non-indigenous and
indigenous people with estimated ambient
PMio.
Summary of Figure 3: Confidence
intervals were wide, but indigenous
people generally had stronger
associations with PMio than non-
indigenous people. Daily PMio exposure
levels were estimated for the population
of the city from visibility data using a
previously validated models.
July 2009
E-282
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Hapcioglu et al. (2006,
093263)
Period of Study: Jan 1,1997-Dec 31,
2001
Location: Istanbul, Turkey
Hospital Admissions
Outcome (ICD-9): C0PD (ICD: NR)
Age Groups: NR
Study Design: Time series
N: 1586 patients
Statistical Analyses: Multiple stepwise
regression, Pearson correlation
Covariates: Humidity, temperature, and
pressure
Season: summer, autumn, winter, spring
Dose-response Investigated? No
Statistical Package: SPSS
Lags Considered: NR
Pollutant: PMio
Averaging Time: 1 month
Mean (SD): NR
Monitoring Stations: 1
Copollutant: NR
PM Increment: NR
Correlation with COPD:
r - 0.28
p - 0.03
Adj for temp: r - 0.16
p - 0.23
Notes: RRs only provided for season, not
PM
Reference: Hwang and Chan (2002,
023222)
Period of Study: 1998
Location: Taiwan
Clinic visits
Outcome: LRI
466, 480-486 (acute bronchitis, acute
bronchiolitis, pneumonia)
Age Groups: 0-14yrs, 15-64, 65+ yrs
Study Design: Cluster analysis of small
study areas
N: 50 communities
Statistical Analyses: GLM to model
temporal patterns, hierarchical model to
obtain estimates across 50 communities
Covariates: day of week, temperature,
dew point, summer/Winter
Season: All
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: 0-2
Pollutant: PMio
Averaging Time: 24 H
Mean (SD): 58.9//gfm3 (14.0)
Range (Min, Max): 33.3, 83.1 //g|m3
PM Component:
Monitoring Stations: 59
Notes: Number Of Stations Estimated
From Figure.
Copollutant: NR
PM Increment: 10% Increase In PMio
(5.9/yg/m3)
Percent Change:
0-14
0.5% (-0.1,0.8] LagO
[-0.3, 0.3] Lag1
0.3 [0.0, 0.6] Lag2
15-64
0.6 [0.2, 0.9] LagO
0.2[-0.1,0.5] Lag 1
0.3 [0.0, 0.6] Lag2
65 +
0.8(0.4,1.1] LagO
0.3 [-0.1, 0.6] Lag1
0.5 [0.1, 0.8] Lag2
All Ages
0.5(0.2, N0.8] LagO
[-0.3, 0.3] Lag1
0.3 (0.0, 0.6] Lag2
July 2009
E-283
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Jaffe et al. (2003, 041957)
Period of Study: 7/1/91-6/30/96
Location: Cincinnati, Cleveland,
Columbus, Ohio
ED visits
Outcome (ICD10): Asthma (493)
Age Groups: Age 5-34 years
Study Design: Time-series
N: 4,416 recipients
Statistical Analyses: Poisson
regression, GAM
Covariates: City, day of week, wk, yr,
minimum temperature, dispersion
parameter
Season: June-August only
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: 0-3 days
Pollutant: PMio
Averaging Time: 24-H
Mean (SD): Cincinnati: 43.0(16.4)
Cleveland: 60.8(28.4)
Columbus: 37.4(16.3)
Range (Min, Max): Cincinnati: (16,90)
Cleveland: (12,183)
Columbus: (7,87)
Monitoring Stations: 3
Copollutant (correlation): Cincinnati:
PMio
0s
r - 0.42
NO?
r - 0.36
SO?
r - 0.31
Cleveland: PMio
0s
r - 0.42
NO?
r - 0.34
SO?
r - 0.29
Columbus: PMio
0s
r - 0.51
NO?
r - Na
SO?
r - 0.42
PM Increment: 50/yg/m3
% Change
Asthma
Cincinnati: -22%[-49,-19] Lag 3
Cleveland: 12%[0,27] Lag 2
Columbus: 32%[-6,-85] Lag 3
Ar Estimate [Lower Ci, Upper Ci]
Lag:
Asthma
Cincinnati: PMio: Nr
Cleveland: PMio: 1.32
Columbus: PMio: 3.62
Notes: dose response was investigated
by assessing the relationship between
odds of ed visit by quintile of PMio.
Results are displayed in figure, "no
consistent effects for all three cities were
observed for PMio." Rate ratios were also
reported for each city.
July 2009
E-284
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Jalaludin et al. (2004,	Doctor Visits
056595)
Outcome (ICD- NR): Respiratory
Period of Study: Feb 1 -Dec 31,1994 symptoms (wheeze, dry cough, and wet
cough), asthma medication use, and
Location: Sydney, Australia	doctor visits for asthma
Age Groups: Primary school children
Study Design: Longitudinal cohort study
N: 125 children
Statistical Analyses: GEE logistic
regression models
Covariates: Temperature, humidity, daily
pollen count, daily alternaria count,
number of h spend outdoors, season
Season: Autumn (Feb-Apr), winter (May-
Aug), springfsummer (Sep-Dec)
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-2 days
Pollutant: PMio
PM Increment: IQR (/L/g|m3)
Averaging Time: 24 h
Same day: 12.0
Mean (SD): 22.8(13.8)
1 -day lag: 12.02
Monitoring Stations: 4
2-day lag: 12.25
Copollutant (correlation):
2-day avg: 11.15
0s: r - 0.13
5-day avg: 10.23
NO2: r - 0.26
OR Estimate [CI]:
Other variables:
Doctor Visits for Asthma
Temp: r - 0.04
Same day: 1.11 [1.04,1.19]
Humidity: r - -0.29
1-day lag: 1.10(1.02,1.19]
Total pollen: r - 0.04
2-day lag: 1.15 [1.06,1.24]
Alternaria: r - 0.04
2-day avg: 1.11 [1.03,1.20]
5-day avg: 1.14(0.98,1.31]
Prevalence of Doctor Visits for
Asthma:
Quartile 1: 0.50 (mean PM - 12.4)
Quartile 2: 0.38 (mean PM - 17.2)
Quartile 3: 0.65 (mean PM - 23.0)
Quartile 4: 0.63 (mean PM - 38.3)
Notes: ORs and prevalence are also
provided for wheeze, dry cough, wet
cough, inhaled (32-agonist use, and
inhaled corticosteroid use. None were
statistically significant.
July 2009
E-285
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Johnston et al. (2007,
155882)
Period of Study: 2000, 2004, 2005
(April-November of each year)
Location: Darwin, Australia
Hospital Admissions/ED visits
Outcome (ICD-10):
All respiratory conditions (J00 - J99),
including asthma (J45 - 46), C0PD (J40
- J44), and respiratory infections (J00 -
J22).
Age Groups Analyzed: All
Study Design: Case-crossover
N: 2466 emergency admissions
Statistical Analyses: Conditional
logistic regression
Covariates: weekly influenza rates,
temperature, humidity, days with rainfall
> 5mm, public holidays, school holiday
periods (for respiratory conditions only)
Season: April - November (dry season)
Dose-response Investigated? No
Statistical package: NR
Lags Considered: 0-3 days
Pollutant: PMio
Averaging Time: 24 h
Median (IQR, 10"1 - 90"1 percentile,
range):
17.4 (13.6-22.3
10.3-27.7
1.1-70.0)
Monitoring Stations: 1
Copollutant (correlation): NR
PM Increment: 10/yg/m3
OR Estimate [95% CI]: All respiratory
conditions: Lag 0: 1.08 [0.98 - 1.18]
Lag 0 (indigenous): 1.17 [0.98 - 1.40]
COPD: Lag 0: 1.21 [1.0-1.47]
Lag 0 (indigenous): 1.98 [1.10 - 3.59]
Asthma: Lag 0: 1.14 [0.90 - 1.44]
Asthma + COPD: Lag 0:1.19 [1.03 -
1.38]
Notes: Figure 1: Adjusted OR and 95%
CI for hospital admissions for all
respiratory conditions per 10 //gfm3 rise
in PMio for the same day and lags up to 3
days, overall and stratified by indigenous
status.
Summary of Figure 1 results:
Marginally significant positive association
at Lag 0 in overall study population.
Larger marginally significant positive
association among indigenous people.
Figure 2: OR and 95% CI for hospital
admissions for COPD. Summary of
Figure 2 results: Marginally significant
positive associations at Lag 0 and Lag 1
in overall study population and among
non-indigenous people. Large, statistically
significant positive association at Lag 0
for indigenous people, with smaller, non-
significant positive associations at Lag 1
and Lag2.
Figure 3: OR and 95% CI for hospital
admissions for asthma.
Summary of Figure 3 results: Positive,
non-significant (sometime marginally
significant) associations at Lag 0, Lag 2,
and Lag 3 for overall population and
indigenous status strata.
Figure 4: OR and 95% CI for hospital
admissions for respiratory infections.
Summary of Figure 4 results: Negative
associations at Lag 2 and Lag 3 in all
population strata.
July 2009
E-286
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kim et al. (2007, 092837)
Period of Study: 2002
Location: Seoul, Korea
Ed Visits
Outcome (ICD10): Asthma (J45), (J46)
Age Groups: All Ages
Study Design: Cass-Crossover
N: 92,535 Visits
Statistical Analyses: Conditional
Logistic Regression, Relative Effect
Modification (Rem)
Covariates: Time Trend, Season, Daily
Mean Temperature, Relative Humidity, Air
Pressure. Sep As Modifier Of Air Pollution
Asthma Visit Association.
Season: All Year
Dose-response Investigated? No
Statistical Package: Nr
Lags Considered: 0-2 Days
Pollutant: PMio
Averaging Time: 8-H
Mean (SD): Daily Concentration: 67.6
(39.0) //g/m3
Relevant Exposure Term (Difference
Between Concentration On Event Day
And Mean Of Concentrations On Control
Days): 26.0(19.7)
Percentiles: 50th(Median): Daily
Concentration: 61.9
Relevant Exposure Term: 21.6
Range (Min, Max): Daily Concentration:
(4.9, 302.0)
Relevant Exposure Term: (0.0,143.1)
Monitoring Stations: 3
Copollutant: Nr
PM Increment: 47.4//g/m3
Rr Estimate For Asthma (Stratified By
Sep):
Individual Level Sep:
Quintile 1-1.06[1.02,1.09]
Quintile 2-1.07[1.04,1.10]
Quintile 3-1.06(1.03,1.10]
Quintile 4-1.03(0.99,1.07]
Quintile 5-1.10[1.05,1.14]
Regional Level Sep:
Quintile 1-1.04(0.99,1.10]
Quintile 2-1.03(1.00,1.07]
Quintile 3-1.05(1.03,1.08]
Quintile 4-1.06(1.02,1.10]
Quintile 5-1.09(1.06,1.13]
Total-1.06(1.04,1.08], 3D Ma
Notes: Relative Effect Modification (Rem)
Estimates Presented In Paper.
Reference: Ko et al. (2007, 091639)
Period of Study: 1/2000-12/2004
Location: Hong Kong, China
Ed Visits
Pollutant: PMio
PM Increment: 10 //g/m3
Outcome (ICD-9): C0PD: chronic
bronchitis (491), emphysema (492),
chronic airway obstruction (496)
Averaging Time: 24-H
Mean (SD): 50.1(23.9)//g/m3
Rr Estimate
C0PD:
Age Groups: All Ages
Percentiles: 25th: 31.9
1.003(1.000,1.005]
Study Design: Time Series
50th(Median): 44.5
Lag 0
N: 15 hospitals, 119,225 admissions
75th: 64.1
1.005(1.002,1.007]
Statistical Analyses: Poisson
regression, gam with stringent
convergence criteria, aphea2 protocol.
Covariates: time trend, season,
temperature, humidity, other cyclical
factors, day, day of wk, holidays
Season: All year, interactions with
season tested
Range (Min, Max): (13.6,172.2)
Monitoring Stations: 14 Stations
Copollutant (correlation): PMio:
SO?
r - 0.436
NO?
Lag 1
1.010(1.007,1.012]
Lag 2
1.011(1.008,1.013]
Lag 3
1.008(1.006,1.011]
Dose-response Investigated? No
r - 0.229
Lag 4
Statistical Package: Splus 4.0
0s
1.007(1.004,1.009]
Lags Considered: 0-5 days
r - 0.421
Lag 5

PM2.6
1.005(1.002,1.008]

r - 0.952
Lag 0-1
1.011(1.008,1.014]
Lag 0-2
1.016(1.013,1.019]
Lag 0-3
1.020(1.017,1.024]
Lag 0-4
1.024(1.021,1.028]
Lag 0-5
July 2009
E-287
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ko et al. (2007, 091639)
Period of Study: 1/2000-12/2004
Location: Hong Kong, China
Hospital Admission
Outcome (ICD-9): Asthma (493)
Age Groups: All, 0-14,15-56, 65 +
Study Design: Time series
N: 69,716 admissions, 15 hospitals
Statistical Analyses: Poisson
regression, with GAM with stringent
convergence criteria.
Covariates: Time trend, season,
temperature, humidity, other cyclical
factors
Season: All year, evaluated effect of
season in analysis
Dose-response Investigated? No
Statistical Package: SPLUS 4.0
Lags Considered: 0-5 days
Pollutant: PMio
Averaging Time: 24-h
Mean (SD): 52.5(27.1)//g/m3
Percentiles: 25th: 30.9
50th(Median): 47.1
75th: 68.8
Range (Min, Max):
(13.4,198.9)
Monitoring Stations:
14 stations
Copollutant (correlation): PMio:
SO?
r - 0.436
NO?
r - 0.761
0s
r - 0.600
PM2.6
r - 0.956
PM Increment: 10.0 //g/m3
RR Estimate: Asthma (Single-pollutant
model): 1.006(1.003,1.010]
lag 0
1.005(1.002,1.009]
lag 1
1.005(1.002,1.009]
lag 2
1.008(1.005,1.012]
lag 3
1.006(1.002,1.009]
lag 4
1.006(0.999,1.006]
lag 5
1.008(1.004,1.012]; lag 0-1
1.012(1.008,1.016]
lag 0-2
1.015(1.011,1.019]
lag 0-3
1.018(1.013,1.022]
lag 0-4
1.019(1.015,1.024]
lag 0-5
Asthma by age group
0-14: 1.023(1.015,1.031]
lag 0-5
14-65: 1.014(1.006,1.022]
lag 0-5
>65: 1.015(1.009,1.022]
lag 0-4
Asthma-Effect of season: 1.148(1.051,
1.245] lag 0-5
Reference: Kuo et al. (2002, 036310)
Period of Study: 1 Yr
Location: central Taiwan
Hospital Admissions
Outcome (ICD-NR): Asthma
Age Groups: 13-16 yrs
Study Design: Cohort
N: 12,926
Statistical Analyses: Multiple logistic
regression, Pearson correlation
Covariates: Sex, age, residential area,
level of parents' education, number of
cigarettes smoked by smokers in the
family, incense burning, frequency of
physical activity
Season: NR
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: NR
Pollutant: PMio
Averaging Time: 1 h
Mean (min-max): NR
Range: (54.1-84.3)
Monitoring Stations: 8
Copollutant: Values NR
Notes: Author states that a positive
correlation was found between NO2 and
PMio
PM Increment: NR
OR Estimate:
PMio < 65.9//g/m3—referent
PMio >65.9//g/m3
Crude OR: 0.837
Adj OR: 0.947
95% CI: (0.640,1.401)
July 2009
E-288
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Langley-Turnbaugh et al.
(2005, 093269)
Period of Study: 2000-2001
Location: Portland, Bridgeton, and
Presque Isle, Maine
Hospital Admissions
Outcome (ICD-9): Asthma (493xx)
Age Groups: 0-18 yrs, 19 + yrs
Study Design: Time series
N: NR
Statistical Analyses: NR
Covariates: NR
Season: Winter, spring, summer, fall
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: NR
Notes: Hospital admissions were used to
determine seasonality of asthma
admissions so that PM components from
those time periods could be analyzed
Pollutant: PMio
Averaging Time: NR
Mean (min-max): NR
Monitoring Stations: NR
Copollutant: NR
PM Increment: NR
RR Estimate [CI]: NR
Notes: Portland filters contained more
PM in the winter (Jan) and Bridgeton
filters contained more PM in the spring
(May)
study analyzed metal components of
PMio (Mn, Cu, Pb, As, V, Ni, Al)
Clinical data shows a strong peak in fall
and weaker peaks in Jan and May for
asthma admissions
Reference: Lee et al. (2002, 034826) Hospital Admissions
Period of Study: 12/1/1997-12/31/1999 Outcome (ICD10): Asthma, J45, J46,
Location: Seoul, Korea
Age Groups: Children < 15 years
Study Design: Time-Series
N: 822 d, 6,436 admissions
Statistical Analyses: Poisson
regression, GAM, LOESS smoothers.
Covariates: Days of the week,
temperature, humidity
Season: All
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-5, 0-1 moving
averages for 1-2, 2-3, and 3-4 days
Pollutant: PMio
Averaging Time: 24-h
Mean (SD): 64.0 (31.8)/yg/m3
Percentiles: 25th: 40.5 /yg/m3
50th(Median): 59.1 /yg/m3
75th: 80.9/yg/m3
Range (Min, Max): NR
Monitoring Stations: 27
Notes: Copollutant (correlation): PMio-
SO?: 0.585
PM 10-NO2: 0.738
PMio-Os: 0.106
PM10-CO: 0.598
PM Increment: IQR: 40.4/yg/m3
RR Estimate:
Single Pollutant:
1.07 (1.04,1.11) lag 1
Two pollutant models:
+SO2: 1.05 (1.01,1.09) lag 1
+NO2: 1.03 (0.99,1.07) lag 1
+O3: 1.06 (1.03,1.10) lag 1
+C0: 1.04 (1.00,1.08) lag 1
Three pollutant models:
+O3 + CO: 1.02 (0.98,1.06), lag 1
Four pollutant models:
+0s + CO +SO2:1.02 (0.98,1.06), lag 1
Five pollutant model:
1.016(0.975,1.059) lag 1
Notes: Investigated the association
between outdoor air pollution and asthma
attacks in children < 15 yrs.
July 2009
E-289
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lee et al. (2006, 090176)
Period of Study: 1/1997-12/2002
Location: Hong Kong, China
Hospital Admission
Outcome: Asthma (493)
Age Groups: < 18 years
Study Design: Time series
N: 26,663 asthma admissions for asthma
and 5821 admissions for influenza
Statistical Analyses: Poisson
regression, GAM
Covariates: Temperature, atmospheric
pressure, relative humidity
Season: All
Dose-response Investigated? No
Statistical Package: SAS 8.02
Lags Considered: 0-5
Notes: Controls were admissions for
influenza ICD9 487
Pollutant: PMio
Averaging Time: 24-hs
Mean (SD): 56.1 (24.2)
Percentiles: 25th: 37.3
50th(Median): 51.1
75th: 70.7
Monitoring Stations: 10
Notes: Copollutant (correlation): PMio-
PM2.6: 0.90
PM10-SO2: 0.39
PM10-NO2: 0.80
PM10-O3: 0.60
PM Increment: IQr - 33.4
Percent Increase:
Single pollutant model:
4.97 [2.96, 7.03], lag 0
5.71 [3.78,7.68], lag 1
6.40 [4.51,8.32], lag 2
7.25 [5.38, 9.16], lag 3
7.45 [5.58, 9.35], lag 4
5.96 [4.11,7.85], lag 5
Multipollutant model (SO2, CO, NO2, O3)
3.67(1.52,5.86] Iag4
Reference: Lin et al. (2005, 087828)
Period of Study: 1998-2001
Location: Toronto, North York, East
York, Etobicoke, Scarborough, and York
(Canada)
Hospital Admissions
Outcome (ICD-9): Respiratory infections
including laryngitis, tracheitis, bronchitis,
bronchiolitis, pneumonia, and influenza
(464,466,480-487)
Age Groups: 0-14 yrs
Study Design: Bidirectional case-
crossover
N: 6782 respiratory infection
hospitalizations
Statistical Analyses: Conditional
logistic regression (Cox proportional
hazards model)
Covariates: Daily mean temp and dew
point temp
Season: NR
Dose-response Investigated? No
Statistical Package: SAS 8.2 PHREG
procedure
Lags Considered: 1-7 day averages
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max):
20.41 (4.00-73.00)
SD - 10.14
Monitoring Stations: 4
Copollutant (correlation): PM2.6:
r - 0.87
PMio2.5: r - 0.76
CO: r - 0.10
SO2: r - 0.48
NO2: r - 0.54
O3: r - 0.54
PM Increment: 12.5/yg/m3
OR Estimate [CI]:
Adjusted for weather
4 day avg: 1.22(1.10,1.34]
6 day avg: 1.25(1.11,1.40]
Adj for weather and other gaseous
pollutants
4 day avg: 1.14 [0.99,1.32]
6 day avg: 1.20 [1.01,1.42]
Notes: OR's were also categorized into
"Boys" and "Girls," yielding similar results
Reference: Lin et al, (2008,126812)
Period of Study: 1991-2001
Location: New York State, US
Outcome: Respiratory hospital
admissions (ICD-9 466, 490-493, 496
Study Design: Time-series
Covariates: Demographic
characteristics, PMio, meteorological
conditions, day of the week, seasonality,
long term trends and different lag periods
Statistical Analysis: GAM and case-
crossover design at the regional level and
Bayesian hierarchical model at the state
level
Pollutant: O3 (PMio is secondary)
Averaging Time: 24h
Mean (SD) Unit: 19.56 (10.92)//g/m3
Range (Min, Max): 1.0, 90.00
Copollutant (correlation): Given in
Figure 3
All PMio results are given in Figure 3
Age Groups: Children 0-17 years
July 2009
E-290
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lin et al. (2002, 026067)
Period of Study: Jan 1,1981 -Dec 31,
1993
Location: Toronto
Hospital Admissions
Outcome (ICD-9): Asthma (493)
Age Groups: 6-12 yrs
Study Design: Uni- and bi directional
case-crossover (UCC, BCC) and time-
series (TS)
N: 7,319 asthma admissions
Statistical Analyses: Conditional
logistic regression, GAM
Covariates: Maximum and minimum
temp, avg relative humidity
Season: Apr-Sep, Oct-Mar
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1-7 day averages
Pollutant: PMio
Averaging Time: 6 days (predicted daily
values)
Mean (min-max):
30.16(3.03-116.20)
SD - 13.61
Monitoring Stations: 1
Copollutant (correlation): PM2.6:
r - 0.87
PMio2.5: r - 0.83
CO: r - 0.38
SO?: r - 0.44
NO2: r - 0.52
O3: r - 0.44
PM Increment: 14.8/yg/m3
RR Estimate [CI]:
Adj for weather and gaseous pollutants
BCC 5 day avg: 0.99 [0.90,1.09]
BCC 6 day avg: 1.01 [0.90,1.12]
TS 5 day avg: 1.03(0.95,1.11]
TS 6 day avg: 1.02(0.94,1.11]
Boys-adj for weather
UCC 1 day avg: 1.10(1.04,1.17]
UCC 2 day avg: 1.10 [1.02,1.17]
BCC 1 day avg: 1.04 [0.98,1.09]
BCC 2 day avg: 1.01 [0.95,1.08]
TS 1 day avg: 1.03 [0.99,1.07]
TS 2 day avg: 1.01 [0.96,1.05]
Girls-adj for weather
UCC 1 day avg: 1.07 [0.99,1.16]
UCC 2 day avg: 1.15 [1.04,1.26]
BCC 1 day avg: 0.99 [0.92,1.06]
BCC 2 day avg: 1.03 [0.95,1.12]
TS 1 day avg: 0.99 [0.94,1.04]
TS 2 day avg: 1.02(0.96,1.08]
Notes: The author also provides RR using
UCC, BCC, and TS analysis for female
and male groups for days 3-7, yielding
similar results
Reference: Linares et al. (2006, 092846) Outcome: Respiratory system diseases
460-519, bronchitis 460-496, pneumonia
Period of Study: Jan 1995-Dec 2000
Location: Madrid, Spain
480-487
Age Groups: < 10 years
Study Design: Time series
N: — 15,000 admissions, 2192 days
Statistical Analyses: Poisson
regression, dummy variables to adjust for SO2: 0-532
season and weather
Pollutant: PMio
Averaging Time: 24-hs
Mean (SD): 33.4//gfm3, (13.7)
Range (Min, Max): 6,109//gfm:l
Monitoring Stations: 24
Notes: Copollutant (correlation): PMio-
Covariates: Temperature, difference in
barometric pressure, relative humidity,
pollen counts, influenza epidemics
Season: All
Dose-response Investigated? Yes
Statistical Package: S-Plus 2000
Lags Considered: 0-13
PMio-Os:-0.289
PMio-NOx: 0.721
PM10-PJO2: 0.711
PM Increment: 10/yg/m3
RR Estimate
Bronchitis
1.09 [1.01,1.16] lag 2
AR% Estimate
Bronchitis
7.9 [CI NR] Iag2
Notes: Only statistically significant
relative and attributable risks were
presented by the authors.
The authors conducted multivariate
modeling using a linear term to represent
PMio. They also report an apparent
estimated PMio effect threshold of
60 /yglm3, based on examination of a
scatter plot of respiratory emergency
hospital admissions and PMio levels.
July 2009
E-291
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Luginaah, et al. (2005,
057327)
Period of Study: Apr 1995-Dec 2000
Location: Windsor, Ontario, Canada
Hospital Admission/ED:
admission
Outcome: All respiratory: 460-519
Age Groups: All, 0-14,15-64, and >65
Study Design: Times-series, bi-
directional case-crossover
N: 4214 admissions
Statistical Analyses: Poisson
regression, GAM wf stringent
convergence criteria or natural splines,
conditional logistic regression
Covariates: Age, sex
Maximum & minimum temperature,
change in barometric pressure from
previous day
Season: All
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 1-3
Pollutant: PMio
Averaging Time: 24 h maximum
Mean (SD): 50.6 ,(35.5)
Range (Min, Max):
9, 349
Monitoring Stations:
4
Notes: Copollutant (correlation): PMio-
NO?: 0.33
PM10-SO2: 0.22
PM10-CO: 0.21
PMio-Os: 0.33
PM Increment: Interquartile range (75th-
25th) 31 /yglm3
RR Estimates (Time Series)
All Age Groups Females
0.996[0.950,1.044], lag 1
1.015[0.963,1.069], lag 2
1.022(0.968,1.078], lag 3
All Age Groups Males
1.008[0.965,1.054], lag 1
1.036(0.986,1.089], lag 2
1.027(0.974,1.083], lag 3
RR Estimates (Case Crossover)
All Age Groups Females
1.034(0.974,1.098], lag 1
1.045(0.972,1.124], lag 2
1.054(0.970,1.145], lag 3
All Age Groups Males
0.997(0.942,1.056], lag 1
1.022(0.953,1.097], lag 2
1.008(0.930,1.092], lag 3
Notes: Results, stratified by age group
available in manuscript.
Reference: Martins et al. (2002,
035059)
Period of Study: May 1996-Sep 1998
Location: Sao Paulo, Brazil
Hospital Admission/ED:
ER visits
Outcome (ICD10): Chronic lower
respiratory disease (CLRD) (40-47)
includes chronic bronchitis, emphysema,
other COPDs, asthma, bronchiectasia
Age Groups: > 64 years
Study Design: Time series
N: 712 for CLRD
1 hospital
Statistical Analyses: Poisson regression
GAM, LOESS smoothers, no mention of
stringent criteria
Covariates: Day of week, time minimum
temperature, relative humidity
Season: All
Statistical Package: S-Plus
Lags Considered: 2-7 3 d ma
Pollutant: PMio
Averaging Time: daily
Mean (SD): 60.0//gfm3' (26.3)
Range (Min, Max):
22.8. 186.5/yg/m3
Unit (i.e./yg/m3):
/yg/m3
PM Component: None
Monitoring Stations:
12
Notes: Copollutant (correlation): PMio-
CO: 0.73
PMio- NO?: 0.83
PMio-SO?: 0.72
PMio-Os: 0.35
PM Increment: 1 /yglm3
Regression Coefficients (SE):
0.0024 (0.0023), 6 d ma
Notes: % Increase (SD) for ER visits per
2435/yglm3 (IQR) PMio (lag 6 d ma)
presented graphically in text.
July 2009
E-292
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Masjedi et al. (2003,
052100)
Period of Study: Sep 1997—Feb 1998
Location: Tehran, Iran
Hospital Admissions
Outcome (ICD-9): Acute asthma and
C0PD exacerbations (ICD: NR)
Age Groups: NR
Study Design: Time series
N: 355 patients
Statistical Analyses: Multiple stepwise
regression, autoregression method (time
series), Pearson correlation
Covariates: NR
Season: NR
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 3, 7, and 10 day mean
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max):
108.41 (14.5-506.60)
SD - 59.55
Monitoring Stations: 3
Copollutant: NR
PM Increment: NR
Results:
Time-series analysis
Asthma: p - 0.002
p - 0.32
C0PD: (3 - 0.004
p - 0.02
Total Acute Resp Conditions: p - 0.006
p - 0.27
Correlation of 3-day mean
Asthma: r - -0.21
P - -0.16
p - 0.08
Correlation of weekly mean
Asthma: r - -0.27
P - -0.008
p - 0.12
Correlation of 10-day mean
Asthma: r - -0.38
P - -0.066
p - 0.089
Reference: McGowan et al. (2002,
030325)
Period of Study: Jun 1988—Dec 1998
Location: Christchurch, New Zealand
Hospital Admissions
Pollutant: PMio
Outcome (ICD-9): Pneumonia (480-487), Averaging Time: 24 h
acute respiratory infections (460-466),
chronic lung diseases (491-492, 494-
496), asthma (493)
Age Groups: < 15 yrs, 15-64, 65 +
Study Design: Time series
N: 20,938 admissions
Statistical Analyses: GAM with log link,
Linear Regression Model
Covariates: Wind speed, relative
humidity, temperature
Season: NR
Dose-response Investigated? No
Statistical Package: S-PLUS
Lags Considered: 0-6 days
Mean (min-max):
25.17 (0-283)
SD - 25.49
Monitoring Stations: 1
Copollutant: NR
PM Increment: 14.8 /yglm3 (IQR)
% Increase [CI]:
Respiratory Admissions (2-day lag)
0-14	yrs: 3.62 [2.34,4.90]
15-64 yrs: 3.39(1.85,4.93]
65+yrs: 2.86 [1.23,4.49]
All ages: 3.37 [2.34,4.40]
Overall
Acute respiratory infections: 4.53
[2.82,6.24]
Pneumonia/influenza: 5.32 [3.46,7.18]
Chronic lung diseases: 3.95 [2.15,5.75]
Asthma: 1.86 [0.48,.3.24]
Total Respiratory Admissions
Same day lag: 2.52 [1.49,3.55]
1-day	lag: 2.56(1.53,3.59]
2-day	lag: 3.37 [2.34,4.40]
3-day	lag: 3.09 [2.06,4.12]
4-day	lag: 3.13(2.10,4.16]
5-day	lag: 3.21 [2.18,4.24];
6-day	lag: 3.09 [2.06,4.12]
July 2009
E-293
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Medina-Ramon et al (2006,
087721)
Period of Study: 1986-99
Location: 36 US Cities
Outcome: 490-496, except 493 (COPD),
480-487 (Pneumonia)
Age Groups: 65 + (US Medicare
beneficiaries)
Study Design: Case crossover
N: 578,006 COPD admissions
1,384,813 Pneumonia admissions
Statistical Analyses: Conditional
logistic regression, Meta-analysis using
REML random effects models
Covariates: Mean and variance of daily
summer apparent temperature index, %
65+ living in poverty,% households with
central air-conditioning mortality rate for
emphysema among 65+(surrogate for
smoking history), % PMio from traffic
Season: WarmfMay -Sepnd ColdfOct-
Apr)
Dose-response Investigated? No
Statistical Package: SAS
STATA
Lags Considered: 0-1 days
Pollutant: PMio
Averaging Time: 24 h avg
Mean (SD): 30.4//gfm3 (5.1)
Monitoring Stations: at least one per
city
Notes: PMio measurements made every
2, 3 or 6 days depending on the city.
Copollutant: NR
PM Increment: 10/yg/m3
% change [Lower CI, Upper CI]
lag:
COPD warm season
0.81 (0.22,1.41) at lag 0
1.47(0.93,2.01) at lag 1
COPD cold season
0.06(-0.40,0.51) at lag 0
0.10(-0.30,0.49) at lag 1
Pneumonia warm season
0.84 (0.50,1.19) at lag 0
0.79(0.45,1.13) at lag 1
Pneumonia cold season
0.30(0.07,0.53) at lag 0
0.14 (-0.17,0.45) at lag 1
Reference: Meng et al., (2007, 093275)
Period of Study: Nov 2000-Sep 2001
Location: Los Angeles and San Dii
counties, California
Outcome (ICD-NR): Poorly controlled
asthma defined as (1) daily or weekly
asthma symptoms or (2) at least 1 ED
visit or hospitalization due to asthma over
the past 12 months
Age Groups: > 18 yrs
Study Design: Time series
N: 1609 asthma patients
Statistical Analyses: Logistic regression
Covariates: Age, sex, race/ethnicity,
poverty level, insurance status, smoking
behavior, employment, asthma medication
use, and county
Season: NR
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: NR
Pollutant: PMio
Averaging Time: 24 h
Mean (25-75th percentile): NR
Monitoring Stations: NR
Copollutant (correlation): PM2.6:
r - 0.84
0s: r - -0.72
NO2: r - 0.83
CO: r - 0.42
Other variables:
Traffic: r - 0.14
PM Increment: 10/yg/m3
OR Estimate [CI]:
All Adults: 1.08 [0.82,1.43]
18-64 yrs: 1.14(0.84,1.55]
65+: 0.84 [0.41,1.73]
Men: 0.72(0.42,1.21]
Women: 1.38(0.99,1.94]
Exposure above 44.01 //g|m3 (annual
concentration)
All Adults: 1.56 [0.96,2.52]
18-64 yrs: 1.40 [0.81,2.41]
65+: 2.23 [0.60,8.27]
Men: 0.80 [0.27,2.41]
Women: 2.06 [1.17,3.61]
Notes: This study focused more on the
relation between poorly controlled asthma
and traffic density.
July 2009
E-294
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Middleton et al. (2008,
156760)
Period of Study: 1995 - 1998, 2000 -
2004
Location: Nicosia, Cyprus
Hospital Admissions/ED visits
Outcome (ICD-NR):
Hospital admissions for all respiratory
disease (ICD-10: J00 - J99).
Age Groups Analyzed: All, also
stratified by age (<15 vs. >15 years)
Study Design: Time series
N: Statistical Analyses:
additive Poisson models
Covariates: seasonality, day of the
week, long- and short-term trend,
temperature, relative humidity
Season: NR
Dose-response Investigated: No
Statistical package: STATA SE 9.0, and
the MGCV package in the R software (R
2.2.0)
Lags Considered: lag 0 -2 days
Pollutant: PMio
Averaging Time: 24 h
Mean (SD
median
5% - 95%
range):
Cold: 57.6 (52.5
50.8
20.0 - 103.0
5.0 - 1370.6)
Warm: 53.4 (50.5
30.7
32.0 - 77.6
18.4- 933.5)
Monitoring Stations: 2
Copollutant (correlation): NR
Other variables:
PM Increment: 10 /yglm3, and across
quartiles of increasing levels of PMio
Percentage increase estimate [CI]: All
age/sex groups (Lag 0): All admissions:
0.85 (0.55,1.15)
Respiratory (all): 0.10 (-0.91,1.11)
Respiratory (cold months): -0.33 (-1.47,
0.82)
Respiratory (warm months): 1.42 (-0.42,
3.31)
CVD +RD: 0.56 (-0.21,1.34)
Nicosia residents (Lag 0): Respiratory
(all): 0.25(-0.84,1.36)
Respiratory (cold months): -0.22 (-1.45,
1.02)
Respiratory (warm months): 1.80 (-0.22,
3.85)
CVD +RD: 0.38(-0.47,1.23)
Males (Lag 0): All admissions: 0.96
(0.54,1.39)
Respiratory (all): -0.06 (-1.37,1.26)
Respiratory (cold months): -0.16 (-1.76,
1.46)
Respiratory (warm months): 1.10
(-1.47,3.74)
CVD +RD: 0.63(-0.34,1.62)
Females (Lag 0): All admissions: 0.74
(0.31,1.18)
Respiratory (all): 0.39 (-1.21, 2.02)
Respiratory (cold months): -0.26 (-2.18,
1.70)
Respiratory (warm months): 3.27 (-0.00,
6.65)
CVD +RD: 0.59 (-0.68,1.87)
Aged < 15 years (Lag 0): All
admissions: 0.47 (-0.13,1.08)
Respiratory (all): -0.35 (-1.77,1.08)
Respiratory (cold months): -0.31 (-2.02,
1.42)
Respiratory (warm months): -0.59 (-3.53,
2.45)
Aged > 15 years (Lag 0): All
admissions: 0.98 (0.63,1.33)
Respiratory (all): 0.59 (-0.87, 2.07)
Respiratory (cold months): 0.02 (-1.76,
1.83)
Respiratory (warm months): 3.89 (1.05,
6.80)
July 2009
E-295
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Moore et al, 2008
Period of Study: 1983-2000
Location: California's South Coast Air
Basin
Outcome: Hospital admissions for
asthma (ICD-9 493)
Study Design: Time-series
Covariates: Income, demographic and
residential variables
Statistical Analysis: HRMSM
Age Groups: Children ages 0-19 years
Pollutant: 03 (PM10 secondary)
Averaging Time: Quarterly
Mean (SD) Unit: NR
Range (Min, Max): NR
Copollutant (correlation):
1 hr 0a: 0.52
8hr O3: 0.46
24hr NO?: 0.53
24hr CO: 0.36
24hr SO2: 0.13
Results given are for O3
Reference: Nascimento et al. (2006,
093247)
Period of Study: May 1, 2000-Dec 31,
2001
Location: Sao Jose dos Campos, Brazil
Hospital Admissions
Outcome (ICD-10): Pneumonia (J12-J18)
Age Groups: 0-10 yrs
Study Design: Time series
N: 1265 admissions
Statistical Analyses: GAM, Poisson
regression
Covariates: Temperature, humidity
Season: NR
Dose-response Investigated? Yes
Statistical Package: S-Plus, SPSS
Lags Considered: 0-7 days
Pollutant: PM10
Averaging Time: 24 h
Mean (min-max):
40.2(3.4-196.6)
SD - 26.9
Monitoring Stations: 2
Copollutant (correlation): SO2: r -
O3: r - 0.09
Other variables:
Admissions: r - 0.21
Temp: r - -0.14
Notes: All p < 0.05
PM Increment: 24.7 /yglm3
Regression coefficients (SE):
Same day: -0.00053 (0.00125)
1-day	lag: 0.00029 (0.00057)
2-day	lag: 0.00089 (0.00069)
3-day	lag: 0.00122 (0.00053)*
0.30 4-day lag: 0.00126 (0.00055)*
5-day	lag: 0.00098 (0.00071)
6-day	lag: 0.00035 (0.00056)
7-day	lag:-0.00067 (0.00123)
*p< 0.05
Notes: Percent increase over all lag days
is displayed in Fig 2
Reference: Neuberger et al. (2004,
093249)
Period of Study: 1999-2000 (1 yr
period)
Location: Vienna and Lower Austria
Hospital Admissions
Outcome (ICD-9): Bronchitis,
emphysema, asthma, bronchiectasis,
extrinsic allergic alveolitis, and chronic
airway obstruction (490-496)
Age Groups: 3.0-5.9 yrs
7-10 yrs
65 +
Study Design: Time series
N: 366 days (admissions NR)
Statistical Analyses: GAM
Covariates: SO2, NO, NO2, O3,
temperature, humidity, and day of the
week
Season: NR
Dose-response Investigated? Yes
Statistical Package: S-Plus 2000
Lags Considered: 0-14 days
Pollutant: PM10
Averaging Time: 24 h
Maximum daily mean:
Vienna: 105
Rural area: NR
Monitoring Stations: NR
Copollutant: NR
PM Increment: 10/yg/m3
Log Relative Rate Estimate (p-value):
Vienna
Male: 2 day lag - 4.217 (0.030)
Association with tidal lung function:
P - -1.067 (p-value - 0.241)
Notes: Effect parameters with significant
coefficients for respiratory health
included: male sex, allergy, asthma in
family, and traffic for Vienna and age,
allergy, asthma in family, and passive
smoking for the rural area. Effect
parameters with significant coefficients
for log asthma score were allergy,
asthma in family, and rain for Vienna and
allergy, asthma in family, and passive
smoking for the rural area.
July 2009
E-296
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Oftedal et al. (2003,
055623)
Period of Study: 1995-2000
Location: Drammen, Norway
Hospital Admissions
Outcome: All Respiratory (460-517)
Age Groups: All
Study Design: Time-series
N: —4,458 admissions
Statistical Analyses: Poisson
regression, GAM wf stringent
convergence criteria
Covariates: Temperature, humidity,
influenza epidemics, summer and
Christmas vacation
Season: All
Dose-response Investigated? Yes
Statistical Package: S-Plus
Lags Considered: 2-3
Pollutant: PMio
Averaging Time: 24-hs
Mean (SD): 16.8//gfm3, (10.2)1994-
1997
16.5/yg|m3, (10.3) 1998-2000
16.6 ,/yg|m3 (10.2) total period
PM Component: Benzene, formaldehyde,
toluene
Monitoring Stations:
NR
Notes: Copollutant (correlation):
Correlation between pollutants ranged
from -0.47-0.78 with the exception of
the VOCs studied
Notes: Benzene, formaldehyde and
toluene also evaluated
PM Increment: IQr - 11.04
RR Estimate
1.035 [0.990,1.083] 1994-1997
0.992 [0.948,1.037] 1998-2000
1.021 [0.990,1.053] 1994-2000
2 Pollutant Model
PMio w/benzene: 1.01 (0.978,1.043)
Reference: Peel et al. (2005, 056305)
Period of Study: Jan 1993-Aug 2000
Location: Atlanta, Georgia
ED visits
Outcome: Asthma (493, 786.09)
C0PD (491,492, 496)
URI (460-466, 477)
Pneumonia (480-486)
Age Groups: All ages. Secondary
analyses conducted by age group: 0-1, 2-
18, >18
Study Design: Time series
N: 31 hospitals
Statistical Analyses: Poisson GEE for
URI, asthma and all RD
Poisson GLM for pneumonia and C0PD)
Covariates: Avg temperature and dew
point, pollen counts
Season: All (secondary analyses of warm
season)
Dose-response Investigated? Yes
Statistical Package: SAS 8.3, S-Plus
2000
Lags Considered: 0-7 d , 3 d ma, 0-13 d
unconstrained distributed lag
Pollutant: PMio
Averaging Time: 24 h avg
Mean (SD): 27.9 (12.3)//g/m3
Percentiles: 10th: 13.2
90th: 44.7
Monitoring Stations:
"Several"
Copollutant (correlation): 8 h 0:i:
r - 0.59
1 h NO?: r - 0.49
1 h CO: r - 0.47
1 h SO?: r - 0.20
24-h PM2.6: 0.84
24 h PMio2.5: r - 0.59
24 h UF: r - -0.13
Components: r ranged from 0.42-0.74
PM Increment: PMio: 10/yg/m3
RR Estimate [Lower CI, Upper CI]
All Respiratory Outcomes:
1.013(1.004-1.021), 3 d ma
URI:
1.014 (1.004-1.025), 3 d ma
1.073 (1.048-1.099), 14-day dist. lag
Asthma:
1.009(0.996-1.022), 3 d ma
1.099(1.065-1.135), 14-day dist. lag:
Pediatric Asthma 2-18yrs):
1.016(0.998 -1.034)
Pneumonia:
1.011 (0.996-1.027), 3 d ma
1.087(1.044-1.132), 14-day dist. lag
C0PD:
1.018(0.994-1.043), 3 d ma
1.092(1.023-1.165), 14-day dist. lag
Notes: RRs obtained using AOS 1993-
2000, AOS 1998-2000 and ARIES data
compared. Infant (0-1 y) and pediatric (2-
18 y) asthma was associated more
strongly with PMio, PM2.6 and 0C than
adult asthma.
July 2009
E-297
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ren et al. (2006, 092824)
Hospital Admissions
Pollutant: PMio
PM Increment: NR
Period of Study: Jan 1,1996-Dec 31,
2001
Location: Brisbane, Australia
Outcome (ICD-9): Respiratory diseases
(460-519) excluding influenza (487.0-
487.8)
Age Groups: NR
Averaging Time: 24 h
Mean (min-max):
15.84 (2.5-60)
Coefficient Estimates:
Respiratory Hospital Admissions
Same day: -0.004296

Study Design: Time series
Monitoring Stations: 1
1-day lag:-0.002474

N: NR
Copollutant: NR
2-day lag:-0.004229

Statistical Analyses: GAM

"all statistically significant

Covariates: Day of week, relative
humidity, influenza outbreaks

Respiratory Emergency Visits
Same day: -0.000887

Season: NR

1-day lag: -0.004209

Dose-response Investigated? Yes

2-day lag: -0.003440

Statistical Package: S-Plus
Lags Considered: 0,1, and 2 days

Notes: Relative risks were provided in
graphical form (Fig 3)
Reference: (Sauerzapf et al„ 2009,
180082)
Period of Study: 3/1/2006-3/2/2007
Location: Norfolk, UK
Outcome: C0PD
Study Design: Case-Crossover
Covariates: Environmental factors and
Influenza
Statistical Analysis: Logistic regression
Statistical Package: SPSS 14
Age Groups: > 18 years
N: 1050 adult C0PD admissions
Pollutant: PMio
Averaging Time: 24h
Mean (SD) Unit:
Control: 19.87 (8.51)//g/m3
Case: 20.47 (9.27)/yg/m3
Range (Min, Max):
Control: 9.77-34.27
Case: 10.04-35.03
Copollutant (correlation): I
Increment: 10/yg/m3
Odds Ratio (95% CI)
Lag 0-7, unadjusted: 1.079 (0.980-1.188)
Lag 0-8, adjusted: 1.101 (0.988-1.226)
Lag 1-8, unadjusted: 1.056 (0.961-1.161)
Lag 1-8, adjusted: 1.054 (0.949-1.170)
Reference: Sinclair and Tolsma (2004,
088696)
Period of Study: 25 Months
Location: Atlanta, Georgia
Outpatient Visits
Outcome: Asthma (493)
URI (460, 461, 462,463, 464, 465, 466,
477)
LRI (466.1, 480, 481,482, 483, 484,
485, 486).
Age Groups: < - 18 y, 18 + y (asthma)
All ages (URI//LRI)
Study Design: Times series
N: 25 months
260,000 to 275,000 health plan
members (August 1998-August 2000)
Statistical Analyses: Poisson GLM
Covariates: Season, Day of week,
Federal Holidays, Study Months
Season: NR
Dose-response Investigated?: No
Statistical Package: SAS
Lags Considered: Three 3 d moving
averages (0-2, 2-5, 6-8)
Pollutant: PMio
Averaging Time: 24 h avg
Mean (SD): PMio mass-29.03/yg/m3
(11.61)
Monitoring Stations:
1
Notes: Copollutant: NR
PM Increment: 11.61 (1 SD)
RR Estimate [Lower CI, Upper CI]
lag:
Child Asthma: 1.049 (S), lag 3-5 d
LRI: 1.074 (S), 3-5 d lag
Notes: Numerical findings for significant
results only presented in manuscript.
Results for all lags presented graphically
for each outcome (asthma, URI, and LRI).
July 2009
E-298
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Slaughter et al. (2005,
073854)
Period of Study: January 1995 through
June 2001
Location: Spokane, WA
Hospital Admissions and ED visits
Outcome: All respiratory (460-519)
Asthma (493)
C0PD (491,492,494,496)
Pneumonia (480-487)
Acute URI not including colds and
sinusitis (464, 466, 490)
Age Groups: All, 15+ years for C0PD
Study Design: Time series
N: 2373 visit records
Statistical Analyses: Poisson
regression, GLM with natural splines. For
comparison also used GAM with
smoothing splines and default
convergence criteria.
Covariates: Season, temperature,
relative humidity, day of week
Season: All
Dose-response Investigated?: No
Statistical Package: SAS, SPLUS
Lags Considered: 1 -3 d
Pollutant: PMio
Averaging Time: 24 h avg
Range (90% of concentrations): 7.9-
41.9/yg/m3
Monitoring Stations:
1
Notes: Copollutant (correlation): PMio
PMir - 0.50
PM2.Br - 0.62
PMio 2.5 r - 0.94
COr - 0.32
Temperature r - 0.11
PM Increment: 25 /yglm3
RR Estimate [Lower CI, Upper CI]
lag:
ER visits -- PMio
All Respiratory
Lag 1: 1.01 [0.99,1.04]
Lag 2: 1.01 [0.98,1.03]
Lag 3: 1.02 [0.99,1.04]
Acute Asthma
Lag 1:1.03 [0.98,1.07]
Lag 2: 1.01 [0.96,1.05]
Lag 3: 1.00 [0.95,1.04]
C0PD (adult)
Lag 1:1.00 [0.93,1.07]
Lag 2: 0.99 [0.92,1.06]
Lag 3: 1.02 [0.95,1.08]
Hospital Admissions - PMio
All Respiratory
Lag 1:0.99 [0.95,1.02]
Lag 2: 0.99 [0.96,1.02]
Lag 3: 1.00 [0.97,1.03]
Asthma
Lag 1:1.03 [0.95,1.12]
Lag 2: 1.01 [0.94,1.10]
Lag 3: 1.00 [0.92,1.09]
C0PD (adult)
Lag 1:0.98 [0.90,1.07]
Lag 2: 1.03 [0.96,1.11]
Lag 3: 1.02 [0.94,1.09]
Reference: Sun et al. (2006, 090768)
ED visits
Pollutant: PMio
Children ED Visits
Period of Study: January 1, 2004 to
Outcome: Asthma (493.xx)
Averaging Time: Monthly avg for 2004
r - 0.626
December 31, 2004

Mean (SD): ~ 60.3//g/m3 (NR)

Age Groups: <55, < 16,16-55 yrs
P - 0.015
Location: Taichung, Taiwan (Central
Study Design: Cross-sectional
(estimated from figure)*
Adult ED Visits
Taiwan)
N: NR
Range (Min, Max): (~ 35, 80)
r - 0.384



Monitoring Stations:
P - 0.109

All diagnoses for all patients at 4 medical

centers
11


Statistical Analyses: Pearson's
Copollutant: NR


correlations, multiple correlation



coefficients from regression analyses.



Covariates: Only copollutants considered



Dose-response Investigated? No



Statistical Package: SPSS



Lags Considered: None


July 2009
E-299
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Szyszkowicz (2007, 092829)
Period of Study: 1 (4/1992-311312002
Location: Edmonton, Canada
Outcome: ED visits for asthma (ICD-
(493)
Study Design: Time-series
Covariates: Temperature, relative
humidity
Statistical Analysis: Poisson regression
Age Groups: All
Pollutant: PMio
Averaging Time: 24h
Mean (SD)Unit: 22.6 (13.1)//gfm3
Median, IQR: 19.4,15.0
Copollutant (correlation): NR
Increment: IQR
Percent Relative Risk (95% CI)
"Only statistically significant results are
presented in the paper*
No lag, a 10 years
April to September, All: 3.7 (1.5-6.0)
April to September, Female: 4.5 (1.8-7.3)
April to September, Male: 3.3 (0.1-6.7)
2d lag, < 10 years
Year round, All: 2.7(0.1-5.4)
April to September, All: 6.3 (2.6-10.2)
April to September, Male: 7.4 (3.1-11.9)
2d lag, a 10 years
April to September, All: 2.4 (0.1-4.7)
April to September, Female: 3.9 (1.1-6.7)
July 2009
E-300
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Tecer et al, (2008,180030) Outcome: ED visits for respiratory
problems (ICD-9 470-478, 493)
Period of Study: 1212004-101
Study Design: Bidirectional Case-
Location: Zonguldak, Turkey	crossover
Covariates: Daily meteorological
parameters
Statistical Analysis: Conditional logistic
regression
Statistical Package: Stata
Age Groups: 0-14 years
Pollutant: PMio
Averaging Time: NR
Mean, Unit: 53.3/yglm3
Range (Min, Max): 12-237.5
Copollutant (correlation):
PM2.6/PM10
Mean: 0.56
Range: 0.17-0.88
Increment: 10 //g/m3
Odds Ratio (95% CI)
Asthma
LagO: 1.14(1.03-1.26)
Lag 1:0.92 (0.83-1.02)
Lag 2: 0.92 (0.81-1.03)
Lag 3: 1.01 (0.92-1.11)
Lag 4: 1.16(1.06-1.26)
Allergic Rhinitis with Asthma
Lag 0:1.07 (1.01-1.13)
Lag 1:0.96 (0.91-1.02)
Lag 2: 0.93 (0.88-0.99)
Lag 3: 0.96 (0.90-1.02)
Lag 4: 1.08 (1.02-1.14)
Allergic Rhinitis
LagO: 1.06 (0.99-1.13)
Lag 1:1.08 (1.01-1.16)
Lag 2: 0.92 (0.87-0.99)
Lag 3: 0.97 (0.92-1.03)
Lag 4: 1.09 (1.03-1.16)
Upper Respiratory Disease
LagO: 0.88 (0.68-1.14)
Lag 1: 1.17 (0.91-1.51)
Lag 2: 1.00 (0.76-1.31)
Lag 3: 0.95 (0.76-1.19)
Lag 4: 1.15(0.97-1.35)
Lower Respiratory Disease
LagO: 1.01 (0.86-1.19)
Lag 1: 1.04 (0.88-1.23)
Lag 2: 1.04 (0.92-1.18)
Lag 3: 1.23 (1.07-1.41)
Lag 4: 0.99 (0.90-1.08)
July 2009
E-301
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Tolbert et al. (2007, 090316) Hospital Admissions/ED visits
Period of Study: 1993 ¦ 2004
Location: Atlanta Metropolitan area,
Georgia
Outcome (ICD-9):
Combined RD group, including:
Asthma (493, 786.07, 786.09), C0PD
(491,492,496), URI (460 - 465,460.0,
477), pneumonia (480 - 486), and
bronchiolitis (466.1, 466.11, and
466.19))
Age Groups Analyzed: All
Study Design: Time series
N: 10,234,490 ER visits (283,360 and
1,072,429 visits included in the CVD and
RD groups, respectively)
Statistical Analyses: Poisson
generalized linear models
Covariates: long-term temporal trends,
season (for RD outcome), temperature,
dew point, days of week, federal
holidays, hospital entry and exit
Season: All
Dose-response Investigated: No
Statistical package: SAS version 9.1
Lags Considered: 3-day moving
averageflag 0 -2)
Pollutant: PMio
Averaging Time: 24 h
Mean (median
IQR, range, 10"1 - 90"1 percentiles):
26.6 (24.8
17.5-33.8
0.5-98.4
12.3-42.8)
Monitoring Stations: NR
Copollutant (correlation): O3: r - 0.59
NO2: r - 0.53
CO: r - 0.51
SO2: r - 0.21
Coarse PM: r - 0.67
PM2.6: r - 0.84
PM2.6 SO4: r - 0.69
PM2.6 EC: r - 0.61
PM2.6 0C: r - 0.65
PM2.6 TC: r - 0.67
PM2.6 water-sol metals: r - 0.73
0HC: r - 0.53
PM Increment: 16.30/yg/m3 (IQR)
Risk ratio [95% CI]:
Single pollutant models:
RD: 1.015(1.006 - 1.024)
Notes: Results of selected multi-pollutant
models for respiratory disease are
presented in Figure 2.
Figure 2: PM10 adjusted for CO, O3, NO2,
or NO2/O3 (non-winter months only)
Summary of results: PM10 remained
predictive of RD in non-winter months
after adjustment for pollutants.
Reference: Tsai et al. (2006, 089768)
Period of Study: 1996 to 2003
Location: Kaohsiung City, Taiwan
Outcome: Asthma (493)
Age Groups: All (universal health care
covers >96% of the population)
Study Design: Case crossover
N: 17,682 admissions
63 hospitals
Statistical Analyses: Conditional
Logistic Regression
Covariates: Temperature, humidity
Season: Warm and cool seasons
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-2 d cumulative
Pollutant: PM10
Averaging Time: 24 h avg
Mean (SD): 76.62/yglm3 (NR)
Percentiles: 25th: 41.73
50th(Median): 74.40
75th: 104.01
Range (Min, Max): (16.70, 232.00)
Monitoring Stations: 6
Copollutant: NR
PM Increment: 62.28/yg/m3
OR Estimate [Lower CI, Upper CI]
lag:
Single-pollutant model, 0-2 d
cumulative lag
>25°C: 1.302 [1.155,1.467]
<	25°C: 1.556(1.398,1.371]
Two-pollutant models, 0-2 d
cumulative lag
PM10 w/ SO2
>	25°C: 1.305 [1.156,1.473]
<	25°C: 1.540(1.374,1.727]
PM10 wI 03
>25°C: 0.985 [0.842,1.152]
<	25°C: 1.581 [1.402,1.783]
PM10 wI NO2
>	25DC: 1.237 [1.052,1.455]
<	25°C: 1.009(0.875,1.163]
PM10 wI CO
>	25°C: 1.156 [1.012,1.320]
<	25°C: 1.300(1.134,1.490]
July 2009
E-302
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ulirsch et al. (2007, 091332)
Period of Study: 1111994 to 312000
Location: Pocatello, Idaho
Chubbuck,Idaho
Outcome: Respiratory Disease (460-499,
509-519)
Reactive Airway Disease (786.09)
Age Groups: All age groups
Study Design: Time series
N: 39,347 visits (TS1)
29,513 visits (TS2)
Statistical Analyses: Poisson
regression, GLM. Sensitivity Analyses
Covariates: Time, Temperature, Relative
Humidity Influenza
Season: WarmfCool
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0 to 4 day lags
Notes: Time series (TS) 1 includes HA,
ED and urgent care visits. TS 2 includes
family practice data available after 1997
Pollutant: PMio
Averaging Time: NR
Mean (SD):TS1: 24.2//gfm3 (NR)
10th: 10.5
90th: 40.7
TS2: 23.2
10th: 10.0
90th: 37.4
Range (Min, Max):
TS1: (3.0,183.0)
TS2: (3.0,183.0)
Monitoring Stations: 4
Notes: Copollutant (correlation): PMio
w/ NO2: r - 0.47. PMio with other
copollutants weakly correlated.
PM Increment: Single Pollutant Models,
TS1: 24.4/yg/m3
Single Pollutant Models: TS2: 23.2//g|m3
Multipollutant Models: TS1/TS2:
50/yg/m3
Mean Percentage Change, lag 0
TS 1: Single Pollutant
All-age (all year): 4.0 [1.4,6.7]
18-64: 3.4[0.2, 6.7]
0-17: 4.3 [-0.1, 8.9]
65+: 5.6 [-1.4,13.1]
0-17/65+: 5.5 [1.4, 9.6]
All age (Cool season): 4.3 [1.3, 7.5]
All age (Warm season): 6.7 [-0.8,14.8]
TS2: Single Pollutant
All-age: 3.3 [0.3, 6.3]
18-64: 3.3[-0.4,7.0]
0-17:5.0(0.1,10.1]
65+: 6.9[-0.4,14.7]
Multipollutant (PMio + SO2)
All-age (all year): TS1 10.8
TS2 17.5
18-64: TS1 8.0
TS2 9.1
0-17: TS1 10.8
TS2 32.7
65+: TS1 8.7
TS2 31.3
0-17/65+: TS1 14.2
TS2 25.3
All age (Cool season) TS1 11.9
Multipollutant (PMio + NO2)
All-age (all year) TS1: TS2 16.3
18-64: TS1 9.3
TS2 17.3
0-17: TS1 4.6
TS2 18.7
65+: TS1 12.4
TS2 32.7
0-17/65+: TS1 9.5
32.7
All age (Cool season): TS1 11.1
TS2 16.8
Notes: Results from multipollutant model
with PMio, SO2 and NO2 also available.
July 2009
E-303
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Vegni and Ros (2004,
Hospital Admissions
Pollutant: PMio
PM Increment: Increase from 5th—95th
087448)


percentile
Outcome (ICD-9): Respiratory, non-
Averaging Time: 24 h
Period of Study: Sep 1, 2001-Sep 31,
infectious admissions
Mean (5th-95th percentile):
Spring: 85 //g|m3
2002


(ICD: NR)
Overall: 41.5(13-98)
summer: 30 //g|m3
Location: Milan area, Italy
Age Groups: NR
SD - 28.2
Autumn: 93 //g|m3


Study Design: Time series
Spring: 29.0 (10-51)
Winter: 115 //g/m3


N: 9881 admissions

RR Estimate [CI]:


SD - 12.6

Statistical Analyses: Poisson regression
summer: 24.8 (10-40)
Overall: 1.10(0.83,1.46]



Covariates: Temperature, wind velocity,
SD - 9.9
Adjusted: 0.97 [0.67,1.41]

relative humidity, week day, holidays


Autumn: 51.8 (21-114)
Notes: 1-day and 2-day lags show similar

Season: Spring, summer, autumn, winter
results, with no association between


SD - 27.1
PMio and daily hospital admissions

Dose-response Investigated? No

Statistical Package: STATA v. 5
Winter: 64.1 (20-135)


Lags Considered: 0,1, and 2-day
SD - 35.7
Monitoring Stations: 1
Copollutant: NR

Reference: Vigotti et al. (2007, 090711
Period of Study: 112000-1212000
Location: Pisa, Italy
ED Visits
Outcome: Asthmatic attack (493), dry
cough (468), acute bronchitis (466)
Age Groups: < 10 y
65 +
Study Design: Time series
N: 966 Emergency room visits
Statistical Analyses: Poisson
regression, GAM, LOESS smoothers,
stringent criteria
Covariates: temperature, humidity,
relative humidity, day of study, rainfall,
influenza, day of-the-wk, holidays, time
trend
Season: All year
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-5 d
Pollutant: PMio
Averaging Time: 24-h
Mean (SD): 35.4 (15.8)//g/m3
Percentiles: 25th: NR
50th(Median): 31.6
75th: NR
Range (Min, Max): (9.5,100.1)
Monitoring Stations:
2
Copollutant (correlation): PMio:
NO?
r - 0.58
CO
r - 0.70
PM Increment: 10 //g/m3
RR Estimate [Lower CI, Upper CI]
lag:
< 10 y: 10%[2.3,18.2]
lag 1
65+: 8.5% [1.5,16.1]
lag 2
July 2009
E-304
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Xirasagar et al. (2006,
093267)
Period of Study: 1998-2001
Location: Taiwan
Hospital Admission/ED:
Outcome: Asthma or Asthmatic
Bronchitis (493)
Age Groups: Less than 2 years old,
2 ~5 years old, 6 —14 years old
Study Design: N:
N - 27, 275 pediatric hospitalizations
Statistical Analyses: ARIMA Modeling
Spearman's Correlations
Covariates: Season, ambient temp., rel.
humidity, atmospheric pressure, rainfall, h
of sunshine
Season: Spring: February to April
summer: May to July
Autumn: August to October
Winter: November to January
Dose-response Investigated? No
Statistical Package: EViews 4
Lags Considered: NR
Pollutant: PMio
Averaging Time: Monthly Means
Mean (SD): 24.4//g/m3 (NR)
Percentiles: NR
Range (Min, Max): NR
PM Component: NR
Monitoring Stations: 44 air quality
monitoring banks. 23 weather
observatories
Notes: Copollutant (correlation): Less
than 2 years old: r - 0.315
2 — 5 years old: r - 0.589
6 —14 years old: r - 0.493
PM Increment: NR
RR Estimate [Lower CI, Upper CI]
lag: NR
AR Estimate [Lower CI, Upper CI]
lag: NR
Notes: Plot of monthly asthma admission
rates per 100,000 population by age
group
Plot of mean monthly concentration
trends of criteria air pollutants
Mean monthly trends of climatic factors
Other Outcomes Assessed? NR
Other Exposures Assessed?
Seasonality
Reference: Wong et al., (2002, 023232) Hospital Admissions
Period of Study: 1995-1997 (Hong
Kong) and 1992-1994 (London)
Location: Hong Kong and London
Outcome (ICD- NR): Asthma (493) for
ages 15-64 and respiratory disease (460-
519) for ages 65 +
Age Groups: 15-64, 65 +
Study Design: Time series
N: NR
Statistical Analyses: Poisson
regression, GAM
Covariates: Temperature, humidity, and
influenza
Season: Warm (Apr-Sep) and cool (Oct-
Mar)
Dose-response Investigated? Yes
Statistical Package: S-Plus
Lags Considered: 0-3 days
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max): Hong Kong: 51.8
(14.1-163.8) SD - 25.0
London: 28.5 (6.8-99.8)
SD - 13.7
Monitoring Stations: NR
Copollutant (correlation): Hong Kong
NO2: r - 0.82
SO2: r - 0.30
O3: r - 0.54
London
NO2: r - 0.68
SO2: r - 0.64
O3: r - 0.17
Other variables: Hong Kong
Temp: r - -0.42
Humidity: r - -0.53
London
Temp: r - 0.02
Humidity: r - -0.05
PM Increment: 10 //g/m3
ER Estimate [CI]:
Single-pollutant excess risk (mean lag 0-1
day)
Asthma-Hong Kong: -1.1 [-2.4,0.1]
Asthma-London: 1.4 [-0.1,3.0]
Respiratory Disease-Hong Kong: 1.0
[0.5,1.5]
Respiratory Disease-London: 0.4 [¦
0.3,1.2]
Warm season
Asthma-Hong Kong: -1.0 [-2.8, 0.8]
Asthma-London: 0.6 [-1.9,3.1]
Respiratory Disease-Hong Kong: 0.8
[0.1,1.4]
Respiratory Disease-London: 1.8
[0.5,3.1]
Cool season
Asthma-Hong Kong: -1.2 [-2.8,0.4]
Asthma-London: 1.6 [-0.3,3.6]
Respiratory Disease-Hong Kong: 1.2
[0.6,1.9]
Respiratory Disease-London: -0.5 [¦
1.5,0.5]
Notes: RRs are shown graphically in Fig
1 and 2. Exposure response curves are
provided in Fig 5 of the article
July 2009
E-305
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Wong et al. (2006, 093266)
Period of Study: 2000-2002
Location: Hong Kong (8 districts)
General Practitioner Visits
Outcome (ICPC-2): Respiratory
diseasesfsymptoms: upper respiratory
tract infections (URTI), lower respiratory
infections, influenza, asthma, COPD,
allergic rhinitis, cough, and other
respiratory diseases
Age Groups: All ages
Study Design: Time series
N: 269,579 visits
Statistical Analyses: GAM, Poisson
regression
Covariates: Season, day of the week,
climate
Season: NR
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0-3 days
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max): Ranged from 43.4-
56.9 (dependent on location)
Monitoring Stations: 1 per district
Copollutant (correlation): PM2.6:
r - 0.94
O3: r - 0.40
SO2: r - 0.28
PM Increment: 10/yg/m3
RR Estimate [CI]:
Overall URTI
1.020[1.016,1.025]
Overall Non-UTRI
1.025[1.018,1.032]
Notes: RRs are also reported for each
individual general practitioner yielding
similar results
Reference: Yang et al. (2007, 092847)
Period of Study: 1996-2003
Location: Taipei, Taiwan
Hospital Admission/ED:
Outcome: Asthma (493)
Age Groups: All ages
Study Design: Case-crossover
N: 25,602 asthma hospital admissions
Statistical Analyses: NR
Covariates: Temperature, humidity, day
of-the-wk, seasonality, long term trends
Season: All year
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-2
Pollutant: PM10
Averaging Time: NR
Mean (SD): 48.99/yglm3
Percentiles: 25th: 32.64
50th(Median): 44.13
75th: 59.05
Range (Min, Max): (14.44,234.91)
PM Component: NR
Monitoring Stations:
6 Stations
Notes: Copollutant: NR
PM Increment: 26.41 /yglm3
OR Estimate [Lower CI, Upper CI]
lag:
Asthma
Single-Pollutant Model: Temperature
>25° C: 1.046(0.971,1.128]
Temperature <25° C: 1.048(1.011,
1.251]
Two-Pollutant Model: Adjusted for SO2:
>25° 0-1.006(0.920,1.099]
<25° 0-1.088(1.040,1.138]
Adjusted for NO?: >25° 0-0.800(0.717,
0.892]
<25° 0-0.982(0.937,1.029]
Adjusted for 00: > 25° 0-0.920(0.844,
1.002]
<25° 0-1.029(0.984,1.076]
Adjusted for O3: > 25° 0-1.038(0.950,
1.134]
<25° 0-1.042(1.004,1.081]
AR Estimate [Lower CI, Upper CI]
lag: NR
Notes: Other Outcomes Assessed? NR
Other Exposures Assessed? SO2, NO2,
00, 0s
July 2009
E-306
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Yang et al. (2007, 092847)
Period of Study: 1996-2003
Location: Taipei, Taiwan
Hospital Admission	Pollutant: PMio
Outcome: C0PD (490-192), (494), (496) Averaging Time: 24 h
Age Groups: All ages	Mean (SD): 48.99//gfm3
Study Design: Case-crossover	25th: 32.64
N: 46,491 C0PD admissions, 47 hospitals 50th(Median): 44.13
75th: 59.05
Statistical Analyses: Conditional
logistic regression
Covariates: Weather, day of-the-wk,
seasonality, long term trends
Season: Warm/Cool
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-2 cumulative
Range (Min, Max):
(14.44,48.99)
Monitoring Stations:
6 Stations
Notes: Copollutant: NR
PM Increment: 26.41 //g|m3
OR Estimate [Lower CI, Upper CI]
Single-Pollutant Model (0-2 d cum lag):
Temperature >20° C: 1.133(1.098,
1.168]
Temperature <20° C: 1.035(0.994,
1.077]
Two-Pollutant Model:
PMio w/ SO2:
>20° C-1.180[1.139,1.223]
<20° C-1.004(0.954,1.057]
PMio wI NO2:
>20° C-1.013[0.973,1.055]
<20° C-1.074(1.022,1.129]
PMio wI CO:
>20° C-1.061(1.023,1.100]
<20° C-1.067(1.016,1.120]
PMio wI O3:
>20° C-1.097(1.062,1.133]
<20° C-1.036(0.996,1.079]
Reference: Yang et al. (2004, 087488) Hospital Admissions
Period of Study: Jun 1,1995-Mar 31,
1999
Location: Vancouver area, British
Columbia
Outcome (ICD-9): Respiratory diseases
(460-519), pneumonia only (480-486),
asthma only (493)
Age Groups: 0-3 yrs
Study Design: Case control, bidirectional
case-crossover (BCC), and time series
(TS)
N: 1610 cases
Statistical Analyses: Chi-square test,
Logistic regression, GAM (time-series),
GLM with parametric natural cubic
splines
Covariates: Gender, socioeconomic
status, weekday, season, study year,
influenza epidemic month
Season: Spring, summer, fall, winter
Dose-response Investigated? No
Statistical Package: SAS (Case control
and BCC), S-Plus (TS)
Lags Considered: 0-7 days
Pollutant: PMio
Averaging Time: 24 h
Mean (min-max):
13.3 (3.8-52.2)
SD - 6.1
Monitoring Stations: NR (data obtained
from Greater Vancouver Regional District
Air Quality Dept)
Copollutant (correlation): PM2.6:
r - 0.83
PMio2.5: r - 0.83
CO: r - 0.46
O3: r - -0.08
NO?: r - 0.54
SO2: r - 0.61
PM Increment: 7.9/yg/m3 (IQR)
OR Estimate [CI]:
Values NR
Notes: Author states that ORs for PMio
increased with lag time up to 3 days for
both single and multiple-pollutant models.
'All units expressed in //gfm3 unless otherwise specified.
July 2009
E-307
DRAFT-DO NOT CITE OR QUOTE

-------
Table E-13. Short-term exposure-respiratory-ED/HA-PMi0 2.5
Reference
Design & Methods
Concentrations
Effect Estimates (95% CI)
Reference: Chen et al. (2005,
087555)
Period of Study: Jun 1,
1995-Mar 31,1999
Location: Vancouver area, BC
Hospital Admissions
Outcome (ICD-9): Acute respiratory
infections (460-466), upper respiratory
tract infections (470-478), pneumonia
and influenza (480-487), C0PD and
allied conditions (490-496), other
respiratory diseases (500-519)
Age Groups: >65 yrs
Study Design: Time series
N: 12,869
Statistical Analyses: GLM
Covariates: Temp and relative
humidity
Season: NR
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 1, 2, 3, 4, 5, 6, and
7-day avg
Pollutant: PMio 2.5 (/yglm3)
Averaging Time: 24 h
Mean (min-max):
5.6(0.1-24.6)
SD - 3.6
Monitoring Stations: 13
Copollutant (correlation):
PM2.6: r - 0.38
PM10: r = 0.83
C0H: r - 0.63
CO: r - 0.53
O3: r - -0.13
NO?: r - 0.54
SO2: r - 0.57
Other variables:
Mean temp: r - 0.13
Rel humidity: r - -0.27
PM Increment: 4.2/yglm3
RR Estimate [CI]:
Adj for weather conditions
Overall admission
1-day	avg: 1.03 [1.00,1.06]
2-day	avg: 1.05(1.02,1.08]
3-day	avg: 1.06 [1.02,1.09]
Adj for weather conditions and copollutants
Overall admission
1-day	avg: 1.02 [0.98,1.06]
2-day	avg: 1.05(1.01,1.10]
3-day	avg: 1.06(1.02,1.11]
Notes: RR's were also provided for lags 4-7 in Table 3, yielding
similar results
July 2009
E-308
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Fung et al. (2006,
089789)
Period of Study: 6/1/95-
3/31/99
Location: Vancouver, Canada
Hospital Admission/ED: Hospital
Admission
Outcome: Respiratory diseases (460-
519)
Age Groups: Age > 65
Study Design: Time series
N: 26,275 individuals admitted
Statistical Analyses: Poisson
regression (spline 12 knots), case-
crossover (controls +/7 d days from
case date), Dewanji and Moolgavkar
(DM) method
Covariates: Long-term trends, day-of-
the-week effect, weather
Season: All year
Dose-response Investigated? No
Statistical Package: SPIus, R
Lags Considered: 0-7 d
Pollutant: PMio 2.5 (/yg/m3)
PM Increment::
Averaging Time: 24-h Avg
4.3/yg/m3
Mean (SD)
RR Estimate (65+ years)
5.6(3.88)/yg/m3
DM method:
Range (Min, Max):
1.011[0.998,1.024]
(-2.9, 27.07)
lagO
Monitoring Stations:
1.016[1.0,1.032]
NR
3 d avg
Notes: Copollutant
1.020(1.001,1.039]
(correlation): PMio2.5
5 d avg

PMio


1.020(0.998,1.042]
r - 0.83


7 d avg
PM2.6


Time series:
r - 0.34


1.0168[1.003,1.031]
CO


lagO
r - 0.51


1.020(1.003,1.037]
CoH


3 d avg
r - 0.61


1.019(0.999,1.039]
0s


5 d avg
r - -0.11


1.018(0.994,1.042]
NO?


7 d avg
r - 0.52


Case-crossover:
SO?


1.019(1.003,1.034]
r - 0.57


lagO

1.019(1.009,1.038]

3 d avg

1.020(0.999,1.042]

5 d avg

1.018(0.994,1.043]

7 d avg
July 2009
E-309
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Halonen et al.
Outcome: Hospital Admissions
Pollutant: PM102 E
PM Increment: Interquartile Range
(2009,180379)



Age Groups: 65+ yrs
Averaging Time: daily
Percent Change (Lower CI, Upper CI):
Period of Study: 1998-2004



Study Design: time series
Mean (SD): NR
All Respiratory Mortality
Location: Helsinki, Finland



N: NR
Min: 0.0
Lag 0: -0.66 (-4.16, 2.97)

Statistical Analyses: Poisson, GAM
25"' percentile: 4.9
Lag 1:2.90 (-0.48, 6.39)*

Covariates: temperature, humidity,
50"1 percentile: 7.5
Lag 2: 0.35 (-3.03, 3.84)

influenza epidemics, high pollen

Lag 3: -0.38 (-3.67, 3.02)

episodes, holidays
75"1 percentile: 12.1

Dose-response Investigated? No
Max: 101.4
5-d mean: 0.36 (-4.54, 5.51)

Statistical Package: R
Monitoring Stations: NR
Pneumonia HA

Lags Considered: lags 0-3 815d (0-4)
Copollutant: PMcora, PM0.030.1,
Lag 0:0.72 (-1.28, 2.77)

mean
PM 0.6
Ranged between r - 0.28 and
r - 0.73 across the six cities.
PM Increment: 10//g|m3, and an 18.8//g|m3 increase
(corresponding to an increase in pollutant levels between the
lowest of the 5th percentiles and the highest of the 95th
percentiles of the cities' distributions)
ERR (excess relative risk) Estimate [CI]: For all respiratory
diseases (10 //g|m3 increase): 0-14 years: 6.2% [0.4,12.3]
15-64 years: 2.6%
[¦0.5,5.8]
a 65 years: 1.9% [-1.9, 5.9]
For all respiratory diseases (18.8/yg/m3 increase): 0-14 years:
12.0(0.8, 24.3]
15-64 years: 5.0 [-0.9,11.1]
a 65 years: 3.7 [-3.6,11.4]
For respiratory infections (10/yg/m3): All ages: 4.4% [0.9,
8.0]
For respiratory infections (18/yglm3): All ages: 8.4% [1.7,
15.5]
July 2009
E-310
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lin et al.
(2005, 087828)
Period of Study: 1998-2001
Location: Toronto, North York,
East York, Etobicoke,
Scarborough, and York
(Canada)
Hospital Admissions
Outcome (ICD-9): Respiratory
infections including laryngitis,
tracheitis, bronchitis, bronchiolitis,
pneumonia, and influenza (464, 466,
480-487)
Age Groups: 0-14 yrs
Study Design: Bidirectional case-
crossover
N: 6782 respiratory infection
hospitalizations
Statistical Analyses: Conditional
logistic regression (Cox proportional
hazards model)
Covariates: Daily mean temp and dew
point temp
Season: NR
Dose-response Investigated? No
Statistical Package: SAS 8.2 PHREG
procedure
Lags Considered: 1-7 day averages
Pollutant: PMio-2.5 (//g|m3)
Averaging Time: 24 h
Mean (min-max):
10.86 (0-45.00)
SD - 5.37
Monitoring Stations: 4
Copollutant (correlation):
PM2.6: r - 0.33
PM10: r - 0.76
CO: r - 0.06
SO2: r - 0.29
NO?: r - 0.40
O3: r = 0.30
PM Increment: 6.5//g|m3
OR Estimate [CI]:
Adjusted for weather
4 day avg: 1.16 [1.07,1.26]
6 day avg: 1.21 [1.10,1.32]
Adj for weather and other gaseous pollutants
4 day avg: 1.13(1.03,1.23]
6 day avg: 1.17(1.06,1.29]
Notes: OR's were also categorized into "Boys" and "Girls,"
yielding similar results
Reference: Lin et al. (2002,
026067)
Period of Study: Jan 1,
1981-Dec 31,1993
Location: Toronto
Hospital Admissions
Outcome (ICD-9): Asthma (493)
Age Groups: 6-12 yrs
Study Design: Uni- and bi-directional
case-crossover (UCC, BCC) and time-
series (TS)
N: 7,319 asthma admissions
Statistical Analyses: Conditional
logistic regression, GAM
Covariates: Maximum and minimum
temp, avg relative humidity
Season: Apr-Sep, Oct-Mar
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1-7 day averages
Pollutant: PMio 2.5 (//g|m3)
Averaging Time: 6 days
(predicted daily values)
Mean (min-max):
12.17 (0-68.00)
SD - 7.55
Monitoring Stations: 1
Copollutant (correlation):
PM2.6: r - 0.44
PM10: r = 0.83
CO: r - 0.17
SO2: r - 0.28
NO2: r - 0.38
O3: r - 0.56
PM Increment: 8.4/yg/m3
RR Estimate [CI]:
Adj for weather and gaseous pollutants
BCC 5 day avg: 1.14(1.01,1.28]
BCC 6 day avg: 1.17 [1.03,1.33]
TS 5 day avg: 1.14 [1.05,1.23]
TS 6 day avg: 1.15 [1.06,1.25]
Boys-adj for weather
UCC 1 day avg: 1.08 [1.01,1.16]
UCC 2 day avg: 1.08 [0.99,1.17]
BCC 1 day avg: 1.06(1.00,1.14]
BCC 2 day avg: 1.06 [0.98,1.14]
TS 1 day avg: 1.08 [1.03,1.12]
TS 2 day avg: 1.07 [1.01,1.13]
Girls-adj for weather
UCC 1 day avg: 1.07 [0.97,1.18]
UCC 2 day avg: 1.16 [1.03,1.31]
BCC 1 day avg: 0.98 [0.90,1.07]
BCC 2 day avg: 1.05 [0.94,1.16]
TS 1 day avg: 1.00 [0.94,1.06]
TS 2 day avg: 1.05 [0.98,1.13]
Notes: The author also provides RR using UCC, BCC, and TS
analysis for female and male groups for days 3-7, yielding
similar results
July 2009
E-311
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Peel et al. (2005,
056305)
Period of Study: Jan 1993-
Aug 2000
Location: Atlanta, Georgia
ED visits
Outcome: Asthma (493, 786.09)
C0PD (491,492, 496)
URI (460-466, 477)
Pneumonia (480-486)
Age Groups: All ages. Secondary
analyses conducted by age group: 0-1,
2-18, >18
Study Design: Time series
N: 31 hospitals
Statistical Analyses: Poisson GEE for
URI, asthma and all RD
Poisson GLM for pneumonia and C0PD)
Covariates: Avg temperature and dew
point, pollen counts
Season: All (secondary analyses of
warm season)
Dose-response Investigated? Yes
Statistical Package: SAS 8.3
S-Plus 2000
Lags Considered: 0-7 d , 3 d ma, 0-13
d unconstrained distributed lag
Pollutant: PMio 2.5 (//g|m3)
Averaging Time: 24 h avg
Mean (SD): 9.7 (4.7)
Percentiles: 10th: 4.4
90th: 16.2
Monitoring Stations:
"Several"
Copollutant (correlation): 24 h
PMio: r - 0.59
8 h O3: r — 0.35
1 h NO?: r - 0.46
1 h CO: r - 0.32
1 h SO2: r - 0.21
24 h PM2.6: r - 0.43
Components: r ranged from 0.23-
0.51
PM Increment: 5
RR Estimate [Lower CI, Upper CI]
All Respiratory Outcomes: 1.003 [0.982,1.025]
URI: 1.013(0.987,1.039]
Asthma: 0.998 [0.987,1.039]
Pneumonia: 0.975 [0.940,1.011]
C0PD: 0.948 [0.897,1.003]
Reference: Peng et al. (2008,
156850)
Period of Study: January 1,
1999-December 31, 2005
Location: 108 U.S. counties in
the following states: Alabama,
Arizona, California, Colorado,
Connecticut, District of
Columbia, Florida, Georgia,
Idaho, Illinois, Indiana,
Kentucky, Louisiana, Maine,
Maryland, Massachusetts,
Michigan, Minnesota, Missouri,
Nevada, New Hampshire, New
Jersey, New Mexico, New
York, North Carolina, Ohio,
Oklahoma, Oregon,
Pennsylvania, Rhode Island,
South Carolina, Tennessee,
Texas, Utah, Virginia,
Washington, West Virginia,
Wisconsin
Outcome (ICD-9): Emergency
hospitalizations for respiratory disease,
including C0PD (490-492) and
respiratory tract infections (464-466,
480 ¦ 487)
Age Groups: 65 + years, 65-74, ,75
Study Design: Time series
N: approximately 12 million Medicare
enrollees (1.4 million RD admissions)
Statistical Analyses: Two-stage
Bayesian hierarchical models: Over
dispersed Poisson models for county-
specific data. Bayesian hierarchical
models to obtain national avg estimate
Covariates: Day of the week, age-
specific intercept, temperature, dew
point temperature, calendar time,
indicator for age of 75 years or older.
Some models were adjusted for PM2.6.
Dose-response Investigated: No
Statistical Package: R version 2.6.2
Lags Considered: 0-2 days
Pollutant: PMio 2.5
Averaging Time: 24 h
Mean (IQR): All counties
assessed: 9.8 (6.9-15.0)
Counties in Eastern US: 9.1
(6.6-13.1)
Counties in Western US: 15.4
(10.3-21.8)
Monitoring Stations: At least 1
pair of co-located monitors
(physically located in the same
place) for PM10 and PM2.6 per
county
Copollutant (correlation):
PM2.6: r - 0.12
PM10: r - 0.75
Other variables: Median within-
county correlations between
monitors: r - 0.60
PM Increment: 10//g|m3
Percentage change [95% CI]: Respiratory disease (RD): Lag 0
(unadjusted for PM2.6): 0.33 [-0.21, 0.86]
Lag 0 (adjusted for PM2.6): 0.26 [-0.32, 0.84]
Most values NR (see note)
Notes: Figure 3: Percentage change in emergency hospital
admissions for RD per 10/yg/m3 increase in PM (single
pollutant model and model adjusted for PM2.6 concentration)
Figure 4: Percentage change in emergency hospital admissions
rate for CVD and RD per a 10//g|m3 increase in PMio-2.5 (0-2
day lags, Eastern vs. Western USA)
July 2009
E-312
DRAFT-DO NOT CITE OR QUOTE

-------
Reference	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Reference: Slaughter et al.
(2005, 073854)
Hospital Admissions and ED visits
Pollutant: PMio 2.5 (/yglm3)
PM Increment: 25/yg/m3
Outcome: All respiratory (460-519)
Averaging Time: 24 h avg
RR Estimate [Lower CI, Upper CI]
Period of Study: January
Asthma (493)
Range (90% of


1995 through June 2001
lag:

COPD (491,492, 494,496)
Concentrations): Reported for
ER visits:

Location: Spokane, WA
PM2.6 and PM10 only

Notes
Pneumonia (480-487)
Monitoring Stations: 1
PMio-2.5


Acute URI not including colds and
Copollutant (correlation): PM10
All Respiratory


sinusitis (464, 466,490)
2.5



Lag 1:1.01 [0.98,1
1.04]

Age Groups: All, 15+ years for COPD
PMir - 0.19
Lag 2:1.01(0.98,1
1.04]

Study Design: Time series
PM2.6 r - 0.31
Lag 3:1.02(0.99,1
1.05]

N: 2373 visit records
PM10 r - 0.94




Acute Asthma


Statistical Analyses: Poisson
CO r - 0.32
Lag 1:1.03(0.98,1
1.08]

regression, GLM with natural splines.

For comparison also used GAM with
Temperature r - 0.11
Lag 2:1.01(0.96,1
1.07]

smoothing splines and default


convergence criteria.

Lag 3: 0.99(0.94,1
1.05]

Covariates: Season, temperature,

COPD (adult)


relative humidity, day of week





Lag 1:1.01 (0.93,1
1.09]

Season: All




Lag 2: 0.98(0.90,1
1.06]

Dose-response Investigated?: No




Lag 3:1.02(0.95,1
1.10]

Statistical Package: SAS, SPLUS


Lags Considered: 1 -3 d



July 2009
E-313
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Tecer et al. (2008, Outcome: ED visits for respiratory
180030)
problems (ICD-9 470-478,493)
Period of Study: 1212004-10| Study Design: Bidirectional Case-
crossover
Location: Zonguldak, Turkey
Covariates: Daily meteorological
parameters
Statistical Analysis: Conditional
logistic regression
Statistical Package: Stata
Age Groups: 0-14 years
Pollutant: PMio-2.b
Averaging Time: NR
Mean, Unit: 24.3/yglm3
Range (Min, Max): 4,195.!
Copollutant (correlation):
PM2.BIPM10.2.B
Mean: 1.49
e: 0.21, 7.53
Increment: 10 //g/m3
Odds Ratio (95% CI)
Asthma
Lag 0:1.18(1.01-1.39)
Lag 1:0.92(0.78-1.08)
Lag 2:0.98(0.84-1.15)
Lag 3:1.11 (0.97-1.27)
Lag 4:1.17(1.05-1.31)
Allergic Rhinitis with Asthma
Lag 0:0.96(0.88-1.04)
Lag 1:1.08(0.99-1.18)
Lag 2: 0.93 (0.86-1.02)
Lag 3:0.94 (0.86-1.03)
Lag 4:1.10(1.03-1.18)
Allergic Rhinitis
Lag 0:1.06(0.95-1.19)
Lag 1:1.17 (1.04-1.31)
Lag 2:0.92(0.84-1.02)
Lag 3:0.99(0.91-1.08)
Lag 4:1.15(1.06-1.25)
Upper Respiratory Disease
Lag 0:0.80(0.54-1.19)
Lag 1:1.22 (0.92-1.61)
Lag 2: 0.97 (0.70-1.33)
Lag 3:0.94 (0.66-1.33)
Lag 4:1.08(0.88-1.32)
Lower Respiratory Disease
Lag 0:0.90(0.71-1.16)
Lag 1:1.20 (0.97-1.50)
Lag 2:1.00(0.84-1.19)
Lag 3:1.26(1.08-1.47)
Lag 4:1.02(0.93-1.13)
July 2009
E-314
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Yang et al„ (2004, Hospital Admissions
087488)
Period of Study: Jun 1,
1995-Mar 31,1999
Location: Vancouver area,
British Columbia
Outcome (ICD-9): Respiratory disi
(460-519), pneumonia only (480-486),
asthma only (493)
Age Groups: 0-3 yrs
Study Design: Case control,
bidirectional case-crossover (BCC), and
time series (TS)
N: 1610 cases
Statistical Analyses: Chi-square test,
Logistic regression, GAM (time-series),
GLM with parametric natural cubic
splines
Covariates: Gender, socioeconomic
status, weekday, season, study year,
influenza epidemic month
Season: Spring, summer, fall, winter
Dose-response Investigated? No
Statistical Package: SAS (Case
control and BCC), S-Plus (TS)
Lags Considered: 0-7 days
Pollutant: PMio-2.5 (/yglm3)
Averaging Time: 24 h
Mean (min-max):
5.6 (0-24.6)
SD - 3.6
PM Increment: 4.2/yglm3 (IQR)
OR Estimate [CI]:
3-day lag
1.12[0.98,1.28]
Adj for gaseous pollutants: 1.22 [1.02,1.48]
Monitoring Stations: NR (data Notes: Author states that ORs for PMio 2.5 increased with lag
obtained from Greater Vancouver time up to 3 days for both single and multiple-pollutant models.
Regional District Air Quality More adjusted ORs and RRs are provided in Fig 1.
Dept)
Copollutant (correlation):
PM2.6: r - 0.39
PM10: r = 0.83
CO: r - 0.33
O3: r - -0.16
NO?: r - 0.37
SO2: r - 0.54
'All units expressed in //gfm3 unless otherwise specified.
July 2009
E-315
DRAFT-DO NOT CITE OR QUOTE

-------
Table E-14. Short-term exposure-respiratory-ED/HA-PM2.5 (including PM components/sources).
Concentrations'
Reference
Design & Methods
Effect Estimates (95% CI)
Reference: Andersen et al. (2008,
189651)
Period of Study: May 2001 - December
2004
Location: Copenhagen, Denmark
Outcome (ICD-10): RD, including chronic
bronchitis (J41-42), emphysema (J43),
other chronic obstructive pulmonary
disease (J44), asthma (J45), and status
asthmaticus (J46). Pediatric hospital
admissions for asthma (J45) and status
asthmaticus (J46).
Age Groups: > 5-18 years (asthma)
Study Design: Time series
N: NR
Statistical Analyses: Poisson GAM
Covariates: Temperature, dew-point
temperature, long-term trend,
seasonality, influenza, day of the week,
public holidays, school holidays (only for
5-18 year olds), pollen (only for pediatric
asthma outcome)
Season: NR
Dose-response Investigated: No
Statistical Package: R statistical
software (gam procedure, mgcv package)
Lags Considered: Lag 0-5 days, 4-day
pollutant avg (lag 0-3) for CVD, 5-day avg
(lag 0-4) for RD, and a 6-day avg (lag 0-5)
for asthma.
Pollutant: PM2.6
Averaging Time: 24 h
Mean |jg/m3 (SD
median
IQR
99th percentile): 10 (5
9
7-12
28)
Monitoring Stations: 1
Copollutant (correlation): NCtot: r -
0.40
NC100: r - 0.29
NCa12: r - 0.07
Nca23: r - -0.25
NCa57: r - 0.51
NCa2i2: r = 0.82
PMio: r - 0.80
CO: r - 0.46
NO?: r - 0.42
Nox: r - 0.40
Noxcurbside: r - 0.28
0a: r - -0.20
Other variables: Temperature: r - -0.01
Relative humidity: r - 0.21
PM Increment: 5 [Jg/nr1 (IQR)
Relative risk (RR) Estimate [CI]: RD
hospital admissions (5 day avg, lag 0 -4),
age 65+:
One-pollutant model: 1.00 [0.95-1.00]
Adj for NCtot: 1.00 [0.95-1.06]
Asthma hospital admissions (6 day avg
lag 0-5), age 5-18:
One-pollutant model: 1.15 [1.00-1.32]
Adj for NCtot: 1.13(0.98-1.32]
Estimates for individual day lags reported
only in figure form (see notes):
Notes: RD: No statistically or marginally
significant associations. Positive
associatons at Lag 4-5.Asthma: Wide
confidence intervals make interpretation
dificult. Positive associations at Lag 1, 2,
3.
Reference: Babin et al. (2007,188476)
Period of Study: 10/2001-9/2004
Location: Washington, DC
ED Visit/Admissions
Outcome: Asthma-493
Age Groups: 1-17 years,1-4, 5-12,13-
17
Study Design: Time-series
N: NR
Statistical Analyses: Poisson
regression, spline w/12 knots to adjust
for long term trend
Covariates: Temperature, mold, pollen,
seasonal trends,
Season: All
Dose-response lnvestigated?No
Statistical Package: STATA
Lags Considered: 0-4
Pollutant: PM2.6
Averaging Time: 24-hs
Mean: "low, never reached code red"
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: 3
Copollutant (correlation): NR
PM Increment: 1 /yglm3
%Change ED Visits
Ages 5-12:
¦0.2 (-0.6,0.2), lag 0
% Change ED Admissions:
Ages 5-12:
¦0.4 (-1.6,0.8), lagO
Ages 1-17:
0.2 (-0.6,1.1), lagO
AR Estimate [Lower CI, Upper CI]
lag:
NR
Notes: No significant interactions
between PM and ozone or other
covariates were observed.
July 2009
E-316
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Barnett et al. (2005,
087394)
Period of Study: 1998-2001
Location: 5 Australian cities (Brisbane,
Canberra, Melbourne, Perth, and Sydney)
and 2 New Zealand cities (Auckland,
Christchurch)
Outcome (ICD: NR): All respiratory
admissions (including asthma, pneumonia,
and acute bronchitis)
Age Groups: Children aged < 1 year, 1-
4 years, and 5-14 years
Study Design: Matched case-crossover
N: ~ 2.4 million children < 15 years old
Statistical Analyses: Random effects
meta-analysis
Covariates: Temperature, current minus
previous day's temperature, relative
humidity, pressure, extremes of hot and
cold, day of the week, public holiday, and
day after public holiday
Season: Warm (Nov-Apr) and Cool (May-
Oct)
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: NR
Pollutant: PM2.6
Averaging Time: 24-hs
Mean (min-max):
Auckland (A): 11.0 (2.1-37.6)
Brisbane (B): 9.7 (3.2-122.8)
Canberra (Ca): NR
Christchurch (Ch): NR
Melbourne (M): 8.9 (2.8-43.3)
Perth (P): 8.1 (1.7-29.3)
Sydney (S): 9.4 (2.4-82.1)
Monitoring Stations: 1-3 per city
Copollutant: NR
PM Increment: 3.8 //g|m3 (IQR)
Percent Increase Estimate [CI]:
Pneumonia & Acute Bronchitis:
Single Pollutant Model
<1 yr (B,M,P,S): 1.7 [0.0,3.4]
1-4 yrs (B,M,P,S): 2.4 [0.1,4.7]
Matched Multipollutant Model
1-4 yrs with 1-h SOz (B,S): 1.9 [-1.7,5.6]
1-4 yrs with temp (B,M,P,S): 2.3 [¦
0.4,5.1]
Respiratory Admissions:
Single Pollutant Model
<1 yr (B,M,P,S): 2.4(1.0,3.8]
1-4 yrs (B,M,P,S): 1.7(0.7,2.7]
Matched Pollutant Model
<1 yr with 1-h SOz (B,S): 3.1 [0.5,5.7]
< 1 yr with temp (B,M,P,S): 1.8
[0.2,3.4]
1-4 yrs with PMio (B,M,P,S): 2.9
[0.2,5.6]
1-4 yrs with 1-h SOz (B,S): 1.3 [-1.8,4.4]
1-4 yrs with 1-h NOz (B,M,P,S): -1.5 [-
3.2,0.2]
1-4 yrs with temp (B,M,P,S,): 1.5 [¦
0.2,3.1]
Reference: Bell et al. (2008, 091268)
Period of Study: 1995 ¦ 2002
Location: Taipei, Taiwan
Outcome (ICD-9): Hospital admissions
for asthma (493), and pneumonia (486).
Age Groups: All
Study Design: Time series
N: 19,966 hospital admissions for
pneumonia, and 10,231 for asthma
Statistical Analyses: Poisson regression
Covariates: Day of the week, time,
apparent temperature, long-term trends,
seasonality
Season: All
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: lags 0-3 days, mean of
lags 0-3
Pollutant: PM2.6
Averaging Time: 24 h
Mean (range
IQR): 31.6(0.50-355.0
20.2)
Monitoring Stations: 2
Copollutant (correlation): NR
PM Increment: 20/yg/m3 (near IQR)
Percentage increase estimate [95% CI]:
Asthma: L0: 0.46 (-2.41, 3.42)
L1:-1.36 (-4.33,1.71)
L2: -0.83 (-3.67, 2.10)
L3: -0.78 (-3.63, 2.16)
L03: -1.75 (-6.21, 2.92)
Pneumonia: L0: 0.06 (-2.74, 2.94)
L1: 0.34(-2.446, 3.20)
L2: -0.59 (-3.38, 2.29)
L3:-0.44 (-3.22, 2.41)
L03: -0.61 (-4.87, 3.85)
Reference: Bell et al. (2008, 091268)
Period of Study: 1999 ¦ 2005
Location: 202 US counties
Outcome (ICD-9): C0PD (490-492),
respiratory tract infections (464 ¦ 466,
480 ¦ 487)
Age Groups: 65 +
Study Design: Time series
N: NR
Statistical Analyses: Two-stage
Bayesian hierarchical model to find
national avg
First stage: Poisson regression (county-
specific)
Pollutant: PM2.6
Averaging Time: 24 h
Mean (/L/g|m3):
Descriptive information presented in
Figure S2 (boxplots):
IQR: 8.7 //g/m3
Monitoring Stations: NR
Copollutant (correlation): NR
PM Increment: 10 //g|m3
Percent increase [95% PI]: Respiratory
admissions: Lag 0 (all seasons): 0.22 [¦
0.12-0.56]
Lag 0 (winter, national): 1.05 [0.29—
1.82]
Lag 0 (winter, northeast): 1.76 [0.60-
2.93]
Lag 0 (winter, southeast): 0.59 [-1.35—
2.58]
Lag 0 (winter, northwest): -0.07 [-6.74-
7.08]
July 2009
E-317
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Covariates: day of the week,
temperature, dew point temperature,
temporal trends, indicator for persons
75+ years, population size
Season: All, June-August (Summer),
September-November (Fall), December-
February (Winter), March-May (Spring)
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: 0-2 day lags
Lag 0 (winter, southwest): 0.03 [-1.25—
1.34]
Lag 0 (spring, national): 0.31 [-0.47—
1.11]
Lag 0 (spring, northeast): 0.34 [-0.66—
1.34]
Lag 0 (spring, southeast): -0.06 [-1.77-
1.68]
Lag 0 (spring, northwest): -8.52 [-25.62-
12.51]
Lag 0 (spring, southwest): 1.87 [-2.00—
5.90]
Lag 0 (summer, national): -0.62 [-1.33—
0.09]
Lag 0 (summer, northeast): -0.8 [-1.65-
0.07]
Lag 0 (summer, southeast): -0.15 [-1.88—
I.61]
Lag 0 (summer, northwest): 0.25 [¦
21.46-27.96]
Lag 0 (summer, southwest): 0.64 [-5.38-
7.04]
Lag 0 (autumn, national): 0.02 [-0.63—
0.67]
Lag 0 (autumn, northeast): -0.01 [-0.87—
0.85]
Lag 0 (autumn, southeast): -0.58 [-2.06—
0.91]
Lag 0 (autumn, northwest): -1.38 [¦
II.84-10.32]
Lag 0 (autumn, southwest): 1.77 [-0.73—
4.33]
Lag 1 (all seasons): 0.05 [-0.29-0.39]
Lag 1 (winter): 0.50 [-0.27—1.27]
Lag 1 (spring): -0.24 [-1.01 -0.53]
Lag 1 (summer): 0.28 [-0.39-0.95]
Lag 1 (autumn): 0.15 [-0.49-0.79]
Lag 2 (all seasons): 0.41 [0.09-0.74]
Lag 2 (winter, national): 0.72 [0.01 —
1.43]
Lag 2 (winter, northeast): 0.79 [-0.21 —
1.80]
Lag 2 (winter, southeast): 0.4 [¦ 1.45,
2.27]
Lag 2 (winter, northwest): -0.06 [-6.52-
6.85]
Lag 2 (winter, southwest): 1.2 [-0.10-
2.52]
Lag 2 (spring, national): 0.35 [-0.29—
0.99]
Lag 2 (spring, northeast): 0.04 [-0.88—
0.97]
Lag 2 (spring, southeast): 0.75 [-0.82—
2.34]
Lag 2 (spring, northwest): 2.29 [-14.26—
22.03]
July 2009
E-318
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Lag 2 (spring, southwest): 1.05 [-2.18-
4.39]
Lag 2 (summer, national): 0.57 [-0.07—
1.23]
Lag 2 (summer, northeast): 0.77 [-0.01 —
1.56]
| Lag 2 (summer, southeast): -0.52 [¦
2.07-1.06]
Lag 2 (summer, northwest): 0.74 [¦
18.73-24.86]
Lag 2 (summer, southwest): 2.41 [-2.61 —
7.69]
Lag 2 (autumn, national): 0.39 [-0.22—
1.01]
Lag 2 (autumn, northeast): 0.12 [-0.82-
1.07]
Lag 2 (autumn, southeast): 0.14 [-1.29—
1.59]
Lag 2 (autumn, northwest): -0.74 [¦
10.08-9.58]
Lag 2 (autumn, southwest): 0.97[-1.36—
3.36]
Reference: Bell et al. (2009,191007)
Period of Study: 1999-2005
Location: 168 US Counties
Outcome: Respiratory hospital
admissions
Study Design: Retrospective Cohort
Covariates: socio-economic conditions,
long term temperature
Statistical Analysis: Bayesian
hierarchical model
Age Groups: £65
Pollutant: PM2.6
Averaging Time: 24h
Mean (SD) Unit: NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 20% of the population
acquiring air conditioning
Percent Change (95% CI) in
community-specific PM health effect
estimates for respiratory hospital
admissions
Any AC, including window units
Yearly health effect: 44.5 (-87.5-176)
Summer health effect: -74.8 (-417-267)
Winter health effect: -32.5 (-245-180)
Central AC
Yearly health effect: 27.6 (-46.7-102)
Summer health effect: -38.6 (-160-82.6)
Winter health effect: 43.8 (-125-213)
July 2009
E-319
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Bell et al. (2009,191007)
Period of Study: 1999-2005
Location: 168 US Counties
Outcome: Respiratory HA
Age Groups: 65 +
Study Design: time series
N: NR
Statistical Analyses: Bayesian
Hierarchical Regression
Covariates: time trend, day of week,
seasonality, dew point, temperature
Statistical Package: NR
Lags Considered: 0-2
Pollutant: PM2.6
Averaging Time: daily
Mean:
EC: 0.715
Ni: 0.002
V: 0.003
Min:
EC: 0.309
Ni: 0.003
V: 0.001
Max:
EC: 1.73
Ni: 0.021
V: 0.010
Interquartile Range:
EC: 0.245
Ni: 0.001
V: 0.001
Interquartile Range of Percents:
EC: 1.7
Ni: 0.01
V: 0.01
Monitoring Stations: NR
Copollutant: Al, NH4+, As, Ca, CI, Cu,
EC, 0MC, Fe, Pb, Mg. Ni, NOs-, K, Si,
Na+, S04-, Ti, V, Zn
Co-pollutant Correlation
Ni, V: 0.48
V, EC: 0.33
Ni, EC: 0.30
Note: Pollutant concentrations available
for all fractions of PM2.6
PM Increment: Interquartile Range in
the fraction of PM2.6
Percent Increase (Lower CI, Upper CI):
EC: 511 (80.7, 941), lag 0
EC + Ni: 399(45.1,843), lag 0
EC + V: 386 (-74.8, 846), lag 0
EC + Ni, V: 362 (-98.0, 823), lag 0
Ni: 223 (36.9,410), lag 0
Ni + EC: 176(-18.7,370), lag 0
Ni + V: 151 (-78.4, 381), lag 0
Ni + EC, V: 136 (-94.9, 368), lag 0
V: 392 (46.3, 738), lag 0
V	+ EC: 279 (-93.2, 651), lag 0
V+Ni: 230(-193.7,653), lag 0
V	+ EC, Ni: 140 (-300, 579), lag 0
EC:-1.5 (80.7, 941), lag 1
EC: 17.5 (-22.3, 57.3), lag 2
Ni:-7.2 (-66.6. 52.1), lag 1
Ni: -4.9 (-22.3,12.5), lag 2
V: -19.6 (-127, 88.3), lag 1
V: 10.5(-21.5,42.4), lag 2
HS education: -77.8 (-390, 234), lag 0
median income: 45.9 (-411, 503), lagO
Percent black: -53.1 (-557, 451), lag 0
Percent living in urban area: -41.9 (¦
774.7, 691), lag 0
Population: -22.9 (-121, 75.3), lag 0
Notes: Interquartile ranges in percent HS
education, median income, percent black,
percent living in urban area, and
population are 5.2 %, $9,223,17.3%,
11.0%, and 549,283 respectively.
July 2009
E-320
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Chardon et al. (2007,
091308)
Period of Study: 2000-2003
Location: Greater Paris Area, France
Doctors house calls
Outcome (ICPC2): Asthma (R96), Upper
respiratory disease (URD R07, R21, R29,
R75, R76, R02), Lower respiratory
disease (LRD, R05, R78)
Age Groups: all
Study Design: Time series
N: 8027 for asthma
52928 for LRD
74845 for URD
Statistical Analyses: Quasi-Poisson,
GAM, parametric penalized spline
smoothers.
Covariates: Lagged and current
temperature, humidity, long term trends,
seasonality, pollen counts, influenza
epidemic, days of the week, holidays,
bank holidays
Season: All
Dose-response Investigated? No
Statistical Package: R
Lags Considered: 0-3 days
Pollutant: PM2.6
Averaging Time: mean of the daily
means
Mean (SD): 14.7(7.34)//g/m3
Percentiles: 25th: 9.5
50th(Median): 12.9
75th: 18.2
Range (Min, Max): (3, 69.6)
Monitoring Stations: 1- 4
Copollutant: PMio: r - 0.95
NO?: r - 0.68
PM Increment: 10 /yglm3
% Change, lag 0-3 d avg
URD
6.0(3.1,9.1)
LRD
5.8 (2.8, 8.9)
Asthma
4.4(-1.3,10.4)
Reference: Chen et al. (2005, 087555) Hospital Admissions
Period of Study: Jun 1,1995-Mar 31,
1999
Location: Vancouver area, BC
Outcome (ICD-9): Acute respiratory
infections (460-466), upper respiratory
tract infections (470-478), pneumonia
and influenza (480-487), C0PD and allied
conditions (490-496), other respiratory
diseases (500-519)
Age Groups: >65 yrs
Study Design: Time series
N: 12,869
Statistical Analyses: GLM
Covariates: Temp and relative humidity
Season: NR
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 1, 2, 3, 4, 5, 6, and 7-
day avg
Pollutant: PM2.6
Averaging Time: 24 h
Mean (min-max):
7.7 (2.0-32.0)
SD - 3.7
Monitoring Stations: 13
Copollutant (correlation): PMio:
r - 0.83
PM 10-2.5: r - 0.38
C0H: r - 0.39
CO: r - 0.23
0a: r - -0.01
NO?: r - 0.36
SO?: r - 0.42
Other variables:
Mean temp: r - 0.41
Rel humidity: r - -0.23
PM Increment: 4.0 /yglm3 (IQR)
RR Estimate [CI]:
Adj for weather conditions
Overall admission
1-day	avg: 1.02 [0.99,1.05]
2-day	avg: 1.02 [0.99,1.06]
3-day	avg: 1.02 [0.98,1.05]
Adj for weather conditions and
copollutants
Overall admission
1-day	avg: 1.01 [0.98,1.06]
2-day	avg: 1.01 [0.98,1.05]
3-day	avg: 1.00 [0.96,1.04]
Notes: RR's were also provided for lags
4-7	in Table 3, yielding similar results
July 2009
E-321
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Chimonas and Gessner
(2007,093261)
Period of Study: January 1,1999-June
30, 2003
Location: Anchorage, Alaska
Outcome (ICD-9): Asthma (493.0-493.9)
Lower respiratory illness-LRI (466.1,
466.0, 480-487, 490,510-511)
Inhaled quick-relief medication
Steroid medication
Age Groups: < 20 years old
Study Design: Time series
N: 42,667 admissions
Statistical Analyses: GEE for
multivariable modeling
Covariates: Season, serial correlation,
year, weekend, temperature,
precipitation, and wind speed
Season: NR
Dose-response Investigated? No
Statistical Package: SPSS (dataset),
SAS (analysis)
Lags Considered: 1 day and 1 week
Pollutant: PM2.6
Averaging Time: 24-hs and 1 week
Mean (min-max):
Daily: 6.1 (0.5-69.8)
Weekly: 5.8 (1.8-45.0)
Monitoring Stations: NR
Copollutant: N/A
PM Increment: 5 /yglm3
RR Estimate [CI]:
Same Day
Outpatient Asthma: 0.992 [0.964,1.024]
Outpatient LRI: 0.952 [0.907,1.001]
Inpatient Asthma: 0.936 [0.798,1.098]
Inpatient LRI: 0.919(0.823,1.027]
Inhaled Steroid Prescriptions: 0.988
[0.902,1.083]
Quick-relief Medication: 0.962
[0.901,1.028]
Weekly (median increase)
Outpatient Asthma: 0.983 [0.935,1.038]
Outpatient LRI: 0.969 [0.874,1.075]
Inpatient Asthma: 0.754 [0.513.1.109]
Inpatient LRI: 0.943 [0.715,1.245]
Inhaled Steroid Prescriptions: 1.018
[0.883,1.175]
Quick-relief Medication: 0.978
[0.882,1.087]
July 2009
E-322
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Delfino et al. (2009,191994) Outcome: Respiratory hospital	Pollutant: PM2.6
admissions
Period of Study: 10/1/2003-	Averaging Time: Hourly
1111512003	Study Design: Time series
Mean (SD) Unit by county:
Location: Southern California	Statistical Analysis: Poisson regression
with GEE	Los Angeles
Age Groups: All	Before Fires: 272 (12-4'/^'m3
During Fires: 54.1 (21.0) //g/m3
After Fires: 15.9 (5.5)//g/m3
Orange
Before Fires: 23.2 (9.6) //g/m3
During Fires: 64.3 (26.5) //g/m3
After Fires: 15.5 (10.2)//g/m3
Riverside
Before Fires: 32.7 (14.7) //g/m3
During Fires: 42.1 (25.5) //g/m3
After Fires: 16.9 (10.2)//g/m3
San Bernadino
Before Fires: 35.7 (16.6) //g/m3
During Fires: 45.3 (28.7) //g/m3
After Fires: 18.5 (8.3)//g/m3
San Diego
Before Fires: 18.5 (6.7) //g/m3
During Fires: 76.1 (66.6)//g/m3
After Fires: 14.2 (7.2)//g/m3
Ventura
Before Fires: 18.4 (8.3) //g/m3
During Fires: 50.1 (50.5)//g/m3
After Fires: 12.9 (4.3)//g/m3
Copollutant (correlation): NR
July 2009
E-323
Increment: 10 //g/m3
Relative Rate (Min CI, Max CI)
All Respiratory, All Ages: All Periods:
1.009 (0.999-1.018)
Pre-Wildfire: 1.022 (1.004-1.040)
Wildfire: 1.028 (1.014-1.041), p - 0.639
Post-Wildfire: 0.999 (0.968-1.031), p -
0.198
All Respiratory, Ages 0-4: All Periods:
0.994 (0.967-1.021)
Pre-Wildfire: 0.982 (0.921-1.046)
Wildfire: 1.045 (1.010-1.082), p - 0.103
Post-Wildfire: 0.894 (0.807-0.991), p -
0.126
All Respiratory, Ages 5-19: All Periods:
1.014(0.983-1.046)
Pre-Wildfire: 1.026 (0.946-1.113)
Wildfire: 1.027 (0.984-1.076), p - 0.990
Post-Wildfire: 0.958 (0.852-1.077), p -
0.354
All Respiratory, Ages 20-64: All Periods:
1.015(1.002-1.029)
Pre-Wildfire: 1.036 (1.007-1.066)
Wildfire: 1.024 (1.005-1.044), p - 0.534
Post-Wildfire: 1.007 (0.960-1.056), p -
0.315
All Respiratory, Ages 65-99: All Periods:
1.009 (0.996-1.022)
Pre-Wildfire: 1.022 (0.994-1.050)
Wildfire: 1.030 (1.011-1.049), p - 0.649
Post-Wildfire: 1.024 (0.967-1.074), p -
0.932
Asthma, All Ages, Male and Female: All
Periods: 1.022 (1.001-1.042)
Pre-Wildfire: 0.998 (0.949-1.050)
Wildfire: 1.048 (1.021-1.076), p - 0.097
Post-Wildfire: 0.986 (0.910-1.068), p -
0.792
Asthma, All Ages, Male: All Periods:
1.010(0.980-1.040)
P re-Wildfire: 1.021 (0.944-1.106)
Wildfire: 1.031 (0.990-1.073), p - 0.848
Post-Wildfire: 1.063 (0.948-1.192), p -
0.553
Asthma, All Ages, Female: All Periods:
1.029 (1.001-1.058)
Pre-Wildfire: 0.979(0.913-1.050)
Wildfire: 1.059 (1.022-1.097), p - 0.056
Post-Wildfire: 0.928 (0.829-1.037), p -
0.412
Asthma, Ages 0-4, Males and Females:
All Periods: 0.996 (0.947-1.048)
Pre-Wildfire: 0.924 (0.824-1.035)
Wildfire: 1.083 (1.021-1.149), p - 0.017
Post-Wildfire: 0.924 (0.767-1.113), p -
O.eQBAFT - DO NOT CITE OR QUOTE
Asthma, Ages 0-4, Males: All Periods:
1.018(0.963-1.076)
Pre-Wildfire: 0.942 (0.815-1.089)

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Dominici et al. (2006,
088398)
Period of Study: 1999 ¦ 2002
Location: 204 US counties, located in:
Alabama, Alaska, Arizona, Arkansas,
California, Colorado, Connecticut,
Delaware, District of Columbia, Florida,
Georgia, Hawaii, Idaho, Illinois, Indiana,
Iowa, Kansas, Kentucky, Louisiana,
Maine, Maryland, Massachusetts,
Michigan, Minnesota, Mississippi,
Missouri, Nevada, New Hampshire, New
Jersey, New Mexico, New York, North
Carolina, Ohio, Oklahoma, Oregon,
Pennsylvania, Rhode Island, South
Carolina, Tennessee, Texas, Utah,
Virginia, Washington, West Virginia,
Wisconsin
Outcome (ICD-9: Daily counts of hospital
admissions for primary diagnosis of
chronic obstructive pulmonary disease
(490-492), and respiratory tract
infections (464-466, 480-487).
Age Groups: >65 years
Study Design: Time series
N: 11.5 million Medicare enrollees
Statistical Analyses: Bayesian 2-stage
hierarchical models.
First stage: Poisson regression (county-
specific)
Second stage: Bayesian hierarchical
models, to produce a national avg
estimate
Covariates: Day of the week,
seasonality, temperature, dew point
temperature, long-term trends
Season: NR
Dose-response Investigated: No
Statistical Package: R statistical
software version 2.2.0
Lags Considered: 0-2 days, avg of days
0-2
Pollutant: PM2.6
Averaging Time: 24 h
Mean U/g/m3) (IQR): 13.4 (11.3-15.2)
Monitoring Stations: NR
Copollutant (correlation): NR
Other variables: Median of pairwise
correlations among PM2.6 monitors within
the same county for 2000: r - 0.91 (IQR:
0.81-0.95)
PM Increment: 10 //gf'nv1 (Results in
figures
see notes)
Percent increase in risk [95% PI]: C0PD
(Lag 0): Age 65 +: 0.91 [0.18,1.64]
Age 65-74:0.42 [-0.64,1.48]
Age 75 + : 1.47 [0.54, 2.40]
Respiratory tract infection: Age 65+:
0.92 [0.41,1.43]
Age 65-74: 0.93 [0.04,1.82]
Age 75 + : 0.92 [0.32,1.53]
Annual reduction in admissions
attributable to a 10 //gf'nv1 reduction in
daily PM2.6 level (95% PI):
Cerebrovascular disease: Annual number
of admissions: 226,641
Annual reduction in admissions: 1836
[680, 2992]
C0PD: Annual number of admissions:
108,812
Annual reduction in admissions: 990
[196,1785]
Respiratory tract infections: Annual
number of admissions: 226,620
Annual reduction in admissions: 2085
[929, 3241]
Reference: Dominici et al. (2006,
088398)
Period of Study: 1999-2002
Location: U.S. (mainland)
Outcome (ICD-9): Respiratory tract
infections (464-466,480-487) and
Chronic Obstructive Pulmonary Disease
(490-492)
Age Groups: All >65 yrs
65-74 yrs
>75 yrs
Study Design: Time series
N: 11.5 million at-risk
Statistical Analyses: Bayesian 2-stage
hierarchical models (day-to-day variation),
Poisson regression (county-specific RRs)
Covariates: Calendar time (seasonality
and year), temperature, dew point
Season: NR
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0,1, 2 days
Pollutant: PM2.6
Averaging Time: daily or every 3 days
(depending on county)
Mean: 13.4 (IQR: 11.3-15.2)
Monitoring Stations: NR (used data
from Air Quality System database)
Copollutant: NR
PM Increment: 10/yg/m3
Percentage Change in Hospital
Admission Rates [PI]:
COPD-Same day
All >65: 0.91 [0.18,1.64]
65-74 yrs: 0.42 [-0.64,1.48]
>75:1.47 [0.54,2.40]
Respiratory Tract lnfections-2-day lag
All >65: 0.92(0.41,1.43]
65-74 yrs: 0.93 [0.04,1.82]
>75:0.92 [0.32,1.53]
Notes: Other lag data shown in Fig 2-4
July 2009
E-324
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Erbas et al. (2005)
Period of Study: Jul 1,1989-Dec 31,
1992
Location: Melbourne, Australia
Outcome (ICD): COPD (490-492, 494,
496)
Asthma (493)
Age Groups: NR
Study Design: Time series
N: NR
Statistical Analyses: GLM, GAM,
Parameter Driven Poisson Regression,
Transitional Regression, Seasonal-Trend
decomposition based on Loess smoothing
for seasonal adjustment
Covariates: Secular trends, seasonality,
relative humidity, dry bulb temp, dew
point temp
Season: NR
Dose-response Investigated? Yes
Statistical Package: S-Plus, SAS
Lags Considered: 0-5 days
Pollutant: PMo.i i (API)
Averaging Time: 24-hs
Mean (min-max): NR
Monitoring Stations: 9
Copollutant (correlation): NR
PM Increment: Increase from the 10th-
90th percentile (value NR)
RR Estimate [CI]:
COPD
GAM:
0.95[0.91,1.00]
GLM, PDM, TRM: NR
Asthma
NR
Notes: This study was used to
demonstrate that conclusions are highly
dependent on the type of model used
Reference: Fung et al. (2006, 089789)
Period of Study: 6/1/95-3/31/99
Location: Vancouver, Canada
Hospital Admission/ED:
Hospital Admission
Outcome: Respiratory diseases (460-
519)
Age Groups: Age >65
Study Design: Time series, case
crossover
N: 26,275 individuals admitted
Statistical Analyses: Poisson regression
(spline 12 knots), case-crossover
(controls +/7 d days from case date),
Dewanji and Moolgavkar (DM) method
Covariates: Long-term trends, day-of-
the-week effect, weather
Season: All year
Dose-response Investigated? No
Statistical Package: SPIus, R
Lags Considered: 0-7 d
Pollutant: PM2.6
Averaging Time: 24-h Avg
Mean (SD): 7.72(3.61)
Range (Min, Max): (2, 32)
Monitoring Stations: NR
Copollutant (correlation): PM2 e:
PM10
r - 0.80
PMio-2.5
r - 0.34
CO
r - 0.23
CoH
r - 0.38
0s
r - -0.03
NO?
r - 0.36
SO?
r - 0.42
PM Increment::
4/yg/m3
RR Estimate (65+ years)
DM method:
1.007[0.994,1.020]
Current
1.007(0.990,1.023]
3 day
0.995(0.979,1.012]
5 day
0.995(0.971,1.020]
7 day
Time series:
1.003(0.989,1.018]
Current
1.000(0.982,1.018]
3 day
0.993(0.972,1.014]
5 day
0.995(0.971,1.020]
7 day
Case-crossover:
1.002(0.986,1.019]
Current
1.001(0.981,1.021]
3 day
0.988(0.966,1.011]
5 day
0.984(0.959,1.010]
7 day
July 2009
E-325
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Hinwood et al. (2006,
088976)
Period of Study: 1/1992-12/1998
Location: Perth, Australia
Hospital Admission
Outcome (ICD-9): COPD (490-496.99,
except asthma), pneumonia /influenza
(480-489.99), asthma
Age Groups: All ages
Study Design: Time stratified case-
crossover
N: NR
Statistical Analyses: Conditional
logistic regression
Covariates: Time trend, season,
temperature, humidity, day of wk,
holidays
Season: All year
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-3 days
Pollutant: PM2.6
Averaging Time: 24-h Avg
Mean (SD): 9.2 (4.3)
Percentiles:
10th: 5.0
90th: 14.5
Monitoring Stations: 13
Notes: Copollutant: NR
Increment: 1 //gf'nr1
Notes: Odds ratio for PM2.6 and all
respiratory, COPD, pneumonia and
asthma. Authors found an elevation in the
odds ratio for lags 2 and 3 reaaching
significance in all age groups for lag 3.
For each increase of 1 /yglm3, the number
of hospitalizations increases 0.2% for
respiratory disease, 0.5% for pneumonia
and 0.3% for asthna. PM2.6
concentrations were also significantly
associated with asthma for those aged
under 15 years with an estimated 0.5%
increase in hospitalizations.
July 2009
E-326
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Hirshon et al. (2008,
180375)
Period of Study: 612002-1112002
Location: Baltimore, Maryland
Outcome: Hospital admissions for
asthma
Study Design: Time-series
Covariates: Spatial distance from
pollution monitor, demographic variation,
long term, seasonal and daily trends,
weather and other pollutants
Statistical Analysis: Overdispersed
Poisson regression
Age Groups: 0-17 years
Pollutant: PM2.6 zinc
Averaging Time: 24h
Mean (SD) Unit: 22.42 (25.14)//gfm3
Range (Min, Max): NR
Copollutant (correlation):
Ni: 0.41
Cr: 0.17
Fe: 0.54
Sulfate: 0.01
CO: 0.40
PM2.6: 0.39
0s: 0.01
NO?: 0.66
Elemental Carbon: 0.48
Increment: NR
Relative Risk (95% CI), Best fit Model
Medium - 8.63-20.76 ng/m3
High - > 20.76 nglm3
No Lag
Medium: 1.12(0.98-1.28)
High: 1.09 (0.91-1.30)
1-day	Lag
Medium: 1.23(1.07-1.41)
High: 1.16 (0.97-1.39)
2-day	Lag
Medium: 1.11 (0.94-1.30)
High: 1.15 (0.96-1.38)
Controlling for Time Trends
No Lag
Medium: 1.08 (0.95-1.23)
High: 0.98 (0.86-1.11)
1-day	Lag
Medium: 1.13(1.003-1.28)
High: 1.03 (0.91-1.16)
2-day	Lag
Medium: 1.13 ()
High: 0.98-1.31
Controlling for Time Trends and
Additional Copollutants
No Lag
Medium: 1.12(0.98-1.29)
High: 1.09 (1.01-1.30)
1-day	Lag
Medium: 1.20(1.04-1.38)
High: 1.12 (0.93-1.35)
2-day	Lag
Medium: 1.12(0.95-1.32)
High: 1.19 (0.98-1.44)
July 2009
E-327
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Host et al. (2007,155851)
Period of Study: 2000 - 2003
Location: Six French cities: Le Havre,
Lille, Marseille, Paris, Rouen, and
Toulouse
Outcome (ICD-10): Daily hospitalizations
for all respiratory diseases (J00-J99),
respiratory infections (J10-J22).
Age Groups: For all respiratory
0-14 years, 15-64 years, and a 65
years.
For respiratory infections: All ages
Study Design: Time series
N: NR (Total population of cities:
approximately 10 million)
Statistical Analyses: Poisson regression
Covariates: Seasons, days of the week,
holidays, influenza epidemics, pollen
counts, temperature, and temporal trends
Season: NR
Dose-response Investigated: No
Statistical Package: MGCV package in
R software (R 2.1.1)
Lags Considered: Avg of 0-1 days
Pollutant: PM2.B
Averaging Time: 24 h
Mean (5th -95th percentile): Le Havre:
13.8(6.0-30.5)
Lille: 15.9(6.9-26.3)
Marseille: 18.8(8.0-33.0)
Paris: 14.7 (6.5-28.8)
Rouen: 14.4 (7.5-28.0)
Toulouse: 13.8 (6.0-25.0)
Monitoring Stations: 13 total: 1 in
Toulouse
4 in Paris
2 each in other cities
Copollutant (correlation): PMio-2.b:
Overall: r > 0.6
Ranged between r - 0.28 and
r - 0.73 across the six cities.
PM Increment: 10 //gf'nv1 increase, and a
27 /yglm3 increase (corresponding to the
difference between the lowest of the 5th
percentiles and the highest of the 95th
percentiles of the cities' distributions)
ERR (excess relative risk) Estimate [CI]:
For all respiratory diseases (27 //g|m3
increase): 0-14 years: 1.1% [-3.1, 5.5]
15-64 years: 2.2% [-1.8, 6.4];
>65 years: 1.3% [-5.3,8.2]
For respiratory infections (10 //g|m3
increase): All ages: 2.5% [0.1, 4.8]
For respiratory infections (27 //g|m3
increase): All ages: 7.0% [0.7,13.6]
Reference: Ko et al. (2007, 091639)
Period of Study: 1/2000-12/2004
Location: Hong Kong, China
ED Visits
Outcome (ICD-9): C0PD: Chronic
bronchitis (491), Emphysema (492),
Chronic airway obstruction (496)
Age Groups: All ages
Study Design: Time series
N: 15 hospitals, 119,225 admissions
Statistical Analyses: Poisson
regression, GAM with stringent
convergence criteria, APHEA2 protocol.
Covariates: Time trend, season,
temperature, humidity, other cyclical
factors, day, day of wk, holidays
Season: All year, interactions with
season tested
Dose-response Investigated? No
Statistical Package: SPLUS 4.0
Lags Considered: 0-5 days
Pollutant: PM2.6
Averaging Time: 24-h
Mean (SD): 35.7 (20.6)
Percentiles:
25th: 19.4
50th(Median): 31.7
75th: 46.7
Range (Min, Max): (6.0,163.2)
Monitoring Stations: 14
Copollutant (correlation): PM2 e:
PM10
r - 0.952
NO2
r - 0.441
0s
r - 0.394
SO2
r - 0.282
PM Increment: PM10
RR Estimate
C0PD
1.002(0.998,1.001]
lag 0
1.003(0.999,1.007]
lag 1
1.011(1.007,1.014]
lag 2
1.013[1.010,1.017]
lag 3
1.011[1.008,1.015]
lag 4
1.009(1.006,1.013]
lag 5
1.004(0.999,1.008]lag 0-1
1.010(1.006,1.015]lag 0-2
1.018(1.013,1.022]lag 0-3
1.024(1.019,1.029]lag 0-4
1.031(1.026,1.036]lag 0-5
4-Pollutant model:
1.014(1.007,1.022]
lag 0-5
3-Pollutant model:
1.011(1.004,1.017]
lag 0-5
July 2009
E-328
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ko et al. (2007, 092844)
Period of Study: 1/2000-12/2005
Location: Hong Kong, China
Hospital Admission
Outcome (ICD-9): Asthma (493)
Age Groups: All, 0-14,15-56, 65 +
Study Design: Time series
N: 69,716 admissions, 15 hospitals
Statistical Analyses: Poisson
regression, with GAM with stringent
convergence criteria.
Covariates: Time trend, season,
temperature, humidity, other cyclical
factors
Season: All year, evaluated effect of
season in analysis
Dose-response Investigated? No
Statistical Package: SPLUS 4.0
Lags Considered: 0-5 days
Pollutant: PM2.6
Averaging Time: 24-h
Mean (SD): 36.4 (21.1)
Percentiles:
25th: 20.0
50th(Median): 32.5
75th: 47.7
Range (Min, Max): (6,163)
Monitoring Stations: 14
Copollutant (correlation): PM2 e:
PM10
r - 0.956
NO?
r - 0.774
0s
r - 0.585
SO?
r - 0.482
PM Increment: 10.0 /yglm3
RR Estimate
Asthma (Single-pollutant model):
1.008(1.004,1.013]
lag 0
1.004(1.000,1.009]
lag 1
1.004(1.000,1.009]
lag 2
1.009(1.005,1.014]
lag 3
1.006(1.001,1.011]
lag 4
1.002(0.998,1.007]
lag 5
1.009(1.004,1.014]
lag 0-1
1.012(1.007,1.018]
lag 0-2
1.017(1.011,1.022]
lag 0-3
1.020(1.014,1.026]
lag 0-4
1.021(1.015,1.028]
lag 0-5
Asthma in Age
0-14: 1.024(1.013,1.034]
lag 0-5
14-65: 1.018(1.008,1.029]
lag 0-5
>65:1.021(1.012,1.030]
lag 0-4
Asthma-Cold Season: 1.139(1.043,
1.244] lag 0-5
July 2009
E-329
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lee et al. (2006, 090176)
Period of Study: 1/1997-12/2002
Location: Hong Kong, China
Hospital Admission
Outcome: Asthma (493)
Age Groups: < 18 years
Study Design: Time series
N: 26,663 asthma admissions for asthma
and 5821 admissions for influenza
Statistical Analyses: Poisson
regression, GAM
Covariates: Temperature, atmospheric
pressure, relative humidity
Season: All
Dose-response Investigated? No
Statistical Package: SAS 8.02
Lags Considered: 0-5
Notes: Controls were admissions for
influenza ICD9 487
Pollutant: PM2.6
Averaging Time: 24-hs
Mean (SD): 45.3/yg/m3, (16.2)
Percentiles: 25th: 33.4
50th(Median): 43.0
75th: 54.0
Range (Min, Max): NR
Monitoring Stations: 10
Copollutant (correlation):
PM2.B-PM10: 0.89
PM2.B-SO2: 0.48
PM2.B-PJO2: 0.74
PM2.B-O3: 0.47
PM Increment: IQr - 20.6/yglm3
Percent increase:
Single pollutant model:
5.10[2.95,7.30], lag 0
5.00[2.88,7.16], lag 1
5.48(2.75, 6.95], lag 2
4.83(2.78, 6.93], lag 3
6.59(4.51,8.72], lag 4
5.24(3.18,7.34 ], lag 5
Multipollutant model (SO2, NO2, CO, O3)
3.24(0.93, 5.60], lag 4
Reference: Letz and Quinn (2005,
088752)
Period of Study: Oct 1, 2001-Aug 24,
2002
Location: San Antonio, Texas
Emergency Dept Visits
Outcome (ICD-9): Asthma or reactive
airway disease (493.0-493.9), wheezing
(786.07), dyspnea (786.01-786.9),
shortness of breath (786.05), bronchitis
(490-496), or cough (786.2)
Age Groups: NR (basic air force trainees)
Study Design: Historic (retrospective)
cohort
N: 149 ED visits
Statistical Analyses: Pearson
correlation
Covariates: NR
Season: NR
Dose-response Investigated? No
Statistical Package: SPSS
Lags Considered: NR
Pollutant: PM2.6
Averaging Time: 24-h AQI
AQI Range (min-max): (
4-109)
Monitoring Stations: Data obtained
from the Texas Commission on
Environmental Quality
Copollutant (correlation): NR
PM Increment: NR
Correlation with Outcomes:
Same-day
All visits: r - 0.082
Proven asthmatic events: r - -0.042
3-day
All visits: r - 0.097
Proven asthmatic events: r - 0.011
Reference: Lin et al. (2005, 087828)
Period of Study: 1998-2001
Location: Toronto, North York, East
York, Etobicoke, Scarborough, and York
(Canada)
Hospital Admissions
Outcome (ICD-9): Respiratory infections
including laryngitis, tracheitis, bronchitis,
bronchiolitis, pneumonia, and influenza
(464, 466, 480-487)
Age Groups: 0-14 yrs
Study Design: Bidirectional case-
crossover
N: 6782 respiratory infection
hospitalizations
Statistical Analyses: Conditional
logistic regression (Cox proportional
hazards model)
Covariates: Daily mean temp and dew
point temp
Season: NR
Dose-response Investigated? No
Statistical Package: SAS 8.2 PHREG
procedure
Lags Considered: 1-7 day averages
Pollutant: PM2.6
Averaging Time: 24 h
Mean (min-max):
9.59 (0.25-50.50)
SD - 7.06
Monitoring Stations: 4
Copollutant (correlation): PMio 2.5:
r - 0.33
PM10: r - 0.87
CO: r - 0.10
SO2: r - 0.47
NO2: r - 0.48
O3: r - 0.56
PM Increment: 7.8/yglm3
OR Estimate [CI]:
Adjusted for weather
4 day avg: 1.11 [1.02,1.22]
6 day avg: 1.11 [1.00,1.24]
Adj for weather and other gaseous
pollutants
4 day avg: 0.94 [0.81,1.08]
6 day avg: 0.90 [0.76,1.07]
Notes: OR's were also categorized into
"Boys" and "Girls," yielding similar results
July 2009
E-330
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lin et al.
(2002, 026067)
Period of Study: Jan 1,1981 -Dec 31,
1993
Location: Toronto
Hospital Admissions
Outcome (ICD-9): Asthma (493)
Age Groups: 6-12 yrs
Study Design: Uni- and bi directional
case-crossover (UCC, BCC) and time-
series (TS)
N: 7,319 asthma admissions
Statistical Analyses: Conditional
logistic regression, GAM
Covariates: Maximum and minimum
temp, avg relative humidity
Season: Apr-Sep, Oct-Mar
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1-7 day averages
Pollutant: PM2.6
Averaging Time: 6 days (predicted daily
values)
Mean (min-max):
17.99 (1.22-89.59)
SD - 8.49
Monitoring Stations: 1
Copollutant (correlation):
PMio: r - 0.87
PM 10-2.5: r - 0.44
CO: r - 0.45
SO?: r - 0.46
NO?: r - 0.50
O3: r - 0.21
PM Increment: 9.3 /yglm3
RR Estimate [CI]:
Adj for weather and gaseous pollutants
BCC 5 day avg: 0.94 [0.85,1.03]
BCC 6 day avg: 0.92 [0.83,1.02]
TS 5 day avg: 0.96 [0.90,1.02]
TS 6 day avg: 0.94 [0.88,1.01]
Boys-adj for weather
UCC 1 day avg: 1.09(1.04,1.15]
UCC 2 day avg: 1.09(1.02,1.16]
BCC 1 day avg: 1.01 [0.97,1.06]
BCC 2 day avg: 0.99 [0.93,1.05]
TS 1 day avg: 1.00 [0.97,1.04]
TS 2 day avg: 0.98 [0.94,1.02]
Girls-adj for weather
UCC 1 day avg: 1.06 [0.99,1.14]
UCC 2 day avg: 1.11 [1.02,1.21]
BCC 1 day avg: 0.99 [0.93,1.06]
BCC 2 day avg: 1.02 [0.94,1.09]
TS 1 day avg: 0.99 [0.95,1.04]
TS 2 day avg: 1.00 [0.95,1.06]
Notes: The author also provides RR using
UCC, BCC, and TS analysis for female
and male groups for days 3-7, yielding
similar results
Reference: Magas et al. (2007, 090714) Hospital Admission/ED: Admissions
Period of Study: 2001-2003
Location: Oklahoma City Metro area,
Oklahoma and Cleveland counties
Outcome: Asthma 493.01-493.99
Age Groups: < 15 yrs
Study Design: Time series
N: 1,270 admissions
Pollutant: PM2.6
Averaging Time: 24 h avg
Mean (SD): NR
Range (Min, Max): NR
Monitoring Stations: 10
Statistical Analyses:
regression
Covariates: Temperature, humidity,
pollen count, mold
Season: All
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1
ive binomial Copollutant (correlation): NR
Notes: Coefficient for PM2.6 was not
significant and thus not reported.
July 2009
E-331
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Mohr et al. (2008,180215) Outcome: Asthma ER Visits
Period of Study: Jun 2001 - May 2003 Age Groups: 2-17 yrs
Location: St. Louis, M0	Study Design: Time series
Statistical Analyses: GEE Poisson
models
Covariates: season, weekend exposure,
ns
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: 1d
Pollutant: PM2.6EC
Averaging Time: 24 h
Std Dev: 0.1
Monitoring Stations: 1
Copollutant: NO., SO2, O3
Co-pollutant Correlation
NO,: 0.68*
SO2: 0.09
O3: -0 06
*pn0.05
PM Increment: 0.1 /yglm3
Relative Risk Effect (Lower CI, Upper
CI):
Weekend Exposure
Summer: 1.05 (1.00,1.11)
Fall: 0.99 (0.97,1.01)
Winter: 0.96(0.92,1.00)
Spring: 0.96 (0.92,1.00)
Weekday Exposure
Summer: 1.01 (0.98,1.03)
Fall: 1.00 (0.99,1.01)
Winter: 0.99(0.96,1.01)
Spring: 0.98 (0.96,1.01)
Reference: Neuberger et al. (2004,
093249)
Period of Study: 1999-2000 (1 yr
period)
Location: Vienna and Lower Austria
Hospital Admissions
Outcome (ICD-9): Bronchitis,
emphysema, asthma, bronchiectasis,
extrinsic allergic alveolitis, and chronic
airway obstruction (490-496)
Age Groups: 3.0-5.9 yrs
7-10 yrs
65 +
Study Design: Time series
N: 366 days (admissions NR)
Statistical Analyses: GAM
Covariates: SO2, NO, NO2, O3,
temperature, humidity, and day of the
week
Season: NR
Dose-response Investigated? Yes
Statistical Package: S-Plus 2000
Lags Considered: 0-14 days
Pollutant: PM2.6
Averaging Time: 24 h
Maximum daily mean:
Vienna: 96.4
Rural area: 48.0
Monitoring Stations: NR
Copollutant (correlation): NR
PM Increment: 10/yg/m3
Log Relative Rate Estimate (p-value):
Vienna
Male: 2 day lag - 5.467(0.019)
Female: 3 day lag - 5.596 (0.009)
Rural
Male: 10 day lag - 9.893 (0.012)
Female: 11 day lag - 10.529 (0.011)
Association with tidal lung functioN:
P - -0.987 (p-value - 0.091)
Notes: Effect parameters with significant
coefficients for respiratory health
included: male sex, allergy, asthma in
family, and traffic for Vienna and age,
allergy, asthma in family, passive
smoking, and PM fraction for the rural
area. Effect parameters with significant
coefficients for log asthma score were
allergy, asthma in family, and rain for
Vienna and allergy, asthma in family, and
passive smoking for the rural area. Cross-
correlation coefficients are provided in Fig
1.
July 2009
E-332
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ostro et al. (2008, 097971)
Period of Study: 2000-2003
Location: Six California Counties
Outcome: Respiratory disease (ICD-9
460-519)
Study Design: Time-Series
Statistical Analysis: Poisson
Regression
Statistical Package: R
Age Groups: Children < 19
Pollutant: PM2.6 and components
Averaging Time: 24h
Mean (SD) Unit: 19.4//gfm3
IQR: 14.6 //g/m3
Copollutants:
EC, 0C, NO?, S04, Cu, Fe, K, Si, Zn
Increment: NR
Relative Risk (Min CI, Max CI)
Lag
Full results are presented graphically in
Figures 1 and 2.
Excess risks for all-year respiratory
hospital admissions in children < 19yrs,
3d lag
PM2.6: 4.1% (1.8-6.4)
EC: 5.4% (0.8-10.3)
Fe: 4.7% (2.2-7.2)
0C: 3.4% (1.1-5.7)
Nitrates: 3.3% (1.1-5.5)
Sulfates: 3.0% (0.4-5.7)
Excess risks for cool season (Oct - Mar)
respiratory hospital admissions in children
< 19yrs, 3d lag
PM2.6: 5.1% (1.6-8.9)
EC: 6.8% (-0.2-14.2)
Fe: 4.8% (1.7-8.0)
K: 4.0% (0.3-7.7)
July 2009
E-333
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Slaughter et al. (2005,
073854)
Period of Study: January 1995 through
June 2001
Location: Spokane, WA
Hospital Admissions and ED visits
Outcome: All respiratory (460-519)
Asthma (493)
C0PD (491,492, 494,496)
Pneumonia (480-487)
Acute URI not including colds and
sinusitis (464, 466,490)
Age Groups: All, 15+ years for C0PD
Study Design: Time series
N: 2373 visit records
Statistical Analyses: Poisson
regression, GLM with natural splines. For
comparison also used GAM with
smoothing splines and default
convergence criteria.
Covariates: Season, temperature,
relative humidity, day of week
Season: All
Dose-response Investigated?: No
Statistical Package: SAS, SPLUS
Lags Considered: 1 -3 d
Pollutant: PM2.6
Averaging Time: 24 h avg
Range (90% of Concentrations):
4.2-20.2 /yglm3
Monitoring Stations:
One
Notes: Copollutant (correlation):
PMir - 0.95
PMior - 0.62
PMio-2.5 r - 0.31
CO r - 0.62
Temperature r - 0.21
PM Increment: 10 /yglm3
RR Estimate [Lower CI, Upper CI]
lag:
ER visits:
PM2.6
All Respiratory
PM2.6 Lag 1: 1.01 [0.98,1.04]
Lag 2: 1.02 [0.99,1.04]
Lag 3: 1.02 [0.99,1.05]
Acute Asthma
Lag 1: 1.03 [0.98,1.09]
Lag 2: 1.00 [0.95,1.05]
Lag 3: 1.01 [0.96,1.06]
C0PD (adult)
Lag 1: 0.96 [0.89,1.04]
Lag 2: 1.01 [0.93,1.09]
Lag 3: 1.00 [0.93,1.08]
Hospital Admissions:
PM2.6
All Respiratory
Lag 1: 0.98[0.94,1.01]
Lag 2: 0.99 [0.96,1.03]
Lag 3: 1.01 [0.98,1.05]
Asthma
Lag 1: 1.01 [0.91, 1.11]
Lag 2: 1.03 [0.94,1.13]
Lag 3: 1.02 [0.93,1.13]
C0PD (adult)
Lag 1: 0.99 [0.91,1.08]
Lag 2: 1.06 [0.98,1.16]
Lag 3: 1.03 [0.94,1.12]
July 2009
E-334
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Tecer et al. (2008,180030) Outcome: ED visits for respiratory
problems (ICD-9 470-478, 493)
Period of Study: 1212004-101
Study Design: Bidirectional Case-
Location: Zonguldak, T urkey	crossover
Covariates: Daily meteorological
parameters
Statistical Analysis: Conditional logistic
regression
Statistical Package: Stata
Age Groups: 0-14 years
Pollutant: PM2.6
Increment: 10 //g/m3
Averaging Time: NR
Odds Ratio (95% CI)
Mean, Unit: 29.1 /yglm3
Asthma
Range (Min, Max): 4.55, 95.65
Lag 0:1.15 (0.99-1.34)
Copollutant (correlation):
Lag 1:0.85 (0.70-1.03)
PM2.6/PM10
Lag 2: 0.87 (0.73-1.04)
Mean: 0.56
Lag 3: 0.93 (0.79-1.10)
Range: 0.17-0.88
Lag 4:1.25 (1.05-1.50)
PM2.6/PM10-2.6
Allergic Rhinitis with Asthma
Mean: 1.49
Lag 0:1.21 (1.10-1.33)
Range: 0.21-7.53
Lag 1:0.84 (0.75-0.93)

Lag 2: 0.89 (0.81-0.98)

Lag 3: 0.99 (0.90-1.09)

Lag 4:1.06 (0.95-1.19)

Allergic Rhinitis

Lag 0:1.08 (0.98-1.20)

Lag 1:1.03 (0.93-1.13)

Lag 2: 0.89 (0.80-0.99)

Lag 3: 0.98 (0.89-1.09)

Lag 4:1.18 (1.00-1.24)

Upper Respiratory Disease

Lag 0: 0.99 (0.49-2.00)

Lag 1:0.52 (0.22-1.20)

Lag 2:1.29 (0.75-2.22)

Lag 3:1.29 (0.69-2.43)

Lag 4:1.47 (0.87-2.50)

Lower Respiratory Disease

Lag 0:1.06 (0.78-1.44)

Lag 1:0.85 (0.59-1.22)

Lag 2:1.08 (0.72-1.61)

Lag 3:1.18 (0.92-1.52)

Lag 4: 0.72 (0.54-0.96)h
July 2009
E-335
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Tolbert et al. (2007,
090316)
Period of Study: August 1998-
December 2004
Location: Atlanta Metropolitan area,
Georgia
Outcome (ICD-9):
Combined RD group, including:
Asthma (493, 786.07, 786.09), C0PD
(491, 492, 496), URI (460-465,460.0,
477), pneumonia (480-486), and
bronchiolitis (466.1, 466.11, and
466.19))
Age Groups: All
Study Design: Time series
N:NR for 1998-2004.
For 1993-2004:10,234,490 ER visits
(283,360 and 1,072,429 visits included
in the CVD and RD groups, respectively)
Statistical Analyses: Poisson
generalized linear models
Covariates: long-term temporal trends,
season (for RD outcome), temperature,
dew point, days of week, federal
holidays, hospital entry and exit
Season: All
Dose-response Investigated: No
Statistical Package: SAS version 9.1
Lags Considered: 3-day moving
0-2)
July 2009
Pollutant: PM2.6
Averaging Time: 24 h
Mean (median
IQR, range, 10th—90th percentiles):
PMz.b: 17.1 (15.6
11.0-21.9
0.8-65.8
7.9-28.8)
PM2.6 sulfate: 4.9 (3.9
2.4-6.2
0.5-21.9
1.7-9.5)
PM2.6 organic carbon: 4.4 (3.8
2.7-5.3
0.4-25.9
2.1-7.2)
PM2.6 elemental carbon: 1.6 (1.3
0.9-2.0
0.1-11.9
0.6-3.0)
PM2.6 water-soluble metals: 0.030 (0.023
0.014-0.039
0.003-0.202
0.009-0.059)
Monitoring Stations: 1
Copollutant (correlation): Between
PM2.6 and:
PM10: r - 0.84
O3: r - 0.62
NO2: r - 0.47
CO: r - 0.47
SO2: r - 0.17
PMio-2.5: r - 0.47;
PM2.6 SO4: r - 0.76;
PM2.6 EC: r - 0.65;
PM2.6 0C: r - 0.70;
PM2.6 TC: r - 0.71;
PM2.6 water-sol metals:
r - 0.69
0HC: r - 0.50
Between PM2.6 SO4 and: PM10: r - 0.69
O3: r - 0.56
NO2: r - 0.14
CO: r - 0.14
SO2: r - 0.09
PMio-2.5: r - 0.32;
PM2.6: r - 0.76;
PM2.6 EC: r - 0.32;
;'3?'fel2.B 0C: r - 0.33;
PM2.6 TC: r - 0.34;
PM2.6 water-sol metals:
PM Increment:
PM2.6:10.96/yglm3 (IQR)
PM2.6 sulfate: 3.82/yg/m3 (IQR)
PM2.6 total carbon: 3.63//g|m3 (IQR)
PM2.6 organic carbon: 2.61 //g|m3 (IQR)
PM2.6 elemental carbon: 1.15 /yg/m3 (IQR)
PM2.6 water-soluble metals: 0.03 /yglm3
(IQR)
Risk ratio [95% CI] (single pollutant
models):
PM2.6:
RD: 1.005[0.995-1.015]
PM2.6 sulfate:
RD: 1.007 [0.996-1.018]
PM2.6 total carbon:
RD: 1.001 [0.993-1.008]
PM2.E organic carbon:
RD: 1.003(0.995-1.011]
PM2.E elemental carbon:
RD: 0.996(0.989-1.004]
PM2.E water-soluble metals:
RD: 1.005 [0.995-1.015]
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Wong et al. (2006, 093266)
Period of Study: 2000-2002
Location: Hong Kong (8 districts)
General Practitioner Visits
Outcome (ICPC-2): Respiratory
diseasesfsymptoms: upper respiratory
tract infections (URTI), lower respiratory
infections, influenza, asthma, COPD,
allergic rhinitis, cough, and other
respiratory diseases
Age Groups: All ages
Study Design: Time series
N: 269,579 visits
Statistical Analyses: GAM, Poisson
regression
Covariates: Season, day of the week,
climate
Season: NR
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0-3 days
Pollutant: PM2.6
Averaging Time: 24 h
Mean (min-max):
35.7(9-120)
SD - 16.7
Monitoring Stations: 1 per district
Copollutant (correlation):
PMio: r - 0.94
PM Increment: 10/yg/m3
RR Estimate [CI]:
Overall URTI
1.021 [1.010,1.032]
Notes: RRs are also reported for each
individual general practitioner yielding
similar results
Reference: Yang Q et al. (2004,
087488)
Period of Study: Jun 1,1995-Mar 31,
1999
Location: Vancouver area, British
Columbia
Hospital Admissions
Outcome (ICD-9): Respiratory (
(460-519), pneumonia only (480-486),
asthma only (493)
Age Groups: 0-3 yrs
Study Design: Case control, bidirectional
case-crossover (BCC), and time series
(TS)
N: 1610 cases
Statistical Analyses: Chi-square test,
Logistic regression, GAM (time-series),
GLM with parametric natural cubic
splines
Covariates: Gender, socioeconomic
status, weekday, season, study year,
influenza epidemic month
Season: Spring, summer, fall, winter
Dose-response Investigated? No
Statistical Package: SAS (Case control
and BCC), S-Plus (TS)
Pollutant: PM2.6
Averaging Time: 24 h
Mean (min-max):
7.7 (2.0-32.0)
SD - 3.7
Monitoring Stations: NR (data obtained
from Greater Vancouver Regional District
Air Quality Dept)
Copollutant (correlation):
PMio: r = 0.83
PM 10-2.5: r - 0.39
CO: r - 0.24
O3: r = -0.03
NO?: r - 0.37
SO2: r - 0.43
PM Increment: 4.0 /yglm3 (IQR)
OR Estimate [CI]:
Values NR
Notes: Author states that no significant
association was found between PM2.6
and respiratory disease hospitalizations.
Lags Considered: 0-7 days
July 2009
E-337
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Zanobetti and Schwartz
(2006,090195)
Period of Study: 1995-1999
Location: Boston, MA
Hospital Admission/ED:
Outcome: Pneumonia (480-487)
Age Groups: >65 y
Study Design: Case-crossover, time
stratified
N: 24,857 for Pneumonia
Statistical Analyses: Condition logistic
regression
Covariates: Season, long term trend, day
of-the-wk, mean temperature, relative
humidity, barometric pressure, extinction
coefficient
Season: All year
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1
Notes: Also looked at Ml cohort
Pollutant: PM non-traffic
Averaging Time: 24 h
Percentiles (pneumonia cohort):
5th:-7.3
25th:-3.28 //g/m3
50th(Median): -0.88
75th: 1.92
95th: 12.11
PM Component: BC
Monitoring Stations: 4-5 monitors
Copollutant (correlation):
PM non-traffic:
PM2.6
r - 0.74
CO
r - -0.01
NO?
r - 0.14
0s
r - -0.47
BC
r - -0.01
PM Increment: PM non-traffic lag 0:
13.44 //g/m3
PM non-traffic lag 0-1 avg: 10.28 /yglm3
% change in Pneumonia:
PM non-traffic-0.57 [-7.51, 6.36]
lag 0
PM non-traffic-0.94 [-7.20, 5.32]
mean lag 1
Reference: Zhong et al. (2006, 093264)
Period of Study: Apr-Oct 2002
Location: Cincinnati, Ohio
Hospital Admissions
Outcome (ICD-9): Asthma (493-493.91)
Age Groups: 1-18 yrs
Study Design: Time series
N: 1254 admissions
Statistical Analyses: Poisson multiple
regression, GAM
Covariates: Season, temperature,
humidity, ozone, day of the week
Season: NR
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: 1-5 days
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD):
Apr: 12.4 (3.8)
May: 13.6(5.8)
Jun: 21.6(9.9)
Jul: 25.8(11.9)
Aug: 20.3 (8.7)
Sep: 19.5(11.1)
Oct: 12.8(6.4)
Monitoring Stations: NR (data obtained
from the National Virtual Data System)
Copollutant (correlation): NR
Notes: Author states all pairwise
correlations were insignificant
PM Increment: NR
RR Estimate [CI]:
Notes: This study focused primarily on
aeroallergens and asthma visits
July 2009
E-338
DRAFT-DO NOT CITE OR QUOTE

-------
Reference
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Zanobetti and Schwartz
(2006,090195)
Period of Study: 1995-1999
Location: Boston, MA
Outcome: Pneumonia (480-487)
Age Groups: >65 y
Study Design: Case-crossover, time
stratified
N: 24,857 for Pneumonia
Statistical Analyses: Condition logistic
regression
Covariates: Season, long term trend, day
of-the-wk, mean temperature, relative
humidity, barometric pressure, extinction
coefficient
Season: All year
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1
Notes: Also looked at Ml cohort
Pollutant: PM2.6
Averaging Time: 24 h
Percentiles (pneumonia cohort):
25th: 7.23//g/m3
50th(Median): 11.10
75th: 16.14
PM Component: Black Carbon (BC), PM
non-traffic
Monitoring Stations:
4-5 monitors
Copollutant (correlation):
PM2.b:
CO
r - 0.52
NO?
r - 0.55
0s
r - 0.20
BC
r - 0.66
PM non-traffic
r - 0.74
PM Increment: PM2.6 lag 0: 17.17 /yglm3
PM2.6 lag 0-1 avg: 16.32 //g/m3
% change in Pneumonia:
6.48[1.13,11.43]
lag 0
5.561-0.45,11.27]
mean lag 1
All units expressed in uillm unless otherwise specified.
July 2009
E-339
DRAFT-DO NOT CITE OR QUOTE

-------
Table E-15. Short-term exposure-respiratory-ED/HA-Other Size Fractions
Concentrations'
Study
Design & Methods
Effect Estimates (95% CI)
Reference: Andersen et al. (2007,
093201)
Period of Study: 2001-2004
Location: Copenhagen, Sweden
Outcome (ICD10): Respiratory disease
(J41-46)
Asthma (J45, 46)
Age Groups: 5-18 and > 65
Study Design: Time-series
N:1327 days
— 1.5 million people at-risk
Statistical Analyses: Poisson
regression, GAM.
Covariates: influenza epidemics, pollen,
temperature, dew point, day-of-week,
holiday, season.
Season: All
Dose-response Investigated? No
Statistical Package: R with gam and
mgcv packages.
Lags Considered: 0-5
Pollutant: Number concentration (NC) of
ultrafine & accumulation mode particles
Averaging Time: 24-h
Mean particles/cm3 (SD): NCtot (total):
8116(3502)
25th: 4959
50th 6243
75th: 8218
99th: 16189
IQR: 3259
NC100 (< 100 nm): 6847 (2864)
25th: 5738
50th (Median): 7358
75th: 9645
99th: 19895
IQR: 3907
Mean particles/cm3 for four size modes
(median diameter (nm) noted):
NCa12: 493(315)
NCa23: 2253(1364)
NCa57: 5104 (2687)
NCa212: 6847 (2864)
Monitoring Stations: 3 (Background,
rural Background, urban Curbside, urban)
Notes: NC exposure data available for
n - 578 days. Information on distribution
of 4 size modes provided in the paper.
Copollutant (correlation):
NCtot and PMio: r - 0.39
NCtot and PM2.6: r - 0.40
NCtot and NO2: r - 0.68
PM10 and PM2.6: r - 0.8
"Low or no" correlations between 4 size
modes
NCa212 and PM2.6: r - 0.8
NCa212 and PM10: r - 0.63
NCa57 and NO2: r - 0.57
Notes: selected correlations reported in
text, all correlations in annex to the
manuscript
PM Increment: Based on the IQR,
specific to metric (see below).
RR Estimate:
Single pollutant results, Asthma, (5-18
yrs), lag 0-5:
PM2.6:1.15 [1,1.32], IQr - 5
NCtot: 1.07 [0.98,1.17], IQr - 3907
NC100:1.06 [0.97,1.16], IQr - 3259
NCa12:1.08 [0.99, 1.18], IQr - 342
NCa212: 1.08 [1,1.17], IQr - 495
NCa23:1.09 [0.98, 1.21], IQr - 1786
NCa57:1.02 [0.94, 1.12], IQr - 3026
2-pollutant results:
NCa212 w/ PM10:1.1 [0.96,1.13], IQr
-	495
NCtot w/ PM10:1.03 [0.92,1.15]
NCtot w/ PM2.6: 1.04 [0.85,1.28]
All RD, (> 65 yrs), lag 0-4, single
pollutant results:
PM2.6: 1 [0.95,1.05]
NCtot: 1.04 [1,1.07] IQr- 3907
NC100:1.03 [0.99,1.07], IQr - 3259
NC12: 1.01 [0.98,1.05], IQr - 342
NC212: 1.04(1.01,1.08], IQr-495
NCa23: 0.99[0.94,1.03], IQr - 1786
NCa57: 1.04 [1,1.08], IQr - 3026
2-pollutant results:
NCa212w/PMio: 1.01 [0.96,1.07], IQr
-	495
NCtot w/ PM2.6: 0.97(0.89,1.05]
NCtot w/ PM10:1 [0.96,1.05]
Notes: Multipollutant model results also
included for models with 4 size modes.
July 2009
E-340
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Agarwal et al. (2006,
099086)
Period of Study: 2000-2003
Location: Safdarjung area of Delhi
Outcome (ICD-NR): C0PD, asthma,
emphysema
Age Groups: NR
Study Design: Time series
N: NR
Statistical Analyses: Kruskal-Wallis
one way analysis, Chi-square,
Multivariate linear regression
Covariates: Temp (min & max), relative
humidity at 0830 and 1730 h, wind
speed
Season: I (Jan-Mar), II (Apr-Jun), III (Jul-
Sep), IV (Oct-Dec)
Dose-response Investigated? Yes
Statistical Package: SPSS
Lags Considered: NR
Pollutant: SPM (Suspended PM)
Averaging Time: 8 h
Mean/zg/nr1 (SD):
Qtr I: 297.5 (34.6)
~tr II: 398.0(85.6)
Qtr III: 220.0(78.0)
Qtr IV: 399.0(54.6)
Monitoring Stations: 2
Copollutant (correlation): RSPM:
r - 0.771
Other variables:
RH0830: r - -0.482
RH1730: r - -0.531
COPD: r - 0.474
PM Increment: NR
RR Estimate [CI]: NR
Notes: This study analyzed seasonal
variation of pollutants and health
outcomes and correlations among the
variables
Reference: Agarwal et al. (2006,
099086)
Period of Study: 2000-2003
Location: Safdarjung area of Delhi
Outcome (ICD-NR): COPD, asthma,
emphysema
Age Groups: NR
Study Design: Time series
N: NR
Statistical Analyses: Kruskal-Wallis
one-way analysis, Chi-square,
Multivariate linear regression
Covariates: Temp (min & max), relative
humidity at 0830 and 1730 h, wind
speed
Season: I (Jan-Mar), II (Apr-Jun), III (Jul-
Sep), IV (Oct-Dec)
Dose-response Investigated? Yes
Statistical Package: SPSS
Lags Considered: NR
Pollutant: RSPM (Respirable Suspended
PM < 10 /ym)
Averaging Time: 8 h
Mean/zg/nr1 (SD):
Qtr I: 119.0(19.8)
Qtr II: 132.0(28.4)
Qtr III: 75.0 (23.4)
Qtr IV: 168.0(40.6)
Monitoring Stations: 2
Copollutant (correlation): SPM:
r - 0.771
Other variables:
Temp (min): r - -0.420
COPD: r - 0.353
PM Increment: NR
RR Estimate [CI]: NR
Notes: This study analyzed seasonal
variation of pollutants and health
outcomes and correlations among the
variables
July 2009
E-341
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Arbex et al. (2007, 091637)
Period of Study: Mar 2003-Jul 2004
Location: Araraquara, Sao Paulo State,
Brazil
Hospital Admission
Outcome (ICD10): Asthma (J15, J45)
Age Groups: All
Study Design: Time-series
N: 493 days, 1 hospital, 640 admissions
Statistical Analyses: Generalized linear
Poisson regression model with natural
cubic spline, Mann-Whitney U Test
Covariates: Temperature and humidity
Season: All
Dose-response Investigated? Yes,
quintile analysis
Statistical Package: SPSS V.11 &
Splus 4.5
Lags Considered: 0-9
Pollutant: TSP
Averaging Time: 24 h
Mean (SD): 46.8 //gfm3 (24.4)
Range (Min, Max):
6.7-137.8 //g/m3
Monitoring Stations:
1
Notes: TSP used as a proxy for fine &
ultrafine particles since it is composed of
85-95% PM2.6.
Copollutant (correlation): NR
PM Increment: 10//g/m3
% Increase
6.96 [1.4-12.86] 2-d ma
9.090 [3.12-15.40]
3 d ma
10.28 [4.05-16.90]
4-d ma
11.63 [5.46-19.318]
5 d ma
12.61 [5.68-20.00]
6-d	ma
12.56 [5.47-20.13]
7-d	ma
% Increase by TSP quintile:
9.25-28.45 //g/m3 1.00
28.46-48.85//g/m3:: 1.55 [045-5.77]
48.86-69.06//gfm3:: 2.46 [1.08-5.60]
69.07-88.44//g/m3 2.77 [1.32-5.84]
88.45-108.9//g/m3 2.94 [1.48-5.85]
Notes: No TSP threshold for asthma
admissions noted. Analysis of lag
structure indicated that the acute effect
of TSP on admissions started 1 day after
TSP concentration increase and remained
unchanged for next 4 days.
Notes: To evaluate the association
between TSP generated from burning
sugar cane and asthma hospital
admissions.
Reference: Bartzokas et al. (2004,
093252)
Period of Study: Jun 1,1992-May 31,
2000
Location: Athens, Greece
Outcome: Respiratory and cardiovascular Pollutant: PM4.5 (black smoke)
diseases (combined)
Age Groups: NR
Study Design: Time series
N: 1554 patients
Statistical Analyses: Simple linear
regression and linear stepwise regression,
Pearson correlation
Covariates: Temperature, atmospheric
pressure, relative humidity, wind speed
Season: Warm (May-Sep) and cold (Nov-
Mar)
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: NR
Averaging Time: 10-day moving avg
Mean//g/m3 (SD): NR
Monitoring Stations: 1
Copollutant (correlation): N
PM Increment: NR
Correlation with Number of
Admissions:
Entire year
Original: r - 0.18
Smoothed: r - 0.31
Warm period
Original: r - 0.19
Smoothed: r - 0.30
Cold period
Original: r - 0.18
Smoothed: r - 0.34
"All above values are statistically
significant
July 2009
E-342
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Erbas et al. (2005, 073849)
Period of Study: Jul 1,1989-Dec 31,
1992
Location: Melbourne, Australia
Outcome (ICD): COPD (490-492,494,
496)
Asthma (493)
Age Groups: NR
Study Design: Time series
N: NR
Statistical Analyses: GLM, GAM,
Parameter Driven Poisson Regression,
Transitional Regression, Seasonal-Trend
decomposition based on Loess smoothing
for seasonal adjustment
Covariates: Secular trends, seasonality,
relative humidity, dry bulb temp, dew
point temp
Season: NR
Dose-response Investigated? Yes
Statistical Package: S-Plus, SAS
Lags Considered: 0-5 days
Pollutant: PM 0.1-1 (API)
Averaging Time: 24-hs
Mean (min-max): NR
Monitoring Stations: 9
Copollutant (correlation): NR
PM Increment: Increase from the 10th-
90th percentile (value NR)
RR Estimate [CI]:
COPD
GAM:
0.95[0.91,1.00]
GLM, PDM, TRM: NR
Asthma
NR
Notes: This study was used to
demonstrate that conclusions are highly
dependent on the type of model used
Reference: Halonen et al. (2008,
189507)
Period of Study: 1998-2004
Location: Helsinki, Finland
Outcome: Respiratory Hospitalizations &
Mortality (ICD 10: J00-99)
Age Groups: 65+ yrs
Study Design: time series
N: NR
Statistical Analyses: Poisson, GAM
Covariates: temperature, humidity,
influenza epidemics, high pollen episodes,
holidays
Dose-response Investigated? No
Statistical Package: R
Lags Considered: lags 0-3 & 5d (0-4)
mean
Pollutant: PM2.6
Averaging Time: daily
Mean (SD): NR
Min: 1.1
25th percentile: 5.5
50th percentile: 9.5
75th percentile: 11.7
Max: 69.5
Monitoring Stations: NR
Copollutant: PMco.03, PM0.030.1, PMco.i
PM
-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Llorca et al. (2005, 087825)
Period of Study: Jan 1,1992-Dec 31,
1995
Location: Torrelavega, Spain
Outcome (ICD-9): Respiratory (460-519)
and cardiac (390-459) admissions
(analyzed combined and individually)
Age Groups: NR
Study Design: Time series
N: 18,137 admissions
Statistical Analyses: Stepwise multiple
linear regression, Poisson regression,
Spearman correlation
Covariates: Influenza, day of week, wind
speed, northeast and southwest winds,
minimum and maximum temperature
Season: NR
Dose-response Investigated? No
Statistical Package: STATA
Intercooled, Release 6
Lags Considered: NR
Pollutant: TSP (total suspended
particles)
Averaging Time: 24 h
Mean//g/m3 (SD):
48.8 (23.7)
Monitoring Stations: 3
Copollutant (correlation):
SO?: r - -0.400
SH2: r - -0.392
NO: r - -0.109
NO2: r - -0.120
Other variables:
Rain: r - -0.339
Max temp: r - 0.071
Min temp: r - -0.003
Avg temp: r - 0.035
Wind speed: r - -0.357
PM Increment: NR
Rate Ratio Estimate [CI]:
Cardiorespiratory Admissions
Single-pollutant model: 0.92 [0.86,0.98]
Five-pollutant model: 1.05 [0.97,1.14]
Respiratory Admissions
Single-pollutant model: 0.98 [0.89,1.08]
Five-pollutant model: 0.91 [0.80,1.02]
Reference: Michaud et al. (2004,
089900)
Period of Study: Jan 1997-May 2001
Location: Hilo, Hawaii
ED visits
Outcome:
Asthma/COPD (490-496)
Respiratory Irritation (506-508)
Age Groups: All
Study Design: Time-series
N: 1,561 ER visits
Statistical Analyses: Multiple linear
regression
Covariates: Hourly temperature,
minimum daily temperature, minimum
daily temperature, humidity, year, month,
day of the week
Season: all
Dose-response Investigated? No
Statistical Package:
STATA 6.0
SAS
Lags Considered: Previous night, 1,2,3
Pollutant: PMi
Averaging Time: 24 h avg
Mean (SD): 1.91 (2.95)//g/m3
Range (Min, Max):
0.0, 56.6/yglm3
Monitoring Stations:
2
Notes: Copollutant (correlation): NR
PM Increment: 10//g/m3
RR Estimate [Lower CI, Upper CI]
lag:
Asthma, C0PD (499-496): Adjusted for
day, month & year:
1.11 (0.92,1.34), 00: 00-6: 00AM
1.14(1.03,1.26), lag 1
1.06(0.83, 0.94), lag 2
0.91 (0.06,1.05), lag 3
Asthma (493, 495): Adjusted for day,
month &year:
1.03(0.90,1.42), 00: 00-6: 00AM
1.02(0.94,1.21), lag 1
1.02(0.99,1.23), lag 2
0.97 (0.69,1.15), lag 3
Bronchitis (490, 491): Adjusted for day,
month &year:
1.02(0.82,1.41), 00: 00-6: 00AM
1.07 (1.18,1.49), lag 1
0.97 (0.60,1.34), lag 2
0.93(0.43,1.18), lag 3
Notes: Crude and estimates adjusted for
month and year only also presented.
Notes: Volcanic fog - vog
July 2009
E-344
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Migliaretti et al. (2005,
088689)
Period of Study: 1997-1999
Location: Turin, Italy
Outcome:
Cases: Asthma (493)
Controls: admissions for non respiratory
or cardiac conditions (460-487, 490-493,
494-496, 500-519, 390-405,410-429)
Age Groups: 0-14,15-64, >64
Study Design: Case-control
N: Cases: 1,401
Controls: 201,071
Statistical Analyses: Logistic regression
Covariates: Gender, age, daily mean
temperature, season, day of week,
holidays, education level
Season: All
Dose-response Investigated? No
Lag: 0-2 d avg
Pollutant: TSP
Averaging Time: Means of daily total
levels at stations
Mean (SD): 105.3//g/m3, (44.2)
Percentiles: 25th: NR
50th(Median): 96.0 //g|m3
75th NR
Monitoring Stations:
10
Notes: Copollutant (correlation): All
seasons:
NOs-TSP - 0.80
Winter:
NOs-TSP - 0.77
summer:
NOs-TSP - 0.69
PM Increment: 10//g|m3 increase
% Increase, lag 0-2 d avg
1	pollutant model:
<15:1.90[0.40,3.40]
15-64:2.30 [-0.01,5.20]
>64:2.30[1.10, 3.60]
Total: 2.30(1.10, 3.60]
% Increase, lag 0-2 d avg
2	pollutant model:
<15:-0.12 [-0.03, 2.50]
15-64:0.90 [-0.04,5.61]
>64:1.21-0.01,4.32]
Total: 0.91 [-0.02,3.11]
Reference: Migliaretti et al. (2004,
087425)
Period of Study: 1997-1999
Location: Turin, Italy
Outcome:
Cases: Asthma (493)
Controls: non-respiratory or cardiac
admissions (460-487, 490-493, 494-496,
500-519, 390-405,410-429)
Age Groups: 0-15
Study Design: Case-control
N: Cases: 1,060
Controls: 25,523
Statistical Analyses: Logistic
regression //g|m3 increase
Covariates: Gender, age, daily mean
temperature, season, day of week,
holidays, solar radiation
Season: All
Lags Considered: 1-3 d avg
Pollutant: Total suspended particulate
Averaging Time: Mean of admission day
and 3 preceding days
Mean(SD): 114.5//g/m3, (42.8)
Percentiles:
25th: NR
50th(Median): 109.9 //g/m3
75th: NR
Monitoring Stations:
10
Notes: Copollutant (correlation): TSP-
NO: 0.76
PM Increment: 10//g/m3
% Increase, lag 1-3 d avg
< 4 yrs: 1.8% [0.00, 3.05]
4-15 yrs: 3.0% [0.01,5.08]
all: 1.8% [0.03, 3.02]
adjusted for all covariates
Notes: Multipollutant models also used
July 2009
E-345
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Neuberger et al. (2004,
093249)
Period of Study: 1999-2000 (1 yr
period)
Location: Vienna and Lower Austria
Outcome (ICD-9): Bronchitis,
emphysema, asthma, bronchiectasis,
extrinsic allergic alveolitis, and chronic
airway obstruction (490-496)
Age Groups: 3.0-5.9 yrs
7-1 0 yrs
65 +
Study Design: Time series
N: 366 days (admissions NR)
Statistical Analyses: GAM
Covariates: SO2, NO, NO2, O3,
temperature, humidity, and day of the
week
Season: NR
Dose-response Investigated? Yes
Statistical Package: S-Plus 2000
Lags Considered: 0-14 days
Pollutant: PM1
Averaging Time: 24 h
Mean/yglnv1 (SD): NR
Monitoring Stations: NR
Copollutant (correlation): I
PM Increment: NR
Effect parameters (Vienna children):
Respiratory Health
Male sex - 0.098
Allergy - 0.238
Asthma in family - 0.190
Traffic - 0.112
Log Asthma Score
Allergy - 0.210
Asthma in family - 0.112
Rain - 0.257
"only significant coefficients are
presented
Association with tidal lung function:
P - -1.059 (p-value - 0.060)
Notes: No significant associations
between PM and respiratory mortality
were found for either sex. Data is also
provided for children in the rural area
where age, allergy, asthma in family,
passive smoking, and PM fraction had
significant coefficients.
Reference: Peel et al. (2005, 056305)
Period of Study: Jan 1993-Aug 2000
Location: Atlanta, Georgia
Hospital Admission/ED:
ED visits
Outcome: Asthma 493, 786.09
C0PD 491,492, 496
URI 460-466, 477
Pneumonia 480-486
Age Groups: All ages. Secondary
analyses conducted by age group: Infants
0-1 yrs
Pediatric asthma 2-18 yrs
Adults > 18 yrs
Study Design: Case-control
All respiratory disease vs. finger wounds
N: 31 hospitals
ED visits NR
Statistical Analyses: Poisson
generalized linear models
General linear models
Covariates: Avg temperature and dew
point, pollen counts
Season: All
Dose-response Investigated? yes
Statistical Package: SAS 8.3
S-Plus 2000
Lags Considered: 0-7 days and 14 day
distributed lag
Pollutant: UF (10-10Onm)
Averaging Time: 24 h avg
Mean (SD): 3800 (40700)
Percentiles:
10th: 11500
90th: 74600
PM Component: Oxygenated
hydrocarbons (OH), sulfate, acidity,
elemental carbon (EC), organic carbon
(0C), water-soluble transition metals
Monitoring Stations: "Several"
Copollutant (correlation):
PM10: r - -0.13
0s: r - -0.13
NO2: r - 0.26
CO: r - 0.10
SO2: r - 0.24
PM2.6: r - -0.16
PM10 2.5: r - 0.13
Increment:
30,000 #/cm3
All Respiratory Disease
0.984 [0.968-1.000]
URI
0.986[0.966,1.006]
Asthma
0.999(0.977,1.021]
Pneumonia
0.997(0.953,1.002]
C0PD
0.982(0.942,1.022]
July 2009
E-346
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Simpson et al. (2005,
087438)
Period of Study: 1996-1999
Location: Brisbane, Sydney, Melbourne,
and Perth, Australia
Outcome: All Respiratory (460-519)
Asthma (493)
COPD (490-492)
Pneumonia, acute bronchitis (466, 480-
486)
Age Groups: All ages, split into f15-64
and > 64 yrs
Study Design: Time-series
N: NR —64,000 admissions
Statistical Analyses: GAM wf LOESS
smoothers
GLM wf natural and penalized spline
smoothers
Covariates: Temperature, relative
humidity, rain, day of the week, public
and school holidays, influenza epidemics,
and controlled burn events
Season: All
Dose-response Investigated? Yes
Statistical Package: S-Plus
R Lags Considered: 1-3, 0-1 avg
Pollutant: BSP (indicator of particles
< 2 /jm in diameter)
(10 ^ m ')
Averaging Time: 24 h avg
Mean (SD): Means only
Brisbane 0.3 10 m
Sydney 0.3 10 m
Melbourne 0.3 10 m ''
Perth 0.3 10 ^ m 1
Range (Min, Max):
Brisbane 0.0, 2.5 10 m ''
Sydney 0.0,1.6 10 ^m 1
Melbourne 0.0, 2.2 10 m
Perth 0.1,1.8 10 ^ m 1
PM Component: Monitoring Stations:
"network of sites across each city"
Notes: Copollutant (correlation): NR
PM Increment: "per unit increase"
RR Estimate [Lower CI, Upper CI]
lag:
Single pollutant model
Respiratory >64 yrs
1.0401 [1.0045,1.0770] Iag1
1.0520(1.0164,1.0889] Iag2;
1.0451 [1.0093,1.0821] Iag3
1.0552(1.0082,1.1045] lag 0-1 avg
Asthma 15-64 yrs
1.0641 [1.0006,1.1315] Iag2
1.0893(1.0240,1.1587] Iag3
Asthma + COPD >64 yrs
1.0713(1.0179,1.1276] Iag3
1.0552(1.0082,1.1045] lag 0-1 avg
Pneumonia & Acute Bronchitis >64 yrs
1.0587(1.0013,1.1193] Iag1
1.0636(1.0056,1.1249] lag 2
1.0769(1.0046,1.1544] lag 0-1 avg
Multipollutant model
Respiratory admissions >64 yrs
No other pollutants: 1.0552 (1.0082,
1.1045] lag 0-1 avg
Max 1 h NO?
1.0028(0.9513,1.0572] lag 0-1 avg
Max 1 h O3
1.0534(1.0058-1.1033] lag 0-1 avg
Reference: Sinclair and Tolsma (2004,
088696)
Period of Study: 25 Months
Location: Atlanta, Georgia
Outpatient Visits
Outcome: Asthma (493)
Pollutant: PM102.B-10 (/yglm3)
Averaging Time: 24 h avg
URI (460, 461, 462, 463,464,465, 466, Mean (SD): PM coarse mass ((2.5-
477)
LRI (466.1,480,481,482, 483, 484,
485, 486).
Age Groups: < - 18 y, 18 + y (asthma)
All ages (URIjjLRI)
Study Design: Times series
N: 25 months
260,000 to 275,000 health plan
members (August 1998-August 2000)
Statistical Analyses: Poisson GLM
Covariates: Season, Day of week,
Federal Holidays, Study Months
Season: NR
Dose-response Investigated?: No
Statistical Package: SAS
Lags Considered: Three 3 d moving
averages (0-2, 2-5, 6-8)
10 //m))—9.67 /yglm (4.74)
Monitoring Stations:
1
Copollutant (correlation): I
PM Increment: 4.74 (1 SD)
RR Estimate [Lower CI, Upper CI]
lag:
Child Asthma:
Coarse PM - 1.053 (S)
3-5 days lag
URI:
Course PM - 1.021 (S)
3-5 days lag
LRI:
Coarse PM - 1.07 (S)
3-5 days lag
Notes: Numerical findings for significant
results only presented in manuscript.
Results for all lags presented graphically
for each outcome (asthma, URI, and LRI).
July 2009
E-347
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Sinclair and Tolsma (2004,
088696)
Period of Study: 25 Months
Location: Atlanta, Georgia
Outpatient Visits
Outcome: Asthma (493)
URI (460, 461, 462, 463,464,465, 466,
477)
LRI (466.1,480,481,482, 483, 484,
485, 486).
Age Groups: < - 18 y, 18 + y (asthma)
All ages (URIjjLRI)
Study Design: Times series
N: 25 months
260,000 to 275,000 health plan
members (August 1998-August 2000)
Statistical Analyses: Poisson GLM
Covariates: Season, Day of week,
Federal Holidays, Study Months
Season: NR
Dose-response Investigated?: No
Statistical Package: SAS
Lags Considered: Three 3 d moving
averages (0-2, 2-5, 6-8)
Pollutant: UF (PMio-ioo nm)
Averaging Time: 24 h avg
Mean (SD): PMio-ioo nm area (/^m2/cm3)
249.33 (244.09)
Monitoring Stations: 1
Copollutant (correlation): NR
PM Increment: NR
RR Estimate [Lower CI, Upper CI]
lag:
Adult Asthma:
Ultrafine PM area - 1.223 (S)
3-5 days lag
URI:
Ultrafine PM: - 1.041 (S)
0-2 days lag
LRI:
Ultrafine PM area - 1.099 (S)
6-8 days lag
Notes: Numerical findings for significant
results only presented in manuscript.
Results for all lags presented graphically
for each outcome (asthma, URI, and LRI).
Reference: Slaughter et al. (2005,
073854)
Period of Study: January 1995-June
2001
Location: Spokane, WA
Hospital Admissions and ED visits
Outcome: All respiratory (460-519)
Asthma (493)
C0PD (491,492, 494,496)
Pneumonia (480-487)
Acute URI not including colds and
sinusitis (464, 466, 490)
Age Groups: All, 15 + years for C0PD
Study Design: Time series
N: 2373 visit records
Statistical Analyses: Poisson
regression, GLM with natural splines. For
comparison also used GAM with
smoothing splines and default
convergence criteria.
Covariates: Season, temperature,
relative humidity, day of week
Season: All
Dose-response Investigated?: No
Statistical Package: SAS, SPLUS
Lags Considered: 1 -3 d
Pollutant: PMi
Averaging Time: 24 h avg
Range (90% of concentrations):
3.3-17.6 //g/m3
Monitoring Stations:
One
Copollutant (correlation): PMi
PM2.6 r - 0.95
PMior - 0.50
PM10 2.5 r - 0.19
COr - 0.63
PM Increment: 10//g/m3
RR Estimate [Lower CI, Upper CI]
lag:
ED visits:
PMi
All Respiratory
Lag 1:1.01 [0.98,1.04]
Lag 2:1.02 [0.99,1.06]
Lag 3:1.02(0.99,1.06]
Acute Asthma
Lag 1:1.03(0.97,1.09]
Lag 2: 0.99 [0.93,1.05]
Lag 3:1.02 [0.96,1.08]
C0PD (adult)
Lag 1:0.96 [0.87,1.05]
Lag 2:1.02 [0.93,1.12]
Lag 3: 0.99 [0.90,1.09]
July 2009
E-348
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Slaughter et al. (2005,
073854)
Period of Study: January 1995 through
June 2001
Location: Spokane, WA
Hospital Admissions and ED visits
Outcome: All respiratory (460-519)
Asthma (493)
C0PD (491,492, 494,496)
Pneumonia (480-487)
Acute URI not including colds and
sinusitis (464, 466, 490)
Age Groups: All, 15 + years for C0PD
Study Design: Time series
N: 2373 visit records
Statistical Analyses: Poisson
regression, GLM with natural splines. For
comparison also used GAM with
smoothing splines and default
convergence criteria.
Covariates: Season, temperature,
relative humidity, day of week
Season: All
Dose-response Investigated?: No
Statistical Package: SAS, SPLUS
Lags Considered: 1 -3 d
Pollutant: PM2.6
Averaging Time: 24 h avg
Range (90% of Concentrations):
4.2-20.2/yg/m3
Monitoring Stations:
One
Notes: Copollutant (correlation): PM2.6
PMir - 0.95
PMior - 0.62
PM10 2.5 r - 0.31
COr - 0.62
Temperature r - 0.21
PM Increment: 10/yg/m3
RR Estimate [Lower CI, Upper CI]
lag:
ER visits:
PM2.6
All Respiratory
Lag 1:1.01 [0.98,1.04]
Lag 2:1.02 [0.99,1.04]
Lag 3:1.02 [0.99,1.05]
Acute Asthma
Lag 1:1.03[0.98,1.09]
Lag 2:1.00 [0.95,1.05]
Lag 3:1.01 [0.96,1.06]
C0PD (adult)
Lag 1:0.96 [0.89,1.04]
Lag 2:1.01 [0.93,1.09]
Lag 3:1.00 [0.93,1.08]
Hospital Admissions:
PM2.6
All Respiratory
Lag 1:0.98 [0.94,1.01]
Lag 2: 0.99 [0.96,1.03]
Lag 3:1.01 [0.98,1.05]
Asthma
Lag 1:1.01 [0.91,1.11]
Lag 2:1.03 [0.94,1.13]
Lag 3:1.02 [0.93,1.13]
C0PD (adult)
Lag 1:0.99 [0.91,1.08]
Lag 2:1.06 [0.98,1.16]
Lag 3:1.03 [0.94,1.12]
July 2009
E-349
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Zanobetti and Schwartz
(2006,090195)
Period of Study: 1995-1999
Location: Boston, MA
Outcome: Pneumonia (480-487)
Age Groups: >65 y
Study Design: Case-crossover, time
stratified
N: 24,857 for Pneumonia
Statistical Analyses: Condition logistic
regression
Covariates: Season, long term trend, day
of-the-wk, mean temperature, relative
humidity, barometric pressure, extinction
coefficient
Season: All year
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1
Notes: Also looked at Ml cohort
Pollutant: PM2.6
Averaging Time: 24 h
Percentiles (pneumonia cohort):
25th: 7.23 //g/m3
50t h(Median): 11.10
75th: 16.14
PM Component: Black Carbon (BC), PM
non-traffic
Monitoring Stations:
4-5 monitors
Copollutant (correlation):
PM2.6:
CO
r - 0.52
NO?
r - 0.55
0s
r - 0.20
BC
r - 0.66
PM non-traffic
r - 0.74
PM Increment: PM2.6 lag 0:17.17 //g|m3
PM2.6 lag 0-1 avg: 16.32 //g/m3
% change in Pneumonia:
6.48[1.13,11.43]
lagO
5.561-0.45,11.27]
mean lag 1
Reference: Zanobetti and Schwartz
(2006,090195)
Period of Study: 1995-1999
Location: Boston, MA
Outcome: Pneumonia (480-487)
Age Groups: >65 y
Study Design: Case-crossover, time
stratified
N: 24,857 for Pneumonia
Statistical Analyses: Condition logistic
regression
Covariates: Season, long term trend, day
of-the-wk, mean temperature, relative
humidity, barometric pressure, extinction
coefficient
Season: All year
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1
Notes: Also looked at Ml cohort
Pollutant: BC
Averaging Time: 24 h
Percentiles (pneumonia cohort):
5th: 0.42
25th: 0.74//gfm3
50th(Median): 1.15
75th: 1.72
95th: 2.83
PM Component: PM non-traffic
Monitoring Stations: 4-5 monitors
Copollutant (correlation):
BC:
PM2.6
r - 0.66
CO
r - 0.82
NO2
r - 0.70
0s
r - -0.25
PM non-traffic
r - -0.01
PM Increment: BC lag 0: 2.05//g|m3
BC lag 0-1 avg: 1.69 /yglm3
% change in Pneumonia:
BC-10.76(4.54,15.89]
lagO
BC-11.71[4.79,17.36]
mean lag 1
'All units expressed in//g/m3 unless otherwise specified.
July 2009
E-350
DRAFT-DO NOT CITE OR QUOTE

-------
E.3. Short-Term Exposure and Mortality
Table E-16. Short-term exposure - mortality - PM10.
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Aga et al. (2003,187122)
Period of Study: ~ 5 yrs for most
cities, during the 1990s
Location: 28 European cities (APHEA2)
Outcome: Non-Accidental Mortality
(< 800)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
LOESS
Age Groups: All ages
>65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): (15, 66)
Copollutant: BS
Note: PMio only measured in 21 cities.
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
All ages
Fixed effects: 0.71 % (0.60,0.83)
0-1
Random effects: 0.67% (0.47,0.87)
0-1
>65
Fixed effects: 0.79% (0.66,0.92)
0-1
Random effects: 0.74% (0.52,0.95)
0-1
Models with effect modifiers (> 65)
24-h NO?:
25th Percentile: 0.30% (0.07,0.53)
75th Percentile: 0.97% (0.82,1.11)
24-h temperature:
25th Percentile: 0.44% (0.25,0.64)
75th Percentile: 0.91% (0.77,1.05)
24-h relative humidity:
25th Percentile: 0.98% (0.82,1.14)
75th Percentile: 0.52% (0.33,0.71)
Age standardized annual mortality rate:
25th Percentile: 0.93% (0.77,1.09)
75th Percentile: 0.61% (0.43,0.79)
Proportion individuals >65
25th Percentile: 0.67% (0.50,0.83)
75th Percentile: 0.85% (0.71,0.99)
Northwest/Central East:
25th Percentile: 0.81% (0.63,0.98)
75th Percentile: 0.26% (-0.05,0.57)
Northwest/South:
25th Percentile: 0.81% (0.63,0.98)
75th Percentile: 1.04% (0.81,1.27)
July 2009
E-351
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Analitis et al. (2006,
088177)
Period of Study: NR
Location: 29 European cities (APHEA2)
Outcome: Mortality: Cardiovascular
diseases (390-459)
Respiratory diseases (460-519)
Study Design: Time-series
Statistical Analyses: 2-stage
hierarchical modeling
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Median (SD) unit: Range: 9-64//gfm3
Range (Min, Max): NR
Copollutant: BS
Note: PMio only measured in 21 cities.
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag: Cardiovascular: Fixed effects: 0.64%
(0.47, 0.80)
0-1
Random effects: 0.76% (0.47,1.05)
0-1
0.90% (0.57,1.23)
0-5
Respiratory: Fixed effects: 0.58% (0.21,
0.95)
0-1
Random effects: 0.71 % (0.22,1.20)
0-1
1.24% (0.49,1.99)
0-5
Reference: Ballester et al. (2002,
030371)
Period of Study: 1990-1996
Location: 13 Spanish cities
Outcome: Mortality:
Non-accidental (< 800)
Cardiovascular diseases (390-459)
Respiratory diseases (460-519)
Study Design: Ecological time series
Statistical Analyses: Poisson GAM,
LOESS
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD):Huelva: 42.5 (15)
Madrid: 37.8 (17.7)
Sevilla: 45.1 (14)
Range (Min, Max): NR
Copollutant: BS
TSP
SO?
Note: PMio only measured in 3 cities.
Increment: 10 //g/m3
Relative Risk (Lower CI, Upper CI)
lag:
Non-accidental:
Random effects: 1.006 (0.998,1.015)
0-1
Fixed Effects: 1.005 (1.001,1.010)
0-1
PM10+SO2:1.013 (1.006,1.020)
0-1
Cardiovascular:
1.012(1.005,1.018)
0-1
PM10+SO2:
Random effects: 1.024 (1.001,1.048)
0-1
Fixed effects: 1.021 (1.007,1.035)
0-1
Respiratory:
1.013(1.001,1.026)
0-1
PM10+SO2:1.003 (0.983,1.023)
0-1
Reference: Bateson and Schwartz
(2004, 086244)
Outcome: Mortality:
Pollutant: PMio
Increment: 10 //g/m3
Heart Disease (390-429)
Averaging Time: 24-h avg
% Increase (Lower CI, Upper CI)
Period of Study: 1988-1991



Respiratory (460-519)
Mean (SE) unit: 37.6 (15.5) //g|m3
lag:
Location: Cook County, Illinois



Study Design: Bi-directional case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: a 65
Range (Min, Max): (3.7,128)
Copollutant: NR
All-cause: 1.14% (0.44,1.85)
0-1
Modification of Effect by Prior Diagnosis
Myocardial Infarction: 1.98% (-0.25,
July 2009
E-352
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Study population:
65,180 elderly residents with history of
hospitalization for heart or lung disease
4.26)
0-1
Diabetes: 1.49% (-0.06, 3.07)
0-1
Congestive heart failure: 1.28% (-0.06,
2.64)
0-1
C0PD: 0.58% (-0.82, 2.00)
0-1
Conduction Disorders: 0.64% (-0.61,
1.90)
0-1
All other heart or lung diseases: 0.74%
(¦0.29,1.79)
0-1
All-cause
Men
65: 2.0% (0.3, 3.8)
0-1
75: 1.5% (-0.2,3.1)
0-1
85: 0.9% (-0.7, 2.5)
0-1
95: 0.3% (-1.3,1.9)
0-1
All: 1.3% (0.4, 2.3)
0-1
Women
65: 0.1% (-1.6,1.9)
0-1
75: 0.7% (-1.1,2.4)
0-1
85: 1.2% (-0.5,3.0)
0-1
95: 1.8% (0.03,3.6)
0-1
All: 1.0% (0.1,1.9)
0-1
Total
65: 1.1% (-0.12, 2.3)
0-1
75: 1.1% (-0.1,2.3)
0-1
85: 1.2% (-0.0,2.4)
0-1
95: 1.2% (0.0, 2.4)
July 2009
E-353
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0-1
All: 1.1% (0.4,1.9)
0-1
Reference: Bell et al. (2009,191007)
Period of Study: 1987-2000
Location: 84 US Counties
Outcome: Mortality
Study Design: Time-series
Covariates: socio-economic conditions,
long term temperature
Statistical Analysis: Bayesian
hierarchical model
Age Groups: All
Pollutant: PMio
Averaging Time: 24h
Mean (SD) Unit: NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 20% of the population
acquiring air conditioning
Percent Change (95% CI) in
community-specific PM health effect
estimates for mortality
Any AC, including window units
Yearly health effect: -30.4 (-80.4-19.6)
Summer health effect: 29.9 (-84-144)
Winter health effect: -573 (-9100-7955)
Central AC
Yearly health effect: -39 (-81.4-3.3)
Summer health effect: 20. (-60.3-64.3)
Winter health effect: -1 111 (-5755-2201)
Reference: Bell et al. (2007, 093256)
Period of Study: 1999-2005
Location: US
Outcome: Mortality
Age Groups: 65 +
Study Design: time series
N: NR
Statistical Analyses: Bayesian
Hierarchical Regression
Covariates: time trend, day of week,
seasonality, dew point, temperature
Statistical Package: NR
Lags Considered: 0-2
Pollutant: PMio
Averaging Time: daily
Mean:
Ni: 0.002
Min:
Ni: 0.003
Max:
Ni: 0.021
Interquartile Range:
Ni: 0.001
Interquartile Range of Percents:
Ni: 0.01
Monitoring Stations: NR
Copollutant: Al, NhU+, As, Ca, CI, Cu,
EC, 0MC, Fe, Pb, Mg. Ni, NOs-, K, Si,
Na+, S04-, Ti, V, Zn
Co-pollutant Correlation
Ni, V: 0.48
Ni, EC: 0.30
Note: Pollutant concentrations available
for all fractions of PM2.6
PM Increment: Interquartile Range in
the fraction of PM2.6
Percent Increase in PMio Health
Effect (Lower CI, Upper CI)
Ni: 14.8 (-8.1, 37.7), lag 0
Ni: 14.7 (4.0,25.3), lag 1
Ni: 14.7 (1.8,27.5), lag 2
HS education: -31.9 (-82.4,18.6)
median income: -12.3 (-62.3, 37.7)
Percent black: 48.7 (-15.8,113)
Percent living in urban area:-20.1 (-102,
61.7)
Population: 5.1 (-14.4, 24.5)
Notes: Interquartile ranges in percent HS
education, median income, percent black,
percent living in urban area, and
population are 5.2 %, $9,223,17.3%,
11.0%, and 549,283 respectively.
July 2009
E-354
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Bellini et al. (2007, 097787) Outcome: Mortality
Period of Study: 1996-2002
Location: 15 Italian cities
All-cause (non-accidental) (< 800)
Cardiovascular (390-459)
Respiratory (460-519)
Study Design: Meta-analysis
Statistical Analyses: Poisson GLM
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant: SO2
NO?
CO
0s
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
All-cause:
0.31% (-0.19, 0.74)
0-1
Winter: 0.08%
0-1
summer: 1.95%
0-1
PMio+Os: 0.30%
0-1
PM10+NO2: 0.08%
0-1
Respiratory:
0.54% (-0.91,1.74)
0-1
Winter: 0.27%
0-1
summer: 3.61%
0-1
PMio+Os: 0.55%
0-1
PM10+NO2: 0.19%
0-1
Cardiovascular:
0.54% (0.02,1.02)
0-1
Winter: 0.20%
0-1
summer: 2.79%
0-1
PMio+Os: 0.57%
0-1
PM10+NO2: 0.39%
0-1
July 2009
E-355
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Burnett et al. (2004,
086247)
Period of Study: 1981-1999
Location: 12 Canadian cities
Outcome: Mortality:
Non-accidental (< 800)
Study Design: Time-series
Statistical Analyses: 1. Poisson,
natural splines
2. Random effects regression model
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): PMz.b: 12.8
PMio-2.5:11.4
Range (Min, Max): NR
Copollutant (correlation): NO2
0s
SO?
CO
Note: PM10 measurement calculated as
the sum of PM2.6 and PMio-2.5
measurements.
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
1981-1999
PM10: 0.57% (0.05, 0.89)
1
PM10+NO2: 0.07% (-0.44,0.58)
1
July 2009
E-356
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Cakmak et al. (2007,
091170)
Period of Study: 1/1997-12/2003
Location: Chile-7 cities
Outcome: Mortality:
Non-accidental (< 800)
Cardiovascular diseases (390-459)
Respiratory diseases (460-519)
Study Design: Time-series
Statistical Analyses: Poisson
Random effects regression model
Age Groups: All age
~ 64
65-74
75-84
>85
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): 84.9
Range (Min, Max): NR
Copollutant (correlation): O3: r — -0.16
to 0.13
SO2: r - 0.37 to 0.77
CO: r-0.49 to 0.82
Note: Correlations are between
pollutants for seven monitoring stations.
July 2009
E-357
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
Non-accidental:
0.97% (-1.09, 2.76)
0
1.31% (-1.56, 3.68)
0-5
PM10+O3+SO2+CO: 0.80% (-0.87, 2.28)
0
~ 64:
0.52% (-0.55,1.51)
0
0.49% (-0.51,1.43)
0-5
65-75:
1.07% (-1.23, 3.03)
0
1.31% (-1.57, 3.69)
0-5
75-84:
1.41% (-1.71,3.94)
0
1.93% (-2.57, 5.30)
0-5
>85:
1.56% (-1.94,4.34)
0
2.14% (-2.97, 5.85)
0-5
April-September:
1.03% (-1.17, 2.93)
0
1.37% (-1.64,3.82)
0-5
October-March:
0.07% (-0.07, 0.21)
0
0.15% (-0.15, 0.44)
0-5
Cardiovascular:
1.14% (-1.31,3.21)
0
1.49% (-1.82, 4.14)
0-5
Respiratory:
2.03% (-2.75, 5.56)
0 DRAFT-DO NOT CITE OR QUOTE
3.11% (-5.25, 8.25)
0-5

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Reference: Chen et al. (2008,190106)
Outcome (ICD9: 2001
Pollutant: PMio
Increment: 10 //g/m3
Period of Study: 2001-2004
ICD10:2002-2004):
Averaging Time: 24-h avg
% Increase (Lower CI, Upper CI)
Location: Shanghai, China
Mortality:
Mean (SD): 102.0
lag:

Non-accidental causes (ICD9 < 800
Range (Min, Max): (14.0-566.8)
Non-accidental

ICD10 A00-R99)
Copollutant (correlation):
Single Pollutant: 0.26% (0.14,0.37)

Cardiovascular (ICD9 390-459
SO?
PM10+SO2: 0.08% (-0.07,0.22)

ICD10 I00-I99)
r - 0.64
PM10+NO2: 0.01% (-0.14,0.17)

Respiratory (ICD9 460-519
NO?
PM10+SO2+NO2: 0.00% (-0.16, 0.16)

ICD10 J00-J98)
r - 0.71
Cardiovascular mortality

Study Design: Time-series

Single Pollutant: 0.27% (0.10, 0.44)

Statistical Analyses: Poisson GAM

PM10+SO2: 0.12% (-0.10,0.34)

Age Groups: All ages

PM10+NO2: 0.01% (-0.22,0.25)
PM10+SO2+NO2: 0.01% (-0.23, 0.25)
Respiratory mortality
Single Pollutant: 0.27% (-0.01, 0.56)
PM10+SO2:-0.04% (-0.41, 0.33)
PM10+NO2: -0.05% (-0.45, 0.34)
PM10+SO2+NO2: -0.10% (-0.50, 0.30)
July 2009
E-358
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Daniels et al. (2004,
087343)
Period of Study: 1987-1994
Location: 20 Largest U.S. cities
Outcome: Mortality:
Pollutant: PM10
Increment: 10 //g/m3
Total (Non-accidental) mortality
Averaging Time: 24-h avg
% Increase (Lower CI, Upper CI)
Cardiovascular-Respiratory (390-448)
Mean (SD):
lag:
(480-486, 487, 490-496, 507)
Los Angeles: 46.0
Total (non-accidental):
Other-cause mortality
New York: 28.8
0.17% (0.03, 0.30)
Study Design: Time-series
Chicago: 35.6
0
Statistical Analyses: City-Specific
Dallas-Ft. Worth: 23.8
0.20% (0.07, 0.33)
Estimates: Poisson GLM, natural cubic


splines
Houston: 30.0
1
Combined Estimates: 2-stage Bayesian
San Diego: 33.6
0.28% (0.16, 0.41)
hierarchical model
Santa Ana-Anaheim: 37.4
0-1 avg
Age Groups: All ages
Phoenix: 39.7
Cardiovascular-Respiratory:

Detroit: 40.9
0.17% (-0.01,0.35)

Miami: 25.7
0

Philadelphia: 35.4
0.27% (0.09, 0.44)

Minneapolis: 26.9
1

Seattle: 25.3
0.30% (0.18, 0.51)

San Jose: 30.4
0-1 avg

Cleveland: 45.1
Other-cause:

San Bernardino: 37.0
0.17% (-0.03, 0.37)

Pittsburgh: 31.6
0

Oakland: 26.3
0.12% (-0.07, 0.31)

Atlanta: 34.4
1

San Antonio: 23.8
0.20% (0.01,0.38)
0-1 avg
Threshold Models: Total Mortality
Threshold - 15//g/m3
0.30% (0.17, 0.42)
0-1 avg
Threshold - 0 //g|m3
0.28% (0.16, 0.41)
0-1 avg
Threshold - 20 //g|m3
0.30% (0.16, 0.43)
0-1 avg
Reference: De Leon et al. (2003,
055688)
Period of Study: 111985-1211994
Location: New York, New York
Outcome: Mortality:
Circulatory (390-459)
Cancer (140-239)
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups: All ages
<75
>75
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD):
33.27 /yglm3
IQR (25th, 75th):
(22.67, 40.83)
Copollutant (correlation): 03
CO
SO?
NO?
Increment: 18.16 //g/m3
Relative Risk (Lower CI, Upper CI)
lag:
All Ages
Cancer: 1.014(1.000,1.029)
0-1
¦w/out respiratory: 1.011 (0.996,1.026)
0-1
¦w/ respiratory: 1.051 (0.998,1.107)
0-1
July 2009
E-359
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI) ~
Circulatory: 1.025(1.014,1.035)
0-1
¦w/out respiratory: 1.022 (1.012,1.033)
0-1
¦w/ respiratory: 1.054(1.022,1.086)
0-1
<75
Cancer: 1.003 (0.985,1.021)
0-1
¦w/out respiratory: 1.002 (0.983,1.022)
0-1
¦w/ respiratory: 1.009(0.943,1.078)
0-1
Circulatory: 1.027 (1.012,1.043)
0-1
¦w/out respiratory: 1.027 (1.011,1.043)
0-1
¦w/ respiratory: 1.033(0.980,1.089)
0-1
>75
Cancer: 1.033 (1.009,1.058)
0-1
¦w/out respiratory: 1.025 (1.000,1.050)
0-1
¦w/ respiratory: 1.129 (1.041,1.225)
0-1
¦w/out pneumonia: 1.026 (1.002,1.050)
0-1
¦w/ pneumonia: 1.183 (1.058,1.323)
0-1
¦w/out COPD: 1.032 (1.008,1.057)
0-1
¦w/ COPD: 1.008 (0.849,1.197)
0-1
Circulatory: 1.025(1.012,1.038)
0-1
¦w/out respiratory: 1.022 (1.008,1.035)
0-1
¦w/ respiratory: 1.066 (1.027,1.106)
0-1
¦w/out pneumonia: 1.023 (1.010,1.036)
0-1
¦w/ pneumonia: 1.078 (1.018,1.141)
0-1
¦w/out COPD: 1.025 (1.012,1.038)
July 2009	E-360	DRAFT - DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0-1
¦w/COPD: 1.058 (0.991,1.130)
0-1
Reference: Dominici et al. (2003,
042804)
Period of Study: 1987-1994
Location: 88 U.S. cities
Outcome: Mortality:
Pollutant: PMio
Increment: 10 //g/m3
All-cause (non-accidental) (< 800)
Cardiac (390-448)
Respiratory (490-496)
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
% Increase (Lower CI, Upper CI)
lag:
Cardio- respiratory

Influenza (487)
Copollutant (correlation): NR
0.31% (0.15, 0.50)

Pneumonia (480-486, 507)

1

Other causes

All-cause

Study Design: Time-series

0.22% (0.10, 0.38)

Statistical Analyses: 2-stage Bayesian
hierarchical model
Age Groups: < 65
65-74

1
Other causes
0.13% (-0.05, 0.29)
1

>75


Reference: Dominici et al. (2004)
Outcome: Mortality:
Pollutant: PMio
Increment: 10 //g/m3
Period of Study: 1987-1994
Total (non-accidental)
Averaging Time: 24-h avg
% Increase (Lower CI, Upper CI)
Location: 90 U.S. cities (NMMAPS)
Study Design: Time-series
Mean (SD): NR
lag:

Statistical Analyses: Poisson. GAM,
GLM
Age Groups: All ages
Range (Min, Max): NR
~ - 3
0.2% (0.05, 0.35)
Reference: Dominici et al. (2004,
096951)
Period of Study: 1986-1993
Location: 10 U.S. cities
Outcome: Mortality:
Pollutant: PMio
Increment: 10 //g/m3
Total (non-accidental)
Study Design: Time-series
Statistical Analyses: 2-stage Bayesian
hierarchical model
Averaging Time: 24-h avg
Mean (SD):
Birmingham 34.8
Canton 28.4
% Increase (Lower CI, Upper CI)
lag:
Combined analysis:
0.26% (-0.37, 0.65)
Age Groups: All;
Colorado Springs 27.5
Minneapolis/St. Paul 28.1
Seattle 32.2
Spokane 42.9
Chicago 36.3
Detroit 36.7
New Haven 28.6
Pittsburgh 36.0
New York: 28.8
0-1
Separate analysis:
0.28% (-0.12, 0.63)
0-1
Notes: A separate analysis assumes the
mortality data does not provide any
information on the log relative rates of
mortality.
Reference: Dominici et al. (2007,
097361)
Period of Study: PMio: 1987-2000
PM2.b: 1999-2000
Location: 100 U.S. counties (NMMAPS)
Outcome: Mortality:
All-cause (non-accidental)
Cardiorespiratory
Other-cause
Study Design: Time-series
Statistical Analyses: 2-stage Bayesian
hierarchical model
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
PMio
All-cause:
East:
1987-1994: 0.29% (0.12,0.46)
1
1995-2000: 0.13% (-0.19,0.44)
1
July 2009
E-361
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)



1987-2000: 0.25% (0.11,0.39)
1



1
West:
1987-1994: 0.12% (-0.07,0.30)
1



1
1995-2000: 0.18% (-0.07,0.44)
1



1
1987-2000: 0.12% (-0.02,0.26)
1



I
National:
1987-1994: 0.21% (0.10, 0.32)
1



I
1995-2000: 0.18% (0.00, 0.35)
1



1
1987-2000: 0.19% (0.10, 0.28)
1



1
Cardiorespiratory:
East:
1987-1994: 0.39% (0.16, 0.63)
1



1
1995-2000: 0.30% (-0.13,0.73)
1



1
1987-2000: 0.34% (0.15, 0.54)
1



I
West:
1987-1994: 0.17% (-0.07,0.40)
1



I
1995-2000: 0.13% (-0.23,0.50)
1



1
1987-2000: 0.14% (-0.05,0.33)
1



1
National:
1987-1994: 0.28% (0.14, 0.43)
1



1
1995-2000: 0.21% (-0.03,0.44)
1



I
1987-2000: 0.24% (0.13, 0.36)
i



i
Other-cause:
East:
1987-1994: 0.21% (-0.03,0.44)
1



I
1995-2000: 0.00% (-0.49,0.50)
1



1
1987-2000: 0.15% (-0.09,0.39)
July 2009
E-362
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1
West:
1987-1994: 0.09% (-0.21,0.38)
1
1995-2000: 0.23% (-0.15,0.62)
1
1987-2000: 0.17% (-0.07,0.41)
1
National:
1987-1994: 0.15% (-0.02,0.32)
1
1995-2000: 0.17% (-0.07,0.41)
1
1987-2000: 0.15% (0.00, 0.29)
1
Reference: Dominici et al. (2007,
099135)
Period of Study: 2000-2005
Location: 72 U.S. counties representing
69 communities
Outcome: Total mortality
Study Design: Time-series
Statistical Analyses: 2-stage Bayesian
hierarchical model
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
The study does not provide results
quantitatively.
Note: The study investigated whether
county-specific short-term effects of
PMio on mortality are modified by long-
term county-specific nickel or vanadium
PM2.6 concentrations.
Reference: Fischer et al. (2003,
043739)
Period of Study: 1986-1994
Location: The Netherlands
Outcome: Mortality:
Non-accidental (< 800)
Pneumonia (480-486)
C0PD (490-496)
Cardiovascular (390-448)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
LOESS
Age Groups: <45
45-64
65-74
>75
Pollutant: PMio
Averaging Time: 24-h avg
Median (SD) unit: 34
Range (Min, Max): (10, 278)
Copollutant: BS
0s
NO?
SO?
CO
Increment: 80 //gfm3
Relative Risk (Lower CI, Upper CI)
lag:
Cardiovascular
<45:0.906(0.728,1.128)
0-6
45-64: 1.023(0.945,1.106)
0-6
65-74: 1.002(0.945,1.062)
0-6
>	75: 1.016 (0.981,1.052)
0-6
C0PD
<45:1.153(0.587,2.268)
0-6
45-64: 1.139(0.841,1.541)
0-6
65-74: 1.166(0.991,1.372)
0-6
>	75:1.066 (0.965,1.178)
0-6
Pneumonia
<45:1.427(0.806,2.525)
0-6
45-64: 1.712(1.042,2.815)
July 2009
E-363
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0-6
65-74: 1.240 (0.879,1.748)
0-6
> 75:1.123 (1.011,1.247)
0-6
Reference: Fischer et al. (2004,
055605)
Period of Study: 6/2003-8/2003
Location: The Netherlands
Outcome: Total mortality
Study Design: NR
Statistical Analyses: NR
Age Groups: All ages
Pollutant: PMio
Averaging Time: Weekly avg
Mean (SD):
2000: 31
2002: 33
2003: 35
IQR (25th, 75th): NR
Copollutant: 0:i
The study does not present quantitative
results.
Notes: The study estimates the number
of deaths attributable to PMio during the
summers of 2000, 2002, and 2003.
Reference: Forastiere et al. (2005,
086323)
Period of Study: 1998-2000
Location: Rome, Italy
Outcome: Mortality:
Pollutant: PMio
Increment: 29.7 /yglm3
Ischemic heart disease (410-414)
Study Design: Time-stratified case-
crossover
Statistical Analyses: Conditional
logistic regression
Averaging Time: 24-h avg
Mean (SD): 52.1 (22.2)
IQR (25th, 75th):
(36.0, 65.7)
% Increase (Lower CI, Upper CI)
lag:
4.8% (0.1,9.8)
0

Age Groups: >35
Copollutant (correlation):
PNC: r - 0.38
CO: r - 0.34
NO?: r - 0.45
SO?: r - 0.23
O3: r - 0.13
4.9% (0.0,10.1)
1
3.8% (-1.0, 8.9)
2
2.8% (-2.0, 7.7)
3
6.1% (0.6,11.9)
0-1
Reference: Forastiere et al. (2007,
090720)
Period of Study: 1998-2001
Location: Rome, Italy
Outcome: Mortality:
Pollutant: PMio
Increment: 10 //g/m3
Natural (< 800)
Malignant neoplasms (140-208)
Diabetes mellitus (250)
Hypertensive disease (401-405)
Previous acute myocardial infarction
(410,412)
Averaging Time: 24-h avg
Mean Range (SD) unit: 51.0
(21.0)/yg/m3
IQR (25th, 75th):
(36.1,63.0)
Copollutant (correlation): NR
% Increase (Lower CI, Upper CI)
lag:
Non-accidental: 1.1% (0.7,1.6)
0-1
Low income: 1.9%
0-1
Low SES: 1.4%

Other ischemic heart diseases (411, 413-
414)


Conduction disorders (426)

0-1

Dysrhythmia (427)

High income: 0.0%

Heart failure (428)

0-1

Cerebrovascular disease (430-438)

High SES: 0.1%

Peripherical artery disease (440-448)

0-1

COPD (490-496)

Low PM Area: 0.9% (-0.4,2.1)

Study Design: Time-stratified case-
crossover

0-1
High PM Area: 1.47% (0.4,2.5)

Statistical Analyses: Conditional
logistic regression

0-1

Age Groups: >35


July 2009
E-364
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Location: 9 Italian cities
Reference: Forastiere et al. (2008,
186937)
Period of Study: 1997-2004
Outcome: Mortality:
Non-accidental (< 800)
Study Design: Time-stratified case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: >35
Pollutant: PMio
Averaging Time: 24-h avg
Mean Range (SD) unit:
35.1 to 71.5
Range (5th, 95th):
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
Total: 0.60% (0.31,0.89)
0-1
Lowest 5th: 14.3
Highest 95th: 147.0
Copollutant (correlation): NR
Age
35-64:-0.20% (-0.77,0.37)
0-1
65-74: 0.51% (0.05, 0.98)
0-1
75-84: 0.59%(0.20, 0.97)
0-1
>	85: 0.97% (0.53,1.42)
0-1
>	65: 0.75% (0.42,1.09)
Sex
Men: 0.72% (0.37,1.07)
0-1
Women: 0.83% (0.33,1.33)
0-1
Median income (by census block)
Low (< 20th percentile): 0.80% (-0.02,
1.62)
0-1
Mid-low (20th-50th percentile): 0.68%
(0.25,1.12)
0-1
Mid-high (51 st-80th percentile): 0.85%
(0.40,1.30)
0-1
High (> 80th percentile): 0.30% (-0.25,
0.86)
0-1
Location of death
Out-of-hospital: 0.71% (0.32,1.11)
0-1
Discharged 2-28 d before death: 1.34%
(0.49, 2.20)
0-1
ln-hospital: 0.65% (0.33, 0.97)
0-1
Nursing home: -0.04% (-1.02, 0.95)
0-1
July 2009
E-365
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Goldberg et al. (2003,
035202)
Period of Study: 1984-1993
Location: Montreal, Quebec, Canada
Outcome: Mortality: Congestive Heart
Failure (428)
Study Design: Time-series
Statistical Analyses: Poisson, natural
splines
Age Groups: a 65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): PMio: 32.2 (17.6)
IQR (25th, 75th): PMio: (19.7,41.1)
Copollutant (correlation):
PM2.6, TSP, Sulfate, CoH, SO2, NO?, CO,
0s
This study does not present results
quantitatively for PMio
Reference: Goldberg et al. (2003,
035202)
Period of Study: 1984-1993
Location: Montreal, Quebec, Canada
Outcome: Mortality:
Diabetes (250)
Study Design: Time-series
Statistical Analyses: Poisson, natural
spline
Age Groups: a 65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD):
PMio: 32.2 (17.6)//g/m3
IQR (25th, 75th):
PMio: (19.7, 41.1)
Copollutant (correlation):
PM2.6, Sulfate, CoH, SO2, NO2, CO, 0s
This study does not present results
quantitatively for PMio
July 2009
E-366
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Reference: Kan and Chen (2003,
087372)
Outcome: Mortality:
Pollutant: PMio
Increment: 10 //g/m3
Non-accidental (< 800)
Averaging Time: 24-h avg
Relative Risk (Lower CI, Upper CI)
Period of Study: 112000-1212001
Cardiovascular (390-459)
Mean (SD): 91.14 (51.85)
lag:
Location: Shanghai, China
COPD (490-496)
Range (Min, Max): (17.0, 385.0)
Non-accidental

Study Design: Time-series
Copollutant (correlation):
All ages: 1.003(1.001,1.005)

Statistical Analyses: Poisson GAM,
SO2: r - 0.71
0

LOESS
NO?: r - 0.73
<65:1.001 (0.997,1.005)

Age Groups: All ages

0

<65

65-75: 1.005(1.001,1.008)

65-75

0

>75

>75:1.003(1.001,1.006)
0
Cardiovascular






All ages: 1.003(1.000,1.006)
0
<65:1.002(0.994,1.010)
0
65-75: 1.003(0.998,1.008)
0
>75:1.003(1.000,1.006)
0
COPD















All ages: 1.005(0.999,1.011)
0
<65:1.004 (0.981,1.027)
0
65-75: 0.996(0.986,1.007)
0
>75:1.006(1.000,1.012)
0
Multipollutant models















SO2:1.001 (0.998,1.003)
0
NO?: 1.001 (0.998,1.003)
0
SO2+NO2: 1.000 (0.997,1.003)









0
July 2009
E-367
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kan and Chen (2003,
087372)
Period of Study: 112000-1212001
Location: Shanghai, China
Outcome: Mortality:
Non-accidental (< 800)
Cardiovascular (390-459)
C0PD (490-496)
Study Design: Case-crossover
Statistical Analyses: Conditional
logistic regression
Averaging Time: 24-h avg
Mean (SD): 91.14 (51.85)
IQR (25th, 75th): (54,114)
Copollutant (correlation):
SO2: r - 0.71
Pollutant: PM10
Increment: 10 //g/m3
Odds Ratio (Lower CI, Upper CI)
lag:
Non-accidental:
Age Groups: All ages
NO?: r - 0.73
Bidirectional referent days:
7 d: 1.000 (0.9988,1.002)
0-1 ma
7 and 14 d: 1.002 (1.000,1.004)
0-1 ma
7,14, and 21 d: 1.003 (1.001,1.005)
0-1 ma
Unidirectional referent days:
7 d: 1.015 (1.012,1.018)
0-1 ma
7 and 14 d: 1.017 (1.015,1.019)
0-1 ma
7,14, and 21 d: 1.019 (1.012,1.021)
0-1 ma
Bidirectional referent days (7,14, and 21
d):
Cardiovascular:
1.004 (1.001,1.007)
0-1 ma
C0PD:
1.006 (0.999,1.013)
0-1 ma
Non-accidental:
PM10+SO2: 0.997 (0.994,1.025)
0-1 ma
PM10+NO2: 0.997 (0.994,1.025)
0-1 ma
PM10+SO2+NO2: 0.995 (0.992,1.025)
0-1 ma
July 2009
E-368
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kan et al. (2005, 087561) Outcome: Mortality:
Period of Study: 4/25/2003-5/31/2003 Severe acute respiratory syndrome
(SARS)
Location: Beijing, China
Study Design: Time-series
Statistical Analyses: Poisson, GAM,
smoothing spline
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): 149.1 (8.1)
Range (Min, Max): (34, 246)
Copollutant:
SO?
NO?
Increment: 10/yg/m3
Relative Risk (Lower CI, Upper CI)
lag:
0.99 (0.96 to 1.03)
0
1.00 (0.97 to 1.04)
1
1.02 (0.98 to 1.06)
2
1.04	(0.99 to 1.09)
3
1.06 (1.00 to 1.11)
4
1.06 (1.00 to 1.12)
5
1.05	(0.98 to 1.12)
6
Reference: Kan et al. (2007, 091267)
Period of Study: 3/2004-12/2005
Location: Shanghai, China
Outcome (ICD10): Mortality:
Total (non-accidental) (A00-R99)
Cardiovascular (I00-I99)
Respiratory (J00-J98)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
penalized splines
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): 107.9 (2.39)/yg/m3
Range (Min, Max): (22.0, 403.0)
Copollutant (correlation):
PMio
PM2.6: r - 0.84
PM 10-2.5: r - 0.88
O3: r - 0.21
Increment: 10/yg/m3
% Increase (Lower CI, Upper CI)
lag:
PMio
Total: 0.16% (0.02, 0.30)
0-1
Cardiovascular: 0.31% (0.10, 0.53)
0-1
Respiratory: 0.33% (-0.08, 0.75)
0-1
Reference: Kan et al. (2008,156621)
Period of Study: 1/2001-12/2004
Location: Shanghai, China
Outcome: Mortality: Total (non-
accidental) (A00-R99)
Cardiovascular (I00-I99)
Respiratory (J00-J98)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural splines
Age Groups: All ages;
0-4
5-44
45-64
>65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD):
Warm season: 87.4 (1.8)
Cool season: 116.7 (2.8)
Entire period: 102.0 (1.7)
Range (Min, Max): NR
Copollutant (correlation): SO2
NO?
03
Increment: 10/yg/m3
% Increase (Lower CI, Upper CI)
lag:
Non-accidental
Warm season: 0.21 (0.09, 0.3)
0-1
Cool season: 0.26 (0.22, 0.30)
0-1
Entire period: 0.25 (0.14, 0.37)
0-1
Female: 0.33(0.18, 0.48)
0-1
Male: 0.17(0.03, 0.32)
0-1
5-44: 0.04 (-0.52, 0.59)
0-1
45-64: 0.17 (-0.11, 0.45)
July 2009
E-369
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0-1
> 65: 0.26(0.15, 0.38)
0-1
Cardiovascular
Warm season: 0.22 (-0.14, 0.58)
0-1
Cool season: 0.25 (0.05, 0.45)
0-1
Entire period: 0.27 (0.10, 0.44)
0-1
Respiratory
Warm season: -0.28 (-0.93, 0.38)
0-1
Cool season: 0.58 (0.25, 0.92)
0-1
Entire period: 0.27 (-0.01, 0.56)
0-1
Stratified by Educational Attainment
Nonaccidental:
Low: 0.33 (0.19, 0.47)
0-1
High: 0.18(0.01,0.36)
0-1
Cardiovascular:
Low: 0.30 (0.10, 0.51)
0-1
High: 0.23 (-0.03, 0.50)
0-1
Respiratory:
Low: 0.36 (0.00, 0.72)
0-1
High: 0.02 (-0.43, 0.47)
0-1
July 2009
E-370
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Keatinge and Donaldson
(2006, 087536)
Period of Study: 1991-2002
Location: London, England
Outcome: Mortality: Total (non-
accidental)
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups: a 65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant: 0:i
SO?
Increment: 10 //g/m3
Mortality per 106 (Lower CI, Upper CI)
lag:
PMio+Temp: 2.1 (0.9, 3.3)
0-2 avg
PMio+Temp+Acclim: 1.6 (0.4, 2.8)
0-2 avg
PMio+Temp+Acclim+Acclim x T: 1.5
(0.3, 2.6)
0-2 avg
PMio+Temp+Acclim+Acclim x T+Sun:
1.4 (0.2,2.5)
0-2 avg
PMio+Temp+Acclim+Acclim x
T+Sun+Wind: 0.8 (-0.4,1.9)
0-2 avg
PMio+Temp+Acclim+Acclim x
T+Sun+Wind+Abs. Humid.: 0.8 (-0.3,
1.9)
0-2 avg
PMio+Temp+Acclim+Acclim x
T+Sun +Wind +Abs. Humid. + Rain: 0.9
(-0.3, 2.0)
0-2 avg
PMio+Temp+Abs. Humid.: 1.9 (0.7, 3.1)
0-2 avg
July 2009
E-371
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kettunen et al. (2007,
091242)
Period of Study: 1998-2004
Location: Helsinki, Finland
Outcome (ICD10): Mortality:
Stroke (160-161,163-164)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
penalized thin-plate splines
Age Groups: a 65
Pollutant: PMio
Averaging Time: 24-h avg
Median (SD) unit:
Cold Season: 16.3
Warm Season: 16.5
Range (Min, Max):
Cold Season: (3.1,136.7)
Warm Season: (3.3, 67.4)
Copollutant:
PM2.6
PM 10-2.6
UFP
0s
CO
NO?
Increment:
Cold Season: 13.8/yg/m3
Warm Season: 9.8/yglm3
% Increase (Lower CI, Upper CI)
lag:
Cold Season
¦0.56% (-3.32, 2.29)
0
¦0.93% (-3.55,1.75)
1
¦1.68% (-4.30,1.00)
2
¦1.53% (-4.14,1.14)
3
Warm Season
10.89% (0.95,21.81)
0
8.56% (-0.88,18.90)
1
2.06% (-6.76,11.71)
2
¦2.89% (-11.32,6.34)
3
Reference: Kim et al. (2003,155899)
Period of Study: 1/1995-12/1999
Location: Seoul, Korea
Outcome (ICD 10): Mortality:
Pollutant: PMic
Non-accidental (all except S01-S99, T01 - Averaging Time: 24-h avg
T98)
Mean (SD): 69.19 (10.36)
Cardiovascular (I00-I52)
IQR (25th, 75th):
Respiratory (J00-J98)
(44.82, 87.95)
Cerebrovascular (I60-I69)
Copollutant (correlation): NR
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups: All ages
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
All cause:
2.8% (1.8,3.7)
0
2.8% (1.9,3.7)
1
1.4% (0.5,2.3)
2
3.7% (2.1,5.4)
distributed lag (6-day)
Respiratory:
8.3% (4.3,12.5)
0
6.4% (2.7,10.2)
1
6.5% (2.7,10.4)
2
13.9% (6.8,21.5)
distributed lag (6-day)
Pneumonia:
July 2009
E-372
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
11.6% (4.2,19.6)
0
9.0% (2.1,16.3)
1
7.7% (0.8,15.2)
2
17.1% (4.1, 31.7)
distributed lag (6-day)
COPD:
4.2% (-1.2,10.0)
0
3.5% (-1.5, 8.9)
1
1.4% (-3.7, 6.8)
2
12.2% (2.5, 22.9)
distributed lag (6-day)
Cardiovascular:
2.0% (-0.9, 5.0)
0
3.3% (0.6,6.2)
1
2.9% (0.1,5.8)
2
4.4% (-0.6, 9.6)
distributed lag (6-day)
Myocardial infarction: 2.6% (-2.3, 7.8)
0
5.8% (1.0,10.7)
1
5.5% (0.7,10.6)
2
4.9% (-3.4,13.9)
distributed lag (6-day)
Cerebrovascular:
3.2% (0.8,5.5)
0
3.1% (0.9,5.3)
1
2.4% (0.1,4.6)
2
6.3% (2.3,10.5)
distributed lag (6-day)
Ischemic stroke:
¦0.6% (-5.6, 4.7)
July 2009
E-373
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0
0.6% (-4.2, 5.7)
1
¦0.1% (-4.9, 5.1)
2
10.3% (1.0, 20.4)
distributed lag (6-day)
Reference: Kimet al. (2004, 087417)
Outcome: Mortality: Non-accidental
Pollutant: PMio
Increment: 42.11 /yglm3
Period of Study: 1/1997-12/2001
Study Design: Time-series
Averaging Time: 24-h avg
Relative Risk (Lower CI, Upper CI)
Location: Seoul, Korea
Statistical Analyses: Poisson GAM,
Mean (SD): 68.23 (36.36)//g/m3
lag:

LOESS
IQR (25th, 75th): (42.56, 84.67)
1.021 (1.009,1.035)

Age Groups: All ages
Copollutant (correlation): NR

Reference: Le Tertre et al. (2005,
087560)
Outcome: Mortality:
Pollutant: PMio
Increment: 1.0 /yglm3
Non-accidental (< 800)
Averaging Time: 24-h avg
P coefficient (SE)
Period of Study: NR
Study Design: Time-series
Mean (SD): NR
lag:
Location: 21 European cities (APHEA-2)
Statistical Analyses: Empirical Bayes
Range (Min, Max): NR
Athens: 0.001311 (0.0003)

Age Groups: All ages
Copollutant: NO2
Barcelona: 0.000575 (0.0002)
Basel: 0.000462 (0.0005)
Birmingham: 0.000305 (0.0003)
Budapest: -0.000248 (0.0005)
Cracow: 0.000155 (0.0004)
Erfurt: -0.000465 (0.0004)
Geneva: -0.000059 (0.0005)
Helsinki: 0.000389 (0.0004)
London: 0.000591 (0.0002)
Lyon: 0.001554 (0.0005)
Madrid: 0.000372 (0.0003)
Milan: 0.000901 (0.0002)
Paris: 0.000411 (0.0003)
Prague: 0.000097 (0.0002)
Rome: (0.001333 (0.0003)
Stockholm: 0.000479 (0.0009)
Tel Aviv: 0.000522 (0.0003)
Teplice: 0.000876 (0.0004)
Torino: 0.000938(0.0002)
Zurich: 0.000365 (0.0004)
Toulouse: NR (NR)
Overall: 0.00055 (0.000098)
July 2009
E-374
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lee et al. (2007, 093042)
Period of Study: 112000-1212004
Location: Seoul, Korea
Outcome (ICD10): Mortality:
Non-accidental (A00-R99)
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD):
w/ Asian dust days: 70.00 (47.80)
w/o Asian dust days: 65.77 (33.60)
Asian dust days only: 188.49 (142.85)
Copollutant:
CO
NO?
SO?
0s
Increment: 41.49 /yglm3
% Increase (Lower CI, Upper CI)
lag:
Model with Asian Dust Days
0.7% (0.2,1.3)
1-3
Model without Asian dust days
1.0% (0.2,1.8)
1-3
Reference: Lee and Shaddick (2007,
156685)
Period of Study: 1/1/1993 -
1213111997
Location: Cleveland, Ohio
Detroit, Michigan
Minneapolis, Minnesota
Pittsburgh, Pennsylvania
Outcome (ICD 10): Mortality:
Non-accidental
Study Design: Time-series
Statistical Analyses: 1. Bayesian,
penalized spline
2. Likelihood, penalized spline
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Increment: 10 //g/m3
Relative Risk (Lower CI, Upper CI)
lag:
Constant model
Cleveland: 1.0049
1
Detroit: 1.0046
1
Minneapolis: 1.0052
1
Pittsburgh: 1.0045
1
Reference: Martins et al. (2004,
087457)Period of Study: 1/1997-
12/1999
Location: Sao Paulo, Brazil
Outcome (ICD 10): Mortality:
Respiratory (J00-J99)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural cubic splines
Age Groups: £ 60
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD):
Cerqueira Cesar: 42.5(22.9)
Santa Amaro: 49.6(32.1)
Central: 52.1(23.5)
Penha: 40.4(23.8)
Santana: 72.6(24.5)
Sao Miguel Paulista: 68.6(31.0)
Range (Min, Max): NR
The study does not present quantitative
results.
July 2009
E-375
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Nawrot et al. (2007,
098619)
Period of Study: 1/1997-12/2003
Location: Flanders, Belgium
Outcome: Mortality:
Non-accidental (< 800)
Cardiovascular (390-459)
Respiratory (460-519)
Study Design: Time-series
Statistical Analyses: Main analysis:
Segmented regression models
Sensitivity analysis: Poisson GAM,
LOESS
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Median (SD) unit:
Winter: 43.3(0.88)
Spring: 39.5(0.88)
summer: 37.7(0.91)
Fall: 37.2(0.88)
Range (Min, Max): NR
Copollutant (correlation): NR
Increment:
Main analysis: NR
Sensitivity analysis: 10/yg/m3
% Increase (Lower CI, Upper CI)
lag:
Highest season-specific PMio quartile
versus the lowest season-specific PMio
quartile
summer: 7.8% (6.1, 9.6)
Spring: 6.3% (4.7, 7.8)
Autumn: 2.2% (0.58, 3.8)
Winter: 1.4% (0.06, 2.9)
Warm months (June, July, August): 7.9%
(6.2, 9.6)
Cold months (December, January,
February): 1.5% (0.22,3.3)
Intermediate months (March, April, May,
September, October, November): 4.2%
(2.9, 5.6)
Warmer Periods (April-September)
Non-accidental: 1.5% (1.1, 2.0)
0
Respiratory: 2.0% (0.6, 3.7)
0
Cardiovascular: 1.8% (1.1, 2.4)
0
Reference: O'Neill et al. (2004, 087429) Outcome: Mortality:
Period of Study: 1996-1998
1994-711995
Location: Mexico City, Mexico
Non-accidental
Study Design: Time-series
Statistical Analyses: Poisson, natural
cubic spline
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Range:
Hi-Vol: 46.3-164.0
TE0M: 48.2-107.5
Predicted: 30.2-162.4
Impactor: 58.4
Range (Min, Max):
Xalostoc
Hi-Vol: (40.0, 335.0)
TE0M: (16.5, 291.2)
Predicted: (60.6, 320.0)
Tlalnepantla
Hi-Vol: (25.0, 264.0)
TE0M: (10.4,275.9)
Predicted: (17.7,175.0)
Merced
Hi-Vol: (17.0, 266.0)
TE0M: (9.4, 318.7)
Predicted: (12.3,160.8)
Cerro de la Estrella
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
TE0M
0.04% (-0.12, 0.20)
0
¦0.02% (-0.18, 0.13)
1
¦0.01% (-0.27, 0.25)
2
¦0.03% (-0.19, 0.13)
3
¦0.03% (-0.19, 0.13)
4
¦0.05% (-0.21,0.11)
5
0.05% (-0.25, 0.35)
0-5
Predicted
¦0.05% (-0.29, 0.19)
0
July 2009
E-376
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Hi-Vol: (15.0, 292.0)
TEOM: (13.7, 268.3)
0.09% (-0.16, 0.34)
Pedregal (1996-1998)
Hi-Vol: (5.0, 226.0)
TEOM: (7.8, 264.4)
Predicted: (-0.5, 86.3)
Pedregal (1994-1995)
Hi-Vol: (24.0,114.0)
TEOM: (8.7,152.5)
Impactor: (15.0,154.0)
Predicted: (3.9, 75.9)
Predicted: (11.2,154.4)
¦0.12% (-0.43, 0.20)
2
¦0.02% (-0.26, 0.21)
3
¦0.14% (-0.37, 0.09)
4
¦0.05% (-0.28, 0.18)
5
0.00% (-0.39, 0.38)
0-5
Sierra-Anderson High Volume Air Sampler
0.02% (-0.29, 0.32)
0
0.13% (-0.27, 0.54)
1
0.21% (-0.10, 0.52)
2
0.53% (0.07, 0.99)
3
0.11% (-0.20, 0.41)
4
0.38% (0.07, 0.70)
5
GAM: 2 LOESS terms, default
convergence
1.68% (0.45, 2.93)
0
¦0.36% (-1.56, 0.86)
1
¦0.21% (-1.40,1.00)
2
¦0.18% (-1.40,1.05)
3
1.31% (0.08, 2.55)
4
1.49% (0.25, 2.73)
5
1.77% (-0.26, 3.83)
0-5
Parametric: cubic splines
5 df
1.45% (0.09, 2.83)
0
July 2009
E-377
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: O'Neill et al. (2005, 098094)
Period of Study: 1996-1998
1996-1999
Outcome: Mortality: Non-accidental
Cardiovascular (390-460)
Respiratory (460-520)
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): Mexico City: 75.8 (31.4)
¦0.71% (-2.06, 0.67)
1
¦0.59% (-1.95, 0.79)
2
¦0.70% (-2.09, 0.71)
3
0.92% (-0.46, 2.32)
4
1.17% (-0.19, 2.55)
5
1.17% (-1.54,3.95)
0-5
10 df
1.60% (0.20, 3.02)
0
¦0.80% (-2.18, 0.60)
1
¦0.73% (-2.11,0.68)
2
¦1.05% (-2.49, 0.40)
3
0.64% (-0.79, 2.10)
4
1.05% (-0.36, 2.48)
5
0.51% (-2.60, 3.71)
0-5
2 df
1.79% (0.48, 3.11)
0
¦0.09% (-1.38,1.22)
1
0.10% (-1.18,1.40)
2
0.20% (-1.10,1.52)
3
1.60% (0.30, 2.91)
4
1.72% (0.43, 3.04)
5
1.90% (-0.36, 4.21)
0-5
The study focuses on the temperature-
mortality relationship and only includes
PMio as a covariate in models.
July 2009
E-378
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Location: Mexico City and Monterrey,
Other-causes
Monterrey: 50.0 (23.5)

Mexico




Study Design: Time-series
Range (Min, Max): Mexico City: (18.0,



233.9)


Statistical Analyses: Poisson, natural


cubic splines
Monterrey: (6.2, 230.8)


Age Groups: All ages, 0-15, £ 65
Copollutant: 0:i

Reference: O'Neill et al. (2008,192314)
Outcome:
Pollutant: PMio
Increment: 10/yg/m3
Period of Study: 1/1/1998 ¦
Study Design: Time-series
Averaging Time: 24h
Percent increase (95% CI) in all-cause
12/30/2002


adult mortality (> 22yrs) by
Covariates: Temperature, Day of Week,
Mean (SD) /vg/m3:
educational level and sex
Location: Mexico City, Mexico
Temporal trends, Sex
Mexico City: 53.8 (24.9)

Mexico City
Santiago, Chile
Statistical Analysis: Poisson regression
Sao Paulo: 48.9 (21.9)
All Adults, Concurrent Day
Sao Paulo, Brazil
Statistical Package: S-Plus
Santiago: 78.7 (33.0)
None: 0.76 (0.17-1.36)

Age Groups: Adults over 21
Range (Min, Max):

Primary: 0.27 (-0.19-0.72)


Mexico City: 1.08-192.2
Secondary: 0.19 (-0.19-0.57)


Sao Paulo: 12.0-171.3
> 12 yrs: 0.83 (0.03-1.63)


Santiago: 8.0-218.6
All Adults, Lag 1


Copollutant: NR
None: 0.62 (0.02-1.22)



Primary: 0.62 (0.17-1.08)



Secondary: 0.29 (-0.09-0.90)



a 12 yrs: 0.58 (-0.21-1.38)



All Adults, Distributed Lags 0-5



None: 0.91 (-0.07-1.89)



Primary: 0.48 (-0.27-1.24)



Secondary: 0.27 (-0.36-0.90)



>12 yrs: 0.75 (-0.49-2.02)



All Adults, df(yrs)



None: 5.4



Primary: 6.0



Secondary: 6.0



a 12 yrs: 3.0



Women, Concurrent Day



None: 0.65 (-0.08-1.38)



Primary: 0.48 (-0.13-1.09)



Secondary: 0.35 (-0.16-0.86)



> 12 yrs: 1.64 (0.69-2.59)



Women, Lag 1



None: 0.62 (-0.12-1.36)



Primary: 1.03 (0.42-1.64)



Secondary: 0.59 (0.08-1.11)



>12 yrs: 1.79 (0.84-2.75)



Women, Distributed Lags 0-5



None: 0.46 (-0.74-1.68)



Primary: 1.39 (0.42-2.36)



Secondary: 0.51 (-0.30-1.33)



> 12 yrs: 1.71 (0.61-2.83)



Women, df (yrs)
July 2009
E-379
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
None: 5.4
Primary: 4.4
Secondary: 4.8
a 12 yrs: 1.0
Men, Concurrent Day
None: 0.75 (-0.21-1.72)
Primary: 0.52 (-0.11-1.15)
Secondary: 0.56 (0.08-1.05)
>	12 yrs: 1.20 (0.25-2.17)
Men, Lag 1
None: 0.45 (-0.51-1.42)
Primary: 0.70 (0.06-1.34)
Secondary: 0.47 (-0.02-0.95)
>	12 yrs: 0.74 (-0.22-1.70)
Men, Distributed Lags 0-5
None: 1.24 (-0.25-2.75)
Primary: 0.65 (-0.39-1.69)
Secondary: 0.88 (0.11-1.66)
>	12 yrs: 1.07 (-0.41-2.57)
Men, rf/(yrs)
None: 3.8
Primary: 5.6
Secondary: 4.6
>	12 yrs: 3.8
Sao Paulo
All Adults, Concurrent Day
None: 0.77 (-0.28-1.82)
Primary: 1.27 (0.78-1.76)
Secondary: 0.93 (-0.07-1.94)
>12 yrs: 2.93 (2.00-2.88)
All Adults, Lag 1
None: 0.70 (-0.34-1.76)
Primary: 1.32 (0.83-1.82)
Secondary: 1.91 (0.58-2.60)
a 12 yrs: 2.20 (1.27-3.15)
All Adults, Distributed Lags 0-5
None: 0.76 (-0.91-2.46)
Primary: 1.34 (0.55-2.14)
Secondary: 1.91 (0.35-2.60)
a 12 yrs: 2.20 (1.27-3.15)
All Adults, df(yrs)
None: 4.0
Primary: 4.0
Secondary: 2.8
a 12 yrs: 1.6
July 2009
E-380
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Women, Concurrent Day
None: 1.93 (0.87-3.00)
Primary: 1.72 (1.04-2.41)
Secondary: 0.85 (-0.21-1.92)
>	12 yrs: 1.84 (0.56-3.13)
Women, Lag 1
None: 1.41 (0.34-2.48)
Primary: 1.64 (0.96-2.33)
Secondary: 1.43 (0.36-2.50)
>	12 yrs: 2.27 (0.99-3.56)
Women, Distributed Lags 0-5
None: 2.00 (0.40-3.63)
Primary: 2.05 (0.96-3.14)
Secondary: 1.61 (0.07-3.17)
>12 yrs: 3.35 (1.49-5.25)
Women, df (yrs)
None: 2.4
Primary: 3.6
Secondary: 1.4
>	12 yrs: 0.8
Men, Concurrent Day
None:-0.43 (-2.15-1.32)
Primary: 1.36 (0.71-2.02)
Secondary: 1.74 (0.77-2.72)
>	12 yrs: 2.81 (1.71-3.92)
Men, Lag 1
None:-0.44 (-2.17-1.33)
Primary: 1.44 (0.79-2.10)
Secondary: 1.52 (0.55-2.49)
>12 yrs: 1.48 (0.38-2.59)
Men, Distributed Lags 0-5
None: -0.30 (-3.09-2.56)
Primary: 1.67 (0.65-2.70)
Secondary: 1.06 (-0.34-2.49)
a 12 yrs: 3.18(1.60-4.79)
Men, rf/(yrs)
None: 4.4
Primary: 3.2
Secondary: 0.8
>	12 yrs: 1.2
Santiago
All Adults, Concurrent Day
None: 1.44(0.53-2.36)
Primary: 0.06 (-0.21-0.34)
Secondary: 0.42 (0.06-0.78)
July 2009
E-381
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
>	12 yrs: 1.32 (0.60-2.05)
All Adults, Lag 1
None: 2.08 (1.16-30.1)
Primary: 0.53 (0.25-0.81)
Secondary: 0.55 (0.19-0.91)
>	12 yrs: 1.31 (0.59-2.04)
All Adults, Distributed Lags 0-5
None: 3.18(1.60-4.78)
Primary: 0.58 (0.10-1.06)
Secondary: 1.10(0.48-1.73)
>12 yrs: 2.00 (0.93-3.07)
All Adults, df(yrs)
None: 3.6
Primary: 5.6
Secondary: 4.0
a 12 yrs: 1.6
Women, Concurrent Day
None: 0.91 (-0.06-1.89)
Primary: 0.31 (-0.06-0.68)
Secondary: 0.84 (0.33-1.36)
a 12 yrs: 0.60 (-0.32-1.52)
Women, Lag 1
None: 1.58 (0.58-2.58)
Primary: 0.79 (0.42-1.17)
Secondary: 0.76 (0.25-1.28)
a 12 yrs: 0.53 (-0.39-1.45)
Women, Distributed Lags 0-5
None: 1.15 (-0.48-2.80)
Primary: 1.05 (0.41-1.69)
Secondary: 1.29(0.40-2.19)
>	12 yrs: 1.06 (-0.27-2.41)
Women, df (yrs)
None: 2.6
Primary: 4.8
Secondary: 4.4
a 12 yrs: 1.0
Men, Concurrent Day
None: 0.05 (-1.02-1.12)
Primary: -0.11 (-0.5-0.28)
Secondary: 0.18 (-0.31-0.68)
>	12 yrs: 1.52 (0.70-2.35)
Men, Lag 1
None: 0.61 (-0.44-1.68)
Primary: 0.23 (-0.16-0.62)
Secondary: 0.49 (0.00-0.98)
July 2009
E-382
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
>	12 yrs: 1.03 (0.21-1.86)
Men, Distributed Lags 0-5
None: 2.08 (0.28-3.91)
Primary: 0.16 (-0.50-0.82)
Secondary: 1.27 (0.43-2.12)
>12 yrs: 1.98 (0.76-3.20)
Men, rf/(yrs)
None: 2.8
Primary: 4.8
Secondary: 4.4
a 12 yrs: 1.6
Percent increase (95% CI) in all-cause
adult mortality (>65yrs) by
educational level and sex
Mexico City
All Adults, Concurrent Day
None: 0.41 (-0.25-1.08)
Primary: 0.40 (-0.15-0.95)
Secondary: 0.50 (-0.01-1.01)
>	12 yrs: 1.51 (0.39-2.63)
All Adults, Lag 1
None: 0.20 (-0.47-0.87)
Primary: 0.80 (0.24-1.36)
Secondary: 0.60(0.09-1.12)
>12 yrs: 1.09 (-0.02-2.22)
All Adults, Distributed Lags 0-5
None: 0.27 (-0.83-1.38)
Primary: 0.99 (0.07-1.91)
Secondary: 0.30 (-0.56-1.16)
>12 yrs: 1.83 (0.09-3.59)
All Adults, df(yrs)
None: 5.6
Primary: 5.4
Secondary: 6.0
>	12 yrs: 3.2
Women, Concurrent Day
None: 0.49 (-0.30-1.29)
Primary: 0.39 (-0.33-1.11)
Secondary: 0.52 (-0.16-1.20)
>	12 yrs: 1.29 (0.12-2.48)
Women, Lag 1
None: 0.73 (-0.07-1.54)
Primary: 1.24 (0.52-1.97)
Secondary: 0.55 (-0.13-1.23)
>12 yrs: 1.50 (0.32-2.70)
Women, Distributed Lags 0-5
July 2009
E-383
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
None: 0.75 (-0.56-2.08)
Primary: 1.43 (0.29-2.59)
Secondary: 0.06 (-1.01-1.15)
>	12 yrs: 1.48 (0.10-2.87)
Women, rf/(yrs)
None: 5.4
Primary: 4.2
Secondary: 4.8
a 12 yrs: 0.6
Men, Concurrent Day
None: 0.90 (-0.23-2.04)
Primary: 0.37 (-0.40-1.16)
Secondary: 0.78 (0.07-1.49)
>12 yrs: 1.66 (0.30-3.04)
Men, Lag 1
None:-0.15 (-1.27-0.98)
Primary: 0.26 (-0.53-1.05)
Secondary: 0.93 (0.22-1.65)
>12 yrs: 0.95 (-0.41-2.32)
Men, Distributed Lags 0-5
None: 0.80 (-0.95-2.58)
Primary: 0.29 (-0.99-1.58)
Secondary: 1.06 (-0.08-2.21)
>	12 yrs: 1.76 (-0.35-3.91)
Men, rf/(yrs)
None: 3.8
Primary: 5.6
Secondary: 4.6
>	12 yrs: 3.8
Sao Paulo
All Adults, Concurrent Day
None: 0.60 (-0.48-1.70)
Primary: 0.59 (1.00-2.19)
Secondary: 1.21 (-0.01-2.44)
>12 yrs: 2.80 (1.67-3.94)
All Adults, Lag 1
None: 0.62 (-0.47-1.72)
Primary: 1.48 (0.89-2.07)
Secondary: 2.31 (1.08-3.55)
>12 yrs: 2.52 (1.40-3.66)
All Adults, Distributed Lags 0-5
None: 0.91 (-0.84-2.69)
Primary: 1.73 (0.79-2.67)
Secondary: 3.25 (1.39-5.16)
>12 yrs: 3.63 (2.01-5.29)
July 2009
E-384
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
All Adults, rf/(yrs)
None: 4.0
Primary: 3.8
Secondary: 2.6
a 12 yrs: 1.6
Women, Concurrent Day
None: 1.82 (0.71-2.94)
Primary: 1.84 (1.05-2.64)
Secondary: 0.62 (-0.55-1.81)
>12 yrs: 1.00 (-0.27-2.29)
Women, Lag 1
None: 1.36 (0.25-2.49)
Primary: 1.76 (0.97-2.56)
Secondary: 1.57 (0.39-2.76)
>	12 yrs: 1.39 (0.12-2.68)
Women, Distributed Lags 0-5
None: 1.80 (0.12-3.51)
Primary: 1.97 (0.73-3.22)
Secondary: 1.89(0.19-3.61)
>12 yrs: 2.53 (0.70-4.40)
Women, df (yrs)
None: 2.4
Primary: 3.4
Secondary: 1.2
>	12 yrs: 0.8
Men, Concurrent Day
None:-0.67 (-2.50-1.19)
Primary: 1.82 (1.00-2.65)
Secondary: 2.46 (1.31-3.63)
>12 yrs: 1.73 (0.47-3.00)
Men, Lag 1
None: -0.59 (-2.42-1.26)
Primary: 1.59 (0.78-2.41)
Secondary: 2.64(1.49-3.80)
>12 yrs: 0.89 (-0.35-2.15)
Men, Distributed Lags 0-5
None: 1.50 (-1.52-4.60)
Primary: 2.46 (1.20-3.74)
Secondary: 2.24 (0.56-3.95)
>12 yrs: 1.45 (-0.34-3.29)
Men, rf/(yrs)
None: 4.6
Primary: 3.0
Secondary: 0.8
a 12 yrs: 1.0
July 2009
E-385
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Santiago
All Adults, Concurrent Day
None: 1.49 (0.54-2.45)
Primary: 0.28 (-0.03-0.59)
Secondary: 0.58 (0.13-1.04)
>	12 yrs: 2.32 (1.50-3.15)
All Adults, Lag 1
None: 2.20 (1.24-3.17)
Primary: 0.74 (0.43-1.05)
Secondary: 0.64 (0.20-1.11)
>12 yrs: 2.20 (1.36-3.04)
All Adults, Distributed Lags 0-5
None: 3.21 (1.54-4.90)
Primary: 0.92 (0.38-1.46)
Secondary: 1.46 (0.67-2.25)
>12 yrs: 4.02 (2.78-5.27)
All Adults, df(yrs)
None: 3.8
Primary: 5.2
Secondary: 4.0
>	12 yrs: 1.8
Women, Concurrent Day
None: 1.39 (0.41-2.39)
Primary: 0.4 (0.01-0.8)
Secondary: 0.91 (0.29-1.53)
>	12 yrs: 0.87 (-0.02-1.78)
Women, Lag 1
None: 1.83 (0.83-2.85)
Primary: 0.98 (0.58-1.38)
Secondary: 0.73 (0.11-1.35)
a 12 yrs: 0.76 (-0.15-1.68)
Women, Distributed Lags 0-5
None: 2.47 (0.85-4.11)
Primary: 1.2(0.52-1.88)
Secondary: 1.71 (0.65-2.78)
>	12 yrs: 0.87 (-0.02-1.78)
Women, df (yrs)
None: 2.4
Primary: 4.8
Secondary: 4.4
a 12 yrs: 0.6
Men, Concurrent Day
None: 0.54 (-0.51-1.61)
Primary: 0.34 (-0.12-0.80)
Secondary: 0.25 (-0.40-0.91)
July 2009
E-386
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
>	12 yrs: 1.97 (1.09-2.86)
Men, Lag 1
None: 0.84 (-0.21-1.91)
Primary: 0.43 (-0.03-0.89)
Secondary: 0.61 (-0.04-1.26)
>	12 yrs: 1.57 (0.67-2.46)
Men, Distributed Lags 0-5
None: 2.41 (0.64-4.22)
Primary: 0.80 (0.02-1.59)
Secondary: 1.58 (0.45-2.71)
>12 yrs: 2.99 (1.66-4.33)
Men, rf/ (yrs)
None: 2.0
Primary: 4.4
Secondary: 4.4
>	12 yrs: 1.8
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
Winter:
¦0.4% (-0.30, 0.21)
0
0.15% (-0.08, 0.39)
1
0.10% (-0.13, 0.33)
2
Spring:
0.32% (0.08, 0.56)
0
0.14% (-0.14,0.42)
1
0.05% (-0.21,0.32)
2
Summer:
0.13% (-0.11,0.37)
0
0.36% (0.11,0.61)
1
¦0.03% (-0.27, 0.21)
2
Fall:
0.05% (-0.16, 0.25)
0
0.14% (-0.06, 0.34)
1
Reference: Peng et al. (2005, 087463)
Period of Study: 1987-2000
Location: 100 U.S. cities (NMMAPS)
Outcome: Mortality:
Non-accidental
Study Design: Time-series
Statistical Analyses: Bayesian
semiparametric hierarchical models
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Median (SD) unit: 27.1
Range (Min, Max): (13.2, 48.7)
Copollutant (correlation): NR
July 2009
E-387
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0.13% (-0.08, 0.35)
2
All Seasons:
0.09% (-0.01,0.19)
0
0.19% (0.10, 0.28)
1
0.08% (-0.03, 0.19)
2
PMio only (45 cities):
Winter: 0.15% (-0.16, 0.45)
Spring: 0.13% (-0.21, 0.48)
Summer: 0.30% (-0.10, 0.69)
Fall: 0.07% (-0.23, 0.37)
PMio + O3 (45 cities):
Winter: 0.18% (-0.16, 0.52)
Spring: 0.10% (-0.30, 0.49)
Summer: 0.33% (-0.14, 0.81)
Fall: 0.08% (-0.25, 0.41)
PMio + O3 (45 cities):
Winter: 0.13% (-0.24, 0.49)
Spring: 0.1% 9( 0.18, 0.56)
Summer: 0.28% (-0.13, 0.70)
Fall:-0.01% (-0.34, 0.31)
PMio + NO2 (45 cities):
Winter: 0.21% (-0.18, 0.60)
Spring: 0.19% (-0.17, 0.54)
Summer: 0.34% (0.01, 0.68)
Fall: 0.13% (-0.12, 0.39)
July 2009
E-388
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Reference: Penttinen et al. (2004,
087432)
Period of Study: 1988-1996
Location: Helsinki, Finland
Outcome: Mortality:
Pollutant: PMio
Increment: 10 //g/m3
Total (non-accidental) (< 800)
Cardiovascular (390-459)
Respiratory (460-519)
Averaging Time: 24-h avg
Median (SD) unit: 21 //gfm3
Range (Min, Max): (0.2, 213)
% Increase (Lower CI, Upper CI)
lag:
Total (non-accidental)

Study Design: Time-series
Copollutant (correlation):
¦0.23% (-1.47,1.01)

Statistical Analyses: Poisson GAM,
LOESS
Age Groups: 15-64
65-74
>75
O3: r - -0.09
NO?: r - 0.50
CO: r - 0.45
SO2: r - 0.61
TSP: r - 0.72
0
0.88% (-0.32, 2.08)
1
0.11 (-0.51,0.73)
0-3 avg
Cardiovascular



¦1.22% (-3.00, 0.56)
0
0.63% (-1.09, 2.35)
1






I
0.08% (-0.96, 0.81)



0-3 avg



Respiratory



3.94% (0.01,7.87)
0
3.96% (0.11,7.81)
1






1
2.13% (0.03, 4.22)



0-3 avg
July 2009
E-389
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Qian et al. (2007, 093054)
Period of Study: 2001-2004
Location: Wuhan, China
Outcome: Mortality:
Total (non-accidental) (< 800)
Cardiovascular (390-459)
Stroke (430-438)
Cardiac Diseases (390-398)
Respiratory (460-519)
Cardiopulmonary
Study Design: Time-series
Statistical Analyses: Poisson GAM,
natural splines
Age Groups: All ages
<45
>45
<65
> 65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): 141.8 3
Range (Min, Max): (24.8, 477.8)
Copollutant (correlation):
NO?
SO?
0s
July 2009
E-390
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
Non-accidental
0.36% (0.19, 0.53)
0
0.28% (0.12, 0.45)
1
0.43% (0.24, 0.62)
0-1
0.08% (-0.15, 0.31)
0-4
<45
0.28% (-0.26, 0.82)
0
0.45% (-0.06, 0.96)
1
0.53% (-0.08,1.13)
0-1
0.41% (-0.31,1.13)
0-4
>45
0.36% (0.19, 0.54)
0
0.27% (0.10, 0.44)
1
0.42% (0.22, 0.62)
0-1
0.05% (-0.18, 0.29)
0-4
<65
0.20% (-0.08, 0.49)
0
0.25% (-0.03, 0.52)
1
0.33% (0.01,0.66)
0-1
0.01% (-0.38, 0.39)
0-4
>65
0.41% (0.21,0.61)
0
0.30% (0.10, 0.49)
1
0.46% (0.24, 0.69)
0-1
0.10% (-0.16, 0.37)
° tRAFT- DO NOT CITE OR QUOTE
Cardiovascular
0.51% (0.28, 0.75)
0

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Qian et al. (2008,156894)
Period of Study: 712001-612004
Location: Wuhan, China
Outcome: Mortality:
Total (non-accidental) (< 800)
Cardiovascular (390-459)
Stroke (430-438)
Cardiac diseases (390-398, 410-429)
Respiratory (460-519)
Cardiopulmonary (390-459,460-519)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural splines and penalized splines
Age Groups: All ages
<65
> 65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD):
Normal temperature: 145.7 (64.6)
Low temperature: 117.3 (49.5)
High temperature: 96.3 (27.9)
Range (Min, Max): NR
Copollutant (correlation):
Normal temperature:
NO?: r - 0172
SO?: r - 0.59
O3: r - 0.06
Low temperature:
NO?: r - 0.83
SO2: r - 0.74
O3: r - 0.19
High temperature:
NO?: r - 0.68
SO2: r - 0.15
O3: r - 0.65
July 2009
E-391
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
Non-accidental:
Normal:
All ages: 0.36 (0.17,0.56)
0-1
<65:0.23(-0.10, 0.56)
0-1
>	65: 0.51 (0.18, 0.64)
0-1
PM10+NO2: 0.07 (-0.17, 0.30)
0-1
PM10+SO2: 0.27 (0.06, 0.47)
0-1
PMio+Os: 0.38 (0.18, 0.58)
0-1
Low:
All ages: 0.62 (-0.09,1.34)
0-1
<65:1.78(0.52, 3.05)
0-1
>	65: 0.22 (-0.61,1.05)
0-1
PM10+NO2: 0.24 (-0.49, 0.97)
0-1
PM10+SO2: 0.45 (-0.27,1.17)
0-1
PMio+Os: 0.72 (0.00,1.44)
0-1
High:
All ages: 2.20 (0.74, 3.68)
0-1
<65:2.34(-0.09, 4.83)
0-1
>	65: 2.14(0.42, 3.89)
0-1
PM10+NO2: 1.87 (0.42, 3.35)
0-1
PM10+SO2: 2.12 (0.67, 3.60)
0-1
PMio+Os: 2.15 (0.55, 3.77)
0-1
Cardiovascular:
Normal:
All ages: 0.39 (0.11,0.66)
0-1
OR QUOTE
0-1
>	65: 0.44 (0.14,0.74)
0-1

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Reference: Ren et al. (2006, 092824)
Outcome: Mortality:
Pollutant: PMio
The study presents quantitative results



associated with an incremental increase
Period of Study: 111996-1212001
Non-accidental
Averaging Time: 24-h avg
in temperature, not PMio.
Location: Brisbane, Australia
Cardiovascular (390-448)
Mean (SD): 15.84


Study Design: Time-series
Range (Min, Max): (2.5, 60)


Statistical Analyses: Poisson GAM,
Copollutant: 0:i


cubic spline



Age Groups: All ages


July 2009
E-392
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Roberts (2004, 087924)
Period of Study: 1987-1994
Location: Cook County, Illinois
Allegheny County, Pennsylvania
Outcome: Mortality:
Pollutant: PMio
Non-accidental (< 800)
Averaging Time: 24-h avg
Study Design: Time-series
Median (SD) unit:
Statistical Analyses: Poisson GAM,
Cook County
smooth splines

Lower Temp.: 29.24
Poisson GLM, natural cubic splines

Middle Temp.: 30.03
Age Groups: a 65

UpperTemp.: 52.76

Allegheny County

Lower Temp.: 16.50

Middle Temp.: 24.97

UpperTemp.: 55.42

Range (10th, 90th):

Cook County

Lower Tem.: (16.42, 46.42)

Middle Temp.: (14.79, 56.33)

UpperTemp.: (30.81, 82.81)

Allegheny County

Lower Temp.: (5.14, 34.54)

Middle Temp.: (8.91,57.91)

UpperTemp.: (30.91, 88.99)
July 2009
E-393
Increment: 10 //g/m3
% Increase (SE)
lag:
GLM
Cook
~	- 0.5
No Interaction: 0.288% (0.157)
0
Low Temp.:-0.272% (0.380)
0
Middle Temp.: 0.344% (0.165)
0
Upper Temp.: 0.281% (0.239)
0
No Interaction: 0.359% (0.149)
1
Low Temp.:-0.168% (0.372)
1
Middle Temp.: 0.361% (0.156)
1
UpperTemp.: 0.616% (0.250)
1
No Interaction: 0.465% (0.176)
0-1 ma
Low Temp.: 0.043% (0.397)
0-1 ma
Middle Temp.: 0.506% (0.184)
0-1 ma
UpperTemp.: 0.464% (0.256)
0-1 ma
No Interaction: 0.633% (0.214)
0-3 ma
Low Temp.: 0.365% (0.419)
0-3 ma
Middle Temp.: 0.638% (0.222)
0-3 ma
UpperTemp.: 0.718% (0.295)
0-3 ma
~	- 1
No Interaction: 0.117% (0.157)
0
Low Temp.: -0.351% (0.406)
0
Middle Temp.: 0.161% (0.165)
0
UpperTemp.: 0.096% (0.264)
0
No^^tjo^g.J^f.^l^R QU0TE
1
Low Temp.: -0.366% (0.397)
1

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Roberts (2004, 087924)
Period of Study: 1987-1994
Location: Cook County, Illinois
Allegheny County, Pennsylvania
Outcome: Mortality:
Non-accidental
Study Design: Time-series
Statistical Analyses: Poisson GLM
Age Groups: a 65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max):
Max - 89
The study does not present quantitative
results.
Reference: Roberts ((2005, 087992)
Period of Study: Cook County: 1987-
2000. Allegheny County: 1987-1998
Location: Cook County, Illinois
Allegheny County, Pennsylvania
Outcome: Mortality:
Non-accidental
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups: a 65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: NR
PISE)
lag:
Standard Model
Cook County
0.000127 (0.000264)
0
¦0.000042 (0.000249)
1
¦0.000441 (0.000246)
2
Allegheny County
0.000693 (0.000437)
0
0.000356 (0.000423)
1
0.000524 (0.000415)
2
Moving Total Model
Cook County
0.000150 (0.000187)
k - 2
¦0.000047 (0.000153)
k - 3
0.000009 (0.000133)
k - 4
Allegheny County
0.000633 (0.000310)
k - 2
0.000542 (0.000255)
k - 3
0.000598 (0.000351)
k - 4
July 2009
E-394
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Roberts (2006, 089762)
Period of Study: 1987-2000
Location: Cook County, Illinois
Suffolk County, Massachusetts
(NMMAPS)
Outcome: Mortality:
Non-accidental
Study Design: Time-series
Statistical Analyses: Poisson GLM
Age Groups: a 65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): Cook County: 33.7 (19.4)
Suffolk County: 25.9 (11.8)
Range (10th, 90th):
Cook County: (13.4, 58.1)
Suffolk County: (14.0, 41.7)
Copollutant (correlation):
Cook County
CO: r - 0.30
NO?: r - 0.53
SO?: r - 0.45
O3: r - 0.44
Suffolk County
CO: r - 0.33
NO?: r - 0.43
SO2: r - 0.23
O3: r = 0.36
Increment:
Cook County: 19.4/yg/m3
Suffolk County: 14.0/yg/m3
% Increase (SD)
lag:
Cook County
Standard Model: 0.49% (0.25)
0
Proposed Model: 0.29% (0.16)
0
Standard Model: 0.67% (0.25)
0-2 avg
Proposed Model: 0.49% (0.25)
0-2 avg
Suffolk County
Standard Model: 0.88% (1.27)
0
Proposed Model: 0.85% (0.84)
0
Standard Model: 1.60% (0.71)
0-2 avg
Proposed Model: 1.35% (0.73)
0-2 avg
Reference: Roberts and Martin (2006,
097799)
Period of Study: 1987-2000
Location: Cook County, Illinois
(NMMAPS)
Outcome: Mortality: Non-accidental
Study Design: Time-series
Statistical Analyses: Dose-response
1.	Piecewise linear relationship (no-
threshold) with change point at 25 //gfm3
and 50 //g|m3
2.	Piecewise linear relationship
(threshold), exposure below 25 /yglm3 no
effect, and exposures above 50 /yglm3
having a different effect then exposures
between 25 /yglm3 and 50 /yglm3
Age Groups: a 65
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD): NR
IQR (25th, 75th):
(23.9, 45.4)
Suffolk County: (14.0, 41.7)
Copollutant (correlation): NR
The study does not present quantitative
results.
Reference: Roberts and Martin (2006,
088670)
Period of Study: 1987-2000
Location: 109 U.S. cities (NMMAPS)
Outcome: Mortality: Non-accidental
Pollutant: PM10
Increment: NR
Cardiorespiratory
Study Design: Time-series
Statistical Analyses: Poisson
Averaging Time: 24-h avg
Mean (SD): NR
IQR (25th, 75th): NR
P x 1000 (SE x 1000)
lag:
Non-accidental

2-stage Bayesian hierarchical model
Copollutant (correlation): NR
Model 1

Age Groups: All ages

Base df: 0.079 (0.050)
0
Double df: 0.044 (0.046)
0
Half df: 0.107 (0.052)
0






July 2009
E-395
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Base df: 0.180 (0.044)
1
Double df: 0.149 (0.047)
1
Half df: 0.254 (0.048)
1
Base df: 0.059 (0.056)
2
Double df: 0.024 (0.056)
2
Half df: 0.143 (0.054)
2
Model 2
Base df: 0.115 (0.037)
0-2 ma
Double df: 0.107 (0.034)
0-2 ma
Half df: 0.145 (0.039)
0-2 ma
Cardio- respiratory
Model 1
Base df: 0.103 (0.068)
0
Double df: 0.056 (0.067)
0
Half df: 0.134 (0.066)
0
Base df: 0.232 (0.060)
1
Double df: 0.179(0.067)
1
Half df: 0.309 (0.059)
1
Base df: 0.210(0.078)
2
Double df: 0.144(0.075)
2
Half df: 0.305 (0.079)
2
Model 2
Base df: 0.168 (0.047)
0-2 ma
Double df: 0.140 (0.044)
0-2 ma
Half df: 0.196 (0.051)
July 2009
E-396
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0-2 ma
Notes: Model 1 uses current day's
mortality count, while Model 2 uses a 3-
day moving total mortality count.
Reference: Roberts and Martin (2007,
156917)
Period of Study: 1987-2000
Location: 8 U.S. cities and > 100 U.S.
cities (NMMAPS)
Outcome: Mortality: Total (non-
accidental)
Cardiorespiratory
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Increment: 10 //g/m3
P x 1000 (SE x 1000)
lag:
8 U.S. cities
Distributed Lag Model: 0.229
0-2
Weighted Model: 0.315
0-2
Standard Model:
0.276
0
¦0.062
1
0.476
2
90 U.S. cities
Total (non-accidental)
Standard Model:
0.078 (0.039)
0
0.182 (0.037)
1
0.108 (0.036)
2
Moving Total Model: 0.131 (0.023)
0-2
Weighted Model: 0.274 (0.075)
0-2
Cardiorespiratory
Standard Model:
0.096 (0.055)
0
0.232 (0.053)
1
0.226 (0.051)
2
Moving Total Model: 0.174 (0.032)
0-2
Weighted Model: 0.389 (0.105)
0-2
Notes: The 8 U.S. cities consist of
Chicago, Cleveland, Denver, El Paso,
^^Houston^JVIashville^^Pittsburgh^^an^SaJ^
July 2009
E-397
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Lake City.
Reference: Roberts and Martin (2007,
156916)
Period of Study: 1987-2000
Location: 10 U.S. cities (NMMAPS)
Outcome: Mortality: Non-accidental
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups: a 65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD):
Anchorage: 27.32
Chicago: 36.95
Cleveland: 39.83
Detroit: 40.78
El Paso: 40.14
Minneapolis/St. Paul: 28.01
Pittsburgh: 35.09
Salt Lake City: 37.40
Seattle: 28.72
Spokane: 34.52
Range (Min, Max): NR
Increment: NR
P Coefficient (SE)
lag:
Pooled Estimates
Combined Model (Unconstrained
Distributed Lag Model + Piecewise Linear
Dose-Response Function)
Change-point: 60//g|m3
Slope below: 0.00130(0.00016)
0-5
Slope above: -0.00163 (0.00026)
0-5
Change-point: 30//g|m3
Slope below: 0.00014 (0.00039)
0-5
Slope above: -0.00003 (0.00015)
0-5
Piecewise Linear Dose-Response Model
Change-point: 60//g|m3
Slope below: 0.00044 (0.00011)
3-day ma
Slope above: -0.00077 (0.00020)
3- day ma
Change-point: 30//g|m3
Slope below: 0.00022 (0.00026)
3-day ma
Slope above: -0.00004 (0.00011)
3-day ma
Polynomial Distributed Lag Model (degree
2)
0.00046 (0.00011)
0-5
Reference: Samoli et al. (2005, 087436) Outcome: Mortality:
Period of Study: 1990-1997	All-cause (non-accidental) (< 800)
Location: 22 European cities (APHEA-2) Cardiovascular (390-459)
Respiratory (460-519)
Study Design: Time-series
Statistical Analyses: Hierarchical
modeling:
1.	Poisson GAM, penalized splines
2.	Multivariate modeling
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Median (SD) unit:
Range: (Stockholm: 14/yg/m3 to Torino:
65/yg/m3)
Percentile (90th):
Range: (Stockholm: 27 /yglm3 to Torino:
129/yg/m3)
Copollutant (correlation): BS
The study does not present quantitative
results.
July 2009
E-398
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Schwartz (2004, 078998)
Period of Study: 1986-1993
Location: 14 U.S. cities
Outcome: Mortality:
Non-accidental (< 800)
Study Design: Case-crossover
Time-series
Statistical Analyses: Conditional
logistic regression
Poisson
Age Groups: All ages
Notes: Case days matched to referent
days that had the same temperature.
Pollutant: PM10
Increment: 10 //g/m3
Averaging Time: 24-h avg
% Increase (Lower CI, Upper CI)
Mean (SD): NR
lag:
Range (Min, Max): NR
Overall:
Copollutant (correlation): NR
Two stage: 0.36% (0.22, 0.50)
1

1
Single stage: 0.33% (0.19, 0.46)
1

More winter temperature lags:

Two Stage: 0.39% (0.23, 0.56)
1

1
One stage: 0.32% (0.19, 0.46)
i

i
Time stratified with temperature

matching:

Two Stage: 0.39% (0.19, 0.58)
i

1
One Stage: 0.53% (0.34, 0.72)
i

i
Poisson regression:

0.40% (0.18, 0.62)
1
Pollutant: PM10
Increment: 10 //g/m3
Averaging Time: 24-h avg
P x 1000 (SE x 1000)
Median (SD) unit: Range: 23 to
lag:
36//g/m3

Matched on CO: 0.527 (0.251)
IQR (25th, 75th):
0-1 avg

Range 25th: 17 to 24//g/m3

Matched on O3: 0.451 (0.170)
Range 75th: 31 to 57 //g/m3
0-1 avg
Copollutant (correlation): CO
Matched on NO2: 0.784 (0.185)
SO2


0-1 avg
NO2


Matched on SO2: 0.811 (0.175)
0s


0-1 avg
Reference: Schwartz (2004, 053506)
Period of Study: 1986-1993
Location: 14 U.S. cities
Outcome: Mortality:
Non-accidental (< 800)
Study Design: Case-crossover
Statistical Analyses: Time-stratified
conditional logistic regression
Age Groups: All ages
Notes: Case days matched to referent
days based on concentration of gaseous
air pollutants. Matched on the following
conditions:
1.	24-h avg SO2 within 1 ppb
2.	Daily-maximum O3 within 2 ppb
3.	24-h avg NO2 within 1 ppb
4.	24-h avg CO within 0.03 ppm
Reference: Sharovsky et al. (2004,
156976)
Period of Study: 7/1996-6/1998
Location: Sao Paulo, Brazil
Outcome: Mortality:
Myocardial infarction
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups: a 35
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD): 58.2 (25.8)
Range (Min, Max): (23,186)
Increment: 10 //g/m3
PISE)
lag:
PM10: 0.001 (0.001)
Copollutant (correlation): CO: r - 0.73 PM10+CO+SO2: 0.0004 (0.0008)
SO2: r - 0.72
July 2009
E-399
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Simpson et al. (2005,
087438)
Period of Study: 1/1996-12/1999
Location: 4 Australian cities
Outcome: Mortality:
Pollutant: PM10
Non-accidental (< 800)
Averaging Time: 24-h avg
Cardiovascular (390-459)
Mean (SD):
Respiratory (460-519)
Brisbane: 16.60
Study Design: Time-series
Sydney: 16.30
meta-analysis
Melbourne: 18.20
Statistical Analyses: Poisson GAM,
Range (Min, Max):
natural splines

Brisbane: (2.6, 57.6)
Poisson GLM, natural splines

Sydney: (3.7, 75.5)
Age Groups: All ages

Melbourne: (3.3, 51.9)

Copollutant:
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
0.2% (-0.8,1.2)
PM2.6
CO
NO?
Reference: Slaughter et al. (2005,
073854)
Period of Study: 1/1995-12/1999
Location: Spokane, Washington
Outcome: Mortality:
Non-accidental (< 800)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural splines
Age Groups: All ages
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD): NR
Range (9th, 95th): (7.9, 41.9)//g/m3
Copollutant (correlation):
PM10
PM 10-2.5: r - 0.94
CO: r - 0.32
Increment:: 25 //g/m3
Relative Risk (Lower CI, Upper CI)
lag:
1.00(0.97,1.03)
1
0.98(0.95,1.01)
2
1.00(0.97,1.03)
3
Reference: Staniswalis et al. (2005,
087473)
Period of Study: 1992-1995
Location: El Paso, Texas
Outcome: Mortality:
Pollutant: PM10
Increment: 10 //g/m3
Non-accidental (< 800)
Study Design: Time-series
Statistical Analyses: Poisson
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max):
% Increase (Lower CI, Upper CI)
lag:
Poisson regression: 1.7%

Principal component analysis (PCA)
(0.2,133.4)
3

Age Groups: All
Notes: The chemical composition and
size distribution of PM was not available,
therefore, the study used wind speed as a
surrogate variable for the PM10
composition.
PCA:
24-hly measurements: 2.06%
3
Daily avg: 1.7%
3
July 2009
E-400
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Stafoggia et al. (2008,
157005)
Period of Study: 1997-2004
Location: 9 Italian cities
Outcome:
Mortality:
Total (non-accidental) (< 800)
Cardiovascular (390-459)
Respiratory (460-519)
Other natural causes
Study Design: Time-stratified
crossover
Statistical Analyses:
Conditional logistic regression
Age Groups: a 35
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD) unit:
Bologna: 50.4 (31.7)
Florence: 37.5 (16.6)
Mestre: 48.1 (26.8)
Milan: 57.9 (38.0)
Palermo: 36.2(21.7)
Pisa: 35.1 (14.9)
Rome: 47.3 (19.9)
Taranto: 59.8 (18.9)
Turin: 71.5(38.1)
Range (Min, Max): NR
Copollutant (correlation): NR
July 2009
E-401
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
Cardiovascular
All year: 0.63% (0.31,1.38)
0-1
Winter: 0.15% (-0.29, 0.59)
0-1
Spring: 0.72% (-0.07,1.52)
0-1
Summer: 2.90% (1.14,4.69)
0-1
Fall: 1.37% (0.43,2.32)
0-1
Apparent Temperature
<	50th Percentile: 0.31% (-0.06, 0.67)
0-1
50th-75th Percentile: 2.05% (0.47, 3.66)
0-1
>	75th Percentile: 2.68% (1.20,4.17)
0-1
Respiratory
All year: 0.98% (0.27,1.70)
0-1
Winter: 0.41% (-0.67,1.51)
0-1
Spring: 2.99% (1.18, 4.83)
0-1
Summer: 3.89% (0.19, 7.73)
0-1
Fall: 0.45% (-1.11,2.03)
0-1
Apparent Temperature
<	50th Percentile: 0.54% (-0.47,1.57)
0-1
50th-75th Percentile: 3.15% (0.64, 5.73)
0-1
>	75th Percentile: 4.12% (0.44,7.93)
0-1
Other natural causes
All year: 0.37% (0.09, 0.66)
0-1
Winter: 0.14% (-0.36, 0.63)
0-1
Spring: 0.29% (-0.47,1.05)
0-1
Summer: 2.15% (0.90,3.42)
01DRAFT - DO NOT CITE OR QUOTE
Fall: 0.70% (-0.41,1.83)
0-1
Annarpnt Tpmnpratnrp

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Stolzel et al. (2007, 091374) Outcome:
Period of Study: 9/1995-8/2001 Mortality:
Location: Erfurt, Germany	Total (non-accidental) (< 800)
Cardio-respiratory (390-459,460-519,
785, 786)
Study Design: Time-series
Statistical Analyses:
Poisson GAM
Age Groups: All ages
Pollutant: PMio	Increment: 23/yg/m3
Averaging Time: 24-h avg	Relative Risk (Lower CI, Upper CI)
Mean (SD) unit:: 31.9 (23.2)	lag:
IQR (25th, 75th):	Total (non-accidental)
(16.5,39.5)	1.004(0.980
Copollutant (correlation):	1.029)
MCo i o.b: r - 0.85	0
MCo.oi-2.b: r - 0.84	1.004(0.981
NO: r- 0.54	1.027)
NO?: r - 0.62	1
CO: r - 0.50	0.998 (0.976
1.021)
2
0.984 (0.962
1.006)
3
0.993 (0.972
1.015)
4
0.990 (0.969
1.012)
5
Cardio- respiratory
1.007 (0.981
1.034)
0
1.006 (0.981
1.032)
1
0.996 (0.971
1.021)
2
0.977 (0.953
1.002)
3
0.994 (0.970
1.018)
4
0.993 (0.969
1.017)
5
July 2009
E-402
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Sullivan et al. (2003,
043156)
Period of Study:
1985-1994
Location: Western Washington
Outcome:
Out-of-hospital cardiac arrest
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups: 19-79
Study Population: Out-of-hospital cardiac
arrests: 1,206
Pollutant: PMio
Averaging Time: 24-h avg
Median (SD) unit:
Lag 0: 28.05
Lag 1: 27.97
Lag 2: 28.40
Range (Min, Max):
(7.38, 89.83)
Copollutant (correlation):
SO?
CO
Notes: Study used nephelometry to
measure particles and equated the
measurements to PM2.6 concentrations.
Increment:: 16.51 /yglm3
Odds Ratio (Lower CI, Upper CI)
lag:
Overall
1.05(0.87,1.27)
0
0.91 (0.75,1.11)
1
1.03(0.82,1.28)
2
Reference: Sunyer et al. (2002, 034835) Outcome: Mortality:
Period of Study: 1985-1995
Location: Barcelona, Spain
Respiratory mortality
Study Design: Case-crossover
Statistical Analyses: Condition logistic
regression
Age Groups: >14
Study population: Asthmatic individuals:
5,610
Pollutant: PM10
Averaging Time: 24-h avg
Median (SD) unit: 61.2
Range (Min, Max): (17.3, 240.7)
Copollutant:
BS
NO?
0s
SO?
CO
Increment: 32.7 /yglm3
Odds Ratio (Lower CI, Upper CI)
lag:
Asthmatic individuals with 1 ED visit
0.884 (0.672,1.162)
0-2 avg
Asthmatic individuals with >1 ED visit
1.084 (0.661,1.778)
0-2 avg
Asthma/COPD individuals with >1 ED
visit
1.011 (0.746,1.368)
0-2 avg
July 2009
E-403
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Touloumi et al. (2005,
087477)
Period of Study: 1990-1997
Location: 7 European cities (London,
Budapest, Stockholm, Zurich, Paris, Lyon,
Madrid) (APHEA2)
Outcome: Mortality:
Total (non-accidental) (< 800)
Cardiovascular (390-459)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
LOESS
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Median (SD) unit:
London: 25.1
Budapest: 40.2
Stockholm: 13.7
Zurich: 27.5
Paris: 22.2
Lyon: 38.5//
Madrid: 33.4
IQR (25th, 75th):
London: (20.3, 33.9)
Budapest: (34.3, 45.8)
Stockholm: (10.3,19.1)
Zurich: (19.2, 38.5)
Paris: (16.0, 33.0)
Lyon: (29.7, 50.4)
Madrid: (27.6, 41.0)
Copollutant (correlation): NR
Increment: 10 //g/m3
P 1x 1000) (SE 1x 1000)):
Total (non-accidental)
No control: 0.4834 (0.1095)
Reported Influenza Data
Count ID: 0.4967 (0.1089)
11 ID: 0.4740(0.1090)
Ml ID: 0.5019 (0.1096)
RI-ID: 0.4735 (0.1091)
SF ID: 0.6714(0.1080)
Estimated Influenza Data
APHEA-2: 0.5550(0.1076)
11 EID: 0.5640 (0.1073)
Ml EID: 0.5872 (0.1100)
Rl EID: 0.5872 (0.1074)
SF EID: 0.6641 (0.1073)
Cardiovascular
No control: 0.8432 (0.1665)
Reported Influenza Data
Count ID: 0.8896 (0.1662)
11 ID: 0.8545(0.1661)
Ml ID: 0.8693 (0.1674)
RI-ID: 0.8649 (0.1665)
SF ID: 1.0107 (0.1659)
Estimated Influenza Data
APHEA-2: 0.9389(0.1654)
11 EID: 0.9485 (0.1648)
Ml EID: 1.0440 (0.1686)
Rl EID: 0.9718 (0.1653)
SF EID: 1.0585 (0.1652)
Notes: 11 - one indicator for all
epidemics
M1 - multiple indicators, one per
epidemic
R1 - indicators for intervals indicating
the range of influenza counts
SF - separate smooth function during
epidemic periods.
July 2009
E-404
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Tsai et al. (2003, 050480)
Period of Study: 1994-2000
Location: Kaohsiung, Taiwan
Outcome: Mortality:
Total (non-accidental) (< 800)
Respiratory (460-519)
Circulatory (390-459)
Study Design: Bidirectional case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): 81.45
Range (Min, Max): (20.50, 232.00)
Copollutant:
SO?
NO?
CO
0s
Increment: 67.00 /yglm3
Odds Ratio (Lower CI, Upper CI)
lag:
Total (non-accidental)
1.000 (0.947,1.056)
0-2 avg
Respiratory
1.023 (0.829,1.264)
0-2 avg
Circulatory
0.971 (0.864,1.092)
0-2 avg
July 2009
E-405
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Vajanapoom et al. (2002,
042542)
Period of Study: 1992-1997
Location: Bangkok, Thailand
Outcome: Mortality:
Total (non-accidental) (< 800)
Respiratory (460-519)
Cardiovascular (390-459)
Other-causes
Study Design: Time-series
Statistical Analyses: Poisson GAM,
LOESS
Age Groups:
All ages
55-64
65-74
>75
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): 68.0 (23.9)
IQR (25th, 75th):
(50.1,80.7)
Copollutant (correlation): NR
Increment: 30 //gfm3
% Increase (Lower CI, Upper CI)
lag:
Total (non-accidental)
All ages: 2.3% (1.3,3.3)
0-4 ma
55-64: 1.5% (-0.8,3.9)
0-4 ma
65-74: 4.2% (2.0, 6.3)
0-4 ma
>75: 3.9% (2.1,5.6)
0-4 ma
Cardiovascular
All ages: 0.8% (-0.9, 2.4)
0
55-64:-2.5% (-6.3,1.3)
0
65-74: 2.9% (-0.7, 6.5)
0
>75:1.6% (-1.8, 5.0)
0
Respiratory
All ages: 5.1% (0.6, 9.6)
0-2 ma
55-64: 1.4% (-11.3,14.2)
0-2 ma
65-74: 2.8% (-9.5,15.2)
0-2 ma
>75: 10.2% (-0.1,20.5)
0-2 ma
Other-causes
All ages: 2.4% (1.3,3.5)
0-4 ma
55-64: 1.7% (-1.1,4.5)
0-4 ma
65-74: 5.6% (3.1,8.1)
0-4 ma
>75: 3.7% (1.8,5.6)
0-4 ma
July 2009
E-406
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Vedal et al. (2003, 039044)
Outcome: Mortality:
Pollutant: PM10
The study does not present quantitative



results
Period of Study: 111994-1211996
Total (non-accidental) (< 800)
Averaging Time: 24-h avg

Location: Vancouver, British Columbia,
Respiratory (460-519)
Mean (SD): 14.4 (5.9)

Canada




Cardiovascular (390-459)
Range (Min, Max): (4.1, 37.2)


Study Design: Time-series
Copollutant (correlation): O3: r
- 0.48

Statistical Analyses: Poisson GAM,
SO2: r - 0.76


LOESS




NO?: r - 0.84


Age Groups: All ages
CO: r - 0.71



Reference: Venners et al. (2003,
089931)
Period of Study: 1/1995-12/1995
Location: Chongqing, China
Outcome: Mortality:
Total (non-accidental) (< 800)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
cubic spline
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): 146.8
Range (Min, Max): (44.7, 666.2)
Copollutant: SO2
Notes: PM10 was measured for only 7
months of the study period.
Increment: 100/yg/m3
Relative Risk (Lower CI, Upper CI)
lag:
1.00(0.93,1.07)
0
0.98(0.91,1.04)
1
1.00(0.93,1.07)
2
0.96(0.90,1.03)
3
0.97 (0.90,1.03)
4
0.99(0.93,1.06)
5
July 2009
E-407
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Vichit-Vadakan et al. (2008,
157095)
Period of Study: 111999-1212003
Location: Bangkok, Thailand
Outcome (ICD10): Mortality:
Non-accidental (A00-R99)
Cardiovascular (I00-I99)
Ischemic heart diseases (I20-I25)
Stroke (I60-I69)
Conduction disorder (I44-I49)
Respiratory (J00-J98)
Lower Respiratory Infection (J10-J22)
C0PD (J40-J47)
Asthma (J45-J46)
Senility (R54)
Study Design: Time-series
Statistical Analyses: Poisson, natural
cubic spline
Age Groups: All ages
0-4
5-44
18-50
45-64
>50
>65
>75
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): 52.1 (20.1)
Range (Min, Max): (21.3,169.2)
Copollutant (correlation): NR
July 2009
E-408
Increment: 10 //g/m3
% Excess Risk (Lower CI, Upper CI)
lag:
Cause-specific mortality:
Nonaccidental: 1.3% (0.8,1.7)
0-1
Cardiovascular: 1.9% (0.8, 3.0)
0-1
Ischemic heart disease: 1.5% (-0.4, 3.5)
0-1
Stroke: 2.3% (0.6, 4.0)
0-1
Conduction disorders: -0.%3 (-5.9, 5.6)
0-1
Cardiovascular:
> 65 1.8 (0.2, 3.3)
0-1
Respiratory
All ages: 1.0 (-0.4, 2.4)
0-1
~ 1:14.6(2.9, 27.6)
0-1
>65: 1.3(-0.8, 3.3)
0-1
LRI:
<5: 7.7(-3.6, 20.3)
0-1
C0PD: 1.3 (-1.8, 4.4)
0-1
Asthma: 7.4 (1.1,14.1)
0-1
Senility: 1.8(0.7,2.8)
0-1
Age-specific for non-accidental
0-4: 0.2 (-2.0, 2.4)
0-1
5-44: 0.9(0.2,1.7)
0-1
18-50:1.2(0.5,1.9)
0-1
45-64: 1.1 (0.4,1.9)
0-1
>50: 1.4 (0.9,1.9)
0-1
>65: 1.5(0.9,2.1)
0-1
>75: 2.2(1.3,3.0)
01DRAFT - DO NOT CITE OR QUOTE
Sex-specific for non-accidental
Male: 1.2(0.7,1.7)
0-1

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Villeneuve et al. (2003,
055051)
Period of Study: 1986-1999
Location: Vancouver, Canada
Outcome: Mortality:
Non-accidental (< 800)
Cardiovascular (401-440)
Respiratory (460-519)
Cancer (140-239)
Study Design: Time-series
Statistical Analyses: Poisson, natural
splines
Age Groups: a 65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD):
Daily 14.0
Every 6th Day 19.6
Range (Min, Max):
Daily (3.8, 52.2)
Every 6th Day (3.5, 63.0)
Copollutant:
SO?
CO
NO?
0s
PM2.6
PMio-2.5
Increment: 15.4/yg/m3
% Increase (Lower CI, Upper CI)
lag:
Non-accidental
3.7% (-0.5, 8.0)
0-2 avg
2.6% (-0.9, 6.1)
0
2.7% (-0.7, 6.2)
1
1.9% (-1.4,5.3)
2
Cardiovascular
3.4% (-2.7, 9.8)
0-2 avg
5.1% (0.0,10.4)
0
1.3% (-3.8, 6.7)
1
0.6% (-4.3, 5.7)
2
Respiratory
PM10
0.1% (-9.5,10.8)
0-2 avg
1.0% (-7.5,10.4)
0
0.4% (-7.7, 9.3)
1
¦1.3% (-8.9,7.1)
2
Cancer
1.2% (-6.9,10.1)
0-2 avg
¦2.5% (-8.8, 4.3)
0
2.3% (-4.6, 9.6)
1
3.3% (-3.7,10.8)
2
July 2009
E-409
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Reference: Welty et al. (2008,157134)
Outcome: Mortality:
Pollutant: PMio
Increment: 10 //g/m3
Period of Study: 1987-2000
Total (non-accidental)
Averaging Time: 24-h avg
% Excess Risk (Lower CI, Upper CI)
Location: Chicago, Illinois
Study Design: Time-series
Mean (SD): NR
lag:

Statistical Analyses: Poisson—Gibbs
Sampler
Bayesian Distributed Lag Model
Range (Min, Max): NR
Poisson-Gibbs Sampler

Copollutant (correlation): NR
0.17% (0.01,0.34)
3

Age Groups: All ages

¦0.24% (-0.73, 0.23)
0-14
Unconstrained: -0.19% (-0.86, 0.48)
0-14
Bayesian Distributed Lag Model
¦0.21% (-0.86, 0.41)
0-14
Reference: Welty and Zeger (2005,
087484)
Period of Study: 1987-2000
Location: 100 U.S. cities (NMMAPS)
Outcome: Mortality:
Pollutant: PMio
Increment: 10 //g/m3
Total (non-accidental) (< 800)
Study Design: Time-series
Statistical Analyses: Bayesian
hierarchical model
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
% Increase (SE)
lag:
Distributed Lag Model: Seasonally-
Temporally Varying
Age Groups: All ages	Temperature variables: 0,1-2, 1-7, 1-14
S(t, 1 x years): 0.229 (0.053)
1
S(t, 2 x years): 0.220 (0.053)
1
S(t, 4 x years): 0.187 (0.050)
1
S(t, 8 x years): 0.178 (0.049)
1
Temperature variables: 0,1-2,1 -7,1-14,
0x1.2,0x1.7,1-2 x 1-7
S(t, 1 x years): 0.195 (0.048)
1
S(t, 2 x years): 0.200 (0.051)
1
S(t, 4 x years): 0.176 (0.050)
1
S(t, 8 x years): 0.149 (0.050)
1
Distributed Lag Model: Nonlinear
Temperature variables: 0,1-2,1-7,1-14
S(t, 4 x years): 0.239 (0.053)
1
Temperature variables: 0,1-2,1 -7,1-14,
0x1-2,0x1-7,1-2 x 1-7
S(t, 4 x years): 0.172 (0.045)
1
Temperature variables: S(0,2), S(1 -2,2),
S(1 -7,2), S(1 -14,2)
July 2009
E-410
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
S(t, 4 x years): 0.186 (0.046)
1
Temperature variables: S(0,2), S(1 -2,2),
S(1 -7,2), S(1 -14,2), S(0x 1-2,2), S(0x1-
7,2), S(1-2 x 1-72)
S(t, 4 x years): 0.189 (0.047)
1
Temperature variables: S(0,4), S(1 -2,4),
S(1 -7,4), S(1 -14,4)
S(t, 4 x years): 0.175 (0.046)
1
Temperature variables: S(0,4), S(1 -2,4),
S(1 -7,4), S(1 -14,4), S(0x 1-2,4), S(0x 1-
7,4), S(1-2 x 1-7,4)
S(t, 4 x years): 0.190 (0.048)
1
Temperature variables: 0,1-2,1-7
S(t, 4 x years): 0.252 (0.053)
1
Temperature variables: 0,1-2,1-7, 0x1-
2, 0x1-7,1-2 x 1-7
S(t, 4 x years): 0.186 (0.044)
1
Temperature variables: S(0,2), S(1 -2,2),
S(1-7,2)
S(t, 4 x years): 0.198 (0.046)
1
Temperature variables: S(0,2), S(1 -2,2),
S(1 -7,2), S(0x 1-2,2), S(0x 1-7,2), S(1-2
x 1-7,2)
S(t, 4 x years): 0.201 (0.047)
1
Temperature variables: S(0,4), S(1 -2,4),
S(1-7,4)
S(t, 4 x years): 0.189 (0.045)
1
Temperature variables: S(0,4), S(1 -2,4),
S(1 -7,4), S(0x 1-2,2), S(0x 1-7,4), S(1-2
x 1-7,2)
S(t, 4 x years): 0.205 (0.047)
1
Temperature variables: S(0,4), S(1 -2,4)
S(t, 4 x years): 0.250 (0.045)
1
Temperature variables: S(0,4), S(1 -2,4),
S(0x 1-2,4)
S(t, 4 x years): 0.253 (0.044)
1
Temperature variables: S(0,4)
S(t, 4 x years): 0.220 (0.045)
July 2009
E-411
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Notes: 0 indicates current-day
temperature
1-r indicates avg of lag 1 through lag r
temperature
S (, ~) indicates a natural spline smooth
with ~ degrees of freedom.
S (t, ~ x years) indicates the natural
spline smooth of time with degrees of
freedom equal to ~ x (number of years of
data).
Reference: Wong et al. (2007, 098391) Outcome: Mortality:
Period of Study: 1/1998-12/1998
Location: Hong Kong, China
Total (non-accidental) (< 800)
Cardiorespiratory (390-519)
Study Design: Main analysis: Time-
series
Sensitivity analysis: Case-crossover,
case-only
Statistical Analyses: Main analysis:
Poisson GAM
Sensitivity analysis: Conditional logistic
regression
Age Groups: £ 30
> 65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD):
48.1 (24.3)
Range (Min, Max):
(15.5,140.5)
Copollutant:
NO?
SO?
0s
Increment: 10 //g/m3
% Excess Risk (Lower CI, Upper CI)
lag:
Main Analysis
Non-accidental
Smokers: > 301:.80% (0.35, 3.26)
0
1.77% (0.46, 3.11)
2
>	65: 3.20% (1.36,5.07)
0
2.42% (0.73, 4.13)
2
Never-smokers
>	30: -0.37% (-2.23,1.52)
0
¦0.03% (-1.72,1.66)
2
>	65P -0.70% (-2.81,1.46)
0
¦0.13% (-2.04,1.80)
2
Cardiorespiratory
Smokers
>	30: 1.43% (-0.86, 3.78)
0
2.32% (0.24, 4.44)
2
>	65: 2.98% (0.47, 5.55)
0
2.61% (0.31,4.95)
2
Never-smokers
>	30: 0.02% (-2.75, 2.87)
0
¦0.79% (-3.33,1.82)
July 2009
E-412
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
2
>	65: 0.25% (-2.62, 3.19)
0
¦0.66% (-3.29, 2.04)
2
Sensitivity Analysis
Poisson Regression
Non-accidental
>30: 1.81% (0.21,3.44)
0
1.93% (0.32, 3.56)
2
1.99% (0.14,3.87)
0-3
>65: 2.31% (0.37, 4.29)
0
2.16% (0.20, 4.15)
2
2.57% (0.30, 4.89)
0-3
Cardiorespiratory
>	30: 1.04% (-1.45, 3.59)
0
2.18% (-0.35, 4.77)
2
1.66% (-1.24,4.64)
0-3
>	65: 1.69% (-0.93, 4.37)
0
2.44% (-0.23, 5.18)
2
2.30% (-0.80, 5.50)
0-3
Case-only: Logistic Regression
Non-accidental
>	30: 1.79% (0.21,3.37)
0
1.94% (0.33, 3.56)
2
>	65: 2.30% (0.42, 4.17)
0
2.16% (0.26, 4.07)
2
Cardiorespiratory
>30: 1.01% (-1.37,3.40)
July 2009
E-413
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0
2.16% (-0.28, 4.61)
2
>	65: 1.65% (-0.96, 4.27)
0
2.42% (-0.27, 5.12)
2
Case-crossover
Non-accidental
>	30: 2.54% (0.35, 4.78)
0
1.35% (-0.81,3.56)
2
>	65: 3.96% (1.37,6.63)
0
2.20% (-0.35, 4.81)
2
Cardiorespiratory
>	30: 0.48% (-2.74, 3.80)
0
3.24% (-0.03, 6.61)
2
>65: 2.17% (-1.40, 5.86)
0
3.43% (-0.13,7.13)
2
Reference: Wong et al. (2007, 093278) Outcome: Mortality:
Period of Study: 1/1998-12/1998
Location: Hong Kong, China
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD):
48.1 (24.3)
Range (Min, Max):
(15.5,140.5)
Copollutant:
Sensitivity analysis: Logistic regression NO2
Age Groups: £ 30
> 65
Total (non-accidental) (< 800)
Cardiorespiratory (390-519)
Study Design: Main analysis: Time-
series
Sensitivity analysis: Case-only
Statistical Analyses: Main analysis:
Poisson GAM, natural cubic spline
SO?
0s
Increment: 10 //g/m3
% Excess Risk (Lower CI, Upper CI)
lag:
Non-accidental
Exercise
>30: 0.13% (-1.16,1.44)
1
>	65: 0.24% (-1.16,1.67)
1
Never-exercise
>	30: 1.04% (0.07,2.02)
1
>	65: 1.26% (0.27,2.27)
1
Cardiorespiratory
Exercise
>	30: 0.46% (-1.43, 2.39)
1
>	65: 0.30% (-1.65, 2.29)
July 2009
E-414
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1
Never-exercise
>	30: 0.97% (-0.36, 2.32)
1
>	65: 0.98% (-0.45, 2.43)
1
Difference in % Excess Risk (Exercise vs.
Never-Exercise)
Non-accidental
Poisson Regression
>	30: -2.86% (-4.03 to-1.67)
1
>	65: -3.06% (-4.37 to-1.74)
1
Case-only
>	30: -2.91% (-4.04 to-1.77)
1
>	65: -3.12% (-4.38 to-1.84)
1
Cardiorespiratory
Poisson regression
>	30: -2.55% (-4.32 to-0.75)
1
>	65: -2.64% (-4.48 to-0.76)
1
Case-only
>	30: -2.63% (-4.32 to-0.92)
1
>	65: -2.73% (-4.50 to-0.92)
1
Adjusted Case-only
Non-accidental
Sex
>	30: -2.88% (-1.73 to-4.01)
1
>	65: -3.09% (-1.82 to-4.35)
1
Education
>	30: -2.94% (-1.80 to -4.07)
1
>	65: -3.18% (-1.90 to-4.44)
1
Job
>	30: -2.88% (-1.74 to -4.02)
1
July 2009
E-415
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
>	65: -3.11% (-1.83 to-4.37)
1
Smoking
>	30: -2.82% (-1.66 to-3.96)
1
>	65: -2.97% (-1.68 to-4.25)
1
Illness time
>	30: -2.94% (-1.80 to -4.07)
1
>	65: -3.16% (-1.88 to-4.42)
1
Cardiorespiratory
Sex
>	30: -2.61% (-0.89 to -4.29)
1
>	65: -2.71% (-0.90 to-4.48)
1
Education
>	30: -2.58% (-0.85 to-4.27)
1
>	65: -2.77% (-0.95 to -4.54)
1
Job
>	30: -2.68% (-0.96 to-4.37)
1
>	65: -2.68% (-0.88 to -4.46)
1
Smoking
>	30: -2.46% (-0.73 to-4.17)
1
>	65: -2.50% (-0.68 to-4.29)
1
Illness Time
>	30: -2.63% (-0.91 to -4.32)
1
>	65: -2.73% (-0.92 to-4.51)
1
Case-only by Exercise Group (Never as
Reference)
Non-accidental
>30
Low:-3.34% (-5.77 to -0.85)
1
Moderate: -6.32% (-8.55 to -4.03)
July 2009
E-416
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
1
High: -1.74% (-3.06 to-0.40)
1
>65
Low:-3.79% (-6.67 to 0.82)
1
Moderate:-7.78% (-10.39 to-5.10)
1
High: -1.77% (-3.21 to-0.31)
1
Cardiorespiratory
>30
Low:-3.95% (-7.77,0.04)
1
Moderate: -8.50% (-11.84 to -5.02)
1
High:-0.62% (-2.58,1.38)
1
>65
Low:-3.97% (-8.17,0.43)
1
Moderate: -9.42% (-13.00 to -5.69)
1
High:-0.68% (-2.71,1.38)
1
Increment: 10 //g/m3
Relative Risk (Lower CI, Upper CI)
lag:
Respiratory
1.008 (1.001 to 1.014)
1
C0PD
1.017 (1.002,1.033)
0-3
Pneumonia & Influenza
1.007 (0.999,1.015)
2
Cardiovascular
1.003 (0.998,1.016)
2
IHD
1.013(1.001,1.025)
0-3
Cerebrovascular
1.007 (0.998,1.016)
Reference: Wong et al. (2002, 025436) Outcome: Mortality:
Period of Study: 1995-1998
Location: Hong Kong, China
Respiratory (461-519)
C0PD (490-496)
Pneumonia & Influenza (480-487)
Cardiovascular (390-459)
IHD (410-414)
Cerebrovascular (430-438)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups: £ 30
> 65
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD):
51.53 (24.79)
Range (Min, Max):
(14.05,163.79)
Copollutant (correlation):
NO?: r - 0.780
SO?: r - 0.344
O3: r - 0.538
July 2009
E-417
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Respiratory
PM10+SO2+O3+NO2: 1.005(0.992,
1.010)
1
COPD
PM10+SO2+O3+NO2: 0.991 (0.968,
1.015)
0-3
PM10+O3+NO2: 0.993 (0.970,1.016)
0-3
Pneumonia 81 Influenza
PM10+SO2+O3+NO2: 1.002(0.991,
1.013)
2
IHD
0.994 (0.978,1.009)
0-3
Reference: Wong et al. (2008,157152) Outcome (ICD10): Mortality:
Period of Study: Bangkok: 1999-2003 Natural causes (A00-R99)
Hong Kong: 1996-2002
Shanghai & Wuhan: 2001-2004
Location: Bangkok, Thailand
Cardiovascular (I00-I99)
Respiratory (J00-J98)
Study Design: Time-series
Hong Kong, Shanghai, and Wuhan, China Statistical Analyses: Poisson GLM,
natural splines
Age Groups: All ages
>	65
>	75
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD):
Bangkok: 52.0
Hong Kong: 51.6
Shanghai: 102.0
Wuhan: 141.8
Range (Min, Max):
Bangkok: (21.3,169.2)
Hong Kong: (13.7,189.0)
Shanghai: (14.0, 566.8)
Wuhan: (24.8, 477.8)
Copollutant:
NO2
SO2
0s
Increment: 10 //g/m3
% Excess Risk (Lower CI, Upper CI)
lag:
Random Effects (4 cities)
Natural causes: 0.55% (0.26, 0.85)
0-1
Cardiovascular: 0.58% (0.22, 0.93)
0-1
Respiratory: 0.62% (0.22,1.02)
0-1
Random Effects (3 Chinese cities)
Natural causes: 0.37% (0.21, 0.54)
0-1
Cardiovascular: 0.44% (0.19, 0.68)
0-1
Respiratory: 0.60% (0.16,1.04)
0-1
Sensitivity Analysis
Random Effects (4 cities)
Omit PM10 > 95th: 0.53% (0.27,0.78)
0-1
Omit PM10 > 75th: 0.53% (0.29,0.78)
0-1
Omit PM10 > 180 /yglm3: 0.65% (0.24,
1.06)
0-1
Omit stations with high traffic source:
0.55% (0.26, 0.85)
0-1
July 2009
E-418
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Warm seasondichotomous variables:
0.86% (0.11,1.60)
0-1
Add temperature at lag 1-2 days: 0.51%
(0.23, 0.79)
0-1
Add temperature at lag 3-7 days: 0.35%
(0.14,0.57)
0-1
Daily PMio defined by centering: 0.54%
(0.26, 0.82)
0-1
Natural spline with (8, 4, 4f: 0.54%
(0.26, 0.81)
0-1
Penalized spline: 0.52% (0.26, 0.77)
0-1
Random Effects (3 Chinese cities)
Omit PMio > 95th: 0.47% (0.21,0.73)
0-1
Omit PMio > 75th: 0.55% (0.24,0.85)
0-1
Omit PMio>180/yg/m3: 0.46% (0.15,
0.76)
0-1
Omit stations with high traffic source:
0.38% (0.20, 0.57)
0-1
Warm season dichotomous variables:
0.43% (0.10, 0.76)
0-1
Add temperature at lag 1-2 days: 0.36%
(0.18, 0.53)
0-1
Add temperature at lag 3-7 days: 0.25%
(0.10, 0.40)
0-1
Daily PMio defined by centering: 0.37%
(0.21,0.53)
0-1
Natural spline with (8, 4, 4f: 0.36%
(0.23, 0.49)
0-1
Penalized spline: 0.34% (0.23, 0.45)
0-1
Increment: 10 //g/m3
% Excess Risk (Lower CI, Upper CI)
lag:
Non-accidental:
Low SDI
Reference: Wong et al. (2008,157151) Outcome (ICD10): Mortality:	Pollutant: PMio
Period of Study: 1/1996-12/2002 Non-accidental (A00-T99	Averaging Time: 24-h avg
Location: Hong Kong	Z00-Z99)	Mean (SD):
Cardiovascular (I00-I99)	51.6 (25.3)
Respiratory (J00-J98)	Range (Min, Max):
July 2009
E-419
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural splines
Age Groups: All ages
(13.5,188.5)
Copollutant:
NO?
SO?
Os
0.37(-0.10, 0.84)
0
0.40 (-0.04, 0.84)
1
0.14(-0.28, 0.57)
2
-0.12(-0.55, 0.30)
3
-0.14(-0.56, 0.28)
4
Middle SDI
0.70(0.34,1.07)
0
0.48 (0.14,0.82)
1
0.35 (0.02, 0.68)
2
0.18(-0.14, 0.51)
3
0.17(-0.16, 0.50)
4
High SDI
0.22 (-0.29, 0.73)
0
0.46(-0.01,0.94)
1
0.29(-0.17,0.75)
2
-0.05(-0.51,0.40)
3
-0.06(-0.51,0.40)
4
All areas
0.45 (0.19,0.72)
0
0.40 (0.15,0.64)
1
0.22 (-0.02, 0.45)
2
0.00 (-0.24, 0.23)
3
0.03 (-0.20, 0.26)
4
Cardiovascular:
Low SDI
July 2009
E-420
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0.14(-0.77,1.06)
0
0.64 (-0.21,1.49)
1
0.24 (-0.58,1.07)
2
-0.27(-1.09, 0.55)
3
0.01 (-0.80, 0.83)
4
Middle SDI
0.66(0.00,1.34)
0
0.49 (-0.13,1.12)
1
0.80(0.20,1.40)
2
0.65(0.06,1.25)
3
0.52(-0.07,1.12)
4
High SDI
0.83(-0.08,1.75)
0
0.89(0.04,1.75)
1
0.12(-0.70, 0.95)
2
-0.09(-0.91,0.73)
3
0.04 (-0.77,0.86)
4
All areas
0.52(0.05,1.00)
0
0.58(0.14,1.03)
1
0.43 (0.00, 0.86)
2
0.14(-0.28, 0.57)
3
0.23 (-0.20, 0.65)
4
Respiratory:
Low SDI
July 2009
E-421
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0 0.691-0.44,1.82)
0
1	0.551-0.50,1.61)
1
2	0.361-0.66,1.39)
2
3-0.24	(-1.25,0.78)
3
4-0.171-1.17,0.85)
4
Middle SDI
0.31 (-0.50,1.13)
0
0.77 (0.01,1.53)
1
0.85(0.12,1.59)
2
0.66(-0.07,1.39)
3
0.69(-0.03,1.42)
4
High SDI
0.27(-0.85,1.40)
0
0.72(-0.32,1.78)
1
1.46 (0.45,2.47)
2
0.70(-0.30,1.71)
3
0.48 (-0.52,1.48)
4
All areas
0.39 (-0.20, 0.99)
0
0.70(0.15,1.26)
1
0.89(0.36,1.42)
2
0.45 (-0.08, 0.98)
3
0.43 (-0.10, 0.96)
4
High SDI vs. Middle SDI
Non-accidental: 0.23 (-0.25, 0.72)
July 2009
E-422
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0-1
Cardiovascular: 0.49 (-0.40,1.40)
0-1
Respiratory: 0.49 (-0.58,1.58)
0-1
High SDI vs. Low SDI
Non-accidental: 0.12 ( 0.42, 0.67)
0-1
Cardiovascular: 0.82 (-0.20,1.86)
0-1
Respiratory: -0.15 (-1.39,1.10)
0-1
Trend Test
Non-accidental: 0.04 (-0.15, 0.22)
0-1
Cardiovascular: 0.27 (-0.07, 0.61)
0-1
Respiratory: -0.04 (-0.46, 0.37)
0-1 SDI - Social Deprivation Index. The
higher the SDI the lower the SES of the
individual.
Reference: Yang et al. (2004, 055603)
Period of Study: 1994-1998
Location: Taipei, Taiwan
Outcome: Mortality:
Non-accidental (< 800)
Circulatory (390-459)
Respiratory (460-519)
Study Design: Bi-directional case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD): 51.99
Range (Min, Max): (13.71,211.30)
Copollutant:
SO?
NO?
CO
0s
Increment: 31.43 /yglm3
Odds Ratio (Lower CI, Upper CI)
lag:
Non-accidental
0.995 (0.971,1.020)
0
Respiratory
0.986 (0.906,1.074)
0
Circulatory
0.988 (0.942,1.035)
Reference: Zanobetti et al. (2003,
042812)
Period of Study: 1990-1997
Location: 10 European cities (APHEA2)
Outcome: Mortality:
Pollutant: PMio
Increment: 10 //g/m3
Non-accidental (< 800)
Averaging Time: 24-h avg
% Increase (Lower CI,
Circulatory (390-459)
Mean (SD):
lag:
Respiratory (460-519)
Athens: 42.7 (12.9)
Cardiovascular
Study Design: Time-series
Budapest: 41 (9.1)
0.69% (0.31,1.08)
Statistical Analyses: Poisson GAM
Lodz: 53.5 (15.5)
0-1 avg
Age Groups: 15-64
London: 28.8(13.7)
40-day distributed lag
65-74
Madrid: 37.8 (17.7)
1.99% (1.44,2.54)
>75
Paris: 22.5(11.5)
4th degree

Prague: 76.2 (45.7)
1.97% (1.38, 2.55)

Rome: 58.7(17.4)
Unrestricted

Stockholm: 15.5 (7.9)
Respiratory

Tel Aviv: 50.3 (57.5)
0.74% (-0.17,1.66)

Range (Min, Max): NR
0-1 avg
July 2009
E-423
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Copollutant (correlation!: NR
40-day distributed lag
4.21% (1.70, 6.79)
4th degree
4.20% (1.08,7.42)
Unrestricted
Unrestricted distributed lags
Cardiovascular
1.34% (0.89,1.79)
20
1.72% (1.20, 2.25)
30
1.97% (1.38, 2.55)
40
Respiratory
1.71% (-0.65, 4.12)
20
2.62% (0.19, 5.11)
30
4.20% (1.08,7.42)
40
40-day lags
Non-accidental
15-64
¦0.25% (-0.87, 0.36)
4th degree
¦0.01 (-0.76, 0.75)
Unrestricted
65-74
0.78% (0.23,1.33)
4th degree
0.74% (0.02,1.45)
Unrestricted
>75
1.84% (0.92, 2.78)
4th degree
1.94% (1.07, 2.81)
Unrestricted
Cardiovascular
65-74
2.06% (1.05, 3.09)
4th degree
1.62(0.54,2.70)
Unrestricted
> 75
2.35% (1.42, 3.29)
July 2009
E-424
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
4th degree
2.52% (1.57, 3.48)
Unrestricted
Respiratory
> 75
4.57% (1.25,7.99)
4th degree
4.52% (0.89, 8.28)
Unrestricted
Reference: Zeka et al. (2005, 088068)
Period of Study: 111989-1212000
Location: 20 U.S. cities
Outcome (ICD10): Mortality:
All-cause (non-accidental) (V01-Y98)
Heart Disease (101-151)
IHD (I20-I25)
Myocardial infarction (121,122)
Dysrhythmias (I46-I49)
Heart failure (I50)
Stroke (I60-I69)
Respiratory (J00-J99)
Pneumonia (J12-J18)
COPD (J40-J44, J47)
Study Design: Time-stratified case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: All ages
Pollutant: PMio
Averaging Time: 24-h avg
Mean (SD):
Birmingham: 31.9 (18.0) //g/m3
Boulder: 22.1 (11.3)
Caton: 26.6 (11.5)
Chicago: 33.7 (16.4)
Cincinnati: 31.4 (13.9)
Cleveland: 37.5(18.7)
Colorado Springs: 24.0 (13.2)
Columbus: 28.5 (12.5)
Denver: 28.5 (12.8)
Detroit: 32.1 (17.7)
Honolulu: 15.9 (6.8)
Minneapolis: 24.7 (12.3)
Nashville: 30.1 (12.1)
New Haven: 25.4 (14.4)
Pittsburgh: 30.2 (18.5)
Provo: 33.7 (22.2)
Seattle: 26.4 (14.7)
Salt lake City: 35.0 (20.8)//
Terra Haute: 29.2 (14.6) fj
Youngstown: 30.8 (13.9)
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
Single-lag model
All-Cause (non-accidental)
0.20% (0.08, 0.32)
0
0.35% (0.21,0.49)
1
0.24% (0.14,0.34)
2
Respiratory
0.34% (-0.07, 0.75)
0
0.52% (0.15, 0.89)
1
0.51% (0.16, 0.86)
2
COPD
¦0.06% (-0.63, 0.51)
0
0.43% (-0.14,1.00)
1
0.39% (-0.16, 0.94)
2
Pneumonia
0.50% (0.09,1.09)
0
0.59% (-0.12,1.30)
1
0.82% (0.25,1.39)
2
Heart disease
0.12% (-0.06, 0.30)
0
0.30% (0.12, 0.48)
July 2009
E-425
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1
0.37% (0.17, 0.57)
2
IHD
0.19% (-0.03, 0.41)
0
0.41% (0.19, 0.63)
1
0.43% (0.10, 0.76)
2
Myocardial Infarction
0.36% (-0.05, 0.77)
0
0.17% (-0.18, 0.52)
1
0.13% (-0.22, 0.48)
2
Heart Failure
0.17% (-0.63, 0.97)
0
¦0.01% (-0.81,0.79)
1
0.78% (-0.004,1.56)
2
Dysrhythmias
¦0.23% (-1.41,0.95)
0
0.37% (-0.47,1.21)
1
0.33% (-0.55,1.21)
2
Stroke
0.09% (-0.49, 0.60)
0
0.41% (-0.02, 0.84)
1
0.14% (-0.27, 0.55)
2
Unconstrained distributed lag model
All-cause (non-accidental)
0.45% (0.25, 0.65)
0-3
Respiratory
0.87% (0.38,1.36)
0-3
July 2009
E-426
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
COPD
0.43% (-0.35,1.21)
0-3
Pneumonia
1.24% (0.46, 2.02)
0-3
Heart Disease
0.50% (0.25, 0.75)
0-3
IHD
0.65% (0.32, 0.98)
Myocardial Infarction
0.36% (-0.25, 0.97)
0-3
Heart Failure
0.60% (-0.50,1.70)
0-3
Dysrhythmias
0.20% (-1.03,1.43)
0-3
Stroke
0.46% (-0.13,1.05)
0-3
Reference: Zeka et al. (2006, 088749)
Outcome (ICD10): Mortality:
Pollutant: PMio
Increment: 10 //g/m3
Period of Study: 111989-1212000
All-cause (non-accidental) (V01-Y98)
Averaging Time: 24-h avg
% Increase (Lower CI, Upper CI)
Location: 20 U.S. cities
Heart Disease (101-151)
Mean (SD):
lag: All-cause (non-accidental)

Myocardial infarction (121,122)
Birmingham: 31.9 (18.0) //g/m3
Male: 0.46% (0.28, 0.64)

Stroke (I60-I69)
Boulder: 22.1 (11.3)
1-2 avg

Respiratory (J00-J99)
Caton: 26.6 (11.5)
Female: 0.37% (0.17, 0.57)

Study Design: Time-stratified case-
Chicago: 33.7 (16.4)
1-2 avg

crossover
Cincinnati: 31.4 (13.9)
White

Statistical Analyses: Conditional
logistic regression
Cleveland: 37.5(18.7)
0.40% (0.22, 0.58)

Age Groups:
Colorado Springs: 24.0 (13.2)
1-2 avg

All ages
Columbus: 28.5 (12.5)
Black: 0.37% (-0.02, 0.76)

<65
Denver: 28.5 (12.8)
1-2 avg

65-75
Detroit: 32.1 (17.7)
Age: <65:0.25% (0.01,0.49)

>75
Honolulu: 15.9 (6.8)
Minneapolis: 24.7 (12.3)
Nashville: 30.1 (12.1)
New Haven: 25.4 (14.4)
Pittsburgh: 30.2 (18.5)
Provo: 33.7 (22.2)
Seattle: 26.4 (14.7)
Salt lake City: 35.0 (20.8)
Terra Haute: 29.2 (14.6)
1-2 avg
75: 0.23% (-0.06, 0.52)
1-2 avg
>75:0.64% (0.44,0.84)
1-2 avg
Educational Attainment: Low (< 8
yrs): 0.62% (0.29, 0.95)
1-2 avg
Medium (8-12 yrs): 0.36% (0.12, 0.60)
July 2009
E-427
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Youngstown: 30.8 (13.9)
Range (Min, Max): NR
Copollutant (correlation): NR
1-2 avg
High (>12 yrs): 0.27% (-0.004, 0.54)
1-2 avg
Location of Death: In hospital: 0.22%
(0.04, 0.40)
1-2 avg
Out of hospital: 0.71% (0.51, 0.91)
1-2 avg
Season: Winter: 0.28% (0.04, 0.52)
1-2 avg
Summer: 0.19% (-0.22, 0.60)
1-2 avg
Transition (spring/fall): 0.49% (0.25,
0.73)
1-2 avg
Respiratory
Male: 0.71% (0.004,1.42)
0-3
Female: 1.04% (0.33,1.75)
0-3
White: 0.88% (0.33,1.43)
0-3
Black: 0.71% (-0.56,1.98)
0-3
Age: <65:0.94% (-0.31,2.19)
0-3
65-75: 0.87% (-0.25,1.99)
0-3
>75:0.88% (0.17,1.59)
0-3
Educational Attainment: Low (< 8
yrs): 0.82% (-0.32,1.96)
0-3
Medium (8-12 yrs): 0.88% (0.12,1.64)
0-3
High (>12 yrs): 0.88% (-0.04,1.80)
0-3
Location of Death: In hospital: 0.78%
(0.17,1.39)
0-3
Out of hospital: 1.09% (0.25,1.93)
0-3
Season: Winter: -0.007% (-0.87, 0.86)
0-3
Summer: 0.69% (-0.68, 2.06)
0-3
^ransjtmn_(s£ring/fall)Mi57%_(0;86;__
July 2009
E-428
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
2.28)
0-3
Heart Disease
Male: 0.54% (0.23, 0.85)
2
Female: 0.46% (0.15, 0.77)
2
White
0.50% (0.25, 0.75)
2
Black: 0.64% (0.13,1.15)
2
Age: <65:0.04% (-0.45,0.53)
2
65-75: 0.60% (0.13,1.07)
2
>75:0.65% (0.30,1.00)
2
Educational Attainment: Low (< 8
yrs): 0.72% (0.23,1.21)
2
Medium (8-12 yrs): 0.38% (0.07, 0.69)
2
High (>12 yrs): 0.54% (0.13,0.95)
2
Location of Death: In hospital: 0.15%
(-0.14,0.44)
2
Out of hospital: 0.93% (0.60,1.26)
2
Season: Winter: 0.41% (-0.002, 0.82)
2
Summer: 0.52 (0.03,1.01)
2
Transition (spring/fall): 0.56% (0.13,
0.99)
2
Myocardial Infarction
Male: 0.21% (-0.40,0.82)
0
Female: 0.59% (0.08,1.10)
0
White
0.24% (-0.27, 0.75)
0
Black: 0.99% (0.05,1.93)
July 2009
E-429
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0
<65:0.12% (-0.76,1.00)
0
65-75: 0.92% (0.21,1.63)
0
>75:0.16% (-0.58, 0.90)
0
Educational Attainment: Low (< 8
yrs): 0.33% (-0.83,1.49)
0
Medium (8-12 yrs): 0.79% (0.28,1.30)
0
High (>12 yrs):-0.13% (-0.82,0.56)
0
Location of Death: In hospital: 0.34%
(-0.11,0.79)
0
Out of hospital: 0.48% (-0.23,1.19)
0
Season: Winter: 0.32% (-0.37,1.01)
0
Summer: 0.30% (-0.82,1.42)
0
Transition (spring/fall): 0.38% -0.31,
1.07)
0
Stroke
Male: 0.11% (-0.58,0.80)
1
Female: 0.59% (-0.04,1.22)
1
White
0.48% (0.01,0.95)
1
Black: 0.13% (-0.87,1.13)
1
Age: <65:0.09% (-1.09,1.27)
1
65-75:-0.46% (-1.42,0.50)
1
>75:0.80% (0.27,1.33)
1
Educational Attainment: Low (< 8
yrs): 0.07% (-1.44,1.58)
1
Medium (8-12 yrs): 0.29% (-0.32, 0.90)
1
July 2009
E-430
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
High (>12 yrs): 0.52% (-0.28,1.32)
1
Location of Death: In hospital: 0.06%
(¦0.49,0.61)
1
Out of hospital: 0.87% (0.05,1.69)
1
Season: Winter: -0.09% (-0.93, 0.75)
1
Summer: 0.67% (-0.31,1.65)
1
Transition (spring/fall): 0.51% (-0.20,
1.22)
1
Contributing causes of disease: All-
cause
Secondary pneumonia present: 0.67%
(0.16,1.18)
1-2 avg
Secondary pneumonia absent: 0.34%
(0.16, 0.52)
1-2 avg
Secondary heart failure present: 0.42%
(0.01,0.83)
1-2 avg
Secondary heart failure absent: 0.37%
(0.19, 0.55)
1-2 avg
Secondary stroke present: 0.85% (0.30,
1.40)
1-2 avg
Secondary stroke absent: 0.32% (0.14,
0.50)
1-2 avg
Diabetes present: 0.57% (0.02,1.12)
1-2 avg
Diabetes absent: 0.34% (0.14, 0.54)
1-2 avg
Respiratory
Secondary pneumonia present: 1.28%
(-0.33, 2.89)
0-3
Secondary pneumonia absent: 0.78%
(0.15,1.41)
0-3
Secondary heart failure present: 1.48%
(0.07, 2.89)
0-3
Secondary heart failure absent: 0.79%
(0.26,1.32)
July 2009
E-431
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0-3
Secondary stroke present: 1.95% (-0.11,
4.01)
0-3
Secondary stroke absent: 0.80% (0.29,
1.31)
0-3
Diabetes present: 1.96% (-0.22, 4.14)
0-3
Diabetes absent: 0.82% (0.31,1.33)
0-3
Heart Disease
Secondary pneumonia present: 0.66%
(¦0.63,1.95)
2
Secondary pneumonia absent: 0.49%
(0.27, 0.71)
2
Secondary stroke present: 0.73% (-0.05,
1.51)
2
Secondary stroke absent: 0.48% (0.24,
0.72)
2
Diabetes present: 0.34% (-0.42,1.10)
2
Diabetes absent: 0.52% (0.28, 0.76)
2
Myocardial Infarction
Secondary pneumonia present: 1.54%
(¦1.05,4.13)
0
Secondary pneumonia absent: 0.42%
(0.05, 0.79)
0
Secondary stroke present: 0.50% (-1.38,
2.38)
0
Secondary stroke absent: 0.36% (-0.05,
0.77)
0
Diabetes present: 0.70% (-0.38,1.78)
0
Diabetes absent: 0.41% (0.04, 0.78)
0
Stroke
Secondary pneumonia present: 1.74%
(0.35, 3.13)
1
^Secondar^neumonia^
July 2009
E-432
DRAFT-DO NOT CITE OR QUOTE

-------
Concentrations'
Effect Estimates (95% CI)
(¦0.16,0.74)
1
Secondary heart failure present: 1.01%
( 0.77, 1.79)
Secondary heart failure absent: 0.38%
(¦0.05,0.81)
1
Diabetes present: 1.02% (-0.53, 2.57)
1
Diabetes absent: 0.37% (-0.08, 0.82)
1
Study
Design & Methods
'All units expressed in//g/m3 unless otherwise specified.
Table E-17. Short-term exposure - mortality - PM102.5.
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Burnett et al.
(2004, 086247)
Period of Study: 1981-
1999
Outcome: Mortality:
Pollutant: P10 2.5
Increment: 10 //g/m3
Non-accidental (< 800)
Study Design: Time-series
Averaging Time: 24-h avg
Mean (SD): 11.4
% Increase (Lower CI, Upper CI)
lag:
Location: 12 Canadian
cities
Statistical Analyses: 1. Poisson,
natural splines
2. Random effects regression
model
Range (Min, Max): NR
Copollutant: NO2
03
1981-1999
PMio-2.5: 0.31% (-0.66,1.33)
1

Age Groups: All ages
SO?
CO
PM10
PM10.2.5+NO2: 0.65% (-0.23,1.59)
1





PM2.6



Note: PM10 measurement
calculated as the sum of PM2.6
and PM10 2.5 measurements.

Reference: Kan et al.
(2007,091267)
Period of Study: 3/2004-
1212005
Outcome (ICD10): Mortality:
Pollutant: PMio 2.5
Increment: 10 //g/m3
Total (non-accidental) (A00-R99)
Cardiovascular (100-199)
Averaging Time: 24-h avg
Mean (SD): 56.4 (1.34)
% Increase (Lower CI, Upper CI)
lag: Total: 0.12% (-0.13,0.36)
Location: Shanghai, China
Respiratory (J00-J98)
Study Design: Time-series
Statistical Analyses: Poisson
GAM, penalized splines
Age Groups: All ages
Range (Min, Max): (8.3,
235.0)
Copollutant (correlation):
PM10: r - 0.88
PM2.6: r - 0.48
0s: r - 0.07
0-1
Cardiovascular: 0.34% (-0.05, 0.73)
0-1
Respiratory: 0.40% (-0.34,1.13)
0-1
July 2009
E-433
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kettunen et al.
(2007,091242)
Period of Study: 1998-
2004
Location: Helsinki, Finland
Outcome (ICD10): Mortality:
Stroke (160-161,163-164)
Study Design: Time-series
Statistical Analyses: Poisson
GAM, penalized thin-plate splines
Age Groups: a 65
Pollutant: PMio-2.5
Averaging Time: 24-h avg
Median (SD) unit: Cold
Season: 6.7
Warm Season: 8.4
Range (Min, Max): Cold
Season: (0.0,101.4)
Warm Season: (0.0, 42.0)
Copollutant: 0:i, CO, NO2
PM10
PM2.6
UFP
Increment:
Cold Season: 8.3 /yglm3
Warm Season: 5.7 /yglm3
% Increase (Lower CI, Upper CI)
lag:
Cold Season: -1.04% (-6.63, 4.89)
0
¦2.49% (-7.57, 2.88)
1. -4.93% (-9.99, 0.41)
2
¦4.33% (-9.32, 0.93)
3
Warm Season: 7.05% (-1.88,16.80)
0
4.38% (-4.26,13.81)
1: -1.19% (-9.45,7.84)
2
1.42% (-6.79,10.34)
3
Reference: Klemm et al.
(2004, 056585)
Outcome: Mortality:
Non-accidental (< 800)
Cardiovascular (390-459)
Location: Fulton and DeKalb Respiratory (460-519)
Period of Study: 8/1998-
712000
counties, Georgia (ARIES)
Cancer (140-239)
Study Design: Time-series
Statistical Analyses: Poisson
GLM, natural cubic splines
Age Groups: < 65,a 65
Pollutant: PMio 2.5
Averaging Time: 24-h avg
Mean (SD): 9.69 (3.94)
Range (Min, Max): (1.71,
25.17)
Copollutant: PM2.6
0s
NO?
CO
SO2
Acid
EC
0C
S04
Oxygenated Hydrocarbons
Nonmethane hydrocarbons
NOs
Increment: NR
PISE)
lag:
Quarterly Knots:
0.00433 (0.00333)
0-1
Monthly Knots:
0.00617 (0.00360)
0-1
Biweekly Knots:
0.00516(0.00381)
0-1
July 2009
E-434
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Perez et al.
(2008,156020)
Period of Study:
3/27/2003-12/31/2005
Location: Barcelona, Spain
Outcome: respiratory mortality
Study Design: cohort
Covariates: temperature,
humidity
Statistical Analysis:
autoregressive Poisson regression
models
Statistical Package: NR
Age Groups: All deaths
Pollutant: PM102 E
Averaging Time: 24 h
Mean (SD) Unit: 14.0 (9.5)
z^g/m3
Range (Min, Max): 0.1, 93.1
Copollutant: PM2.6-i, PMi
Increment: 10/yg/m3
Odds Ratio |95%CI)
Lag
Single Pollutant Model
Avg L0-1: 1.000 (0.944-1.060), p - 0.991
L1:1.002(0.955-1.052), p - 0.931
L2:1.070(1.023-1.118), p - 0.003
Multi-pollutant Model
Avg L0-1: 1.002 (0.937-1.071), p - 0.958
L1: 0.998 (0.943-1.056), p - 0.0.936
L2:1.033(0.980-1.089), p - 0.226
Reference: Perez et al.
(2008,156020)
Period of Study:
3/27/2003-12/31/2005
Location: Barcelona, Spain
Outcome: cardiovascular mortality Pollutant: PM102 e
Study Design: cohort	Averaging Time: 24 h
Covariates: temperature,
humidity
Mean (SD) Unit: 14.0 (9.5)
Z^g/m3
Statistical Analysis:	Range (Min, Max): 0.1, 93.1
autoregressive Poisson regression
mo(|e|s	Copollutant: PM2.6-i, PMi
Statistical Package: NR
Age Groups: All deaths
Increment: 10/yg/m3
Odds Ratio |95%CI)
Lag
Avg L0-1: 1.054(1.019-1.089), p - 0.002
L1:1.059(1.031-1.072), p - 0.000
L2:1.044(1.017-1.072), p - 0.001
Multi-pollutant Model
Avg L0-1: 1.053 (1.013-1.094), p - 0.009
L1:1.059(1.026-1.094), p - 0.001
L2:1.044(1.012-1.078), p - 0.007
Reference: Perez et al.
(2008,156020)
Period of Study:
3/27/2003-12/31/2005
Location: Barcelona, Spain
Outcome: cerebrovascular
mortality
Study Design: cohort
Covariates: temperature,
humidity
Pollutant: PM102 E
Averaging Time: 24 h
Mean (SD) Unit: 14.0 (9.5)
/yg/m3
Range (Min, Max): 0.1, 93.1
Statistical Analysis:
autoregressive Poisson regression Copollutant: PM2.6-i, PMi
models
Statistical Package: NR
Age Groups: All deaths
Increment: 10/yg/m3
Odds Ratio |95%CI)
Lag
Avg L0-1: 1.087 (1.018-1.161), p - 0.013
L1:1.086(1.030-1.145), p - 0.002
L2:1.051 (0.997-1.108), p - 0.064
Multi-pollutant Model
Avg L0-1: 1.103 (1.022-1.191), p - 0.011
L1:1.098(1.030-1.171), p - 0.004
L2:1.076(1.010-1.146), p - 0.023
Reference: Slaughter et al.
Outcome: Mortality: Non-
Pollutant: PMio-2.5 This study does not present quantitative results for PM10-2.B.
(2005, 073854)
accidental (< 800)

Averaging Time: 24-h avg
Period of Study: 1/1995-
Study Design: Time-series
Mean (SD) unit: NR
12/1999
Statistical Analyses: Poisson
Range (9th, 95th): NR
Location: Spokane,
GLM, natural splines
Washington
Age Groups: All ages
Copollutant (correlation):

PM1: r - 0.19


PM2.6: r - 0.31


PM10: r - 0.94


CO: r - 0.32
July 2009
E-435
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% ClJ
Reference: Stieb et al. Outcome: Mortality: All-cause
Pollutant: PMio 2.5
Increment: 13.0/yg/m3
(2002,025205) (non-accidental)


Averaging Time: NR
% Increase (Lower CI, Upper CI)
Period of Study: Study Design: Meta-analysis
Mean (SD): NR
lag:
Publication dates of studies:
1985-12/2000 Mortality Statistical Analyses: Random
series: 1958-1999 effects model
Range (Min, Max): NR
Single-pollutant models: 10 studies
Location: 40 cities (11 A9e Groups: All ages
Canadian cities, 19 U.S.
Copollutant: Varied between
studies: PM2.6, O3, SO2, NO2,
CO
PMio-2.5: 1.2% (0.5,1.9)
Multipollutant models: 6 studies
cities, Santiago, Amsterdam,
Erfurt, 7 Korean cities)

PMio-2.5: 0.9% (-0.3, 2.0)
July 2009
E-436
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Villeneuve et al. Outcome: Mortality:
(2003,055051)
Period of Study: 1986-
1999
Location: Vancouver,
Canada
Non-accidental (< 800)
Cardiovascular (401-440)
Respiratory (460-519)
Cancer (140-239)
Study Design: Time-series
Statistical Analyses: Poisson,
natural splines
Age Groups: a 65
Pollutant: PMio 2.5
Averaging Time: 24-h avg
Mean (SD):
Daily: 6.1
Every 6th Day
8.3
Range (Min, Max):
Daily: (0.0, 72.0)
Every 6th Day: (0.7, 35.0)
Copollutant:
PM2.6
PM10
SO?
CO
NO?
0s
Increment: 11.0/yg/m3
% Increase (Lower CI, Upper CI)
lag:
Non-accidental
1.4% (-2.5,5.4)
0-2 avg
1.0% (-1.9,4.0)
0
¦1.1% (-4.0,1.8)
1
2.0% (-1.0,5.1)
2
Cardiovascular
5.9% (-0.2,12.4)
0-2 avg
5.9% (1.1,10.8)
0
1.4% (-3.3,6.4)
1
2.2% (-2.0,6.7)
2
Respiratory
¦1.0% (-9.8, 8.8)
0-2 avg
¦1.5% (-9.4,7.1)
0
¦1.5% (-8.4,6.0)
1
0.1% (-6.4,6.9)
2
Cancer
4.4% (-3.6,13.1)
0-2 avg
3.1% (-2.9,9.4)
0
¦1.0% (-6.9, 5.3)
1
4.0% (-2.1,10.4)
2
July 2009
E-437
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Wilson et al.
(2007,157149)
Period of Study: 1995-
1997
Location: Phoenix, Arizona
Outcome: Mortality:
Cardiovascular
Study Design: Time-series
Statistical Analyses: Poisson
GAM, nonparametric smoothing
spline
Age Groups:
>25
Pollutant: PMio 2.5
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Increment: 10 //g/m3
% Excess Risk (Lower CI, Upper CI)
lag:
Central Phoenix: 2.4% (-1.2, 6.1)
Copollutant (correlation): NR 0-5 ma
Middle Phoenix:
3.8% (0.3, 7.5)
0-5 ma
3.4% (1.0, 5.8)
1
3.0% (0.7, 5.4)
2
Outer Phoenix: 1.1
0-5 ma
[¦1.9, 5.2)
'All units expressed in/yglm3 unless otherwise specified.
July 2009
E-438
DRAFT-DO NOT CITE OR QUOTE

-------
Table E-18. Short-term exposure - mortality - PM2.5 (including PM components/sources).
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Basu et al. (2008, 098716)
Outcome (ICD10): Mortality:
Pollutant: PM2.6
The study does not provide results



quantitatively.
Period of Study: 5/1999-9/2003
Non-accidental (V01-Y98)
Averaging Time: 24-h avg
Location: 9 California counties
Study Design: (1) Main analysis: Case-
Mean (SE) unit:


crossover




Contra Costa: 8.6


(2) Sensitivity analysis: Time-series
Fresno: 7.6


Statistical Analyses: (1) Main analysis:
Kern: 11.3


conditional logistic regression


(2) Sensitivity analysis: Poisson GAM
Los Angeles: 19.8


Age Groups: All ages
Orange: 17.0



Riverside: 28.4



Sacramento: 8.8



San Diego: 13.4



Santa Clara: 10.8



IQR (25th, 75th):



Contra Costa: (5.8,10.1)



Fresno: (3.8, 9.8)



Kern: (8.0,13.5)



Los Angeles: (14.7, 23.3)



Orange: (11.8, 21.0)



Riverside: (17.9, 36.1)



Sacramento: (5.8,10.1)



San Diego: (10.3,15.8)



Santa Clara: (7.2,13.8)



Copollutant (correlation): PMio



r - 0.45



0s (1 hr)



r - 0.28



03 (8hr)



r - 0.22



CO



r - 0.45



NO?



r - 0.43

Reference: Dominici et al. (2007,
097361)
Period of Study: PM10: 1987-2000.
PM2.1,: 1999-2000
Outcome: Mortality:
Pollutant: PM2.6
Increment: 10 //g/m3
All-cause (non-accidental)
Cardiorespiratory
Averaging Time: 24-h avg
Mean (SD): NR
% Increase (Lower CI, Upper CI)
lag:
Location: 100 U.S. counties (NMMAPS)
Other-cause
Range (Min, Max): NR
1999-2000:

Study Design: Time-series
Copollutant (correlation): NR
All-cause: 0.29% (0.01,0.57)

Statistical Analyses: 2-stage Be
hierarchical model
Age Groups: All ages
lyesian
1
Cardiorespiratory: 0.38% (-0.07, 0.82)
1
July 2009
E-439
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Dominici et al. (2007,
099135)
Period of Study: 2000-2005
Location: 72 U.S. counties representing
69 communities
Reference: Franklin et al. (2007,
091257)
Period of Study: 1997-2002
Location: 27 U.S. communities
Outcome: Total mortality
Study Design: Time-series
Statistical Analyses: 2-stage Bayesian
hierarchical model
Age Groups: All ages
Outcome: Mortality:
All-cause (non-accidental (< 800)
Cardiovascular (390-429)
Respiratory (460-519)
Stroke (430-438)
Study Design: Time-stratified case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: All ages
Pollutant: PM2.6, Nickel, speciated fine
PM, and Vanadium
Averaging Time: Annual avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD):15.7/yglm3
Range (Min, Max): NR
Copollutant (correlation): NR
The study does not provide results
quantitatively.
Note: The study investigated whether
county-specific short-term effects of
PMio on mortality are modified by long-
term county-specific nickel or vanadium
PM2.6 concentrations.
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag: All-cause (non-accidental): 0.67%
(-0.12,1.46)
0
1.21% (0.29, 2.14)
10.82% (0.02,1.63)
0-1
Respiratory: 1.31% (-0.10, 2.73)
0
1.78% (0.20, 3.36)
1
1.67% (0.19, 3.16)
0-1
Cardiovascular: 0.34% (-0.61,1.28)
0
0.94%	(-0.14,2.02)
1.
0.54%	(-0.47,1.54)
0-1
Stroke: 0.62% (-0.69,1.94)
0
1.03% (0.02, 2.04)
1.
0.67% (-0.23,1.57)
0-1
Age> 75: All cause: 1.66% (0.62, 2.70)
1
Respiratory: 1.85% (0.27, 3.44)
1
Cardiovascular: 1.29% (0.15,2.42)
1
Stroke: 1.52% (0.37, 2.67)
1
Age <75: All cause: 0.62% (-0.30,1.55)
1
Respiratory: 1.53% (-0.67,3.74)
1
Cardiovascular: 0.26% (-1.04,1.56)
1
July 2009
E-440
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Stroke:-0.78% (-2.32, 0.76)
1
Male: All cause: 1.06% (0.07,2.06)
1
Respiratory: 1.90% (0.14, 3.65)
1
Cardiovascular: 0.52% (-0.63,1.66)
1
Stroke: 0.79% (-0.42, 2.02)
1
Female: All cause: 1.34% (0.40, 2.27)
1
Respiratory: 1.57% (-0.22,3.35)
1
Cardiovascular: 1.30% (0.14,2.46)
1
Stroke: 0.79% (-0.51,2.09)
1
East: All cause: 1.95% (0.50, 3.40)
1
Respiratory: 2.66% (0.33, 5.00)
1
Cardiovascular: 1.52% (0.06,2.98)
1
Stroke: 1.16% (-0.40, 2.73)
1
West: All cause: 0.05% (-1.80,1.89)
1
Respiratory: 0.67% (-2.00, 3.34)
1|
Cardiovascular: 0.11% (-2.03,2.24)
1|
Stroke: 0.94% (-0.38, 2.26)
1
PM2.6 > 15 /yg/m3: All cause: 1.10%
(-0.43, 2.64)
1
Respiratory: 1.42% (-0.84, 3.68)
1
Cardiovascular: 0.88% (-0.87, 2.62)
1
Stroke: 0.91% (-0.28, 2.10)
1
PM2.bd 15/yg/m3: All cause: 1.41 %
(-0.49, 3.30)
1
July 2009
E-441
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Respiratory: 2.46% (-0.49, 5.42)
1
Cardiovascular: 1.09% (-1.15,3.32)
1
Stroke: 1.36% (-0.56, 3.27)
1
Effect of A/C at percentile of air
conditioning prevalence: 25th percentile
(45% prevalence of A/C): All cause:
1.50% (0.13, 2.88)
1
Respiratory: 2.27% (0.27, 4.27)
1
Cardiovascular: 1.04% (-0.54,2.63)
1
Stroke: 1.04% (-0.44, 2.53)
1
75th percentile (80% prevalence of A/C):
All cause: 0.85% (-0.64, 2.35)
1
Respiratory: 1.04% (-1.29,3.37)
1
Cardiovascular: 0.81% (-0.93, 2.61)
1
Stroke: 1.03% (-0.76, 2.83)
1
Effect of A/C at percentile of air
conditioning prevalence in cities with
summer peaking PM2.6 concentrations:
25th percentile (45% prevalence of A/C):
All cause: 1.01% (-0.30, 2.32)
1
Respiratory: 0.76% (-1.38, 2.90)
1
Cardiovascular: 0.43% (-0.86,1.72)
1
Stroke:-0.18% (-2.08,1.73)
1
75th percentile (77% prevalence of A/C):
All cause:-0.55% (-1.95,0.85)
1
Respiratory: -2.08% (-4.47, 0.31)
1
Cardiovascular: -1.02% (-2.44, 0.41)
1
Stroke: 0.69% (-1.19, 2.57)
1
July 2009
E-442
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Franklin et al. (2008,
097426)
Period of Study: 2000-2005
Location: 25 U.S. communities
Outcome (ICD10): Mortality:
Non-accidental (V01-Y98)
Respiratory (J00-J99)
Cardiovascular (101152)
Stroke (I60-J69)
Study Design: Time-series
Statistical Analyses: 1st stage:
Poisson, cubic spline
2nd stage: Random effects meta-analysis
Age Groups: All ages
Pollutant: PM2.6
Averaging Time: 24-h avg
Range Mean (SD):
Winter: 9.6 to 34.4
Spring: 6.7 to 27.6
summer: 7.6 to 26.0
Fall: 9.5 to 32.1
Range (Min, Max): NR
Copollutant:
Al, As, Br, Cr, EC, Fe, K, Mn, Na+, Ni,
NOs-, NH4, 0C, Pb, Si, S042-, V, Zn
July 2009
E-443
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
Non-accidental: 0.74% (0.41,1.07)
0-1
Cardiovascular: 0.47% (0.02, 0.92)
0-1
Respiratory: 1.01% (-0.03,2.05)
1-2
Stroke: 0.68% (-0.21,1.57)
0-1
Winter: 0.15% (-0.42, 0.72)
0-1
Spring: 1.88% (1.29, 2.48)
0-1
Summer: 0.99% (0.35,1.68)
0-1
Fall: 0.19% (-0.25, 0.64)
0-1
West: 0.51% (0.10, 0.92)
0-1
East & Central: 0.92% (0.44,1.39)
0-1
% Increase per 10//g/m3 increase in
PM2.E for an IQR increase in species to
PM2.E mass proportion
Univariate analysis
Al: 0.58%
As: 0.55%
Br: 0.38
Cr: 0.33%
EC: 0.06%
Fe: 0.12%
K: 0.41%
Mn: 0.14%
Na+: 0.20%
Ni: 0.37%
NOs-: -0.49%
NFk 0.04%
0C:-0.02%
Pb: 0.17%
Si: 0.41%
SO42-: 0.51 %
V: 0.30%
Zn: 0.23%
Multivariate (1)
Al: 0.79%
Ni: 0.34%
S0|fRAH53tD0 NOT CITE OR QUOTE
Multivariate (2)
Al: 0.61%

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Holloman et al. (2004,
087375)
Period of Study: 1999-2001
Location: 7 North Carolina counties
Outcome (ICD10): Mortality:
Cardiovascular (I00-I99)
Study Design: Time-series
Statistical Analyses: 3-stage E
hierarchical model
Age Groups: >16
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD): NR
esian Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
2.5% (-3.9 to 9.6)
0
4.0% (-3.3 to 12.2)
1
11.4% (2.8 to 19.8)
2
¦1.1% (-7.5 to 5.2)
3
Reference: Hopke et al. (2006, 088390)
Outcome: Mortality:
Pollutant: Source-apportioned PM2.6:
The study does not present quantitative



results.
Period of Study: Washington, DC:
Total (non-accidental)
Washington, DC: Soil

8/1988-12/1997. Phoenix, Arizona:



3/1995-6/1998
Cardiovascular
Traffic

Location: Washington, DC and
Cardiovascular-Respiratory
Secondary Sulfate

surrounding counties
Study Design: Source-apportionment
Nitrate

Phoenix, Arizona
Statistical Analyses: Receptor modeling
Residual Oil


Age Groups: All ages
Wood Smoke



Sea Salt



Incinerator



Primary Coal



Phoenix, Arizona: Crustal



Traffic



Vegetation and Wood Burning



Secondary Sulfate



Metals



Sea Salt



Primary Coal



Averaging Time: 24-h avg



Mean (SD): NR



Range (Min, Max): NR



Copollutant (correlation): NR

July 2009
E-444
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ito et al. (2006, 088391)
Period of Study: 811988-1211997
Location: Washington, DC and
surrounding counties
Outcome: Mortality:
Total (non-accidental)
Cardiovascular
Cardiovascular-Respiratory
Study Design: Time-series
Source-apportionment
Statistical Analyses: Poisson GLM,
natural splines
Age Groups: All ages
Pollutant: Source-apportioned PM2.6:
Soil
Traffic
Secondary Sulfate
Nitrate
Residual Oil
Wood Smoke
Sea Salt
Incinerator
Primary Coal
Averaging Time: 24-h avg
Mean (SD):
17.8(8.7)
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: PM2.6 - 28.7 /yglm3
PM2.6 Sources 5-95th - Not reported
% Increase (Lower CI, Upper CI)
lag:
Secondary sulfate (variance-weighted
mean percent excess mortality)
6.7% (1.7,11.7)
3
Primary coal-related PM2.6 (mean percent
excess mortality)
5.0% (1.0,9.1)
3
Residual oil (mean percent excess
mortality)
2.7% (-1.1, 6.5)
2
Traffic-related PM2.6 (mean percent
excess mortality)
2.6% (-1.6, 6.9)
NR
Soil-related PM2.6 (mean percent excess
mortality)
2.1% (-0.8, 4.9)
PM2.5 Sensitivity analysis:
2	df/year: 7.9% (3.3,12.6)
3
4	df/year: 8.3% (3.7,13.1)
3
8 df/year: 8.3% (3.7,13.2)
3
16 df/year: 8.1% (3.1,13.2)
3
Reference: Kan et al. (2007, 091267)
Period of Study: 312004-1212005
Location: Shanghai, China
Outcome (ICD10): Mortality:
Total (non-accidental) (A00-R99)
Cardiovascular (I00-I99)
Respiratory (J00-J98)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
penalized splines
Age Groups: All ages
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD): 52.3(1.57)
Range (Min, Max): (2.0, 330.3)
Copollutant (correlation):
PM10: r - 0.84
PM 10-2.5: r - 0.48
O3: r - 0.31
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
Total: 0.36% (0.11,0.61)
0-1
Cardiovascular: 0.41% (0.01, 0.82)
0-1
Respiratory: 0.95% (0.16,1.73)
0-1
July 2009
E-445
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kettunen et al. (2007,
091242)
Period of Study: 1998-2004
Location: Helsinki, Finland
Outcome (ICD10): Mortality:
Pollutant: PM2.6
Increment:
Stroke (160-161,163-164)
Averaging Time: 24-h avg
Cold Season: 6.7 /yglm3
Study Design: Time-series
Median (SD) unit:
Warm Season: 5.7 /yglm3
Statistical Analyses: Poisson GAM,
penalized thin-plate splines
Age Groups: a 65
Cold Season: 8.2
Warm Season: 7.8
Range (Min, Max):
% Increase (Lower CI, Upper CI)
lag:
Cold Season

Cold Season: (1.1, 69.5)
¦0.19% (-3.77, 3.51)

Warm Season: (1.1,41.5)
0

Copollutant: 0:i
¦0.17% (-3.73, 3.52)

CO
1

NO?
0.59% (-2.95, 4.26)

PMio
2

PMio-2.5
0.46% (-3.10, 4.15)

UFP
3
Warm Season
6.86% (0.37,13.78)
0
7.40% (1.33,13.84)
1
4.01% (-1.79,10.14)
2
¦1.72% (-7.38, 4.29)
3
Period of Study: 8/1998-7/2000
Location: Fulton and DeKalb counties,
Georgia (ARIES)
Outcome: Mortality:
Pollutant: PM2.6
Increment: NR
Non-accidental (< 800)
Averaging Time: 24-h avg
PISE)
Cardiovascular (390-459)
Mean (SD):
lag:
Respiratory (460-519)
19.62 (8.32)
Quarterly Knots:
Cancer (140-239)
Range (Min, Max):
PM2.b: 0.00398 (0.00161)
Study Design: Time-series
(5.29,48.01)
0-1
Statistical Analyses: Poisson GLM,
Copollutant:
Monthly Knots:
natural cubic splines


PM 10-2.6
PM2 b: 0.00544 (0.00184)
Age Groups: < 65
0s
0-1
>65



NO?
Biweekly Knots:

CO
PM2.b: 0.00369 (0.00201)

SO?
0-1

Acid


EC


OC


S04


Oxygenated Hydrocarbons


Nonmethane hydrocarbons


NOs

July 2009
E-446
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lippmann et al. (2006,
Outcome: Mortality:
Pollutant: Speciated Fine PM:
The study does not present quantitative
091165)


results.
Non-accidental (< 800)
Al, Ar, Cr, Cu, EC, Fe, Mn, Ni, Nitrate,

Period of Study: 2000-2003

0C, Pb, Se, Si, Sulfate, V, Zn

Study Design: Time-series

Location: 60 U.S. cities (NMMAPS)

Averaging Time: Annual avg

Statistical Analyses: Poisson GLM



Mean (SD): R


Age Groups: All ages



Range (Min, Max): NR

Reference: Maret al. (2005, 087566)
Period of Study: 1995-1997
Location: Phoenix, Arizona
Outcome: Mortality:
Non-accidental (< 800)
Cardiovascular (390-448)
Study Design: Time-series
Statistical Analyses: Poisson GLM
Age Groups: a 65
Pollutant: Source-apportioned PM2.6:
Soil
Traffic
Secondary Sulfate
Nitrate
Residual Oil
Wood Smoke
Sea Salt
Incinerator
Primary Coal
Increment: PM2.6 Sources 5-95th - NR
% Increase (median percent excess
risk)
lag:
Secondary sulfate: 16.0%
0
Traffic: 13.2%
1
Copper (Cu) smelter: 12.0%
0

Averaging Time: 24-h avg
Sea salt: 10.2%

Mean (SD): NR
Range (Min, Max): NR
5
Biomassfwood combustion: 8.6%
3
Outcome (ICD10): Mortality:
Pollutant: PM2.6
Increment: 10 //g/m3
Total mortality (respiratory,
cardiovascular, ischemic heart disease,
diabetes)
Averaging Time: 24-h avg
Mean (SD):
% Increase (Lower CI, Upper CI)
lag:
Respiratory (J00-J98)
Contra Costa: 14
Penalized splines
Cardiovascular (I00-I99)
Fresno: 23
All ages:
Ischemic heart disease (I20-I25)
Kern: 22
All-cause:
Diabetes (E10-E14)
Los Angeles: 21
0.2% (-0.2, 0.7)
Study Design: Time-series
Orange: 21
2
Statistical Analyses: Poisson, natural
splines and penalized splines
Riverside: 29
Sacramento: 14
0.6% (0.2,1.0)
0-1
Age Groups: All ages
Santa Clara: 15
Cardiovascular:
>65
San Diego: 16
0.3% (-0.1, 0.7)

Range (Min, Max):
2

Contra Costa: (1, 77)
0.6% (0.0,1.1)

Fresno: (1,160)
0-1

Kern: (1,155)
Respiratory:

Los Angeles: (4, 85)
1.3% (0.1,2.6)

Orange: (4,114)
2

Riverside: (2,120)
2.2% (0.6,3.9)

Sacramento: (1,108)
0-1

Santa Clara: (2, 74)
>65:

San Diego: (0, 66)
All-cause:

Copollutant (correlation):
0.2% (-0.2, 0.7)

NO?
2

r - 0.56
0.7% (0.2,1.1)
Reference: Ostro et al. (2006, 087991)
Period of Study: 111999-1212002
Location: 9 California counties
(CALFINE)
July 2009
E-447
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
CO
0-1
r - 0.60
Ischemic heart disease: 0.3% (-0.5,1.0)
Os (1h)
0-1
r - -0.14
Males: 0.5% (-0.2,1.2)
Os (8h|
0-1
r - 0.22
Females: 0.8% (0.3,1.3)

0-1

Whites: 0.8% (0.2,1.3)

0-1

Blacks: 0.1% (-0.9,1.2)

0-1

Hispanics: 0.8% (-0.1,1.6)

0-1

In hospital: 0.6% (-0.1,1.3)

0-1

Out of hospital: 0.6% (0.1,1.1)

0-1

High school graduates: 0.4% (0.0, 0.8)

0-1

Non-high school graduates: 0.9% (-0.1,

1.9)

0-1

Natural splines

All cause

4 df: 0.5% (-0.1,1.1)

0-1

8 df: 0.4% (-0.1,0.9)

0-1

12 df: 0.3% (-0.1,0.7)

0-1

Cardiovascular

4 df: 0.4% (-0.2, 0.9)

0-1

8 df: 0.1% (-0.5, 0.6)

0-1

12 df: 0.0% (-0.6, 0.6)

0-1

Respiratory

4 df: 2.1% (0.2, 4.1)

0-1

8 df: 1.6% (-0.5, 3.6)

0-1

12 df: 1.3% (-0.3, 2.9)

0-1

>65
July 2009
E-448
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
All cause
4 df: 0.7% (0.0,1.3)
0-1
8 df: 0.4% (-0.1,0.9)
0-1
12 df: 0.3% (-0.1,0.8)
0-1
Reference: Ostro et al. (2007, 091354)
Period of Study: PM2.6 speciation
analysis: 1/2000-12/2003. PM2.1,
analysis: 1/1999-12/2003
Location: 6 California counties (2000-
2003). 9 California counties (1999-
2003) (CALFINE)
Outcome (ICD10): Mortality:
Total (non-accidental) mortality
Respiratory (J00-J98)
Cardiovascular (I00-I99)
Study Design: Time-series
Statistical Analyses: Poisson, natural
splines
Age Groups: >65
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD):
2000-2003: 19.28
1999-2003: 18.6
Range (Min, Max): NR
Copollutant (correlation):
EC: r - 0.53
0C: r - 0.62
NOs: r - 0.65
S(k r - 0.32
Al: r - 0.02
Br: r - 0.54
Ca:r - 0.23
CI: r - 0.15
Cu:r - 0.23
Fe: r - 0.38
K: r - 0.52
Mn: r - 0.21
Ni: r - 0.11
Pb: r - 0.27
S: r - 0.35
Si: r - 0.16
Ti: r - 0.24
V: r - 0.20
Zn: r - 0.51
Increment: 14.6 /yglm3
% Increase (Lower CI, Upper CI)
lag:
Cardiovascular
1.6% (0.0,3.1)
3
Notes: The study does not present all
estimates quantitatively.
July 2009
E-449
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ostro et al. (2008, 097971) Outcome (ICD10): Mortality:
Period of Study: 112000-1212003
Location: 6 California counties
Cardiovascular (100199)
Study Design: Time-series
Statistical Analyses: Poisson, natural
cubic splines and natural splines
Age Groups:
Pollutant: PM2.6, EC, 0C, NO3, SO4, Ca,
CI, Cu, Fe, K, S, Si, Ti, Zn
Averaging Time: 24-h avg
Mean (SD): PMz.b: 19.28
EC: 0.966
0C: 7.129
NO3: 5.415
S(k 1.908
Ca: 0.080
CI: 0.094
Cu: 0.007
Fe: 0.124
K: 0.117
S: 0.648
Si: 0.168
Ti: 0.009
Zn: 0.012
Range (95th): PM21,: 46.91
EC: 2.57
0C: 15.91
NOs: 17.46
S(k 5.18
Ca: 0.20
CI: 0.41
Cu: 0.02
Fe: 0.34
K: 0.26
S: 1.70
Si: 0.43
Ti: 0.02
Zn: 0.04
The study does not present quantitative
results.
Reference: Perez et al. (2008,156020) Outcome: respiratory mortality
Study Design: cohort
Period of Study: 3/27/2003-
1213112005
Location: Barcelona, Spain
Covariates: temperature, humidity
Statistical Analysis: autoregressive
Poisson regression models
Statistical Package: NR
Age Groups: All deaths
Pollutant: PM2.6-i
Averaging Time: 24 h
Mean (SD) Unit: 5.5 (3.8)//gfm3
Range (Min, Max): 0.6, 45.5
Copollutant: PMwj,, PMi
Increment: 10 //g/m3
Odds Ratio |95%CI)
Lag
Avg L0-1: 0.998 (0.849-1.174), p -
0.981
L1: 1.014 (0.886-1.161), p - 0.838
L2: 1.295 (1.141-1.470), p - 0.000
Multi-pollutant Model
Avg L0-1: 0.987 (0.806-1.208), p -
0.898
L1: 1.022 (0.859-1.214), p - 0.
L2: 1.206 (1.028-1.416), p - 0.022
July 2009
E-450
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Perez et al. (2008,156020)
Outcome: cardiovascular mortality
Pollutant: PM2.6-i
Increment: 10/yg/m3
Period of Study: 3/27/2003-
1213112005
Location: Barcelona, Spain
Study Design: cohort
Averaging Time: 24 h
Odds Ratio |95%CI)
Covariates: temperature, humidity
Statistical Analysis: autoregressive
Poisson regression models
Statistical Package: NR
Age Groups: All deaths
Mean (SD) Unit: 5.5 (3.8)/yg/m3
Range (Min, Max): 0.6, 45.5
Copollutant: PMwj,, PMi
Lag
Avg L0-1:1.100 (1.002-1.207), p -
0.046
L1: 1.112(1.031-1.200),p - 0.006
L2: 1.078 (0.999-1.163), p - 0.052
Multi-pollutant Model
Avg L0-1: 0.994 (0.885-1.116), p -
0.920
L1: 0.984 (0.892-1.086), p - 0.754
L2: 0.981 (0.891-1.079), p - 0.688
Reference: Perez et al. (2008,156020)
Outcome: cerebrovascular mortality
Pollutant: PM2.6-i
Increment: 10/yg/m3
Period of Study: 3/27/2003-
12/31/2005
Location: Barcelona, Spain
Study Design: cohort
Averaging Time: 24 h
Odds Ratio |95%CI)
Covariates: temperature, humidity
Statistical Analysis: autoregressive
Poisson regression models
Statistical Package: NR
Age Groups: All deaths
Mean (SD) Unit: 5.5 (3.8)/yg/m3
Range (Min, Max): 0.6, 45.5
Copollutant: PM102 E, PMi
Lag
Avg L0-1:1.083 (0.897-1.307), p -
0.406
L1: 1.121 (0.964-1.303), p - 0.140
L2: 0.984 (0.841-1.152), p - 0.839
Multi-pollutant Model
Avg L0-1: 0.899 (0.712-1.135), p -
0.371
L1: 0.905 (0.743-1.102), p - 0.321
L2: 0.868 (0.711-1.060), p - 0.165
Reference: Rainham et al. (2005,
088676)
Period of Study: 1981-1999
Location: Toronto, Canada
Outcome: Mortality:
Pollutant: PM2.6
Increment: NR
Total (non-accidental) (< 800)
Cardiorespiratory (390-459
480-519)
Averaging Time: 24-h avg
Mean (SD):
All years: 17.0 (8.7)
% Increase (Lower CI, Upper CI)
lag:
Winter and Winter Synoptic Events

Other-causes
Winters (Dec-Feb): 17.2 (6.8)
Winter

Study Design: Time-series
Summers (June-Aug): 18.8 (10.2)
Total: 0.998% (0.997,1.000)

Statistical Analyses: Poisson GLM,
natural splines
Age Groups: All ages
Range (Min, Max): NR
Copollutant:
CO
NO2
SO2
0s
2
Cardiorespiratory: 0.998 (0.996,1.000)
2
Other: 0.998% (0.996,1.000)
2
Dry Moderate
Total: 1.001% (0.996,1.007)
1
Cardiorespiratory: 1.005(0.998,1.011)



Other: 0.997% (0.989,1.006)
0
Dry Polar






Total: 0.998% (0.995,1.001)



Cardiorespiratory: 0.995 (0.991, 0.999)
2
July 2009
E-451
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Other: 1.002% (0.998,1.005)
1
Moist Moderate
Total: 0.998% (0.993,1.002)
2
Cardiorespiratory: 1.003(0.995,1.010)
1
Other: 0.997% (0.991,1.004)
1
Moist Polar
Total: 1.001% (0.998,1.005)
1
Cardiorespiratory: 1.002(0.997,1.007)
2
Other: 1.003% (0.999,1.007)
0
Moist Tropical
Total: 1.007% (0.965,1.203)
0
Cardiorespiratory: 1.123 (1.031,1.224)
2
Other: 1.248% (1.123,1.387)
0
Transition
Total: 1.003% (0.996,1.009)
1
Cardiorespiratory: 0.996 (0.987,1.004)
0
Other: 0.997% (0.990,1.004)
0
summer and summer Synoptic Events
summer
Total: 1.000% (1.000,1.001)
0
Cardiorespiratory: 1.001 (1.000,1.002)
0
Other: 1.001% (1.000,1.002)
0
Dry Moderate
Total: 1.001% (0.999,1.002)
2
Cardiorespiratory: 1.002(0.999,1.004)
2
Other: 0.999% (0.997,1.002)
0
July 2009
E-452
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Dry Polar
Total: 1.002% (0.999,1.005)
2
Cardiorespiratory: 0.996 (0.991,1.000)
0
Other: 1.003% (0.999,1.007)
2
Dry Tropical
Total: 1.016% (1.006,1.027)
0
Cardiorespiratory: 1.017 (1.005,1.030)
2
Other: 1.017% (1.003,1.031)
0
Moist Moderate
Total: 1.002% (1.000,1.004)
2
Cardiorespiratory: 1.003(0.999,1.006)
2
Other: 1.004% (1.001,1.006)
0
Moist Polar
Total: 1.005% (0.998,1.011)
1
Cardiorespiratory: 1.008(0.997,1.018)
0
Other: 1.003% (0.995,1.011)
1
Moist Tropical
Total: 0.999% (0.997,1.001)
2
Cardiorespiratory: 0.996 (0.993,1.000)
2
Other: 0.998% (0.995,1.001)
1
Transition
Total: 1.005% (0.996,1.014)
1
Cardiorespiratory: 1.007 (0.994,1.020)
1
Other: 1.002% (0.996,1.008)
2
July 2009
E-453
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Rosenthal et al. (2008,
156925)
Period of Study: 712002-71 2006
Location: Indianapolis, Indiana
Outcome: Non-Dead on Arrival (D0A)
Out-of-Hospital Cardiac Arrests (0HCA)
Witnessed non-DOA OHCA
Study Design: Case-crossover
Statistical Analyses: Time-stratified
conditional logistic regression
Age Groups: All ages
Study Population: Non-DOA OHCA: 1,374
Witnessed non-DOA OHCA: 511
Pollutant: PM2.6
Averaging Time: 24-h avg
Hourly
Mean (SD):
NR
IQR (25th, 75th):
All non-DOA
All heart rhythms: (9.4,19.5)
OHCA: (9.6,19.5)
Referents: (9.3,19.5)
Asystole: (9.2,19.4)
OHCA: (9.2,19.7)
Asystole: (9.2,19.2)
Witnessed non-DOA hourly
All heart rhythms: (8.8, 20.7)
OHCA: (8.8,21.9)
Referents: (8.8, 20.4)
Asystole: (8.5,19.8)
OHCA: (9.4,21.3)
Referents: (8.3,19.1)
Copollutant (correlation): NR
July 2009
E-454
Increment: 10 //g/m3
Hazard Ratio (Lower CI, Upper CI)
lag:
Out-of-Hospital non-DOA Cardiac Arrests
All
1.02(0.94,1.11)
0
1.00(0.92,1.08)
1
0.98(0.90,1.06)
2
1.00(0.92,1.08)
3
1.02(0.92,1.12)
0-1 avg
1.01 (0.91,1.12)
0-2 avg
1.02(0.91,1.14)
0-3 avg
Asystole
1.03(0.91,1.17)
0
1.00(0.89,1.13)
1
1.01 (0.90,1.13)
2
0.98(0.87,1.10)
3
1.03(0.90,1.18)
0-1 avg
1.05(0.90,1.22)
0-2 avg
1.04 (0.88,1.22)
0-3 avg
Vfib
1.08(0.92,1.28)
0
1.02(0.87,1.21)
1
0.96(0.80,1.14)
2
1.10(0.93,1.31)
3
1.06(0.88,1.28)
0-1 avg
1.01 (0.82,1.25)
0-2 avg
1tM#-1f$N0T CITE OR QUOTE
0-3 avg
PEA
0 9210 77.1 08)

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Schwartz et al. (2002,
025312)
Period of Study: 1979-Late 1980's
Location: 6 U.S. cities
Outcome: Mortality:
Total (non-accidental) (< 800)
Study Design: Time-series
Statistical Analyses: Hierarchical
modeling:
1.	Poisson GAM, LOESS
2.	Multivariate modeling
Age Groups: All ages
Pollutant: PM2.6, PM2.6 sources (Traffic,
Coal, Residual Oil)
Averaging Time: 24-h avg
Mean (SD):
PM2.6 Range: (Madison: 11.3 to
Steubenville: 30.5]
Traffic Range: (Steubenville: 1.5 to
Boston: 4.8)
Coal Range: (Madison: 4.9 to
Steubenville: 19.2)
Residual Oil Range: (Boston: 0.5 to
Steubenville: 0.9)
Range (Min, Max): NR
The study does not present quantitative
results.
Reference: Simpson et al. (2005,
087438)
Period of Study: 1/1996-12/1999
Location: 4 Australian cities
Outcome: Mortality:
Non-accidental (< 800)
Cardiovascular (390-459)
Respiratory (460-519)
Study Design: Time-series
meta-analysis
Statistical Analyses: Poisson GAM,
natural splines
Poisson GLM, natural splines
Age Groups: All ages
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD):
Brisbane: PM2.6: 7.50
Sydney: PM2.6: 9.00
Melbourne: PM2.6: 9.30
Perth: PM2.6: 9.0//g/m3
Range (Min, Max):
Brisbane: PM2.6: (1.9,19.7)
Sydney: PM2.6: (2.4, 35.3)
Melbourne: PM2.6: (2.7, 35.1)
Perth: PM2.1,: (2.8, 37.3)
Copollutant: CO, NO2
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
lag:
PM2.6
0.9% (-0.7, 2.5)
Reference: Slaughter et al. (2005,
073854)
Period of Study: 1/1995-12/1999
Location: Spokane, Washington
Outcome: Mortality:
Non-accidental (< 800)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural splines
Age Groups: All ages
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD): NR
Range (9th, 95th):
PM2.6: (4.2, 20.2)
Copollutant (correlation):
PM2.6: r - 0.95
PM10: r - 0.62
PMio-2.5: r - 0.31
CO: r - 0.62
Increment:
PM2.6:10 //g/m3
PM10: 25 //g/m3
Relative Risk (Lower CI, Upper CI)
lag:
PM2.6
(0.97,1.04)
1
0.99(0.96,1.03)
2
1.00(0.97,1.03)
3
July 2009
E-455
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Stieb et al. (2002, 025205)
Period of Study: Publication dates of
studies: 1985-12/2000 Mortality series:
1958-1999
Location: 40 cities (11 Canadian cities,
19 U.S. cities, Santiago, Amsterdam,
Erfurt, 7 Korean cities)
Outcome: Mortality:
All-cause (non-accidental)
Study Design: Meta-analysis
Statistical Analyses: Random effects
model
Age Groups: All ages
Pollutant: PM2.6
Averaging Time: NR
Mean (SD): NR
Range (Min, Max): NR
Copollutant: Varied between studies:
0s
SO?
NO?
CO
Increment:
PM2.6:18.3/yg/m3
% Increase (Lower CI, Upper CI)
lag:
Single-pollutant models
18 studies
PM2.6: 2.0% (1.2, 2.7)
Multipollutant models
8 studies
PMz.b: 1.3% (0.6,1.9)
Reference: Sullivan et al. (2003,
043156)
Period of Study: 1985-1994
Location: Western Washington
Outcome: Out-of-hospital cardiac arrest Pollutant: PM2.6
Study Design: Case-crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: 19-79
Study Population: Out-of-hospital cardiac
arrests: 1,206
Averaging Time: 24-h avg
Median (SD) unit:
PM10
Lag 0: 28.05
Lag 1:27.97
Lag 2: 28.40
Range (Min, Max):
PM10: (7.38,89.83)
Copollutant (correlation): SO2, CO
Notes: Study used nephelometry to
measure particles and equated the
measurements to PM2.6 concentrations.
Increment:
PM10: 16.51 /yglm3
PM2.6:13.8/yg/m3
Odds Ratio (Lower CI, Upper CI)
lag:
Overall
PM10
1.05(0.87,1.27)
0
0.91 (0.75,1.11)
1
1.03(0.82,1.28)
2
PM2.6
0.94 (0.88,1.01)
0.94 (0.88,1.02)
1
1.00(0.93,1.08)
2
PM2.6: Stratified by subject
characteristics
~ 55
0.95(0.76,1.18)
0
0.89(0.71,1.12)
1
0.95(0.75,1.20)
2
>55
0.94 (0.88,1.02)
0
0.95, (0.88,1.03)
1
1.01 (0.93,1.10)
2
July 2009
E-456
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Male
0.95(0.87,1.03)
0
0.96(0.88,1.04)
1
1.01 (0.93,1.10)
2
Female
0.93(0.82,1.06)
0
0.92(0.80,1.07)
1
0.98(0.83,1.15)
2
White
0.93(0.86,1.01)
0
0.95(0.88,1.03)
1
1.03(0.95,1.12)
2
Non-White
1.09(0.88,1.36)
0
0.96(0.75,1.22)
1
0.88(0.68,1.14)
2
Current Smoker
1.05(0.92,1.19)
0
0.98(0.86,1.12)
1
1.06(0.92,1.22)
2
Nonsmoker
0.93(0.85,1.01)
0
0.93(0.85,1.02)
1
0.97 (0.89,1.07)
2
Drinker
1.13(0.92,1.39)
0
July 2009
E-457
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1.15(0.94,1.41)
1
1.16(0.92,1.45)
2
Nondrinker
0.94 (0.86,1.03)
0
0.93(0.85,1.02)
1
1.00(0.92,1.10)
2
Activity Level-Unrestricted
0.96(0.89,1.03)
0
0.96(0.89,1.04)
1
1.01 (0.93,1.10)
2
Activity Level-Limited
0.82(0.56,1.20)
0
0.70(0.45,1.09)
1
0.97 (0.65,1.43)
2
PM2.e: Stratified by disease state
Heart disease
0.95(0.87,1.04)
0
0.97 (0.89,1.07)
1
1.06(0.96,1.16)
2
Ischemic Heart Disease
0.91 (0.80,1.04)
0
0.97 (0.84,1.11)
1
1.09(0.95,1.26)
2
Active Angina
0.98(0.81,1.20)
0
1.07 (0.88,1.31)
1
July 2009
E-458
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1.08(0.89,1.32)
2
Congestive Heart Failure
0.91 (0.80,1.03)
0
0.99(0.87,1.13)
1
1.11 (0.97,1.26)
2
Supraventricular tachycardia
1.41 (0.97,2.04)
0
1.55(1.07,2.25)
1
1.23(0.84,1.82)
2
Bradycardia
0.97 (0.64,1.46)
0
1.29(0.85,1.96)
1
1.30(0.84,2.01)
2
Asthma
(0.80,1.27)
0
0.92(0.71,1.19)
1
0.93(0.71,1.22)
2
COPD
1.00(0.86,1.17)
1.04 (0.88,1.23)
1
1.08(0.92,1.28)
2
PM2.6: Persons with prior recognized
heart disease stratified by smoking status
All heart disease
Current smoker
1.08(0.92,1.26)
0
1.06(0.89,1.26)
1
1.29(1.06,1.55)
July 2009
E-459
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
2
Nonsmoker
0.91 (0.82,1.02)
0
0.94 (0.84,1.05)
1
0.99(0.88,1.11)
2
Ischemic Heart Disease
Current smoker
1.06(0.84,1.34)
0
0.99(0.75,1.30)
1
1.39(1.04,1.86)
2
Nonsmoker
0.86(0.73,1.02)
0
0.93(0.78,1.11)
1
0.99(0.83,1.18)
2
Active Angina
Current smoker
1.28(0.88,1.86)
0
1.26(0.79,2.01)
1
1.57 (0.99,2.48)
2
Nonsmoker
0.87 (0.68,1.12)
0
0.93(0.72,1.21)
1
0.91 (0.70,1.17)
2
Congestive Heart Failure
Current smoker
1.00(0.79,1.28)
0
1.03(0.78,1.35)
1
1.46 (1.10,1.96)
July 2009
E-460
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
2
Nonsmoker
0.88(0.76,1.03)
0
0.96(0.82,1.12)
1
0.99(0.84,1.17)
2
Supraventricular tachycardia
Current smoker
12.80 (1.05,156.57)
0
2.56 (0.82, 7.99)
1
1.15(0.46,2.86)
2
Nonsmoker
1.19(0.74,1.90)
0
1.35(0.87,2.10)
1
1.15(0.73,1.82)
2
Bradycardia
Nonsmoker
0.84 (0.14,4.95)
0
0.42 (0.03, 5.34)
1
0.51 (0.05, 5.79)
2
Nonsmoker
0.99(0.63,1.55)
0
1.42 (0.90,2.24)
1
1.39(0.88,2.20)
2
Reference: Thurston et al. (2005,
Outcome: Mortality:
Pollutant: PM2.6, and source
apportioned Increment: 10/yg/m3
097949)

PM2.b:

Total (non-accidental) (< 800)

% Increase:
Period of Study: Washington, DC:

Crustal

8/1988-12/1997. Phoenix, Arizona:
Cardiovascular (390-448)

Total (non-accidental):
1995-1997

Traffic

Study Design: Time-series

Secondary sulfate:
Location: Washington, DC and
surrounding counties
Source-apportionment
Secondary S04
Secondary NO3
Wood
Phoenix: 5.2%
Phoenix, Arizona
Statistical Analyses: Poisson GLM,
natural splines
Washington, DC: 3.8%


Motor vehicles:
July 2009
E-461
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Age Groups: Washington, DC: All ages
Phoenix, Arizona: a 65
Oil
Salt
Incinerator
Averaging Time: 24-h avg
Median (SD) unit: NR
Range (Min, Max): NR
Copollutant: PM2.6 species (Na, Mg, Al,
Si, P, S, CI, K,Ca, Sc, Ti, V,Cr, Mn, Fe,
Co, Ni, Cu, Zn, Ga, Ge, As, Se, Br, Rb, Sr,
Y, Zr, Mo, Rh, Pd, Ag, Cd, Sn, Sb, Te, I,
Cs, Ba, La, W, Au, Hg, Pb, OC, EC)
Phoenix: 0.9%
Washington, DC: 4.2%
Reference: Villeneuve et al. (2003,
055051)
Period of Study: 1986-1999
Location: Vancouver, Canada
Outcome: Mortality:
Non-accidental (< 800)
Cardiovascular (401-440)
Respiratory (460-519)
Cancer (140-239)
Study Design: Time-series
Statistical Analyses: Poisson, natural
splines
Age Groups: a 65
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD):
Daily
PM2.b: 7.9
Every 6th Day
PM2.b: 11.6
Range (Min, Max):
Daily
PM2.b: (2.0, 32.0)
Every 6th Day
PM2.6: (1.8, 43.0)
Copollutant:
SO2
CO
NO2
0s
Increment:
PM2.6 (Daily): 9.0/yg/m3
PM2.6 (6th Day): 15.7 /yglm3
% Increase (Lower CI, Upper CI)
lag:
Non-accidental
PM2.6 (Daily)
¦0.1% (-5.1, 5.2)
0-2 avg
¦0.1% (-4.1,4.1)
0
¦0.3% (-4.2, 3.7)
1
0.5% (-3.3, 4.4)
2
PM2.6 (6th Day)
¦2.8% (-7.5, 2.1)
0
2.0% (-2.6, 7.0)
1
4.5% (-0.3, 9.5)
2
Cardiovascular
PM2.6 (Daily)
1.5% (-6.1, 9.7)
0-2 avg
4.3% (-1.7,10.7)
0
¦1.0% (-7.0, 5.4)
1
¦0.5% (-6.5, 5.9)
2
PM2.6 (6th Day)
¦1.5% (-8.9, 6.5)
0
¦2.0% (-9.3, 5.8)
July 2009
E-462
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1
3.0% (-4.2,10.8)
2
Respiratory
PM2.6 (Daily)
¦0.7% (-13.1,13.4)
0-2 avg
6.7% (-3.7,18.3)
0
¦3.0% (-12.8,7.9)
1
¦5.8% (-15.2, 4.7)
2
PM2.6 (6th Day)
10.0% (-4.7, 26.8)
0
8.3% (-5.4, 24.0)
1
0.3% (-12.4,14.9)
2
Cancer
PM2.6 (Daily)
¦0.3% (-9.4, 9.8)
0-2 avg
¦4.5% (-11.2, 2.8)
0
2.7% (-5.0,11.0)
1
2.5% (-5.1,10.7)
2
PM2.6 (6th Day)
¦5.1% (-13.8, 4.5)
0
¦0.3% (-9.7,11.0)
1
0.2% (-9.1,10.4)
2
July 2009
E-463
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Wilson et al. (2007,157149) Outcome: Cardiovascular
Period of Study: 1995-1997
Location: Phoenix, Arizona
Study Design: Time-series
Statistical Analyses: Poisson GAM,
nonparametric smoothing spline
Age Groups: >25
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10 //g/m3
% Excess Risk (Lower CI, Upper CI)
lag: Central Phoenix: 11.5% (2.8, 20.9)
0-5 ma
6.6% (1.1,12.5)
1
2.0% (-3.2, 7.5)
2
Middle Phoenix: 2.9% (-4.9,11.4)
0-5 ma
6.4% (1.1,11.9)
2
Outer Phoenix: 1.6% (-6.2,10.0)
0-5 ma
'All units expressed in/yglm3 unless otherwise specified.
Table E-19. Short-term exposure - mortality - other PM size fractions.
Study
Design & Methods
Concentrations
Effect Estimates (95% CI)
Reference: Perez et al. (2008,156020) Outcome: respiratory mortality
Period of Study: 3/27/2003-12/31/2005 Study Design: cohort
Location: Barcelona, Spain
Covariates: temperature, humidity
Statistical Analysis: autoregressive
Poisson regression models
Statistical Package: NR
Age Groups: All deaths
Pollutant: PMi
Averaging Time: 24 h
Mean (SD)Unit: 20.0 (10.3)/yg/m3
Range (Min, Max): 1.9, 80.1
Copollutant: PM102 E, PMyj. i
Increment: 10 //g/m3
Odds Ratio |95%CI)
Lag
Avg L0-1: 1.005 (0.960-1.053), p -
0.824
L1: 1.012(0.969-1.056), p - 0.599
L2: 1.042 (0.998-1.087), p - 0.063
Multi-pollutant Model
Avg L0-1: 1.007 (0.957-1.059), p -
0.799
L1: 1.008 (0.961-1.058), p - 0.739
L2: 1.010(0.963-1.059), p - 0.678
Reference: Perez et al. (2008,156020) Outcome: cardiovascular mortality
Period of Study: 3/27/2003-12/31/2005 Study Design: cohort
Location: Barcelona, Spain
Covariates: temperature, humidity
Statistical Analysis: autoregressive
Poisson regression models
Statistical Package: NR
Age Groups: All deaths
Pollutant: PMi
Averaging Time: 24 h
Mean (SD)Unit: 20.0 (10.3)/yg/m3
Range (Min, Max): 1.9, 80.1
Copollutant: PM102 E, PMyj. i
Increment: 10 //g/m3
Odds Ratio |95%CI)
Lag
Avg L0-1: 1.028 (1.000-1.057), p -
0.054
L1: 1.029(1.003-1.056), p - 0.030
L2: 1.023 (0.996-1.050), p - 0.091
Multi-pollutant Model
Avg L0-1: 1.025 (0.995-1.057), p -
0.688
L1: 1.028(1.000-1.058), p - 0.053
L2: 1.024 (0.995-1.053), p - 0.110
July 2009
E-464
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Pollutant: PMi	Increment: 10/yg/m3
Averaging Time: 24 h	Odds Ratio (95%CI)
Mean (SD) Unit: 20.0 (10.3) //g/m3 Lag
Range (Min, Max): 1.9, 80.1	Avg L0-1: 1.037 (0.981-1.097), p -
0.202
Copollutant: PMubj,, PMyj. i
L1: 1.056(1.003-1.113), p - 0.039
L2: 1.020 (0.968-1.075), p - 0.460
Multi-pollutant Model
Avg L0-1: 1.042 (0.981-1.107), p -
0.179
L1: 1.063(1.004-1.124), p - 0.035
L2: 1.034 (0.976-1.095), p - 0.255
Reference: Slaughter et al. (2005,
Outcome: Mortality: Non-accidental
Pollutant: PMi This study does not present quantitative
073854)
(< 800)
results for PMi.
Averaging Time: 24-h avg
Period of Study: 1/1995-12/1999
Study Design: Time-series
Mean (SD): NR
Location: Spokane, Washington
Statistical Analyses: Poisson GLM,
Range (9th, 95th)
natural splines

Age Groups: All ages
PMi: (3.3,17.6)


Copollutant (correlation): PMi


PM2.6: r - 0.95


PM10: r - 0.50


PMio-2.5: r - 0.19


CO: r - 0.63
July 2009	E-465	DRAFT - DO NOT CITE OR QUOTE
Reference: Perez et al. (2008,156020)
Period of Study: 3/27/2003-12/31/2005
Location: Barcelona, Spain
Outcome: cerebrovascular mortality
Study Design: cohort
Covariates: temperature, humidity
Statistical Analysis: autoregressive
Poisson regression models
Statistical Package: NR
Age Groups: All deaths

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Stolzel et al. (2007, 091374) Outcome: Mortality:
Period of Study: 9/1995-8/2001
Location: Erfurt, Germany
Total (non-accidental) (< 800)
Cardiorespiratory (390-459, 460-519,
785, 786)
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups: All ages
Pollutant: MCo.i-ob, MCo.oi-2.b
Averaging Time: 24-h avg
Mean (SD):
MCo i ob: 17.6 (14.8)
MC0.01-2.B: 22.3(19.2)
IQR (25th, 75th):
MCo.i-ob: (8.4, 21.5)
MCo.oi-2.b: (10.5,27.3)
Copollutant (correlation):
MCo.i-ob
NO: r - 0.52
NO?: r - 0.60
CO: r - 0.58
MCo.01-2.B
NO: r - 0.51
NO?: r - 0.58
CO: r - 0.57
July 2009
E-466
Increment:
MCoi-ob: 13.1 /yglm3
MCo.oi-2.b: 16.8 /yglm3
Relative Risk (Lower CI, Upper CI)
lag:
Total (non-accidental)
MC0.10.B
1.010(0.986
1.034)
0
1.006(0.983
1.029)
1
1.007(0.985
1.029)
2
0.994 (0.973
1.016)
3
1.002(0.981
1.023)
4
0.997 (0.976
1.018)
5
MCo.01-2.B
1.007 (0.985
1.030)
0
1.005 (0.984
1.026)
1
1.003	(0.983
1.023)
2
0.989 (0.970
1.009)
3
1.002 (0.982
1.022)
4
0.998 (0.979
1.018)
5
Cardiorespiratory
MCoi-ob
1.004	(0.977
WAfr- DO NOT CITE OR QUOTE
0
1.004 (0.979
1	029)

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Yamazaki et al. (2007,
090748)
Period of Study: 1995-1998
Location: Hong Kong, China
Outcome: Mortality:
Intracerebral hemorrhage (431)
Ischaemic stroke (434)
Study Design: Time-stratified case-
crossover
Statistical Analyses: Conditional logistic
regression
Age Groups: a 65
Pollutant: PM7
Averaging Time: 1-h avg
Mean (SD):
Warmer Months (April-September):
40.3
Colder Months (October-March):
39.4
Range (Min, Max): NR
Copollutant (correlation): Warmer
Months
NO2: r - 0.46 to 0.63
Ox: r - -0.14 to 0.20
Colder Months
NO2: 0.42 to 0.79
Ox: r - -0.36 to -0.14
Increment: 30//gfm3
Odds Ratio (Lower CI, Upper CI)
lag:
24-h avg concentrations
Intracerebral hemorrhage
Warmer months: 1.041 (0.984,1.102)
0
Colder months: 1.005 (0.951,1.061)
0
Ischaemic stroke
Warmer months: 1.027 (0.993,1.062)
0
Colder months: 1.005 (0.973,1.039)
0
Exposure measured jointly as 24-h and 1 -h
mean concentrations
Warmer months
Intracerebral hemorrhage
1	-h with 200 /yglm3 threshold: 2.397
(1.476,3.892)
2	h
24-h: 1.019 (0.960,1.082)
0
Ischaemic stroke
1	-h with 200 /yglm3 threshold: 1.051
(0.750,1.472)
2	h
24-h: 1.018 (0.983,1.055)
0
Warmer months
Intracerebral hemorrhage
1	-h with 200/yglm3 threshold: 0.970
(0.712,1.322)
2	h
24-h: 1.015 (0.958,1.075)
0
Ischaemic stroke
1	-h with 200 /yglm3 threshold: 1.040
(0.855,1.265)
2	h
24-h: 1.003 (0.968,1.039)
0
'All units expressed in/yglm3 unless otherwise specified.
July 2009
E-467
DRAFT-DO NOT CITE OR QUOTE

-------
E.4. Long-Term Exposure and Cardiovascular Outcomes
Table E-20. Long-term exposure - cardiovascular morbidity outcomes - PM10
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Baccarelli et al. (2008,
157984)
Period of Study: 1995-2005
Location: Italy (Lombardy region)
Outcome (ICD9 and ICD10): Deep Vein
Thrombosis (DVT)
prothrombin time (PT)
activated partial thromboplastin time
(aPTT)
Age Groups: 18-84yrs
Study Design: Case-control (DVT
outcome)
Cross-sectional (PT and aPTT outcomes)
N: 871 cases
1210 controls (randomly selected from
friends and nonblood relatives of cases
frequency matched by age to cases)
Statistical Analyses: Unconditional
logistic regression (DVT outcome)
linear regression (PT and aPTT outcomes)
Covariates: sex, area of residence,
education, factor V Leiden or G2021 OA
prothrombin mutation, current use of oral
contraceptives or hormone therapy
(variables controlled using penalized
regression splines with 4 df) age, BMI,
day of year (for seasonality), index date,
ambient temperature
Season: covariate
Dose-response Investigated? Yes
Statistical Package: STATA v9.0 and R
V2.2.0
Pollutant: PMio
Averaging Time: 1 year (immediately
preceding the diagnosis date for cases or
the date of examination for controls)
assessed other averaging periods
presented in supplements (90 days, 180
days, 270 days, 2 yrs)
Mean (SD): NR
Percentiles: NR
Range (Min, Max):
Range for tertiles of exposure:
1: 12.0-44.2
2: 44.3-48.1
3: 48.2-51.5
Monitoring Stations: Monitors from 53
sites
exposure assigned by dividing area into 9
regions
Copollutant (correlation): NR
PM Increment: 10 /yglm3
Effect Estimate [Lower CI, Upper CI]:
Estimated changes of PT associated
with PMio:
Among DVT cases: -0.12 (-0.23, 0.00), p
-	0.04
Among Controls: -0.06 (-0.11, 0.00), p -
0.04
Estimated changes of aPTT
associated with PMio: Among Controls:
¦0.09 (-0.19, 0.01), p - 0.07
Among DVT cases: 0.01 (-0.03, 0.04), p
-	0.78
Risk of DVT associated with PMio (avg
of 1 yr preceding diagnosis/exam
date): |
All subjects: 1.70 (1.30, 2.23), p <
0.001
Sex: Male: 2.07 (1.50, 2.84), p <
0.001
Female: 1.40(1.02,1.92), p - 0.04
P for interaction: p - 0.02
Age: 18-35yrs: 1.57 (1.11, 2.24), p -
0.01
36-50yrs: 1.97 (1.41, 2.77), p < 0.001
51 -84yrs: 1.54 (0.90, 2.63), p - 0.12
P for interaction: p - 0.99
Premenopausal women with current
use of oral contraceptives: No: 1.53
(0.86, 2.72), p - 0.14Yes: 0.87(0.46,
1.67), p - 0.68
P for interaction: p - 0.11
Postmenopausal women with current
use of hormone therapy: No: 1.60
(0.72, 3.54), p - 0.24
Yes: 0.85 (0.29, 2.45), p - 0.76
P for interaction: p - 0.27
Current use of oral contraceptive or
hormone replacement therapy: No:
1.64(1.05, 2.57), p - 0.03
Yes: 0.97 (0.58,1.61), p - 0.89
P for interaction: p - 0.048
Body Mass Index: 13.3-22.0: 1.47
(0.97, 2.23), p - 0.07;
22.1-24.9: 1.72 (1.17, 2.54), p - 0.006
25.0-53.3: 1.83(1.03, 3.24), p - 0.04
P for interaction: p - 0.37
_Education^_EJementar^/m^ddle_school^^^
July 2009
E-468
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
1.93(1.35, 2.76), p < 0.001
High school: 1.72(1.24, 2.39), p -
0.001
College: 1.35 (0.74, 2.45), p - 0.33
P for interaction: p - 0.21
Deficiencies of natural anticoagulant
proteins: None: 1.66 (1.26, 2.18), p <
0.001
Any: 2.56 (0.91, 7.18), p - 0.07
P for interaction: p - 0.41
Factor V Leiden or G20210A
prothrombin mutation: None: 1.69
(1.27,2.23), p < 0.001
Any: 1.79 (1.05, 3.05), p - 0.03
P for interaction: p - 0.83
Hyperhomocysteinemia: No: 1.66
(1.26,2.19), p < 0.001
Yes: 2.19(1.33, 3.61), p - 0.002
P for interaction: p - 0.25
Any cause of thrombophilia: No: 1.59
(1.19,2.13), p - 0.002
Yes: 1.96 (1.34, 2.87), p < 0.001
P for interaction: p - 0.27
Year of diagnosis: 1995-97: 1.61 (1.06,
2.46), p - 0.03
1998-00: 1.34 (0.90,1.99), p - 0.15
2001-05: 2.14(1.04,4.39), p - 0.04
P for interaction: p - 0.12
Risk of DVT associated with PMio
over varying averaging times: 90 days:
0.91 (0.80,1.03), p - 0.12
180 days: 0.96 (0.82,1.13), p - 0.63
270 days: 1.26(1.01,1.57), p - 0.04
365 days: 1.70(1.30, 2.23), p - 0.0001
2 years: 1.47 (1.01, 2.14), p - 0.04
Risk of DVT associated with PMio
(year preceding diagnosis/exam date)
sensitivity analysis to evaluate the
effect of different methods for
adjusting for long-term trends:
Handling of long-term time trends:
Ignored: 1.13 (0.89,1.42), p - 0.31
Dummy variable for each year: 1.78
(1.31,2.44), p - 0.0003
Linear term: 1.32 (1.02,1.69), p - 0.03
Penalized spline, 2 df: 1.54 (1.19, 2.00),
p - 0.001
Penalized spline, 3 df: 1.64 (1.26, 2.14),
p - 0.0002
Penalized spline, 4 df: 1.70 (1.30, 2.23),
p - 0.0001
Penalized spline, 5 df: 1.70(1.29, 2.22),
p - 0.0002
July 2009
E-469
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Penalized spline, 6 df: 1.66 (1.26, 2.19),
p - 0.0003
Penalized spline, 7 df: 1.60 (1.21,2.13),
p - 0.001
Penalized spline, 8 df: 1.55 (1.15, 2.10),
p - 0.004
Reference: Baccarelli et al. (2009,
188183)
Period of Study: 111995-912005
Location: Lombardia Region, Italy
Outcome: Deep Vein Thrombosis
Study Design: Case-control
Covariates: age, sex, area of residence,
BMI, education, medication use
Statistical Analysis: logistic regression
Statistical Package: Stata
Pollutant: PMio
Risk of DVT measured with regards to
distance of residence from major road.
Specific levels of PMio not given.
Increment: NA
Relative Risk |95%CI) of DVT
All subjects, age-adjusted: 1.33 (1.03-
1.71), p - 0.03
All subjects, adjusted for covariates: 1.47
(1.10-1.96), p - 0.008
All subjects, adjusted for covariates and
background PMio exposure: 1.47 (1.11-
1.96), p - 0.008
Reference: Calderon-Garciduenas et al.
(2008,156317)
Period of Study: Children recruited
between Jul 2003 and Dec 2004
Location: Mexico (northeast or
southwest Mexico city or Polotitlan)
Outcome (ICD9 and ICD10): Plasma
Endothelin-1 (ET-1) and pulmonary
arterial pressure (PAP)
Age Groups: 6-13 years
7.9 ± 1.3 years
Study Design: Cross-sectional
N: 81 children
Statistical Analyses: Analysis of
variance by parametric one-way analysis
of variance and the Newman-Keuls
multiple comparison post test, Pearson's
correlation
Covariates: doesn't appear to have
performed multivariable analyses
however, collected information on age,
place and length of residency, daily
outdoor time, household cooking
methods, parents' occupational history,
family history of atopic illnesses and
respiratory disease, and personal history
of otolaryngologic and respiratory
symptoms
Season: No
Dose-response Investigated? No
Statistical Package: STATA v8.3, or
GraphPad Software, Inc.
Pollutant: PMio(|jg/m3)
Exposures assessed quantitatively in
Mexico City only
no monitors in Polotitlan
Averaging Time: 1, 2, and 7 days before
the exam
pollutant concentrations between 0700
and 1900 h were used for the estimates
Mean (SD): Presented only in figures
Percentiles: NR
Range (Min, Max): Presented only in
figures
Monitoring Stations: 4 (2 in northeast
and 2 in southwest Mexico City
residence and school within 5 miles of
one of these monitors)
Copollutant (correlation): 03
PM Increment: NA
Effect Estimate [Lower CI, Upper CI]:
No health effects models with measured
PM concentrations were presented
used city of residence to assign exposure
no multivariable analyses presented
Authors presented (statistically
significantly) elevated ET-1 levels among
children residing in both areas of Mexico
City as compared to Polotitlan (control
city):
Mean ± SE (pg/mL)
Control: 1.23 ± 0.06
Southwest Mexico City: 2.40 ± 0.14
Northeast Mexico City: 2.09 ± 0.10
Mexico City (overall): 2.24 ± 0.12
Authors presented (statistically
significantly) elevated PAP levels among
children residing in both areas of Mexico
City as compared to Polotitlan (control
city):
Mean ± SE (mmHg)
Control: 14.6 ± 0.4
Southwest Mexico City: 16.7 ± 0.6
Northeast Mexico City: 18.6 ± 0.9
Mexico City (overall): 17.3 ± 0.5
Correlation between ET-1 and time spent
outdoors: r - 0.31, p - 0.0012
Correlation between PAP and time spent
outdoors: r - 0.42, p - 0.0008
July 2009
E-470
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Diez Roux et al. (2008,
156401)
Period of Study: Baseline data collected
June 2000-Aug 2002
Exposure assessed retrospectively
between Aug 1982 and baseline date
Location: USA (6 field centers:
Baltimore, MD
Chicago, IL
Forsyth Co, NC
Los Angeles, CA
New York, NY
St. Paul, MN
Outcome (ICD9 and ICD10): Three
measures of subclinical atherosclerosis
(common carotid intimal-medial thickness
(CIMT), coronary artery calcification, and
ankle-brachial index (ABI))
Age Groups: 44-84 yrs (MESA cohort)
Study Design: Cross-sectional
retrospective cohort
N: 5172 for coronary calcium analysis
5037 for CIMT analysis
5110 for ABI analysis
Statistical Analyses: Generalized
Additive Models (Binomial regression:
presence of calcification
Linear regression: CIMT, ABI, amount of
calcium among persons with non-zero
calcification)
Covariates: age, sex, race/ethnicity,
socioeconomic factors, cardiovascular
risk factors (BMI, hypertension, high
density lipoprotein and low density
lipoprotein cholesterol, smoking, diabetes,
diet, physical activity
models presented with and without
adjustment for cardiovascular RFs)
Season: NA
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PMio
(|jg/m3)
Averaging Time: 20-yr imputed mean
Mean (SD): 34.1 (7.5)
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: A spatio-temporal
model was used to predict monthly PM2.6
exposures based on the geographic
location of each participant's residence.
Copollutant (correlation with 20-year
imputed mean): PM10 20-yr observed
mean
r - 0.93
PM2.6 20-yr imputed mean
r - 0.73
PM10 2001 imputed mean
r - 0.75
PM10 2001 observed mean
r - 0.80
PM2.6 2001 mean
r - 0.86
PM Increment: 21.0 |Jgfm3 (approx.
10th-90th percentile)
Effect Estimate [Lower CI, Upper CI]:
CIMT:
Relative difference (95% CI):
1.01 (1.00,1.02)
Adj. for additional CVD RFs:
1.02(1.00,1.03)
ABI:
Mean difference (95% CI):
0.002 (-0.005, 0.009)
Adj. for additional CVD RFs:
0.001 (-0.006, 0.009)
Coronary calcium:
Relative prevalence (95% CI):
1.02(0.96,1.07)
Adj. for additional CVD RFs:
1.02(0.96,1.08)
Coronary calcium (in those with
calcium):
Relative difference (95% CI):
0.98(0.84,1.13)
Adj. for additional CVD RFs:
1.01 (0.86,1.18)
Found no clear heterogeneity by age, sex,
lipid status, smoking status, diabetes
status, BMI, education or study site.
Reference: Maheswaran et al. (2005,
088683)
Period of Study: 1994-1998
Location: Sheffield, United Kingdom
Outcome (ICD9 and ICD10): Stroke
mortality (ICD9: 430-438) and Emergency
hospital admissions (ICD10:160-I69)
Age Groups: a 45 years
Study Design: Small area ecological
cross-sectional
N: 1030 census enumeration districts
(CEDs)
108 CEDs excluded from PM analyses
due to artifacts in the modeled emissions
data. The analysis was based on 2979
deaths, 5122 admissions and a
population of 199,682
Statistical Analyses: Poisson regression
Covariates: age, sex, socioeconomic
deprivation, and smoking prevalence
(some models also included age-by-
deprivation interaction)
Season: NA
Dose-response Investigated? Yes,
examined quintiles of exposure
Statistical Package: SAS
Pollutant: PM10
(|jg/m3)
Averaging Time: 5-yr avg
Mean (SD): Presented mean values and
ranges for each quintile of exposure:
1: 16.0 (< 16.8)
2: 17.5 (> 16.8, <18.2)
3: 18.8 (> 18.2, <19.3)
4: 19.8 (> 19.3, <20.6)
5: 23.3 (> 20.6)
Monitoring Stations: Used air pollution
model incorporating point, line and grid
sources of pollution and meteorological
data.
Copollutant (correlation): CO (r - 0.82)
NOx (r - 0.87)
PM Increment: NA
Effect Estimate [Lower CI, Upper CI]:
Rate Ratios (95%CI) for stroke
mortality adjusted for overdispersion
by quintile of PM10 level
Adjusted for sex and age:
1: 1 (ref)
2: 0.95 (0.84,1.08)
3: 1.12(0.99,1.27)
4: 1.16(1.03,1.32)
5: 1.39(1.23,1.58)
Adjusted for sex, age, deprivation, and
smoking:
1: 1 (ref)
2: 0.94 (0.83,1.07)
3: 1.08 (0.94,1.24)
4: 1.12(0.97,1.29)
5: 1.33(1.14,1.56)
Rate Ratios (95%CI) for emergency
hospital admissions because of stroke
by quintile of PM10 level
Adjusted for sex and age:
July 2009
E-471
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1: 1 (ref)
2: 1.06(0.95,1.17)
3: 1.10(0.99,1.23)
4: 1.25(1.12,1.38)
5: 1.40(1.26,1.55)
Adjusted for sex, age, deprivation, and
smoking:
1: 1 (ref)
2: 1.01 (0.91,1.13)
3: 0.98(0.87,1.10)
4: 1.08(0.96,1.22)
5: 1.13(0.99,1.29)
Rate Ratios (95%CI) for stroke
mortality in relation to spatially
smoothed (using a Mini radius)
modeled outdoor air pollution
quintiles
Adjusted for sex, age, socioeconomic
deprivation, age by deprivation
interaction, and smoking prevalence:
1: 1 (ref)
2: 0.86(0.75, 0.98)
3: 1.05(0.92,1.21)
4: 1.03(0.89,1.19)
5: 1.24(1.05,1.47)
Rate Ratios (95%CI) for emergency
hospital admissions because of stroke
in relation to spatially smoothed
modeled outdoor air pollution
quintiles
Adjusted for sex, age, socioeconomic
deprivation, age by deprivation
interaction, and smoking prevalence:
1: 1 (ref)
2: 1.05(0.94,1.17)
3:1.07(0.95,1.20)
4: 1.06(0.94,1.20)
5: 1.15(1.01,1.31)
Reference: Maheswaran et al. (2005,
090769)
Period of Study: 1994-1998
Location: Sheffield, United Kingdom
Outcome (ICD9 and ICD10): Coronary
Heart Disease (CHD) mortality (ICD9:
410-414) and Emergency hospital
admissions (ICD10:120-I25)
Age Groups: a 45 years
Study Design: Small area ecological
cross-sectional
N: 1030 census enumeration districts
(CEDs)
108 CEDs excluded from PM analyses
due to artifacts in the modeled emissions
data. Results based on 6857 deaths,
11407 hospital admissions and 199,682
people aged a 45 years
Statistical Analyses: Poisson regression
Covariates: age, sex, socioeconomic
JejjNvatioiijjin^^
Pollutant: PMio
Mm3)
Averaging Time: 5-yr avg
Mean (SD): Presented mean values and
ranges for each quintile of exposure:
1: 16.0 (< 16.8)
2: 17.5 (> 16.8, <18.2)
3: 18.8 (> 18.2, <19.3)
4: 19.8 (> 19.3, <20.6)
5: 23.3 (> 20.6)
Monitoring Stations: Study used an air
pollution model incorporating points,
lines, and grids as sources of pollution,
and meteorological data.
PM Increment: NA
Effect Estimate [Lower CI, Upper CI]:
Rate Ratios (95%CI) for CHD mortality
in relation to modeled outdoor air
pollution quintiles, adjusted for
overdispersion
Adjusted for sex and age:
1: 1 (ref)
2: 1.06(0.98,1.16)
3: 1.10(1.01,1.21)
4: 1.23(1.13,1.35)
5: 1.30(1.19,1.43)
Adjusted for sex, age, deprivation, and
smoking:
July 2009
E-472
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
(some models also included age-by-
deprivation interaction)
Season: NA
Dose-response Investigated? Yes,
examined quintiles of exposure
Statistical Package: SAS
Copollutant (correlation): CO (r - 0.82) 1: 1 (ref)
NOx (r - 0.87)
2: 1.03 (0.94,1.12)
3: 1.00 (0.90,1.11)
4: 1.08 (0.98,1.20)
5: 1.08 (0.96,1.20)
Adjusted for sex, age, deprivation, and
smoking (spatially smoothed using a 1km
radius):
1: 1 (ref)
2: 0.97 (0.89,1.07)
3: 1.00 (0.90,1.10)
4: 1.03 (0.93,1.15)
5:1.07 (0.96,1.21)
Rate Ratios (95%CI) for emergency
hospital admissions from CHD in
relation to modeled outdoor air
pollution quintiles
Adjusted for sex and age:
1: 1 (ref)
2: 1.08 (0.98,1.19)
3: 1.11 (1.01,1.22)
4:1.17(1.07,1.29)
5: 1.36(1.23,1.50)
Adjusted for sex, age, deprivation, and
smoking:
1: 1 (ref)
2: 1.03 (0.93,1.13)
3: 0.96 (0.86,1.07)
4: 0.97 (0.87,1.08)
5: 1.01 (0.90,1.14)
Adjusted for sex, age, deprivation, and
smoking (spatially smoothed using a 1km
radius):
1: 1 (ref)
2: 1.01 (0.92,1.11)
3: 1.04 (0.93,1.15)
4: 0.97 (0.87,1.08)
5:1.07 (0.95,1.20)
July 2009
E-473
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: O'Neill et al. (2007,156006)
Period of Study: 2000-2004
Location: USA (6 field centers:
Baltimore, MD
Chicago, IL
Forsyth Co, NC
Los Angeles, CA
New York, NY
St. Paul, MN
Outcome (ICD9 and ICD10): Creatinine
adjusted urinary albumin excretion
Assessed 2 ways: continuous log urinary
albumin/creatine ration (UACR) and
clinically defined micro- or macro-
albuminuria (UACR a 25 mg/g) versus
normal levels
Age Groups: 44-84 yrs
Study Design: Cross-sectional analyses
and prospective cohort analyses
N: 3901 participants free of clinical CVD
at baseline
Statistical Analyses: At baseline:
multiple linear regression (continuous
outcome)
binomial regression (dichotomous
outcome)
3-year change: repeated measures model
with random subject effects (estimate 3-
yr change in log UACR by levels of
exposure)
Covariates: age, gender, race, BMI,
cigarette status, ETS, percent dietary
protein
for repeated measures models: time
time x PMio
Season: NA
Dose-response Investigated? Yes,
examined quartiles of exposure
Statistical Package: SAS
Pollutant: PMio
N/m3)
Averaging Time: avg of previous month,
avg of previous 2 months (recent
exposures)
20-yr directly monitored PMio avg, 20-yr
imputed PMio avg (longer-term exposures)
Mean (SD):
Previous 20 years: 34.7 (7.0)
Previous month: 27.5 (7.9)
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: NR (used closest
monitor to residence to assign exposure
20 year imputed PMio was derived using
a space-time model)
Copollutant (correlation): PM2.6
PM Increment: 10/yglm3
Effect Estimate [Lower CI, Upper CI]:
Adjusted mean differences in log
UACR (mg/g) per increase in PMio
among participants seen at baseline
Previous 30 days
Full sample: -0.42 (-0.085, 0.002)
Within 10 km:-0.023 (-0.079, 0.034)
Previous 60 days
Full sample: -0.056 (-0.106 to -0.005)
Within 10 km:-0.040 (-0.106, 0.025)
20 yr PMio (nearest monitors)
Full sample:-0.019 (-0.072, 0.033)
Within 10 km: 0.009 (-0.067,0.085)
Imputed 20 yr exposure
Full sample:-0.002 (-0.038, 0.035)
Within 10 km: 0.016 (-0.033, 0.066)
Adjusted relative prevalence of
microalbuminuria vs high-normal and
normal levels (below 25 mg/g) per
increase in PMio among participants
without macroalbuminuria during the
baseline visit
Previous 30 days: 0.88 (0.76,1.02)
Previous 60 days: 0.83 (0.70, 0.99)
20 yr PMio (nearest monitors): 0.92
(0.77,1.08)
Imputed 20 yr exposure: 0.98 (0.87,
1.10)
Adjusted mean 3-yr change (SE) in log
UACR (mg/g) by quartiles of 1982-
2002 exposure to PMio from ambient
monitors among participants seen in
2000-20004
Full sample
~uartile:
18.5 to <29.3:0.147 (0.024)
29.3 to <33.1:0.159 (0.024)
33.1 to <36.3:0.163 (0.024)
36.3 to 55.7: 0.174 (0.023)
p-trend: 0.42
Within 10 km
~uartile:
18.5to <29.3:0.159 (0.030)
29.3 to <33.1:0.155 (0.031)
33.1 to <36.3:0.167 (0.028)
36.3 to 55.7: 0.152 (0.036)
p-trend: 0.99
Interactions with either 20 year or
shorter-term PM exposure were not
significant (p < 0.01) by gender, age,
city, race/ethnicity or study site.
July 2009
E-474
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Puett et al, (2008,156891)
Outcome: Nonfatal myocardial infarction
Pollutant: PM10
Increment: 10 |Jgfm3
Period of Study: 1992-2002
Study Design: Cohort
Averaging Time: 3,12, 24, 36 and 48
Hazard Ratio, 95% CI, 12 month


month moving averages
moving average
Location: Northeastern metropolitan US
Covariates: age in months, state of
residence, year and season
Mean (SD) Unit: NR
0.94 (0.77-1.15)

Statistical Analysis: Cox proportional
Range (Min, Max): NR


hazard




Copollutant (correlation): NR


Statistical Package: SAS



Age Groups: 30-55


Reference: Rosenlund et al. (2006,
114678)
Period of Study: 1992-1994
Location: Stockholm County, Sweden
Outcome (ICD9 and ICD10): Myocardial
infarction (Ml)
Age Groups: 45-70 yrs
Study Design: Case-control
N: 1397 cases
1870 controls
Statistical Analyses: Logistic
regression (main analysis)
also performed multinomial logistic
regression to assess cases as nonfatal,
fatal in the hospital within 28 days, and
out-of-hospital death within 28 days with
all controls as reference
Covariates: age, sex, and hospital
catchment area (frequency matched
variables)
smoking, physical inactivity, diabetes,
SES
also assessed but did not include
hypertension, BMI, job strain, diet,
passive smoking, alcohol consumption,
coffee intake, and occupational exposure
to motor exhaust and other combustion
products
Season: NA
Dose-response Investigated? No
Statistical Package: STATA v8.2
Pollutant: PMio
(modeled traffic-related pollution
also modeled PM2.6, but since the PM
correlation was high (r - 0.998) only
PM10 results were presented) (|jg/m )
Averaging Time: 30 yrs (PM only
assessed during 2000, thus assumed
constant levels during 1960-2000)
Median (5th—95th percentile):
Cases: 2.6 (0.5-6.0)
Controls: 2.4 (0.6-5.9)
Range (Min, Max): NR
Monitoring Stations: NR
Copollutant (correlation): NO2
(r - 0.93)
CO (r - 0.66)
SO?
PM Increment: 5 /yglm3 (5th to 95th
percentile distribution among controls)
Effect Estimate [Lower CI, Upper CI]:
Association of 30-yr avg exposure to air
pollution from traffic with Ml
Logistic regression
All cases: 1.00 (0.79,1.27)
Multinomial logistic regression
Nonfatal cases: 0.92 (0.71,1.19)
Fatal cases: 1.39 (0.94, 2.07)
ln-hospital death: 1.21 (0.75,1.94)
Out-of-hospital death: 1.84 (1.00, 3.40)
After adjustment for heating-related SO2,
the estimate for fatal Ml was 1.40 (0.86-
2.26) for PM10.
Reference: Zanobetti 81 Schwartz (2007,
091247)
Period of Study: 1985-1999
Location: 21 US cities (Birmingham,
Alabama
Boulder, Colorado
Canton, Ohio
Chicago, Illinois
Cincinnati, Ohio
Cleveland, Ohio
Colorado Springs, Colorado
Columbus, Ohio
Denver, Colorado
Detroit, Michigan
Honolulu, Hawaii
Houston, Texas
Minneapolis-St. Paul, Minnesota
Nashville, Tennessee
Outcome (ICD9 and ICD10): Death,
subsequent myocardial infarction (Ml
ICD9 codes 410.0-410.9), and a first
admission for congestive heart failure
(CHF
ICD9 code 428)
Age Groups: a 65 yrs
Study Design: Cohort
N: 196,000 persons discharged alive
following an acute Ml
Statistical Analyses: Cox's Proportional
Hazards Regression
Pollutant: PM10
Averaging Time: Yearly averages of
pollution for that year and lags up to the
3 previous years (distributed lag)
Mean (SD): 28.8 (all cities
SD not reported)
Percentiles: 10, 50, and 90 percentiles
listed individually for each city (Table 2)
Range (Min, Max): NR
Monitoring Stations: NR (obtained data
from the US EPA Aerometric Information
Retrieval System)
Meta-regression for city-specific results Copollutant (correlation): None
Covariates: age, sex, race, type of Ml,
number of days of coronary care and
intensive care, previous diagnoses for
atrial fibrillation, and secondary or
previous diagnoses for C0PD, diabetes,
and hypertension, and for season of initial
event (time period, and, sex, race, and
type of Ml were treated as stratification
variables)
PM Increment: 10 /yglm3
Effect Estimate [Lower CI, Upper CI]:
Hazard ratio (95%CI) for an increase in
PM for the year of failure and for the
distributed lag from the year of failure up
to 3 previous years
Death
PM10 annual: 1.11 (1.05,1.19), p -
0.001
Distributed lag model
Lag 0:1.04 (0.96,1.14), p - 0.336
Lag 1:1.07 (0.99,1.14), p - 0.070
Lag 2:1.14 (1.10,1.18), p - 0.000
Lag 3:1.06 (0.99,1.12), p - 0.077
Sum lags 0-3: 1.34 (1.14,1.52), p -
0.000
CHF
PM10 annual: 1.11 (1.03,1.21), p -
0.009
July 2009
E-475
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
New Haven, Connecticut
Pittsburgh, Pennsylvania
Provo Orem, Utah
Salt Lake City, Utah
Seattle, Washington
Steubenville, Ohio
and Youngstown, Ohio)
Season: Assessed as a confounder
Dose-response Investigated? No
Statistical Package: NR
Distributed lag model
Lag 0
Lag 1
Lag 2
Lag 3
1.09 (1.01,1.18), p - 0.030
1.09 (1.01,1.19), p - 0.038
1.13(1.02,1.25), p - 0.014
1.04 (0.97,1.12), p - 0.260
Sum lags 0-3: 1.41 (1.19,1.66), p -
0.000
PMio annual: 1.17 (1.05,1.31), p ¦
0.003
Distributed lag model
Lag 0
Lag 1
Lag 2
Lag 3
1.09 (0.92,1.30), p - 0.325
1.12(0.97,1.30), p - 0.108
1.15(1.08,1.23), p - 0.000
1.01 (0.94,1.09), p - 0.783
Sum lags 0-3: 1.43 (1.12,1.82), p -
0.005
Hazard Ratio (95%CI) for an increase in
PM (sum of the previous 3 yrs distributed
lag) for the sensitivity analyses
Death
Subjects with follow-up starting after 2nd
Ml:
1.33(1.15,1.55), p - 0.000
Subjects admitted between 1985-1996:
1.45(1.26,1.68), p - 0.000
2nd cohort definition (year defined at time
of Ml):
1.29(1.15,1.44), p - 0.000
CHF
Subjects with follow-up starting after 2nd
Ml:
1.42(1.22,1.65), p - 0.000
Subjects admitted between 1985-1996:
1.51 (1.26,1.81), p - 0.000
2nd Ml
Subjects admitted between 1985-1996:
1.62(1.23, 2.13), p - 0.001
Note: Age and sex effect modification
results presented in Figure 1
used meta-regression to examine
predictors of heterogeneity across city
and found that most predictors were not
significant modifiers of PM (Table 7)
'All units expressed in //gfm3 unless otherwise specified.
Table E-21. Long-term effects-cardiovascular- PM2.5 (including PM components/sources)
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Allen et al. (2009,189644)
Outcome (ICD9 and ICD10): Abdominal
Pollutant: PM2.6 (/yglm3)
PM Increment: 10/yg/m3
July 2009
E-476
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Period of Study: Oct 2000—Sep 2002
(exposure averaging period)
outcome assessed in 2002
Location: 5 US communities (Chicago,
Illinois
Forsyth County, North Carolina
Los Angeles, California
Northern Manhattan and the Bronx, New
York
and St. Paul, Minnesota)
part of MESA (Multi-ethnic Study of
Atherosclerosis)
aortic calcium (AAC), a marker of
systemic atherosclerosis (quantitative
measure of interest was the Agatston
score)
Age Groups: 46-88 years
Study Design: Cross-sectional
N: 1,147 participants (sensitivity analysis
among 1,269 participants)
Statistical Analyses: 2-part modeling
approach:
1)	modeled relative risk of having any
AAC using a log link and a Gaussian error
model
sensitivity analysis used modified Poisson
regression with robust error variance
2)	multiple linear regression of the log-
transformed AAC Agatston score (among
those with AAC >0)
sensitivity analysis modeled all
participants by adding 1 prior to log-
transforming
Covariates: age, gender, race/ethnicity,
BMI, smoking status, pack-year of
smoking, diabetes, education, annual
income, blood lipid concentration, blood
pressure, and medications
assessed impact of gender, age, diabetes,
obesity, use of lipid-lowering medications,
education, income, race/ethnicity, and
employment status on heterogeneity of
effects (or in sensitivity analyses)
Season: NA
Dose-response Investigated? NR
Statistical Package: SAS v9.1
Averaging Time: 2 year averaging period
(Oct 2000-Sep 2002)
Mean (SD): 15.8 (3.6) (Jg/m3
Percentiles: NR
Range (Min, Max): 10.6-24.7/yglm3
Monitoring Stations: All monitors with
1) the objective of "population exposure,"
"regional transport," or
"generalfbackground;) and 2) at least
50% data reporting in each of 8 3-month
periods over the averaging time
used monitors located within 50 km of a
study participant's residence
Copollutant (correlation): (assessed
traffic by roadway proximity)
Effect Estimate [Lower CI, Upper CI]:
Results for fully adjusted models under
different participant inclusion,
employment status, and roadway
proximity criteria.
Presence/Absence of Calcium
RR (95% CI)
Inclusion criteria: < 10yrs at address:
1.04 (0.89,1.22)
a 10yrs at address: 1.06 (0.96,1.16)
a 10yrs at address & < 10km from
monitor: 1.08 (0.98,1.18)
a 20yrs at address: 1.10 (0.99,1.22)
a 20yrs at address & < 10km from
monitor: 1.11 (1.00,1.24)
<	10yrs at address & employed: 1.02
(0.87,1.20)
a 20yrs at address & employed: 1.07
(0.89,1.27)
<	10yrs at address & not employed:
1.10(1.00,1.22)
a 20yrs at address & not employed: 1.16
(1.02,1.31)
<	10yrs at address & near major road:
0.85(0.69,1.05)
a 20yrs at address & not near major
road: 1.10(0.99,1.23)
Log-transformed Agatston Score
(Agatston >0)
% Change (95% CI)
Inclusion criteria: < 10yrs at address: ¦
6.6 (-64.0, 50.9)
a 10yrs at address: 8.0 (-29.7,45.7)
a 10yrs at address & < 10km from
monitor: 19.7 (-19.6, 58.9)
a 20yrs at address: 14.4 (-32.8, 61.7)
a 20yrs at address & < 10km from
monitor: 24.6 (-24.6, 73.8)
<	10yrs at address & employed: 29.1
(-25.7,83.8)
a 20yrs at address & employed: 43.8
(-32.4,119.9)
<	10yrs at address & not employed: ¦
15.1 (-66.3,36.1)
a 20yrs at address & not employed: ¦
14.1 (-72.6,44.4)
<	10yrs at address & near major road:
34.0 (-44.2,112.1)
a 20yrs at address & not near major
road: 3.9 (-39.9, 47.8)
Log-transformed Agatston Score (all)
% Change (95% CI)
Inclusion criteria: < 10yrs at address: ¦
8.5(-81.3, 64.2)
a 10yrs at address: 40.7 (-11.5, 92.8)
July 2009
E-477
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
a 10yrs at address & < 10km from
monitor: 60.7 (5.9,115.4)
a 20yrs at address: 64.1 (-1.73,129.9)
a 20yrs at address & < 10km from
monitor: 79.2 (10.1,148.3)
<	10yrs at address & employed: 33.5
(¦35.9,102.9)
a 20yrs at address & employed: 55.8
(¦37.2,148.7)
<	10yrs at address & not employed:
54.8 (-23.8,133.4)
a 20yrs at address & not employed: 89.3
(¦3.7,182.3)
<	10yrs at address & near major road: -
30.6 (-141.3, 80.1)
a 20yrs at address & not near major
road: 51.3 (-8.3,110.8)
Exploratory/sensitivity analyses (also
presented in figures): Detectable AAC
RR (95%CI): Among women: 1.14 (1.00,
1.30)
Among persons >65yrs: 1.10 (1.01,
1.19)
Among users of lipid-lowering
medications: 1.14 (1.00,1.30)
Among Hispanics: 1.22 (1.03,1.45)
Imputing missing covariates among
residentially stable participants: 1.08
(0.98,1.19)
Agatston score
% change (95%CI): Among Hispanics: 64
(-4,133)
Among persons earning > $50,000: 72
(5,139)
Agatston score including those with
Agatston = 0
% change (95%CI): Fully adjusted model:
41 (-12, 93)
Among persons >65yrs: 75 (8,143)
Among diabetics: 149 (29, 270)
Among users of lipid-lowering
medications: 121 (25, 217)
Among Hispanics: 141 (45, 236)
Imputing missing Covariates: 49 (1.3,
100.1)
Reference: Auchincloss et al. (2008,
156234)
Period of Study: Jul 2000-Aug 2002
Location: 6 US communities (Baltimore
City and Baltimore County, Maryland
Chicago, Illinois
Forsyth County, North Carolina
Los Angeles, California
Northern Manhattan and the Bronx, New
York
Outcome (ICD9 and ICD10): Blood
pressure: systolic (SBP), diastolic (DBP),
mean arterial (MAP), pulse pressure (PP)
Avg of 2nd and 3,d BP measurement used
for analyses
Age Groups: 45-84 years
Study Design: Cross-sectional (Multi-
Ethnic Study of Atherosclerosis baseline
examination)
N: 5,112 persons (free of clinically
apparent cardiovascular disease)
Pollutant: PM2.6 (/L/g|m3)
Averaging Time: 5 exposure metrics
constructed: prior day, avg of prior 2
days, prior 7 days, prior 30 days, and
prior 60 days
Mean (SD): Prior day: 17.0(10.5)
Prior 2 days: 16.8 (9.3)
Prior 7 days: 17.0 (6.9)
Prior 30 days: 16.8 (5.0)
Prior 60 days: 16.7 (4.4)
PM Increment: 10/yg/m3 (approx.
equivalent to difference between 90th
and 10th percentile for prior 30 day
mean)
Effect Estimate [Lower CI, Upper CI]:
Adjusted mean difference (95% CI) in PP
and SBP (mmHg) per 10 /yglm3 increase
in PM2.E (averaged for the prior 30 days)
Pulse Pressure
Adjustment variables: Person-level
Covariates: 1.04 (0.25,1.84)
Person-level cov., weather: 1.12 (0.28,
July 2009
E-478
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
and St. Paul, Minnesota)
part of MESA (Multi-ethnic Study of
Atherosclerosis)
Statistical Analyses: Linear regression Percentiles: NR
secondary analyses used log binomial
models to fit a binary hypertension
outcome
Covariates: age, sex, race/ethnicity, per
capita family income, education, BMI,
diabetes status, cigarette smoking
status, exposure to ETS, high alcohol use,
physical activity, BP medication use,
meteorology variables, and copollutants
Range (Min, Max): NR
Monitoring Stations: Used monitor
nearest the participant's residence to
calculate exposure metrics
Copollutant (correlation):
SO?
NO?
examined site as a potential confounder qq
and effect modifier
Traffic-related exposures (straight-line
heterogeneity of effects also examined by distance to a highway
traffic-related exposures, age, sex, type 2
diabetes, hypertensive status, cigarette total road length around a residence)
use
Season: Adjusted for temperature and
barometric pressure to adjust for
seasonality (because seasons vary by the
study sites)
Also performed sensitivity analyses
adjusting for season to examine the
potential for residual confounding not
accounted for by weather variables
Dose-response Investigated? Assessed
nonlinear relationships-no evidence of
strong threshold/nonlinear effects for
PM2.6
Statistical Package: NR
1.97)
Person-level cov., weather, gaseous
copollutants: 2.66 (1.61, 3.71)
Person-level cov., study site: 0.93 (-0.04,
1.90)
Person-level cov., study site, weather:
1.11 (0.01,2.22)
Person-level cov., study site, weather,
gaseous copollutants: 1.34 (0.10, 2.59)
Systolic Blood Pressure
Adjustment variables: Person-level
Covariates: 0.66 (-0.41,1.74)
Person-level cov., weather: 0.99 (-0.15,
2.13)
Person-level cov., weather, gaseous
copollutants: 2.8 (1.38, 4.22)
Person-level cov., study site: 0.86 (-0.45,
2.17)
Person-level cov., study site, weather:
1.32 (-0.18, 2.82)
Person-level cov., study site, weather,
gaseous copollutants: 1.52 (-0.16, 3.21)
Additional results: Associations became
stronger with longer averaging periods up
to 30 days. For example: Adjusted
(personal covariates and weather) mean
differences in PP: Prior day: -0.38 (-0.76,
0.00)
Prior 2 days: -0.22 (-0.65, 0.21)
Prior 7 days: 0.52 (-0.08,1.11)
Prior 30 days: 1.12(0.28,1.97)
Prior 60 days: 1.08 (0.11,2.05)
(Pattern held for additional adjustments
and for SBP results
therefore, only results for 30-day mean
differences were presented)
Additional results (not presented):
None of DBP results were statistically
significant
results for MAP were similar to SBP,
though weaker and generally not
significant
Effect modification: associations
between PM2.6 and BP were stronger for
persons taking medications, with
hypertension, during warmer weather, in
the presence of high NO2, residing £
300m from a highway, and surrounded by
a high density of roads (Fig 1)
associations were not modified for age,
sex, diabetes, cigarette smoking, study
site, high levels of CO or SO2, season ,
nor residence £ 400m fro a highway
Note: supplementary material available
on-line
Reference: Calderon-Garciduenas et al.
(2009,192107)
Period of Study: 9/11/2004-1/6/2005
Location: Mexico City and Polotitlan,
Outcome: Flow cytometry
Study Design: Panel
Covariates: NR
_Statistical_^nal^sis^_Pearson^s_
Pollutant: PM2.6
Averaging Time: 1, 2 and 7 day
averages
Mean (SD) Unit: 35.89 ± 0.93 |Jg/m3
Increment: NR
Flow cytometry results and their
statistical significance in control
versus exposed children
July 2009
E-479
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Mexico
Correlation
Statistical Package: Stata
Age Groups: 9.7 ± 1.2 years
Range (Min, Max): NR
Copollutant: PMm, 0:i
CD3
Exposed: 62.9 ±1.8
Control: 67.1 ±1.7
P - 0.1
CD4
Exposed: 39.3 ±1.3
Control: 38.2±1.4
P - 0.57
CD8
Exposed: 24.0±0.95
Control: 27.3±1.0
P - 0.02
CD4/CD8
Exposed: 1.7 ±0.14
Control: 1.4 ±0.07
P - 0.09
CD3-/CD19 +
Exposed: 11.8 ±1.0
Control: 14.8 ± 1.0
P - 0.04
CD56 +
Exposed: 11.5 ±1.2
Control: 12.4±1.5
P - 0.63
CD56+/CD3-NK
Exposed: 14.0 ±9.5
Control: 7.0±2.7
P - 0.003
HLA-DR +
Exposed: 27.5±4.2
Control: 17.0±2.4
P - 0.04
mCD14+
Exposed: 66.5±2.3
Control: 80.6±1.8
P - <0.001
CD14/CD69
Exposed: 0.20±0.07
Control: 1.0 ±0.26
P - <0.001
CD4/CD69
Exposed: 0.08±0.03
Control: 3.1 ±0.65
P - <0.001
July 2009
E-480
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Diez Roux et al. (2008,
156401)
Period of Study: Baseline data collected
June 2000-Aug 2002
Exposure assessed retrospectively
between Aug 1982 and baseline date
Location: USA (6 field centers:
Baltimore, MD
Chicago, IL
Forsyth Co, NC
Los Angeles, CA
New York, NY
St. Paul, MN
Outcome (ICD9 and ICD10): Three
measures of subclinical atherosclerosis
(common carotid intimal-medial thickness
(CIMT), coronary artery calcification, and
ankle-brachial index (ABI))
Age Groups: 44-84 yrs
Study Design: Cross-sectional
N: 5172 for coronary calcium analysis
5037 for CIMT analysis
5110 for ABI analysis
Statistical Analyses: Generalized
Additive Models (Binomial regression:
presence of calcification
Linear regression: CIMT, ABI, amount of
calcium)
Covariates: age, sex, race/ethnicity,
socioeconomic factors, cardiovascular
risk factors (BMI, hypertension, high
density lipoprotein and low density
lipoprotein cholesterol, smoking, diabetes,
diet, physical activity
models presented with and without
adjustment for cardiovascular RFs)
Season: NA
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM2.6 (/yglm3)
Averaging Time: 20-yr imputed mean
Mean (SD): 21.7 (5.0)
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: NR
Long-term exposure to PM estimated
based on residential history reported
retrospectively
all addresses geocoded
ambient AP obtained from US EPA
Copollutant (correlation): PM10 20-yr
observed mean
r - 0.64
PM10 20-yr imputed mean
r - 0.73
PM10 2001 mean
r - 0.43
PM2.6 2001 mean
r - 0.64
Due to high correlation among PM
exposures, only results of mean 20-yr
exposures are reported.
PM Increment: 12.5 |Jgfm3 (approx.
10th-90th percentile)
Effect Estimate [Lower CI, Upper CI]:
CIMT:
Relative difference (95% CI):
1.01 (1.00,1.01)
Adj. for additional CVD RFs:
1.01 (1.00,1.02)
ABI:
Mean difference (95% CI):
0.000 (-0.006, 0.006)
Adj. for additional CVD RFs:
¦0.001 (-0.006,0.006)
Coronary calcium:
Relative prevalence (95% CI):
1.01 (0.96,1.05)
Adj. for additional CVD RFs:
1.01 (0.96,1.06)
Coronary calcium (in those with
calcium):
Relative difference (95% CI):
0.99(0.88,1.12)
Adj. for additional CVD RFs:
1.01 (0.89,1.14)
July 2009
E-481
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Hoffman et al. (2007,
091163)
Period of Study: 2000-2003
Location: Ruhr area of Germany (3 large
cities: Essen, Mulheim, and Bochum)
Outcome (ICD9 and ICD10): Coronary
artery calcification (CAC)
Age Groups: 45-74 years
Study Design: Cross-sectional
N: 4494 participants
Statistical Analyses: Linear regression
(outcome - natural logarithm of CAC
score + 1)
logistic regression (outcome - CAC score
abovefbelow the age- and gender-specific
75th percentile)
Covariates: city and area of residence,
age, sex, education, smoking, ETS,
physical inactivity, waist-to-hip ratio,
diabetes, blood pressure, and lipids (and
household income in a subset)
Season: NA
Dose-response Investigated? Yes, PM
was also categorized into quartiles for
analyses
Statistical Package: NR
Pollutant: PM2.6 (/L/g|m3)
Averaging Time: One year (2002,
midpoint of the study)
Mean (SD): Total:
22.8(1.5)
High traffic exposure (£ 100m):
22.9(1.4)
Low traffic exposure (> 100m):
22.8(1.5)
Percentiles:
Q1: 21.54
Q2: 22.59
Q3: 23.75
10tl,-90"' percentile: 3.91
Range (Min, Max): NR
Monitoring Stations: Daily mean PM2.6
values for 2002 were estimated with the
EURAD model using data from official
emission inventories, meteorological
information, and regional topographical
data.
Copollutant (correlation): None
(Traffic was assessed using distance to
roadways)
Correlation between modeled daily
averages of PM2.6 and measured PM2.6:
0.86-0.88, depending on season.
PM Increment: 3.91 /yglm3 (10th-90th
percentile)
Effect Estimate [Lower CI, Upper CI]:
Percent change (95%CI) in CAC
associated with an increase in PM2.6
Unadjusted:
12.7(-7.0, 36.4)
Model 1 (adjusted for distance to major
road):
12.3 (-7.3, 35.9)
Model 2 (model 1 + city and area of
residence):
29.7 (0, 68.3)
Model 3 (model 2 + age, sex, education):
24.2(0, 55.1)
Model 4 (model 3 + smoking, ETS,
physical inactivity, waist-to-hip ratio):
17.9 (-5.3, 46.7)
Model 5 (model 4 + diabetes, blood
pressure, LDL, HDL, triglycerides):
17.2 (-5.6, 45.5)
Adjusted ORs (95%CI) for the
association between the top quarter
of PM exposure vs. the low quarter of
PM exposure and a CAC score above
the age- and sex-specific 75th
percentiles
All: 1.22 (0.96,1.54)
No CHD: 1.22 (0.95,1.57)
Men: 1.09(0.78,1.53)
Women: 1.34 (0.97,1.87)
Age < 60 yrs: 1.18 (0.83,1.68)
Age >60 yrs: 1.27 (0.93,1.75)
Nonsmokers: 1.17 (0.89,1.53)
Current smokers: 1.30 (0.83, 2.05)
Educational level
Low: 1.16 (0.86,1.57)
Medium: 1.30 (0.83, 2.05)
High: 1.62(0.81,3.25)
Additional notes:
No clear dose-response relationship
demonstrated when exposure assessed in
quartiles (Figure 2)
Participants who had not been working
full-time during the last 5 years showed
stronger effects, with possible dose-
response between PM2.5 and CAC (results
presented in Figure 3)
July 2009
E-482
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Hoffman et al. (2006,
091162)
Period of Study: Dec 2000—Jul 2003
Location: Ruhr area of Germany (2 large
cities: Essen, Mulheim)
Outcome (ICD9 and ICD10): Clinically
manifest CHD (defined as self-reported
history of a 'hard' coronary event, i.e.
myocardial infarction or application of a
coronary stent or angioplasty or bypass
surgery)
Age Groups: 45-75 years
Study Design: Cross-sectional (German
Heinz Nixdorf RBCALL study)
N: 3399 participants
Statistical Analyses: Multivariable
logistic regression
Covariates: sex, diabetes, hypertension,
smoking status, ETS, educational level,
physical activity, BMI, triglycerides, age,
cigarettes smoked per day, WHR, LDL,
HDL, HbAlc, indicator variable for cities,
indicator variable for living in northern
part of cities.
Statistical Package: SAS v8.2
Pollutant: PM2.6 (/L/g|m3)
Averaging Time: Yearly mean estimated
with model for year 2002 (on a spatial
scale of 5 km)
Mean (SD): Total: 23.3(1.4)
High traffic: 23.4 (1.4)
Low traffic: 23.3 (1.4)
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: NR
Copollutant (correlation): None (Traffic
was assessed using distance to
roadways)
Effect Estimate [Lower CI, Upper CI]:
Model 1: PM2.6 + high traffic exposure
0.92 (0.36, 2.39)
Model 2: model 1 + age, sex
0.83(0.31,2.27)
Model 3: model 2 + education, diabetes,
HbAlc, BMI, WHR, smoking status, ETS,
physical activity, city, area of residence
0.56(0.16, 2.01)
Model 4: model 3 + hypertension, lipids
0.55(0.14,2.11)
Modeled vs. Measured: r - 0.86-0.88,
depending on season
Reference: Hoffmann et al, (2009,
190376)
Period of Study: 2000-2003
Location: Ruhr area, Germany
Outcome: Peripheral Arterial Di:
Study Design:
Covariates: height, weight, medication
use, diabetes, physical activity level,
smoking, socioeconomic status,
education, population density
Statistical Analysis: NR
Statistical Package: NR
Age Groups: 45-75 years
Pollutant: PM2.6
Averaging Time: daily
Mean (SD) Unit: 22.96 (0.85)
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 3.91 |Jgfm3
Odds Ratio (95%CI) for prevalence of
peripheral arterial disease
0.87(0.57-1.34)
July 2009
E-483
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kunzli et al. (2005, 087387)
Period of Study: 1998-2003
Location: Los Angeles Basin
Outcome (ICD9 and ICD10): Carotid
intima-media thickness (CIMT)
Age Groups: Less than 40 yrs excluded
mean age - 59.2 ± 9.8
Study Design: Cross-sectional
N: 798 participants
Statistical Analyses: Linear regression
Covariates: age, sex, education, income,
smoking, ETS, blood pressure, LDL
cholesterol, treatment with
antihypertensives or lipid-lowering
medications
Season: NA
Dose-response Investigated? Yes,
assessed PM2.6 in quartiles
Statistical Package: NR
Pollutant: PM2.6 (/L/g|m3)
Averaging Time: GIS/geostatics model
to estimate 'long-term mean ambient
concentrations of PM2.6' derived from
data collected in 2000, including data
from 23 state and local monitoring
stations.
Mean (SD): 20.3 ± 2.6
Percentiles: NR
Range (Min, Max): 5.2, 26.9
Monitoring Stations: 23 monitors
Copollutant (correlation): None
PM Increment: 10/yg/m3
Effect Estimate [Lower CI, Upper CI]:
Percent change (95%CI) in CIMT
associated with an increase in PM2.6
concentration
based on a linear model with log intima-
media thickness as dependent variable
Total population: Unadjusted: 5.9 (1.0,
10.9), p - 0.018
Adjusted for age, sex, education, income:
4.4 (0.0, 9.0), p - 0.056
Adjusted for above + smoking, ETS,
multivitamins, alcohol:
4.2(-0.2, 8.9), p - 0.064
Among Females > 60 years:
Unadjusted: 19.2 (8.8, 30.5), p - 0.001
Adjusted for age, sex, education, income:
15.7(5.7, 26.6), p - 0.002
Adjusted for above + smoking, ETS,
multivitamins, alcohol:
13.8(4.0, 24.5), p - 0.002
Among those taking lipid-lowering
therapy:
Unadjusted: 15.8 (2.1, 31.2), p - 0.024
Adjusted for age, sex, education, income:
13.3(0, 28.5), p - 0.031
Adjusted for above + smoking, ETS,
multivitamins, alcohol:
13.3(-0.3, 28.8), p - 0.060
For the observed contrast between
lowest and highest exposure:
Approximately 20/yg/ m3 ->12.1% (2.0-
231%) increase in CIMT.
Among nonsmokers: 6.6% (1.0-12.3%).
The estimate was small and not
significant in current and former smokers.
Women: In the range of 6-9% per
10/yg/m3
Unadjusted means of CIMT across
quartiles of exposure were 734, 753,
758, and 774 |Jm
adjusted means trend across exposure
groups, p - 0.041
stratified results presented in figures
Reference: Miller et al. (2007, 090130)
Period of Study: 1994-2003
Location: 36 US metropolitan areas
(Women's Health Initiative)
Outcome (ICD9 and ICD10): First
cardiovascular event (myocardial
infarction, coronary revascularization,
stroke, and death from either coronary
heart disease [categorized as "definite" or
"possible"] or cerebrovascular disease)
Age Groups: 50-79 years (median age at
enrollment: 63)
Study Design: Cohort (median follow-up
of 6 yrs)
Pollutant: PM2.6 (/L/g|m3)
Averaging Time: Annual avg
concentration in 2000 (used to represent
long-term exposure)
Mean (SD): Individual exposure: 13.5
(3.7)
Citywide avg exposure: 13.5 (3.3)
Median: 13.4
PM Increment: 10 /yglm3
Effect Estimate [Lower CI, Upper CI]:
Estimated Hazards Ratio (95%CI) for the
time to the first cardiovascular event or
death associated with an increase in
PM2.6
Any cardiovascular event (first event)
Overall: 1.24 (1.09,1.41)
July 2009
E-484
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
N: 65,893 postmenopausal women
without previous cardiovascular disease
Statistical Analyses: Cox-proportional
hazards regression
Covariates: age, race/ethnicity, smoking
status, the number of cigarettes smoked
per day, the number of years of smoking,
systolic blood pressure, education level,
household income, BMI, and presence or
absence of diabetes, hypertension, or
hypercholesterolemia (also evaluated
ETS, occupation, physical activity, diet,
alcohol consumption, waist circum-
ference, waist-to-hip ratio, medical
history, medications, and presence or
absence of a family history of cardiovas-
cular disease as possible confounders in
extended models)
Season: NA
Dose-response Investigated?
Statistical Package: SAS v8.0, STATA
v8.0
Percentiles: Quintile ranges:
1: 3.4,10.9
2: 11.0,12.4
3: 12.5,14.2
4: 14.3,16.4
5: 16.5, 28.3
IQR: 11.6-18.3
10",-90Bl
Personal: 9.1-18.3
City-wide: 9.3-17.8
Range (Min, Max): Personal exposure:
3.4, 28.3
Citywide exposure: 4.0,19.3
Monitoring Stations: 573 monitors
the nearest monitor to the location of
each residence was used to assign
exposure (monitor within 30 mi of
residence
median of 20 monitors per city (range: 4-
78))
Copollutant (correlation): PMio
SO?
NO?
CO
Os
Between cities: 1.15 (0.99,1.32)
Within cities: 1.64 (1.24,2.18)
Coronary heart disease (first event):
Overall: 1.21 (1.04,1.42)
Between cities: 1.13 (0.95,1.35)
Within cities: 1.56 (1.11,2.19)
Cerebrovascular disease (first event):
Overall: 1.35 (1.08,1.68); Between
cities: 1.20 (0.94,1.54)
Within cities: 2.08 (1.28,3.40)
Ml (first event): Overall: 1.06 (0.85,1.34)
Between cities: 0.97 (0.75,1.25)
Within cities: 1.52 (0.91, 2.51)
Coronary revascularization (first event):
Overall: 1.20 (1.00,1.43)
Between cities: 1.14 (0.93,1.39)
Within cities: 1.45 (0.98, 2.16)
Stroke (first event): Overall: 1.28 (1.02,
1.61)
Between cities: 1.12 (0.87,1.45)
Within cities: 2.08 (1.25,3.48)
Any death from cardiovascular cause:
Overall: 1.76(1.25,2.47)
Between cities: 1.63 (1.10, 2.40)
Within cities: 2.28 (1.10, 4.75)
Coronary heart disease death (definite
diagnosis): Overall: 2.21 (1.17,4.16)
Between cities: 2.22 (1.06,4.62)
Within cities: 2.17 (0.60, 7.89)
Coronary heart disease death (possible
diagnosis): Overall: 1.26 (0.62, 2.56)
Between cities: 1.20 (0.54, 2.63)
Within cities: 1.57 (0.29, 8.51)
Cerebrovascular disease death: Overall:
1.83(1.11,3.00)
Between cities: 1.58 (0.90, 2.78)
Within cities: 2.93 (1.03, 8.38)
Estimated Hazard Ratios for
cardiovascular events associated with an
increase in PM2.6 according to selected
characteristics (presented adjusted H and
adjusted H including adjustment for city)
Any cardiovascular event: H: 1.24 (1.09,
1.41)
H (city): 1.69 (1.26,2.27)
Household income < $20,000: H: 1.30
(1.10,1.53)
H (city): 1.75(1.28,2.40)
Household income $20,000-49,999: H:
1.23(1.08,1.41)
H (city): 1.69 (1.25,2.27)
Household income a $50,000: H: 1.20
July 2009
E-485
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
(1.02,1.40)
6
H (city): 1.66 (1.22,2.26)
P for trend: HR: p - 0.34
HR (city): p - 0.54
Education: Not high-school graduate: H:
1.40 (1.11,1.75)
H (city): 1.88 (1.32,2.67)
Education: High school grad/trade
school/GED: H: 1.33 (1.14,1.55)
H (city): 1.79(1.32,2.44)
Education: Some college or associate
degree: H: 1.26(1.09,1.44)
H (city): 1.74 (1.29,2.34)
Education: Bachelor's degree or higher: H:
1.11 (0.94,1.31)
H (city): 1.54 (1.13,2.10)
P for trend: H: p - 0.07
H (city): p - 0.15
Age < 60 yr: H: 1.21 (0.84,1.73)
H (city): 1.66 (1.05,2.61)
Age 60-69 yr: H: 1.14 (0.93,1.39)
H (city): 1.53 (1.09,2.14)
Age > 70 yr: H: 1.34 (1.11,1.63)
H (city): 1.85 (1.34,2.56)
P for trend: H: p - 0.20
H (city): p - 0.20
Current smoker: H: 1.68 (1.06, 2.66)
H (city): 2.28 (1.33,3.92)
Former smoker: H: 1.24 (1.01,1.52)
H (city): 1.71 (1.23,2.39)
Never smoked: H: 1.18 (0.99,1.40)
H (city): 1.60 (1.16,2.21)
Living with smoker currently: H: 1.28
(0.84,1.97)
H (city): 1.65 (0.99,2.76)
Living with smoker formerly: H: 1.18
(1.00,1.38)
H (city): 1.59 (1.16,2.16)
Living with smoker never: H: 1.39 (1.07,
1.80)
H (city): 1.90 (1.31,2.78)
BMI <22.5: H: 0.99 (0.80,1.21)
H (city): 1.35 (0.96,1.88)
BMI 22.5-24.7: H: 1.16(0.96,1.40)
H (city): 1.58 (1.14,2.19)
BMI 24.8-27.2: H: 1.24(1.05,1.45)
H (city): 1.69 (1.24,2.30)
July 2009
E-486
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
BMI 27.3-30.9: H: 1.38(1.18,1.61)
H (city): 1.88 (1.38,2.56)
BMI >30.9: H: 1.35 (1.12,1.64)
H (city): 1.84 (1.33,2.55)
P for trend: H: p - 0.003
H (city): p - 0.007
Waist-to-hip ratio < 0.74: H: 1.07 (0.90,
1.29)
H (city): 1.45 (1.05,2.00)
Waist-to-hip ratio 0.74-0.77: H: 1.12
(0.95,1.31)
H (city): 1.51 (1.11,2.06)
Waist-to-hip ratio 0.78-0.80: H: 1.24
(1.07,1.44)
H (city): 1.68 (1.23,2.27)
Waist-to-hip ratio 0.81-0.86: H: 1.30
(1.13.1.50)
H (city): 1.76(1.30,2.38)
Waist-to-hip ratio >0.86: H: 1.29 (1.11,
1.50)
H (city): 1.75(1.29,2.37)
Waist circumference < 73 cm: H: 1.05
(0.86,1.27)
H (city): 1.43 (1.02,1.99)
Waist circumference 73-78 cm: H: 1.20
(1.02,1.41)
H (city): 1.63 (1.19,2.23)
Waist circumference 79-85 cm: H: 1.22
(1.05,1.41)
H (city): 1.66 (1.22,2.24)
Waist circumference 86-95 cm: H: 1.33
(1.15,1.53)
H (city): 1.80 (1.33,2.43)
Waist circumference >95 cm: H: 1.27
(1.07.1.51)
H (city): 1.73(1.26,2.36)
P for trend: H: p - 0.06
H (city): p - 0.07
Hormone-replacement therapy-Current
Use: H: 1.33 (1.09,1.61)
H (city): 1.85 (1.32,2.58)
Hormone-replacement therapy-No Current
Use: H: 1.16 (0.98,1.39)
H (city): 1.57 (1.14,2.17)
Diabetes-yes: H: 0.96 (0.67,1.37)
H (city): 1.24 (0.78,1.96)
Diabetes-no: H: 1.28 (1.12,1.47)
H (city): 1.75(1.30,2.36)
Hypertension-yes: H: 1.22 (1.02,1.45)
H (city): 1.65 (1.09,2.27)
July 2009
E-487
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Hypertension-no: H: 1.26 (1.05,1.51)
H (city): 1.74 (1.25,2.40)
Hypercholesterolemia-yes: H: 1.25 (0.94,
1.67)
H (city): 1.71 (1.15,2.54)
Hypercholesterolemia-no: H: 1.23 (1.07,
1.42)
H (city): 1.69 (1.25,2.28)
Family history of CVD- yes: H: 1.30
(1.12,1.51)
H (city): 1.80 (1.32,2.44)
Family history of CVD- no: H: 1.07 (0.83,
1.37)
H (city): 1.46 (1.00,2.12)
Time lived in current state: a 20 yr: H:
1.21 (1.06,1.39)
H (city): 1.66 (1.23,2.23)
Time lived in current state: 10-19 yr: H:
1.39(1.12,1.72)'H (city): 1.97 (1.40,
2.79)
Time lived in current state: £ 9 yr: H:
1.54 (1.06,2.26)
H (city): 2.24 (1.39,3.59)
Health insurance coverage-yes: H: 1.22
(1.07,1.39)
H (city): 1.71 (1.27,2.30)
Health insurance coverage-no: H: 1.82
(0.81,4.10)
H (city): 2.65 (1.12,6.28)
Time spent outdoors: < 30 min: H: 1.09
(0.86,1.39)
H (city): 1.56 (1.05,2.31)
Time spent outdoors: a 30 min
H: 1.26(1.05,1.50)
H (city): 1.82 (1.29,2.57)
July 2009
E-488
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: O'Neill et al. (2007,156006)
Period of Study: 2000-2004
Location: USA (6 field centers:
Baltimore, MD
Chicago, IL
Forsyth Co, NC
Los Angeles, CA
New York, NY
St. Paul, MN
Outcome (ICD9 and ICD10): Creatinine
adjusted urinary albumin excretion
Assessed 2 ways: continuous log urinary
albumin/creatine ration (UACR) and
clinically defined micro- or macro-
albuminuria (UACR a 25 mg/g) versus
normal levels
Age Groups: 44-84 yrs
Study Design: Prospective cohort
analyses (MESA cohort)
N: 3901 participants, free of clinical CVD
at baseline
Statistical Analyses: Multiple linear
regression (continuous outcome)
binomial regression (dichotomous
outcome)
Covariates: age, gender, race, BMI,
cigarette status, ETS, percent dietary
protein
Season: NA
Dose-response Investigated? Yes,
examined quartiles of exposure
Statistical Package: SAS
Pollutant: PM2.6 (/L/g|m3)
Averaging Time: avg of previous month,
avg of previous 2 months (recent
exposures)
20-yr imputed PM2.6avg (longer-term
exposures)
Mean (SD): Previous month:
16.5(4.8)
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: NR (used closest
monitor to residence to assign value for
recent exposures
20 year PM2.6 exposures were imputed
using a space-time model.)
Copollutant (correlation): PM10
PM Increment: 10 //g|m3
Effect Estimate [Lower CI, Upper CI]:
Adjusted mean differences in log
UACR (mg/g) per increase in PM2.5
among participants seen at baseline
Previous 30 days
Full sample:-0.017 (-0.087, 0.052)
Within 10 km: 0.026 (-0.067,0.119)
Previous 60 days
Full sample:-0.040 (-0.121, 0.042)
Within 10 km: -0.013 (-0.122, 0.097)
Imputed 20 yr exposure
Full sample: 0.002 (-0.048, 0.052)
Within 10 km:-0.012 (-0.076, 0.053)
Adjusted relative prevalence of
microalbuminuria vs high-normal and
normal levels (below 25 mg/g) per
increase in PM2.6 among participants
without macroalbuminuria during the
baseline visit
Previous 30 days: 0.94 (0.77,1.16)
Previous 60 days: 0.90 (0.71,1.14)
Imputed 20 yr exposure: 0.98 (0.84,
1.14)
Reference: Solomon et al. (2003,
156994)
Period of Study: Exposures measures
1966-1969
Health endpoints assessed via
questionnaire, year not reported but
apparently 30 years after exposure
assessment (given the 30 yr residency
requirement)
Location: United Kingdom
Outcome (ICD9 and ICD10): Ischemic
heart disease (a self-reported history of
medically diagnosed angina or heart
attack)
Age Groups: 45 yrs and older
Study Design: Cross-sectional
N: 1,166 women
Statistical Analyses: Log linear
modeling
Covariates: smoking, passive smoking in
childhood, tenancy, social class, worked
in industry with respiratory hazard,
childhood hospital admission for chest
problem, diabetes, BMI
Season: NA
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: Black smoke (/L/g|m3)
Averaging Time: Exposure measures
performed 1966-1969
women had to live within 5 miles of their
current address for the past 30 years to
be included
Mean (SD): 11 wards with pollution
measures were categorized into high
(mean >120 //g/m3) and low (mean
< 50 //g/m3) exposure categories when
classified according to their black smoke
levels during 1966-69
SD not reported
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: NR
Copollutant (correlation): SO2 (health
results not presented)
PM Increment: Categorical
Effect Estimate [Lower CI, Upper CI]:
Association of particulate pollution in
place of residence and ischemic heart
disease
Low (ref): 1.0
High: 1.0(0.7,1.4)
All units expressed in uillm unless otherwise specified.
July 2009
E-489
DRAFT-DO NOT CITE OR QUOTE

-------
E.5. Long-Term Exposure and Respiratory Outcomes
Table E-22. Long-term exposure - respiratory morbidity outcomes - PM10.
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ackermann-Liebrich et al.
(1997,077537)
Period of Study: 1991-1993
Location: Switzerland (Aarau, Basel,
Davos, Geneva, Lugano, Montana,
Payerne, Wald)
Outcome: Pulmonary function
Age Groups: 18-60 yrs
Study Design: Cross-sectional
N: 9651 people
Statistical Analyses: Regression
analysis
Covariates: Age, sex, height, weight,
education level, nationality, workplace
exposure
Season: NR
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PMio
Averaging Time: Continuously
measured, 12 mo. avg. used
Mean (SD): 21.2 (7.4)
Range: (10.1-33.4)
Copollutant (correlation): SO2: r ¦
NO2: r - 0.91
O3: r - -0.55
Summer Daytime O3:
r - 0.31
Excess O3: r - 0.67
Altitude: r - -0.77
0.93
PM Increment: 10/yglm3
Regression Coefficient fB (Lower CI,
Upper CI) for air pollutants as predictors
of pulmonary function
FVC: -0.0345 (-0.0407 to -0.0283)
p< 0.001
FEVi: -0.0160 (-0.0225 to -0.0095)
p< 0.001
Percent Change (Lower CI, Upper CI)
associated with increase in avg annual air
pollution concentration
Healthy Never-smokers
FVC: -3.39
p< 0.001
FEVi:-1.59
p< 0.001
All Never-smokers
FVC: -3.14
p< 0.001
FEVi:-1.06
p< 0.001
Former Smokers
FVC: -3.03
p< 0.001
FEVi: -0.42
Current Smokers
FVC: -3.21
p< 0.001
FEVi:-1.35
p< 0.001
All
FVC: -3.14
p< 0.001
FEVi:-1.03
p< 0.001
Long-term Residents
FVC:-3.16
p< 0.001
FEVi: -0.96
p< 0.001
July 2009
E-490
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Avol et al. (2001, 020552)
Period of Study: 1993-1998
Location: Southern California
Outcome: FVC, FEVi, MMEF, PEFR
Age Groups: 10 yrs
Study Design: cohort
N: 110
Statistical Analyses: Linear regression
Covariates: Sex, race, cohort entry year,
annual avg change in height, weight, BMI
Dose-response Investigated? No
Pollutant: PMio
Averaging Time: 24 h PMio averag
over 1994
Mean (SD): 15.0-66.2
PM Increment: 10/yglm3
Mean Change (Lower CI, Upper CI)
FVC:-1.8 (-9.1, 5.5)
FEVi:-6.6 (-13.5, 0.3)
MMEF:-16.6 (-32.1 to-1.1)
PEFR:-34.9 (-59.8 to-10.0)
Reference: Bayer-Oglesby et al. (2005,
086245)
Period of Study: 1992-2001
Location: Switzerland (Lugano, Zurich,
Bern, Geneva, Anieres, Biel, Langnau,
Payerne, & Montana)
Outcome: Respiratory symptoms (chronic
cough, bronchitis, cold, dry cough,
conjunctivitis, wheeze, sneezing, asthma,
& hay fever)
Age Groups: 6-15 yrs
Study Design: cross-sectional
N: 9,591 children
Statistical Analyses: Logistic
regression models
Covariates: age, sex, nationality,
parental education, number of siblings,
farming status, low birth weight, breast
feeding, smoking, family history of
asthma, bronchitis and/or atopy, mother
who smokes, indoor humidity, mode of
cooking & heating, carpeting, pets,
removal of carpets/pets for health
reasons, completed questionnaire &
month, days max temperature < 0°C,
mother's belief of association between
environmental exposures & respiratory
health
Dose-response Investigated? Yes
Statistical Package: STATA
Pollutant: PMio
Averaging Time: 12 month avg
Mean (SD): NR
Range (Min, Max): NR
Monitoring Stations: 9
Copollutant (correlation): NR
PM Increment: 10 /yglm3
"Figure 2 shows that declining levels of
PMio were associated with declining
prevalence of chronic cough, bronchitis,
common cold, nocturnal dry cough, and
conjunctivitis symptoms. For wheezing,
sneezing, asthma, and hay fever, no
significant association could be seen with
declining PMio levels."
"Figure 3 illustrates that, on an aggregate
level, across regions the mean change in
PMio levels (r pearson - 0.81,
p - 0.008). The strongest decline of
adjusted prevalence of nocturnal dry
cough was observed in Geneva, Lugano,
and Anieres, where the strongest
reduction of PMio had also been
achieved."
Reference: Burret al. (2004, 087809)
Period of Study: 3 weeks in July and
Jan 1997 and 2 weeks in Nov 1996 and
April 1997
Location: North Wales, England
Outcome: Self-report of symptoms only
for wheeze, cough, phlegm, rhinitis, and
itchy eyes.
Age Groups: all
Study Design: Repeated measures
N: 386 persons in congested streets and
425 in the uncongested streets in
199611997. Of these, 165 and 283
completed the second phase of the study.
Pollutant: PMio
Averaging Time: Mean hourly
concentrations
Mean (SD): SD NR
Congested streets -
1996-97 35.2
1998-99 27.2
Uncongested Streets
1996-97 11.6
1998-99 8.2
Monitoring Stations: 1 in con
street and 1 in uncongested
Percent change PMio in congested
streets: 22.7
Percent change PMio in uncongested
streets: 28.9
Uncongested street sampling site was 20
m from the congested street sampler.
The opening of the by-pass produced a
reduction in pollution in the congested
streets. The health effects of these
changed is likely to be greater for nasal
and ocular symptoms than for lower
respiratory symptoms. Uncertainty about
the causality arises from low response
rates and conflicting trends in respiratory
and nasal symptoms.
July 2009
E-491
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Calderon-Garciduenas et al.
(2006,091253)
Period of Study: 1999, 2000
Location: Southwest Mexico City &
Tlaxcala, Mexico
Outcome: Hyperinflation, interstitial
markings-measured by chest radiograph,
and lung function-FVC, FEVi, PEF,
FEF25-75, measured using spirometry
tests
Age Groups: 5-13 yrs
Study Design: Cohort
N: 249 (total), 230 (Southwest Mexico
City), 19 (Tlaxcala)
Statistical Analyses: Bayes test,
Spearman rank correlation, multiple
regression
Covariates: /
e, sex
Dose-response Investigated? No
Statistical Package: SAS 8.2
Pollutant: PMio
Averaging Time: 1 yr
Mean (SD):
Mexico City
1999-48
2000-45
Tlaxacala:
1994-2000: < NAAQS std
Monitoring Stations: Southwest
Mexico City—2
Tlxacala-periodic air monitoring data
Copollutant: 0:i
PM Increment: NR
% Change:
% of children with FEVi < 80% expected
value:
Mexico City (n - 77): 7.8%
Tlaxacala (n - 19): 0%
% children with hyperinflation: Mexico
City: 65.6%
No hyperinflation: 79
Mild: 72
Moderate: 56
Severe: 23
Tlaxacala: 5.3%
No hyperinflation: 18
Mild: 1
Moderate: 0
Severe: 0
% children with interstitial markings:
Mexico City: 52.6%
Number with:
No interstitial markings: 19
Mild: 0
Moderate: 0
Severe: 0
Tlaxacala: 0%
No interstitial markings: 109
Mild: 112
Moderate: 9
Severe: 0
Reference: Calderon-Garciduenas, et al.
(2003,156316)
Period of Study: Jan 1999-Jun 2000
Location: Mexico City, Tuxpam, and
Tlaxcala, Mexico
Outcome: Respiratory system changes
Age Groups: 5-17 yrs
Study Design: Case-control of subjects
examined for this study
N: 174 cases, 27 controls, children
Statistical Analyses: Chi-square test
with Yates correction, Spearman's rank
correlation test.
Dose-response Investigated? No
Statistical Package: SAS 8.2
Pollutant: PMio
Averaging Time: 12 h (daytime 08: 00-
20: 00) and nighttime (20: 00-08: 00)
Mean (SD): Mexico City
~ay/Night
Jan-Jun 1999 76.0/50.0
Jul-Dec 1999 42.8122.5
Jan-Jun 2000 75.2147.5
Daily ambient exposure of children to a
complex mixture of air pollutants
produces significant chest X-ray
abnormalities, a decrease in predicted
values of FEF25-75, FEF75, and the
FEVilFVC ratio in association with
interstitial marking on chest X-rays, a
mild restrictive pattern by spirometry,
peripheral blood abnormalities, and an
imbalance of serum cytokines.
July 2009
E-492
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Cavanagh et al. (2007,
098618)
Period of Study: Mar-Aug 2004
Location: Christchurch, New Zealand
Outcome: A clinical study of excretion of
1 -hydroypyrene (1-0HP) as a marker of
PAH exposure
Age Groups: non smoking males aged
12-18 yr
Study Design: Comparison of 2 high
pollution events and 2 low pollution
events
N: 89 male students in a boarding school
Statistical Analyses: Wilcoxon signed
rank test for paired observations, Mann-
Whitney U test
Season: Winter
Dose-response Investigated? No
Pollutant: PMio
Averaging Time: 24 h
Mean (SD):
Autumn Low
Outdoor 19 Indoor NA
Winter I
Outdoor 43 Indoor 38
Winter II
Outdoor 72 Indoor 84
Winter Low
Outdoor 12 Indoor 16
Monitoring Stations: One inside the
boarding house, and one outside
Urinary 1-0HP were raised after high-
pollutions events. Peaks were slightly
higher than for US non-smokers of similar
ages and slightly lower than for German
non-smokers of similar ages. Urinary 1-
0HP was slightly higher in asthmatics
compared to non-asthmatics.
There were no indoor sources of PAHs
(wood-burning stoves, tobacco smoke).
Diet is another source of PAHs, but all
students ate in the boarding house.
These results suggest 1-OHP could be
used as a biomarker of ambient air
pollution.
Reference: Downs et al. (2007, 092853) Outcome: FEVi, FEVi as % of FVC,
FEF25-75
Period of Study: 1991, 2002
Age Groups: 18-60 years
Location: Switzerland
Study Design: Prospective Cohort
N: 4742 people
Statistical Analyses: Linear random
effects models
Covariates: Age, sex, height, parental
smoking, season, education, nationality,
occupational exposure, smoking (status,
pack-years), atopy, BMI
Dose-response Investigated? Yes-
linear fit best
Statistical Package: SAS 9.1, STATA
8.2, R 2.4
Pollutant: PMio
Averaging Time: Annual
Mean: Mean interval exposure: 238
/yg/m3lyears
Percentiles: 25th: 197
75th: 287
PM Increment: 10/yg/m3 reduction in
annual mean
Percent I absolute reduction in annual
decline in lung function over 11-year
period (95% CI):
Annual decline in FEVi reduced by 9% I
3.1 mL (0.03-6.2)
Annual decline in FEF2E-7E reduced by 16%
111.3 mL/second (4.3-18.2)
Annual decline in FEVi as a
percentage of FVC of 0.06 (0.01-0.12)
A reduction in interval exposure of 109
|jg per m3 cubic meter-years (equivalent
to a reduction of 10 |Jgfm3 in the annual
avg during the mean follow-up time of
10.9 years) was associated with:
A reduction of 6.9 mL (95% CI, 2.1 to
11.7) in the annual decline in FEVi
A 22% reduction in the annual decline in
FEF26-76 (i.e., by 14.0 mL per second
95% CI, 3.1 to 24.8)
July 2009
E-493
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Gauderman et al. (2000,
012531)
Period of Study: 1993-1997
Location: Southern California
Outcome: FVC, FEVi, MMEF, FEFjb
Age Groups: fourth, seventh, or tenth
Study Design: cohort
N: 3035 subjects
Statistical Analyses: Linear regression
Covariates: Height, weight, BMI,
asthma, smoking, exercise, room
temperature, barometric pressure
Dose-response Investigated? Yes
Statistical Package: SAS
Pollutant: PMio
Averaging Time: 24 h avg PMio
Mean (SD): PMio 51.5
Copollutant (correlation): PM2.6
r - 0.96
0s r - -0.32
PMio-2.5 r - 0.92
NO? r - 0.65
Inorg. Acid r - 0.68
PMio Increment: 51.5/yg/m3
% Change (Lower CI, Upper CI)
PMio-4th grade
FVC-0.58 (-1.14 to-0.02)
FEVi-0.85 (-1.59 to-0.10)
MMEF-1.32 (-2.43 to-0.20)
FEF75-1.63 (-3.14 to -0.11)
PMio-7th grade
FVC-0.45 (-1.03, 0.13)
FEVi-0.44 (-1.10, 0.23)
MMEF-0.48 (-2.51,1.59)
FEFjb -0.50 (-2.26,1.29)
PMio-10th grade
FVC 0.07 (-0.99,1.13)
FEVi-0.46 (-1.84, 0.94)
MMEF-0.71 (-4.87,3.63)
FEF7B-1.54 (-5.61, 2.71)
Reference: Gauderman et al. (2002,
026013)
Period of Study: 1996-2000
Location: Southern California
Outcome: Lung function development:
FEVi, maximal midexpiratory flow
(MMEF)
Age Groups: Fourth grade children (avg
age - 9.9 yrs)
Study Design: Cohort study
N: 1678 children, 12 communities
Statistical Analyses: Mixed model
linear regression
Covariates: Height, BMI, doctor-
diagnosed asthma and cigarette smoking
in previous year, respiratory illness and
exercise on day of test, interaction of
each of these variables with sex,
barometric pressure, temperature at test
time, indicator variables for field
technician and spirometer
Dose-response Investigated? Yes
Statistical Package: SAS (10)
Pollutant: PMio
Averaging Time: Annual 24 h averages
Mean (SD): The avg levels were
presented in an online data supplement
(Figure E1)
Monitoring Stations: 12
Copollutant (correlation): O3 (10 AM to
6 PM) r - 0.13
0s r - -0.37
NO? r - 0.64
Acid vapor r - 0.79
PM2.6 r - 0.95
PMio-2.5 r - 0.95
EC r - 0.86
0C r - 0.97
PM Increment: 51.5/yg/m3
Association Estimate:
None of the pulmonary function tests had
a statistically significant correlation with
PMio
FEVi r - -0.12 p - 0.63
MMEF r - -0.22 p - 0.30
July 2009
E-494
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Gauderman et al. (2004,
056569)
Period of Study: Air pollution data
ascertainment: 1994-2000. Spirometry
testing: spring 2001- spring 2003
Location: 12 Communities in Southern
California
Outcome: Lung function
FVC, FEVi, MMEF (Maximal
midexpiratory flow rate)
Age Groups: Children, Avg age 10 years
Study Design: Prospective Cohort Study
N: 12 Communities
2,034 Children
24,972 child-months
Statistical Analyses: Linear regression
of changes in sex-and-community specific
lung growth function and PM
Covariates: Random effect for
communities
Season: ALL (except for PM2.6)
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PMio
Averaging Time: 24-h measurements
over each year used to create annual avg
Mean: Means are presented in figures
only.
Range (Min, Max): —15, —65
Monitoring Stations: 12
Copollutant (correlation):
0s:r - 0.18
NO?: r - 0.67
PM2.6: r - 0.95
EC: r - 0.85
0C:r - 0.97
PM Increment: Most to least polluted
community Range:
PM10: 51.4/yg/m3
EC: 1.2/yg/m3
0C: 10.5/yg|m3
Difference in Lung Growth [Lower CI,
Upper CI];
FVC-60.2 (-190.6 to 70.3)
FEVi-82.1 (-176.9 to 12.8)
MMEF-154.2 (-378.3 to 69.8)
EC:
FVC -77.7 (-166.7 to 11.3)
FEVi-87.9 (-146.4 to-29.4)
MMEF-165.5 (-323.4 to-7.6)
0C:
FVC-58.6 (-196.1 to 78.8)
FEVi-86.2 (-185.6 to 13.3)
MMEF -151.2 (-389.4 to 87.1)
Correlation with % below 80% predicted
Lung function (p-value)
PM10: 0.66(0.02)
EC: 0.74 (0.006)
Reference: Gauderman et al. (2007,
090121)
Period of Study: 1993-2004
Location: 12 Southern California
Communities
Outcome: pulmonary function tests FVC, Pollutant: PM10
FEVi, MMEF/FEF26.76
Age Groups: Children (mean age 10 at
recruitment, followed for 8 years)
Study Design: Cohort Study (Children's
Health Study)
N: 3677 children
(1718 in cohort 1 recruited 1993 and
1959 in cohort 2 recruited 1996)
22686 pulmonary function tests.
Statistical Analyses: Hierarchical mixed
effects model with linear splines
Covariates: Adjustments for height,
height squared, BMI, BMI squared,
present asthma status, exercise or
respiratory illness on day of test, smoking
in previous year, field technician, traffic
indicator (distance from freeway,
distance from major roads), random
effects for participant and community.
Dose-response Investigated? no
Statistical Package: SAS
Monitoring Stations: 1 in each
community
PM Increment: 51.4/yg/m3
Pollutant effect reported as difference in
8 year lung function growth from least to
most polluted community. Negative
difference indicates growth deficits
associated with exposure. For PM10 FEV
growth deficit is -111
July 2009
E-495
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Goss et al. (2004, 055624)
Period of Study: 1999-2000
Location: USA
Outcome: Cystic Fibrosis pulmonary
exacerbations, FEVi
Age Groups: > 6
Study Design: cohort
N: 11484 patients
Statistical Analyses: Logistic
regression, t-tests, Mann-Whitney tests,
Chi-squared tests, polytomous regression,
multiple linear regression
Covariates: Age, sex, lung function,
weight, insurance status, pancreatic
insufficiency, airway colonization,
genotype, median household income by
census tract, zipcode.
Dose-response Investigated? No
Statistical Package: STATA, SAS
Pollutant: PMio
Averaging Time: annual mean of 24 h
averages
Mean (SD): 24.8(7.8) mglm3
Percentiles: 25th: 20.3
50th(Median): 24.0
75th: 28.9
Monitoring Stations: 626
PM Increment: 10/yg/m3
Odds Ratio Estimate [Lower CI, Upper
CI]:
Odds of having 2 or more pulmonary
exacerbations as compared to 1 or less in
2000
1.08(1.02-1.15)
Odds of having 2 or more pulmonary
exacerbations as compared to no
exacerbations in 2000
1.09(1.02-1.17)
Decrease in FEVi 38ml(18-58)
Reference: Hanigan et al, (2008,
156518)
Period of Study: Fire Season (April-
November) from 1996-2005
Location: Darwin, Australia
Outcome: Respiratory admissions
Study Design: time-series
Covariates: Race, age
Statistical Analysis: Over-dispersed
Poisson generalized linear models
Statistical Package: R
Age Groups: All
Pollutant: PMio
Averaging Time: Daily levels estimated
from visibility data
Mean Unit: "Only reported for 2005*
15.31 /yglm3
Range (Min, Max): 6.93, 31.12
Copollutant (correlation): NR
Increment: 10 //g/m3
Percent Increase (95% CI)
"Full results reported visually in Figure 3*
Total Respiratory Admissions
4.81 % (-1.04-11.01)
Indigenous Respiratory Admissions, No
Lag
9.40% (1.04-18.46)
Non-Indigenous Respiratory Admissions,
No Lag
3.14% (-2.99-9.66)
Indigenous Respiratory Admissions, Lag 3
15.02% (3.73-27.54)
Non-Indigenous Respiratory Admissions,
Lag 3
0.67% (-7.55-9.61)
Indigenous Asthma Admissions, Lag 1
16.27% (3.55-40.17)
Non-Indigenous Asthma Admissions, Lag
1
8.54% (-5.60-24.80)
July 2009
E-496
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ho et al. (2007, 093265)
Period of Study: Oct 1995-Mar 1996
Location: Taiwan, Republic of China
Outcome: Asthma
Age Groups: 10-17 yrs
Study Design: Screened junior high
students for asthma, collected
meteorological data to determine the
relationship.
N: 69,367
Statistical Analyses: Logistic
regression model, the maximum likelihood
estimation with Fisher's scoring
algorithm, stepwise regression model,
Wald statistic, Akaike criteria. GEE,
GENMOD
Covariates: Wind, barometric pressure,
temperature, rain, humidity
Season: Fall-spring
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PMio
Averaging Time: Monthly
Monitoring Stations: 72
Odds Ratio from stepwise regression
model:
Females (n - 32, 648)
0.993 [0.990-0.997]
Males: NS
Higher PMio concentration resulted in
less asthma prevalence. However, a
higher number of rain days seemed to
reduce asthma prevalence
rain days might interact with PMio.
Reference: Hong et al. (2004,156565)
Period of Study: 2001
Location: Kerinci, SP7, and Pelalawan,
Indonesia
Outcome: Respiratory symptoms
Age Groups: < 12 yrs
Study Design: Disproportionate random
sampling was used to select 100
households from each village. An
interviewer interviewed all children
through the caregiver/parent to obtain
symptoms in the past 2 weeks (cough,
cold, phlegm) and the last 12 months.
N: 382 children
Statistical Analyses: Chi-square test,
analysis of variance, prevalence rates,
adjusted odds ratios, multivariate
adjusted odds ratios from multiple logistic
regression models, allowing for
clustering.
Covariates: Age, gender, no. of children
in household, household income, floor
area of house, fuel for cooking, no. of
smokers in household, personal and
family medical history.
Dose-response Investigated? No
Statistical Package: SPSS STATA v.7
Pollutant: PMio
Averaging Time: 24 h measurements
were taken daily from 2
weeks before the field survey to 1 month
after the survey
Mean (SD): Kerinci 102.9 (49.6)/yg/m3
SP7 73.7 (41.7)
Pelalawan 26.1 (14.5)
P< 0.01
Range (Min, Max):
Kerinci 25,184
SP7 13,138
Pelalawan 10, 66
Monitoring Stations: 3
PM Increment: Low (Pelalawan),
Medium (SP7), & High (Kerinci) PM
Exposure
Odds Ratios (95% CI) for Symptoms by
village:
Cough/cold past 2 wks
Pelalawan 1.00
SP7 2.03(1.04,3.96)
Kerinci 3.17 (1.43, 7.07)
Respiratory symptoms last 12 months
Pelalawan 1.00
SP7 1.15(0.58, 2.26)
Kerinci 1.42 (0.62, 3.25)
Ever had rhinitis wfo flu
Pelalawan 1.00
SP7 2.17(0.57, 8.29)
Kerinci 0.56 (0.11,2.83)
Ever had wheezing
Pelalawan 1.00
SP7 0.85 (0.35, 2.08)
Kerinci 1.18 (0.46, 3.01)
Reference: Horak et al. (2002, 034792) Outcome:
Period of Study: 1994-1997
Location: Lower Austria
Lung function growth measured by
changes in: 1. FVC (forced vital capacity)
2.	FEVi
3.	MEF21, -/!, (midexpiratory flow between
25-75% of the forced vital capacity)
Age Groups: 2-3 grade schoolchildren
(mean age - 8)
Study Design: Prospective cohort with
repeated measures
N: 975 children
Statistical Analyses: Linear regression
GEE, nonstationary M-dependent
correlation structure
Pollutant: PMio
Mean (SD): Winter: 21.0 (4.8)
summer: 17.4 (2.8)
Range (Min, Max):
Winter: 9.4-30.5
summer: 11.7-28.9
Monitoring Stations:
NR, stations were located in the
immediate vicinity of each of the 8
elementary schools
Copollutant (correlation): Winter
0a: (r - -0.581)
PM Increment: 1 /yglm3
Mean per unit increase in PM (p-value)
Outcome: difference per day of FVC
(mL/day)
Summer: 0.001 (0.938)
Winter: 0.008 (0.042)
Controlling for temperature:
Summer: -0.007 (0.417)
Winter: -0.003 (0.599)
Controlling for O3:
Summer: 0.001 (0.911)
July 2009
E-497
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Covariates: Gender, atopy, ETS
SO? (r - 0.520)
Winter: 0.010(0.019)
exposure, baseline lung function, first
NO? (r - 0.595)
Controlling for NO2:
height, height difference, school site
Season: Winter, summer
summer
Summer: -0.018 (0.056)
Dose-response Investigated? No
Os (r - -0.429)
Winter: 0.015(0.000)

SO? (r - 0.335)
Controlling for SO2:

NO? (r - 0.412)
Summer: 0.005 (0.575)


Winter: 0.004 (0.492)


In non-asthmatic children:


Summer: -0.003 (0.710)


Winter: 0.009 (0.030)


In group not exposed to ETS:


Summer: 0.014 (0.154)


Winter: 0.012(0.0018)


In group exposed to ETS:


Summer: 0.022 (0.088)


Winter: 0.003 (0.656)


Outcome: difference per day of FEVi


(mL/day)


Summer: -0.023 (0.003)


Winter: 0.001 (0.885)


Controlling for temperature:


Summer: -0.034 (0.000)


Winter: -0.011 (0.016)


Controlling for O3:


Summer: -0.022 (0.008)


Winter: 0.004 (0.338)


Controlling for NO2:


Summer: -0.038 (0.000)


Winter: 0.011 (0.005)


Controlling for SO2:


Summer: -0.022 (0.010)


Winter: -0.005 (0.358)


Outcome: difference per day MEF25-75


(mL/day)


Summer: -0.090 (0.000)


Winter: -0.008 (0.395)


Controlling for temperature:


Summer: -0.112 (0.000)


Winter: -0.013 (0.295)


Controlling for O3:


Summer: -0.087 (0.000)


Winter: -0.008 (0.434)


Controlling for NO2:


Summer: -0.102 (0.000)


Winter: 0.005 (0.610)


Controlling for SO2:
July 2009
E-498
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)



Summer: -0.095 (0.000)



Winter:-0.011 (0.474)
Reference: Hwang et al. (2006, 088971) Outcome: Peak expiratory flow rate
(PEFR), Forced Expiratory Volume in 1
Period of Study: 2001
Location: Taiwan
second (FEVi), Forced Vital Capacity
(FVC), Self reported "frequent coughing,"
Self reported "shortness of breath," Self
reported " irritation of respiratory tract"
Age Groups: 24-55 years (mean - 40)
Study Design: Cohort
N: 120 men (60 traffic policemen and 60
controls)
Statistical Analyses: AN0VA, odds
ratios calculated from 2X2 table
Dose-response Investigated? No
Pollutant: PMio
Mean (SD): 55.58 (16.57)
Percentiles: 25th: 42.96
50th(Median): 53.81
75th: 70.37
Range (Min, Max): 29.36, 99.58
Monitoring Stations: 22
Copollutant (correlation): NOx
(r - 0.34)
SO? (r - 0.58)
CO (r - 0.27)
0s (r - 0.28)
PM Increment: 10//g/m3
RR Estimate [Lower CI, Upper CI]
Single pollutant model: 1.00 [0.99,1.02]
Controlling for NOx: 0.99 [0.97,1.00]
Controlling for CO: 1.00 [0.99,1.01]
Controlling for O3:1.00 [0.99,1.02]
Reference: Hwang et al, (2008,134420)
Outcome: Oral Cleft
Pollutant: PM10
Increment: 10 //g/m3
Period of Study: 2001-2003
Study Design: Case-control
Averaging Time: hourly
Odds Ratio (Min CI, Max CI);
Location: Taiwan
Covariates: Maternal age, plurality,
Mean (SD) Unit:
Single Pollutant Model

gestational age, population density and
Average: 54.83 ± 13.07 //g/m3
Month 1: 1.01 (0.96-1.06)

season of conception

Statistical Analysis: logistic regression
Spring: 64.44 ± 16.21 //g/m3
Month 2: 1.00 (0.95-1.05)

Age Groups: Infants
Summer: 39.11 ± 8.31 //g/m3
Month 3: 0.99 (0.95-1.05)


Fall: 47.76 ± 11.77 //g/m3
Two Pollutant Model (O3 + PM10)


Winter: 68.00 ± 21.88//g/m3
Month 1:0.99 (0.94-1.04)


Range (Min, Max):
Month 2: 0.99 (0.94-1.04)


Average: 20.75-78.05 //g/m3
Month 3: 0.98 (0.93-1.04)


Spring: 23.33-94.33//g/m3
Two Pollutant Model (CO + PM10)


Summer: 17.33-60.00 //g/m3
Month 1: 1.01 (0.96-1.06)


Fall: 21.00-72.00 //g/m3
Month 2: 1.00 (0.95-1.05)


Winter: 21.33-116.00 //g/m3
Month 3: 0.99 (0.95-1.05)


Copollutant (correlation):
Two Pollutant Model (NOx + PM10)


CO:-0.19
Month 1:1.02 (0.97-1.08)


NO,: 0.56
Month 2: 1.01 (0.95-1.07)


0s: 0.39
Month 3: 1.01 (0.95-1.07)


SO2: 0.50
Three Pollutant Model (O3 + CO + PM10)



Month 1:0.99 (0.94-1.04)



Month 2: 0.99 (0.94-1.04)



Month 3: 0.99 (0.93-1.04)



Three Pollutant Model (O3 + NOx + PM10)



Month 1:1.00 (0.94-1.06)



Month 2: 0.98 (0.92-1.05)



Month 3: 1.00 (0.93-1.06)
July 2009
E-499
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ingle et al. (2005, 089014)
Period of Study: May 2003-April 2004
Location: Jalgaon City, India
Outcome: Peak expiratory flow rate
(PEFR), Forced Expiratory Volume in 1
second (FEVi), Forced Vital Capacity
(FVC), Self reported "frequent coughing,"
Self reported "shortness of breath," Self
reported " irritation of respiratory tract"
Age Groups: 24-55 years (mean - 40)
Study Design: Cohort
N: 120 men (60 traffic policemen and 60
controls)
Statistical Analyses: AN0VA, odds
ratios calculated from 2X2 table
Dose-response Investigated? No
Pollutant: PMio
Mean (SD): Location-specific means:
Prabhat: 224 (27)
Ajanta: 269 (41)
Icchdevi: 229 (24)
Monitoring Stations: 3
OR Estimate [p-value]
Self reported frequent coughing
2.96 [p< 0.05]
Self reported shortness of breath
1.22 [p< 0.05]
Self reported irritation in respiratory tract
7.5 [p < 0.05]
Observed/expected lung function
p-value for difference between groups:
FVC (L)
Traffic policemen: 0.82
Controls: 0.99
Traffic policemen:
Obs - 3.03 ± 1.7 Exp - 3.70 ± 2.8
Controls:
Obs - 3.18 ± 0.91 Exp - 3.19 ± 1.71
FEV, (L)
Traffic policemen: 0.73
Controls: 1.18
Traffic policemen:
Obs - 2.27 ± 1.05 Exp - 3.08 ± 2.7
Controls:
Obs - 3.61 ± 0.90 Exp - 3.06 ± 0.91
PEFR (Us)
Traffic policemen: 0.66
Controls: 0.92
Traffic policemen:
Obs - 6.05 ± 2.15 Exp - 9.21 ± 0.47
Controls:
Obs - 5.54 ± 1.85 Exp - 6.11 ± 2.31
Reference: Islam et al. (2007, 090697)
Outcome: Respiratory symptoms,
Pollutants: PMio
The study doesn't present quantitative

Asthma

results on PMio.
Period of Study: 2006

Averaging Time: 24-h avg


Study Design: Longitudinal study cohort


Location: 12 California communities
Copollutants (correlation): O3


Statistical Analyses: Cox proportional
NO?


hazards regression


Age Groups: 7-9
EC


10-11
OC


>11


July 2009
E-500
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Janssen et al. (2003,
133555)
Period of Study: 4/1997-7/1998
Location: Netherlands-24 schools
Outcome: Symptoms of asthma and
allergic disease (asthma, conjunctivitis,
hay fever, itchy rash, eczema, phlegm,
bronchitis), skin prick test (SPT) reaction
to allergens, lung function (forced vital
capacity [FVC], forced expiratory volume
in one second [FEVi], and positive test for
fall in FEVi a 15% after inhalation of
maximal 23 mL hypertonic saline [BHR -
bronchial hyper-responsiveness])
Age Groups: 7-12 years old
Study Design: Cohort
N: 24 schools (see notes)
Statistical Analyses: Multilevel model
Covariates: Age, sex, non-Dutch
nationality, cooking on gas, current
parental smoking, current pet possession,
parental education level, number of
persons in the household, presence of an
unvented water heater in kitchen,
questionnaire not filled out by the mother,
presence of mold stains in kitchen or
living room or bedroom, parental
respiratory symptoms, distance of home
to motorway, cough or cold at time of
lung function measurement, bronchitis or
severe cold or flu in 3 weeks preceding
measurement, season
Dose-response Investigated? No
Statistical Package: MLwiN
Pollutant: PM2.6
Averaging Time: Annual
Mean (SD): 20.5//gfm3 (2.2)
Percentiles:
25th: 18.6
50th (Median): 20.4
75th: 22.1
Range (Min, Max):
17.3, 24.4
PM Increment: 'Difference between the
maximum and the minimum of the
exposure indicator' (3.5/yg/m3)
RR Estimate [Lower CI, Upper CI]
lag:
Current wheeze 1.51 (0.90, 2.53)
Asthma ever 1.03 (0.59,1.82)
Current conjunctivitis 2.08 (1.17, 3.71)
Hay fever ever 2.28 (1.13, 4.57)
Current itchy rash 1.63 (0.91, 2.89)
Eczema ever 1.31 (0.94,1.83)
Current phlegm 1.53 (0.74, 3.19)
Current bronchitis 1.71 (0.84, 3.50)
Elevated total IgE 1.45 (0.74, 2.84)
Any allergen (spt reactivity) 1.33 (0.83,
2.11)
Indoor allergens (spt reactivity) 1.17
(0.70,1.94)
Outdoor allergens (spt reactivity) 1.90
(1.06, 3.40)
FVC < 85% predicted 0.54 (0.29,1.00)
FEVi < 85% predicted 0.88 (0.37, 2.09)
BHR 0.93 (0.51,1.68)
Notes:
Figure 1 of the article illustrates the
association between exposures, including
PM2.6, and various respiratory symptoms
among children with and without a
positive SPT and positive BHR. In
general, the association between PM2.6
and respiratory symptoms were higher for
children with a positive SPT or BHR,
except for the outcome of current
phlegm. This effect appeared to be the
strongest for children with a positive
BHR, particularly for current wheeze and
current bronchitis.
The authors also reported separate
analyses for children with SPT reactivity
for indoor and outdoor allergens, but did
not report any clear differences between
the two groups. The authors did report, in
the text, that the OR of PM2.6 exposure
for children sensitized for outdoor
allergens was 7.64 for current itchy rash
(p < 0.05).
July 2009
E-501
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kan, et al. (2007, 091383)
Period of Study: 1987-1992
Location: Four Communities in the U.S.:
Forsyth County, North Carolina
Jackson, Mississippi
northwest suburbs of Minneapolis,
Minnesota
and Washington County, Maryland.
Outcome: FEVi and FVC
Age Groups: Middle-:
54.2 years)
Pollutant: PMio
i (mean age was Averaging Time: 24-h PMio averag
over study period
Study Design: Hierarchical regression
N: 15,792
Statistical Analyses: SAS PROC MIXED Copollutant:
NO?
PM Component: Vehicle emissions
Monitoring Stations: 0
Covariates: Distance to major roads,
traffic exposure, age, ethnicity, sex,
smoking, environmental tobacco smoke
exposure, occupation, education, medical
history, BMI.
Dose-response Investigated? No
Statistical Package:
SPSS Version 11 for traffic density,
SAS Version 9.1.2 for statistical analysis
0s
RR Estimate (Lower CI, Upper CI):
(Note: for ARIC participants living < 150
meters from major roads)
Women
FEVi(mL)
Age-adjusted model
-29.5 (-52.2 to -6.9)
Multivariate model
¦15.7 (-34.4 to-2.9)
FVC (mL)
Age-adjusted model
-33.2 (-60.4 to -5.9)
Multivariate model
¦24.2 (-46.2,-2.3)
FEVi/FVC (%)
Age-adjusted model
-0.1 (-0.5,0.2)
Multivariate model
0.1 (-0.3,0.4)
Men
FEVi(mL)
Age-adjusted model
-38.4 (-76.7,0.6)
Multivariate model
¦6.4 (-38.1,25.3)
FVC (mL)
Age-adjusted model
¦17.01-62.0,28.0)
Multivariate model
10.9(-24.7,46.5)
FEVi/FVC (%)
Age-adjusted model
¦0.05 (-0.9,0.0)
Multivariate model
¦0.3 (-0.7,0.2)
Reference: Kimet al. (2005, 087418)
Period of Study: Mar and Dec 2000
Location: Incheon & Ganghwa, Korea
Outcome: lung function (FEVi, FVC)
Age Groups: middle school students
Study Design: Panel
N: 368 children
Statistical Analyses: Generalized liner
model
Covariates: c
Season: Spring and fall
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PMio
Averaging Time: monthly
Mean (SD):
Incheon
March 64
December 54
Ganghwa
March 64
December 53
Range (Min, Max): NR
PM Increment: NR
OR Estimate [Lower CI, Upper CI):
"The present study showed that the
values of FEVi and FVC were greater in
December than in March for both male
and female students at all academic
years...Because only the level of PMio
was significantly higher for March than
for December in both areas, the authors
suggest that decrements of pulmonary
function in March for both areas are
associated with the increased level of
PMio"
July 2009
E-502
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kim et al. (2004, 087383)
Period of Study: Mar-June (spring) 2001
Sep-Nov (fall) 2001
Location: Alameda County, CA
Outcome: Asthma, bronchitis
Age Groups: Children (in grades 3-5)
Study Design: Cross-sectional
N: 1109 children, 871 (long term resident
children), 462 (long term related females),
403 (long term related males)
Statistical Analyses: 2-stage multiple
logistic regression model
Covariates: respiratory illness before
age of 2, household mold/moisture, pests,
maternal history of asthma (for asthma)
Season: Spring and fall
Dose-response Investigated? Yes
Statistical Package: SAS 8.2
Pollutant: PMio
Averaging Time: 9 weeks
Mean (SD): Study Avg 30
Monitoring Stations: 10
Copollutant (correlation): r2 is
approximately 0.9 for all copollutants-
Black Carbon (BC), PM2.6, NOx, NO2, NO
(NOx-NO?)
PM Increment: 1.4 (IQR)
OR Estimate [Lower CI, Upper CI]:
Bronchitis
All subjects: 1.03 [0.99,1.07]
LTR subjects: 1.02(0.98,1.07]
LTR females: 1.04 [1.01,1.09]
LTR males: 1.01 [0.95,1.06]
Asthma
All subjects: 1.02 [0.96,1.09]
LTR subjects: 1.04(0.97,1.12]
LTR females: 1.09 [0.92,1.29]
LTR males: 1.02(0.94,1.10]
Asthma excluding outlier school having ;
larger proportion of Hispanics
All subjects: 1.06[0.97,1.16]
LTR subjects: 1.08 [0.98,1.19]
LTR females: 1.09 [0.96,1.24]
LTR males: 1.08[0.97,1.19]
Reference: Kumar et al. (2004, 089873) Outcome: Chronic respiratory symptoms Pollutant: PM10
81 Spirometric ventilatory defect
Period of Study: 1999-2001
Location: Mandi Gobindgarh and
Morinda, Punjab State, northern India
Age Groups: > 15 yrs
Study Design: Cross-sectional
N: 3603 individuals
Statistical Analyses: Logistic
regression
Covariates: Age, gender, migration, SES,
smoking, type of cooking fuel use
Dose-response Investigated? No
Mean (SD): Study town 112.8 (17.9)
Reference town 75.8 (2.9)
PM10 Increment:
Low vs. High
OR (Lower CI, Upper CI)
p-value
Chronic respiratory symptoms
Low 1.00 (ref)
High 1.5(1.2,1.8)
<0.001
Spirometric ventilatory defect
Low 1.00 (ref)
High 2.4 (2.0-2.9)
<0.001
July 2009
E-503
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Leonardi et al. (2000,
010272)
Period of Study: 1996
Location: 17 cities of Central Europe
(Bulgaria, Czech Republic, Hungary ,
Poland, Romania, Slovakia)
Outcome: Immune biomarkers
Age Groups: 9-11
Study Design: Cross-sectional
N: 366 school children
Statistical Analyses: Linear regression
Covariates: Age, gender, parental
smoking, laboratory of analysis, recent
respiratory illness
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PMio
Averaging Time: annual PMio
Mean (SD): PMio: 65 (14)
Range (Min, Max):
PMio: (41, 96)
5th, median, & 95th percentile
PMio: 41,63, 90
% Change (Lower CI, Upper CI)
p-value
PMio
Neutrophils -5 (-33, 36)
>.20
Total lymphocytes 20 (-6, 54); .150
B lymphocytes 42 (-3,107); .067
Total T lymphocytes 30 (-2, 73); .072
CD4+ 28 (-10, 82);.177
CD8+ 29 (-5, 75); .097
CD4/CD8 7 (-20,43)
>.20
NK 33 (-10, 97);. 157
Total IgG 11 (-10, 38)
>.20
Total IgM 5 (-21, 39)
>.20
Total IgA 11 (-16,46)
>.20
Total IgE-8 (-62,123)
>.20
Reference: Lichtenfels et al, (2007,
097041)
Period of Study: 2001-2003
Location: Sao Paulo, Brazil
Outcome: Secondary sex ratio
Study Design: Retrospective Cohort
Covariates: NR
Statistical Analysis: Correlation
Coefficient
Age Groups: Infants
Pollutant: PMio
Averaging Time: Annual
Mean (SD) Unit:
2001: 49.8(10.5)/yglm3
2002: 48.5 (11.4) //g/m3
2003: 49.4 (14.4) //g/m3
Range (Min, Max): 31.71 -60.96 /yglm3
Copollutant (correlation): NR
Increment: NR
Correlation Coefficient:
R2 - 0.7642, P - 0.13
July 2009
E-504
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lubinski, et al. (2005,
087563)
Period of Study: 1993-1997
Location: Poland
Outcome: Pulmonary function
TLC: total lung capacity
ITGV: interthoracic gas volume
ITGV%TLC: ITGV percent total lung
capacity
Raw: airway resistance
FVC: forced vital capacity
FEVi: forced expiratory volume, 1 second
FEVi%FVC: FEVi percent forced vital
capacity
PEF: peak expiratory flow
FEFbo: forced expiratory flow
Age Groups: 18-23 males, healthy
Study Design: ecological cross-sectional
study
N: 1278 subjects
Statistical Analyses: Multiple linear
regression, ANOVA
Covariates: report unclear on whether or
not there was covariate control, but may
include NO2 and SO2
Dose-response Investigated? No
Pollutant: PM10
Averaging Time: 12 mo
Mean (SD):
A: Highest Pollution Region
Katowice 67-125
Krakow 41-49
B: Moderate Pollution Region
Bielsko-Biala 29-48
Opole 18-45
Lodz 23-38
Warsaw 35-45
Wroclaw 28-76
Zagan 5-35
C: Lowest Pollution Region
Gizycko 5-18
Hel 12-18
Ostroda 23-33
Swinoujscie 7-16
Ustka 12-26
Copollutant: NO2, SO2
PM Increment: 1 //gfm3
Slope, multiple regression
TLC
FEVi
PM10: -0.05
PM10: 0.031
+SO2: 0.03
+SO2: -0.08
+NO2: -0.06
+NO2: -0.12
ITGV
FEVi%FVC
PM10: 0.01
PM10: 0.00
+SO2: -0.07
+SO2: -0.14
+NO2: -0.07
+NO2:-0.048
ITGV%TLC
PEF
PM10: -0.06
PM10: -0.18
+SO2: 0.08
+SO2: 0.056
+NO2: 0.00
+NO2: -0.09
Raw
FEF50
PM10: 0.075
PM10: 0.031
+SO2: -0.08
+SO2: -0.11
+NO2: 0.127
+NO2: -0.04
FVC

PM10: 0.045

+SO2: 0.045

+NO2: -0.14

Reference: McConnell et al. (1999,
007028)
Period of Study: 1993
Location: Southern California
Outcome: Bronchitis, chronic cough,
phlegm
Age Groups: Children: 4th, 7th, 8110th
Study Design: Cross-sectional
N: 3676 people
Statistical Analyses: Logistic
regression
Covariates: Age, sex, race, grade, health
insurance
Dose-response Investigated?
Yes
Pollutant: PM10
Averaging Time: yearly avg 24 h PM10
Mean (SD): 34.8
Range (Min, Max): 13.0, 70.7
Copollutant (correlation): NO2
r - 0.74
0s
r - 0.32
Acid
r - 0.54
PM2.6
r - 0.90
NO2
r - 0.83
0s
r - 0.50
Acid
r - 0.71
PM10 Increment: 19/yg/m3
Children wf asthma
Bronchitis: 1.4 (1.1,1.8)
Phlegm: 2.1 (1.4,3.3)
Cough: 1.1 (0.8,1.7)
Children wf wheeze, no asthma
Bronchitis: 0.9 (0.7,1.3)
Phlegm: 0.9 (0.6,1.4)
Cough: 1.2 (0.9,1.8)
Children wf no wheeze, no asthma
Bronchitis: 0.7 (0.4,1.0)
Phlegm: 0.8 (0.6,1.3)
Cough: 0.9 (0.7,1.2)
July 2009
E-505
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: McConnell et al. (2003,
049490)
Period of Study: 1993-99
Location: 12 Southern CA communities
Outcome: bronchitis symptoms
Age Groups: 9-19
Study Design: communities selected on
basis of historic levels of criteria
pollutants and low residential mobility.
N: 475 children
Statistical Analyses: 3 stage regression
combined to give a logistic mixed effects
model
Covariates: sex, ethnicity, allergies
history, asthma history, SES, insurance
status, current wheeze, current exposure
to ETS, personal smoking status,
participation in team sports, in utero
tobacco exposure through maternal
smoking, family history of asthma,
amount of time routinely spent outside by
child during 2-6 pm.
Dose-response Investigated? No
Statistical Package: SAS Glimmix
macro
Pollutant: PMio
Averaging Time: 4 year averages
Mean (SD): .30.8(13.4)//g/m3
Range (Min, Max): 15.7-63.5
PM Component: particulate organic
carbon and elemental carbon
Copollutant (correlation): PM2 e: r -
0.79
PM 10-2.5: r - 0.79
Inorganic acid: r - 0.72
Organic Acid: r - 0.59
Elemental carbon: r - 0.71
Organic Carbon: r - 0.70
NO?: r - 0.20
O3: r - 0.64
PM Increment:
Between community range 47.8 /yglm3
Between community unit 1 /yglm3
Within community 1 /yglm3
OR Estimate [Lower CI, Upper CI]
Between community per range 1.72(0.93-
3.20) |
Between Community per unit 1.01 (1.00-
1.02)|
Within community per unit 1.04(0.99-
1.10)
Reference: McConnell et al. (2002,
023150)
Period of Study: 1993-1998
Location: 12 communities in Southern
California (grouped into either high and
low pollution communities)
Outcome: Asthma (new diagnosis)
Age Groups: 9-12 yrs, 12-13 yrs, 15-16
yrs
Study Design: Cohort
N:3535
Statistical Analyses: Multivariate
proportion hazard model
Covariates: Sex, age, ethnic origin, BMI,
child history of allergies and asthma
history, SES, maternal smoking, time
spent outside, history of wheezing,
ownership of insurance (yes/no), number
and type of sports played
Dose-response Investigated? Yes
Statistical Package: SAS 8.1
Pollutant: PM10
Averaging Time: 4 yrs
Mean (SD): Low pollution communities:
21.6(3.8)
High pollution communities: 43.3
(12.0)
Percentiles: Low pollution communities:
50th(Median): 20.8
High pollution communities:
50th(Median): 43.3
Range (Min, Max): Low pollution
communities: 16.62, 27.3
High pollution communities: 33.5, 66.9
Monitoring Stations: 12
Copollutant (correlation): PM2 e:
r - 0.96
NO?: r - 0.65
0s
RR Estimate [Lower CI, Upper CI]
lag:
Low PM communities: 1.0 [ref] 0 sport
1.5 [1.0,2.211 sport
1.2 [0.7,1.9] 2 sports
1.7 [0.9, 3.2] a 3 sports
High PM communities: 1.0 [ref] 0 sport
1.1 [0.7, 1.7] 1 sport
0.9[0.5,1.7] 2 sports
2.0 [1.1, 3.6] £3 sports
High vs Low PM10 communities: 0.8
(0.6, 1.0)
lncidence-l\l (incidence) number of
sports: Low PM communities: 49 (0.023)
0
54 (0.032) 1
22 (0.024) 2
13 (0.033) >3
High PM communities: 55 (0.021) 0
36 (0.021)1
14(0.018)2
16 (0.033) >3
July 2009
E-506
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Reference: McConnell, et al. (2006,
180226)
Period of Study: 1996-1999
Location: 12 Southern California
communities
	Design & Methods	
Outcome: Prevalence of bronchitic
symptoms (yearly).
Age Groups: 10-15-years-old
Study Design: longitudinal cohort
N: 475 asthmatic children
Statistical Analyses: Multilevel logistic
mixed effects models.
Covariates: age, second hand smoke
personal smoking history
sex, race.
Dose-response Investigated? No
Statistical Package: SAS with
GLIMMIX macro
Concentrations'
Pollutant: PMio
Averaging Time: 365 days
Percentiles: Community by year
(n - 48 - 12 communities ~ 4 years)
25th: NR
50th(Median): 3.4
75th: NR
Range (Min, Max):
Community by year (n - 48 - 12
communities ~ 4 years):
(0.89,8.7)
Monitoring Stations: 12
Copollutant: 0:i, NO2, EC, 0C
Acid vapor (acetic and formic acid)
Effect Estimates (95% CI)
PM Increment: 6.1 |Jg|m3
OR Estimate [Lower CI, Upper CI]
PM10
Dog (n - 292): 1.60 [1.12: 2.30]
No dog (n - 183): 0.89(0.57: 1.39]
PMio*Dog interaction p-value: 0.02
Cat (n - 202): 1.47 [0.96: 2.24]
No Cat (n - 273): 1.20(0.83: 1.73]
PMio*Cat interaction p-value: 0.41
Neither pet (n - 112): 0.91 (0.53: 1.56]
Cat only (n - 71): 0.84(0.42: 1.66]
Dog only (n - 161): 1.41 (0.91: 2.19]
Both pets (n - 131): 1.89(1.15:3.10]
Results suggest that dog ownership, a
source of residential exposure to
endotoxin, may worsen the severity of
respiratory symptoms from exposure to
air pollutants in asthmatic children.
Reference: Meng et al. (2007, 093275)
Period of Study: November 2000 and
September 2001 (collection of health
data)
Location: Los Angeles and San
counties
Outcome: Poorly controlled asthma vs.
controlled asthma
Age Groups: 18-64, 65 +
Study Design: Long-term exposure study
comparison of cases and controls
N: 1,609 adults (represented individuals
age 18 + who reported ever having been
diagnosed as having asthma by a
physician and had their address
successfully geocoded)
Statistical Analyses: Logistic
regression to evaluate associations
between TD (traffic density) and annual
avg air pollution concentrations and
poorly controlled asthma. Used sample
weights that adjusted for unequal
probabilities of selection into the CHIS
sample.
Covariates: Age, sex, race/ethnicity,
family federal poverty level, county,
insurance status, delay in care for
asthma, taking medications, smoking
behavior, self-reported health status,
employment, physical activity
Dose-response Investigated? yes
Pollutant: PM10
Averaging Time: 24 over 1 year
Copollutant (correlation):
Oa: r - -0.72
NO?: r - 0.83
PM2.6: r - 0.84
CO: r - 0.42
TD: r - 0.14
PM Increment: Continuous data: per
10/yg/m3
OR Estimate [Lower CI, Upper CI]
All Adults: 1.08[0.82,1.43]
Non-Elederly Adults: 1.14(0.84,1.55]
Elderly: 0.84 [0.41,1.73]
Women: 1.38 [0.99,1.94]
July 2009
E-507
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Millstein et al. (2004,
088629)
Period of Study: Mar-Aug, 1995, and
Sep, 1995 to Feb, 1996
Data were taken from the Children's
Health Study
Location: Alpine, Atascadero, Lake
Arrowhead, Lake Elsinore, Lancaster,
Lompoc, Long Beach, Mira Loma,
Riverside, San Dimas, Santa Maria, and
Upland, CA
Outcome: Wheezing & asthma
medication use (ICD9 NR)
Age Groups: 4th grade students, mostly
9 yrs at the time of the study
Study Design: Cohort Study, stratified
into 2 seasonal groups/
N: 2081 enrolled, 2034 provided parent-
completed questionnaire.
Statistical Analyses: Multilevel, mixed-
effects logistic model.
Covariates: Contagious respiratory
disease, ambient airborne pollen and
other allergens, temperature, sex, age
race, allergies, pet cats, carpet in home,
environmental tobacco smoke, heating
fuel, heating system, water damage in
home, education level of questionnaire
signer, physician diagnosed asthma.
Season: Mar-Aug, 1995, and Sep, 1995
to Feb, 1996
Statistical Package: GLIMMIX SAS
8.00 macro for generalized linear mixed
models.
Lags Considered: 14
Pollutant: PMio
Averaging Time: Monthly means for
PMio.
PM Component: Nitric acid, formic acid,
acetic acid
Monitoring Stations:
1 central location in each community
Copollutant (correlation):
Os:r - 0.76
NO?: r - 0.39
PM2.6: r - 0.91
PM Increment: IQR 13.39 //g/m3
Odds Ratio [lower CI, Upper CI]
Annual
PMio: 0.93(0.67,1.27]
March-August
PMio: 0.91 (0.46,1.80]
Sep-Feb
PMio: 0.65(0.40,1.06]
Reference: Neuberger et al. (2004,
093249)
Period of Study: 6/1999-6/2000
Location: Austria (Vienna and a rural
area near Linz)
Outcome: Questionnaire derived asthma
score, and a 1-5 point respiratory health
rating by parent
Age Groups: 7-10 years
Study Design: Cross-sectional survey
N: about 2000 children
Statistical Analyses: mixed models
linear regression-used factor analysis to
develop the "asthma score"
Covariates: Pre-existing respiratory
conditions, temperature, rainy days, #
smokers in household, heavy traffic on
residential street, gas stove or heating,
molds, sex, age of child, allergies of child,
asthma in other family members
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 4 week avg (preceding
interview)
Pollutant: PMio
Averaging Time: 24-h
Copollutant (correlation): PM2.6
(r - 0.94) in Vienna
PM Increment: 10 /yglm3
Change in mean associated unit
increase in PM (p-value)
lag
Respiratory Health score
Vienna: 0.005 (p > 0.05)
lag 4 week avg
Rural area: 0.008 (p > 0.05)
lag 4 week avg
Asthma score
Vienna: 0.006 (p > 0.05)
lag 4 week avg
Rural area: -0.001 (p > 0.05)
lag 4 week avg
July 2009
E-508
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Oftedal et al. (2008,
093202)
Period of Study: 2001-2002
Location: Oslo, Norway
Outcome: Lung function (PEF, FEF2b%,
FEFbo%, FEVi, FVC)
Age Groups: 9-10 yrs
Study Design: Cross-sectional
N: 1847 children
Statistical Analyses: Linear regression
Covariates: Height, age, BMI, birth
weight, temperature, maternal smoking,
sex
Dose-response Investigated?
Yes
Statistical Package: SPSS, STATA, S-
Plus
Lags Considered: 1-3
Pollutant: PMio
IQR:
PMio in 1st yr of life: 10.3
PMio lifetime: 5.8
PM Increment: Per IQR
fB (Lower CI, Upper CI)
PMio in 1st yr of life
PEF-72.5 (-122.3 to -22.7)
FEF2B* -77.4 (-133.4 to -21.4)
FEFbo*-53.9 (-102.6. to-5.2)
FEVi-6.7 (-24.1,10.7)
FVC 0.5(-18.5,19.6)
PMio lifetime exposure
PEF-66.4 (-109.5 to -23.3)
FEF2B* -61.5 (-110.0 to -13.1)
FEFbo*-45.6 (-87.7 to-3.5)
FEVi-7.3 (-22.4, 7.7)
FVC-2.1 (-18.6,14.4)
Reference: Parker et al. (2009,192359)
Outcome: Respiratory allergyfhayfever
Pollutant: PMio
Increment: 10//gfm3
Period of Study: 1999-2005
Study Design: Cohort
Averaging Time: NR
Odds Ratio (95% CI)
Location: US
Covariates: survey year, age, family
Median: 24.1 /yglm3
Single Pollutant Model, variable N

structure, usual source of care, health



insurance, family income relative to
IQR: 20.8-28.7
Adjusted: 1.04 (0.99-1.09)

federal poverty level, race/ethnicity
Copollutant (correlation):


Statistical Analysis: logistic regression
Summer O3: 0.26


Statistical Package: SUDAAN
SO2: -0.19


Age Groups: 73,198 children aged 3-17
NO?: 0.48


years



PM2.6: 0.51



PM 10-2.B: 0.86

July 2009
E-509
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Penard-Morand et al. (2005,
087951)
Period of Study: 03/1999 - 10/2000
Mean concentrations of NO2, SO2, PM10,
and O3 were taken from 01/01/1998 to
1213112000
Location: 6 French cities: Bordeaux,
Clermont-Ferrand, Creteil, Marseille,
Strasbourg, Reims.
Outcome:
Flexural dermatitis
Asthma (493)
Rhinoconjunctivitis
Atopic dermatitis
Wheeze
Allergic rhinitis
Atopy
EIB (exercise-induced bronchial reactivity)
Age Groups: 9-11 years
Study Design: Cross-sectional
N: 9615 Children (6672 complete
examination and questionnaire info)
Statistical Analyses: Logistic
regression
Marginal Model (GENM0D)
Covariates: Age, Sex, Family history of
allergy, Passive smoking
Parental education
Season: All
Excluding end of spring and during
summer for clinical examinations
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PM10
Averaging Time: 3 years
Mean (SD): Low concentrations: 26.9
High Concentrations: 23.8
Range (Min, Max):
Low concentrations: 10-20
High concentrations: 21.5-29.5
Copollutant (correlation): NO2: r -.46
SO2: r -.76
O3: r - -.02
Monitoring Stations: 16
PM Increment: 10/yg/m3 (IQR)
OR Estimate [Lower CI, Upper CI]:
EIB (during exam): 1.43 (1.02-2.01)
Flexural dermatitis (during exam): 0.79
(0.59-1.07)
Wheeze (past year): 1.05 (0.72-1.54)
Asthma (past year): 1.23 (0.77-1.95)
Rhinoconjunctivitis (past year): 1.17
(0.86-1.59)
Atopic dermatitis (past year): 1.28 (0.96-
1.71)
Asthma (lifetime): 1.32 (0.96-1.81)
Allergic rhinitis (lifetime): 1.32 (1.04-
1.68)
Atopic dermatitis (lifetime): 1.09 (0.88-
1.36)
Atopy (lifetime): 0.98(0.80-1.22)
Pollen: 1.14(0.85-1.53)
Indoor: 0.91 (0.72-1.15)
Moulds: 1.00(0.53-1.88)
Highest correlated pollutant adjustments:
EIB (during exam): 1.16 (0.72-1.85)
Flexural dermatitis (during exam): 0.93
0(60-1.43)
Wheeze (past year): 1.31 (0.71-2.36)
Asthma (past year): 1.25 (0.66-2.37)
Rhinoconjunctivitis (past year): 1.22
(0.98-1.68)
Atopic dermatitis (past year): 1.63 (1.07-
2.49)
Asthma (lifetime): 1.11 (0.70-1.74)
Allergic rhinitis (lifetime): 1.19 (0.94-
1.59)
Atopic dermatitis (lifetime): 1.47 (1.07-
2.00)
Atopy (lifetime): 0.93(0.69-1.26)
Pollen: 1.30 (0.98-1.57)
Indoor: .83 (0.63-1.12)
Moulds: 1.62(0.64-4.09)
July 2009
E-510
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Peters et al„ (1999, 087237)
Period of Study: 1986-1990,1994
Location: Southern California
Outcome: Asthma, cough, bronchitis,
wheeze
Age Groups: 4th, 7th, & 10th graders
Study Design: cohort
l\l: 3676 children
Statistical Analyses: Stepwise logistic
regression
Covariates: Community, grade, race,
sex, height, BMI, asthma in parents, hay
fever, health insurance, plants in home,
mildew in home, passive smoke exposure,
pest infestation, carpet, vitamin
supplements, active smoking, pets, gas
stove, air conditioner
Dose-response Investigated? Yes
Pollutant: PMio
Averaging Time: 24 h PMio averaged
over 1994
Mean based on data collected during
1986-1990,1994:
Alpine 37.4,21.3
Atascadero 28.0, 20.7
Lake Elsinore 59.5, 34.7
Lake Gregory 38.3, 24.2
Lancaster 47.0, 33.6
Lompoc 30.0,13.0
Long Beach 49.5, 38.8
Mira Loma 84.9, 70.7
Riverside 84.9, 45.2
San Dimas 67.0, 36.7
Santa Maria 28.0, 29.2
Upland 75.6, 49.0
PM Increment: 25/yg/m3
OR (Lower CI, Upper CI) for respiratory
illness
Based on 1986-1990 pollutant levels
Ever asthma 0.93 (0.76,1.13)
Current asthma 1.09 (0.86,1.37)
Bronchitis 0.94 (0.74,1.19)
Cough 1.06(0.93,1.21)
Wheeze 1.05 (0.89,1.25)
Based on 1994 pollutant levels
Ever asthma 0.87 (0.67,1.14)
Current asthma 1.11 (0.81,1.54)
Bronchitis 0.90 (0.65,1.26)
Cough 1.14(0.96,1.35)
Wheeze 1.01 (0.79,1.29)
July 2009
E-511
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Pierse, et al. (2006, 088757) Outcome: Cough without a cold	Pollutant: PMio
Period of Study: 2 years (once in 1998
and once in 2001-surveys)
Location: Leicestershire, UK
Night time cough
Current wheeze
Age Groups: 1-5 years
Study Design: Cross-sectional (cohorts)
N: 4400 children
Statistical Analyses: Binomial
generalized linear models (compared with
likelihood ratio tests)
Spatial variograms (due to the spatial
concerns)
Covariates: Age, Gender
Mother/father has asthma
Coal heating the home, Smoking by
household member in the home, Either
parent continued education past 16 years
of age, Pre term birth, Breast feeding,
Gas cooking, Presence of pets, Number of
cigarettes smoked by mother,
Overcrowding, Single parenthood, Diet
Dose-response Investigated? Yes (Fig.
2 shows evidence of dose-response
effect based on surveys, states in
discussion).
Statistical Package: SAS 8.2
S-Plus 6.1
Averaging Time: annual PMio
Mean (SD): 1998: 1.47
2001: 1.33
Percentiles: 25th: 1998 (.73) and 2001
(.8)
75th: 1998(1.93) and 2001 (1.84)
PM Increment: 10/yg/m3 (IQR)
Unadjusted OR estimates [Lower CI,
Upper CI]:
Cough without cold (1998): 1.22 (1.10 to
1.36)
Cough without cold (2001): 1.46 (1.27 to
1.68)
Night-time cough (1998): 1.11 (1.01 to
1.23)
Night-time cough (2001): 1.25 (1.09 to
1.43)
Current wheeze (1998): 0.99 (0.89 to
1.10)
Current wheeze (2001): 1.09 (0.93 to
1.30)
Adjusted OR Estimate [Lower CI,
Upper CI]:
Cough without cold (1998): 1.21 (1.07 to
1.38)
Cough without cold (2001): 1.56 (1.32 to
1.84)
Night-time cough (1998): 1.06 (0.94 to
1.19)
Night-time cough (2001): 1.25 (1.06 to
1.47)
Current wheeze (1998): 0.99 (0.88 to
1.12)
Current wheeze (2001): 1.28 (1.04 to
1.58)
When the child was originally
asymptomatic in 1998:
Unadjusted OR estimates [Lower CI,
Upper CI]:
Cough without cold (2001): 1.68 (1.39 to
2.03)
Night-time cough (2001): 1.21 (1.00 to
1.46)
Current wheeze (2001): 1.22 (0.92 to
1.62)
Adjusted OR Estimate [Lower CI, Upper
CI]:
Cough without cold (2001): 1.62 (1.31 to
2.00)
Night-time cough (2001): 1.19 (0.96 to
1.47)
Current wheeze (2001): 1.42 (1.02 to
1.97)
July 2009
E-512
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Qian et al. (2005, 093283)
Period of Study: 1990-1992
Location: Forsythe, NC
Minneapolis, MN
Jackson, MS.
Outcome: FVC, FEVi, FEVilFVC
Age Groups: middle aged (avg 56.8
years)
Study Design: cross-sectional
N: 10,240 people
Statistical Analyses: regression
equations, multiple linear regression
analyses
Covariates: Smoking status, recent use
of respiratory medication, current
respiratory symptoms, chronic lung
diseases, field center
Dose-response Investigated? No
Statistical Package: SAS software,
version 9.1
Pollutant: PMio
Averaging Time: Annual
Mean (SD): 27.9 (2.8)
Percentiles: 25th: 25.8
50th(Median): 27.5
75th: 30.2
Range (Maximum-Minimum): 12.2
Monitoring Stations: 3 (Minneapolis,
MN)
5 (Jackson, MS)
and 9 (Forsythe, NC)
Copollutant: 0:i
PM Increment: 2.8 //g|m3 (1 SD)
Effect Estimate:
In Never Smokers
FVC B - -0.0108, SE - 0.0026,
p -.0001
FEVi B - -0.0082, SE - 0.0029,
p - .0047
FEVilFVC B - -0.0024, SE - 0.0023,
p -.2787
Smoking status
Current n - 2377, FVC - -1.96, FEVi - ¦
2.23, FEVilFVC - -0.94
Former n - 3858, FVC - -1.25, FEVi - ¦
1.10, FEVilFVC - -0.30
Never n - 4005, FVC - -1.12, FEVi - ¦
0.63, FEVilFVC - 0.06
Recent Use of Respiratory Medication
Yes n - 424, FVC - -2.65, FEVi - ¦
3.89, FEVilFVC - -3.00
Non - 9816, FVC --1.41, FEVi -¦
1.20, FEVilFVC - -0.24
Current Respiratory Symptoms
Yes n- 4340, FVC --1.68, FEVi -¦
1.70, FEVilFVC - -0.63
Non - 5900, FVC --1.05, FEVi -¦
0.63, FEVilFVC - 0.05
Chronic Lung Diseases
Yes n - 1374, FVC --1.95, FEVi -¦
2.31, FEVilFVC - -1.18
Non - 8866, FVC --1.35, FEVi -¦
1.10, FEVilFVC - -0.19
Field Center
Forsythe, NC n - 3504, FVC - -0.03,
FEVi - 0.05, FEVilFVC - -0.33
Minneapolis, MN n - 3793, FVC - 0.50,
FEVi - 0.54, FEVilFVC - -0.30
Jackson, MS n - 2943, FVC - -0.01,
FEVi - 0.17, FEVilFVC - -0.32
July 2009
E-513
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Rios et al. (2004, 087800)
Period of Study: 1998-2000
Location: the metropolitan area of Rio de
Janiero, Brazil, Duque de Caxias (DC) and
Seropedica (SR)
Outcome: wheezing, asthma, cough at
night
Age Groups: 13-14 yrs
Study Design: cohort
N: 4064 students
Statistical Analyses: chi-squared
Covariates: sex, type of school, time of
residence, domestic smoking, residents
per home
Dose-response Investigated? Yes
Statistical Package: Epilnfo
Pollutant: PMio
Averaging Time: weekly measurements
used to create annual PM estimate
Mean (SD): DC
1998: 147
1999: 115
2000: 110
Total: 124
SR
1998: 37
1999: 31
2000: 37
Total: 35
Monitoring Stations: NR
PM Increment: High vs. Low
Global Cut-Off Score %, p-val:
DC
Male: 15.0
Female: 22.3, p< ,05t
Private School: 16.6
Public School: 19.4, p < .05*
<	5yr residence: 20.9
>5yr residence: 16.8
No domestic smoking exposure: 17.6
Domestic smoking exposure: 20.4,
p< .051"
<	5 residents per home: 18.4
5+ residents per home: 19.5
SR
Male: 12.3
Female: 19.7, p< ,05t
Private School: 28.3, p < .05*1"
Public School: 14.7
<	5yr residence: 10.8
>5yr residence: 16.5
No domestic smoking exposure: 14.8
Domestic smoking exposure: 18.3
<	5 residents per home: 15.6
5+ residents per home: 17.4
Notes: The Global Cut-off Score
encompasses replies to the asthma
component of ISAAC'S written
questionnaire that establishes a cut-off
from which is defined the presence of
asthma for the Brazilian population.
"comparing the cities in the same
controlled variable
tcomparing the controlled variable in the
same city
Reference: Rojas-Martinez et al. (2007,
091064)
Period of Study: 1996-1999
Location: Mexico City, Mexico
Outcome: Lung function: FEVi, FVC,
FEF25-75%
Age Groups: Children 8 years old at time
of cohort recruitment
Study Design: school-based "dynamic"
cohort study
N: 3170 children
14,545 observations
Statistical Analyses: Three-level
generalized linear mixed models with
unstructured variance-covariance matrix
Covariates: age, body mass index,
height, height by age, weekday spent
outdoors, environmental tobacco smoke,
previous-day mean air pollutant
concentration, time since first test
Dose-response Investigated? No
Pollutant: PMio
Averaging Time: 6-mo
Mean (SD): 6-mo averaging
SD: NR
Mean: 75.6
Percentiles: 6-mo averaging
25th: 55.8
50th(Median): 67.5
75th: 92.2
Monitoring Stations: 5 sites for PMio,
10 for other pollutants
Copollutant:
Os
NO?
PM Increment: IQR 6-LC: 36.4
Slope [Lower CI, Upper CI]
Girls
One-pollutant model
FVC: -39 [-47:-31]
FEV: -29 [-36: -21]
FEF2e-7e«:-17 [-36: 1]
FEVi/FVC: 0.12(0.07: 0.17]
Two-pollutant model: PMio, 6-LC & O3
FVC: -30 [-39:-22]
FEV:-24 [-31:-16]
FEF2b-7b%: -9 [-26: 9]
FEVi/FVC: 0.10(0.06: 0.15]
July 2009
E-514
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Statistical Package: SA
PMio, 6-LC & NO?
FVC: -21 [-30:-13]
FEV: -17 [-25: -8]
FEF2B-7B*: -23 [-43: -4]
FEVi/FVC: 0.07 [0.02: 0.13]
Multipollutant model: PMio, 6-LC, O3, &
NO?
FVC:-14[-23:-5]
FEV:-11 [-20: -3]
FEF2B.7B*: -7 [-27: 12]
FEVi/FVC: 0.08 [0.03: 0.13]
Boys
One-pollutant model
FVC:-33 [-41:-25]
FEV: -27 [-34: -19]
FEF2b-7b%:-18 [-34: -2]
FEVi/FVC: 0.04 [-0.01:0.09]
Two-pollutant model: PMio, 6-LC & O3
FVC: -28 [-36:-19]
FEV: -22 [-30:-15]
FEF2b-7b%:-10 [-27: 7]
FEVi/FVC: 0.04 [-0.01:0.09]
FEVi/FVC: 0.24 [0.13: 0.34]
PMio, 6-LC & NO?
FVC: -16 [-26:-7]
FEV: -19 [-27:-10]
FEF2b-7b%: -26 [-44: -9]
FEVi/FVC: 0.005 [-0.06: 0.05]
Multipollutant model PMio, 6-LC, O3, &
NO2
FVC: -12 [-22: -3]
FEV: -15 [-23: -6]
FEF2b-7b%:-12 [-30: 6]
FEVi/FVC:-0.002 [-0.06: 0.05]
July 2009
E-515
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Schikowski et al. (2005,
088637)
Period of Study: 1985-1994
Location: Rhine-Ruhr Basin of Germany
[Dortmund (1985,1990), Duisburg
(1990), Gelsenkirchen (1986,1990), and
Heme (1986)]
Outcome: Respiratory symptoms &
pulmonary function
Age Groups: age 54-55
Study Design: Cross-sectional
N:4757 women
Statistical Analyses: Linear & Logistic
regressions, including random effects
model
Covariates: age, smoking, SES,
occupational exposure, form of heating,
BMI, height
Season: NR
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: NR
Pollutant: PMio
Averaging Time: NR
Min, P25, Median, Mean, P75, Max
Annual Mean
35,40, 43, 44, 47, 53
Five year Mean
39,43, 47, 48, 53, 56
Monitoring Stations: 7
Copollutant (correlation): NR
PM Increment: 7 /yglm3
OR (Lower CI, Upper CI) for asthma
symptoms
Annual means
Chronic bronchitis 1.00 (0.85,1.18)
Chronic cough 1.03 (0.87,1.23)
Frequent cough 1.01 (0.93,1.10)
C0PD 1.37 (0.98,1.92)
p< 0.1
FEVi 0.953 (0.916, 0.989)
p< 0.1
FVC 0.966 (0.940, 0.992)
p< 0.1
FEVi/FVC 0.989 (0.978,1.000)
p< 0.1
Five year means
Chronic bronchitis 1.13 (0.95,1.34)
Chronic cough 1.11 (0.93,1.31)
Frequent cough 1.05 (0.94,1.17)
C0PD 1.33(1.03,1.72)
p< 0.1
FEVi 0.949 (0.923, 0.975)
p < 0.05
FVC 0.963 (0.945, 0.982)
p < 0.05
FEVi/FVC 0.989 (0.980, 0.997)
p< 0.1
Reference: Schindler et al, (2009,
191950)
Period of Study: 1991-2002
Location: Switzerland
Outcome: Respiratory Symptoms
Study Design: Prospective Cohort
Statistical Analysis: Logistic
Regression Model
Age Groups: Adults, 18-60 years of age
at start of study
Covariates: sex, age, level of education,
Swiss citizenship, BMI, parental smoking,
parental history of asthmafatopy, early
respiratory infection, smoking status,
pack years, daily number of cigarettes,
years since smoking cessation, passive
smoking in general/at work, occupational
exposure to airbourne irritants
Pollutant: PMio
Averaging Time: Annual
Mean (SD) Unit:
Range (Min, Max):
Copollutant (correlation): NR
Increment: 10//gfm3
Odds Ratio (95%CI) of reporting
symptoms at second interview
Entire Sample, New Reports
Regular Cough: 0.77 (0.62-0.97)
Regular Phlegm: 0.74 (0.56-0.99)
Chronic Cough or Phlegm: 0.78 (0.62-
0.98)
Wheezing: 1.01 (0.74-1.39)
Wheezing with Dyspnea: 0.70 (0.49-1.01)
Wheezing without Cold: 1.06 (0.76-1.50)
Entire Sample, Persistent Reports
Regular Cough: 0.55 (0.39-0.78)
Regular Phlegm: 0.82 (0.52-1.33)
Chronic Cough or Phlegm: 0.67 (0.40-
1.15)
Wheezing: 0.50 (0.32-0.80)
Wheezing with Dyspnea: 0.59 (0.30-1.23)
Wheezing without Cold: 0.61- (0.35-1.12)
Persistent Non-Smokers, New Reports
July 2009
E-516
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Regular Cough: 0.86 (0.63-1.19)
Regular Phlegm: 0.70 (0.49-0.99)
Chronic Cough or Phlegm: 0.71 (0.52-
0.99)
Wheezing: 0.93(0.60-1.46)
Wheezing with Dyspnea: 0.77 (0.50-1.20)
Wheezing without Cold: 1.11 (0.66-1.92)
Persistent Non-Smokers, Persistent
Reports
Regular Cough: 0.28 (0.14-0.60)
Regular Phlegm: 0.87 (0.43-1.84)
Chronic Cough or Phlegm: 0.35 (0.16-
0.81)
Wheezing: 0.53(0.28-1.08)
Wheezing with Dyspnea: 0.76 (0.30-
2012)
Wheezing without Cold: 0.61 (0.26-1.52)
Gender-specific odds ratio (95%CI) of
reporting symptoms at second
interview
New Reports
Regular Cough, p - 0.73
Men: 0.75(0.53-1.06)
Women: 0.81 (0.58-1.15)
Regular Phlegm, p - 0.41
Men: 0.85(0.60-1.20)
Women: 0.68 (0.46-1.00)
Chronic Cough or Phlegm: 0.36
Men: 0.87(0.63-1.21)
Women: 0.71 (0.51-0.97)
Wheezing, p - 0.20
Men: 0.83(0.57-1.20)
Women: 1.20 (0.78-1.87)
Wheezing with Dyspnea, p - 0.11
Men: 0.56(0.36-0.87)
Women: 1.00.57-1.842
Wheezing without Cold, p - 0.43
Men: 0.95(0.63-1.42)
Women: 1.25 (0.72-2.17)
Persistent Reports
Regular Cough, p - 0.02
Men: 0.75(0.48-1.18)
Women: 0.31 (0.17-0.56)
Regular Phlegm, p - 0.33
Men: 0.65(0.37-1.12)
Women: 1.04 (0.47-2.34)
Chronic Cough or Phlegm: 0.47
July 2009
E-517
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Men: 0.68(0.39-1.20)
Women: 0.47 (0.20-1.11)
Wheezing, p - 0.29
Men: 0.34 (0.17-0.72)
Women: 0.57 (0.32-1.01)
Wheezing with Dyspnea, p - 0.63
Men: 0.56(0.16-1.95)
Women: 0.37 (0.13-1.05)
Wheezing without Cold, p - 0.57
Men: 0.34 (0.12-0.91)
Women: 0.49 (0.21-1.15)
Odds Ratio (95%CI) of reporting
symptoms at second interview with
additional adjustment for annual
outdoor PM exposure at baseline
Entire Sample
Regular Cough, p - 0.0003
New Reports: 0.77 (0.61-0.97)
Persistent Reports: 0.55 (0.39-0.78)
Regular Phlegm, p - 0.13
New Reports: 0.77 (0.59-1.02)
Persistent Reports: 0.79 (0.46-1.33)
Chronic Cough or Phlegm, p - 0.02
New Reports: 0.78 (0.62-0.98)
Persistent Reports: 0.64 (0.40-1.02)
Wheezing, p - 0.002
New Reports: 0.91 (0.63-1.33)
Persistent Reports: 0.47 (0.31-0.72)
Wheezing with Dyspnea, p - 0.03
New Reports: 0.65 (0.43-0.98)
Persistent Reports: 0.55 (0.28-1.10)
Severe Wheezing, p - 0.28
New Reports: 0.96 (0.66-1.40)
Persistent Reports: 0.62 (0.34-1.12)
Non-Smokers
Regular Cough, p < 0.001
New Reports: 0.87 (0.63-1.19)
Persistent Reports: 0.29 (0.16-0.52)
Regular Phlegm, p - 0.07
New Reports: 0.70 (0.50-0.99)
Persistent Reports: 0.67 (0.34-1.33)
Chronic Cough or Phlegm, p - 0.008
New Reports: 0.72 (0.52-0.99)
Persistent Reports: 0.38 (0.17-0.84)
Wheezing, p - 0.07
New Reports: 0.87 (0.52-1.48)
July 2009
E-518
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Persistent Reports: 0.48 (0.25-0.91)
Wheezing with Dyspnea, p - 0.36
New Reports: 0.76 (0.48-1.19)
Persistent Reports: 0.70 (0.27-1.82)
Severe Wheezing, p - 0.57
New Reports: 1.11 (0.64-1.93)
Persistent Reports: 0.64 (0.26-1.54)
Reference: Sharma et al. (2004,
156974)
Period of Study: 1112002-412003
Location: 3 sections in Kanpur City,
India: 1) Indian Institute of Technology
Kanpur (IITK)
2)	Vikas Nagar (VN)
3)	Juhilal Colony (JC)
Outcome: Lung function
Age Groups: 20-55 years
Study Design: Cohort
N: 91 people
Statistical Analyses: Linear regression
Covariates: NR
Season: Fall, Winter, spring
Dose-response Investigated? No
Statistical Package: Microsoft Excel
Lags Considered: 1 d lag & 5d mov avg
Pollutant: PMio
Averaging Time: 24-h
Mean (SD): IITK 184 (40)
VN 295 (58)
JC 293 (90)
PM Component: Lead, Nickel, Cadmium,
Chromium, Iron, Zinc
Benzene soluble fraction (includes
polycyclic aromatic hydrocarbons [PAHs])
Copollutant (correlation): nPEF - mean
daily deviations in PEF
PMio-nPEF: (-0.52)
PMio-PM2.b: (0.67)
PMio-PMio (1-day lag): (0.45)
PMio-PM2.b (1-day lag): (0.46)
PM Increment: 1 /yglm3
~PEF (difference or change in peak
expiratory flow)
¦0.0318 L/min
Reference: Tager et al. (2005, 087538)
Period of Study: 412000- 612000,
212001-612001,
212002-612002
Location:
Los Angeles, California
San Francisco, California
Outcome: Lung Function (FEVi, FVC,
PEFR, FEF75, FEF25-75, FEF2b jb/FVC ratio)
Age Groups: 16-21 + yfo
College Freshman
Study Design: Retrospective cohort
N: 255 students
108 Men (M)
147 Women (W)
Statistical Analyses: Multivariate
Linear Regression
Covariates: Sex, height, weight, area of
residence, age, race, ETS exposure,
respiratory disease history
Dose-response Investigated? No
Pollutant: PM10
Averaging Time: Cumulative lifetime
exposure
Median: Prior to 1987: M: 73
W: 71
1987 and later: M: 36
W: 34
Lifetime: M: 48
W: 45
Range (Min, Max): Prior to 1987: M: 34,
117
W:31,124
1987 and later: M:18, 68
W: 20, 61
Lifetime: M: 21, 80
W: 18,71
Monitoring Stations: Between 1 and 3
Copollutant (correlation): O3 prior to
1987: r - 0.68
O3 1987 and later: r - 0.81
O3—Lifetime: r - 0.57
PM Increment: 1 /yglm3
Parameter Estimates (SD)
(Lifetime PM10, Interaction PM10
FEF2b-7b|FVC)
LnFEF75:
M:-0.009 (0.0009), 0.009(0.007)
W:-0.010(0.0007), 0.008(0.0005)
July 2009
E-519
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Tamura et a. (2003, 087445) Outcome: non specific respiratory
disease (Chronic bronchitis, acute
Period of Study: 1998-1999
Location: Bangkok, Thailand
Pollutant: PMio
bronchitis, bronchial asthma, dyspnea and Averaging Time: 24 h
Mean (SD):
wheezing)
Age Groups: adults
Study Design: Cross-sectional
N: 1603 policemen
Statistical Analyses: Multiple logistic
regression
Covariates: age, smoking status
Dose-response Investigated? Yes
Statistical Package: SPSS
Heavily Polluted 80-190
Moderately Polluted 60-69
Control 59
Monitoring Stations: 13
PM Increment: Heavily Polluted vs.
Moderately Polluted vs. Control
Number and Prevalence (%) of respiratory
disease among heavily polluted,
moderately polluted, and control areas.
Heavily Polluted
Chronic bronchitis 16 (3.0)
Acute bronchitis 19 (3.5)
Bronchial asthma 5 (0.9)
Dyspnea & wheezing 49 (9.2)
Any 1 of above 69 (13.0)
Persistent cough 11 (2.1)
Persistent phlegm 27 (1.3)
Cough & phlegm 6 (1.1)
Moderately Polluted
Chronic bronchitis 8 (2.4)
Acute bronchitis 12 (9.0)
Bronchial asthma 2 (0.6)
Dyspnea & wheezing 23 (6.8)
Any 1 of above 37 (10.9)
Persistent cough 1 (0.3)
Persistent phlegm 11 (3.3) |
Cough & phlegm 1 (0.3)
Control
Chronic bronchitis 6 (1.9)
Acute bronchitis 11 (3.3)
Bronchial asthma 0 (0.0)
Dyspnea & wheezing 23 (7.2)
Any 1 of above 31 (9.4)
Persistent cough 1 (0.3)
Persistent phlegm 8 (2.4)
Cough & phlegm 1 (0.3)
Reference: Wheeler and Ben-Schlomo
(2005, 089860)
Period of Study: 1995-1997
Location: England
Outcome: FEVi
Age Groups: 16-79 yrs
Study Design: Data from Health Survey
for England were coupled geographically
with air pollution measurements on a 1
km grid.
N: 26,426 households with 39,251
adults
Statistical Analyses: Logistic
regression, least squares regression
Covariates: Age, sex, height, body mass
index, smoking status, household passive
smoke exposure, inhaler use in the
previous 24-hs, doctor diagnosis of
asthma.
Pollutant: PMio
Averaging Time: 1996 annual mean
Mean (SD): 23.95 (3.58)
Range (Min, Max): 17.87-43.37
fB (95%CI) for Height-age standardized
FEVi by ambient air quality index
p-value
Male
Good (ref)
Poor-0.023 (-0.030 to-0.016)
<0.001
Female
Good (ref)
Poor-0.019 (-0.026 to -0.013)
<0.001
Dose-response Investigated? No
July 2009
E-520
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Zhang et al„ (2002, 034814)
Period of Study: 1993-1996
Location: 4 Chinese cities (urban and
suburban location in each city):
Guangzhou, Wuhan, Lanzhou, Chongqing
Outcome: Interview-self reports of
symptoms: Wheeze (ever wheezy when
having a cold)
Asthma (diagnosis by doctor)
Bronchitis (diagnosis by doctor),
Hospitalization due to respiratory disease
(ever)
Persistent cough (coughed for at least 1
month per year with or apart from colds)
Persistent phlegm (brought up phlegm or
mucus from the chest for at least 1
month per year with or apart from colds)
Age Groups: Elementary school students
age range: 5.4-16.2
Study Design: Cross-sectional
N: 7,557 returned questionnaires
7,392 included in first stage of analysis
Statistical Analyses: 2-stage
regression approach: Calculated odds
ratios and 95% CIs of respiratory
outcomes and covariates Second stage
consisted of variance-weighted linear
regressions that examined associations
between district-specific adjusted
prevalence rates and district-specific
ambient levels of each pollutant.
Covariates: Age, gender, breast-fed,
house type, number of rooms, sleeping in
own or shared room, sleeping in own or
shared bed, home coal use, ventilation
device used, homes smokiness during
cooking, eye irritation during cooking,
parental smoking, mother's education
level, mother's occupation, father's
occupation, questionnaire respondent,
year of questionnaire administration,
season of questionnaire administration,
parental asthma prevalence
Pollutant: PMio
Averaging Time: 2 years
Mean (SD): 151 (56)
IQR: 87
Range (Min, Max):
Gives range (max.-min.):
80
Monitoring Stations:
2 types: municipal monitoring stations
over a period of 4 years (1993-1996)
schoolyards of participating children over
a period of 2 years (1995-1996)
PM Increment: Interquartile range
corresponded to 1 unit of change.
RR Estimate [Lower CI, Upper CI]
lag:
Association between persistent phlegm
and PMio: 3.21 (1.55, 6.67)
p < 0.05
Between and within city modeled ORs,
scaled to interquartile range of
concentrations for each pollutant.
No associations between any type of
respiratory outcome and PMio
When scaled to an increment of 50 /yglm3
of PMio, ORs were:
Wheeze: 1.07
Asthma: 1.18
Bronchitis: 1.53
Hospitalization: 1.17
Persistent cough: 1.20
Persistent phlegm: 1.95
'All units expressed in/yglm3 unless otherwise specified.
Table E-23. Long-term exposure - respiratory morbidity outcomes - PM102.5.
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Chattopadhyay et al. (2007,
Outcome: pulmonary function tests
Pollutant: PM0304
PM Increment: NR
147471)
(respiratory impairments)
Averaging Time: 8 h
Respiratory impairments (SD):
Period of Study: NR
Age Groups: All ages


Mean (SD):
North Kolkata
Location: Three different points in
Study Design: Cross-sectional
North Kolkata: 266.1
Male (n — 137)
Kolkata, India: North, South, and Central
N: 505 people studied for PFT




Central Kolkata: 435.3
Restrictive: 4 (2.92)

total population of Kolkata not given



South Kolkata: 449.1
Obstructive: 5 (3.64)

Statistical Analyses:



Unit (i.e./yg/m3):/yg/m3
Combined Res. And Obs.: 6 (4.37)

Frequencies



Monitoring Stations: 1
Total: 15(10.95)

Covariates: Meteorologic data (i.e.



temperature, wind direction, wind speed
I Copollutant (correlation):
Female (n — 152)

and humidity)
PMio
Restrictive: 3 (1.97)

Dose-response Investigated? No
PM < 10-3.3
Obstructive: 5 (3.28)



Combined Res. And Obs.: 0
July 2009
E-521
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Total: 8 (5.26)
Total (n — 289)
Restrictive: 7 (2.42)
Obstructive: 10 (3.46)
Combined Res. And Obs.: 6 (2.07)
Total: 23 (7.96)
Central Kolkata
Male (n-44)
Restrictive: 6 (13.63)
Obstructive: 1 (2.27)
Combined Res. And Obs.: 1 (2.27)
Total: 8 (18.18)
Female (n — 50)
Restrictive: 3 (6.00)
Obstructive: 2 (4.00)
Combined Res. And Obs.: 0
Total: 5 (10.00)
Total (n — 94)
Restrictive: 9 (9.57)
Obstructive: 3 (3.19)
Combined Res. And Obs.: 1 (1.06)
Total: 13(13.82)
South Kolkata
Male (n — 52)
Restrictive: 1 (1.92)
Obstructive: 2 (3.84)
Combined Res. And Obs.: 3 (5.76)
Total: 6 (11.53)
Female (n — 70)
Restrictive: 2 (2.85)
Obstructive: 1 (1.42)
Combined Res. And Obs.: 0
Total: 3 (4.28)
Total (n — 122)
Restrictive: 3 (2.45)
Obstructive: 3 (2.45)
Combined Res. And Obs.: 3 (2.45)
Total: 9 (7.37)
Reference: Chattopadhyay et al. (2007,
Outcome: pulmonary function tests
Pollutant: PM< 10-3.3
PM Increment: NR
147471)
(respiratory impairments)
Averaging Time: 8 h
Respiratory impairments (SD):
Period of Study: NR
Age Groups: All ages


Mean (SD):
North Kolkata
Location: Three different points in
Study Design: Cross-sectional
North Kolkata: 269.8
Male (n — 137)
Kolkata, India: North, South, and Central
N: 505 people studied for PFT
Central Kolkata: 679.2
Restrictive: 4 (2.92)

total population of Kolkata not given



South Kolkata: 460.1
Obstructive: 5 (3.64)
July 2009
E-522
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Statistical Analyses: Frequencies
Covariates: Meteorologic data (i.e.
temperature, wind direction, wind speed,
and humidity)
Dose-response Investigated? No
Combined Res. And Obs.: 0
Total: 8 (5.26)
Total (n — 289)
Restrictive: 7 (2.42)
Obstructive: 10 (3.46)
Combined Res. And Obs.: 6 (2.07)
Total: 23 (7.96)
Central Kolkata
Male (n-44)
Restrictive: 6 (13.63)
Obstructive: 1 (2.27)
Combined Res. And Obs.: 1 (2.27)
Total: 8 (18.18)
Female (n — 50)
Restrictive: 3 (6.00)
Obstructive: 2 (4.00)
Combined Res. And Obs.: 0
Total: 5 (10.00)
Total (n — 94)
Restrictive: 9 (9.57)
Obstructive: 3 (3.19)
Combined Res. And Obs.: 1 (1.06)
Total: 13(13.82)
South Kolkata
Male (n — 52)
Restrictive: 1 (1.92)
Obstructive: 2 (3.84)
Combined Res. And Obs.: 3 (5.76)
Total: 6 (11.53)
Female (n — 70)
Restrictive: 2 (2.85)
Obstructive: 1 (1.42)
Combined Res. And Obs.: 0
Total: 3 (4.28)
Total (n — 122)
Restrictive: 3 (2.45)
Obstructive: 3 (2.45)
Combined Res. And Obs.: 3 (2.45)
Total: 9 (7.37)
Unit (i.e./yg/m ):/yg/m
Monitoring Stations: 1
Copollutant (correlation):
PMio
PM< 3.3-0.
Combined Res. And Obs.: 6 (4.37)
Total: 15(10.95)
Female (n — 152)
Restrictive: 3 (1.97)
Obstructive: 5 (3.28)
July 2009
E-523
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Dales et al., (2008,156378) Outcome: Pulmonary function and
inflammation
Period of Study: Location: Windsor, ON
Age Groups: grades 4-6
Study Design: cross-sectional
prevalence design
Statistical Analyses: multivariate linear 95th: 8.23
Pollutant: PMio-2.5
Averaging Time: Annual
Mean: 7.25
5th: 6.02
regression
Covariates: Ethnic background, smokers
at home, pets at home, acute respiratory
illness, medication use
Copollutant:
SO?
NO?
Increment: Tertiles of exposure
FEVi:
<7.04: 2.18 ± 0.01
7.04-7.53: 2.19 ± 0.02
>7.53: 2.14 ± 0.01
FVC:
<7.04: 2.52 ± 0.02
7.04-7.53: 2.53 ± 0.02
>7.53: 2.48 ± 0.02
eNO:
<7.04: 15.48 ± 0.63
7.04-7.53: 16.73 ± 0.76
>7.53: 16.59 ± 0.79
Reference: Gauderman et al. (2000,
012531)
Period of Study: 1993-1997
Location: Southern California
Outcome: FVC, FEVi, MMEF, FEFjb
Age Groups: fourth, seventh, or tenth
Study Design: cohort
N: 3035 subjects
Statistical Analyses: Linear regression
Covariates: Height, weight, BMI,
asthma, smoking, exercise, room
temperature, barometric pressure
Dose-response Investigated? Yes
Statistical Package: SAS
Pollutant: PMio-2.5
Averaging Time: 24 h avg PMio &
annual avg of 2-week avg PM2.6
Mean (SD): PMio-2.5 25.6
Copollutant (correlation):
0s
r - -0.29
NO? r - 0.44
Inorg. Acid r - 0.43
Increment: 25.6 /yglm3
% Change (Lower CI, Upper CI)
PMio-2.5-4th grade
FVC-0.57 (-1.20 to-0.06)
FEVi-0.90 (-1.71 to-0.09)
MMEF-1.37 (-2.57 to-0.15)
FEFjb -1.62 (-3.24, 0.04)
PMio-2.5-7th grade
FVC-0.35 (-1.02, 0.31)
FEVi-0.49 (-1.21,0.24)
MMEF-0.64 (-2.83,1.60)
FEFjb-0.74 (-2.65,1.20)
PMio-2.5-10th grade
FVC-0.17 (-1.32, 0.99)
FEVi-0.68 (-2.15, 0.81)
MMEF-1.41 (-5.85, 3.25)
FEFjb-2.32 (-6.60, 2.17)
Reference: Gauderman et al. (2002,
026013)
Period of Study: 1996-2000
Location: Southern California
Outcome: Lung function development:
FEVi, maximal mid-expiratory flow
(MMEF)
Age Groups: Fourth grade children (avg
age - 9.9 yrs)
Study Design: Cohort study
N: 1678 children, 12 communities
Statistical Analyses: Mixed model
linear regression
Covariates: Height, BMI, doctor-
diagnosed asthma and cigarette smoking
in previous year, respiratory illness and
exercise on day of test, interaction of
each of these variables with sex,
barometric pressure, temperature at test
time, indicator variables for field
technician and spirometer
Dose-response Investigated? Yes
Statistical Package: SAS (10)
Pollutant: PMio-2.5
Averaging Time: Annual 24 h averages
Mean (SD): The avg levels were
presented in an online data supplement
(Figure E1)
Monitoring Stations: 12
Copollutant (correlation):
0s (10 AM to 6 PM) r - 0.10
O3 r - -0.31
NO? r - 0.46
Acid vapor r - 0.63
PM10 r - 0.95
PMio-2.5 r - 0.81
EC r - 0.71
0C r - 0.96
PM Increment: 29.1 /yglm3
Association Estimate:
PMio-2.5 was not correlated with any of
the pulmonary function tests that were
analyzed
July 2009
E-524
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Leonardi et al. (2000,
010272)
Period of Study: 1996
Location: 17 cities of Central Europe
(Bulgaria, Czech Republic, Hungary ,
Poland, Romania, Slovakia)
Outcome: Immune biomarkers
Age Groups: 9-11
Study Design: Cross-sectional
N: 366 school children
Statistical Analyses: Linear regression
Covariates: Age, gender, parental
smoking, laboratory of analysis, recent
respiratory illness
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PMio-2.5
Averaging Time: subtracting PM2.6 from
PM10 provides avg PM10 2.5
Mean (SD): PMio-2.5: 20 (5)
Range (Min, Max):
PM 10-2.5: (12,38)
5th, median, 81 95th percentile
PMio-2.5:12,19, 29
% Change (Lower CI, Upper CI)
p-value
PMio-2.5
Neutrophils 1 (-27, 38)
>.20
Total lymphocytes 8 (-15, 38)
>.20
B lymphocytes 22 (-16, 76)
>.20
Total T lymphocytes 2 (-25, 37)
>.20
CD4+-1 (-30,41)
>.20
CD8+ 3 (-25, 41)
>.20
CD4/CD8 0 (-23, 30)
>.20
NK 1 (-33,51)
>.20
Total IgG-3 (-21,18)
>.20
Total IgM 19 (-9, 55)
>.20
Total IgA 16(-12, 52)
>.20
Total IgE-29 (-70, 70)
>.20
Reference: McConnell et al. (2003,
049490)
Period of Study: 1993-99
Location: 12 Southern CA communities
Outcome: bronchitic symptoms
Age Groups: 9-19
Study Design: communities selected on
basis of historic levels of criteria
pollutants and low residential mobility.
N: 475 children
Statistical Analyses: 3 stage regression
combined to give a logistic mixed effects
model
Covariates: sex, ethnicity,
history, asthma history, SES, insurance
status, current wheeze, current exposure
to ETS, personal smoking status,
participation in team sports, in utero
tobacco exposure through maternal
smoking, family history of asthma,
amount of time routinely spent outside by
child during 2-6 pm.
Dose-response Investigated? No
Statistical Package: SAS Glimmix
macro
Pollutant: PMio-2.5
Averaging Time: 4 year avg
Mean (SD): 17.0(6.4)
Range (Min, Max): 10.2-35.0
Copollutant (correlation):
PM2.6: r - 0.24
PM10: r - 0.79
Inorganic acid: r - 0.38
Organic Acid: r - 0.35
EC: r - 0.30
0C: r - 0.27
NO2: r - -0.22
O3: r - 0.29
PM Increment: Between community
range 24.8 /yglm3
Between community unit 1 /yglm3
Within community 1 /yglm3
OR Estimate [Lower CI, Upper CI]
Between community per range
1.38(0.65-2.92)
Between Community per unit
1.01(0.98-1.04)
Within community per unit
1.02(0.95-1.10)
July 2009
E-525
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Millstein et al. (2004,
088629)
Period of Study: Mar-Aug, 1995, and
Sep, 1995 to Feb, 1996
Data were taken from the Children's
Health Study
Location: Alpine, Atascadero, Lake
Arrowhead, Lake Elsinore, Lancaster,
Lompoc, Long Beach, Mira Loma,
Riverside, San Dimas, Santa Maria, and
Upland, CA
Outcome: Wheezing & asthma
medication use
Age Groups: 4th grade students, mostly
9 yrs at the time of the study
Study Design: Cohort Study, stratified
into 2 seasonal groups/
N: 2081 enrolled, 2034 provided parent-
completed questionnaire.
Statistical Analyses: Multilevel, mixed-
effects logistic model.
Covariates: Contagious respiratory
disease, ambient airborne pollen and
other allergens, temperature, sex, age
race, allergies, pet cats, carpet in home,
environmental tobacco smoke, heating
fuel, heating system, water damage in
home, education level of questionnaire
signer, physician diagnosed asthma.
Season: Mar-Aug, 1995, and Sep, 1995
to Feb, 1996
Statistical Package: SAS 8.00
Lags Considered: 14
Pollutant: PMio-2.5
Averaging Time: monthly
PM Component: Nitric acid, formic acid,
acetic acid
Monitoring Stations: 1 central location
in each community
Copollutant (correlation):
NO?: r - 0.29
Os:r - 0.77
PM2.6: r - -0.08
PM Increment: IQR 11.44/yg/m3
Odds Ratio [lower CI, Upper CI]
Annual
PMio-2.5: 0.96 [0.74,1.25]
March-August
PMio-2.5: 0.93 [0.54,1.59]
Sep-Feb
PMio-2.5: 0.68 [0.46,1.01]
Reference: (Parker et al., 2009,
192359)
Period of Study: 1999-2005
Location: US
Outcome: Respiratory allergyfhayfever
Study Design: Cohort
Covariates: survey year, age, family
structure, usual source of care, health
insurance, family income relative to
federal poverty level, race/ethnicity
Statistical Analysis: logistic regression
Statistical Package: SUDAAN
Age Groups: 73,198 children aged 3-17
years
Pollutant: PM10-2.E
Averaging Time: NR
Median: 11.2 /yglm3
IQR: 8.2-15.2
Copollutant (correlation):
Summer O3: 0.16
SO2: -0.33
NO?: 0.29
PM2.6: 0.02
PM10: 0.86
Increment: 10//gfm3
Odds Ratio (95% CI)
Single Pollutant Model, variable N
Adjusted: 1.01 (0.95-1.07)
Single Pollutant Model, constant N
Adjusted: 1.13(1.04-1.46)
Multi-pollutant Model: 1.16 (1.06-1.24)
July 2009
E-526
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Zhang et al. (2002, 034814)
Period of Study: 1993-1996
Location: 4 Chinese cities (urban and
suburban location in each city):
Guangzhou, Wuhan, Lanzhou, Chongqing
Outcome: Interview-self reports of
symptoms: Wheeze (ever wheezy when
having a cold)
Asthma (diagnosis by doctor)
Bronchitis (diagnosis by doctor),
Hospitalization due to respiratory disease
(ever)
Persistent cough (coughed for at least 1
month per year with or apart from colds)
Persistent phlegm (brought up phlegm or
mucus from the chest for at least 1
month per year with or apart from colds)
Age Groups: Elementary school students
age range: 5.4-16.2
Study Design: Cross-sectional
N: 7,557 returned questionnaires
7,392 included in first stage of analysis
Statistical Analyses: 2-stage
regression approach: Calculated odds
ratios and 95% CIs of respiratory
outcomes and covariates Second stage
consisted of variance-weighted linear
regressions that examined associations
between district-specific adjusted
prevalence rates and district-specific
ambient levels of each pollutant.
Covariates: Age, gender, breast-fed,
house type, number of rooms, sleeping in
own or shared room, sleeping in own or
shared bed, home coal use, ventilation
device used, homes smokiness during
cooking, eye irritation during cooking,
parental smoking, mother's education
level, mother's occupation, father's
occupation, questionnaire respondent,
year of questionnaire administration,
season of questionnaire administration,
parental asthma prevalence
Pollutant: PMio-2.5
Averaging Time: 2 years
Mean (SD): 59 (28)
Percentiles: 25th: NR
50th(Median): NR
75th: NR
IQR: 42
Range (Min, Max):
Gives range (max.-min.):
80
Monitoring Stations:
2 types: municipal monitoring stations
over a period of 4 years (1993-1996)
schoolyards of participating children over
a period of 2 years (1995-1996)
PM Increment: Interquartile range
corresponded to 1 unit of change.
RR Estimate [Lower CI, Upper CI]
lag:
Association between bronchitis and PMio-
2.5:2.20(1.14,4.26)
p < 0.05
Association between persistent cough
and PMio-2.5:1.46 (1.12,1.90)
p < 0.05
Between and within city associations:
Bronchitis: 3.18 (between city)
Persistent phlegm (between city): 2.78
When scaled to an increment of 50 /yglm3
of PMio-2.5 associations (ORs) between
respiratory outcome and PMio-2.5 were:
Wheeze: 1.14
Asthma: 1.34
Bronchitis: 2.56
Hospitalization: 1.58
Persistent cough: 1.57
Persistent phlegm: 3.45
'All units expressed in //gfm3 unless otherwise specified.
July 2009
E-527
DRAFT-DO NOT CITE OR QUOTE

-------
Table E-24. Long-term exposure - respiratory morbidity outcomes - PM2.5 (including PM
components/sources).
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Annesi-
Maesano et al.(2007,
091348)
Outcome: EIB, Flexural atopic
dermatitis, asthma, rhiniconjuctivitis,
allergic rhinitis
Averaging Time: 5-day mean (Mon.- Allergic and respiratory morbidity OR Estimate (Lower CI, Upper
Fri.) over a 13-week to 24-week span CI)
Residential Proximity Level	Proximity Level
Pollutant: PM2.6
PM Increment: High vs. Low
Period of Study: Mar Age Groups: Children mean
1999-Oct 2000	10.4 ± 0.7 yrs
Location: France	Study Design: Semi-individual design
(Bordeaux, Clermont-
Ferrand, Creteil, Marseille, N: 5338
Strasbourg,, & Reims)	c. .. .. . . . . ...
0	Statietiral Analveec nnictir
Statistical Analyses: Logistic
regression
Covariates: Age, sex, family history
of allergy, passive smoking
Season: NR
Dose-response Investigated?
No
Statistical Package: SAS
Mean (SD): Low cone: 8.7
High cone: 20.7
Range (Min, Max):
Low cone: (1.6,12.2)
High cone: (12.5, 54.0)
City Level
Mean (SD): Low cone: 9.6
High cone: 23.0
Range (Min, Max):
Low cone: (4.7,12.7)
High cone: (13.0, 54.5)
EIB (0 1.35 (1.10,1.67)
Fl. Atopic dermatitis (C) 2.51 (2.06, 3.06)
Asthma (P) 1.11 (0.88,1.39)
Atopic asthma (P) 1.43 (1.07,1.91)
Non-atopic asthma (P) 0.73 (0.49,1.07)
Rhiniconjunctivitis (P) 0.94 (0.77,1.15)
Atopic dermatitis (P) 1.05 (0.88,1.27)
Asthma (L) 1.00 (0.82,1.22)
Allergic Rhinitis (L) 1.09 (0.93,1.27)
Atopic dermatitis (L) 0.94 (0.82,1.09)
City Level
EIB (0 1.43 (1.15,1.78)
FL Atopic dermatitis (C) 2.06 (1.69, 2.51)
Asthma (P) 1.31 (1.04,1.66)
Atopic asthma (P) 1.58 (1.17, 2.14)
Non-atopic asthma (P) 1.00 (0.68,1.49)
Rhiniconjunctivitis (P) 0.98 (0.80,1.20)
Atopic dermatitis (P) 1.08 (0.90,1.30)
Asthma (L) 1.09 (0.89,1.33)
Allergic Rhinitis (L) 1.13 (0.97,1.33)
Atopic dermatitis (L) 0.95 (0.82,1.09)
Notes: C - Current
P - Past year
L - Lifetime
Allergic sensitization OR Estimate (Lower CI, Upper CI)
Proximity Level
All allergens 1.19(1.04,1.36)
Indoor allergens 1.29 (1.11,1.50)
Outdoor allergens 1.02 (0.85,1.23)
Moulds 1.13 (0.78,1.65)
City Level
All allergens 1.32(1.15,1.51)
Indoor allergens 1.51 (1.29,1.76)
Outdoor allergens 1.06 (0.88,1.28)
Moulds 1.00 (0.69,1.46)
July 2009
E-528
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Bakke et al.
(2004,156246)
Period of Study: January
1,1989-June 31, 2002
Location: One of
Norway's major
construction companies
Outcome: Spirometric measurements
Age Groups: All ages, mean - 39
yrs
Study Design: cohort
N: 651 male construction workers
Statistical Analyses: Multiple linear
regression models
Covariates: Age, years for non-
smokers and ever smokers
Dose-response Investigated? No
Statistical Package: SYSTAT 10.0
and SPSS 11.0
Pollutant: Respirable dust
Averaging Time: 5-8 h
Mean (SD): Drill and blast workers:
6.3(2.8)
Tunnel concrete workers: 6.1 (3.1)
Shotcreting operators: 19(11)
TBM workers: 16 (6.6)
Outdoor concrete workers: 1.4 (0.73)
Foremen: 0.28 (0.48)
Engineers: 0.09 (0.28)
Unit (i.e. //g/m3): mgny/m3
Monitoring Stations: 16 tunnel
sites visited with sampling equipment
Copollutant (correlation): Total
dust: r - 0.99
~ quartz: r - 0.48
NO2: r - 0.75
CO: r - 0.61
Oil mist: r - 0.83
Oil vapor: r - 0.68
V0C: r - 0.89
PM Increment: NR-exposure respirable dust
Effect Estimate (Lower CI, Upper CI):
Lung function changes predicted by multiple linear regression
models using one exposure variable adjusted for age and
observation time by non-smokers and ever smokers
Non-smokers: B - -16.0
(-24- -6.8)
SE - 4.5
Ever smokers: B - -9.3
(-17-1.6)
SE - 4.0
Reference: Bakke et al.
(2004,156246)
Period of Study: January
1,1989-June 31, 2002
Location: One of
Norway's major
construction companies
Outcome: Spirometric measurements
Age Groups: All ages, mean - 39
yrs
Study Design: cohort
N: 651 male construction workers
Statistical Analyses: Multiple linear
regression models
Covariates: Age, years for non-
smokers and ever smokers
Dose-response Investigated?
No
Statistical Package:
SYSTAT 10.0 and SPSS 11.0
Pollutant: Total dust
Averaging Time: 5-8 h
Mean (SD): Drill and blast workers:
18(7.8)
Tunnel concrete workers: 21 (11)
Shotcreting operators: 73 (41)
TBM workers: 48 (20)
Outdoor concrete workers: 6.5 (3.4)
Foremen: 0.78 (1.3)
Engineers: 0.27 (0.78)
Unit (i.e. //g/m3): mgny/nr1
Monitoring Stations: 16 tunnel
sites visited with sampling equipment
Copollutant (correlation):
Respirable dust: r - 0.99
~ quartz: r - 0.42
NO?: r - 0.67
CO: r - 0.49
Oil mist: r - 0.81
Oil vapor: r - 0.64
V0C: r - 0.91
PM Increment: NR-exposure expirable dust
Lung function changes predicted by multiple linear regression
models using one exposure variable adjusted for age and
observation time by non-smokers and ever smokers
Non-smokers: B - -4.0 (-6.5-1.4)
SE - 1.3
Ever smokers: B - -2.0 (-4.2-0.23)
SE - 1.1
July 2009
E-529
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Bennett et al.
(2007,156268)
Period of Study: 1992-
2005
Location: Melbourne,
Australia
Outcome: Respiratory symptoms
(from questionnaire)
Age Groups: All ages, mean - 37.2
yrs
Study Design: cohort
N:1446
Statistical Analyses: Logistic
regression models
Covariates: age, gender, use of B2-
agonists, use of inhaled
corticosteroids, smoking, year of
data collection, and avg daily
exposure to PM2.6 in the 12 months
corresponding to the time frame of
symptoms
Dose-response Investigated?
No
Statistical Package: STATA,
version 9
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD): 6.8
Range (Min, Max): (1.8-73.3)
Monitoring Stations: up to 3
PM Increment: NR
Effect Estimate [Lower CI, Upper CI]:
Respiratory symptoms in last 12 months and exposure to
ambient PM2.6 over the same period
Within-person (longitudinal) effects
Wheeze: OR - 1.08 (0.79-1.48), p - 0.62
SOB on waking: OR - 1.34 (0.84-2.16), p - 0.22
Cough (AM): OR - 0.74 (0.47-1.15), p - 0.18
Phlegm (AM): OR - 1.55 (0.95-2.53), p - 0.08
Cough w/ phlegm (AM): OR - 1.28 (0.70-2.33), p - 0.42
Asthma attack: OR - 0.91 (0.55-1.49), p - 0.69
Between-person (cross-sectional) effects
Wheeze: OR - 1.32 (0.82-2.10), p - 0.25
SOB on waking: OR - 1.29 (0.46-3.60), p - 0.63
Cough (AM): OR - 0.21 (0.07-0.62), p - 0.01
Phlegm (AM): OR - 0.49 (0.16-1.44), p - 0.19
Cough w/ phlegm (AM): OR - 0.28 (0.08-0.97), p - 0.05
Asthma attack: OR - 0.52 (0.17-1.59), p - 0.26
July 2009
E-530
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Allergen sensitivity (any, indoor,
Period of Study: 1999- outdoor, food, total) IgE > 100 lll/mL
Location: The
Netherlands
Asthma (probable, MDdiagnosed,
Location: The	ever MD diagnosed)
Reference: (Brauer et al„ Outcome:
2007,090691)
Dry cough at night
Itchy rash
Itchy rashfeczema
Ear/Nose/Throat (ENT) infection
Eczema, MD diagnosed
Eczema, ever MD diagnosed
Flu/serious cold, MDdiagnosed
Wheeze (ever, early, early frequent,
persistent)
Age Groups: very young children
(< 4-years-old) enrolled prenatally
Study Design: prospective birth
cohort study
N: —4000 subjects
Statistical Analyses: multiple
logistic regression
Dose-response Investigated? No
diagnosed)
Bronchitis (MD diagnosed, ever MD-
Averaging Time: 12 months
Mean (SD): SD: NR
16.9
Percentiles: 25th: 14.8
Pollutant: PM2.6
50th(Median): 17.3
75th: 18.1
PM Increment: IQR 3.3 (xglm3
Notes: Traffic-related pollution (PM2.6, soot, NO2) was associated
with respiratory infections, asthma, and allergic sensitization in
children during the first four years of life.
Symptom At 4-Years-0ld
Wheeze
4-years-old: 1.23(1.00: 1.51]
Early-life: 1.20(0.99: 1.46]
Range (Min, Max): (13.5, 25.2) Asthma, MD-diagnosed
Monitoring Stations: 40	4-years-old: 1.15 (0.82: 1.62]
Copollutant (correlation): Soot: r - Early-life: 1.32 (0.96: 1.83]
0.97
4-years-old: 1.11 (0.94: 1.31]
Early-life: 1.14(0.98: 1.33]
Bronchitis, MD-diagnosed
4-years-old: 0.88 (0.66: 1.18]
Early-life: 0.86(0.66: 1.11]
ENT infection
4-years-old: 1.13(0.98: 1.31]
Early-life: 1.17(1.02: 1.34]
Flu/serious cold, MD-diagnosed
4-years-old: 1.21 (1.02: 1.42]
Early-life: 1.25(1.07: 1.46]
Itchy rash
4-years-old: 0.96 (0.82: 1.11]
Early-life: 0.98(0.85: 1.14]
Eczema, MD-diagnosed
4-years-old: 1.00 (0.88: 1.21]
Early-life: 0.98(0.82: 1.17]
Allergen Sensitivity At 4-Yr-0ld
Allergen, any: 1.55(1.13:2.11]
Allergen, indoor: 1.03(0.69: 1.55]
Allergen, outdoor: 0.93(0.54: 1.58]
Allergen, food: 1.75 (1.23: 2.47]
Allergen, total IgE > 100 lU/mL: 0.84 (0.59: 1.18]
Cumulative Allergy/Asthma Symptoms At 4-Years-0ld
Wheeze, ever: 1.22 (1.06: 1.41]
Asthma, ever MD diagnosed: 1.32(1.04: 1.69]
Asthma, probable: 1.08 (0.90:1.30]
Wheeze, early: 1.16(1.00: 1.34]
Wheeze, persistent: 1.19(0.96: 1.48]
Wheeze, early frequent: 1.19 (0.96:1.47]
Bronchitis, ever MD-diagnosed: 0.96 (0.81: 1.13]
Itchy rash/eczema: 0.99 (0.88: 1.13]
Eczema, ever MD diagnosed: 0.98 (0.85:1.13]
July 2009
E-531
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Braueret al„
2007,090691)
Period of Study: 1999-
2000
Location: The
Netherlands
Outcome: Allergen sensitivity (any,
indoor, outdoor, food, total)
IgE > 100 lU/mL
Asthma (probable, MDdiagnosed,
ever MD diagnosed)
Bronchitis (MD diagnosed, ever MD-
Dry cough at night
Itchy rash
Itchy rashfeczema
Ear/Nose/Throat (ENT) infection
Eczema, MD diagnosed
Eczema, ever MD diagnosed
Flu/serious cold, MDdiagnosed
Wheeze (ever, early, early frequent,
persistent)
Age Groups: very young children
(< 4-years-old) enrolled prenatally
Study Design: prospective birth
cohort study
N: —4000 subjects
Statistical Analyses: multiple
logistic regression
Dose-response Investigated? No
Pollutant: Soot (as PM2.6
absorbance)
Averaging Time: 12 months
Mean (SD): 1.71
Percentiles: 25th: 1.33
50th(Median): 1.78
75th: 1.91
Range (Min, Max):
(0.77,3.68)
Unit (i.e. /yg/m3): 1 E-5/m
Monitoring Stations: 40
Copollutant (correlation):
NO?: r - 0.96
PM2.b: r - 0.97
PM Increment: IQR 0.58 E-5/m
Notes: Traffic-related pollution (PM2.6, soot, NO2) was associated
with respiratory infections, asthma, and allergic sensitization in
children during the first four years of life.
Symptom At 4-Years-0ld
Wheeze
4-years-old: 1.18(0.98: 1.41]
Early-life: 1.18(1.00: 1.40]
Asthma, MD diagnosed
4-years-old: 1.15(0.85: 1.55]
Early-life: 1.30(0.98: 1.71]
Dry cough at night
4-years-old: 1.13(0.97: 1.30]
Early-life: 1.14(1.00: 1.31]
Bronchitis, MD diagnosed
4-years-old: 0.90 (0.69: 1.16]
Early-life: 0.88(0.69: 1.11]
ENT infection
4-years-old: 1.15(1.01: 1.31]
Early-life: 1.16(1.03: 1.31]
Flu/serious cold, MD-diagnosed
4-years-old: 1.18(1.02: 1.36]
Early-life: 1.19(1.04: 1.37]
Itchy rash
4-years-old: 0.94 (0.82: 1.08]
Early-life: 0.97(0.85:1.10]
Eczema, MD-diagnosed
4-years-old: 0.99 (0.84: 1.17]
Early-life: 0.97(0.83:1.14]
Allergen Sensitivity At 4-Yrs-0ld
Allergen, any: 1.45 (1.11:1.91]
Allergen, indoor: 1.02(0.71: 1.46]
Allergen, outdoor: 0.95 (0.59: 1.52]
Allergen, food: 1.64(1.21: 2.23]
Allergen, total IgE > 100 lU/mL: 0.80 (0.59: 1.09]
Cumulative Allergy/Asthma Symptoms At 4-Years-0ld
Wheeze, ever: 1.18 (1.04: 1.34]
Asthma, ever MD diagnosed: 1.26(1.02: 1.56]
Asthma, probable: 1.06 (0.90:1.24]
Wheeze, early: 1.11 (0.97: 1.26]
Wheeze, persistent: 1.18(0.98: 1.42]
Wheeze, early frequent: 1.14 (0.95:1.37]
Bronchitis, ever MD-diagnosed: 0.95 (0.82: 1.10]
Itchy rash/eczema: 0.99 (0.89: 1.11]
Eczema, ever MD diagnosed: 0.99 (0.87:1.12]
July 2009
E-532
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Braueret al.
(2002,035192)
Period of Study: NR
Location: The
Netherlands
Outcome: Questionnaire derived
wheezing, dry nighttime cough, ear,
nose and throat infections, skin rash
Physician diagnosed asthma,
bronchitis, influenza, eczema
Age Groups: age 2
Study Design: Prospective cohort
N: 4146 children
Statistical Analyses: Logistic
regression
Covariates: Maternal age, maternal
smoking, mattress cover (allergen-
free), maternal education, paternal
education, gender, gas stove, gas
water heater, any other siblings,
ethnicity, breastfeeding, mold at
home, pets, allergies in mother,
allergies in father
Dose-response Investigated? No
Pollutant: PM2.6
Averaging Time: 4 2-week periods
dispersed throughout 1 year, adjusted
for temporal trend
Mean (SD): 16.9
Percentiles: 10th: 14.0
25th: 15.0
50th(Median): 17.3
75th: 18.2
90th: 19.1
Range (Min, Max): 13.5, 25.2
Monitoring Stations: 40
Copollutant (correlation):
Soot: r - 0.99
NO?: r - 0.97
PM Increment: 3.2/yglm3
OR Estimate [Lower CI, Upper CI];
Unadjusted
Wheeze 1.14(0.99-1.30)
Asthma 1.08(0.84-1.37)
Dry cough at night 1.10 (0.95-1.27)
Bronchitis 1.00 (0.85-1.18)
E, N,T infections 1.14 (0.99-1.33)
Flu 1.15 (1.03-1.28)
Itchy rash 1.07 (0.95-1.20)
Eczema 1.02 (0.90-1.16)
Adjusted
Wheeze 1.14(0.98-1.34)
Asthma 1.12(0.84-1.50)
Dry cough at night 1.04 (0.88-1.23)
Bronchitis 1.04 (0.85-1.26)
E, N,T infections 1.20 (1.01-1.42)
Flu 1.12 (1.00-1.27)
Itchy rash 1.01 (0.88-1.16)
Eczema 0.95 (0.83-1.10)
Reference: Brauer et al.
(2002,035192)
Period of Study: NR
Location: The
Netherlands
Outcome: Questionnaire derived
wheezing, dry nighttime cough, ear,
nose and throat infections, skin rash
Physician diagnosed asthma,
bronchitis, influenza, eczema
Age Groups: age 2
Study Design: Prospective cohort
N: 4146 children
Statistical Analyses: Logistic
regression
Covariates: Maternal age, maternal
smoking, mattress cover (allergen-
free), maternal education, paternal
education, gender, gas stove, gas
water heater, any other siblings,
ethnicity, breastfeeding, mold at
home, pets, allergies in mother,
allergies in father
Dose-response Investigated? No
Pollutant: PM2.6 "soot"
Averaging Time: 4 2-week periods
dispersed throughout 1 year, adjusted
for temporal trend
Mean (SD): 16.9 10-5/m
Percentiles: 10th: 1.16
25th: 1.38
50th(Median): 1.78
75th: 1.92
90th: 2.19
Range (Min, Max):
0.77,3.68
Unit (i.e. /yg/m3): 10-5/m
Monitoring Stations: 40
Copollutant (correlation): PM2.6
(r - 0.99)
NO? (r - 0.96)
PM Increment: 0.54 x 10-5/m
(equivalent to 0.8 /yg/m3 elemental carbon)
OR Estimate [Lower CI, Upper CI]
Unadjusted
Wheeze 1.11 [0.99-1.24]
Asthma 1.07(0.87-1.31]
Dry cough at night 1.08 [0.95-1.21]
Bronchitis 0.98 [0.85-1.12]
E, N,T infections 1.12 [0.99-1.27]
Flu 1.13 [1.03-1.23]
Itchy rash 1.07(0.97-1.19]
Eczema 1.01 [0.91-1.13]
Adjusted
Wheeze 1.11 [0.97-1.26]
Asthma 1.12(0.88-1.43]
Dry cough at night 1.02 [0.88-1.17]
Bronchitis 0.99 [0.84-1.17]
E, N,T infections 1.15 [1.00-1.33]
Flu 1.09 [0.98-1.21]
Itchy rash 1.02(0.91-1.15]
Eczema 0.96 [0.85-1.08]
July 2009
E-533
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Braueret al.
(2006,090757)
Period of Study: 1997-
2001
Location: Germany
The Netherlands
Outcome:
Otitis Media (parental report of
doctor's diagnosis prior to age 2
years)
Age Groups: 0-2 years
Study Design: Prospective Cohort
Study
N: 4,379 children total
The Netherlands: 3,714
Germany: 665
Statistical Analyses: Logistic
regression
Covariates: Sex, parental atopy,
maternal education, siblings,
maternal smoking during pregnancy,
ETS exposure at home, use of gas for
cooking, indoor moulds and
dampness, number of siblings, breast-
feeding, and presence of pets in the
home
Season: All
Dose-response Investigated?
No
Pollutant: PM2.6
PM Component: Elemental Carbon
(EC)
Averaging Time: 8 weeks (4 2-week
periods dispersed throughout 1 year,
adjusted for temporal trends)
Mean: The Netherlands:
PM2.b: 16.9
EC: 1.72
Germany:
PM2.b: 13.4
EC: 1.76
Range (Min, Max):
The Netherlands:
PM2.6:13.5, 25.2
EC: 0.77,3.68
Germany:
PM2.B: 12.0, 21.9
EC: 1.40,4.39
Monitoring Stations: 80 (40 for
each cohort)
PM Increment: PM2b: 3/yglm3 (-
EC: ~ 0.5 //g/m3 (~ IQR)
OR Estimate [Lower CI, Upper CI]
The Netherlands:
PM2.6:
At age 1:1.13 (0.98-1.32)
At age 2:1.13 (1.00-1.27)
EC:
At age 1:1.11 (0.98-1.26)
At age 2:1.10 (1.00-1.22)
Germany:
PM2.6:
At age 1:1.19 (0.73-1.92)
At age 2:1.24 (0.84-1.83)
EC:
At age 1:1.12 (0.83-1.51)
At age 2:1.10 (0.86-1.41)
IQR)
Reference: (Burr et al.,
2004,189788)
Period of Study: 3
weeks in July and Jan
1997 and 2 weeks in Nov
1996 and April 1997
Location: North Wales,
England
Outcome: Self-report of symptoms
only for wheeze, cough, phlegm,
rhinitis, and itchy eyes.
Age Groups: all
Study Design: Repeated measures
N: 386 persons in congested streets
and 425 in the uncongested streets
in 1996/1997. Of these, 165 and
283 completed the second phase of
the study.
Pollutant: PM2.6
Averaging Time: Mean hourly
concentrations
Mean (SD):
Congested Streets
1996-97 21.2
1998-99 16.2
Uncongested Streets
1996-97 6.7
1998-99 4.9
Monitoring Stations: 1 in
congested street and 1 in
uncongested
% change PM10 in congested streets: 23.6
% change PM10 in uncongested streets: 26.6
Uncongested street sampling site was 20 m from the congested
street sampler.
The opening of the by-pass produced a reduction in pollution in
the congested streets. The health effects of these changed are
likely to be greater for nasal and ocular symptoms than for lower
respiratory symptoms. Uncertainty about the causality arises
from low response rates and conflicting trends in respiratory and
nasal symptoms.
July 2009
E-534
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Calderon-
Garciduehas et al. (2006,
091253)
Period of Study: 1999,
2000
Location: Southwest
Mexico City &Tlaxcala,
Mexico
Outcome: Hyperinflation, interstitial
markings-measured by chest
radiograph, and lung function-FVC,
FEVi, PEF, FEF25-75, measured using
spirometry tests
Age Groups: 5-13 yrs
Study Design: Cohortl 999-
N: 249 (total), 230 (Southwest
Mexico City), 19 (Tlaxcala)
Statistical Analyses: Bayes test,
Spearman rank correlation, multiple
regression
Covariates: Age, sex
Dose-response Investigated? No
Statistical Package: SAS 8.2
Pollutant: PM2.6
Averaging Time: 1 yr
Mean (SD): 21
2000-19
Tlaxacala:
1994-2000: < NAAQS std
Mexico City
Monitoring Stations:
Southwest Mexico City-2
Tlxacala-periodic air monitoring data
Copollutant: 0:i
PM Increment: NR
% Change:
% of children with FEVi < 80% expected value:
Mexico City (n - 77): 7.8%
Tlaxacala (n - 19): 0%
% children with hyperinflation: Mexico City: 65.6%
Number with:
No hyperinflation: 79
Mild: 72
Moderate: 56
Severe: 23
Tlaxacala: 5.3%
Number with:
No hyperinflation: 18
Mild: 1
Moderate: 0
Severe: 0
% children with interstitial markings:
Mexico City: 52.6%
Number with:
No interstitial markings: 19
Mild: 0
Moderate: 0
Severe: 0
Tlaxacala: 0%
Number with:
No interstitial markings: 109
Mild: 112
Moderate: 9
Severe: 0
July 2009
E-535
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Cesaroni et al.
(2008,156326)
Period of Study: Data on
PM emissions collected in
2002
cross-sectional survey
carried out in 1995
Location: Rome, Italy
Outcome: Self-reported chronic
bronchitis or emphysema, asthma,
and rhinitis
Age Groups: 25-59 yrs
Study Design: Cross-sectional
N: 9,488 subjects who had been
residents in same place for at least 3
yrs and who had participated in an
extension of the ISAAC initiative in
Italy in 1994 & 1995
Statistical Analyses: GEE with a
logit link
Covariates: sex, age, smoking
habits, education level, and variable
to account for correlation of data for
members of the same family
Effect Modifiers: stratified analysis
by smoking status (only presented for
the traffic score variable)
also stratified by education level
(data not shown)
Dose-response Investigated: Wald
test to calculate p for trend
Pollutant: PM emissions (estimated)
Emissions estimated using a
model/method based on factors such
as vehicle park, driving conditions,
emission factors, fuel consumption,
fuel properties, road gradients, and
climatic conditions
Mean: 0.12 kg/km2
SD: 0.081
Odds Ratios for quartiles of PM emissions:
Chronic bronchitis or emphysema (n - 397):
1st: 1.00
2nd: 0.96 (0.71,1.30)
3rd: 0.90(0.66,1.23)
4th: 1.05 (0.77,1.42)
p-trend - 0.871
Asthma (n - 472):
1st: 1.00
2nd: 1.10(0.84,1.44)
3rd: 0.94 (0.71,1.24)
4th: 1.06 (0.80,1.39)
p-trend - 0.980
Rhinitis (n - 1227):
1st: 1.00
2nd: 1.41 (1.17,1.69)
3rd: 1.11 (0.92,1.34)
4th: 1.37 (1.14,1.64)
p-trend - 0.018
Reference: Dales et al.,
(2008,156378)
Period of Study:
Location: Windsor, ON
Outcome: Pulmonary function and
inflammation
Age Groups: grades 4-6
Study Design: cross-sectional
prevalence design
Statistical Analyses: multivariate
linear regression
Covariates: Ethnic background,
smokers at home, pets at home,
acute respiratory illness, medication
use
Pollutant: PM2.6
Averaging Time: Annual
Mean: 15.4
5th: 14.2
95th: 17.2
Copollutant:
SO?
NO?
Increment: Tertiles of exposure
FEVi:
<15.19: 2.16 ± 0.01
15.19-15.96: 2.17 ± 0.02
>15.96: 2.18 ± 0.01
FVC:
<15.19: 2.51 ± 0.02
15.19-15.96: 2.50 ± 0.02
>15.96: 2.52 ± 0.02
eNO:
<15.19: 16.08 ± 0.70
15.19-15.96: 15.80 ± 0.76
>15.96: 16.79 ± 0.72
July 2009
E-536
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Gauderman et Outcome: FVC, FEVi, MMEF, FEF76
al. (2000, 012531)
Age Groups: fourth, seventh, or
tenth graders
Period of Study: 1993
1997
Location: Southern
California
Study Design: cohort
N: 3035 subjects
Statistical Analyses: Linear
regression
Covariates: Height, weight, BMI,
asthma, smoking, exercise, room
temperature, barometric pressure
Dose-response Investigated? Yes
Statistical Package: SAS
Pollutant: PM2.6
Averaging Time: annual avg of 2-
week avg PM2.6
Mean (SD): PM21, 25.9
Copollutant (correlation):
0a: r - -0.32
PM 10-2.5: r - 0.76
NO?: r - 0.74
Inorg. Acid: r - 0.79
Increment: 25.9/yg/m3
% Change (Lower CI, Upper CI)
PM2.B-4th grade
FVC-0.47 (-0.94,0.01)
FEVi-0.64 (-1.28, 0.01)
MMEF-1.03 (-1.95 to-0.09)
FEF76 -1.31 (-2.57 to-0.03)
PM2.B-7th grade
FVC -0.42 (-0.89, 0.05)
FEVi-0.32 (-0.88, 0.24)
MMEF-0.29 (-1.99,1.44)
FEFjb -0.26 (-1.75,1.25)
PM2.B-10th grade
FVC 0.19 (-0.68,1.07)
FEVi-0.25 (-1.41, 0.93)
MMEF-0.17 (-3.66, 3.46)
FEFjb-0.79 (-4.27, 2.82)
Reference: Gauderman et Outcome: Lung function
al. (2002, 026013)
Period of Study: 1996-
2000
Location: Southern
California
development: FEVi, maximal
midexpiratory flow (MMEF)
Age Groups: Fourth grade children
(avg age - 9.9 yrs)
Study Design: Cohort study
N: 1678 children, 12 communities
Statistical Analyses: Mixed model
linear regression
Covariates: Height, BMI, doctor-
diagnosed asthma and cigarette
smoking in previous year, respiratory
illness and exercise on day of test,
interaction of each of these variables
with sex, barometric pressure,
temperature at test time, indicator
variables for field technician and
spirometer
Dose-response Investigated? Yes
Statistical Package: SAS (10)
Pollutant: PM2.6
Averaging Time: Annual 24 h
averages
Mean (SD): The avg levels were
presented in an online data
supplement (Figure E1)
PM Component: Elemental carbon
and organic carbon.
Monitoring Stations: 12
Copollutant (correlation):
O3: (10 AM to 6 PM) r - 0.14
O3: r = -0.39
NO2: r - 0.77
Acid vapor: r - 0.87
PM10: r - 0.95
PM 10-2.5: r - 0.81
EC: r - 0.93
0C: r - 0.89
PM Increment: 22.2/yglm3
Association Estimate:
Non-statistically significant negative correlation between PM2.6
and FEViand FVC growth rates were observed. MMEF growth
rates had a negative correlation with PM2.6 (r - -0.43 p - 0.05).
PM2.6 was not significantly correlated to FEVi (r - -0.31
p - 0.25)
July 2009
E-537
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Gauderman
et al., 2004, 056569)
Period of Study: Air
pollution data
ascertainment: 1994-
2000. Spirometry testing:
spring 2001- spring 2003
Location: 12
Communities in Southern
California
Outcome: Lung function
FVC, FEVi, MMEF (Maximal
midexpiratory flow rate)
Age Groups: Children, Avg age 10
years
Study Design: Prospective Cohort
Study
N: 12 Communities
2,034 children
24,972 child-months
Statistical Analyses: Linear
regression of changes in sex-and-
community specific lung growth
function and PM
Correlation between % with low
attained FEVi and PM.
Covariates: Random effect for
communities
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PM2.6
PM Increment: Most to least polluted community Range:
22.8 /yglm3
Difference in Lung Growth [Lower CI, Upper CI];
Mean: Means are presented in figures FVC -60.1 (-166.1 to 45.9)
FEVi -79.7 (-153.0 to j6.4)
MMEF-168.9 (-345.5 to 7.8)
Correlation with % below 80% predicted Lung function (p-value)
PM2.b: 0.79 (0.002)
Averaging Time: 2-week
measurements used to create annual
averages
only
Range (Min, Max): ~6, —27
Monitoring Stations: 12
Copollutant (correlation): PMio: r
- 0.95
0s:r - 0.18
NO2: r - 0.79
EC: r - 0.91
0C: r - 0.91
Reference: Gauderman et
al. (2007, 090121)
Period of Study: 1993-
2004
Location: 12 Southern
California Communities
Outcome: pulmonary function tests
FVC, FEVi, MMEF/FEF26.76
Age Groups: Children (mean age 10
at recruitment, followed for 8 years)
Study Design: Cohort Study
(Children's Health Study)
N: 3677 children (1718 in cohort 1
recruited 1993 and 1959 in cohort 2
recruited 1996)
22686 pulmonary function tests.
Statistical Analyses: Hierarchical
mixed effects model with linear
splines
Covariates: Adjustments for height,
height squared, BMI, BMI squared,
present asthma status, exercise or
respiratory illness on day of test,
smoking in previous year, field
technician, traffic indicator (distance
from freeway, distance from major
roads), random effects for participant
and community.
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PM2.6
Monitoring Stations: 1 in each
community
PM Increment: 22.8/yglm3
Pollutant effect reported as difference in 8 year lung function
growth from least to most polluted community. Negative
difference indicate growth deficits associated with exposure. For
PM2.6 FEV growth deficit is ¦ 100
Reference: Gehring et al.
(2002, 036250)
Period of Study: 1995-
2002
Location: Munich,
Germany
Outcome: wheezing, cough without
infection, dry cough at night,
obstructive, spastic or asthmoid
bronchitis, respiratory infections,
sneezing, runny/stuffed nose
Age Groups: 0-2 years
Study Design: Prospective cohort
N: 1756 infants
Statistical Analyses: Logistic
regression
Covariates: sex, parental atopy
(yes/no), maternal education, siblings
(y/n), environmental tobacco smoke
_atJiome_J^/n);_use_of_2as_^oi^cookin2_
Pollutant: PM2.6
Mean (SD): PM2.6 mass: 13.4
PM2.6 absorb. 1.77 * 10-5/m
Percentiles: PM2.6 mass:
10th: 12.2
25th: 12.5
50th(Median): 13.1
75th: 14.0
90th: 14.9
PM2.6 absorbance:
PM Increment: PM2.6 mass: 1.5/yglm3
PM2.6 absorb. 0.4 * 10-5/m (IQR)
RR Estimate [Lower CI, Upper CI]
Wheeze (PM2.5 mass)
Age of 1 yr: All: 0.91 (0.76-1.09)
Males: 0.91 (0.72-1.16)
Females: 0.94 (0.70-1.27)
Age of 2 years: All: 0.96 (0.83-1.12)
Males: 0.93 (0.76-1.14)
Females: 1.04 (0.83-1.30)
July 2009
E-538
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
(yfn), home dampness (y/n), indoor
moulds (y/n), keeping of dogs (y/n)
and cats (y/n) study (GINI or LISA)
Dose-response Investigated? No
10th: 1.47 * 10-5
25th: 1.54* 10-5
50th(Median): 1.70 * 10-5
75th: 1.88 *10-5
90th: 2.13*10-5
Range (Min, Max):
PM2.6 mass: 11.9, 21.9
PM2.6 absorbance:
1.38 to 4.39 * 10-5
PM2.6 mass:
PM2.6 absorbance: 1/m
PM Component: PM2.B mass
PM2.E absorbance (as a marker of
diesel soot)
Monitoring Stations: 40
Copollutant (correlation):
NO2: r - 0.99
PM2.6 absorbance and NO2: r - 0.95
PM2.6 mass and PM2.6 absorbance:
r - 0.96
Cough W/O Infection (PM2.B mass)
Age of 1 yr: All: 1.34 (1.11-1.61)
Males: 1.43 (1.14-1.80)
Females: 1.19(0.84-1.70)
Dry Cough At Night (PM2.B mass)
Age of 1 yr: All: 1.31 (1.07-1.60)
Males: 1.39 (1.08-1.78)
Females: 1.17 (0.81-1.68)
Age of 2 years: All: 1.20 (1.02-1.42)
Males: 1.25 (1.01-1.55)
Females: 1.13(0.86-1.48)
Bronchitis (PM2.B mass)
Age of 1 yr: All: 0.98 (0.80-1.20)
Males: 0.97 (0.76-1.25)
Females: 0.98 (0.68-1.41)
Age of 2 years: All: 0.92 (0.78-1.09)
Males: 0.92 (0.74-1.14)
Females: 0.91 (0.68-1.21)
Resp Infections (PM2.B mass)
Age of 1 yr: All: 1.04 (0.91-1.19)
Males: 1.04 (0.87-1.25)
Females: 1.06 (0.87-1.31)
Age of 2 years: All: 0.98 (0.80-1.20)
Males: 0.99 (0.74-1.31): Females: 0.98 (0.73-1.31)
Sneezing/Runny Nose (PM2.5 mass)
Age of 1 yr: All: 1.01 (0.85-1.20)
Males: 0.97 (0.77-1.24)
Females: 1.08 (0.84-1.41)
Age of 2 years: All: 0.96 (0.82-1.12)
Males: 0.91 (0.73-1.12)
Females: 1.04 (0.83-1.31)
Wheeze (PM2.B absorbance)
Age of 1 yr: All: 0.93 (0.78-1.12)
Males: 0.91 (0.71-1.15)
Females: 1.01 (0.74-1.37)
Age of 2 years: All: 0.98 (0.84-1.14)
Males: 0.92 (0.75-1.13)
Females: 1.07 (0.85-1.36)
Cough W/0 Infection (PM2.B absorbance)
Age of 1 yr: All: 1.32 (1.10-1.59)
Males: 1.38 (1.11-1.71)
Females: 1.25 (0.87-1.78)
Dry Cough At Night (PM2 e absorbance)
Age of 1 yr: All: 1.27 (1.04-1.55)
Males: 1.31 (1.04-1.67)
July 2009
E-539
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Females: 1.16(0.79-1.71)
Age of 2 years: All: 1.16 (0.98-1.37)
Males: 1.17 (0.95-1.44)
Females: 1.12(0.84-1.48)
Bronchitis (PM2.6 absorbance)
Age of 1 yr: All: 0.99 (0.81-1.22)
Males: 1.00 (0.78-1.27)
Females: 0.94 (0.63-1.39)
Age of 2 years: All: 0.94 (0.79-1.12)
Males: 0.91 (0.72-1.13)
Females: 0.95 (0.71-1.28)
Resp Infections (PM2.5 absorbance)
Age of 1 yr: All: 1.03 (0.90-1.18)
Males: 1.03 (0.86-1.23)
Females: 1.05 (0.85-1.30)
Age of 2 years: All: 0.99 (0.80-1.22)
Males: 0.96 (0.73-1.26)
Females: 1.04 (0.75-1.43)
Sneezing/Runny Nose (PM2.5 absorbance)
Age of 1 yr: All: 0.95 (0.79-1.14)
Males: 0.90 (0.70-1.16)
Females: 1.06 (0.80-1.39)
Age of 2 years: All: 0.92 (0.78-1.09)
Males: 0.83 (0.66-1.05)
Females: 1.06 (0.83-1.34))
Reference: Goss et al. Outcome: Cystic Fibrosis pulmonary
(2004, 055624)	exacerbations, FEVi
Period of Study: 1999- Age Groups: Children and adults
2000	over the age of 6
Location: USA	Study Design: cohort
N: 11484 patients
Statistical Analyses: Logistic
regression, t-tests, Mann-Whitney
tests, Chi-squared tests, polytomous
regression, multiple linear regression
Covariates: Age, sex, lung function,
weight, insurance status, pancreatic
insufficiency, airway colonization,
genotype, median household income
by census tract, zipcode.
Dose-response Investigated? No
Statistical Package: STATA, SAS
Pollutant: PM2.6
Averaging Time: annual mean of
24 h averages
Mean (SD): 13.7(4.2)
Percentiles: 25th: 11.8
50th(Median): 13.9
75th: 15.9
Monitoring Stations: 713
PM Increment: 10/yg/m3
Odds Ratio Estimate [Lower CI, Upper CI]:
Odds of having 2 or more pulmonary exacerbations as compared
to 1 or less in 2000
1.21 (1.07-1.33)
Odds of having 1 pulmonary exacerbation as compared to no
exacerbations in 2000
0.70 (0.59-0.98)
Decrease in FEVi 155ml(115-194)
Decrease in FEVi in 2000 after adjusting for FEVi in 1999
24ml(7-40)
July 2009
E-540
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Hertz-
Picciotto et al. (2005,
088678)
Period of Study: May
1994 to March 1999
Location: Teplice and
Prachatice, Czech
Republic
Outcome: Developmental
immunotoxicity as assessed by
neonatal immunophenotypes
Age Groups: Not specified: every
woman who delivered in the two
aforementioned districts were asked
to participate
Study Design: Cohort study
N: 1397 mother-infant pairs
Statistical Analyses: Multiple linear
regression with lymphocyte
percentage as responding variable
and pollutant exposure to 14day
averaging period before the date of
cord blood collection
Covariates: Season, length of labor,
parity, number of previous stillbirths,
medication during delivery, working
status of mother, maternal education,
exposure to active and secondhand
smoke, family history of allergy, self-
reports of workplace exposure to
dust during pregnancy, self-reported
maternal chronic or severe
respiratory diseases during
pregnancy. Ambient temperature and
season were controlled for.
Dose-response Investigated? Yes
Statistical Package: SUDAAN
(version 8)
Pollutant: PM2.6
Averaging Time: 24 h
14 day averages
Mean (SD): Overall 24 h: 24.8
14 day avg:
Teplice: 30.1
Prachatice 19.8
PM Component: PAHs
Monitoring Stations: 2 stations:
Teplice and Prachatice
PM Increment: 25//g|m3
Adjusted for 3-day temperature and season, PM2.6 exposure
during the 14 days before birth was associated with reduced T-
lymphocyte fractions CD4+, CD3+ and an increase in B-
lymphocyte fraction (CD19+).
The associations were not quantitatively reported anywhere else
in the paper other than in Figure 2 and Table 3
Reference: (Hertz-
Picciotto et al., 2007,
135917)
Period of Study: 1994-
98 + follow-ups at up to
4.5 years of age for child
Location: Czech Republic
districts of Teplice and
Prachatice
Outcome: Lower respiratory
illnesses, majority being acute
laryngitis, tracheitis, bronchitis.
ICD10 codes J04 and J20
Age Groups: Birth-4.5 years of
Pollutant: PM2.6
Averaging Time: Used 3, 7,14, 30
and 45 day averages
Mean (SD): daily mean 22.3
(sd 16 for 3 day avg, 11 for 45 day
Study Design: longitudinal follow up avg)
of a stratified random sample of
mother-infant pairs from previous
Pregnancy Outcome Study. Low birth
weight and preterm births sampled at
higher fractions.
N: 1133 children
Statistical Analyses: Generalized
linear longitudinal models, GEE to
adjust for within subject correlations,
robust variance estimates were
obtained. Model fit judged using
Akaike Information criterion.
Covariates: age of child, breast
feeding, environmental tobacco
smoke, season, day of week, year of
birth, gender, birth weight, pregnancy
data including age at delivery, length
of gestation, maternal hypertension
and diabetes, infant APGAR score,
maternal work history, demographics,
lifestyle, reproductive and medical
histories, temperature, fuel type,
other children in household
Dose-response Investigated? No
Statistical Package: SUDAAN
version 8
PM Increment: 25//g|m3
RR Estimate [Lower CI, Upper CI]
lag:
Bronchitis, birth-23 months of age
Categorical model
High 30 day avg PM2.6 (greater than 50 //g|m3)
2.26(1.81-2.82)
Medium 30 day avg PM2.6 (between 25 and 50 /yg/m3)
1.48(1.32-1.65)
Continuous model
1.30(1.08-1.58)
Bronchitis, 2-4.5 years of age
Categorical model
High 30 day avg PM2.6 (greater than 50 /yg/m3)
3.66(2.07-6.48)
Medium 30 day avg PM2.6 (between 25 and 50 /yglm3)
1.60(1.41-1.82)
Continuous model
1.23(0.94-1.62)
Notes: Results of other averaging periods shown in plots.
July 2009
E-541
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Hogervorst et
al„ 2006,156559)
Period of Study: NR
Location: Maastricht, the
Netherlands (six schools
selected)
Outcome:
Decreased lung function
Age Groups: 8-13 years old
Study Design: Multivariate linear
regression (enter method) analysis
N: 342 children
Statistical Analyses: ANOVA, Chi
square
Covariates: Independent variables:
Age, height, gender, smoking at home
by parents, pets, use of ventilation
hoods during cooking, presence of
unvented geysers, tapestry in the
home, indoor/outdoor time, education
level of parents.
Dependent variables: lung function
indices
Dose-response Investigated? No
Pollutant: PM2.6
PM Increment: 10/yg/m3
Averaging Time: Daily
RR Estimate [Lower CI, Upper CI]
Mean (SD): 19.0 (3.2)
lag:
Monitoring Stations: 6
FEV
Copollutant:
3.62(0.50,7.63]
PM10
FVC
TSP
1.80 [-2.10, 5.80]

FEF

5.93 [-2.34,14.89]
Reference: Islam et al.
(2007,090697)
Period of Study: 1993-
2001
Location: 12 communities
in Southern California,
U.S.
Outcome: New onset asthma
Age Groups: 9-10 years
Study Design: cohort
N:2057
Statistical Analyses: Cox
proportional hazard model
Covariates: Community, sex,
race/ethnicity
Season: all
Dose-response Investigated? No
Statistical Package: SAS V 9.1
Lags Considered: 0-2 years
Pollutant: PM2.6
Range (Min, Max):
"Low" PM2.6 Communities
(5.7-8.5)
"High" PM2.6 Communities
(13.7-29.5)
Monitoring Stations: 12
Copollutant: NO2, acid vapor, PM10
and elemental and organic carbon
correlated as a "non-ozone package"
of pollutants with a similar pattern
relative to each other across the 12
communities.
PM Increment: NR
IR Estimate [Lower CI, Upper CI]
Low PM
FVC ~ 90: 19.4 (7.5,50.5)
FVC 90-110: 16.8(7.0, 40.1)
FVC >110: 7.9(2.9,21.9)
FEVi ~ 90: 23.7 (9.4, 59.4)
FEVi 90-110:15.6(6.5, 37.4)
FEVi >110: 6.5(2.3,18.7)
FEF25-75 ~ 90: 21.1 (8.8, 50.5)
FEF2B.7B 90-110:11.9 (4.7, 30.0)
FEF25-75 >110: 6.4(2.3,18.2)
Overall: 14.2(7.0, 28.7)
High PM
FVC ~ 90: 14.2(5.1,39.6)
FVC 90-110: 25.6(11.1,59.2)
FVC >110: 16.7 (6.5,42.9)
FEVi ~ 90: 20.8 (8.0, 54.0)
FEVi 90-110: 23.1 (10.0, 53.7)
FEVi >110: 18.8(7.5,47.3)
FEF25-75~ 90: 23.8(10.2,55.6)
FEF2B7B 90-110: 23.9 (9.9, 57.7)
FEF25-75 >110:15.9(6.3,40.5)
Overall: 18.4(9.4, 35.9)
July 2009
E-542
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Karr et al.
(2007,090719)
Period of Study: 1995 to
2000
Location: South Coast Air
Basin of southern
California
Outcome: Bronchioloitis
Study Design: Case-control. Cases
included subjects with a record of a
single hospitalization with a dis-
charge diagnosis of acute bronchio-
litis.10 controls per case were
matched on birth date and gesta-
tional age.
N: 18,595 cases
169,472 controls
Statistical Analyses: Conditional
logistic regression to estimate
relative risk of hospitalization for
bronchiolitis.
Covariates: Confounders included in
the model were: gender, parity,
chronic lung disease, cardiac and
pulmonary anomalies, SES covariates
Age, gestational age, and season of
birth were controlled for by matching
Dose-response Investigated? Yes
Statistical Package: STATA
(Version 8)
Pollutant: PM2.6
Averaging Time: 24 h (lifetime
monthly avg from birth & 30 days
preceding cases hospitalization)
Mean (SD): 25
Percentiles: 25th: 19
50th(Median): 23
75th: 29
Range (Min, Max): 6 to 111
Monitoring Stations: 17
PM Increment: 10/yg/m3
RR Estimate [Lower CI, Upper CI]
Sub-chronic and chronic exposure: OR - 1.09 (1.04-1.14)
Adjusted for adjusted: Sub-chronic OR - 1.10(1.04,1.16)
Chronic OR - 1.09 (1.03-1.15)
Adjusted for CO and NO2: Sub-chronic OR - 1.14 (1.07,1.21)
Chronic OR - 1.12(1.06,1.20)
Adjusted for O3, CO, and NO2: Chronic OR - 1.15 (1.08,1.22)
Sub-chronic OR - 1.13(1.06,1.21)
Reference: (Kim et al.,
2004, 087383)
Period of Study: Mar-
June (spring) 2001
Sep-Nov (fall) 2001
Location: Alameda
County, CA
Outcome: Asthma, bronchitis
Age Groups: Children (grades 3-5)
Study Design: Cross-sectional
N: 1109 children, 871 (long term
resident children), 462 (long term
related females), 403 (long term
related males)
Statistical Analyses: 2-stage
multiple logistic regression model
Covariates: respiratory illness before
age of 2, household mold/moisture,
pests, maternal history of asthma
(for asthma) Season: spring and fall
Dose-response Investigated? Yes
Statistical Package: SAS 8.2
Pollutant: PM2.6
Averaging Time: 10 weeks
Mean (SD): Study Avg 12
Monitoring Stations: 10
Copollutant (correlation): r2 is
approximately 0.9 for all
copollutants-Black Carbon (BC),
PM10, NOx, NO2, NO (NOX-NO2)
PM Increment: 0.7 (IQR)
OR Estimate [Lower CI, Upper CI]:
Bronchitis
All subjects: 1.02(1.00,1.08]
LTR subjects: 1.03 [1.01,1.08]
LTR females: 1.04(1.02,1.05]
LTR males: 1.02(0.99,1.05]
Asthma
All subjects: 1.00 [0.96,1.12]
LTR subjects: 1.01 [0.97,1.06]
LTR females: 1.06 [0.99,1.15]
LTR males: 0.99(0.95,1.04]
Asthma excluding outlier school having a larger proportion of
Hispanics
All subjects: 1.04[0.96,1.12]
LTR subjects: 1.03 [0.94,1.13]
LTR females: 1.03 [0.91,1.17]
LTR males: 1.03[0.94,1.18]
July 2009
E-543
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Leonardi et al.
(2000,010272)
Period of Study: 1996
Location: 17 cities of
Central Europe (Bulgaria,
Czech Republic, Hungary,
Poland, Romania,
Slovakia)
Outcome: Immune biomarkers
Age Groups: 9-11
Study Design: Cross-sectional
N: 366 school children
Statistical Analyses: Linear
regression
Covariates: Age, gender, parental
smoking, laboratory of analysis,
recent respiratory illness
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PM2.6
Averaging Time: annual PM2.6
Mean (SD): PMz.b: 46 (10)
Range (Min, Max):
PM2.6: (29, 67)
5th, median, 81 95th percentile
PM2.6: 29, 44, 67
% Change (Lower CI, Upper CI)
p-value
PM2.6
Neutrophils -10 (-45, 46)
>.20
Total lymphocytes 49 (11,101); .008
B lymphocytes 63 (4,155); .034
Total T lymphocytes 72 (32
123)
<.001
CD4+ 80 (34
143)
<.001
CD8+61 (17,119); .003
CD4/CD8 16 (-17, 62)
>.20
NK 63 (3,158); .035
Total IgG 24 (2, 52); .034
Total IgM-9 (-32, 22)
>.20
Total IgA-1 (-25, 32)
>.20
Total IgE -4 (-61,137)
>.20
Reference: McConnell
(1999,007028)
Period of Study: 1993
Location: Southern
California
Outcome: Bronchitis, chronic cough,
phlegm
Age Groups: Children: 4th, 7th, 81
10th graders
Study Design: Cross-sectional
N: 3676 people
Statistical Analyses: Logistic
regression
Covariates: Age, sex, race, grade,
health insurance
Dose-response Investigated? Yes
Pollutant: PM2.6
Averaging Time: Yearly 2 wk avg
Mean (SD): 15.3
Range (Min, Max): 6.7, 31.5
Copollutant (correlation):
NO2
r - 0.83
0s
r - 0.50
Acid
r - 0.71
Child Respiratory symptoms OR Estimate (Lower CI, Upper
CI)
PM2.5 Increment: 15 //gfm3
Children w) asthma
Bronchitis: 1.4 (0.9, 2.3)
Phlegm: 2.6 (1.2, 5.4)
Cough: 1.3 (0.7, 2.4)
Children wf wheeze, no asthma
Bronchitis: 0.9 (0.6,1.4)
Phlegm: 1.0 (0.6,1.8)
Cough: 1.1 (0.6,1.9)
Children wf no wheeze, no asthma
Bronchitis: 0.5 (0.3,1.0)
Phlegm: 0.8 (0.4,1.5)
Cough: 0.9 (0.6,1.3)
July 2009
E-544
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: McConnell et
al. (2003, 049490)
Period of Study: 1993-
99
Outcome: bronchitic symptoms
Age Groups: 9-19
Study Design: communities selected
on basis of historic levels of criteria
Location: 12 Southern CA pollutants and low residential
communities	mobility.
N: 475 children
Statistical Analyses: 3 stage
regression combined to give a logistic
mixed effects model
Covariates: sex, ethnicity, allergies
history, asthma history, SES,
insurance status, current wheeze,
current exposure to ETS, personal
smoking status, participation in team
sports, in utero tobacco exposure
through maternal smoking, family
history of asthma, amount of time
routinely spent outside by child
during 2-6 pm.
Dose-response Investigated? No
Statistical Package: SAS Glimmix
macro
Pollutant: PM2.6
Averaging Time: 4 year averages
Mean (SD): 13.8(7.7)
Range (Min, Max): 5.5-28.5
Copollutant (correlation): PMio: r
- 0.79
PM 10-2.5: r - 0.24
Inorganic acid: r - 0.76
Organic Acid: r - 0.58
EC: r - 0.83
0C:r - 0.84
NO?: r - 0.54
0s:r - 0.72
PM Increment: Between community range 23//g|m3
Between community unit 1 //g|m3
Within community 1 //g|m3
OR Estimate [Lower CI, Upper CI]
Between community per range
1.81(1.14-2.88)
Between Community per unit
1.03(1.01-1.05)
Within community per unit
1.09(1.01-1.17)
Reference: McConnell et
al. (2003, 049490)
Period of Study: 1993-
99
Outcome: bronchitic symptoms
Age Groups: 9-19
Study Design: communities selected
on basis of historic levels of criteria
Location: 12 Southern CA pollutants and low residential
communities	mobility.
N: 475 children
Statistical Analyses: 3 stage
regression combined to give a logistic
mixed effects model
Covariates: sex, ethnicity, allergies
history, asthma history, SES,
insurance status, current wheeze,
current exposure to ETS, personal
smoking status, participation in team
sports, in utero tobacco exposure
through maternal smoking, family
history of asthma, amount of time
routinely spent outside by child
during 2-6 pm.
Dose-response Investigated? No
Statistical Package: SAS Glimmix
macro
Pollutant: Elemental Carbon
Averaging Time: 4 year avg
Mean (SD): 0.71(0.41)
Range (Min, Max): 0.1-1.2
Copollutant (correlation): PM2 e: r
- 0.83
PM10: r - 0.71
PM 10-2.5: r - 0.30
Inorganic acid: r - 0.82
Organic Acid: r - 0.66
Organic Carbon: r - 0.88
NO?: r - 0.54
O3: r - 0.68
PM Increment: Between community range 1.1 //g|m3
Between community unit 1 /yglm3
Within community 1 /yglm3
OR Estimate [Lower CI, Upper CI]
Between community per range
1.64(1.06-2.54)
Between Community per unit
1.55(1.05-2.30)
Within community per unit
2.63(0.83-8.33)
July 2009
E-545
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: McConnell et
al. (2003, 049490)
Period of Study: 1993-
99
Outcome: bronchitic symptoms
Age Groups: 9-19
Study Design: communities selected
on basis of historic levels of criteria
Location: 12 Southern CA pollutants and low residential
communities	mobility.
N: 475 children
Statistical Analyses: 3 stage
regression combined to give a logistic
mixed effects model
Covariates: sex, ethnicity, allergies
history, asthma history, SES,
insurance status, current wheeze,
current exposure to ETS, personal
smoking status, participation in team
sports, in utero tobacco exposure
through maternal smoking, family
history of asthma, amount of time
routinely spent outside by child
during 2-6 pm.
Dose-response Investigated? No
Statistical Package: SAS Glimmix
macro
Pollutant: Organic Carbon
Averaging Time: 4 year avg
Mean (SD): 4.5(2.7)
Range (Min, Max): 1.4-11.6
Copollutant (correlation): PM2 e: r
- 0.84
PM10: r - .70
PM 10-2.5: r - 0.27
Inorganic acid: r - 0.83
Organic Acid: r - 0.69
EC: r - 0.88
NO?: r - 0.67
O3: r - 0.81
PM Increment: Between community range 10.2//g|m3
Between community unit 1 //g|m3
Within community 1 //g|m3
OR Estimate [Lower CI, Upper CI]
Between community per range
1.74(0.89-3.4)
Between Community per unit
1.06(0.99-1.13)
Within community per unit
1.41(1.12-1.78)
Reference: McConnell, et Outcome:
al. (2006,180226)
Period of Study: 1996-
1999
Location: 12 Southern
California communities
Prevalence of bronchitic symptoms
(yearly).
Age Groups: 10-15-years-old
Study Design: longitudinal cohort
N: 475 asthmatic children
Statistical Analyses: Multilevel
logistic mixed effects models.
Covariates: age, second-hand smoke
personal smoking history
sex, race.
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PM2.6
Averaging Time: 365 days
Percentiles: Community by year
(n - 48 - 12 communities ~ 4
years)
25th: NR
50th(Median): 3.4
75th: NR
Range (Min, Max):
Community by year (n - 48 - 12
communities ~ 4 years):
(0.89,8.7)
Monitoring Stations: 12
Copollutant:
0s
NO?
EC
0C
Acid vapor (acetic and formic acid)
PM Increment: 3.4 |Jg|m3
OR Estimate [Lower CI, Upper CI]
PM2.6
Dog (n - 292): 1.56(1.15: 2.12]
No dog (n - 183): 1.03(0.71: 1.49]
PM2.6*Dog interaction p-value: 0.06
Cat (n - 202): 1.30 [0.90:1.88]
No Cat (n - 273): 1.36 [0.99:1.83]
PM2.6*Cat interaction p-value: 0.87
Neither pet (n - 112): 1.11 [0.71:1.74]
Cat only (n - 71): 0.85 [0.46:1.57]
Dog only (n - 161): 1.53 [1.04: 2.25]
Both pets (n - 131): 1.58 [1.02: 2.46]
Results suggest that dog ownership, a source of residential
exposure to endotoxin, may worsen the severity of respiratory
symptoms from exposure to air pollutants in asthmatic children.
Although PM2.6 was associated at a statistically significant level
with ownership of both cats and dogs, it appears that dog
ownership (with or without a cat) specifically worsens the
association between PM2.6 and respiratory symptoms in
asthmatic children.
July 2009
E-546
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Meng et al„
2007, 093275)
Outcome: Poorly controlled asthma
vs. controlled asthma
Period of Study:
November 2000 and
September 2001
ICD9NR
Age Groups: 18-64, 65 +
Location: Los Angeles and studY Design: Long-term exposure
San Diego counties	stuc'V
comparison of cases and controls
N: 1,609 adults (represented
individuals age 18+ who reported
ever having been diagnosed as having
asthma by a physician and had their
address successfully geocoded)
Statistical Analyses: Logistic
regression to evaluate associations
between TD (traffic density) and
annual avg air pollution
concentrations and poorly controlled
asthma. Used sample weights that
adjusted for unequal probabilities of
selection into the CHIS sample.
Covariates: Age, sex, race/ethnicity,
family federal poverty level, county,
insurance status, delay in care for
asthma, taking medications, smoking
behavior, self-reported health status,
employment, physical activity
Dose-response Investigated? yes
Pollutant: PM2.6
Averaging Time: 24-hs
Copollutant (correlation):
0s: r - -0.76
NO?: r - 0.87
PMio: r - 0.84
CO: r - 0.52
TD: r - 0.13
Results for PM2.6 were nonsignificant and not reported
quantitatively.
Reference: Millstein, J et
al. (2004, 088629)
Period of Study: Mar-
Aug, 1995, and Sep, 1995
to Feb, 1996
Data were taken from the
Children's Health Study
Location: Alpine,
Atascadero, Lake
Arrowhead, Lake Elsinore,
Lancaster, Lompoc, Long
Beach, Mira Loma,
Riverside, San Dimas,
Santa Maria, and Upland,
CA
Outcome: Wheezing 81 asthma
medication use (ICD 9 NR)
Age Groups: 4th grade students,
mostly 9 yrs at the time of the study
Study Design: Cohort Study,
stratified into 2 seasonal groups/
N: 2081 enrolled, 2034 provided
parent-completed questionnaire.
Statistical Analyses: Multilevel,
mixed-effects logistic model.
Covariates: Contagious respiratory
disease, ambient airborne pollen and
other allergens, temperature, sex, age
race, allergies, pet cats, carpet in
home, environmental tobacco smoke,
heating fuel, heating system, water
damage in home, education level of
questionnaire signer, physician
diagnosed asthma.
Season: Mar-Aug, 1995, and Sep,
1995 to Feb, 1996
Statistical Package: GLIMMIX SAS
8.00 macro for generalized linear
mixed models.
Pollutant: PM2.6
Averaging Time: Integrated values
for successive 2-wk periods
PM Component: Nitric acid, formic
acid, acetic acid
Monitoring Stations: 1 central
location in each community
Copollutant (correlation):
O3: r - 0.09
NO?: r - 0.28
PM10: r = 0.33
PM 10-2.5: r - -0.08
PM Increment: IQR: 5.24/yg/m3
Odds Ratio [lower CI, Upper CI]
Annual
PM2.6: 1.04(0.83,1.29]
March-August
PM2.6: 0.91 [0.64,1.30]
Sep-Feb
PM2.6: 1.18(0.89,1.58]
Lags Considered: 14
July 2009
E-547
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Morgenstern
et al. (2007, 090747)
Period of Study: Mar
1999-Jul 2000
Location: Munich,
Germany
Outcome: Asthma, wheezing,
spastic/obstructive bronchitis. Dry
cough at night, respiratory infections,
sneezing, runny/stuffed nose without
a cold.
Age Groups: at 1 yr & at 2 yrs
Study Design: Cohort
N: 3577 children for the prediction
models. Respiratory data available for
3129 children at 1 yr.
Statistical Analyses: Pearson's
correlation coefficient, prediction
error expressed as root mean squared
error (RMSE), multiple logistic
regression with confounding factors,
odds ratios
Covariates: Sex, Parental atopy
(genetic predisposition to allergies),
environmental tobacco smoke at
home, maternal education > or <12
yrs, sibling, gas stove, home
dampness, indoor mold, pets. Since it
was not feasible to measure personal
exposure to NO2, PM2.6, and PM2.6
absorbance, exposure modeling was
used.
Statistical Package: SAS V.8.02
Pollutant: PM2.6
Averaging Time: annual
Mean (SD): 12.8
Percentiles: 25th: 12.5
50th(Median): 12.9
75th: 13.3
Range (Min, Max): 6.8,15.3
Monitoring Stations: 40: traffic,
n - 17 and background, n - 23.
Copollutant (correlation):
PM2.6 absorbance r - 0.49
NO2 r - 0.45
PM Increment: 1.04/yg/m3
Odds Ratio [Lower CI, Upper CI]
Adjusted OR for PM2.6 and: sneezing, runny/stuffed nose during
the first year of life was 1.16 [1.01,1.34]
At age 1 yr
For wheezing 1.01 [0.87,1.18]
For cough without infection 1.05 [0.88,1.25]
For dry cough at nightl .08 [0.86,1.27]
For asthmatic, spastic, or obstructive bronchitis
1.04 [0.90,1.29]
For respiratory infectionl .05 [0.88,1.22]
For sneezing, runny or stuffed nose 1.16 [1.01,1.34]
At age 2 yrs
For wheezing 1.10 [0.96,1.25]
For cough without infection NA, insufficient sample
For dry cough at night 1.03 [0.86,1.19]
For asthmatic, spastic, or obstructive bronchitis
1.05[0.92,1.20]
For respiratory infection 1.09 [0.94,1.07]
For sneezing, runny or stuffed nose 1.19 [1.04,1.36]
Reference: Morgenstern
et al. (2007, 090747)
Period of Study: Ma4
1999-Jul 2000
Location: Munich,
Germany
Outcome: Asthma, wheezing,
spastic/obstructive bronchitis. Dry
cough at night, respiratory infections,
sneezing, runny/stuffed nose without
a cold.
Age Groups: at 1 yr & at 2 yrs
Study Design: Cohort
N: 3577 children for the prediction
models. Respiratory data were
available for 3129 children at 1 yr.
Statistical Analyses: Pearson's
correlation coefficient, prediction
error expressed as root mean squared
error (RMSE), multiple logistic
regression with confounding factors,
odds ratios
Covariates: Sex, Parental atopy
(genetic predisposition to allergies),
environmental tobacco smoke at
home, maternal education > or <12
yrs, sibling, gas stove, home
dampness, indoor mold, pets. Since it
was not feasible to measure personal
exposure to NO2, PM2.6, and PM2.6
absorbance, exposure modeling was
used.
Statistical Package: SAS V.8.02
Pollutant: PM2.6 Absorbance (PM2.6
ab)
Averaging Time: annual
Mean (SD): 1.7 10 -5 m -1,
Percentiles: 25th: 1.610 —5 m -1
50th(Median): 1.7 10 -5 m -1
75th: 1.8 10 -5 m -1
Range (Min, Max):
1.3, 3.2 10 -5 m -1
Unit (i.e. /yg/m3): 10 -5 m -1
Monitoring Stations: 40: traffic,
n - 17 and background, n - 23.
PM Increment: 0.22 x 10 -5
Odds Ratio [Lower CI, Upper CI]
no lag
At age 1 yr
For wheezing 0.97 [0.77,1.23]
For cough without infection 1.16 [0.87,1.54]
For dry cough at nightl .09 [0.78,1.51 ]
For asthmatic, spastic, or obstructive bronchitis
1.14[0.88,1.48]
For respiratory infectionsl .03 [0.86,1.24]
For sneezing, runny or stuffed nose 1.30 [1.03,1.65]
At age 2 yrs
For wheezing 1.09 [0.90,1.33]
For cough without infection NR insufficient data
For dry cough at nightl. 18 [0.93,1.50]
For asthmatic, spastic, or obstructive bronchitis
0.85 [0.30, 2.34]
For respiratory infectionsl .05 [0.79,1.39]
For sneezing, runny or stuffed nose
1.27(1.04,1.56]
July 2009
E-548
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Oftedal et al.
(2008, 093202)
Period of Study: 2001-
2002
Location: Oslo, Norway
Outcome: Lung function (PEF,
FEF25%, FEF50%, FEVi, FVC)
Age Groups: 9-10 yrs
Study Design: Cross-sectional
l\l: 1847 children
Statistical Analyses: Linear
regression
Covariates: Height, age, BMI, birth
weight, temperature, maternal
smoking, se
Dose-response Investigated? Yes
Statistical Package: SPSS, STATA,
S-Plus
Lags Considered: 1-3
Pollutant: PM2.6
IQR:
PM2.6 in 1st yr of life: 6.2
PM2.6 lifetime: 3.6
PM Increment: Per IQR
fB (Lower CI, Upper CI)
PM2.6 in 1st yr of life
PEF-76.1 (-122.2 to-30.0)
FEF25%-75.6 (-127.4 to-23.8)
FEF 50% -62.4 (-107.4 to-17.4)
FEVi-12.7 (-28.8, 3.4)
FVC-2.9 (-20.5,14.7)
PM2.E lifetime exposure
PEF-57.7 (-94.4 to-21.1)
FEF25% -51.8 (-93.1 to-10.6)
FEF 50%-48.4 (-84.2 to-12.6)
FEVi-10.4 (-23.2, 2.4)
FVC -3.9 (-17.9,10.1)
Reference: (Parker et al.,
Outcome: Respiratory
Pollutant: PM2.6
Increment: 10//gfm3
2009,192359)
allergy/hayfever


Averaging Time: NR
Odds Ratio (95% CI)
Period of Study: 1999-
Study Design: Cohort


2005
Median: 13.1
Single Pollutant Model, variable N
Covariates: survey year, age, family
IQR: 10.9-15.2
Adjusted: 1.16(1.04-1.30)
Location: US
structure, usual source of care,

health insurance, family income
relative to federal poverty level,
Copollutant (correlation):
Single Pollutant Model, constant N

race/ethnicity
Summer O3: 0.10
Adjusted: 1.23 (1.04-1.46)

Statistical Analysis: logistic
SO2: 0.21
Multi-pollutant Model: 1.29 (1.07-1.56)

regression
NO2: 0.53

Statistical Package: SUDAAN
PMio-2.b: 0.02


Age Groups: 73,198 children aged



3-17 years
PM10: 0.51

July 2009
E-549
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Sekine et al.
(2004, 090762)
Period of Study: 1987-
1994
Location: Nine districts in
the Tokyo, Japan
metropolitan area: Chuo
ward, Ohta ward, Shibuya
ward, Itabashi ward,
Hachioji City, Tachikawa
City, Ome City, Machida
City, Tanashi City
Outcome: pulmonary function tests
Age Groups: 30-59 yrs
Study Design: Cross-sectional and
longitudinal
N: 500 females
Statistical Analyses: Multiple
logistic regression analysis
Pollutant: Suspended PM (SPM)
Averaging Time: measured each
month for three consecutive days (72
Mean (SD): 28.1-63.3
Range (Min, Max):
3.4-140.6
Covariates: group (classification by Copollutant (correlation): NO2
air pollution level), pulmonary
function at initial test, age and height
at the time of the initial test, number
of years investigated, years of
residence in the area, type of heater,
housing structure, and job status.
Dose-response Investigated? No
Statistical Package: SAS
Results of multiple logistic regression analysis for respiratory
symptoms
Persistent cough
Group 3: OR - 1.00
Group 2: OR - 1.02 (0.70-1.48)
Group 1: OR - 1.07 (0.67-1.70)
Persistent phlegm
Group 3: OR - 1.00
Group 2: OR - 1.51 (1.11-2.04)
Group 1: OR - 1.78 (1.26-2.53)
Asthma
Group 3: OR - 1.00
Group 2: OR - 1.99 (0.82-4.83)
Group 1: OR - 2.66 (0.98-7.19)
Wheeze
Group 3: OR - 1.00
Group 2: OR - 1.39 (0.95-2.01)
Group 1: OR - 1.34(0.85-2.11)
Breathlessness
Group 3: OR - 1.00
Group 2: OR - 0.84(0.47-1.50)
Group 1: OR - 2.70(1.48-4.91)
Reference: Sharma et al.
(2004,156974)
Period of Study:
1112002-412003
Location: 3 sections in
Kanpur City, India
1)	Indian Institute of
Technology Kanpur (IITK)
2)	Vikas Nagar (VN)
3)	Juhilal Colony (JC)
Outcome: Lung function
Age Groups: 20-55 years
Study Design: Cohort
N: 91 people
Statistical Analyses: Linear
regression
Covariates: NR
Season: Fall, Winter, spring
Dose-response Investigated? No
Statistical Package:
Microsoft Excel
Lags Considered: 1 d lag & 5d mov
avg
Pollutant: PM2.6
Averaging Time: 24-h
Mean (SD): IITK 158 (22)
VN 85 (30)
JC 59 (9)
PM Component: Lead, Nickel,
Cadmium, Chromium, Iron, Zinc
Benzene soluble fraction (includes
polycyclic aromatic hydrocarbons
[PAHs])
Copollutant (correlation):
~PEF - mean daily deviations in PEF
PM2.b-dPEF: -0.30
PM2.B-PM10: 0.67
PM2.b-PMio (1-day lag): 0.49
PM2.5-PM2.il (1-day lag): 0.88
PM Increment: 1 /yglm3
~PEF (difference or change in peak expiratory flow)
¦0.0297 Umin
July 2009
E-550
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Singh et al„
2003, 052686)
Period of Study: NR
Location: Jaipur, India
Outcome: Lung function (peak
expiratory flow variability)
Age Groups: Medical school-aged
students
Study Design: Cross sectional
N: 313 nonsmoker students
Statistical Analyses: Amplitude %
mean was used as the measure of
PEF variability. Mean value of
amplitude % mean of peak flow
variability were compared for in the
two groups by application of
Student's t-test. The two groups
were: living on campus and
commuters.
Dose-response Investigated? Yes
Pollutant: Respirable su
(RSPM)
Averaging Time: 8 h
Mean (SD): Roadside: 1,666
Campus: 177
Monitoring Stations: 2
PM It appears that no associations between particulates and the
outcome of interest were calculated and reported in this study
Reference: (Solomon et
al„ 2003, 087441)
Period of Study: 1966 to
1997
Location: United
Kingdom: Northern
England, North-West
Midlands, and Wales.
Outcome: Cardio respiratory morbidity Pollutant: Black Smoke
Age Groups: 45 yrs and older	Averaging Time: Annual
Study Design: Cross-sectional
N: 1,166 women
Statistical Analyses: Prevalence
ratios were reported for ischemic heart
disease, asthma, productive cough,
wheeze, and use of an inhaler for
asthma or other breathing problems.
Covariates: Smoked, passive smoking
in childhood, tenancy, SES, worked in
industry with respiratory hazards,
childhood admission to hospital for
chest problem, diabetes, BMI were all
controlled for as potential
confounders.
Dose-response Investigated? yes
Statistical Package: STATA
RR Estimate [Lower CI, Upper CI]
The findings provide no indication that prolonged residence in
places that have had relatively high levels of particulate air
pollution causes an important increase in cardio respiratory
morbidity.
Prevalence ratios are based on high vs. low pollution with low as
referent.
Particulate pollution in place of residence:
Rr - 1.0 (0.7-1.4) for ischemic heart disease:
Rr - 0.7 (0.5-1.0) for asthma
Rr - 1.0 (0.7 -1.5) for productive cough
Reference: Suglia et al.
(2008,157027)
Period of Study: March
1986-October 1992
Location: Boston, MA
Outcome: lung function
Age Groups: 18-42
Study Design: Prospective cohort
N: 272 women of childbearing age
Statistical Analyses: Linear
regression
Covariates: Height, age, weight,
race/ethnicity, year, education
Dose-response Investigated? yes-
tertiles of exposure
Statistical Package: SAS v. 9.0
Pollutant: Black Carbon (BC)
Averaging Time: Annual
Mean (SD): 0.62 (0.15)
PM Increment: 0.22/yg/m3 (IQR)
Effect Estimate [Lower CI, Upper CI]
FEVi:-1.08 (-2.5, 0.3)
FVC:-0.62 (-1.9, 0.6)
FEF2b-7b%: -2.97 (-5.8 to -0.2)
Current Smokers:
FEVi: 0.62 (-2.1, 3.4)
FVC: 0.64 (-2.0, 3.3)
FEF2B-7B*: -2.63 (-3.7, 8.9)
Former Smokers:
FEVi:-4.40 (-7.8 to-1.0)
FVC:-3.11 (-6.1 to-0.2)
FEF2B7B%: -8.78 (-14.7 to -2.9)
Nonsmokers:
FEVi:-0.98 (-2.9, 0.9)
FVC:-0.32 (-2.0,1.4)
FEF2b-7b%: -4.39 (-8.1 to -0.6)
Exposure-response relationship presented graphically in Figure 1:
the highest BC exposure group had decreases in FEVi, FVC, and
FEF2B-7B% compared with the lowest tertile group, although these
differences were not statistically significant.
July 2009
E-551
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Sunyeret al„
2006,089771)
Period of Study: initial
selection: 1991-1993,
follow-up June 2000-
December 2001
Location: 21 centers in
10 European countries
Outcome: Chronic bronchitis
Age Groups: Mean age (range)
Males- 42.62 (38.12-45.62)
Females- 42.57 (39.92-45.69)
Study Design: Hierarchical models
N:6924
Statistical Analyses: General
additive models (GAM)
Covariates: Smoking, age at end of
education, occupational group,
occupational exposures, respiratory
infections during childhood, rhinitis,
asthma, traffic intensity at household
level.
Statistical Package: STATA-8
Pollutant: PM2.6
Averaging Time: 18 months
Mean (SD): 3.7-44.9
Copollutants: NO2, SO2
PM Increment: NR
Odds ratio [Lower CI, Upper CI]
Chronic phlegm prevalence at follow up
Males: 0.97 [0.70,1.35]
Reference: Zhang et al.
(2002,034814)
Period of Study: 1993-
1996
Location: 4 Chinese cities
(urban and suburban
location in each city):
Guangzhou, Wuhan,
Lanzhou, Chongqing
Outcome: Interview-self reports of
symptoms: Wheeze (ever wheezy when
having a cold)
asthma (diagnosis by doctor)
bronchitis (diagnosis by doctor)
hospitalization due to respiratory
disease (ever)
persistent cough (coughed for at least
1 month per year with or apart from
colds)
persistent phlegm (brought up phlegm
or mucus from the chest for at least 1
month per year with or apart from
colds).
Age Groups: Elementary school
students
age range: 5.4-16.2
Study Design: Cross-sectional
N: 7,557 returned questionnaires
7,392 included in first stage of
analysis
Statistical Analyses: 2-stage
regression approach:
Calculated odds ratios and 95% CIs of
respiratory outcomes and covariates
Second stage consisted of variance-
weighted linear regressions that
examined associations between
district-specific adjusted prevalence
rates and district-specific ambient
levels of each pollutant.
Covariates: Age, gender, breast-fed,
house type, number of rooms, sleeping
in own or shared room, sleeping in own
or shared bed, home coal use,
ventilation device used, homes
smokiness during cooking, eye
irritation during cooking, parental
smoking, mother's education level,
mother's occupation, father's
occupation, questionnaire respondent,
year of questionnaire administration,
season of questionnaire
administration, parental asthma
prevalence.
Pollutant: PM2.6
Averaging Time: 2 years
Mean (SD): 92 (31)
Percentiles:
25th: NR
50th(Median): NR
75th: NR
IQR: 39
Range (Min, Max):
Gives range (max.-min.):
PM2.B-98
Monitoring Stations: 2 types:
municipal monitoring stations over a
period of 4 years (1993-1996)
schoolyards of participating children
over a period of 2 years (1995-
1996)
PM Increment: Interquartile range corresponded to 1 unit of
change.
RR Estimate [Lower CI, Upper CI]
lag:
No association between PM2.6 and any type of respiratory
morbidity.
No between or within city association between PM2.6 and any
type of respiratory morbidity.
When scaled to an increment of 50 /yglm3 increase in PM2.6,
association (ORs) between respiratory outcome and PM2.6 was:
Wheeze: 1.06
Asthma: 1.29
Bronchitis: 1.68
Hospitalization: 1.08
Persistent cough: 1.24
Persistent phlegm: 3.09
'All units expressed in //gfm3 unless otherwise specified.
July 2009
E-552
DRAFT-DO NOT CITE OR QUOTE

-------
Table E-25. Long-term exposure - respiratory morbidity outcomes - other PM size fractions.
Study
Design & Methods
Concentrations
Effect Estimates (95% CI)
Reference: El-Zein et al. (2007, 093043) ED Admissions
Period of Study: 2000-2004
Location: Beirut, Lebanon
Outcome: Acute respiratory symptoms:
asthma, URTI, pneumonia, bronchitis
Age Groups: < 17
Study Design: Ecological (natural
experiment comparing admissions before
and after ban on diesel fuel)
N: 5 hospitals, 7573 admissions Oct-Feb,
4303 admissions Oct-Dec
Statistical Analyses: t-test, Poisson
regression
Covariates: Month of Year, temperature,
humidity, orthogonalized rainfall
Season: Oct-Dec (excluding flu season)
and Oct-Feb
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1-2 years before the
ban compared to 1-2 years after the ban
Pollutant: PM from diesel
Range (Min, Max): NR
PM Component: NR
Monitoring Stations: 1
Notes: Did not look at specific exposure
data
looked at outcome with respect to a
timeline that plotted admissions before
and after a ban on diesel fuel.
Copollutant: NR
PM Increment: NA
p (p-value):
2 years pre-ban vs. 2 years post-ban
Oct to Feb
All Resp: 0.128 (0.32)
Asthma: -0.176 (0.16)
Bronchitis: 0.505 (0.02)
Pneumonia: 0.287 (0.17)
URTI:-0.265 (0.41)
Oct to Dec
All Resp:-0.022 (0.87)
Asthma: -0.21 (0.07)
Bronchitis: 0.2 (0.35)
Pneumonia: -0.065 (0.78)
URTI:-0.628 (0.05)
2 years pre-ban vs. 1 year post-ban
Oct-Feb
All Resp:-0.093 (0.45)
Asthma: -0.208 (0.05)
Bronchitis: 0.286 (0.32)
Pneumonia: -0.07 (0.76)
URTI:-0.715 (0.11)
Oct to Dec
All Resp:-0.147 (0.02)
Asthma: -0.147 (0.00)
Bronchitis: -0.011 (0.96)
Pneumonia: -0.214 (0.15)
URTI:-0.885 (0.06)
1 years pre-ban vs. 1 year post-ban
Oct-Feb
All Resp:-0.165 (0.04)
Asthma: -0.212 (0.09)
Bronchitis: 0.059 (0.85)
Pneumonia: -0.034 (0.84)
URTI:-1.023 (0.00)
Oct to Dec
All Resp:-0.17 (0.00)
Asthma: -0.131 (0.00)
Bronchitis: -0.145 (0.001)
Pneumonia: -0.168 (0.12)
URTI:-1.036 (0.00)
July 2009
E-553
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kasamatsu et al. (2006,
156627)
Period of Study: 2001-2002
Location: Shenyang, China
Outcome: FVC, FEVi, PEF, FEFjb
Age Groups: School Children aged 8-10
Study Design: Children in three schools
in three types of areas (commercial city
area, residential city area, residential
suburban area) invited to participate
N: 322 children participated, 244 have
complete data.
Statistical Analyses: Genralized
estimating equations
Covariates: age, height,
Dose-response Investigated? no
Statistical Package: SAS
Lags: Considered: previous quarter.
Pollutant: PM7
Averaging Time: avg of 4 separate 2-7
consecutive day measurements within
each designated measurement month of
the quarter
Mean (SD): School A
712001 86.4(14.2)
1012001 114.1(35.1)
112002 118.2(28.2)
412002 182.7(102.1)
School B
712001 90.1(8.3)
1012001 161.5(45.7)
112002 118.8(28.2)
412002 152.0(31.3)
School C
712001 78.1(16.9)
1012001 131.2(29.6)
112002 142.2(37.6)
412002 173.6(121.5)
PM Component: mainly pollutants
associated with coal heating
Monitoring Stations: 1 at each location
PM Increment: 63.0/yg/m3
Mean change of pulmonary function
value [Lower CI, Upper CI] at lag 0
Boys
FVC -0.095( 0.170,-0.019)
FEV1 -0.088(0.158,-0.019)
PEF -0.170( 0.365,0.032)
FEFjb-0.063(0.183,0.050)
Girls
FVC -0.082( 0.145,-0.019)
FEV1 -0.069(0.126,-0.006)
PEF 0.095(-0.095,0.290)
FEFjb-0.032(0.151,0.082)
Mean change of pulmonary function value
[Lower CI, Upper CI] at lag 1 (previous
quarter)
Boys
FVC-0.145(-0.189,-0.095)
FEV1 -0.095(0.139,-0.057)
PEF -0.082( 0.208,0.050)
FEFjb 0.013(-0.063,0.088)
Girls
FVC-0.126(-0.170,-0.088)
FEV1 -0.101(0.139,-0.063)
PEF -0.101( 0.227,0.025)
FEFjb-0.057(0.132,0.019)
July 2009
E-554
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kasamatsu et al.(2006,
156627)
Period of Study: 2001-2002
Location: Shenyang, China
Outcome: FVC, FEVi, PEF, FEFjb
Age Groups: School Children aged 8-10
Study Design: Children in three schools
in three types of areas (commercial city
area, residential city area, residential
suburban area) invited to participate
N: 322 children participated, 244 have
complete data.
Statistical Analyses: Genralized
estimating equations
Covariates: age, height,
Dose-response Investigated? no
Statistical Package: SAS
Lags: Considered: previous quarter.
Pollutant: PM2.1
Averaging Time: avg of 4 separate 2-7
consecutive day measurements within
each designated measurement month of
the quarter
Mean (SD):
School A
712001 47.6(6.4)
1012001 54.2(20.5)
112002 68.9(15.8)
412002 115.8(76.7)
School B
712001 45.6(6.5)
1012001 74.4(27.1)
112002 63.3(17.9)
412002 96.3(27.6)
School C
712001 42.5(9.5)
1012001 59.7(13.1)
112002 76.4(22.1)
412002 123.0(100.9)
PM Component: mainly pollutants
associated with coal heating
Monitoring Stations: 1 at each location
PM Increment: 42.1 /yglm3
Mean change of pulmonary function value
[Lower CI, Upper CI] at lag 0
Boys
FVC -0.126( 0.181,-0.076)
FEV1 -0.122(0.173,-0.076)
PEF -0.164( 0.303, 0.025)
FEFjb-0.046(0.131,0.038)
Girls
FVC-0.110(-0.156,-0.067)
FEV1 -0.101(0147,-0.059)
PEF 0.008(-0.131,0.147)
FEFjb-0.055(0.139,0.030)
Mean change of pulmonary function value
[Lower CI, Upper CI] at lag 1 (previous
quarter)
Boys
FVC -0.099( 0.145,-0.053)
FEV1 -0.059(0.106,-0.020)
PEF -0.040( 0.158,0.086)
FEFjb 0.026(-0.046,0.092)
Girls
FVC -0.086( 0.125,-0.046)
FEV1 -0.066(0.106,-0.026)
PEF -0.079( 0.198,0.040)
FEFjb-0.033(0.106,0.040)
'All units expressed in //gfm3 unless otherwise specified.
E.6. Long-Term Exposure and Cancer
Table E-26. Long-term exposure - cancer outcomes - PM10.
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Abbey et al., 1999, 047559) Outcome (ICD9): Lung Cancer Mortality
(162)
Period of Study: 1977-1992
Location: California
Age Groups: 27-95 at baseline
Study Design: Cohort (AHSMOG)
N: 6,338 nonsmoking CA Seventh-Day
Adventists
Statistical Analyses: time-dependent,
gender-specific, Cox proportional hazards
regression models
Covariates: age, smoking, education,
occupation, BMI
Pollutant: PMio
Averaging Time: monthly estimates
from 1966-1992
Mean (SD): 51.24 (16.63)
Percentiles: IQR: 24.08
Range (Min, Max): 0, 83.9
Correlations:
S04: r - 0.68)
SO2: r - 0.31
O3: r - 0.77
NO?: r - 0.56
Lag: 3 years
PM Increment: 24.08 (IQR)
RR, males: 3.36(1.57,7.19]
RR, females: 1.33 [0.60, 2.96]
PM10 above 10O/jg/m3 (days per year)
IQR: 43 days/year
Males: 2.38 (1.42, 3.97)
Females: 1.08 (0.55, 2.13)
July 2009
E-555
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Beeson et al. (1998,
048890)
Period of Study: 1977-1992
Location: California
Reference: Binkova et al. (2007,
156273)
Period of
Location:
Reference: (Liu et al., 2009,190292)
Period of Study: 1995-2005
Location: Taiwan
Outcome (ICD9: Lung Cancer Mortality
(ICD0-1: 162, ICDO-2: C34.0-C34.9)
Age Groups: 27-95 at baseline
Study Design: Cohort (AHSM0G)
N: 6,338 nonsmoking CA Seventh-Day
Adventists (non-Hispanic white)
Statistical Analyses: time-dependent,
gender-specific, Cox proportional hazards
regression models
Covariates: Smoking, Education, Age,
Alcohol
Statistical Package: SAS
Lags Considered: 3 yr
Outcome: Total DNA adducts (bulky
aromatic PAH-DNA adducts and ...
N: 53 occupational^ exposed policemen
and 52 control policemen
Statistical Analyses: Multivariate
logistic regression, Mann-Whitney u-test
Covariates: Smoking. Vitamin C,
polymorphisms of XPD repair gene in
exon 23 and 6 and GSTM 1 and XRCC1
genes
Season: Winter
Outcome: Bladder Cancer Mortality (ICD-
9188)
Age Groups: 50-69
Study Design: case-crossover
Statistical Analysis: Multiple Logistic
Regression
Statistical Package: NR
Covariates: none
Dose-response Investigated? No
Pollutant: PMio
Averaging Time: averaged monthly
estimates from 1966-1992
Mean (SD): 51 (16.52)
Percentiles: IQR: 24
Range (Min, Max): 0, 84
Pollutant: PMio
Range (Min, Max): 32-55
Monitoring Stations: 2 (and personal
monitors)
Pollutant: PMio
Averaging Time: annual mean of 24h
avg
Tertiles (median):
T1:n52.80
T2: 53.04-71.72
T3: 72.24-90.29
Copollutant: 0:i, CO, NO2, SO2
Copollutant (correlation): NR
Monitoring Sattions: 64
PM Increment: 24 (IQR)
RR, males: 5.21 [1.94,13.99]
RR, females: Positive, but not
statistically significant
No relationship between short term
exposure to C-PAHs evaluated by
personal monitors and DNA adduct level.
Genetic damage was observed in city
policemen working in winter outdoors in
the Prague downtown area
they had slightly elevated aromatic DNA
adduct levels, which was statistically
significant for a distinct DNA adduct spot
that could originate from ambient
exposure to B[n]P.
Total PAH-DNA adducts: p - 0.065
Exposed: 0.92 ± 0.28 adducts/108
nucleotids
Control: 0.82 ± 0.23 adducts/108
nucleotids
B[D]P-like adducts:
Exposed: 0.122 ± 0.36 adducts/108
nucleotids
Control: 0.099 ± 0.035 adducts/108
nucleotids
Multiple regression "like" B[n]P-DI\IA
adduct for air pollution exposure group: B
- 0.016, p - 0.01
Increment:
Odds Ratio (Min CI, Max CI)
Lag
T1 vs. T1:1.00 (ref)
T2 vs. T1: 1.08 (0.83-1.41)
T3 vs. T1: 1.39 (1.06-1.83)
P for trend - .020
Study: February 6-20, 2001 Age Groups: 22-50 yrs
Prague, Czech Republic	Study Design: Case Control
July 2009
E-556
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Pope CA et al., 2002,
024689)
Period of Study: 1982-1998
Location: 50 US states, District of
Columbia, and Puerto Rico
Outcome (ICD9): Lung cancer mortality
(162)
Age Groups: Ages > 30 years Study
Design: Longitudinal cohort (Cancer
Prevention Study II)
N: 1.2 million people
Statistical Analyses: Cox proportional
hazard, generalized additive
Covariates: Age, sex, race, education,
smoking status, marital status,
occupational exposure, diet, body-mass
index, alcohol consumption
Pollutant: PMio
Mean (SD): 1982-1998: 28.8(5.9)
Effect estimates: Effect estimates were
recorded in Figure 5 and not presented
quantitatively anywhere else
Reference: Sram et al, (2007,188457) Outcome: Chromosomal aberrations
Period of Study: January and March of
2004
Location: Prague, Czech Republic
Study Design: Panel
Covariates: urinary cotinine, plasma
levels of vitamins A, E and C, folate, total
cholesterol, HDL and LDL cholesterols,
and triglycerides
Statistical Analysis: bivariate
correlations, ANOVA, Mann-Whitney,
Kruskal-Wallis and Spearman rank
correlation
Statistical Package: STATISTICA
Age Groups: 61 city policemen, aged 34
± 8 years, spending 8+ hours outdoors
Pollutant: PMio
Averaging Time: NR
Mean (SD) Unit:
January: 55.6/yg/m3
March: 36.4/yg/m3
Copollutant: PM2.6
Results not given by PM increment.
Reference: Sram et al, (2007,188457) Outcome: Chromosomal aberrations
Period of Study: January and March of
2004
Location: Prague, Czech Republic
Study Design: Panel
Covariates: urinary cotinine, plasma
levels of vitamins A, E and C, folate, total
cholesterol, HDL and LDL cholesterols,
and triglycerides
Statistical Analysis: bivariate
correlations, ANOVA, Mann-Whitney,
Kruskal-Wallis and Spearman rank
correlation
Statistical Package: STATISTICA
Age Groups: 61 city policemen, aged 34
± 8 years, spending 8+ hours outdoors
Pollutant: PM2.6
Averaging Time: NR
Mean (SD) Unit:
January: 44.4/yg/m3
March: 24.8 /yglm3
Copollutant: PMio
Results not given by PM increment.
Reference: (Tarantini et al., 2009,
192010
Period of Study: NR
Location: Brescia, Italy
Outcome: DNA methylation content
estimated by Alu, LINE-1 and/TITOS
analysis
Study Design: Panel
Covariates: age, BMI, smoking, number
of cigarettes/day
Statistical Analysis: mixed effects
models
Statistical Package: NR
Age Groups: 63 male workers between
27 and 55 years, mean age 44.
Pollutant: PMio
Averaging Time: NR
Mean (SD) Unit: NR
Individual Exposure Range: 73.4-1220
/yg/m3
Copollutant (correlation): NR
Difference in DNA Methylation before
and after work exposure, mean (SE)
Alu (%5mC): 0.00 (0.08), p - 0.99
LINE-1 (%5mC): 0.02 (0.11), p - 0.89
ilMOS (%5mC): -0.61 (0.26), p - 0.02
July 2009
E-557
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Vineis et al„ 2006,192089) Outcome: lung cancer
Period of Study: 1990-1999
Location: 10 European countries
Study Design: nested case-control
Covariates: age, sex, country, smoking
status, time since recruitment, education,
BMI, physical activity, intake of fruit,
vegetables, meat, alcohol and energy
Statistical Analysis: conditional logistic
regression models
Statistical Package: NR
Age Groups: 35-74 at recruitment
Pollutant: PMio
Averaging Time: NR
Mean by Country (/vg/m3):
France
lie-de-France
1990-1994: 22.3
1995-1999: 19.9
Northeast France
1990-1994: 30.2
1995-1999: 29.5
Italy
Turin
1990-1994: 73.4
1995-1999: 61.1
Florence
1990-1994: 40.4
1995-1999: 33.3
United Kingdom
Oxford
1990-1994: 29.0
1995-1999: 25.5
Cambridge
1990-1994: NR
1995-1999: 25.4
The Netherlands
Utrecht
1990-1994: 42.8
1995-1999: 40.0
Bilthoven
1990-1994: 39.0
1995-1999: 37.2
Germany
1990-1994: NR
1995-1999: 27.0
Potsdam
1990-1994: 32.0
1995-1999: 28.9
Range (Min, Max): NR
Copollutant: NO2, O3, SO2
Increment: 10 //g/m3
Odds Ratios (Min CI, Max CI) for
increase in lung cancer per increment
increase in PM10
0.91 (0.70-1.18)
July 2009
E-558
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Wei et al„ 2009,192361)
Period of Study: 11/17/2006 -
111312007
Location: Peking, China
Outcome: Urinary 8-OHdG increase
Study Design: Panel
Covariates: NR
Statistical Analysis: analysis of
variance model with autoregressive terms
Statistical Package: SAS
Age Groups: Two nonsmoking security
guards, ages 18 and 20
Pollutant: PM2.6
Averaging Time: 24h
Median: 154.87 /yglm3
IQR: 166.29
Copollutant (correlation): NA
Increment: 166.29/yg/m3
8-OHdG Concentrations, pre and post-
work shift, subjects averaged
Pre-work: 1.83
Post-work: 6.92
Concentration Changes (95%CI) of 8-
OHdG per IQR Increase
Pre-work: 0.256 (0.040, 0.472), p -
0.021
Post-work: 2.370 (0.907, 3.833), p -
0.002
'All units expressed in /yglm3 unless otherwise specified.
Table E-27. Long-term exposure - cancer outcomes - PM2.5 (including PM components/sources).
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Baccarelli et al, (2009,
188183)
Period of Study: 111999-612007
Location: Boston, Massachusetts
Outcome: DNA methylation of LINE-1
and Alu
Study Design: Panel
Covariates: age, BMI, smoking status,
pack-years, statin use, fasting blood
glucose, diabetes mellitus, percent
lymphocytes and neutrophils in
differential blood count, day of the week,
season, temperature
Statistical Analysis: mixed effects
models
Statistical Package: SAS
Age Groups: 719 elderly individuals,
mean age 73.3, range 55-100 years
Pollutant: PM2.6
Averaging Time: NR
Mean (SD) Unit:
4h: 12.2 (7.7)/yg/m3
1d: 10.9 (6.3) /ygfm3
2d: 10.6 (5.2)/ygfm3
3d: 10.4 (4.8)/yg/m3
4d: 10.3 (4.3) /ygfm3
5d: 10.2 (3.9) /ygfm3
6d: 10.3 (3.5) /ygfm3
7d: 10.3 (3.3) /ygfm3
Copollutants: Black carbon, Sulfate
Increment: SD for each lag
Correlation Coefficient (95% CI)
Lag for LINE-1 Methylation
4h: -0.07 (-0.13,-0.01), p - 0.03
1d: -0.09 (-0.16, -0.02), p - 0.008
2d: -0.10 (-0.17,-0.03), p - 0.003
3d: -0.10 (-0.17,-0.04), p - 0.003
4d: -0.10 (-0.16, -0.03), p - 0.004
5d: -0.10 (-0.16, -0.03), p - 0.004
6d: -0.11 (-0.17,-0.04), p - 0.001
7d: -0.13 (-0.19, -0.06), p < 0.001
Correlation Coefficient (95% CI)
Lag for Alu Methylation
4h: 0.03 (-0.03, 0.09), p - 0.28
1d: -0.01 (-0.07,0.05), p - 0.74
2d: -0.01 (-0.07,0.05), p - 0.82
3d:-0.01 (-0.07,0.05), p - 0.78
4d: -0.01 (-0.07,0.05), p - 0.75
5d: -0.01 (-0.07,0.05), p - 0.84
6d: -0.01 (-0.07,0.05), p - 0.74
7d: -0.01 (-0.07,0.05), p - 0.71
Correlation Coefficient (95% CI)
LINE-1 Methylation and moving
averages of pollutant levels
4h: -0.04 (-0.11, 0.03), p - 0.24
7d: -0.11 (-0.18, -0.05), p - 0.001
July 2009
E-559
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Binkova et al. (2007,
156273)
Period of Study: February 6-20, 2001
Location: Prague, Czech Republic
Outcome: Bulky aromatic PAH-DNA
adducts
Age Groups: 22-50 yrs
Study Design: Case Control
N: 53 exposed policemen and 52 control
policemen
Statistical Analyses: Multivariate
logistic regression, Mann-Whitney, Rank-
Sum Litest
Covariates: Smoking. Vitamin C,
polymorphisms of XPD repair gene in
exon 23 and 6 and GSTM 1 gene
Season: Winter
Pollutant: PM2.6
Range (Min, Max): 27-38
c-PAHs: range - 18-22 ng/m3
B[a]P: range - 2.5-3.1 ng/m3
Monitoring Stations: 2
Genetic damage was observed in city
policemen working in winter outdoors in
the Prague downtown area
they had slightly elevated aromatic DNA
adduct levels, which was more
pronounced for a distinct DNA adduct
spot that could originate from ambient
exposure to B[n]P.
Total DNA-adduct level
Exposed: 0.92±0.28 adducts/108
nucleotides
Control: 0.82±0.23 adducts/108
nucleotides
p - 0.065
"Like" B[D]P-derived DNA adducts
Exposed: 0.122 ±0.036
Control: 0.101 ±0.035
p < 0.01
Multiple Regression (exposed vs.
control)
B - 0.016, p - 0.011
Reference: Brunekreef et al, (2009,
191947)
Period of Study: 1987-1996
Location: The Netherlands
Outcome: Air pollution related lung
cancer deaths (ICD-9 162)
Study Design: Case-cohort
Covariates
Individual: sex, age, Quetelet index,
smoking status, passive smoking status,
educational level, occupation,
occupational exposure, marital status,
alcohol use, intake of vegetables, fruits,
energy, saturatured and monounsaturated
fatty acids, trans fatty acids, total fiver,
folic acid and fish
Area-level: Percent of population with
income below the 40th percentile and
above the 80th percentile
Statistical Analysis: Cox proportional
hazards
Statistical Package: Stata, SPSS, R
Age Groups: 120,000 adults aged 55-69
years at enrollment
Pollutant: PM2.6, estimated from PM10
levelsf
Averaging Time: 24hr
50"1 Percentile: 28 /yg/m3
Range (Min, Max): 23-37
Copollutant (correlation):
NO2: 0.75
Black Smoke: 0.84
NO: 0.69
SO2: 0.43
Increment: 10/yg/m3
Relative Risk (95% CI) for
associations between PM2.sand lung
cancer incidence
Case Cohort
Unadjusted: 0.93 (0.71-1.22)
Adjusted: 0.67(0.41-1.10)
Unadjusted Complete: 0.87 (0.60-1.25)
Full Cohort
Unadjusted: 0.96 (0.79-1.18)
Adjusted: 0.81 (0.63-1.04)
Unadjusted Complete: 0.92 (0.74-1.15)
Reference: Liu et al. (2008,156708)
Period of Study: 1995-2005
Location: Taiwan
Outcome: Brain cancer deaths
ICD9: 191
Age Groups: 29 yrs of age or younger
Study Design: matched case-control by
sex, years of birth and death
N: 340 matched pairs
Statistical Analyses: Conditional
logistic regression
Covariates: age, gender, urbanization
level of residence, nonpetrochemical air
pollution exposure level
No direct measures of pollutants
used an index to assign petrochemical air
pollution exposure (each municipality was
assigned an exposure by dividing the
number of workers per municipality
employed in the petrochemical industry
by the municipalities total population).
Study participants divided into tertiles
based on this index.
People who lived in the group of
municipalities with the highest levels of
air pollutants arising from petrochemical
sources were at a statistically significant
increased risk for brain cancer
development compared to the group living
in municipalities with the lowest
petrochemical air pollution exposure
index.
Effect Measure: OR (95%CI)
Tertile 1: 1. ?0
Tertile 2: 1.54 (0.98-2.42)
Tertile 3: 1.65 (1.00-2.73)
P for trend < 0.01
July 2009
E-560
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Nafstad et al. (2004,
087949)
Period of Study: 1972/73-1998
Location: Oslo, Norway
Outcome: Lung cancer
ICD7 162.1-162.9
Age Groups: 40-49 yr old men
Study Design: Cohort
PM values had small variations in No effect estimates for PM
exposure level, and strong correlations
with another pollutant of interest (SO2)
and were not considered in analyses.
Copollutants: SO2

N: 16,209 males
NOx

Statistical Analyses: Cox regression
models (proportional hazards)


Covariates: age at inclusion, smoking
habits, education


Season: all year

Reference: (Pope CA and Burnett, 2007,
090928)
Period of Study: 1982-1998
Location: 50 US states, District of
Columbia, and Puerto Rico
Outcome: Lung cancer mortality (162)
Age Groups: >30 years
Study Design: Longitudinal cohort
(Cancer Prevention II Study)
N: 415,000 CPS II patients with
information involving PMio
Statistical Analyses: Cox proportional
hazard, incorporating a spatial random-
effects component
Covariates: Age, sex, race, education,
ETS, smoking status, marital status,
occupational exposure, diet, body-mass
index, alcohol consumption
Pollutant: PM2.6
Mean (SD): 1979-1983: 21.1(4.6)
1999-2000: 14.0(3.0)
Average: 17.7(3.7)
Averaging time: 1982-1998
PM Increment: 10 /yglm3
RR Estimate [Lower CI, Upper CI]
lag:
Lung Cancer: 1979-1983: 1,08[1.01,
1.16]
1999-2000: 1.13(1.04,1.22]
Avg: 1.14[1.04,1.23]
RR results were also presented in Figures
2-5. Authors found that PM2.6 had the
strongest association with increased risk
of all-cause, cardiopulmonary, and lung
cancer mortality.
Reference: Sram et al, (2007,188457)
Period of Study: 2/6/2001-2/20/2001
Location: Prague, Czech Republic
Outcome: Chromosomal aberrations
Study Design: Panel
Covariates: urinary cotinine, plasma
levels of vitamins A, E and C
Statistical Analysis: bivariate
correlations, AN0VA, Mann-Whitney,
Kruskal-Wallis and Spearman rank
correlation
Statistical Package: STATISTICA, SAS
Age Groups: 53 city policemen, aged 22-
50 years, spending 8+ hours outdoors
Pollutant: PM10
Averaging Time: NR
Range: 32-55/yg/m3
Copollutant: PM2.6
Results not given by PM increment.
Reference: Sram et al, (2007,188457)
Period of Study: 2/6/2001-2/20/2001
Location: Prague, Czech Republic
Outcome: Chromosomal aberrations
Study Design: Panel
Covariates: urinary cotinine, plasma
levels of vitamins A, E and C
Statistical Analysis: bivariate
correlations, AN0VA, Mann-Whitney,
Kruskal-Wallis and Spearman rank
correlation
Statistical Package: STATISTICA, SAS
Age Groups: 53 city policemen, aged 22-
50 years, spending 8+ hours outdoors
Pollutant: PM2.6
Averaging Time: NR
Range: 27-38/yg/m3
Copollutant: PM10
Results not given by PM increment.
July 2009
E-561
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Tovalin et al. (Tovalin et al.,
2006, 091322)
Period of Study: April-May 2002
Location: Mexico City and Puebla
Outcome: DNA damage (comet tail
length)
Age Groups: 18-60
Study Design: Panel Study
N: 55 male workers
Statistical Analyses: Mann-Whitney
test, Chi-square, Spearman's correlation,
logistic regression
Statistical Package: SPSS and STATA
Pollutant: PM2.6
Personal monitoring values observed in
this study reported in Tovalin et al. 2003
Median Personal Exposure to PM2.5:
Mexico City
Outdoor Worker: 133/yg/m3
Indoor Worker: 86.6 //g|m3
Puebla
Outdoor Worker: 122/yg/m3
Indoor Worker: 78.3 //g|m3
OR for being a highly damaged worker:
1.02 (1.01-1.04), p - 0.03
Correlation between comet tail length and
PM 2.5: 0.57, p - 0.000
OR for being a highly damaged worker:
1.03, pn 0.07
Comet Tail Length
Outdoor Worker: 46.80 /jm
Indoor Worker: 30.11 fjm
p < 0.01
Percent Highly DNA Damaged Cells
Outdoor Worker: 68%
Indoor Worker: 20%
All units expressed in uglm unless otherwise specified.
Table E-28. Long-term exposure - cancer outcomes - other PM size fractions.
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Pope CA et al., 2002,
024689)
Period of Study: 1982-1998
Location: 50 US states, District of
Columbia, and Puerto Rico
Outcome: Lung cancer mortality (162) Pollutant: PMie
Age Groups: Ages >30 years who were Mean (SD): 1979-1983: 40.3(7.7)
members of a household with at least one
individual £45yrs.
Relative risks effect estimates were
recorded in Figure 5 and not presented
quantitatively anywhere else.
Study Design: Longitudinal cohort
(Cancer Prevention Study II)
N: 359,000 CPS II participants with
information regarding PMie and PMie -
PM2.6
Statistical Analyses: Cox proportional
hazard, incorporating a spatial random-
effects component
Covariates: Age, sex, race, education,
ETS, smoking status, marital status,
occupational exposure, diet, body-mass
index, alcohol consumption
Smoking covariates adjusted for:
Indicator: current smoker, former smoker,
pipe or cigar smoker, started smoking
before or after age 18
Continuous, current and former smokers:
years smoked, years smoked squared,
cigarettes per day, cigarettes per day
squared, number of hours per day
exposed to passive cigarette smoke.
Pollutant: PMib 2b
Mean (SD): 1979-1983: 19.2(6.1)
Averaging Time: 1979-1983
'All units expressed in //gfm3 unless otherwise specified.
July 2009
E-562
DRAFT-DO NOT CITE OR QUOTE

-------
E.7. Long-Term Exposure and Reproductive Effects
Table E-29. Long-term exposure - reproductive outcomes - PM10.
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Bell at al. (2007, 091059)
Period of Study: 1999-2002
Location: Connecticut-Fairfield,
Hartford, New Haven, New London,
Windham, Massachusetts-Barnstable,
Berkshire, Bristol, Essex, Hampden,
Middlesex, Norfolk, Plymouth, Suffolk,
Worcester
Outcome: Low birth weight
Age Groups: Neonates
Study Design: Cross-sectional
N: 358,504 births
Statistical Analyses: Multiple logistic
and linear regressions
Covariates: Child's sex, mother's
education, tobacco use, mother's marital
status, mother's race, time prenatal care
began, mother's age, birth order,
gestation length
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): 22.3 (5.3)
Monitoring Stations: NR
Copollutant: NO2, CO, SO2
Gestation exposure correlation:
PM2.6: r - 0.77
NO?: r - 0.55
PM Increment: 7.4/yg/m3 (IQR)
Difference in birth weight [Lower CI,
Upper CI]
per IQR for the gestational period: -
8.21-11.1 to-5.3]
Difference in birth weight by race of
mother [Lower CI, Upper CI]:
Black: -7.9 [-16.0, 0.2]
White:-9.0 [-12.2 to-5.9]
Range among trimester models for
change in birth weight per IQR
increase (min, max)
trimester: -6.6 to -4.7
3rd
OR Estimate for birth weight < 2500
g [Lower CI, Upper CI]
per IQR for the gestational period:
1.027 [0.991,1.064]
Notes: Analyses using first births alone
yielded similar results. Two pollutant
models for uncorrelated pollutants were
analyzed but not presented quantitatively.
Reference: Brauer et al. (2008,156292) Outcome: Preterm birth, SGA, LBW
Period of Study: 1999-2002
Location: Vancouver, BC
Age Groups: Study Design: Cross-
sectional
N: 70,249 births
Statistical Analyses: Logistic
regression
Covariates: Sex, parity, month and year
of birth, maternal age and smoking,
neighborhood level income and education
Statistical Package: SAS
Pollutant: PM10
Averaging Time: 24-h
Mean (SD): 12.7
Range (Min, Max): 5.6, 35.4
Monitoring Stations: 19
Copollutant: NO
NO2
CO
SO2
0s
PM Increment: 1 /yglm3
Effect Estimate [Lower CI, Upper CI]
pollutant assessed for entire
pregnancy period:
SGA: 1.02 (0.99,1.05)
LBW: 1.01 (0.95,1.08)
Preterm (< 30 weeks): 1.13 (0.95,1.35)
July 2009
E-563
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Chen et al. (2002, 024945)
Period of Study: 1991-1999
Location: Washoe County, Nevada
Outcome: birth weight
Age Groups: single births with
gestational age between 37-44 weeks
and maternal all ages
Study Design: cross-sectional
N: 33,859 single births
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): 31.53 (22.32)
Percentiles: 25th: 16.80
50th(Median): 26.30
Statistical Analyses: multiple linear and 75th: 39.35
logistic regression
Covariates: infant sex, maternal
residential city, education, medical risk
factors, active tobacco use, drug use,
alcohol use, prenatal care, mother's age,
race and ethnicity of mothers and weight
gain of mothers
Dose-response Investigated? No
Statistical Package: SPSS 10.0
Range (Min, Max): (0.97-157.32)
Monitoring Stations: 4
Copollutant: CO
0s
July 2009
E-564
PM Increment: 10 /yglm3
Effect Estimate [Lower CI, Upper CI]:
Using continuous pollutant variables
Model one-PMio
1	trimester
Crude model: B - -0.186 (0.225)
Adjusted model: B - -0.082 (0.221)
2	trimester
Crude model: B - 0.045 (0.223)
Adjusted model: B - -0.020 (0.221)
3	trimester
Crude model: B - -0.509 (0.231)
Adjusted model: B - -0.395 (0.227)
Whole
Crude model: B - -0.823 (0.459)
Adjusted model: B - -0.726 (0.483)
Model two
CO and PMio
3 trimester
Crude model: B - -1.044 (0.457)
Adjusted model: B - -1.078 (0.445)
O3 and PMio
3 trimester
Crude model: B - -1.035 (0.385)
Adjusted model: B - -0.966 (0.378)
Model three
PMio, O3, and CO
3 trimester
Crude model: B - -1.070 (0.458)
Adjusted model: B - -1.102 (0.446)
Whole
Crude model: B - -1.413 (0.733)
Adjusted model: B - -1.332 (0.738)
Using categorical pollutant variables-3
trimester
Model 1-PMio
Adjusted model: B - -10.243 (5.235)
Model 2
PMio and CO
Adjusted model: B - -11.883 (6.108)
PMio and O3 Adjusted model: B - -9.144
(5.860)
Model 3
PMio, CO, and O3 Adjusted model: B - ¦
10.937 (6.222)
Using logistic regression (ref value -
< 19.72 //g/m3
Exposure to PMio at 3 trimester at
>44.74/yglm3: OR - 1.105 (0.714-
1.709)
BeJMk^TO™ QUOTE
- 1.050 (0.811-1.360)
Notes: Crude model: model with air-
pollutant variables controlled with

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Dales et al. (2004, 087342)
Period of Study: Jan 1,1984-Dec 31,
1999
Location: Canada (12 cities)
Outcome: SIDS (a sudden, unexplained Pollutant: PM
death of a child < 1 year of age for
which a clinical investigation and autopsy
fail to reveal a cause of death)
Averaging Time: 24-hs (PM measures
every 6 days
Age Groups: Infants < 1 yr
Study Design: Time-series
N: Total population of 12 cities:
10,310,309
1556 cases of SIDS over study period
Statistical Analyses: Random-effects
regression model for count data (a linear
association between air pollution and the
incidence of SIDS was assumed on the
logarithmic scale)
Covariates: weather factors (daily mean
temp, daily mean relative humidity,
maximum change in barometric pressure,
all measured on the day of death), length
of time-period adjustment, seasonal
indicator variables, and size-fractionated
PM
Season: Used piece-wise constant
functions in time that varied by 3, 6, or
12 months
Dose-response Investigated? No
Statistical Package: NR
gaseous pollutants every day)
Mean (IQR): PMio: 23.43 (15.56)
Range (Min, Max): IQR presented above
Monitoring Stations: When data were
available from more than one monitoring
site, they were averaged
Copollutant: PM2.6
PM10
CO
NO?
0s
SO?
Notes: The abstract reports no
association between increased daily rates
of SIDS and fine particles measured every
sixth day. However, no effect estimates
presented for PM (only gaseous
pollutants adjusted for PM).
Reference: Dugandzic et al. (2006,
088681)
Period of Study: 1/1/1988-12/31/2000
Location: Nova Scotia, Canada
Outcome: Low birth weight (LBW)
(< 2500 grams)
Age Groups: Babies born a 37 weeks
(full term)
Study Design: cross-sectional
N: 74,284 births
Statistical Analyses: Logistic
regression
Covariates: Maternal age, parity, prior
fetal death, prior neonatal death, prior
low birth weight infant, smoking during
pregnancy, neighborhood family income,
infant gender, gestational age, weight
change, year of birth
Season: All
Dose-response Investigated? Yes
Statistical Package: SAS
Pollutant: PM10
Averaging Time: 24-h
Mean (SD):
Percentiles: 25th: 14
50th(Median): 16
75th: 19
Range (Min, Max): Max: 53
Monitoring Stations: 18
Copollutant: SO2, O3
Notes: Only three stations monitored
more than one pollutant. Daily data were
available for gaseous pollutants while
particulate levels were measured every
sixth day.
PM Increment: 1) IQR (5/yglm3)
2) Quartiles (first quartile is the
reference)
Exposure period: first trimester
Unadjusted model
2nd quartile: 1.24 (0.95,1.62)
3rd quartile: 1.25 (0.96,1.62)
4th quartile: 1.28 (1.00,1.65)
Per IQR: 1.09 (1.00,1.18)
Adjusted model
2nd quartile: 1.24 (0.94,1.64)
3rd quartile: 1.24 (0.95,1.64)
4th quartile: 1.33 (1.02,1.74)
Per IQR: 1.09 (1.00,1.19)
Adjusted for Birth Year model
2nd quartile: 1.14 (0.86,1.52)
3rd quartile: 1.08 (0.82,1.44)
4th quartile: 1.11 (0.84,1.48)
Per IQR: 1.03 (0.94,1.14)
Exposure period: second trimester
Unadjusted model
2nd quartile: 0.98(0.76,1.28)
3rd quartile: 1.09 (0.84,1.40)
4th quartile: 1.00 (0.77,1.28)
Per IQR: 1.00 (0.91,1.09)
Adjusted model
July 2009
E-565
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
2nd quartile: 1.02 (0.77,1.34)
3rd quartile: 1.16 (0.89,1.51)
4th quartile: 1.09 (0.83,1.42)
Per IQR: 1.02 (0.93,1.12)
Adjusted for Birth Year model
2nd quartile: 0.99(0.75,1.31)
3rd quartile: 1.10(0.84,1.45)
4th quartile: 1.01 (0.76,1.34)
Per IQR: 1.00 (0.90,1.10)
Exposure period: third trimester
Unadjusted model
2nd quartile: 0.93(0.72,1.20)
3rd quartile: 1.07 (0.83,1.37)
4th quartile: 0.92 (0.71,1.18)
Per IQR: 0.95 (0.87,1.05)
Adjusted model
2nd quartile: 0.96(0.73,1.26)
3rd quartile: 1.14 (0.88,1.48)
4th quartile: 1.03 (0.79,1.35)
Per IQR: 0.99 (0.89,1.09)
Adjusted for Birth Year model
2nd quartile: 0.92(0.70,1.21)
3rd quartile: 1.04 (0.80,1.36)
4th quartile: 0.92 (0.69,1.22)
Per IQR: 0.94 (0.85,1.05)
Reference: Gilboa, et al. (2005, 087892) Outcome: Birth defects
Age Groups: newborn babies
Period of Study: January 1,1996-
December 31, 2000
Location: Seven Counties in Texas, USA:
(Bexar, Dallas, El Paso, Harris, Hidalgo,
Tarrant, Travis)
Study Design: Case-control
l\l: 5,338 newborn babies
4574 controls
Statistical Analyses: Logistic
regression
Covariates: alcohol consumption during
pregnancy, attendant of delivery (i.e., the
person who delivered the baby
(physician/nursemaid-wife vs. other)),
gravidity, marital status, maternal age,
maternal education, maternal illness,
maternal race/ethnicity, parity, place of
delivery, plurality, prenatal care, season
of conception, and tobacco use during
pregnancy
control frequency matched to cases by
vital status, year and maternal county of
residence
Season: covariate in model
Dose-response Investigated? Yes
Statistical Package: SAS v 8.2
Pollutant: PMio
Averaging Time: NR
Percentiles: 25th: < 19.5
50th(Median): 19.5-<23.8
75th: 23.8- < 29.0
100th: > 29.0
Monitoring Stations: The Environmental
Protection Agency provided raw data
for hourly (for gases) or daily (for PM) air
pollution concentrations for the seven
study counties
Copollutant: CO, NO2, O3, SO2
PM Increment: calculated as quartiles of
avg concentration during weeks 3-8 of
pregnancy
Isolated Cardiac Defects
Aortic artery and valve defects:
25th: 0.40 (0.15,1.03)
50th: 0.45 (0.18,1.13)
75th: 0.68 (0.28,1.65)
Atrial septal defects:
25th: 1.41 (0.86, 2.31)
50th: 2.13 (1.34, 3.37)
75th: 2.27 (1.43, 3.60)
Pulmonary artery and valve defects:
25th: 1.14(0.62, 2.10)
50th: 0.79 (0.41,1.55)
75th: 0.68 (0.33,1.40)
Ventricular septal defects:
25th: 0.83 (0.61,1.11)
50th: 1.12 (0.85,1.48)
75th: 0.98 (0.73,1.32)
Multiple Cardiac Defects
July 2009
E-566
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Conotruncal defects:
25th: 1.13 (0.79,1.62)
50th: 1.20 (0.84,1.72)
75th: 1.26 (0.86,1.84)
Endocardial cushion and mitral valve
defects:
25th: 0.82 (0.54,1.25)
50th: 0.66 (0.42,1.05)
75th: 0.63 (0.38,1.03)
Isolated Oral Clefts
Cleft lip with or without palate:
25th: 1.29 (0.90,1.85)
50th: 1.45 (1.01, 2.07)
75th: 1.37 (0.94,2.00)
Cleft palate:
25th: 0.99 (0.55,1.78)
50th: 1.14 (0.64, 2.03)
75th: 1.11 (0.60, 2.06)
Individual Birth Defects
Aortic valve stenosis:
25th: 0.91 (0.53,1.57)
50th: 0.86 (0.50,1.50)
75th: 1.12 (0.63,1.99)
Atrial septal defects:
25th: 1.10 (0.89,1.35)
50th: 1.28 (1.04,1.57)
75th: 1.26 (1.03,1.55)
Coarctation of the aorta:
25th: 0.78 (0.53,1.15)
50th: 0.68 (0.45,1.02)
75th: 0.75 (0.48,1.15)
Endocardial cushion defects:
25th: 0.87 (0.49,1.55)
50th: 1.12 (0.64,1.96)
75th: 0.89 (0.47,1.65)
Ostium secundum:
25th: 1.15 (0.85,1.55)
50th: 1.13 (0.83,1.53)
75th: 1.06 (0.77,1.48)
Pulmonary artery atresia without
ventricular septal defects:
25th: 1.93 (1.08, 3.45)
50th: 2.01 (1.11,3.64)
75th: 0.86 (0.41,1.83)
Pulmonary valve stenosis:
25th: 1.16 (0.88,1.55)
July 2009
E-567
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
50th: 1.25 (0.94,1.66)
75th: 1.27 (0.94,1.71)
Tetralogy of Fallot:
25th: 1.21 (0.72, 2.01)
50th: 1.40 (0.84,2.33)
75th: 1.45 0.85, 2.48)
Ventricular septal defects:
25th: 1.06 (0.90,1.24)
50th: 1.10 (0.94,1.29)
75th: 1.08 (0.92,1.27)
Reference: Gouveia et al. (2004,
055613)
Period of Study: 1997
Location: Sao Paulo, Brazil
Outcome: birth weight
Age Groups: singleton full term live
births within 1000 g to 5500 g
Study Design: Cross sectional study
N: 179,460 live births
Monitoring Stations: maximum of 12
Statistical Analyses: GAM and Logistic sites
regression models
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): 60.3 (25.2)
Range (Min, Max): (25.5-153.0)
Covariates: maternal age, length of
gestation, season, infant gender,
maternal education, number of antenatal
care visits, parity, and the type of
delivery	q3
Season: All seasons
Dose-response Investigated? Yes
Statistical Package: S-Plus 2000
Copollutant (correlation): CO: r - 0.9
SO?
PM Increment: 10 /yglm3
Mean [Lower CI, Upper CI]:
Changes in birth weight (in g)
First trimester - -13.7 (-27.0, -0.4)
Second trimester - -4.4 (-18.9,10.1)
Third trimester - 14.6 (0.0, 29.2)
RR Estimate [Lower CI, Upper CI]:
(RR estimates are adjusted odds ratios
for low birth weight according to
quartiles of air pollution in each trimester
of pregnancy.)
1st quartile
First trimester - 1 (REF)
Second trimester - 1 (REF)
Third trimester - 1 (REF)
2nd quartile
First trimester - 1.105 (0.994,1.229)
Second trimester - 1.003 (0.904,1.113)
Third trimester - 1.004 (0.914,1.104)
3,d quartile
First trimester - 1.049 (0.903,1.219)
Second trimester - 1.074 (0.920,1.254)
Third trimester - 1.003 (0.861,1.169)
4th quartile
First trimester - 1.144 (0.878,1.491)
Second trimester - 1.252 (1.028,1.525)
Third trimester - 0.970 (0.780,1.205)
Multiple linear regression coefficients
(SE) obtained from single, dual, and three
pollutant models
Single pollutant model - -1.37 (0.68)
Two pollutant (PMio and CO) - -0.51
(0.87)
Two pollutant (PMio and SO2) - -0.94
(0.75)
Three pollutant - -0.47 (0.88)
July 2009
E-568
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ha et al. (2003, 042552)
Period of Study: Jan 1995-Dec 1999
Location: Seoul, South Korea
Outcome: Postneonate total and
respiratory mortality
Age Groups: 1 month-1 yr
2 yr-65 yr, > 65 yr
Study Design: Time-series
N: 1045 post neonate deaths, 67,597 2-
65 yr old deaths, 100,316 > 65 yr old
deaths
Statistical Analyses: Generalized
additive model
Covariates: Seasonality, temperature,
relative humidity, day of the week
Dose-response Investigated? No
Statistical Package: S Plus
Lags Considered: 0,1, 2, 3, 4, 5, 6, 7,
moving averages from 1-5 days
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): 69.2 (31.6)
Percentiles: 25th: 44.8
50th(Median): 64.2
75th: 87.7
Range (Min, Max): 10.5/yglm3,
245.4 /yg/m3
Monitoring Stations: 27
Copollutant (correlation):
NO?: r - 0.73
S02: r - 0.62
0a: r - -0.02
CO: r - 0.63
PM Increment: 42.9/yglm3
RR Estimate [Lower CI, Upper CI]
lag:
Total Mortality:
1	month-1 yr (post-neonates):
1.142(1.096,1.190] lag 0
2	yr-65 yr:
1.008(1.006,1.010] lag 0
>65 yr (elderly):
1.023(1.023,1.024] lag 0
Respiratory Mortality:
1	month-1 yr (post-neonates):
2.018(1.784, 2.283] lag 0
2	yr-65 yr:
1.066(1.044,1.090] lag 0
>65 yr (elderly):
1.063(1.055,1.072] lag 0
Reference: Hansen, et al. (2006,
089818)
Period of Study: July 1, 2000- June 30,
2003
Location: Brisbane, Australia
Outcome: Pre term birth (< 37 weeks)
Age Groups: newborn babies
Study Design: Cross-sectional
N: 1583 live pre terms births
28,200 singleton live births
Statistical Analyses: Multiple logistic
regression models
Covariates: Neonate gender, mother's
age, parity, indigenous status, number of
antenatal visits, marital status, number of
previous abortions/miscarriages, type of
delivery, and index of SES
Season: all
Dose-response Investigated? Yes
Statistical Package: SAS version 8.2
Pollutant: PM10
Averaging Time: recorded hourly,
averaged daily
Mean (SD): 19.6 (9.4)
Range (Min, Max): 4.9,171.7
Monitoring Stations: 5
Copollutant (correlation):
Fine PM or bsp, 0.1 to < 2.5 /yg in
diameter (0.58 to 0.76)
O3 (0.54 to 0.83)
NO? (0.54 to 0.75)
PM10 (0.80 to 0.93)
Note: Correlations presented are for the
individual pollutant across monitoring
stations (not correlations between PM10
and the pollutant.)
PM Increment: Trimester One
4.5/yg/m3
Last 90 days prior to birth
5.7/yg/m3
Odds Ratio [Lower CI, Upper CI]:
Trimester one
1.15 [1.06,1.25]
Last 90 days prior to birth
1.04 [0.92,1.16]
Reference: Hansen et al. (2007,
090703)
Period of Study: Jul 2000-Jun 2003
Location: Brisbane, Australia
Outcome: Birth weight and Small for
Gestational Age (SGA
< 10th percentile for age and gender)
head circumference (HC) and crown-heel
length (CHL) among subsample
Study Design: Cross-sectional
N: 26,617 births (birth weight analysis)
and 21,432 (HC and CHL analyses)
Statistical Analyses: Logistic (SGA) and
linear (birth weight, HC, CHL) regressions
Covariates: gender, gestational age
(with a quadratic term), maternal age,
parity, number of previous
abortions/miscarriages, marital status,
indigenous status, number of antenatal
visits, type of delivery, an index of SES,
and season of birth
Season:
as a covariate
Pollutant: PM10
Averaging Time: Trimester and monthly
averages were used in analyses
(calculated as the mean of daily values
hourly data was use to calculate daily
means
city-wide avg used)
Mean (SD): 19.6 (9.4)
Percentiles: 25th: 14.6
50th: 18.1
75th: 22.7
Range (Min, Max): (4.9,171.7)
Monitoring Stations: 5
Copollutant (correlation): By
trimesters:
PM10TI:
PM Increment: IQR (8.1 /yg/m3)
Effect Estimate [Lower CI, Upper CI]:
Change (0) in mean birth weight (g)
associated with trimester-specific
exposures
Trimester 1:
Continuous exposure: -3.2 (-11.9, 5.5)
Quartiles of exposure:
1: Ref
2: -4.7 (-19.7,10.2)
3: 4.2(-12.9, 21.3)
4:-0.2 (-19.2,18.8)
p-trend: 0.864
Trimester 2:
Continuous exposure: 0.4 (-9.4,10.2)
July 2009
E-569
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Dose-response Investigated? Yes,
assessed exposures as quartiles
Statistical Package: SAS v8.2
PMio T2: r - 0.12
PM10 T3: r - -0.55
O3 T1: r - 0.77
O3 T2: r - 0.28
O3 T3: r - -0.61
NO2TI: r - 0.32
NO2 T2: r - -0.65
NO2 T3: r - -0.17
visibility reducing particles (bsp) T1:
r - 0.82
visibility reducing particles (bsp) T2: r — -
.15
visibility reducing particles (bsp) T3: r — -
0.50
PM10TI: r - 0.12
PM10 T2:
PM10 T3: r - 0.04
O3 T1: r - -0.11
O3 T2: r - 0.80
O3 T3: r - 0.18
NO2TI: r - 0.77
NO2T2: r - 0.25
NO2 T3: r - -0.72
visibility reducing particles (bsp) T1:
r - 0.23
visibility reducing particles (bsp) T2:
r - 0.80
visibility reducing particles (bsp) T3: r — -
0.24
PM10TI: r - -0.55
PM10 T2: r - 0.04
PM10 T3:
O3 T1: r - -0.56
O3 T2: r - -0.18
O3 T3: r - 0.81
NO2TI: r - -0.20
NO2T2: r - 0.75
NO2 T3: r - 0.22
visibility reducing particles (bsp) T1: r — -
0.62
visibility reducing particles (bsp) T2:
r - 0.19
visibility reducing particles (bsp) T3:
r - 0.79
Quartiles of exposure:
1: Ref
2: 12.7 (-2.3, 27.6)
3: 7.6 (-10.6, 25.7)
4: 1.0(-18.7, 20.7)
p-trend: 0.922
Trimester 3:
Continuous exposure: 3.6 (-6.9,14.0)
Quartiles of exposure:
1: Ref
2: 2.9 (-12.8,18.7)
3: 18.5(0.0, 36.9)
4: 4.3(-15.8, 24.4)
p-trend: 0.524
ORs for SGA associated with
trimester-specific exposures
Trimester 1:
Continuous exposure: 1.04 (0.96,1.12)
Quartiles of exposure:
1: Ref
2: 1.23(1.07,1.42)
3: 1.12(0.95,1.31)
4: 1.12(0.94,1.34)
p-trend: 0.361
Trimester 2:
Continuous exposure: 0.95 (0.88,1.04)
Quartiles of exposure:
1: Ref
2: 0.96 (0.83,1.11)
3: 1.06 (0.89,1.25)
4: 0.98 (0.81,1.18)
p-trend: 0.962
Trimester 3:
Continuous exposure: 0.93 (0.85,1.03)
Quartiles of exposure:
1: Ref
2: 0.90 (0.78,1.04)
3: 0.81 (0.68, 0.96)
4: 0.86 (0.71,1.04)
p-trend: 0.098
Change (0) in mean head
circumference (HC
cm) associated with trimester-specific
exposures
Trimester 1:
Continuous exposure: -0.01 (-0.04, 0.02)
July 2009
E-570
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
~uartiles of exposure:
1: Ref
2: -0.02 (-0.07, 0.04)
3: -0.02 (-0.08, 0.04)
4: -0.02 (-0.08, 0.05)
p-trend: 0.605
Trimester 2:
Continuous exposure: -0.01 (-0.04, 0.02)
~uartiles of exposure:
1: Ref
2: 0.03 (-0.02, 0.08)
3: 0.00 (-0.06, 0.06)
4: -0.01 (-0.08, 0.05)
p-trend: 0.538
Trimester 3:
Continuous exposure: 0.02 (-0.02, 0.05)
~uartiles of exposure:
1: Ref
2: 0.02 (-0.04, 0.07)
3: 0.07(0.01,0.13)
4: 0.04 (-0.03,0.11)
p-trend: 0.171
Change (0) in mean crown-heel length
(CHL
cm) associated with trimester-specific
exposures
Trimester 1:
Continuous exposure: 0.00 (-0.05, 0.05)
~uartiles of exposure:
1: Ref
2: 0.02 (-0.07, 0.11)
3: 0.01 (-0.10,0.11)
4: 0.04 (-0.07,0.16)
p-trend: 0.511
Trimester 2:
Continuous exposure: 0.07 (0.01, 0.13)
~uartiles of exposure:
1: Ref
2: 0.10(0.01,0.18)
3: 0.11 (0.00, 0.21)
4: 0.13(0.01,0.24)
p-trend: 0.049
Trimester 3:
Continuous exposure: -0.01 (-0.07, 0.05)
~uartiles of exposure:
1: Ref
July 2009
E-571
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
2:
¦0.02 (-0.11,0.07)
3:
0.10 (-0.01, 0.21)
4:
¦0.01 (-0.13, 0.10)
p-trend: 0.883
Reference: (Hansen et al„ 2009,
192362)
Period of Study: 1/1997-12/2004
Location: Brisbane, Australia
Outcome: birth defects- artery and valve, Pollutant: PMio
atrial and ventricular septal, conotruncal,
endocardial cushion and mitral valve, cleft Averaging Time: daily
lip and palate
Study Design: Case-control
Mean (SD) Unit: 18.0//gfm
Range (Min, Max): (4.4,151.7)
Covariates: mother s age, marital status, „ „ .
... , , ¦	. Copollutant correlation: NR
indigenous status, previous pregnancies, r
last menstrual period, area-level
socioeconomic status, distance to a
pollution monitor
Statistical Analysis: Conditional logistic
regression
Statistical Package: R
Age Groups: neonates
Increment: 4/yg/m3
Odds Ratios (95% CI) for risk of defect
Aortic Artery and Valve Defects
All Births, Matched: 1.10(0.76-1.56)
Births ~ 12km to Monitor: 1.83 (1.16-
2.98)
Births ~ 6km to Monitor: 1.43 (0.73-
2.90)
All Births, Unmatched: 1.09 (0.84-1.39)
Atrial Septal Defects
All Births, Matched: 1.06 (0.86-1.30)
Births ~ 12km to Monitor: 1.07 (0.84-
1.37)
Births ~ 6km to Monitor: 0.88 (0.60-
1.27)
All Births, Unmatched: 1.14(0.98-1.33)
Pulmonary Artery and Valve Defects
All Births, Matched: 0.90 (0.61-1.29)
Births ~ 12km to Monitor: 0.69 (0.43-
1.08)
Births ~ 6km to Monitor: 1.46 (0.76-
2.73)
All Births, Unmatched: 0.99 (0.78-1.24)
Ventricular Septal Defects
All Births, Matched: 0.87 (0.73-1.04)
Births ~ 12km to Monitor: 0.85 (0.69-
1.03)
Births ~ 6km to Monitor: 0.90 (0.68-
1.18)
All Births, Unmatched: 1.15 (1.02-1.30)
Conotruncal Defects
All Births, Matched: 0.80 (0.54-1.19)
Births ~ 12km to Monitor: 0.94 (0.55-
1.49)
Births ~ 6km to Monitor: 0.66 (0.27-
1.45)
All Births, Unmatched: 0.97 (0.74-1.24)
Endocardial Cushion and Mitral Valve
Defects
All Births, Matched: 1.29 (0.82-2.04)
Births ~ 12km to Monitor: 1.28 (0.75-
2.19)
Births ~ 6km to Monitor: 0.90 (0.44-
1.86)
All Births, Unmatched: 0.94 (0.68-1.26)
Cleft Lip
July 2009
E-572
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
All Births, Matched: 1.05 (0.72-1.51)
Births ~ 12km to Monitor: 1.16 (0.72-
1.82)
Births ~ 6km to Monitor: 1.03 (0.56-
1.82)
All Births, Unmatched: 1.01 (0.79-1.27)
Cleft Palate
All Births, Matched: 0.69 (0.50-0.93)
Births ~ 12km to Monitor: 0.53 (0.29-
0.87)
Births ~ 6km to Monitor: 0.71 (0.49-
1.00)
All Births, Unmatched: 0.89 (0.72-1.10)
Cleft Lip with or without Cleft Palate
All Births, Matched: 1.05 (0.84-1.30)
Births ~ 12km to Monitor: 1.03 (0.79-
1.34)
Births ~ 6km to Monitor: 0.83 (0.58-
1.19)
All Births, Unmatched: 1.04 (0.89-1.21)
July 2009
E-573
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Jalaludin et al. (2007,	Outcome: Gestational age (categorized:
156601)	preterm birth: < 37 weeks
Period of Study: 1998-2000	term birth: a 37 weeks but < 42 weeks)
Location: Sydney, Australia	Age Groups: infants
Study Design: Cross-sectional
N: 123,840 singleton births of > 20
weeks gestation
Statistical Analyses: Logistic
regression
Covariates: sex of child, maternal age,
maternal smoking during pregnancy,
gestational age at first antenatal visit,
whether mother identifies as being
Aboriginal or Torres Strait Islander,
whether first pregnancy, season of
conception, SES, (temperature and
relative humidity were not significant in
single variable models and therefore,
were not included)
Season: examined as covariate and
effect modifier
Dose-response Investigated? No
Statistical Package: SAS v8
Pollutant: PMio
Averaging Time: 24 h averages used to
calculate the mean concentration over
the first trimester, the 3 months
preceding birth, the first month after the
estimated date of conception, and the
month prior to delivery
Mean (SD): (24 hr averages)
All year: 16.3 (6.38)
summer: 18.2 (7.20)
Autumn: 17.0 (6.23)
Winter: 14.5(5.57)
Spring: 15.7 (5.82)
Monitoring Stations: 14 stations within
the Sydney metropolitan area (levels
averaged to provide one estimate for the
entire study area)
Copollutant (correlation): PMio
PM2.B (r - 0.83)
CO (r - 0.28)
N0?(r - 0.48)
O3 (r - 0.50)
SO2 (r - 0.42)
Notes: Correlations between monitoring
stations measuring PMio ranged from
0.67 to 0.91
PM Increment: 1 /yglm3
Effect Estimate [Lower CI, Upper CI]:
ORs (air pollutant concentration during
the 1st trimester and preterm birth by
season)
Autumn: 1.462 (1.267,1.688)
Winter: 1.343 (1.190,1.516)
Spring: 1.119(0.973,1.288)
Summer: 0.913 (0.889, 0.937)
ORs (air pollutant concentrations during
different exposure periods and preterm
birth
for all of Sydney and among only those
residing within 5 km of a monitoring
station)
1 month preceding birth
Sydney: 0.991 (0.979,1.003)
5km: 1.008 (0.993,1.022)
3 months preceding birth
Sydney: 0.989 (0.975,1.004)
5km: 1.012 (0.995,1.030)
1st month of gestation
Sydney: 0.983 (0.973, 0.993)
5km: 0.957 (0.914,1.002)
1st trimester
Sydney: 0.987 (0.973,1.001)
5km: 1.009 (0.978,1.041)
Notes: Authors note that effect of PMio
on preterm birth for infants conceived
during the autumn did not remain in 2
pollutant models (ORs between 0.77 and
1.04)
July 2009
E-574
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Kaiser et al. (2004, 076674)
Period of Study: 1995-1997
Location: 25 US counties (23
metropolitan areas): Jackson, AL
Fresno, CA
Los Angeles, CA
Sacramento, CA
San Diego, CA
San Francisco, CA
Denver, CO
Hartford, CT
Cook, IL
Baltimore, MD
Wayne, Ml
St. Louis, MO
Bronx, NY
Kings, NY
New York, NY
Philadelphia, PA
El Paso, TX
Harris, TX
Dallas, TX
Oklahoma, OK
Tulsa, OK
Providence, Rl
Salt Lake City, LIT
King, WA
Milwaukee, Wl
Outcome: Postneonatal death:
All cause, SIDS (798.0)
Respiratory disease (460-519)
Age Groups: infants between 1-12
months
Study Design: Attributable risk
assessment
N: 700,000 infants (# deaths NR)
Statistical Analyses: Risk assessment
methods described in: Kunzli et al. Public-
health impact of outdoor and traffic-
related air pollution: a European
assessment. Lancet 2000, 356: 795-801.
Covariates: Maternal education,
maternal ethnicity, parental marital
status, maternal smoking during
pregnancy, infant's month and year of
birth, avg temperature in the first 2
months of life
Season: All
adjusted for monthfyear of birth
Dose-response Investigated? NR
Statistical Package: NR
Lags Considered: Annual, county-level
mean
Pollutant: PMio
Averaging Time: "annual mean levels" in
each county
Mean (SD): 28.4
Range (Min, Max):
County range: 18.0, 44.8
Monitoring Stations: NR
Notes: 14 out of 25 counties had PMio
levels >25//g/m3
PM Increment: Analysis 1:
16.4//g/m3 (difference between reference
level of 12 //g/m3 and observed mean
level of 28.4 //g/m3)
Analysis 2:
13 //g/m3 (difference between reference
level of 12 //g/m3 and 25 //g/m3)
AR Estimate [Lower CI, Upper CI]:
Analysis 1:
All cause 6% [3,11]
SIDS 16% [9, 23]
Respiratory 24% [7, 44]
Attributable # deaths per 100,000
infants:
All cause 14.7 [7.3, 25.6]
SIDS 11.7 [6.8,16.6]
Respiratory 2.3 [0.7, 4.1]
Analysis 2:
All cause 5% [2, 8]
SIDS 12% [7,18]
Respiratory 19% [6, 34]
Attributable # deaths per 100,000
infants:
All cause 10.9 [5.5,19.1]
SIDS 9.0 [5.3,12.8]
Respiratory 1.8 [0.5, 3.2]
Notes: -Authors did not extrapolate
attributable cases below 12 //g/m3 (i.e.,
reference level was set at 12 //g/m3)
-Attributable risks are based on the RRs
reported by Woodruff et al, 1997 for a
10//g/m3 increase:
All cause 1.04(1.02-1.07]
SIDS 1.12 [1.07,1.17]
Respiratory 1.20 [1.06,1.36]
Reference: (Kim et al., 2007,156642)
Period of Study: May 1, 2001-May 31,
2004
Location: Seoul, Korea
Outcome (ICD9 and ICD10): LBW (low
birth weight, less than 2500 g at later
than gestational week 37), premature
delivery (birth before the completion of
the 37th week), stillbirth (intrauterine
fetal death), IUGR (birth weight lower
than the 10th percentile for the given
gestational age), and congenital anomaly
(a defect in the infant's body structure)
Age Groups: Infants
Study Design: Cross-sectional (women
visiting the clinic for prenatal care were
recruited with follow-up until discharge
after delivery)
N: 1514 observations (births)
Statistical Analyses: multiple logistic
and linear regression (in addition, for birth
weight, used generalized additive model
to account for long-term trends and
nonlinear relationships between the
response variable and the predictors, and
Pollutant: PMio
Averaging Time: Used hourly exposure
levels to calculate avg exposure levels at
each trimester, each month of pregnancy,
and 6 weeks before delivery from the
nearest monitoring station (based on
home address of mother)
also created categories within each
pregnancy period (<25th percentile
[referent], 25th to 50th percentile, and
> 50th percentile)
Mean (SD): Range of PM means across
pregnancy periods: 88.7-89.7
Monitoring Stations: 27 stations
PM increment: 10//g/m3
Preterm:
1st Trimester Odds Ratios:
Crude: 0.95 (0.90,1.01)
Adj 1:0.93 (0.87,1.00)
Adj 2: 0.93 (0.85,1.01)
2nd Trimester Odds Ratios:
Crude: 0.99 (0.94,1.06)
Adj 1: 0.98 (0.92,1.04)
Adj 2: 1.00 (0.93,1.07)
3,d Trimester Odds Ratios:
Crude: 1.02 (0.98,1.06)
Adj 1:1.05 (1.00,1.10)
Adj 2: 1.05 (0.99,1.11)
July 2009
E-575
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
to produce smoothed plots of the
relationship between PM and birth
weight)
Covariates: Adjustment 1: infant sex,
infant order, maternal age and education,
paternal education, season of birth
Adjustment 2: adjustment 1 factors plus
alcohol, maternal BMI, maternal weight
prior to delivery
(collected information on smoking, ETS,
parity, past history of illnesses, history of
illnesses during pregnancy but did not use
in analyses due to small numbers or non-
significance)
Season: adjusted for season of delivery
Dose-response Investigated? Yes
Statistical Package: SAS 8.01, S-Plus
2000
LBW:
T'Trimester Odds Ratios:
Crude: 1.02 (0.93,1.12)
Adj 1:1.03 (0.93,1.14)
Adj 2: 1.07 (0.96,1.19)
2nd Trimester Odds Ratios:
Crude: 1.03 (0.94,1.14)
Adj 1: 1.04 (0.93,1.17)
Adj 2: 1.07 (0.94,1.22)
3,d Trimester Odds Ratios:
Crude: 1.04 (0.97,1.11)
Adj 1:1.05 (0.97,1.14)
Adj 2: 1.05 (0.96,1.16)
IUGR:
T'Trimester Odds Ratios:
Crude: 1.07 (0.97,1.19)
Adj 1: 1.07 (0.95,1.21)
Adj 2: 1.14 (0.99,1.31)
2nd Trimester Odds Ratios:
Crude: 0.97 (0.85,1.12)
Adj 1:0.97 (0.82,1.13)
Adj 2: 0.93 (0.77,1.13)
3,d Trimester Odds Ratios:
Crude: 0.82 (0.68,0.99)
Adj 1:0.88 (0.72,1.08)
Adj 2: 0.85 (0.67,1.08)
Birth defect:
T'Trimester Odds Ratios:
Crude: 1.08 (0.98,1.20)
Adj 1:1.12 (1.00,1.25)
Adj 2: 1.08 (0.95,1.22)
2nd Trimester Odds Ratios:
Crude: 1.09 (0.99,1.21)
Adj 1: 1.11 (0.98,1.26)
Adj 2: 1.16 (1.00,1.34)
3,d Trimester Odds Ratios:
Crude: 1.00 (0.90,1.11)
Adj 1: 0.97 (0.86,1.08)
Adj 2: 0.97 (0.87,1.10)
Stillbirth:
T'Trimester Odds Ratios:
Crude: 0.83 (0.76,0.90)
Adj 1:0.93 (0.85,1.02)
Adj 2: 0.95 (0.85,1.02)
2nd Trimester Odds Ratios:
July 2009
E-576
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Crude: 0.99 (0.93,1.05)
Adj 1: 1.03 (0.95,1.11)
Adj 2: 1.07 (0.98,1.17)
3,d Trimester Odds Ratios:
Crude: 1.14(1.10,1.18)
Adj 1:1.09 (1.04,1.15)
Adj 2: 1.08 (1.02,1.14)
LBW (categorical PM exposure):
1" Trimester ORs:
<	25th: 1.0
25th-50th: 0.5(0.1,3.2)
>	50th: 1.0(0.3,3.8)
3,d Trimester ORs:
<	25th: 1.0
25th-50th: 1.3(0.2,10.4)
>	50th: 3.0(0.5,18.5)
6 wk before birth ORs:
<	25th: 1.0
25th-50th: 3.2(0.3,33.7)
>	50th: 5.2(0.6,47.6)
Changes in Birth Weight (95%CI) per
10 //gf'nr1 increase in PM
concentration: 1st trimester: 7.8 (1.2,
14.5)
2nd trimester:-0.3 (-7.3, 6.8)
3rd trimester: -2.1 (-7.5, 3.4)
1st month: 4.4 (-1.0, 9.8)
2nd month: 6.4 (0.6,12.2)
3,d month: 4.3 (-1.5,10.2)
4th month: 3.0 (-3.7, 9.6)
5th month: -3.9 (-10.5, 2.7)
6th month: 0.1
(-5.7, 5.8)
7th month: 0.1 (-5.1, 5.3)
8th month: 0.0 (-4.5,4.5)
9th month: 1.8 (-2.3, 5.9)
Last 6 wk: -4.8 (-9.9, 0.4)
July 2009
E-577
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lee et al. (2003, 043202)
Period of Study: Jan 1,1996-Dec 31
1998
Location: Seoul, South Korea
Outcome: Low birth weight (LBW),
< 2500 g
Age Groups: child-bearing age women
and their newborn children - delivered at
37-44 gestational weeks
Study Design: Cross-section
N: 388,905 full-term single births
Statistical Analyses: Generalized
additive model, LOESS, Akaike's criterion,
Covariates: Infant sex, birth order,
maternal age, parental education level,
time trend and gestational age.
Season: All
Dose-response Investigated? Yes
Statistical Package: NR
Pollutant: PMio
Averaging Time: Arithmetic avg of
hourly measurements at 20 stations
Mean (SD): 71.1 (30.1)
Percentiles: 25th: 47.4
50th(Median): 67.6
75th: 89.3
Range (Min, Max): 18.4, 236.9
Monitoring Stations: 20
Copollutant (correlation):
1st trimester:
PMio—C0: 0.47
PM10-SO2: 0.78
PM10-NO2: 0.66
2nd trimester:
PM10-CO: 0.68
PM10-SO2: 0.82
PM10-NO2: 0.81
3,d trimester:
PM10-CO: 0.69
PM10-SO2: 0.85
PM10-NO2: 0.80
PM Increment: IQR, 41.9
RR Estimate [Lower CI, Upper CI]:
1 st trimester: 1.03 [1.00,1.07]
2nd trimester: 1.04(1.00,1.08]
3rd trimester: 1.00 (0.95,1.04]
All trimesters: 1.06 (1.01,1.10]
Low exposure in last 5 months using IQR
during last 5 months: 0.94 (0.85,1.05]
Low exposure in first 5 months using IQR
during first 5 months: 1.04 [1.01,1.08]
Notes: Birth weight was decreased by
19.6 g for an IQR increase in the 2nd
trimester.
The OR for LBW increased for female
children, fourth or higher order child,
mother < 20 yrs of age, and low
parental education level.
July 2009
E-578
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Leem et al. (2006, 089828)
Period of Study: 2001-2002
Location: Incheon, Korea
Outcome (ICD9 and ICD10): Age
Groups: Pre term delivery
Study Design: Cross-sectional
N: Cases: 2,082
Controls: 50,031
Statistical Analyses: Log-binomial
regression (corrected for overdispersion
used the log link function)
Covariates: Maternal age, parity, sex,
season of birth, and education level of
each parent
Season: Controlled as a covariate
Dose-response Investigated? Yes,
assessed quartiles of exposure
Statistical Package: NR
Pollutant: PMio
Averaging Time: Trimesters (daily
hourly data used to calculate)
Range (Min, Max): Reported ranges
within quartiles by trimester:
1st Trimester:
4: 64.57-106.39
3: 53.84-64.56
2: 45.95-53.83
1: 26.99-45.94
3,d Trimester:
4: 65.63-95.91
3: 56.07-65.62
2: 47.07-56.06
1: 33.12-47.06
Monitoring Stations: 27 monitoring
stations
pollutant levels for each area were
predicted from the levels recorded at the
monitors using ordinary block kriging
Copollutant (correlation):
SO? (r - 0.13)
NO? (r - 0.37)
CO (r - 0.27)
Effect Estimate [Lower CI, Upper CI]:
Crude and Adjusted RR for preterm
delivery and exposure during the 1st
trimester
Crude
Quartiles of exposure:
4:1.07(0.95,1.21)
3: 1.02(0.90,1.15)
2: 1.06(0.94,1.20)
1: 1.00
Adjusted
Quartiles of exposure:
4:1.27(1.04,1.56)
3: 1.13(0.94,1.37)
2: 1.14(0.97,1.34)
1: 1.00
p-trend: 0.39
Crude and Adjusted RR for preterm
delivery and exposure during the 3,d
trimester
Crude
Quartiles of exposure:
4: 1.06(0.94,1.20)
3: 1.06(0.94,1.19)
2: 1.05(0.93,1.18)
1: 1.00
Adjusted
Quartiles of exposure:
4: 1.09(0.91,1.30)
3: 1.04 (0.90,1.21)
2: 1.05(0.91,1.20)
1: 1.00
p-trend: 0.33
July 2009
E-579
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lin et al. (2004, 095787)
Period of Study: 1198-12100
Location: Sao Paulo, Brazil
Outcome: Neonatal death
Age Groups: Neonates (infants 0-28
days after birth)
Study Design: Time series
N: 1096 days, 6697 deaths
Statistical Analyses: Poisson regression
(GAM)
Covariates: Non-parametric LOESS
smoothers to control for: time (long term
trend), temperature, humidity, and day of
week
Also controlled for holidays with linear
term
Season: All
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: Lag 0, "moving
averages from 2 to 7 days"
Notes: No explicit control for season
apart from temperature
Pollutant: PMio
Averaging Time: Daily values
Mean (SD): 48.62 (21.18)
Range (Min, Max): 13.9,157.3
Monitoring Stations: NR (indicated
more than 1)
Copollutant (correlation):
CO r - 0.71
NO2 r - 0.76
SO? r - 0.80
O3 r = 0.36
PM Increment: 1 /yglm3
Log relative rate (standard error)
lag
Single pollutant model
0.0017 (0.0008)
lag 0
This translates to a 4.0% [95% CI: 0.3,
7.9] increase in neonatal mortality for a
23.3/yg/m3 increase in PM10
Two-pollutant model
0.0000 (0.0011)
lag 0
Notes: -In two pollutant model with PM10
and SO2 (which are highly correlated),
effect of PM disappeared and effect of
SO2 remained constant
-	results from pollutant moving averages
from 2 to 7 days not reported, authors
indicate effects only found for lag 0
(same day levels)
-	confidence intervals reported in abstract
are incompatible with fBsfstandard errors
and plotted results in text: abstract
indicates a 4% increase in mortality with
95% CI: 2-6 for a 23.3/yglm3 increase in
PM10
July 2009
E-580
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Lin et al„ 2004, 089827)
Period of Study: 1995-1997
Location: Taipei and Kaoshiung, Taiwan
Outcome: Low birth weight (<2500
grams)
Age Groups: newborns
Study Design: Cross-sectional
N: 92,288 infants
Statistical Analyses: Logistic
regression
Covariates: Gender, birth order,
gestational weeks, season of birth,
maternal age, maternal education,
copollutants
Season: All
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: The 9-month
pregnancy period for each infant, and
each trimester
Pollutant: PMio
Averaging Time: NR, "daily
measurements"
Mean (SD): Reported by monitoring
station:
1.48.78
2.	46.29
3.	48.79
4.	50.80
5.	52.54
Kaohsiung
1.	69.99
2.	63.39
3.	64.89
4.	75.79
5.	77.27
Monitoring Stations:
10 (5 in each city)
Notes: All pregnant womenfinfants
included in study lived within 3 km of an
air quality monitoring station
Pollution assigned based on nearest air
quality station to the maternal residence
Co-pollutant:
CO, SO?, 03, NO?
PM Increment: Tertiles
Entire pregnancy
T1: <46.4 ppb
T2: 46.4-63.1 ppb
T3: >63.1 ppb
First trimester
T1: <45.8 ppb
T2: 45.8-67.6 ppb
T3: > 67.6 ppb
Second trimester
T1: <44.6 ppb
T2: 44.6-64.2 ppb
T3: >64.2 ppb
Third trimester
T1: <43.7 ppb
T2: 43.7-63.7 ppb
T3: >63.7 ppb
RR Estimate [Lower CI, Upper CI]
Entire pregnancy
T1: 1.00
T2: 0.96 [0.83,1.11]
T3: 0.87 [0.71,1.05]
First trimester
T1: 1.00
T2: 0.96 [0.84,1.09]
T3: 0.97 [0.80,1.17]
Second trimester
T1: 1.00
T2: 1.03 [0.90,1.17]
T3: 1.00 [0.83,1.21]
Third trimester
T1: 1.00
T2: 1.02 [0.90,1.16]
T3: 0.97 [0.81,1.17]
Notes: RR for births in Kaoshiung vs.
Taipei: 1.13 [1.03,1.24]
Reference: Lipfert et al. (2000, 004103) Outcome: Infant mortality
Period of Study: 1990
Location: U.S.
Pollutant: PM10
including respiratory mortality (traditional Averaging Time: Yearly avg used
definition, ICD9 460-519), expanded
definition (adds ICD9 769 and 770)
Age Groups: Infants
Study Design: Cross-sectional
N: 2,413,762 infants in 180 counties (Ns
differ for various models)
Statistical Analyses: Logistic
regression
Covariates: mother's smoking,
Mean (SD): 33.1 (9.17) (based on 180
counties)
Range (Min, Max): (16.9, 59)
Monitoring Stations: NR
Copollutant (correlation):
PM10
SO42- (r - 0.10)
J\ISPJVho^norvsulfate_^ortion_o^PMio__
PM Increment: NR (present regression
coefficients)
Effect Estimate [Lower CI, Upper CI]:
Presented regression coefficients
(standard errors)
(3 PM exposures regressed jointly)
bold - p < 0.05
Cause of death: All
Birth weight: All
PM10: 0.0114(0.0015)
July 2009
E-581
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
education, marital status, and race
month of birth
and county avg heating degree days
Dose-response Investigated? NR
Statistical Package: NR
(r - 0.91)
CO (r - 0.27)
SO? (r - 0.04)
Notes: TSP-based sulfate was adjusted
for compatibility with the PMio-based
data
S042 :-0.0002 (0.0061)
NSPMio: 0.0115(0.0014)
Cause of death: All
Birth weight: LBW
PMio: 0.0088 (0.0019)
S042-: 0.0265 (0.0080)
NSPMio: 0.0086 (0.0020)
Cause of death: All
Birth weight: normal
PMio: 0.0092 (0.0024)
S042-: -0.0488 (0.0098)
NSPMio: 0.0096 (0.0024)
Cause of death: All neonatal
Birth weight: All
PMio: 0.0126 (0.0018)
S042-: 0.0267 (0.0076)
NSPMio: 0.0126 (0.0018)
Cause of death: All neonatal
Birth weight: LBW
PMio: 0.0086 (0.0022)
S042-: 0.0388 (0.0088)
NSPMio: 0.0093 (0.0022)
Cause of death: All neonatal
Birth wt: normal
PMio: 0.0123 (0.0041)
S042 :-0.0334 (0.0169)
NSPMio: 0.0125 (0.0040)
Cause of death: All post neonatal
Birth wt: All
PMio: 0.0091 (0.0024)
S042 :-0.0474 (0.0100)
NSPMio: 0.0096 (0.0024)
Cause of death: All post neonatal
Birth wt: LBW
PMio: 0.0096 (0.0043)
S042 :-0.0247 (0.0173)
NSPMio: 0.0101 (0.0042)
Cause of death: All post neonatal
Birth wt: normal
PMio: 0.0074 (0.0030)
S042 :-0.0569 (0.0121)
NSPMio: 0.0080 (0.0029)
Cause of death: SIDS
Birth weight: All
PMio: 0.0138 (0.0038)
July 2009
E-582
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
SO42 :-0.1078 (0.0151)
NSPM10: 0.0149 (0.0037)
Cause of death: SIDS
Birth weight: LBW
PM10: 0.0115(0.0088)
SO42 :-0.1378 (0.0337)
NSPM10: 0.0146 (0.0085)
Cause of death: SIDS
Birth weight: normal
PM10: 0.0137 (0.0042)
SO42 :-0.0995 (0.0168)
NSPM10: 0.0147 (0.0041)
Cause of death: All respiratory (ICD9:
460-519,769,770)
Birth weight: All
PM10: 0.0168 (0.0034)
SO42 :0.0706 (0.0146)
NSPM10: 0.0166 (0.0034)
Cause of death: All respiratory (ICD9:
460-519,769,770)
Birth weight: LBW
PM10: 0.0144(0.0038)
SO42 :0.0821 (0.0158)
NSPM10: 0.0139 (0.0038)
Cause of death: All respiratory (ICD9:
460-519,769,770)
Birth weight: normal
PM10: 0.0177 (0.0091)
SO42-: 0.0001 (0.0392)
NSPM10: 0.0118(0.0090)
Cause of death: Respiratory disease
(ICD9: 460-519)
Birth weight: All
PM10: 0.0133 (0.0089)
SO42-: 0.0093 (0.0384)
NSPM10: 0.0134 (0.0089)
Cause of death: Respiratory disease
(ICD9: 460-519)
Birth weight: LBW
PM10: 0.0092 (0.0137)
SO42-: 0.0434 (0.0580)
NSPM10: 0.0089 (0.0138)
Cause of death: Respiratory disease
(ICD9: 460-519)
Birth weight: normal
PM10: 0.0126 (0.0120)
SO42 :-0.0177 (0.0509)
July 2009
E-583
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Maisonet et al. (2001,
016624)
Period of Study: 1994-1996
Location: Northeastern U.S. (6 cities:
Boston, Hartford, Philadelphia,
Pittsburgh, Springfield, Washington DC)
Outcome: Low birth weight (LBW):
infants with a birth weight < 2,500 g
and having a gestational age between 37
and 44 weeks
Age Groups: Term live births (singleton)
Study Design: Cross-sectional
N: 89,557 infants
Statistical Analyses: Logistic
regression (LBW) and linear regression
(for reductions in birth weight)
Covariates: gestational age, gender,
birth order, maternal age, race/ethnicity,
years of education, marital status,
adequacy of prenatal care, previous
induced or spontaneous abortions, weight
gain during pregnancy, maternal prenatal
smoking, and alcohol consumption
season
Season: Yes, as covariate
Dose-response Investigated? Yes,
categorical exposure variables assessed
Pollutant: PMio
Averaging Time: Trimester averages
calculated using 24-h measurements
taken every 6 days
Range (Min, Max): Ranges for
categories of exposure:
1st Trimester
<	25th: <24.821
25 to < 50th: 24.821, 30.996
50 to < 75th: 30.997, 36.142
75 to < 95th: 36.143,46.547
> 95th: > 46.548
2nd Trimester
<	25th: <24.702
25 to < 50th: 24.702, 30.294
50 to < 75th: 30.295, 35.410
75 to < 95th: 35.411,43.928
NSPMio: 0.0128 (0.0119)
Associations with SIDS by smoking
status
Smoking status: Yes
Birth weight: Normal
PMio: 0.0202 (0.0073)
S042-: -0.0722 (0.0284)
NSPMio: 0.0206 (0.0071)
Smoking status: No
Birth weight: Normal
PMio: 0.0104 (0.0051)
S042 :-0.1 14 (0.021)
NSPMio: 0.0117 (0.005)
Smoking status: Yes
Birth weight: LBW
PMio: 0.0322 (0.0130)
S042-: -0.0958 (0.0483)
NSPMio: 0.0345 (0.0125)
Smoking status: No
Birth weight: LBW
PMio:-0.0044 (0.012)
S042 :-0.0172 (0.047)
NSPMio:-0.0007 (0.012)
Mean risks (95%CI) between post
neonatal SIDS among normal birth weight
babies
pollutants regressed one at a time
PMio: 1.20(1.02,1.42)
S042 :0.43 (0.37, 0.51)
NSPMio: 1.33 (1.18,1.50)
PM Increment: 10 /yglm3 for analyses
assessing exposures continuously
Effect Estimate [Lower CI, Upper CI]:
ORs for term LBW by trimester
T'Trimester Crude
<	25th: 1.00
25 to < 50th: 1.02 (0.90,1.14)
50 to < 75th: 0.90 (0.65,1.24)
75 to < 95th: 0.87 (0.58,1.30)
>	95th: 0.89 (0.60,1.33)
Continuous: 0.93 (0.77,1.13)
T'Trimester Adjusted
<	25th: 1.00
25 to < 50th: 1.02 (0.94,1.11)
50 to < 75th: 0.90 (0.78,1.03)
75 to < 95th: 0.85 (0.73,1.00)
>	95th: 0.83 (0.70,0.97)
July 2009
E-584
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Statistical Package: STATA
> 95th: > 43.929
Continuous: 0.93 (0.85,1.00)

3,d Trimester
2nd Trimester Crude

< 25th: <24.702
< 25th: 1.00

25 to < 50th: 24.702, 30.162
25 to < 50th: 1.01 (0.93,1.10)

50 to < 75th: 30.163, 35.642
50 to < 75th: 0.90 (0.66,1.21)

75 to < 95th: 35.643,43.588
75 to < 95th: 0.92 (0.62,1.34)

> 95th: > 43.589
> 95th: 0.90 (0.61,1.33)

Monitoring Stations: 3-4 per city
Continuous: 0.95 (0.78,1.16)

Copollutants: CO, SO2
2nd Trimester Adjusted


< 25th: 1.00


25 to < 50th: 1.06 (0.97,1.15)


50 to < 75th: 0.95 (0.85,1.07)


75 to < 95th: 0.91 (0.79,1.05)


> 95th: 0.77 (0.63,0.95)


Continuous: 0.93 (0.85,1.02)


3,d Trimester Crude


< 25th: 1.00


25 to < 50th: 0.94 (0.85,1.05)


50 to < 75th: 0.86 (0.58,1.25)


75 to < 95th: 0.86 (0.57,1.29)


> 95th: 0.92 (0.61,1.38)


Continuous: 0.95 (0.75,1.20)


3,d Trimester Adjusted


< 25th: 1.00


25 to < 50th: 0.98 (0.87,1.10)


50 to < 75th: 0.92 (0.76,1.11)


75 to < 95th: 0.88 (0.75,1.04)


> 95th: 0.91 (0.77,1.07)


Continuous: 0.96 (0.88,1.06)


Adjusted ORs by race/ethnicity


Whites:


1st Trimester


< 25th: 1.00


25 to < 50th: 1.13 (0.96,1.33)


50 to < 75th: 1.00 (0.92,1.08)


75 to < 95th: 1.00 (0.91,1.09)


> 95th: 0.92 (0.81,1.04)


Continuous: 0.94 (0.90, 0.98)


2nd Trimester


< 25th: 1.00


25 to < 50th: 0.88 (0.77,1.02)


50 to < 75th: 0.95 (0.89,1.02)


75 to < 95th: 0.95 (0.84,1.07)


> 95th: 0.89 (0.64,1.26)


Continuous: 0.96 (0.89,1.04)
July 2009
E-585
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
3,d Trimester
<	25th: 1.00
25 to < 50th: 0.84 (0.64,1.11)
50 to < 75th: 0.91 (0.83,1.01)
75 to < 95th: 0.80 (0.71,0.90)
>	95th: 1.03 (0.86,1.24)
Continuous: 0.95 (0.90,1.00)
African Americans:
1st Trimester
<	25th: 1.00
25 to < 50th: 1.01 (0.98,1.05)
50 to < 75th: 0.88 (0.79, 0.98)
75 to < 95th: 0.83 (0.70, 0.97)
>	95th: 0.81 (0.67,0.99)
Continuous: 0.93 (0.85,1.01)
2nd Trimester
<	25th: 1.00
25 to < 50th: 1.10 (0.93,1.30)
50 to < 75th: 0.95 (0.80,1.12)
75 to < 95th: 0.88 (0.69,1.11)
>	95th: 0.75(0.54,1.03)
Continuous: 0.92 (0.80,1.05)
3,d Trimester
<	25th: 1.00
25 to < 50th: 1.08 (0.92,1.27)
50 to < 75th: 0.89 (0.70,1.12)
75 to < 95th: 0.94 (0.75,1.18)
>	95th: 0.83 (0.71,0.97)
Continuous: 0.99 (0.87,1.11)
Hispanics:
1st Trimester
<	25th: 1.00
25 to < 50th: 0.83 (0.64,1.06)
50 to < 75th: 0.86 (0.70,1.05)
75 to < 95th: 0.79 (0.68, 0.93)
>	95th: 1.36 (1.06,1.75)
Continuous: 0.96 (0.84,1.09)
2nd Trimester
<	25th: 1.00
25 to < 50th: 1.16 (0.84,1.61)
50 to < 75th: 0.86 (0.63,1.19)
75 to < 95th: 0.98 (0.71,1.34)
>	95th: 0.68 (0.38,1.21)
Continuous: 0.92 (0.81,1.05)
3,d Trimester
July 2009
E-586
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
< 25th: 1.00
25 to < 50th: 0.77 (0.55,1.07)
50 to < 75th: 1.12 (0.76,1.66)
75 to < 95th: 0.93 (0.65,1.31)
> 95th: 0.90 (0.55,1.47)
Continuous: 0.96 (0.80,1.15)
Reference: Mannes et al.(2005,
087895)
Period of Study: January 1,1998-
December 31, 2000
Outcome: Risk of SGA and birth weight
Age Groups: all singleton births > 20
weeks and a 400 grams birth weight and
maternal all ages
Location: Metropolitan Sydney, Australia Study Design: Cross-sectional
N: 138,056 singleton births
Statistical Analyses: Logistic and linear
regression models
Covariates: sex of child, maternal age,
gestational age, maternal smoking,
gestational age at first antenatal visit,
maternal indigenous status, whether first
pregnancy, season of birth,
socioeconomic status
Season: All seasons
included as covariate
Dose-response Investigated? No
Statistical Package: SAS v8.02
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): 16.8 (7.1)
25th: 12.3
50th(Median): 15.7
75th: 19.9
Range (Min, Max): (3.8-104.0)
Monitoring Stations: up to 14
Copollutants (correlations): CO:
r - 0.26
NO?: r - 0.47
0s: r - 0.52
PM2.6: r - 0.81
PM Increment: 1 /yglm3
Risk of SGA
All births
One month before birth: OR - 1.01
(1.00-1.03)
Third trimester: OR - 1.00 (0.99-1.013)
Second trimester: OR - 1.01 (1.00-1.04)
First trimester: OR - 1.00 (0.98-1.02)
5 km births
One month before birth: OR - 1.00
(0.99-1.02)
Third trimester: OR - 1.01 (0.99-1.02)
Second trimester: OR - 1.02 (1.01-1.03)
First trimester: OR - 1.01 (0.99-1.02)
Change in birth weight
All births
One month before birth: B - -1.21 (-2.31-
¦0.11)
Third trimester: B - -0.95 (-2.30-0.40)
Second trimester: B - -2.05 (-3.36- ¦
0.74)
First trimester: B - -0.14 (-1.37- 1.09)
5 km births
One month before birth: B - -2.98 (-4.25-
¦1.71)
Third trimester: B - -3.84 (-5.35- -2.33)
Second trimester: B - -4.28 (-5.79- ¦
2.77)
First trimester: B - -2.57 (-4.04- -1.10)
Key second trimester findings
Single pollutant model: B - -4.28 (-5.79- ¦
2.77)
2 pollutant (PM10 and CO): B - -3.72
(-6.29-1.15)
2 pollutant (PM10 and NO2): B - -2.65
(-4.32- -0.98)
2 pollutant (PM10 and Osi: B - -5.47
(-7.06-3.88)
4 pollutant (PM10, NO2, CO and O3): B — -
3.27 (-7.05-0.51)
Controlling for exposures in other
pregnancy periods: B - -3.03 (-4.85- ¦
1.21)
July 2009
E-587
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Pereira et al. (1998, 007264) Outcome: Intrauterine mortality (fetuses
over 28 weeks of pregnancy)
Period of Study: Jan 1991-Dec 1992
Location: Sao Paulo, Brazil
Notes: Paper does not focus on PM as a
pollutant of interest.
Study Design: Time-series
N: 730 days with PM measures
Statistical Analyses: Poisson regression
Covariates: Season, day of the week
and weather (temperature and relative
humidity)
Season: Assessed by including 24
indicator variables for month and year
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: Paper focuses on
other pollutants (lags for PM not
reported)
Pollutant: PMio
Averaging Time: 24 hr mean
Mean (SD): 65.04 (27.28)
Range (Min, Max): (14.80,192.80)
Monitoring Stations: 13 (averaged to
provide city-wide pollutant level)
Copollutants (correlation): NO2
(r - 0.45)
SO? (r - 0.74)
CO (r - 0.41)
0s (r - 0.25)
PM Increment: NR (reported only
regression coefficients for PM)
Effect Estimate [Lower CI, Upper CI]:
Regression coefficients (standard errors)
for pollutants when considered separately
and simultaneously in the completed
model:
Separately: 0.0008 (0.0006)
Simultaneously: -0.0005 (0.0010)
Reference: Ritz et al. (2000, 012068)
Period of Study: 1989-1993
Location: Southern California
Outcome: Preterm birth (treated
dichotomously as birth at < 37 weeks
gestation
also analyzed continuously)
Age Groups: infants (born vaginally
between 26-44 weeks of gestation)
Study Design: Cross-sectional
N: 97,158 births
Statistical Analyses: Logistic and linear
regression
Covariates: maternal age, race,
education, parity, interval since the
previous live birth, access to prenatal
care, infant sex, previous low weight or
preterm births, smoking (reported as
"pregnancy complications")
to examine effect modification, authors
conducted stratified analysis by region,
birth and conception seasons, maternal
age, race, education, and infant gender
Season: Some models included season of
birth or conception
also assessed as effect modifier in
stratified analyses
Dose-response Investigated? Examined
adequacy of linear or log-linear relation
using indicator terms for pollutant-avg
quartiles
results presented in Fig 2 (dose-response
demonstrated for last 6 weeks exposure
period)
Statistical Package: NR
Pollutant: PM10
Averaging Time: 24-h averages at 6 day
intervals
averaged pollutant measures for 1, 2, 4,
6, 8,12, and 26 weeks before birth and
the whole pregnancy period
Mean (SD): 6 weeks before birth: 47.5
(15.0)
1st month of pregnancy: 49.3 (16.9)
Range (Min, Max): 6 weeks before birth:
12.3-152.3
1st month of pregnancy: 9.5-178.8
Monitoring Stations: 17 stations (PM
measured at only 8 stations)
Copollutants (correlations):
6 weeks before birth: CO (r - 0.43)
NO2 (r - 0.74)
O3 (r - 0.20)
1st month of pregnancy: CO (r - 0.37)
NO2 (r - 0.71)
O3 (r - 0.23)
Notes: Averaged pollutant measures
taken at the air monitoring station closest
to the residence
PM Increment: 50/yg/m3
Effect Estimate [Lower CI, Upper CI]:
All 8 stations
6 weeks before birth
Crude: 1.20 (1.09,1.33)
2 exposure periods: 1.18 (1.07,1.31)
Other risk factors: 1.15 (1.04,1.26)
Other RFs plus season: 1.15 (1.03,1.29)
Multipollutant model: 1.19 (1.01,1.40)
1st month of pregnancy
Crude: 1.16(1.06,1.26)
2 exposure periods: 1.13 (1.04,1.24)
Other risk factors: 1.09 (1.00,1.19)
Other RFs plus season: 1.09 (0.99,1.20)
Multipollutant model: 1.12 (0.97,1.29)
Coastal stations only
6 weeks before birth
Crude: 1.22 (1.00,1.49)
2 exposure periods: 1.28 (1.04,1.56)
Other risk factors: 1.13 (0.93,1.38)
Other RFs plus season: 1.18 (0.92,1.51)
Multipollutant model: 1.42 (097, 2.01)
1st month of pregnancy
Crude: 1.28 (1.06,1.54)
2 exposure periods: 1.32 (1.09,1.59)
Other risk factors: 1.17 (0.97,1.40)
Other RFs plus season: 0.99 (0.79,1.24)
Multipollutant model: 1.09 (0.83,1.41)
Inland stations only
6 weeks before birth
Crude: 1.27 (1.12,1.44)
2 exposure periods: 1.27 (1.11,1.44)
July 2009
E-588
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Other risk factors: 1.19 (1.05,1.35)
Other RFs plus season: 1.27 (1.10,1.48)
Multipollutant model: 1.18 (0.97,1.43)
1st month of pregnancy
Crude: 1.16(1.04,1.29)
2 exposure periods: 1.16 (1.04,1.29)
Other risk factors: 1.09 (0.98,1.21)
Other RFs plus season: 1.09 (0.97,1.24)
Multipollutant model: 1.11 (0.93,1.33)
Crude estimates for last 6 weeks
exposure by season
Fall: 1.08 (0.88,1.31)
Summer: 1.06 (0.87,1.29)
Winter: 1.33(1.07,1.65)
Spring: 1.81 (1.41,2.31)
Reduction in mean gestation length
for each increase in PMio during last 6
weeks before birth (linear regression
analysis)
Crude: 0.66 (± 0.24) days
Adj: 0.90 (± 0.27) days
Notes: Effect estimates remain stable
when excluding SGA or LBW children or
when restricting preterm births to SGA or
LBW children only (results not presented)
July 2009
E-589
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ritz, et al. (2002, 023227)
Period of Study: 1987-1993
Location: Southern California
(July 1990—July 1993 for Los Angeles,
1989 for Riverside, 1988-1989 for San
Bernardino, and 1987-1989 for Orange
counties
Outcome: 1) aortic defects
2)	defects of the atrium and atrium
septum
3)	endocardial and mitral valve defects
4)	pulmonary artery and valve defects
Pollutant: PMio
Averaging Time: 24 h (every 6 days)
PM Component: vehicle emissions
Monitoring Stations: 11 (for PMio)
Notes: The authors did not observe
consistently increased risks and dose-
response patterns for PMio after
controlling for the effects of CO and
ozone on these cardiac defects.
(Quantitative results not shown).
Copollutants (correlations): CO: r ¦
5) conotruncal defects including tetralogy 0.32
of Fallot, transposition of great vessels,
truncus arteriosus communis, double
outlet right ventricle, and
aorticopulmonary window
and 6) ventricular septal defects not
included in the conotruncal category.
Age Groups: all live born infants and
fetal deaths diagnosed between 20
weeks of gestation and 1 year after birth
Study Design: case-control
N: 10,649 infants and fetuses
Statistical Analyses: hierarchical (two-
level) regression model, polytomous
logistic regression, linear model
Covariates: gender, no prenatal care,
multiple births, no siblings, maternal race,
maternal age, maternal education, born
before 1990, season of conception,
Season: all
Dose-response Investigated? Yes, for
ozone and CO, study found a clear dose-
response pattern for aortic septum and
valve and ventricular septal defects and
possibly for conotruncal and pulmonary
artery and valve defects
Statistical Package: SAS
N0?(NR)
03 (NR)
Reference: Ritz et al. (2006, 089819)
Period of Study: 1989-2000
Location: 389 South Coast Air Basin
(SoCAB) zip codes
Outcome: total infant deaths during the
first year of life as well as all respiratory
causes of death (ICD-9 codes 460-519,
769, 770.4,770.7, 770.8, and 770.9
and ICD-10 codes J00-J98, P22.0,
P22.9, P27.1, P27.9, P28.0, P28.4,
P28.5, and P28.9) and sudden infant
death syndrome (SIDS) (ICD-9 code 798.0
and ICD-10 code R95).
Age Groups: infants 0-1 yr
Study Design: Case-control
N: 2,975,059 births and 19,664 infant
deaths
Cases, n - 13,146
Controls, n - 151,015
Statistical Analyses: Conditional
logistic regression analysis
Covariates: risk factors available on
birth and/or death certificates (maternal
age, race/ethnicity, and education, level
of prenatal care, infant gender, parity,
birth country, and death season)
Season: Death season (spring, summer,
autumn, winter)
Dose-response Investigated? Yes
Statistical Package: NR
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): Two weeks before death:
46.2
One month before death: 46.3
Two months before death: 46.3
Six months before death: 46.3
Range (Min, Max): Two weeks before
death: (21.0-83.5)
One month before death: (25.0-77.2)
Two months before death: (27.6-74.2)
Six months before death: (31.3-69.5)
Monitoring Stations: maximum of 31
Copollutants (correlation): Two weeks
before death
CO: r - 0.33
NO?: r - 0.48
Os:r - 0.12
One month before death
CO: r - 0.33
NO?: r - 0.48
Os:r - 0.12
PM Increment: 10 /yglm3
Effect Estimate [Lower CI, Upper CI]:
All-cause death
2 mo before death
Single-pollutant model:
<	25th - 1.04(1.01-1.06)
25,h-75,h - 0.96 (0.89-1.04)
>75,h - 1.14(1.03-1.27)
Multiple-pollutant model:
<	25th - 1.02 (0.99-1.05)
25th-75th - 0.92 (0.84-1.00)
>75,h - 1.07 (0.95-1.20)
SIDS
2 mo before death:
Single-pollutant model:
<	25th - 1.03 (0.99-1.08)
25th-75th - 0.94 (0.81-1.08)
> 75th - 1.13(0.93-1.36)
Multiple-pollutant model:
<	25th - 1.01 (0.95-1.07)
25th-75th - 0.90 (0.76-1.06)
July 2009
E-590
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Two months before death
> 75th - 0.99 (0.80-1.24)
CO: r - 0.32
Respiratory death
NO?: r - 0.48
2 wl< before death
Os:r - 0.12
Postneonatal deaths (28 d to 1 y)
Six months before death
Single-pollutant model:
CO: r - 0.29
< 25th - 1.05 (1.01-1.10)
NO?: r - 0.44
25th-75th - 1.13(1.01-1.10)
Os:r - 0.16
> 75th - 1.46 (1.13-1.88)

Multiple-pollutant model:

< 25th - 1.04(0.98-1.09)

25th-75th - 1.09 (0.86-1.38)

> 75th - 1.40 (1.03-1.89)

Postneonatal deaths (28 d to 3 mo)

Single-pollutant model:

< 25th - 1.01 (0.95-1.08)

25th-75th - 1.16(0.82-1.63)

>75,h - 1.44(0.96-2.17)

Multiple-pollutant model:

< 25th - 1.00 (0.92-1.09)

25,h-75,h - 0.97 (0.67-1.42)

> 75th - 1.23 (0.76-2.00)

Post neonatal deaths (4-12 mo)

Single-pollutant model:

< 25th - 1.12(1.02-1.23)

25th-75th - 1.08 (0.81-1.44)

>75,h - 1.41 (1.02-1.96)

Multiple-pollutant model:

< 25th - 1.07 (1.00-1.15)

25th-75th - 1.02 (0.75-1.40)

> 75th - 1.36 (0.92-2.01)
July 2009
E-591
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Rogers et al. (2006, 091232)
Period of Study: 1986-1988
Location: Georgia, USA
Outcome: VLBW
Term, AGA, Preterm AGA, Preterm, SGA
Age Groups: Newborns and their
mothers (< 19 to a 35-years-old)
Study Design: case-control
N: 325 infants (69 preterm SGA
59 preterm AGA
197 term AGA) and their mothers
Statistical Analyses: logistic regression
Covariates: maternal age, maternal race,
maternal education, active and passive
smoking, birth season, prepregnancy
weight, pregnancy weight gain, maternal
toxemia, anemia, asthma
Dose-response Investigated? Yes, used
Statistical Package: SUDAAN
Cochran-Armitage test for trend to
determine whether the observed
proportions of cases and controls differed
in a linear manner across exposure
categories.
Pollutant: PMio
Averaging Time: annual
Preterm SGA:
50th(Median): 3.38
Preterm AGA:
50th(Median): 7.84
Term AGA:
50th(Median): 3.23
Monitoring Stations: NR
Percent Mothers Residing In County
With Industrial Point Source
Preterm SGA: 60.9%
Preterm AGA: 79.7%
Term AGA: 60.4%
Percent Mothers Residing In Pmio
Quartile (based on environmental
transport model)
Preterm SGA
1stquartile!<1.48): 31.9%
2nd quartile (1.48-3.74): 18.8%
3rd quartile (3.75-15.07): 26.1%
4th quartile (>15.07): 23.2%
Preterm AGA
1st quartile! <1.48): 16.9%
2nd quartile (1.48-3.74): 22.1%
3rd quartile (3.75-15.07): 28.8%
4th quartile (>15.07): 32.2%
Term AGA
1st quartile (< 1.48): 24.7%
2nd quartile (1.48-3.74): 28.4%
3rd quartile (3.75-15.07): 27.9%
4th quartile (> 15.07): 19.3%
PM Increment: Quartile
Notes: Statistically significant increases
in the odds of VLBW and preterm AGA
births are associated with living in a
county with a PMio point source. Preterm
AGA births are also associated with living
in an area with very high (4th quartile)
estimated PMio exposure.
Delivery of VLBW vs. Term AGA infant
County with point source
2.54(1.46, 4.22]
PMio quartile
1st quartile: reference
2nd quartile:
0.81 [0.42,1.55]
3rd quartile:
0.85(0.45,1.16]
4th quartile:
1.94 [0.98,3.83]
Delivery of Preterm AGA vs. Term AGA
infant
County with point source
4.31 [1.88: 9.87]
PMio quartile
1st quartile: reference
2nd quartile:
1.56(0.56: 4.35]
3rd quartile:
1.19(0.44: 3.23]
4th quartile:
3.68(1.44: 9.44]
Delivery of Preterm AGA vs. Preterm SGA
infant
County with point source
2.07(0.83: 5.16]
PMio quartile
1st quartile: reference
2nd quartile:
1.96(0.59: 6.43]
3rd quartile:
2.10(0.66: 6.73]
4th quartile:
2.58(0.78: 8.51]
July 2009
E-592
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Romieu et al. (2004,
093074)
Period of Study: 1997 to 2001
Location: Ciudad Juarez, Mexico
Outcome: Respiratory-related infant
mortality ICD9 (460-519)
ICD10 (J00-J99)
Age Groups: 1 month to 1 year
Study Design: Case crossover
N: 216 respiratory-related deaths
N - 412 other causes and N - 628 total
deaths
Statistical Analyses: The acute effects
of air pollution was modeled on both total
and respiratory-related mortality as a
function of the pollution levels on the
same day and preceding days and over
two- and three-day averages before the
date of death. Case-crossover with semi-
symmetric bidirectional referent selection
was the approach used. Data were
stratified by day of the week and
calendar month. Data were analyzed with
conditional logistic regression. Second
and third polynomial distributed lag
models were used to study lag structure.
BIC was used to determine lag length.
Covariate: temperature, season
Dose-response Investigated? Yes
Statistical Package: STATA 7.0
Lags Considered: 1-15 days
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): 1997: 33.04 (20.67)//gfm3
1998: 35.25 (17.32) //g/m3
1999: 45.92 (28.69)//g/m3
2000: 43.38 (23.77)//g/m3
2001:39.46 (29.43) //g/m3
Monitoring Stations: 5 stations in
Ciudad Juarez
2 stations in El Paso (close to US-Mexico
border)
Copollutant (correlation): 0:i: r - 0.01
Notes: Ciudad Juarez monitors measured
PMio every 6 days while El Paso monitors
measured on a daily basis.
PM Increment: 20//g/m3
RR Estimate [Lower CI, Upper CI]
lag:
Total mortality:
OR - 1.02 (0.94-1.11) lag 1
OR - 1.03 (0.95-1.12) lag 2
OR - 1.03 (0.94-1.13 ac2
OR - 1.04(0.95-1.15) ac3
Respiratory mortality
OR - 0.95 (0.83-1.09) lag 1
OR - 1.04 (0.91-1.19) lag 2
OR - 0.98(0.81-1.19) ac2
OR - 0.97 (0.74-1.26) ac3
Higher SES
OR - 0.82 (0.59,1.14) lag 1
OR - 1.08 (0.84,1.40) lag 2
OR - 0.89 (0.58,1.35) ac2
OR - 0.97 (0.52,1.82) ac3
Medium SES
OR - 0.99 (0.79,1.27) lag 1
OR - 1.11 (0.86,1.43) lag 2
OR - 1.03 (0.73,1.45) ac2
OR - 1.17(0.72,1.90) ac3
Lower SES
OR - 1.61 (0.97-2.66) lag 1
OR - 1.07 (0.65,1.75) lag 2
OR - 2.56 (1.06-6.17) ac2
OR - 1.76(0.59, 5.23) ac3
Notes:
ac2 and ac3 represent cumulative PMio
ambient levels over two or three days
before death.
July 2009
E-593
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Sagiv et al. (2005, 087468)
Period of Study: 1/1/1997-12/31/2001
Location: Allegheny county, Beaver
county, Lackawanna county, Philadelphia
county, Pennsylvania, U.S.A.
Outcome: Preterm birth (< 36 weeks)
Age Groups: Babies born between 20
and 44 weeks
Study Design: Time series
N: 3704 observation days, 187,997
births
Statistical Analyses: Poisson regression
multivariable mixed-effects model with a
random intercept for each county to
incorporate count-level information.
Covariates: Temperature, dew point
temperature, mean 6-week level of
copollutants (CO, NO2, and SO2), long-
term preterm birth trends
Season: All
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: 1, 2, 3, 4, 5, 6, 7
Pollutant: PM10
Averaging Time: Daily used to calculate
6-week period
Mean (SD): 6-week period
27.1 (8.3)
Daily
25.3(14.6)
Percentiles: 6-week period
50th (Median): 26.0 Daily
50th (Median): 21.6
Range (Min, Max): 6-week period: 8.7,
68.9
Daily: 2.0,156.3
Monitoring Stations: NR
Copollutant (correlation): Daily PM10-
daily SO2: r - 0.46
Also considered CO, NO2 and O3 as
copollutants.
PM Increment: 1) 50 /yglm3 2) Quartiles
(first quartile is the reference)
Exposure period: 6 weeks before birth
Per 50 //g/m3:1.07 (0.98, 1.18)
2nd quartile: 1.00 (0.95,1.05)
3rd quartile: 1.04 (0.99,1.09)
4th quartile: 1.03 (0.98,1.08)
Exposure period: 1-day acute time
windows Per 50/yglm3: 2-day lag: 1.10
(1.00,1.21)
5-day lag: 1.07 (0.98,1.18)
Notes: Within the article, authors provide
a Figure 1 displaying a graph of the
relative risk (RR) and 95% confidence
intervals (CI) for 1- to 7-day lags. While
the authors report the 2- and 5-day lag
RRs and 95% CIs in the text, the others
are not specifically reported. However,
the figure shows the approximate RRs
per 50 /yglm3 as indicated below: 1 -day
lag: 1.05
3-day	lag: 1.05
4-day	lag: 1.00
6-day	lag: 0.97
7-day	lag: 1.03
July 2009
E-594
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Salam et al„ 2005, 087885) Outcome: Birth weight
Period of Study: 1975-1987	Low birth weight (LBW
Location: Southern California	< 2500 g)
Intrauterine growth retardation (IUGR)
Age Groups: Children born full-term
(between 37 and 44 weeks)
Study Design: Cohort study
N: 3901 children
Statistical Analyses: Linear mixed-
effects
Logistic regression
Covariates: Maternal age, months since
last live birth, parity, maternal smoking
during pregnancy, SES, marital status at
childbirth, gestational diabetes, child's
sex, child's race/ethnicity, child's grade in
school (4th, 7th, and 10th), Julian day of
birth
Season: All
Dose-response Investigated? Yes
Statistical Package: SAS
Pollutant: PMio
Averaging Time: Monthly
Mean (SD): Entire pregnancy: 45.8
(12.9)
First trimester: 46.6 (15.9)
Second trimester: 45.4 (14.8)
Third trimester: 45.4 (15.5)
Monitoring Stations: 1 or 3 (See notes)
Copollutant (correlation): Entire
pregnancy
PMio-03[io-6i: r - 0.54
PM 10-03[24 hr): X = 0.20
PMio-NO?: r - 0.55
PMio-CO: r - 0.41
First trimester
PMio-03[io-6i: r - 0.54
PM 10-03[24 hr]: r = 0.34
PMio-NO?: r - 0.48
PMio-CO: r - 0.29
Second trimester
PMio-03[io-6i: r - 0.50
PM 10-03(24 hr): x = 0.27
PMio-NO?: r - 0.53
PMio-CO: r - 0.35
Third trimester
PMio-03[io-6i: r - 0.52
PM 10-03[24 hr]: x = 0.31
PMio-NO?: r - 0.52
PMio-CO: r - 0.37
Notes: Exposure estimates were
calculated by spatially interpolated
monthly averages which were based off
of three monitoring stations located
within 50 km of the ZIP code region of
maternal birth residences.
PM Increment: IQR (interquartile range)
Outcome: birth weight (g)
Single-pollutant model
Entire pregnancy
18//gfm3: -19.9 (-43.6, 3.8)
First trimester
20 //g/m3: -3.0 (-22.7,16.7)
Second trimester
19//gfm3: -15.7 (-36.1, 4.7)
Third trimester
20/yg/m3: -21.7 (-42.2 to -1.1)
Multipollutant model
(included O3
(24 hr) in model
third trimester exposure)
20//gfm3: -10.8 (-31.8,10.2)
Outcome: IUGR (ORs)
Single-pollutant model
Entire pregnancy
18/yg/m3: 1.1 (0.9,1.3)
First trimester
20/yg/m3: 1.0(0.9,1.2)
Second trimester
19/yg/m3: 1.0(0.9,1.2)
Third trimester
20/yg/m3: 1.1 (0.9,1.3)
Outcome: LBW
Single-pollutant model
Entire pregnancy
18/yg/m3: 1.3(0.8,2.2)
First trimester
20/yg/m3: 1.0(0.7,1.5)
Second trimester
19/yg/m3: 1.2(0.8,1.7)
Third trimester
20/yg/m3: 1.3(0.9,1.9)
Notes: Numbers reported for birth weight
outcome are the effects on birth weight
outcome (the change in birth weight in
grams) across the IQR (which vary
depending on air pollutant and duration of
exposure measurement).
July 2009
E-595
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Pollutant: PMio
Averaging Time: 0-9d, 10-14d and 70-
90d
PMio specific results are given in Figures
3-5. PMio was not significantly
correlated with sperm quality.
Reference: (Sokol et al„ 2006, 098539)
Period of Study: 1/1996-12/1998
Location: Los Angeles, California
Reference: (Suh et al„ 2007,157028)
Period of Study: 2001-2004
Location: Seoul, Korea
Outcome: Semen Quality
Study Design: Panel
Statistical Analysis: Univariate and
Multivariate Regression
Statistical Package: SAS
Age Groups: Males ranging 19-35 in age
Outcome: Birth weight
Age Groups: prenatal follow-up for
newborns
Study Design: based prospective cohort
study
N: 199 pregnant mothers
Statistical Analyses: ANC0VA,
generalized linear models
Covariates: infant's sex, maternal age,
maternal and paternal education, parity,
presence of illness during pregnancy,
delivery month, gestational age (squared)
Dose-response Investigated? Yes
Statistical Package: SAS
Mean (SD) Unit: 35.74 ± 13.83//gfm3
Copollutant (correlation): 03, NO2, CO
Pollutant: PMio
Averaging Time: 24-h
Mean (SD): 1st trimester: 76.41 (28.80)
2nd trimester: 77.84 (31.63)
3rd trimester: 95.61 (26.15)
Percentiles: 1st trimester
25th: 55.28
50th(Median): 71.09
75th: 92.38
2nd trimester
25th: 48.65
50th(Median): 72.36
75th: 108.00
3rd trimester
25th: 77.10
50th(Median): 96.35
75th: 116.68
Range (Min, Max):
1st trimester (21.00,151.65)
2nd trimester (31.45,139.13)
3rd trimester (23.45,172.75)
Monitoring Stations: 27
Copollutant:
CO
SO?
NO?
PM Increment: Trimester a 90th%ile
compared to < 90th%ile
Least-square (ANC0VA) mean (SE)
All Genotypes
1st trimester
< 90th%ile, N(%): 158 (90.3%): 3253
(37)
>	90th%ile, N(%): 17 (9.7%): 2841 (145)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.009
Adjusted, with CO: 0.041
Adjusted, with NO2: 0.092
Adjusted, with SO2: 0.012
2nd trimester
<90th%ile, N(%): 153 (89.5%): 3253
(39)
>90th%ile, N(%): 18(10.5%): 3026
(157)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.177
Adjusted, with CO: 0.203
Adjusted, with NO2: 0.151
Adjusted, with SO2: 0.151
3rd trimester
<90th%ile, N(%): 162 (90.5%): 3226
(38)
>	90th%ile, N(%): 17 (9.5%): 3122 (140)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.487
Adjusted, with CO: 0.748
Adjusted, with NO2: 0.420
Adjusted, with SO2: 0.466
Genotype Mspl TT
1st trimester
<90th%ile,N(%): 60 (34.3%): 3350
(64)
>	90th%ile, N(%): 5 (2.9%): 3001 (229)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.147
Adjusted, with CO: 0.186
July 2009
E-596
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Adjusted, with NO2: 0.430
Adjusted, with SO2: 0.155
2nd trimester
<90th%ile, N(%): 59 (34.5%): 3335
(66)
>	90th%ile, N(%): 6 (3.5%): 3281 (249)
p-Value for mean birth weight for a
90th%ile PM10 vs. for < 90th%ile PM10
Adjusted: 0.833
Adjusted, with CO: 0.833
Adjusted, with NO2: 0.778
Adjusted, with SO2: 0.806
3rd trimester
<	90th%ile, N(%): 61 (34.1%): 3327
(65)
>	90th%ile, N(%): 6 (3.4%): 3227 (300)
p-Value for mean birth weight for a
90th%ile PM10 vs. for < 90th%ile PM10
Adjusted: 0.749
Adjusted, with CO: 0.980
Adjusted, with NO2: 0.635
Adjusted, with SO2: 0.687
Genotype Mspl TC/CC
1st trimester
<90th%ile,N(%): 98 (56.0%): 3193
(48)
>	90th%ile, N(%): 12 (6.9%): 2799 (169)
p-Value for mean birth weight for a
90th%ile PM10 vs. for < 90th%ile PM10
Adjusted: 0.033
Adjusted, with CO: 0.073
Adjusted, with NO2: 0.150
Adjusted, with SO2: 0.036
2nd trimester
<	90th%ile, N(%): 94 (55.0%): 3200
(52)
>	90th%ile, N(%): 12 (7.0%): 2933 (176)
p-Value for mean birth weight for a
90th%ile PM10 vs. for < 90th%ile PM10
Adjusted: 0.161
Adjusted, with CO: 0.172
Adjusted, with NO2: 0.152
Adjusted, with SO2: 0.158
3rd trimester
<90th%ile, N(%): 101 (56.4%): 3165
(49)
>90th%ile, N(%): 11 (6.2%): 3087 (147)
p-Value for mean birth weight for a
90th%ile PM10 vs. for < 90th%ile PM10
July 2009
E-597
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Adjusted: 0.626
Adjusted, with CO: 0.978
Adjusted, with NO2: 0.551
Adjusted, with SO2: 0.614
Genotype Ncol llelle
1st trimester
<90th%ile,N(%): 87 (49.7%): 3244
(52)
>	90th%ile, N(%): 7 (4.0%): 2983 (232)
p-Value for mean birth weight for a
90th%ile PM10 vs. for < 90th%ile PM10
Adjusted: 0.289
Adjusted, with CO: 0.344
Adjusted, with NO2: 0.641
Adjusted, with SO2: 0.293
2nd trimester
<90th%ile,N(%): 82 (48.0%): 3243
(55)
>	90th%ile, N(%): 11 (6.4%): 3185 (207)
p-Value for mean birth weight for a
90th%ile PM10 vs. for < 90th%ile PM10
Adjusted: 0.790
Adjusted, with CO: 0.783
Adjusted, with NO2: 0.707
Adjusted, with SO2: 0.733
3rd trimester
<90th%ile,N(%): 90 (50.3%): 3239
(53)
>	90th%ile, N(%): 9 (5.0%): 2944 (198)
p-Value for mean birth weight for a
90th%ile PM10 vs. for < 90th%ile PM10
Adjusted: 0.161
Adjusted, with CO: 0.279
Adjusted, with NO2: 0.134
Adjusted, with SO2: 0.150
Genotype Ncol MeVal/ValVal
1st trimester
<	90th%ile, N(%): 71 (40.6%): 3262
(56)
>90th%ile, N(%): 10(5.7%): 2773 (171)
p-Value for mean birth weight for a
90th%ile PM10 vs. for < 90th%ile PM10
Adjusted: 0.009
Adjusted, with CO: 0.031
Adjusted, with NO2: 0.058
Adjusted, with SO2: 0.010
2nd trimester
<	90th%ile, N(%): 71 (41.5%): 3264
(61)
July 2009
E-598
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
>90th%ile, N(%): 7 (4.1%): 2862 (208)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.076
Adjusted, with CO: 0.093
Adjusted, with NO2: 0.063
Adjusted, with SO2: 0.061
3rd trimester
<90th%ile,N(%): 72 (40.2%): 3207
(58)
a 90th%ile, N(%): 8 (4.5%): 3262 (180)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.777
Adjusted, with CO: 0.607
Adjusted, with NO2: 0.843
Adjusted, with SO2: 0.791
Reference: Tsai et al. (2006, 098312)
Period of Study: 1994-2000
Location: Kaohsiung, Taiwan
Outcome: post neonatal mortality
Age Groups: infants more than 27 days
and less than 1 year
Study Design: Case-crossover study
N: 207 deaths
Statistical Analyses: Conditional
logistic regression
Covariates: temperature, humidity
Dose-response Investigated? No
Statistical Package: SAS, version 8.2
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): 81.45 /yglm3
Percentiles: 25th: 44.50
50th(Median): 79.20
75th: 111.50
Range (Min, Max): (20.50-232.00)
Monitoring Stations: 6
Copollutant: SO2
NO2
CO
0s
PM Increment: 67.00 /yglm3
Effect Estimate [Lower CI, Upper CI]:
OR - 1.040(0.340-3.177)
Note: Air pollution levels at the dates of
infant death were compared with air
pollution levels 1 week before and 1
week after death
a cumulative lag up to 2 previous days
was used to assign exposure.
Reference: (Wilhelm and Ritz, 2005,
188761)
Period of Study: 1994-2000
Location: Los Angeles County, California,
U.S.
Outcome: Term low birth weight (LBW)
(< 2500 g at a 37 completed weeks
gestation), Vaginal birth <37 completed
weeks gestation
Age Groups: LBW: a 37 com|
weeks
Preterm births: < 37 completed weeks
Study Design: Cross-sectional
N: For LBW: 136,134
For preterm birth:
106,483
Statistical Analyses: Logistic
regression
Covariates: Maternal age, maternal race,
maternal education, parity, interval since
previous live birth, level of prenatal care,
infant sex, previous LBW or preterm
infant, birth season, other pollutants (CO,
NO2, 0 i, PMio), gestational age (in birth
weight analysis)
Dose-response Investigated? Yes
Statistical Package: NR
Pollutant: PMio
Averaging Time:
24 hr (every 6 days)
Entire pregnancy
Trimesters of pregnancy
Months of pregnancy
6 weeks before birth
Mean (SD): First trimester: 42.2
Third trimester: 41.5
6 weeks before birth: 39.1
Range (Min, Max):
First trimester: 26.3, 77.4
Third trimester: 25.7, 74.6
6 weeks before birth: 13.0,103.7
Monitoring Stations:
Zip-code-level analysis: 8
PM Increment: 1) 10 /yg/m3
2) 3 levels:
a)	< 25%ile (reference)
b)	25%-75%ile
c)	> 75%ile
Incidence of LBW (third trimester
exposure)
<	32.8/yg/m3: 2.0(1.8, 2.2)
32.8	to < 43.4//g/m3: 2.0 (1.9, 2.1)
a 43.4/yg/m3: 2.2(2.0, 2.4)
Incidence of preterm birth (first
trimester exposure)
<	32.9/yglm3: 8.7 (8.3, 9.2)
32.9	to < 43.9//g/m3: 8.8 (8.5, 9.1)
>43.9/yg|m3: 8.6(8.1,9.0)
Incidence of preterm birth (6 weeks
before birth exposure)
<	31.8//g/m3: 8.8(8.4,9.3)
July 2009
E-599
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Address-level analysis: 6
Copollutant (correlation!:
First trimester: PM10 CO: r - 0.12
PMio-NO?: r - 0.29
PMio-03: r - -0.01
PMio-PM2.b: r - 0.43
Third trimester: PMio CO: r - 0.32
PMio-NO?: r - 0.45
PMio-(k r - -0.08
PMio-PM2.b: r - 0.52
6 weeks before birth: PMio CO: r -
0.36
PMio-NO?: r - 0.49
PM10-O3: r - -0.16
PM10-PM2.B: r - 0.60
31.8	to <44.1 //g/m3: 8.6 (8.3, 8.9)
> 44.1 //g/m3: 8.8 (8.4, 9.2)
Outcome: LBW
Exposure Period: Third trimester
Address-level analysis:
Single-pollutant model:
Distance ~ 1 mile
Per 10 //g/m3:1.22 (1.05,1.41)
33.4 to < 44.7 //g/m3:1.08 (0.76, 1.52)
>44.7//g/m3: 1.48 (1.00,2.19)
Multipollutant model:
Distance ~ 1 mile
Per 10 //g/m3:1.36 (1.12, 1.65)
33.4 to <44.7 //g/m3:1.16 (0.77, 1.74)
>44.7//g/m3: 1.58 (0.95, 2.62)
Single-pollutant model:
1 < distance ~ 2 mile
Per 10 //g/m3: 0.98 (0.90, 1.06)
33.4 to < 44.7 //g/m3: 0.95 (0.80, 1.13)
>44.7 //g/m3: 0.96 (0.78, 1.18)
Multipollutant model:
1	< distance ~ 2 mile
Per 10 //g/m3:1.02(0.92,1.14)
33.4 to < 44.7 //g/m3: 0.93 (0.77, 1.12)
>44.7 //g/m3: 1.02 (0.79, 1.32)
Single-pollutant model:
2	< distance ~ 4 mile
Per 10 //g/m3:1.03 (0.99, 1.08)
33.9	to < 45.0//g/m3:1.04 (0.96,1.14)
>45.0//g/m3: 1.08 (0.97, 1.20)
Multipollutant model:
2 < distance ~ 4 mile
Per 10//g/m3:1.04 (0.98,1.09)
33.9 to <45.0//g/m3:1.02 (0.92, 1.12)
>45.0//g/m3: 1.06 (0.93, 1.21)
Zip-code-level analysis
Single-pollutant model:
Per 10//g/m3:1.03 (0.97, 1.09)
33.2 to <43.6//g/m3: 0.98 (0.86, 1.1 1)
>43.6//g/m3: 1.03 (0.88, 1.21)
Multipollutant model:
Per 10 //g/m3:1.07 (0.99, 1.15)
33.2 to < 43.6//g/m3: 0.97 (0.85, 1.12)
>43.6//g/m3: 1.09 (0.90, 1.31)
Outcome: LBW
July 2009
E-600
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Exposure Period: Entire pregnancy
period
Address-level analysis:
Multipollutant model:
Per 10//g/m3:1.24 (0.91, 1.70)
Outcome: Preterm Birth
Exposure Period: First trimester of
pregnancy
Address-level analysis:
Single-pollutant model:
Distance ~ 1 mile
Per 10 //g/m3:1.00 (0.93, 1.09)
33.3 to <45.1 //g/m3:1.07 (0.90, 1.26)
>45.1 //g/m3: 1.12(0.91,1.38)
Multipollutant model:
Distance ~ 1 mile
Per 10 //g/m3:1.00 (0.90, 1.12)
33.3 to <45.1//g/m3:1.12 (0.92, 1.36)
>45.1//g/m3: 1.17 (0.90, 1.50)
Single-pollutant model:
1 < distance ~ 2 mile
Per 10 //g/m3:1.01 (0.97, 1.05)
33.7 to <45.3//g/m3:1.03 (0.95, 1.12)
>45.3//g/m3: 1.07 (0.97, 1.19)
Multipollutant model:
1	< distance ~ 2 mile
Per 10//g/m3:1.04 (0.99,1.10)
33.7 to <45.3//g/m3:1.07 (0.98, 1.17)
a 45.3 //g/m3: 1.13 (1.00, 1.27)
Single-pollutant model:
2	< distance ~ 4 mile
Per 10 //g/m3:1.01 (0.99, 1.03)
34.1 to <45.5//g/m3:1.03 (0.99, 1.08)
>45.5//g/m3: 1.02 (0.96, 1.07)
Multipollutant model:
2 < distance ~ 4 mile
Per 10//g/m3: 0.99 (0.97, 1.02)
34.1 to <45.5//g/m3: 0.99 (0.95,1.04)
>45.5//g/m3: 0.94 (0.89,1.01)
Zip-code-level analysis
Single-pollutant model:
Per 10//g/m3: 0.99 (0.96, 1.01)
33.3 to <44.2//g/m3:1.01 (0.95, 1.08)
>44.2//g/m3: 0.98 (0.90, 1.05)
Multipollutant model:
Per 10//g/m3: 0.99 (0.96, 1.03)
July 2009
E-601
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
33.3 to <44.2//g/m3:1.03 (0.97, 1.1 1)
a 44.2 //g/m3: 1.01 (0.92,1.11)
Outcome: Preterm birth
Exposure Period: 6 weeks before birth
Address-level analysis:
Single-pollutant model:
Distance ~ 1 mile
Per 10 //g/m3:1.02 (0.95, 1.10)
32.5 to < 44.8//g/m3:1.09 (0.92, 1.29)
>44.8//g/m3: 1.12 (0.92, 1.37)
Multipollutant model:
Distance ~ 1 mile
Per 10 //g/m3:1.06 (0.97, 1.16)
32.5 to < 44.8//g/m3:1.09 (0.90, 1.31)
>44.8//g/m3: 1.17 (0.91,1.49)
Single-pollutant model:
1 < distance ~ 2 mile
Per 10//g/m3:1.00 (0.96, 1.03)
32.3 to < 45.3//g/m3: 0.99 (0.91, 1.07)
>45.3//g/m3: 0.99 (0.89, 1.10)
Multipollutant model:
1	< distance ~ 2 mile
Per 10 //g/m3:1.01 (0.97, 1.06)
32.3 to <45.3//g/m3:1.00 (0.92, 1.10)
>45.3//g/m3: 1.02 (0.91,1.16)
Single-pollutant model:
2	< distance ~ 4 mile
Per 10 //g/m3: 0.99 (0.98, 1.01)
33.1 to <45.3//g/m3:1.00 (0.96, 1.05)
>45.3//g/m3: 0.98 (0.93, 1.03)
Multipollutant model:
2 < distance ~ 4 mile
Per 10 //g/m3:1.00 (0.98, 1.02)
33.1 to <45.3//g/m3:1.01 (0.96, 1.05)
>45.3//g/m3: 0.98 (0.92,1.04)
Zip-code-level analysis
Single-pollutant model:
Per 10 //g/m3:1.02(0.99,1.04)
32.1 to <44.3//g/m3:1.01 (0.95, 1.07)
>44.3//g/m3: 1.04 (0.96,1.12)
Multipollutant model:
Per 10 //g/m3:1.02 (0.99, 1.06)
32.1 to <44.3//g/m3:1.02 (0.95, 1.09)
a 44.3 //g/m3: 1.04 (0.95,1.14)
Notes: multipollutant model adds
	CO,NO?, and 03 in addition to the main
July 2009
E-602
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
pollutant of interest, PMio.
Reference: Woodruff et al. (1997,
084271)
Period of Study: 1989-1991
Location: 86 Metropolitan Statistical
Areas in the US (counties with
populations less than 100,000 were
excluded)
Outcome: Postneonatal mortality (death Pollutant: PM
of an infant between 1 month and 1 yr of
1)	all post neonatal deaths
2)	normal birth weight (NBW, a 2500 g)
SIDS deaths
3)	NBW respiratory deaths
4)	low birth weight (LBW) respiratory
death
Respiratory deaths: ICD9 codes 460-519
SIDS: ICD9 code 798.0
Age Groups: infants (1 month-1yr of
age)
Study Design: Cross-sectional
N: 3,788,079 infants
Statistical Analyses: Logistic
regression
Covariates: maternal education,
maternal race, parental marital status,
maternal smoking during pregnancy
avg temperature during the first 2 months
of life
infant's month and year of birth
assessed race as an effect modifier (p-val
for interaction terms > 0.2)
Dose-response Investigated? Yes
Statistical Package: NR
Averaging Time: Mean of 1st 2 months
of life
analyzed as tertiles of exposure and as
continuous exposure
Mean (SD): 31.4 (7.8)
Range (Min, Max):
Overall: 11.9-68.8
Low category: < 28.0
Medium category: 28.1-40.0
High category: >40.0
Monitoring Stations: NR
PM Increment: 10 //gf'nv1 (for continuous
exposure analysis)
Adjusted ORs for cause-specific post
neonatal mortality by pollution
category (tertiles)
All causes
Low: Ref
Medium: 1.05 (1.01,1.09)
High: 1.10(1.04,1.16)
SIDS, NBW:
Low: Ref
Medium: 1.09 (1.01,1.17)
High: 1.26(1.14,1.39)
Respiratory death, NBW:
Low: Ref
Medium: 1.08 (0.87,1.33)
High: 1.40(1.05,1.85)
Respiratory death, LBW:
Low: Ref
Medium: 0.93 (0.73,1.18)
High: 1.18(0.86,1.61)
All other causes:
Low: Ref
Medium: 1.03 (0.97,1.08)
High: 0.97(0.90,1.04)
Adjusted ORs for a continuous
10//g/m3change in exposure
All causes: 1.04 (1.02,1.07)
SIDS, NBW: 1.12 (1.07,1.17)
Respiratory death, NBW: 1.20 (1.06,
1.36)
Respiratory death, LBW: 1.05 (0.91,
1.22)
All other causes: 1.00 (0.99,1.00)
July 2009
E-603
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Woodruff et al (2008,
098386)
Period of Study: 1999-2002
Location: US counties with >250,000
residents (96 counties)
Outcome: Postneonatal deaths
Respiratory mortality (ICD10: J000-99,
plus bronchopulmonary dysplasia [BPD]
P27.1)
SIDS (ICD10: R95)
Ill-defined causes (R99);
All other deaths evaluated as a control
category
Age Groups: Infants aged > 28 days
and <1 yr
Study Design: Cross-sectional
N: 3,583,495 births (6,639 post neonatal
deaths)
Statistical Analyses: Logistic GEE
(exchangeable correlation structure)
Covariates: maternal race/ethnicity,
marital status, age, education,
primiparity, county-level poverty and per
capita income levels, year and month of
birth dummy variables to account for time
trend and seasonal effects, and region of
the country
sensitivity analyses performed among
only those mothers with smoking
information (adjustment for smoking had
no effect on the estimates)
Season: Adjusted for year and month of
birth dummy variables to account for time
trend and seasonal effects
Dose-response Investigated? Evaluated
the appropriateness of a linear form from
analysis based on quartiles of exposure
and concluded that linear form was
appropriate (data not shown)
Statistical Package: SAS
Pollutant: PMio
Averaging Time: Measured continuously
for 24 h once every 6 days
exposure assigned by calculating avg
concentration of pollutant during first 2
months of life
Median and IQR (25th-75th
percentile): Survivors: 28.9 (23.3-34.4)
All causes of death: 29.1 (23.9-34.5)
Respiratory: 29.8 (24.3-36.5)
SIDS: 28.6 (23.5-33.8)
SIDS + ill-defined: 28.8 (23.9-33.9)
Other causes: 29.2 (23.9-34.5)
Percentiles: see above
PM Component: Not assessed, but
controlled for region of the country to
account for PM composition variation
Monitoring Stations: NR
Copollutant (correlation): PMio
PM2.6 (r - 0.34)
CO (r - 0.18)
SO2 (r - 0.00)
O3 (r - 0.20)
Notes: Monthly averages calculated if
there were at least 3 available measures
for PM
Assigned exposures using the avg
concentration of the county of residence
PM Increment: IQR (11 /yglm3)
Effect Estimate [Lower CI, Upper CI]:
Adjusted ORs for single pollutant models
All causes: 1.04 (0.99,1.10)
Respiratory: 1.18 (1.06,1.31)
SIDS: 1.02(0.89,1.16)
Ill-defined + SIDS: 1.06 (0.97,1.16)
Other causes: 1.02 (0.96,1.07)
Adjusted ORs for multipollutant models
(including CO, O3, SO2)
Respiratory: 1.16 (1.04,1.30)
SIDS: 1.02(0.90,1.16)
OR for deaths coded as BPD per increase
in IQR: 1.19 (0.85,1.65)
OR for respiratory post neonatal death
stratified by birth weight
NBW only: 1.19(1.05,1.36)
LBW only: 1.12(0.95,1.31)
OR for respiratory deaths removing region
of US as a confounding variable: 1.30
(1.04,1.61)
OR for respiratory deaths assessing
exposure as quartiles
Highest vs Lowest quartile: 1.31 (1.00,
1.71)
OR for respiratory deaths among only
those deaths that occurred during the
first 90 days (most closely matched
exposure metric of the avg over the first
2 months of life):
1.25(1.06,1.47)
Reference: (Suh et al., 2007,157028)
Period of Study: 2001-2004
Location: Seoul, Korea
Outcome: Birth weight
Age Groups: prenatal follow-up for
newborns
Study Design: based prospective cohort
study
N: 199 pregnant mothers
Statistical Analyses: ANCOVA,
generalized linear models
Covariates: infant's sex, maternal age,
maternal and paternal education, parity,
presence of illness during pregnancy,
delivery month, gestational age (squared)
Dose-response Investigated? Yes
Statistical Package: SAS
Pollutant: PMio
Averaging Time: 24-h
Mean (SD): 1st trimester: 76.41 (28.80)
2nd trimester: 77.84 (31.63)
3rd trimester: 95.61 (26.15)
Percentiles: 1st trimester
25th: 55.28
50th(Median): 71.09
75th: 92.38
2nd trimester
25th: 48.65
50th(Median): 72.36
75th: 108.00
3rd trimester
25th: 77.10
50th(Median): 96.35
75th: 116.68
Range (Min, Max):
1st trimester (21.00,151.65)
PM Increment: Trimester a 90th%ile
compared to < 90th%ile
Least-square (ANCOVA) mean (SE)
All Genotypes
1st trimester
< 90th%ile, N(%): 158 (90.3%): 3253
(37)
> 90th%ile, N(%): 17 (9.7%): 2841 (145)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.009
Adjusted, with CO: 0.041
Adjusted, with NO2: 0.092
Adjusted, with SO2: 0.012
2nd trimester
<90th%ile, N(%): 153 (89.5%): 3253
(39)
>90th%ile, N(%): 18(10.5%): 3026
(157)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
July 2009
E-604
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
2nd trimester (31.45,139.13)
3rd trimester (23.45,172.75)
Monitoring Stations: 27
Copollutant:
CO
SO?
NO?
Adjusted: 0.177
Adjusted, with CO: 0.203
Adjusted, with NO2: 0.151
Adjusted, with SO2: 0.151
3rd trimester
<	90th%ile, N(%): 162 (90.5%): 3226
(38)
>	90th%ile, N(%): 17 (9.5%): 3122 (140)
p-Value for mean birth weight for a
90th%ile PM10 vs. for < 90th%ile PM10
Adjusted: 0.487
Adjusted, with CO: 0.748
Adjusted, with NO2: 0.420
Adjusted, with SO2: 0.466
Genotype Mspl TT
1st trimester
<90th%ile,N(%): 60 (34.3%): 3350
(64)
>	90th%ile, N(%): 5 (2.9%): 3001 (229)
p-Value for mean birth weight for a
90th%ile PM10 vs. for < 90th%ile PM10
Adjusted: 0.147
Adjusted, with CO: 0.186
Adjusted, with NO2: 0.430
Adjusted, with SO2: 0.155
2nd trimester
<90th%ile,N(%): 59 (34.5%): 3335
(66)
>	90th%ile, N(%): 6 (3.5%): 3281 (249)
p-Value for mean birth weight for a
90th%ile PM10 vs. for < 90th%ile PM10
Adjusted: 0.833
Adjusted, with CO: 0.833
Adjusted, with NO2: 0.778
Adjusted, with SO2: 0.806
3rd trimester
<	90th%ile, N(%): 61 (34.1%): 3327
(65)
>	90th%ile, N(%): 6 (3.4%): 3227 (300)
p-Value for mean birth weight for a
90th%ile PM10 vs. for < 90th%ile PM10
Adjusted: 0.749
Adjusted, with CO: 0.980
Adjusted, with NO2: 0.635
Adjusted, with SO2: 0.687
Genotype Mspl TC/CC
1st trimester
<90th%ile,N(%): 98 (56.0%): 3193
(48)
July 2009
E-605
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
>	90th%ile, N(%): 12 (6.9%): 2799 (169)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.033
Adjusted, with CO: 0.073
Adjusted, with NO2: 0.150
Adjusted, with SO2: 0.036
2nd trimester
< 90th%ile, N(%): 94 (55.0%): 3200
(52)
>	90th%ile, N(%): 12 (7.0%): 2933 (176)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.161
Adjusted, with CO: 0.172
Adjusted, with NO2: 0.152
Adjusted, with SO2: 0.158
3rd trimester
<90th%ile, N(%): 101 (56.4%): 3165
(49)
>90th%ile, N(%): 11 (6.2%): 3087 (147)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.626
Adjusted, with CO: 0.978
Adjusted, with NO2: 0.551
Adjusted, with SO2: 0.614
Genotype Ncol llelle
1st trimester
<90th%ile,N(%): 87 (49.7%): 3244
(52)
>	90th%ile, N(%): 7 (4.0%): 2983 (232)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.289
Adjusted, with CO: 0.344
Adjusted, with NO2: 0.641
Adjusted, with SO2: 0.293
2nd trimester
<90th%ile,N(%): 82 (48.0%): 3243
(55)
>	90th%ile, N(%): 11 (6.4%): 3185 (207)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.790
Adjusted, with CO: 0.783
Adjusted, with NO2: 0.707
Adjusted, with SO2: 0.733
3rd trimester
July 2009
E-606
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
<90th%ile, N(%): 90 (50.3%): 3239
(53)
>	90th%ile, N(%): 9 (5.0%): 2944 (198)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.161
Adjusted, with CO: 0.279
Adjusted, with NO2: 0.134
Adjusted, with SO2: 0.150
Genotype Ncol MeVal/ValVal
1st trimester
<	90th%ile, N(%): 71 (40.6%): 3262
(56)
>90th%ile, N(%): 10(5.7%): 2773 (171)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.009
Adjusted, with CO: 0.031
Adjusted, with NO2: 0.058
Adjusted, with SO2: 0.010
2nd trimester
<	90th%ile, N(%): 71 (41.5%): 3264
(61)
>	90th%ile, l\l(%): 7 (4.1%): 2862 (208)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.076
Adjusted, with CO: 0.093
Adjusted, with NO2: 0.063
Adjusted, with SO2: 0.061
3rd trimester
<90th%ile,N(%): 72 (40.2%): 3207
(58)
>	90th%ile, l\l(%): 8 (4.5%): 3262 (180)
p-Value for mean birth weight for a
90th%ile PMio vs. for < 90th%ile PMio
Adjusted: 0.777
Adjusted, with CO: 0.607
Adjusted, with NO2: 0.843
Adjusted, with SO2: 0.791
July 2009
E-607
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Tsai et al. (2006, 098312)
Period of Study: 1994-2000
Location: Kaohsiung, Taiwan
Outcome: post neonatal mortality
Age Groups: infants more than 27 days
and less than 1 year
Study Design: Case-crossover study
N: 207 deaths
Statistical Analyses: Conditional
logistic regression
Covariates: temperature, humidity
Dose-response Investigated? No
Statistical Package: SAS, version 8.2
Pollutant: PMio
Averaging Time: 24 h
Mean (SD): 81.45 //g/m3
Percentiles: 25th: 44.50
50th(Median): 79.20
75th: 111.50
Range (Min, Max): (20.50-232.00)
Monitoring Stations: 6
Copollutant: SO2
NO?
CO
0s
PM Increment: 67.00 //g/m3
Effect Estimate [Lower CI, Upper CI]:
OR - 1.040(0.340-3.177)
Note: Air pollution levels at the dates of
infant death were compared with air
pollution levels 1 week before and 1
week after death
a cumulative lag up to 2 previous days
was used to assign exposure.
Reference: (Wilhelm and Ritz, 2005,
188761)
Period of Study: 1994-2000
Location: Los Angeles County, California,
U.S.
Outcome: Term low birth weight (LBW)
(< 2500 g at a 37 completed weeks
gestation), Vaginal birth <37 completed
weeks gestation
Age Groups: LBW: a 37 com|
weeks
Preterm births: < 37 completed weeks
Study Design: Cross-sectional
N: For LBW: 136,134
For preterm birth:
106,483
Statistical Analyses: Logistic
regression
Covariates: Maternal age, maternal race,
maternal education, parity, interval since
previous live birth, level of prenatal care,
infant sex, previous LBW or preterm
infant, birth season, other pollutants (CO,
NO2, 0 i, PM in), gestational age (in birth
weight analysis)
Dose-response Investigated? Yes
Statistical Package: NR
Pollutant: PM10
Averaging Time:
24 hr (every 6 days)
Entire pregnancy
Trimesters of pregnancy
Months of pregnancy
6 weeks before birth
Mean (SD): First trimester: 42.2
Third trimester: 41.5
6 weeks before birth: 39.1
Range (Min, Max):
First trimester: 26.3, 77.4
Third trimester: 25.7, 74.6
6 weeks before birth: 13.0,103.7
Monitoring Stations:
Zip-code-level analysis: 8
Address-level analysis: 6
Copollutant (correlation):
First trimester: PM10 CO: r - 0.12
PMio-NO?: r - 0.29
PMio-03: r - -0.01
PMio-PM2.b: r - 0.43
Third trimester: PM10 CO: r - 0.32
PMio-NO?: r - 0.45
PMio-03: r - -0.08
PMio-PM2.b: r - 0.52
6 weeks before birth: PMio-CO: r -
0.36
PMio-NO?: r - 0.49
PMio-03: r - -0.16
PMio-PM2.b: r - 0.60
PM Increment: 1) 10 //g/m3
2) 3 levels:
a)	< 25%ile (reference)
b)	25%-75%ile
c)	> 75%ile
Incidence of LBW (third trimester
exposure)
<	32.8//g/m3: 2.0(1.8, 2.2)
32.8	to < 43.4//g/m3: 2.0 (1.9, 2.1)
a 43.4//g/m3: 2.2(2.0, 2.4)
Incidence of preterm birth (first
trimester exposure)
<	32.9//g/m3: 8.7 (8.3, 9.2)
32.9	to < 43.9//g/m3: 8.8 (8.5, 9.1)
>43.9//g/m3: 8.6(8.1,9.0)
Incidence of preterm birth (6 weeks
before birth exposure)
<	31.8//g/m3: 8.8(8.4,9.3)
31.8 to <44.1 //g/m3: 8.6 (8.3, 8.9)
> 44.1 //g/m3: 8.8 (8.4, 9.2)
Outcome: LBW
Exposure Period: Third trimester
Address-level analysis:
Single-pollutant model:
Distance ~ 1 mile
Per 10 //g/m3:1.22 (1.05,1.41)
33.4 to < 44.7 //g/m3:1.08 (0.76, 1.52)
>44.7//g/m3: 1.48 (1.00,2.19)
Multipollutant model:
Distance ~ 1 mile
Per 10 //g/m3:1.36 (1.12, 1.65)
33.4 to <44.7 //g/m3:1.16 (0.77, 1.74)
>44.7//g/m3: 1.58 (0.95, 2.62)
Single-pollutant model:
July 2009
E-608
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1 < distance ~ 2 mile
Per 10 //g/m3: 0.98 (0.90, 1.06)
33.4 to < 44.7 //g/m3: 0.95 (0.80, 1.13)
>44.7//g/m3: 0.96 (0.78, 1.18)
Multipollutant model:
1	< distance ~ 2 mile
Per 10 //g/m3:1.02 (0.92,1.14)
33.4 to < 44.7 //g/m3: 0.93 (0.77, 1.12)
>44.7//g/m3: 1.02 (0.79, 1.32)
Single-pollutant model:
2	< distance ~ 4 mile
Per 10 //g/m3:1.03 (0.99, 1.08)
33.9 to < 45.0//g/m3:1.04 (0.96,1.14)
>45.0 //g/m3: 1.08 (0.97, 1.20)
Multipollutant model:
2 < distance ~ 4 mile
Per 10 //g/m3:1.04 (0.98,1.09)
33.9 to <45.0//g/m3:1.02 (0.92, 1.12)
>45.0 //g/m3: 1.06 (0.93, 1.21)
Zip-code-level analysis
Single-pollutant model:
Per 10//g/m3:1.03 (0.97, 1.09)
33.2 to <43.6//g/m3: 0.98 (0.86, 1.1 1)
>43.6//g/m3: 1.03 (0.88, 1.21)
Multipollutant model:
Per 10 //g/m3:1.07 (0.99, 1.15)
33.2	to < 43.6//g/m3: 0.97 (0.85, 1.12)
>43.6//g/m3: 1.09 (0.90, 1.31)
Outcome: LBW
Exposure Period: Entire pregnancy
period
Address-level analysis:
Multipollutant model:
Per 10 //g/m3:1.24 (0.91, 1.70)
Outcome: Preterm Birth
Exposure Period: First trimester of
pregnancy
Address-level analysis:
Single-pollutant model:
Distance ~ 1 mile
Per 10//g/m3:1.00 (0.93, 1.09)
33.3	to <45.1 //g/m3:1.07 (0.90, 1.26)
>45.1 //g/m3: 1.12(0.91,1.38)
Multipollutant model:
Distance ~ 1 mile
Per 10//g/m3:1.00 (0.90, 1.12)
July 2009
E-609
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
33.3 to <45.1 //g/m3:1.12 (0.92, 1.36)
>	45.1 //g/m3: 1.17 (0.90, 1.50)
Single-pollutant model:
1 < distance ~ 2 mile
Per 10 //g/m3:1.01 (0.97, 1.05)
33.7 to < 45.3//g/m3:1.03 (0.95, 1.12)
>	45.3 //g/m3: 1.07 (0.97, 1.19)
Multipollutant model:
1	< distance ~ 2 mile
Per 10 //g/m3:1.04 (0.99,1.10)
33.7 to <45.3//g/m3:1.07 (0.98, 1.17)
>45.3//g/m3: 1.13 (1.00, 1.27)
Single-pollutant model:
2	< distance ~ 4 mile
Per 10 //g/m3:1.01 (0.99, 1.03)
34.1 to <45.5//g/m3:1.03 (0.99, 1.08)
a 45.5 //g/m3: 1.02 (0.96, 1.07)
Multipollutant model:
2 < distance ~ 4 mile
Per 10 //g/m3: 0.99 (0.97, 1.02)
34.1 to <45.5//g/m3: 0.99 (0.95,1.04)
>45.5//g/m3: 0.94 (0.89,1.01)
Zip-code-level analysis
Single-pollutant model:
Per 10 //g/m3: 0.99 (0.96, 1.01)
33.3 to <44.2//g/m3:1.01 (0.95, 1.08)
>44.2//g/m3: 0.98 (0.90, 1.05)
Multipollutant model:
Per 10 //g/m3: 0.99 (0.96, 1.03)
33.3 to <44.2//g/m3:1.03 (0.97, 1.1 1)
>44.2//g/m3: 1.01 (0.92,1.11)
Outcome: Preterm birth
Exposure Period: 6 weeks before birth
Address-level analysis:
Single-pollutant model:
Distance ~ 1 mile
Per 10 //g/m3:1.02 (0.95, 1.10)
32.5 to < 44.8//g/m3:1.09 (0.92, 1.29)
a 44.8 //g/m3: 1.12 (0.92, 1.37)
Multipollutant model:
Distance ~ 1 mile
Per 10 //g/m3:1.06 (0.97, 1.16)
32.5 to <44.8//g/m3:1.09 (0.90, 1.31)
>44.8//g/m3: 1.17 (0.91,1.49)
Single-pollutant model:
July 2009
E-610
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1 < distance ~ 2 mile
Per 10 //g/m3:1.00 (0.96, 1.03)
32.3 to < 45.3//g/m3: 0.99 (0.91, 1.07)
>45.3//g/m3: 0.99 (0.89, 1.10)
Multipollutant model:
1	< distance ~ 2 mile
Per 10 //g/m3:1.01 (0.97, 1.06)
32.3 to < 45.3//g/m3:1.00 (0.92, 1.10)
>45.3 //g/m3: 1.02 (0.91,1.16)
Single-pollutant model:
2	< distance ~ 4 mile
Per 10//g/m3: 0.99 (0.98, 1.01)
33.1 to <45.3//g/m3:1.00 (0.96, 1.05)
>45.3//g/m3: 0.98 (0.93, 1.03)
Multipollutant model:
2 < distance ~ 4 mile
Per 10//g/m3:1.00 (0.98, 1.02)
33.1 to <45.3//g/m3:1.01 (0.96, 1.05)
>45.3//g/m3: 0.98 (0.92,1.04)
Zip-code-level analysis
Single-pollutant model:
Per 10//g/m3:1.02(0.99,1.04)
32.1 to <44.3//g/m3:1.01 (0.95, 1.07)
>44.3//g/m3: 1.04 (0.96,1.12)
Multipollutant model:
Per 10 //g/m3:1.02 (0.99, 1.06)
32.1 to <44.3//g/m3:1.02 (0.95, 1.09)
a 44.3 //g/m3: 1.04 (0.95,1.14)
Notes: multipollutant model adds
C0,N02, and 03 in addition to the main
pollutant of interest, PM10.
July 2009
E-611
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Woodruff et al. (1997,
084271)
Period of Study: 1989-1991
Location: 86 Metropolitan Statistical
Areas in the US (counties with
populations less than 100,000 were
excluded)
Outcome: Postneonatal mortality (death
of an infant between 1 month and 1 yr of
age)
1)	all post neonatal deaths
2)	normal birth weight (NBW, a 2500 g)
SIDS deaths
3)	NBW respiratory deaths
4)	low birth weight (LBW) respiratory
death
Respiratory deaths: ICD9 codes 460-519
SIDS: ICD9 code 798.0
Age Groups: infants (1 month-1yr of
age)
Study Design: Cross-sectional
N: 3,788,079 infants
Statistical Analyses: Logistic
regression
Covariates: maternal education,
maternal race, parental marital status,
maternal smoking during pregnancy
avg temperature during the first 2 months
of life
infant's month and year of birth
assessed race as an effect modifier (p-val
for interaction terms > 0.2)
Dose-response Investigated? Yes
Statistical Package: NR
Pollutant: PMio
Averaging Time: Mean of 1st 2 months
of life
analyzed as tertiles of exposure and as
continuous exposure
Mean (SD): 31.4 (7.8)
Range (Min, Max):
Overall: 11.9-68.8
Low category: < 28.0
Medium category: 28.1-40.0
High category: >40.0
Monitoring Stations: NR
PM Increment: 10 //gf'nv1 (for continuous
exposure analysis)
Adjusted ORs for cause-specific post
neonatal mortality by pollution
category (tertiles)
All causes
Low: Ref
Medium: 1.05 (1.01,1.09)
High: 1.10(1.04,1.16)
SIDS, NBW:
Low: Ref
Medium: 1.09 (1.01,1.17)
High: 1.26(1.14,1.39)
Respiratory death, NBW:
Low: Ref
Medium: 1.08 (0.87,1.33)
High: 1.40(1.05,1.85)
Respiratory death, LBW:
Low: Ref
Medium: 0.93 (0.73,1.18)
High: 1.18(0.86,1.61)
All other causes:
Low: Ref
Medium: 1.03 (0.97,1.08)
High: 0.97(0.90,1.04)
Adjusted ORs for a continuous
10//g/m3change in exposure
All causes: 1.04 (1.02,1.07)
SIDS, NBW: 1.12 (1.07,1.17)
Respiratory death, NBW: 1.20 (1.06,
1.36)
Respiratory death, LBW: 1.05 (0.91,
1.22)
All other causes: 1.00 (0.99,1.00)
July 2009
E-612
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Woodruff et al (2008,
098386)
Period of Study: 1999-2002
Location: US counties with >250,000
residents (96 counties)
Outcome: Postneonatal deaths
Respiratory mortality (ICD10: J000-99,
plus bronchopulmonary dysplasia [BPD]
P27.1)
SIDS (ICD10: R95)
Ill-defined causes (R99);
All other deaths evaluated as a control
category
Age Groups: Infants aged > 28 days
and <1 yr
Study Design: Cross-sectional
N: 3,583,495 births (6,639 post neonatal
deaths)
Statistical Analyses: Logistic GEE
(exchangeable correlation structure)
Covariates: maternal race/ethnicity,
marital status, age, education,
primiparity, county-level poverty and per
capita income levels, year and month of
birth dummy variables to account for time
trend and seasonal effects, and region of
the country
sensitivity analyses performed among
only those mothers with smoking
information (adjustment for smoking had
no effect on the estimates)
Season: Adjusted for year and month of
birth dummy variables to account for time
trend and seasonal effects
Dose-response Investigated? Evaluated
the appropriateness of a linear form from
analysis based on quartiles of exposure
and concluded that linear form was
appropriate (data not shown)
Statistical Package: SAS
Pollutant: PMio
Averaging Time: Measured continuously
for 24 h once every 6 days
exposure assigned by calculating avg
concentration of pollutant during first 2
months of life
Median and IQR (25th-75th
percentile): Survivors: 28.9 (23.3-34.4)
All causes of death: 29.1 (23.9-34.5)
Respiratory: 29.8 (24.3-36.5)
SIDS: 28.6 (23.5-33.8)
SIDS + ill-defined: 28.8 (23.9-33.9)
Other causes: 29.2 (23.9-34.5)
Percentiles: see above
PM Component: Not assessed, but
controlled for region of the country to
account for PM composition variation
Monitoring Stations: NR
Copollutant (correlation): PMio
PM2.6 (r - 0.34)
CO (r - 0.18)
SO2 (r - 0.00)
O3 (r - 0.20)
Notes: Monthly averages calculated if
there were at least 3 available measures
for PM
Assigned exposures using the avg
concentration of the county of residence
PM Increment: IQR (11 /yglm3)
Effect Estimate [Lower CI, Upper CI]:
Adjusted ORs for single pollutant models
All causes: 1.04 (0.99,1.10)
Respiratory: 1.18 (1.06,1.31)
SIDS: 1.02(0.89,1.16)
Ill-defined + SIDS: 1.06 (0.97,1.16)
Other causes: 1.02 (0.96,1.07)
Adjusted ORs for multipollutant models
(including CO, O3, SO2)
Respiratory: 1.16 (1.04,1.30)
SIDS: 1.02(0.90,1.16)
OR for deaths coded as BPD per increase
in IQR: 1.19 (0.85,1.65)
OR for respiratory post neonatal death
stratified by birth weight
NBW only: 1.19(1.05,1.36)
LBW only: 1.12(0.95,1.31)
OR for respiratory deaths removing region
of US as a confounding variable: 1.30
(1.04,1.61)
OR for respiratory deaths assessing
exposure as quartiles
Highest vs Lowest quartile: 1.31 (1.00,
1.71)
OR for respiratory deaths among only
those deaths that occurred during the
first 90 days (most closely matched
exposure metric of the avg over the first
2 months of life):
1.25(1.06,1.47)
July 2009
E-613
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Jedrychowski, et al„ (2007,
156607)
Period of Study: Jan 2001-Feb 2004
Location: Krakow, Poland
Outcome: Birth weight (grams), birth
length (cm)
Age Groups: pregnant women 18-35
years
Study Design: Prospective cohort
N:493 women
Statistical Analyses: Linear regression
Covariates: Environmental tobacco
smoke (# cigarettes smoked daily in
presence of pregnant woman), season of
birth, size of mother, parity, gestational
age, gender of child, vitamin A intake
Season: All
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: Two consecutive days
in the second trimester
Pollutant: PM2.6
Averaging Time: 48 h period
Percentiles: 50th(Median): 35.3
Range (Min, Max): 10.3, 294.9
Monitoring Stations: No stations,
personal monitoring
Notes: PM measured during a two day
period in the second trimester by Personal
Environmental Monitoring Sampler
(PEMS)
PM Increment: in 1 //gf'nr1 and tertiles
T1: < 27.0 /yglm3
T2: 27.0-46.2 //g/m3
T3: > 46.2/ygfm3
Mean [Lower CI, Upper CI]:
Birth weight (g)
For In unit PM: p - -172.39 (p - 0.02)
Tertiles:
T1: ref
T2: p - -16.510 [-94.630, 61.610]
T3: p - -109.956 [-196.649 to-23.263]
In low Vitamin A group (< 1,378/yg)
T1: ref
T2: p - -68.354 [-165.643, 28.935]
T3: p - -185.070 [-293.393 to-76.747]
In high Vitamin A group (> 1,378/yg)
T1: ref
T2: p - 64.262 [-70.464,198.988]
T3: p - 38.593[-109.853,187.039]
Birth length (cm)
For In unit PM: p — -1.39 (p — 0.00)
Tertiles:
T1: ref
T2: p - -0.288 [-0.790, 0.214]
T3: p - -0.810 [-1.367 to -0.253]
In low Vitamin A group (< 1,378/yg)
T1: ref
T2: p - -0.514 [-1.114, 0.086]
T3: p - -1.100 [-1.768 to-0.432]
In high Vitamin A group (> 1,378/yg)
T1: ref
T2: p - 0.039 [-0.896, 0.974]
T3: p - -0.301 [-1.326, 0.724]
July 2009
E-614
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Lipfert et al„ 2000,
004103)
Period of Study: 1990
Location: U.S.
Outcome (ICD9 and ICD10): Infant
mortality
including respiratory mortality (traditional
definition, ICD9 460-519), expanded
definition (adds ICD9 769 and 770)
Age Groups: Infants
Study Design: Cross-sectional
N: 2,413,762 infants in 180 counties (Ns
differ for various models)
Statistical Analyses: Logistic
regression
Covariates: mother's smoking,
education, marital status, and race
month of birth
and county avg heating degree days
Dose-response Investigated? NR
Statistical Package: NR
Pollutant: SO421 NSPM10 (regressed
jointly)
Averaging Time: Yearly avg used
Mean (SD): 33.1 (9.17) (based on 180
counties)
Range (Min, Max): (16.9, 59)
Monitoring Stations: NR
Copollutant:
PM10
NSPM10
CO
SO?
Notes: TSP-based sulfate was adjusted
for compatibility with the PMio-based
data
July 2009
E-615
PM Increment: NR (present regression
coefficients)
Effect Estimate [Lower CI, Upper CI]:
Presented regression coefficients
(standard errors)
(3 PM exposures regressed jointly)
bold - p < 0.05
Cause of death: All
Birth weight: All
SO42 :-0.0002 (0.0061)
NSPM10: 0.0115(0.0014)
Cause of death: All
Birth weight: LBW
SO42-: 0.0265 (0.0080)
NSPM10: 0.0086 (0.0020)
Cause of death: All
Birth weight: normal
SO42-: -0.0488 (0.0098)
NSPM10: 0.0096 (0.0024)
Cause of death: All neonatal
Birth weight: All
SO42-: 0.0267 (0.0076)
NSPM10: 0.0126 (0.0018)
Cause of death: All neonatal
Birth weight: LBW
SO42-: 0.0388 (0.0088)
NSPM10: 0.0093 (0.0022)
Cause of death: All neonatal
Birth wt: normal
SO42 :-0.0334 (0.0169)
NSPM10: 0.0125 (0.0040)
Cause of death: All post neonatal
Birth wt: All
PM10: 0.0091 (0.0024)
SO42 :-0.0474 (0.0100)
NSPM10: 0.0096 (0.0024)
Cause of death: All post neonatal
Birth wt: LBW
SO42 :-0.0247 (0.0173)
NSPM10: 0.0101 (0.0042)
Cause of death: All post neonatal
Birth wt: normal
SO42 :-0.0569 (0.0121)
NSPM10: 0.0080 (0.0029)
Cause of death: SIDS
Birth weight: All
SO42 :-0.1078 (0.0151)
NSPM10: 0.0149 (0.0037)
CapiAFTleatBtfNpgT CITE OR QUOTE
Birth weight: LBW
SO42 :-0.1378 (0.0337)

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Liu et al„ 2007, 090429)
Period of Study: 1985-2000
Location: 3 Canadian cities: Calgary,
Edmonton, and Montreal
Outcome: intrauterine growth restriction
(IUGR)
Age Groups: singleton term live births
(37-42 wks gestation)
Study Design: retrospective cohort
N: 386,202 singleton live births
Statistical Analyses: Multiple logistic
regression
Covariates: maternal age, parity, infant
gender, season, and city of residence at
time period of birth
Season: All seasons
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM2.6
Averaging Time: 24 h (6-day schedule)
Mean (SD): 12.2
Percentiles: 25th: 6.3
50th(Median): 9.7
75th: 15
PM Component: metals and organic
matter such as polycyclic aromatic
hydrocarbons
Monitoring Stations: Calgary (4),
Edmonton (2), and Montreal (8)
Copollutant (correlation): SO2:
r- 0.44, p< 0.0001
NO?: r - 0.41, p< 0.0001
CO: r - 0.31, p< 0.0001
0s: r - -0.14, p< 0.0001
PM Increment: 10 /yglm3
Effect Estimate
single-pollutant model [Lower CI,
Upper CI]:
1st trimester
OR - 1.07(1.03-1.10)
2nd trimester
OR - 1.06 (1.03-1.10)
3,d trimester
OR - 1.06(1.03-1.10)
Effect Estimate
multi-pollutant model [Lower CI,
Upper CI]:
1st trimester
OR- 1.03 (0.99-1.06)
2nd trimester
OR- 1.01 (0.98-1.05)
3,d trimester
OR- 1.03 (0.99-1.06)
Note: ORs and CIs estimated from Figs. 6
and 7
July 2009
E-616
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Loomis et al. (1999, 087288)
Period of Study: Jan 1,1993-Jul
31,1995
Location: Mexico City (southwestern
section)
Study Design: Time-series
N: 942 deaths (days were the unit of
observation)
Statistical Analyses: Poisson regression
(generalized additive model)
Covariates: Final models controlled for
mean temp of 3 days before death and
nonparametrically smoothed periodic
cycles
Season: Yes (considered)
Dose-response Investigated? Loess
smoother
Statistical Package: NR
Lags Considered: 0-5 (also considered
lags with avg exposure levels during
"windows" of 2 to 4 days)
Outcome (ICD9 and ICD10): Infant
mortality (daily counts of deaths)
All ICD9 codes, excluding accidents,
poisoning, and violence (ICD9 £800)
Age Groups: Children < 1 yr of age
Pollutant: PM2.6
Averaging Time: 24-h
Mean (SD): 27.4 (10.5)
Percentiles: Lower quartile: 20
Median: 26
Upper quartile: 34
Range (Min, Max): 4, 85
Monitoring Stations: one
Copollutant: 0:i
NO?
NO
NOx
SO?
Notes: Pearson correlation coefficients
ranging from 0.52 to 0.71
PM Increment: 10/yg/m3
Effect Estimate [Lower CI, Upper CI]:
%Change in infant mortality
Lags 0-5 (single day) presented in Figure
1:
Lag0,1,2: No association (results not
presented)
Lag3: 4.8 (0.97, 8.61)
Lag4: 4.2 (0.37, 7.93)
%Change in mortality when avg exposure
levels during "windows" of 2 to 4 days
were considered
Two Days:
No lag: -1.36 (-5.51, 2.8)
Lag1: -0.95 (-5.10, 3.20)
Lag2: 2.78 (-1.33, 6.89)
Lag3: 4.93 (0.86, 9.01)
Three Days:
No lag: -0.81 (-5.29, 3.67)
Lag1: 1.99 (-2.46, 6.45)
Lag2: 4.54 (0.12, 8.96)
Lag3: 6.87(2.48,11.26)
Four Days:
No lag: 1.95 (-2.76, 6.66)
Lag1: 3.74 (-0.95,8.42)
Lag2: 5.87(1.21,10.53)
Multipollutant models (3-day mean wf 3-
day lag)
1	pollutant model:
6.87(2.48,11.26)
2	pollutant models:
wI 03:6.24(1.35,11.14)
wI NO?: 5.91 (-0.76,12.59)
3	pollutant model (wf O3 and NO2): 6.30
(-0.54,13.15)
July 2009
E-617
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Mannes et al. (2005,
087895)
Period of Study: January 1,1998-
December 31, 2000
Location: metropolitan Sydney, Australia
Outcome: risk of small for gestational
age (SGA) and birth weight
Age Groups: all singleton births > 20
weeks and a 400 grams birth weight and
maternal all ages
Study Design: cross-sectional
N: 138,056 singleton births
Statistical Analyses: Logistic and linear
regression models
Covariates: sex of child, maternal age,
gestational age, maternal smoking,
gestational age at first antenatal visit,
maternal indigenous status, whether first
pregnancy, season of birth, and
socioeconomic status (SES)
Season: All seasons
included as covariate.
Dose-response Investigated? No
Statistical Package: SAS System for
Windows v8.02
Pollutant: PM2.6
Averaging Time: 24 h
Mean (SD): 9.4 (5.1)
Percentiles: 25th: 6.5
50th(Median): 8.4
75th: 11.2
Range (Min, Max): (2.4- 82.1)
Monitoring Stations: up to 14
Copollutant (correlation):
CO: r - 0.53
NO?: r - 0.66
O3: r = 0.36
PM10: r - 0.81
PM Increment: 1 /yglm3
Risk of SGA
All births
One month before birth: OR - 1.01
(0.99-1.03)
Third trimester: OR - 0.99(0.97-1.02)
Second trimester: OR - 1.03 (1.01-1.05)
First trimester: OR - 0.99 (0.97-1.01)
5 km births
One month before birth: OR - 1.01
(0.97-1.04)
Third trimester: OR - 1.00(0.95-1.05)
Second trimester: OR - 1.00 (0.96-1.05)
First trimester: OR - 0.99 (0.94-1.04)
Change in birth weight
All births
One month before birth: B - -2.48 (-4.58-
¦0.38)
Third trimester: B - -0.98 (-3.74—1.78)
Second trimester: B - -4.10 (-6.79- ¦
1.41)
First trimester: B - 0.36 (-2.29- 3.01)
5 km births
One month before birth: B - -2.70 (-6.80-
1.40)
Third trimester: B - -2.83 (-9.00-3.34)
Second trimester: B - 1.54 (-4.59-7.67)
First trimester: B - 1.89 (-1.99-5.77)
July 2009
E-618
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Parker et al. (2005, 087462) Outcome: small for gestational age (SGA) Pollutant: PM2
and birth weight
Period of Study: 1999-2000
Location: California
Age Groups: infants delivered at 40
weeks gestation
maternal all ages
Study Design: cross-sectional
N: 18,247 singleton births
Statistical Analyses: Linear and logistic
regression models
Covariates: maternal race, maternal
Hispanic origin, marital status, parity,
maternal education, and maternal age
Season: season of delivery (covariate)
Dose-response Investigated? Yes
Statistical Package: STATA
Averaging Time: NR (measurement
taken every 6 days)
Mean (SD): 15.42 (5.08)
PM Component: metals, polycyclic
aromatic hydrocarbons
Monitoring Stations: 40
Copollutant (correlation):
PM2.B-CO: r - 0.6
Notes: Mean calculated for 9-month
exposure. The following means (SDs) are
calculated for trimester:
First: 15.70(6.26)
Second: 15.40 (6.53)
Third: 14.29 (6.35)
PM categorized into quartiles:
Q1: <11.9
Q2: 11.9-13.9
Q3: 13.9-18.4
Q4: >18.4
PM Increment: < 11.9 /yg/m3
Referent PM Increment: 11.9-
13.9/yg/m3
Effect Estimate [Lower CI, Upper CI]:
First Trimester
Birth weight: B - -5.7 (-27.9-16.5)
SGA: OR - 1.02 (0.84-1.23)
Second Trimester
Birth weight: B - 11.3 (-12.2-34.9)
SGA: OR - 0.89 (0.73-1.09)
Third Trimester
Birth weight: B - 8.3 (-13.1-29.8)
SGA: OR - 1.00 (0.83-1.19)
PM Increment: 13.9-18.4 //g|m3
Effect Estimate [Lower CI, Upper CI]:
First Trimester
Birth weight: B - -2.5 (-24.5-19.5)
SGA: OR - 1.12(0.93-1.34)
Second Trimester
Birth weight: B - -17.2 (-39.4-4.9)
SGA: OR - 1.05 (0.88-1.26)
Third Trimester
Birth weight: B - -8.1 (-30.2-13.9)
SGA: OR - 0.98 (0.82-1.18)
PM Increment: > 18.4//g/m3
Effect Estimate [Lower CI, Upper CI]:
First Trimester
Birth weight: B - -35.8 (-58.4-13.3)
SGA: OR - 1.26 (1.04-1.51)
Second Trimester
Birth weight: B - -46.6 (-68.6- -24.6)
SGA: OR - 1.24(1.04-1.49)
Third Trimester
Birth weight: B - -31.6 (-52.0- -11.1)
SGA: OR - 1.21 (1.02-1.43)
July 2009
E-619
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Parker and Woodruff, 2008, Outcome: Low birth weight
189095)
Period of Study: 2001-2003
Location: US
Study Design: cohort
N: 785,965 Singleton births delivered at
40 weeks gestation
Statistical Analyses: GEE regression
models
linear and logistic regression
Covariates: race/ethnicity, parity,
maternal age
Season: season of delivery
Statistical Package: SUDAAN
Pollutant: PM2.6
Averaging Time: 9-months
Mean (SD): 14.5
25th: 12.1
75th: 17.6
Copollutant (correlation): SO2, NO2 CO
0s
PM Increment: 10 /yglm3
Change in Birth weight (9 month
exposure): Unadjusted: 19.4 (9.8, 29.0)
Adjusted for maternal factors: 18.4 (9.2,
27.7)
Stratified by region: Industrial
Midwest:-15.3 (-43.4,12.9)
Northeast: -9.8 (-11.9, 26.6)
Northwest: 27.5 (5.5, 49.4)
Southern CA: 5.5 (-9.6, 20.5)
Southeast: 7.3 (-11.9, 26.6)
Southwest: 72.3 (34.0,110.5)
Upper Midwest: -0.7 (-62.0, 60.6)
Multipollutant models: PM2.6 +PM10
2.5:14.2(4.3, 24.1)
PM2.6 +PMio-2.5+S02+CO+N02+03:
28.6(14.2, 43.0)
Reference: Rich et al. (2009,180122)
Period of Study: 1999-2003
Location: New Jersey, United States
Outcome: Small for gestational age
Study Design: Retrospective Cohort
Covariates: month and calendar year of
birth, apparent temperature, pregnancy
complications
Statistical Analysis: Polytomous
logistic regression
Pollutant: PM2.6
Averaging Time: 24h
Mean (SD) Unit:
"All values are for first trimester, other
trimesters are available in paper
Reference Births: 13.8 (2.5)
SGA Births: 13.9(2.5)
Statistical Package: SAS
Age Groups: Gestational age 37-42 wks VSGA Births: 13.9 (2.4)
Range (Min, Max): 2.0, 29.0
Copollutant (correlation):
"All values are for first trimester, other
trimesters are available in paper
NO2: 0.01
SO2: 0.17
CO: 0.25
"All values are for first trimester, other
trimesters are available in paper
Increment: 4/yg/m3
Percent Change in Risk (95% CI)
SGA: 4.5 (0.5-8.7)
VSGA: 2.6 (-4.4-10.0)
Percent Change in Risk (95% CI) for
single and two pollutant models
Single, SGA: 4.6 (-0.3-9.8)
Single, VSGA: 4.5 (-4.0-13.4)
Two (PM2.6 & NO2), SGA: 4.5 (-0.4-9.7)
Two (PM2.6 & NO2), VSGA: 3.2 (-5.2-12.4)
Percent Change in Risk (95% CI) by
pregnancy complication in third
trimester
SGA
Any Complication
No: 4.7 (0.6-9.0)
Yes: 2.2 (-6.1-11.3)
Placental Abruption
No: 4.0 (0.3-7.9)
Yes: 11.7 (-21.7-59.5)
Placental Praevia
No: 3.9 (0.2-7.8)
Yes: 23.2 (-20.9-91.9)
Preeclampsia
No: 4.2 (0.4-8.2)
Yes: 2.7 (-13.8-22.3)
Gestational Hypertension
No: 4.3 (0.4-8.4)
Yes: 3.9 (-7.8-17.1)
Premature Rupture of the Membrane
July 2009
E-620
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
No: 3.7 (-0.1-7.7)
Yes: 14.6 (-3.3-35.9)
Gestational Diabetes
No: 4.6 (0.8-8.6)
Yes: -9.3 (-24.7-9.3)
VSGA
Any Complication
No: 1.5 (-6.1-9.7)
Yes: 12.6 (0.1-26.7)
Placental Abruption
No: 4.1 (-2.6-11.2)
Yes: 7.6 (-29.8-64.9)
Placental Praevia
No: 4.1 (-2.5-11.2)
Yes: 3.2 (-43.0-86.9)
Preeclampsia
No: 4.4 (-2.6-11.9)
Yes: 3.9 (-15.7-28.1)
Gestational Hypertension
No: 3.2 (-4.0-10.9)
Yes: 12.9 (-3.3-31.9)
Premature Rupture of the Membrane
No: 3.3 (-3.5-10.5)
Yes: 21.9 (-3.6-54.2)
Gestational Diabetes
No: 4.3 (-2.5-11.5)
Yes: 1.4 (-27.0-40.9)
July 2009
E-621
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Ritz et al. (2007, 096146)
Period of Study: Jan 1, 2003-Dec 31,
2003
Location: Los Angeles, California
Outcome: Preterm births (infants
delivered before 37 weeks)
Age Groups: Births
Study Design: Case-control nested
within a birth cohort (cases and controls
matched on zip code and birth month)
Phase 1: cross-sectional including all birth
cohort
Phase 2: nested case-control of survey
respondents
N: Phase 1: Birth cohort consisted of
58,316 eligible births. Phase II: 2,543
Statistical Analyses: Logistic
regression
Covariates: Birth certificate information:
maternal age, race/ethnicity, parity,
education, season of birth
survey information: maternal smoking,
alcohol consumption, living with a
smoker, and marital status during
pregnancy
income (imputed)
occupation and pregnancy weight gain
considered but not included in final
models
Season:Yes
Dose-response Investigated? Yes,
examined categories of exposure
Statistical Package: NR
Pollutant: PM2.6
Averaging Time: daily or every 3>d day
used to calculate the entire pregnancy,
the first trimester, and the last 6 weeks
before delivery
only reported first trimester exposures for
PM
Range (Min, Max): NR
Ranges for 3 categories reported:
Low (ref): ~ 18.63
Mid: 18.64-21.36
High: >21.36
Monitoring Stations: Each zip code was
linked to the nearest monitoring station
(number not reported)
Copollutant (correlation): CO
NO?
Os
Notes: Daily or every 3,d day
measurements used for mean calculations
PM Increment: Reported analyses using
exposure categories
Effect Estimate [Lower CI, Upper CI]:
Birth cohort (phase I)
Crude: Low: 1.0
Mid: 0.96 (0.90,1.03)
High: 1.05(0.99,1.12)
Adj for birth cert Covariates: Low: 1.0
Mid: 1.01 (0.93,1.09)
High: 1.10(1.01,1.20)
Survey respondents (phase II)
Crude: Low: 1.0' Mid: 1.11 (0.90,1.36)
High: 1.27 (1.06,1.53)
Adj for birth cert Covariates: Low: 1.0
Mid: 1.14(0.90,1.46)
High: 1.27 (0.99,1.64)
Adj for all Covariates: Low: 1.0
Mid: 1.15(0.90,1.47)
High: 1.29(1.00,1.67)
Two-phase model: * Low: 1.0
Mid: 0.98 (0.84,1.15)
High: 1.07 (0.85,1.35)
"method to reduce potential selection
bias and increase statistical efficiency
Reference: (Slama et al., 2007, 093216) Outcome: Birth weight offspring at term
Period of Study: 111998-111999
Location: Munich, Germany
Study Design: Cohort study
N: 1016 births
Statistical Analyses: Poisson model
Covariates: Maternal passive smoking,
maternal age, gestational duration, sex of
child, parity, maternal education, maternal
size, prepregnancy weight, other
pollutants (PM2.6, PM2.6 absorbance, NO2),
season of conception
Dose-response Investigated? Yes
Statistical Package: STATA
Pollutant: PM2.6 (estimated based on
larger PM size fractions)
Averaging Time: Entire pregnancy period
and trimesters
Mean (SD): 14.4
Percentiles: 25th: 13.5
50th (Median): 14.4
75th: 15.4
Monitoring Stations: Spatial component:
40
Temporal component: 1
Copollutant (correlation):
p.a. - pregnancy avg
trim. - trimester
PM2.6 (p.a.)—PM2.E (1st trim.): 0.85
PM2.6 (p.a.)—PM2.B (2nd trim.): 0.77
PM2.6 (p.a.)—PM2.B (3,d trim.): 0.87
PM2.6 (p.a.)—NO2 (p.a.): 0.45
PM2.6 (p.a.)—NO2 (1st trim.): 0.18
PM2.6 (p.a.)—NO2 (2nd trim.): 0.32
PM2.6 (p.a.)—NO2 (3rd trim.): 0.37
PM2.6 (1s,trim.)-PM2.B (2nd trim.): 0.40
PM2.6 (1st trim.)—PM2.E (3,d trim.): 0.68
PM Increment: 1) 1 //g|m3
2) Quartiles: a) 1st (reference) (7.2-
13.5//g/m3)
b)	2nd (13.5-14.4 //gfm3)
c)	3rd (14.4-15.4 //g/m3)
d)	4th (15.41-17.5//g/m3)
Prevalence ratios (PRs) of birth weight
< 3000 g during exposure over the
whole pregnancy
Single-pollutant models
Unadjusted models
2nd quartile: 1.07 (0.65,1.73); 3rd
quartile: 1.38 (0.91, 2.09)
4th quartile: 1.45 (0.92, 2.25)
Per 1 //g/m3: 1.06 (0.95, 1.19)
Adjusted models
2nd quartile: 1.08 (0.63,1.82); 3rd
quartile: 1.34 (0.86, 2.13)
4th quartile: 1.73 (1.15, 2.69); Per
1 //g/m3: 1.13 (1.00, 1.29)
Multipollutant models
Adjusted models
2nd quartile: 1.01 (0.57,1.85)
3rd quartile: 1.12(0.64,1.87)
_^th_^uartileM_;36_(0i^2;_2;45h_Pei^^^^^_
July 2009
E-622
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1 //g|mT 1.07 (0.91,1.26)
PM2.6
PM2.6
PM2.B
PM2.B
PM2.B
PM2.B
PM2.6
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.6
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
0.29
PM2.B
PM2.6
PM2.B
PM2.B
PM2.B
1st trim.)-l\l02 (p a.): 0.48
1st trim.)—IMO2 (1st trim.): 0.15
1st trim.)-l\l02 (2nd trim.): 0.41
1st trim.)-PJ02 (3rd trim.): 0.39
2nd trim.)—PM2.B (3,d trim.): 0.51
2nd trim.)—NO2 (p.a.): 0.23
2nd trim.)-l\l02 (1st trim.): -0.03
2nd trim.)—NO2 (2nd trim.): 0.17
2nd trim.)—NO2 (3,d trim.): 0.30
3,d trim.)-N02 (p a.): 0.39
3,d trim.)-N02 (1st trim.): 0.33
3,d trim.)-N02 (2nd trim.): 0.21
3,d trim.)-N02 (3rd trim.): 0.23
p.a.)- PM2.B absorbance (p.a.): 0.69
p.a.)- PM2.B abs (1st trim.): 0.33
p.a.)- PM2.6 abs (2nd trim.): 0.48
p.a.)- PM2.B abs (3,d trim.): 0.52
1st trim.)- PM2.B abs (p.a.): 0.68
1st trim.)- PM2.6 abs (1st trim.): 0.27
1st trim.)- PM2.B abs (2nd trim.): 0.53
1st trim.)- PM2.6 abs (3rd trim.): 0.51
2nd trim.)- PM2 b abs(p.a.): 0.41
2nd trim.)- PM2.6 abs (1st trim.): 0.08
2nd trim.)- PM2.6 abs (2nd trim.):
2nd trim.)- PM2.6 abs (3,d trim.): 0.41
3,d trim.)- PM2.B abs (p.a.): 0.62
3,d trim.)- PM2.1, abs (1st trim.): 0.48
3,d trim.)- PM2.B abs (2nd trim.): 0.36
3,d trim.)- PM2.B abs (3rd trim.): 0.37
Single-pollutant models (restricted
analysis to PM2.6 absorbance below the
median)
Per 1 //g/m3: 1.15 (0.89, 1.52)
Prevalence ratios (PRs) of birth weight
<3000 g
Multipollutant models (simultaneous
adjustment of 3rd trimester PM2.6 and
whole pregnancy PM2.6)
PM2.B (whole pregnancy)
Per 1 /yg/m3: 0.96 (0.75, 1.19)
PM2.B (3rd trimester)
Per 1 //g/m3: 1.17 (0.98, 1.40)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over the
whole pregnancy (adjustment for
season of conception)
4th quartile: 1.68 (1.05, 2.75); Per
1 //g/m3: 1.12 (0.97, 1.28)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over first
trimester of pregnancy
Each trimester separately
2nd quartile: 1.14 (0.74,1.96); 3rd
quartile: 1.28(0.84,2.10)
4th quartile: 1.65 (1.02, 2.60)
Per 1 /yglm3: 1.10 (0.99, 1.20)
All trimesters adjusted simultaneously
2nd quartile: 0.97 (0.60,1.73); 3rd
quartile: 0.98 (0.57,1.75)
4th quartile: 1.22(0.71,2.18)
Per 1 //g/m3: 1.03 (0.90, 1.17)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over second
trimester of pregnancy
Each trimester separately
2nd quartile: 0.83 (0.52,1.32); 3rd
quartile: 1.08 (0.71,1.60)
4th quartile: 0.94 (0.61,1.47)
Per 1 //g/m3: 1.01 (0.92,1.12)
All trimesters adjusted simultaneously
2nd quartile: 0.75 (0.46,1.24)
3rd quartile: 0.86 (0.56,1.30);
4th quartile: 0.75(0.48,1.23)
Per 1 //g/m3: 0.94 (0.84,1.06)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over third
trimester of pregnancy
Each trimester separately
2nd quartile: 1.30 (0.80, 2.17)
3rd quartile: 1.44(0.85,2.27)
4th quartile: 1.90(1.20, 2.82)
July 2009
E-623
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Slama et al„ 2007, 093216) Outcome: Birth weight offspring at term
Period of Study: 111998 -111999	Study Design: Cohort study
Location: Munich, Germany	N: 1016 births
Statistical Analyses: Poisson model
Covariates: Maternal passive smoking,
maternal age, gestational duration, sex of
child, parity, maternal education, maternal
size, prepregnancy weight, other
pollutants (PM2.6, PM2.6 absorbance, NO2),
season of conception
Dose-response Investigated? Yes
Statistical Package: STATA
Pollutant: PM2.6 absorbance (estimated)
Averaging Time: Entire pregnancy period
and trimesters
Mean (SD): 1.76*
Percentiles: 25th: 1.61*
50th (Median): 1.72*
75th: 1.89*
Unit (i.e. /yg/m3): 106/m
Monitoring Stations: Spatial component:
40
Temporal component: 1
Copollutant (correlation):
p.a. - pregnancy avg
trim. - trimester
abs -
absorbance
PM2.6
abs
p.a.)—PM2.E abs (1st trim.): 0.54
PM2.6
abs
p.a.)—PM2.B abs (2nd trim.): 0.84
PM2.6
abs
p.a.)—PM2.B abs (3,d trim.): 0.55
PM2.6
abs
p.a.)—PM2.B (p.a.): 0.69
PM2.6
abs
p.a.)—PM2.B (1st trim.): 0.68
PM2.6
abs
p.a.)—PM2.B (2nd trim.): 0.41
PM2.6
abs
p.a.)—PM2.B (3,d trim.): 0.62
PM2.6
abs
p.a.)—NO2 (p.a.): 0.67
PM2.6
abs
p.a.)-N02 (1st trim.): 0.34
PM2.6
abs
p.a.)-N02 (2nd trim.): 0.63
PM2.6
abs
p.a.)—NO2 (3,d trim.): 0.36
PM2.6
0.32
abs
1st trim.)—PM2.B abs (2nd trim.):
PM2.6
abs
1st trim.)—PM2.B abs (3,d trim.): ¦
Per 1 //g/m3: 1.14 (1.02,1.24)
All trimesters adjusted simultaneously
2nd quartile: 1.34 (0.79, 2.30)
3rd quartile: 1.48 (0.86, 2.58)
4th quartile: 1.91 (1.00, 3.20)
Per 1 //g/m3: 1.14 (0.99,1.29)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over third
trimester of pregnancy (adjustment for
season of conception)
All trimesters adjusted simultaneously
Per 1 //g/m3 1.25 (1.04,1.50)
Sensitivity analysisfbootstrapped PR)
2nd quartile: 0.98 (0.63,1.61); 3rd
quartile: 1.22(0.82, 2.02)
4th quartile: 1.57(1.02, 2.57)
Per 1 //g/m3: 1.1 1 (0.98, 1.27)
Estimated increments in prevalence of
birth weight of < 3000 g during
exposure 9 months after birth Per
1 //g/m3: 7% (-7%, 22%)
PM Increment: 1) 0.5 * 106/m 2)
Quartiles: a) 1st (reference) (1.29-1.61)
b)	2nd (1.61-1.72)
c)	3rd (1.72-1.89)
d)	4th (1.89-3.10)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over the
whole pregnancy
Single-pollutant models Unadjusted
models
2nd quartile: 1.19 (0.74,1.99)
3rd quartile: 1.56 (0.98, 2.50);
4th quartile: 1.52 (0.96, 2.46)
Per 0.5 * 10B/m: 1.25 (0.90,1.70)
Adjusted models
2nd quartile: 1.21 (0.73,1.97)
3rd quartile: 1.63 (0.98, 2.57);
4th quartile: 1.78(1.10, 2.70)
Per 0.5 * 10B/m: 1.45 (1.06,1.87)
Multipollutant models Adjusted models
2nd quartile: 1.19 (0.70, 2.01)
3rd quartile: 1.55 (0.80, 2.80);
4th quartile: 1.46 (0.67, 2.90)
Per 0.5 * 10B/m: 1.33 (0.76, 2.38)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over the
whole pregnancy (adjustment for
season of conception)
4th quartile: 1.72(1.08, 2.73)
Per 0.5 * 10B/m: 1.38 (0.96,1.86)
July 2009
E-624
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
0.26
PM2.6 abs
PM2.6 abs
PM2.6 abs
PM2.6 abs
PM2.6 abs
PM2.6 abs
PM2.B abs
PM2.B abs
PM2.6 abs
0.31
PM2.B abs
PM2.6 abs
PM2.B abs
PM2.6 abs
PM2.B abs
PM2.B abs
PM2.6 abs
PM2.B abs
PM2.6 abs
PM2.B abs
PM2.B abs
PM2.B abs
PM2.B abs
PM2.B abs
PM2.B abs
PM2.B abs
" trim.
" trim.
" trim.
" trim.
" trim.
" trim.
" trim.
" trim.
2nd trim.
2nd trim.
2nd trim.
2nd trim.
2nd trim.
2nd trim.
2nd trim.
2nd trim.
2nd trim.
3,d trim.
3,d trim.
3,d trim.
3,d trim.
3,d trim.
3,d trim.
3,d trim.
3,d trim.
-PM2.B (p.a.): 0.33
-PM2.B (1st trim.): 0.27
-PM2.B (2nd trim.): 0.08
-PM2 B (3,d trim ): 0.48
-NO2 (p.a.): 0.29
-NO2 (1st trim.): 0.84
-NO2 (2nd trim.): 0.16
-NO2 (3rd trim.): -0.39
PM2.Babs (3,d trim.):
PM2.B (p.a.): 0.48
PM2.B (1st trim.): 0.53
PM2.B (2nd trim.): 0.29
PM2.B (3rd trim.): 0.36
NO2 (p.a.): 0.61
NO2 (1st trim.): 0.19
NO2 (2nd trim.): 0.85
NO2 (3,d trim.): 0.17
-PM2.B (p.a.): 0.52
-PM2.B (1st trim.): 0.51
—PM2.B (2nd trim.): 0.41
-PM2 B (3,d trim ): 0.37
-NO2 (p.a.): 0.40
-NO2 (1st trim.): -0.34
-NO2 (2nd trim.): 0.21
-NO2 (3rd trim.): 0.88
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over the
whole pregnancy
Single-pollutant models
(restricted analysis to PM2.6 below the
median)
Per 0.5 * 10B/m: 1.67 (0.66, 3.73)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over first
trimester of pregnancy
Each trimester separately
2nd quartile: 1.15 (0.73,1.80)
3rd quartile: 1.01 (0.61,1.53);
4th quartile: 1.04 (0.70,1.57)
Per 0.5 * 10B/m: 1.03 (0.82,1.28)
All trimesters adjusted simultaneously
2nd quartile: 0.90 (0.52,1.58)
3rd quartile: 0.82 (0.45,1.31);
4th quartile: 0.88 (0.53,1.42)
Per 0.5 * 10B/m: 1.02 (0.77,1.29)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over second
trimester of pregnancy
Each trimester separately
2nd quartile: 1.33 (0.85, 2.22)
3rd quartile: 1.76(1.07,2.91);
4th quartile: 1.83(1.11,2.81)
Per 0.5 * 10B/m: 1.27 (1.04,1.54)
All trimesters adjusted simultaneously
2nd quartile: 1.30 (0.77, 2.16)
3rd quartile: 1.63 (0.93, 2.73);
4th quartile: 1.99(1.12, 3.33)
Per 0.5* 10B/m: 1.21 (0.93,1.54)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over third
trimester of pregnancy
Each trimester separately
2nd quartile: 1.30 (0.85, 2.09)
3rd quartile: 0.92 (0.55,1.50);
4th quartile: 1.50(1.00, 2.27)
Per 0.5 * 10B/m: 1.20 (0.98,1.44)
All trimesters adjusted simultaneously
2nd quartile: 0.99 (0.64,1.62)
3rd quartile: 0.71 (0.40,1.20);
4th quartile: 1.14(0.68,1.91)
Per 0.5 * 10B/m: 1.15 (0.92,1.42)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over first
trimester of pregnancy
(adjustment for season of conception)
July 2009
E-625
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
All trimesters adjusted simultaneously
4th quartile: 0.73 (0.38,1.38)
Per 0.5 * 10'6/m: 0.93(0.41,1.32)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over second
trimester of pregnancy (adjustment for
season of conception)
All trimesters adjusted simultaneously
4th quartile: 2.45(1.22, 4.77)
Per 0.5 * 10B/m: 1.14(0.70,1.64)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over third
trimester of pregnancy
(adjustment for season of conception)
All trimesters adjusted simultaneously
4th quartile: 1.19 (0.60, 2.48)
Per 0.5 * 10B/m: 1.29 (0.90,1.75)
Sensitivity analysis (bootstrapped PR)
2nd quartile: 1.19 (0.76,1.91)
3rd quartile: 1.52 (0.99, 2.34);
4th quartile: 1.62 (1.06, 2.55)
Per 0.5 * 10B/m: 1.35 (1.01,1.83)
Estimated increments in prevalence of
birth weight < 3000 g during
exposure 9 months after birth Per 0.5
* 10'6|m: 18% (-16%, 57%)
Reference: Wilhelm et al. (2005,
088668)
Outcome: Term low birth weight (LBW)
(<2500 g at a 37 completed weeks
gestation)
Period of Study: 1994-2000
Vaginal birth < 37 completed weeks
Location: Los Angeles County, California, gestation
U.S.
Age Groups: LBW: a 37 completed
weeks
Preterm births: <37 completed weeks
Study Design: cross-sectional study
N: For LBW: 136,134
For preterm birth:
106,483
Statistical Analyses: Logistic regression
Covariates: Maternal age, maternal race,
maternal education, parity, interval since
previous live birth, level of prenatal care,
infant sex, previous LBW or preterm
infant, birth season, other pollutants (not
specified in birth weight analyses, also
adjusted for gestational age)
Dose-response Investigated? Yes
Statistical Package: NR
Pollutant: PM2.6
Averaging Time: 24hr (every 3 days)
Entire pregnancy
Trimesters of pregnancy
Months of pregnancy
6 weeks before birth
Mean (SD): First trimester: 21.9
Third trimester: 21.0
6 weeks before birth: 21.0
Range (Min, Max):
First trimester: 11.8-38.9
Third trimester: 11.8-.38.9
6 weeks before birth: 9.9-48.5
Monitoring Stations:
Zip-code-level analysis: 9
Address-level analysis: 8
Copollutant (correlation): First trimester
PM2.B-C0: 0.57
PM2.B-N02: 0.73
PM2.B-03: -0.55
PM2.b-PMio: 0.43
Third trimester: PM2.B-C0: 0.67
PM Increment: 1) 10/yg/m3
2) 3 levels: a) < 25%ile (reference)
b)	25%-75%ile
c)	> 75%ile
Incidence of LBW (third trimester
exposure)
<	17.1 //g/m3: 2.4(2.0, 2.8)
17.1 to < 24.0/yglm3: 2.2 (2.0, 2.5)
a 24.0/yg/m3: 2.1 (1.7, 2.4)
Incidence of preterm birth (first
trimester exposure)
<	18.0//g/m3: 10.6 (9.6,11.7)
18.0 to < 25.4/yglm3: 8.8 (8.1, 9.5)
a 25.4/yg/m3: 9.0(8.1,10.0)
Incidence of preterm birth (6 weeks
before birth exposure)
<	16.5/yg/m3: 8.2(7.4,9.1)
16.5 to < 24.7/yg/m3: 8.8 (8.2, 9.4)
>24.7/yg|m3: 9.6(8.7,10.5)
Outcome: Preterm birth
Exposure Period: First trimester of
pregnancy
Address-level analysis: Single-pollutant
model: Distance ~ 1 mile
July 2009
E-626
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
PM2.B-N02: 0.78
PM2.B-03: -0.60
PM2.b-PMio: 0.52
6 weeks before birth: PM2.B-C0: 0.63
PM2.B-N02: 0.74
PM2.B-03: -0.60
PM2.b-PMio: 0.60
Per 10//g/m3: 0.85 (0.70, 1.02)
18.1	to < 25.2//g/m3: 0.91 (0.72,1.16)
>	25.2//g/m3: 0.83 (0.60,1.14)
Single-pollutant model: 1 < distance ~ 2
mile
Per 10//g/m3: 0.85 (0.74,0.99)
18.3 to < 25.2//g/m3: 0.81 (0.69, 0.94)
>25.2//g/m3: 0.79 (0.65, 0.97)
Multipollutant modeM < distance ~ 2
mile
Per 10//g/m3: 1.18(0.84,1.65)
Single-pollutant model: 2 < distance ~ 4
mile
Per 10//g/m3: 0.83 (0.78, 0.88)
18.5 to < 24.9//g/m3: 0.79 (0.74, 0.85)
>	24.9//g/m3: 0.76 (0.70, 0.84)
Zip-code-level analysis: Single-pollutant
model: Per 10//g/m3: 0.73 (0.67, 0.80)
18.0 to < 25.4//g/m3: 0.70 (0.61, 0.80)
a 25.4//g/m3: 0.64 (0.53, 0.76)
Outcome: Preterm birth
Exposure Period: 6 weeks before birth
Address-level analysis:
Single-pollutant model: Distance ~ 1 mile
Per 10//g/m3: 1.09 (0.91, 1.30)
16.8 to <24.1 //g/m3:1.21 (0.97,1.51)
>	24.1 //g/m3:1.25 (0.93, 1.68)
Single-pollutant model: 1 < distance ~ 2
mile
Per 10//g/m3: 1.08 (0.97, 1.21)
17.2	to < 24.5//g/m3: 0.94 (0.82,1.08)
>	24.5//g/m3:1.04 (0.87,1.24)
Single-pollutant model: 2 < distance ~ 4
mile
Per 10//g/m3: 1.05 (0.99, 1.10)
17.3	to < 24.6//g/m3:1.06 (1.00, 1.13)
>	24.6//g/m3:1.08 (0.99, 1.17)
Zip-code-level analysis
Single-pollutant model: Per 10//g/m3:
1.10(1.00,1.21)
16.5 to < 24.7//g/m3:1.06 (0.94,1.20)
>	24.7//g/m3:1.19(1.02,1.40)
(See Notes1)
Multipollutant model
Per 10 //g/m3: 1.12(0.90,1.40)
>	24.6 //g/m3:1.12 (0.82, 1.52)
Notes:11n the table, the 75%ile is noted
as 24.7 /yglm3. However, the text notes
the 75%ile as 24.3/yglm3.
July 2009
E-627
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Woodruff et al. (2006,
088758)
Period of Study: 1999-2000
Location: California
Outcome (ICD10): SIDS (R95)
Respiratory mortality (J00-J99)
Bronchopulmonary dysplasia (P27.1)
External accidents (V01-Y98)
Ill-defined and unspecified causes of
mortality (R99)
Age Groups: > 28 days old
Study Design: Matched case-control
(matched on date of birth and birth
weight)
N: 3877 infants
Statistical Analyses: Conditional logistic
regression
Covariates: Maternal race, education,
parity, age, marital status
Dose-response Investigated? Yes
Statistical Package: STATA
Pollutant: PM2.6
Averaging Time: 24 hrs (every 6 days)
(time period between birth and post
neonatal death for the infant who died
and the same period for its four matched
surviving infants) Percentiles: Infants
who died of all causes (cases)
25th: 13.4
50th (Median): 19.2
75th: 23.6
Matched controls
25th: 13.5
50th (Median): 18.4
75th: 22.7
Monitoring Stations:
73 (from 39 counties)
PM Increment: 10/yg/m3
RR Estimate [Lower CI, Upper CI]
lag:
All-cause mortality: Unadjusted: 1.15
(1.00,1.32)
Adjusted: 1.07 (0.93,1.24)
Cause-specific mortality: Respiratory (all):
Unadjusted: 2.15(1.15, 4.02)
Adjusted: 2.13(1.12,4.05)
Respiratory (excluding deaths due to
BPD): Adjusted: 1.42 (0.66, 3.03)
Respiratory (BPD alone): Unadjusted: 6.00
(1.40, 27.76)
Respiratory (low birth weight infants
only): Unadjusted: 3.09 (1.14, 8.40)
Respiratory (normal birth weight infants
only): Unadjusted: 1.66 (0.74, 3.70)
Respiratory (with matched PM2.6 averaged
over all monitors in county)
Adjusted: 2.28 (0.94, 5.52)
Respiratory (averaging all PM2.6
measurements in county over the 2-year
study period): Adjusted: 2.26 (0.83, 6.21)
SIDS: Unadjusted: 0.86 (0.61,1.22)
Adjusted: 0.82 (0.55,1.23)
SIDS (includes ICD10 code R99: ill-defined
and unspecified causes of mortality):
Adjusted: 1.03 (0.79,1.35)
External causes: Unadjusted: 0.91 (0.56,
1.47)
Adjusted: 0.83 (0.50,1.39)
Compare against the lowest quartile,
estimates for respiratory-specific
mortality were provided: 2nd quartile: 1.28
(0.47,3.51)
3,d quartile: 1.75(0.65, 4.72)
4th quartile: 2.35 (0.85, 6.54)
July 2009
E-628
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Woodruff et al. (2008,
098386)
Period of Study: 1999-2002
Location: US counties with >250,000
residents (96 counties)
All other deaths evaluated as a control
category
Age Groups: Infants aged > 28 days and
<1 yr
Study Design: Cross-sectional
N: 3,583,495 births (6,639 post neonatal
deaths)
Statistical Analyses: Logistic GEE
(exchangeable correlation structure)
Covariates: maternal race/ethnicity,
marital status, age, education, primiparity,
county-level poverty and per capita
income levels, year and month of birth
dummy variables to account for time trend
and seasonal effects, and region of the
country
sensitivity analyses performed among only
those mothers with smoking information
(adjustment for smoking had no effect on
the estimates)
Season: Adjusted for year and month of
birth dummy variables to account for time
trend and seasonal effects
Dose-response Investigated? Evaluated
the appropriateness of a linear form from
analysis based on quartiles of exposure
and concluded that linear form was
appropriate (data not shown)
Statistical Package: SAS
'All units expressed in //gfm3 unless otherwise specified.
Outcome (ICD10): Postneonatal deaths:
Respiratory mortality (J000-99, plus
bronchopulmonary dysplasia [BPD] P27.1)
SIDS (R95)
Ill-defined causes (R99)
Pollutant: PM2.6
Averaging Time: Measured continuously
for 24 h once every 6 days
exposure assigned by calculating avg
concentration of pollutant during first 2
months of life
Median and IQR (25th-75th percentile):
Survivors: 14.8(11.7-18.7)
All causes of death: 14.9 (12.0-18.6)
Respiratory: 14.8(11.5-18.5)
SIDS: 14.5(12.0-17.5)
SIDS + ill-defined: 14.8(12.1-18.5)
Other causes: 14.9 (12.0-18.6)
Percentiles: See above
PM Component: Not assessed, but
controlled for region of the country to
account for PM composition variation
Monitoring Stations: NR
Copollutant (correlation):
PMio (r - 0.34)
PM2.6
CO (r - 0.35)
SO2 (r - 0.21)
O3 (r - -0.10)
Notes: Monthly averages calculated if
there were at least 3 available measures
for PM
Assigned exposures using the avg
concentration of the county of residence
PM Increment: IQR (7/yglm3)
Effect Estimate [Lower CI, Upper CI]:
Adjusted ORs for single pollutant models
All causes: 1.04 (0.98,1.11)
Respiratory: 1.11 (0.96,1.29)
SIDS: 1.01 (0.86,1.20)
Ill-defined + SIDS: 1.06(0.97,1.17)
Other causes: 1.03 (0.96,1.12)
Adjusted ORs for multipollutant models
(including CO, O3, SO2)
Respiratory: 1.05 (0.89,1.24)
SIDS: 1.04 (0.87,1.23)
OR for respiratory deaths assessing
exposure as quartiles
Highest vs Lowest quartile: 1.39 (1.04,
1.85)
July 2009
E-629
DRAFT-DO NOT CITE OR QUOTE

-------
E.8. Long-Term Exposure and Mortality
Table E-30. Long-term exposure - mortality - PM10.
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Breitner et al„ 2009,
188439)
Period of Study: 10/1 /1991 to
313112002
Location: Efurt, Germany
Outcome: Mortality, excluding infants
and ICD-9 > 800
Study Design: Time-series
Covariates: seasonal and weekday
variations, influenza epidemics, air
temperature, relative humidity
Statistical Analysis: semiparametric
Poisson regression, polynomial distributed
lag (PDL)
Statistical Package: R
Age Groups: All
Pollutant: PMio
Averaging Time: daily
Mean (SD) Unit:
1 (10/1/1991-8/31/1995): 50.6 ± 32.2
/yg/m3
2(9/1/1995-2/28/19981:41.1 ±28.4
/yg/m3
3(3/1/1998-3/31/2002): 24.3 ± 15.4
/yg/m3
Total: 38.0 ±28.3/yg/m3
Range (Min, Max): NR
Copollutant: NO2, CO, UFP
Increment: IQR
Relative Risk (95% CI)
Lag
New City Limits
6-day IQR: 17.2
PDL: 0.997 (0.972-1.022)
Mean of lags 0-5: 0.995 (0.971-1.019)
Old City Limits
6-day IQR: 17.2
PDL: 1.004 (0.978-1.031)
Mean of lags 0-5: 1.001 (0.976-1.027)
New City Limits
15-day IQR: 14.5
PDL: 1.008 (0.982-1.036)
Mean of lags 0-14: 1.006 (0.981-1.032)
Old City Limits
15-day IQR: 14.5
PDL: 1.019(0.991-1.048)
Mean of lags 0-14:1.017(0.990-1.044)
Multiday Moving Averages, 6-day
Overall IQR: 24.2
Overall RR (95% CI): 0.998 (0.976-
1.021)
Period 1: 0.996 (0.969-1.024)
Period 2: 1.013 (0.972-1.056)
Period 3: 0.949 (0.897-1.004)
Multiday Moving Averages, 15-day
Overall IQR: 22.3
Overall RR (95% CI): 1.020 (0.993-
1.093)
Period 1: 1.017 (0.984-1.051)
Period 2: 1.012(0.973-1.071)
Period 3: 0.978(0.911-1.051)
Reference: (Slama et al„ 2007, 093216) Outcome: Birth weight offspring at term
Period of Study: 1/1998-1/1999
Location: Munich, Germany
Study Design: Cohort study
N: 1016 births
Statistical Analyses: Poisson model
Covariates: Maternal passive smoking,
maternal age, gestational duration, sex of
child, parity, maternal education, maternal
size, prepregnancy weight, other
pollutants (PM2.6, PM2.6 absorbance, NO2),
Pollutant: PM2.6 (estimated based on
larger PM size fractions)
Averaging Time: Entire pregnancy period
and trimesters
Mean (SD): 14.4
Percentiles: 25th: 13.5
50th (Median): 14.4
75th: 15.4
PM Increment: 1) 1 /yg/m3
2) Quartiles: a) 1st (reference) (7.2-
13.5/yg/m3)
b)	2nd (13.5-14.4 /yg/m3)
c)	3rd (14.4-15.4/yg/m3)
d)	4th (15.41-17.5/yg/m3)
Prevalence ratios (PRs) of birth weight
< 3000 g during exposure over the
July 2009
E-630
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
season of conception
Dose-response Investigated? Yes
Statistical Package: STATA
Monitoring Stations: Spatial component:
40
Temporal component: 1
Copollutant (correlation!:
p.a. - pregnancy avg
trim. - trimester
PM2.6
PM2.6
PM2.6
PM2.6
PM2.B
PM2.6
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.6
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.6
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
PM2.B
0.29
PM2.B
PM2.6
PM2.B
PM2.B
PM2.B
p.a.)—PM2.B (1st trim.): 0.85
p.a.)—PM2.B (2nd trim.): 0.77
p.a.)—PM2.B (3,d trim.): 0.87
p.a.)—IMO2 (p.a.): 0.45
p.a.)—IMO2 (1st trim.): 0.18
p.a.)—NO2 (2nd trim.): 0.32
p.a.)—NO2 (3rd trim.): 0.37
1st trim.)—PM2.B (2nd trim.): 0.40
1st trim.)-PM2.B (3rd trim.): 0.68
1st trim.)-PJ02 (p.a.): 0.48
1st trim.)—IMO2 (1st trim.): 0.15
1st trim.)-PJ02 (2nd trim.): 0.41
1st trim.)-PJ02 (3rd trim.): 0.39
2nd trim.)—PM2.B (3,d trim.): 0.51
2nd trim.)-N02 (p.a.): 0.23
2nd trim.)-N02 (1st trim.): -0.03
2nd trim.)—NO2 (2nd trim.): 0.17
2nd trim.)—NO2 (3,d trim.): 0.30
3,d trim.)-N02 (p.a.): 0.39
3,d trim.)-N02 (1st trim.): 0.33
3,d trim.)-N02 (2nd trim.): 0.21
3,d trim.)-N02 (3rd trim.): 0.23
p.a.)- PM2.6 absorbance (p.a.): 0.69
p.a.)- PM2.6 abs (1st trim.): 0.33
p.a.)- PM2.6 abs (2nd trim.): 0.48
p.a.)- PM2.B abs (3,d trim.): 0.52
1st trim.)- PM2.6 abs (p.a.): 0.68
1st trim.)- PM2.6 abs (1st trim.): 0.27
1st trim.)- PM2.B abs (2nd trim.): 0.53
1st trim.)- PM2.6 abs (3rd trim.): 0.51
2nd trim.)- PM2.B abs(p.a.): 0.41
2nd trim.)- PM2.B abs (1st trim.): 0.08
2nd trim.)- PM2.6 abs (2nd trim.):
2nd trim.)- PM2.B abs (3,d trim.): 0.41
3,d trim.)- PM2.6 abs (p.a.): 0.62
3,d trim.)- PM2.B abs (1st trim.): 0.48
3,d trim.)- PM2.6 abs (2nd trim.): 0.36
3,d trim.)- PM2.1, abs (3rd trim.): 0.37
whole pregnancy
Single-pollutant models
Unadjusted models
2nd quartile: 1.07 (0.65,1.73); 3rd
quartile: 1.38 (0.91, 2.09)
4th quartile: 1.45 (0.92, 2.25)
Per 1 /yglm3: 1.06 (0.95, 1.19)
Adjusted models
2nd quartile: 1.08 (0.63,1.82); 3rd
quartile: 1.34 (0.86, 2.13)
4th quartile: 1.73 (1.15, 2.69); Per
1 //g/m3: 1.13 (1.00, 1.29)
Multipollutant models
Adjusted models
2nd quartile: 1.01 (0.57,1.85)
3rd quartile: 1.12(0.64,1.87)
4th quartile: 1.36 (0.72, 2.45); Per
1 //g/m3: 1.07 (0.91, 1.26)
Single-pollutant models (restricted
analysis to PM2.6 absorbance below the
median)
Per 1 //g/m3: 1.15 (0.89, 1.52)
Prevalence ratios (PRs) of birth weight
<3000 g
Multipollutant models (simultaneous
adjustment of 3rd trimester PM2.6 and
whole pregnancy PM2.6)
PM2.B (whole pregnancy)
Per 1 //g/m3: 0.96 (0.75, 1.19)
PM2.6 (3rd trimester)
Per 1 //g/m3: 1.17 (0.98, 1.40)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over the
whole pregnancy (adjustment for
season of conception)
4th quartile: 1.68 (1.05, 2.75); Per
1 //g/m3: 1.12 (0.97, 1.28)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over first
trimester of pregnancy
Each trimester separately
2nd quartile: 1.14(0.74,1.96); 3rd
quartile: 1.28(0.84,2.10)
4th quartile: 1.65 (1.02, 2.60)
Per 1 //g/m3: 1.10 (0.99, 1.20)
All trimesters adjusted simultaneously
2nd quartile: 0.97 (0.60,1.73); 3rd
quartile: 0.98 (0.57,1.75)
4th quartile: 1.22(0.71,2.18)
Per 1 //g/m3: 1.03 (0.90, 1.17)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over second
trimester of pregnancy
July 2009
E-631
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Each trimester separately
2nd quartile: 0.83 (0.52,1.32); 3rd
quartile: 1.08 (0.71,1.60)
4th quartile: 0.94 (0.61,1.47)
Per 1 //g/m3: 1.01 (0.92,1.12)
All trimesters adjusted simultaneously
2nd quartile: 0.75 (0.46,1.24)
3rd quartile: 0.86 (0.56,1.30);
4th quartile: 0.75(0.48,1.23)
Per 1 //g/m3: 0.94 (0.84,1.06)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over third
trimester of pregnancy
Each trimester separately
2nd quartile: 1.30 (0.80, 2.17)
3rd quartile: 1.44(0.85,2.27)
4th quartile: 1.90(1.20, 2.82)
Per 1 //g/m3: 1.14 (1.02,1.24)
All trimesters adjusted simultaneously
2nd quartile: 1.34 (0.79, 2.30)
3rd quartile: 1.48 (0.86, 2.58)
4th quartile: 1.91 (1.00, 3.20)
Per 1 //g/m3: 1.14 (0.99,1.29)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over third
trimester of pregnancy (adjustment for
season of conception)
All trimesters adjusted simultaneously
Per 1 /yg|m3:1.25 (1.04,1.50)
Sensitivity analysisfbootstrapped PR)
2nd quartile: 0.98 (0.63,1.61); 3rd
quartile: 1.22(0.82, 2.02)
4th quartile: 1.57(1.02, 2.57)
Per 1 //g/m3: 1.1 1 (0.98, 1.27)
Estimated increments in prevalence of
birth weight of < 3000 g during
exposure 9 months after birth Per
1 /yglm3: 7% (-7%, 22%)
Reference: (Slama et al„ 2007, 093216) Outcome: Birth weight offspring at term
Period of Study: 111998 -111999	Study Design: Cohort study
Location: Munich, Germany	N: 1016 births
Statistical Analyses: Poisson model
Covariates: Maternal passive smoking,
maternal age, gestational duration, sex of
child, parity, maternal education, maternal
size, prepregnancy weight, other
pollutants (PM2.6, PM2.6 absorbance, NO2),
season of conception
Dose-response Investigated? Yes
Statistical Package: STATA
Pollutant: PM2.6 absorbance (estimated)
Averaging Time: Entire pregnancy period
and trimesters
Mean (SD): 1.76*
Percentiles: 25th: 1.61*
50th (Median): 1.72*
75th: 1.89*
Unit (i.e. /yg/m3): 106/m
Monitoring Stations: Spatial component:
40
Temporal component: 1
Copollutant (correlation):
PM Increment: 1) 0.5 * 106/m 2)
Quartiles: a) 1st (reference) (1.29-1.61)
b)	2nd (1.61-1.72)
c)	3rd (1.72-1.89)
d)	4th (1.89-3.10)
Prevalence ratios (PRs) of birth weight
< 3000 g during exposure over the
whole pregnancy
Single-pollutant models Unadjusted
models
2nd quartile: 1.19 (0.74,1.99)
3rd quartile: 1.56 (0.98, 2.50);
4th quartile: 1.52 (0.96, 2.46)
July 2009
E-632
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
p.a. - pregnancy avg
Per 0.5 * 10B/m: 1.25 (0.90,1.70)
trim. - trimester
Adjusted models
abs - absorbance
2nd quartile: 1.21 (0.73,1.97)
PM2.6 abs
p.a.)—PM2.B abs (1st trim.): 0.54
3rd quartile: 1.63 (0.98, 2.57);
PM2.6 abs
p.a.)—PM2.B abs (2nd trim.): 0.84
4th quartile: 1.78(1.10, 2.70)
PM2.6 abs
p.a.)—PM2.B abs (3,d trim.): 0.55
Per 0.5 * 10B/m: 1.45 (1.06,1.87)
PM2.6 abs
p.a.)—PM2.B (p.a.): 0.69
Multipollutant models Adjusted models
PM2.6 abs
p.a.)—PM2.B (1st trim.): 0.68
2nd quartile: 1.19 (0.70, 2.01)
PM2.6 abs
p.a.)—PM2.B (2nd trim.): 0.41
3rd quartile: 1.55 (0.80, 2.80);
PM2.6 abs
p.a.)—PM2.B (3,d trim.): 0.62
4th quartile: 1.46 (0.67, 2.90)
PM2.6 abs
p.a.)—NO2 (p.a.): 0.67
Per 0.5 * 10B/m: 1.33 (0.76, 2.38)
PM2.B abs
p.a.)—NO2 (1st trim.): 0.34
Prevalence ratios (PRs) of birth weight
PM2.6 abs
p.a.)—NO2 (2nd trim.): 0.63
< 3000 g during exposure over the

whole pregnancy (adjustment for
PM2.B abs
p.a.)—NO2 (3,d trim.): 0.36
season of conception)
PM2.6 abs
0.32
1st trim.)—PM2.B abs (2nd trim.):
4th quartile: 1.72(1.08, 2.73)

Per 0.5 * 10B/m: 1.38 (0.96,1.86)
PM2.B abs
1st trim.)—PM2.B abs (3,d trim.): ¦
Prevalence ratios (PRs) of birth weight
0.26

< 3000 g during exposure over the
PM2.6 abs
1st trim.)-PM2.B (p.a.): 0.33
whole pregnancy
PM2.B abs
1st trim.)-PM2.B (1st trim.): 0.27
Single-pollutant models
PM2.6 abs
1st trim.)—PM2.B (2nd trim.): 0.08
(restricted analysis to PM2.E below the

median)
PM2.B abs
1st trim.)—PM2.B (3,d trim.): 0.48


Per 0.5 * 10B/m: 1.67 (0.66, 3.73)
PM2.B abs
1st trim.)-PJ02 (p.a ): 0.29

Prevalence ratios (PRs) of birth weight
PM2.6 abs
1st trim.)—IMO2 (1st trim.): 0.84
< 3000 g during exposure over first
PM2.B abs
1st trim.)—IMO2 (2nd trim.): 0.16
trimester of pregnancy
PM2.6 abs

Each trimester separately
1st trim.)-l\l02 (3,d trim.): -0.39
PM2.B abs
2nd trim.)—PM2.B abs (3,d trim.):
2nd quartile: 1.15 (0.73,1.80)
0.31

3rd quartile: 1.01 (0.61,1.53);
PM2.6 abs
2nd trim.)-PM2.B (p.a.): 0.48
4th quartile: 1.04 (0.70,1.57)
PM2.B abs
2nd trim.)—PM2.B (1st trim.): 0.53
Per 0.5 * 10B/m: 1.03 (0.82,1.28)
PM2.B abs
2nd trim.)—PM2.B (2nd trim.): 0.29
All trimesters adjusted simultaneously
PM2.6 abs
2nd trim.)-PM2 b (3,d trim ): 0.36
2nd quartile: 0.90 (0.52,1.58)
PM2.B abs
2nd trim.)—IMO2 (p.a.): 0.61
3rd quartile: 0.82 (0.45,1.31);
PM2.6 abs
2nd trim.)—IMO2 (1st trim.): 0.19
4th quartile: 0.88 (0.53,1.42)
PM2.B abs
2nd trim.)-l\l02 (2nd trim ): 0.85
Per 0.5 * 10B/m: 1.02 (0.77,1.29)
PM2.B abs
2nd trim.)—IMO2 (3,d trim.): 0.17
Prevalence ratios (PRs) of birth weight
PM2.B abs
3,d trim.)—PM2.B (p.a.): 0.52
< 3000 g during exposure over second
trimester of pregnancy
PM2.B abs
3,d trim.)-PM2B (1st trim ): 0.51
Each trimester separately
PM2.B abs
3,d trim.)—PM2.B (2nd trim.): 0.41
2nd quartile: 1.33 (0.85, 2.22)
PM2.B abs
3,d trim.)—PM2.B (3,d trim.): 0.37
3rd quartile: 1.76(1.07,2.91);
PM2.B abs
3,d trim.)—NO2 (p.a.): 0.40
4th quartile: 1.83(1.11,2.81)
PM2.B abs
3,d trim.)—NO2 (1st trim.): -0.34
Per 0.5 * 10B/m: 1.27 (1.04,1.54)
PM2.B abs
3,d trim.)—NO2 (2nd trim.): 0.21
All trimesters adjusted simultaneously
PM2.B abs
3,d trim.)—NO2 (3,d trim.): 0.88
2nd quartile: 1.30 (0.77, 2.16)
3rd quartile: 1.63 (0.93, 2.73);
4th quartile: 1.99(1.12, 3.33)
July 2009
E-633
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Per 0.5* 10B/m: 1.21 (0.93,1.54)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over third
trimester of pregnancy
Each trimester separately
2nd quartile: 1.30 (0.85, 2.09)
3rd quartile: 0.92 (0.55,1.50);
4th quartile: 1.50(1.00, 2.27)
Per 0.5 * 10B/m: 1.20 (0.98,1.44)
All trimesters adjusted simultaneously
2nd quartile: 0.99 (0.64,1.62)
3rd quartile: 0.71 (0.40,1.20);
4th quartile: 1.14(0.68,1.91)
Per 0.5 * 10B/m: 1.15 (0.92,1.42)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over first
trimester of pregnancy
(adjustment for season of conception)
All trimesters adjusted simultaneously
4th quartile: 0.73 (0.38,1.38)
Per 0.5 * 10B/m: 0.93(0.41,1.32)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over second
trimester of pregnancy (adjustment for
season of conception)
All trimesters adjusted simultaneously
4th quartile: 2.45(1.22, 4.77)
Per 0.5 * 10B/m: 1.14(0.70,1.64)
Prevalence ratios (PRs) of birth weight
<	3000 g during exposure over third
trimester of pregnancy
(adjustment for season of conception)
All trimesters adjusted simultaneously
4th quartile: 1.19 (0.60, 2.48)
Per 0.5 * 10B/m: 1.29 (0.90,1.75)
Sensitivity analysis (bootstrapped PR)
2nd quartile: 1.19 (0.76,1.91)
3rd quartile: 1.52 (0.99, 2.34);
4th quartile: 1.62 (1.06, 2.55)
Per 0.5 * 10B/m: 1.35 (1.01,1.83)
Estimated increments in prevalence of
birth weight < 3000 g during
exposure 9 months after birth Per 0.5
* 10'6|m: 18% (-16%, 57%)
Reference: Wilhelm et al. (2005,
088668)
Outcome: Term low birth weight (LBW)
(<2500 g at a 37 completed weeks
gestation)
Period of Study: 1994-2000
Vaginal birth < 37 completed weeks
Location: Los Angeles County, California, gestation
U.S.
Age Groups: LBW: a 37 completed
weeks
Preterm births: <37 completed weeks
Pollutant: PM2.6
Averaging Time: 24hr (every 3 days)
Entire pregnancy
Trimesters of pregnancy
Months of pregnancy
6 weeks before birth
PM Increment: 1) 10/yg/m3
2) 3 levels: a) < 25%ile (reference)
b)	25%-75%ile
c)	> 75%ile
Incidence of LBW (third trimester
exposure)
July 2009
E-634
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Study Design: cross-sectional study
N: For LBW: 136,134
For preterm birth:
106,483
Statistical Analyses: Logistic regression
Covariates: Maternal age, maternal race,
maternal education, parity, interval since
previous live birth, level of prenatal care,
infant sex, previous LBW or preterm
infant, birth season, other pollutants (not
specified in birth weight analyses, also
adjusted for gestational age)
Dose-response Investigated? Yes
Statistical Package: NR
Mean (SD): First trimester: 21.9
Third trimester: 21.0
6 weeks before birth: 21.0
Range (Min, Max):
First trimester: 11.8-38.9
Third trimester: 11.8-.38.9
6 weeks before birth: 9.9-48.5
Monitoring Stations:
Zip-code-level analysis: 9
Address-level analysis: 8
Copollutant (correlation): First trimester
PM2.B-C0: 0.57
PM2.B-N02: 0.73
PM2.B-03: -0.55
PM2.b-PMio: 0.43
Third trimester: PM2.B-C0: 0.67
PM2.B-N02: 0.78
PM2.B-03: -0.60
PM2.b-PMio: 0.52
6 weeks before birth: PM2.B-C0: 0.63
PM2.B-N02: 0.74
PM2.B-03: -0.60
PM2.b-PMio: 0.60
<	17.1 //g/m3: 2.4(2.0, 2.8)
17.1 to < 24.0//g/m3: 2.2 (2.0, 2.5)
a 24.0//g/m3: 2.1 (1.7, 2.4)
Incidence of preterm birth (first
trimester exposure)
<	18.0//g/m3: 10.6 (9.6,11.7)
18.0	to < 25.4//g/m3: 8.8 (8.1, 9.5)
a 25.4//g/m3: 9.0(8.1,10.0)
Incidence of preterm birth (6 weeks
before birth exposure)
<	16.5//g/m3: 8.2(7.4,9.1)
16.5 to < 24.7//g/m3: 8.8 (8.2, 9.4)
>24.7//g/m3: 9.6(8.7,10.5)
Outcome: Preterm birth
Exposure Period: First trimester of
pregnancy
Address-level analysis: Single-pollutant
model: Distance ~ 1 mile
Per 10//g/m3: 0.85 (0.70, 1.02)
18.1	to < 25.2//g/m3: 0.91 (0.72,1.16)
>	25.2//g/m3: 0.83 (0.60,1.14)
Single-pollutant model: 1 < distance ~ 2
mile
Per 10//g/m3: 0.85 (0.74,0.99)
18.3 to < 25.2//g/m3: 0.81 (0.69, 0.94)
>25.2//g/m3: 0.79 (0.65, 0.97)
Multipollutant modeM < distance ~ 2
mile
Per 10//g/m3: 1.18(0.84,1.65)
Single-pollutant model: 2 < distance ~ 4
mile
Per 10//g/m3: 0.83 (0.78, 0.88)
18.5 to < 24.9//g/m3: 0.79 (0.74, 0.85)
>24.9//g/m3: 0.76 (0.70, 0.84)
Zip-code-level analysis: Single-pollutant
model: Per 10//g/m3: 0.73 (0.67, 0.80)
18.0 to < 25.4//g/m3: 0.70 (0.61, 0.80)
a 25.4//g/m3: 0.64 (0.53, 0.76)
Outcome: Preterm birth
Exposure Period: 6 weeks before birth
Address-level analysis:
Single-pollutant model: Distance ~ 1 mile
Per 10//g/m3: 1.09 (0.91, 1.30)
16.8 to <24.1 //g/m3:1.21 (0.97,1.51)
>	24.1 //g/m3:1.25 (0.93, 1.68)
Single-pollutant model: 1 < distance ~ 2
mile
Per 10//g/m3: 1.08 (0.97, 1.21)
17.2	to < 24.5//g/m3: 0.94 (0.82,1.08)
July 2009
E-635
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
>	24.5//gfm3:1.04 (0.87,1.24)
Single-pollutant model: 2 < distance ~ 4
mile
Per 10//g/m3: 1.05 (0.99, 1.10)
17.3 to < 24.6//gfm3:1.06 (1.00, 1.13)
>	24.6//g/m3:1.08 (0.99, 1.17)
Zip-code-level analysis
Single-pollutant model: Per 10//g/m3:
1.10(1.00,1.21)
16.5 to < 24.7//g/m3:1.06 (0.94,1.20)
>	24.7//gfm3:1.19(1.02,1.40)
(See Notes1)
Multipollutant model
Per 10 //g/m3: 1.12(0.90,1.40)
>	24.6 //g/m3:1.12 (0.82, 1.52)
Notes:11n the table, the 75%ile is noted
as 24.7 //g|m3. However, the text notes
the 75%ile as 24.3//g|m3.
July 2009
E-636
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Woodruff et al. (2006,
088758)
Period of Study: 1999-2000
Location: California
Outcome (ICD10): SIDS (R95)
Respiratory mortality (J00-J99)
Bronchopulmonary dysplasia (P27.1)
External accidents (V01-Y98)
Ill-defined and unspecified causes of
mortality (R99)
Age Groups: > 28 days old
Study Design: Matched case-control
(matched on date of birth and birth
weight)
N: 3877 infants
Statistical Analyses: Conditional logistic
regression
Covariates: Maternal race, education,
parity, age, marital status
Dose-response Investigated? Yes
Statistical Package: STATA
Pollutant: PM2.6
Averaging Time: 24 hrs (every 6 days)
(time period between birth and post
neonatal death for the infant who died
and the same period for its four matched
surviving infants) Percentiles: Infants
who died of all causes (cases)
25th: 13.4
50th (Median): 19.2
75th: 23.6
Matched controls
25th: 13.5
50th (Median): 18.4
75th: 22.7
Monitoring Stations:
73 (from 39 counties)
PM Increment: 10/yg/m3
RR Estimate [Lower CI, Upper CI]
lag:
All-cause mortality: Unadjusted: 1.15
(1.00,1.32)
Adjusted: 1.07 (0.93,1.24)
Cause-specific mortality: Respiratory (all):
Unadjusted: 2.15(1.15, 4.02)
Adjusted: 2.13(1.12,4.05)
Respiratory (excluding deaths due to
BPD): Adjusted: 1.42 (0.66, 3.03)
Respiratory (BPD alone): Unadjusted: 6.00
(1.40, 27.76)
Respiratory (low birth weight infants
only): Unadjusted: 3.09 (1.14, 8.40)
Respiratory (normal birth weight infants
only): Unadjusted: 1.66 (0.74, 3.70)
Respiratory (with matched PM2.6 averaged
over all monitors in county)
Adjusted: 2.28 (0.94, 5.52)
Respiratory (averaging all PM2.6
measurements in county over the 2-year
study period): Adjusted: 2.26 (0.83, 6.21)
SIDS: Unadjusted: 0.86 (0.61,1.22)
Adjusted: 0.82 (0.55,1.23)
SIDS (includes ICD10 code R99: ill-defined
and unspecified causes of mortality):
Adjusted: 1.03 (0.79,1.35)
External causes: Unadjusted: 0.91 (0.56,
1.47)
Adjusted: 0.83 (0.50,1.39)
Compare against the lowest quartile,
estimates for respiratory-specific
mortality were provided: 2nd quartile: 1.28
(0.47,3.51)
3,d quartile: 1.75(0.65, 4.72)
4th quartile: 2.35 (0.85, 6.54)
July 2009
E-637
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Woodruff et al. (2008,
098386)
Period of Study: 1999-2002
Location: US counties with >250,000
residents (96 counties)
All other deaths evaluated as a control
category
Age Groups: Infants aged > 28 days and
<1 yr
Study Design: Cross-sectional
N: 3,583,495 births (6,639 post neonatal
deaths)
Statistical Analyses: Logistic GEE
(exchangeable correlation structure)
Covariates: maternal race/ethnicity,
marital status, age, education, primiparity,
county-level poverty and per capita
income levels, year and month of birth
dummy variables to account for time trend
and seasonal effects, and region of the
country
sensitivity analyses performed among only
those mothers with smoking information
(adjustment for smoking had no effect on
the estimates)
Season: Adjusted for year and month of
birth dummy variables to account for time
trend and seasonal effects
Dose-response Investigated? Evaluated
the appropriateness of a linear form from
analysis based on quartiles of exposure
and concluded that linear form was
appropriate (data not shown)
Statistical Package: SAS
'All units expressed in //gfm3 unless otherwise specified.
Outcome (ICD10): Postneonatal deaths:
Respiratory mortality (J000-99, plus
bronchopulmonary dysplasia [BPD] P27.1)
SIDS (R95)
Ill-defined causes (R99)
Pollutant: PM2.6
Averaging Time: Measured continuously
for 24 h once every 6 days
exposure assigned by calculating avg
concentration of pollutant during first 2
months of life
Median and IQR (25th-75th percentile):
Survivors: 14.8(11.7-18.7)
All causes of death: 14.9 (12.0-18.6)
Respiratory: 14.8(11.5-18.5)
SIDS: 14.5(12.0-17.5)
SIDS + ill-defined: 14.8(12.1-18.5)
Other causes: 14.9 (12.0-18.6)
Percentiles: See above
PM Component: Not assessed, but
controlled for region of the country to
account for PM composition variation
Monitoring Stations: NR
Copollutant (correlation):
PMio (r - 0.34)
PM2.6
CO (r - 0.35)
SO2 (r - 0.21)
O3 (r - -0.10)
Notes: Monthly averages calculated if
there were at least 3 available measures
for PM
Assigned exposures using the avg
concentration of the county of residence
PM Increment: IQR (7/yglm3)
Effect Estimate [Lower CI, Upper CI]:
Adjusted ORs for single pollutant models
All causes: 1.04 (0.98,1.11)
Respiratory: 1.11 (0.96,1.29)
SIDS: 1.01 (0.86,1.20)
Ill-defined + SIDS: 1.06(0.97,1.17)
Other causes: 1.03 (0.96,1.12)
Adjusted ORs for multipollutant models
(including CO, O3, SO2)
Respiratory: 1.05 (0.89,1.24)
SIDS: 1.04 (0.87,1.23)
OR for respiratory deaths assessing
exposure as quartiles
Highest vs Lowest quartile: 1.39 (1.04,
1.85)
July 2009
E-638
DRAFT-DO NOT CITE OR QUOTE

-------
E.9. Long-Term Exposure and Mortality
Table E-31. Long-term exposure - mortality - PM10.
Reference: (Breitner et al„ 2009,
188439)
Period of Study: 10/1 /1991 to
313112002
Location: Efurt, Germany
	Design & Methods	
Outcome: Mortality, excluding infants
and ICD-9 > 800
Study Design: Time-series
Covariates: seasonal and weekday
variations, influenza epidemics, air
temperature, relative humidity
Statistical Analysis: semiparametric
Poisson regression, polynomial distributed
lag (PDL)
Statistical Package: R
Age Groups: All
Concentrations' ~
Pollutant: PMio
Averaging Time: daily
Mean (SD) Unit:
1 (10/1/1991-8/31/1995): 50.6 ± 32.2
Z^g/m3
2(9/1/1995-2/28/19981:41.1 ± 28.4
/yg/m3
3(3/1/1998-3/31/20021:24.3 ± 15.4
/yg/m3
Total: 38.0 ±28.3/yg/m3
Range (Min, Max): NR
Copollutant: NO2, CO, UFP
__Effect_EstimatesJ95%_CI)__
Increment: IQR
Relative Risk (95% CI)
Lag
New City Limits
6-day IQR: 17.2
PDL: 0.997 (0.972-1.022)
Mean of lags 0-5: 0.995 (0.971-1.019)
Old City Limits
6-day IQR: 17.2
PDL: 1.004 (0.978-1.031)
Mean of lags 0-5: 1.001 (0.976-1.027)
New City Limits
15-day IQR: 14.5
PDL: 1.008 (0.982-1.036)
Mean of lags 0-14: 1.006(0.981-1.032)
Old City Limits
15-day IQR: 14.5
PDL: 1.019(0.991-1.048)
Mean of lags 0-14: 1.017(0.990-1.044)
Multiday Moving Averages, 6-day
Overall IQR: 24.2
Overall RR (95% CI): 0.998 (0.976-
1.021)
Period 1: 0.996 (0.969-1.024)
Period 2: 1.013 (0.972-1.056)
Period 3: 0.949 (0.897-1.004)
Multiday Moving Averages, 15-day
Overall IQR: 22.3
Overall RR (95% CI): 1.020 (0.993-
1.093)
Period 1: 1.017 (0.984-1.051)
Period 2: 1.012(0.973-1.071)
Period 3: 0.978(0.911-1.051)
July 2009
E-639
DRAFT-DO NOT CITE OR QUOTE

-------
Table E-32. Long-term exposure - mortality - PMio2.5.
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: (Chen et al., 2005, 087942)
Period of Study: 1973-1998
Outcome: Mortality: CHD
Study Design: Cohort
Location: San Francisco, San Diego, Los Statistical Analyses: Cox proportion
Angeles, CA	hazards model
Age Groups: >25
Pollutant: PMio-2.5
Averaging Time: 25 years
Mean (SD): 25.4
Range (Min, Max): NR
Copollutant: NO2
0s
SO?
Increment: 10 //g/m3
Relative Risk (Lower CI, Upper CI)
lag: Males
PMio-2.5: 0.93 (0.68,1.29)
0-1
PMio-2.5+NO2: 0.86 (0.62,1.20)
0-1
PMio-2.5+SO2: 0.90 (0.64,1.27)
0-1
PMio.2.5+03: 1.01 (0.67,1.51)
0-1
Females
PMio-2.5:1.20 (0.95,1.53)
0-1
PMio-2.5+NO2:1.19 (0.92,1.54)
0-1
PM10.2.5+SO2:1.31 (1.03,1.68)
0-1
PMio-2.5+0s: 1.47 (1.10,1.96)
0-1
Reference: Goss et al. (2004, 055624)
Period of Study: 1999-2000
Location: United States
Outcome: Mortality
Study Design: Cohort Study (Cystic
Fibrosis Cohort)
Statistical Analyses: Logistic
Regression
Age Groups: >6 yrs
Pollutant: PM2.6
Averaging Time: Annual avg
Mean (SD) unit: PM2.6: 13.7(4.2)
IQR: PM2 b: 11.8-15.9
Copollutant: O3
NO2
SO2
CO
Increment: 10 //g/m3
PM2.E: 1.32 (0.91 - 1.93)
Reference: (Lipfert et al., 2006,
189271)
Period of Study: 1989-1996
Location: Various parts of the Untied
States
Outcome: Mortality
Study Design: Retrospective Cohort
Statistical Analyses: Cox proportional
hazards regression
Age Groups: Male US veterans between
ages of 39 and 63 (Avg. age: 51)
Pollutant: PMio-2.5
Mean (SD): 16.0 (5.1
Increment: 12
1.07 (1.01,1.13)
Reference: McDonnell et al. (2000,
Outcome: Mortality
Pollutant: PMio-2.5
Increment: IQR
010319)


Study Design: Cohort (AHSM0G airport
Averaging Time: monthly averages
All Cause: 1.05 (0.92-1.20)
Period of Study: 1973-1977
cohort)


Mean (SD): PMio-2.5: 27.3 (8.6)
Resp: 1.19(0.88,1.62)
Location: California
Statistical Analyses: Cox regression



models
IQR: 9.7
Lung Cancer: 1.25 (0.63-2.49)

Age Groups: Males, 27 yrs +
Copollutant: O3: 0.70



SO2: 0.31



NO2: 0.23



SO4: 0.47

'All units expressed in //gfm3 unless otherwise specified.
July 2009
E-640
DRAFT-DO NOT CITE OR QUOTE

-------
Table E-33. Long-term exposure - mortality - PM2.5 (including PM components/sources).
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Abrahamowicz et al. (2003,
086292)
Period of Study: 1982-1989
Location: 151 Cities
Outcome: Mortality: All-causes
Study Design: Case-cohort study
Statistical Analyses: Cox proportion-
hazards model flexible regression spline
generalization
Age Groups: >18
Pollutant: PM2.6
Averaging Time: Annual
Mean (SD): 18.2
Range (Min, Max): (9.0, 33.5)
Copollutant: Sulfates
Relative Risk (Min CI, Max CI)
Estimated from graph (Figure 1): log
HR for a 24.5//g|m3 increase in PM2.6
over time
Years
0: 0.5(-1.1,1.6)
2: 0.6 (0.2, 0.9)
4: 0.6 (0.3, 0.8)
6: 0.8(0.3,1.1)
8:-1.0 (-1.5,1.0)
Reference: Abrahamowicz et al. (2003,
086292)
Period of Study: 1982-1989
Location: 151 Cities
Outcome: Mortality: All-causes
Study Design: Case-cohort study
Statistical Analyses: Cox proportion-
hazards model flexible regression spline
generalization
Age Groups: >18
Pollutant: Sulfates
Averaging Time: Annual
Mean (SD): 18.2
Range (Min, Max): (9.0, 33.5)
Copollutant: PM2.6
Relative Risk (Min CI, Max CI)
Estimated from graph (Figure 1): Log
HR for a 19.9//g|m3 increase in Sulfates
over time
Years
0: 0.1 (-0.2, 0.7)
2: 0.1 (-0.2, 0.4)
4: 0.0 (-0.4, 0.3)
6: 0.3 (-0.1, 0.5)
8: 0.4(-0.4,1.6)
Reference: Ballester et al. (2008,
Outcome: Mortality- All-causes
Pollutant: PM2.6
Potential Reduction in the total
189977)

burden of mortality (min CI, max CI)
Study Design: Health Impact
Averaging Time: Annual
for four different decreases in annual
Period of Study: 2001-2002
Assessment
Mean (SD): NR
PM2.5 using a conservative estimate
Location: Europe
Statistical Analyses: Aphesis Network
Range (Min, Max): NR
Reduction to 25/yg/m3 ¦ 0.4 (0.1, 0.8)

Age Groups: >30

Reduction to 20 //g|m3 ¦ 0.8 (0.2,1.6)



Reduction to 15/yg/m3 ¦ 1.6 (0.4, 3.1)



Reduction to 10 /yg/m3 ¦ 3.0 (0.8, 5.8)
July 2009
E-641
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Beelen et al. (2008,156263)
Period of Study: 1987-1996
Location: Netherlands
Outcome: Mortality: Total (non-
accidental) (< 800)
Cardiorespiratory (390-448,490-496,
487, 480-486, 507)
Pulmonary (460-519)
Cardiovascular (400-440)
Lung Cancer (162)
Other-causes
Study Design: Case-cohort study and
prospective cohort
Statistical Analyses: Cox proportion-
hazards model
Age Groups: 55-69
Pollutant: PM2.6
Averaging Time: Annual
Mean (SD): 28.3 (2.1)//g/m3
Range (Min, Max): (23.0, 36.8)
Copollutant (correlation): NO2: (>0.8)
BS: (>0.8)
SO2: (>0.6)
Increment: 11 /yglm3
Relative Risk (Min CI, Max CI)
RR for the association between
exposures to PM2.5 and cause specific
mortality
Natural Cause: Full cohort: 1.06 (0.97,
1.16)
Case cohort: 0.86 (0.66,1.13)
Cardiovascular: Full cohort: 1.04 (0.90,
1.21)
Case cohort: 0.83 (0.60,1.15)
Respiratory: Full cohort: 1.07 (0.75,
1.52)
Case cohort: 1.02 (0.56,1.88)
Lung Cancer: Full cohort: 1.06 (0.82,
1.38)
Case cohort: 0.87 (0.52,1.47)
Other cause: Full cohort: 1.08 (0.96,
1.23)
Case cohort: 0.85 (0.65,1.12)
RR for the association between
exposures to BS and cause specific
mortality
Natural Cause: Full cohort: 1.05 (1.00,
1.11)
Case cohort: 0.97 (0.83,1.13)
Cardiovascular: Full cohort: 1.04 (0.95,
1.13)
Case cohort: 0.98 (0.81,1.18)
Respiratory: Full cohort: 1.22 (0.99,
1.50)
Case cohort: 1.29 (0.91,1.83)
Lung Cancer: Full cohort: 1.03 (0.88,
1.20)
Case cohort: 1.03 (0.77,1.38)
Other cause: Full cohort: 1.04 (0.97,
1.12)
Case cohort: 0.91 (0.78,1.07)
July 2009
E-642
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Breitner et al. (2009,
188439)
Period of Study: 10/1 /1991 to
313112002
Location: Efurt, Germany
Covariates: seasonal and weekday
variations, influenza epidemics, air
temperature, relative humidity
Outcome: Mortality, excluding infants
and ICD-9 > 800
Study Design: Time-series
1/1991-8/31/1995): 50.6 ± 32.2 New City Limits
6-day IQR: 13.3
2(9/1/1995-2/28/19981:41.1 ± 28.4
Pollutant: PM2.6
Averaging Time: daily
Mean (SD) Unit:
Increment: IQR
Relative Risk (95% CI)
Lag
Statistical Analysis: semiparametric ^ I"1!;1
Poisson regression, polynomial distributed ^m
PDL: 1.009 (0.984-1.035)
Mean of lags 0-5: 1.004 (0.981-1.027)
Old City Limits
6-day IQR: 13.3
PDL: 1.017 (0.990-1.044)
Mean of lags 0-5: 1.010(0.986-1.035)
lag (PDL)
3(3/1/1998-3/31/20021:24.3 ± 15.4
Z^g/m3
Total: 38.0 ±28.3/yg/m3
Range (Min, Max): NR
Copollutant: NO2, CO, UFP
Statistical Package: R
Age Groups: All
New City Limits
15-day IQR: 11.5
PDL: 1.019(0.988-1.050)
Mean of lags 0-14: 1.017(0.992-1.042)
Old City Limits
15-day IQR: 11.5
PDL: 1.030 (0.997-1.063)
Mean of lags 0-14: 1.025(0.999-1.052)
Multiday Moving Averages, 6-day
Overall IQR: 13.3
Overall RR (95% CI): 1.004 (0.981-
1.027)
Period 1: NR
Period 2: 1.017 (0.990-1.044)
Period 3: 0.974 (0.937-1.013)
Multiday Moving Averages, 15-day
Overall IQR: 11.5
Overall RR (95% CI): 1.017 (0.992-
1.042)
Period 1: NR
Period 2: 1.016(0.988-1.045)
Period 3: 1.016(0.971-1.063)
July 2009
E-643
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Brunekreef et al. (2009,
191947)
Period of Study: 1987-1996
Location: The Netherlands
Outcome: All cause mortality (ICD-9
400440,460-519, > 800)
Study Design: Case-cohort
Covariates
Individual: sex, age, Quetelet index,
smoking status, passive smoking status,
educational level, occupation,
occupational exposure, marital status,
alcohol use, intake of vegetables, fruits,
energy, saturatured and monounsaturated
fatty acids, trans fatty acids, total fiver,
folic acid and fish
Area-level: Percent of population with
income below the 40th percentile and
above the 80th percentile
Statistical Analysis: Cox proportional
hazards
Statistical Package: Stata, SPSS, R
Age Groups: 120,000 adults aged 55-69
years at enrollment
Pollutant: PM2.6, estimated from PM10
levelsf
Averaging Time: 24hr
50"1 Percentile: 28 /yglm3
Range (Min, Max): 23-37
Copollutant (correlation):
NO2: 0.75
Black Smoke: 0.84
NO: 0.69
SO2: 0.43
Increment: 10//gfm3
Relative Risk (95 % CI) for PM2.5
concentrations and cause specific
mortality
Case Cohort
Natural Cause: 0.86 (0.66-1.13)
Cardiovascular: 0.83(0.60-1.15)
Respiratory: 1.02 (0.56-1.88)
Lung Cancer: 0.87 (0.52-1.47)
Noncardiopulmonary, non-lung cancer:
0.85(0.65-1.23)
Full Cohort
Natural Cause: 1.06 (0.97-1.16)
Cardiovascular: 1.04 (0.90-1.21)
Respiratory: 1.07 (0.75-1.52)
Lung Cancer: 1.06 (0.82-1.38)
Noncardiopulmonary, non-lung cancer:
1.08(0.72-1.19)
Relative Risk (95%CI) for PM2.6
concentrations and cause specific
mortality in full cohort analysis by
confounder model
Natural Cause Mortality
Unadjusted: 1.11 (1.04-1.20)
Smoking: 1.04 (0.96-1.13)
Smoking, area-level income: 1.06 (0.97-
1.16)
Cardiovascular Mortality
Unadjusted: 1.09 (0.97-1.23)
Smoking: 1.02 (0.90-1.16)
Smoking, area-level income: 1.04 (0.90-
1.21)
Respiratory Mortality
Unadjusted: 1.23 (0.92-1.65)
Smoking: 1.10 (0.81-1.50)
Smoking, area-level income: 1.07 (0.75-
1.52)
Lung Cancer Mortality
Unadjusted: 1.17 (0.95-1.46)
Smoking: 1.06 (0.85-1.33)
Smoking, area-level income: 1.06 (0.82-
1.38)
Noncardiopulmonary, Non-Lung Cancer
Mortality
Unadjusted: 1.10 (1.00-1.22)
Smoking: 1.05 (0.94-1.16)
Smoking, area-level income: 1.08 (0.96-
1.22)
July 2009
E-644
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Reference: Chen et al. (2005, 087942)
Outcome: Mortality: CHD
Pollutant: PM2.6
Increment: 10 //g/m3
Period of Study: 1973-1998
Study Design: Cohort
Averaging Time: 25 years
Relative Risk (Lower CI, Upper CI)
Location: San Francisco, San Diego, Los
Statistical Analyses: Cox proportion
Mean (SD): 29.0
lag: Males
Angeles, CA
hazards model



Range (Min, Max): NR
PM2.6: 0.89 (0.69,1.17)

Age Groups: >25



Copollutant: NO2, O3, SO2
0-1



PM2.B+NO2: 0.82 (0.61,1.10): 0-1



PM2.6+SO2: 0.86 (0.65,1.14)



0-1



PM2.B+03:0.92 (0.65,1.29)



0-1



Females



PM2 b: 1.19 (0.96,1.47)



0-1



PM2.B+NO2:1.18(0.95,1.47): 0-1



PM2.B+SO2:1.36 (1.05,1.74)



0-1



PM2.B+O3:1.61 (1.17,2.22)



0-1
Increment: 10 //g/m3
% Increase in Mortality for overall
exposure period and individual year
(95%CI Min, 95%CI Max):
ACS (adjusted for age, sex)
Overall: 10.8 (8.6,13.0)
2000:10.9(7.3,14.6)
2001:9.1 (5.3,12.7)
2002: 10.1 (6.0,14.3)
SCS (adjusted for age, sex)
Overall: 20.8 (14.8, 27.1)
2000: 17.8(9.8, 26.4)
2001: 16.5(7.4,25.0)
2002: 33.5(19.2, 49.3)
Reference: Eftim et al. (2008, 099104)
Period of Study: 2000-2002
Location: USA, Same cities as six cities
and ACS cohorts
Outcome (ICD-9): All non-accidental
causes (< 800)
Study Design: Cross-sectional
Statistical Analyses: Log-linear
regression, Poisson
Age Groups: >65
Pollutant: PM2.6
Averaging Time: Annual avg
Mean (SD): |
ACS: 13.6 (2.8)
SCS: 14.1 (3.1)
Range (Min, Max): ACS: (6.0, 25.1
SCS: (9.6,19.1)
July 2009
E-645
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Enstrom et al. (2005,
087356)
Period of Study: 1973-2002
Location: 25 California Colonies
11 California Colonies (EPA IPN study)
Outcome: Mortality: Cardiovascular-
respiratory (390-448)
(480-486, 487, 490-496, 507)
Study Design: Retrospective cohort
Statistical Analyses: Cox proportional
hazards regression model, SAS PHREG
Age Groups: 35 or older
Pollutant: PM2.6
Averaging Time: Annual
Mean (SD): 23.4
Range (Min, Max): (13.1 /yglm3, 36.1)
Relative Risk (Lower CI, Upper CI)
RR from causes for both sexes by
county from 1973-2002
Alameda: 0.962 (0.926,0.999)
Butte: 0.999 (0.910,1.096)
Contra Costa: 0.999 (0.943,1.058)
Fresno: 0.935 (0.872,1.002)
Humboldt: 0.992 (0.900,1.092)
Kern: 0.944(0.872,1.023)
Marin: 0.939 (0.867,1.016)
Napa: 0.949 (0.868,1.038)
Orange: 0.990 (0.948,1.034)
Riverside: 0.959 (0.906,1.015)
Sacramento: 0.998 (0.944,1.055)
San Bernardino: 0.992 (0.938,1.049)
San Diego: 0.992 (0.954,1.033)
San Francisco: 0.963 (0.914,1.014)
San Joaquin: 0.925 (0.816,1.049)
San Mateo: 0.949 (0.899,1.003)
Santa Barbara: 0.968 (0.878,1.068)
Santa Clara: 0.955 (0.910,1.003)
Santa Cruz: 0.890 (0.793,0.999)
Solano: 0.901 (0.815,0.995)
Sonoma: 0.968 (0.884,1.060)
Stanislaus: 0.984 (0.904,1.072)
Tulare: 1.047 (0.979,1.119)
Ventura: 0.967(0.872,1.072)
RR from all causes for 11 counties for
both sexes (EPA IPN study)
Santa Barbara: 0.968 (0.878,1.068)
Contra Costa: 0.999 (0.943,1.058)
Alameda: 0.962 (0.926,0.999)
Butte: 0.999 (0.910,1.096)
San Francisco: 0.963 (0.914,1.014)
Santa Clara: 0.955 (0.910,1.003)
Fresno: 0.935 (0.872,1.002)
San Diego: 0.992 (0.954,1.033)
Kern: 0.944(0.872,1.023)
Riverside: 0.959 (0.906,1.015)
July 2009
E-646
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Filleul et al. (2005, 087357) Outcome: Non-accidental causes
(< 800), cardiopulmonary disease (401-
Period of Study: 1974-1976
Location: 7 cities in France
440 and 460-519), lung cancer (162)
Age Groups: 25-59 years
Study Design: Cohort
N: 14,284 people
Statistical Analyses: Cox proportional
hazard, regression
Covariates: Sex, smoking habits,
educational level, body-mass index (BMI),
occupational exposure
Statistical Package: Proc
SAS
Pollutant: Total suspended particles
(TSP)
Averaging Time: NR
Mean (SD): NR
Range (Min, Max): (45, 243)
PM Component: NR
Monitoring Stations: 1 station
Copollutant (correlation): BS
r - 0.87
SO?
r - 0.17
NO
r - 0.84
NO?
r - 0.60
Increment: 10 //g/m3
Adjusted mortality rate ratios: 24 areas:
All non-accidental causes: 1.00(0.99,
1.01]
Lung cancer: 0.97(0.94,1.01]
Cardiopulmonary disease: 1.01(0.99,
1.03]
18 areas: All non-accidental causes:
1.05(1.02,1.08]
Lung cancer: 1.00(0.92,1.10]
Cardiopulmonary disease: 1.06(1.01,
1.12]
Reference: Fuentes et al. (2006,
097647)
Period of Study: June 2000
Location: Conterminous U.S.
Outcome: Mortality:
Study Design: Time-series
Statistical Analyses: Generalized
Poisson Regression
Age Groups: 0-14,15-64, >65
Covariates: temperature, pressure, dew
point, wind speed, elevation, age,
ethnicity
Pollutant: PM2.6
Averaging Time: monthly
Mean (SD): 6.60 (0.76)
Copollutant: PMio, 0:i
Increment: 10 //g/m3
PMib: 1.066 (1.064,1.069)
PMio: 1.030 (1.028,1.032)
July 2009
E-647
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Janes et al. (2007, 090927) Outcome: Mortality:
Period of Study: 2000 to 2002
Location: 113 US counties
Study Design: Time-series
Pollutant: PM2.6
Averaging Time: Annual Avg
Statistical Analyses: Cox proportional Mean (SD): NR
hazards model
Range (Min, Max): NR
Age Groups: 65-74
75-84
85 +
Increment: 1 //gf'nr1
% Increase (Lower CI, Upper CI)
lag: Overall % Increase by age-sex
stratum
Age Category
65-74: Male: 1.48 (0.93,2.03)
Female: 0.83(0.24,1.43)
75-84: Male: 0.85 (0.34,1.35)
Female: 0.77(0.28,1.27)
85 + : Male: 0.70 (0.03,1.38)
Female: 0.59(0.05,1.12)
National Trend % Increase by age-sex
stratum
Age Category
65-74: Male: 3.55 (2.77,4.34)
Female: 1.97(1.12,2.83)
75-84: Male: 2.48 (1.83,3.14)
Female: 2.29(1.66,2.93)
85 + : Male: 1.38 (0.52,2.26)
Female: 1.65(1.01,2.29)
Local Trend % Increase by age-sex
stratum
Age Category
65-74: Male: 0.04 (-0.58,0.67)
Female:-0.03 (-0.71,0.66)
75-84: Male:-0.34 (-0.87,0.19)
Female:-0.31 (-0.82, 0.21)
85 + : Male: <0.01 (-0.71,0.73)
Female:-0.22 (-0.74,0.31)
"Local trends are county specific
deviations from national trends
Reference: Jerrett et al. (2003, 087380) Outcome: Mortality
Period of Study: 1982
Location: 151 cities from ACS
Study Design: multilevel, individual-
ecologic analysis
Statistical Analysis: Cox proportional
hazards model
Covariates: Smoking, education,
occupational exposures, BMI, marital
status, alcohol consumption, gender
Pollutant: Sulfates
Mean (SD): 10.6
Range (Min, Max): 3.6,23.5
Increment: 19.9 (Range)
All Cause: SO4:1.17 (1.07, 1.27)
S04 + CO: 1.16 (1.10,1.23)
S04 + NO?: 1.16(1.08,1.24)
S04 + 0s: 1.17 (1.11,1.24)
SO4 + SO2: 1.05 (0.98,1.12)
CPD: SO4: 1.25 (1.16,1.35)
S04 +CO: 1.28 (1.18,1.39)
S04 + NO?: 1.29(1.17,1.42)
S04 + 0s: 1.27 (1.17,1.38)
SO4 + SO2: 1.13 (1.03,1.24)
Lung Cancer: SO4: 1.31 (1.09,1.58)
S04 +CO: 1.26 (1.03,1.53)
S04 + NO2: 1.31 (1.05,1.65)
SO4 + O3: 1.30 (1.07,1.59)
S04 + SO2: 1.37 (1.08,1.73)
July 2009
E-648
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Jerrett et al. (2005, 087600) Outcome: Mortality: Non- accidental
(< 800)
Period of Study: 1982-2000
Location: Los Angeles, California
IHD (410-414)
Cardiopulmonary (400-440,460-519)
Lung Cancer (162)
Other Cancers (140-149,160, 161,163-
239)
Other causes
Study Design: Retrospective Cohort
Statistical Analyses: Cox regression
hazards model
kriging, radial basis function multiquadric
interpolator
Age Groups: All ages
Pollutant: PM2.6
Averaging Time: Annual avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant: 0:i
Increment: 10/yg/m3
Relative Risk (Lower CI, Upper CI)
All Causes-PM2.b Only: 1.24 (1.11,1.37)
44 Ind. Covariates together+PIVh.B: 1.17
(1.03.1.32)
44 Ind. Covariates together+ PM2.6+O3:
1.20(1.07,1.34)
44 Ind. Covariates together+intersection
within freeways within 500 m +
PM2.B+O3:1.17 (1.05,1.31)
IHD ¦ PM2.B Only: 1.49 (1.20,1.85)
44 Ind. Covariates together+PM2.6: 1.39
(1.12,1.73)
44 Ind. Covariates together+PM2.B+O3:
1.45(1.15,1.82)
44 Ind. Covariates together+intersection
within freeways within 500 m +
PM2.B+O3:1.38 (1.11,1.72)
Cardiopulmonary - PM2.6 Only: 1.20
(1.04,1.39)
44 Ind. Covariates together+ PM2.6+O3:
1.19(1.02,1.38)
44 Ind. Covariates together+intersection
within freeways within 500 m +
PM2.B+O3:1.13 (0.97,1.31)
Lung Cancer - PM2.6 Only: 1.60
(1.09.2.33)
44 Ind. Covariates together+PM2.6: 1.44
(0.98,2.11)
44 Ind. Covariates together+intersection
within freeways within 500 m +
PM2.B+O3:1.46 (0.99,2.16)
Other Cancers ¦ PM2.6 Only: 1.09
(0.85,1.40)
44 Ind. Covariates together+ PM2.6+O3:
1.08(0.83,1.39)
44 Ind. Covariates together+intersection
within freeways within 500 m +
PM2.B+O3:1.08 (0.83,1.39)
All Other Causes ¦ PM2.6 Only: 1.11
(0.74,1.67)
44 Ind. Covariates together+ PM2.6+O3:
0.95(0.64,1.39)
44 Ind. Covariates together+intersection
within freeways within 500 m +
PM2.B+O3:1.02 (0.71,1.48)
July 2009
E-649
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Laden et al. (2006, 087605)
Period of Study: 1974-1998
Period 1: 1974-1989
Period 2: 1990-1998
Location: Nine US Cities
Watertown, MA
Kingston, TN
Harriman, TN
St. Louis, M0
Steubenville, OH
Portage, Wl
| Wyocena, Wl
Pardeeville, Wl
Topeka, KS
Outcome: Total mortality
Non-accidental (< 800)
Cardiovascular (400-440)
Respiratory (485-496)
Lung Cancer (162)
Other
Study Design: Prospective Cohort
Statistical Analyses: Cox proportional
hazards regression
Age Groups: 25-74
Pollutant: PM2.B
Averaging Time: Annual avg
Mean (SD): Period 1
Portage: 11.4
Topeka: 12.4
Watertown: 15.4
Harriman: 20.9
St Louis: 19.2
Steubenville: 29.0
Period 2
Portage: 10.2
Topeka: 13.1
Watertown: 12.1
Harriman: 18.1
St. Louis: 13.4
Steubenville: 22.0
Increment: 10 //g/m3
Relative Risk (Lower CI, Upper CI)
lag:
Period 1: Portage: 1.00
Topeka: 1.06 (0.86,1.31)
Watertown: 1.06 (0.87,1.28)
Harriman: 1.19(0.98,1.44)
St Louis: 1.15 (0.96,1.38)
Steubenville: 1.31 (1.10,1.57)
Period 2: Portage: NR
Topeka: 1.01 (0.83,1.22)
Watertown: 0.82 (0.67,1.00)
Harriman: 1.10 (0.91,1.33)
St Louis: 0.96 (0.80,1.15)
Steubenville: 1.06 (0.89,1.27)
Complete Period: Portage: 1.00
Topeka: 1.03 (0.89,1.19)
Watertown: 0.95 (0.83,1.08)
Harriman: 1.15(1.01,1.32)
St. Louis: 1.05 (0.93,1.20)
Steubenville: 1.18(1.04,1.34)
RR for complete follow up Avg. PM2.6
Total Mortality: 1.16 (1.07,1.26)
Cardiovascular: 1.28(1.13,1.44)
Respiratory: 1.08 (0.79,1.49)
Lung Cancer: 1.27 (0.96,1.69)
Other: 1.02 (0.90,1.17)
RR for period one Avg. PM2.6
Total Mortality: 1.18 (1.09,1.27)
Cardiovascular: 1.28 (1.14,1.43)
Respiratory: 1.21 (0.89,1.66)
Lung Cancer: 1.20 (0.91,1.58)
Other: 1.05 (0.93,1.19)
Decrease in Avg. PM2.6 over the two
periods
Total Mortality: 0.73 (0.57, 0.95)
Cardiovascular: 0.69 (0.46,1.01)
Respiratory: 0.43 (0.16,1.13)
Lung Cancer: 1.06 (0.43, 2.62)
Other: 0.85 (0.56,1.27)
Reference: Lipfert et al. (2006, 088756) Outcome: Mortality
Period of Study: 1989-1996	Study Design: Retrospective Cohort
Pollutant: Sulfate
Mean (SD) from 1976-81:10.7 (3.6)
Increment: 8
1.045 (0.944,1.157)
Location: Various parts of the Untied
States
Statistical Analyses: Cox proportional
hazards regression
Age Groups: Male US veterans between
ages of 39 and 63 (Avg. age: 51)
July 2009
E-650
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
Reference: Lipfert et al. (2006, 088756)
Outcome: Mortality
Pollutant: PM2.6
Increment: 8
Period of Study: 1989-1996
Study Design: Retrospective Cohort
Mean (SD): 14.3 (3.2)
1.118(1.038,1.203)
Location: Various parts of the Untied
Statistical Analyses: Cox proportional


States
hazards regression



Age Groups: Male US veterans between



ages of 39 and 63 (Avg age 51)


July 2009
E-651
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Lipfert et al. (2006, 088218) Outcome: Mortality: Non- accidental
(< 800)
Period of Study: 1997-2002
Location: Various parts of the Untied
States
Study Design: Retrospective cohort
Statistical Analyses: Cox proportional
hazards regression
AIC
Pollutant: PM2.6
Averaging Time: Annual avg
Mean (SD): 15.02 (4.80)//g/m3 (2000-
2003)
Range (Min, Max): (3.29, 24.96)
Copollutant (correlation): As:
Age Groups: Male US veterans between r - 0.443
ages of 39 and 63 (Avg. age: 51)
Cr: r - 0.379
Cu: r - 0.530
Fe: r - 0.379;
Pb: r - 0.489
Mn: r - 0.389;
Ni: r - 0.140
Se: r - 0.312;
V: r - 0.197
Zn: r - 0.420;
0C:r - 0.620
EC: r - 0.544; |
S(k r - 0.827
NOs: r - 0.649
NO?: r - 0.641
Peak CO: r - 0.040
Peak 0s: r - 0.222
Peak SO2: r - 0.714
Increment: 10 //g/m3
% Increase per 10//g/m3 increase in
PM2.6
Single-Pollutant Model
As:-5.23%
Cr: -2.11%
Cu: 2.12%
Fe: 2.81%
Pb: -2.40%
Mn: -1.20%
Ni: 3.75%
Se: -0.30%
V: 5.08%
Zn: 1.52%
0C:-0.02%
EC: 9.16%
SO4: 3.04%
NOs: 6.60%
NO?: 6.92%
Peak CO:-0.61%
Peak O3: 4.95%
Peak SO2: -4.20%
Multiple Pollutants model- Pollutant
with traffic density
NOs: 3.42%
SO4: -2.73%
EC: 6.27%
Ni: 2.51%
V: 3.27%
Pollutant with NO3
EC: 5.93%
Ni: 2.31%
V: 3.11%
Pollutant with Peak O3
Traffic density: 2.40%
EC: 10.79%
Fe: 5.94%
NOs: 7.57%
PM2.6: 8.97%
V: 4.93%
Ni: 3.65%
SO4: 6.75%
Cu: 1.55%
0C: 0.21%
July 2009
E-652
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Krewski et al. (2009,
191193)
Period of Study: 1979-2000
Location: 48 contiguous states US
Outcome: Death
Study Design: cohort
Covariates: demographic, socioeconomic
and ecologic characteristics
Statistical Analysis: Cox proportional-
hazards model
Statistical Package: NR
Age Groups: Adults of at least 30 years
Pollutant: PM2.6
Averaging Time: NR
Mean Unit:
1979-1983: 21.20//gfm3
1999-2000: 14.02 //g/m3
Range (Min, Max):
1979-1983: 10.77-30.01
1999-2000: 5.80-22.20
Copollutant: SO42, SO2, PMib, TSP, O3,
no2,co
Increment: 10//gfm3
Hazard Ratio (95% CI)
MSA&DIFF
Increment Change: 10.78 (1.043-1.115)
Change 5-15//gfm3: 1.128 (1.077-1.183)
Change 10-20/yglm3: 1.079 (1.048-
1.112)
HR (95% CI)
Los Angeles
Parsimonious ecologic covariates: 1.126
(1.014-1.251)
HR (95% CI)
15 year time window
Group A: 0.98 (0.92-1.06)
Group B: 1.01 (0.99-1.02)
HR (95% CI)
Third follow-up, 7 Ecologic Variables
1979-1983: 1.044(1.028-1.060)
1999-2000: 1.057 (1.036-1.079)
HR (95% CI)
Nationwide analysis, 1999-2000
Standard Cox: 1.03 (1.01-1.05)
Random Effects Cox: 1.06 (1.04-1.08)
Increment: 1.5 /yglm3
HR (95% CI)
28 County, 3 year model
All 7 ecologic covariates: 0.977 (0.932-
1.025)
July 2009
E-653
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Krewski et al. (2009,
191193)
Period of Study: 1979-2000
Location: 48 contiguous states US
Outcome: Death from cardiopulmonary
disease
Study Design: cohort
Covariates: demographic, socioeconomic
and ecologic characteristics
Statistical Analysis: Cox proportional-
hazards model
Statistical Package: NR
Age Groups: Adults of at least 30 years
Pollutant: PM2.6
Averaging Time: NR
Mean Unit:
1979-1983: 21.20//gfm3
1999-2000: 14.02 //g/m3
Range (Min, Max):
1979-1983: 10.77-30.01
1999-2000: 5.80-22.20
Copollutant: SO42, SO2, PMib, TSP, O3,
no2,co
Increment: 10//gfm3
Hazard Ratio (95% CI)
MSA&DIFF
Increment Change: 1.078 (1.077-1.182)
Change 5-15 //g/m3: 1.208 (1.132-1.290)
Change 10-20 //g/m3: 1.127 (1.081-
1.174)
HR (95% CI)
Los Angeles
Parsimonious ecologic covariates: 1.086
(0.939-1.285)
HR (95% CI)
15 year time window
Group A: 1.00 (0.90-1.11)
Group B: 1.05 (1.03-1.07)
HR (95% CI)
Third follow-up, 7 Ecologic Variables
1979-1983: 1.094 (1.070-1.118)
1999-2000: 1.138 (1.106-1.172)
HR (95% CI)
Nationwide analysis, 1999-2000
Standard Cox: 1.09 (1.06-1.12)
Random Effects Cox: 1.13 (1.10-1.16)
Increment: 1.5 /yglm3
HR (95% CI)
28 County, 3 year model
All 7 ecologic covariates: 0.940 (0.875-
1.011)
July 2009
E-654
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Krewski et al. (2009,
191193)
Period of Study: 1979-2000
Location: 48 contiguous states US
Outcome: Death from ischemic heart
disease
Study Design: cohort
Covariates: demographic, socioeconomic
and ecologic characteristics
Statistical Analysis: Cox proportional-
hazards model
Statistical Package: NR
Age Groups: Adults of at least 30 years
Pollutant: PM2.6
Averaging Time: NR
Mean Unit:
1979-1983: 21.20//gfm3
1999-2000: 14.02 //g/m3
Range (Min, Max):
1979-1983: 10.77-30.01
1999-2000: 5.80-22.20
Copollutant: SO42, SO2, PMib, TSP, O3,
no2,co
Increment: 10//gfm3
Hazard Ratio (95% CI)
MSA&DIFF
Increment Change: 1.196 (1.177-1.407)
Change 5-15 //g/m3: 1.484 (1.311-1.680)
Change 10-20//gfm3: 1.283 (1.186-
1.387)
HR (95% CI)
Los Angeles
Parsimonious ecologic covariates: 1.263
(10.22-1.563)
HR (95% CI)
Third follow-up, 7 Ecologic Variables
1979-1983: 1.184 (1.146-1.222)
1999-2000: 1.242 (1.191-1.295)
HR (95% CI)
Nationwide analysis, 1999-2000
Standard Cox: 1.15 (1.11-1.20)
Random Effects Cox: 1.24 (1.19-1.29)
Increment: 1.5 /yglm3
HR (95% CI)
28 County, 3 year model
All 7 ecologic covariates: 1.072 (0.980-
1.172)
July 2009
E-655
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Krewski et al. (2009,
191193)
Period of Study: 1979-2000
Location: 48 contiguous states US
Outcome: Death from lung cancer
Study Design: cohort
Pollutant: PM2.6
Averaging Time: NR
Covariates: demographic, socioeconomic Mean Unit:
and ecologic characteristics
Statistical Analysis: Cox proportional-
hazards model
1979-1983: 21.20//gfmJ
1999-2000: 14.02 //g/m3
Range (Min, Max):
Statistical Package: NR
Age Groups: Adults of at least 30 years 1979-1983: 10.77-30.01
1999-2000: 5.80-22.20
Copollutant: SO42, SO2, PMib, TSP, O3,
no2,co
Increment: 10//g/m3
Hazard Ratio (95% CI)
MSA&DIFF
Increment Change: 1.142 (1.057-1.234)
Change 5-15 //gfm3: 1.236 (1.114-1.372)
Change 10-20//gfm3: 1.143 (1.071-
1.221)
HR (95% CI)
Los Angeles
Parsimonious ecologic covariates: 1.311
(0.897-1.915)
HR (95% CI)
15 year time window
Group A: 1.08 (0.87-1.35)
Group B: 1.07 (1.02-1.13)
HR (95% CI)
Third follow-up, 7 Ecologic Variables
1979-1983: 1.092 (1.033-1.154)
1999-2000: 1.138 (1.057-1.225)
HR (95% CI)
Nationwide analysis, 1999-2000
Standard Cox: 1.11 (1.04-1.18)
Random Effects Cox: 1.14 (1.06-1.23)
Increment: 1.5 //g|m3
HR (95% CI)
28 County, 3 year model
All 7 ecologic covariates: 0.985 (0.832-
1.166)
Reference: Krewski et al. (2009,
Outcome: Death from diabetes
Pollutant: PM2.6
Increment: 1.5 //g/m3
191193)



Study Design: cohort
Averaging Time: NR
HR (95% CI)
Period of Study: 1979-2000



Covariates: demographic, socioeconomic
Mean Unit:
28 County, 3 year model
Location: 48 contiguous states US
and ecologic characteristics


1979-1983: 21.20//g/m3
All 7 ecologic covariates: 1.083 (0.723-

Statistical Analysis: Cox proportional-
1999-2000: 14.02 //g/m3
1.621)

hazards model


Statistical Package: NR
Range (Min, Max):


Age Groups: Adults of at least 30 years
1979-1983: 10.77-30.01



1999-2000: 5.80-22.20



Copollutant: SO42', SO2, PMib, TSP, O3,



N02.C0

July 2009
E-656
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Krewski et al. (2009,
Outcome: Death from endocrine disease
Pollutant: PM2.B
Increment: 1.5 //g/m3
191193)


Study Design: cohort
Averaging Time: NR
HR (95% CI)
Period of Study: 1979-2000



Covariates: demographic, socioeconomic
Mean Unit:
28 County, 3 year model
Location: 48 contiguous states US
and ecologic characteristics


1979-1983: 21.20//g/m3
All 7 ecologic covariates: 1.143 (0.835-

Statistical Analysis: Cox proportional-
1999-2000: 14.02 //g/m3
1.564)

hazards model


Statistical Package: NR
Range (Min, Max):


Age Groups: Adults of at least 30 years
1979-1983: 10.77-30.01



1999-2000: 5.80-22.20



Copollutant: SO42, SO2, PMib, TSP, O3,



N02,C0

Reference: Krewski et al. (2009,
Outcome: Death from digestive cancer
Pollutant: PM2.B
Increment: 10 //g/m3
191193)



Study Design: cohort
Averaging Time: NR
HR (95% CI)
Period of Study: 1979-2000



Covariates: demographic, socioeconomic
Mean Unit:
Los Angeles
Location: 48 contiguous states US
and ecologic characteristics


1979-1983: 21.20//g/m3
Parsimonious ecologic covariates: 1.199

Statistical Analysis: Cox proportional-
1999-2000: 14.02 //g/m3
(0.817-1.758)

hazards model


Statistical Package: NR
Range (Min, Max):


Age Groups: Adults of at least 30 years
1979-1983: 10.77-30.01



1999-2000: 5.80-22.20



Copollutant: SO42', SO2, PMib, TSP, O3,



N02.C0

Reference: Krewski et al. (2009,
Outcome: Death cancers other than lung
Pollutant: PM2.B
Increment: 10 //g/m3
191193)
and digestive

Averaging Time: NR
HR (95% CI)
Period of Study: 1979-2000
Study Design: cohort
Mean Unit:
Los Angeles
Location: 48 contiguous states US
Covariates: demographic, socioeconomic


and ecologic characteristics
1979-1983: 21.20//g/m3
Parsimonious ecologic covariates: 1.012


(0.788-1.299)

Statistical Analysis: Cox proportional-
1999-2000: 14.02 //g/m3

hazards model
Range (Min, Max):


Statistical Package: NR
1979-1983: 10.77-30.01


Age Groups: Adults of at least 30 years
1999-2000: 5.80-22.20



Copollutant: SO42', SO2, PMib, TSP, O3,



N02,C0

Reference: Krewski et al. (2009,
Outcome: Deaths from causes other than
Pollutant: PM2.B
Increment: 10 //g/m3
191193)
CPD, IHD and lung cancer

Averaging Time: NR
Hazard Ratio (95% CI)
Period of Study: 1979-2000
Study Design: cohort
Mean Unit:
MSA&DIFF
Location: 48 contiguous states US
Covariates: demographic, socioeconomic


and ecologic characteristics
1979-1983: 21.20//g/m3
Increment Change: 1.010 (0.968-1.055)

Statistical Analysis: Cox proportional-
1999-2000: 14.02//g/m3
Change 5-15//g/m3: 1.026 (0.970-1.085)

hazards model
Range (Min, Max):
Change 10-20//g/m3: 1.016 (0.981 -

Statistical Package: NR
1979-1983: 10.77-30.01
1.053)

Age Groups: Adults of at least 30 years
1999-2000: 5.80-22.20
HR (95% CI)


Copollutant: SO42', SO2, PMib, TSP, O3,
Third follow-up, 7 Ecologic Variables




N02,C0
1979-1983: 0.983 (0.960-1.007)
1999-2000: 0.953 (0.923-0.984)
July 2009
E-657
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: McDonnell et al. (2000,
Outcome: Mortality
Pollutant: PM2.6
Increment: IQR
010319)



Study Design: Cohort (AHSMOG airport
Averaging Time: monthly averages
All Cause: 1.22 (0.95-1.58)
Period of Study: 1973-1977
cohort)


Mean (SD): 31.9 (10.7)
Resp: 1.64 (0.93-2.90)
Location: California
Statistical Analyses: Cox regression



models
IQR: 24.3
Lung Cancer: 2.23 (0.56-8.94)

Age Groups: Males, 27 yrs +
Copollutants (correlation): O3: 0.68



SO2: 0.18



NO2: -0.08;



S(k 0.33

Reference: Miller et al. (2007, 090130)
Outcome: CVD Mortality
Pollutant: PM2.6
Increment: 10 //g/m3
Period of Study: 1994-1998
Study Design: Prospective Cohort (WHI)
Averaging Time: annual avg (2000)
CVD Death: 1.76(1.25, 2.47)
Location: 36 US Metropolitan Areas
Statistical Analyses: Cox proportional
Mean (SD): 13.4
CHD Death: 2.21 (1.17,4.16)

hazards regression



IQR: 11.6,18.3
CV Death: 1.83(1.11,3.00)

Age Groups: postmenopausal women
Range: 3.4, 28.3


ages 50-79

July 2009
E-658
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Naess et al. (2007, 090736) Outcome: Mortality: Non-accidental
(< 800)
Period of Study: 1992-1998
Location: Oslo, Norway
Lung cancer (162)
C0PD (490-496)
Cardiovascular (390-459)
Study Design: Prospective Cohort
Statistical Analyses: Cox proportional
hazards regression model
Age Groups: 51-70, 71-90
Pollutant: PM2.B
Averaging Time: 4 year avg
Mean (SD): PMz.b: 15
Range (Min, Max): PM2 e: (7, 22)
Copollutant (correlation): NO2:
r - 0.95
Relative Risk (CI min, CI max)
RR for deaths from all causes
Men (ages 51-70) PM2.6 exposure
(in/yg/m3)
6.56-11.45: 1.00
11.46-14.25:0.96 (0.89,1.04)
14.26-18.43:1.12(1.03,1.22)
18.44-22.34:1.48 (1.36,1.60)
Men (ages 71-90) PM2.6 exposure
(in/yg/m3)
6.56-11.45: 1.00
11.46-14.25:0.99 (0.93,1.06)
14.26-18.43:1.10(1.03,1.17)
18.44-22.34:1.19(1.12,1.27)
Women (ages 51-70) PM2.6 exposure
(in/yg/m3)
6.56-11.45: 1.00
11.46-14.25:0.96 (0.87,1.07)
14.26-18.43:1.08 (0.98,1.20)
18.44-22.34:1.44(1.30,1.59)
Women (ages 71-90) PM2.6 exposure
(in/yg/m3)
6.56-11.45: 1.00
11.46-14.25:1.03 (0.97,1.09)
14.26-18.43:1.07 (1.01,1.12)
18.44-22.34:1.11 (1.05,1.16)
Increment: 10/yg/m3
RR for death from CVD and lung cancer
Men (ages 51-70)
CVD-PM2.6:1.11 (1.06,1.16)
C0PD- PM2.6: 1.32 (1.17,1.49)
Lung Cancer- PM2.6: 1.07 (0.98,1.17)
Women (ages 51-70)
CVD: PMib: 1.16 (1.09,1.24)
C0PD: PM2 b: 1.18 (1.03,1.34)
Lung Cancer: PM2.6: 1.23 (1.10,1.37)
Men (ages 71-90)
CVD: PM2.B: 1.06 (1.03,1.09)
C0PD: PM10: 1.13 (1.04,1.24)
PMib: 1.14(1.04,1.24)
Lung Cancer: PM2.6: 1.08 (0.98,1.19)
Women (ages 71-90)
CVD: PM2.B: 1.02 (1.00,1.05)
C0PD: PM2.6: 1.09 (1.00,1.18)
Lung Cancer: PM2.B: 1.16 (1.03,1.31)
Reference: Naess et al. (2007, 090736) Outcome: Mortality: Lung cancer (162) Pollutant: PM2 e
Period of Study: 1992-1998	C0PD (490-496)	Averaging Time: (Month-year) avg
Relative Risk (CI min, CI max)
RR on All-cause mortality of PM2.5 in
July 2009
E-659
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Location: Oslo, Norway
Cardiovascular (390-459)
Psychiatric causes (290, 292-302, 304,
306-319)
Stomach cancer (151)
Violence (800-999)
Study Design: Multilevel cohort
Statistical Analyses: WinBUGS
Age Groups: 50-74
Range Mean (SD): 14.2 (3.6)
IQ Range (1st, 4th): (6.6, 22.3)
Copollutant (correlation): PMio:
r - 0.951
NO?: r - 0.87
Men Age 50-74
Primary Education: PM2.6: 1.06 (1.00,
1.11)
Individual: 1.34 (1.24,1.43)
Neighborhood: 1.22 (1.16,1.28)
Manual Class: PM2 B: 1.06 (1.01,1.12)
Individual: 1.28(1.20,1.37)
Neighborhood: 1.20(1.14,1.26)
Income below median: PM2.6: 1.05 (1.00,
1.12)
Individual: 1.44(1.35,1.53)
Neighborhood: 1.16 (1.11,1.21)
Not owner occupied: PM2.6: 1.06(1.00,
1.13)
Individual: 1.24 (1.12,1.36)
Neighborhood: 1.11 (1.05,1.17)
Lives in flat dwelling: PM2.6: 1.04 (0.98,
1.11)
Individual: 1.19(1.09,1.31)
Neighborhood: 1.10(1.04,1.17)
More than one person per room in
dwelling: PMib: 1.10 (1.02,1.18)
Individual: 1.05(0.98,1.13)
Neighborhood: 1.01 (0.96,1.05)
RR on All-cause mortality of PM2.5 in
Women Age 50-74
Primary Education Only: PM2.6: 1.05
(1.00,1.11)
Individual: 1.32(1.23,1.42)
Neighborhood: 1.18(1.12,1.24)
Manual Class: PM2 B: 1.07 (1.01,1.13)
Individual: 1.27 (1.18,1.36)
Neighborhood: 1.18(1.12,1.24)
Income below median: PM2.6: 1.05 (1.01,
1.10)
Individual: 1.52(1.41,1.63)
Neighborhood: 1.13 (1.09,1.18)
Not owner occupied: PM2.6: 1.07 (1.01,
1.14)
Individual: 1.24 (1.12,1.38)
Neighborhood: 1.08 (1.02,1.14)
Lives in a flat dwelling: PM2.6: 1.05 (0.99,
1.11)
Individual: 1.21 (1.09,1.34)
Neighborhood: 1.09 (1.02,1.15)
More than one person per room in
dwelling: PM2.1,: 1.11 (1.04,1.19)
Individual: 1.07 (0.99,1.14)
Neighborhood: 1.01 (0.96,1.05)
JlRJoMnter2uartNeJncreaseJMnjn__
July 2009
E-660
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI) ~
PM2.5 for different causes of death
CVD: Age and sex adjusted: 1.11 (1.07,
1.15)
Primary education only: M1 + Individual:
1.07 (1.04,1.11)
M1 +Neighborhood: 1.03 (1.00,1.07)
Manual Class: M1 + Individual: 1.08
(1.04,1.11)
M1 +Neighborhood: 1.06 (1.02,1.10)
Income below Median: M1 + Individual:
1.07 (1.03,1.11)
M1 +Neighborhood: 1.02 (0.98,1.05)
Not owner occupied: M1 + Individual:
1.05(1.01,1.09)
M1 +Neighborhood: 1.03 (0.99,1.07):
Living in a Flat dwelling
M1 +Individual: 1.04 (1.00,1.08)
M1 +Neighborhood: 1.01 (0.97,1.05)
Crowded household: M1 + Individual:
1.10(1.05,1.14)
MUNeighborhood: 1.10 (1.06,1.15)
Pulmonary Cancer: Age and sex adjusted:
1.12(1.05,1.19)
Primary education only: M1 + Individual:
1.09(1.01,1.17)
MUNeighborhood: 1.05 (0.98,1.13)
Manual Class: M1 + Individual: 1.09
(1.01,1.17)
MUNeighborhood: 1.10 (1.06,1.13)
Income below Median: M1 + Individual:
1.09(1.01,1.17)
M1 +Neighborhood: 1.02 (0.95,1.10)
Not owner occupied: M1 + Individual:
1.07 (1.00,1.15)
M1 +Neighborhood: 1.04 (0.97,1.12)
Living in a Flat dwelling: M1 + Individual:
1.03(0.96,1.11)
M1 +Neighborhood: 1.00 (0.92,1.08)
Crowded household: M1 + Individual:
1.10(1.03,1.14)
MUNeighborhood: 1.11 (1.04,1.20)
C0PD: Age and sex adjusted: 1.17 (1.09,
1.25)
Primary education only: M1 + Individual:
1.13(1.05,1.22)
M1 +Neighborhood: 1.09 (1.01,1.19)
Manual Class: M1 + Individual: 1.14
(1.05,1.23)
MUNeighborhood: 1.12 (1.04,1.22)
Income below Median: M1 + Individual:
1.13(1.04,1.22)
M1 +Neighborhood: 1.06 (0.97,1.15)
July 2009
E-661
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
1.10(1.02,1.19)
M1 + Neighborhood: 1.07 (0.99,1.16)
Living in a Flat dwelling: M1 + Individual:
1.08(1.00,1.18)
M1 + Neighborhood: 1.03 (0.95,1.13)
Crowded household: M1 + Individual:
1.16(1.07,1.26)
MUNeighborhood: 1.16 (1.07,1.26)
Estimates for psychiatric diseases,
genetic cancer and violent death
Reference: Nerriere et al. (2005,
088630)
Period of Study: Grenoble (2001)
Paris (2002)
Rouen (2002-2003)
Strasbourg (2003)
Location: Four French Cities- Grenoble,
Rouen, Paris, and Strasbourg
Outcome: Mortality: Lung Cancer (162)
Study Design: Time-series
Statistical Analyses: GIS
Pollutant: PM2.6
Averaging Time: 48-h avg
Mean Range:
Age Groups: 30-71 year old nonsmoking 17 to 49//gfm3
adults
Increment: 10 //g/m3
% Increase (Lower CI, Upper CI)
% increase in lung cancer deaths
attributable to PM2.6 exposure
France: 8 (1,16)
Grenoble: 10 (3,19)
Rouen: 10 (2,19)
Strasbourg: 24 (4, 40)
Reference: Ozkaynak and Thurston
(1987,072960)
Period of Study: 1980
Location: U.S.
Outcome: Total Mortality
Study Design: Cross-sectional
Statistical Analyses: Multiple
regression analysis
Pollutant: Sulfate
Averaging Time: Annual avg
Mean Range: Sulfate: 11.1 (3.5)
Range of estimated total mortality
effects of air pollutions: Sulfate: 4-9%
"Sulfate concentration was consistently
found to be a significant predictor of
mortality in the models considered. Fine
particle mass coefficients were also often
found to be statistically significant in the
mortality regressions."
July 2009
E-662
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Pope et al. (2004, 055880)
Period of Study: 1982-2000
Location: Metropolitan areas in all 50
states in the US
Outcome: Mortality: Cardiovascular
Diseases (390-459)
Diabetes (250)
Respiratory Disease (460-519)
Study Design: Prospective Cohort
Statistical Analyses: Cox proportional
hazards regression
Age Groups: >30
Pollutant: PM2.6
Averaging Time: Annual avg
Mean (SD): 17.1 (3.7)
Range (Min, Max): NR
Increment: 10 //g/m3
Relative Risk (Lower CI, Upper CI)
All cardiovascular disease plus diabetes:
PMz.b: 1.12 (1.08,1.15)
Former Smoker: 1.26 (1.23,1.28)
Current Smoker: 1.94 (1.90,1.99)
Ischemic Heart Disease: PM2.6: 1.18
(1.14,1.23)
Former Smoker: 1.33 (1.29,1.37)
Current Smoker: 2.03 (1.96, 2.10)
Diabetes: PM2.6: 0.99 (0.86,1.14)
Former Smoker: 1.05 (0.94,1.16)
Current Smoker: 1.35 (1.20,1.53)
All other Cardiovascular Diseases: PM2.6:
0.84 (0.71,0.99)
Former Smoker: 1.22 (1.09,1.38)
Current Smoker: 1.78 (1.56, 2.04)
Diseases of the respiratory system:
PM2.6: 0.92 (0.86, 0.98)
Former Smoker: 2.16 (2.04, 2.28)
Current Smoker: 3.88 (3.66, 4.11)
C0PD: PM2.6: 0.84 (0.77, 0.93)
Former Smoker: 4.93 (4.48, 5.42)
Current Smoker: 9.85 (8.95,10.84)
All other respiratory diseases: PM2.6:
0.86(0.73,1.02)
Former Smoker: 1.54 (1.36,1.74)
Current Smoker: 1.83 (1.57, 2.12)
Reference: Pope et al. (2007, 091256)
Period of Study: 1960-1975
Location: New Mexico, Arizona, Utah,
and Nevada
Outcome (ICD7&8):
Mortality: Cardiovascular (ICD 7: 400-
468, 331, 332 ICD 8: 390-458)
Respiratory (ICD 7: 470-527 ICD 8: 460-
519)
Influenza/ pneumonia (ICD 7: 480-483,
490-493, ICD 8: 470-474, 480-486)
Study Design: Retrospective Cohort
Statistical Analyses: Poisson regression
model
GAM
SAS
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
The study does not present quantitative
results
results are presented in figures. The
References found that the strike-related
estimated percent decrease in mortality
was 2.5% (1.1-4.0),
: Groups: All smelter workers > 18
July 2009
E-663
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Pope et al. (2009,190107)
Period of Study: 1978-1982,1997-
2001
Location: 211 US counties and 51
metropolitan areas
Outcome: Increased life expectancy
Study Design: Cross-sectional
Statistical Analysis: Cross-sectional
regression
Age Groups: Adults £45 years
Pollutant: PM2.6
Averaging Time: daily, quarterly and
annual
Mean (SD) Unit:
1979-1983: 20.61 ± 4.36//gfm3
1999-2000: 14.10 ± 2.86//gfm3
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10 //g/m3
Regression Coefficient ± SD
211 County Units
Intercept: 1.75 ± 0.27
Reduction in PM2.6: 0.61 ± 0.20
Change in Income: 0.13 ± 0.01
Change in Population: 0.06 ± 0.02
Change in Black Population: -2.70 ± 0.64
Change in Lung Cancer Mortality Rate: -
0.06 ± 0.02
Change in C0PD Mortality Rate: -0.08 ±
0.02
R: 0.53
51 Metropolitan Areas
Intercept: 2.09 ± 0.36
Reduction in PM2.6: 0.95 ± 0.23
Change in Income: 0.11 ± 0.02
Change in Population: 0.05 ± 0.02
Change in Black Population: -5.98 ± 1.99
Change in Lung Cancer Mortality Rate:
0.02 ± 0.03
Change in C0PD Mortality Rate: -0.19 ±
0.05
R: 0.74
Reference: Rainham et al. (2005,
088676)
Period of Study: 1981-1999
Location: Toronto, Canada
Outcome: Total deaths (ICD9 < 800),
cardiorespiratory (390-459), non-
cardiorespiratory (ICD9-NR)
Study Design: Time-series
Statistical Analyses: Generalized linear
models were used
Season: Winter (December-February)
Summer (June-August)
Statistical Package: S-Plus 6.1
Pollutant: PM2.6
Averaging Time: NR
Mean (SD): All years: 17.0 (8.7)//g|m3
Winters: 17.2 (6.8)
Summers: 18.8 (10.2)
Avg Winter values: Dry Moderate: 17.0
(1.0)
Dry Polar: 17.5(0.5)
Dry Tropical: No Comparison
Moist Moderate: 17.1 (0.8)
Moist Polar: 17.5 (0.6)
Moist Tropical: 16.5 (3.6)
Transition: 16.7 (1.0)
Avg summer values: Dry Moderate: 18.4
(0.9)
Dry Polar: 19.0(1.2)
Dry Tropical: 18.5 (2.4)
Moist Moderate: 19.2 (1.2)
Moist Polar: 17.5 (2.0)
Moist Tropical: 19.8 (1.1)
Transition: 17.6 (1.5)
Mortality risk for winter season and
within winter synoptic weather
categories
RR Estimate [Lower CI, Upper CI]:
Winter: Total: 0.998(0.997,1.000]
Cardioresp: 0.998(0.996,1.000]
Other: 0.998 [0.996,1.000]
Dry Moderate: Total: 1.001(0.996,
1.007]
Cardioresp: 1.005(0.998,1.011]
Other: 1.002 [0.998,1.005]
Dry Polar: Total: 0.998(0.995,1.001]
Cardioresp: 0.995(0.991, 0.999]
Other: 1.002 (0.998,1.005]
Dry Tropical: NA
Moist Moderate: Total: 0.998(0.993,
1.002]
Cardioresp: 1.003(0.995,1.010]
Other: 0.997 [0.991,1.004]
Moist Polar: Total: 1.001(0.998,1.005]
Cardioresp: 1.002(0.997,1.007]
Other: 1.003 (0.999,1.007]
Moist Tropical: Total: 1.007(0.965,
1.203]
Cardioresp: 1.123(1.031,1.224]
July 2009
E-664
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Other: 1.248 [1.123,1.387]
Transition Total: 1.003[0.996,1.009]
Cardioresp: 0.996(0.987,1.004]
Other: 0.997 [0.990,1.004]
Mortality risk for summer season and
within summer synoptic weather
categories
RR Estimate [Lower CI, Upper CI]:
Summer: Total: 1.000(1.000,1.001]
Cardioresp: 1.001(1.000,1.002]
Other: 1.001(1.000,1.002]
Dry Moderate: Total: 1.001(0.999,
1.002]
Cardioresp: 1.002(0.999,1.004]
Other: 0.999(0.997,1.002]
Dry Polar: Total: 1.002(0.999,1.005]
Cardioresp: 0.996(0.991,1.000]
Other: 1.003[ 0.999,1.007]
Dry Tropical: Total: 1.016(1.006,1.027]
Cardioresp: 1.017(1.005,1.030]
Other: 1.017 (1.003,1.031]
Moist Moderate: Total: 1.002(1.000,
1.004]
Cardioresp: 1.003(0.999,1.006]
Other: 1.004(1.001,1.006]
Moist Polar: Total: 1.005(0.998,1.011]
Cardioresp: 1.008(0.997,1.018]
Other: 1.003 (0.995,1.011]
Moist Tropical: Total: 0.999(0.997,
1.001]
Cardioresp: 0.996(0.993,1.000]
Other: 0.998 (0.995,1.001]
Transition: Total: 1.005(0.996,1.014]
Cardioresp: 1.007(0.994,1.020]
Other: 1.002 (0.996,1.008]
Reference: Roman et al. (2008,156921) Outcome: Mortality
Period of Study: 2006	Study Design: Expert Judgment Study
Location: U.S.
Statistical Analyses: Standard best
practices for expert elicitation
Pollutant: PM2.6
Averaging Time: annual average
Mean (SD): 4-30
Quantitative results are not presented in
the text, but can be found graphically in
Figure 3.
"Most of the experts' central estimates
fall at or above the 2002 ACS median
(0.6% per//g/m3) and below the original
Six Cities median (1.2% per/yg/m3)."
July 2009
E-665
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Schwartz, et al (2008,
156921)
Period of Study: 1979-1988
Location: Six U.S. metropolitan areas:
Boston, Massachusetts
Outcome: Mortality
Pollutant: PM2.6
PM Increment: 10//g/m3
Study Design: Poisson regression with
GAM
Statistical Analyses: Weighted linear
regression
Averaging Time: daily
Mean (SD): Boston-16.5
Knoxville-21.1
The difference between mean PM2.6
concentrations of 10 |Jg/m3 and 20
|jg/m3 is associated with about a 1.5%
increase in deaths.
Knoxville, Tennessee
Season: all
St. Louis-19.2

St. Louis, Missouri
Dose-response Investigated? No
Steubenville-30.5

Steubenville, Ohio
Statistical Package: S-plus
Madison-11.3

Madison, Wisconsin

T opeka-12.2

and Topeka, Kansas

SD not reported
Range (Min, Max): (0,35)
Monitoring Stations: 6

Reference: (Schwartz et al„ 2008,
156963)
Period of Study: 1979-1998
Location: Watertown, MA
Kingston and Harriman, TN
St Louis, MO
Steubenville, OH
Portage, Wyocena
Pardeeville Wl
Topeka, KS
Outcome: Mortality: Non-accidental
(< 800)
Study Design: Cross-sectional
Statistical Analyses: Cox proportional
hazards regression
penalized splines
Bayesian Model Averaging
Age Groups: >18
Pollutant: PM2.6
Averaging Time: Annual avg
Mean (SD): 17.5 (6.8)
Range (Min, Max): (8, 40)
Increment: 10 //g/m3
Relative Risk (Lower CI, Upper CI)
Estimated from Figure 4: All Cause
Mortality ¦ Year before Death
0: 1.10(1.00,1.21)
1: 1.03(0.98,1.08)
2: 1.01 (1.00,1.02)
3: 1.00(0.99,1.01)
4: 1.00(0.99,1.01)
5: 1.00
Lung Cancer Mortality ¦ Year Before
Death
Estimated from Figure 5
0: 1.18(1.00,1.48)
1: 1.12(0.98,1.33)
2: 1.08(0.92,1.22)
3: 1.02(1.01,1.03)
4: 1.01 (1.00,1.02)
5: 1.01
RR per 10 //g/m3 increase of PM2.6
exposure
Level Of Increase
Estimated from Figure 3
10 //g/m3: 1.15
20 //g/m3: 1.29
30//g/m3: 1.46
40 //g/m3: 1.64
July 2009
E-666
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Reference: Tainio et al. (2005, 087444) Outcome (ICD10): Mortality:
Cardiopulmonary (111-170 and J15-J47)
Period of Study: 1997-Present
Location: Helsinki, Finland
Lung Cancer (C34)
Other causes
Study Design: Time-series simulation
Statistical Analyses: Monte Carlo
Simulation
Age Groups: All ages
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD): 10.7
Range (Min, Max): NR
Estimated Deaths Per Year (Min CI,
Max CI) Associated with Primary
PM2.5 Emissions from buses in
Helsinki in 2020 for different bus
strategies
Cardiopulmonary Mortality
Current Fleet: 15.9 (0, 46.6)
Modern Diesel: 7.9 (0, 23.0)
Diesel with particle trap: 3.9 (0,12)
Natural gas bus: 2.3 (0, 6.8)
Lung Cancer Mortality
Current Fleet: 2.2 (0, 6.1)
Modern Diesel: 1.1 (0, 3.0)
Diesel with particle trap: 0.6 (0,1.6)
Natural gas bus: 0.3 (0, 0.9)
Total Mortality
Current Fleet: 18.1 (0, 55.0)
Modern Diesel: 9.0 (0, 27.0)
Diesel with particle trap: 4.4 (0,14.1)
Natural Gas Bus: 2.6 (0, 8.0)
Reference: Villeneuve et al. (2002,
042576)
Period of Study: 1974-1991
Location: Six US Cities: Steubenville,
OH, St. Louis, M0, Portage, Wl, Topeka,
KS, Watertown, MA, Kingston/ Harriman,
TN
Outcome (ICD 10): Mortality: Non-
accidental (< 800)
Study Design: Prospective Cohort
Statistical Analyses: Poisson, EPICURE
Age Groups: All ages
<60
>60
Pollutant: PM2.6
Averaging Time: 24-h avg
Mean (SD): Portage: 10.9 (7.2)
Topeka: 12.1 (7.1)
Harriman: 20.7 (9.4)
Watertown: 14.9 (8.4)
St. Louis: 18.7 (10.6)
Steubenville: 28.6 (21.0)
Overall: 18.6
Range (Min, Max): NR
Increment: 18.6/yg/m3
Relative Risk (Min CI, Max CI)
RR of all cause mortality for exposure of
PM2.E by age group
Exposure to PM2.6 remained fixed over
entire study period
<60:1.89 (1.32, 2.69)
>60:1.21 (1.02,1.43)
Total: 1.31 (1.12,1.52)
Exposure to PM2.6 was defined according
to 13 calendar periods* (no smoothing)
<60:1.52 (1.15, 2.00)
>60:1.11 (0.95,1.29)
Total: 1.19(1.04,1.36)
Exposure to PM2.6 was defined according
to 13 calendar periods* (smoothed)
<60:1.43 (1.10,1.85)
>60:1.09 (0.93,1.26)
Total: 1.16(1.02,1.32)
Time dependent estimate of PM2.6
received during the previous two years
<60:1.42 (1.09,1.82)
>60:1.08 (0.94,1.25)
Total: 1.16(1.02,1.31)
Time dependent estimate of PM2.6
received 3-5 years before current year
<60:1.35 (1.08,1.67)
>60:1.08 (0.95,1.22)
Total: 1.14(1.02,1.27)
July 2009
E-667
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Time dependent estimate of PM2.6
received >5 years before current year
<60:1.34 (1.11,1.59)
>60:1.09 (0.99,1.20)
Total: 1.14(1.05,1.23)
* The calendar periods used were: 1970-
1978,1979,1980,1981,1982,1983,
1984,1985,1986,1987,1988,1989,
and 1990+.
RR of all cause mortality and PM2.6
exposure by city
Portage: 1.16 (0.96,1.39)
Topeka: 1.06 (0.89,1.27)
Harriman
Men: 1.04 (0.79,1.36)
Women: 0.96 (0.69,1.31)
All: 1.13 (0.95,1.35)
Watertown
Men: 1.20(0.95,1.51)
Women: 1.06 (0.78,1.43)
All: 1.32 (1.11,1.51)
St. Louis
Men: 0.97(0.76,1.24)
Women: 1.13 (0.86,1.49)
Steubenville
Men: 1.39(1.11,1.74)
Women: 1.22 (0.93,1.61)
Reference: Willis et al. (2003, 089922)
Outcome: Mortality: All causes
Pollutant: Sulfates
All Cause, Metropolitan Scale: 1.25



(1.13,1.37)
Period of Study: 1982-1989
Lung Cancer (162)
Averaging Time: Annual avg



All Cause, County Scale: 1.50 (1.30,
Location: US Metropolitan areas in all 50
Cardiopulmonary (401-440,460-519)
Mean (SD): 10.6/yg/m3
1.73)
states



Study Design: Prospective Cohort
Range (Min, Max): 3.6, 23.5
CPD, Metropolitan Scale: 1.29 (1.15,

Statistical Analyses: Cox proportional
Copollutant: CO, I\I02,03,S02
1.46)

hazards model

CPD, County Scale: 1.75 (1.48, 2.08)

Age Groups: All ages


Reference: Zanobetti and Schwartz
Outcome: Mortality, all causes, excluding
Pollutant: PM2.6
Increment: 10//gfm3
(2009,188462)
ICD codes S00-U99



Averaging Time: 24h
Percent Increase (95% CI) in mortality
Period of Study: 1999-2005
Study Design: Time-series
Mean (SD) Unit://g/m3
by increment of PM2.5, combined by
season
Location: 112 US Cities
Covariates: Region, season



Birmingham AL ¦ 16.5
All Cause Mortality

Statistical Analysis: Poisson regression
Phoenix AZ - 11.4

Overall: 0.98 (0.75-1.22)

Age Groups: All


LittleRock AR ¦ 14.3
Winter: 0.56(0.17-0.94)


Fresno CA ¦ 19.4
Spring: 2.57 (1.96-3.19)


Bakersfield CA ¦ 21.7
Summer: 0.25 (-0.13-0.63)


Los Angeles CA ¦ 19.9
Autumn: 0.95 (0.56-1.34)


Anaheim CA ¦ 16.3
CVD


Rubidoux CA ¦ 24.9
Overall: 0.85 (0.46-1.24)


Sacramento CA ¦ 13.0
Winter: 0.70(0.04-1.36)


El Cajon CA ¦ 13.5
Spring: 2.18 (1.22-3.15)
July 2009
E-668
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Denver CO ¦ 10.3
Hartford CT ¦ 11.6
New Haven CT ¦ 13.7
Wilmington DE ¦ 15.1
Davie FL ¦ 8.4
Miami FL ¦ 9.4
Jacksonville FL ¦ 10.6
Pensacola FL ¦ 12.4
Tampa FL- 11.9
Orlando FL -10.3
Palm beach FL ¦ 7.9
Pinellas FL ¦ 10.4
Atlanta GA ¦ 17.6
Chicago IL ¦ 15.9
Gary IN - 15.3
Indianapolis IN -16.3
Cedar Rapids IA - 11.0
Des Moines IA - 10.5
Davenport IA ¦ 12.3
Louisville KY ¦ 15.9
Baton Rouge LA ¦ 13.4
Avondale LA ¦ 12.3
New Orleans LA ¦ 12.6
Baltimore MD ¦ 15.6
Springfield MA ¦ 12.3
Boston MA ¦ 12.4
Worcester MA ¦ 11.3
Holland Ml - 12.1
Grand Rapids Ml -13.6
Detroit Ml ¦ 16.2
Minneapolis MN - 11.1
Kansas MO ¦ 12.0
St Louis MO ¦ 14.5
Omaha NE ¦ 10.4
Elizabeth NJ -14.7
Albuquerque NM - 6.7
New York NY - 14.8
Bath NY - 9.6
Durham NC ¦ 14.3
Winston NC -14.7
Greensborough NC ¦ 14.2
Charlotte NC ¦ 15.3
Raleigh NC - 14.3
Middletown OH ¦ 16.4
Youngstown OH ¦ 15.6
Summer: -0.03 (-0.75-0.69)
Autumn: 0.92 (0.17-1.68)
Ml
Overall: 1.18 (0.48-1.89)
Winter: 1.29 (-0.14-2.75)
Spring: 2.12 (0.53-3.74)
Summer: -0.03 (-1.46-1.42)
Autumn: 1.24 (0.12-2.36)
Stroke
Overall: 1.78 (0.96-2.62)
Winter: 1.93(0.34-3.54)
Spring: 2.04 (-0.02-4.13)
Summer: 1.64 (0.05-3.26)
Autumn: 1.69 (0.06-3.35)
Respiratory
Overall: 1.68 (1.04-2.33)
Winter: 0.86 (-0.16-1.88)
Spring: 4.62 (3.08-6.18)
Summer: 0.78 (-0.49-2.06)
Autumn: 1.45 (0.19-2.72)
Percent Increase (95% CI) in mortality
by increment in PM2.5 combined by
region
All Cause Mortality
Humid Subtropical and Maritime: 1.02
(0.65-1.38)
Warm Summer Continental: 1.19 (0.73-
1.64)
Hot Summer Continental: 1.14 (0.55-
1.73)
Dry: 1.18 (-0.70-3.10)
Dry, Continental: 1.26 (-0.21-2.76)
Mediterranean: 0.50 (0.00-1.01)
CVD
Humid Subtropical and Maritime: 0.78
(0.05-1.51)
Warm Summer Continental: 1.43 (0.67-
2.19)
Hot Summer Continental: 0.43 (-0.53-
1.40)
Dry: 3.11 (-0.02-6.33)
Dry, Continental: 1.67 (-0.75-4.16)
Mediterranean: 0.16 (-0.46-0.79)
Ml
Humid Subtropical and Maritime: 0.97 (¦
0.29-2.26)
Warm Summer Continental: 1.50 (0.05-
2.97)
Hot Summer Continental: 0.64 (-0.96-
2.28)
July 2009
E-669
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Cleveland OH - 16.4
Columbus OH ¦ 16.2
Cincinnati OH ¦ 17.1
Steubenville OH ¦ 17.0
Toledo OH - 14.9
Dayton OH ¦ 16.2
Akron OH - 16.0
Warren OH - 15.3
Oklahoma OK ¦ 9.9
Tulsa OK - 11.1
Bend OR ¦ 7.8
Medford OR ¦ 9.9
Klamath OR - 10.6
Eugene OR ¦ 8.0
Portland OR ¦ 8.8
Gettysburg PA ¦ 13.4
Pittsburgh PA ¦ 15.7
State College PA ¦ 13.2
Carlisle PA - 15.1
Harrisburg PA ¦ 15.5
Erie PA - 13.1
Scranton PA - 11.8
Allentown PA ¦ 14.2
Wilkes Barre PA - 12.8
Mercer PA ¦ 14.1
Easton PA ¦ 14.0
Philadelphia PA ¦ 14.5
Washington PA ¦ 14.7
Providence Rl - 11.5
Charleston SC ¦ 12.1
Taylors SC ¦ 15.3
Columbia SC ¦ 14.0
Spartanburg SC ¦ 14.2
Nashville TIM - 14.0
Knoxville TIM - 16.0
Memphis TN ¦ 13.5
San Antonio TX ¦ 9.4
Dallas TX - 12.9
El Paso TX ¦ 9.2
Houston TX ¦ 12.9
Port Arthur TX-11.5
Ft Worth TX - 12.2
Austin TX -10.4
Salt Lake UT ¦ 11.5
Provo UT ¦ 9.5
Dry: 4.25 (-2.38-11.33)
Dry, Continental: 0.60 (-7.42-9.32)
Mediterranean: 1.85 (-0.66-4.41)
Stroke
Humid Subtropical and Maritime: 2.94
(1.59-4.32)
Warm Summer Continental: 1.85 (0.04-
3.69)
Hot Summer Continental: 0.77 (-1.77-
3.38)
Dry: 1.82 (-6.98-11.45)
Dry, Continental: 2.49 (-2.32-7.53)
Mediterranean: 0.95 (-0.66-2.59)
Respiratory
Humid Subtropical and Maritime: 0.91 (¦
0.25-2.08)
Warm Summer Continental: 2.12 (0.89-
3.36)
Hot Summer Continental: 3.36 (1.95-
4.79)
Dry: 5.81 (-0.04-12.00)
Dry, Continental: -0.31 (-5.89-5.61)
Mediterranean: 1.06 (-0.36-2.50)
July 2009
E-670
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
WDCVA- 15.2
Annandale VA ¦ 14.0
Dumbarton VA ¦ 13.6
Chesapeake VA ¦ 12.7
Norfolk VA-12.7
Richmond VA ¦ 14.3
Seattle WA -10.1
Tacoma WA ¦ 11.2
Spokane WA ¦ 9.1
Dodge Wl- 11.1
Milwaukee Wl -13.2
Waukesha Wl - 13.2
Range (Min, Max): NR
Copollutant (correlation): NR
Reference: Zeger et al. (2007,157176) Outcome: Mortality
Period of Study: 2000-2002
Location: 250 largest US counties
Study Design: Retrospective Cohort
(MCAPS)
Pollutant: PM2.6
Averaging Time: 3 year avg
Statistical Analyses: log-linear
regression models (GAM)
Covariates: age, gender, race, county-
level SES, education and C0PD SMR
Age Groups: 65 +
65-74,75-84,85 +
Increment: 10 //g/m3
65+: 1.076(1.044,1.108)
Eastern US: 1.125 (1.091,1.159)
Central US: 1.196 (1.115,1.277)
Western US: 1.029 (0.994,1.064)
65-74: 1.156(1.117,1.196)
75-84: 1.081 (1.042,1.121)
85+: 0.995 (0.956,1.035)
Reference: Zeger et al. (2008,191951) Outcome: Mortality
Period of Study: 2000-2005
Study Design: Retrospective Cohort
Location: 4568 zip codes in urban areas Statistical Analysis: Log-linear
regression model
Age Groups: £65
Pollutant: PM2.6
Averaging Time: Annual
Median (SD) Unit:
Eastern: 14.0 //g|m3
Central: 10.7 //g|m3
Western: 13.1 //g|m3
All: 13.2/yg|m3
Range (IQR):
Eastern: 12.3-15.3
Central: 9.8-12.2
Western: 10.4-18.5
All: 11.1-14.9
Copollutant (correlation): NR
Increment: 10 //g/m3
Relative Risk (Min CI, Max CI)
Lag
Risk estimate for increase in mortality per
increase in PM2.6, all ages
Eastern Region
Age: 1.155 (1.130-1.180)
Age + SES: 1.105 (1.084-1.125)
Age + SES + C0PD: 1.068 (1.049-
1.087)
Central Region
Age: 1.178 (1.133-1.222)
Age + SES: 1.089 (1.052-1.125)
Age + SES + C0PD: 1.132 (1.095-
1.169)
Western Region
Age: 1.003 (0.981-1.025)
Age + SES: 0.997 (0.978-1.016)
Age + SES + C0PD: 0.989 (0.970-
1.008)
Risk estimate for increase in mortality per
increase in PM2.6, ages 65-74
Eastern Region
Age: 31.1 (26.8-35.5)
Age + SES: 17.3(14.6-20.0)
July 2009
E-671
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Age + SES + COPD: 11.4(8.8-14.1)
Central Region
Age: 39.0 (29.7-48.2)
Age + SES: 16.5(10.9-22.1)
Age + SES + COPD: 20.4(15.0-25.8)
Western Region
Age: 6.0 (2.3-9.6)
Age + SES: -2.1 (-5.0-0.8)
Age + SES + COPD:-1.5 (-4.2-1.1)
Risk estimate for increase in mortality per
increase in PM2.6, ages 75-84
Eastern Region
Age: 17.6(14.9-20.4)
Age + SES: 12.4 (10.1-14.6)
Age + SES + COPD: 8.9 (6.8-11.0)
Central Region
Age: 17.5 (12.7-22.2)
Age + SES: 8.8(4.6-13.0)
Age + SES + COPD: 12.0(7.6-16.4)
Western Region
Age: 0.4 (-2.0-2.7)
Age + SES: 0.3 (-1.8-2.5)
Age + SES + COPD:-0.2 (-2.2-1.9)
Risk estimate for increase in mortality per
increase in PM2.6, aged £85
Eastern Region
Age:-1.4 (-3.5-0.8)
Age + SES: 1.4 (-0.7-3.5)
Age + SES + COPD: 1.7 (-0.3-3.7)
Central Region
Age:-2.1 (-5.9-1.6)
Age + SES: -0.7 (-4.2-2.8)
Age + SES + COPD: -0.3 (-4.0-3.3)
Western Region
Age: -5.2 (-7.2-3.2)
Age + SES: 0.9 (-0.8-2.7)
Age + SES + COPD:-0.5 (-2.5-1.5)
'All units expressed in //gfm3 unless otherwise specified.
Table E-34. Long-term exposure - central nervous system outcomes - PM.
Study
Design & Methods
Concentrations
Effect Estimates (95% CI)
Author: Calderon-Garciduenas et al.
(2008,192369)
Period of Study: NR
Outcome (ICD9 and ICD10):
PM Size: No measure of PM
hippocampus, substantia nigrae,
Location: Mexico City (polluted city) and periaqueductal gray and vagus nerves
C0X2 (cyclooxygenase), IL-1 (3, CD14 in used Mexico City as the "polluted city"
lungs, OB (olfactory bulb), frontal cortex, and Tlaxcala and Veracruz as the "control
cities"
PM Increment: NA
Effect Estimate [Lower CI, Upper CI]:
RT-PCR sample results from Control and
Mexico City (MC) lung, CNS, PNS
July 2009
E-672
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
T'axcala and Veracruz (control cities), fl G Analyzed:
Mexico	a M '
Subjects 2-45 yrs of age
mean-25.1 ± 1.5 yrs
Study Design:
Cross-sectional
N: 47 deceased subjects with complete
autopsies and neuropathological
examinations (each subject had to be
considered clinically healthy and
cognitively and neurologically intact prior
to death) (primarily cause of death:
accidents resulting in immediate death)
Statistical Analyses: NR
likely used T-tests
in addition, stated using "parametric
procedure that considers the differences
among variances of the variables of
interest"
Covariates: Age, gender, place of birth,
place of residency, occupation, smoking
habits, clinical histories, cause of death,
and time between death and autopsy
Season: NR
Dose-response Investigated? (Yes/No):
No
Statistical package: Stata
Averaging Time: NA
Mean (SD): NA
Percentiles: NA
Range (Min, Max): NA
Unit (i.e. /yg/m3): NA
Number of Monitoring Stations: NA
Co-pollutant (correlation):
NA
(peripheral nervous system) tissues and p-
value for the difference between the
means
concentrations are normalized to the
amount of GAPDH cDNA
C0X2 (cyclooxygenase-2) lung
Controls: 15.9±6.7 x 10e
MC residents: 42.3±7.4 x 10e
p-value: 0.015
IL-1 (3 lung
Controls: 3.08 ± 1.87 x 106
MC residents: 4.51 ±2.6 x 10e
p-value: 0.60
C0X2 0B (olfactory bulb)
Controls: 12.9±3.0 x 106
MC residents: 38.7±5.5x106
p-value: 0.0002
IL-1 (3 0B
Controls: 3.4±0.8 x 10"
MC residents: 7.7±1.0 x 10"
p-value: 0.003
CD140B
Controls: 0.01 ±0.001
MC residents: 0.04±0.01
p-value: 0.04
C0X2 frontal
Controls: 2.6±0.4 x 106
MC residents: 5.0 ± 0.7 x 106
p-value: 0.008
IL-1 p frontal
Controls: 0.6±0.2 x 10"
MC residents: 6.2 ± 1.3 x 10"
p-value: 0.0002
C0X2 hippocampus
Controls: 1.9±0.5 x 106
MC residents: 1.6 ± 8.7 x 106
p-value: 0.1
IL-1 (3 hippocampus
Controls: 1.8 ±0.2x10"
MC residents: 3.0 ± 0.5 x 10"
p-value: 0.06
C0X2 substantia nigrae
Controls: 0.16±0.06
MC residents: 0.97±0.2
p-value: 0.03
IL-1 (3 substanita nigrae
July 2009
E-673
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Controls: 0.01 ±0.005
MC residents: 0.09±0.03
pvalue: 0.06
CD14 substantia nigrae
Controls: 0.02±0.005
MC residents: 0.03±0.007
p value: 0.7
C0X2 periaqueductal gray
Controls: 0.10 ±0.03
MC residents: 0.45±0.12
p value: 0.12
IL-1 (3 periaqueductal gray
Controls: 0.009±0.003
MC residents: 0.07±0.02
p value: 0.09
C0X2 left vagus
Controls: 0.65±0.18
MC residents: 2.68±0.82
p value: 0.03
C0X2 right vagus
Controls: 0.43±0.09
MC residents: 3.68±0.8
p value: 0.0002
IL-1 (3 left vagus
Controls: 0.1 ±0.03
MC residents: 1.3±0.73
p value: 0.06
IL-1 fB right vagus
Controls: 0.15 ±0.09
MC residents: 0.87±0.53
p value: 0.66
CD14 left vagus
Controls: 0.07±0.01
MC residents: 0.79±0.41
p value: 0.01
CD14 right vagus
Controls: 0.05±0.01
MC residents: 0.31 ±0.1
p value: 0.02
Distribution of subjects with expression
of Af342 as a function of age and
residency
Groups: No (%) with Af342 expression
Controls < 25yr AP0E 3/3 (n — 6): 0 (0)
Controls > 25yr AP0E 3/3 (n — 3): 0 (0)
MC E2 or E3 < 25yr (n-17): 10 (58.82)
July 2009
E-674
DRAFT-DO NOT CITE OR QUOTE

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CI)
MC E2 or E3 >25yr (n-10): 8 (80)
MC E4 (n — 8): 8 (100)
Controls E4 (n — 3): 2 (66)
Distribution of subjects with expression
of ~ synuclein as a function of age and
residency
Groups: No (%) with synuclein
expression
Controls < 25yr APOE 3/3 (n — 6): 0 (0)
Controls >25yr APOE 3/3 (n — 3): 0 (0)
MC E2 or E3 < 25yr (n-17): 4 (23.5)
MC E2 or E3 >25yr (n-10): 3 (30)
MC E4 (n — 8): 2 (25)
Controls E4 (n — 3): 0 (0)
Increment: 10/yg/m3
Regression Coefficient p (95% CI)
Crude
SRTT: 2.14 (-0.084.36)
SDST: 0.08(0.04-0.13)
SDLT Trials: 0.22(0.13-0.31)
SDLT Total: 0.44 (0.23-0.65)
Model 1: adjusted for age, sex,
race/ethnicity
SRTT: 2.03 (-0.15-4.20)
SDST: 0.10(0.05-0.15)
SDLT Trials: 0.23(0.14-0.32)
SDLT Total: 0.48 (0.27-0.68)
Model 2: Model 1 + socioeconomic
factors
SRTT: -0.11 (-2.38-2.16)
SDST: 0.01 (-0.04-0.06)
SDLT Trials: 0.01 (-0.08-0.10)
SDLT Total:-0.07 (-0.27-0.13)
Model 3: Model 2 + lifestyle factors
SRTT:-0.36 (-2.58-1.85)
SDST: 0.00 (-0.04-0.05)
SDLT Trials: 0.09(0.00-0.17)
SDLT Total: 0.12 (-0.07-0.31)
Reference: Chen and Schwartz (2009,
179945)
Period of Study: 1989-1991
Location: US
Outcome: change in central nervous
system function
Study Design: Panel
Covariates: age, sex, race/ethnicity,
individual socioeconomic position,
lifestyle factors, household and
neighborhood characteristics,
conventional CVD risk factors
Statistical Analysis: Pearson Chi-
square tests and t-tests, as appropriate
Statistical Package: STATA
Age Groups: 20-59 years
Pollutant: PMio
Averaging Time: 1 year
Mean (SD) Unit: 37.2 ± 12.8/yg/m3
Copollutant: 0:i
July 2009
E-675
DRAFT-DO NOT CITE OR QUOTE

-------
Study
Design & Methods
Concentrations'
Effect Estimates (95% CI)
Author: Suglia et al. (2008,157027)
Period of Study: 1986-2001
Location: Boston, Massachusetts
Outcome (ICD9 and ICD10):
Cognition:
Kaufman Brief Intelligence Test, K BIT
(vocabulary and matrices subscales and
composite IQ score)
Wide Range Assessment of Memory and
Learning, WRAML (psychometric
instrument with subscales on verbal
memory, visual memory, learning, and
overall general index scale)
All cognition scores have a mean of 100
and SD-15.
Age Groups Analyzed: Cognitive tests
administered when children were 8-11
yrs of age
Study Design: Cross-sectional
N: 202 children
Statistical Analyses: Linear regression
Covariates: Child's age at cognitive
assessment, gender, primary language
spoken at home, and maternal education
(model 1
"demographic factors")
sensitivity analyses performed with
further adjustment for in-utero and
postnatal secondhand tobacco smoke
exposure (via questionnaire during follow-
ups and urinary cotinine levels) (model 2)
birth weight (model 3)
and blood lead level (model 4)
Season: Separate land-use regression
models were fit for the warm (May-Oct)
and cold (Nov-Apr) seasons
used average of two seasons as measure
of average lifetime BC exposure
Dose-response Investigated? (Yes/No):
No
Statistical package: SAS (v9.0)
PM Size: Black carbon (BC)
Averaging Time: Lifetime exposure
Estimated 24hr measures of traffic using
a spatiotemporal land-use regression
model using data from > 80 locations in
Greater Boston (6021 pollution
measurements from 2127 unique
exposure days)
Predictors in the land-use regression
analysis were the BC level at a central
station (to capture average
concentrations on that day),
meteorological conditions,
weekdayfweekend, and measure of
traffic activity (GIS-based measures of
cumulative traffic density within 100m,
population density, distance to nearest
major roadway, % urbanization)
Used the average of the cold and warm
seasons as the measure of average
lifetime BC exposure
Mean (SD): 0.56(0.13) (Jg/m3
Percentiles: NR
Range (Min, Max): NR
Unit (i.e. /yg/m3):
Number of Monitoring Stations: >80
locations
Co-pollutant (correlation): NA
July 2009
E-676
PM Increment: 0.4/yg/m3
Effect Estimate [Lower CI, Upper CI]:
Change in subscale score (95%CI) per IQR
(0.4 |jg/m3) increase in log BC level
KBIT
Vocabulary:
Adj for demographic factors: -2.0 (-5.3,
1.3)
Adj for above factors + secondhand
smoke: -2.0 (-5.3,1.4)
Adj for above factors + birth weight: -2.0
(-5.4,1.3)
Adj for above factors + blood lead level: -
2.2 (-5.5,1.1)
Matrices:
Adj for demographic factors: -4.2 (-7.7, ¦
0.7)
Adj for above factors + secondhand
smoke: -4.0 (-7.5, -0.4)
Adj for above factors + birth weight: -4.0
(-7.6, -0.5)
Adj for above factors + blood lead level: -
4.0 (-7.6,-0.5)
Composite:
Adj for demographic factors: -3.4 (-6.6, ¦
0.3)
Adj for above factors + secondhand
smoke: -3.3 (-6.4, -0.1)
Adj for above factors + birth weight: -3.3
(-6.5, -0.2)
Adj for above factors + blood lead level: -
3.4 (-6.6, -0.3)
WRAML
Verbal:
Adj for demographic factors: -1.1 (-4.6,
2.3)
Adj for above factors + secondhand
smoke: -1.2 (-4.7, 2.3)
Adj for above factors + birth weight: -1.3
(-4.7, 2.2)
Adj for above factors + blood lead level: -
1.3(-4.8, 2.2)
Visual:
Adj for demographic factors: -5.2 (-8.6, ¦
1.7)
Adj for above factors + secondhand
smoke: -5.3 (-8.8, -1.8)
Adj for above factors + birth weight: -5.3
(-8.8,-1.8)
Adj for above factors + blood lead level: -
5.4 (-8.9,-1.9)
Learning:
Adj for demographic factors: -2.7 (-6.5,
1.1)
Adj for above factors + secondhand
smoke: -2.6 (-6.5,1.2)
Adj for above factors + birth weight: -2.6
' 6DRAFT - DO NOT CITE OR QUOTE
Adj for above factors + blood lead level: -
2.8(-6.6,1.1)
General:

-------
Study	Design & Methods	Concentrations'	Effect Estimates (95% CTT
'All units expressed in //gfm3 unless otherwise specified.
July 2009
E-677
DRAFT-DO NOT CITE OR QUOTE

-------
Annex E References
Abbey DE; Nishino N; McDonnell WF; Burchette RJ; Knutsen SF; Beeson WL; Yang JX. (1999). Long-term
inhalable particles and other air pollutants related to mortality in nonsmokers. Am J Respir Crit Care
Med, 159: 373-382. 047559
Abrahamowicz Mi Schopflocher Ti Leffondre Ki Du Berger Ri Krewski D. (2003). Flexible modeling of exposure-
response relationship between long-term average levels of particulate air pollution and mortality in the
American Cancer Society study. J Toxicol Environ Health A, 66: 1625-1654. 086292
Ackermann-Liebrich Ui Leuenberger Pi Schwartz J; Schindler Ci Monn Ci Bolognini Bi Bongard JPi Brandli Oi
Domenighetti Gi Elsasser Si Grize Li Karrer Wi Keller Ri Keller-Wossidlo Hi Kunzli Ni Martin BWi
Medici TCi Pliger Bi Wuthrich Bi Zellweger JPi Zemp E. (1997). Lung function and long term exposure to
air pollutants in Switzerland. Am J Respir Crit Care Med, 155: 122-129. 077537
Adamkiewicz Gi Ebelt Si Syring Mi Slater Ji Speizer FEi Schwartz Ji Suh Hi Gold DR. (2004). Association
between air pollution exposure and exhaled nitric oxide in an elderly population. Thorax, 59: 204-209.
087925
Adar SDi Adamkiewicz Gi Gold DRi Schwartz Ji Coull BAi Suh H. (2007). Ambient and microenvironmental
particles and exhaled nitric oxide before and after a group bus trip. Environ Health Perspect, 115: 507"
12. 098635
Adar SDi Gold DRi Coull BAi Schwartz Ji Stone PHi Suh H. (2007). Focused exposures to airborne traffic
particles and heart rate variability in the elderly. Epidemiology, 18: 95-103. 001458
Aekplakorn Wi Loomis Di Vichit-Vadakan Ni Shy Ci Plungchuchon S. (2003). Acute effects of S02 and particles
from a power plant on respiratory symptoms of children, Thailand. , 34: 906-914. 089908
Aga Ei Samoli Ei Touloumi Gi Anderson HRi Cadum Ei Forsberg Bi Goodman Pi Goren Ai Kotesovec Fi Kriz Bi
Macarol-Hiti Mi Medina Si Paldy Ai Schindler Ci Sunyer Ji Tittanen Pi Wojtyniak Bi Zmirou Di Schwartz
Ji Katsouyanni K. (2003). Short-term effects of ambient particles on mortality in the elderly: results from
28 cities in the APHEA2 project. , 40: 28s-33s. 187122
Agarwal Ri Jayaraman Gi Anand Si Marimuthu P. (2006). Assessing respiratory morbidity through pollution
status and meteorological conditions for Delhi. Environ Monit Assess, 114: 489-504. 099086
Allen RWi Criqui MHi Diez Roux AVi Allison Mi Shea Si Detrano Ri Sheppard Li Wong Ni Hinckley Stukovsky
Ki Kaufman JD. (2009). Fine particulate air pollution, proximity to traffic, and aortic atherosclerosis:
The Multi-Ethnic Study of Atherosclerosis. Epidemiology, 20: 254-264. 189644
Allen RWi Mar Ti Koenig Ji Liu LJi Gould Ti Simpson Ci Larson T. (2008). Changes in lung function and airway
inflammation among asthmatic children residing in a woodsmoke-impacted urban area. Inhal Toxicol,
20: 423-433. 156208
Analitis Ai Katsouyanni Ki Dimakopoulou Ki Samoli Ei Nikoloulopoulos AKi Petasakis Yi Touloumi Gi
Schwartz Ji Anderson HRi Cambra Ki Forastiere Fi Zmirou Di Vonk JMi Clancy Li Kriz Bi Bobvos Ji
Pekkanen J. (2006). Short-term effects of ambient particles on cardiovascular and respiratory mortality. ,
17: 230-233. 088177
Andersen ZJi Wahlin Pi Raaschou-Nielsen Oi Ketzel Mi Scheike Ti Loft S. (2008). Size distribution and total
number concentration of ultrafine and accumulation mode particles and hospital admissions in children
and the elderly in Copenhagen, Denmark. , 65: 458-66. 189651
Andersen ZJi Wahlin Pi Raaschou-Nielsen Oi Scheike Ti Loft S. (2007). Ambient particle source apportionment
and daily hospital admissions among children and elderly in Copenhagen. J Expo Sci Environ Epidemiol,
17: 625-636. 093201
Anderson HRi Atkinson RWi Bremner SAi Marston L. (2003). Particulate air pollution and hospital admissions
for cardiorespiratory diseases: are the elderly at greater risk?. Eur Respir J, 40: 39s-46s. 054820
Anderson MEi Bogdan GM. (2007). Environments, indoor air quality, and children. Pediatr Clin North Am, 54:
295-307, viii. 156214
Annesi-Maesano, I.i Forastiere, F.i Kunzli, N.i Brunekref, B.i Environment and Health Committee of the
European Respiratory Society. (2007). Particulate matter, science and EU policy. Eur Respir J, 29: 428"
431. 091348
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health
and Environmental Research Online) at http: //epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in
the process of developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk
Information System (IRIS).
July 2009
E-678
DRAFT-DO NOT CITE OR QUOTE

-------
Arbex MA; Martins LC; De Oliveira RC; Pereira LA; Arbex FF; Cancado JE; Saldiva PH; Braga AL. (2007). Air
pollution from biomass burning and asthma hospital admissions in a sugar cane plantation area in
Brazil. J Epidemiol Community Health, 61- 395-400. 091637
Auchincloss AH; Roux AV; Dvonch JT; Brown PL; Barr RG; Daviglus ML; Goff DC; Kaufman JD; O'Neill MS.
(2008). Associations between Recent Exposure to Ambient Fine Particulate Matter and Blood Pressure in
the Multi-Ethnic Study of Atherosclerosis (MESA). Environ Health Perspect, 116- 486-491. 156234
Avol EL; Gauderman WJ; Tan SM; London SJ; Peters JM. (2001). Respiratory effects of relocating to areas of
differing air pollution levels. Am J Respir Crit Care Med, 164: 2067-2072. 020552
Babin SM; Burkom HS; Holtry RS; Tabernero NR; Stokes LD; Davies-Cole JO; DeHaan K; Lee DH. (2007).
Pediatric patient asthma-related emergency department visits and admissions in Washington, DC, from
2001-2004, and associations with air quality, socio-economic status and age group. , 6: 9. 188476
Baccarelli A; Martinelli I; Pegoraro V; Melly S; Grillo P; Zanobetti A; Hou L; Bertazzi PA; Mannucci PM;
Schwartz J. (2009). Living near Major Traffic Roads and Risk of Deep Vein Thrombosis. , x: x. 188183
Baccarelli A; Martinelli I; Zanobetti A; Grillo P; Hou LF; Bertazzi PA; Mannucci PM; Schwartz J. (2008).
Exposure to Particulate Air Pollution and Risk of Deep Vein Thrombosis. Arch Intern Med, 168: 920-927.
157984
Baccarelli A; Zanobetti A; Martinelli I; Grillo P; Hou L; Giacomini S; Bonzini M; Lanzani G; Mannucci PM;
Bertazzi PA; Schwartz J. (2007). Effects of exposure to air pollution on blood coagulation. J Thromb
Haemost, 5: 252-260. 090733
Baccarelli A; Zanobetti A; Martinelli I; Grillo P; Hou L; Lanzani G; Mannucci PM; Bertazzi PA; Schwartz J.
(2007). Air pollution, smoking, and plasma homocysteine. Environ Health Perspect, 115: 176-181. 091310
Bakke B; Ulvestad B; Stewart P; Eduard W. (2004). Cumulative exposure to dust and gases as determinants of
lung function decline in tunnel construction workers. Occup Environ Med, 61: 262-269. 156246
Ballester F; Medina S; Boldo E; Goodman P; Neuberger M; Iniguez C; Kunzli N. (2008). Reducing ambient levels
of fine particulates could substantially improve health: a mortality impact assessment for 26 European
cities. J Epidemiol Community Health, 62: 98-105. 189977
Ballester F; Rodriguez P; Iniguez C; Saez M; Daponte A; Galan I; Taracido M; Arribas F; Bellido J; Cirarda FB;
Canada A; Guillen JJ; Guillen-Grima F; Lopez E; Perez-Hoyos S; Lertxundi A; Toro S. (2006). Air
pollution and cardiovascular admisisons association in Spain: results within the EMECAS project. J
Epidemiol Community Health, 60: 328-336. 088746
Ballester F; Saez M; Perez-Hoyos S; Iniguez C; Gandarillas A; Tobias A; Bellido J; Taracido M; Arribas F;
Daponte A; Alonso E; Canada A; Guillen-Grima F; Cirera L; Perez-Boillos MJ; Saurina C; Gomez F;
Tenias JM. (2002). The EMECAM project: a multicentre study on air pollution and mortality in Spain:
combined results for particulates and for sulfur dioxide. Occup Environ Med, 59: 300-308. 030371
Barclay JL; Egred M; Kruszewski K; Nandakumar R; Norton MY; Stirrat C; Redpath TW; Walton S; Hillis GS.
(2007). The relationship between transmural extent of infarction on contrast enhanced magnetic
resonance imaging and recovery of contractile function in patients with first myocardial infarction
treated with thrombolysis. Cardiology, 108: 217-22. 192229
Barclay JL; Miller BG; Dick S; Dennekamp M; Ford I; Hillis GS; Ayres JG; Seaton A. (2009). A panel study of
air pollution in subjects with heart failure: negative results in treated patients. Occup Environ Med, 66:
325-334. 179935
Barnett AG; Williams GM; Schwartz J; Neller AH; Best TL; Petroeschevsky AL; Simpson RW. (2005). Air
pollution and child respiratory health: a case-crossover study in Australia and New Zealand. Am J Respir
Crit Care Med, 171: 1272-1278. 087394
Barraza-Villarreal A; Sunyer J; Hernandez-Cadena L; Escamilla-Nunez MC; Sienra-Monge JJ; Ramirez"Aguilar
M; Cortez-Lugo M; Holguin F; Diaz-Sanchez D; Olin AC; Romieu I. (2008). Air pollution, airway
inflammation, and lung function in a cohort study of Mexico City schoolchildren. Environ Health
Perspect, 116: 832-838. 156254
Bartzokas A; Kassomenos P; Petrakis M; Celessides C. (2004). The effect of meteorological and pollution
parameters on the frequency of hospital admissions for cardiovascular and respiratory problems in
Athens. Indoor Built Environ, 13: 271-275. 093252
Basu R; Feng WY; Ostro BD. (2008). Characterizing temperature and mortality in nine California counties. , 19:
138-45. 098716
Bateson TF; Schwartz J. (2004). Who is sensitive to the effects of particulate air pollution on mortality? A case-
crossover analysis of effect modifiers. Epidemiology, 15: 143-149. 086244
Bayer-Oglesby L; Grize L; Gassner M; Takken-Sahli K; Sennhauser FH; Neu U; Schindler C; Braun-Fahrlander
C. (2005). Decline of ambient air pollution levels and improved respiratory health in Swiss children.
Environ Health Perspect, 113: 1632-1637. 086245
July 2009
E-679
DRAFT-DO NOT CITE OR QUOTE

-------
Beelen R; Hoek Gi van den Brandt PA; Goldbohm RA; Fischer P; Schouten LJ; Jerrett Mi Hughes Ei Armstrong
Bi Brunekreef B. (2008). Long-term effects of traffic-related air pollution on mortality in a Dutch cohort
(NLCS-AIR study). Environ Health Perspect, 116- 196-202. 156263
Beeson WLi Abbey DE; Knutsen SF. (1998). Long-term concentrations of ambient air pollutants and incident
lung cancer in California adults' results from the AHSMOG study. Environ Health Perspect, 106- 813-
823.048890
Bell Mi Ebisu K> Peng R; Samet J; Dominici F. (2009). Hospital Admissions and Chemical Composition of Fine
Particle Air Pollution. Am J Respir Crit Care Med, 179- 1115-1120. 191997
Bell ML; Ebisu K; Belanger K. (2007). Ambient air pollution and low birth weight in Connecticut and
Massachusetts. Environ Health Perspect, 115: 1118-24. 091059
Bell ML; Ebisu K; Peng RD; Dominici F. (2009). Adverse health effects of particulate air pollution: modification
by air conditioning. Epidemiology, in press: in press. 191007
Bell ML; Kim JY; Dominici F. (2007). Potential confounding of particulate matter on the short-term association
between ozone and mortality in multisite time-series studies. Environ Health Perspect, 115: 1591-1595.
093256
Bell ML; Levy JK; Lin Z. (2008). The effect of sandstorms and air pollution on cause-specific hospital admissions
in Taipei, Taiwan. Occup Environ Med, 65: 104-111. 091268
Bellini P; Baccini M; Biggeri A; Terracini B. (2007). The meta-analysis of the Italian studies on short-term
effects of air pollution (MISA): old and new issues on the interpretation of the statistical evidences.
Environmetrics, 18: 219-229. 097787
Bennett CM; Simpson P; Raven J; Skoric B; Powell J; Wolfe R; Walters EH; Abramson MJ. (2007). Associations
between ambient PM2.5 concentrations and respiratory symptoms in Melbourne, 1998-2005. J Toxicol
Environ Health A, 70: 1613-1618. 156268
Binkova B; Chvatalova I; Lnenickova Z; Milcova A; Tulupova E; Farmer PB; Sram RJ. (2007). PAH-DNA
adducts in environmentally exposed population in relation to metabolic and DNA repair gene
polymorphisms. , 620: 49-61. 156273
Boezen HM! Van Der Zee SC; Postma DS; Vonk JM; Gerritsen J; Hoek G; Brunekreef B; Rijcken B; Schouten
JP. (1999). Effects of ambient air pollution on upper and lower respiratory symptoms and peak
expiratory flow in children. Lancet, 353: 874-878. 040410
Boezen HM; Vonk JM; Van Der Zee SC; Gerritsen J; Hoek G; Brunekreef B; Schouten JP; Postma DS. (2005).
Susceptibility to air pollution in elderly males and females. Eur Respir J, 25: 1018-1024. 087396
Bourotte C', Curi-Amarante A"P; Forti M"C; APereira LA; Braga AL; Lotufo PA. (2007). Association between
ionic composition of fine and coarse aerosol soluble fraction and peak expiratory flow of asthmatic
patients in Sao Paulo city (Brazil). Atmos Environ, 41: 2036-2048. 150040
Brauer M; Gehring U; Brunekreef B; De Jongste J; Gerritsen J; Rovers M; Wichmann H"E; Wijga A; Heinrich J.
(2006).	Traffic-related air pollution and otitis media. Environ Health Perspect, 114: 1414-1418. 090757
Brauer M; Hoek G; Smit HA; De Jongste JC; Gerritsen J; Postma DS; Kerkhof M; Brunekreef B. (2007). Air
pollution and development of asthma, allergy and infections in a birth cohort. Eur Respir J, 29: 879-888.
090691
Brauer M; Hoek G; Van Vliet P; Meliefste K; Fischer PH; Wijga A; Koopman LP; Neijens HJ; Gerritsen J;
Kerkhof M; Heinrich J; Bellander T; Brunekreef B. (2002). Air pollution from traffic and the development
of respiratory infections and asthmatic and allergic symptoms in children. Am J Respir Crit Care Med,
166: 1092-1098. 035192
Brauer M; Lencar C; Tamburic L; Koehoorn M; Demers P; Karr C. (2008). A cohort study of traffic-related air
pollution impacts on birth outcomes. Environ Health Perspect, 116: 680-686. 156292
Breitner S; Stolzel M; Cyrys J; Pitz M; Wolke G; Kreyling W', Kiichenhoff H; Heinrich J; Wichmann H; Peters A.
(2009). Short-Term Mortality Rates during a Decade of Improved Air Quality in Erfurt, Germany.
Environ Health Perspect, 117: 448. 188439
Briet M; Collin C; Laurent S; Tan A; Azizi M; Agharazii M; Jeunemaitre X; Alhenc-Gelas F; Boutouyrie P.
(2007).	Endothelial function and chronic exposure to air pollution in normal male subjects.
Hypertension970"976. 093049
Brunekreef B; Beelen R; Hoek G; Schouten L; Bausch-Goldbohm S; Fischer P; Armstrong B; Hughes E; Jerrett
M; van den Brandt P. (2009). Effects of long-term exposure to traffic-related air pollution on respiratory
and cardiovascular mortality in the Netherlands: The NLCS-AIR Study. Health Effects Institute. Boston,
MA. 139. 191947
Burnett RT; Stieb D; Brook JR; Cakmak S; Dales R; Raizenne M; Vincent R; Dann T. (2004). Associations
between short-term changes in nitrogen dioxide and mortality in Canadian cities. Arch Environ Occup
Health, 59: 228-236. 086247
July 2009
E-680
DRAFT-DO NOT CITE OR QUOTE

-------
Burr ML; Karani G> Davies B; Holmes BA; Williams KL. (2004). Effects on respiratory health of a reduction in
air pollution from vehicle exhaust emissions. , 61- 212. 189788
Burr ML; Karani GJ Davies B; Holmes BA; Williams KL. (2004). Effects on respiratory health of a reduction in
air pollution from vehicle exhaust emissions. Occup Environ Med, 61- 212-218. 087809
Cakmak S; Dales RE; Vidal CB. (2007). Air pollution and mortality in Chile- susceptibility among the elderly.
Environ Health Perspect, 115: 524"7. 091170
Calderon-Garciduenas L; Maronpot RR; Torres-Jardon R; Henriquez-Roldan C; Schoonhoven R; Acuna-Ayala H;
Villarreal-Calderon A; Nakamura J; Fernando R; Reed W; Azzarelli B; Swenberg JA. (2003). DNA
damage in nasal and brain tissues of canines exposed to air pollutants is associated with evidence of
chronic brain inflammation and neurodegeneration. Toxicol Pathol, 31: 524-538. 156316
Calderon-Garciduenas L; Mora-Tiscareno A; Fordham LA; Chung CJ; Valencia-Salazar G; Flores-Gomez S; Solt
AC; Gomez-del Campo A; Jardon-Torres R; Henriquez-Roldan C; Hazucha MJ; Reed W. (2006). Lung
radiology and pulmonary function of children chronically exposed to air pollution. Environ Health
Perspect, 114: 1432-1437. 091253
Calderon-Garciduenas L; Macias-Parra M; Hoffmann HJ; Valencia-Salazar G; Henriquez-Roldan C; Osnaya N,
Monte OC; Barragan-Mejia G; Villarreal-Calderon R; Romero L; Granada-Macias M; Torres-Jardon R;
Medina-Cortina H; Maronpot RR. (2009). Immunotoxicity and environment: immunodysregulation and
systemic inflammation in children. Toxicol Pathol, 37: 161-169. 192107
Calderon-Garciduenas L; Mora-Tiscareno A; Ontiveros E; Gomez-Garza G; Barragan-Mejia G; Broadway J;
Chapman S; Valencia-Salazar G; Jewells V; Maronpot RR; Henriquez-Roldan C; Perez-Guille B; Torres-
Jardon R; Herrit L; Brooks D; Osnaya-Brizuela N; Monroy M. (2008). Air pollution, cognitive deficits and
brain abnormalities: A pilot study with children and dogs. Brain Cognit, 68: 117-127. 156317
Calderon-Garciduenas L; Solt A; Henriquez-Roldan C; Torres-Jardon R; Nuse B; Herritt L; Stone I. (2008).
Long-term air pollution exposure is associated with neuroinflammation , an altered innate immune
response, disruption of the blood-brain-barrier, ultrafine particle deposition, and accumulation of amyloid
642 and a synuclein in children and young adults. Toxicol Pathol, 36: 289-310. 192369
Cavallari JM; Eisen EA; Chen JC; Fang SC; Dobson CB; Schwartz J; Christiani DC. (2007). Night Heart Rate
Variability and Particulate Exposures among Boilermaker Construction Workers. Environ Health
Perspect, 115: 1046-1051. 157425
Cavanagh JA; Brown L; Trought K; Kingham S; Epton MJ. (2007). Elevated concentrations of 1 "hydroxypyrene
in schoolchildren during winter in Christchurch, New Zealand. Sci Total Environ, 374: 51"9. 098618
Cesaroni G; Badaloni C; Porta D; Forastiere F; Perucci CA. (2008). Comparison between several indices of
exposure to traffic-related air pollution and their respiratory health impact in adults. Occup Environ
Med, 65: 683-690. 156326
Chahine T; Baccarelli A; Litonjua A; Wright RO; Suh H; Gold DR; Sparrow D; Vokonas P; Schwartz J. (2007).
Particulate air pollution, oxidative stress genes, and heart rate variability in an elderly cohort. Environ
Health Perspect, 115: 1617-1622. 156327
Chan CC; Chuang KJ; Chen WJ; Chang WT; Lee CT; Peng CM. (2008). Increasing cardiopulmonary emergency
visits by long-range transported Asian dust storms in Taiwan. Environ Res, 106: 393-400. 093297
Chan MN; Chan CK. (2007). Mass transfer effects on the hygroscopic growth of ammonium sulfate particles
with a water-insoluble coating. Atmos Environ, 41: 4423-4433. 147787
Chang SY; Fang GC. (2007). Springtime soluble particles in a suburban area of Taichung in central Taiwan. ,
86: 30-41. 147621
Chardon B; Lefranc A; Granados D; Gremy I. (2007). Air pollution and doctors' house calls for respiratory
diseases in the greater Paris area (2000-3). Occup Environ Med, 64: 320-4. 091308
Chattopadhyay BP; Mukherjee A; Mukherjee K; Roychowdhury A. (2007). Exposure to vehicular pollution and
assessment of respiratory function in urban inhabitants. Lung, 185: 263-70. 147471
Chen J; Schwartz J. (2009). Neurobehavioral effects of ambient air pollution on cognitive performance in US
adults. Neurotoxicology, 30: 231-239. 179945
Chen JC; Schwartz J. (2008). Metabolic syndrome and inflammatory responses to long-term particulate air
pollutants. Environ Health Perspect, 116: 612-617. 190106
Chen L; Yang W', Jennison BL; Goodrich A; Omaye ST. (2002). Air pollution and birth weight in northern
Nevada, 1991-1999. Inhal Toxicol, 14: 141-157. 024945
Chen LH; Knutsen SF; Shavlik D; Beeson WL; Petersen F; Ghamsary M; Abbey D. (2005). The association
between fatal coronary heart disease and ambient particulate air pollution: Are females at greater risk?.
Environ Health Perspect, 113: 1723-1729. 087942
Chen Y; Yang Q; Krewski D; Burnett RT; Shi Y; McGrail KM. (2005). The effect of coarse ambient particulate
matter on first, second, and overall hospital admissions for respiratory disease among the elderly. Inhal
Toxicol, 17: 649-655. 087555
July 2009
E-681
DRAFT-DO NOT CITE OR QUOTE

-------
Cheng M-F; Tsai S"S; Wu T"N; Chen P-S; Yang C-Y. (2007). Air pollution and hospital admissions for
pneumonia in a tropical city- Kaohsiung, Taiwan. J Toxicol Environ Health A, 70- 2021-6. 093034
Chimonas MA; Gessner BD. (2007). Airborne particulate matter from primarily geologic, non-industrial sources
at levels below National Ambient Air Quality Standards is associated with outpatient visits for asthma
and quick-relief medication prescriptions among children less than 20 years old enrolled in Medicaid in
Anchorage, Alaska. Environ Res, 103- 397-404. 093261
Chiu H; Tiao Mi Ho S> Kuo Hi Wu Ti Yang C. (2008). Effects of Asian Dust Storm events on hospital admissions
for chronic obstructive pulmonary disease in Taipei, Taiwan. Inhal Toxicol, 20- 777-781. 191989
Chiu HFi Cheng MHi Yang CY. (2009). Air pollution and hospital admissions for pneumonia in a subtropical
city: Taipei, Taiwan. , 21: 32-7. 190249
Choi J-H Xu Q-S; Park S-Y> Kim J-H; Hwang S-S> Lee K-H> Lee H-J; Hong Y-C. (2007). Seasonal variation of
effect of air pollution on blood pressure. J Epidemiol Community Health314"318. 093196
Chuang K-Ji Chan C"Ci Su T"C> Lee C-Ti Tang C-S. (2007). The effect of urban air pollution on inflammation,
oxidative stress, coagulation, and autonomic dysfunction in young adults. Am J Respir Crit Care Med,
176: 370-6. 091063
Chuang KJ; Chan CC; Su TC; Lin LY; Lee CT. (2007). Associations between particulate sulfate and organic
carbon exposures and heart rate variability in patients with or at risk for cardiovascular diseases. J
Occup Environ Med, 49: 610-7. 098629
Cardenas Mi Vallejo Mi Romano-Riquer Pi Ruiz-Velasco Si Ferreira-Vidal ADi Hermosillo AG. (2008). Personal
exposure to PM2.5 air pollution and heart rate variability in subjects with positive or negative head-up
tilt test. Environ Res, 108: 1-6. 191900
D'Ippoliti Di Forastiere Fi Ancona Ci Agabiti Ni Fusco Di Michelozzi Pi Perucci CA. (2003). Air pollution and
myocardial infarction in Rome: a case-crossover analysis. Epidemiology, 14: 528-535. 074311
Dales Ri Burnett RTi Smith-Doiron Mi Stieb DM; Brook JR. (2004). Air pollution and sudden infant death
syndrome. Pediatrics, 113: 628-631. 087342
Dales R; Liu L; Szyszkowicz M; Dalipaj M; Willey J; Kulka R; Ruddy TD. (2007). Particulate air pollution and
vascular reactivity: the bus stop study. Int Arch Occup Environ Health, 81: 159-164. 155743
Dales R; Wheeler A; Mahmud M; Frescura AM; Smith-Doiron M; Nethery E; Liu L. (2008). The Influence of
Living Near Roadways on Spirometry and Exhaled Nitric Oxide in Elementary Schoolchildren. Environ
Health Perspect, 116: 1423. 156378
Dales RE; Cakmak S; Doiron MS. (2006). Gaseous air pollutants and hospitalization for respiratory disease in
the neonatal period. Environ Health Perspect, 114: 1751-1754. 090744
Daniels MJ; Dominici F; Zeger SL; Samet JM. (2004). The national morbidity, mortality, and air pollution study
Part IIP PM10 concentration-response curves and thresholds for the 20 largest US cities. 087343
De Hartog JJ; Hoek G; Peters A; Timonen KL; Ibald-Mulli A; Brunekreef B; Heinrich J; Tiittanen P; Van
Wijnen JH; Kreyling W; Kulmala M; Pekkanen J. (2003). Effects of fine and ultrafine particles on
cardiorespiratory symptoms in elderly subjects with coronary heart disease: the ULTRA study. Am J
Epidemiol, 157: 613-623. 001061
De Leon SF; Thurston GD; Ito K. (2003). Contribution of respiratory disease to nonrespiratory mortality
associations with air pollution. Am J Respir Crit Care Med, 167: 1117-1123. 055688
Delfino R; Brummel S; Wu J; Stern H; Ostro B; Lipsett M; Winer A; Street D; Zhang L; Tjoa T. (2009). The
relationship of respiratory and cardiovascular hospital admissions to the southern California wildfires of
2003. Occup Environ Med, 66: 189. 191994
Delfino RJ; Gong H; Linn WS; Hu Y; Pellizzari ED. (2003). Respiratory symptoms and peak expiratory flow in
children with asthma in relation to volatile organic compounds in exhaled breath and ambient air. J
Expo Sci Environ Epidemiol, 13: 348-363. 090941
Delfino RJ; Quintana PJE; Floro J; Gastanaga VM; Samimi BS; Kleinman MT; Liu L-JS; Bufalino C; Wu C"F;
McLaren CE. (2004). Association of FEV1 in asthmatic children with personal and microenvironmental
exposure to airborne particulate matter. Environ Health Perspect, 112: 932-941. 056897
Delfino RJ; Staimer N; Gillen D; Tjoa T; Sioutas C; Fung K; George SC; Kleinman MT. (2006). Personal and
ambient air pollution is associated with increased exhaled nitric oxide in children with asthma. Environ
Health Perspect, 114: 1736-1743. 090745
Delfino RJ; Staimer N; Tjoa T; Polidori A; Arhami M; Gillen DL; Kleinman MT; Vaziri ND; Longhurst J;
Zaldivar F; Sioutas C. (2008). Circulating biomarkers of inflammation, antioxidant activity, and platelet
activation are associated with primary combustion aerosols in subjects with coronary artery disease.
Environ Health Perspect, 116: 898-906. 156390
Delfino RJ; Zeiger RS; Seltzer JM; Street DH. (1998). Symptoms in pediatric asthmatics and air pollution:
differences in effects by symptom severity, anti-inflammatory medication use and particulate averaging
time. Environ Health Perspect, 106: 751-761. 051406
July 2009
E-682
DRAFT-DO NOT CITE OR QUOTE

-------
Delfino RJ; Zeiger RS; Seltzer JM; Street DH; McLaren CE. (2002). Association of asthma symptoms with peak
particulate air pollution and effect modification by anti"inflammatory medication use. Environ Health
Perspect, 110: A607-A617. 093740
DeMeo DL; Zanobetti A; Litonjua AA; Coull BA; Schwartz J; Gold DR. (2004). Ambient air pollution and oxygen
saturation. Am J Respir Crit Care Med, 170- 383-387. 087346
Desqueyroux H; Pujet J-C> Prosper Mi Squinazi Fi Momas I. (2002). Short-term effects of low-level air pollution
on respiratory health of adults suffering from moderate to severe asthma. Environ Res, 89- 29"37. 026052
de Hartog JJi Lanki Ti Timonen KLi Hoek Gi Janssen NAi Ibald-Mulli Ai Peters Ai Heinrich Ji Tarkiainen THi
van Grieken Ri van Wijnen JHi Brunekreef Bi Pekkanen J. (2009). Associations between PM2.5 and
heart rate variability are modified by particle composition and beta-blocker use in patients with coronary
heart disease. Environ Health Perspect, 117: 105-111. 191904
Diette GBi Hansel NNi Buckley TJi Curtin-Brosnan Ji Eggleston PAi Matsui ECi McCormack MCi Williams DLi
Breysse PN. (2007). Home indoor pollutant exposures among inner-city children with and without
asthma. Environ Health Perspect, 115: 1665-1669. 156399
Diez Roux AVi Auchincloss AHi Franklin TGi Raghunathan Ti Barr RGi Kaufman Ji Astor Bi Keeler J. (2008).
Long-term exposure to ambient particulate matter and prevalence of subclinical atherosclerosis in the
Multi-Ethnic Study of Atherosclerosis. Am J Epidemiol, 167: 667-675. 156401
Diez-Roux AVi Auchincloss AHi Astor Bi Barr RGi Cushman Mi Dvonch Ti Jacobs DR Jri Kaufman Ji Lin Xi
Samson Px. (2006). Recent exposure to particulate matter and C-reactive protein concentration in the
multi-ethnic study of atherosclerosis x. Am J Epidemiol, 164: 437-448. 156400
Dominici Fi Daniels Mi McDermott Ai Zeger SLi Samet JM. (2003). Shape of the exposure-response relation and
mortality displacement in the NMMAPS database. 042804
Dominici Fi Peng RDi Bell MLi Pham Li McDermott Ai Zeger SLi Samet JL. (2006). Fine particulate air
pollution and hospital admission for cardiovascular and respiratory diseases. JAMA, 295: 1127-1134.
088398
Dominici Fi Peng RDi Ebisu Ki Zeger SLi Samet JMi Bell ML. (2007). Does the effect of PM10 on mortality
depend on PM nickel and vanadium content? A reanalysis of the NMMAPS data. Environ Health
Perspect, 115: 1701-3. 099135
Dominici Fi Peng RDi Zeger SLi White RHi Samet JM. (2007). Particulate air pollution and mortality in the
United States: did the risks change from 1987 to 2000?. Am J Epidemiol, 166: 880-8. 097361
Dominici Fi Zanobetti Ai Zeger SLi Schwartz Ji Samet JM. (2004). Hierarchical bivariate time series models: a
combined analysis of the effects of particulate matter on morbidity and mortality. , 5- 341-60. 096951
Downs SHi Schindler C; Liu L-JSi Keidel Di Bayer-Oglesby Li Brutsche MHi Gerbase MWi Keller Ri Kunzli Ni
Leuenberger Pi Probst-Hensch NMi Tschopp J-Mi Zellweger J-Pi Rochat Ti Schwartz Ji Ackermann-
Liebrich U>. (2007). Reduced exposure to PM10 and attenuated age-related decline in lung function. , 15:
185-204. 092853
Dubowsky SDi Suh Hi Schwartz Ji Coull BA; Gold DR. (2006). Diabetes, obesity, and hypertension may enhance
associations between air pollution and markers of systemic inflammation. Environ Health Perspect, 114:
992-998. 088750
Dugandzic RDodds L; Stieb D; Smith-Doiron M. (2006). The association between low level exposures to ambient
air pollution and term low birth weight: a retrospective cohort study. , 5- 3. 088681
Ebelt ST; Wilson WE; Brauer M. (2005). Exposure to ambient and nonambient components of particulate
matter: a comparison of health effects. Epidemiology, 16: 396-405. 056907
Eftim SE; Samet JM; Janes H; McDermott A; Dominici F. (2008). Fine particulate matter and mortality: a
comparison of the six cities and American Cancer Society cohorts with a medicare cohort. Epidemiology,
19: 209-216. 099104
El-Zein A; Nuwayhid I; El-Fadel M; Mroueh S. (2007). Did a ban on diesel-fuel reduce emergency respiratory
admissions for children?. Sci Total Environ, 384: 134-140. 093043
Enstrom JE. (2005). Fine particulate air pollution and total mortality among elderly Californians, 1973-2002.
Inhal Toxicol, 17: 803-816. 087356
Erbas B; Kelly A"M; Physick B; Code C; Edwards M. (2005). Air pollution and childhood asthma emergency
hospital admissions: estimating intra-city regional variations. Int J Environ Health Res, 15: 11-20.
073849
Fan Z; Meng Q; Weisel C; Laumbach R; Ohman-Strickland P; Shalat S; Hernandez M; Black K. (2008). Acute
exposure to elevated PM (2.5) generated by traffic and cardiopulmonary health effects in healthy older
adults. J Expo Sci Environ Epidemiol, 19: 525-533. 191979
Farhat SCL; Paulo RLP; Shimoda TM; Conceicao GMS; Lin CA; Braga ALF; Warth MPN; Saldiva PHN. (2005).
Effect of air pollution on pediatric respiratory emergency room visits and hospital admissions. Braz J
Med Biol Res, 38: 227-235. 089461
July 2009
E-683
DRAFT-DO NOT CITE OR QUOTE

-------
Ferdinands JM; Crawford CA; Greenwald R; Van Sickle D; Hunter E; Teague WG. (2008). Breath acidification
in adolescent runners exposed to atmospheric pollution- a prospective, repeated measures observational
study. Environ Health, 7- 10. 156433
Filleul L; Rondeau V; Vandentorren S; Le Moual N> Cantagrel A; Annesi-Maesano I; Charpin D; Declercq Ci
Neukirch F> Paris Ci Vervloet Di Brochard Pi Tessier JFi Kauffmann Fi Baldi I. (2005). Twenty five year
mortality and air pollution - results from the French PAARC survey. Occup Environ Med, 62- 453-460.
087357
Fischer Pi Hoek Gi Brunekreef Bi Verhoeff Ai van Wijnen J. (2003). Air pollution and mortality in the
Netherlands' are the elderly more at risk?. Eur Respir J, 40- 34S-38S. 043739
Fischer PHi Brunekreef Bi Lebret E. (2004). Air pollution related deaths during the 2003 heat wave in the
Netherlands. Atmos Environ, 38- 1083-1085. 055605
Fischer SLi Koshland CP. (2007). Daily and peak 1 h indoor air pollution and driving factors in a rural Chinese
village. Environ Sci Technol, 41- 3121-3126. 156435
Folino AFi Scapellato ML; Canova Ci Maestrelli Pi Bertorelli Gi Simonato Li Iliceto Si Lotti M. (2009).
Individual exposure to particulate matter and the short-term arrhythmic and autonomic profiles in
patients with myocardial infarction. Eur Heart J, 30: 1614-1620. 191902
Forastiere Fi Stafoggia Mi Berti Gi Bisanti Li Cernigliaro Ai Chiusolo Mi Mallone Si Miglio Ri Pandolfi Pi
Rognoni Mi Serinelli Mi Tessari Ri Vigotti Mi Perucci C. (2008). Particulate Matter and Daily Mortality:
A Case-Crossover Analysis of Individual Effect Modifiers. Epidemiology, 19: 571-580. 186937
Forastiere Fi Stafoggia Mi Picciotto Si Bellander Ti D'Ippoliti Di Lanki Ti Von Klot Si Nyberg Fi Paatero Pi
Peters Ai Pekkanen Ji Sunyer Ji Perucci CA. (2005). A case-crossover analysis of out-of-hospital coronary
deaths and air pollution in Rome, Italy. Am J Respir Crit Care Med, 172: 1549-1555. 086323
Forastiere Fi Stafoggia Mi Tasco Ci Picciotto Si Agabiti Ni Cesaroni Gi Perucci CA. (2007). Socioeconomic status,
particulate air pollution, and daily mortality: differential exposure or differential susceptibility. Am J Ind
Med, 50: 208-216. 090720
Forbes LJi Patel MDi Rudnicka ARi Cook DGi Bush Ti Stedman JRi Whincup PHi Strachan DPi Anderson RH.
(2009). Chronic exposure to outdoor air pollution and markers of systemic inflammation. Epidemiology,
20: 245-253. 190351
Forsberg Bi Segerstedt Bi Stjernberg Ni Roemer W. (1998). Air pollution and respiratory health of children: the
PEACE panel study in Umea, Sweden. , 8: 12-19. 051714
Franklin Mi Koutrakis Pi Schwartz J. (2008). PM25 composition & daily mortality in 25 US communities.
097426
Franklin Mi Zeka Ai Schwartz J. (2007). Association between PM2.5 and all-cause and specific-cause mortality
in 27 US communities. J Expo Sci Environ Epidemiol, 17: 279-287. 091257
Fuentes Mi Song HRi Ghosh SKi Holland DMi Davis JM. (2006). Spatial association between speciated fine
particles and mortality. Biometrics, 62: 855-63. 097647
Fung KYi Khan Si Krewski Di Chen Y. (2006). Association between air pollution and multiple respiratory
hospitalizations among the elderly in Vancouver, Canada. Inhal Toxicol, 18: 1005-1011. 089789
Fung KYi Luginaah IKMGi Webster G. (2005). Air pollution and daily hospitalization rates for cardiovascular
and respiratory diseases in London, Ontario. Int J Environ Stud, 62: 677-685. 093262
Galan L Tobias Ai Banegas JRi Aranguez E. (2003). Short-term effects of air pollution on daily asthma
emergency room admissions. Eur Respir J, 22: 802-808. 087408
Gauderman WJi Avol Ei Gilliland Fi Vora Hi Thomas Di Berhane Ki McConnell Ri Kuenzli Ni Lurmann Fi
Rappaport Ei Margolis Hi Bates Di Peters J. (2004). The effect of air pollution on lung development from
10 to 18 years of age. N Engl J Med, 351: 1057-1067. 056569
Gauderman WJi Gilliland GFi Vora Hi Avol Ei Stram Di McConnell Ri Thomas Di Lurmann Fi Margolis HGi
Rappaport EBi Berhane Ki Peters JM. (2002). Association between air pollution and lung function
growth in southern California children: results from a second cohort. Am J Respir Crit Care Med, 166:
76-84. 026013
Gauderman WJi McConnell Ri Gilliland Fi London Si Thomas Di Avol Ei Vora Hi Berhane Ki Rappaport EBi
Lurmann Fi Margolis HGi Peters J. (2000). Association between air pollution and lung function growth in
southern California children. Am J Respir Crit Care Med, 162: 1383-1390. 012531
Gauderman WJi Vora Hi McConnell Ri Berhane Ki Gilliland Fi Thomas Di Lurmann Fi Avol Ei Kunzli N.
(2007). Effect of exposure to traffic on lung development from 10 to 18 years of age: a cohort study.
Lancet, 369: 571-577. 090121
Gehring Ui Cyrys Ji Sedlmeir Gi Brunekreef Bi Bellander Ti Fischer Pi Bauer CP; Reinhardt Di Wichmann HE;
Heinrich J. (2002). Traffic-related air pollution and respiratory health during the first 2 yrs of life. Eur
Respir J, 19: 690-698. 036250
July 2009
E-684
DRAFT-DO NOT CITE OR QUOTE

-------
Gent JF; Koutrakis P; Belanger K; Triche E; Holford TR; Bracken MB; Leaderer BP. (2009). Symptoms and
medication use in children with asthma and traffic-related sources of fine particle pollution. Environ
Health Perspect, In Press: 1-41. 180399
Gent JF; Triche EW; Holford TR; Belanger K; Bracken MB; Beckett WS; Leaderer BP. (2003). Association of
lowlevel ozone and fine particles with respiratory symptoms in children with asthma. JAMA, 290: 1859"
1867. 052885
Gilboa SM; Mendola P; Olshan AF; Langlois PH; Savitz DA; Loomis D; Herring AH; Fixler DE. (2005). Relation
between ambient air quality and selected birth defects, seven county study, Texas, 1997-2000. Am J
Epidemiol, 162: 238-252. 087892
Girardot SP; Ryan PB; Smith SM; Davis WT; Hamilton CB; Obenour RA; Renfro JR; Tromatore KA; Reed GD.
(2006). Ozone and PM25 exposure and acute pulmonary health effects: a study of hikers in the Great
Smoky Mountains National Park. Environ Health Perspect, 113: 612-617. 088271
Goldberg MS; Burnett RT; Valois M"F; Flegel K; Bailar JC III; Brook J; Vincent R; Radon K. (2003).
Associations between ambient air pollution and daily mortality among persons with congestive heart
failure. Environ Res, 91: 8-20. 035202
Goldberg MS; Giannetti N; Burnett RT; Mayo NE; Valois MF; Brophy JM. (2008). A panel study in congestive
heart failure to estimate the short-term effects from personal factors and environmental conditions on
oxygen saturation and pulse rate. Occup Environ Med, 65: 659-666. 180380
Goncalves FLT; Carvalho LMV Conde FC; Latorre MRDO; Saldiva PHN; Braga ALF. (2005). The effects of air
pollution and meteorological parameters on respiratory morbidity during the summer in Sao Paulo City.
Environ Int, 31: 343-349. 089884
Gordian ME; Choudhury AH. (2003). PM10 and asthma medication in schoolchildren. Arch Environ Occup
Health, 58: 42-47. 054842
Goss CH; Newsom SA; Schildcrout JS; Sheppard L; Kaufman JD. (2004). Effect of ambient air pollution on
pulmonary exacerbations and lung function in cystic fibrosis. Am J Respir Crit Care Med, 169: 816-821.
055624
Gouveia N> Bremner SA; Novaes HMD. (2004). Association between ambient air pollution and birth weight in
Sao Paulo, Brazil. J Epidemiol Community Health, 58: 11-17. 055613
Ha E-H; Lee J-T; Kim H; Hong Y-C; Lee. (2003). Infant susceptibility of mortality to air pollution in Seoul,
South Korea. , 111: 284-290. 042552
Hajat S; Anderson HR; Atkinson RW! Haines A. (2002). Effects of air pollution on general practitioner
consultations for upper respiratory diseases in London. Occup Environ Med, 59: 294-299. 030358
Halonen JI; Lanki T; Yli-Tuomi T; Kulmala M; Tiittanen P; Pekkanen J. (2008). Urban air pollution, and
asthma and COPD hospital emergency room visits. Thorax, 63: 635-41. 189507
Halonen JI; Lanki T; Yli-Tuomi T; Tiittanen P; Kulmala M; Pekkanen. (2009). Particulate air pollution and
acute cardiorespiratory hospital admissions and mortality among the elderly. Epidemiology, 20: 143-153.
180379
Hanigan IC; Johnston FH; Morgan GG. (2008). Vegetation fire smoke, indigenous status and cardio respiratory
hospital admissions in Darwin, Australia, 1996-2005: a time-series study. , T- 42. 156518
Hansen C; Neller A; Williams G; Simpson R. (2006). Maternal exposure to low levels of ambient air pollution
and preterm birth in Brisbane, Australia. BJOG, 113: 935-941. 089818
Hansen C; Neller A; Williams G; Simpson R. (2007). Low levels of ambient air pollution during pregnancy and
fetal growth among term neonates in Brisbane, Australia. Environ Res, 103: 383-389. 090703
Hansen CA; Barnett AG; Jalaludin B; Morgan G. (2009). Ambient air pollution and birth defects in brisbane,
Australia. , 4- e5408. 192362
Hapcioglu B; Issever H; Kocyigit E; Disci R; Vatansever S; Ozdilli K. (2006). The effect of air pollution and
meteorological parameters on chronic obstructive pulmonary disease at an Istanbul hospital. Indoor Built
Environ, 15: 147-153. 093263
Harre ESM; Price PD; Ayrey RB; Toop LJ; Martin IR; Town GI. (1997). Respiratory effects of air pollution in
chronic obstructive pulmonary disease: a three month prospective study. Thorax, 52: 1040-1044. 095726
Hastings DL; Jardine S. (2002). The relationship between air particulate levels and upper respiratory disease in
soldiers deployed to Bosnia (1997-1998). Mil Med, 167: 296-303. 030344
Henrotin JB; Besancenot JP; Bejot Y; Giroud M. (2007). Short-term effects of ozone air pollution on ischaemic
stroke occurrence: a case-crossover analysis from a 10-year population-based study in Dijon, France.
Occup Environ Med, 64: 439-445. 093270
Hertz-Picciotto I; Baker RJ; Yap PS; Dostal M; Joad JP; Lipsett M; Greenfield T; Herr CE; Benes I; Shumway
RH; Pinkerton KE; Sram R. (2007). Early childhood lower respiratory illness and air pollution, [see
comment]. Environ Health Perspect, 115: 1510-8. 135917
July 2009
E-685
DRAFT-DO NOT CITE OR QUOTE

-------
Hertz-Picciotto I; Herr CE; Yap PS; Dostal Mi Shumway RHi Ashwood Pi Lipsett Mi Joad JPi Pinkerton KEi
Sram RJ. (2005). Air pollution and lymphocyte phenotype proportions in cord blood. Environ Health
Perspect, 113: 1391-1398. 088678
Hinwood ALi De Klerk Ni Rodriguez Ci Jacoby Pi Runnion Ti Rye Pi Landau Li Murray Fi Feldwick Mi Spickett
J. (2006). The relationship between changes in daily air pollution and hospitalizations in Perth, Australia
1992-1998: a case-crossover study. Int J Environ Health Res, 16: 27"46. 088976
Hirshon JMi Shardell Mi Alles Si Powell JLi Squibb Ki Ondov Ji Blaisdell CJ. (2008). Elevated ambient air zinc
increases pediatric asthma morbidity. Environ Health Perspect, 116: 826-831. 180375
Ho W-Ci Hartley WRi Myers Li Lin M-Hi. (2007). Air pollution, weather, and associated risk factors related to
asthma prevalence and attack rate. Environ Res, 104: 402-409. 093265
Hoffmann Bi Moebus Si Kroger Ki Stang Ai Mohlenkamp Si Dragano Ni Schmermund Ai Memmesheimer Mi
Erbel Ri Jockel K"H. (2009). Residential exposure to urban pollution, ankle "brachial index, and
peripheral arterial disease. Epidemiology, 20: 280-288. 190376
Hoffmann Bi Moebus Si Mohlenkamp Si Stang Ai Lehmann Ni Dragano Ni Schmermund Ai Memmesheimer Mi
Mann Ki Erbel Ri Jockel K"Hi Heinz Nixdorf Recall Study Investigative Group. (2007). Residential
exposure to traffic is associated with coronary atherosclerosis. Circulation, 116: 489-496. 091163
Hoffmann Bi Moebus Si Stang Ai Beck E"Mi Dragano Ni Mohlenkamp Si Schmermund Ai Memmesheimer Mi
Mann Ki Erbel Ri Jockel K"Hi Heinz Nixdorf RECALL Study Investigative Group. (2006). Residence
close to high traffic and prevalence of coronary heart disease. Eur Heart J, 27: 2696-2702. 091162
Hogervorst JGi de Kok TMi Briede JJi Wesseling Gi Kleinjans JCi van Schayck CP. (2006). Relationship
between radical generation by urban ambient particulate matter and pulmonary function of school
children. , 69: 245-62. 189460
Hogervorst JGi de Kok TMi Briede JJi Wesseling Gi Kleinjans JCi van Schayck CP. (2006). Relationship
between radical generation by urban ambient particulate matter and pulmonary function of school
children. J Toxicol Environ Health A, 69: 245-262. 156559
Holguin Fi Flores Si Ross Zi Cortez Mi Molina Mi Molina Li Rincon Ci Jerrett Mi Berhane Ki Granados Ai
Romieu I. (2007). Traffic-related exposures, airway function, inflammation, and respiratory symptoms in
children. Am J Respir Crit Care Med, 176: 1236-42. 099000
Holloman CHi Bortnick SMi Morara Mi Strauss WJi Calder CA. (2004). A Bayesian hierarchical approach for
relating PM2.5 exposure to cardiovascular mortliaty in North Carolina. Environ Health Perspect, 112:
1282-1288. 087375
Hong CYi Chia SE; Widjaja D; Saw SM; Lee J; Munoz C; Koh D. (2004). Prevalence of Respiratory Symptoms in
Children and Air Quality by Village in Rural Indonesia. J Occup Environ Med, 46: 1174. 156565
Hong Y-C; Hwang S-S; Kim JH; Lee K-H; Lee H-J; Lee K-H; Yu S-D; Kim D-S. (2007). Metals in particulate
pollutants affect peak expiratory flow of schoolchildren. Environ Health Perspect, 115: 430-434. 091347
Hopke PE Ito E Mar T; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Liu H;
Neas L; Pinto J; Stolzel M; Suh H; Paatero P; Thurston GD. (2006). PM source apportionment and health
effects: 1 Intercomparison of source apportionment results. J Expo Sci Environ Epidemiol, 16: 275-286.
088390
Horak F Jri Studnicka M; Gartner C; Spengler JD; Tauber E; Urbanek R; Veiter A; Frischer T. (2002).
Particulate matter and lung function growth in children: a 3-yr follow-up study in Austrian
schoolchildren. Eur Respir J, 19: 838-845. 034792
Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V;
Prouvost H; Lefranc A. (2007). Short-term Associations between Fine and Coarse Particles and
Cardiorespiratory Hospitalizations in Six French Cities. Occup Environ Med. 155851
Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V;
Prouvost H; Lefranc A. (2008). Short-term associations between fine and coarse particles and hospital
admissions for cardiorespiratory diseases in six French cities. Occup Environ Med, 65: 544-551. 155852
Hwang B-F; Jaakkola JJK; Lee Y"L; Lin Y"C; Guo Y-LL. (2006). Relation between air pollution and allergic
rhinitis in Taiwanese schoolchildren. Respir Res, T- 23. 088971
Hwang J-S; Chan C"C. (2002). Effects of air pollution on daily clinic visits for lower respiratory tract illness. Am
J Epidemiol, 155: 1-10. 023222
Hwang K-W; Lee J-H; Jeong D"Y; Lee C"H; Bhatnagar A; Park J-M; Kim S"H. (2008). Observation of difference
in the size distribution of carbon and major inorganic compounds of atmospheric aerosols after the long-
range transport between the selected days of winter and summer. Atmos Environ, 42: 1057-1063. 134420
Ibald-Mulli A; Timonen KL; Peters A; Heinrich J; Wolke G; Lanki T; Buzorius G; Kreyling WG; De Hartog J;
Hoek G; Ten Brink HM; Pekkanen J. (2004). Effects of particulate air pollution on blood pressure and
heart rate in subjects with cardiovascular disease: a multicenter approach. Environ Health Perspect, 112:
369-377. 087415
July 2009
E-686
DRAFT-DO NOT CITE OR QUOTE

-------
Ingle ST; Pachpande BG; Wagh ND; Patel VS; Attarde SB. (2005). Exposure to vehicular pollution and
respiratory impairment of traffic policemen in Jalgaon City, India. , 43- 656-662. 089014
Islam T; Gauderman WJ; Berhane K; McConnell R; Avol E; Peters JM; Gilliland FD. (2007). The relationship
between air pollution, lung function and asthma in adolescents. Thorax, 62- 957-963. 090697
Issever H; Disci R; Hapcioglu B; Vatansever S> Karan M A; Akkaya V; Erk O. (2005). The effect of air pollution
and meteorological parameters in Istanbul on hospital admissions for acute coronary syndrome. Indoor
Built Environ, 14: 157-164. 097736
Ito E Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Neas L; Hopke PK;
Thurston GD. (2006). PM source apportionment and health effects: 2 An investigation of intermethod
variability in associations between source-apportioned fine particle mass and daily mortality in
Washington, DC. J Expo Sci Environ Epidemiol, 16: 300-310. 088391
Jaffe DH; Singer ME; Rimm AA. (2003). Air pollution and emergency department visits for asthma among Ohio
Medicaid recipients, 1991-1996. Environ Res, 91: 21-28. 041957
Jalaludin B; Mannes T; Morgan G; Lincoln D; Sheppeard V; Corbett S. (2007). Impact of ambient air pollution
on gestational age is modified by season in Sydney, Australia. , 6: 16. 156601
Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Simpson R; Corbett S. (2006). Associations between ambient
air pollution and daily emergency department attendances for cardiovascular disease in the elderly (65+
years), Sydney, Australia. , 16: 225-37. 189416
Jalaludin BB; O'Toole BI; Leeder SR. (2004). Acute effects of urban ambient air pollution on respiratory
symptoms, asthma medication use, and doctor visits for asthma in a cohort of Australian children.
Environ Res, 95: 32-42. 056595
Janes H; Dominici F; Zeger SL. (2007). Trends in air pollution and mortality: an approach to the assessment of
unmeasured confounding. Epidemiology, 18: 416-23. 090927
Jansen KL; Larson TV; Koenig JQ; Mar TF; Fields C; Stewart J; Lippmann M. (2005). Associations between
health effects and particulate matter and black carbon in subjects with respiratory disease. Environ
Health Perspect, 113: 1741-1746. 082236
Janssen NAH; Brunekreef B; van Vliet P; Aarts F; Maliefste K; Harssema H; Fischer P. (2003). The relationship
between air pollution from heavy traffic and allergic sensitization, bronchial byperresponsiveness, and
respiratory symptoms in Dutch schoolchildren. Epidemiology, 111: 1512-1518. 133555
Jedrychowski W; Masters E; Choi H; Sochacka E; Flak E; Mroz E; Pac A; Jacek R; Kaim I; Skolicki Z; Spengler
JD; Perera F. (2007). Pre-pregnancy dietary vitamin A intake may alleviate the adverse birth outcomes
associated with prenatal pollutant exposure: epidemiologic cohort study in Poland. Int J Occup Environ
Health, 13: 175-180. 156607
Jerrett M; Burnett RT; Ma R; Pope CA III! Krewski D; Newbold KB; Thurston G; Shi Y; Finkelstein N; Calle
EE; Thun MJ. (2005). Spatial analysis of air pollution and mortality in Los Angeles. Epidemiology, 16:
727-736. 087600
Jerrett M; Burnett RT; Willis A; Krewski D; Goldberg MS; DeLuca P; Finkelstein N. (2003). Spatial analysis of
the air pollution-mortality relationship in the context of ecologic confounders. J Toxicol Environ Health
A, 66: 1735-1777. 087380
Johnston FH; Bailie RS; Pilotto LS; Hanigan IC. (2007). Ambient biomass smoke and cardio-respiratory hospital
admissions in Darwin, Australia. , !'¦ 240. 155882
Johnston FH; Webby RJ; Pilotto LS; Bailie RS; Parry DL; Halpin SJ. (2006). Vegetation fires, particulate air
pollution and asthma: a panel study in the Australian monsoon tropics. Int J Environ Health Res, 16:
391-404. 091386
Just J; Segala C; Sahraoui F; Priol G; Grimfeld A; Neukirch F. (2002). Short-term health effects of particulate
and photochemical air pollution in asthmatic children. Eur Respir J, 20: 899-906. 035429
Kaiser R; Romieu I; Medina S; Schwartz J; Krzyzanowski M; Kunzli N. (2004). Air pollution attributable
postneonatal infant mortality in US metropolitan areas: a risk assessment study. Retrieved , from .
076674
Kan H-D; Chen B"H. (2003). Air pollution and daily mortality in Shanghai: a time series study. Arch Environ
Occup Health, 58: 360-367. 087372
Kan H-D; Chen B"H; Fu C-W; Yu S"Z; Mu L-N. (2005). Relationship between ambient air pollution and daily
mortality of SARS in Beijing. Biomed Environ Sci, 18: 1-4. 087561
Kan H; Heiss G; Rose KM; Whitsel E; Lurmann F; London SJ. (2007). Traffic exposure and lung function in
adults: the Atherosclerosis Risk in Communities study. Thorax, 62: 873-879. 091383
Kan H; London SJ; Chen G; Zhang Y; Song G; Zhao N; Jiang L; Chen B. (2007). Differentiating the effects of
fine and coarse particles on daily mortality in Shanghai, China. Environ Int, 33: 376-384. 091267
July 2009
E-687
DRAFT-DO NOT CITE OR QUOTE

-------
Kan H; London SJ; Chen G> Zhang Yi Song G> Zhao N> Jiang L> Chen B. (2008). Season, sex, age, and education
as modifiers of the effects of outdoor air pollution on daily mortality in Shanghai, China- The Public
Health and Air Pollution in Asia (PAPA) Study. Environ Health Perspect, 116: 1183-1188. 156621
Karr Ci Lumley Ti Schreuder A; Davis R; Larson Ti Ritz B> Kaufman J. (2007). Effects of subchronic and chronic
exposure to ambient air pollutants on infant bronchiolitis. Am J Epidemiol, 165: 553-560. 090719
Kasamatsu J; Shima M> Yamazaki Si Tamura Ki Sun G. (2006). Effects of winter air pollution on pulmonary
function of school children in Shenyang, China. Int J Hyg Environ Health, 209: 435-444. 156627
Kaufman Y. (1987). Satellite sensing of aerosol absorption. J Geophys Res, 92: 4307-4317. 190960
Keatinge WRi Donaldson GC. (2006). Heat acclimatization and sunshine cause false indications of mortality due
to ozone. Environ Res, 100: 387-393. 087536
Kettunen Ji Lanki Ti Tiittanen Pi Aalto PPi Koskentalo Ti Kulmala Mi Salomaa Vi Pekkanen J. (2007).
Associations of fine and ultrafine particulate air pollution with stroke mortality in an area of low air
pollution levels. Stroke, 38: 918-922. 091242
Kim CGi Bell JNBi Power SA. (2003). Effects of soil cadmium on Pinus sylvestris L. seedlings. Plant Soil, 257:
443-449. 155899
Kim Hi Lee J-T; Hong Y-C; Yi S-MJ Kim Y. (2004). Evaluating the effect of daily PM10 variation on mortality.
Inhal Toxicol, l: 55-58. 087417
Kim JH; Lim DH; Kim JK; Jeong SJ; Son BK. (2005). Effects of particulate matter (PM10) on the pulmonary
function of middle-school children. J Korean Med Sci, 20: 42-45. 087418
Kim JJ; Smorodinsky Si Lipsett Mi Singer BCi Hodgson ATi Ostro B. (2004). Traffic-related air pollution near
busy roads: the East Bay children's Respiratory Health Study. Am J Respir Crit Care Med, 170: 520-526.
087383
Kim OJi Ha EH; Kim BM; Park HS; Jung WJ; Lee BE; Suh YJ; Kim YJ; Lee JT; Kim H; Hong YC. (2007). PM10
and pregnancy outcomes: a hospital-based cohort study of pregnant women in Seoul. J Occup Environ
Med, 49: 1394-1402. 156642
Kim SY; O'Neill MS; Lee JT; Cho YJ Kim J; Kim H. (2007). Air pollution, socioeconomic position, and emergency
hospital visits for asthma in Seoul, Korea. Int Arch Occup Environ Health, 80: 701-710. 092837
KLemm RJ; Lipfert FW; Wyzga RE; Gust C. (2004). Daily mortality and air pollution in Atlanta: two years of
data from ARIES. Inhal Toxicol, 16 Suppl l: 131-141. 056585
Ko FWS; Tam W; Wong TW; Chan DPS. (2007). Temporal relationship between air pollutants and hospital
admissions for chronic obstructive pulmonary disease in Hong Kong. Thorax, 62: 780-785. 091639
Ko FWS; Tam W; Wong TW; Lai CKW;. (2007). Effects of air pollution on asthma hospitalization rates in
different age groups in Hong Kong. Clin Exp Allergy, 37: 1312-1319. 092844
Koenig JQ; Jansen K; Mar TF; Lumley T; Kaufman J; Trenga CA; Sullivan J; Liu LJ; Shapiro GG; Larson TV.
(2003). Measurement of offline exhaled nitric oxide in a study of community exposure to air pollution.
Environ Health Perspect, 111: 1625-1629. 156653
Koken PJM; Piver WT; Ye F; Elixhauser A; Olsen LM; Portier CJ. (2003). Temperature, air pollution, and
hospitalization for cardiovascular diseases among elderly people in Denver. Environ Health Perspect,
111: 1312-1317. 049466
Kongtip P; Thongsuk WJ Yoosook W; Chantanakul S. (2006). Health effects of metropolitan traffic-related air
pollutants on street vendors. Atmos Environ, 40: 7138-7145. 096920
Krewski D; Jerrett M; Burnett RT; Ma R; Hughes E; Shi YJ Turner MC; Pope AC III; Thurston G; Calle EE;
Thun MJ. (2009). Extended follow-up and spatial analysis of the American Cancer Society study linking
particulate air pollution and mortality. Health Effects Institute. Cambridge, MA. 140. 191193
Kulkarni N; Pierse N; Rushton L; Grigg J. (2006). Carbon in airway macrophages and lung function in children.
, 355: 21-30. 089257
Kumar R; Sharma M; Srivastva A; Thakur JS; Jindal SK; Parwana HK. (2004). Association of outdoor air
pollution with chronic respiratory morbidity in an industrial town in northern India. Arch Environ Occup
Health, 59: 471-477. 089873
Kunzli N; Jerrett M; Mack WJ; Beckerman B; LaBree L; Gilliland F; Thomas D; Peters J; Hodis HN. (2005).
Ambient air pollution and atherosclerosis in Los Angeles. Environ Health Perspect, 113: 201-206. 087387
Kuo HW; Lai JS; Lee MC; Tai RC; Lee MC. (2002). Respiratory effects of air pollutants among asthmatics in
central Taiwan. Arch Environ Occup Health, 57: 194-200. 036310
Laden F; Schwartz J; Speizer FE; Dockery DW. (2006). Reduction in fine particulate air pollution and mortality:
extended follow-up of the Harvard Six Cities study. Am J Respir Crit Care Med, 173: 667-672. 087605
Lagorio S; Forastiere F; Pistelli R; Iavarone I; Michelozzi P; Fano V; Marconi A; Ziemacki G; Ostro BD. (2006).
Air pollution and lung function among susceptible adult subjects: a panel study. ,5' 11. 089800
Langley-Turnbaugh SJ; Gordon NR; Lambert T. (2005). Airborne particulates and asthma: a Maine case study.
Toxicol Ind Health, 21: 75-92. 093269
July 2009
E-688
DRAFT-DO NOT CITE OR QUOTE

-------
Langrish JP; Mills NL; Chan JK; Leseman DL; Aitken RJ; Fokkens PH; Cassee FR; Li J; Donaldson Ki Newby
DE; Jiang L. (2009). Beneficial cardiovascular effects of reducing exposure to particulate air pollution
with a simple facemask. Part Fibre Toxicol, 6- 8. 191908
Lanki Ti De Hartog JJ; Heinrich J; Hoek Gi Janssen NAH; Peters A; Stolzel Mi Timonen KLi Vallius Mi
Vanninen Ei Pekkanen J. (2006). Can we identify sources of fine particles responsible for exercise-
induced ischemia on days with elevated air pollution? The ULTRA study. Environ Health Perspect, 114:
655-660. 088412
Lanki Ti Hoek Gi Timonen Ki Peters Ai Tiittanen Pi Vanninen Ei Pekkanen J. (2008). Hourly variation in fine
particle exposure is associated with transiently increased risk of ST segment depression. Br Med J, 65-
782. 191984
Lanki Ti Pekkanen Ji Aalto Pi Elosua Ri Berglind Ni D'Ippoliti Di Kulmala Mi Nyberg Fi Peters Ai Picciotto Si
Salomaa Vi Sunyer Ji Tiittanen Pi Von KLot Si Forastiere Fi for the HEAPSS Study Group. (2006).
Associations of traffic-related air pollutants with hospitalisation for first acute myocardial infarction- the
HEAPSS study. Occup Environ Med, 63- 844-851. 089788
Larrieu Si Jusot J-Fi Blanchard Mi Prouvost Hi Declercq Ci Fabre Pi Pascal Li Le Tertre Ai Wagner Vi Riviere
Si Chardon Bi Borelli Di Cassadou Si Eilstein Di Le franc A. (2007). Short term effects of air pollution on
hospitalizations for cardiovascular diseases in eight French cities' The PSAS program. Sci Total Environ,
387: 105-112. 093031
Laurent Oi Pedrono Gi Segala Ci Filleul Li Havard Si Deguen Si Schillinger Ci Riviere Ei Bard D. (2008). Air
pollution, asthma attacks, and socioeconomic deprivation: a small-area case-crossover study. Am J
Epidemiol, 168: 58-65. 156672
Le Tertre Ai Medina Si Samoli Ei Forsberg Bi Michelozzi Pi Boumghar Ai Vonk JMi Bellini Ai Atkinson Ri Ayres
JGi Sunyer Ji Schwartz Ji Katsouyanni K. (2002). Short term effects of particulate air pollution on
cardiovascular diseases in eight European cities. J Epidemiol Community Health, 56: 773-779. 023746
Le Tertre Ai Schwartz Ji Touloumi G. (2005). Empirical Bayes and adjusted estimates approach to estimating
the relation of mortality to eposure of PM10. Risk Anal, 25: 711-718. 087560
Lee BE; Ha EH; Park HS; Kim YJ; Hong YC; Kim H; Lee JT. (2003). Exposure to air pollution during different
gestational phases contributes to risks of low birth weight. Hum Reprod, 18: 638-643. 043202
Lee D; Shaddick G. (2007). Time-varying coefficient models for the analysis of air pollution and health outcome
data. Biometrics, 63: 1253-1261. 156685
Lee IM; Tsai SS; Ho CK; Chiu HF; Wu TN; Yang CY. (2008). Air pollution and hospital admissions for
congestive heart failure: are there potentially sensitive groups?. Environ Res, 108: 348-353. 192076
Lee J-T; Kim H; Song H; Hong Y-C; Cho Y-S; Shin S-Y; Hyun Y-J; Kim Y-S. (2002). Air pollution and asthma
among children in Seoul, Korea. , 13: 481-484. 034826
Lee J-T; Son J-Y; Cho Y"S. (2007). A comparison of mortality related to urban air particles between periods with
Asian dust days and without Asian dust days in Seoul, Korea, 2000-2004. Environ Res, 105: 409-13.
093042
Lee JT; Kim H; Cho YS; Hong YC; Ha EH; Park H. (2003). Air pollution and hospital admissions for ischemic
heart diseases among individuals 64+ years of age residing in Seoul, Korea. , 58: 617-623. 095552
Lee SL; Wong WHS; Lau YL. (2006). Association between air pollution and asthma admission among children in
Hong Kong. Clin Exp Allergy, 36: 1138-1146. 090176
Leem J-H; Kaplan BM; Shim YK; Pohl HR; Gotway CA; Bullard SM; Rogers JF; Smith MM; Tylenda CA. (2006).
Exposures to air pollutants during pregnancy and preterm delivery. Environ Health Perspect, 114: 905"
910. 089828
Leonardi GS; Houthuijs D; Steerenberg PA; Fletcher T; Armstrong B; Antova T; Lochman I; Lochmanova A;
Rudnai P; Erdei E; Musial J; Jazwiec-Kanyion B; Niciu EM; Durbaca S; Fabianova E; Koppova K; Lebret
E; Brunekreef B; Van Loveren H. (2000). Immune biomarkers in relation to exposure to particulate
matter: a cross-sectional survey in 17 cities of central Europe. Inhal Toxicol, 12: 1-14. 010272
Letz AG; Quinn JM. (2005). Relationship of basic military trainee emergency department visits for asthma and
San Antonio air quality. Allergy Asthma Proc, 26: 463-467. 088752
Lewis TC; Robins TG; Dvonch JT; Keeler GJ; Yip FY; Mentz GB; Lin X; Parker EA; Israel BA; Gonzalez L; Hill
Y. (2005). Air pollution-associated changes in lung function among asthmatic children in Detroit. Environ
Health Perspect, 113: 1068-1075. 081079
Liao D; Duan Y; Whitsel EA; Zheng Z-J; Heiss G; Chinchilli VM; Lin H-M. (2004). Association of higher levels of
ambient criteria pollutants with impaired cardiac autonomic control: a population-based study. Am J
Epidemiol, 159: 768-777. 056590
Liao D; Heiss G; Chinchilli VM; Duan Y; Folsom AR; Lin HM; Salomaa V. (2005). Association of criteria
pollutants with plasma hemostatic/inflammatory markers: a population-based study. J Expo Sci Environ
Epidemiol, 15: 319-328. 088677
July 2009
E-689
DRAFT-DO NOT CITE OR QUOTE

-------
Liao KJ; Tagaris E; Manomaiphiboon K; Napelenok SL; Woo JH; He S; Amar P; Russell AG. (2007).
Sensitivities of Ozone and Fine Particulate Matter Formation toEmissions under the Impact of Potential
Future Climate Change. Environ Sci Technol, 41- 8355-8361. 180272
Lichtenfels AJFC; Gomes JB; Pieri PC; Miraglia SGEK; Hallak J; Saldiva PHN. (2007). Increased levels of air
pollution and a decrease in the human and mouse male-to-female ration in Sao Paulo, Brazil. Fertil
SterH, 87: 230-232. 097041
Lin C-M; Li C-Y> Mao I-F. (2004). Increased risks of term low "birth-weight infants in a petrochemical industrial
city with high air pollution levels. Arch Environ Occup Health, 59: 663-668. 089827
Lin CA; Pereira LAA; Nishioka DC; Conceicao GMS; Graga ALF; Saldiva PHN. (2004). Air pollution and
neonatal deaths in Sao Paulo, Brazil. Braz J Med Biol Res, 37: 765-770. 095787
Lin M; Chen Y; Burnett RT; Villeneuve PJ; Krewski D. (2002). The influence of ambient coarse particulate
matter on asthma hospitalization in children: case-crossover and time-series analyses. Environ Health
Perspect, 110: 575-581. 026067
Lin M; Stieb DM; Chen Y. (2005). Coarse particulate matter and hospitalization for respiratory infections in
children younger than 15 years in Toronto: a case-crossover analysis. , 116: 235-240. 087828
Lin Y-C; Lee C-FJ Fang T. (2008). Characterization of particle size distribution from diesel engines fueled with
palm-biodiesel blends and paraffinic fuel blends. Atmos Environ, 42: 1133-1143. 126812
Linares CJ Diaz J; Tob?as A; Migue JMDei Otero A. (2006). Impact of urban air pollutants and noise levels over
daily hospital admissions in children in Madrid: a time series analysis. Int Arch Occup Environ Health,
79: 143-152. 092846
Lipfert FW; Baty JD; Miller JP; Wyzga RE. (2006). PM2.5 constituents and related air quality variables as
predictors of survival in a cohort of U.S. military veterans. , 18: 645-57. 189271
Lipfert FW; Baty JD; Miller JP; Wyzga RE. (2006). PM2.5 constituents and related air quality variables as
predictors of survival in a cohort of U.S. military veterans. Inhal Toxicol, 18: 645-657. 088756
Lipfert FW; Wyzga RE; Baty JD; Miller JP. (2006). Traffic density as a surrogate measure of environmental
exposures in studies of air pollution health effects: long-term mortality in a cohort of US veterans. Atmos
Environ, 40: 154-169. 088218
Lipfert FW; Zhang J; Wyzga RE. (2000). Infant mortality and air pollution: a comprehensive analysis of US data
for 1990. J Air Waste Manag Assoc, 50: 1350-1366. 004103
Lippmann M; Ito K; Hwang JS; Maciejczyk P; Chen LC. (2006). Cardiovascular effects of nickel in ambient air.
Environ Health Perspect, 114: 1662-9. 091165
Lipsett MJ; Tsai FC; Roger L; Woo M; Ostro BD. (2006). Coarse particles and heart rate variability among older
adults with coronary artery disease in the Coachella Valley, California. Environ Health Perspect, 114:
1215-1220. 088753
Lisabeth LD; Escobar JD; Dvonch JT; Sanchez BN; Majersik JJ; Brown DL; Smith MA; Morgenstern LB. (2008).
Ambient air pollution and risk for ischemic stroke and transient ischemic attack. , 64: 53"59. 155939
Liu CC; Chen CC; Wu TN; Yang CY. (2008). Association of brain cancer with residential exposure to
petrochemical air pollution in Taiwan. J Toxicol Environ Health A, 71: 310-314. 156708
Liu CC; Tsai SS; Chiu HF; Wu TN; Chen CC; Yang CY. (2009). Ambient exposure to criteria air pollutants and
risk of death from bladder cancer in Taiwan. , 21: 48"54. 190292
Liu L; Poon R; Chen L; Frescura AM; Montuschi P; Ciabattoni G; Wheeler A; Dales R. (2009). Acute effects of
air pollution on pulmonary function, airway inflammation, and oxidative stress in asthmatic children.
Environ Health Perspect, 117: 668-674. 192003
Liu L; Ruddy TD; Dalipaj M; Szyszkowicz M; You H; Poon R; Wheeler A; Dales R. (2007). Influence of personal
exposure to particulate air pollution on cardiovascular physiology and biomarkers of inflammation and
oxidative stress in subjects with diabetes. J Occup Environ Med, 49: 258-265. 156705
Liu S; Krewski D; Shi Y; Chen Y; Burnett R. (2007). Association between maternal exposure to ambient air
pollutants during pregnancy and fetal growth restriction. J Expo Sci Environ Epidemiol, 17: 426-432.
090429
Ljungman P; Bellander T; Schneider A; Breitner S; Forastiere F; Hampel R; Illig T; Jacquemin B; Katsouyanni
K; von Klot S. (2009). Modification of the interleukin-6 response to air pollution by interleukin-6 and
fibrinogen polymorphisms. Environ Health Perspect, 0800370: 1-31. 191983
Ljungman PLS; Berglind N; Holmgren C; Gadler F; Edvardsson N; Pershagen G; Rosenqvist M; Sjogren B;
Bellander T. (2008). Rapid effects of air pollution on ventricular arrhythmias. Eur Heart J, 29: 2894"
2901. 180266
Llorca J; Salas A; Prieto-Salceda D; Chinchon-Bengoechea V; Delgado-Rodriguez M. (2005). Nitrogen dioxide
increases cardiorespiratory admissions in Torrelavega (Spain). J Environ Health, 68: 30-35. 087825
Loomis D; Castillejos M; Gold DR; McDonnell W', Borja-Aburto VH. (1999). Air pollution and infant mortality in
Mexico City. Epidemiology, 10: 118-123. 087288
July 2009
E-690
DRAFT-DO NOT CITE OR QUOTE

-------
Lubinski W> Toczynska L Chcialowski A; Plusa T. (2005). Influence of air pollution on pulmonary function in
healthy young men from different regions of Poland. Ann Agric Environ Med, 12- 1"4. 087563
Luginaah IN; Fung KY; Gorey KM; Webster G; Wills C. (2005). Association of ambient air pollution with
respiratory hospitalization in a government designated "area of concern"' the case of Windsor, Ontario.
Environ Health Perspect, 113- 290-296. 057327
Luttmann-Gibson H; Suh HH; Coull BA; Dockery DW; Sarnet SE; Schwartz J; Stone PH; Gold DR. (2006).
Short-term effects of air pollution on heart rate variability in senior adults in Steubenville, Ohio. J Occup
Environ Med, 48: 780-788. 089794
Magas OK; Gunter JT; Regens JL. (2007). Ambient air pollution and daily pediatric hospitalizations for asthma.
Environ Sci Pollut Res Int, 14: 19-23. 090714
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005). Outdoor
air pollution and stroke in Sheffield, United Kingdom: a small-area level geographical study. , 36: 239"
243. 088683
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005). Outdoor
air pollution, mortality, and hospital admissions from coronary heart disease in Sheffield, UK: a small-
area level ecological study. Eur Heart J, 26: 2543-2549. 090769
Maisonet M; Bush TJ; Correa A; Jaakkola JJK. (2001). Relation between ambient air pollution and low birth
weight in the northeastern United States. Environ Health Perspect, 1093: 351-356. 016624
Mann JK; Tager IB; Lurmann F; Segal M; Quesenberry CP Jr; Lugg MM; Shan J; Van den Eeden SK. (2002).
Air pollution and hospital admissions for ischemic heart disease in persons with congestive heart failure
or arrhythmia. Environ Health Perspect, 110: 1247-1252. 036723
Mannes T; Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Corbett S. (2005). Impact of ambient air pollution
on birth weight in Sydney, Australia. Occup Environ Med, 62: 524-530. 087895
Mar TF; Jansen K; Shepherd K; Lumley T; Larson TV; Koenig JQ. (2005). Exhaled nitric oxide in children with
asthma and short-term PM25 exposure in Seattle. Environ Health Perspect, 113: 1791-1794. 088759
Mar TF; Koenig JQ; Jansen K; Sullivan J; Kaufman J; Trenga CA; Siahpush SH; Liu L-JS; Neas L. (2005). Fine
particulate air pollution and cardiorespiratory effects in the elderly. , 16: 681-687. 087566
Mar TF; Larson TV; Stier RA; Claiborn C; Koenig JQ. (2004). An analysis of the association between respiratory
symptoms in subjects with asthma and daily air pollution in Spokane, Washington. Inhal Toxicol, 16:
809-815. 057309
Martins LC; Latorre MRDO; Saldiva PHN; Braga ALF. (2002). Air pollution and emergency room visits due to
chronic lower respiratory diseases in the elderly: an ecological time-series study in Sao Paulo, Brazil. J
Occup Environ Med, 44: 622-627. 035059
Martins MCH; Fatigati FL; Vespoli TC; Martins LC; Martins MA; Saldiva PHN; Braga ALF. (2004). Influence of
socioeconomic conditions on air pollution effects in elderly people an analysis of six regions in Sao Paolo,
Brazil. J Epidemiol Community Health, 58: 41-46. 087457
Masjedi MR; Jamaati HR; Dokouhaki P; Ahmadzadeh Z; Taheri SA; Bigdeli M; Izadi S; Rostamian A; Aagin K;
Ghavam SM. (2003). The effects of air pollution on acute respiratory conditions. Respirology, 8: 213-230.
052100
McConnell R; Berhane K; Gilliland F; London SJ; Islam T; Gauderman WJ; Avol E; Margolis HG; Peters JM.
(2002).	Asthma in exercising children exposed to ozone: a cohort study. Lancet, 359: 386-391. 023150
McConnell R; Berhane K; Gilliland F; London SJ; Vora H; Avol E; Gauderman WJ; Margolis HG; Lurmann F;
Thomas DC; Peters JM. (1999). Air pollution and bronchitic symptoms in southern California children
with asthma. Environ Health Perspect, 107: 757-760. 007028
McConnell R; Berhane K; Gilliland F; Molitor J; Thomas D; Lurmann F; Avol E; Gauderman WJ; Peters JM.
(2003).	Prospective study of air pollution and bronchitic symptoms in children with asthma. Am J Respir
Crit Care Med, 168: 790-797. 049490
McConnell R; Berhane K; Molitor J; Gilliland F; Kunzli N; thorne PS; Thomas D; Gauderman WJ; Avol E;
Lurmann F; Rappaport E; Jerrett M; Peters JM. (2006). Dog ownership enhances symptomatic responses
to air pollution inchildren with asthma. Environ Health Perspect, 114: 1910-1915. 180226
McCreanor J; Cullinan P; Nieuwenhuijsen MJ; Stewart-Evans J; Malliarou E; Jarup L; Harrington R;
Svartengren M; Han I"K; Ohman-Strickland P; Chung KF; Zhang J. (2007). Respiratory effects of
exposure to diesel traffic in persons with asthma. , 357: 2348-2358. 092841
McDonnell WF; Nishino-Ishikawa N; Petersen FF; Chen LH; Abbey DE. (2000). Relationships of mortality with
the fine and coarse fractions of long-term ambient PM10 concentrations in nonsmokers. J Expo Sci
Environ Epidemiol, 10: 427-436. 010319
McGowan JA; Hider PN; Chacko E; Town GI. (2002). Particulate air pollution and hospital admissions in
Christchurch, New Zealand. Aust N Z J Public Health, 26: 23-29. 030325
July 2009
E-691
DRAFT-DO NOT CITE OR QUOTE

-------
Medina-Ramon Mi Zanobetti Ai Schwartz J. (2006). The effect of ozone and PM10 on hospital admissions for
pneumonia and chronic obstructive pulmonary disease- a national multicitv study. Am J Epidemiol, 163:
579-588. 087721
Meng YYi Wilhelm Mi Rull RP; English Pi Ritz B. (2007). Traffic and outdoor air pollution levels near
residences and poorly controlled asthma in adults. Ann Allergy Asthma Immunol, 98- 455-63. 093275
Metzger KB; Klein M; Flanders WD; Peel JL; Mulholland JA; Langberg JJ; Tolbert PE. (2007). Ambient air
pollution and cardiac arrhythmias in patients with implantable defibrillators. Epidemiology, 18- 585-592.
092856
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd KH; Mulholland JA; Ryan PB; Frumkin H.
(2004). Ambient air pollution and cardiovascular emergency department visits. , 15- 46"56. 044222
Michaud J-P; Grove JS; Krupitsky D. (2004). Emergency department visits and "vog"-related air quality in Hilo,
Hawaii. Environ Res, 95: 11-19. 089900
Middleton N; Yiallouros P; KLeanthous S; Kolokotroni 0; Schwartz J; Dockery DW; Demokritou P; Koutrakis P.
(2008). A 10-year time-series analysis of respiratory and cardiovascular morbidity in Nicosia, Cyprus: the
effect of short-term changes in air pollution and dust storms. , T- 39. 156760
Migliaretti G; Cadum E; Migliore E; Cavallo F. (2005). Traffic air pollution and hospital admission for asthma: a
case-control approach in a Turin (Italy) population. Int Arch Occup Environ Health, 78: 164-169. 088689
Migliaretti G; Cavallo F. (2004). Urban air pollution and asthma in children. Pediatr Pulmonol, 38: 198-203.
087425
Miller KA; Siscovick DS; Sheppard L; Shepherd K; Sullivan JH; Anderson GL; Kaufman JD. (2007). Long-term
exposure to air pollution and incidence of cardiovascular events in women. , 356: 447-458. 090130
Millstein J; Gilliland F; Berhane K; Gauderman WJ; McConnell R; Avol E; Rappaport EB; Peters JM. (2004).
Effects of ambient air pollutants on asthma medication use and wheezing among fourth-grade school
children from 12 Southern California communities enrolled in The Children's Health Study. Arch
Environ Occup Health, 59: 505-514. 088629
Min KB; Min JY; Cho SI; Paek D. (2008). The relationship between air pollutants and heart-rate variability
among community residents in Korea. Inhal Toxicol, 4- 435-444. 191901
Mohr LB; Luo S; Mathias E; Tobing R; Homan S; Sterling D. (2008). Influence of season and temperature on the
relationship of elementalcarbon air pollution to pediatric asthma emergency room visits. J Asthma, 45:
936-943. 180215
Morgenstern V; Zutavern A; Cyrys J; Brockow I; Gehring U; Koletzko S; Bauer CP; Reinhardt D; Wichmann H"
E; Heinrich J. (2007). Respiratory health and individual estimated exposure to traffic-related air
pollutants in a cohort of young children. Occup Environ Med, 64: 8"16. 090747
Mortimer K; Neugebauer R; Lurmann F; Alcorn S; Balmes J; Tager I. (2008). Early-Lifetime exposure to air
pollution and allergic sensitizationin children with asthma. J Asthma, 45: 874-881. 187280
Mortimer KM; Neas LM; Dockery DW; Redline S; Tager IB. (2002). The effect of air pollution on inner-city
children with asthma. Eur Respir J, 19: 699-705. 030281
Moshammer H; Hutter H"P; Hauck H; Neuberger M. (2006). Low levels of air pollution induce changes of lung
function in a panel of schoolchildren. Eur Respir J, 27: 1138-1143. 090771
Moshammer H; Neuberger M. (2003). The active surface of suspended particles as a predictor of lung function
and pulmonary symptoms in Austrian school children. Atmos Environ, 37: 1737-1744. 041956
Murata A; Kida K; Hasunuma H; Kanegae H; Ishimaru Y; Motegi T; Yamada K; Yoshioka H; Yamamoto K;
Kudoh S. (2007). Environmental influence on the measurement of exhaled nitric oxide concentration in
school children: special reference to methodology. , 74: 30-36. 156787
Naess O; Nafstad P; Aamodt G; Claussen B; Rosland P. (2007). Relation between concentration of air pollution
and cause-specific mortality: four-year exposures to nitrogen dioxide and particulate matter pollutants in
470 neighborhoods in Oslo, Norway. Am J Epidemiol, 165: 435-443. 090736
Nafstad P; Haheim LL; Wisloff T; Gram F; Oftedal B; Holme I; Hjermann I; Leren P. (2004). Urban air pollution
and mortality in a cohort of Norwegian men. Environ Health Perspect, 112: 610-605. 087949
Nascimento LF; Pereira LA; Braga AL; Modolo MC; Carvalho JA Jr. (2006). Effects of air pollution on children's
health in a city in southeastern Brazil. Rev Saude Publica, 40: 77-82. 093247
Nawrot TS; Torfs R; Fierens F; De Henauw S; Hoet PH; Van Kersschaever G; De Backer G; Nemery B. (2007).
Stronger associations between daily mortality and fine particulate air pollution in summer than in
winter: evidence from a heavily polluted region in western Europe. J Epidemiol Community Health, 61:
146-9. 098619
Nerriere E; Zmirou-Navier D; Desqueyroux P; Leclerc N; Momas I; Czernichow P. (2005). Lung cancer risk
assessment in relation with personal exposure to airborne particles in four French metropolitan areas. J
Occup Environ Med, 47: 1211-1217. 088630
July 2009
E-692
DRAFT-DO NOT CITE OR QUOTE

-------
Neuberger M> Schimek MG; Horak F Jr', Moshammer H; Kundi Mi Frischer Ti Gomiscek Bi Puxbaum Hi Hauck
Hi AUPHEP-Team. (2004). Acute effects of particulate matter on respiratory diseases, symptoms and
functions' epidemiological results of the Austrian Projects on Health Effects of Particulate Matter
(AUPHEP). Atmos Environ, 38: 3971-3981. 093249
O'Connor GTi Neas Li Vaughn Bi Kattan Mi Mitchell Hi Crain EFi Evans R, 3rd; Gruchalla R; Morgan W', Stout
J; Adams GK; Lippmann M. (2008). Acute respiratory health effects of air pollution on children with
asthma in US inner cities. J Allergy Clin Immunol, 121: 1133-1139 e!131. 156818
O'Neill MS; Bell ML; Ranjit N; Cifuentes LA; Loomis D; gouveia N; Borja-Aburto VH. (2008). Air pollution and
mortality in Latin America: The role of education. Epidemiology, 19: 810-819. 192314
O'Neill MS; Diez-Roux AV; Auchincloss AH; Franklin TG; Jacobs Jnr DR; Astor BC; Dvonch JT; Kaufman J.
(2007). Airborne particulate matter exposure and urinary albumin excretion: The Multi-Ethnic Study of
Atherosclerosis. Occup Environ Med, 65: 534-540. 156006
O'Neill MS; Hajat S; Zanobetti A; Ramirez-Aguilar M; Schwartz J. (2005). Impact of control for air pollution and
respiratory epidemics on the estimated associations of temperature and daily mortality. Int J
Biometeorol, 50: 121-9. 098094
O'Neill MS; Loomis D; Borja Aburto VH; Gold D; Hertz-Picciotto I; Castillejos M. (2004). Do associations
between airborne particles and daily mortality in Mexico City differ by measurement method, region, or
modeling strategy?. J Expo Sci Environ Epidemiol, 14: 429-439. 087429
O'Neill MS; Veves A; Sarnat JA; Zanobetti A; Gold DR; Economides PA; Horton ES; Schwartz J. (2007). Air
pollution and inflammation in type 2 diabetes: a mechanism for susceptibility. Occup Environ Med, 64:
373-379. 091362
O'Neill MS; Veves A; Zanobetti A; Sarnat JA; Gold DR; Economides PA; Horton ES; Schwartz J. (2005).
Diabetes enhances vulnerability to particulate air pollution-associated impairment in vascular reactivity
and endothelial function. , 111: 2913-2920. 088423
Odajima H; Yamazaki S; Nitta H. (2008). Decline in peak expiratory flow according to hourly short-term
concentration of particulate matter in asthmatic children. Inhal Toxicol, 20: 1263. 192005
Oftedal B; Brunekreef B; Nystad W', Madsen C; Walker S"E; Nafstad P. (2008). Residential outdoor air pollution
and lung function in schoolchildren. Epidemiology, 19: 129-137. 093202
Oftedal B; Nafstad P; Magnus P; Bjorkly S; Skrondal A. (2003). Traffic related air pollution and acute hospital
admission for respiratory diseases in Drammen, Norway 1995-2000. Eur J Epidemiol, 18: 671-675.
055623
Ostro B; Broadwin R; Green S; Feng W"Y; Lipsett M. (2006). Fine particulate air pollution and mortality in nine
California counties: results from CALFINE. Environ Health Perspect, 114: 29"33. 087991
Ostro B; Feng W"Y; Broadwin R; Green S; Lipsett M. (2007). The effects of components of fine particulate air
pollution on mortality in California: results from CALFINE. Environ Health Perspect, 115: 13-9. 091354
Ostro BD; Feng WY; Broadwin R; Malig BJ; Green RS; Lipsett MJ. (2008). The impact of components of fine
particulate matter on cardiovascular mortality in susceptible subpopulations. Occup Environ Med, 65:
750-756. 097971
Ozkaynak H; Thurston GD. (1987). Associations between 1980 US mortality rates and alternative measures of
airborne particle concentration. Risk Anal, T- 449-461. 072960
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Schwartz J. (2005). Effects of air pollution on heart rate
variability: the VA normative aging study. Environ Health Perspect, 113: 304-309. 057331
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Spiro A, 3rd; Tucker KL; Suh H; Hu H; Schwartz J. (2008).
Traffic-related particles are associated with elevated homocysteine: the VA normative aging study. Am J
Respir Crit Care Med, 178: 283-289. 156845
Park SK; O'Neill MS; Wright RO; Hu H; Vokonas PS; Sparrow D; Suh H; Schwartz J. (2006). HFE genotype,
particulate air pollution, and heart rate variability A gene-environmental interaction. , 114: 2798-2805.
091245
Parker JD; Akinbami LJ; Woodruff TJ. (2009). Air Pollution and Childhood Respiratory Allergies in the United
States. Environ Health Perspect, 117: 140-147. 192359
Parker JD; Woodruff TJ. (2008). Influences of study design and location on the relationship between particulate
matter air pollution and birthweight. , 22: 214-27. 189095
Parker JD; Woodruff TJ; Basu R; Schoendorf KC. (2005). Air pollution and birth weight among term infants in
California. , 115: 121-128. 087462
Peacock JL; Symonds P; Jackson P; Bremner SA; Scarlett JF; Strachan DP; Anderson HR. (2003). Acute effects
of winter air pollution on respiratory function in schoolchildren in southern England. Occup Environ
Med, 60: 82-89. 042026
July 2009
E-693
DRAFT-DO NOT CITE OR QUOTE

-------
Peel JL; Metzger KB; Klein M; Flanders WD; Mulholland JA; Tolbert PE. (2007). Ambient air pollution and
cardiovascular emergency department visits in potentially sensitive groups. Am J Epidemiol, 165- 625"
633. 090442
Peel JL; Tolbert PE; Klein M; Metzger KB; Flanders WD; Knox T; Mulholland JA; Ryan PB; Frumkin H. (2005).
Ambient air pollution and respiratory emergency department visits. , 16- 164-174. 056305
Pekkanen J; Peters A; Hoek G; Tiittanen P; Brunekreef B; de Hartog J; Heinrich J; Ibald-Mulli A; Kreyling WG;
Lanki T; Timonen KL; Vanninen E. (2002). Particulate air pollution and risk of ST-segment depression
during repeated submaximal exercise tests among subjects with coronary heart disease- the exposure and
risk assessment for fine and ultrafine particles in ambient air (ULTRA) study. , 106- 933-938. 035050
Peled R; Friger M; Bolotin A; Bibi H; Epstein L; Pilpel D; Scharf S. (2005). Fine particles and meteorological
conditions are associated with lung function in children with asthma living near two power plants. Public
Health, 119: 418-425. 156015
Penard-Morand C; Charpin D; Raherison C; Kopferschmitt C; Caillaud D; Lavaud F; Annesi-Maesano I. (2005).
Long-term exposure to background air pollution related to respiratory and allergic health in
schoolchildren. Clin Exp Allergy, 35: 1279-1287. 087951
Peng RD; Chang HH; Bell ML; McDermott A; Zeger SL; Samet JM; Dominici F. (2008). Coarse particulate
matter air pollution and hospital admissions for cardiovascular and respiratory diseases among Medicare
patients. JAMA, 299: 2172-2179. 156850
Peng RD; Dominici F; Pastor-Barriuso R; Zeger SL; Samet JM. (2005). Seasonal analyses of air pollution and
mortality in 100 US cities. Am J Epidemiol, 161: 585-594. 087463
Penttinen P; Tiittanen P; Pekkanen J. (2004). Mortality and air pollution in metropolitan Helsinki, 1988-1996.
Scand J Work Environ Health, 2: 19-27. 087432
Penttinen P; Vallius M; Tiittanen P; Ruuskanen J; Pekkanen J. (2006). Source-specific fine particles in urban
air and respiratory function among adult asthmatics. Inhal Toxicol, 18: 191-198. 087988
Pereira LAA; Loomis D; Conceicao GMS; Braga ALF; Areas RM; Kishi HS; Singer JM; Bohm GM; Saldiva PHN.
(1998). Association between air pollution and intrauterine mortality in Sao Paulo, Brazil. Environ Health
Perspect, 106: 325-329. 007264
Perez L; Tobias A; Querol X; Kunzli N> Pey J; Alastuey A; Viana M; Valero N; Gonzalez-Cabre M; Sunyer J.
(2008).	Coarse particles from Saharan dust and daily mortality. Epidemiology, 19: 800-807. 156020
Peters A; Greven S; Heid I; Baldari F; Breitner S; Bellander T; Chrysohoou C; Illig T; Jacquemin B; Koenig W.
(2009).	Fibrinogen Genes Modify the Fibrinogen Response to Ambient Particulate Matter. Am J Respir
Crit Care Med, 179: 484-491. 191992
Peters A; MONICA/KORA-Studiengruppe. (2005). Partikel in der Aussenluft erhoren das Risiko fur Herz-
Kreislauf-Erkrankungen [Aerosol particles increase the risk of cardiovascular diseases].
Gesundheitswesen, l: S79-S85. 095747
Peters A; von Klot S; Heier M; Trentinaglia I; Cyrys J; Hormann A; Hauptmann M; Wichmann HE; Lowel H.
(2005). Particulate air pollution and nonfatal cardiac events. Part I. Air pollution, personal activities, and
onset of myocardial infarction in a case-crossover study. 156859
Peters JM; Avol E; Gauderman WJ; Linn WS; Navidi W', London SJ; Margolis H; Rappaport E; Vora H; Gong H
Jr; Thomas DC. (1999). A study of twelve southern California communities with differing levels and
types of air pollution II Effects on pulmonary function. Am J Respir Crit Care Med, 159: 768-775. 087237
Pierse N> Rushton L; Harris RS; Kuehni CE; Silverman M; Grigg J. (2006). Locally-generated particulate
pollution and respiratory symptoms in young children. Thorax, 61: 216-220. 088757
Pino P; Walter T; Oyarzun M; Villegas R; Romieu I. (2004). Fine particulate matter and wheezing illnesses in
the first year of life. , 15: 702-708. 050220
Pitard A; Zeghnoun A; Courseaux A; Lamberty J; Delmas V; Fossard JL; Villet H. (2004). Short-term
associations between air pollution and respiratory drug sales. Environ Res, 95: 43"52. 087433
Pope C; Renlund D; Kfoury A; May H; Home B. (2008). Relation of heart failure hospitalization to exposure to
fine particulate air pollution. , 102: 1230-1234. 191969
Pope CA 3rd; Ezzati M; Dockery DW. (2009). Fine-particulate air pollution and life expectancy in the United
States. N Engl J Med, 360: 376-386. 190107
Pope CA III; Burnett RT. (2007). Confounding in air pollution epidemiology: the broader context. , 18: 424-426.
090928
Pope CA III; Burnett RT; Thun MJ; Calle EE; Krewski D; Ito K; Thurston GD. (2002). Lung cancer,
cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA, 287: 1132"
1141. 024689
Pope CA III; Burnett RT; Thurston GD; Thun MJ; Calle EE; Krewski D; Godleski JJ. (2004). Cardiovascular
mortality and long-term exposure to particulate air pollution: epidemiological evidence of general
pathophysiological pathways of disease. , 109: 71-77. 055880
July 2009
E-694
DRAFT-DO NOT CITE OR QUOTE

-------
Pope CA IIL Rodermund DL; Gee MM. (2007). Mortality effects of a copper smelter strike and reduced ambient
sulfate particulate matter air pollution. Environ Health Perspect, 115- 679-683. 091256
Pope CA; Hansen ML; Long RW; Nielsen KR; Eatough NL; Wilson WE; Eatough DJ. (2004). Ambient
particulate air pollution, heart rate variability, and blood markers of inflammation in a panel of elderly
subjects. Environ Health Perspect, 112- 339-345. 055238
Pope CA IIL Muhlestein JB; May HT; Renlund DG; Anderson JL; Home BD. (2006). Ischemic heart disease
events triggered by short-term exposure to fine particulate air pollution. Circulation, 114- 2443-2448.
091246
Preutthipan A; Udomsubpayakul U; Chaisupamongkollarp T; Pentamwa P. (2004). Effect of PM10 pollution in
Bangkok on children with and without asthma. Pediatr Pulmonol, 37- 187-192. 055598
Puett RC; Schwartz J; Hart JE; Yanosky JD; Speizer FE; Suh H; Paciorek CJ; Neas LM; Laden F. (2008).
Chronic particulate exposure, mortality, and coronary heart disease in the nurses' health study. Am J
Epidemiol, 168: 1161-1168. 156891
Qian Z; He Q; Lin H"M; Kong L; Liao D; Dan J; Bentley CM; Wang B. (2007). Association of daily cause-specific
mortality with ambient particle air pollution in Wuhan, China. Environ Res, 105: 380-389. 093054
Qian Z; He Q; Lin HM; Kong L; Bentley CM; Liu W; Zhou D. (2008). High temperatures enhanced acute
mortality effects of ambient particle pollution in the "oven" city of Wuhan, China. Environ Health
Perspect, 116: 1172-1178. 156894
Qian Z; Liao D; Lin H"M; Whitsel EA; Rose KM; Duan Y. (2005). Lung function and long-term exposure to air
pollutants in middle-aged American adults. Arch Environ Occup Health, 60: 156-163. 093283
Rabinovitch N; Strand M; Gelfand EW. (2006). Particulate levels are associated with early asthma worsening in
children with persistent disease. Am J Respir Crit Care Med, 173: 1098-1105. 088031
Rabinovitch N; Zhang LN; Murphy JR; Vedal S; Dutton SJ; Gelfand EW. (2004). Effects of wintertime ambient
air pollutants on asthma exacerbations in urban minority children with moderate to severe disease. J
Allergy Clin Immunol, 114: 1131-1137. 096753
Rainham DG; Smoyer-Tomic KE; Sheridan SC; Burnett RT. (2005). Synoptic weather patterns and modification
of the association between air pollution and human mortality. Int J Environ Health Res, 15: 347-360.
088676
Ranzi A; Gambini M; Spattini A; Galassi C; Sesti D; Bedeschi M; Messori A; Baroni A; Cavagni G; Lauriola P.
(2004).	Air pollution and respiratory status in asthmatic children: hints for a locally based preventive
strategy AIRE study. Eur J Epidemiol, 19: 567-576. 089500
Ren C; Tong S. (2006). Temperature modifies the health effects of particulate matter in Brisbane, Australia. Int
J Biometeorol, 51: 87-96. 092824
Rich DQ; Demissie K; Lu SE; Kamat L; Wartenberg D; Rhoads GG. (2009). Ambient air pollutant
concentrations during pregnancy and the risk of fetal growth restriction. J Epidemiol Community
Health, In Press: 1-9. 180122
Rich DQ; Freudenberger RS; Ohman-Strickland P; Cho Y; Kipen HM. (2008). Right heart pressure increases
after acute increases in ambient particulate concentration. Environ Health Perspect, 116: 1167-1171.
156910
Rich DQ; Kim MH; Turner JR; Mittleman MA; Schwartz J; Catalano PJ; Dockery DW. (2006). Association of
ventricular arrhythmias detected by implantable cardioverter defibrillator and ambient air pollutants in
the St Louis, Missouri metropolitan area. Occup Environ Med, 63: 591-596. 089814
Rich DQ; Mittleman MA; Link MS; Schwartz J; Luttmann-Gibson H; Catalano PJ; Speizer FE; Gold DR;
Dockery DW. (2006). Increased risk of paroxysmal atrial fibrillation episodes associated with acute
increases in ambient air pollution. Environ Health Perspect, 114: 120-123. 088427
Rich DQ; Schwartz J; Mittleman MA; Link M; Luttmann-Gibson H; Catalano PJ; Speizer FE; Dockery DW.
(2005).	Association of short-term ambient air pollution concentrations and ventricular arrhythmias. Am J
Epidemiol, 161: 1123-1132. 079620
Rich KE; Petkau J; Vedal S; Brauer M. (2004). A case-crossover analysis of particulate air pollution and cardiac
arrhythmia in patients with implantable cardioverter defibrillators. Inhal Toxicol, 16: 363-372. 055631
Riediker M; Devlin RB; Griggs TR; Herbst MC; Bromberg PA; Williams RW; Cascio WE. (2004). Cardiovascular
effects in patrol officers are associated with fine particulate matter from brake wear and engine
emissions. Part Fibre Toxicol, l: 2. 091261
Riojas-Rodriguez H; Escamilla-Cejudo JA; Gonzalez-Hermosillo JA; Tellez-Rojo MM; Vallejo M; Santos-Burgoa
C; Rojas-Bracho L. (2006). Personal PM2.5 and CO exposures and heart rate variability in subjects with
known ischemic heart disease in Mexico City. J Expo Sci Environ Epidemiol, 16: 131-137. 156913
Rios JLM; Boechat JL; Sant'Anna CC; Franca AT. (2004). Atmospheric pollution and the prevalence of asthma:
study among schoolchildren of 2 areas in Rio de Janeiro, Brazil. Ann Allergy Asthma Immunol, 92: 629"
634. 087800
July 2009
E-695
DRAFT-DO NOT CITE OR QUOTE

-------
Ritz Bi Wilhelm Mi Hoggatt KJi Ghosh JK. (2007). Ambient air pollution and preterm birth in the environment
and pregnancy outcomes study at the University of California, Los Angeles. Am J Epidemiol, 166- 1045"
52.096146
Ritz Bi Wilhelm Mi Zhao Y. (2006). Air pollution and infant death in southern California, 1989-2000. , 118: 493-
502. 089819
Ritz Bi Yu Fi Chapa Gi Fruin S. (2000). Effect of air pollution on preterm birth among children born in Southern
California between 1989 and 1993. , 11: 502-511. 012068
Ritz B> Yu Fi Fruin Si Chapa Gi Shaw GMi Harris JA. (2002). Ambient air pollution and risk of birth defects in
Southern California. Am J Epidemiol, 155: 17-25. 023227
Roberts S. (2004). Interactions between particulate air pollution and temperature in air pollution mortality time
series studies. Environ Res, 96: 328-337. 087924
Roberts S. (2005). Using moving total mortality counts to obtain improved estimates for the effect of air
pollution on mortality. Environ Health Perspect, 113: 1148-1152. 087992
Roberts S. (2006). A new model for investigating the mortality effects of multiple air pollutants in air pollution
mortality time-series studies. J Toxicol Environ Health A, 69: 417-435. 089762
Roberts Si Martin MA. (2006). Applying a moving total mortality count to the cities in the NMMAPS database
to estimate the mortality effects of particulate matter air pollution. Occup Environ Med, 63: 193-197.
088670
Roberts Si Martin MA. (2006). The question of nonlinearity in the dose-response relation between particulate
matter air pollution and mortality: can Akaike's Information Criterion be trusted to take the right turn?.
Am J Epidemiol, 164: 1242-50. 097799
Roberts Si Martin MA. (2007). A distributed lag approach to fitting non-linear dose-response models in
particulate matter air pollution time series investigations. Environ Res, 104: 193-200. 156916
Roberts Si Martin MA. (2007). Methods for bias reduction in time-series studies of particulate matter air
pollution and mortality. J Toxicol Environ Health A, 70: 665-675. 156917
Rodriguez Ci Tonkin Ri Heyworth Ji Kusel Mi De Klerk Ni Sly PDi Franklin Pi Runnion Ti Blockley Ai Landau
Li Hinwood AL. (2007). The relationship between outdoor air quality and respiratory symptoms in young
children. Int J Environ Health Res. 092842
Rogers JFi Dunlop AL. (2006). Air pollution and very low birth weight infants: a target population?. , 118: 156-
164. 091232
Rojas-Martinez Ri Perez-Padilla Ri Olaiz-Fernandez Gi Mendoza-Alvarado Li Moreno-Macias Hi Fortoul Ti
McDonnell Wi Loomis Di Romieu I. (2007). Lung function growth in children with long-term exposure to
air pollutants in Mexico City. Am J Respir Crit Care Med, 176: 377-384. 091064
Roman HAi Walker KDi Walsh TLi Conner Li Richmond HMi Hubbell BJi Kinney PL. (2008). Expert judgment
assessment of the mortality impact of changes in ambient fine particulate matter in the U.S. Environ Sci
Technol, 42: 2268-2274. 156921
Romieu Ii Garcia-Esteban Ri Sunyer Ji Rios Ci Alcaraz-Zubeldia Mi Velasco SRi Holguin F. (2008). The effect of
supplementation with omega-3 polyunsaturated fatty acids on markers of oxidative stress in elderly
exposed to PM(2.5). Environ Health Perspect, 116: 1237-1242. 156922
Romieu Ii Ramirez-Aguilar Mi Moreno-Macias Hi Barraza-Villarreal Ai Miller Pi Hernandez-Cadena Li
Carbajal"Arroyo LAi Hernandez-Avila M. (2004). Infant mortality and air pollution: modifying effect by
social class. J Occup Environ Hyg, 46: 1210-1216. 093074
Romieu Ii Tellez-Rojo MMi Lazo Mi Manzano-Patino Cortez-Lugo Mi Julien Pi Belanger MCi Hernandez-Avila
MHolguin F. (2005). Omega-3 fatty acid prevents heart rate variability reductions associated with
particulate matter. Am J Respir Crit Care Med, 172: 1534-1540. 086297
Rosenlund Mi Bellander Ti Nordquist Ti Alfredsson L. (2007). Long-Term Exposure to Air Pollution and Cancer.
, 18(5): S66. 114679
Rosenlund Mi Bellander Ti Nordqvist Ti Alfredsson L. (2006). A register-based case-control study of air
pollution and myocardial infarction. Epidemiology, 17: S240. 114678
Rosenthal FSi Carney JPi Olinger ML. (2008). Out-of-hospital cardiac arrest and airborne fine particulate
matter: a case-crossover analysis of emergency medical services data in Indianapolis, Indiana. Environ
Health Perspect, 116: 631-636. 156925
Ruckerl Ri Greven Si Ljungman Pi Aalto Pi Antoniades Ci Bellander Ti Berglind Ni Chrysohoou Ci Forastiere Fi
Jacquemin Bi von Klot Si Koenig Wi Kuchenhoff Hi Lanki Ti Pekkanen Ji Perucci CAi Schneider Ai
Sunyer Ji Peters A. (2007). Air pollution and inflammation (interleukin-6, C-reactive protein, fibrinogen)
in myocardial infarction survivors. Environ Health Perspect, 115: 1072-1080. 156931
Ruckerl Ri Ibald-Mulli Ai Koenig WSchneider Ai Woelke Gi Cyrys Ji Heinrich Ji Marder Vi Frampton Mi
Wichmann HEi Peters A. (2006). Air pollution and markers of inflammation and coagulation in patients
with coronary heart disease. Environ Health Perspect, 173: 432-441. 088754
July 2009
E-696
DRAFT-DO NOT CITE OR QUOTE

-------
Ruckerl R; Phipps RP; Schneider A; Frampton M; Cyrys J; Oberdorster G; Wichmann HE; Peters A. (2007).
Ultrafine particles and platelet activation in patients with coronary heart disease -- results from a
prospective panel study. Part Fibre Toxicol, 4- 1. 091379
Sagiv SE Mendola P; Loomis D; Herring AH; Neas LM; Savitz DA; Poole C. (2005). A time-series analysis of air
pollution and preterm birth in Pennsylvania, 1997-2001. Environ Health Perspect, 113- 602-606. 087468
Sakai M; Sato YJ Sato S; Ihara SJ Onizuka M; Sakakibara Y; Takahashi H. (2004). Effect of relocating to areas
of reduced atmospheric particulate matter levels on the human circulating leukocyte count. J Appl
Physiol, 97: 1774-1780. 087435
Salam MT; Millstein J; Li Y-F; Lurmann FW; Margolis HG; Gilliland FD. (2005). Birth outcomes and prenatal
exposure to ozone, carbon monoxide, and particulate matter: results from the Children's Health Study.
Environ Health Perspect, 113: 1638-1644. 087885
Samoli E; Analitis A; Touloumi G; Schwartz J; Anderson HR; Sunyer J; Bisanti L; Zmirou D; Vonk JM;
Pekkanen J; Goodman P; Paldy A; Schindler C; Katsouyanni K. (2005). Estimating the exposure-
response relationships between particulate matter and mortality within the APHEA multicity project.
Environ Health Perspect, 113: 88-95. 087436
Sanchez-Carrillo CI; Ceron-Mireles P; Rojas-Martinez MR; Mendoza-Alvarado L; Olaiz-Fernandez G; Borja-
Aburto VH. (2003). Surveillance of acute health effects of air pollution in Mexico City. , 14: 536-44.
098428
Santos U; Terra-Filho M; Lin C; Pereira L; Vieira T; Saldiva P; Braga A. (2008). Cardiac arrhythmia emergency
room visits and environmental air pollution in Sao Paulo, Brazil. J Epidemiol Community Health, 62:
267. 192004
Sarnat JA; Marmur A; Klein M; Kim E; Russell AG; Sarnat SE; Mulholland JA; Hopke PK; Tolbert PE. (2008).
Fine particle sources and cardiorespiratory morbidity: an application of chemical mass balance and factor
analytical source-apportionment methods. Environ Health Perspect, 116: 459-66. 097972
Sarnat SE; Suh HH; Coull BA; Schwartz J; Stone PH; Gold DR. (2006). Ambient particulate air pollution and
cardiac arrhythmia in a panel of older adults in Steubenville, Ohio. Occup Environ Med, 63: 700-706.
090489
Sauerzapf V; Jones AP; Cross J. (2009). Environmental factors and hospitalisation for chronic obstructive
pulmonary disease in a rural county of England. J Epidemiol Community Health, 63: 324-328. 180082
Schikowski T; Sugiri D; Ranft U; Gehring U; Heinrich J; Wichmann HE; Kramer U. (2005). Long-term air
pollution exposure and living close to busy roads are associated with COPD in women. Respir Res, 22:
152-161. 088637
Schildcrout JS; Sheppard L; Lumley T; Slaughter JC; Koenig JQ; Shapiro GG. (2006). Ambient air pollution and
asthma exacerbations in children: an eight-city analysis. Am J Epidemiol, 164: 505-517. 089812
Schindler C; Keidel D; Gerbase MWi Zemp E; Bettschart R; Brandli O; Brutsche MH; Burdet L; Karrer W;
Knopfli B; Pons M; Rapp R; Bayer-Oglesby L; Kunzli N; Schwartz J; Liu L-JS; Ackermann-Liebrich U;
Rochat T; the SAPALDIA Team. (2009). Improvements in PM10 exposure and reduced rates of
respiratory symptoms in a cohort of swiss adults (SAPALDIA). , 179: 579-587. 191950
Schneider A; Neas L; Herbst M; Case M; Williams R; Cascio W', Hinderliter A; Holguin F; Buse J; Dungan K.
(2008). Endothelial dysfunction: associations with exposure to ambient fine particles in diabetic
individuals. Environ Health Perspect, 116: 1666. 191985
Schreuder AB; Larson TV; Sheppard L; Claiborn CS. (2006). Ambient woodsmoke and associated respiratory
emergency department visits in Spokane, Washington. Int J Occup Environ Health, 12: 147-53. 097959
Schwartz J. (2004). Is the association of airborne particles with daily deaths confounded by gaseous air
pollutants? An approach to control by matching. Environ Health Perspect, 112: 557-561. 053506
Schwartz J. (2004). The effects of particulate air pollution on daily deaths: a multi-city case crossover analysis.
Occup Environ Med, 61: 956-961. 078998
Schwartz J; Coull B; Laden F; Ryan L. (2008). The effect of dose and timing of dose on the association between
airborne particles and survival. Environ Health Perspect, 116: 64"69. 156963
Schwartz J; Laden F; Zanobetti A. (2002). The concentration-response relation between PM2.5 and daily deaths.
Environ Health Perspect, 110: 1025-1029. 025312
Schwartz J; Litonjua A; Suh H; Verrier M; Zanobetti A; Syring M; Nearing B; Verrier R; Stone P; MacCallum G;
Speizer FE; Gold DR. (2005). Traffic related pollution and heart rate variability in a panel of elderly
subjects. Thorax, 60: 455-461. 074317
Sekine K; Shima M; Nitta Y; Adachi M. (2004). Long term effects of exposure to automobile exhaust on the
pulmonary function of female adults in Tokyo, Japan. Occup Environ Med, 61: 350-357. 090762
Sharma M; Kumar VN; Katiyar SE Sharma R; Shukla BP; Sengupta B. (2004). Effects of particulate air
pollution on the respiratory health of subjects who live in three areas in Kanpur, India. , 59: 348-358.
156974
July 2009
E-697
DRAFT-DO NOT CITE OR QUOTE

-------
Sharovsky R; Cesar LA; Ramires JA. (2004). Temperature, air pollution, and mortality from myocardial
infarction in Sao Paulo, Brazil. Braz J Med Biol Res, 37- 1651-1657. 156976
Silkoff PE; Zhang L; Dutton Si Langmack ELi Vedal Si Murphy Ji Make B. (2005). Winter air pollution and
disease parameters in advanced chronic obstructive pulmonary disease panels residing in Denver,
Colorado. J Allergy Clin Immunol, 115: 337-344. 087471
Simpson Ri Williams Gi Petroeschevsky Ai Best Ti Morgan Gi Denison Li Hinwood Ai Neville G. (2005). The
short-term effects of air pollution on hospital admissions in four Australian cities. Aust N Z J Public
Health, 29: 213-221. 087438
Sinclair AHi Tolsma D. (2004). Associations and lags between air pollution and acute respiratory visits in an
ambulatory care setting: 25-month results from the aerosol research and inhalation epidemiological
study. J Air Waste Manag Assoc, 54: 1212-1218. 088696
Singh Vi Khandelwal Ri Gupta AB. (2003). Effect of air pollution on peak expiratory flow rate variability. J
Asthma, 40: 81-86. 052686
Sivacoumar Ri Jayabalou Ri Swarnalatha Si Balakrishnan K. (2006). Particulate Matter from Stone Crushing
Industry: Size Distribution and Health Effects. , 132: 405-414. 111115
Slama Ri Morgenstern Vi Cyrys Ji Zutavern Ai Herbarth Oi Wichmann HE; Heinrich J; LISA Study Group.
(2007). Traffic-related atmospheric pollutants levels during pregnancy and offspring's term birth weight:
a study relying on a land-use regression exposure model. Environ Health Perspect, 115: 1283-1292.
093216
Slaughter JC; Kim E; Sheppard L; Sullivan JH; Larson TV; Claiborn C. (2005). Association between particulate
matter and emergency room visits, hospital admissions and mortality in Spokane, Washington. J Expo
Sci Environ Epidemiol, 15: 153-159. 073854
Slaughter JC; Lumley T; Sheppard L; Koenig JQ; Shapiro GG. (2003). Effects of ambient air pollution on
symptom severity and medication use in children with asthma. Ann Allergy Asthma Immunol, 91: 346"
353. 086294
Sokol RZ; Kraft P; Fowler IM; Mamet R; Kim E; Berhane KT. (2006). Exposure to environmental ozone alters
semen quality. Environ Health Perspect, 114: 360-5. 098539
Solomon C; Poole J; Jarup L; Palmer K; Coggon D. (2003). Cardio-respiratory morbidity and long-term exposure
to particulate air pollution. Int J Environ Health Res, 13: 327-335. 087441
Solomon P; Baumann K; Edgerton E; Tanner R; Eatough D; Modey W', Marin H; Savoie D; Natarajan S; Meyer
MB. (2003). Comparison of integrated samplers for mass and composition during the 1999 Atlanta
supersites project. J Geophys Res, 108: 8423. 156994
Sorensen M; Daneshvar B; Hansen M; Dragsted LO; Hertel O; Knudsen L; Loft S. (2003). Personal PM25
exposure and markers of oxidative stress in blood. Environ Health Perspect, 111: 161-165. 042700
Sorensen M; Loft S; Andersen HV; Raaschou-Nielsen O; Skovgaard LT; Knudsen LE; Nielsen IV; Hertel O.
(2005). Personal exposure to PM25, black smoke and N02 in Copenhagen: relationship to bedroom and
outdoor concentrations covering seasonal variation. J Expo Sci Environ Epidemiol, 15: 413-422. 089428
Sram R; Beskid 0; Binkova B; Chvatalova I; Lnenickova Z; Milcova A; Solansky I; Tulupova E; Bavorova H;
Ocadlikova D. (2007). Chromosomal aberrations in environmentally exposed population in relation to
metabolic and DNA repair genes polymorphisms. , 620: 22"33. 188457
Stafoggia M; Schwartz J; Forastiere F; Perucci CA. (2008). Does temperature modify the association between air
pollution and mortality? A multicity case-crossover analysis in Italy. Am J Epidemiol, 167: 1476-1485.
157005
Staniswalis JG; Parks NJ; Bader JO; Maldonado YM. (2005). Temporal analysis of airborne particulate matter
reveals a dose-rate effect on mortality in El Paso: indications of differential toxicity for different particle
mixtures. J Air Waste Manag Assoc, 55: 893-902. 087473
Steinvil A; Kordova-Biezuner L; Shapira I; Berliner S; Rogowski O. (2008). Short-term exposure to air pollution
and inflammation-sensitive biomarkers. , 106: 51-61. 188893
Stieb DM; Judek S; Burnett RT. (2002). Meta-analysis of time-series studies of air pollution and mortality:
effects of gases and particles and the influence of cause of death, age, and season. J Air Waste Manag
Assoc, 52: 470-484. 025205
Strand M; Vedal S; Rodes C', Dutton SJ; Gelfand EW; Rabinovitch N. (2006). Estimating effects of ambient
PM25 exposure on health using PM25 component measurements and regression calibration. J Expo Sci
Environ Epidemiol, 16: 30-38. 089203
Stolzel M; Breitner S; Cyrys J; Pitz M; Wolke G; Kreyling W; Heinrich J; Wichmann H"E; Peters A. (2007).
Daily mortality and particulate matter in different size classes in Erfurt, Germany. J Expo Sci Environ
Epidemiol, 17: 458-467. 091374
Su TC; Chan CC; Liau CS; Lin LY; Kao HL; Chuang KJ. (2006). Urban air pollution increases plasma fibrinogen
and plasminogen activator inhibitor-1 levels in susceptible patients. , 13: 849-852. 157022
July 2009
E-698
DRAFT-DO NOT CITE OR QUOTE

-------
Suglia SF; Gryparis A; Wright RO; Schwartz J; Wright RJ. (2008). Association of black carbon with cognition
among children in a prospective birth cohort study. Am J Epidemiol, 167- 280-286. 157027
Suh YJ; Kim BM; Park BH; Park H; Kim YJ; Kim H; Hong YC; Ha EH. (2007). Cytochrome P450IA1
polymorphisms along with PM(10) exposure contribute to the risk of birth weight reduction. Reprod
Toxicol, 24: 281-288. 157028
Sullivan J; Ishikawa N> Sheppard L; Siscovick D; Checkoway H; Kaufman J. (2003). Exposure to ambient fine
particulate matter and primary cardiac arrest among persons with and without clinically recognized
heart disease. Am J Epidemiol, 157: 501-509. 043156
Sullivan JH; Hubbard R; Liu SL; Shepherd Ki Trenga CA Koenig JQ; Chandler WL; Kaufman JD. (2007). A
community study of the effect of particulate matter on blood measures of inflammation and thrombosis in
an elderly population. Environ Health Perspect, 6: 3. 100083
Sullivan JH; Schreuder AB; Trenga CA; Liu SL; Larson TV; Koenig JQ; Kaufman JD. (2005). Association
between short term exposure to fine particulate matter and heart rate variability in older subjects with
and without heart disease. Thorax, 60: 462"6. 109418
Sun H-L; Chou M-C; Lue K-H. (2006). The relationship of air pollution to ED visits for asthma differ between
children and adults. Am J Emerg Med, 24: 709-713. 090768
Sunyer J; Basagana X; Belmonte J; Anto JM. (2002). Effect of nitrogen dioxide and ozone on the risk of dying in
patients with severe asthma. Thorax, 57: 687-693. 034835
Sunyer J; Jarvis D; Gotschi T; Garcia-Esteban R; Jacquemin B; Aguilera I; Ackerman U; De Marco R; Forsberg
B; Gislason T; Heinrich J; Norback D; Villani S; Kunzli N. (2006). Chronic bronchitis and urban air
pollution in an international study. Occup Environ Med, 63: 836-843. 089771
Symons JM; Wang L; Guallar E; Howell E; Dominici F; Schwab M; Ange BA; Samet J; Ondov J; Harrison D;
Geyh A. (2006). A case-crossover study of fine particulate matter air pollution and onset of congestive
heart failure symptom exacerbation leading to hospitalization. Am J Epidemiol, 164: 421-33. 091258
Szyszkowicz M. (2007). Air pollution and emergency department visits for depression in Edmonton, Canada. Int
J Occup Med Environ Health, 20: 241-245. 092829
Tager IB; Balmes J; Lurmann F; Ngo L; Alcorn S; Kunzli N. (2005). Chronic exposure to ambient ozone and
lung function in young adults. , 16: 751-759. 087538
Tainio M; Tuomisto JT; Hanninen O; Aarnio P; Koistinen KJ; Jantunen MJ; Pakkanen J. (2005). Health effects
caused by primary fine particulate matter (PM2.5) emitted from buses in the Helsinki metropolitan area,
Finland. Risk Anal, 25: 151-160. 087444
Tamura K; Jinsart W', Yano E; Karita K; Boudoung D. (2003). Particulate air pollution and chronic respiratory
symptoms among traffic policemen in Bangkok. Arch Environ Occup Health, 58: 201-207. 087445
Tang C-S; Chang L"T; Lee H-C; Chan C-C. (2007). Effects of personal particulate matter on peak expiratory flow
rate of asthmatic children. Sci Total Environ, 382: 43-51. 091269
Tarantini L; Bonzini M; Apostoli P; Pegoraro V; Bollati V; Marinelli B; Cantone L; Rizzo G; Hou L; Schwartz J;
Bertazzi PA; Baccarelli A. (2009). Effects of particulate matter on genomic DNA methylation content and
iNOS promoter methylation. Environ Health Perspect, 117: 217-222. 192010
Tecer LH; Alagha O; Karaca F; Tuncel G; Eldes N. (2008). Particulate Matter (PM2.5, PM10-2.5, and PM10) and
Children's Hospital Admissions for Asthma and Respiratory Diseases: A Bidirectional Case "Crossover
Study. J Toxicol Environ Health A, 71: 512-520. 180030
Thurston G; Ito K; Mar T; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Liu H;
Neas L; Pinto J; Stolzel M; Suh H; Hopke PK. (2005). Results and implications of the workshop on the
source apportionment of PM health effects. Epidemiology, 16: S134-S135. 097949
Timonen KL; Hoek G; Heinrich J; Bernard A; Brunekreef B; De Hartog J; Hameri K; Ibald-Mulli A; Mirme A;
Peters A; Tiittanen P; Kreyling WG! Pekkanen J. (2004). Daily variation in fine and ultrafine particulate
air pollution and urinary concentrations of lung Clara cell protein CC16. Occup Environ Med, 61: 908"
914. 087915
Timonen KL; Vanninen E; De Hartog J; Ibald-Mulli A; Brunekreef B; Gold DR; Henrich J; Hoek G; Lanki T;
Peters A; Tarkiainen T; Tiittanen P; Kreyling W', Pekkanen J. (2006). Effects of ultrafine and fine
particulate and gaseous air pollution on cardiac autonomic control in subjects with coronary artery
disease: the ULTRA study. J Expo Sci Environ Epidemiol, 16: 332-341. 088747
Tolbert PE; Kleina M; Peelb JL; Sarnata SE; Sarnata JA. (2007). Multipollutant modeling issues in a study of
ambient air quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 24: 938"
945. 090316
Touloumi G; Samoli E; Quenel P; Paldy A; Anderson RH; Zmirou D; Galan I; Forsberg B; Schindler C; Schwartz
J; Katsouyanni K. (2005). Short-term effects of air pollution on total and cardiovascular mortality: the
confounding effect of influenza epidemics. , 16: 49"57. 087477
July 2009
E-699
DRAFT-DO NOT CITE OR QUOTE

-------
Tovalin H; Valverde M; Morandi MT; Blanco S; Whitehead L; Rojas E. (2006). DNA damage in outdoor workers
occupationally exposed to environmental air pollutants. Occup Environ Med, 63- 230-236. 091322
Trenga CA; Sullivan JH; Schildcrout JS; Shepherd KP; Shapiro GG; Liu LJ; Kaufman JD; Koenig JQ. (2006).
Effect of particulate air pollution on lung function in adult and pediatric subjects in a Seattle panel
study. , 129: 1614-22. 155209
Tsai CJ; Chang CT; Huang CH. (2006). Direct field observation of the relative humidity effect on the beta-gauge
readings. J Air Waste Manag Assoc, 56: 834-40. 098312
Tsai S-S; Cheng M"H; Chiu H-F; Wu T-N; Yang C-Y. (2006). Air pollution and hospital admissions for asthma in
a tropical city: Kaohsiung, Taiwan. Inhal Toxicol, 18: 549-554. 089768
Tsai S-S; Goggins WB; Chiu H"F; Yang C"Y. (2003). Evidence for an association between air pollution and daily
stroke admissions in Kaohsiung, Taiwan. , 34: 2612-2616. 080133
Tsai S-S; Huang C-H; Goggins WB; Wu T"N; Yang C"Y. (2003). Relationship between air pollution and daily
mortality in a tropical city: Kaohsiung, Taiwan. J Toxicol Environ Health A, 66: 1341-1349. 050480
Ulirsch GV; Ball LM; Kaye W', Shy CM; Lee CV; Crawford-Brown D; Symons M; Holloway T. (2007). Effect of
particulate matter air pollution on hospital admissions and medical visits for lung and heart disease in
two southeast Idaho cities. J Expo Sci Environ Epidemiol, 17: 478-487. 091332
Vajanapoom N> Shy CM; Neas LM; Loomis D. (2002). Associations of particulate matter and daily mortality in
Bangkok, Thailand. Southeast Asian J Trop Med Public Health, 33: 389-399. 042542
Vallejo M; Ruiz S; Hermosillo AG; Borja-Aburto VH; Cardenas M. (2006). Ambient fine particles modify heart
rate variability in young healthy adults. J Expo Sci Environ Epidemiol, 16: 125-130. 157081
Van Hee VC; Adar SD; Szpiro AA; Barr RG; Bluemke DA; Diez Roux AV; Gill EA; Sheppard L; Kaufman JD.
(2009). Exposure to traffic and left ventricular mass and function: the Multi-Ethnic Study of
Atherosclerosis. Am J Respir Crit Care Med, 179: 827-834. 192110
Vedal S; Brauer M; White R; Petkau J. (2003). Air pollution and daily mortality in a city with low levels of
pollution. Environ Health Perspect, 111: 45-51. 039044
Vedal S; Rich K; Brauer M; White R; Petkau J. (2004). Air pollution and cardiac arrhythmias in patients with
implantable cardiovascular defibrillators. Inhal Toxicol, 16: 353-362. 055630
Vegni FE; Ros O. (2004). Hospital accident and emergency burden is unaffected by today's air pollution levels.
Eur J Emerg Med, 11: 86-88. 087448
Venners SA; Wang B; Xu Z; Schlatter Y; Wang L; Xu X. (2003). Particulate matter, sulfur dioxide, and daily
mortality in Chongqing, China. Environ Health Perspect, 111: 562-567. 089931
Vichit-Vadakan N; Vajanapoom N; Ostro B. (2008). The Public Health and Air Pollution in Asia (PAPA) Project:
estimating the mortality effects of particulate matter in Bangkok, Thailand. Environ Health Perspect,
116: 1179-1182. 157095
Vigotti MA; Chiaverini F; Biagiola P; Rossi G. (2007). Urban air pollution and emergency visits for respiratory
complaints in Pisa, Italy. J Toxicol Environ Health A, 70: 266-269. 090711
Villeneuve PJ; Burnett RT; Shi Y; Krewski D; Goldberg MS; Hertzman C; Chen Y; Brook J. (2003). A time-
series study of air pollution, socioeconomic status, and mortality in Vancouver, Canada. J Expo Sci
Environ Epidemiol, 13: 427-435. 055051
Villeneuve PJ; Chen L; Stieb D; Rowe BH. (2006). Associations between outdoor air pollution and emergency
department visits for stroke in Edmonton, Canada. Eur J Epidemiol, 21: 689-700. 090191
Villeneuve PJ; Goldberg MS; Krewski D; Burnett RT; Chen Y. (2002). Fine particulate air pollution and all-
cause mortality within the Harvard six-cities study: variations in risk by period of exposure. Ann
Epidemiol, 12: 568-576. 042576
Vineis P; Hoek G; Krzyzanowski M; Vigna-Taglianti F; Veglia F; Airoldi L; Autrup H; Dunning A; Garte S;
Hainaut P; Malaveille C; Matullo G; Overvad K; Raaschou-Nielsen O; Clavel-Chapelon F; Linseisen J;
Boeing H; Trichopoulou A; Palli D; Peluso M; Krogh V; Tumino R; Panico S; Bueno-De-Mesquita HB;
Peeters PH; Lund EE; Gonzalez CA; Martinez C; Dorronsoro M; Barricarte A; Cirera L; Quiros JR.;
Berglund G; Forsberg B; Day NE; Key TJ; Saracci R; Kaaks R; Riboli E. (2006). Air pollution and risk of
lung cancer in a prospective study in Europe. Int J Cancer, 119: 169-174. 192089
Von KLot S; Peters A; Aalto P; Bellander T; Berglind N; D'Ippoliti D; Elosua R; Hormann A; Kulmala M; Lanki
T; Lowel H; Pekkanen J; Picciotto S; Sunyer J; Forastiere F; Health Effects of Particles on Susceptible
Subpopulations (HEAPSS) Study Group. (2005). Ambient air pollution is associated with increased risk
of hospital cardiac readmissions of myocardial infarction survivors in five European cities. Circulation,
112: 3073-3079. 088070
Von Klot S; Wolke G; Tuch T; Heinrich J; Dockery DW; Schwartz J; Kreyling WG; Wichmann HE; Peters A.
(2002). Increased asthma medication use in association with ambient fine and ultrafine particles. Eur
Respir J, 20: 691-702. 034706
July 2009
E-700
DRAFT-DO NOT CITE OR QUOTE

-------
Ward DJ; Roberts KT; Jones N; Harrison RM; Ayres JG; Hussain S; Walters S. (2002). Effects of daily variation
in outdoor particulates and ambient acid species in normal and asthmatic children. Thorax, 57- 489-502.
025839
Wei Y; Han I"K; Shao M; Hu M; Zhang J; Tang X. (2009). PM2.5 Constituents and Oxidative DNA Damage in
Humans. Environ Sci Technol, 43- 4757—4762. 192361
Wellenius GA; Bateson TF; Mittleman MA; Schwartz J. (2005). Particulate air pollution and the rate of
hospitalization for congestive heart failure among medicare beneficiaries in Pittsburgh, Pennsylvania.
Am J Epidemiol, 161: 1030-1036. 087483
Wellenius GA; Schwartz J; Mittleman MA. (2005). Air pollution and hospital admissions for ischemic and
hemorrhagic stroke among medicare beneficiaries. , 36: 2549-2553. 088685
Wellenius GA; Schwartz J; Mittleman MA. (2006). Particulate air pollution and hospital admissions for
congestive heart failure in seven United States cities. Am J Cardiol, 97: 404-408. 088748
Wellenius GA; Yeh GY; Coull BA; Suh HH; Phillips RS; Mittleman MA. (2007). Effects of ambient air pollution
on functional status in patients with chronic congestive heart failure: a repeated-measures study. , 6: 26.
092830
Welty LJ; Peng RD; Zeger SL; Dominici F. (2008). Bayesian Distributed Lag Models: Estimating Effects of
Particulate Matter Air Pollution on Daily Mortality. Biometrics, 65: 282-291. 157134
Welty LJ; Zeger SL. (2005). Are the acute effects of particulate matter on mortality in the National Morbidity,
Mortality, and Air Pollution study the result of inadequate control for weather and season? A sensitivity
analysis using flexible distributed lag models. Am J Epidemiol, 162: 80-88. 087484
Wheeler A; Zanobetti A; Gold DR; Schwartz J; Stone P; Suh HH. (2006). The relationship between ambient air
pollution and heart rate variability differs for individuals with heart and pulmonary disease. Environ
Health Perspect, 114: 560-566. 088453
Wheeler BW; Ben-Shlomo Y. (2005). Environmental equity, air quality, socioeconomic status, and respiratory
health: a linkage analysis of routine data from the Health Survey for England. J Epidemiol Community
Health, 59: 948-954. 089860
Whitsel E; Quibrera P; Christ S; Liao D; Prineas R; Anderson G; Heiss G. (2009). Heart rate variability,
ambient particulate matter air pollution, and glucose homeostasis: the environmental epidemiology of
arrhythmogenesis in the Women's Health Initiative. Am J Epidemiol, x: x. 191980
Wilhelm M; Ritz B. (2005). Local variations in CO and particulate air pollution and adverse birth outcomes in
Los Angeles County, California, USA. Environ Health Perspect, 113: 1212-1221. 088668
Wilhelm M; Ritz B. (2005). Local variations in CO and particulate air pollution and adverse birth outcomes in
Los Angeles County, California, USA. Environ Health Perspect, 113: 1212-21. 188761
Willis A; Jerrett M; Burnett RT; Krewski D. (2003). The association between sulfate air pollution and mortality
at the county scale: an exploration of the impact of scale on a long-term exposure study. J Toxicol
Environ Health A, 66: 1605-1624. 089922
Wilson WE; Mar TF; Koenig JQ. (2007). Influence of exposure error and effect modification by socioeconomic
status on the association of acute cardiovascular mortality with particulate matter in Phoenix. J Expo Sci
Environ Epidemiol, 17: Sll. 157149
Wong C-M; Atkinson RW; Anderson HR; Hedley AJ; Ma S; Chau PY-K; Lam T"H. (2002). A tale of two cities:
effects of air pollution on hospital admissions in Hong Kong and London compared. Environ Health
Perspect, 110: 67-77. 023232
Wong C-M; Ou C-Q; Thach T-Q; Chau Y-K; Chan K-P; Ho S-Y; Chung RY; Lam T-H; Hedley AJ. (2007). Does
regular exercise protect against air pollution-associated mortality?. Prev Med, 44: 386-392. 093278
Wong CM; Ou CQ; Chan KP; Chau YE Thach TQ; Yang U Chung RY; Thomas GN; Peins JS; Wong TW;
Hedley AJ; Lam TH. (2008). The effects of air pollution on mortality in socially deprived urban areas in
Hong Kong, China. Environ Health Perspect, 116: 1189-1194. 157151
Wong CM; Ou CQ; Lee NW Chan KP; Thach TQ; Chau YE Ho SY; Hedley AJ; Lam TH. (2007). Short-term
effects of particulate air pollution on male smokers and never-smokers. , 18: 593"8. 098391
Wong CM; Vichit-Vadakan N; Kan H; Qian Z. (2008). Public Health and Air Pollution in Asia (PAPA): a
multicity study of short-term effects of air pollution on mortality. Environ Health Perspect, 116: 1195-
1202. 157152
Wong TW; Tam W', Tak Sun Yu I; Wun TY; Wong AH; Wong CM. (2006). Association between air pollution and
general practitioner visits for respiratory diseases in Hong Kong. Thorax, 61: 585-591. 093266
Wong TW; Tam WS; Yu TS; Wong AHS. (2002). Associations between daily mortalities from respiratory and
cardiovascular diseases and air pollution in Hong Kong, China. Occup Environ Med, 59: 30-35. 025436
Woodruff TJ; Darrow LA; Parker JD. (2008). Air pollution and postneonatal infant mortality in the United
States, 1999-2002. Environ Health Perspect, 116: 110-5. 098386
July 2009
E-701
DRAFT-DO NOT CITE OR QUOTE

-------
Woodruff TJ; Grillo J; Schoendorf KC. (1997). The relationship between selected causes of postneonatal infant
mortality and particulate air pollution in the United States. Environ Health Perspect, 105- 608-612.
084271
Woodruff TJ; Parker JD; Schoendorf KC. (2006). Fine particulate matter (PM25) air pollution and selected
causes of postneonatal infant mortality in California. Environ Health Perspect, 114- 785-790. 088758
Xirasagar Si Lin HCi Liu TC. (2006). Seasonality in pediatric asthma admissions' the role of climate and
environmental factors. Eur J Pediatr, 165- 747-752. 093267
Yamazaki Si Nitta Hi Ono Mi Green Ji Fukuhara S. (2007). Intracerebral haemorrhage associated with hourly
concentration of ambient particulate matter- case-crossover analysis. Occup Environ Med, 64- 17-24.
090748
Yang C-Yi Chang C"Ci Chuang H"Yi Tsai S"Si Wu T"Ni Ho C"K. (2004). Relationship between air pollution and
daily mortality in a subtropical city- Taipei, Taiwan. Environ Int, 30- 519-523. 055603
Yang C-Yi Chen Y"Si Yang C"Hi Ho S"C. (2004). Relationship between ambient air pollution and hospital
admissions for cardiovascular diseases in Kaohsiung, Taiwan. J Toxicol Environ Health A, 67- 483-493.
094376
Yang CY. (2008). Air pollution and hospital admissions for congestive heart failure in a subtropical city- Taipei,
Taiwan. J Toxicol Environ Health A, 71- 1085-1090. 157160
Yang CYi Chen CJ. (2007). Air pollution and hospital admissions for chronic obstructive pulmonary disease in a
subtropical city: Taipei, Taiwan. J Toxicol Environ Health A, 70: 1214-9. 092847
Yang Q; Chen Y; Krewski D; Shi Y; Burnett RT; McGrail KM. (2004). Association between particulate air
pollution and first hospital admission for childhood respiratory illness in Vancouver, Canada. Arch
Environ Occup Health, 59: 14-21. 087488
Yeatts K; Svendsen E; Creason J; Alexis N; Herbst M; Scott J; Kupper L; Williams R; Neas L; Cascio W', Devlin
RB; Peden DB. (2007). Coarse particulate matter (PM25-10) affects heart rate variability, blood lipids,
and circulating eosinophils in adults with asthma. Environ Health Perspect, 115: 709-714. 091266
Yue W; Schneider A; Stolzel M; Ruckerl R; Cyrys J; Pan X; Zareba W; Koenig W; Wichmann HE; Peters A.
(2007). Ambient source-specific particles are associated with prolonged repolarization and increased
levels of inflammation in male coronary artery disease patients. , 621: 50-60. 097968
Zanobetti A; Canner MJ; Stone PH; Schwartz J; Sher D; Eagan-Bengston E; Gates KA; Hartley LH; Suh H;
Gold DR. (2004). Ambient pollution and blood pressure in cardiac rehabilitation patients. , 110: 2184"
2189. 087489
Zanobetti A; Schwartz J. (2002). Cardiovascular damage by airborne particles: are diabetics more susceptible?. ,
13: 588-592. 034821
Zanobetti A; Schwartz J. (2003). Multicity assessment of mortality displacement within the APHEA2 project.
042812
Zanobetti A; Schwartz J. (2005). The effect of particulate air pollution on emergency admissions for myocardial
infarction: a multicity case-crossover analysis. Environ Health Perspect, 113: 978-982. 088069
Zanobetti A; Schwartz J. (2006). Air pollution and emergency admissions in Boston, MA. J Epidemiol
Community Health, 60: 890-895. 090195
Zanobetti A; Schwartz J. (2007). Particulate air pollution, progression, and survival after myocardial infarction.
Environ Health Perspect, 115: 769-775. 091247
Zanobetti A; Schwartz J. (2009). The effect of fine and coarse particulate air pollution on mortality: A national
analysis. Environ Health Perspect, 117: 1-40. 188462
Zeger S; Dominici F; McDermott A; Samet J. (2008). Mortality in the Medicare population and chronic exposure
to fine particulate air pollution in urban centers (2000-2005). Environ Health Perspect, 116: 1614.
191951
Zeger S; McDermott A; Dominici F; Samet J. (2007). Mortality in the medicare population and chronic exposure
to fine particulate air pollution. Environ Health Perspect, 116: 1614-1619. 157176
Zeka A; Sullivan JR; Vokonas PS; Sparrow D; Schwartz J. (2006). Inflammatory markers and particulate air
pollution: characterizing the pathway to disease. Int J Epidemiol, 35: 1347-1354. 157177
Zeka A; Zanobetti A; Schwartz J. (2005). Short term effects of particulate matter on cause specific mortality:
effects of lags and modification by city characteristics. Occup Environ Med, 62: 718-725. 088068
Zeka A; Zanobetti A; Schwartz J. (2006). Individual-level modifiers of the effects of particulate matter on daily
mortality. Am J Epidemiol, 163: 849-859. 088749
Zhang J; Hu W', Wei F; Wu G; Korn LR; Chapman RS. (2002). Children's respiratory morbidity prevalence in
relation to air pollution in four Chinese cities. Environ Health Perspect, 110: 961-967. 034814
Zhang Z; Whitsel E; Quibrera P; Smith R; Liao D; Anderson G; Prineas R. (2009). Ambient Fine Particulate
Matter Exposure and Myocardial Ischemia in the Environmental Epidemiology of Arrhythmogenesis in
the Women?s Health Initiative (EEAWHI) Study. Environ Health Perspect, 117: 751-756. 191970
July 2009
E-702
DRAFT-DO NOT CITE OR QUOTE

-------
Zhong W> Levin L; Reponen T> Hershey GE Adhikari A; Shukla R; LeMasters G. (2006). Analysis of short-term
influences of ambient aeroallergens on pediatric asthma hospital visits. Sci Total Environ, 370- 330-336.
093264
July 2009
E-703
DRAFT-DO NOT CITE OR QUOTE

-------
Annex F. Source Apportionment Studies
Table F-1. Epidemiologic studies of ambient PM sources, factors, or constituents.
Reference: Andersen
et al. (2007, 093201)
Location: 1 monitor in
Copenhagen, Denmark!
6 years, but
apportionment done for
1.5 year only (2002-
2003)
Particle Size: PMio
Subjects: NR
Exposure: NR
n: NR
Number of
constituents
considered for
grouping: 31
Grouping method: Groups/Factors/ Sources: PM variables used:
PCA + PMF/CMB
hybrid (C0PREM)
ft of groups: 12,
but only 6 used in
relating to health
effects, and CO,
NO?
Road, Vehicle, Salt,
Biomass, Oil, Coal, Rock,
Lime, NaNOs, NH4NO3,
(NH4)2S03, (NH4SO4
Mass contribution of
sources
Results: Single pollutant models: Biomass, secondary compounds, oil, and crustal significantly associated with CVD HA (4 day
moving ave). Biomass and secondary components significantly associated with respiratory HA (5 day moving average). No significant
effects for asthma HA in children (6 day moving average).
Two pollutant models: Crustal effect for CVD admissions remained robust. Biomass effect for respiratory admissions was highest.
Effect of vehicle source remained robust for asthma admissions in children in presence of other PM10 sources.
Reference: Bell et al.
(Bell et al., 2009,
191007)
Location: PM2.B:
2000-2005 (6
years)/106 US
counties/EPA
composition data
PM10: 1987-20001100
counties/EPA
composition data
Particle Size: PM10,
PM2.6
Subjects: NR n: NR	Number of	Grouping method: Groups/Factors/ Sources: PM variables used:
Exposure'NR	constituents	NR	NR	Every component (16
considered for	^ 0j gr0UpS- |\|r	elements + NO3, S04,EC,
grouping: 16	0C)
elements + NO3,
S04, EC, 0C
Results: Mortality: Ni significantly increased PM10 mortality risks. However, effect of Ni was not significant when NY City was
removed, in a sensitivity analysis conducted by selectively removing cities from the overall estimate.
Hospital Admissions: CVD and respiratory HAs higher in counties with higher EC, Ni, and V PM2.6. In CVD association between
PM2.E, RR and V robust to inclusion of EC or V, and V robust to inclusion of EC.
Reference: Cakmak et Subjects: NR
n: NR
Number of
al. (2009,191995)
Exposure: 1998-
Location: 1 monitor in 2009 (8.3 years)
Santiago, Chile
Particle Size: PM2.6
Grouping method: Groups/Factors/ Sources: PM variables used:
constituents PCA
considered for
grouping: 16
elements + CO,
N02,S02,EC,0C
tt of groups: 4
Vehicle (CO, NO2, EC, 0C), individual components,
Soil (Al, Ca, Fe, Si),	then groupings
Combustion (Cr, Cu, Fe, Mn,
Zn), Factor 4 (Br, CI, Pb)
Results: Individual components: EC, 0C only stat. sign, risk estimates for total, cardiac, and respiratory mortality for 1 day lag
after adjustment for other elements.
Groupings.: Lag 1. Vehicle factor: Increased total mortality, cardiac mortality, and respiratory mortality. Soil factor: increased
cardiac mortality and respiratory mortality (but smaller than vehicle factor RRs). Combustion factor: greatest RR for respiratory
mortality, but significant for total and cardiac mortality. Factor 4: increased total, cardiac, and respiratory mortality. Point estimates
for Factor 1 significantly different from Factors 3 and 4. Elderly had higher risk estimates for combustion and soil sources. No
significant effect modification by gender or season.
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database
(Health and Environmental Research Online) at http-//epa.gov/hero. HERO is a database of scientific literature
used by U.S. EPA in the process of developing science assessments such as the Integrated Science Assessments
(ISA) and the Integrated Risk Information System (IRIS).
July 2009	F-1	DRAFT - DO NOT CITE OR QUOTE

-------
Reference: Franklin et
al. (2008,155779)
Location: STNI25
communities/2000-
2005 (6 years)
Particle Size: PM2.B
Subjects: NR
Exposure: NR
n: NR
Number of
constituents
considered for
grouping: 15
elements + EC,
0C, NOs
Grouping method:
# of groups: NR
Groups/Factors/ Sources:
NR
PM variables used:
Every component
Results: The PM2.B-mortality association was significantly modified by Al, As, Sulfate, Ni, and Si. When including a combination of
species proportions and using backwards elimination Al, sulfate, and Ni remained significant. Al and Ni explained most of the residual
heterogeneity.
Reference: Gent et al.
(2009,180399)
Location: 2 monitors
in New Haven, CT/ 3.5
years
Particle Size: PM2.6
Subjects:
Children with
physician
diagnosed asthma
and symptoms or
medication use in
previous 12
months, and
resided within
30km of New
Haven county
monitor
Exposure: NR
n: 149
children
Number of
constituents
considered for
grouping: 17
elements + EC
Grouping method: Groups/Factors/ Sources: PM variables used:
PCA
ft of groups: 6
Vehicle (EC, Zn, Pb, Cu, Se),
road dust (Si, Fe, Al, Ca, Ba,
Ti), sulfur (S, P), biomass
burning, (K) oil (V, Ni), sea
salt (Na, CI)
In addition, effects of NO2,
CO, SO2, and O3 were
included in the health
outcomes model
Groupings and individual
elements
Results: Overall: Trace elements originating from motor vehicle, road dust, biomass burning, and oil sources associated with
symptoms and/or medication use. No associations with S or sea salt.
Specific Results: PM2.6 mass from motor vehicle or road dust associated with increased odds of respiratory symptoms or inhaler use.
Reduced odds of wheeze or inhaler use with same day S. Significant reductions odds of wheeze with biomass burning.
Co-pollutant: Positive effects of motor vehicles and road dust on wheeze were robust to the inclusion of gaseous copollutants.
However, NO2 increases association with wheeze.
Reference: Ito et al. Subjects: NR
(2006,188554) Exposure: NR
Location: Washington,
DC
Particle Size: PM2.6
n: NR
Number of
constituents
considered for
grouping: NR
Grouping method:
Comparison of:
PMF; (absolute)
PCA; UNMIX
# of groups: 6-10
Groups/ Factors/
Sources: Different
research groups
gave different
names to sources
Sources for which
association with health
was analyzed: Soil, traffic,
Secondary SO4, NO3 (Wash
DC only), residual oil (Wash
DC only), Wood smoke/
biomass combustion, Sea
salt, incinerator (Wash DC
only), primary coal (Wash
DC only), Cu smelter
(Phoenix only)
PM variables used:
Mass contribution of
Results: Overall, PM2.6 effects observed at lag 3. Lag structure of association varied across source types, but consistent across
investigators for total (non-accidental mortality): soil factor - mostly positive at various lags (not significant); secondary sulfate -
strongest association at lag 3; nitrate - mostly negative except at lag 3; residual oil - strongest association at lag 2 (not significant);
wood-burning - increasing association as lag increases (not significant); incinerator - significant negative associations at lag 0; primary
coal - significant association at lag 3.
Reference: Laden et
al. (2000, 012102)
Location: Monitors in
6 Eastern US cities
(Harvard Six Cities
Subjects: NR
Exposure: NR
n: NR
Number of
constituents
considered for
grouping: 15
elements
Grouping method:
PCA
# of groups: 8
Groups/ Factors/ Sources:
Soil/Crustal (PM fine),
Mobile vehicle exhaust (PM
fine), Coal (PM fine), Fuel
oil; Metals, Salt'
Manganese, Residual
PM variables used:
Tracers: Si, V, CI, Pb, Se
July 2009	F-2	DRAFT - DO NOT CITE OR QUOTE

-------
'	Results: Lag 0-1 average for all results. Over all 6 cities, mobile source factor (using Pb as tracer) had greatest association with daily
Particle Size: NR mortality (3.4%) with 10 /yglm3 increase. The greatest effects for mortality due to mobile sources were observed in Madison
(Portage), Knoxville (Kingston-Harriman), and St. Louis, although the Madison results were not statistically significant. The coal source
factor was only significant in Boston (Watertown) - 2.8% increase in mortality and the overall percent increase was also significant
(1.1 %). Deaths from pneumonia attributable to coal combustion sources was 7.9% (CI 3.1-12.7%) and statistically significant. The
crustal factor was not associated with mortality in any city, although this factor was not a significant predictor in the regression
model for Boston (Watertown) due to its low contribution to PM2.B mass. For specific elements included simultaneously, S, Pb, and Ni
were significantly associated with overall mortality (3.0,1.6,1.5%, respectively). Boston had the greatest percent increase in
mortality for S (7.9%), Knoxville for Pb (15.0%), and Steubenville for Ni (8.2%), although the CIs are all quite large.
Reanalysis results (Schwartz, 2003): Effects changed slightly. New percent increases in mortality for combined cities are 3.5 and
0.79 for traffic and coal, respectively. The coal factor in Boston decreased to 2.1% increased mortality. A residual oil factor in Boston
and Steubenville resulted in at 22.9% increase in daily deaths (but was not significant in the original analysis).
Reference: Lanki et al.
(2006,088412)
Location: Monitors in
Helsinki, Finland,
Amsterdam,The
Netherlands and Erfurt,
Germany
Particle Size:
UF/PM2.6
Subjects: NR n: NR
Exposure: NR
Number of
constituents
considered for
grouping: 13
elements
Grouping method:
Absolute PCA
ft of groups: 5
Groups/ Factors/ Sources:
Crustal; long range
transported; oil combustion;
soil; traffic
PM variables used:
Tracers: Si (crustal); S
(long-range transport); Ni
(oil combustion); CI (salt);
ABS (local traffic).
Results: Highest observed effects were for crustal sources and salt at lag 3 (when analyzing sources), but not consistent or
significant. In multipollutant models only ABS associated with ST-segment depression, but wide CIs. When examining indicator
elements of a source, local traffic found to be the most toxic, but when examined per IQR long-range transport and traffic had similar
effects.
Results: All had significant associations with mortality. Traffic density and EC had the largest effects.
Reference: Lippmann
et al. (2006, 091165)
Location: U.S.
Particle Size: PM10
for risk estimates,
PM2.6 for speciation
data
Subjects: NR
Exposure: NR
n: NR
Number of
constituents
considered for
grouping: NR
Grouping method:
No grouping was
performed
# of groups: NR
Groups/ Factors/ Sources: PM variables used:
NR	Mass contribution of 16
constituents
Results: The strongest predictions of the variation in PM10 risk estimates across the 90 NMMAPs MSAs was for Ni and V. Elevated,
but nonsignificant increases were associated with EC, Zn, SOr', Cu, Pb, and OC. Al and Si had the lowest values.
Reference: Mar et al.
(2000,001760)
Location: 1 monitor in
Phoenix, AZ
Particle Size: NR
Subjects: Elderly n: NR
only
Exposure: NR
Number of
constituents
considered for
grouping: 10
elements, OC, EC,
CO, NO?; SO2
Grouping method: Groups/ Factors/ Sources: PM variables used:
Unspecified type of Motor exhaust/road dust,
factor analysis soil, vegetative burning,
tt of groups: 3 or 5 local SO2, regional S04
First used individual
constituents: S, Zn, Pb,
K, OC, EC, TC
(AL+Si+Ca+Fe+Ti),
then factor scores
Results: Cardiovascular mortality associated with PM2.6 mass on lag 1 and 4 (6 and 4%, respectively). EC and TC associated with CV
mortality for lag 1 (RR - 1.05); OC was weakly associated with CV mortality for lags 1 and 3. For total mortality, regional sulfate
was positively associated at lag 0, but negatively associated at lag 3. The local SO2 and the soil factors were negatively associated
with total mortality. For CV mortality, secondary sulfate was positively associated at lag 0, motor vehicle at lag 1, and vegetative
burning at lag 3.
Reanalysis results (Mar, 2003): Similar associations were observed.
Reference: Mar et al. Subjects: NR n: NR
(2006,086143) Exposure: NR
Location: Phoenix, AZ
Particle Size: PM2.6
Number of	Grouping method:
constituents Comparison of: PMF
considered for (absolute); PCA;
grouping: NR UNMIX
# of groups: 6-10
Groups/ Factors/
Sources: Different
labs gave different
names to sources
(see Hopke et al,
table 2)
Sources for which	PM variables used:
association with health Mass contribution of
was analyzed: Soil, Traffic, sources
secondary SO4, NO3, (Wash
DC only), residual oil (Wash
DC only), woodsmokef
biomass combustion, sea
salt, incinerator (Wash DC
only); primary coal (Wash
DC only); Cu smelter
(Phoenix only)
Results: Using daily PM2.6 data found the following associations with cardiovascular mortality: Secondary sulfate - greatest effect
observed for all sources and at lag 0; traffic - associated at lag 1; copper smelter associated at lag 0; sea salt - had the greatest
statistical significance and observed at lag 5; biomassfwood burning - less consistent lag structure but greatest association at lag 3;
soil - did not show an association or consistent lag structure. For total (non-accidental) mortality associations were weaker and
consistently observed for only: copper smelter - lag 0; sea salt - lag 5.
July 2009
F-3
DRAFT-DO NOT CITE OR QUOTE

-------
Reference: Ostro et
al. (2007, 091354)
Location: Monitors in
6 CA counties, some
with 2 monitors, for 4
years
Particle Size: PM2.B
Subjects NR n: NR
Exposure: NR
Number of con-
stituents consi-
dered for
grouping: 15
elements, EC, 0C;
NO3; SO4, PM2.6
mass
Grouping method:
No grouping was
performed
# of groups: NA
Groups/ Factors/ Sources: PM variables used:
NR	Mass contribution of
every constituent
Results: Effects were greater during the winter months. In the all year analysis, at 3-day lag associations observed for EC, 0C, NO3
and Zn. During winter months (Oct -March) effects observed for most species for both all-cause and cardiovascular mortality at lag 3
(EC, 0C, SO4, Ca, Fe, K, Mn, Pb, S, Si, Ti, Zn) and (0C, NO3, SO4, Fe, Mn, S, V, Zn), respectively.
Reference: Ostro et
al. (2009,191971)
Location: Monitors in
6 CA counties, some
with 2 monitors/4
years
Particle Size: PM2.B
Subjects NR
Exposure: NR
n: NR
Number of
constituents
considered for
grouping: 9
elements,EC, 0C,
PM2.6 mass, SO4,
NOs
Grouping method:
No grouping was
performed
# of groups: NA
Groups/ Factors/ Sources:
NR
PM variables used:
Mass contribution of
every constituent
Results: The following associations were observed with cardiovascular mortality: PM2.6 (lag 3); EC (lag 2); NO3 (lag 3); SOi (lag 3); Fe
(lag 2); K (lag 2); S (lag 3); Ti (lag 2); Zn (lag 3).
Reference: Peng et al.
(2009,191998)
Location: 119 urban
communities STN
data/2000-2006
Particle Size: PM2.6
Subjects:	n: NR	Number of	Grouping method: Groups/Factors/ Sources: PM variables used:
Medicare enrollees	constituents NR	Only suggested in discussion Tracers
65 or older	considered for #ofgroUpS:NR
Exposure: NR	grouping: S04,
NOs, Si, EC, 0CM,
Na, NH4
Results: CVD HA's: EC associated with same-day CVD HA's in single and multi-pollutant models. In single pollutant models
associations also observed for sulfate, nitrate, 0CM, and ammonium. However, the sulfate, nitrate, 0CM, and ammonium associations
were reduced in the multi-pollutant models.
Respiratory HA's: 0CM associated with same-day respiratory HA's in single and multi-pollutant models. Some evidence for sulfate
associations at one and two-day lag.
Reference: Penttinen Subjects: Adult n: 78
et al. (2006, 087988) asthma subjects,
Location: Helsinki
max 2 km from
1996-1997 (7 months) sin9le monitor
Particle Size: PM2 B Exposure: NR
Number of
constituents
considered for
grouping:
Unknown
Grouping method: Groups/Factors/ Sources: PM variables used:
PCA
ft of groups: 6
Long range (PM mass, S, K, every component
Zn), local combustion-traffic individually, then
(Cu, Zn, Mn, Fe), soil (Si, Al, groupings
Ca, Fe, Mn), oil (V, Ni), salt
(Na, CI), unidentified
Results: Long range PM2.6 associated with decreased mean PEF in the morning at lag 1. Local combustion PM2.6 associated with
decreased mean PEF in the evening for lag 1. Local combustion PM2.6 associated with decreased mean PEF in the afternoon and
evening for 5-day mean lag. Negative significant association between long-range PM2.6 and asthma symptom prevalence at lag 3. Sea-
salt PM2.E negatively associated with bronchodilator use at lag 3 and 5-day mean lag. Sea-salt PM2.6 negatively associated with
corticosteroid use for 5-day mean lag. Unidentified PM2.6 negatively associated with corticosteroid use at lag 1. Most consistent
negative responses for local combustion, although not always significant. No consistent or significant associations between 5-day
average concentrations of elements and PEF, cough, asthma symptoms, or medication use.
Reference: Riediker et
al. (2004, 091261)
Location: Inside 9
state police patrol cars
Particle Size: PM2.6
Subjects: Healthy n: 9
male young police
officers
Exposure: 4
consecutive days
Number of
constituents
considered for
grouping: 10
elements; 3
gaseous
pollutants; 2
physical variables
Grouping method:
PCA
# of groups: 4
when 13 + 2
constituents
included; 3 when
only 9
"PM-associated"
constituents
included
Groups/ Factors/ Sources:
Soil; automotive steel wear;
gasoline combustion;
speed-changing traffic
PM variables used:
Mass contribution or
score of sources
Results: Using two different factor analysis models found most significant effects (MCL, SDNN, PNN50, supraventricular ectopic
beats, % neutrophils, % lymphocytes, MCV, von Willebrand Factor, and protein C) were for "speed-change factor" (i.e., Cu, S,
aldehydes). Some associations observed for "crustal" and none for "steel wear" and "gasoline."
July 2009
F-4
DRAFT-DO NOT CITE OR QUOTE

-------
Reference: Sarnat et
al. (2008, 097972)
Location: 1 monitor in
Atlanta. GA for 2 yrs
Particle Size: PM2.6
Subjects: NR n: NR
Exposure: NR
Number of
constituents
considered for
grouping: NR
Grouping method:
Comparison of:
PMF, CMB-LG0,"a
priori decision"
# of groups: 9,11
(6 of them common
between methods)
Groups/ Factors/ Sources:
gasoline, diesel, wood
smokef biomass burning,
soil, secondary
S04fammonium sulfate,
seconddary nitrate!
ammonium nitrate, metal
processing, railroad, bus and
highway, cement kiln,
power plants, other 0C,
ammonium bisulfate
PM variables used:
Mass contribution or
score of sources, and
tracers
Results: Sulfate secondary associated with 1.2 - 2.0% increase in RD visits, significant negative association RD visits and primary
emissions from coal-fired power plants. CVD significantly associated with other OC (1.014), biomass (1.033), diesel and gas for
CMB-LGO. For PMF and CVD visits: diesel (1.025), gas, wood smoke, metal processing (1.013). Year long associations: PMF diesel,
EC, CMB LGO gas, Zn and biomass combustion sources (CMB LGO biomass burning, PMF wood smoke, and K). Diesel and gas sources
association with RD in the warm season (1.2-2.1 % per IQR).
Reference: Schreuder Subjects: NR n: NR
et al. (2006, 097959)
Location: 1 monitor in
Spokane, WA for 7
years
Particle Size: PM2.6
Reference: Tsaiet al.
(2000,006251)
Location: 3 NJ sites
for 2 summers (ATEOS
study)
Particle Size: NR
Exposure: NR
Number of
constituents
considered for
grouping: 11
elements, TC, NO3
Grouping method: Groups/ Factors/ Sources: PM variables used:
Comparison of:
PMF, UNMIX,
Multilinear Engine
# of groups: 8
Vegetative burning: As-rich Tracers: TC (vegetative
Vehicle: SO4; NO3; Soil; burning): As (As-rich); Zn
Cu-rich; Marine	(vehicle); Si (soil)
Results: Si, As, and Zn were not associated with any health outcomes; while an IQR increase in TC (vegetative burning) was
associated with a 2% increase in respiratory ED visits.
Subjects: NR
Exposure: NR
N: NR
Number of
constituents
considered for
grouping: 8
metals, IPM, FPM,
S04, CX, DCM,
ACE, CO
Grouping method: Groups/ Factors/ Sources: PM variables used:
Unspecified type of Oil burning, motor
factor analysis
# of groups: 5
emissions, resuspended
dust, secondary aerosol,
industrial sources
individual constituents,
then factor scores, then
tracers
Results: RR associated with 10 //gfm3 increases: Newark -1.03 for industrial and total daily deaths; 1.02 for sulfate and total daily
deaths; 1.04 for sulfate and cardiopulmonary deaths. Camden -1.11 for oil burning sources and total daily deaths; 1.10 industrial and
total daily deaths; 1.12 for oil burning sources and cardiopulmonary daily deaths; 1.02 for sulfate and cardiopulmonary daily deaths
Reference: Yue et al. Subjects: Adult n: 56, data Number of
Grouping method: Groups/ Factors/ Sources: PM variables used:
(2007,097968)
Location: 1 monitor in
German city, 30.000
samples
Particle Size:
UF/PM2.6
males
Exposure: CAD
collected 12
times over 6
month for
every
subject, but
extended
period of
missing PM
data
constituents
considered for
grouping:
Apportionment
based on particle
size distribution.
PMF
# of groups: 5
Airborne soil, local traffic,
local fuel combustion,
remote traffic (diesel),
secondary aerosols
Mass contribution or
score of sources
Results: Overall, repolarization parameters influenced by traffic-related particles; vWF increased in response to traffic-related
particles and combustion-generated aerosols. All source factors contributed to increasing CRP levels.
Table F-2. Human clinical studies of ambient PM sources, factors, or constituents.
Study: Gong et al.
(2003, 042106)
Location: Los
Angeles, CA
Particle Size: PM2.6
Subjects: Adult
18-45, healthy vs.
asthmatic
Exposure: CAPs,
healthy and
asthmatic subjects
exposed at different
times
N: 12 healthy, 12
asthmatic
Constituents
considered for
grouping: 7
elements, EC, NO3,
S04
Grouping method:
PCA
ft of groups: 4 (note:
0C data was
unavailable)
Groups/ Factors/
Sources: Crustal (Al
Si CA K Fe),S (2
metrics of SO4 +
elemental S), Total
Mass+N03, EC
PM variables used:
Total mass, then
tracers: SO4, EC, Fe
July 2009
F-5
DRAFT-DO NOT CITE OR QUOTE

-------
Results: Fe and EC associated with a decrease in ST-segment voltage 2 days post exposure. EC associated with an increase in ST-
segment voltage immediately following exposure. Sulfate content associated with a decrease in systolic BP 4 h post exposure.
Study: Gong et al.
(2005,087921)
Location: Los
Angeles, CA
Particle Size: PM2
Reference: Huang
et al. (2003,
087377)
Location: Chapel
Hill, NC
Particle Size: PM2.6
Subjects: Elderly, N: 6 healthy, 18
COPD vs. healthy/ COPD
CAPs
Exposure: NO2 (full
factorial)
Constituents
considered for
grouping: 7
elements
+ EC
Grouping method:
PCA
ft of groups: 3 (note:
Si CA K Fe),
0C was unavailable) ® S(M,Na
Groups/ Factors/ PM variables used:
Sources: Crustal (Al Total mass, then
tracers:, SO4, Si, Fe,
EC
Results: Mass concentration of CAPs not observed to significantly affect lung function. However, sulfate content was associated with
a decrease lung function (FEVi and FVC), which was enhanced by coexposure to NO2.
Subjects: Healthy
adults
Exposure: CAPs
N: 35 male; 2 female Constituents
considered for
grouping: 8
elements and SO4
Grouping method:
PCA
# of groups: 2
Groups/ Factors/
Sources:
Fe/S04/Se/V/Zn/Cu
PM variables used:
Factor scores, then
mass contribution of
all 9 constituents
Results: Associations observed between sulfate,Zn, and Se content and increases in BAL neutrophils. Increases in fibrinogen associated
with Cu, Zn, and V content.
Reference: Urch et
al. (2004, 055629)
Location: Toronto,
Canada
Particle Size: PM2.6
Subjects: Healthy N: 23
adults 19-50
yrs/CAPs
Exposure: O3
Constituents	Grouping method: Groups/ Factors/
considered for No grouping was Sources: NR
grouping: unknown performed
# of groups: NA
PM variables used:
Every constituent in
univariate analysis,
then 0C and SO4 in
multivariate analysis
Results: CAPs-induced increase in diastolic BP significantly associated with carbon content of the particles.
Reference: Urch et
al. (2004, 055629)
Location: Toronto,
Canada
Particle Size: PM2.6.
Subjects: Healthy
adults/CAPs
Exposure: O3
N: 24
Constituents
considered for
grouping: 14
elements, EC, 0C
Grouping method: Groups/ Factors/
No grouping was Sources: NR
performed
# of groups: NA
PM variables used:
Every constituent in
univariate analysis,
then 0C and SO4 in
multivariate analysis
Results: Both organic and EC content of CAPs associated with an increase in brachial artery vasoconstriction.
Table F-3. Toxicological studies of ambient PM sources, factors, or constituents
Reference:
Batalha et al.
(2002,
088109)
Location:
Boston, MA
Particle Size:
PM2.6
Subjects: Rats
Exposure: CAPs
(3-day mean CAPs
concentration range:
126.1 -481.0/yg|m3)
CAPs (3-day mean
CAPs concentration
range: 126.1-481.0
/yg/m3)
n: 7-10 rats x 2
levels CAPs x 2
levels SO2 x 6
runs in different
seasons
Constituents
considered for
grouping: 20
elements: 0C; EC
Grouping
method: Previous
study in same city
(Clarke et al.), and
PCA of this
experiment's data
# of groups:4
Groups/ Factors/ Sources:
V/Ni, S, Al/Si, Br/Pb
PM variables used:
4 tracers (Si, SO4, V,
Pb) and EC, 0C in
univariate step. 4
tracers (Si, SO4, V,
Pb) in multivariate
step
Results: Univariate analyses for first day not significant for L/W ratio. Univariate analyses for second and third day and second+third day
mean were similar. Presented second+third day mean regression data. CAPs mass, Si, Pb, SO4, EC, 0C significant for decreased L/W ratio in
normal+CB rats exposed to CAPs. Si, SO4 significant for decreased L/W ratio in normal rats. Si, 0C significant for decreased L/W ratio in CB
rats. Multivariate analysis using normal + CB rats for Si, SO4, V, Pb - only Si remained significant with decreased L/W ratio.
Reference:
Subjects: Normal n: NR
Constituents
Grouping
Groups/ Factors/ Sources:
PM variables used:
Becker et al.
human bronchial
considered for
method: PCA
Cr/AI/Si/Ti/Fe/Cu ("crustal"),
NR
(2005,
epithelial and human
grouping: 12
# of groups: 2
Zn/As/V/Ni/Pb/

088590)
AM
elements
Se

Location:
Exposure: (2-3X106




Chapel Hill,
cells/mL; 11 or 50




NC; repeated
/yg/mL)




July 2009
F-6
DRAFT-DO NOT CITE OR QUOTE

-------
samp ing or	Resu|ts: Cr/AI/Si/Ti/Fe/Cu associated with IL-8 release in normal human bronchial epithelial cells and IL-6 release in AM. Zn/As/V/Ni/Pb/Se not
^8ar	associated with any endpoints. Stepwise linear regression with individual constituents Fe and Si associated with IL-6 release in AM. Cr
Particle Size:	associated with IL-8 release in NHBE cells.
PMio
Reference:
Clarke et al.
(2000,
013252)
Location:
Boston, MA
Particle Size:
PM2.6
Subjects: [
n: 10 dogs, 20 Constituents
Exposure: CAPs Paired
(average for all studies, exPosures' 24
paired: 203.4,	crossover
crossover: 360.8
/yg/m3) repeated
exposure with several
weeks in between
considered for
grouping: 19
elements, black C
Grouping
method: PCA
ft of groups: 4
for exposure in
paired runs,6 for
exposure in
crossover runs
Groups/ Factors/ Sources: PM variables used:
V/Ni, S, Al/Si, Br/Pb, S, Na/CI
Cr
All elements, then
factor scores
Results: No significant differences between baseline, sham, or CAPs group for BAL cell differential percentages. Total BAL protein increased
with CAPs compared to sham. No significant hematological effects with CAPs exposure. Mixed linear regression analyses (statistics not
provided): Al and Ti (3-day avg. concentrations) associated with dose-dependent decreases in BAL AM and increases in BAL PMN percentages.
Sulfate associated with increased WBC. BC, Al, Mn, Si, Zn, Ti, V, Fe, Ni associated with increased blood PMN. Na associated with increased
blood lymphocytes. Al, Mn, Si associated with decreased blood lymphocytes. CAPs mass and BC associated with decreased blood eosinophils.
CAPs mass associated with decreased platelet count. Regression using results of factor analysis: None for 3-day avg. concentration for BAL
parameters. V/I\li for increased AM percentage and Br/Pb for increased PMN percentage for 3rd-day only concentration. V/I\li and Al/Si for
increased blood PMN percentage and decreased blood lymphocyte percentage. Al/Si also for increased WBC counts. Na/CI for increased blood
lymphocyte percentage. S for decreased RBC and hemoglobin.
Reference:
Duvall et al.
(2008,
097969)
Location: 5
US cities
Particle Size:
PM2.6
Reference:
Godleski et al.
(2002,
156478)
Location:
Boston, MA
Particle Size:
NR
Subjects Primary
human airway
epithelial cells
(100,000 cells/mL;
dose not provided)
Exposure: NR
n: NR
Constituents
considered for
grouping: NR
Grouping
method: CMB,
but not on coarse
and ultrafine
ft of groups: 6 or
7
Groups/ Factors/ Sources: PM variables used:
Mobile, residual, oil, wood, Mass contribution of
soil, secondary SO4, secondary constituents, then
NO3	mass contribution of
sources
Results: Linear regression with individual constituents: Sulfate associated with increased IL-8 mRNA expression. Sr associated with
increased COX-2 and decreased H0-1 mRNA expressions. K associated with decreased HO-1 mRNA expression.
Linear regression with sources: Significance levels not provided.
Subjects: Rats
Exposure: CAPs
(3-day mean CAPs
concentration range:
126.1 -481.0/yg/m3)
n: 7-10 rats x 2
levels CAPs x 2
levels SO2 x 6
runs in different
seasons
Constituents
considered for
grouping: 20
elements, 0C, EC
Grouping
method: Previous
study in same city
(Clarke et al.), and
PCA of this
experiment's data
# of groups: 4
Groups/ Factors/ Sources:
V/Ni, S, Al/Si/Ca, Br/Pb
PM variables used:
4 tracers (I, S04,V,
Pb) and EC, 0C
Results: Increased percent of PMNs in BALF in CAPs-exposed rats at 24 h. CAPs affected lung tissue mRNA involved in pro-inflammation,
immune, and vascular endothelial responses. Linear regression: Increased PMN associated with CAPs mass, Br, Pb, SO4, EC, and 0C.
Reference:
Gurgueira et
al. (2002,
036535)
Location:
Boston, MA carbon black and R0FA
Subjects Rats
(Sprague Dawley)
Exposure: CAPs (avg.
mass concentration
600/yg/m3); also
n: 13	Constituents
experiments (1	considered for
rat/group at each	grouping: 20
time point)	elements
Grouping
method: No
grouping was
performed
# of groups: NA
Groups/ Factors/ Sources:
NR
PM variables used:
Mass contribution of
every constituent
Particle Size:
PM2.6
Results: Increased oxidative stress in heart and lungs following CAPs exposure (and R0FA exposure).
Univariate regression: Mn, Zn, Fe, Cu, and Ca most significant responses for lung (r2> 0 .40). Al, Si, Ti, Fe, and total mass most significant
response for heart (r2 > 0.49).
Reference: Subjects Rats (SH and n: 6 1 -day,
Constituents
Kodavanti et
al. (2005,
087946)
Location:
RTP.NC
Particle Size:
PM2.6
WKY)
Exposure: CAPs
(144-2758 jjg/m3)
1 -strain runs, 7 considered for
2-day, 2-strain grouping: NR
runs, 4-9 rats
per run.
Grouping
method: No
grouping was
performed
# of groups: NA
Groups/ Factors/ Sources:
NR
PM variables used:
Mass contribution of
every constituent
Results: No significant correlations between biologic responses and exposure variables (i.e., CAP mass, 0C, inorganic C, sulfate, and other
major elemental constituents). Al, Cu, Zn correlated with biologic responses when constituents normalized per unit mass of CAP (/L/g/mg). Zn
correlated with plasma fibrinogen in SH rats (p - 0.0023).
July 2009	F-7	DRAFT - DO NOT CITE OR QUOTE

-------
Reference:
Lippmann et al.
(2005,
087453)
Location:
Rural location
upwind from
NYC
Particle Size:
PM2.6
Subjects: Mice (C57
and ApoE)
Exposure: CAPs (avg.
mass concentration
113/yg/m3)
n: C57: 3-6
mice/group
ApoE1": 9-10
mice/group
Constituents
considered for
grouping: 19
elements + 0C, EC,
NOs
Grouping
method:
(Absolute) PCA
ft of groups: 4
Groups/ Factors/ Sources:
Regional SO4 (S/Si/OC);
Resuspended soil	sources
(CA/Fe/AI/Si);R0 power plants
(V/Ni/Se);
Traffic and unknown
PM variables used:
Mass contribution of
Results: ApoE null mice: Resuspended soil associated with decreased HR during exposure, but increased HR after exposure. Secondary
sulfate associated with decreased HR after exposure. Residual oil associated with increased RMSSD and SDNN in afternoon following
exposure. Secondary sulfate associated with decreased RMSSD and SDNN in night following exposure. Resuspended soil associated with
increased RMSSD at night following exposure. PM mass associated with decreased HR during exposure and decreased RMSSD at night
following exposure.
C57 mice: Motor vehicle/other source category associated with decrease in RMSSD in afternoon following exposure
Subjects: Mice (ApoE'
)
Exposure: CAPs (avg.
mass concentration
85.6 jjg/m3)
n: 12 ApoE1
mice (6/group)
Number of
constituents
considered for
grouping: NR
Grouping
method: No
grouping was
performed
# of groups: NR
Groups/ Factors/ Sources:
NR
Reference:
Lippmann et al.
(2006,
091165)
Location:
Rural location
upwind from
NYC
Results: Lag for HR elevations on 14 days with wind from NW was same day. Lag for SDNN reduction on 14 days with wind from NW was
Particle Size: 0,1 and 2.
PM26	ana|ysjs: b coefficient significant for Ni and HR (but not V, Cr, or Fe). B coefficient significant for Ni and log SDNN (but not V, Cr, or
Fe).
PM variables used:
Mass contribution of
every constituent in
CAPs portion of
study, contribution of
16 constituents in epi
portion
Reference:
Maciejczyk
and Chen
(2005,
087456)
Location:
Rural; upwind
from NYC
Particle Size:
PM2.6
Subjects: NR
Exposure: CAPs
(90000/well; 300
/yg/mL)
n: 110 samples Constituents
considered for
grouping: 19
elements + 0C, EC,
NOs
Grouping
method:
(Absolute) PCA
# of groups: 4
Groups/ Factors/ Sources:
Regional SO4
Soil; Unknown
Oil combustion
PM variables used:
Mass contribution of
sources
Results: Correlation: V and Ni positively correlated with NF-nB. Oil combustion correlated the greatest with NF-nB (0.316). Significance not
provided. Only 2% of mass contribution originates from this source.
Reference:
Nikolovet al.
(2008,
156808)
Location:
Boston, MA
Particle Size:
NR
Subjects: D
Exposure:
n: 8 dogs, 24
exposure-days in
1997-98:4
dogs, 21
exposure-days in
2001-02
Constituents
considered for
grouping: 13
elements, BC, EC,
0C
Grouping
method:
Compared 3
factor-analytic
models within a
SEM model
# of groups: 4
Groups/ Factors/ Sources:
Oil Combustion VfNi; Power
Plants S ;Road dust
Al/Si ;Motor vehicles
BC/0C/EC
PM variables used:
Mass contribution of
every constituent
Results: Univariate response for respiratory outcomes: road dust and oil combustion associated with decreased respiratory frequency;
motor vehicles associated with increased respiratory frequency; motor vehicles associated with increased PEF; road dust associated with
decreased penh and motor vehicles associated with increased penh.
Multivariate responses for respiratory outcome: Road dust associated with decreased respiratory rate; Motor vehicles associated with
increased airway irritation.
Reference:
Rhoden et al.
(2004,
087969)
Location:
Boston, MA
Particle Size:
PM2.6
Subjects Rats
(Sprague Dawley)
Exposure: CAPs (avg.
mass concentration
range 150-2520
/yg/m3) acetylcysteine
full factorial
n: 4-8 rats (1 -2 Constituents
per group - sham, considered for
CAPs,
sham/NAC,
CAP/NAC)
10 exposures
grouping: 20
elements
Grouping
method: No
grouping was
performed
# of groups: NA
Groups/ Factors/ Sources:
NR
PM variables used:
Mass contribution of
every constituent
Results: Increased oxidative stress and inflammation in lungs of CAPs animals that was attenuated with NAC.
Univariate regression: Al, Si, Fe, K, Pb, and Cu most significantly correlated with lung TBARS. No significant correlations for lung carbonyls
or lung PMN.
July 2009	F-8	DRAFT - DO NOT CITE OR QUOTE

-------
Reference:
Saldiva et al.
(2002,
025988)
Location:
Boston, MA
Particle Size:
PM2.6
Subjects: Rats
(Sprague Dawley
Exposure: CAPs
(3-day avg. mass
concentration range
126.1-481 /yglm3)
n: 7-10
Constituents
Grouping
Groups/ Factors/ Sources:
PM variables used:
rats/group
considered for
method: Previous
V/Ni
Mass contribution of
(air/sham.
grouping: 15
study in same city
S
8 elelments in
S02/sham,
elements (used Clarke
(Clarke et al.
Al/Si
univariate step.
air/CAP,
2000 to select
2000)
Br/Pb
Tracers (Si, SO4, V,
SO2/CAP) X 6
tracers)
# of groups: 6

Pb, Br, CI) and EC, OC
runs in different

Na/CI
in multivariate step.
seasons


Cr
Results: Increased percent and number of PMN in majority of air and SO2 rats exposed to CAPs, but significance levels not provided.
Other responses (protein, LDH, NAG) were variable and depended upon the CAPs exposure. No CAPs effect on histopathology.
Linear regression: V, Br, Pb, SOi, EC, 0C, Si, CAP mass associated with increased PMN and lymphocytes for normal+CB rats. Only V not
associated with PMN in normal rats. Lymphocyte response due to CB rats, but not observed for SO4, Si, or mass in this group. Br, Pb, SO4,
EC, OC, Si associated with increased total protein in CB rats. CI and V associated with decreased LDH in CB rats. No BAL effects in normal
rats exposed to CAPs. V, Br, Pb, EC, OC, and CI associated with increased neutrophil density in lungs of normal rats.
Reference:
Seagrave et al.
(2006,
091291)
Location: 4
SE US sites for
2 seasons
Particle Size:
PM2.6
Subjects Rats (Fisher n: 5 rats/dose
344)
Exposure: 0.75,1.5
and 3 mg/rat via
intratracheal
instillation
Constituents
considered for
grouping: NR
Grouping
method: CMB
ft of groups: 13
Groups/ Factors/ Sources:
secondary NO3; secondary
NH4; secondary SO4; coke
production; vegetative detri-
PM variables used:
Mass contribution of
every constituent,
then mass
tus; natural gas combust; road contribution of
dust; wood combust; meat sources
cooking gasoline; diesel other
OM; other mass
Results: Potency depended upon season and site of sample collection. In general, effects were greater in the winter.
PLS analysis: 2 major constituents identified (OC, Pb, hopanesfsteranes, nitrate, As for first and major metal oxides for the second), gasoline
most important predictor for both constituents, with diesel influencing second constituent and nitrate influencing first constituent. First
constituent affected cytotoxic responses, second constituent affected inflammatory responses.
Reference:
Veranth et al.
(2006,
087479)
Location: 8
sites in the
Western US
Particle Size:
PM2.6
Subjects BEAS-2B
cells (35000 cells/cm2;
10,20, 40,80
/yg/cm2)
Exposure: Loose
surface soil sweepings
through mechanical
tumbler and cascade
impactor
n: 6; 16 runs
over 6 months.
Constituents
considered for
grouping: 13
elements, TC, 5 OC
variables, 4 EC
variables, 2 ions, EU,
one ratio (Ca: Al), OP,
C03
Grouping
method: PLS
# of groups: P
Groups/ Factors/ Sources: PM variables used:
NR	Mass contribution
every constituent (?)
Results: Dose-related increase in IL-6 and decreases in cell viability for all soil types. IL-8 responses more variable and dependent upon soil
type. Univariate correlations. Low correlations for all constituents tested with IL-6. Highest correlations for EC1 (R2 - 0.50) and pyrolyzed
OC (R2 - 0.46), then Ca/AI (R2 - 0.21). Carbonate carbon, EC3, and Sr correlated with IL-8 (R2 - 0.27, 0.13, and 0.25, respectively). EC
and Ni correlated with IL-8 trend over the range of 10-80//g/cm2 (R2 - 0.39 and 0.27, respectively). Multivariate redundancy analysis 0C1,
0C3, 0C2, EC2, Br, EC1, Ni correlated with IL-8 release, decreased viability, and decreased IL-6 at low and high doses. Ni, EC1, and EC2
correlated with IL-6 release at the high dose, decreased IL-6 at the low dose, decreased IL-8 release, and decreased viability. Br was
negatively associated.
Reference:
Wellenius et al.
(2003,
055691)
Location:
Boston, MA
Particle Size:
PM2.6
Subjects: Dogs
Exposure: CAPs (avg.
mass concentration
range 161.3-957.3
/yg/m3) repeated
exposure with several
weeks in between
n: 6 dogs, 20 Constituents
Grouping
Groups/ Factors/ Sources: PM variables used:
exposures	considered for	method: Previous	VfNi
grouping: 15	study in same city	S
elements (+EC OC?)	(Clarke et al.	Al/Si
(used Clarke et al.	2000)	Br/Pb
2000)
tt of groups: 6
(but did not use all
in analysis of
health effects)
Na/CI
Cr
Univariate: Mass
Number
Ni, S, Si, BC
Multivariate: Ni, S,
Si, BC
Results: ST-segment elevation increased with CAPs.
Univariate regression: Si and Pb associated with peak ST-segment elevation and integrated ST-segment change.CAPs mass or number
concentration were not associated with any change.
Multivariate regression: Si associated with peak ST-segment elevation and integrated ST-segment change.
July 2009
F-9
DRAFT-DO NOT CITE OR QUOTE

-------
Reference:
Zhang et al.
(2008,
192008)
Location:
Metro area of
Denver, CO I
45 samples
through 1 year
Particle Size:
2.5; filtered to
0.22 um
Subjects: Alveolar
macrophage cell line
(NR8383); 1x106
cells/ml
Exposure: Soluble
components exposure
concentration range
from 20-200 pg of
PM/cell
n: 45 PM
samples, 3 runs
Constituents
considered for
grouping: 43 +
0C
EC,
Grouping
method: PMF
ft of groups: 9
Groups/Factors/ Sources:
Mobile, Water soluble carbon,
Sulfate, Soil, Iron, Cd and Zn
point source, Pb,
Pyrotechniques, Platinum
PM variables used:
Mass contribution of
Results: Started with regression on 9 sources, then 3 (water-soluble carbon factor, soil dust source, iron source). Soil dust source was not
significant. Final regression model excluded 3 days of outliers (Fe source most significant, then water-soluble carbon factor, then soil dust
source) for R0S effects, with adjusted R2 of 0.774. Fe source likely associated with industrial source and includes high loadings of water-
soluble Fe and Ti (not identified); water-soluble C factor derived from both secondary organic aerosol and biomass smoke (largely consists of
polar organic compounds); soil dust source identified by water-soluble resuspended dust elements and contains Mg and Ca.
July 2009
F-10
DRAFT-DO NOT CITE OR QUOTE

-------
Annex F References
Andersen ZJ> Wahlin Pi Raaschou-Nielsen 0> Scheike T> Loft S. (2007). Ambient particle source
apportionment and daily hospital admissions among children and elderly in Copenhagen. J Expo
Sci Environ Epidemiol, 17- 625-636. 093201
Batalha JR; Saldiva P H; Clarke RW! Coull BA; Stearns RC; Lawrence J; Murthy GG; Koutrakis Pi
Godleski JJ. (2002). Concentrated ambient air particles induce vasoconstriction of small
pulmonary arteries in rats. Environ Health Perspect, 110- 1191-1197. 088109
Becker Si Mundandhara Si Devlin RBi Madden M. (2005). Regulation of cytokine production in human
alveolar macrophages and airway epithelial cells in response to ambient air pollution particles-
further mechanistic studies. Toxicol Appl Pharmacol, 207: 269-275. 088590
Bell MLi Ebisu K> Peng RDi Dominici F. (2009). Adverse health effects of particulate air pollution:
modification by air conditioning. Epidemiology, in press: in press. 191007
Cakmak Si Dales Ri Vida C. (2009). Components of particulate air pollution and mortality in Chile. Int
J Occup Environ Health, 15: 152. 191995
Clarke RWi Coull Bi Reinisch Ui Catalano Pi Killingsworth CRi Koutrakis Pi Kavouras Ii Murthy GGKi
Lawrence Ji Lovett Ei Wolfson JMi Verrier RLi Godleski JJ. (2000). Inhaled concentrated
ambient particles are associated with hematologic and bronchoalveolar lavage changes in
canines. Environ Health Perspect, 108: 1179-1187. 013252
Duvall RMi Norris GAi Dailey LAi Burke JMi McGee JKi Gilmour ML Gordon Ti Devlin RB. (2008).
Source apportionment of particulate matter in the US and associations with lung inflammatory
markers. Inhal Toxicol, 20: 671-83. 097969
Franklin Mi Koutrakis Pi Schwartz J. (2008). The role of particle composition on the association
between PM2.5 and mortality. Epidemiology, 19: 680-689. 155779
Gent JFi Koutrakis Pi Belanger Ki Triche Ei Holford TRi Bracken MBi Leaderer BP. (2009). Symptoms
and medication use in children with asthma and traffic-related sources of fine particle pollution.
Environ Health Perspect, In Press: 1-41. 180399
Godleski JJi Clarke RWi Coull BAi Saldiva PHNi Jiang NFi Lawrence Ji Koutrakis P. (2002).
Composition of inhaled urban air particles determines acute pulmonary responses. Ann Occup
Hyg, 46: 419-424. 156478
Gong H Jri Linn WSi Clark KWi Anderson KRi Geller MDi Sioutas C. (2005). Respiratory responses to
exposures with fine particulates and nitrogen dioxide in the elderly with and without COPD.
Inhal Toxicol, 17: 123-132. 087921
Gong H Jri Linn WSi Sioutas Ci Terrell SLi Clark KWi Anderson KRi Terrell LL. (2003). Controlled
exposures of healthy and asthmatic volunteers to concentrated ambient fine particles in Los
Angeles. Inhal Toxicol, 15: 305-325. 042106
Gurgueira SAi Lawrence Ji Coull Bi Murthy GGKi Gonzalez-Flecha B. (2002). Rapid increases in the
steady-state concentration of reactive oxygen species in the lungs and heart after particulate air
pollution inhalation. Environ Health Perspect, 110: 749-755. 036535
Huang Y-CTi Ghio AJi Stonehuerner Ji McGee Ji Carter JDi Grambow SCi Devlin RB. (2003). The role
of soluble components in ambient fine particles-induced changes in human lungs and blood.
Inhal Toxicol, 15: 327-342. 087377
Ito Ki Christensen WFi Eatough DJi Henry RCi Kim Ei Laden Fi Lall Ri Larson TVi Neas Li Hopke PKi
Thurston GD. (2006). PM source apportionment and health effects: 2. An investigation of
intermethod variability in associations between source-apportioned fine particle mass and daily
mortality in Washington, DC. , 16: 300-10. 188554
Kodavanti UPi Schladweiler MCi Ledbetter ADi McGee JKi Walsh Li Gilmour PSi Highfill JWi Davies
Di Pinkerton KEi Richards JHi Crissman Ki Andrews Di Costa DL. (2005). Consistent
pulmonary and systemic responses from inhalation of fine concentrated ambient particles: roles
of rat strains used and physicochemical properties. Environ Health Perspect, 113: 1561-1568.
087946
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database
(Health and Environmental Research Online) at http://epa. gov/hero. HERO is a database of scientific literature
used by U.S. EPA in the process of developing science assessments such as the Integrated Science Assessments
(ISA) and the Integrated Risk Information System (IRIS).
July 2009
F-11
DRAFT-DO NOT CITE OR QUOTE

-------
Laden Fi Neas LMi Dockery DW! Schwartz J. (2000). Association of fine particulate matter from
different sources with daily mortality in six US cities. Environ Health Perspect, 108- 941-947.
012102
Lanki Ti De Hartog JJi Heinrich J; Hoek Gi Janssen NAHi Peters A; Stolzel Mi Timonen KLi Vallius
Mi Vanninen Ei Pekkanen J. (2006). Can we identify sources of fine particles responsible for
exercise-induced ischemia on days with elevated air pollution? The ULTRA study. Environ
Health Perspect, 114- 655-660. 088412
Lipfert FWi Baty JDi Miller JPi Wyzga RE. (2006). PM2.5 constituents and related air quality variables
as predictors of survival in a cohort of U.S. military veterans. Inhal Toxicol, 18- 645-657. 088756
Lippmann Mi Hwang Ji Maclejczyk Pi Chen L. (2005). PM source apportionment for short-term cardiac
function changes in ApoE-/- mice. Environ Health Perspect, 113: 1575-1579. 087453
Lippmann Mi Ito Ki Hwang JSi Maciejczyk Pi Chen LC. (2006). Cardiovascular effects of nickel in
ambient air. Environ Health Perspect, 114: 1662-9. 091165
Maciejczyk Pi Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles
(CAPs) in mice^ VIII source-related daily variations in in vitro responses to CAPs. Inhal Toxicol,
17: 243-253. 087456
Mar TFi Ito Ki Koenig JQi Larson TVi Eatough DJi Henry RCi Kim Ei Laden Fi Lall Ri Neas Li Stolzel
Mi Paatero Pi Hopke PKi Thurston GD. (2006). PM source apportionment and health effects: 3
Investigation of inter-method variations in associations between estimated source contributions
of PM2.5 and daily mortality in Phoenix, AZ. J Expo Sci Environ Epidemiol, Epidemiol.16: 311-
320. 086143
Mar TFi Norris GAi Koenig JQi Larson TV. (2000). Associations between air pollution and mortality in
Phoenix, 1995-1997. Environ Health Perspect, 108: 347-353. 001760
Nikolov MCi Coull BAi Catalano PJi Diaz Ei Godleski JJ. (2008). Statistical methods to evaluate health
effects associated with major sources of air pollution: a case-study of breathing patterns during
exposure to concentrated Boston air particles. J Roy Stat Soc C Appl Stat, 57: 357-378. 156808
Ostro Bi Feng W"Yi Broadwin Ri Green Si Lipsett M. (2007). The effects of components of fine
particulate air pollution on mortality in California: results from CALFINE. Environ Health
Perspect, 115: 13-9. 091354
Ostro Bi Roth Li Malig Bi Marty M. (2009). The effects of fine particle components on respiratory
hospital admissions in children. Environ Health Perspect, 117: 475. 191971
Peng Ri Bell Mi Geyh Ai McDermott Ai Zeger Si Samet Ji Dominici F. (2009). Emergency admissions
for cardiovascular and respiratory diseases and the chemical composition of fine particle air
pollution. Environ Health Perspect, 117: 957-963. 191998
Penttinen Pi Vallius Mi Tiittanen Pi Ruuskanen Ji Pekkanen J. (2006). Source-specific fine particles in
urban air and respiratory function among adult asthmatics. Inhal Toxicol, 18: 191-198. 087988
Rhoden CRi Lawrence Ji Godleski JJi Gonzalez-Flecha B. (2004). N-acetylcysteine prevents lung
inflammation after short-term inhalation exposure to concentrated ambient particles. Toxicol
Sci, 79: 296-303. 087969
Riediker Mi Devlin RBi Griggs TRi Herbst MCi Bromberg PAi Williams RWi Cascio WE. (2004).
Cardiovascular effects in patrol officers are associated with fine particulate matter from brake
wear and engine emissions. Part Fibre Toxicol, l: 2. 091261
Saldiva PHNi Clarke RWi Coull BAi Stearns RCi Lawrence Ji Krishna-Murthy GGi Diaz Ei Koutrakis
Pi Suh Hi Tsuda Ai Godleski JJ. (2002). Lung inflammation induced by concentrated ambient air
particles is related to particle composition. Am J Respir Crit Care Med, 165: 1610-1617. 025988
Sarnat JAi Marmur Ai Klein Mi Kim Ei Russell AGi Sarnat SEi Mulholland JAi Hopke PKi Tolbert PE.
(2008). Fine particle sources and cardiorespiratory morbidity: an application of chemical mass
balance and factor analytical source-apportionment methods. Environ Health Perspect, 116:
459-66. 097972
Schreuder ABi Larson TVi Sheppard Li Claiborn CS. (2006). Ambient woodsmoke and associated
respiratory emergency department visits in Spokane, Washington. Int J Occup Environ Health,
12: 147-53. 097959
Seagrave JCi McDonald JDi Bedrick Ei Edgerton ESi Gigliotti APi Jansen JJi Ke Li Naeher LPi Seilkop
SKi Zheng Mi Mauderley JL. (2006). Lung toxicity of ambient particulate matter from
southeastern US sites with different contributing sources: relationships between composition
and effects. Environ Health Perspect, 114: 1387-93. 091291
Tsai FCi Apte MGi Daisey JM. (2000). An exploratory analysis of the relationship between mortality
and the chemical composition of airborne particulate matter. Inhal Toxicol, 12: 121-135. 006251
July 2009
F-12
DRAFT-DO NOT CITE OR QUOTE

-------
Urch B; Brook JR; Wasserstein D; Brook RD; Rajagopalan S> Corey P; Silverman F. (2004). Relative
contributions of PM2.5 chemical constituents to acute arterial vasoconstriction in humans. Inhal
Toxicol, 16: 345-352. 055629
Veranth JM; Moss TA; Chow JC; Labban R; Nichols WE Walton JC; Walton JG; Yost GS. (2006).
Correlation of in vitro cytokine responses with the chemical composition of soil-derived
particulate matter. Environ Health Perspect, 114: 341-349. 087479
Wellenius GA; Coull BA; Godleski JJ; Koutrakis P; Okabe K; Savage ST. (2003). Inhalation of
concentrated ambient air particles exacerbates myocardial ischemia in conscious dogs. Environ
Health Perspect, 111: 402-408. 055691
Yue W; Schneider A; Stolzel M; Ruckerl R; Cyrys J; Pan X; Zareba W; Koenig W', Wichmann HE; Peters
A. (2007). Ambient source-specific particles are associated with prolonged repolarization and
increased levels of inflammation in male coronary artery disease patients. Mutat Res Fund Mol
Mech Mutagen, 621: 50-60. 097968
Zhang Y, Schauer JJ, Shafer MM, Hannigan MP, Dutton SJ. (2008). Source apportionment of in vitro
reactive oxygen species bioassay activity from atmospheric particulate matter. Environ Sci
Technol, 42: 7502-7509. 192008
July 2009
F-13
DRAFT-DO NOT CITE OR QUOTE

-------